Allocating Title I and II funds for HIV/AIDS services in Minnesota

1. Allocating Title I and II funds for HIV/AIDS services in Minnesota An ethical analysis of the prioritization process Stuart Rennie, PhD University of North Carolina-Chapel Hill September 20, 2006

2. Presentation outline • Background • Customizing an evaluation tool • Applying evaluation tool to ‘crucial phases’ • Discussion of key issues • Recommendations

3. 1. Background • Prior ethics consultation (March-May 2005) • Main issues: • Conceptual framework for rationing AIDS treatment • Pros and cons of various rationing programs • Ethical management of waiting lists and cost sharing • Creating processes of decision-making for ADAP • Balancing ethics, health outcomes and economics in regard to formulary tiering

4. Background (continued) • Information offered to guide future prioritization: • Framing concepts (dignity, autonomy, distributive justice, utility, efficiency, maximin principle, etc.) • 4 Rationing models (utilitarian, egalitarian, libertarian, maximin) • Pros and cons of implicit and explicit rationing • ‘Accountability for reasonableness’ model of priority-setting processes (Daniels)

5. Scope of this consultation • Retrospective ethical analysis and evaluation of 2006 process of prioritizing service areas/activities by Minnesota HIV Services Planning Council • Recommendations to further operationalize ethical principles in future prioritization processes

6. Two fundamental decisions • The Minnesota Planning Council must make two decisions which impact significantly on persons living with HIV/AIDS • Prioritization: given that there are not enough funds to support all health services/activities on HRSA CARE Act list that ideally would maximally extend and improve quality of life, a ‘fundability’ line must be drawn • Allocation: need to decide how to allocate available funds among service areas/activities chosen by prioritization process

7. HRSA Guidance on priority-setting • In its CARE Act Title II Manual (2003), HIV/AIDS Bureau of HRSA offers a 15 step template for priority-setting (see handout) • Assumption that there is no single ‘correct’ way for planning bodies to prioritize and allocate, but a range of possible acceptable solutions • Focus of this consultation: particular choices made by Planning Council in dealing with ethical challenges of prioritizing service areas/activities

8. 2. Customizing an evaluation tool • Evaluation by definition is an appraisal of something in the light of relevant values • Desired properties of evaluation tool: • Specific to and relevant for Planning Council’s prioritization process • Possessing some degree of validity beyond this particular context

9. Principles and procedures • Ethical principles • Discussed during earlier ethics consultation • Main guiding principles: equity, cultural responsiveness, efficient use of resources • Ethical procedures • Partly validated in scientific literature on health care priority-setting • Publicity, relevance, inclusiveness, revisions and appeals, accountability

10. Issues to be evaluated • Guidance by values Has the actual prioritization process adequately adhered to relevant and agreed-upon ethical principles? • Adherence to procedures Has the actual prioritization process adequately embodied relevant and agreed-upon procedural values? • Acceptability of outcomes Are the outcomes of the prioritization process in harmony with ethical principles?

11. Evaluation strategy Ethical principles: equity, cultural responsiveness, efficiency Ethical procedures: publicity, relevance, inclusiveness, appeals, accountability Constitution and community involvement Data collection, Training and assimilation Prioritization via forced choice voting Allocation process Grievance and Appeals mechanisms Crucial phases of prioritization process

12. 3. Applying the evaluation tool • Crucial phases • Constitution of board and community participation • Data collection, training and assimilation • Prioritization via forced choice voting • Allocation process (including cost-share issue) • Grievance and appeals process

13. (a) Constitution and community participation • Procedural value involved: inclusiveness • Two distinct issues: • Was the constitution of the board sufficiently inclusive? • Did the views of the community adequately impact on the priority-setting decisions?

14. Board constitution • Mandated categories • Meeting them = issue of compliance • Nearly all targets met or surpassed • Observations: • It is not ethically relevant if membership along some variable ‘reflects the epidemic’ • Some categories not filled – is this a problem? • African-born PLWHA on board? • Why inclusiveness matters

15. Community participation • Procedural value involved: inclusiveness • To what extent did the values/voices of community impact importantly on the prioritization process? • Mechanisms: • Activities of committees (Community Participation and Voice): impact on prioritization? • Community forums (Camp Benedict, Aliveness Project) • Does need assessment count as ‘community participation’?

16. What’s out there Greater adherence to procedural value of inclusiveness by means of • Citizens’ juries • User consultation panels • Focus groups • Questionnaire surveys • Opinion surveys of standing panels • Two problems: • Lack of evidence on effectiveness of community participation • Ethical conflict between inclusiveness and efficiency (these mechanisms cost money)

17. A community/Council gap Does it matter if there is a substantial gap between the service areas/activities community members think are important and the priority-setting decisions of the council? • Example: In a study by Kwait et. al. (1999), researchers examined the similarities and differences between the services selected as priorities for funding by Baltimore Ryan White CARE Act Title I Planning Council and services perceived by HIV/AIDS service providers and PLWHA to be unmet needs. • One striking difference: the Baltimore Planning Council ranked health care services (primary health care, mental health services, and home health care) very high, providers and PLWHA were more likely to give the highest priority to direct financial voucher assistance, housing related services, and transportation. • Who lacks information here? The community members or the Planning Council? Can and should the gap be narrowed?

18. (b) Data collection, training and assimilation • Procedural value involved: relevance • Basic issue: did the Council’s process ensure decisions about priority-setting were adequately supported by sufficient, relevant, impartial and understandable information? • Use of extensive and diverse data sources • Quantitative (Epidemiological data, needs assessments) • Qualitative (community forums) • Regulatory (Conflict of interest) • Normative (Continuum of Prevention andCare)

19. Questions to inform evaluation • Were there significant gaps in information that were felt to have compromised the priority setting process to some extent? • Was there enough qualitative information, i.e. information that, for example, could help explain why HIV-positive persons (knowing their status) did not utilize certain services? • Did Planning Council members find it challenging to make comparisons between service areas/activities using these very diverse data sets? • Did the training sessions during the lead-up to the process enable all Planning Council members to understand and make competent use of all the data provided to make their decisions? • Was the information presented in an unbiased way in regard to priority-setting decisions, or was it presented in a way that ‘promoted’ certain service areas/activities?

20. Key challenges In the scientific literature, the following challenges concerning the use of data in priority-setting under the CARE act have been noted (Hayes et. al. 2005): • Devaluation of qualitative data relative to quantitative data • Inaccessibility of some health data due to proprietary and confidentiality restrictions • Difficulties making comparisons across diverse data sets, especially from clients receiving services from multiple systems • Data may be limited only on people already getting HIV care • Data from different sources may use varying definitions of services • Varying levels of quality in regard to reporting and data entry • Challenges of quantifying unmet need among all HIV-infected individuals when significant numbers may not know that they are infected or may not be in care • Comprehensiveness of data may be weakest precisely among members of historically underserved populations and with greatest need for CARE Act services

21. Living with incomplete data • It is unreasonable to think that priority-setting decisions are only ethical if based on complete and perfect information • It would be unethical to drain resources in the pursuit of complete and perfect information if it is at the expense of the welfare of PLWHA • It is reasonable to try to improve data collection and assimilation whenever feasible • HRSA accepts the incompleteness of data if there is commitment to identifying gaps of knowledge and if other parts of the prioritization process are sound

22. Case study: letting the data do (all) the talking • The Brazos Valley Council of Governments is the Ryan White Title II Administrative Agency for the Central Texas HIV/AIDS planning area, and uses the methodology for setting priorities among service areas/activities described in the handout • Note that the prioritization of service areas/activities seems to calculated straight from the 4 data sources: no voting takes place • Is this an ideally ‘evidence-based’ way of setting priorities for Ryan White Titles I and II services? • Would you want to adopt this model?

23. (c) Prioritization via forced choice voting method HRSA guidance on prioritization methods HRSA does not dictate a specific method, but gives the following as (non-exhaustive) examples of acceptable methods: • Group discussion and consensus: decisions to be made are listed, discussed formally or informally, and decisions reached without a formal vote • Aggregate checklists or score sheets: aggregation of individual rankings into final priority list • Nominal group process: small group discussions, individual votes, group discussion, final voting • Delphi method: mailed questionnaires, rankings sent back for feedback and refinement

24. Forced choice voting method • Forced choice instrument to set priorities was first introduced by a former co-chair of the Planning Council in 2004 • Adapted from a forced choice instrument used by a school district’s parents committee to prioritize programs • 2006 is the second year that the forced choice instrument has been used • Early July 2006, Planning Council members were given a forced choice worksheet comparing service areas for PLWHA (some areas have multiple service activities and these were ranked among each other using separate sheets) + MAI worksheets • Two weeks to individually complete the worksheets – using the data they had received over the previous months • Individual prioritization worksheets returned • Results of all individual worksheets were aggregated into a final priority ranking of service areas/activities

25. Evaluation of voting method Procedural values involved: publicity and relevance (a) Methodological issue: Are the choices made partly an artifact of the way (or order in which that the choices) are presented? (b) Ethical issue: Can the forced choice voting method be reconciled with the procedural values of publicity and relevance?

26. Conflict with publicity • Decisions regarding both direct and indirect limits to care and their rationales should be publicly accessible. • This does not just mean that the results of the priority-setting process should be made public; it means also that the reasons behind the decisions made should be made public as well. • However, the public visibility of the rationale behind the individual choices, and hence the rationale behind the final priority list of services/areas, is highly restricted when the forced choice voting method is employed. • Ultimately, only the individual members can know what drove their own choices.

27. Conflict with relevance • Priority-setting decisions should be based on reasons that stakeholders can agree are relevant to meeting health needs under resource constraints • But in the forced choice vote process, the reasons why any individual member chooses one area/activity over another remain hidden from view • The individual’s rankings may well be based on good (data-based, relevant, ethically sound) reasons • But when voting takes place in a ‘black box’, it is hard to know …

28. A critical view of forced choice voting Settling moral disputes simply by aggregating preferences seems to ignore some fundamental differences between the nature of values and commitments to them and tastes or preferences . . . The aggregative conception seems insensitive to how we ideally would like to resolve moral disputes, namely through argument and deliberation. We expect people to offer reasons and arguments for their moral views, and we hope that the better arguments will prove persuasive . . . Merely aggregating preferences through voting or surveying ends up making majority might determine what is right, with no real constraints on the kinds of reasons that play a role in the decision.” (Daniels and Sabin 2002: 37-39)

29. A counterargument “If the Council members used all the relevant data, and were committed to applying ethical values when making their decisions, then their choices cannot be called mere ‘preferences’. There is a need for trust that the outcomes reflect proper judgment.” Response: • Everyone has biases, so it may be better to air reasons for priority choices in open discussion • Given the impact of the decision of PLWHA, the burden for stating reasons for priorities would seem to fall on the Council • Forced choice voting also hampers the appeals and revisions part of the process

30. Questions to ponder • What motivated the adoption of the forced choice method in the first place? This method is crucial to the prioritization process, so how was the original decision to use this particular method made? • Should the prioritization of service areas/activities be ‘fixed’ by the aggregation of forced choice voting sheets alone? • Why is there an extensive, data-based and public discussion about allocations, but no similar discussion about prioritization of service areas/services after the results of the forced choice voting are reported? • Given the criticisms of the forced choice voting method mentioned above, does the Planning Council still want to use the method next year, in the same way?

31. Possible alternatives • Re-assessing the current use of the forced choice method does not necessarily mean rejecting the method altogether. Look for complex methods that can tap the strengths of the forced choice voting method while avoiding its shortcomings • One possibility: combine elements of the ‘group discussion and consensus’ and the forced choice voting methods. Reconceive voting as an important input within a public discussion devoted to prioritization, rather than the final word on priorities • To operationalize this method, Council members would have to keep a record of what considerations were important to them while making their choices, and be prepared to discuss the reasons behind their choices during a Planning Council meeting

32. The San Francisco approach Council members are asked to complete a questionnaire concerning which ‘operating values’ and ‘justice paradigm’ they are using during the prioritization process (See handout HIV Services Planning Council, Yerba Buena Center) The San Francisco Council understands the defining of core values as an iterative process, partly driven by changes in the epidemic Seems to provide a window on the reasons driving the priority decision, while stopping short of going into group discussion about reasons

33. 4. Allocations process Allocations process lies outside the scope of work of this consultation On the other hand, the allocations is partly determined by the prioritization process – ethical issues that affect prioritization affect the allocation process too The policy of cost-sharing and patient closures: probably one of the most ethically challenging aspects of the Planning Council’s activities • Cost-sharing pits the value of equity (in terms of the ability to pay) against efficient use of resources (in terms of cost-recovery). • If priority-setting decisions can be made that favor cost-assistance programs, without undue draining of other important resources, it seems ethically obligatory to do so

34. Recommendations (provisional) May need more conceptual clarity in regard to fundamental value concepts, used to guide and evaluate Council activities (such as equity) There may be need for more extensive community consultation to adhere to the value of inclusiveness, but this needs to be balanced against concerns about planning costs The data collected seems relevant but more needs to be known about how Council members tackle the challenge of channeling data into values and choices More qualitative studies that answer ‘why’? questions may help Council members link data with values and choices more easily The forced choice voting method may need to be revisited so the process better embodies the values of publicity and relevance