Web-based Tool for Assessing and Improving the Usefulness of Community Health Assessments Michael A. Stoto, Georgetown University With Susan Straus, RAND; and Cate Bohn and Priti Irani, NYSDOH Assessment Initiative Conference Atlanta, GA, August 22, 2007
25
Embed
A web-based tool for assessing and improving the usefulness of community health assessments
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Web-based Tool for Assessing and Improving the Usefulness of
Community Health AssessmentsMichael A. Stoto, Georgetown University
With Susan Straus, RAND; andCate Bohn and Priti Irani, NYSDOH
Assessment Initiative Conference Atlanta, GA, August 22, 2007
Stoto 2007 AI Conference 2
Background
• Project goal: Develop a web-based tool to assess and improve the usefulness of Community Health Assessments (CHA) in New York State and partners
• Partners– New York State Department of Health
Public Health Information Group– RAND Health
• Supported by CDC Assessment Initiative
Stoto 2007 AI Conference 3
Project Tasks
• Literature review of studies assessing the CHA process to identify factors/criteria for a useful CHA (www.rand.org/publications/TR/TR314/)
• Build consensus among stakeholders on the definition of “usefulness” of CHAs
• Develop and pilot test a web-based tool to– Assess the usefulness of CHAs to users– Feed the results back to CHA developers
• Refine, roll out, and evaluate the web-based tool to assess the usefulness of CHAs
Stoto 2007 AI Conference 4
Development of Criteria
• From literature review, created list of 85 criteria and 2 open-ended questions
• Obtained feedback from NY and national experts and practitioners
• Reduced list to 32 close-ended criteria in 3 categories plus 3 open-ended questions– CHA content, format, and impact
• Conducted e-mail survey of 5 NY counties• Reduced to 21 close-ended criteria in same
3 categories, plus 3 open-ended questions
Presenter
Presentation Notes
Technical advisory panel including researchers, practitioners from other states, and CDC
Stoto 2007 AI Conference 5
CHA Document Content1. Clearly states the goals and purpose of the CHA2. Includes the most important aspects of the
community’s health3. Allows comparisons with data from other
communities or other appropriate benchmarks4. Allows comparisons over time 5. Presents data in meaningful subgroups of
population (e.g. to assess health disparities)6. Provides sufficient focus on positive
characteristics, e.g., community assets, as well as negative characteristics, e.g., death rates
7. Sufficiently documents the process and methods used to create the CHA
Presenter
Presentation Notes
Answered on a 4-point scale as follows: 1. Strongly disagree, 2. Somewhat disagree, 3. Somewhat agree, 4. Strongly agree. Changed to a 6-point scale for national roll-out.
Stoto 2007 AI Conference 6
CHA Document Format8. Uses a consistent format to present information on
different topics9. Includes both summary and detailed versions to be
useful for a variety of audience10. Is well organized; it is easy to find content 11. Is easy to understand 12. Clearly indicates the relationships among related
indicators 13. Includes narrative and graphic representation of key
findings to meet the needs of varying audiences 14. Uses a similar structure or data elements as other
community planning tools in use 15. a) Is available online; b) includes appropriate links 16. Can be reproduced easily by photocopy 17. Clearly identifies data sources
Presenter
Presentation Notes
#15 is yes/no for both parts
Stoto 2007 AI Conference 7
CHA Document Impact18. Serves as a resource to prioritize and plan
services19. Serves as a resource for writing grant
applications20. Serves as a resource to guide a comprehensive
health promotion strategy
Open Ended Questions21. Are there any indicators that are not useful?
Why? How could they be made useful?22. Are there any useful indicators that are not
included in the CHA? What are they? Why?23. How will you use information from the CHA?24. Comments about this web-based survey
Stoto 2007 AI Conference 8
Pilot Test Procedures
• Five NY “counties”– Cortland, Dutchess, Monroe, Schenectady – Tri-County area: Clinton, Essex, and Franklin
• Process– Counties identified actual and potential users– County representatives sent e-mail invitation– RAND sent follow up e-mail with password– County representatives sent follow-up e-mail
• Response rate: 39% (70/180)
Stoto 2007 AI Conference 9
CHA Document Content
Stoto 2007 AI Conference 10
Statistical properties• Scales are Reliable
– Coefficient α
indicates whether responses to items on a scale are conceptually related
• Users Find Survey Easy to Use– “This survey has been easy to complete. It is
well thought out and presented. I am looking forward to receiving feedback on our Community Health Assessment.”
– Very easy and user friendly. Easy to navigate and respond to the questions.”
– “One of the better ones I have done--ease to use and follow. Thanks.”
Stoto 2007 AI Conference 11
Average Score by Question (1/3)
CHA Document Content 3.551. Clearly states goals and purpose 3.552. Includes important aspects of health 3.723. Allows comparisons with appropriate 3.524. Allows comparisons over time. 3.645. Presents data in meaningful subgroups 3.556. Focus on positive characteristics 3.437. Documents process and methods 3.43
Presenter
Presentation Notes
Blue indicates dimensions on which the average county scored relatively well. Red indicates dimensions on which the average county scored relatively poorly.
Stoto 2007 AI Conference 12
Average Score by Question (2/3)
CHA Document Format 3.298. Uses a consistent format 3.759. Includes summary and detailed versions 3.4310. Well organized. 3.6911. Easy to understand. 3.6612. Indicates relationships among indicators 3.2113. Includes narrative and graphics 3.4614. Similar to other community planning tools 3.3816. Can be reproduced by photocopy 3.6517. Clearly identifies data sources 3.75
Stoto 2007 AI Conference 13
Average Score by Question (3/3)
Impact of the CHA Document 3.7018. Resource to prioritize and plan services 3.7319. Resource for writing grant applications 3.7520. Resource to guide health promotion 3.61
CHA Document Content 3.55CHA Document Format 3.29Impact of the CHA document 3.70
Average Score by Category
Presenter
Presentation Notes
Note that users gave CHAs good scores on impact, which really means “immediate changes” as in Priti’s report.
Stoto 2007 AI Conference 14
Average Score by County (NY Pilot)
2.75
3.00
3.25
3.50
3.75
4.00Q
1
Q2
Q3
Q4
Q5
Q6
Q7
Con
tent Q8
Q9
Q10
Q11
Q12
Q13
Q14
Q16
Q17
Form
at
Q18
Q19
Q20
Impa
ct
CHA Content CHA Format CHA Impact
Ave
rage
sco
re
Overall mean Cortland Dutchess Monroe Schenectady Tri county
Presenter
Presentation Notes
Note that users gave CHAs good scores on impact, which really means “immediate changes” as in Priti’s report.
Stoto 2007 AI Conference 15
Overall Results• Most useful aspects of CHA’s reviewed
Content2. Includes important aspects of health (3.72)
Format8. Uses a consistent format (3.75)
Impact18. Resource to prioritize and plan services (3.73) 19. Resource for writing grant applications (3.75)
• Higher ratings by those involved in preparing the CHA
• No or small differences by affiliation
Presenter
Presentation Notes
Overall results suggest aspects of the NY State required format that lead to useful CHAs Standardized differences focus on difference between county CHA where some may be more or less useful than others in certain respects.
Stoto 2007 AI Conference 16
Sample Open Ended ResponsesAre there any indicators in the CHA that are not useful? Which ones? Why? How could they be made useful? (21 responses)– “Everything in the CHA is quite useful”– DWI arrests and cirrhosis are poor measures of
substance abuse– Break down birth indicators by age of mother– Break out data for the City vs. rest of county– Include break down for all immigrant groups– Include community input data– Indicate whether change is due to change in health
status or better/worse reporting
Stoto 2007 AI Conference 17
Sample Open Ended ResponsesAre there any useful indicators that are not included in the CHA? What are they? Why? (24 responses)– “I found it to be entirely comprehensive”– Substance abuse and related behaviors– Communicable disease rates (HIV, STD, etc.) – Environmental health– Mental health, dementia– Domestic violence, sexual assault– Intendedness of pregnancy– Differentials in vulnerable populations– Lack of health insurance– Poverty, homelessness, crime
Stoto 2007 AI Conference 18
Sample Open Ended Responses
How will you use information from the CHA in developing, implementing, and evaluating your programs? (45 responses)– Planning (e.g. primary health care, services for
baby boomers, Consolidated Services Plan, Communities that Care, etc.)
– Prioritize goals and objectives– Advocacy (Coalition findings regarding youth)– Grant proposals (identify areas of unmet need,
state breast and cervical cancer grant)– Program management and improvement (MAPP)– Source of information for comparisons
Stoto 2007 AI Conference 19
Sample Feedback Report: High scoring CHA
2.75
3.00
3.25
3.50
3.75
4.00Q
1Q
2Q
3Q
4Q
5Q
6Q
7C
onte
nt Q8
Q9
Q10
Q11
Q12
Q13
Q14
Q16
Q17
Form
at
Q18
Q19
Q20
Impa
ct
CHA Content CHA Format CHA Impact
Ave
rage
sco
re
Overall mean Tri county
Presenter
Presentation Notes
Note that some strengths and weaknesses are common to all NY counties because of the common state format, while others are county specific. This format attempts to sort these out. Questions marked in red are significantly different from the state average.
Stoto 2007 AI Conference 20
Sample Feedback Report : Low scoring CHA
2.75
3.00
3.25
3.50
3.75
4.00Q
1Q
2Q
3Q
4Q
5Q
6Q
7C
onte
nt Q8
Q9
Q10
Q11
Q12
Q13
Q14
Q16
Q17
Form
at
Q18
Q19
Q20
Impa
ct
CHA Content CHA Format CHA Impact
Ave
rage
sco
re
Overall mean Dutchess
Presenter
Presentation Notes
No strong evidence that this county’s strengths and weaknesses are different from other counties.
Average results similar to pilot test, despite different states. No counties significantly different from average on any item (probably due to small sample size). Note that “impact” scores remain high.
Stoto 2007 AI Conference 24
National roll out results
• Increase in number of scale categories and loss in sample size operate in opposite directions
– Need larger samples of users!• “National” average similar to NY pilot after
adjusting for number of categories– Unexpected, given differences in state
CHA formats• No significant differences in usefulness
between counties in the national sample
Stoto 2007 AI Conference 25
Conclusions• Literature review and CHA user process developed
– Consensus criteria for CHA usefulness – Easy-to-use web-based tool to obtain feedback
• Users rate NY CHAs favorably– May have improved since 1998 critique
• More similarities across communities than differences– Lack of significant differences limits value of
feedback• Low participation rate
– Resulting small sample limits reliability– May reflect
• Logistical difficulties• Lack of interest in CHAs• Lack of a community involvement process