practical approaches to evidence-based evaluation practice in public health
DESCRIPTION
Practical Approaches to Evidence-Based Evaluation Practice in Public Health. Joseph Telfair, DrPH, MSW/MPH Professor Department of Public Health School of Health and Human Performance University of North Carolina at Greensboro Greensboro, NC (USA) j [email protected] ♦ (336) 334 - 3240. - PowerPoint PPT PresentationTRANSCRIPT
Practical Approaches to Evidence-Based Evaluation Practice in Public Health
Joseph Telfair, DrPH, MSW/MPHProfessor
Department of Public HealthSchool of Health and Human Performance University of North Carolina at Greensboro
Greensboro, NC (USA)[email protected] ♦ (336) 334 - 3240
OVERVIEW OF PRESENTATIONBest Practices/Evidence: MCHB
Perspective Setting the Stage: Why Important,
Definitions and Key ConceptsPerformance Measurement: Selecting and
Constructing MeasuresProcess Monitoring: Developing a
Monitoring SystemConcluding RemarksQuestions and Discussion
Tell me ....I Forget
Show me....I remember
Involve me....I understand
Chinese Proverb
Of RelevanceThe MCHB has developed key strategies that
are the broad, cross-cutting approaches the Bureau uses in order to reach its five-year (and beyond) goals in the Bureau Strategic Plan. Goal 4 of the Strategic Plan is:“Improve the Health Infrastructure and Systems
of Care.” One key strategy used to support this goal is:
“Using the best available evidence, develop and promote guidelines and practices that improve services and systems of care.”
Best Practices/EvidenceMCHB Perspective
(http://mchb.hrsa.gov/about/stratplan03-07.htm)
Best Practices/Evidence (1)MCHB/AMCHP defines “best practices”
as a continuum of practices, programs and policies ranging from promising to evidence-based to science-based
EVALUATION of best practices requires the identification and establishment of evidence
Evaluating Evidence Evidence can be evaluated in four categories
ResearchExpert OpinionField LessonsTheoretical Rationale
All best practice approaches reported have a strong conceptual/theoretical rationaleHowever, the strength of evidence from
research, expert opinion and field lessons fall within a spectrum
Strength of Evidence Spectrum
Promising
Best Practice Approaches
Research +
Expert Opinion +
Field Lessons +
Theoretical Rationale +++
Proven
Best Practice Approaches
Research +++
Expert Opinion +++
Field Lessons +++
Theoretical Rationale +++
Strength of Evidence Spectrum
Promising
Best Practice Approaches
Little researchA beginning of agreement
in expert opinionVery few field lessons
evaluating effectiveness
Proven
Best Practice Approaches
Supported by strong research
Extensive expert opinion from multiple authoritative sources
Solid field lessons evaluating effectiveness
Grading Evidence (1)
Research/Evaluation• + A few studies in public health reporting
effectiveness (Promising)• ++ Descriptive review of scientific literature
supporting effectiveness (Promising/Proven)• +++ Systematic review of scientific literature
supporting effectiveness (Proven)
Grading Evidence (2)
Expert Opinion• + An expert group or general professional
opinion supporting the practice (Promising)• ++ One authoritative source (such as a national
organization or agency) supporting the practice (Promising/Proven)
• +++ Multiple authoritative sources (including national organizations, agencies or initiatives) supporting the practice (Proven)
Grading Evidence (3)
Field Lessons/Promising Practices• + Successes in state practices reported
without evaluation documenting effectiveness (promising)
• ++ Evaluation by a few states separately documenting effectiveness
(promising/proven)• +++ Cluster evaluation of several states (group
evaluation) documenting effectiveness (proven)
Grading Evidence (4)
• Practice-based Conceptual/Theoretical Rationale
• +++ Only practices which are linked by strong causal reasoning to the desired outcome of improving health and total well-being of priority populations will be reported (proven)
Best Practices/Evidence (2)MCHB has established that family and
community participation and engagement are key to the development of effective, quality health systems and services
Testing of Best Practices to Build Evidence – Deduction to Verification to Induction - Repeats
Requires a Practical Approach to Evaluation
Setting the Stage
WHY? (1)Four Primary reasons:
To develop and maintain an effective program and service delivery
process at the state and local levelTo enhance staff’s understanding of the factors
that contribute to the extent to which and in what ways the specific aimsprogram service targetsevaluation objectives
are being followed
WHY? (2)
Four Primary reasons (cont):To assure staff and stakeholders
by putting in place a process for determining whether or not the program and service delivery activities are succeeding as planned
WHY? (3)Four Primary reasons (cont):
To Build best practices data by Assessing the application of ‘the best available evidence’ from (MCHB modified) (4 levels): Existing Research/EvaluationExpert OpinionField Lessons/Promising PracticesPractice-based
Conceptual/Theoretical Rationale
Definitions and Key Concepts (1)Definition: “Evaluation or program
measurement (PM) is a systematic process for staff and institutions to obtain information on the service delivery process, its outcomes, and the effectiveness of its work, so that they can improve the process and describe its accomplishments” Mattessich, PW (2003) (p. 3) [modified]
Definitions and Key Concepts (2)
Definition: Program monitoring is the process of assessing progress toward achievement of a service delivery process’s objectives to determine whether the process was implemented as planned (Peoples-Sheps & Telfair (2005 – See Handout)
Definitions and Key Concepts (3)
Evaluation or PM involves a comparison of the staff’s planned processes and outcomes with selected standards in order to assess accomplishments
Evaluation or PM involves the application of social science methods to determine whether assessed efforts are the cause of observed results
Definitions and Key Concepts (4)
Evaluation or PM relies on both qualitative and quantitative methods, and often a triangulation of the two, to produce informative results
Definitions and Key Concepts (5)
Program monitoring is carried out by assessing the extent to which a program is implemented as designed that involves tracking progress toward achievement of a program’s objectives (Peoples-Sheps & Telfair (2005)
It is a very traditional form of assessment that is generally considered an administrative function and integral to the ongoing operations of every program (Kettner, et al., 1999).
Definitions and Key Concepts (6)Definition: A performance measure is a
specific, quantitative or qualitative representation (measure) of a capacity, process, or outcome deemed relevant to the assessment of program performance (Peoples-Sheps & Telfair (2005))
Definitions and Key Concepts (7)
Both program monitoring and performance measures depend on strong, meaningful
measures of program and service delivery process performance
PRACTICE EXERCISE
Questions 1- 4
Performance Measurement
Selecting or Constructing Measures (1)
Deciding what to measure is an essential first step
The aspects the service delivery process that are measured attract attention and generate action (Hatry, 1999).
Conversely, aspects not measured may go unnoticed until a crisis brings them to the surface (e.g., discovery of inadequate data collection efforts that did not allow for population or service targets to be met)
Selecting or Constructing Measures (2)
If the staff takes the time to think through what is needed, they are much less likely to miss something important
To cover all of the bases, start with the monitoring and evaluation’s specific aim(s) or hypothesis(es) to identify the main program and service delivery efforts and expected outcomes
Selecting or Constructing Measures (3)
To construct performance measures, three tasks must be undertaken: identifying concepts to be measuredselecting or constructing measureslocating or developing data sources
Selecting or Constructing Measures (4)
Performance measures can be formulated in many different ways. They may be:numbers (number of TB deaths)rates (TB mortality rate)proportions or percentages
(percentage of days missed at work among person with TB)
Selecting or Constructing Measures (5)
Performance measures can be formulated in many different ways. They may be (cont):averages (average number of emergency
department visits per person 18 to 44 years of age in a given year)
Categories (team meetings held)Numbers, percentages, and rates are the
most frequently used in MCH
Selecting or Constructing Measures (6)
Numbers, percentages, and rates are the most frequently used in MCH
Least used, but often just as critical are Qualitative indicators such as consensus measures, aggregated (agreement/ disagreement) statements, archival text-based descriptors (e.g., policy statements and group opinions from advisor or consumer groups
Selecting or Constructing Measures (7)
It is often helpful to include numbers and qualitative indicators along with rates and percentages so that the latter measures can be understood in the context of the type of service focus for which they were derived
To select or develop high-quality performance measures, candidate measures are generally assessed according to criteria that represent both scientific rigor and practical relevance
Selecting or Constructing Measures (8)
Responsive measures are able to detect a change
Measures need to be understandable to the audience to whom they will be presented
Regardless of how it is formulated, a measure should have very precise wording, a specific timeframe, and a clearly defined research population (e.g., persons with TB - Quant) or set of tasks (e.g., steps for securing needed sample - Qual)
Selecting or Constructing Measures (9)
A performance measure should be meaningful, valid, reliable, responsive, and understandable and should allow for risk adjustments (errors)
Selecting or Constructing Measures (10)
A valid measure is one that measures what it intends to measure. Validity, like all of the qualities in
this list, is measured on a continuum, meaning that some measures have greater validity than others
Selecting or Constructing Measures (11)
Reliable performance measures can be reproduced regardless of who collects the data or when they are collected (assuming the true results have not changed)Like validity, reliability is viewed as
a continuum
Selecting or Constructing Measures (12)
The selection of measures is closely tied to the data or research project information available to construct them
Data or information sources shouldBe of high quality, with standardized
definitions (as defined and agreed upon by the research team) and data collection methods and
Have acceptable levels of validity and reliability on the items of interest
Selecting or Constructing Measures (13)
Data or information sources should (cont)Be available within the program service
delivery timeframe (e.g., 3 years)Have cost conforming to budgetary
constraints of the programIt is more efficient, but not essential, to
construct measures from existing, or secondary, data sources, rather than to collect new data specifically for a given set of performance measures
PROCESS MONITORING
Source: Mattessich, PW (2003). p. 10
Developing a Monitoring System (1)
Development of a monitoring system is an essential component of program and service delivery process measurement plan
The monitoring process described in this presentationidentifies the program’s objectives the base from which formulas to
measure progress are developed
Developing a Monitoring System (2)
The monitoring process described in this presentation (cont)relative strength or emphasis of a
measure is assigned as necessary data collection plans are developedachievement scores are calculated at
predetermined intervals
Developing a Monitoring System (3)
Start with the Aim-linked Objectives The objectives of a Specific Aim, each of
which consists of a performance measure and a target, serve as the foundation for project monitoring
Fully developed, measurable objectives must correspond with the program or service purpose
Performance measures must be developed as the program is being planned
Developing a Monitoring System (4)
Each objective should have an explicit date by which the target is to be achieved (see example next slide)
With objectives clearly and precisely stated, the next challenge is to develop a system through which progress towards meeting the program’s targets can be monitored
Performance Measure Target
Percentage of adults in village 2 by desired gender and age within normal range
A 7% increase over baseline (estimated at 80%)
Average amount of time spent collecting staff comments per week by program assistants
Four hours
Number adults from Village 2 in the project shuttled to and from the city for the purpose of data gathering
Thirty adults sampled 80% of the allocated study days per month
Developing a Monitoring System (5)
The information derived from monitoring shows which program objectives need more attention in the future and whether any of them require less intensive work
If the process has fallen short on some objectives, this information should trigger an in-depth search for the reasons expected targets were not achieved
Developing a Monitoring System (6)
The Table on the slide to come shows the components of a monitoring system
The first two columns are identical to those in the previous slide showing performance measures and targets
Developing a Monitoring System (7)
The remaining three columns represent the basic elements of a monitoring system, as it builds on the program’s Specific Aims linked objectives
See Expanded Matrix (Handout)
Performance Measure Target Formula to Measure Progress
Results at End of Year 1
Achievement Score
Percentage of adults in village 2 by desired gender and age within normal range
A 7% increase over baseline (estimated at 80%)
Percentage over baseline with BMI within normal range
7
1.75% 0.25
Number adults from Village 2 in the project shuttled to and from the city for the purpose of data gathering
Thirty adults sampled 80% of the allocated study days per month
Number of adults
sampled 80% of study days
30
24 0.80
Average amount of time spent collecting comments per week by program assistants
Four hours Number of hours spent in collecting comments
4
3.2 hours 0.80
Developing a Monitoring System (8)
FormulasThe first step in developing a
monitoring system is to construct formulas to reflect progress toward achievement of the objectives’ targets.
The formula is based on the principle that a score of 1.00 is complete accomplishment
Developing a Monitoring System (9)
Formulas (cont)For example, A score of 0.99 or lower signifies that the performance measure fell short of the target; a score that exceeds 1.00 indicates greater than expected achievement
Developing a Monitoring System (10)
Formulas (cont)Three types of formulas can serve this
purposeWhen the target is a percentage, proportion,
or a simple count, the most informative and frequently used formula involves dividing the level of actual achievement at a specified time with the level given in the target -
Actual value
Targeted value
Developing a Monitoring System (11)
Data collection planThe first three columns of the
previous Table should be completed with the project’s initial plan.
To create a fully operational monitoring system, one more step is required:data items and sources necessary
to construct performance measures should be identified
Developing a Monitoring System (12)
This step should not be missed even if some data sources seem obvious since it is far too common to discover that researchers had incorrectly assumed the necessary data would be available and accessible when needed
Developing a Monitoring System (13)
Analyses/Interpretation of ResultsThe information derived from
monitoring shows which objectives need more attention in subsequent years and whether any of them require less intensive work
Adjustments in resource allocations can be based on the needs of specific objectives for more or less effort
Developing a Monitoring System (14)
Analyses/Interpretation of Results (cont)Careful assessment of the reasons for
shortfalls on objectives should be conducted before any reallocation decisions are made.
A review of end of year achievement scores provides helpful information for further investigation and subsequent adjustments to the process
PRACTICE EXERCISE
Questions 5 - 9
IN CONCLUSION
IN CONCLUSION (1)Service Programs may not reach their
targets for a number of reasons A primary reason is inadequate
resources, which may take the form of insufficient funds across the board or misallocation of funds across Specific Aims linked objectives
It may be possible to detect misallocation if some targets are overachieved, whereas others fall short
IN CONCLUSION (2)Other commonly cited reasons why programs
may fall short in achieving objectives include a lack of adequate knowledge about feasible
target levelsexternal factors that make it difficult or
impossible to reach the target (e.g., inability to find or retain clients that meet the program criteria)
inaccurate measurement of the objectivea conceptual error in the program purpose
IN CONCLUSION (3)As an evaluation strategy, monitoring has
three important shortcomingsFirst, it does not produce evidence of
cause–effect relationships; only evaluation research can do that.
Second, the results of monitoring are limited to a single program; they cannot be extrapolated from one program to another
IN CONCLUSION (4)As an evaluation strategy, monitoring has
three important shortcomings (cont)Finally, there are no firm guidelines for
interpretation of the scoresAlthough a score of 0.70 might be
considered good and 0.90 might be superior, the most useful interpretations depend on the program’s context and purpose (Peoples-Sheps, Rogers, & Finerty, 2002).
IN CONCLUSION (5)Advantages and Disadvantages of
Monitoring as an Evaluation effortProgram monitoring is a valuable tool
for building and establishing evidenceProgram monitoring is a valuable tool
for planning and management decisions
The process is inexpensive and can be applied readily by anyone with entry-level training or experience
IN CONCLUSION (6)Advantages and Disadvantages of
Monitoring as an Evaluation effort (cont)It includes a flexible set of methods that
can be modified to accommodate the needs of each service program at both the state and local level
Monitoring requires staffs to develop objectives that serve as the basis of the service delivery process and then to plan for necessary data so that the capability for tracking progress is assured
IN CONCLUSION (7)
Another important advantage is that it encourages the production of information for critical management decisions, identifying and assessing Promising/Best Practices in both short- and long-term time frames and across all levels of the service delivery process
Thus, it is compatible with most governmental programmatic guidelines
“Just because you can quantify something, doesn’t mean you
understand it” (Aubel, 1993, p. 10)
“Not everything that counts can be counted and not everything
that be counted counts”Anonymous
QUESTIONSand
Discussion
References (1)Peoples-Sheps, M. D., Byars, E., Rogers, M. M.,
Finerty, E. J., & Farel, A. (2001). Setting objectives (revised). Chapel Hill, NC: The University of North Carolina at Chapel Hill.
Peoples-Sheps, M. D., & Telfair, J (2005). Maternal and Child Health Program Monitoring and Performance Appraisal in J. Kotch (ed). Maternal And Child Health: Programs, Problems And Policies In Public Health, 2nd. Edition (Chapter 16). Boston, MA: Jones & Bartlett Publishers
References (2)Grembowski, D. (2001). The practice of
health program evaluation. Thousand Oaks, CA: Sage Publications. Hatry, H. P. (1999). Performance measurement: Getting results. Washington, DC: The Urban Institute Press.
Kettner, P. M., Moroney, R. M., & Martin, L. L. (1999). Designing and managing programs: An effectiveness-based approach (2nd ed.). Thousand Oaks, CA: Sage Publications.
References (3)Durch, J. S., Bailey, L. A., & Stoto, M. A.
(Eds.). (1997). Improving health in the community: A role for performance monitoring. Washington, DC: National Academy Press.
Mattessich, PW (2003). The Manager’s Guide to Program Evaluation. Saint Paul, MN: Wilder Publishing Center
References (4)Roberts, A. R. & Yeager, K. (2004),
Evidence-based Practice Manual: Research and Outcome Measures in Health and Human Services, Oxford University Press
Aubel, J (1993), Participatory program Evaluation: A manual for involving program stakeholders in the evaluation process. The Gambia: Catholic Relief Services – USCC.
References (5)• Telfair, J., & Mulvihill, B.A. (2000),
Bridging science and practice: The integrated model of community-based evaluation. Journal of Community Practice, 7(3), 37-65.