You are on page 1of 81

For Public Comment March 15April 6, 2012 Comments due 5:00 pm EST

Technical Specifications for ACO Measures Public Comment Overview

HEDIS is a registered trademark of the National Committee for Quality Assurance (NCQA).

Note: This publication is protected by U.S. and international copyright laws. You may reproduce this document for the sole purpose of facilitating public comment. 2012 by the National Committee for Quality Assurance 2012 by the National Committee for Quality Assurance 1100 13th Street NW, Suite 1000 Washington, DC 20005 All rights reserved. Printed in U.S.A. NCQA Customer Support: 888-275-7585 www.ncqa.org

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012

Synopsis
NCQA seeks feedback on proposed technical specifications for performance measures for Accountable Care Organizations (ACO). Public Comment documents are: Public Comment Overview. Provides background and rationale for NCQAs approach to measuring ACO performance and explains how NCQA arrived at some of the decisions reflected in the accompanying documents. Technical Specifications for ACO MeasuresOverview and General Guidelines. The first two sections of the Technical Specifications for ACO Measures publication provide guidance on defining the population for ACO measurement, reporting measure results to NCQA, data collection and sampling methods for a core set of performance measures. The publication also includes three appendices for reference. Selected measures Revised specifications for five selected measures that illustrate the application of the general guidelines for ACO measurement to the actual measure specification: Colorectal Cancer Screening. Comprehensive Adult Diabetes CareLDL Control. Use of Appropriate Medications for People With Asthma. Use of Imaging Studies for Low Back Pain. Follow-Up After Hospitalization for Mental Illness. Submit comments by April 6 at 5:00 p.m. EST.

Background NCQAs Mission: Improve the Quality of Health Care


For over 20 years, NCQA has been driving improvement throughout the health care system and helping to raise the issue of health care quality to the top of the national agenda. NCQA accredits and certifies a wide range of health care organizations. It also recognizes clinicians and practices in key areas of performance. NCQAs Healthcare Effectiveness Data and Information Set (HEDIS) is the most widely used performance measurement tool in health care. NCQA is committed to providing health care quality information for consumers, purchasers, health care providers and researchers.

What Is an Accountable Care Organization?


Accountable Care Organizations (ACO) are provider-based organizations that take responsibility for meeting the health care needs of a defined population, with the goal of simultaneously improving health, improving patient experience and reducing per capita costs. How providers organize themselves as accountable entities may vary, based on existing practice structures in a region, on population needs or on local environmental factors. Within the ACO structure itself (i.e., subject to the direct authority of ACO governance) ACOs are likely to differ with respect to the components of care delivery directly included. Some may include a full range of services with a variety of subspecialists, hospitals, home care agencies and insurance products. Others may be more narrowly constructed, but maintain active relationships and formal contracts with providers across the spectrum of care necessary to meet the needs of their patients.

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012 Although there is room for much innovation and experimentation, NCQAs definition of an ACO requires the organization to include a group of physicians with a strong primary care base and enough specialty physicians to support the care needs of a defined population of patients. A well-run ACO should align the clinical and financial incentives of its providers. Providers will need to be clinically integrated and work together to seamlessly coordinate care for assigned patients. Given the evidence that alignment of clinical and financial incentives is desirable, ACOs will also need the administrative infrastructure to manage budgets; collect data; report performance; make payments related to performance; and organize providers around shared goals. By coordinating and integrating care, ACOs have the potential to simplify the care process for patients, enhance quality and reduce costs. ACOs have the potential to get better quality at lower cost by aligning incentives to promote coordination and transform health care delivery. However, not every group of providers that wants to call itself an ACO has what it takes to accomplish this vital mission. That is why NCQA, working with a broad array of stakeholders, developed clear criteria and standards for guiding ACOs to success. NCQAs ACO Accreditation provides independent evaluation of organizations abilities to coordinate and be accountable for the high -quality, efficient, patientcentered care expected from ACOs. Like NCQAs Patient-Centered Medical Home (PCMH) program, ACO accreditation provides a roadmap to help providers make the challenging and much-needed transformation into ACOs. The program aligns with many of the expectations that the Centers for Medicare & Medicaid Services (CMS) has for the Medicare Shared Savings Program, as well as common expectations of private purchasers.

About ACO Measure Specifications


In November 2011, NCQA released Standards and Guidelines for the Accreditation of Accountable Care Organizations. The ACO standards include a set of 40 recommended core performance measures drawn from the HEDIS measurement set. Currently, the program allows organizations to report on core performance measures using a variety of methods. However, over time NCQA plans to move to standardized measures and performance-based scoring (where results and standards are used together to score performance). NCQAs development of the Technical Specifications for ACO Measures (ACO Measure Specifications) was guided by the following principles: Leverage HEDIS Health Plan Measurement to build on its support of another NCQA accreditation program. Align with CMS Medicare Shared Savings Program (MSSP) rules, while considering ACO national leadership. Build from the knowledge and experience of Californias IHA. Strive for consistent measurement between ACOs for future benchmarking. Encourage ACOs to develop the capacity for electronic evaluation of quality of care across populations. Meet most ACOs where they will be in the near term, with respect to their ability to measure. Allow ACO measurement to evolve as NCQA, ACOs and others gain experience. NCQA strives to balance the need to accommodate a range of organizational and contractual structures as ACOs initiate operations with the need to encourage standardization in measurement activities for the purpose of comparing results over time. The ACO Measure Specifications are an extension of NCQAs significant experience and record of accomplishment with establishing and maintaining provider quality measures, many of which build from more than 17 years of developing and testing HEDIS measures. Both the HEDIS 2012 Technical Specifications for Health Plans (HEDIS Health Plan Measurement) and HEDIS Physician Measurement were the foundation for ACO Measure Specifications. The ACO Measure Specifications publication specifically supports PR 1: Performance Reporting, Element A Core Performance Measures of the Standards and Guidelines for Accreditation of ACOs . It enumerates specific requirements for reliable, valid, equitable ACO reporting, and is intended for use by ACOs or by third-party organizations conducting population health analytics on behalf of the ACO.

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012

Measure specification complexity varies because of the amount of data available to ACOs. ACOs with payer contracts that include the exchange of paid claims (pharmacy and medical), enrollment and laboratory data for the ACO population, as well as fully functioning EHRs across all settings of care, may have a rich array of data available. Other ACOs may need to use manual billing systems and paper records that differ in level of completeness. Recognizing that measurement programs applied across a diverse set of ACOs must accommodate the least advanced measurement capabilityat least for nowNCQA has attempted to provide specifications that apply to all levels of data sophistication. NCQA believes that satisfying performance reporting requirements for ACO standards will differentiate ACOs based on their ability to effectively improve the quality of care for the populations they serve. Our development process identified a number of key policy issues for applying HEDIS specifications to an ACO environment. NCQA convened an external ACO Advisory Panel to provide input on these issues and discussed them with a variety of standing committees, including the Committee on Performance Measurement. NCQA now seeks Public Comment on our proposals addressing these issues.

ACO Measurement Policy Considerations


This section provides the rationale for several key considerations on which NCQA seeks Public Comment: A. Defining the Population for ACO Measurement ACO Contracts and Measurement Attributing Patients to ACOs B. Measure Reporting to NCQA Stratifying the Population for ACO Measurement ACO vs. Payer or Collaborative Reporting Phase-In, Periodicity and Requirements for ACO Data Submission Audit Requirements for ACO Submission C. Methods and Definitions for Compiling Measures Include Hybrid Method for All Measures Systematic Sampling and Sample Size Requirements Exclusions for ACO Measurement Pharmacy Data for ACO Measurement For each consideration, we provide a description and our proposed resolution. Page numbers in the heading of each section refer to the draft ACO Measure Specifications.

A. Defining the Population for ACO Measurement


A-1. ACO Contracts and Measurement (page 11) NCQAs ACO Accreditation standards allow organizations to evaluate: The assigned or attributed ACO population, or All patients seen by participating ACO providers. NCQA intends to require ACOs to report measures on their aggregate ACO population (i.e., the entire population under contract with the ACO), but recognizes that in the early years of ACO accreditation, health care providers transforming toward accountable care may not yet have formal ACO contracts. NCQA developed three alternatives for defining the population for which an ACO submits measures to NCQA.

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012 Method 1 1A 1B If the ACO has at least one ACO contract in force as of January 1 of the measurement year, it has two options: Report on the aggregate ACO contract population, as defined through contractual attribution models or assignment methods, or Report on the aggregate ACO contract population, plus all or at least 80 percent of the patient population (consistent with how the population is defined for NCQA ACO Accreditation), as defined by NCQAs Recommended Patient Attribution Model. If the ACO has no ACO contract in force as of January 1 of the measurement year, it reports on all or at least 80 percent of its patient population (consistent with how the population is defined for NCQA ACO Accreditation), using NCQAs Recommended Patient Attribution Model.

Method 2

For the purposes of Methods 1B and 2, two steps apply when using NCQAs Recommended Patient Attribution Model. The first step is to define the universe of an ACOs patient population as all active patients as of January 1 of the measurement year (and for purposes of actual measurement, defined as of December 31 of the measurement year) with any provider participating in the ACO as contractually defined by the ACO (e.g., primary and specialty care, acute care, surgical care, post-acute care, long term care). The second step is applying NCQAs Recommended Patient Attribution Model to at least 80 percent of that universe of patients. The intent is for an ACOs entire patient population to be considered through this process. NCQA expects that ACOs will use a consistently defined population across all performance measures. NCQA requires that ACOs disclose the attribution models used to identify their population for measurement, regardless of the model used to define the population. Refer to Disclosure of Patient Attribution Model. A-2. NCQAs Recommended Patient Attribution Model (pages 1112) Background. Accountable care populations are generally defined by patients receiving medical care from providers participating in an ACO or by assigning people (sometimes voluntarily) to a defined ACO. Neither approach defines a broad community of individuals for whom an ACO can be held accountable for achieving the Triple Aim. With this in mind, NCQA intends to define patient attribution in terms of where the industry is today and will be in the near futureto account for ACOs in their current status. Patient attribution or assignments will be contractually defined and may therefore differ by ACO and by contract; therefore, NCQAs approach to recommending an attribution model must accommodate different attribution models currently in the field and encourage standardization. NCQA will require ACOs to report the attribution models they use to identify their population for measurement and expects that the population will remain consistent across all performance measures. NCQA considered several patient attribution models, including the CMS Physician Group Practice and Medicare Shared Savings Program (MSSP), the Brookings-Dartmouth ACO Collaborative and organization-specific methodologies (e.g., BCBS of Illinois). Several of these models required ACOs to have access to payeradjudicated claims (e.g., use of episode groupers or allowed charges), which raised concerns about broad applicability across ACOs. While none of these models identify all individuals (e.g., those not seeking health care) for which an ACO may be accountable, specifying an attribution methodology is a reasonable starting place. NCQA encourages long-term standardization of patient attribution models across the health care industry and seeks harmony with CMS accountable care evaluation methods. NCQA also recognizes the ongoing empirical contributions to this model by researchers at The Dartmouth Institute, and therefore proposes to specify an attribution method for organizations opting to report on ACO populations as follows.

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012 Recommended Attribution Method: Use the patient attribution model as defined by CMS MSSP, with the following modifications:

1. Apply a 24 month period, rather than the CMS 12-month period, given the broader demographic of all ages, rather than only CMS beneficiaries who use more services than younger cohorts use. 2. Apply the plurality of primary care visits, rather than the CMS plurality of primary care allowed charges, given the probable dependence on payer paid claims to determined allowed charges.

Recommended Patient Attribution Model


This model is annotated from the CMS MSSP Final Rule, with brackets in bold type indicating the modifications for use with ACO measurement as defined in this publication. Patients are assigned to an ACO through a step-wise process. The ACO starts by identifying all patients that had at least one primary care service (defined below) during the prior [24-month] period with a physician who is an ACO provider or supplier of that ACO. Then the ACO identifies all of those primary care services rendered by primary care physicians. Patients are assigned to an ACO if the patients [visits] for primary care services to all primary care physicians who are ACO providers/suppliers in the ACO are greater than the [number of visits] for primary care services furnished by primary care physicians who are ACO providers/suppliers in any other ACO or not affiliated with any ACO. The second step considers the remainder of the patients who received at least one primary care service from an ACO physician, but who have not had a primary care service rendered by any primary care physician, either inside or outside the ACO. The patient will be assigned to an ACO if their [visits] for primary care services furnished to the patient by all ACO professionals who are ACO providers/suppliers in the ACO are greater than the [number of visits] for primary care services furnished by all ACO professionals who are ACO providers/suppliers in any other ACO or other physicians, nurse practitioners, physician assistants, clinical nurse specialists who are unaffiliated with an ACO.
Source: Annotated from the final CMS MSSP rules published in the Federal Register, Vol. 76, No. 212, Wednesday, November 2, 2011, Rules and Regulations, 425.400, page 67983, available at: http://www.gpo.gov/fdsys/pkg/FR-2011-11-02/pdf/2011-27461.pdf

Primary Care Service Codes for Attribution. Publish and maintain specific primary care service HCPCS codes, G-codes and corresponding revenue codes (as required for inclusion of Federally Qualified Health Centers and rural health centers), in accordance with CMS MSSP.

Primary Care Services


CPT/HCPCS Level I Codes HCPCS Level II G Codes UB Revenue Center Codes 9920199215; 9930499340; 9934199350 G0402 (the code for the Welcome to Medicare visit), G0438 and G0439 (the codes for the annual wellness visits). 0521, 0522, 0534, 0525 submitted by FQHCs (for services furnished prior to January 1, 2011), or by rural health centers.

Retrospective Attribution for Measurement. Although the populations targeted for measurement are determined as of January 1 of the measurement year (refer to A-1, above), patient attribution to those populations, for the purposes of ACO measurement, should be determined as of December 31 of the measurement year. This is consistent with CMS MSSP, whereby there is an initial prospective attribution at the start of the measurement year to determine the populations to be evaluated and a final attribution at the close of the year for measurement.

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012

Submission of ACO Attribution Models. NCQA proposes that ACOs submit all models of attribution they use to define the population. This will allow NCQA to collect information for exploring the potential for future standardization in models across ACOs. Where ACOs use multiple attribution models, NCQA will request information on the model and the proportion of the aggregate ACO population identified using that model. ACO specifications will not include continuous enrollment or patient inclusion criteria, because patient attribution methodologies will adequately assign the accountability of patients to ACOs.

B. Measure Reporting to NCQA


B-1. Stratifying the Population for ACO Measurement (page 16) ACOs are designed to focus on the care of populations rather than on specific product lines or age groups. Providers generally deliver care regardless of payer and products. NCQA believes is it important not to artificially compartmentalize measures through stratification, but recognizes that there are Medicaid programs entering into ACO contracts and that Medicaid populations historically perform differently on quality measures than do other populations. Other considerations are the impact on measures with age bands that cross over broad age strata (e.g., Colorectal Cancer Screening is ages 5075), and providing value to all stakeholders. Population Stratification. NCQA will require reporting on the ACO aggregate population as a whole, with an option to report on Medicaid separately for those measures that include a Medicaid specification for health plan calculations. Refer to Appendix 3 for the list of measures with a Medicaid specification. B-2. ACO vs. Payer or Collaborative Reporting (page 16) Payers may provide data to the ACO for quality reporting. For an ACO, this could occur with a single payer, or with multiple payers if there is more than one ACO contract in force. The ACO retains responsibility for reporting performance measure results on its population following ACO measure specifications. ACOs with multiple payers providing measure results could report to NCQA on their aggregate ACO population only if they could compile a composite score, using weighted rates from each payer population. The intent of PR 1: Performance Reporting is to expect ACOs to conduct quality measurement across their populations using their own capabilities and not relying exclusively on payers. ACO Reporting Accountability. The ACO is responsible for reporting to NCQA on the ACO population using ACO Measure Specifications. Even ACOs with payers that can report quality measures are responsible for submitting the ACO measures they compiled on their own or with the assistance of an analytical third party. Substitute Measures from Collaboratives. PR 1 A (see Appendix 2) allows an ACO to substitute up to 15 out of 40 possible ACO measures for its aggregate ACO population, to accommodate measurement collaborative and other collective reporting initiatives. B-3. Phase-In, Periodicity and Requirements for ACO Data Submission (page 16) Beginning in 2013, the ACO accreditation standards require annual monitoring of core performance measures. Periodicity of Measurement Reports. Consistent with NCQAs health plan HEDIS measurement for accreditation, an annual reporting cycle will be instituted whereby ACOs report in June of the reporting year on the prior calendar measurement year. Phase-In of ACO Data Submission. Beginning in 2013, NCQA expects ACOs to submit measures annually while they are in the accreditation program.

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012 Data Elements for ACO Submission. Consistent with NCQAs requirement for health plan HEDIS reporting, ACOs will be asked to annually submit a set of data elements stipulated for each measure and information about their organization and populations. NCQA expects to phase in requirements for data submission (e.g., begin with data collection methodology, numerator, denominator, rate and attribution methods and expand to requirements similar to health plan HEDIS data submission). B-4. Audit Requirement for ACO Measurement (page 17) A key facet of the NCQA Accreditation programs is their ability to benchmark performance, which requires optimum consistency in measurement and scoring. Consistency is achieved when audits are instituted as a requirement for reporting. Because NCQA encourages emerging ACOs to develop the internal capacity for reporting quality measures across their aggregate ACO population, measure audits will be a future requirement. At this time, NCQA does not anticipate an audit requirement before the June 2014 submission on the 2013 measurement year.

C. Methods and Definitions for Compiling Measures


C-1. Include Hybrid Method for All Measures (page 18) It will be important to accommodate ACOs with a range of data and analytical sophistication. ACO Accreditation standards set the expectation for ACOs to electronically manage health data. HEDIS 2012 Technical Specifications for Physician Measurement allows the Hybrid Method for the use of manually abstracted data for all numerators, for event-driven denominators and all population-based denominators. The Hybrid Method allows manually abstracted medical record data to supplement electronic data in the numerator, increasing compliance estimates. HEDIS Health Plan Measurement allows the use of manually abstracted data for the numerator of a select set of hybrid measures and a single denominator (Controlling High Blood Pressure). Data Collection Methods. NCQA proposes two data collection methods for all measures: the Electronic Method and the Hybrid Method. C-2. Systematic Sampling and Sample Size Requirements (pages 2627) For the Hybrid Method, NCQA proposes to use the HEDIS health plan systematic sampling method, which also requires a minimum sample size from the eligible population. Since the early years of HEDIS measurement in support of health plan accreditation standards, a standard sample size of 411 has been required, to ensure adequate statistical power for comparison between measurement periods in and among organizations. To reduce the measurement burden over time of health plan HEDIS hybrid measures, NCQA provides measurespecific sample minimums based on past performance of more than 250 organizations. NCQAs analysts advise that ACO measurement should start with a sample size of 411 and when adequate experience data are available, offer reductions in sample sizes. Note: The CMS MSSP program uses a minimum sample size of 411 for its measures. Systematic Sampling and Sample Size. For the Hybrid Method, NCQA requires use of the HEDIS systematic sampling method, which requires a minimum sample size from the eligible population. Building from the HEDIS measurement experience, NCQA is adopting the minimum 411 sample size for ACO measures, but allows use of the Hybrid Method in cases where the eligible population is less than the sample size requirement (this may frequently occur, given the relatively small populations measured).

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012 C-3. Exclusions for ACO Measurement (page 15)

For many HEDIS measure specifications, exclusions are cited as the result of careful consideration by expert measurement panels assigned to each measure. Optional exclusions also allow organizations to receive credit for services received, even when a member meets the exclusion criteria. HEDIS 2012 Technical Specifications for Physician Measurement takes a different approach and requires exclusions to ensure consistent measurement and comparisons between providers. NCQA believes that ACO performance measurement is more closely aligned with physician measurement. Both the CMS MSSP and National Quality Forum have shown preference for requiring exclusions. Require Exclusions for ACO Measurement. NCQA proposes to diverge from HEDIS Health Plan Measurement and align with HEDIS 2012 Technical Specifications for Physician Measurement by requiring exclusions to provide more consistent and accurate measurement. C-4. Pharmacy Data for ACO Measurement (pages 2021) Eleven core measures for NCQA ACO Accreditation rely on pharmacy or prescription data in the numerator, in the denominator, or both. An ACO that does not have access to the requisite pharmacy data can opt not to pursue the 11 measures. HEDIS Health Plan Measurement requires the use of dispensed medications (fills) , in order to have the most accurate estimate of care received by the health plan members. In NCQAs HEDIS 2012 Technical Specifications for Physician Measurement, records of prescriptions (scrips) count the same as records of dispensed medications, because of providers general inability to access payers paid pharmacy claims or dispensed data from e-prescription services for their patients. NCQA expects ACOs to have greater access to information about dispensed medications through paid pharmacy claims from their ACO payer and vendor partners. It is critical that payers exchange data with ACO partners, as CMS does through the MSSP. This would give ACOs full knowledge of patients medications; of medications received vs. prescribed by participant providers; and of medications prescribed by providers outside the ACO, in order to effectively manage medications. In this context, NCQA and its external advisory sources considered three options with respect to pharmacy data requirements for ACO measures. Option A Option B Specify the use of only dispensed medications for both denominators and numerators. Specify the use of either dispensed medications or scrips for both denominators and numerators (allowing the greatest flexibility for ACOs without paid pharmacy claims).

Pharmacy Data Requirements. For the first year of ACO Measure Specifications , NCQA would require Option B. For subsequent years and over time, NCQA will raise expectations toward Option A, with transparency in NCQAs measure guidelines.

NCQA 2012

Technical Specifications for ACO MeasuresPublic Comment OverviewObsolete After April 6, 2012

Whats in Public Comment?


Interested parties should review all documents: Public Comment Overview. Provides background and rationale for NCQAs approach to measuring ACO performance and explains how NCQA arrived at some of the decisions reflected in the accompanying documents. Technical Specifications for ACO MeasuresOverview and General Guidelines. The first two sections of the Technical Specifications for ACO Measures publication provide guidance on defining the population for ACO measurement, measure reporting to NCQA, data collection and sampling methods for a core set of performance measures. The publication also includes three appendices for reference. Selected measures Revised specifications for five selected measures that illustrate the application of the general guidelines for ACO measurement to the actual measure specification: Colorectal Cancer Screening. Comprehensive Adult Diabetes CareLDL Control. Use of Appropriate Medications for People With Asthma. Use of Imaging Studies for Low Back Pain. Follow-Up After Hospitalization for Mental Illness.

Submitting Comments
Submit all comments through NCQAs Public Comment Web site (http://publiccomments.ncqa.org). NCQA does not accept comments via mail, e-mail or fax. All comments are due by Friday, April 6, by 5 p.m. ET. To enter comments: 1. Go to the Public Comment database. 2. Enter your e-mail address and contact information. 3. Select Technical Specifications for Accountable Care Organizations (ACO). 4. Select the Topic and Issue on which you would like to comment. 5. Select your support option (e.g., Support, Do not support, Support with modifications, Other). If you choose Do not support, include your rationale in the text box. If you choose Support with modifications, enter the suggested modification in the text box. There is an 1,800 character limit for each comment. Comments are cut off at 1,800 characters. Please try to be brief and to the point in your feedback. We suggest that you develop your comments in Word, in order to check your character limit and save a copy for reference. Use the cut and paste function to copy your comment into the text box.

Thank you for your feedback on the Technical Specifications for ACO Measures!

NCQA 2012

Technical Specifications for ACO Measures


Overview and General Guidelines

This publication is protected by U.S. and international copyright laws. You may reproduce this document for the sole purpose of facilitating public comment. 2012 by the National Committee for Quality Assurance 1100 13th Street NW, Suite 1000 Washington, DC 20005 All rights reserved. Printed in U.S.A. NCQA Customer Support: 888-275-7585 www.ncqa.org

Table of Contents

Table of Contents
Overview Background.................................................................................................................................................... 1 Accountable Care Organization Accreditation ........................................................................................ 1 Provider Quality Measurement in Context .............................................................................................. 1 NCQA and Provider Performance Measurement ................................................................................... 2 1 HEDIS .................................................................................................................................................. 2 Standardizing Measurement ................................................................................................................... 3 Selecting Quality of Care Measures ....................................................................................................... 4 About ACO Measure Specifications .............................................................................................................. 5 Principles of ACO Measurement ............................................................................................................. 6 How Are ACO Measure Specifications Used? ........................................................................................ 6 Cautionary Statements............................................................................................................................ 7 How This Publication Is Organized ......................................................................................................... 8 If You Have Questions About the Specifications .................................................................................... 8 Additional Resources .............................................................................................................................. 8 General Guidelines for ACO Measure Specifications Defining the Population for ACO Measurement .......................................................................................... 11 About ACO Measure Specifications ............................................................................................................ 14 ACO Measure Reporting to NCQA .............................................................................................................. 16 Data Collection Methods ............................................................................................................................. 18 The Electronic Method ................................................................................................................................. 18 The Hybrid Method ...................................................................................................................................... 21 Sampling Methods ....................................................................................................................................... 26 Coding Conventions .................................................................................................................................... 31 References .................................................................................................................................................. 34

Appendices
Appendix 1: Modifications from HEDIS Health Plan Measurement Appendix 2: ACO Standards and GuidelinesPR 1 A: Performance Reporting Appendix 3: Core Measures Included in ACO Measure Specifications

_______
1

HEDIS is a registered trademark of the National Committee for Quality Assurance (NCQA).

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview

Background Accountable Care Organization Accreditation


Accountable care organizations (ACO) are provider-based entities that have come together with the shared goal of taking the responsibility for improving quality of care and reducing cost growth for a group of people. Today, most health care is organized around a site of care and the services provided there. ACOs should transcend the boundaries of particular sites to care for populations over time and across settings. The key tools needed for this approach are careful analysis of patterns of care, to identify high-risk populations, and opportunities for improvement and targeting resources devoted to care management. These will help drive out waste and unnecessary care by identifying and addressing unwarranted variation. Successful ACOs will have committed leaders who set aligned goals and incentives across diverse providers of care and actively engage their patient population. ACOs have the potential to attain better quality at lower cost by aligning incentives to promote coordination and transform health care delivery. However, not every group of providers that wants to call itself an ACO has what it takes to accomplish this vital mission. That is why NCQA, working with a broad array of expert stakeholders, has developed clear criteria and standards for guiding ACOs to success. ACO accreditation provides independent evaluation of organizations abilities to coordinate and be accountable for the high quality, efficient, patient-centered care expected from ACOs. Like the highly successful NCQA Patient Centered Medical Home (PCMH) program, ACO accreditation provides a roadmap for providers, to help them make the challenging and much-needed transformation into ACOs. The NCQA ACO Accreditation program aligns with many of the expectations that the Centers for Medicare & Medicaid Services (CMS) has for the Medicare Shared Savings Program, as well as with common expectations of private purchasers and consumers. NCQA ACO Accreditation will help organizations meet the needs of multiple payers, both public and private. NCQA identified 40 core measures for initial reporting. To reflect the early stages of ACO development, NCQA allowed ACOs to have flexibility in the measures they can report. ACOs may substitute some NCQA measures if the initiatives in which they are involved require it, and if the ACOs can demonstrate their initiatives sophistication and rigor. NCQA plans to move to standardized measures and performancebased scoring (where results and standards are used together to score performance) over time.

Provider Quality Measurement in Context


The Institute of Medicines (IOM) Crossing the Quality Chasm report noted major gaps in quality at all levels of the American health care system. With the exception of trauma and a few acute illnesses, most health care begins in a physicians office, and the course and outcome of most chronic illness is influenced by the effectiveness of a physicians care. Physician practices also provide a critical link to other health care organizations such as insurers, hospitals, nursing homes, medical networks and government agencies. ACOs offer the promise of integrating the continuum of care around patients needs and services and provid ing a stronger focus on prevention; care coordination and management; and greater efficiency. Measurement of provider (e.g., physicians, hospitals, ACOs) care is a subject of intense interest and vigorous discussion among health care and policy leaders. There are barriers to measurement efforts: a lack of wellspecified, accurate measures; access to the requisite administrative and clinical data; problems obtaining appropriate sample size; and questions about the ability to influence measure results through physician intervention. Because of continued concerns about the quality and cost of care and the advent of innovative reimbursement and health insurance programs, providers, payers (public and private) and consumers are more interested in expanded quality and cost measurement and reporting. As joint efforts to develop standardized measures have evolved, there have been many changes in physician performance measurement. Notable among them are NCQAs work with the AQA , the National Quality Forum (NQF), the American Medical AssociationPhysician Consortium for Performance Improvement (AMA-PCPI) and several physician specialty societies.

Draft 2013 Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview

NCQA released updated physician and hospital quality (PHQ) standards for organizations in June 2008. The release of HEDIS 2012 Technical Specifications for Physician Measurement (HEDIS Physician Measurement) further encourages and supports standardized provider performance measurement and data reporting, the demand for which has never been higher. These products reflect the call for rigorous measurement criteria, the options for overcoming the methodological and logistical challenges associated with obtaining meaningful provider performance measurement and the need for data to support informed decision making and pay-for-performance efforts.

NCQA and Provider Performance Measurement


Recognition Programs. NCQA Recognition programs enjoy broad support among purchasers, physicians and plans, and provide tools to measure physician performance in critical areas of care. These measurement and reporting programs include information on patients across a physicians entire practice, as opposed to only patients enrolled in a particular health plan. A substantial number of health plans and some employers report recognition and actively promote physician involvement in NCQA Recognition Programs through financial assistance and rewards. NCQA has expanded HEDIS Physician Measurement to include nearly all measures and methods from these programs. Diabetes Recognition Program (DRP), launched in 1997, recognizes physicians and other clinicians, who use evidence-based measures and provide excellent care to their patients with diabetes. Heart/Stroke Recognition Program (HSRP), launched in 2003, recognizes physicians who deliver high-quality care to their patients with cardiovascular disease or who have suffered a stroke using key quality performance measures that are based on national evidence based guidelines. Physician Practice Connections (PPC) launched in 2004, recognizes physician practices that use that use systematic processes and information technology to enhance the quality of patient care. Back Pain Recognition Program (BPRP) launched in 2007, recognize physicians and chiropractors who provide superior care to their patients with back pain based on widely accepted medical evidence, with significant input from physician specialists and health plan and employer representatives. Patient Centered Medical Home (PCMH) launched in 2011 as a new version of the Physician Practice Connections-Patient-Centered Medical Home program, to identify physicians who deliver superior care using standards firmly rooted in medical evidence. Integrated Healthcare Associations Pay for Performance Program in California. NCQA works with the Integrated Healthcare Association (IHA), which launched the California Pay for Performance (P4P) program, a statewide collaboration that creates the business case for physician organization (PO) quality. Data are integrated from a variety of sources and used by health plans to determine the financial reward paid to participating California Provider Organizations. NCQA serves as the data aggregator for IHAs P4P program and as technical advisor and consultant in P4P measurement.

HEDIS
HEDIS, the most widely used set of performance measures for inpatient and ambulatory care, is developed and maintained by NCQA, a private, nonprofit organization dedicated to improving health care quality through measurement, transparency and accountability. NCQA accredits and certifies a wide range of health care organizations; recognizes physicians and physician practices in key clinical areas; and provides health care quality information in order to help consumers, employers and others make informed health care choices. Originally designed to address private employers needs as purchasers of health care, HEDIS measurement specifications have been adapted for use by public purchasers, regulators, consumers and clinicians. Quality improvement activities, health management systems and physician profiling efforts have often based their measurement efforts on HEDIS.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview

HEDIS has evolved into a critical component of an emerging measurement system to establish accountability in health care. HEDIS measures are equally useful as part of a quality improvement effort in a physician practice; as a purchaser request for quality information at the health plan or clinician level; as elements of NCQAs numerous Accreditation, Certification or Recognition Programs; and as the basis of consume r report cards. HEDIS is also the model for emerging systems of performance measurement in other areas of health care delivery.

Standardizing Measurement
NCQA addresses the following key concepts to create standardized, reliable performance measures some of these are directly addressed in this publication and others are explicit in our measure development process. Engage in a consensus process Involve multiple health care stakeholders. Model national organizations best practice processes. Review measures against explicit criteria. Include Public Comment mechanisms to enrich perspectives on measures proposed for consensus. Technical construction of rates Ensure that measures reflect evidence-based medicine. Define the eligible population as the denominator (age, gender, diagnoses, and exclusions). Define the desired event as the numerator (processes, outcomes, time windows). Define data elements (e.g., electronic codes, diagnoses, clinical procedures/tests, medications). Field-test to refine validity of measures. Define audit or data validation specifications. Define implementation rules and guidelines Specify accountable parties and to whom measures apply. Define units of observation (groups, sites, individuals, specialties). Define how patients are assigned to observation units. Specify required statistical and technical considerations. Analysis of information Use descriptive statistics (distribution, confidence intervals, benchmarking) in reporting. Promote understanding through analytical statistics (correlation, regression, and modeling). Annual maintenance and review of measures and guide-lines Perform regular refinement (updating codes, clinical logic, evidence re-review, evidence-based medicine). Perform ad hoc review (take action based on emerging, ground-breaking evidence). Maintain and update technical specifications and data collection tools.

Draft 2013 Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview

Selecting Quality of Care Measures


The defining question for all performance measurementWhere can measurement make a difference? can only be answered after carefully considering a number of factors. NCQA and NQF have established three areas of desirable attributes for HEDIS measures. 1. Relevance. 2. Scientific Soundness. 3. Feasibility. Whether considering the value of a new measure or the continuing worth of an existing one, we must define what makes a measure desirable, as follows. Measures should address features of health care that are relevant to purchasers or consumers for making health care choices, or which will stimulate internal efforts toward quality improvement. Measures should be meaningful to at least one HEDIS audience (patients, purchasers or health care systems). Decision makers should be able to understand a measures clinical and economic significance. The type of measure (e.g., outcome or process), prevalence of medical condition addressed by the measure, and seriousness of affected health outcomes are all considered during the development of quality measures. Measures should relate to activities that have high financial impact. Measures should encourage the use of cost-effective activities or discourage the use of activities that have low cost-effectiveness. Measures should encourage activities that use resources most efficiently to maximize health. Health care systems should be able to take action to improve their performance. For outcome measures, there should be at least one process that has an important effect on outcome and can be controlled. For process measures, there should be a strong link between process and desired outcome. There should be the potential for wide variation across systems. There should be substantial room to improve performance.

Relevance
Meaningfulness

Health importance

Financial importance Cost effectiveness Strategic importance Controllability

Variance among systems Potential for improvement

Scientific Soundness
Clinical evidence Reproducible Valid

Perhaps in no other industry is scientific soundness as important as in health care. Scientific soundness must be a core value of our health care system. There should be evidence that documents the link between clinical process and outcome addressed by the measure. Measures should produce the same results when repeated in the same population and setting. Measures should make sense logically and clinically. They should correlate well with other measures of the same aspects of care.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview Accurate Risk adjustment Measures should precisely evaluate what is actually happening. Measure variables should not differ appreciably beyond the health care system s control, or variables should be known and measurable. Risk stratification, or a validated model for calculating an adjusted result, can be used for measures with confounding variables. Accuracy, reproducibility and validity should not be affected if different systems use different data sources for a measure. The goal is not only to include feasible measures, but also to catalyze a process wherein relevant measures can be made feasible. Measures should have clear specifications for data sources and methods for data collection and reporting. Measures should not impose an inappropriate burden on health care systems. Data collection should not violate accepted standards of patient confidentiality. Required data should be available. Measures should not be susceptible to manipulation or gaming that would be undetectable in an audit.

Comparability of data sources

Feasibility
Precise specification Reasonable cost Confidentiality Logistical feasibility Auditability

About ACO Measure Specifications


In November 2011, NCQA released Standards and Guidelines for the Accreditation of Accountable Care Organizations. NCQA ACO Accreditation includes a set of performance measures for reporting and, over time, performance-based scoring. The measures are drawn from HEDIS. NCQA anticipates eventual opportunities for standardization in ACO measurement across the industry. The Technical Specifications for ACO Measures (ACO Measure Specifications) are an extension of NCQAs significant experience and record of accomplishment with establishing and maintaining provider quality standards, many of which build from more than 17 years of developing and testing HEDIS measures. Both the HEDIS 2012 Technical Specifications for Health Plans (HEDIS Health Plan Measurement) and HEDIS Physician Measurement were used as the foundation for the ACO Measure Specifications publication because of its inherent focus on evaluating provider performance. Appendix 1 contains a table that compares ACO Measure Specifications to HEDIS Health Plan Measurement, because of its broad use to measure and report performance. These specifications are a component of NCQA ACO Accreditation, and in turn support the variety of accompanying innovative accountable care reimbursement programs that commonly require quality performance thresholds to attain financial incentives. Development of ACO Measure Specifications was guided by the following goals: Align with HEDIS Health Plan Measurement to build on its support of another NCQA accreditation program. Attain synergy with CMS Medicare Shared Savings Program (MSSP) rules since CMS is emerging as a national ACO leader. Leverage the knowledge and experience of Californias IHA. Strive for consistent measurement between ACOs, for future benchmarking. Encourage ACOs to develop the capacity for electronic evaluation of quality of care across populations.

Draft 2013 Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview Respect the current limitations of most ACOs with respect to their ability to measure quality of care across populations. Allow ACO measurement to evolve as NCQA, ACOs and others gain experience.

The ACO Measure Specifications publication specifically supports PR 1 A: Performance Reporting, Element A. Core Performance Measures (Appendix 2). It enumerates specific requirements for reliable, valid, equitable ACO reporting, and is intended for use by ACOs or by third-party organizations conducting population health analytics on behalf of the ACO for all populations served by the ACO. Measure specification complexity varies because of the varying amount of data available to ACOs. ACOs with payer contracts that include the exchange of paid claims (pharmacy and medical), enrollment and laboratory data for the ACO population, as well as fully functioning EHRs across all settings of care, may have a rich array of data available. Other ACOs may need to use manual billing systems and paper records that differ in level of completeness. Recognizing that measurement programs applied across a diverse set of ACOs must accommodate the least advanced measurement capabilityat least for nowNCQA has attempted to provide specifications that apply to all levels of data sophistication. For comparative purposes, the ability to satisfy performance reporting requirements for NCQA ACO Accreditation will distinguish the capacity of ACOs to improve the quality of care for the populations they serve.

Principles of ACO Measurement


The content of ACO Measure Specifications was developed with four key principles in mind. 1. Increase provider measurement standardization. Standards support leading provider measurement activitiesincluding multiple private sector collaborative measurement initiatives and NCQA Recognition programs, which both enjoy broad purchaser support. When organizations pool data on standardized measures, they can produce results with greater statistical reliability. Collaboration can also produce results that physicians and other providers trust. 2. All measurement activities follow sound, rigorous, transparent methodology. Commitment to critical methodological principles and defined processes and transparency (particularly with respect to performance reporting) must form the foundation of any measurement effort. ACOs should allow provider input into all measurement programs, so providers can understand how they are affected (e.g., physician tiering, submitting questions, requesting corrections before data are used). Similarly, ACOs should clearly communicate how they measure and make decisions, so consumers and purchasers understand the meaning of ratings or rankings and their limitations. 3. Measurement standards address methodological challenges of distinguishing ACO performance fairly. Purchasers, providers and patients are eager to receive meaningful performance information about ACOs, to support ACO and provider selection. 4. Quality is the centerpiece of meaningful ACO measurement. Quality must not be sacrificed in the name of reduced cost. Often, increasing quality can actually decrease overall cost and resource use.

How Are ACO Measure Specifications Used?


Quality improvement. ACOs can use performance information to reduce variation and to track progress in achieving goals defined by evidence-based guidelines. Utilization improvement. In conjunction with quality measurement, ACOs use performance information to help provider participants increase underutilized services, decrease overutilized services or prevent misuse.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview Sharing data. By using the underlying denominator and numerator definitions for population measures, ACOs can also provide data to help physicians, hospitals, practice sites or medical groups identify and engage patients who have missed appropriate services or need follow-up. Establishing contract performance standards. Payers will include performance standards in ACO contracts with financial consequences, withholding a portion of payment for low performance or providing additional payment for high performance. Managing ACO physician networks. Some ACOs may use performance information to identify high- and low-performing physicians in selecting or tiering their networks. Benefit design. To address significant gaps in quality performance, ACOs may collaborate with payer partners to structure benefits to promote patients lifestyle choices, provider choices and compliance with treatment plans through differential copay or coinsurance for specific services, provider types or drugs. Pay for performance. These measures can be the basis for rewarding high levels of performance or improvement in performance. External reporting for consumer decision making. In accordance with NCQA ACO Accreditation, ACOs will be expected to offer publicly available performance reports.

Cautionary Statements
NCQA cannot guarantee results. The measures included in the ACO Measure Specifications are not clinical guidelines and do not establish a standard of medical care. Gaps in the breadth and completeness of health care quality measurement affect the ability to provide performance measures applicable to a variety of value-based outcomes. Some gaps exist because of a lack of scientific knowledge regarding treatments or interventions; others exist because of a lack of consensus regarding treatment approaches; still others exist because of a lack of consensus around measures of value to stakeholders, such as patients. As methodology and technology advance the capability to effectively address these issues, so too will the ACO Measure Specifications. Because of the complexity of methods and technical issues related to using measurement to evaluate performance, this publication cannot replace consultation with statisticians and quality measurement experts, which is necessary to ensure that measures and measurement methods described in this publication are both appropriate to and implemented in the most effective manner for a specific program. This is especially critical where measures are used outside of internal quality improvement. Give careful attention to the limitations of the data and methodology before comparing ACO performance with external benchmarks. Use caution when comparing results obtained through ACO Measure Specifications with other HEDIS measurement sets, such as NCQAs HEDIS Health Plan Measurement. The methodology used to compile the measures may vary (e.g., the type and source of data) While this release of ACO Measure Specifications does not currently provide explicit auditing guidelines for measurement, ensuring the validity and integrity of data and methodology is an important component of any measurement program.

Draft 2013 Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Overview

How This Publication Is Organized


Overview General Guidelines for ACO Measurement Specifications Appendices Provides general background; an overview of ACO Measure Specifications, including cautionary statements; and information about this release. General guidelines for ACO Measure Specifications, include specifications which apply to all measures namely: defining the population for ACO measurement, ACO measure reporting to NCQA, data collection and sampling methods and coding conventions. Appendix 1: Modifications from HEDIS Health Plan Measurement Appendix 2: ACO Standards and GuidelinesPR 1 A: Performance Reporting Appendix 3: Core Measures Included in ACO Measure Specifications

If You Have Questions About the Specifications


NCQA provides different types of policy support to customers, including a function that allows customers to submit specific policy interpretation questions to NCQA staff: the Policy Clarification Support (PCS) system. The PCS can be accessed on the NCQA Web site (www.ncqa.org/pcs).

Additional Resources
In addition to its publications, NCQA provides a variety of resources to help user organizations understand and use these measure guidelines and specifications. E-pubs. All HEDIS publications are available as easy-to-use electronic publications (e-pubs) that contain the complete text of NCQA printed publications and are sold by user licenses. E-pubs are protected Microsoft Word and Excel files sent to the purchaser via e-mail. E-pubs are simple to download onto a PC, network or intranet. HEDIS Electronic Coding Table. To save programming hours, eliminate the manual search for codes and reduce keying errors, NCQA offers the HEDIS Electronic Coding Table (ECT). Available in 2, Microsoft Excel, the ECT provides an easy way to incorporate CPT HCPCS, ICD-9-CM, UB Revenue, 3 UB Type of Bill, DRG and Logical Observation Identifiers Names and Codes (LOINC ) into a data collection program. Educational Seminars. NCQA educational seminars provide valuable information on NCQA quality measurement programs. Course offerings range from a basic introduction to HEDIS and NCQA standards to advanced techniques for quality improvement. For information about NCQA conferences, go to www.ncqa.org/ Education or call NCQA Customer Support at 888-275-7585. Publications on Quality Measurement. NCQA produces many publications that are relevant to quality measurement. To obtain a list or to order publications, call the NCQA Publications Center at Customer Support at 888-275-7585 or go to www.ncqa.org.

____________
2

CPT is a trademark of the American Medical Association. Current Procedural Terminology (CPT) is copyright 2012 American Medical Association. All Rights Reserved. No fee schedules, basic units, relative values, or related listings are included in CPT. The AMA assumes no liability for the data contained herein. Applicable FARS/DFARS restrictions apply to government use. LOINC is a registered trademark of the Regenstrief Institute.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications

11

Defining the Population for ACO Measurement 1. Minimum Population for Measurement
An ACO must have a minimum of 5,000 patients in the population targeted for ACO measurement as of January 1 of the measurement year.

2. ACO Contracts and Measurement


NCQA requires ACOs to report measures on their aggregate ACO population (i.e., the entire population under contract between payers and the ACO). The term contract means a formal, accountable care reimbursement contract (e.g., shared savings) between a legal entity that defines the ACO and with public or private payers (including employers). NCQA recognizes that in the early years of ACO accreditation, health care providers transforming toward the capacity for population accountability may not yet have formal ACO contracts; therefore, ACOs will use the following alternatives to define the population for measurement as of January 1 of each measurement year. Method 1 1A 1B If the ACO has at least one ACO contract in force, it has two options: Report on the aggregate ACO contract population, as defined through contractual attribution models or assignment methods, or Report on the aggregate ACO contract population, plus all or at least 80 percent of the patient population (consistent with how the population is defined for NCQA ACO Accreditation), as defined by NCQAs recommended patient attribution model. If the ACO has no ACO contract in force, the ACO reports on all or at least 80 percent of its patient population (consistent with how the population is defined for NCQA ACO Accreditation), using NCQAs recommended patient attribution model.

Method 2

For the purposes of Methods 1B and 2, two steps apply when using NCQAs Recommended Patient Attribution Model. Step 1 Define the universe of an ACOs patient population as all active patients as of January 1 of the measurement year (and for purposes of actual measurement, defined as of December 31 of the measurement year) with any provider participating in the ACO as contractually defined by the ACO (e.g., primary and specialty care, acute care, surgical care, post-acute care, long term care). Apply NCQAs Recommended Patient Attribution Model to at least 80 percent of that universe of patients. The intent is for an ACOs entire patient population to be considered through this process.

Step 2

NCQA expects that ACOs will consistently define the population across all performance measures. NCQA requires ACOs to disclose the attribution models used to identify their population for measurement, regardless of the model used to define the population. Refer to Disclosure of Patient Attribution Models.

3. Recommended Patient Attribution Model


NCQA recognizes that patient attribution or assignment is contractually defined and may differ by ACO contract with public and private payers. As indicated in ACO Contracts and Measurement, NCQA does not intend to interfere with existing ACO contracts and how those populations are defined, but encourages longterm standardization of patient attribution models across the health care industry, and seeks harmony with CMS accountable care evaluation methods. NCQA also recognizes the ongoing empirical contributions to this model made by researchers at The Dartmouth Institute.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

12

General Guidelines for ACO Measure Specifications

To satisfy the requisite specifications in this publication, NCQA recommends the attribution model as defined by the CMS MSSP, with two adaptations as follows. 1. Apply a 24-month period, rather than the CMS 12-month period, given the broader demographic of all ages, rather than only CMS beneficiaries, who use more services than younger cohorts. 2. Apply the plurality of primary care visits, rather than the CMS plurality of primary care allowed charges, given the probable dependence on payer paid claims to determine allowed charges. This applies to ACOs that opt to use Method 1B or Method 2 to define their ACO population, in accordance with ACO Contracts and Measurement, above.

Patient Attribution Model


This model is annotated from the CMS MSSP Final Rule, with brackets in bold type indicating the adaptations, described above, for use with ACO measurement as defined in this publication. Patients are assigned to an ACO through a step-wise process. The ACO starts by identifying all patients that had at least one primary care service (defined below) during the prior [24-month] period with a physician who is an ACO provider or supplier of that ACO. Then the ACO identifies all of those primary care services rendered by primary care physicians (defined below). Patients are assigned to an ACO if the patients [visits] for primary care services to all primary care physicians who are ACO providers/suppliers in the ACO are greater than the [number of visits] for primary care services furnished by primary care physicians who are ACO providers/suppliers in any other ACO or not affiliated with any ACO. The second step considers the remainder of the patients who received at least one primary care service from an ACO physician, but who have not had a primary care service rendered by any primary care physician, either inside or outside the ACO. The patient will be assigned to an ACO if their [visits] for primary care services furnished to the patient by all ACO professionals who are ACO providers/suppliers in the ACO are greater than the [number of visits] for primary care services furnished by all ACO professionals who are ACO providers/suppliers in any other ACO or other physicians, nurse practitioners, physician assistants, clinical nurse specialists who are unaffiliated with an ACO.
Source: Annotated from the final CMS MSSP rules published in the Federal Register, Vol. 76, No. 212, Wednesday, November 2, 2011, Rules and Regulations, 425.400, page 67983, available at: http://www.gpo.gov/fdsys/pkg/FR2011-11-02/pdf/2011-27461.pdf

4. Defining Primary Care Providers for the Attribution Model


Physician means a doctor of medicine or osteopathy.

Primary care physician includes general or family practice physicians, geriatricians, general internal medicine physicians, general pediatricians and obstetricians/gynecologists (OB/GYN). Primary care professional includes physician or nonphysician (e.g., physician assistant, nurse practitioner) who offers primary care services. Licensed practical nurses and registered nurses are not included in this definition.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications

13

5. Code Set for the Attribution Model


NCQA defines the applicable visit codes for primary care services in accordance with the CMS MSSP. According to CMS, Revenue Center Codes are included to allow for the participation of FQHCs and rural health centers. These codes are contained in the ACOs participating providers billing and claims data, as well as the ACOs payer partners adjudicated claims and encounter data. This code set will be maintained annually by NCQA as this publication is published, reflecting updates by CMS.

Primary Care Services


CPT/HCPCS Level I Codes HCPCS Level II G Codes UB Revenue Center Codes 9920199215; 9930499340; 9934199350 G0402 (the code for the Welcome to Medicare visit), G0438 and G0439 (codes for the annual wellness visits). 0521, 0522, 0534, 0525 submitted by FQHCs (for services furnished prior to January 1, 2011), or by Rural Health Centers.

Source: Final CMS MSSP rules published in the Federal Register, Vol. 76, No. 212, Wednesday, November 2, 2011, Rules and Regulations, 425.20 page 67975, available at: http://www.gpo.gov/fdsys/pkg/FR-2011-11-02/pdf/2011-27461.pdf

6. Retrospective Patient Attribution for Measurement


Although the patient populations targeted for ACO measurement are determined as of January 1 of the measurement year, patient attribution to those populations for ACO measurement should be determined as of December 31 of the measurement year. This is consistent with the CMS MSSP, whereby there is an initial, prospective attribution at the start of the measurement year to determine the populations to be included, with a final attribution at the close of the year for evaluation and measurement. This specification applies to any method used by an ACO, in accordance with ACO Contracts and Measurement, above. While this defines the aggregated patient population earmarked for measurement to evaluate the ACO, specific eligibility requirements for each measure, as outlined in the measure specification in this publication, will further define the subset of the aggregate population that will be eligible for each measure.

7. Disclosure of Patient Attribution Models


NCQA requires that ACOs disclose the models of patient attribution used to define the population used for measurement. This allows NCQA to collect information on ACO attribution models, for exploration of the potential for future standardization in methods across ACOs. When ACOs use multiple attribution models because they have more than one contract with a payer, NCQA requests information on the model used and the proportion of the aggregate ACO population identified using those models. Disclosure will occur through measure submission to NCQA. Refer to Required Data Elements for ACO Report Submission, below.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

14

General Guidelines for ACO Measure Specifications

About ACO Measure Specifications 8. Continuous Enrollment and Patient Attribution


In NCQAs HEDIS Health Plan Measurement, each measure has a defined set of continuous enrollment criteria to ensure each member in the eligible population is appropriately attributed to the health plan, and that the health plan has adequate time to affect the members health care. In HEDIS Physician Measurement, patient inclusion criteria are provided as an option to continuous enrollment for provider-generated measures, and define the number of patient encounters with the provider for patient inclusion in the denominator of each measure. NCQA recognizes that ACOsnot individual providersare broadly accountable for care that patients receive across the ACO and its participating providers, and believes that ACO attribution models adequately define patient inclusion criteria for the ACO population being measured. Therefore, NCQAs ACO Measure Specifications will not include continuous enrollment or patient inclusion criteria within each measure specification.

Example: Breast Cancer Screening


Publication HEDIS Health Plan Measurement (Continuous Enrollment) Measure Specific Patient Criteria Continuous medical benefit enrollment for the measurement year and the year prior to the measurement year, with no more than one gap in continuous enrollment of up to 45 days during each year of continuous enrollment. To determine continuous enrollment for a Medicaid beneficiary for whom enrollment is verified monthly, there may not be more than a 1-month gap in coverage during each year of continuous enrollment. The member must be enrolled as of December 31 of the measurement year. Health Plan: Continuous Enrollment. Non-Health Plan: Any enrollment, claim or encounter transaction any time during the measurement year and the year prior to the measurement year. None.

HEDIS Physician Measurement (Patient Inclusion Criteria) ACO Measure Specifications

9. Measurement Period
The previous calendar year is the standard measurement year for HEDIS data and is to be used for ACO measurement. . Although measurement periods can be adjusted for internal use as long as all date-related specifications are also adjusted and data completeness (e.g., run out to allow claim processing time) is carefully considered, non-calendar year reporting periods do not satisfy ACO Measure Specifications.

10. Measure Format


Each measure specification has nine sections. 1. Modifications from HEDIS 2012 Volume 2: Technical Specifications for Health Plans. 2. Description. 3. Calculation (if applicable). 4. Definitions (if applicable). 5. Eligible population (defining the population and required exclusions). 6. Electronic Specification. 7. Hybrid Specification. 8. Note (if applicable). 9. Data Elements for Reporting.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications

15

11. Eligible Population Criteria


The eligible population includes all patients of the aggregate ACO population who meet the specified criteria. Four criteria define the eligible population for a specific measure. 1. Age and gender specifies any age group and gender requirement (e.g., Cervical Cancer Screening is collected for women 24-64 years of age). 2. Defining the population points to related specifications within the General Guidelines and determines whether Medicaid stratification is allowed for the measure. 3. Event/diagnosis specifies the medical event or diagnosis requirements for the eligible population (e.g., patients must have a diagnosis of AMI for inclusion in the Persistence of Beta-Blocker Treatment After a Heart Attack measure). In some cases, steps are provided to define the event/diagnosis and tables are provided that define codes to use for the steps. Note: This measure is used as an example only and is not included in this publication. 4. Exclusion specifies criteria for any mandatory exclusion in the measure using either the Electronic Method or the Hybrid Method.

12. Exclusions
For many HEDIS measure specifications, exclusions are the result of careful consideration by expert measurement panels assigned to each measure. For HEDIS Health Plan Measurement and California Integrated Healthcare Associations measure specifications, exclusions are optional unless otherwise specified. Optional exclusions allow ACOs to receive credit for services received even when a member meets the exclusion criteria. HEDIS Physician Measurement takes a different approach and requires exclusions to ensure consistent measurement and comparisons between providers. NCQA believes that ACO performance measurement more closely aligns with physician measurement. Therefore, for ACO Measure Specifications, NCQA requires exclusions to provide for more consistent measurement and future benchmarking of ACO performance. Exclusion criteria are specified for both the Electronic Method and the Hybrid Method.

13. Data Considerations for Measure Denominators


There are two types of denominators; each requires a different type of data. It is important to note the distinction, since it may affect the ACOs ability to use a particular data collection method. 1. Population-dependent denominators (e.g., age and gender for preventive care measures) require an electronic data source that identifies all eligible patients of the ACOs aggregate population defined for measurement, rather than only those who use services (e.g., utilization of health care services, specific diagnoses). 2. Event-dependent denominators (e.g., acute or chronic care measures), require identification of specific health care services or diagnoses. Health services data from various data sources are required to identify eligible patients. The ACOs ability to access or collect different types of data will influence its ability to measure performance using a given method, so an ACO should consider these questions when preparing to start its performance measurement: 1. What data collection methodologies are specified for each measure? 2. What are our available data collection methodology options? 3. What data collection methodology meets the needs of our program? 4. How can our organization effectively access information needed to produce the measures? 5. Where, and how, is the information stored?

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

16

General Guidelines for ACO Measure Specifications

ACO Measure Reporting to NCQA 14. Stratifying the Population for ACO Measurement
ACOs are designed to focus on the care of populations, rather than on specific managed care product lines or age groups. NCQA recognizes that providers generally deliver care regardless of payer or product, and believes is it important not to artificially compartmentalize measures through stratification. NCQA also recognizes that Medicaid programs are entering into ACO contracts and that Medicaid populations historically perform differently on quality measures than other populations. ACOs are required to report on their aggregate population as a whole, but have the option to report on Medicaid separatelybut only for measures with Medicaid specifications defined for the corresponding measure in HEDIS Health Plan Measurement. Refer to Appendix 3: Core Measures Included in ACO Measure Specifications. This publication can be used outside the NCQA ACO Accreditation process for discrete populations attributed to the ACO. For example, an ACO may desire to have separate measurement reports for each ACO population it manages, by payer contract, but this section specifies the allowed population stratification for measurement for NCQA ACO Accreditation.

15. ACO vs. Payer or Collaborative Reporting


Payers may provide data or reports to the ACO for quality reporting. For an ACO, this could occur with a single payer, or with multiple payers if there is more than one ACO contract in force, but even if ACOs have a payer that can report quality measures on their behalf, ACOs are responsible for submitting measures compiled on their own or with the assistance of an analytical third party. This does not preclude an ACO from having a separate arrangement with payer partners to receive quality measurement reports for the populations for which it is accountable.

16. Reporting Accountability


Although ACOs are responsible for reporting on the defined ACO population using the measurement specifications defined in this publication, NCQA ACO Accreditation Standards offer the use of substitute measures from collaboratives. This allows an ACO to substitute ACO measures for its aggregate ACO population from measurement collaboratives or other collective reporting initiatives in which the ACO participates. Refer to Appendix 2: ACO Standards and Guidelines ExcerptPR 1 A: Performance Reporting for more information.

17. Reporting Periodicity


ACOs must monitor core performance measures annually beginning in 2013. ACOs in NCQAs accreditation program report on the prior calendar year in June of the reporting year. .

18. Required Data Elements for ACO Report Submission


ACOs must annually submit a set of data elements stipulated for each measure, as well as information about their organization and populations. NCQA expects to phase in requirements for data submission, beginning with requiring information about the ACOs data collection methodology, numerator, denominator, rate and attribution methods, and expanding to requirements similar to health plan HEDIS data submission.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications

17

19. Reporting Rate Stratification


Several measures provide specifications for stratifying the indicator by categories such as age. While this analytical methodology is intended to provide more actionable and meaningful information to the ACO, some measures only require the total or overall rate for submission of rates to NCQA during the initial phase. Refer to the individual measure specifications for additional guidance.

20. Small Numbers


When fewer than 100 patients are in the eligible population for a reported measure during the entire measurement year, include in the denominator all patients who meet the criteria and report a 95 percent confidence interval. If the denominator used to calculate a measure is smaller than 30 using either the Electronic or Hybrid Method, the ACO is not required to report the rate but must provide all other information, including the following. A count of all patients eligible for the measure, as defined by the measure specification (the eligible population/denominator). A count of all patients who received the treatment or service as indicated (the numerator).

21. Audit Requirements for ACO Measurement


A key part of NCQA Health Plan Accreditation is benchmarking performance, which requires optimum consistency in measurement and scoring. Consistency is achieved when audits are required for reporting. NCQA encourages emerging ACOs to develop the internal capacity to accurately report quality measures across the aggregate ACO population. NCQA will develop an ACO measurement audit program in the future to ensure accuracy and integrity of measurement and reporting. NCQA does not anticipate requiring an audit before the June 2014 submission for the 2013 measurement year.

22. Benchmarking
In this context, benchmarking is defined as comparing one ACO quality performance with that of another ACO, or against national averages for ACOs. Benchmarking ACO measurement is valuable for providing context to measurement for ACOs and their stakeholders, but it is not yet viable. Until standardization evolves, benchmarking results across all ACOs at a national level will be difficult because of the numerous variations of such efforts. However, establishing national ACO measurement consistency is one of the key goals of NCQA ACO Accreditation standards, in order to support benchmarking in the future. NCQA advises caution when comparing results with other quality measurement programs whose specifications differ, even if the differences seem minimal; this includes comparing HEDIS health plan results to ACO results.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

18

General Guidelines for ACO Measure Specifications

Data Collection Methods 23. Data Collection Methods


There are two primary data collection methods included in this publication and outlined in detail, below. 1. Electronic Method uses data routinely captured during the course of care delivery for either payment or care documentation purposes. The Electronic Method relies on standardized, systematic sources of electronic data such as medical claims, pharmacy claims, patient registration data, EMR or EHR databases, laboratory results, and radiology results. 2. Hybrid Method leverages the Electronic Method and supplements numerator data with data manually abstracted from medical records. The Hybrid Method, relies on the Electronic Method to identify the eligible population (denominator) and then allows medical record data to supplement electronic data when identifying the numerator of a measure. The Hybrid Method is an option for all measures in NCQAs ACO Measure Specifications.

The Electronic Method 24. Using the Electronic Method


The Electronic Method requires ACOs to identify the eligible population (denominator) and numerator using standard, systematic data sets. ACOs report a rate based on all patients from the aggregate ACO population who meet the eligible population/denominator criteria. Step 1 Step 2 Step 3 Identify the eligible population. This number is the denominator for the measure. Search electronic data to identify numerator events for all patients in the eligible population. Calculate the rate.

25. Electronic Data Definition


In this publication, electronic data refers to both electronic administrative data and electronic clinical data. NCQA considers electronic data to include all data, whether derived from claims or from clinical data that are stored, retrieved and analyzed electronically. For the purposes of this section, the term claims used below refers to either paid claims from a payer or claims found in the ACOs billing systems. Further, the term encounters refers to services typically captured by a payer as delivered, but not directly paid for, commonly the artifact of capitation or bundled payments for services. Demographics of patients served by an ACO found in health plan enrollment systems or provider practice management systems. Professional claims and encounters. Inpatient facility claims. Outpatient facility claims. Mental health service claims. Radiology procedure claims. Pharmacy claims, including data from a pharmacy benefit manager. Laboratory claims or encounters, including electronic data from a lab vendor.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications Laboratory results, including electronic data from a lab vendor. Retrievable coded data from EMR or EHR databases and patient registries. Some EMRs and EHRs do not meet this definition and are considered medical records from which data must be manually abstracted.

19

26. Cautions for Using Electronic Data


Traditional administrative data includes only transactional data from claims submitted by providers for payment (and often includes diagnoses) and are then adjudicated by payers. The expanded set of electronic data now used by many payers, and some providers, includes laboratory, pharmacy, mental health and pharmacy claims. Electronic data restricted to only visit and procedure claims impose substantial limitations on defining a broad set of performance measures that can lead to valid and reliable results. Many meaningful, accurate and reliable performance measures can be computed relying on the expanded definition of electronic data, although even expanded electronic data have limits for measurement of clinical performance. To address some performance aspects appropriately, detailed clinical data contained in (multiple) medical records may be critical. When using data from an electronic data source for quality of care measurement, remember: Information from an automated appointment scheduler must distinguish between scheduled and completed services. Only kept appointments or completed services count toward measure compliance. The date of service is required for all services rendered. All data elements specified in the measure (e.g., date of service, procedure codes, diagnosis codes, prescriptions) must be identified as rendered by the time specified (e.g., a mammogram during the measurement year or the year prior to the measurement year for the Breast Cancer Screening measure).

27. Nontransactional Databases


Internally developed clinical databases Externally developed databases Required data elements of nontransactional databases ACOs may create and use internal databases for ACO measurement, such as case management databases, utilization management (UM) databases or patient registries populated with medical record information. NCQA considers these electronic data, as opposed to manually abstracted medical record data. Any automated data supplied by contracted practitioners, vendors or public agencies (e.g., pharmacies, labs, immunization registries, schools, public health agencies) constitute an external electronic database, and may be used as such. The database should include all data elements specified in the measure (e.g., date of service, procedure codes, diagnosis codes, prescription, practitioner type), and services should be identified as having been rendered within the time frame specified in the measure. All database entries should be made and all services rendered by the deadline for delivery of the service established in the measure.

28. EMR and EHR Databases


An electronic medical record (EMR) is usually synonymous with a paper medical record, which can be episodic and not integrated across episodes and sites of care. An electronic health record (EHR) is a longitudinal, electronic record of patient health information (including information from the patient) that integrates data from various sites and episodes of health care services ideally, over a lifetimeinto one record. EHRs are not yet widely available although Meaningful Use criteria and incentives from the federal Department of Health and Human Services have spurred rapid adoption of this technology.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

20

General Guidelines for ACO Measure Specifications

To be considered a source of electronic data for purposes of this publication, EMR/EHR information must be aggregated into a database that stores multiple records and can be programmed for electronic evaluation according to ACO Measure Specifications. Historical patient data that have been entered into the EMR or EHR manually or electronically are considered viable for use in ACO measurement. NCQA urges ACOs electronically accessing data from EMR or EHR databases to adhere to the following guidelines. EMR and EHR information must distinguish between ordered and completed appointments, procedures, and lab and radiology orders. Only completed events count toward measures. Some EMRs and EHRs capture and record CPT codes when a practitioner completes a service order on the Order screen. A CPT code found as a stand-alone on the ordered list does not comply with numerator criteria for quality of care measurement. All data elements specified in a measure (e.g., date of service, procedure codes, diagnosis codes, prescriptions) must be identified as rendered by the time specified in the measure (e.g., a mammogram during the measurement year or the year prior to the measurement year for the Breast Cancer Screening measure). The ACO may count a service for a measure if the EMR or EHR contains the following information: Claim data captured through a patient management system linked to the EMR, or A note indicating the date on which the service was rendered, and the result (when applicable). Be cautious when using an EMR or EHR database to identify the denominator for a measure for which a population-based sample is desired, unless the ACO is confident that the EMR database contains all patients of the aggregate ACO population, rather than a subset of the eligible population. EMR and EHR databases may generally be more reliable for identifying numerator-compliant events, such as the delivery of a specific service.

Note
An EMR or EHR that entails manual abstraction of data elements from the electronic record for reporting purposes is not considered electronic data. ACO Measure Specifications provides the option of manually abstracting data from electronic or paper medical records for the numerator of all measures included in this publication.

29. Measures That Use Pharmacy Data


Several measures for NCQA ACO Accreditation rely on pharmacy data, either in the numerator or in the denominator, or in both. NCQAs HEDIS Health Plan Specifications requires the use of dispensed medications (fills) to optimize the accuracy of estimating care received by health plan members. In NCQAs HEDIS 2012 Technical Specifications for Physician Measurement, records of prescriptions count the same as records of dispensed medications for numerators and some denominators, because of providers g eneral inability to access payers paid pharmacy claims for their patient. An ACO that does not have access to the requisite pharmacy data can opt not to pursue those measures, but NCQA expects ACOs to have access to information about dispensed medications through paid pharmacy claims from their ACO payer and vendor partners (including through CMS MSSP). It is critical that payers exchange these data with their ACO partners, as CMS does through MSSP, to give ACOs full knowledge of and help them effectively manage their patients medications; medications actually received vs. those prescribed by participant providers; and medications prescribed by providers outside of the ACO.

______________
Current Procedural Terminology 2012 American Medical Association. All rights reserved.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications Given the current state of the health care industry, for ACO Measure Specifications, either prescriptions or dispensed medications can be used for both denominators and numerators of measures that require pharmacy data. This may evolve over time as ACOs gain greater access to data on dispensed medications from payer partners or pharmacy vendors.

21

For measures that rely on medications, measure specifications include medication tables that must be referenced. When using dispensed medications, tables must be used in conjunction with the National Drug Code (NDC) lists posted to NCQAs Web site. The tables include a Description column that indicates the therapeutic category, and a Medication column that includes all appropriate medications in their generic form. ACOs must use the NDC lists for each pharmacy-dependent measure. Final NDC lists for pharmacy-related measures are posted to the NCQA Web site annually in mid-November.

30. Identifying Events/Diagnoses Using Laboratory Data


Laboratory data may not be used to identify an event, disease or condition (e.g., acute myocardial infarction [AMI], diabetes) unless listed in a code table that contains LOINC codes. Many ACOs find a high rate of false positives when they use laboratory data to identify patients with a disease or condition. Diagnosis codes are frequently reported on laboratory tests in cases where the condition is being ruled out; therefore, ACOs should not use laboratory data or claims when identifying the following criteria. Eligible Population Event/Diagnosis. Negative Diagnosis History. Negative Competing Diagnosis. Negative Comorbid Condition History. Exclusions. Numerator events. Laboratory claims and data may be used for any code tables that contain LOINC codes.

The Hybrid Method 31. The Hybrid Method


The Hybrid Method allows ACOs to identify the numerator through electronic and medical record data. The denominator consists of a systematic (population based) sample of patients drawn from the measures eligible population, identified through electronic data. The rate reported using the Hybrid Method is based on patients in the sample who are found through either electronic or manually abstracted medical record data to have received the service identified in the numerator. There are two approaches to conducting the Hybrid Method; they differ only in the timing of identification of patients in the denominator who have a valid exclusion. 1. The ACO first pulls a sample and then searches for valid exclusions. 2. The ACO searches for valid exclusions on the entire eligible population before it pulls a sample. The method used depends largely on data availability.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

22

General Guidelines for ACO Measure Specifications

Note
The Hybrid Method can be used for all measures in this publication. If the eligible population is smaller than the required sample size for a measure, the ACO may apply the Hybrid Method without drawing a systematic sample from the population.

32. Cautions for Using Manually Abstracted Medical Record Data


The medical record is the collective accumulation of notes kept by all practitioners who treat a patient. It constitutes the legal record of patient visits and treatments and includes all test results (e.g., reports from the laboratory, pathology, radiology). The complete record may be located in more than one place (e.g., in the doctors office, in specialists offices). Even medical records within EMRs and EHRs can require manual abstraction when there is no capacity to manage or report on those data electronically. While manually collected medical record data offer the potential for richer clinical detail than paid claims data from payers, medical record data abstraction is very difficult to perform accurately and consistently. It requires careful attention to data collection protocols; the development of data collection forms and databases; staff training and technical support; and audits to ensure accuracy and interrater reliability. For paper medical records, EMRs and EHRs, there are issues related to data accuracy and burden of data collection. When evaluating data sources to use for performance measurement, ACOs must consider: 1. Accuracy problems with medical record data, including incomplete documentation; missing information about care provided by other physicians; inter-provider variability in documentation; lack of appreciation for the complexity of the EMR or EHRs data model and where data are captured in the application by various providers using the system; and illegibility in paper medical records. 2. Medical records (including many EMRs and EHRs) lack key measurement information (e.g., records of when, and whether, prescriptions are filled and laboratory events) more readily available through paid claims from payers. 3. The burden of abstracting data from paper records, EMRs or EHRs (when the EMR or EHR does not provide access to a database for use as electronic data) especially when a patients records exist in multiple locations or on multiple systems. Not only can manual abstraction from multiple medical records be costly and inefficient, but it can also be perceived as intrusive by office staff who manage the records. These data collection processes are not an intrinsic part of the care delivery process; they are an addition and should not be considered long-term solutions to accessing clinical data across the ACO. 4. Measure requirements, such as negative medication or diagnosis histories, duration of medications or complexity of methods and data sources used to identify the denominator or numerator.

33. Date Specificity in Medical Records


ACO Measure Specifications require a date to be specific enough to determine that an event occurred during the time established in the measure. For example, in the Childhood Immunization Status measure, patients should receive three hepatitis B vaccines, with different dates of service on or before the second birthday. Example: Assume a child was born on February 5, 2009. Documentation that the first hepatitis B vaccine was given at birth is specific enough to determine that it was given prior to the deadline for this measure (i.e., the childs second birthday). However, if the documentation states that the third hepatitis B vaccine was given in February 2011, the ACO cannot count the immunization because the date is not specific enough to confirm that it occurred prior to the childs second birthday. Sometimes, documentation of the year alone is adequate; for example, for most required exclusions and measures that look for events in the measurement year or in the year prior to the measurement year. Terms such as recent, most recent or at a prior visit are not acceptable. Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications

23

For documented history of an event (e.g., documented history of a disease), undated documentation may be used if it is specific enough to determine that the event occurred during the time frame specified in the measure. For example, for the Breast Cancer Screening measure, undated documentation on a problem list stating bilateral mastectomy in 1999 is specific enough to determine that this exclusion occurred prior to December 31 of the measurement year.

34. Obtaining Medical Record Information Manually


ACOs (and their analytical contractors) that use the Hybrid Method are responsible for determining compliance with NCQAs ACO Measure Specifications. Information may be manually abstracted from paper medical records or from automated systems such as EMRs or EHRs, patient registries, billing systems or payer claims systems. Abstraction may be performed by one of the following groups: The ACO or any of its participant providers. Contractors hired to conduct manual record abstraction. Processes used to determine the validity and integrity of abstracted data, including interrater reliability, quality control and rater-to-standard tests, should be subject to review during the ACOs audit review. ACO or contracted abstractor Count a service if the medical record or EMR contains the date of the service and evidence that the service occurred. All services must be rendered by the deadline established in the measure (e.g., by the childs second birthday, for the Childhood Immunization Status measure). Because NCQA requires that the systematic sample be stable and reproducible, the ACO may not change the sample after it is created. If an ACO refreshes the data repository used to compile ACO measures after the sample is drawn and manual medical record abstraction is in progress, it should follow the guidelines below to use the newer data. Compare only numerator-negative patients in the sample with electronic data such as payer claims or billing data. The ACO may use an automated process that loads the entire sample and compares it to the refreshed data. All data must be used consistently in the samples. If recent data contradict numerator compliance, recent data must be used. If recent data exclude a patient, recent data must be used and the oversample must provide the substitute patient. If the oversample is exhausted, the ACO must use the Sampling Guidelines to ensure a denominator of 411. The ACOs internal audit process should include a review and approval of the abstraction timing, processes and results. EMR abstraction Whether manually abstracting data from paper medical records or electronic medical record systems, ACOs must be able to distinguish between ordered and completed visits, procedures, lab tests and radiology tests. Only completed events count toward compliance. Some EMRs or EHRs record CPT codes when the practitioner enters a service order in the order screen. A CPT code found on the order list alone d oes not comply with the numerator criteria.

Sample frame integrity

Manual updates Automated updates

____________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

24

General Guidelines for ACO Measure Specifications Measure-specific results collected by practitioners and provided in an abstraction tool are acceptable. ACOs must provide guidelines for practitioner abstraction. Verbal or written guidance should include clear instructions for applying the technical specifications to medical record review. Abstraction tools must have all necessary data elements and must be subject to quality control processes, such as interrater reliability or rater-to-standard reliability validation. Note: Do not use practitioner attestation forms; they do not require practitioners to verify services using the medical record.

Guidelines for practitioner abstraction

35. Using the Hybrid Method Approach 1


Step 1 Identify the eligible population through electronic data, according to the measure specification, and draw a systematic sample. Refer to Drawing a Systematic Sample for the Hybrid Method, below. If the eligible population is smaller than the required sample size for a measure, the ACO may apply the Hybrid Method without drawing a sample. Search the electronic data for numerator events for the sampled patients. Starting with the primary list, follow the Electronic Method measure specification to search electronic systems for numerator events. Step 3 Search the electronic data for an exclusion. For patients for whom electronic data does not show a positive numerator event, search electronic data for an exclusion to the service/procedure being measured, if applicable. Step 4 Step 5 From the oversampled population, substitute for the excluded records. Refer to Substituting Medical Records, below. Search the medical records for numerator events for the sampled population. Review the medical records of patients in the sample for whom a numerator event or exclusion to the service/procedure being measured was not identified using electronic data. Refer to The Electronic Method, above, and Using The Hybrid Method to Collect Measures With Multiple Numerator Events, below, for guidance. Step 6 Search the record for an exclusion to the service/procedure being measured. For patients for whom the medical record does not show a positive numerator event, search the record for an exclusion to the service/procedure being measured, if applicable. Step 7 Remove patients with medical record identified exclusions. Remove from the denominator patients for whom the medical record identified an exclusion. Step 8 Step 9 Step 10 From the oversampled population, substitute for the excluded records. Refer to Substituting Medical Records, below. Use either electronic or medical record data to identify the numerator event for additional patients used as replacements. Determine the numerator. Include only patients in the sample identified through either electronic data or medical record review as having had received or experienced the numerator event.

Step 2

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications Step 11 Calculate the rate.

25

The rate may be calculated using the denominator (MRSS with substitutions) or the denominator plus the entire oversample (FSS), but the method should be consistent across measures. Refer to Substituting Medical Records.

Approach 2
Step 1 Step 2 Identify the eligible population through electronic data, according to the measure specification. Search the electronic data for numerator events for the entire eligible population. Follow the electronic specifications to search electronic data for numerator events for the eligible patient population identified in step 1. Step 3 Search for electronically identified exclusions. For patients for whom electronic data does not show a positive numerator event, search electronic data for an exclusion to the service/procedure being measured, if applicable. Step 4 Exclude patients with electronically identified exclusions. Remove from the eligible population patients from step 3 for whom electronic data identified an exclusion to the service/procedure being measured. Step 5 Step 6 Determine the FSS and draw a systematic sample from the eligible population. Search the electronic data for numerator events for the sampled patients. From the patients identified as compliant for numerator events from step 2, pull the patients in the FSS. Step 7 Search the medical records for numerator events for the sampled population. Review the medical records of patients in the sample for whom a numerator event or exclusion to the service/procedure being measured was not identified using electronic data. Refer to Using The Hybrid Method to Collect Measures With Multiple Numerator Events. Step 8 Search for medical record exclusions. For patients for whom electronic data or the medical record does not show a positive numerator event, search the medical record for an exclusion to the service/procedure being measured, if applicable. Step 9 Exclude patients with medical record identified exclusions. Remove from the denominator patients for whom the medical record identified an exclusion. Step 10 Step 11 Step 12 From the oversampled population, substitute for the excluded records. Refer to Substituting Medical Records, below. Use either electronic or medical record data to identify the numerator event for additional patients used as replacements. Determine the numerator. Include only patients in the sample who were identified through either electronic data or medical record review as having received or experienced the numerator event.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

26 Step 13

General Guidelines for ACO Measure Specifications Calculate the rate. The rate may be calculated using the denominator (MRSS with substitutions) or the denominator plus the entire oversample (FSS), but the method should be consistent across measures. Refer to Substituting Medical Records.

36. Using the Hybrid Method to Collect Measures With Multiple Numerator Events
The Childhood Immunization Status measure may be collected using the Hybrid Method, and requires more than one event to satisfy the numerator. Three separate vaccinations are collected to report the IPV rate for Childhood Immunization Status. For this measure, ACOs may use a combination of electronic data and medical record data to satisfy the numerator for patients in the denominator, if the events across medical record data and electronic data are at least 14 days apart. ACOs may count two IPV vaccines identified through electronic data and the third IPV identified through medical record review for a patient in the denominator, if the medical record date of service is not within 14 days of either electronic date of service. ACOs may also count three IPVs identified through electronic data and four DTaPs identified through medical record data for the same patient. The ACO must use only the medical record event if it has one event from the medical record and one from electronic data and cannot determine if the dates are at least 14 days apart.

Sampling Methods 37. Population-Based Sampling for the Hybrid Method


Drawing the sample Like the Electronic Method, the eligible population for measures collected via the Hybrid Method is based on either: Enrollment or demographic data (e.g., women between 24 and 64 years of age for Cervical Cancer Screening), or Claims or service data (e.g., patient who had a discharge with a cardiac condition for Cholesterol Management for Patients With Cardiovascular Conditions). Demographic(enrollment) dependent denominators Claim-(service) dependent denominators For measures in which the eligible population is determined through demographic or enrollment data, do not draw the sample prior to one month before the end of measurement year. If a sample is drawn in the month before the measurement year, oversample, if possible, to replace patients included in the sample who were found to be noncompliant with the denominator criteria after the end of the measurement year. For measures in which the eligible population is determined through claims data drawn from a complete eligible population, select the sample no earlier than January of the year after the measurement year. Allow claims incurred through the end of the measurement year to be captured in administrative systems (which may include electronic clinical billing systems) before identifying the eligible population and drawing the sample. This may require a lag time of up to three months, in some situations. Use of the Hybrid Method requires the sample to be drawn from the identified complete eligible population.

Determining the required sample size

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications Population definition In some cases, the size of the eligible population for a measure may be smaller than the recommended sample size. For example, there may be very few attributable patients for a measure such as Follow-Up After Hospitalization for Mental Illness. In this case, use the entire eligible population and report the data with a 95 percent confidence interval.

27

Statistical assumptions for sample size

Consistent with HEDIS Health Plan Measurement, a sample size of 411 is required when the Hybrid Method is used for direct comparison between health plans or, in this case, between ACOs on a single measure. The sample size was calculated assuming a two-tailed test of significance between two proportions ( = .05, 80 percent power, two-tailed test of significance). A normal approximation to the binomial with a continuity correction was employed in the sample size calculation. The worst-case 50 percent expected value was assumed. The detectable difference for most measures applied to health plans is 10 percentage points. This was chosen by NCQA because the difference is large enough to be actionable using a single measure and directly comparing one health plan or ACO to another, it is not unduly burdensome for data collection and it is not so small that it will be swamped by nonsampling error.

38. Drawing a Systematic Sample for the Hybrid Method


NCQA recommends a systematic sampling scheme. Although the directions appear complex, they create a standardized framework for drawing a sample that represents the aggregate population. For each measure where the Hybrid Method is used, complete each of the following steps to draw a systematic sample. Step 1 Determine the eligible population (EP). Develop a list of eligible patients, including full name (last, first), date of birth and event (if applicable). Step 2 Step 3 Determine the MRSS. If the EP MRSS, proceed to step 4. Determine the FSS. The FSS includes the MRSS (from step 2), plus an adequate number of additional records to make substitutions. The user ACO should attempt to oversample only enough to ensure that the MRSS is met. Keep incomplete records and other exclusion criteria in mind when making this decision. Common oversampling rates are in the 5 percent20 percent range. The FSS is calculated by the following formula. FSS = MRSS + (MRSS oversampling rate) (round up to the next whole number), where MRSS equals the minimum required sample size (step 2). Step 4 If EP >FSS, go to step 5. If EP MRSS, all eligible patients are included in the sample. If MRSS <EP FSS, proceed to step 8. Step 5 Step 6 Sort the list of eligible patients in alphabetical order by the last name, first name, date of birth and event (if applicable). Determine N, which is used in the formula to determine the patient who will start your sample. N is calculated using the equation: N = EP/FSS (round down to a whole number), where EP equals the eligible patient population (step 1) and FSS equals the final sample size (step 3).

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

28

General Guidelines for ACO Measure Specifications Step 7 Before choosing patients, determine the starting patient (START). It is important that the sample be selected from a single pass through the patient list. START can have many values and still allow only one pass. Determine a Random Number (RAND) value between 0 and 1 for each measure where the Hybrid Method is applied. The random number for each respective measure should be used to calculate the starting point from which to draw the final sample. Many analytical and data management software packages can generate a random number between 0 and 1. For example, in Microsoft Excel, the formula = RAND( ) typed into a cell and entered (press the Enter key) will replace the formula with a random number between 0 and 1. Calculate the number from which to start drawing the final sample as follows. START = (RAND N) (round per the .5 rule to the nearest whole number greater than 0), where RAND equals the random number for each respective measure identified. Step 8 Select the sample, choosing every i patient using the formula: i patient = START + [(i-1) x (EM/FSS)] (rounding [(i-1) x (EM/FSS)] per the .5 rule to the nearest whole number greater than 0). For i = 2,3,4, , FSS where EM equals the eligible patient population (step 1). FSS equals the final sample size (step 3). Starting with the patient corresponding to the number START, choose every i patient until the MRSS is met. This becomes the primary list of sampled patients. Continue choosing every i patient until the FSS is met. This set of patients becomes the auxiliary list of sampled patients (i.e., the oversample). The ACO can stop once the FSS is achieved, or use all patients in the primary and auxiliary list.
th th th th

Note: From step 4, if MRSS <EP FSS, sort the eligible patients in alphabetical order by the last name, first name and date of birth. Choose the first MRSS eligible patients as the primary sample and the remaining eligible patients as the auxiliary sample.

Example 1
The eligible population for Cervical Cancer Screening is 500. Based on experience, estimate a 5 percent oversample rate. Following the systematic sampling scheme: Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 EP = 500. MRSS = 55. FSS = 55 + (55 .05) = 57.75 (the next whole number above 57.75 is 58, so FSS = 58). Since 500 >58, go to step 5. Sort the list alphabetically: last name, first name, date of birth. N = 500/58 = 8.62 (the next whole number below 8.62 is 8, so N = 8).

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications Step 7 For this example, assume that RAND = 0.66, so START = 0.66 x 8 = 5.28. Rounding using the .5 rule, START = 5. The 5th sorted patient is chosen first. The 2nd patient chosen is the 5 + [(2-1) x 8] = 5 + 8 = 13th sorted patient. The 3rd patient chosen is the 5 + [(3-1) x 8] = 5 + 16 = 21st sorted patient. The 55th patient (the last one in the primary list) is the 5 + [(55-1) x 8] = 5 + 432 = 437th sorted patient. The last patient in the auxiliary sample is the 18 + [(58-1) x 8] = 5 + 456 = 461st sorted patient.

29

Example 2
The eligible population for Cholesterol Management for Patients With Cardiovascular Conditions is 100. Following the systematic sampling scheme: Step 1 Step 2 Step 3 Step 4 EP = 100. MRSS = 150. Since 100 <150, skip to step 4. Skip this step. Include all 150 patients on your primary list.

Example 3
The eligible population for Childhood Immunization Status is 175. Following the systematic sampling scheme: Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 EP = 175. MRSS = 155. FSS = 155 + (155 .10) = 170.5 (the next whole number above is 170.5, so FSS = 171). Since 155 <171, skip to step 8. Skip this step. Sort the list and choose the first 155 as the primary list. The remaining 20 patients become the auxiliary list.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

30

General Guidelines for ACO Measure Specifications

39. Calculating a Confidence Interval at the 95 Percent Level


The formula for calculating the 95 percent confidence interval around an ACO measure Specifications rate is as follows. p (1 p ) 1 lower = p 1.96 n 2n
upper = p 1.96 p (1 p ) n

2n

Where p = the ACOs rate and n = the sample size. For example, suppose the ACO has a sample size of 96 eligible women for the Cervical Cancer Screening rate. Of these, 50 received a Pap test during the year. The calculation would proceed as follows.
p = 50 96 = 52%

lower = .52 1.96

.52 (1 .52)
96

1 192

= 41.5%

upper = .52 1.96

.52 (1 .52) 1 = 62.5%


96 192

Thus, the user can be 95 percent certain that the physicians true Pap test rate is between 41.5 percent and 62.5 percent.

40. Complex Probability Sampling


The ACOs responsibility Properly applied, techniques such as stratified sampling, cluster sampling and other complex probability approaches can improve precision and increase sampling efficiency. Consult a methodology expert before implementing a complex sampling method.

41. Substituting Medical Records


Unless otherwise noted in the specifications for a particular measure, do not exclude patients from the sample or make substitutions, except under the three circumstances described below. Acceptable circumstances for substitution: 1. Errors in sampling data An entity collecting data from medical records should not substitute patients who are noncompliant because they refused the service or because the medical record cannot be located. For quality control, use one of the three options below to document why a sample patient record is substituted for another. Medical record abstraction reveals that the patient was included in error and does not meet the eligibility criteria for inclusion in the sample. Data errors can be caused by incorrect patient or clinical information; for example, if a patient in the sample of Childhood Immunization Status is discovered to be 22 years old, not 2 years old.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications The medical record must have evidence that a patient does not meet the criteria for the measure; a medical record without any notation either substantiating or refuting the diagnosis is not evidence that the patient does not have the condition being measured. 2. Exclusion to treatment being measured 3. Selecting an employee/ dependent for the sample A patient has a valid exclusion to the treatment being measured; for example, a patient with a diagnosis of colorectal cancer or total colectomy is a valid exclusion in the denominator for the Colorectal Cancer Screening measure.

31

An employee of the ACO or the vendor, or the employees dependent, was selected for the sample, and the employee or dependents medical record must be reviewed to determine compliance with the measure. The ACO or vendor may exclude employees and their dependents in this situation only.

Coding Conventions
Note: This section does not pertain to the codes used in Code Set for the Attribution Model.

42. Coding Systems Included in ACO Measure Specifications


ACO Measure Specifications includes codes from the following coding systems. CMS Place of Service (POS). Current Procedural Terminology (CPT). Medicare Severity Diagnosis-Related Group (MS-DRG). Healthcare Common Procedure Coding System (HCPCS) Level II. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). Logical Observation Identifiers Names and Codes (LOINC). Uniform Bill (UB) revenue and type of bill (TOB).

43. Code Presentation


Unless otherwise noted, codes in this publication are stated to the minimum specificity required. If a code is presented to the third digit, any valid fourth or fifth digit may be used for reporting in accordance with ACO Measure Specifications. For example, when a table lists ICD-9-CM Diagnosis code 401, the codes 401.0, 401.1 and 401.9 are used for ACO measure reporting. When necessary, a code may be specified with an x, which represents a valid digit that can be used for reporting. For example, ICD-9-CM Diagnosis code 640.x1 indicates that any valid fourth digit can be used if the fifth digit is 1. When reading ACO Measure Specifications coding tables, assume there is an or in between each column; each code set is not dependent on another code set unless otherwise noted.

____________
Current Procedural Terminology 2012 American Medical Association. All rights reserved.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

32

General Guidelines for ACO Measure Specifications

44. Principal vs. Secondary Diagnoses


Principal and secondary diagnoses are mentioned throughout ACO Measure Specifications. A principal diagnosis or primary diagnosis is the diagnosis given at discharge and the one listed in the first position on a claim form. A secondary diagnosis is a diagnosis listed on a claim or encounter form that is not classified as the principal diagnosis. A claim form may contain several secondary diagnoses. The ACO should follow the specifications stated within each particular measure to determine whether a diagnosis must be principal or may be secondary. Some measures require a specific principal diagnosis for eligibility; others allow any diagnosis (principal or secondary). For example, the Persistence of Beta-Blocker Treatment After a Heart Attack measure specifies that any diagnosis of an initial AMI is eligible. If a patients claim lists the principal diagnosis as severe head injury trauma, but an initial AMI is listed as a second, third, fourth or fifth diagnosis on the same claim form, the patient would still be included in the measure. (Note: this measure is used for illustrative purposes only as this is not a measure currently included in this publication) On a UB-04 claim form, the principal diagnosis is listed in Form Locator 67, Principal Diagnosis Code, and secondary diagnoses are listed in Form Locators 67AQ, Other Diagnosis Codes. Data in Form Locators 69, Admitting Diagnosis Code, and 70ac, Patients Reason for Visit, should not be included in reporting in accordance with ACO Measure Specifications. On a CMS1500 claim form, the principal diagnosis is listed in Item Number 21, line 1, and secondary diagnoses are listed in Item Number 21, lines 24.

45. CPT Code Modifiers


CPT modifiers are two- or five-digit extensions that, when added to CPT codes, provide additional information about a service or procedure. The same procedure should never be counted twice for the same date of service. The ACO should follow the guidelines below when procedure codes in its databases have modifiers (xxxxx denotes the five-digit CPT code). xxxxx-26 indicates the professional component of a service (xxxxx-TC is used by some ACOs to indicate the technical component of the same service). For a given procedure, the ACO should count one or the other of these codes, but not both. xxxxx-54 denotes surgical care only; xxxxx-55 denotes postoperative management only; xxxxx-56 denotes preoperative management only. For a given procedure, the ACO should count only one of these codes. xxxxx-80 and xxxxx-82 indicate charges for surgical assistant services; xxxxx-81 indicates a charge for minimum surgical assistant services. If the primary surgeon does not submit a claim for a given procedure, the user ACO should count only one of these codes. If a primary surgeon submits a claim, the ACO should not count any of these codes. In general, a CPT code specified in ACO Measure Specifications that appears in the ACOs database with any modifier other than those specified above may be counted in the measure.

46. Uniform Bill Code Specificity


Uniform Bill (UB) codes, primarily Type of Bill and Revenue codes, are used to identify services. As with the ICD-9-CM codes, an x may be used in place of a digit. ACO Measure Specifications specifies UB Type of Bill Codes using three digits. The ACO may also use the equivalent four-digit version of the code (the three-digit code plus a leading zero); for example, to identify skilled nursing facility (SNF) encounters, it may use 21x or 021x.
____________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

General Guidelines for ACO Measure Specifications

33

47. Mapping Proprietary or Other Codes


An ACO that does not use the coding systems specified (e.g., CPT or ICD-9-CM) must map the codes it uses to the codes specified in ACO Measure Specifications. The ACO may map proprietary codes, Level III and state-specific Level II HCPCS codes and NDC codes; it may not map standard codes or deleted codes to the codes used in the measures. When mapping codes, it is important to match the clinical specificity required for ACO Measure Specifications. NDC code mapping should be linked to the generic name, strength/dose and route indicated in the HEDIS NDC lists posted on the NCQA Web site (www.ncqa.org).

48. Retired Codes


NCQA annually tracks billing, diagnostic and procedure codes that are designated obsolete. NCQA does not remove codes in the year in which they receive the designation of obsolete because of the look-back period in many ACO Measure Specifications measures. Obsolete codes are deleted from the specifications one year after the look-back period is exhausted. For example, since the Breast Cancer Screening measure counts a mammogram in the measurement year or the year prior to the measurement year, mammography codes have a two-year look-back period. A mammogram code that is designated obsolete effective January 1, 2010, is deleted from the specifications in the 2013 specification publication after the two-year look-back period (2011, 2012) plus one additional year (2010) is exhausted. NCQA uses the NDC system. Obsolete NDC codes are phased out of the specifications three years after the look-back period to allow pharmacies and organizations to use their inventory and change their systems. NCQA encourages ACOs to update their information systems and to ensure that complete, accurate and consistent coding is used for all encounters and claims so that ACO Measure Specifications can be followed. This will help the industry move toward a uniform system of performance measurement.

49. Table Names


Measure specifications contain two types of tables: one to present specification requirements and one used by ACOs to submit data. Tables use a standardized naming system. For ACO Measure Specifications, table names begin with the letter A followed by the measures three-character abbreviation; for example, Childhood Immunization Status tables begin with ACIS. Specification tables Tables that are part of the specifications (i.e., coding and pharmaceutical tables) begin with the measure abbreviation and end with a hyphen (-) and a capital letter to distinguish their order in the measures specifications. For example, the first table in the Childhood Immunization Status measure is ACIS-A. Data element tables begin with the measure abbreviation. Each product line is assigned a number; for example: ACIS-ACO1 (aggregate ACO population). ACIS-ACO2a (Medicaid, if applicable). ACIS-ACO2b (balance of aggregate ACO population if Medicaid is stratified).

Reporting tables

____________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

34

General Guidelines for ACO Measure Specifications

References
Bynum, J.P, E. Bernal-Delgado, D. Gottlieb, E. Fisher. 2007. Assigning ambulatory patients and their physicians to hospitals: a method for obtaining population-based provider performance measurements. Health Services Research. 42:1, Part I: 4562. Deming, W.E. On the interpretation of censuses as samples. 1941. Journal of the American Statistical Association. 36:459. Fleiss, L. Statistical Methods for Rates and Proportions. 1981. 2nd Ed. (New York: John Wiley & Sons, Inc.): 3842. Fleiss, J.L., B. Levin, M.C. Paik. 2003. Statistical Methods for Rates & Proportions. Wiley-Interscience: Indianapolis, IN. Greenfield, S., S.H. Kaplan, R. Kahn, J. Ninomiya, J.L. Griffith. January 15, 2002. Profiling care provided by different groups of physicians: effects of patient case-mix (bias) and physician-level clustering on quality assessment results. Ann Intern Med. 136(2):11121. Kaplan, S.H., J.L. Griffith, L.L. Price, L.G. Pawlson, S. Greenfield. Assessing Physician Performance: addressing the science and recognizing the future. Unpublished. Krein, S.L., T.P. Hofer, E.A. Kerr, R.A. Hayward. 2002. Whom Should We Profile? Examining Diabetes Care Practice Variation among Primary Care Providers, Provider Groups, and Health Care Facilities. Health Services Research. 37(5):115980. Safran D.G., M. Karp, K. Coltin, H. Li A. Chang, J. Ogren, W.H. Rogers. 2006. Measuring patients' experiences with individual primary care physicians. Results of a statewide demonstration project. Journal of General Internal Medicine. 21(1):1321. Scholle, S.H., et al. NCQA PB report. Unpublished. Tucker, J.L., 3rd. 2000. The theory and methodology of provider profiling. Int J Health Care Qual Assur Inc Leadersh Health Serv. 13(6-7):31621.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Appendix 1Modifications From HEDIS Health Plan Measurement

1-1

APPENDIX 1
MODIFICATIONS FROM HEDIS HEALTH PLAN MEASUREMENT
Key Specification HEDIS Health Plan Measurement Defining the Population for ACO Measurement The Organization Evaluated Managed care organizations (Commercial HMO, POS, PPO, Medicare Advantage, Medicaid) that administer health insurance benefits on behalf of their private and public customers. Individuals in the Population Members of the health plan. ACO Contracts and Measurement Commercial, Medicare and Medicaid populations defined by insurance licenses held by the managed care organizations. ACO Measure Specifications ACOs that provide health care services to defined populations through a specified legal entity that can contract with payers. Patients of the ACO. Methods: 1. If an ACO has at least one contract in force, it has two options: 1A. Report on the aggregate ACO contract population as defined through contractual attribution models or assignment methods, or 1B. Report on the aggregate ACO contract population plus all or at least 80 percent of its patient population, as defined by NCQAs recommended attribution model. 2. If the ACO has no ACO contract in force, it reports on all or at least 80 percent of its patient population, using NCQAs recommended attribution model. For method 1B and 2, NCQA recommends the CMS MSSP method of attributing patients to ACO provider participants based on a two-step process of primary care first and then other providers of care, with two modifications: Use a 24-month period (vs. CMS 12-month period). Use a plurality of primary care visits (vs. CMS plurality of allowed charges for primary care visits). Minimum population of 5,000 as of January 1 of the measurement year, regardless of whether the population is stratified for measurement (see below). Continuous Enrollment is not provided in ACO measure specifications.

Attribution of Patients to the Organization

Enrollment in a health plan.

Minimum Population Continuous Enrollment Measure Reporting to NCQA Population Stratification

Suggested minimum is 15,000 per product reported. Defined for each measure to ensure adequate opportunity for health plans to intervene. Separate commercial (HMO, POS, PPO), Medicaid and Medicare reports.

The ACO aggregate population is reported as a whole; with an option to report Medicaid separately for measures for which HEDIS Health Plan Measurement offers Medicaid specifications. Refer to Appendix 2.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

1-2

Appendix 1Modifications From HEDIS Health Plan Measurement ACO Measure Specifications The ACO retains responsibility for reporting ACO measures according to ACO measure specifications. Same. Similar to HEDIS Health Plan Measurement (June) with an ACO defined set of data elements. Phased in audit requirement. No audit required before the June 2014 submission on the 2013 measurement year. Electronic and Hybrid. Allows manually abstracted medical record data for all numerators. Same. To start, ACO measures will require a base minimum sample of 411. Once ACOs establish historical results using standard measure specifications, NCQA will explore sample size adjustments based on the reported rate. Same. Require measure exclusions similar to NCQAs HEDIS Physician Measurement. Use of pharmacy claims or scripts for both denominators and numerators to allow for the greatest flexibility for ACOs without paid pharmacy claims. Defined in alignment with HEDIS Physician Measurement: Data routinely captured for payment or care delivery including payer claims (medical and pharmacy), provider claims (billing systems); patient registration, lab results, and EMR databases used by practices or health systems. Same with respect to what is defined for select measures, as well as the option to stratify for Medicaid populations with applicable measures. However, no further population stratification by type of insurance will be allowed. Same.

Key Specification HEDIS Health Plan Measurement Measure Reporting to NCQA Reporting to NCQA Reporting can be done by the Health Plan, the Certified HEDIS Vendor or contracted medical record abstractors, in accordance with HEDIS Specifications. Measurement Year Generally the calendar year unless otherwise specified. Report Submission HEDIS submission required annually (June 15 of the reporting year, on the prior calendar measurement year) by a defined set of requirements. Measurement Audit HEDIS Compliance Audit is required with each annual measure submission to NCQA. Methods and Definitions for Compiling Measures Methods Offered Administrative and Hybrid. Hybrid Method Allows manually abstracted medical record data for select numerators. Systematic sampling. Sampling for Hybrid Method Sample Sizes for Hybrid There is a base minimum sample size of 411, but specific sample size Method parameters are based on previous years result. Medical Records for Denominator Exclusions Pharmacy Data Electronic Data Allows manually abstracted medical record data to confirm the denominator for Controlling High Blood Pressure. Specified for each measure; many measures have optional exclusions and some are required. Use of pharmacy claims only, where dispensed medications (or fills) are used for denominator and numerators. Not a defined term in HEDIS, similar to administrative data (payer enrollment and claims); with specific supplemental data allowed per the guidelines.

Methods and Definitions for Compiling Measures Risk Adjustment Product stratification and what is defined within select measures. Coding Tables As specified for each measure and maintained annually.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Appendix 2ACO Standards and Guidelines Excerpt

2-1

APPENDIX 2
ACO STANDARDS AND GUIDELINES EXCERPT PR 1 A: Performance Reporting
The organization measures and reports clinical quality of care, patient experience and resource stewardship. Intent The organization strives to improve the quality of services by evaluating its performance using valid measures and making results available to the public and participating providers. Element A: Core Performance Measures At least annually, the organization monitors core performance measures. 100% The organization submits 40 core performance measures Reports NCQA scores this element once for the organization. 75% The organization submits 30 core performance measures 50% The organization submits 20 core performance measures 25% No scoring option 0% The organization submits 0-19 core performance measures 2.40 points

Scoring

Data source Scope of review Look-back period Explanation

NCQA reviews reports detailing performance on measures within the 12 months prior to survey. This is a must-pass element. Organizations must earn a score of 50% or higher. All four must-pass elements are required to reach Level 2 accreditation. Core performance measures Performance monitoring is a necessary component of an organizations quality improvement strategy. At least annually, the organization assesses the population it serves by collecting and monitoring core performance measures. Core performance measures are the minimum set of measures that organizations must monitor, and include indicators of clinical quality and cost. Refer to Appendix 2: Core Performance Measures for a description of measure domains. For each measure, the organization must use valid data and methodology to produce a quantitative result. To calculate core performance measures, organizations may evaluate: Patients for whom the organization has accepted accountability for the total quality, costs and overall care (i.e., the assigned or attributed ACO population), or All patients seen by participating ACO providers.

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Appendix 3Core Measures Included in ACO Measure Specifications

3-1

APPENDIX 3
CORE MEASURES INCLUDED IN ACO MEASURE SPECIFICATIONS
ACO Core Measure Prevention and Screening Adult BMI Assessment (AABA) Weight Assessment and Counseling for Nutrition and Physical Activity for Children/ Adolescents (AWCC) Childhood Immunization Status (ACIS) Immunizations for Adolescents (AIMA) Breast Cancer Screening (ABCS) Cervical Cancer Screening (ACCS) Colorectal Cancer Screening (ACOL) Chlamydia Screening in Women (ACHL) Care for Older Adults (ACOA) Respiratory Conditions Appropriate Testing for Children With Pharyngitis (ACWP) Appropriate Treatment for Children With Upper Respiratory Infection (AURI) Avoidance of Antibiotic Treatment in Adults With Acute Bronchitis (AAAB) Use of Spirometry Testing in the Assessment and Diagnosis of COPD (ASPR) Use of Appropriate Medications for People With Asthma (AASM) Cardiovascular Conditions Cholesterol Management for Patients With Cardiovascular Conditions (ACMC) Controlling High Blood Pressure (ACBP) Diabetes Comprehensive Adult Diabetes Care (ACDC) Note: This measure includes seven core ACO measures as indicators. Musculoskeletal Conditions Disease Modifying Anti-Rheumatic Drug Therapy for Rheumatoid Arthritis (AART) Osteoporosis Management in Women (AOMW) Use of Imaging Studies for Low Back Pain (ALBP) Behavioral Health Antidepressant Medication Management (AAMM) Follow-Up Care for Children Prescribed ADHD Medication (AADD) Follow-Up After Hospitalization for Mental Illness (AFUH) Medication Management Annual Monitoring for Patients on Persistent Medications (AMPM) Medication Reconciliation Post-Discharge (AMRP) Potentially Harmful Drug-Disease Interactions in the Elderly (ADDE) Use of High-Risk Medications in the Elderly (ADAE) Access/Availability of Care Initiation and Engagement of Alcohol and Other Drug Dependence Treatment (AIET) Utilization and Relative Resource Use Medicaid Option Measure Status In Development In Development In Development In Development In Development In Development Available In Development In Development In Development In Development In Development In Development Available In Development In Development Available In Development In Development Available In Development In Development Available In Development In Development In Development In Development In Development

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

3-2

Appendix 3Core Measures Included in ACO Measure Specifications Medicaid Option

ACO Core Measure All-Cause Readmissions (AACR) Relative Resource Use for People With Diabetes (ADRI) Relative Resource Use for People With Asthma (ARAS) Relative Resource Use for People With Cardiovascular Conditions (ARCA) Relative Resource Use for People With Hypertension (ARHY) Relative Resource Use for People With COPD (ARCO)

Measure Status In Development On Hold On Hold On Hold On Hold On Hold

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Technical Specifications for ACO Measures Selected Measures

Note: This publication is protected by U.S. and international copyright laws. You may reproduce this document for the sole purpose of facilitating public comment. 2012 by the National Committee for Quality Assurance 2012 by the National Committee for Quality Assurance 1100 13th Street NW, Suite 1000 Washington, DC 20005 All rights reserved. Printed in U.S.A. NCQA Customer Support: 888-275-7585 www.ncqa.org

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Modifications for ACO Measures


The ACO measure specifications are based upon both HEDIS Health Plan Measurement and HEDIS Physician Measurement. The table below displays how components of the specifications were adapted from HEDIS Health Plan Measurement to apply to the ACO environment. The table describes the modifications that apply to all of the ACO measure specifications; measure-specific modifications are described within each measure specification.

Specification Component Eligible Population

ACO Measure Specifications Removed: Product Lines. Continuous Enrollment. Allowable Gap. Anchor Date. Benefit. Added Defining the Population, which includes Attribution and Medicaid Stratification criteria. Renamed Administrative to Electronic. Exclusion criteria are required and moved to Eligible Population, when appropriate. Provided electronic and hybrid specifications to identify exclusions. Added hybrid specification to measures that only allow the Administrative Method in HEDIS Health Plan Measurement. Removed option to reduce the sample size from the hybrid specification. For measures that in HEDIS Health Plan Measurement only allow the Administrative Method, added a Hybrid column with the following data element: Numerator events by medical records. For measures in HEDIS Health Plan Measurement that have a hybrid specification, removed the following from the Data Elements table: Number of required exclusions (if applicable). Number of numerator events by electronic data in eligible population (before exclusions). Current years electronic rate (before exclusions). Minimum required sample size (MRSS) or other sample size. Oversampling rate. Final sample size (FSS). Number of numerator events by electronic data in FSS. Electronic rate on FSS. Number of original sample records excluded because of valid data errors. Number of electronic data records excluded. Number of medical records excluded. Number of employee/dependent medical records excluded. Records added from the oversample list. For measures that require pharmacy data, prescribed medications can also be used for denominators and numerators.

Specification Method Exclusions Hybrid Method Sample Sizes for Hybrid Method Data Elements

Pharmacy Data

NCQA 2012

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012

Colorectal Cancer Screening (ACOL)


MODIFICATIONS FROM HEDIS 2012 VOLUME 2 TECHNICAL SPECIFICATIONS FOR HEALTH PLANS
No measure-specific modifications.

Description
The percentage of adults 5075 years of age who had appropriate screening for colorectal cancer.

Eligible Population
Ages Defining the population Event/ diagnosis Exclusion 5175 years as of December 31 of the measurement year. Attribution. Report on the aggregate ACO population; refer to the General Guidelines. Medicaid stratification. No option to report Medicaid separately. None. Electronic Method Specification A diagnosis of colorectal cancer or total colectomy (Table ACOL-B). Look for evidence of colorectal cancer or total colectomy as far back as possible in the patients history, using electronic data or medical record review. Note: If using the Electronic Method, medical record documentation may not be used to identify exclusions. Hybrid Method Specification Apply the exclusion specifications for the Electronic Method to the electronic steps of the Hybrid Method. Exclusionary evidence in the medical record must include a note indicating a diagnosis of colorectal cancer or total colectomy, which must have occurred by December 31 of the measurement year. Note: If using the Hybrid Method, both medical records and electronic data may be used to identify exclusions.

Electronic Specification
Denominator Numerator The eligible population. One or more screenings for colorectal cancer. Appropriate screenings are defined by one of the following criteria. Fecal occult blood test (FOBT) during the measurement year. Regardless of FOBT type, guaiac (gFOBT) or immunochemical (iFOBT), assume that the required number of samples was returned. Flexible sigmoidoscopy during the measurement year or the four years prior to the measurement year.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012 Colonoscopy during the measurement year or the nine years prior to the measurement year. A patient had an appropriate screening if a submitted claim/encounter contains any code in Table ACOL-A.

Table ACOL-A: Codes to Identify Colorectal Cancer Screening


Description FOBT CPT 82270, 82274 HCPCS G0328 ICD-9-CM Procedure LOINC 2335-8, 12503-9, 12504-7, 14563-1, 14564-9, 14565-6, 27396-1, 27401-9, 27925-7, 27926-5, 29771-3, 56490-6, 56491-4, 57905-2, 58453-2

Flexible sigmoidoscopy Colonoscopy

45330-45335, 45337-45342, 45345 44388-44394, 44397, 45355, 45378-45387, 45391, 45392

G0104 G0105, G0121

45.24 45.22, 45.23, 45.25, 45.42 ,45.43

Table ACOL-B: Codes to Identify Exclusions


Description Colorectal cancer
Total colectomy

CPT

HCPCS G0213-G0215, G0231

ICD-9-CM Diagnosis 153, 154.0, 154.1, 197.5, V10.05

ICD-9-CM Procedure

44150-44153, 4415544158, 44210-44212

45.8

Hybrid Specification
Denominator Numerator A systematic sample drawn from the eligible population; refer to the General Guidelines. One or more screenings for colorectal cancer. Appropriate screenings are defined by any one of the four criteria below. FOBT during the measurement year. Flexible sigmoidoscopy during the measurement year or the four years prior to the measurement year. Colonoscopy during the measurement year or the nine years prior to the measurement year. Electronic Refer to Electronic Specification to identify positive numerator hits from the electronic data.

______________
Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public CommentObsolete After April 6, 2012 Medical Record Documentation in the medical record must include a note indicating the date when the colorectal cancer screening was performed. A result is not required if the documentation is clearly part of the medical history section of the record. If it is unclear whether the documentation is part of the medical history, then the result or finding must be present (this ensures that the screening was performed and not merely ordered). There are two types of FOBT tests: guaiac (gFOBT) and immunochemical (iFOBT). Depending on the type of FOBT test, a certain number of samples are required for numerator compliance. Follow the instructions below to determine patient compliance. If the medical record does not indicate the type of test and there is no indication how many samples were returned, assume the required number was returned. The patient meets the screening criteria for inclusion in the numerator. If the medical record does not indicate the type of test and the number of returned samples is specified, the patient would only meet the screening criteria if the number of samples specified is greater than or equal to three samples. If the number of samples is less than three, the patient does not meet the screening criteria for inclusion in the numerator. iFOBT tests may require fewer than three samples. If the medical record indicates that an iFOBT was done, the patient meets the screening criteria for inclusion in the numerator regardless of the number of returned samples. If the medical record indicates that a gFOBT was done, follow the scenarios below. If the medical record does not indicate the number of returned samples, assume the required number was returned. The patient meets the screening criteria for inclusion in the numerator. If the medical record indicates that three or more samples were returned, the patient meets the screening criteria for inclusion in the numerator. If the medical record indicates that fewer than three samples were returned, the patient does not meet the screening criteria. Do not count digital rectal exam as evidence of a colorectal screening because it is not specific or comprehensive enough to screen for colorectal cancer .

Data Elements for Reporting


Organizations that submit ACO data to NCQA must provide the following data elements.

Table ACOL-ACO 1: Data Elements for Colorectal Cancer Screening


Electronic Measurement year Data collection methodology (Electronic or Hybrid) Eligible population Denominator Numerator events by electronic data Numerator events by medical records Reported rate Lower 95% confidence interval Upper 95% confidence interval Hybrid

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Use of Appropriate Medications for People With Asthma (AASM)


MODIFICATIONS FROM HEDIS 2012 VOLUME 2 TECHNICAL SPECIFICATIONS FOR HEALTH PLANS
Added Hybrid Method Specification.

Description
The percentage of patients 564 years of age during the measurement year who were identified as having persistent asthma and were appropriately prescribed or dispensed medication during the measurement year.

Definitions
Note: While not explicitly stated here, the intent is that specifications about the use of pharmacy data refer to both prescribed and dispensed medications Oral medication dispensing event One prescription of an amount lasting 30 days or less. To calculate dispensing events for prescriptions longer than 30 days, divide the days supply by 30 and round down to convert. For example, a 100-day prescription is equal to three dispensing events (100/30 = 3.33, rounded down to 3). The organization should allocate the dispensing events to the appropriate year based on the date when the prescription is filled. Multiple prescriptions for different medications dispensed on the same day should be assessed separately. If multiple prescriptions for the same medication are dispensed on the same day, the organization should sum the days supply and divide by 30. Use the Drug ID to determine if the prescriptions are the same or different. Two prescriptions for different medications dispensed on the same day, each with a 60-day supply, equals four dispensing events (two prescriptions with two dispensing events each). Two prescriptions for different medications dispensed on the same day, each with a 15-day supply, equals two dispensing events (two prescriptions with one dispensing event each). Two prescriptions for the same medication dispensed on the same day, each with a 15-day supply, equals one dispensing event (sum the days supply for a total of 30 days). Two prescriptions for the same medication dispensed on the same day, each with a 60-day supply, equals four dispensing events (sum the days supply for a total of 120 days). Inhaler/ Injection dispensing event Inhalers and injections count as one dispensing event. For example, an inhaler with a 90-days supply is considered one dispensing event. In addition, multiple inhalers or injections of the same medication (as identified by Drug ID in the NDC list) filled on the same date of service should be counted as one dispensing event. For example, a patient may obtain two inhalers on the same day (one for home and one for work), but intend to use both during the same 30-day period. The organization should allocate the dispensing events to the appropriate year based on the date when the prescription is filled.

Multiple prescriptions dispensed on the same day

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Eligible Population
Ages 564 years by December 31 of the measurement year. Report the total rate. 511 years 1218 years 19-50 years The total is the sum of the four age stratifications. Defining the population Event/ diagnosis Step 1 Attribution. Report on the aggregate ACO population; refer to the General Guidelines. Medicaid stratification. Option to report Medicaid separately. Follow the steps below to identify the eligible population for the measure. Identify patients as having persistent asthma who met at least one of the following criteria during both the measurement year and the year prior to the measurement year. Criteria need not be the same across both years. At least one ED visit (Table AASM-B), with asthma as the principal diagnosis (Table AASM-A). At least one acute inpatient discharge (Table AASM-B), with asthma as the principal diagnosis (Table AASM-A). At least four outpatient visits (Table AASM-B), with asthma as one of the listed diagnoses (Table AASM-A) and at least two asthma medication prescribing or dispensing events (Table AASM-C). At least four asthma medication prescribing or dispensing events (i.e., an asthma medication was prescribed or dispensed on four occasions) (Table AASM-C). 51-64 years Total

Table AASM-A: Codes to Identify Asthma


Description Asthma ICD-9-CM Diagnosis 493.0, 493.1, 493.8, 493.9

Table AASM-B: Codes to Identify Visit Type


Description Outpatient Acute inpatient ED CPT 99201-99205, 99211-99215, 99217-99220, 99241-99245, 99341-99345, 99347-99350, 99382-99386, 99392-99396, 99401-99404, 99411, 99412, 99420, 99429 99221-99223, 99231-99233, 99238, 99239, 9925199255, 99291 99281-99285 UB Revenue 051x, 0520-0523, 0526-0529, 057x- 059x, 0982, 0983 010x, 0110-0114, 0119, 0120-0124, 0129, 0130-0134, 0139, 0140-0144, 0149, 01500154, 0159, 016x, 020x,021x, 072x, 0987 045x, 0981

______________
Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Table AASM-C: Asthma Medications


Description Antiasthmatic combinations Antibody inhibitor Inhaled steroid combinations Inhaled corticosteroids Dyphylline-guaifenesin Omalizumab Budesonide-formoterol Beclomethasone Budesonide Ciclesonide Montelukast Aformoterol Formoterol Cromolyn Aminophylline Dyphylline Albuterol Levalbuterol Prescriptions Guaifenesin-theophylline Potassium iodide-theophylline

Fluticasone-salmeterol Flunisolide Fluticasone CFC free Mometasone Zafirlukast Indacateroll Salmeterol Nedocromil Oxtriphylline Theophylline Metaproterenol Pirbuterol

Mometasone-formoterol Triamcinolone

Leukotriene modifiers Long-acting, inhaled beta2 agonists Mast cell stabilizers Methylxanthines Short-acting, inhaled beta2 agonists

Zileuton

Note: NCQA will post a comprehensive list of medications and NDC codes to www.ncqa.org by November 15, 2012. Step 2 A patient identified as having persistent asthma because of at least four asthma medication prescribing or dispensing events, where leukotriene modifiers were the sole asthma medication dispensed in that year, must also have at least one diagnosis of asthma, in any setting, in the same year as the leukotriene modifier (i.e., measurement year or year prior to the measurement year). Electronic Method Specification Exclude from the eligible population all patients diagnosed with emphysema, COPD, cystic fibrosis or acute respiratory failure (Table AASM-E) any time on or prior to December 31 of the measurement year. Look as far back as possible in the patients history through December 31 of the measurement year. Note: If using the Electronic Method, medical record documentation may not be used to identify exclusions. Hybrid Method Specification Apply the exclusion specifications for the Electronic Method to the electronic steps of the Hybrid Method. Exclusionary evidence in the medical record must include a note indicating a diagnosis of emphysema, COPD, cystic fibrosis or acute respiratory failure any time on or prior to December 31 of the measurement year. Note: If using the Hybrid Method, both medical records and electronic data may be used to identify exclusions.

Step 3: Exclusion

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Table AASM-E: Codes to Identify Exclusions


Description Emphysema COPD Cystic fibrosis Acute respiratory failure ICD-9-CM Diagnosis 492, 506.4, 518.1, 518.2 491.2, 493.2, 496, 506.4 277.0 518.81

Electronic Specification
Denominator Numerator The eligible population. Prescribed or dispensed at least one prescription for a preferred therapy during the measurement year (Table AASM-D).

Table AASM-D: Preferred Asthma Therapy Medications


Description Antiasthmatic combinations Antibody inhibitor Inhaled steroid combinations Inhaled corticosteroids Dyphylline-guaifenesin Omalizumab Budesonide-formoterol Beclomethasone Budesonide Ciclesonide Montelukast Cromolyn Aminophylline Dyphylline Prescriptions Guaifenesin-theophylline Fluticasone-salmeterol Flunisolide Fluticasone CFC free Mometasone Zafirlukast Nedocromil Oxtriphylline Theophylline Potassium iodide-theophylline Mometasone-formoterol Triamcinolone

Leukotriene modifiers Mast cell stabilizers Methylxanthines

Zileuton

Note: NCQA will post a comprehensive list of medications and NDC codes to www.ncqa.org by November 15, 2012.

Hybrid Specification
Denominator Numerator Electronic Medical Record A systematic sample drawn from the eligible population; refer to the General Guidelines. Prescribed or dispensed at least one prescription for a preferred therapy during the measurement year (Table AASM-D). Refer to Electronic Specification to identify positive numerator hits from the electronic data. The medical record must include a note indicating the name of the preferred therapy and the date when the prescription was written or filled during the measurement year.

Note
The HEDIS age strata for asthma measures are designed to align with both clinical practice guidelines and reporting requirements for child health quality improvement programs. Clinical guidelines specify appropriate age cohorts for measuring use of asthma medications as 5 11 years of age and 1250 years of age to account for the differences in medication regimens for children vs. for adolescents and adults. Implementation requires further stratification of the age ranges to enable creation of comparable cohorts that align with child health populations.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Data Elements for Reporting


Organizations that submit ACO data to NCQA must provide the following data elements.

Table AASM-1/2a and 2b: Data Elements for Use of Appropriate Medications for People With Asthma
Electronic Measurement year Data collection methodology (Electronic or Hybrid) Eligible population Denominator Numerator events by electronic data Numerator events by medical records Reported rate Lower 95% confidence interval Upper 95% confidence interval
For total only For total only For total only For total only

Hybrid


For total only For total only For total only For total only For total only For total only

NCQA 2012

10

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Comprehensive Adult Diabetes Care (ACDC)


MODIFICATIONS FROM HEDIS 2012 VOLUME 2 TECHNICAL SPECIFICATIONS FOR HEALTH PLANS
No measure-specific modifications.

Description
The percentage of patients 1875 years of age with type 1 or type 2 diabetes who had the following: LDL-C control (<100 mg/dL).

Eligible Population
Ages Defining the population Event/ diagnosis 1875 years as of December 31 of the measurement year. Attribution. Report on the aggregate ACO population; refer to the General Guidelines. Medicaid stratification. Option to report Medicaid separately. Diabetes. There are two ways to identify patients with diabetes: by pharmacy data and electronic data. The organization must use both methods to identify the eligible population, but a patient only needs to be identified by one method to be included in the measure. Patients may be identified as having diabetes during the measurement year or the year prior to the measurement year. Pharmacy data. Patients who were dispensed or prescribed insulin or oral hypoglycemics/antihyperglycemics on an ambulatory basis during the measurement year or the year prior to the measurement year (Table ACDC-A).

Table ACDC-A: Medications to Identify Patients With Diabetes


Description Alpha-glucosidase inhibitors Amylin analogs Antidiabetic combinations Medication Acarbose Miglitol Pramlinitide Glimepiride-pioglitazone Glimepiride-rosiglitazone Glipizide-metformin Glyburide-metformin Insulin aspart Insulin aspart-insulin aspart protamine Insulin detemir Insulin glargine Insulin glulisine Insulin inhalation Insulin isophane beef-pork Insulin isophane human Insulin isophane pork Insulin isophane-insulin regular Nateglinide Repaglinide

Metformin-pioglitazone Metformin-rosiglitazone Metformin-sitagliptin Insulin lispro Insulin lispro-insulin lispro protamine Insulin regular beef-pork Insulin regular human Insulin regular pork Insulin zinc beef-pork Insulin zinc extended human Insulin zinc human Insulin zinc pork

Insulin

Meglitinides

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012 Description Miscellaneous antidiabetic agents Sulfonylureas Thiazolidinediones Medication Sitagliptin Glyburide Tolazamide Tolbutamide

11

Exenatide Acetohexamide Chlorpropamide Pioglitazone

Liraglutide

Glimepiride Glipizide Rosiglitazone

Note: Glucophage/metformin is not included because it is used to treat conditions other than diabetes; patients with diabetes on these medications are identified through diagnosis coding only. NCQA will post a complete list of medications and NDC codes to www.ncqa.org by November 15, 2012. Electronic data. Patients who had two face-to-face encounters in an outpatient setting or nonacute inpatient setting on different dates of service, with a diagnosis of diabetes (Table ACDC-B), or one face-to-face encounter in an acute inpatient or ED setting, with a diagnosis of diabetes, during the measurement year or the year prior to the measurement year. Services that occur over both years may be counted. Refer to Table ACDC-C for codes to identify the visit type.

Table ACDC-B: Codes to Identify Diabetes


Description Diabetes ICD-9-CM Diagnosis 250, 357.2, 362.0, 366.41, 648.0

Table ACDC-C: Codes to Identify Visit Type


Description Outpatient CPT 92002, 92004, 92012, 92014, 99201-99205, 9921199215, 99217-99220, 99241-99245, 99341-99345, 99347-99350, 99384-99387, 99394-99397, 9940199404, 99411, 99412, 99420, 99429, 99455, 99456 99304-99310, 99315, 99316, 99318, 99324-99328, 99334-99337 99221-99223, 99231-99233, 99238, 99239, 9925199255, 99291 99281-99285 UB Revenue 051x, 0520-0523,0526-0529, 057x-059x, 082x-085x, 088x, 0982, 0983 0118, 0128, 0138, 0148, 0158, 019x, 0524, 0525, 055x, 066x 010x, 0110-0114, 0119, 0120-0124, 0129, 0130-0134, 0139, 0140-0144, 0149, 0150-0154, 0159, 016x, 020x,021x, 072x, 080x, 0987 045x, 0981

Nonacute inpatient Acute inpatient ED Exclusions

Electronic Method Specification


Patients with a diagnosis of polycystic ovaries (Table ACDC-D) who did not have a face-toface encounter, in any setting, with a diagnosis of diabetes (ACDC-B) during the measurement year or the year prior to the measurement year. Diagnosis may occur at any time in the patients history, but must have occurred by December 31 of the measurement year. Patients with gestational or steroid-induced diabetes (ACDC-D) who did not have a face-toface encounter, in any setting, with a diagnosis of diabetes (ACDC-B) during the measurement year or the year prior to the measurement year. Diagnosis may occur during the measurement year or the year prior to the measurement year, but must have occurred by December 31 of the measurement year. Note: If using the Electronic Method, medical record documentation may not be used to identify exclusions.

______________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

12

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012 Hybrid Method Specification Apply the exclusion specifications for the Electronic Method to the electronic steps of the Hybrid Method. Exclusionary evidence in the medical record must include a note indicating polycystic ovaries at any time in the patients history, but must have occurred by December 31 of the measurement year. The patient must not have a face-to-face encounter in any setting, with any diagnosis of diabetes, during the measurement year or the year prior to the measurement year. Exclusionary evidence in the medical record must include a note indicating a diagnosis of gestational or steroid-induced diabetes during the measurement year or the year prior to the measurement year. The patient must not have a face-to-face encounter in any setting, with a diagnosis of diabetes, during the measurement year or the year prior to the measurement year.

Note: If using the Hybrid Method, both medical records and electronic data may be used to identify exclusions. Table ACDC-D: Codes to Identify Exclusions
Description Polycystic ovaries Steroid induced Gestational diabetes ICD-9-CM Diagnosis 256.4 249, 251.8, 962.0 648.8

Electronic Specification
Denominator Numerator LDL-C Control <100 mg/dL Identify the most recent LDL-C test (Table ACDC-E) during the measurement year using claim/encounter or automated laboratory data. The organization may use a calculated or direct LDL for this indicator. The patient is numerator compliant if the most recent LDL-C level is <100 mg/dL. The patient is not numerator compliant if the automated result for the most recent LDL-C test during the measurement year is 100 mg/dL or is missing, or if an LDL-C test was not done during the measurement year. An organization that uses CPT Category II codes to identify numerator compliance for this indicator must search for all codes in Table ACDC-F and use the most recent code during the measurement year to evaluate whether the patient is numerator compliant. The eligible population.

Table ACDC-E: Codes to Identify LDL-C Screening


CPT 80061, 83700, 83701, 83704, 83721 CPT Category II 3048F, 3049F, 3050F LOINC 2089-1, 12773-8, 13457-7, 18261-8, 18262-6, 22748-8, 39469-2, 49132-4, 55440-2, 60678-0, 63474-1.

______________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

13

Table ACDC-F: Codes to Identify LDL-C Levels


Description Numerator compliant (LDL-C <100 mg/dL) Not numerator compliant (LDL-C 100 mg/dL) CPT Category II 3048F 3049F, 3050F

Hybrid Specification
Denominator Numerators LDL-C Control <100 mg/dL Electronic Medical Record The most recent LDL-C level performed during the measurement year is <100 mg/dL, as documented through automated laboratory data or medical record review. Refer to Electronic Specification to identify positive numerator hits from the electronic data. Documentation in medical record must include, at a minimum, a note indicating the date when the LDL-C test was performed and the result. The organization may use a calculated LDL for these indicators. LDL-C levels may be calculated from total cholesterol, HDL-C and triglycerides using the Friedewald equation if the triglycerides are 400 mg/dL. (LDL-C) = (total cholesterol) (HDL) (triglycerides/5) If lipoprotein (a) is measured, this calculation is: (LDL-C) = (total cholesterol) (HDL) (triglycerides/5) 0.3[lipoprotein (a)] These formulae are used when all levels are expressed in mg/dL and may not be used if triglycerides >400 mg/dL. A systematic sample drawn from the eligible population; refer to the General Guidelines.

Data Elements for Reporting


Organizations that submit ACO data to NCQA must provide the following data elements.

Table ACDC-1/2a and 2b: Data Elements for Comprehensive Adult Diabetes Care LDL-C control
Electronic Measurement year Data collection methodology (Electronic or Hybrid) Eligible population Denominator Numerator events by electronic data Numerator events by medical records Reported rate Lower 95% confidence interval Upper 95% confidence interval Hybrid

______________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

14

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Use of Imaging Studies for Low Back Pain (ALBP)


MODIFICATIONS FROM HEDIS 2012 VOLUME 2 TECHNICAL SPECIFICATIONS FOR HEALTH PLANS
Added Hybrid Method Specification.

Description
The percentage of patients with a primary diagnosis of low back pain who did not have an imaging study (plain X-ray, MRI, CT scan) within 28 days of the diagnosis.

Calculations
The measure is reported as an inverted rate [1 (numerator/denominator)]. A higher score indicates appropriate treatment of low back pain (i.e., proportion for whom imaging studies did not occur).

Definitions
Intake Period IESD Negative Diagnosis History January 1December 3 of the measurement year. The Intake Period is used to identify the first outpatient or ED encounter with a primary diagnosis of low back pain. Index Episode Start Date. The earliest date of service for any outpatient or ED encounter (Table ALBP-B) during the Intake Period with a primary diagnosis of low back pain. A period of 180 days (6 months) prior to the IESD when the patient had no claims/ encounters with any diagnosis of low back pain.

Eligible Population
Ages Defining the population Event/ diagnosis Step 1 18 years as of January 1 of the measurement year to 50 years as of December 31 of the measurement year. Attribution. Report on the aggregate ACO population; refer to the General Guidelines. Medicaid stratification. Option to report Medicaid separately. Outpatient or ED visit with a primary diagnosis of low back pain . Follow the steps below to identify the eligible population: Identify all patients in the specified age range who had an outpatient or ED encounter (Table ALBP-B) with a primary diagnosis of low back pain (Table ALBP-A) during the Intake Period.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

15

Table ALBP-A: Codes to Identify Low Back Pain


ICD-9-CM Diagnosis 721.3, 722.10, 722.32, 722.52, 722.93, 724.02, 724.03, 724.2, 724.3, 724.5, 724.6, 724.7, 738.5, 739.3, 739.4, 846, 847.2

Table ALBP-B: Codes to Identify Visit Type


Description Outpatient CPT 98925-98929, 98940-98942, 99201-99205, 99211-99215, 9921799220, 99241-99245, 99341-99345, 99347-99350, 99385, 99386, 99395, 99396, 99401-99404, 99411, 99412, 99420, 99429, 99455, 99456 99281-99285 UB Revenue 051x, 0520-0523, 0526-0529, 057x-059x, 0982, 0983 045x, 0981

ED*

*Do not include ED visits that result in an inpatient admission. Step 2 Step 3 Step 4 Determine the IESD. For each patient identified in step 1, determine the earliest episode of low back pain. If the patient had more than one encounter, include only the first encounter. Test for Negative Diagnosis History. Exclude patients with any low back pain diagnosis during the 180 days (6 months) prior to the IESD. Test for clinically appropriate imaging studies. Identify patients who have a diagnosis for which an imaging study (Table ALBP-C) in the presence of low back pain is clinically indicated. Electronic Method Specification Exclude patients with a diagnosis of cancer. Look as far back as possible in the patients history through 28 days after the Episode Start Date. Exclude patients diagnosed with recent trauma, intravenous drug abuse, or neurological impairment in the 12 months (1 year) prior to the IESD through 28 days after the Episode Start Date. Note: If using the Electronic Method, medical record documentation may not be used to identify exclusions. Hybrid Method Specification

Exclusions

Apply the Exclusion specifications for the Electronic Specification, above, to the electronic steps of the Hybrid Specification.
Exclusionary evidence in the medical record must include a note indicating a diagnosis of cancer. Look as far back as possible in the patients history through 28 days after the Episode Start Date. Exclusionary evidence in the medical record must include a note indicating diagnosis of recent trauma, intravenous drug abuse or neurological impairment in the 12 months (1 year) prior to the IESD through 28 days after the Episode Start Date. Note: If using the Hybrid Method, both medical records and electronic data may be used to identify exclusions.

_____________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

16

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Table ALBP-C:

Codes to Identify Exclusions (Clinically Appropriate Indications for Low Back Imaging)
ICD-9-CM Diagnosis 140-209, 230-239, V10 800-839, 850-854, 860-869, 905-909, 926.11, 926.12, 929, 952, 958-959 304.0-304.2, 304.4, 305.4-305.7 344.60, 729.2

Description Cancer Trauma IV drug abuse Neurologic impairment

Electronic Specification
Denominator Numerator The eligible population. An imaging study conducted on the IESD or in the 28 days following the IESD. Refer to Table ALBP-D for codes to identify imaging studies that count toward the numerator. A diagnosis code from Table ALBP-A must be used in conjunction with an imaging study code in Table ALBP-D.

Table ALBP-D: Codes to Identify Imaging Studies


Description Imaging studies CPT 72010, 72020, 72052, 72100, 72110, 72114, 72120, 7213172133, 72141, 72142, 72146-72149, 72156, 72158, 72200, 72202, 72220 UB Revenue 0320, 0329, 0350, 0352, 0359, 0610, 0612, 0614, 0619, 0972

Hybrid Specification
Denominator Numerator Electronic Medical Record A systematic sample drawn from the eligible population; refer to the General Guidelines. Patients with documentation in the medical record of an imaging study (plain x-ray, MRI, CT scan) of the lower back conducted on the IESD or in the 28 days following the IESD. Refer to Electronic Specification to identify positive numerator hits from the electronic data. Documentation of an imaging study (plain x-ray, MRI, CT scan) of the lower back must include the date and result of the study.

_____________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

17

Data Elements for Reporting


Organizations that submit ACO data to NCQA must provide the following data elements.

Table ALPB-ACO 1/2a and 2b: Data Elements for Use of Imaging Studies for Low Back Pain
Electronic Measurement year Data collection methodology (Electronic or Hybrid) Eligible population Denominator Numerator events by electronic data Numerator events by medical records Reported rate Lower 95% confidence interval Upper 95% confidence interval Hybrid

NCQA 2012

18

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Follow-Up After Hospitalization for Mental Illness (AFUH)


MODIFICATIONS FROM HEDIS 2012 VOLUME 2 TECHNICAL SPECIFICATIONS FOR HEALTH PLANS
Added Hybrid Specification.

Description
The percentage of discharges for patients 6 years of age and older who were hospitalized for treatment of selected mental health disorders and had an outpatient visit, intensive outpatient encounter or partial hospitalization with a mental health provider. Two rates are reported. 1. The percentage of discharges for which the patient received follow-up within 30 days of discharge. 2. The percentage of discharges for which the patient received follow-up within 7 days of discharge.

Eligible Population
Ages Defining the population Event/diagnosis 6 years and older as of the date of discharge. Attribution. Report on the aggregate ACO population; refer to the General Guidelines. Medicaid stratification. Option to report Medicaid separately. Discharged alive from an acute inpatient setting (including acute care psychiatric facilities) with a principal mental health diagnosis (Table AFUH-A) on or between January 1 and December 1 of the measurement year. The denominator for this measure is based on discharges, not on patients. If patients have more than one discharge, include all discharges on or between January 1 and December 1 of the measurement year.

Table AFUH-A: Codes to Identify Mental Health Diagnosis


ICD-9-CM Diagnosis 295299, 300.3, 300.4, 301, 308, 309, 311314

Mental health readmission or direct transfer

If the discharge is followed by readmission or direct transfer to an acute facility for any mental health principal diagnosis (Table AFUH-B, AFUH-C) within the 30-day follow-up period, count only the readmission discharge or the discharge from the facility to which the patient was transferred. Although rehospitalization might not be for a selected mental health disorder, it is probably for a related condition. Exclude both the initial discharge and the readmission/direct transfer discharge if the readmission/direct transfer discharge occurs after December 1 of the measurement year. Exclude discharges followed by readmission or direct transfer to a nonacute facility for a mental health principal diagnosis (Tables AFUH-B, AFUH-C) within the 30-day follow-up period. These discharges are excluded from the measure because readmission or transfer may prevent an outpatient follow-up visit from taking place. Refer to Table AFUH-D for codes to identify nonacute care.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012 Non-mental health readmission or direct transfer Exclude discharges in which the patient was transferred directly or readmitted within 30 days after discharge to an acute or nonacute facility for a non-mental health principal diagnosis. This includes any ICD-9-CM Diagnosis code or DRG code other than those listed in AFUH-B and AFUH-C. These discharges are excluded from the measure because rehospitalization or transfer may prevent an outpatient follow-up visit from taking place.

19

Table AFUH-B: Codes to Identify Mental Health Diagnosis


ICD-9-CM Diagnosis 290, 293-302, 306-316 Note: DSM-IV codes mirror ICD-9-CM codes. An organization that has access only to DSM-IV codes should use and document them. Follow the specifications outlined above for ICD-9-CM codes.

Table AFUH-C: Codes to Identify Inpatient Services


MSDRG 876, 880-887; exclude discharges with ICD-9-CM Principal Diagnosis codes 317-319

Table AFUH-D: Codes to Identify Nonacute Care


Description Hospice SNF Hospital transitional care, swing bed or rehabilitation Rehabilitation HCPCS UB Revenue 0115, 0125, 0135, 0145, 0155, 0650, 0656, 0658, 0659 019x UB Type of Bill 81x, 82x 21x, 22x, 28x 18x POS 34 31, 32

Respite Intermediate care facility 1002 Residential substance abuse treatment facility Psychiatric residential treatment center T2048, H0017-H0019 1001 Comprehensive inpatient rehabilitation facility Other nonacute care facilities that do not use the UB Revenue or Type of Bill codes for billing (e.g., ICF, SNF)

0118, 0128, 0138, 0148, 0158 0655

54 55 56 61

NCQA 2012

20

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Electronic Specification
Denominator Numerator 30-Day Follow-Up 7-Day Follow-Up An outpatient visit, intensive outpatient encounter or partial hospitalization (Table AFUH-E) with a mental health practitioner within 30 days after discharge. Include outpatient visits, intensive outpatient encounters or partial hospitalizations that occur on the date of discharge. An outpatient visit, intensive outpatient encounter or partial hospitalization (Table AFUH-E) with a mental health practitioner within 7 days after discharge. Include outpatient visits, intensive outpatient encounters or partial hospitalizations that occur on the date of discharge. The eligible population.

Table AFUH-E: Codes to Identify Visits


CPT HCPCS Follow-up visits identified by the following CPT or HCPCS codes must be with a mental health practitioner. 90804-90815, 98960-98962, 99078, 99201-99205, 99211-99215, G0155, G0176, G0177, G0409G0411, H0002, H0004, 99217-99220, 99241-99245, 99341-99345, 99347-99350, 99383-99387, H0031, H0034-H0037, H0039, H0040, H2000, H2001, 99393-99397, 99401-99404, 99411, 99412, 99510 H2010-H2020, M0064, S0201, S9480, S9484, S9485 CPT POS Follow-up visits identified by the following CPT or POS codes must be with a mental health practitioner. 90801, 90802, 90816-90819, 90821-90824, 90826-90829, 90845, WITH 03, 05, 07, 09, 11, 12, 13, 14, 15, 20, 22, 24, 33, 90847, 90849, 90853, 90857, 90862, 90870, 90875, 90876 49, 50, 52, 53, 71, 72 99221-99223, 99231-99233, 99238, 99239, 99251-99255 WITH 52, 53 UB Revenue The organization does not need to determine practitioner type for follow-up visits identified by the following UB Revenue codes. 0513, 0900-0905, 0907, 0911-0917, 0919 Visits identified by the following revenue codes must be with a mental health practitioner or in conjunction with any diagnosis code from Table AFUH-A. 0510, 0515-0517, 0519-0523, 0526-0529, 0982, 0983

Note
Organizations may have different methods for billing intensive outpatient encounters and partial hospitalizations. Some methods may be comparable to outpatient billing, with separate claims for each date of service; others may be comparable to inpatient billing, with an admit date, a discharge date and units of service. Organizations whose billing methods are comparable to inpatient billing may count each unit of service as an individual visit. The unit of service must have occurred during the period required for the rate (e.g., within 30 days after discharge or within 7 days after discharge). Visits for substance abuse are not compliant events for the numerator. Refer to the last page of this specification for the definition of mental health practitioner.

___________ Current Procedural Terminology 2012 American Medical Association. All rights reserved.

NCQA 2012

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

21

Hybrid Specification
Denominator Numerator 30-Day Follow-Up 7-Day Follow-Up Electronic Medical Record Documentation of an outpatient visit, intensive outpatient encounter or partial hospitalization with a mental health practitioner within 30 days after discharge. Documentation of an outpatient visit, intensive outpatient encounter or partial hospitalization with a mental health practitioner within 7 days after discharge. Refer to Electronic Specification to identify positive numerator hits from the electronic data. Documentation in the medical record of an outpatient visit, intensive encounter or partial hospitalization with a mental health practitioner within the specified timeframe. Include outpatient visits, intensive outpatient encounters or partial hospitalizations that occur on the date of discharge. Documentation must include the date of the visit. A systematic sample drawn from the eligible population; refer to the General Guidelines.

Data Elements for Reporting


Organizations that submit ACO data to NCQA must provide the following data elements.

Table AFUH-1/2a and 2b: Data Elements for Follow-Up After Hospitalization for Mental Illness
Electronic Measurement year Data collection methodology (Electronic or Hybrid) Eligible population Denominator Numerator events by electronic data Numerator events by medical records Reported rate Lower 95% confidence interval Upper 95% confidence interval Each of the 2 rates Each of the 2 rates Each of the 2 rates Each of the 2 rates Hybrid


Each of the 2 rates Each of the 2 rates Each of the 2 rates Each of the 2 rates Each of the 2 rates Each of the 2 rates

NCQA 2012

22

Draft Technical Specifications for ACO Measures for Public Comment Obsolete After April 6, 2012

Definitions
Mental health practitioner A practitioner who provides mental health services and meets any of the following criteria. An MD or doctor of osteopathy (DO) who is certified as a psychiatrist or child psychiatrist by the American Medical Specialties Board of Psychiatry and Neurology or by the American Osteopathic Board of Neurology and Psychiatry; or, if not certified, who successfully completed an accredited program of graduate medical or osteopathic education in psychiatry or child psychiatry and is licensed to practice patient care psychiatry or child psychiatry, if required by the state of practice. An individual who is licensed as a psychologist in his/her state of practice. An individual who is certified in clinical social work by the American Board of Examiners; who is listed on the National Association of Social Workers Clinical Register; or who has a masters degree in social work and is licensed or certified to practice as a social worker, if required by the state of practice. A registered nurse (RN) who is certified by the American Nurses Credentialing Center (a subsidiary of the American Nurses Association) as a psychiatric nurse or mental health clinical nurse specialist, or who has a masters degree in nursing with a specialization in psychiatric/mental health and two years of supervised clinical experience and is licensed to practice as a psychiatric or mental health nurse, if required by the state of practice. An individual (normally with a masters or a doctoral degree in marital and family therapy and at least two years of supervised clinical experience) who is practicing as a marital and family therapist and is licensed or a certified counselor by the state of practice, or if licensure or certification is not required by the state of practice, who is eligible for clinical membership in the American Association for Marriage and Family Therapy. An individual (normally with a masters or doctoral degree in counseling and at least two years of supervised clinical experience) who is practicing as a professional counselor and who is licensed or certified to do so by the state of practice, or if licensure or certification is not required by the state of practice, is a National Certified Counselor with a Specialty Certification in Clinical Mental Health Counseling from the National Board for Certified Counselors (NBCC).

NCQA 2012

You might also like