Implementing Consumer Financial Incentive Programs
Consumer Financial Incentives: A Decision Guide for Purchasers
Question 9. Should consumer financial incentives be structured as rewards, penalties, or a combination of these two approaches?
There is no specific evidence from health services research to address whether consumer financial incentives should be structured as rewards, penalties, or a combination of the two. In economic situations other than health care, it has been shown that people are less responsive to potential financial gains than they are to potential financial losses, even when the gains and losses are of equal dollar amounts.95 In the midst of a medical problem, moreover, patients are even less likely than usual to adhere to economists' standard assumptions about rational choices.
With these caveats, it is likely that both penalties and rewards can be used in creating incentive programs. The identification of any specific price differential as a reward or a penalty could be situational. On one hand, if during open enrollment an employer offers a choice of health plans in three different performance tiers with three different levels of employee contribution, the employee who previously has chosen a plan in the lowest cost tier but now wishes to change plans might see any additional contribution to join a middle-tier plan as a penalty. On the other hand, employees in the highest contribution plans may perceive only "rewards" if they were to move into the same middle-tier plan, especially if it is clear that providers in the middle tier plan provide a similar level of quality.
If, conversely, the approach is to use point-of-care incentives, purchasers' preferences for rewards or penalties may depend on the likelihood of underuse versus overuse of services. For example, multiple small rewards—like the slow accumulation of frequent flier miles—may be effective to reduce underuse of chronic disease interventions, just as airlines are trying to reduce consumers' underuse of their planes and credit cards. In other circumstances, penalties may be appropriate—such as charging a very high copayment for the patient who insists on getting an MRI during the first week of onset of low-back pain, even though he has no neurological symptoms suggesting a serious cause.
Question 10. What are the options for phasing in consumer incentives?
One approach to phasing in an incentive program is to start with public reporting of health care providers' quality ratings combined with a program of public education about differences in quality of care. After a certain level of consumer awareness of quality and cost variation has been achieved, the consumer incentives can then be introduced, for instance, through tiering. The program sponsors might start the incentive program with a limited subset of the measures that had been in the public report—focusing, for example, on one or a few clinical areas, such as maternity care—and then broaden the incentive set over time.
That was the approach used by Wisconsin employers, working through the Alliance—an employer cooperative based in Madison. In 2001, the Alliance issued its first public report of hospital quality, which was limited to maternity care and orthopedic and cardiac surgery. The release of the report card was followed by an extensive consumer education campaign orchestrated through news media, employers' human resource offices, and labor unions. Initial evaluation of this program showed that consumers who had seen the reports were aware that there were differences among hospitals, and they correctly identified high and low performance hospitals. After an apparently adequate level of consumer understanding of the issues was achieved, some Wisconsin employers began pursuing consumer incentives.17
Other options for phasing in consumer incentives include doing a pilot test focusing on a limited geographic area or on specific groups of patients, such as maternity patients only.
Still another approach to phasing in incentives is being taken in Phoenix. The initial goal of the Phoenix Healthcare Value Measurement Initiative (PHVMI) is to integrate existing data from a variety of sources, such as Medicare, Medicaid, hospitals, health plans, and laboratories, and with those data, to develop measures that a broad range of stakeholders agree can be used to improve the quality and cost of health care in Arizona.96
These stakeholders include providers, payers, and consumers. PHVMI will proceed in two phases over 18 months. First, PHVMI partner data will be integrated and baseline reports regarding local standards of practice and disease profiles will be developed to provide a context for phase two. During phase two, stakeholder work groups will evaluate the accuracy, fairness, and completeness of proposed measures of quality, efficiency, and value. After this period of development, it is anticipated that the measures will be used in a variety of ways, which include sharing the data with consumers and offering provider or consumer incentive programs.
Advantages of phasing in incentives are that it allows time to educate consumers about quality differences and permits testing of consumers' responses to various types of incentives before full-scale implementation of an incentive program. Phasing in the incentives also gives providers time to understand the impact of the program and enables the employer or government sponsors to evaluate the small-scale impact before applying an incentive program more broadly.
Question 11. How much money should be put into consumer incentives? How big does the incentive need to be to effect a change, and does the level of incentive necessary vary by the specific behavior that is the object of the incentive?
There is no single answer to the question of how much money is needed for consumer incentives to effect change. Characteristics of both the consumer and the decision being targeted by the incentive clearly matter. Consumers with a lower income are likely to be more responsive to incentives. In terms of the decision being targeted by the incentive, a smaller incentive is generally needed to influence consumers to comply with goals that are easier to accomplish—for example, it is easier to get a flu shot than to quit smoking.
A smaller incentive also can influence consumers to comply with goals that are more obviously beneficial to their health—for example, patients with diabetes may respond better to incentives to get their blood sugar checked than they would to incentives to use a doctor who is "more efficient" in their diabetes care, as patients might assume that "more efficient" means the doctor would spend less time with them.
In terms of incentives to select high value providers, incentive program sponsors may need to factor in issues of proximity and geography. If a plan's preferred providers are all clustered in one part of town, inconvenience may prevent consumers who have to travel across town from choosing providers that perform better, unless the incentives offered are sufficient to compensate adequately for the time and travel required.
In the Tufts Navigator plan, hospitals are tiered based on their performance, but among tiers, the difference to consumers is only $150 to $200—just a fraction of the cost of any hospitalization.97 Of course, from the consumer perspective, an increase in copayment of $150 may be quite large, which is why the designers of this program expect consumers to respond to the incentives. Nonetheless, since the incentive is small relative to total cost, this program and others like it maintain much of the key insurance and risk-spreading benefits of being in a health plan in the first place.
Question 12. How should we think about consumer financial incentives and their relationship to public reporting of quality scores and provider incentives such as pay-for-performance?
No studies have compared the effects of consumer incentives like tiering relative to public reporting of quality scores or the use of provider incentives. There is evidence that providers respond to both public reports about their performance98, 99 and to direct financial incentives.100
The first step is the same for each of the three approaches: the collection of provider performance data, which can then be used for multiple purposes, even simultaneous provision of public reporting and provider and consumer incentives if desired. Thus, the approaches may best be viewed as complementary rather than mutually exclusive.
A potential connection between consumer and provider incentives is their impact on the patient-provider relationship. Although this connection has not been studied directly, it is logical to anticipate that incentives for patients and providers each may be more powerful if they align the goals of the patient and the provider.68 For example, some purchasers in the Bridges to Excellence Diabetes Care Link program—a provider incentive program—also offer their employees rewards for participating in improving the management of their diabetes.101
Acceptance of Consumer Incentive Programs by Consumers and Providers
Question 13. Are consumers in our community ready for financial incentives?
Consumer incentive programs are currently underway in a number of different types of communities and involve large and small health plans, for-profit and not-for-profit providers, public and private payers, and all types of market structures. These myriad programs suggest that consumer incentives could be implemented in a wide range of communities.
As consumers have shown the ability to respond appropriately to data about quality of care in so many situations and regions of the country,4-14 we can assume that a certain number of consumers are informed enough to be able to understand and make decisions about quality of care. Therefore, a relevant question is whether the community is ready to assist patients' health care decisionmaking in a way that would improve the value of care.
The answer to this question turns on whether the community, or the employer or program offering the incentives, is able to: put the quality-of-care information together in a way that enhances consumers' understanding of performance differences,102 design financial incentives that better align the goals of consumers with those of the program sponsor,103 and disseminate the information about health care quality and cost in such a way that consumers find it credible.
Of these, the issue of information dissemination has been studied the least. Consumers are much more likely to use data if they believe the data to be fair and accurate. Trusting the source of the information is key,104 so friends and family members are often relied upon for health care information. Physicians are another trusted source. However, the distribution of performance data needs to be better organized to make incentives most effective.
To date, too little attention has been paid to private and public community organization as additional distribution channels.102 For instance, churches and labor unions—because unions represent members' interests on other issues and sometimes serve as health plan purchasers—are good candidates for the dissemination of information about health care value.102
Question 14. Will consumers believe that the incentives are designed to improve quality, or will they suspect the only goal is to cut costs?
Consumers' response to performance data and any associated incentive is likely to be influenced substantially by the extent to which they are convinced that the data presented are fair and accurate. Consumers' conviction will depend on the other types of information they receive about the measures and the sources of the information. A provider who is not rated "high performance" may tell patients that the measurement is wrong in some way, or that the health plan just does not want to pay for the best providers—leaving consumers with conflicting signals about the data regarding quality performance. These instances become issues of credibility: does the consumer believe the source of the report card is more credible than the provider who rejects being labeled a poor performer?
Similar, but probably more severe, will be situations in which the incentive is to avoid overuse of health care services or to choose a provider who is more efficient. In these cases, consumers may suspect that the goal is to reduce cost to the employer or government program, regardless of the impact on quality of care. In such situations, consumers' resistance can be expected to be significant unless the program's sponsor can provide very credible evidence that the lower utilization targeted will not harm the consumer's health.
One obvious way to counter this potential consumer suspicion is to clearly incorporate quality as a predominant decision factor in designing a tiering system or other incentive.
Another strategy to address the credibility issue is for purchasers and health plans to partner with other stakeholders in the accrual of performance information. For instance, in the Phoenix Healthcare Value Measurement Initiative (PHVMI), providers, health plans, employers, and government representatives are collaborating to decide what factors to measure and how to report their findings in the Phoenix area.96
Basing ratings on a collaboration such as this will lessen the likelihood that providers will question the accuracy of the data made available. In addition, because PHVMI itself is a non-profit, multi-stakeholder collaborative housed at a provider—St. Luke's Health Initiatives—and Arizona State University, the data produced by PHVMI are likely to be viewed as more credible than if similar results were reported by only one health plan or employer.
Question 15. When and how should we engage consumers in discussions about financial incentives?
The timing and mechanism of consumers' engagement in discussions about financial incentives depend on the decisions to be influenced. If the goal is to encourage consumers to select high performance health plans or providers, the first issue to be addressed is when and how to educate them about variations in quality of care. Available data suggesting that consumers can learn and retain information about the quality of providers' care over time5,16 imply that offering such information can be effective, even before the consumer has a clinical event. Thus, although people suffering chest pain seldom have time to consult a hospital's report card about mortality rates from heart attack, they may use the report card in deciding which doctors and hospitals to use before they ever know they have heart disease.
Once consumers appreciate the fact that quality varies among providers in their community, then when and how they are engaged in the financial implications of medical care decisions may vary, depending on both the individual consumer and the decision.
It clearly matters how information about providers' performance is presented to consumers.92 The information should be presented in as simple a way as possible, so that consumers can easily evaluate providers' performance. For instance, clinicians are used to thinking of outcome rates with 95 percent confidence intervals, and they often think that failing to include confidence intervals in presentations of performance data is misleading.
However, consumers find confidence intervals confusing and actually make better decisions when they are presented with data without confidence intervals. Moreover, many of the potential adverse outcomes in health care, such as death from a minor procedure, happen only rarely, and in such cases even physicians are better able to interpret data presented as frequencies—3 out of 100—rather than percentages—3 percent.16
In encouraging consumers to choose high value treatment options in high-deductible health plans with savings options, there are some advantages to targeting the moment when the consumer has a defined clinical problem and is deciding whether having a procedure or another option is worth the cost. Few people pay attention to or retain information that is not relevant to their current health status—even if they realize it could one day be important.105 In choosing the best approach to engage consumers in selecting among treatment options, it is important to realize that often patients are facing clinical consequences they have never experienced.
Accordingly, they do not know how they would feel about different outcomes. For example, it may be difficult for a woman considering mastectomy versus lumpectomy for breast cancer to know how she would feel about losing her breast. An effective approach is to offer decisionmaking tools that include the stories of other people who have experienced the various potential outcomes because these narratives give patients greater context for their own circumstances.106 In addition, patients generally are most able to interpret stories from "people like me," so it may be helpful for decision tools to include narratives from several patients of different backgrounds, even when discussing the same treatment option or clinical result. Expressing data in terms of risks rather than positive results—that is, citing a 15 percent mortality rate rather than an 85 percent survival rate—also seems to be helpful.16
A different approach may be more effective if the goal is to reduce underuse of preventive and chronic care services. In this case, frequent reminders may be helpful—for example, a note sent on a woman's birthday to remind her to get a mammogram.
Question 16. How do consumer financial incentives fit within the broader construct of consumers' engagement?
The consumerism movement in health care has generated substantial interest among purchasers of health care.107 Proponents of this approach want to put consumers in charge of their care on the assumption that consumers can make the best choices for themselves—and they want consumers to share at least some component of the cost implications of those choices. Engaged consumers need tools to help them make good decisions. They need information about the quality of plans and providers and decisionmaking aids when choosing among treatment options.108 If a goal of consumer engagement is to enhance the value of health care provided, then financial incentives may focus consumers' attention on the differences in value among the available options.108
A Social Marketing Perspective on Engaging Consumers in Value-Based Health Care Purchasing
The experiences of social marketers in recent campaigns to reduce lifestyle risks shed light on how consumer incentives should be designed and presented to consumers. Social marketing has been used to reduce the prevalence of smoking—the Truth Campaign;109 increase compliance with seat belt laws—Click It or Ticket;110 and change social norms regarding binge drinking on campus.111 It developed as a way to apply the experience of marketing products and services to the needs of motivating behavioral changes with a societal benefit. The practice of both traditional marketing and social marketing suggests that people change behavior voluntarily when they perceive the new behavior to:
It is easier to market a product or behavior that is actually superior to currently used products and behaviors than one that offers no superior benefit. In terms of consumer incentives in health care, this means that the consumer's response to incentives will be greater if responding clearly makes the overall health care experience better, easier, or more consistent with what other people they respect are doing.
How consumers decide whether or not to respond to an incentive. A fundamental reason for viewing consumer incentive programs through a social marketing lens is that people's perceptions of "better," "easier," and"popular" are governed as much by emotional decisionmaking as by objective fact. Perceptions do not always equal fact, but it is perceptions that govern behavior. In addition, behaviors compete; therefore, any new behavior must be judged by how its benefits and barriers compare to those of a person's current behavior.
In designing a consumer incentive, studies of perceptions of the benefits of current behaviors—what consumers are getting out of their current interactions with the health care system—are just as important as studies establishing the perceived benefits of new behaviors. Moreover, perceptions of what is better, easier, and more popular vary among individual people, both across behaviors and over time.
Finally, while a cash incentive may be attractive to one person, another person might believe easier access to services is more important. For that person, a financial incentive to enroll in a high quality provider network during open enrollment may not be as salient as considering whether to use quality-of-care information to choose a provider in an acute care situation. The perception of better, easier, and more popular is not stable but varies over time as new information and experience are accumulated.
Designing successful consumer incentive programs. Based on these observations, social marketers have adopted a variety of audience and market research tactics to achieve four important objectives:
Divide large populations into segments of people who share common perceptions of what is better, easier, and more popular for a particular behavior.
In trying to identify population segments or to establish what incentives would appeal to consumers, single market research tactics—such as focus groups, surveys, observational studies, or intercept and in-depth interviews—are not as reliable when they are used alone as when they are used in combination. Experience suggests, for example, that the use of focus groups alone is insufficient to understand a large population segment's perceptions accurately.
Incentives are not one-size-fits-all, and some people may never try them on. A key point is that it is difficult to design incentives that will change the behavior of all consumers. In health care, there are some obvious potential targets. For instance, anyone with a chronic disease will naturally think about his or her interactions with the health care system on many occasions, and so they may already be aware of some of the issues that incentives are being used to address.
In addition, they may be more willing to grapple with technical information—as long as it is not too technical—over time, and their repeated attention to the topic may improve their response when incentives are offered. This is fortunate, because a small percentage of all patients—most of them with chronic illness—account for a very large percentage of total care received, costs incurred, and quality deficiencies created.
Another implication of this observation is that many consumers can be ignored in designing consumer incentive programs. This may seem counterintuitive at first, but in health care, there is often a large segment of the population that is healthy and appropriately uses very few services. It may not matter whether this large segment of the population recognizes, understands, and responds to an incentive program because they have little impact on overall clinical or financial results. Instead, attention should be focused on those consumers who are more likely to respond and whose responses matter more.
Communicating with consumers about an incentive program. In communicating information, several basic principles are widely accepted, including the importance of keeping a message simple, repeating it often, stimulating conversation about it as an aid to adoption, and selecting channels and spokespersons who have high credibility with the audience. The recent report from the Institute of Medicine—Health Literacy: A Prescription to End the Confusion112—suggests that as many as 90 million Americans have trouble understanding written health information. This fact, coupled with expanding communication technologies, suggests that oral communication may have important advantages over written materials.
Similarly, it has been shown that unplanned "communication noise"—such as news coverage, public debate, and demonstrations of polarizing behaviors—has a big effect on personal perceptions of public issues. For example, these types of communications had a major impact on parents' responses to educational campaigns about preventing Sudden Infant Death Syndrome. Indeed, the concept of "confirmation bias"113 suggests that whatever information people are given about subject matter for which they already have a strongly held belief is used to strengthen their existing prejudice.
It is often difficult to alter significantly people's perceptions about benefits of or barriers to a new behavior. The use of social norms—making a behavior appear to be popular—has been shown to compensate for this weakness; for example, social "norming" of binge drinking or waste recycling behavior. The pressure to do what others are doing, or what we think others are doing can be an important factor in the adoption of a behavior, even if the behavior is not clearly better or easier. This pressure can accelerate adoption by leading to a "tipping point" in social networks, where a behavior that previously was not common becomes widespread.114
Putting it all together. Table 4 offers some hypothetical examples of consumer population segments that might be amenable to incentives. Also described are some benefit design changes to the insurance product that would represent incentives— financial or non-financial—that might make receiving optimal care more attractive to the targeted consumers. In addition, we propose some communication strategies tailored to these segments.
Although all of these are hypothetical programs—real programs should be based on careful assessment of what your beneficiaries want—they illustrate the general approach that social marketing research suggests should be adopted in designing incentive programs for consumers.