Current Medical Research and Opinion

ISSN: 0300-7995 (Print) 1473-4877 (Online) Journal homepage: http://www.tandfonline.com/loi/icmo20

Compliance with clinical guidelines and adherence

to antiretroviral therapy among patients living
with HIV

Neeta Tandon, Jianbin Mao, Adam Shprecher, Amy J Anderson, Feng Cao,
Xiaolong Jiao & Kimberley Brown

To cite this article: Neeta Tandon, Jianbin Mao, Adam Shprecher, Amy J Anderson, Feng Cao,
Xiaolong Jiao & Kimberley Brown (2018): Compliance with clinical guidelines and adherence to
antiretroviral therapy among patients living with HIV, Current Medical Research and Opinion, DOI:
10.1080/03007995.2018.1519499

To link to this article: https://doi.org/10.1080/03007995.2018.1519499

View supplementary material

Accepted author version posted online: 03

Sep 2018.

Submit your article to this journal

Article views: 1

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=icmo20
Compliance with clinical guidelines and adherence to antiretroviral therapy among patients living with
HIV
Authors: Neeta Tandona; Jianbin Maob*; Adam Shprechera; Amy J Andersonb; Feng Caob; Xiaolong Jiaoa;
and Kimberley Browna
Author affiliations: aJanssen Scientific Affairs, LLC, Titusville, NJ, USA; bOptum, Eden Prairie, MN, USA

*Corresponding author: Jianbin Mao, Optum, MN101-E300, 11000 Optum Circle, Minneapolis, MN
55440-9472. Email: jianbin.mao@optum.com

t
ip
Declaration of funding

cr
This study was sponsored by Janssen Scientific Affairs, LLC.

us
Declaration of financial/other relationships
N. Tandon, A. Shprecher, and K. Brown are employees of Janssen. X. Jiao is a former employee of
Janssen (when the study was conducted). J. Mao, A. Anderson, and F. Cao are employees of Optum,
an
which was contracted by Janssen to conduct this research.

Author contributions
M
N. Tandon, A. Shprecher, and K. Brown were involved in conception and design of the study,
interpretation of the data, and drafting and revising the manuscript. J. Mao was involved in conception
and design of the study, analysis and interpretation of the data, and drafting and revising the
d

manuscript. A. Anderson and F. Cao were involved in analysis and interpretation of the data, and
e

drafting and revising the manuscript. X. Jiao was involved in study design, analysis and interpretation of
the data, and drafting and revising the manuscript. All authors provided final approval of the manuscript
pt

to be published, and agree to be accountable for all aspects of the work. CMRO peer reviewers on this
manuscript have received an honorarium from CMRO for their review work, but have no other relevant
ce

financial relationships to disclose.

Acknowledgments
Ac

Medical writing support was provided by Yvette Edmonds, PhD (Optum). The authors would also like to
thank Shiqiang Li, MS; Yiyu Fang, PhD; Damon Van Voorhis; and Vincent Peichel (all of Optum) for
assistance with programming, study dataset creation, and table verification; and Jamie Forlenza,
PharmD (Janssen) for contributions to the study design, data analysis, and study report.
Abstract
Objective: Evaluation of provider compliance with antiretroviral (ARV) treatment guidelines and patient
adherence to ARVs is important for HIV care quality assessment; however, there are few current real-
world data for guideline compliance and ARV adherence in the US. We evaluated provider compliance
with US Department of Health and Human Services (DHHS) guidelines and patient adherence to ARVs in
a US population of patients with HIV.
Methods: This was a retrospective claims study of adults with HIV-1 receiving ARV treatment between
January 2010–December 2014. Follow-up began at first ARV treatment and ended at health plan
disenrollment or study end. ARV regimens for treatment-naïve patients were categorized as
’preferred/recommended,’ ’alternative,’ or ’non–preferred/recommended/alternative’ according to
DHHS guidelines. ARV adherence was evaluated using proportion of days covered (PDC) and medication

t
ip
possession ratio (MPR).
Results: The analysis included 25,320 patients (84.4% male, mean age 45.3 years) and 39,071 regimens.

cr
Preferred/recommended regimens were most common during each study year, but the proportion of
non–preferred/recommended/alternative regimens was substantial (15.9%–20.6%). Only 53.6% of

us
patients had optimal adherence by PDC ≥0.95, and 57.9% by MPR ≥0.95. Guideline noncompliance and
suboptimal adherence were more prevalent among female vs male patients (22.6% vs 14.8% [in 2014]
and 65.9% vs 53.7%, respectively).
an
Conclusions: Provider noncompliance with DHHS guidelines and suboptimal ARV adherence among
patients with HIV remain common in real-world practice, particularly for female patients. Healthcare
providers should follow the latest clinical guidelines to ensure that patients receive recommended
M
therapy, and address nonadherence when selecting ARV regimens.

Keywords: adherence, HIV, treatment, guidelines, DHHS, noncompliance, nonadherence

e d

Short title: Compliance with clinical guidelines and adherence to antiretroviral therapy
pt
ce
Ac
Introduction
HIV infection constitutes a substantial healthcare burden; it is estimated that in the US there are more
than 1.1 million people living with HIV, and nearly 40,000 new infections occur each year [1,2]. Although
there is no cure for HIV, the advent of highly active antiretroviral (ARV) therapy has greatly improved
patient survival, transforming HIV from an acute disease to a chronic condition and allowing infected
individuals to live a near-normal lifespan [3].
To optimize clinical outcomes among those with HIV, the US Department of Health and Human Services
(DHHS) has developed guidelines to assist healthcare providers (HCPs) in choosing initial ARV regimens
for HIV-1–infected patients [4]. These guidelines are updated frequently as preferred treatment
strategies evolve due to updated information regarding drug safety and efficacy, as well as the
introduction of new drug combinations with more convenient formulations and fewer adverse effects.

t
ip
Recommended regimens include 2 nucleos(t)ide reverse transcriptase inhibitors (NRTIs) and a third
agent from another class of antiretroviral drugs (an integrase strand transfer inhibitor [INSTI], a non-

cr
nucleoside reverse transcriptase inhibitor [NNRTI], or a protease inhibitor with a booster agent). While
modern ARV regimens are highly effective, stringent medication adherence is requisite for sustained

us
viral suppression and improved survival [4]; poor and suboptimal adherence to ARV therapy has been
associated with increased risk of virologic failure, development of drug resistance, and death [5], as well
as higher healthcare utilization and related costs [6,7].
an
Given that compliance with recommended treatment guidelines and adherence to ARV therapy are
associated with better clinical outcomes [5,8], evaluation of compliance and adherence is an important
aspect of HIV care quality assessment. However, there are few real-world data regarding current ARV
M
treatment patterns and guideline compliance in the US [9,10], as many previous studies were conducted
before modern therapies were available [11-13]. Furthermore, recent studies of ARV adherence were
performed among patient groups that may not be representative of the US HIV-infected population
d

[9,14-16]. To help address these gaps, we performed a claims data analysis to evaluate healthcare
e

providers’ compliance with DHHS-recommended ARV regimens and ARV adherence in a large US
population of patients with HIV, the majority of whom were commercially insured. To our knowledge,
pt

this is the first study to simultaneously address patient- and provider-related aspects of ARV adherence
and guideline compliance.
ce

Methods
Ac

Data sources
This was a retrospective observational study using administrative claims data from the Optum Research
Database and Impact National Benchmark Database. These databases include medical and pharmacy
claims with linked enrollment information for commercial and Medicare Advantage health plan
members and are geographically diverse across the US. Medical claims included diagnosis and procedure
codes from the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM);
Current Procedural Terminology (CPT) or Healthcare Common Procedure Coding System (HCPCS) codes;
and other information. Outpatient pharmacy claims included National Drug Codes (NDC) for dispensed
medications, quantity dispensed, dose, and number of days’ supply. Because no identifiable protected
health information was extracted or accessed during the course of the study, institutional review board
approval or waiver of authorization was not required.

Study sample
Patients
The study population comprised patients 18 years or older who had at least 1 pharmacy claim for an
ARV (Supplementary Table) between 01 January 2010, and 31 December 2014 (identification period),
and a claim with an HIV diagnosis code (ICD-9-CM 042, 795.71, or V08) in any position from 01 January
2006 (or the start date of the patient’s continuous health plan enrollment, if later), through 31
December 2014 (study period). The date of the first ARV claim was set as the index date. Only patients
who had continuous health plan enrollment with both medical and pharmacy coverage for at least 6

t
ip
months before the index date (baseline period) and 1 day after the index date (follow-up period) were
included in the study sample. The follow-up period began on the index date and ended on the earlier of

cr
health plan disenrollment or the study end date. Patients with claims for only nucleos(t)ide reverse
transcriptase inhibitors (N(t)RTIs) during the identification period were excluded, as were those who had

us
a claim with an HIV-2 diagnosis code (ICD-9-CM 079.53) at any time during the study period. The study
was focused on HIV-1 to reflect the majority of patients with HIV infection, and because many ARVs that
were guideline recommended during the study period are not indicated for HIV-2.
Definition of ARV regimens
an
The start date of each ARV treatment regimen was defined as the latest of the index date; the end date
of a previous regimen plus 1 day; or, if there was a ≥ 90-day gap in treatment, the onset of a new
M
treatment regimen. All ARV medications filled within 14 days of the regimen start date were considered
to be part of the regimen. A regimen continued until switching of a third agent (INSTI, NNRTI, protease
inhibitor, fusion inhibitor, or entry inhibitor); stopping of all NRTIsa; or a gap in therapy (defined as ≥ 90
d

days without access to any ARV). The end date of each regimen was defined as the earliest of
e

discontinuation of one of the third agent medications, discontinuation of all NRTIs, addition of a new
third agent medication, the end of the follow-up period, or the end of the study period.
pt

Study measures
ce

Patient characteristics
Demographic and clinical characteristics assessed for all patients during the baseline period included
age, sex, US Census region, insurance type, comorbidity burden based on both the Quan–Charlson index
Ac

[17] (excluding HIV/AIDS) and Clinical Classifications Software from the Agency for Healthcare Research
and Quality (AHRQ) [18], and prior treatment status; patients who were on their first regimen and had
no fills for an ARV medication in the 6 months prior to the regimen start date were defined as
treatment-naïve.
HCP compliance with DHHS guidelines among treatment-naïve patients
For treatment-naïve patients, the frequency and duration of ARV regimens were described, and HCP
compliance of regimens with the DHHS guidelines was assessed as follows. During each study year,
a
For the purpose of regimen determination, NRTIs were treated as a single medication; ie, the
appearance or disappearance of NRTIs as a whole from a regimen triggered a regimen change, but
switching from one NRTI to another did not.
regimens were categorized as ’preferred/recommended regimens,’ ’alternative regimens,’ or ’non–
preferred/recommended/alternative regimens’ (all regimens not listed in either of the above categories)
according to the ’What to Start’ section of the 2010-2014 DHHS guidelines for ARV treatment of HIV-
infected adults and adolescents [19]. If a regimen’s status changed in one study year because of an
update to the DHHS guidelines, a single category was assigned to the regimen for that year, as follows. A
regimen was categorized as ’preferred/recommended’ if it had been listed as such in any of the
guidelines applicable to that year (eg, the 2014 analysis included guidelines issued on February 12, 2013;
May 1, 2014; and November 13, 2014). Regimens not categorized as ’preferred/recommended’ were
categorized as ’alternative’ if they had been listed as such in any of the applicable guidelines, and all
remaining regimens were categorized as ’non–preferred/recommended/alternative.’ To simplify
analysis, changes in boosting agents (ritonavir or cobicistat; BAs) were not considered.

t
ip
Patient adherence to ARV treatment
ARV adherence was evaluated for all patients from ARV treatment initiation through the earlier of

cr
health plan disenrollment or December 31, 2014. Adherence was assessed using 2 different measures:
proportion of days covered (PDC) and medication possession ratio (MPR) [20]. Although MPR has been

us
commonly used, it can overestimate adherence for multiple-drug regimens and among patients who
routinely refill their medications early; PDC, which is more conservative, is now the preferred metric for
ARV adherence [21]. PDC was calculated by dividing the number of days during which the drug(s) in
an
question were available (based on prescription fill dates and days’ supply) by the number of days in the
follow-up period. The proportion of time during follow-up in which a patient was in possession of any
ARV was designated “PDC1” and the proportion of time during follow-up in which a patient was in
M
possession of 2 or more ARVs was designated ’PDC2.’ MPR was calculated by dividing the total number
of days’ supply of any ARV during the follow-up period (except the last fill) by the number of days
between the first and last fill; only patients with ≥ 2 ARV fills were included. Poor medication adherence
d

was defined as PDC or MPR < 0.80, suboptimal as PDC or MPR ≥ 0.80 but < 0.95, and optimal as PDC or
e

MPR ≥ 0.95 [5,20,21].

pt

Statistical analysis
Descriptive analyses were conducted for all baseline and study outcome variables. Frequencies and
ce

percentages were provided for categorical variables, and means and standard deviations (SDs) were
provided for continuous variables. All statistical analyses were performed using SAS 9.4 (SAS Institute,
Cary, NC, USA).
Ac
Results

Study sample
There were 25,320 patients (Figure 1) and 39,071 regimens included in the analysis. The majority of
patients (94.1%) were commercially insured; 84.4% were men, mean (SD) age was 45.3 (10.8) years, and
35.7% (9,030/25,320) were treatment naïve (Table 1). Excluding viral infections, the most common
comorbid conditions were lipid disorders (27.4% of patients), respiratory infections (20.8%), and
hypertension (19.5%). Clinical and demographic characteristics were similar between men and women
(Table 1).

t
Treatment characteristics

ip
Mean (SD) follow-up after ARV therapy initiation was 753.6 (558.7) days. Of the 25,320 patients in the
study, 30.7% had more than 1 regimen during the follow-up period. The average regimen duration was

cr
489.6 ±507.8 days (median = 290 days).
Of the 38,663 observed regimens with a third agent, 29.4% contained protease inhibitors, 42.1%

us
NNRTIs, 11.3% INSTIs, and < 0.5% fusion and entry inhibitors; 16.9% of regimens contained multiple ARV
classes (Table 2). The most common third agents (observed in > 5% of regimens) included efavirenz
(36.3%), raltegravir (18.0%,) atazanavir (15.5%), darunavir (13.5%), lopinavir (8.8%), nevirapine (6.3%),
an
etravirine (5.3%), and rilpivirine (5.3%). Compared with regimens for male patients, those for female
patients had a higher prevalence of protease inhibitors (male, 27.9%; female, 36.8%) and a lower
M
prevalence of NNRTIs (male, 43.4%; female, 35.8%; Table 2). With regard to individual drugs, female
patient regimens were less likely to contain efavirenz (male, 37.5%; female, 30.6%) and more likely to
contain atazanavir (male, 14.7%; female, 19.7%) and lopinavir (male, 8.0%; female, 12.7%).
d

HCP compliance with DHHS guidelines among treatment-naïve patients

e

Among the 9,030 treatment-naïve patients, preferred/recommended regimens were prescribed more
pt

commonly than other categories over each of the 5 years in the study period, representing 72.5%,
68.2%, 58.1%, 47.1%, and 81.7% of total regimens used during 2010, 2011, 2012, 2013, and 2014,
ce

respectively (Figure 2A). During these same years, alternative regimens comprised 6.9%, 12.1%, 24.0%,
35.2%, and 2.3% of the regimens used, respectively. The use of non–
preferred/recommended/alternative regimens exhibited a small decline over time, from 20.6% in 2010
Ac

to 15.9% in 2014 (Figure 2A). The proportion of non–preferred/recommended/alternative regimens

among men also declined during the study period, from 20.3% in 2010 to 14.8% in 2014 (Figure 2B);
however, among women, this proportion ranged from 22.1% in 2010 to 29.3% in 2012 and exceeded
that of alternative regimens in every study year except 2013 (Figure 2C).

Patient adherence to ARV treatment

A substantial number of patients had poor or suboptimal adherence as assessed by both PDC and MPR.
The proportions of patients with PDC1 < 0.80, PDC2 < 0.80, and MPR < 0.80 were 25.1%, 29.4%, and
20.4%, respectively, while the proportions of patients with PDC1, PDC2, and MPR ≥ 0.80 but < 0.95 were
21.3%, 26.2%, and 21.7%, respectively (Table 3). Mean (SD) PDC1 and PDC2 were 0.85 (0.23) and 0.82
(0.23), respectively; mean (SD) MPR was 0.88 (0.18). A higher proportion of women than men had poor
or suboptimal adherence; PDC2 was < 0.95 in 65.9% of women compared with 53.7% of men, while MPR
was < 0.95 in 50.5% of women compared with 40.5% of men.

Discussion

Healthcare provider compliance with treatment guidelines is associated with improved outcomes
among patients with HIV, including greater odds of viral suppression, higher CD4 cell count, and better
medication adherence [8,10,13]. It is important for HCPs to follow the latest guidelines because
preferred treatment strategies evolve frequently with the availability of new information regarding drug

t
safety and efficacy as well as more potent and conveniently formulated drug combinations. However,

ip
our real-world analysis of ARV treatment patterns among US patients with HIV-1 revealed that from 12%
to 21% of the total treatment-naïve study population received initial regimens that were not listed as

cr
preferred/recommended or alternative in contemporary DHHS guidelines.
Our findings are consistent with those of other assessments of DHHS guideline compliance among HCPs

us
in the US [9,10]. In an analysis of 3,593 adults initiating ARV treatment between 2007 and 2011,
Johnston et al. [10] found that 45% of patients initiated regimens that were not guideline preferred. A
similar analysis by Cooke et al. [9] revealed that among 3,528 HIV-infected adults with at least 1 claim
an
for an ARV drug, 31% were on regimens in 2010–2011 that were not listed in the 2011 guidelines, while
another 1% were on contraindicated regimens. Although nonrecommended regimens were substantially
M
more prevalent in these earlier studies compared with ours, this discrepancy may be attributable to
differences in methodology. First, the Johnston et al. [10] analysis did not differentiate between
regimens listed as alternative and those not listed in the guidelines at all, which would increase the
d

number of regimens categorized as nonrecommended. Perhaps more importantly, Johnston et al.

categorized regimens according to the March 2012 DHHS guidelines only, and Cooke et al. [9] did not
e

require study patients to be treatment-naïve. The DHHS guidelines are updated frequently and changing
pt

the treatment of patients who are doing well on older regimens is not generally encouraged; therefore,
applying a single set of guidelines across all regimens regardless of year—or categorizing the current
ce

regimens of patients who are treatment-experienced—may result in an inflated assessment of guideline

noncompliance that may not be clinically meaningful. This effect was mitigated in the present analysis
by limiting the guideline compliance analysis to treatment-naïve patients, and by categorizing each
Ac

regimen according to the year-appropriate version of the DHHS guidelines.

Our data also suggest an inverse relationship between the prevalence of recommended and alternative
regimens across the study years, which may reflect the tendency of HCPs to adopt new treatment
paradigms informed by clinical trial results or clinical experience prior to their addition to the guidelines
[12]. In the present study, guideline compliance reached a nadir in 2013 only to peak in 2014. While
2014 was the first year in which single-tablet initial regimens including the INSTIs dolutegravir and
elvitegravir were categorized as recommended [22], the 2014 guideline update was based in part on
clinical trial results that were already becoming available in 2013 [23-26]. Thus, it is possible that the low
guideline compliance observed in 2013 resulted from HCPs taking action on compelling new data in
advance of an official guideline change. This suggestion is supported by findings from another recent
claims analysis which demonstrated that many patients were being initiated on the new single-tablet
regimens in 2013, the year before these drugs were recommended in the guidelines [27].
Notably, we found that a higher proportion of women had non–preferred/recommended/alternative
regimens compared with men. This result reflects a history of evidence for gender disparity in ARV
treatment patterns; multiple studies have indicated that women are less likely to receive ARV therapy
overall [28,29] and more likely to receive regimens that are not guideline recommended [30]. These
phenomena may arise from a combination of clinical and socioeconomic factors that function to limit
healthcare access and affect medication-taking behavior among women. For example, women are more
likely than men to experience adverse events secondary to ARV therapy, and many women living with
HIV face additional challenges such as poverty, lack of education, drug use, and racial/gender
discrimination [31,32]. Although a considerable amount of research has been directed toward

t
ip
uncovering sex disparities in the clinical management of patients with HIV, our findings suggest that
continued effort is necessary to understand and address the specific needs of HIV-infected women.

cr
Receiving the right ARV regimen is only the first step in successful management of HIV infection;
patients also must adhere to their regimens to achieve and maintain viral load control. Studies

us
performed in the early days of ARV therapy indicated that an adherence level of 95% or greater was
required to maintain viral suppression [33]. The minimum adherence threshold has been tempered
somewhat over the years as more potent drugs and more convenient single-tablet regimens have been
an
introduced [34-37]. However, maintaining the highest possible level of adherence remains imperative
because the risk of viral rebound and development of drug-resistant HIV strains decreases with better
adherence [38-42], and treatment failure must be prevented in order to avoid limiting future treatment
M
options. Suboptimal ARV adherence was common in the present study, with 56% of patients exhibiting
PDC2 below 0.95. This finding reflects the results of other recent US claims analyses, in which
proportions of nonadherent patients ranged from 31% to 68% depending on the stringency of the
d

adherence definition (ie, MPR ≥ 0.80 [14], MPR ≥ 0.90 [9], or PDC ≥ 0.95) [15,16,43]. Earlier studies were
e

limited by restriction to an older Medicare-only population [15], claims from only 1 pharmacy chain [16],
or a narrow geographic region [9,14]; therefore, our analysis represents an important contribution to
pt

current understanding of ARV adherence in a representative US population of HIV-infected patients.

In parallel to the results of our assessment of DHHS guideline compliance, suboptimal ARV compliance
ce

was more prevalent among women than men in the present study. This finding is consistent with
existing evidence that women tend to have lower ARV adherence and higher discontinuation compared
with men [8,9,44,45], and highlights the need for strategies to promote medication adherence and/or
Ac

select ARV regimens that reduce the risk of drug resistance among women living with HIV.

Limitations
Our findings should be considered in light of several limitations. This analysis was conducted using data
from 2010 through 2014; therefore, ARV treatment patterns may have evolved since the data were
collected. Because this study was conducted in a managed care population among patients meeting the
study criteria, the results may not be generalizable to uninsured persons or other populations similar to
those excluded from the analysis. The hierarchical regimen categorization scheme utilized for instances
in which multiple guidelines were available in a given year may have shifted regimens into a
preferred/recommended or alternative categorization that may not have been applicable at the time of
guideline implementation. Furthermore, because the reasons for regimen selection are not observable
in the claims data, the categorization scheme did not account for circumstances that might make an
alternative regimen the best choice for a particular patient (eg, disease characteristics or comorbidities).
Finally, while patients with evidence of prior ARV use were excluded from the analysis, it is possible that
some patients categorized as treatment naïve had previously received ARV agents outside of the study
period (eg, prior to joining the health plan). This study is also subject to certain limitations inherent to
claims-based analyses: medications provided as samples by an HCP or as part of a clinical trial are not
accounted for in claims data, and pharmacy claims do not indicate whether the medication was taken as
prescribed.

Conclusion

t
ip
The results of this real-world analysis suggest that provider noncompliance with DHHS guidelines for

cr
initial ARV regimens and suboptimal patient ARV adherence remain common in US clinical practice,
particularly for female patients. It is critical that HCPs play an active role in addressing this issue by

us
complying with current treatment recommendations; by implementing strategies to improve patient
adherence, such as prescribing single-tablet regimens when appropriate [9]; and by striving to minimize
the clinical impact of poor patient adherence—for example, by using regimens with high genetic barriers
an
to resistance when treating patients with uncertain adherence [46]. Further studies are warranted to
elucidate the relationship between guideline compliance/patient adherence and clinical/economic
M
outcomes among patients with HIV infection.

References
d

[1] Centers for Disease Control and Prevention. HIV in the United States: at a glance. 2017 [cited 2017
Jun 19]. Available at: https://www.cdc.gov/hiv/statistics/overview/ataglance.html.
e

[2] Centers for Disease Control and Prevention. HIV surveillance report. 2016 [cited 2017 Jun 19].
pt

Available at: https://www.cdc.gov/hiv/pdf/library/reports/surveillance/cdc-hiv-surveillance-report-

2015-vol-27.pdf.
ce

[3] Samji H, Cescon A, Hogg RS, et al. Closing the gap: increases in life expectancy among treated HIV-
positive individuals in the United States and Canada. PLoS One. 2013;8(12):e81355.
[4] US Department of Health & Human Services, Panel on Antiretroviral Guidelines for Adults and
Ac

Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and
adolescents. 2016 [cited 2016 Feb 3]. Available at:
https://aidsinfo.nih.gov/guidelines/html/1/adult-and-adolescent-arv-guidelines/0.
[5] McMahon JH, Jordan MR, Kelley K, et al. Pharmacy adherence measures to assess adherence to
antiretroviral therapy: review of the literature and implications for treatment monitoring. Clin
Infect Dis. 2011;52(4):493–506.
[6] Nachega JB, Leisegang R, Bishai D, et al. Association of antiretroviral therapy adherence and health
care costs. Ann Intern Med. 2010;152(1):18–25.
[7] Gardner EM, Maravi ME, Rietmeijer C, et al. The association of adherence to antiretroviral therapy
with healthcare utilization and costs for medical care. Appl Health Econ Health Policy. 2008;6(2–
3):145–155.
[8] Cocohoba J, Wang QJ, Cox C, et al. Consistency of initial antiretroviral therapy with HIV treatment
guidelines in a US cohort of HIV-infected women. J Acquir Immune Defic Syndr. 2008;47(3):377–
383.
[9] Cooke CE, Lee HY, Xing S. Adherence to antiretroviral therapy in managed care members in the
United States: a retrospective claims analysis. J Manag Care Pharm. 2014;20(1):86–92.
[10] Johnston SS, Juday T, Farr AM, et al. Comparison between guideline-preferred and nonpreferred
first-line HIV antiretroviral therapy. Am J Manag Care. 2014;20(6):448–455.
[11] Ghani AC, Donnelly CA, Anderson RM. Patterns of antiretroviral use in the United States of America:

t
ip
analysis of three observational databases. HIV Med. 2003;4(1):24–32.
[12] Holodniy M, Hornberger J, Rapoport D, et al. Relationship between antiretroviral prescribing

cr
patterns and treatment guidelines in treatment-naive HIV-1-infected US veterans (1992-2004). J
Acquir Immune Defic Syndr. 2007;44(1):20–29.

us
[13] Cocohoba JM, Althoff KN, Godfrey R, et al. Influence of gender on receipt of guideline-based
antiretroviral therapy in the era of HAART. AIDS Care. 2012;24(1):20–29.
[14] Hirsch JD, Gonzales M, Rosenquist A, et al. Antiretroviral therapy adherence, medication use, and
an
health care costs during 3 years of a community pharmacy medication therapy management
program for Medi-Cal beneficiaries with HIV/AIDS. J Manag Care Pharm. 2011;17(3):213–223.
[15] Abara WE, Adekeye OA, Xu J, et al. Adherence to combination antiretroviral treatment and clinical
M
outcomes in a Medicaid sample of older HIV-infected adults. AIDS Care. 2017;29(4):441–448.
[16] Murphy P, Cocohoba J, Tang A, et al. Impact of HIV-specialized pharmacies on adherence and
persistence with antiretroviral therapy. AIDS Patient Care STDS. 2012;26(9):526–531.
d

[17] Quan H, Li B, Couris CM, et al. Updating and validating the Charlson comorbidity index and score for
e

risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol.
2011;173(6):676–682.
pt

[18] HCUP CCS [computer program]. Rockville (MD): Agency for Healthcare Research and Quality; 2015.
Available at: https://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp.
ce

[19] US Department of Health & Human Services, Panel on Antiretroviral Guidelines for Adults and
Adolescents. Archived adult and adolescent guidelines. 2016 [cited 2017 Apr 12]. Available at:
https://aidsinfo.nih.gov/guidelines/archive/adult-and-adolescent-guidelines/.
Ac

[20] Andrade SE, Kahler KH, Frech F, et al. Methods for evaluation of medication adherence and
persistence using automated databases. Pharmacoepidemiol Drug Saf. 2006;15(8):565–574;
discussion 575–567.
[21] Nau DP. Proportion of days covered as a preferred method of measuring medication adherence.
[cited 2017 May 19]. Available at:
http://www.pqaalliance.org/images/uploads/files/PQA%20PDC%20vs%20%20MPR.pdf.
[22] US Department of Health & Human Services, Panel on Antiretroviral Guidelines for Adults and
Adolescents. Archived adult and adolescent guidelines. 2014 [cited 2018 Aug 7]. Available at:
http://aidsinfo.nih.gov/contentfiles/AdultandAdolescentGL003392.pdf.
[23] Raffi F, Jaeger H, Quiros-Roldan E, et al. Once-daily dolutegravir versus twice-daily raltegravir in
antiretroviral-naïve adults with HIV-1 infection (SPRING-2 study): 96 week results from a
randomised, double-blind, non-inferiority trial. Lancet Infect Dis. 2013;13(11):927-935.
[24] Walmsley SL, Antela A, Clumeck N, et al. Dolutegravir plus abacavir-lamivudine for the treatment of
HIV-1 infection. N Engl J Med. 2013;369(19):1807-1818.
[25] Feinberg J, Clotet B, Khuong-Josses MA. Once-daily dolutegravir (DTG) is superior to
darunavir/ritonavir (DRV/r) in antiretroviral-naïve adults: 48 week results from FLAMINGO
(ING114915). Presented at: 53rd Interscience Conference on Antimicrobial Agents and
Chemotherapy. 2013; Denver, CO.
[26] Clumeck N, Molina J, Henry K, Gathe J. Elvitegravir/cobicistat/emtricitabine/tenofovir DF (STB) has
durable efficacy and differentiated safety compared to atazanavir boosted by ritonavir plus

t
ip
emtricitabine/tenofovir DF at week 144 in treatment-naive HIV-1 infected patients. Presented at:
14th European AIDS Conference. 2013; Brussels, Belgium.

cr
[27] Mujeebuddin S, Mao J, Bengtson LGS, Buikema AR. HIV treatment guidelines and medication
initiation patterns in treatment naive adults, 2008-2016. Value Health. 2018;18(8).

us
[28]King WD, Minor P, Ramirez Kitchen C, et al. Racial, gender and geographic disparities of
antiretroviral treatment among US Medicaid enrolees in 1998. J Epidemiol Community Health.
2008;62(9):798–803.
an
[29] Fleishman JA, Yehia BR, Moore RD, et al. Disparities in receipt of antiretroviral therapy among HIV-
infected adults (2002-2008). Med Care. 2012;50(5):419–427.
[30] Wandeler G, Keiser O, Hirschel B, et al. A comparison of initial antiretroviral therapy in the Swiss
M
HIV Cohort Study and the recommendations of the International AIDS Society-USA. PLoS One.
2011;6(12):e27903.
[31] Loutfy MR, Sherr L, Sonnenberg-Schwan U, et al. Caring for women living with HIV: gaps in the
d

evidence. J Int AIDS Soc. 2013;16:18509.

e

[32] Monforte A, Anderson J, Olczak A. What do we know about antiretroviral treatment of HIV in
women? Antivir Ther. 2013;18 (Suppl 2):27–34.
pt

[33] Paterson DL, Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy and outcomes in
patients with HIV infection. Ann Intern Med. 2000;133(1):21–30.
ce

[34] Shuter J. Forgiveness of non-adherence to HIV-1 antiretroviral therapy. J Antimicrob Chemother.

2008;61(4):769–773.
[35] Bangsberg DR. Less than 95% adherence to nonnucleoside reverse-transcriptase inhibitor therapy
Ac

can lead to viral suppression. Clin Infect Dis. 2006;43(7):939–941.

[36] Viswanathan S, Detels R, Mehta SH, et al. Level of adherence and HIV RNA suppression in the
current era of highly active antiretroviral therapy (HAART). AIDS Behav. 2015;19(4):601–611.
[37] Sutton SS, Magagnoli J, Hardin JW. Odds of viral suppression by single-tablet regimens, multiple-
tablet regimens, and adherence level in HIV/AIDS patients receiving antiretroviral therapy.
Pharmacotherapy. 2017;37(2):204–213.
[38] Cambiano V, Lampe FC, Rodger AJ, et al. Use of a prescription-based measure of antiretroviral
therapy adherence to predict viral rebound in HIV-infected individuals with viral suppression. HIV
Med. 2010;11(3):216–224.
[39] Nachega JB, Hislop M, Dowdy DW, et al. Adherence to nonnucleoside reverse transcriptase
inhibitor-based HIV therapy and virologic outcomes. Ann Intern Med. 2007;146(8):564–573.
[40] King MS, Brun SC, Kempf DJ. Relationship between adherence and the development of resistance in
antiretroviral-naive, HIV-1-infected patients receiving lopinavir/ritonavir or nelfinavir. J Infect Dis.
2005;191(12):2046–2052.
[41] Maggiolo F, Airoldi M, Kleinloog HD, et al. Effect of adherence to HAART on virologic outcome and
on the selection of resistance-conferring mutations in NNRTI- or PI-treated patients. HIV Clin Trials.
2007;8(5):282–292.
[42] von Wyl V, Klimkait T, Yerly S, et al. Adherence as a predictor of the development of class-specific
resistance mutations: the Swiss HIV Cohort Study. PLoS One. 2013;8(10):e77691.
[43] Lafeuille M, Woodruff K, Dunn K, et al. Adherence to antiretrovirals in Medicaid-insured patients

t
ip
living with human immunodeficiency virus. J Manag Care Pharm. 2017;23 (Suppl 3-a):S26.
[44] Gonzalez-Serna A, Chan K, Yip B, et al. Temporal trends in the discontinuation of first-line

cr
antiretroviral therapy. J Antimicrob Chemother. 2014;69(8):2202–2209.
[45] Hughes AJ, Mattson CL, Scheer S, et al. Discontinuation of antiretroviral therapy among adults

us
receiving HIV care in the United States. J Acquir Immune Defic Syndr. 2014;66(1):80–89.
[46] Capetti A, Cossu MV, Rizzardini G. Darunavir/cobicistat for the treatment of HIV-1: a new era for
compact drugs with high genetic barrier to resistance. Expert Opin Pharmacother.
2015;16(17):2689–2702.
an
M
Figure legends
Figure 1. Sample selection and attrition flow diagram. ARV, antiretroviral; NRTI, nucleoside reverse
transcriptase inhibitors.
e d

Figure 2. First-line ARV regimens used by year.* *Total treatment-naïve population, n = 9,030;
treatment-naïve men, n = 7,463; treatment-naïve women, n = 1,387. ARV, antiretroviral.
pt
ce
Ac
Ac
ce
pt
ed
M
an
us
cr
ip
t
Ac
ce
pt
ed
M
an
us
cr
ip
t
Table 1. Baseline Demographic and Clinical Characteristics

Total Men Women

Characteristic
(n = 25,320) (n = 21,371) (n = 3,949)

Age, years, mean (SD) 45.3 (10.8) 45.3 (10.8) 45.2 (11.2)

Male sex, n (%) 21,371 (84.4) 21,371 (100.0) 0 (0.0)

Geographic region, n (%)

Northeast 8,320 (32.9) 6,878 (32.2) 1,442 (36.5)

t
ip
Midwest 3,290 (13.0) 2,837 (13.3) 453 (11.5)

cr
South 10,985 (43.4) 9,118 (42.7) 1,867 (47.3)

West 2,720 (10.7) 2,534 (11.9) 186 (4.7)

us
Other 5 (0.0) 4 (0.0) 1 (0.0)

Insurance type, n (%) an

Commercial 23,819 (94.1) 20,246 (94.7) 3,573 (90.5)
M
Medicare Advantage 1,501 (5.9) 1,125 (5.3) 376 (9.5)

Modified Quan–Charlson comorbidity

d

a
score category, n (%)
e

0 20,454 (80.8) 17,321 (81.1) 3,133 (79.3)

pt

1–2 3,936 (15.6) 3,300 (15.4) 636 (16.1)

ce

3–4 708 (2.8) 576 (2.7) 132 (3.3)

≥5 222 (0.9) 174 (0.8) 48 (1.2)

Ac

Top 10 AHRQ comorbid conditions, n (%)

Viral infection 23,102 (91.2) 19,478 (91.1) 3,624 (91.8)

Disorders of lipid metabolism 6,924 (27.4) 6,044 (28.3) 880 (22.3)

Hypertension 4,937 (19.5) 4,008 (18.8) 929 (23.5)

Respiratory infections 5,261 (20.8) 4,377 (20.5) 884 (22.4)

Diseases of the urinary system 3,408 (13.5) 2,771 (13.0) 637 (16.1)
 Diseases of the heart 3,441 (13.6) 2,808 (13.1) 633 (16.0)

Other lower respiratory disease 3,702 (14.6) 2,993 (14.0) 709 (18.0)

Other connective tissue disease 3,271 (12.9) 2,680 (12.5) 591 (15.0)

Other skin disorders 3,563 (14.1) 3,097 (14.5) 466 (11.8)

Spondylosis; intervertebral disc

2,791 (11.0) 2,286 (10.7) 505 (12.8)
disorders; other back problems

Treatment-naïve, n (%) 9,030 (35.7) 7,643 (35.8) 1,387 (35.1)

t
ip
AHRQ, Agency for Healthcare Research and Quality; SD, standard deviation.
a
Excluding HIV/AIDS.

cr
us
an
M
e d
pt
ce
Ac
Table 2. Regimens by Third Agent Class and Drug

Male patient Female patient

Total regimens
Drug class or name regimens regimens
a
(n = 38,663)
a a
(n = 32,065) (n = 6,598)

Third agent class, n (%)

Protease inhibitors 11,359 (29.4) 8,930 (27.9) 2,429 (36.8)

t
NNRTIs 16,294 (42.1) 13,930 (43.4) 2,364 (35.8)

ip
Multiple classes 6,517 (16.9) 5,414 (16.9) 1,103 (16.7)

cr
INSTIs 4,358 (11.3) 3,676 (11.5) 682 (10.3)

us
Entry inhibitors 125 (0.3) 110 (0.3) 15 (0.2)

Fusion inhibitors 10 (0.0) an 5 (0.0) 5 (0.1)

b,c
Major third agent, n (%)
M
Efavirenz 14,053 (36.3) 12,032 (37.5) 2,021 (30.6)

Raltegravir 6,962 (18.0) 5,786 (18.0) 1,176 (17.8)

d

Atazanavir 6,006 (15.5) 4,707 (14.7) 1,299 (19.7)

e

Darunavir 5,217 (13.5) 4,320 (13.5) 897 (13.6)

pt

Lopinavir 3,392 (8.8) 2,556 (8.0) 836 (12.7)

ce

Nevirapine 2,417 (6.3) 2,097 (6.5) 320 (4.8)

Ac

Etravirine 2,052 (5.3) 1,729 (5.4) 323 (4.9)

Rilpivirine 2,040 (5.3) 1,678 (5.2) 362 (5.5)

INSTI, integrase strand transfer inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; NRTI, nucleoside/

nucleotide reverse transcriptase inhibitor.

a
Data are presented at the regimen level; numbers given are regimen counts, not patient counts. Regimens that were

categorized as non-recommended according to US Department of Health and Human Services guidelines are not

included (n = 408).
b
Includes only drugs used in > 5% of total regimens.
c
As of the March 30, 2018 DHHS guideline update, only third agents in the INSTI class (dolutegravir, elvitegravir, and

raltegravir) are recommended for initial therapy.

Table 3. Follow-up Antiretroviral Therapy Adherencea

Total Male Female

Metric

t
(n = 25,320) (n = 21,371) (n = 3,949)

ip
PDC1

cr
PDC1, mean (SD) 0.85 (0.23) 0.86 (0.22) 0.80 (0.26)

us
PDC1 < 0.80, n (%) 6,359 (25.1) 5,014 (23.5) 1,345 (34.1)

0.80 ≤ PDC1 < 0.95 5,396 (21.3) an 4,539 (21.2) 857 (21.7)

PDC1 ≥ 0.95, n (%) 13,565 (53.6) 11,818 (55.3) 1,747 (44.2)

M
PDC2

b
PDC2, mean (SD) 0.82 (0.23) 0.83 (0.23) 0.76 (0.76)
d

PDC2 < 0.80, n (%) 7,408 (29.4) 5,857 (27.5) 1,551 (39.5)
e

0.80 ≤ PDC2 < 0.95 6,598 (26.2) 5,561 (26.1) 1,037 (26.4)
pt

PDC2 ≥ 0.95, n (%) 11,190 (44.4) 9,853 (46.3) 1,337 (34.1)

ce

MPR
Ac

c
MPR, mean (SD) 0.88 (0.18) 0.89 (0.18) 0.84 (0.22)

MPR < 0.80, n (%) 4,935 (20.4) 3,899 (19.1) 1,036 (27.8)

0.80 ≤ MPR < 0.95 5,238 (21.7) 4,388 (21.5) 850 (22.8)

MPR ≥ 0.95, n (%) 14,002 (57.9) 12,157 (59.5) 1,845 (49.5)

ARV, antiretroviral; MPR, medication possession ratio; PDC1, proportion of days covered for any ARV; PDC2,

proportion of days covered for regimens with at least 2 ARVs; SD, standard deviation.
a
PDC1, PDC2, and MPR calculations were based on all ARV medications except for boosting agents (ritonavir and

cobicistat).
b
PDC2 calculations included only regimens with at least 2 medications overlapping by at least 1 day; total n = 25,196

(21,271 men and 3,925 women).

c
MPR calculations included only regimens with at least 2 fills; total n = 24,175 (20,444 men and 3,731 women).

t
ip
cr
us
an
M
e d
pt
ce
Ac