Development and evaluation of a computer decision exercise for consumer participation in insurance benefit planning
October 2005, Vol. 10, No. 2
ISSN 1540 5273
Marion Danis, M.D., Mike Nowak, MILS,
Ellen Benavides, MHA, Susan Dorr Goold, MD, MHSA, MA.
The need for cost containment often forces prioritization of insurance benefits, yet consumers, who have much at stake, have limited options for determining these priorities. We designed and evaluated a computerized, interactive decision tool that groups of employees and other consumers can use to learn about and participate in the design of health insurance benefit packages.
The exercise includes a game board displaying benefit options and markers representing monthly premiums to be used to select benefits, information describing benefit options, and randomly distributed health events to illustrate the consequences of selections. An accompanying program allows modification of the benefit options, premium, and health events. Users with a range of computer experience, education, and income found the exercise easy to use, informative, and enjoyable. The majority reported being satisfied with the group's benefit package decision. The exercise offers a helpful strategy for educating and involving consumers in selection of health insurance benefits.
Keywords: consumer participation, decision making, insurance benefits, managed care programs, software
The call for affordable health insurance benefits is a pressing problem, particularly as health care costs rise. The design of their own insurance benefits greatly affects consumers, yet they have few avenues for participating directly that design. The predominant strategy for consumer choice involves allowing them to select among existing health plans.
A compelling case can be made for directly involving consumers in the design of health insurance benefits through both commercially and publicly funded insurance programs. An employer's purchase of health insurance programs and services without attention to employees' needs results in higher health care costs (Miller and Miller 1993). It has been suggested that small businesses could avoid many problems in insurance purchase decisions if they carefully analyzed the needs and expectations of their employees (Miller and Miller 1993). "Consumer-driven" health plans are a necessary and potentially cost-saving approach to the next generation of health plans (Gupta 2003). This is particularly true as health insurance shifts from comprehensive coverage for a restricted set of benefits to limited coverage for a broader set of benefits (Robinson 2002). It has been suggested that many employers believe that the public is capable of expressing their preferences and choices in healthcare but are concerned that the public needs to understand that resources are limited and trade-offs are necessary (Scandlen 2002). Regarding the public sector, policy makers have increasingly sought to involve the public in health-care priority setting as they realize the value-laden nature of the decisions they face (Abelson et al 2003). The information gained from enrollees in publicly funded programs can serve to effectively shape benefits (Sisk et al 1996).
In today's health insurance environment, the internet and other technological advances have made dissemination of information regarding insurance plans easier. Web sites such as those run by EHealthInsurance (KPMG 2003) and Insure.com (Quotesmith 2003) allow consumers to view and compare the benefits of various plans and choose plans that best suit their needs. Several national organizations offer resources to explain health insurance and ease the cognitive burden on the average consumer making health insurance choices. The Health Insurance Association of American (HIAA 2003) provides guides to managed care, definitions of types and characteristics of insurance packages, and directories of providers. The Agency for Healthcare Research and Quality provides on-line checklists and worksheets to assist consumers in selecting among insurance plans (AHRQ 2003). However, such approaches to consumer involvement remain a step removed from active participation in the design of health insurance benefits. Defined-contribution plans may be the exception that offers an avenue for consumer involvement, but they threaten to undermine the pooling of risk that is necessary to insure large populations of patients (Taylor 2002).
We have therefore designed and developed a group exercise, Choosing Healthplans All Together (CHAT), that involves consumers in shaping health insurance benefits typically found in managed care (Goold et al. 2005; Danis, Biddle, and Goold 2002; Danis, Biddle, and Goold 2004). The exercise helps consumers understand the need for constraints and trade-offs and allows them to compose a heath insurance package using a limited premium. Here we report the design and evaluation of an electronic version of the CHAT exercise that can be useful to health service researchers, purchasers, and health plan managers to study and accommodate the particular preferences and financial needs of various insured populations.
CHAT software design and development
The chat software was designed and developed based on the original chat game (Goold et al. 2005), a simulation exercise designed to allow groups of laypeople to construct health plans within the constraint of limited resources. Like the original exercise, chat software uses a game board shaped like a pie chart in which insurance benefit categories are represented in slices around the pie (Figure 1). Participants select their insurance package by distributing markers among the holes on the board. Participants can select Basic, Medium, High, or Premium options for each benefit category, or they can forgo a category. Participants are given 50 markers to permit them to allocate a quantity of funds comparable to a typical per-member-per-month (PMPM) premium paid by U.S. employers for managed care plans, excluding administrative costs (based on 2002 estimates). Each marker represents 2 percent of the premium. (See Derivation of Actuarial Estimates below.)
During the exercise, the chat board is used in a sequence of four rounds to allow participants to make choices and face consequences (1) alone, (2) in groups of three, (3) as an entire group, and (4) once again alone. This sequence promotes group decision making and allows comparison of individual and group choices.
After participants choose benefits in each of the first and second rounds, the program randomly assigns health events that describe illness scenarios and the associated consequences of their coverage choices including out-of-pocket payment responsibilities, access, and choice of provider or treatment (Figure 2). The group facilitator leads a discussion of these events. During the entire-group round, the facilitator manages the choice of a benefit package by having participants take turns nominating benefits for inclusion in the benefit package, conducting a non-directive discussion of the pros and cons of different benefits, and taking a vote on the use of the last few markers if the group cannot come to easy consensus.
CHAT text is written for participants with a sixth-grade education and the text was tested using Microsoft software for Flesch-Kincaid readability statistics. The exercise is also designed for individuals with little or no experience with computers. To facilitate easy use by individuals without computer experience, a tutorial teaching participants how to roll and click the computer mouse is included at the beginning of the exercise. The software program also includes on-line help, on-line dictionary, and on-line benefit compendium. The software can run under Windows and Macintosh operating systems (OS). The CHAT software is available and can be licensed from the University of Michigan by contacting the authors directly or at firstname.lastname@example.org.
The CD includes a planner program that offers the capacity to vary the premium, benefit categories, health events, dictionary terms, and pre- and post-exercise questionnaires (Figure 3). The CD includes documents including a brochure explaining the purpose of the CHAT exercise, a Planner's Guide for CHAT exercises, and a Facilitator's Manual to instruct those leading CHAT exercises. The Planner runs under both Windows and Mac OS.
Data collection features of the CHAT CD
All data are stored without personally identifying information and are automatically entered into a log. A log utility feature allows the facilitator to save data by downloading from each participant's computer to the facilitator's computer.
Derivation of actuarial estimates
The major data source for cost estimates in the CHAT exercise is derived from the Milliman USA's Health Cost Guidelines (Brookfield, Wisconsin). The average annual premium per employee for 2002 was projected to be $5,360, based on the Mercer/Foster Higgins National Survey of Employer-sponsored Health Plans figure for 2000 ($4,430) and a rate of 10 percent inflation/year. Employee costs (claims and administrative costs) were calculated assuming 2.1 family members per employee. These amounts included any enrollee contributions. The cost of each benefit as derived from the Milliman USA Health Cost Guidelines was converted to markers for representation in the exercise. Because there were 50 markers to spend, each marker was assigned a value of 2 percent of the total premium ($5,360/50 = $107). Policy makers interested in substituting alternative benefits and cost estimates can use the planner software to make such substitutions.
Facilitation of the CHAT exercise
Details about computer system requirements and instructions for setting up and facilitating the CHAT exercise are available in the facilitator's manual contained in the CD. These instructions include a facilitator's script so that those unfamiliar with the exercise can guide participants through the exercise in a uniform and unbiased manner.
Chat user evaluation sessions were held in Minneapolis during July 2002 in a conference room in a non-clinical setting. The room contained a conference table that accommodated 12 laptop computers and an LCD projector for projection of the chat board during the group round. The sessions averaged 2.4 hours including de-briefing time. Sessions were led by one of the authors (EB), a health care consultant who has conducted CHAT exercises, and an assistant, using the CHAT facilitator's script contained in the CD. Pre-exercise questions included items about demographics, health status, educational status, and computer use. Post-exercise questions asked users to rate their enjoyment, understanding, ease of use, and informativeness of CHAT using 4- point response items that were each similarly worded as follows: Was doing CHAT very enjoyable, fairly enjoyable, fairly unenjoyable, very unenjoyable? In addition, we used a 5-point Likert scale to measure participant agreement with statements (Table 3) to determine affective response to the exercise and perceptions of the group process, outcome of decision making, informational adequacy, and range of choices available. These items were adapted from existing instruments (Lind and Kanfer, Tyler 1994, Tyler and Dogoey 1996, Lind and Tyler 1997) and used in other studies of CHAT (Goold et al. 2005). Finally, users were asked if they were satisfied with their group's benefit package and whether they would recommend the CHAT exercise to others.
Participants were recruited to reflect a range of computer use experience, education and income levels, and self-identified ethnicity and health status. Half were recruited through clinical and social service settings including programs for persons with chronic illnesses and uninsured persons, and half were recruited from non-clinical settings. Participants were paid a $50 incentive fee.
The Office of Human Subjects Research at the Clinical Center of the National Institutes of Health and the University of Michigan's Institutional Review Board approved the exercise. All participants were assured of their confidentiality, which was achieved by assigning each participant a letter of the alphabet to serve as their identification during the exercise.
The test population included eighty-seven participants in eight groups. Participants had a mean age of 45 years (range 21-81 years); half were female; one-fifth African American; about one-sixth had no more than a high school education, and three-quarters possessed an education at a college level or higher; one-quarter had incomes under $23,000; about one-third were married; 10 percent reported that their health was poor; 69 percent reported using a computer less than once a month (Table 1).
A large majority of participants found the CHAT exercise either very or fairly enjoyable, easy to understand, easy to use, and informative (Table 2). Nearly all (93 percent) reported that they would recommend CHAT to others, although a quarter of participants found the exercise frustrating (Table 3). Regarding adequacy of information, most believed they learned a lot, that the information was clear and believable (Table 3, items 3-5). The decision tool was perceived as offering realistic choices and sufficient time to make decisions, although many would have preferred more information in order to make good decisions (items 6-8). Participants found the group decision-making process fair, respectful, and attentive to their views (items 10-16). While some found the resulting group plan more favorable to others than themselves, participants were fairly (67 percent) or very (13 percent) satisfied with the group's decision (items 17-19).
The evaluation suggests that the CHAT exercise can serve as a user-friendly tool for facilitating consumer deliberation and participation in selection of health insurance benefits. The majority of study subjects, including those who reported using a computer less than once a month, found CHAT informative and easy to use, and would recommend it to others. While some found the exercise of setting priorities and forgoing some benefits frustrating, as one might expect when "you can't have it all," the large majority found the process fair and reported that the health insurance benefit package thereby chosen was acceptable to them.
Several limitations of the CHAT software warrant attention. While the exercise works effectively with small groups, those planning to use it may need to complement the exercise with strategies for collecting information from larger populations, as recruitment for groups invariably violates the principles of random selection that allow generalization of findings to a population. The development of a web-based version of the CHAT exercise is in progress to address this limitation. Additionally, the extent to which choices made in CHAT, a hypothetical exercise, validly reflect the choices that insurance enrollees would make in a non-hypothetical context remains to be examined.
Another drawback of CHAT is the time commitment required of participants. Such involvement is inherently necessary for meaningful participation in a deliberative process. A variety of strategies can be used to minimize the time required, such as skipping steps of the exercise and eliminating any pre- and post-exercise questions, however a minimum of 1½ hours is needed. Nonetheless, we believe a participatory exercise such as this affords the possibility of providing more thoughtful opinions than surveys may provide (Dolan, Cook, and Ferguson 1999).
The CHAT exercise is a decision tool intended to educate and engage consumers in making health insurance benefit choices. It can serve a variety of educational, research, and policy purposes. As an educational tool, it conveys the concepts of limited resources and rationing in a manner that differs from existing types of consumer education programs. In a project that used CHAT to educate employees of 41 public and private employers about health care trade-offs, employees increased their acceptance of the need to set limits on healthcare coverage, and their willingness to accept tightly managed benefits for the sake of a broad benefit package (Sacramento Healthcare Decisions 2004). As a research and policy tool, CHAT offers a means to assess consumer perspectives on prioritizing benefits. Toward this end, the exercise has been used to survey Medicare enrollees (Danis et al 2004), and it showed their interest in adding benefits such as dental coverage and long-term care along with their interest in pharmacy benefits. It also showed their willingness to expand coverage to the population under age 65. It was also possible to see what trade-offs Medicare enrollees would make to incorporate such additional benefits (Danis et al. 2004). The exercise has been used to determine the benefit priorities of disabled adults who are enrolled in a Medicaid program facing budget cuts (Ginsburg and Glasmire 2004). The exercise has also been used to identify the benefit preferences of the uninsured (Danis et al. 2002). Thus, CHAT has been used as a tool to ascertain the insurance benefit preferences of privately and publicly insured populations, as well as uninsured or underinsured populations that are in need of affordable coverage.
As a practical matter, CHAT would be a useful tool for Cooperative Extension programs and other educators in family economics and consumer health decision making who seek tools for teaching about household budgeting and for reinforcing the lessons of scarcity and trade-offs. The structured nature of the CHAT exercise requires participants to make choices within constraints, which necessitates their recognition of limits. The process thereby increases their acceptance of tight benefit management that is not otherwise possible (Sacramento Healthcare Decisions 2004).
Additionally, human resource consultants and benefits managers may find it useful to ascertain employees' preferences for insurance benefits and gain an understanding of the rationale for their choices. Greater understanding of their priorities can assist employers in evaluating the status of the plans they offer and determine potential directions for the future that are financially feasible and responsive to consumer needs.
How does the CHAT exercise compare to other consumer-centered insurance such as defined-contribution health insurance products? Defined-contribution plans typically offer a consumer health spending account, a major medical policy, and the use of internet support for consumer decision making. (Christianson, Parente, and Taylor 2002; Kelly 2003) Like such plans, CHAT shifts the risks and rewards of managed care to the consumer. Unlike such plans, CHAT is intended to help consumers understand:
- the concept of priority setting through the experience of distributing a limited number of markers
- the need to think prudently about otherwise unforeseen events through the use of health events
- and the nature of insurance as a means of pooling risk by makings decisions in a group.
The impact of consumer involvement in benefit design on the quality and cost of managed care remains to be determined. The literature indicates substantial improvement from the standpoint of consumer satisfaction. One might be concerned that the opportunity to select benefits will be associated with higher usage of those benefits by enrollees. The attendant costs will need to be addressed by narrowing the scope of coverage according to enrollee preferences and adjusting the anticipated costs of selected benefits. Nonetheless, we anticipate that by helping consumers understand the need to ration and elucidating their benefit preferences, the CHAT exercise can increase the feasibility of designing health benefits that are affordable and consonant with consumer choice.
The original CHAT game was jointly designed and developed by Marion Danis at the National Institutes of Health and Susan Dorr Goold at the University of Michigan with valuable contributions from Charlie Hall and Richard Duke. The CHAT software was designed and developed by Marion Danis and Susan Goold with valuable contributions from Mike Nowak, Lesa Monroe-Gatrell, and Ed Saunders of the Health Media Research Laboratory at the University of Michigan. The testing sessions of the electronic version of the CHAT exercise that are reported here were conducted by Ellen Benavides and John Klein of Cirdan Health Systems. CHAT© is a registered copyright owned by the Board of Regents of the University of Michigan. We wish to acknowledge the important contributions of Steve Cigich, F.S.A., and Shelly Brandel, F.S.A., at Milliman USA who prepared the actuarial estimates for CHAT. Their estimates are based on actuarial modeling and analysis regarding the cost of employer purchased health care. Several assumptions were made in facilitating the estimates and each was assessed relative to the degree of precision required for use in CHAT. Actual costs may be different.
We also wish to acknowledge Sailaja Paidipaty, a summer intern in the Department of Clinical Bioethics, at the Clinical Center of the National Institutes of Health.
The opinions expressed here are the authors' and do not reflect official policy of the National Institutes of Health or the Department of Health and Human Services.
Marion Danis, M.D.
Head, Section on Ethics and Health Policy
Department of Clinical Bioethics, National Institutes of Health, Building 10 Rm. 1C118,
Bethesda, MD, 20892-1156,
Fax 301 496-0760
Mike Nowak, MILS
Center for Health Communications Research, University of Michigan
300 N. Ingals 5D-04
Ann Arbor, MI 48109-0471
Fax 734 647-7343
Ellen Benavides, MHA
Health Policy Consultant
2349 Commonwealth Ave.
St. Paul, MN 55108
Susan Dorr Goold, MD, MHSA, MA
Associate Professor of Medicine,
Director, Bioethics Program, University of Michigan
300 North Ingalls Street, 7D20
Ann Arbor, MI 48109-0429
Fax 734 936-8944
Abelson, Julia, John Eyles, Christopher B. McLeod, Patricia Collins, Colin McMullan, Pierre-Gerlier Forest. 2003. Does deliberation make a difference? Results from a citizens' panel study of health goals priority setting. Health Policy 66(1):95-106.
Agency for Healthcare Research and Quality. 2003. Checkup on Health Insurance Choices. Agency for Healthcare Research and Quality. http://www.ahcpr.gov/consumer/insurance.htm,
Christianson, Jon B, Stephen T. Parente, Ruth Taylor. 2002. Defined-contribution health insurance products: development and prospects. Heath Affairs 21:49-64.
Danis, Marion, Andrea Biddle, Susan Dorr Goold. 2002. Insurance benefit preferences of the low-income uninsured. Journal of General Internal Medicine 17:125-133.
Danis Marion, Andrea Biddle, Susan Dorr Goold. 2004. Enrollees choose priorities for Medicare. The Gerontologist 44:58-67.
Dolan, Paul, Richard Cookson, Brian Ferguson. 1999. Effect of discussion and deliberation on the public's views of priority setting in health care: focus group study. British Medical Journal 318:916-919.
Ginsburg, Marjorie, and Kathy Glasmire. 2004. Making Tough Choices: Adults with Disabilities Prioritize their Medi-Cal Options. California Healthcare Foundation. http://www,sachealthdecisions.org/docs/MediCalCHAT.pdf
Goold, Susan Dorr, Andrea K. Biddle, Glenn Klipp, Charles N. Hall, Marion Danis. 2005. Choosing Healthplans All Together: A deliberative exercise for allocating limited health care resources. Journal of Health Politics, Policy, and Law 30:563-602.
Gupta, Ashish. The arrival of consumer-centric healthcare. 2003. Managed care quarterly 2003(11):20-23.
Health Insurance Association of America. http://www.hiaa.org/customer.
Kelly, Paul M. 2003. Will employee choice and defined contribution health plans salvage the embattled managed care system? Benefits Q 19:32-50.
KPMG. 2003. eHealthInsurance.com, KPMG LLP.
Lind, E. Allen, Ruth Kanfer, and Christopher P. Earley. 1990. Voice, control, and procedural justice: Instrumental and non-instrumental concerns in fairness judgments. Journal of Personality and Social Psychology 59(5).
Lind, E. Allan, Tom R. Tyler, and Yuen J. Huo.1997. Procedural context and culture: Variation in the antecedents of procedural justice judgments. Journal of Personality and Social Psychology 73:767-780.
Miller, L., and J. Miller. 1993. Options for small employers. Managed Care Quarterly 1:20-28.
Robinson, J. 2002. Renewed emphasis on consumer cost sharing in health insurance benefit design. Health Affairs Supplement Web Exclusives: W139-154.
Quotesmith.com. Insure.com. 2003. Quotesmith.com, Inc http://www.insure.com,
Sacramento Health Decisions. 2003. "When Options Exceed Resources." http://www.sachealthdecisions.org/CHAT%20report.pdf.
Scandlen, Greg. 2002. A responsible method of making healthcare choices. Managed Care Quarterly 10:13-17.
Sisk, Jane E., Sheila A. Gorman, Anne L. Reisinger, Sherry A. Glied, William H. DuMouchel, Margaret M. Hynes. 1996. Evaluation of Medicaid managed care. Satisfaction, access, and use. JAMA 276(1):50-55.
Taylor, Humphrey. 2002. How and why the health insurance system will collapse. Health Affairs 21:195-197.
Tyler, Tom R. 1994. Psychological models of the justice motive: antecedents of distributive and procedural justice. Journal of Personality and Social Psychology 67:850-863.
Tyler, Tom T, Peter Degoey, and Heather Smith. 1996. Understanding why the justice of group procedures matters: A test of the psychological dynamics of the group-value. Journal of Personality and Social Psychology 70:913-930.
Table 1. Characteristics of the study sample (N=87)
Mean +/- SD
|Age (in years)||-||45 +/_ 11|
|American Indian or
or Pacific Islander
|No health insurance||5||7%|
|Self, spouse/partner or
|VA / military||3||4%|
|Marital status||Single or never married||30||35%|
|8th grade or less||4||5%|
|Some high school
but did not graduate
|High school graduate or GED||9||10%|
|Some college or 2-year degree||12||14%|
|4-year college graduate||23||26%|
education or degree
|$0 to less than $15,500||19||23%|
|$15,000 to less than $35,000||19||23%|
|$35,000 to less than $60,000||11||13%|
|$60,000 or more||35||42%|
|Once or more times a day||4||5%|
|About once a week||4||5%|
|About once a month||10||12%|
|Less than once a month||34||39%|
|Less than a year||4||5%|
|More than 3 years||66||77%|
1 Percentages may exceed 100 because respondents could choose more than one option.
Table 2. Assessment of CHAT
|Assessment||Very +||Fairly +||Fairly -||Very -||Missing||Mean +/- SD|
|Enjoyable||63%||33%||2%||2%||1%||3.56 +/- 1.17|
|60%||39%||1%||0%||5%||3.59 +/- 2.04|
|Easy to do||65%||29%||6%||0%||1%||3.59 +/- 1.14|
|Informative||51%||44%||5%||1%||2%||3.44 +/- 1.52|
Results on a 4-point scale: 4 = Very +; 3 = Fairly +, 2 = Fairly -, 1 = Very -.
Table 3. Statements about CHAT
|1||Thinking about CHAT
makes me feel angry
|2||Thinking about CHAT
makes me feel frustrated
|3||I learned a lot playing
the CHAT game
|4||The information presented
in CHAT was clear
|5||Information given to us
|6||We had enough information
to make good decisions
|7||Health care choices in
the game were realistic
|8||We had enough time to
make good decisions
|9||Game choices included
choices I wanted
|10||I had lots of chances to
share my views
|11||The way the group reached
its decision was fair
|12||The way the group reached
its decision was fair to each
|13||During the game I was
treated with respect
|14||My choice was not
considered by the group
|15||The group tried to be fair||0%||2%||60%||37%||5%|
|16||Discussion during game
was open and honest
|17||The group decision was
favorable to others than me
|18||The group's decision was
favorable for me
|19||I was satisfied with
the group's decision