Professional Documents
Culture Documents
HEALTHCARE INFORMATION
MANAGEMENT
COURSE OBJECTIVES
At the end of this course, you’ll be able to:
1
5/13/2018
AGENDA
Introduction and important definitions Bias and confounding factors
Information management principles KPIs
Confidentiality of records / reports Tools for displaying data
Data collection and tool designing Quality improvement process tools
Introduction to statistics Run charts and SPC
Sampling techniques Additional quality improvement tools
Descriptive statistics Pop Quiz
Inferential statistics
INTRODUCTION…
DEFINITIONS
Information management:
2
5/13/2018
HEALTHCARE ANALYTICS
Consists of the systems, tools, and
techniques that help healthcare
organizations gain insight into current
performance, and guide future actions,
by identifying patterns and relationships
in data and using that understanding to
guide decision making.
First, no harm
INTRODUCTION… Doing the right thing right, 1st time and every time
QUALITY
Having high degree of excellence
3
5/13/2018
INSTITUTE OF MEDICINE
(IOM) DEFINED QUALITY AS
The extent to which health services provided to individuals
and patient populations improve desired health outcomes.
The care should be based on the strongest clinical evidence
and provided in technically and culturally competent manner
with good communication and shared decision making.
Appreciative Quality
ASPECTS OF
QUALITY (MAP) Perceptive Quality
Measurable Quality
4
5/13/2018
ASPECTS OF
QUALITY (MAP)
Measurable quality: is equal to
compliance with or adherence to
standards, practice guidelines, or
protocols (provider’s aspect of care).
INFORMATION MANAGEMENT
Quality must be defined in quantifiable terms to enable measurement,
monitoring, analysis, and most important, decision making and action.
5
5/13/2018
Wisdom
Understanding, Integrated, Actionable
Given Insight
Becomes
Knowledge
Contextual, Synthesized, Learning
Given Meaning
Becomes
Information
Useful, Organized, Structured
Given Context
Becomes
Data
Signals, Know-nothing
INFORMATION MANAGEMENT
GOAL
To support decision making to:
6
5/13/2018
Assure
Assure timely, easy access, accurate, comparable
data/information, balance with security
requirements.
Redesign
Redesign information-related processes for
efficiency and effectiveness then share it.
Education Research
7
5/13/2018
What is noticed
gets acted on
Accurate
Timely
Relevant
Analyzed
“
Ensure that the information that is being used for
decision making and QI has the following attributes:
Visualized
8
5/13/2018
INFORMATION MANAGEMENT
Identify current
Identify critical Define data
available data
information needs elements
sources
PROCESSES
Determine data Aggregate and
Acquire / collect data
collection plans display data
DECISION MAKING
Traditional model
In the past, without good data, we relied upon
opinion, logic, rationalization or hearsay to lead us
to recommendations and hopefully appropriate
action or improvement.
9
5/13/2018
CONFIDENTIALITY
It is a fact to recognize that patient file is a
facility property while the information in the
file is a patient property.
CONFIDENTIALITY
Written authorization is needed for any purpose
unrelated to treatment (PTO); payment,
treatment, operations.
✓ Patient name.
“
CEO, governing body, QPI responsible, and direct care
providers may use patient information without written
authorization.
release.
10
5/13/2018
DATA COLLECTION
11
5/13/2018
DATA COLLECTION
TOOLS
TYPES OF
facilitate interpretation directly
from the form, collecting data
in real time and at the location
DATA where the data is generated.
12
5/13/2018
KEEP PROVIDE
Keep the tool as short and simple as Provide appropriate definition of terms for
possible. using the tool.
CONSIDER INCLUDE
Consider computerizing whenever Include all data elements necessary to
possible. monitor the specified issue / indicator.
CHECKLIST FOR
ADEQUACY OF THE
TOOL
Does the tool really measure the
process or aspect of care and its
indicator?
13
5/13/2018
RELIABILITY
RELIABILITY
Types of reliability:
14
5/13/2018
VALIDITY
Types of reliability:
“
it with another measure or procedure that has
been demonstrated to be valid. It has different
types; predictive, concurrent, convergent and
discriminant. It’s the degree to which an instrument
measures what it is intended to measure.
VALIDITY
Types Of Validity:
“
test provides a significant improvement in
addition to the use of another approach. A test
has incremental validity if it helps more than
not using it. It’s the degree to which an instrument
measures what it is intended to measure.
15
5/13/2018
STATISTICS
BASIC TERMS
Population: A population describes a large set Sample: A portion of the population selected for
or collection of items that have something in study is referred to as a sample. Representative
common. It consists of all elements being sample contains the characteristics of the
studied. (N) population as closely as possible. (n)
Statistic: is a descriptive measure derived from Parameter: is a descriptive measure derived from
a sample. a population.
16
5/13/2018
BASIC TERMS
Data set: is a collection of observation on a Bias: is the errors related to the ways targeted
variable. and sampled population differ.
17
5/13/2018
http://www.macorr.com/sample-size-calculator.htm
18
5/13/2018
SAMPLING
Probability Sampling
Simple Random
Systematic
Systematic
Cluster
Non-Probability Sampling
Purposive
Expert
Consecutive
Convenience
Quota
Snowball
PROBABILITY
SAMPLING
Simple Random Sample: Every element in
the population has the same probability of
being chosen for sampling as all other
elements in the population.
19
5/13/2018
PROBABILITY
SAMPLING
Systematic Sampling: divide population into
groups of K individuals: K= N/n. Randomly
select one, then select every Kth.
NON-PROBABILITY SAMPLES
(INTENTIONALLY BIASED)
Purposive or judgmental sampling: it involves hand-picking from the accessible population those individuals judged
most appropriate for the study. Means that selection will be based on a desired characteristic. E.g. All patients above
age of 60 with THR.
Expert sampling: a type of purposive sampling in which experts in a giving area selected because of their access to
relevant information. Used with Delphi technique.
Consecutive sampling: it involves taking every patient who meets the selection criteria over a specified time interval or
number of patients. It’s the best of non-probability techniques and one that is very often practical.
20
5/13/2018
NON-PROBABILITY SAMPLES
(INTENTIONALLY BIASED)
Convenience sampling: it’s the process of taking those members of the accessible population who are easily available.
Example: all patient seen in ER in a given week (it’s widely used in clinical research)
Snowball sampling: it’s a subtype of convenience sampling involve subjects suggesting other subjects for inclusion in the
study. E.g. Drug addicts.
TYPES OF DATA
Data
21
5/13/2018
TYPES OF DATA
CATEGORICAL DATA
Nominal
Ordinal
22
5/13/2018
QUANTITATIVE DATA
Discrete Variable
Continuous Variable
TYPES OF DATA
Categorical
- Nominal Binary (Male / Female)
Nominal (Unordered) - Nominal Multi-category (Single,
married, divorced)
Ordinal (Ordered) Cancer stage I, II, III, IV
Quantitative
23
5/13/2018
QUANTITATIVE DATA
Although quantitative data has more statistical
power, It can be converted into categorical
data which is easy to analyze.
DESCRIPTIVE
STATISTICS
24
5/13/2018
Categorical Quantitative
Mode, Frequency
25
5/13/2018
CATEGORICAL DATA
Mode
The mode of a set of values is that value which occurs
most frequently.
Frequency
The number in each category
▪Near miss = 2
▪Minor event = 5
▪Major event = 3
26
5/13/2018
Categorical Quantitative
Mode, Frequency
NORMAL DISTRIBUTION
“Bell Curve”
Data can be distributed in different ways, but continuous data tends to be around a
central value with no bias left or right, and it gets close to a “Normal distribution”
27
5/13/2018
NORMAL DISTRIBUTION
NORMALITY TEST
Use Minitab
P value Ho HA Distribution
28
5/13/2018
DISTRIBUTED DATA
Mean
NORMALLY
The most commonly used measure of location.
Example:
29
5/13/2018
DISTRIBUTED DATA
Standard Deviation:
The most commonly used measure of spread.
NORMALLY
The SD is a measure of the spread of data about their
mean.
Suppose the exam score for 4 students are 82, 95, 67, and 92. Calculate SD
2 2
Values X Mean X X-X (X – X) ∑(X – X) / n-1 √
95 84 11 121
67 84 - 17 289
92 84 8 64
∑ = 478
30
5/13/2018
Categorical Quantitative
Mode, Frequency
Median
The median is the middle number after they have been
NON-NORMALLY
sorted in ascending or descending order.
31
5/13/2018
Median
E.g. what’s the median of the following?
NON-NORMALLY
5, 22, 13, 16, 24, 12, 20, 8, 10
Sort: 5, 8, 10, 12, 13, 16, 20, 22, 24
Median= 13
Range:
Very simple, very rough, and the weakest
E.g.
MI patients weight: 53, 47, 60, 74, 66
Range = 74 – 47 = 27
32
5/13/2018
Interquartile Range:
DISTRIBUTED DATA
NON-NORMALLY
It’s the difference between the 75th and 25th percentiles.
IQR = Q3 – Q1
62 63 64 64 70 72 76 77 81 81
Median =
Q1 = 64
Q3 = 77
IQR =
Measures of Central
Tendency
Mean
Mode
Median
Measures of Dispersion
Standard Deviation
Range
Variance
Interquartile Range
Coefficient of variation
(CV= SD/Mean%)
33
5/13/2018
Measures of Position
Quartiles
Deciles
Percentiles
Ordering (Organizing)
Proportion
Ratio
MEASURES OF POSITION
Quartiles:
It’s the value which divides the distribution into 4 parts subsets of equal size, each comprising 25%
of the observations
E.g.: Weight of patients (165, 250, 140, 155, 160, 170, 180, 200, 200, 225, 230, 260)
Q1 Q2 Q3 Q4
=(160+165)/2=162.52 =(180+200)/2=190 =(225+230)/2=227.5
34
5/13/2018
MEASURES OF POSITION
Deciles:
The values which divide the data into 10 equal parts are
called deciles and are denoted by D1, D2, …. , D9, D10
Percentiles
The values dividing the data into 100 parts are called
percentiles and are denoted by P1, P2, P3, …. P100. It’s the
most common used among measures of position.
E.g.: 70th percentile (P70) is the value such that 70% of the
observations are less or equal to.
ORDERING (ORGANIZING)
Proportion:
Is the number of observations with a given characteristic, “Part of a whole”
e.g. in hydrocortisone group, the proportion of died patients = P = a/a+b =
50/200 = 0.25
Percentage:
Is a proportion multiplied by 100%, e.g. in hydrocortisone group, the
percentage of died patients = (a/a+b) X 100 = 25%
35
5/13/2018
ORDERING (ORGANIZING)
Rate:
Hydrocortisone Group Placebo Group
Died 50 (a) 80 (c)
Survived 150 (b) 122 (d)
Total 200 202
ORDERING (ORGANIZING)
Ratio
Is the number of observations in a given group with a given
characteristic, “ 2 different groups or population” e.g. in
hydrocortisone group, the ratio of died patients to survived
= a/b = 50/150 = 0.33
Hydrocortisone Group Placebo Group
Died 50 (a) 80 (c)
Survived 150 (b) 122 (d)
Total 200 202
36
5/13/2018
INFERENTIAL
STATISTICS
INFERENTIAL STATISTICS
Generalization
Inference
37
5/13/2018
INFERENTIAL STATISTICS
There are 2 main methods used in inferential statistics:
Estimation: using CIs
Hypothesis Testing: Null hypothesis (HO) vs. Alternative
hypothesis (HA)
38
5/13/2018
ESTIMATION
➢ Confidence level = 95%
➢ Confidence interval (CI)= 5
✓ We are 95% certain that the true population
mean is something between (25 – 35)
Hypothesis Testing:
Is a branch of statistics that specifically
determines:
1- whether a particular value of interest is
contained within calculated range (CI).
2- whether you conclude that 2 things are
the same or different.
3- if 2 means are equal (or mean = target).
4- if differences you observe are likely due to
true differences in the underlying
populations or random variation.
39
5/13/2018
HYPOTHESIS TESTING
Null Hypothesis: HO
Is a statement being tested to determine whether
or not it is true.
HO: ϻ1 = ϻ2 or ϻ1-ϻ2 = 0
It means that the means from 2 sets of data are
the same.
Alternative Hypothesis: HA
Is a statement that represents reality if there’s
enough evidence to reject Ho.
HA: ϻ1 ≠ ϻ2 or ϻ1-ϻ2 ≠ 0
It means that the means from 2 sets of data are
not the same.
40
5/13/2018
HYPOTHESIS TESTING
Statistical Significance:
Group A: Mean = 20
Group B: Mean = 40
Is the difference real or could be due to chance?
41
5/13/2018
P-Value Ho HA Difference
Conclusion Drawn
42
5/13/2018
P-Value
If we reject Ho, the P-value is the probability of being wrong.
“The P-value is the probability of
making a Type I error”
BIAS &
CONFOUNDING FACTORS
IN
INFERENTIAL STATISTICS
43
5/13/2018
Bias types:
1- Selection
2- Performance
3- Observation
Types of Bias:
Selection Bias: occurs through the
identification and/or recruitment
of an unrepresentative sample
population, such that any drawn
conclusions cannot be generalized
to the target population.
44
5/13/2018
45
5/13/2018
PERFORMANCE MEASURES
&
KEY PERFORMANCE
INDICATORS (KPIs)
46
5/13/2018
KEY
a process. By themselves, indicators do not directly measure quality.
PERFORMANCE performance that are the most critical for the current and future
success of the organization.
Types:
1- Structure: e.g. % licensed
physicians.
47
5/13/2018
Prevalence:
To meet
customer needs
& expectations
To improve To manage the
the process organization
48
5/13/2018
IMPROVEMENT
“
If you can’t measure it,
you can’t improve it.
Lord Kelvin
Page 97
FOCUS:
WHAT GETS
MEASURED GETS
DONE
“Without measurement,
Sustained outcomes are
unlikely.”
Page 98
49
5/13/2018
CLARITY
Page 99
You have to
keep your eye
on the ball
or risk
getting hit!
50
5/13/2018
Target setting
Selection Documentation
Planning Implementation
1- Planning
Identify the need / rational
Form a team
Identify measurement
Objective
Clear definitions
102
51
5/13/2018
2- KPI Selection
Requirements: regulatory, accreditation.
Selection Criteria:
✓ Relevant: Finding meaningful KPIs for what you
want to measure.
✓ Clearly defined: Concise, precise.
✓ Balanced: More than 1 KPI to reduce risk of
negative behavior.
52
5/13/2018
Do not overburden
your self with
MEASURES
3- KPI
Target setting Best practices:
benchmarking
History
53
5/13/2018
Non achievable
Target
✓ Challenging
✓ Achievable (Not impossible)
✓ Non conflicting
Target
Demotivating
Effort
Generate positive behavior
Frustrating
Employees
107
BENCHMARKING
Types of benchmarking:
Risk adjustment:
Technique used to control the fact that
different patients with the same diagnosis
may have additional conditions that can
affect how they respond to treatment
(comparing the outcome without risk
adjustment ca be misleading)
“
It’s the comparison of an organization’s or
practitioner’s results against a proper
reference point (Best practice) which give
shape for the shapeless raw data.
Healthcare Information Management
54
5/13/2018
4- Documentation:
109
General
Category Clinical
Standard(s)
Type
Measuring unit %
Target population
Sample size
All patients with prescribed medications
100%
34% Target / Threshold
Reporting
31% Unique identification code
Reporting time 11th working day
Data aggregation & analysis Intital analysis by thePharmacy and the final analysis by Quality department.
Revision date
110
55
5/13/2018
5- Implementation
Data analysis
Information reporting
Action plan
111
DATA COLLECTION
56
5/13/2018
DATA ANALYSIS
INFORMATION REPORTING
57
5/13/2018
ACTION PLAN
Attach
Measurement
to Coaching
58
5/13/2018
KPI Scorecard
3. Performance measurement
Target Dashboard
Program
Portfolio of
Initiative/ Project initiatives
4. Performance improvement
Milestone Initiative
description form
Task
118
59
5/13/2018
1. Scorecard
2. Dashboard
11
9
SCORECARD
12
0
60
5/13/2018
SCORECARD - SPORTS
12
1
SCORECARD - SPORTS
12
2
61
5/13/2018
STRUCTURE OF SCORECARD
12
3
STRUCTURE OF SCORECARD
Strategic
objectives
12
4
62
5/13/2018
STRUCTURE OF SCORECARD
Strategic
objectives
KPIs
12
5
STRUCTURE OF SCORECARD
Strategic
objectives
KPIs
Targets
12
6
63
5/13/2018
STRUCTURE OF SCORECARD
Level of results
Strategic
objectives
KPIs
Targets
12
7
Analyze
12
8
64
5/13/2018
Analyze
12
9
Analyze
• Country Scorecards annually
13
0
65
5/13/2018
SCORECARD
• Provide performance
measures in relation to
strategic plan (mission, vision,
goals & objectives)
DASHBOARD
13
2
66
5/13/2018
DASHBOARD
13
3
DASHBOARD
• Better decisions on a
daily / monthly basis.
13
4
67
5/13/2018
DASHBOARD
• Data verification
• Data analysis
• Decision making
• Action
• Feedback
13
5
SCORECARDS
• Reflect on strategy
• Indicate level of performance in
reaching the strategy
• Monitor objectives
DASHBOARDS
• Operational tools
• Indicate the status of key processes
and Key areas of interest
• Monitor processes
13
6
68