To address the rising cost of chronic conditions, health systems
must nd effective ways to get people to adopt healthier
behaviors. A new person-centric approach to behavior change is likely to improve the odds of success. Changing patient behavior: the next frontier in healthcare value Health International is published by McKinseys Healthcare Systems and Services Practice. Copyright 2012. McKinsey & Company. 65 Designing and implementing programs that enable people to achieve sustainable behavior change is hard. Few programs tried in the past achieved sustained impact. However, many of these interventions were rooted in the old model of healthcare, focusing on the treatment of clinical problems after an acute event. Too often, the interventions had poor program design, insufcient measurement rigor, and implementation issues. The failures led many health system leaders to be skeptical about whether any behavior change program can achieve long-term impact. We believe that behavior change programs can succeed, but only if their design paradigm is rethought. This article describes an emerging approacha person-focused paradigm that uses a behaviorally based rather than disease- based orientation to drive sustainable behavior change. Instead of assuming that individuals are fully rational, it recognizes that human decision making is afected by systematic cognitive biases, habits, and social norms. Instead of focusing exclusively on the clinician- patient relationship, it seeks to create a sup- portive ecosystem that engages individuals and those closest to them. Our perspectives draw on an analysis of global trends, our extensive experience working with clients throughout the healthcare industry on this topic, and interviews with leading experts. They are grounded in emerging insights from the behavioral sciences that shed light on how individuals actually make decisions, as well as new technological advances. Leveraging these insights, we have developed an integrated frame- work to help healthcare organizations across the value chain understand the new paradigm and how they can design and implement high- impact, patient-focused interventions. Changing individual behavior is increas- ingly at the heart of healthcare. The old model of healthcarea reactive system that treats acute illnesses after the factis evolving to one more centered on patients, prevention, and the ongoing management of chronic conditions. This evolution is essential. Across the globe, a fundamental shift in healthcare risk is taking place, driven by an aging population and the increasing incidence of behaviorally induced chronic conditions. Health systems are inno- vating on the delivery side to meet this chal- lenge through a growing emphasis on primary care, integrated care models, and pay-for-value reimbursement. Yet more must be done to reorient health systems toward prevention and the long-term management of chronic conditions. In an ana ly- sis we conducted of US healthcare costs (which are now nearing $3 trillion annually), 31 per- cent of those costs could be directly attributed to behaviorally infuenced chronic conditions. Fully 69 percent of total costs were heavily infuenced by consumer behaviors. Poor medi- cation adherence alone costs the United States more than $100 billion annually in avoidable healthcare spending. 1 The burden consumer choices place on low- and middle-income coun- tries is similarly staggering: Harvard and the World Economic Forum have estimated that noncommunicable diseases result in economic losses for developing economies equivalent to 4 percent or 5 percent of their GDP per annum. 2 Unless health systems fnd ways to get people to change their behavior (in terms of both making healthier lifestyle choices and seeking and receiving appropriate preventive and pri- mary care to manage their health conditions), they will fail in their quest to tame healthcare costs without impairing care quality or access. 65 Sundiatu Dixon- Fyle, PhD; Shonu Gandhi; Thomas Pellathy; and Angela Spatharou, PhD 1 Osterberg L, Blaschke T. Ad- herence to medication. N Engl J Med. 2005;353:487-97. 2 Bloom DE et al. The Global Economic Burden of Non- communicable Diseases. World Economic Forum and Harvard School of Public Health. 2011. 66 Health International 2012 Number 12 Utilizing remote and self-care-oriented technologies to support and empower individuals, and connect them to clinicians and other infuencers Adopting a multi-stakeholder approach, which includes public-private partnerships, to support high-impact societal and pri- mordial prevention interventions Engaging individuals Insights from behavioral sciences are being widely used in fnancial services, retail, and other sectors to infuence what we buy, how we save, and other aspects of our behavior. Yet the Elements of the paradigm The new person-focused paradigm for be- havior change has fve major components (Exhibit 1): Engaging individuals more efectively by taking advantage of new insights from behavioral psychology and behavioral economics Integrating behavior change as a core component of new care delivery models Using the power of infuencers and networks to support behavior change Exhibit 1 New paradigm for patient behavior Health International #12 December 2012 Patient Behavior Exhibit 1 of 2 Integrating into care delivery Leveraging networks Creating broader coalitions for change Using technology to enable self-care Engaging individuals 67 Changing patient behavior: the next frontier in healthcare value than reaching those who need help before they can take proactive steps to improve their health. What does good design look like? With regard to behavior-change interventions, three inno- vations appear to be most important. Behaviorally based segmentation should be used to deepen insights into specifc groups. Current approaches to patient segmentation and predictive modeling tend to center on clinical conditions. However, change inter- ventions are more likely to be successful if they take into account additional factors, such as a persons behavioral profle or motivation to change. These insights enable more focused targeting of the groups of people for whom impact is most likely to be achieved. They also make it possible to design programs that more efectively address practical barriers to change. design of most health-related products, services, and interventions remains remark- ably unafected by these discoveries into how humans make decisions. For example, tradi- tional clinically driven interventions assume that individuals understand their own health issues and usually act rationally to address them; however, this is often far from the case. In a survey we recently conducted, 76 percent of the participants with high-risk clinical conditions described themselves as being in excellent, very good, or good health (Exhibit 2). Programs that fail to account for this gap be- tween individuals actual health status and how they understand and experience their health on a day-to-day basis (and thus how willing they are to change their behavior) miss the boat in terms of design. Often, these programs simply attract individuals who are already activated to change their behavior, rather Exhibit 2 Most people think they are signicantly healthier than they are 7 24 33 12 28 29 43 57 31 26 41 34 42 66 27 Respondents self-assessment of their health status by different risk categories, 1 % 1 Based on derived health profile. Source: McKinsey Retail Healthcare Consumer Survey BMI >30 BMI <30 Chronic High-risk Low-risk Excellent/very good Good Fair/poor Health International #12 December 2012 Patient Behavior Exhibit 2 of 2 68 Health International 2012 Number 12 Person-focused pathways should be used to support people as they attempt to alter their behavior. Most disease management programs remain rooted in a clinically based view of the world. For example, they may correctly identify a patient with diabetes or another chronic condition, but do not fully address the fact that the same patient may also be overweight, sufer from heart disease, have mild-to-moderate depression, mistrust his clinician, and be socially isolated. Clinical insights are critical, but our experience shows that program designs are more efective when they directly address the root causes and barriers to behavior change and provide interactions with the right timing and frequency to ensure impact. In essence, these designs translate clinical insights into person-focused pathways that support individuals from the point at which they decide to make changes to the point that the new behaviors are sustained. A simple example demonstrates the impact of guiding patients to the behavior-change inter- ventions that are most suited to them, based on their needs. In England, we worked with a regional payor to improve diabetes care by defning behavioral segments among afected patients and then matching the right portfolio of support programs to each segment. General practitioners were trained to identify which segment patients belonged to by asking a few For example, most programs geared to ER frequent fiers or people with high hospital admission rates target patients through risk-, disease-, or condition-based retrospective reviews of high-cost episodes. Incorporating additional behavioral insights permits a more nuanced approach. In a recent project for a large US payor, we used demographic, family structure, and consumer purchase data (e.g., nature of purchases, car ownership, etc.) to construct a social isolation index (a variable intended to measure each indi vi- duals degree of social connection) for the target population. When combined with claims data, this index enabled us to more efectively predict, among groups with equivalent at-risk chronic con ditions, which people were likely to have a high-cost emergency room admission or inpatient event. We found, for example, that hospital costs were 24 percent higher for socially isolated indivi duals than for socially connected indi- viduals with an equivalent level of clinical risk, and that the socially isolated individuals also had lower prescription drug use. Such insights can help identify key patient subgroups before high-cost episodes occur by typing members against defned predictors; interventions targeted toward these subgroups can then be designed with the right focus (e.g., feld-based extender services and medication adherence interventions for socially isolated individuals). Incentives that take peoples cognitive biases (e.g., loss aversion, regret aversion, optimism, and present-biased preferences) into account are more efective than direct cash rewards. 69 Changing patient behavior: the next frontier in healthcare value ery estimates that the program has lowered participants overall healthcare costs (on a risk- adjusted basis) by about 15 percent. 5 Innovative corporate wellness programs, such as those ofered by Limeade, are also gaining traction. The structure of the rewards matters. Incen- tives that take peoples cognitive biases (e.g., loss aversion, regret aversion, optimism, and present-biased preferences) into account are more efective than direct cash rewards. We recently tested behaviorally based incentives using a regret lottery design. 6 The goal was to get a companys employees to complete a health risk assessment. Half the employees were given cash incentives directly; the others were divided into small teams that were then enrolled in a lottery. Each week, one team would win the lottery, but rewards were distributed only to team members who had completed the assess- ment. The winning teams were widely publi- cized to leverage anticipated regret (peoples disinclination to miss their chance of winning the big prize the week their team was selected). The result: 69 percent of the employees in the lottery completed their assessments, compared with 43 percent of those given direct incentives. simple questions and then to direct them to the behavior change intervention that best met their needs. This simple steerage led to a nine- fold increase in program enrollment (from 7 percent to 63 percent) within six months and, more importantly, to a higher rate of program completion. Similarly, even very simple de- faults, such as automatic mail-order enrollment for prescription renewals, can help address pa- tients barriers to adherence. Active communication along the pathway is also critical, because frequent feedback en- courages behavior change. A study on weight loss we conducted with leading behavioral economists suggests that giving people frequent, automated feedback helps improve weight loss. 3 Text messaging is being increas- ingly used to support patients with diabetes or other chronic conditions and to send them educational materials, medication reminders, and tips on disease management; preliminary results are encouraging. Behaviorally based incentives should be used to encourage change. Incentives are an increasing part of the toolkit for addressing behavior change. Two-thirds of US companies, for example, now ofer employees fnancial incentives to encourage healthy behaviors. 4
Well-designed incentive programs have de monstrated impact. Discoverys Vitality program, for example, informs members about their health status, encourages them to set behavior-dependent health goals, and then rewards them for attaining those goals. Members earn points for behaviors ranging from under- going diabetes screening to healthy purchases in supermarkets, and in turn receive a mixture of short- and longer-term rewards, including cinema tickets and discounted fights. Discov- 3 In press. 4 Performance in an Era of Un- certainty. 2012 Tower Watson employer survey results. 5 Morris G. Presentation about Discoverys Vitality program. Oxford Health Alliance Summit. 2010 6 Haisley E et al. The impact of alternative incentive schemes on completion of health risk assessments. Am J Health Promot. 2012;26:184-188. 70 Health International 2012 Number 12 house calls by physicians and nurse practitio- ners, tailored ftness centers, and an intervention team that goes to patients homes to investigate nonclinical problems). CareMore reports that its risk-adjusted costs are 15 percent lower than the regional average for comparable patients and its clinical outcomes are above average. For example, its amputation rate among diabetes patients with wounds is 78 percent below the national average, and its rate of hospitalization for end-stage renal disease is 42 percent below that average. 7 Using the power of influencers and networks Health choices are not made in a vacuum. Our research shows that when faced with a health event, people follow the treatment advice of friends and family 86 percent of the time. Some health promotion eforts already recognized the importance of these infuencers. For example, adult smoking cessation programs in the United Kingdom and elsewhere are increasingly targeting young children, because parents who smoke are more likely to respond to their childrens concerns than to the prospect of their own poor health. Payors and providers have also come to appreci- ate the power of infuencers to support behavior change and have used peer programs with considerable success. In Philadelphia, for example, the US Veterans Afairs (VA) Medical Center created a peer program to encourage better diabetes self-management among African- Americans (a group with a higher-than-average prevalence of diabetes and a signifcantly increased risk for complications). The program frst identifed mentorsother diabetes patients who were already keeping their glucose levels under good controland gave them train- ing. Program participants were then assigned Integrating behavior change into new care delivery models Many health systems are putting increased emphasis on primary care, especially through the use of integrated care delivery models designed to improve the health of the popu- lation. To succeed, these new models must extend their reach outside of the four walls of a clinicians ofce so that they can support patient behavior change beyond traditional clinician-patient interactions. This requires new capabilities, including clinical workfow tools to support patient targeting, care alerts sent to both clinicians and patients, enhanced communication and care manage- ment support for patients, and remote moni- toring. More fundamentally, clinicians must adopt a patient-centered approach when they interact with patients, one that focuses on understanding the whole person and their barriers to change. A good example of this kind of model is CareMore, a California provider that focuses on seniors. One of its primary goals is to encourage behavior changes crucial for efec- tively managing chronic conditions. CareMore combines technological innovations, including electronic medical records (EMRs) and remote monitoring, with a wide array of nontraditional services (e.g., caregiver support, preventive podiatry, no-cost transportation to its ofces, Health choices are not made in a vacuum. When faced with a health event, people follow the treatment advice of friends and family 86 percent of the time. 7 Reuben DB. Physicians in sup- porting roles in chronic disease care: the CareMore model. J Am Geriatr Soc. 2011;59:158-60. 71 Changing patient behavior: the next frontier in healthcare value Utilizing remote and self-care-oriented technologies Frequent, real-time communication and feedback are important in supporting change eforts. Traditional models of care delivery have, at their core, face-to-face interactions between clinicians and patients. New techno- logies, however, are augmenting this interac- tion model and fundamentally transforming the ways in which clinicians deliverand indi- viduals and their friends and family consume care. Mobile apps, for example, can facilitate tracking and monitoring. Wireless devices can transmit adherence information directly from pill boxes, scales, or even ingested smart pills. Webcams enable remote consultations. Ulti- mately, these remote and self-care-oriented technologies may help create a truly interactive healthcare ecosystem for patients. Many of these new technologies are gaining traction, particularly in developing countries, mentors with the same demographic back- ground (gender, age, etc.). The participants and mentors interacted on a weekly basis, primarily by telephone. After six months, the participants had achieved an 11 percent drop in their average glucose levels (from 9.8 percent to 8.7 percent), a change sufcient to decrease their risk of disease-related complications. 8
In contrast, a control group of patients who did not have mentors experienced no improvement in their glucose levels during the study. Nearly two-thirds of the participants in the peer pro- gram said that having a mentor who also had dia betes was important in helping them control their own glucose levels. As the VA program demonstrates, peer-based networks can be relatively easy to implement. As long as the peer matching is done in a way that resonates with participants, these networks can provide an additional support system to help sustain behavior change. 8 Long JA et al. Peer mentoring and fnancial incentives to improve glucose control in African American veterans: a randomized trial. Ann Intern Med. 2012;156:416-424. 72 Health International 2012 Number 12 Adopting a multi-stakeholder approach There is increasing recognition that if health systems are to address the full range of issues adversely afecting patients health, healthcare leaders will need to partner with a broader set of stakeholders to create an environment conducive to driving healthier behaviors and achieving impact. We have worked closely with clients attempting to create such broad coali- tions, which we believe are crucial for achieving strong, sustained behavior changes. For example, we worked with major retailers and food manufacturers in one country to ad- dress the challenge of obesity by creating a movement to raise awareness and spur con- sumers, employ ers, children, communities, and organizations to action. With the support of a multi-stakeholder coalition, a plan was developed in which the CEOs of participating retailers and food manufacturers committed their organizations to certain targets and actions. These ranged from healthy school partnership programs, workplace ftness and nutrition programs, and joint manu- facturer/retailer initiatives to lower caloric intake and increase caloric transparency. Although the economic impact and health consequences of these types of eforts are hard to quantify, they are critical in creating an envi- ronment that supports more direct interventions. More direct impact can be achieved through appropriately focused government interventions and public-private partnerships. A classic example is increased taxation on cigarettes, but more creative interventions are also possible. In Argentina, for example, a govern- ment-sponsored conditional-transfer program aims to reduce average sodium intake; bakers have been asked to decrease the amount of salt in their bread but are directly compensated for lost revenues from lower sales. where access remains an issue. However, they are also being increasingly used in more developed countries. In the United Kingdom, for example, a large trial of telehealth devices for patients with social care needs and chronic conditions has produced positive results. Participants received either home monitoring equipment or a set-top box that could be connected to their TVs; the devices enabled patients to ask questions about their symptoms, gave them visual or audio reminders when measurements were due, showed educational videos, and charted a graphical history of recent clinical readings. In the trial, telehealth device use appeared to reduce the number of emergency room visits and hospital admissions, as well as one-year mortality rates. 9 Studies among US Medicare and VA patients have also shown that telehealth devices decrease healthcare utilization. In these studies, use of the devices has produced savings of up to 13 percent. 10 9 Steventon A et al. Efect of telehealth on use of secondary care and mortality: fndings from the Whole System Demon- strator cluster randomized trial. BMJ. 2012;344:e3874. 10 Baker LC et al. Integrated telehealth and care manage- ment program for Medicare benefciaries with chronic dis- ease linked to savings. Health Afairs. 2011;30:1689-1697. 73 Changing patient behavior: the next frontier in healthcare value attitudes hinder the fact-based evaluation of behavior change programs and the adoption of proven successes. Re-orienting health systems around a model focused on prevention, long-term management, and patient-centered care will require top-down leadership and advocacy. Such leadership is necessary if health systems are to meet the coming wave of healthcare challenges. . . . If health systems are to address the shifts in healthcare risk now taking placeespecially those resulting from chronic conditions they must fnd ways to get individuals to adopt healthier behaviors. New behavior change programs based on a person-focused, rather than disease-focused, paradigm are proving that it is possible to achieve strong, sustained results. However, a change in mindset is required if these programs are to gain widespread use.
Sundiatu Dixon-Fyle, PhD, a practice expert
in McKinseys London ofce, supports payors and providers on strategies for patient-centered care. Shonu Gandhi, an engagement manager in the Washington, DC ofce, works extensively on con- sumer engagement with payors. Thomas Pellathy, a partner in the Pittsburgh ofce, concentrates on helping health systems improve healthcare value. Angela Spatharou, PhD, a partner in the London ofce, focuses on strategic issues, including integrated care, for healthcare providers. Impact and implementation We believe that the new person-focused paradigm described here is likely to deliver stronger results than traditional behavior change programs have produced. Disease management programs rooted in the old model of healthcare typically achieve savings in the range of 2 per- cent to 5 percent of medical costs. Based on our experience and the studies published to date, we estimate that programs designed under the new paradigm could deliver a 10 percent to 15 percent reduction in those costs in target populations, in addition to productivity gains, better outcomes, and better quality of life. Implementation of the new paradigm is chal- lenging, though. One signifcant issue is scal- ability: while many of the needed elements exist and pilots abound, there are few instances of anyone applying all of the design elements at scale. The cost of building the underlying infrastructure (e.g., platforms to administer incentives and provider EMR systems to enable efective patient insights) is also an issue although, in most cases, low-tech, cost-efective approaches exist, and ongoing innovation is simplifying and lowering the price of many technologies. The biggest obstacle, however, is the mind- sets of healthcare leaders and clinicians. Most remain rooted in the old model of healthcare. Many are highly skeptical of behavior change programs; some do not even consider behavior change as part of a health systems remit. These Re-orienting health systems around a model focused on prevention, long-term management, and patient-centered care will require top-down leadership and advocacy.