Appendix E: Technical Expert Panel Meeting Summary

Design and Evaluation of Three Administration on Aging (AoA) Programs: Chronic Disease Self-Management Program Evaluation Design—Final Evaluation Design Report (continued)

Contract #: HHSA290200710071T

CDSMP Evaluation Design Technical Expert Panel (TEP) Conference Call
March 30, 2011, 2pm—4pm


William Spector, Project Officer

Jennifer Klocinski, Aging Services Program Specialist
Saadia Greenberg, Director, Office of Performance and Evaluation
Jane Tilly, Health/Dementia Team Lead
Susan Jenkins, Social Science Analyst, Office of Performance and Evaluation

IMPAQ International
Daver Kahvecioglu, Project Director and Co-PI
Marios Michaelides, Research Associate
Sarah Ruiz, Research Associate
Leanne Clark, Gerontology Research Consultant
Carolyn Corea, Research Analyst

Abt Associates
Terry Moore, Co-PI
Rosanna Bertrand, Associate
Johanna Hudgens

Technical Expert Panel Members
Lynda Anderson, PhD, Director, Healthy Aging Programs, CDC—Atlanta, Georgia
Candace Goehring, MN RN, Unit Manager, Home and Community Programs, Aging and Disability Services Administration—Washington
Diana Scully, MSW, Director, Office of Elder Services—Maine


Opening Remarks by Administration on Aging

  • OMB and Congress are very interested in how CDSMP has actually worked in the aging network. OMB expects an experimental evaluation.
  • Despite a lot of research done on CDSMPs, there are some gaps in areas of: 1) the effect of the program on older adults, 2) the effect of the program as it is implemented by the aging network, 3) the effect of the program on healthcare utilization and cost.

TEP Member Introductions and Project Overview by Daver Kahvecioglu

  • Daver Kahvecioglu, from IMPAQ international, gave a brief overview of efforts to date and described where this call fits in the evaluation design project.
  • Daver requested that the group focuses feedback on the methodology and feasibility of implementing the rigorous design.

Discussion of the Design

  • Use of Medicaid data:
    • Even though it does not seem feasible to conduct at this point, it would be of great interest to states to assess the effects of CDSMP on the Medicaid population (or the dually eligible population). There are a number of new Medicaid initiatives funded by Affordable Care Act that potentially incorporate CDSMPs:  Health Homes (Section 2703), Medicaid Incentives for Prevention of Chronic Diseases Program (Section 4108), and State Demonstrations to Integrate Care for Dual Eligible Individuals (Section 2602).
    • The TEP members agreed that there would be great interest in making the case that the CDSMP program will save Medicare and Medicaid dollars and that CDSMP could be integrated into ongoing programs.
    • Due to long delays in Medicaid data becoming available through CMS, it does not seem feasible to use Medicaid data from MAX or MSIS. However, States do have this data current, but the challenges are getting individual states permission and commitment of state resources to pull the data and also creating a data system that is comparable across states.  A TEP member suggested that if the states were shown the value of sharing the data (i.e., having an estimate of potential Medicaid savings unique to State), they may be more willing to provide it. Results of any cost savings may also lead to Medicare funding of CDSMP. Furthermore, some states are considering integrating CDSMP in their Medicaid programs.
    • Another Medicaid-relevant variable that would be interesting to explore is diversion from (1) Medicaid and/or (2) nursing homes. Insurance status (dually eligibles) could be tracked for this purpose.
    • In light of difficulty to obtain and use Medicaid data, there may be an indirect method of assessing impact on Medicaid outcomes using Medicare outcome data. Some Medicare outcomes could be similar and relevant to Medicaid.
    • If significant number of participants turn out to be Medicaid, then maybe we can later pursue Medicaid data.
    • The more buy-in from the States, the better; we should be careful that this does not appear to be an imposition from AoA. Strategies to buy-in should be considered in the evaluation design.
  • Follow-up Period Recommendations:
    • The report currently recommends a 6 month period where the control group receives no CDSMP services. 
    • The TEP agreed that 6 months seems to be a reasonable amount of time to delay access to CDSMP services.  The TEP members stated that they thought 12 months may be too long, and may pose ethical issues for denying the program for longer than necessary.
    • If a site receives a lot of participants from doctor referrals, the possibility of being placed in a control group may not be appealing.  However, physicians may not mind if their patients are guaranteed a spot in 6 months.
    • A TEP member suggested identifying the pros and cons of using a 6 month control versus a 12 month control from the literature.  It should be determined if there are benefits to keeping the control at 6 months (e.g., attrition).
  • Participant Exclusion Factors:
    • Exclusion criteria should be detailed in the report.  TEP members identified the following potential exclusion criteria: participation in similar programs, proxies/caregiver participants, and cognitive screening.
      • If using the original 24 grantees, the evaluator should avoid recruiting repeat enrollees; they should recruit individuals who are new to the program.
      • NCOA collects repeat enrollee data.  They are aware that sites allow participants to repeat the workshop 2 or 3 times.  At the national level, 2% of participants are repeaters.
      • The evaluator should make a point to track other courses participants may have taken.
      • Using appropriate exclusion criteria will help to avoid criticism of the outcomes.
      • TEP members reported that family caregivers, who may be 60+, attend CDSMP workshops, and should be taken into consideration. 
      • CDSMP is not designed for people with dementia.  Given that a high percentage of older adults in nursing homes have dementia and a high percentage of folks in assisted living do as well, we need to be careful about inclusion of participants from nursing homes or assisted living.  This may be an issue that needs to be addressed for each state.
      • The TEP members agreed that CDSMP programs generally focus on populations that do not need extensive, long-term services.  The percent of participants from long term care institutions should be identified, especially if the CDSM program is targeting these populations.
  • Sample Expansion:
    • Currently the report recommends an evaluation of the original 24 grantees.
    • If using the original 24 grantees, the evaluator should avoid recruiting repeat enrollees.
  • Inclusion of Dual Eligibles:
    • We may want to recommend that sites identify dual eligibles as part of intake process and suggest an optional task and analysis.
      • TEP members agreed that dual eligibles are being served by CDSMP programs.
      • The evaluator should identify dual eligibles at intake and track them.  Outcomes should be identified that would depend on the use of Medicaid data.
      • The participant survey should include questions about what services the participants receive.  These questions should include generic descriptions of service categories.  This will help to determine if participants are receiving Medicaid waiver services.
  • Project Challenge:
    • ARRA funding ends in 2012.  AoA requested that a Plan B, or a “contingency plan” be recommended in the event sites are not able to enroll a sufficient number of participants.
  • Additional Control Variables:
    • Insurance status of participants should be tracked.  If we find a significant number of participants are part of the Medicaid population, we can decide to collect Medicaid data to determine the effects on Medicaid.
    • There are workshops that are provided in languages other than English and Spanish.  For these workshops the lay-leaders and their master trainers are bilingual and they deliver the workshop in another language using the English curriculum.  These workshops are typically recorded as generic CDSMPs.
    • The evaluator may want to determine via the survey, if participants receive Medicaid waiver and/or OAA services.  It could be that those who get these types of services actually have better health and better, more appropriate use of healthcare services.
  • Population Characteristics:
    • The evaluation report should include an analysis of participant characteristics and how they compare to other populations.
    • TEP members agreed that this analysis would be useful for marketing purposes as well as explaining the variation in outcomes.
    • Asking questions about other Older Americans Act, Medicaid waiver services, or other local aging network services used is a relatively simple but interesting addition to data collection.
  • Outcome Measures:
    • The goal of achieving cost neutrality may be an appropriate in determining the percent reduction in Medicare spending that would constitute a significant enough benefit.
    • The TEP members suggested looking into the average cost one hospitalization and/or ER visit.  Average costs may contribute to the development of target cost reductions.
    • Dr. Lorig included a self reported cost savings of $500 in her research.  The recommended sample size would not detect this level of savings.
    • One possible option would be to acquire CMS data for the regions where the evaluation is occurring to get an average estimate of costs for the types of people participating in CDSMPs.
    • The people who are likely to participate in a CDSMP have chronic conditions.  Their cost of care will be higher than average so it is more likely that reductions in cost would be found in the cost analysis.
    • Overall costs (not including medications/Part D) are recommended to be included in the evaluation.  The desired/required reduction in costs may vary based on the different outcomes.
  • Inclusion of Diverse CDSMP Sites:
    • TEP members felt strongly that both rural and Tomando sites should be included in the evaluation, to reflect diversity in the CDSMP program.
    • TEP members noted that there are no major differences between the CDSMP and Tomando workshops.  The main difference is that the Tomando program includes a longer nutrition class. 
    • Nationwide, Tomando accounts for only 8% of the CDSMP.  IMPAQ/Abt pointed out that there are two options: Include Tomando in the study (can’t identify variability) or do a sub-study.
    • TEP Members pointed out that the English version of the CDSMP is provided in multiple languages.  Instructors implement the same training but they use translated materials.  Tomando is the actually modified slightly.  Workshops delivered in other languages may be recorded in the same manner as English CDSMPs.
    • If CDSMPs are being provided in other languages, the evaluation materials (survey and forms) will need to be translated.  Budgeting for translation materials will need to be included in the evaluation report.
  • Discussion of Site Selection and Coordination across Sites:
    • In order to get an adequate sample size, it appears more than 20 sites will need to be included in the study. If there are certain sites that are serving most of the clients, they need to be included but the sample must also be representative.
    • Selecting only the largest sites would likely create an urban bias in the sample.
    • TEP members pointed out that they have been targeting underserved populations.  These populations are often in rural areas.  These populations would be missed if the analysis only includes sites from urban areas.
  • Class Zero:
    • A TEP member suggested that the evaluation contractor may be interested in recommending sites hold an introductory session where participants understand what they are signing up for.  This would also be a good opportunity to collect baseline data.
    • During the site interview calls IMPAQ/Abt did not find that class zeros were being implemented.  Some sites reported that they read a script to interested persons who call in.  The script explains what the workshop entails and it is thought that this may help to prevent drop-outs.  The process was not conducted in a systematic manner.
    • TEP members noted that not all participants show up for the first workshop.  Participants often show up for the second class after hearing about the program from a friend/relative.
    • A participant who shows up for the second workshop and attends at least 4 sessions is counted as a completer.

Return to Contents

Page last reviewed December 2014
Page originally created May 2011
Internet Citation: Appendix E: Technical Expert Panel Meeting Summary. Content last reviewed December 2014. Agency for Healthcare Research and Quality, Rockville, MD.