Patient Engagement in Health Care

People's interactions with health care are now widely acknowledged to be a central focus of health services research. In the past several decades the research community has made great strides in developing and testing frameworks and influences on numerous aspects of individuals' engagement at multiple points in the increasingly complicated matter of seeking and using health care services. Individuals are expected to decide whether and when to seek care, which plans and providers meet their needs, how to manage their health, and how to cope with sometimes conflicting advice from providers and friends and family, all amplified by advances in communications and information technology. To evaluate these increased responsibilities and expectations, researchers have used an array of methods and designs, drawing on economics, psychology, sociology, and other fields to enhance our understanding of how individuals participate at these and other decision points. Moreover, publication of the Institute of Medicine's landmark report Crossing the Quality Chasm in 2001 formally articulated patient-centered care as an essential dimension of high-quality care, a clear focus of new models of delivering care such as primary care patient-centered medical homes (PCMH) and accountable care organizations (ACO) (Committee on Quality of Health Care in America, Institute of Medicine 2001).

At the same time, while the use of researcher-developed tools to assess patient experiences of care (e.g., CAHPS) is now considered routine, and numerous provisions of the Patient Protection and Affordable Care Act reinforce patient-centered care as pivotal to achieving high-quality, affordable care, it is also clear that individuals' tasks are increasingly complex. A recent report on patient engagement in health care describes an Engagement Behavior Framework and multiple behaviors that individuals must master to benefit optimally from their care (Center for Advancing Health 2010). Drawing on previous research, this report represents a precursor to estimating the scale and complexity of challenges confronting individual patients and caregivers, from navigation to managing increasingly sophisticated medical technology at home, often magnified by limited health literacy or past experiences. In short, effective engagement requires considerable skill and motivation, and well-intended initiatives often appear to fall short of collective aspirations to build a system responsive to the needs of patients and families.

In this issue of the journal, four papers address distinct tasks of patient engagement in health care, and one presents a patient-centered approach to assessing health care expenditures. Papers by Boonen, Donkers and colleagues, and Sinaiko examine the impact of financial incentives and limited quality information on consumers' decisions (Boonen et al. 2011; Sinaiko 2011;). The first paper is a discrete choice experiment conducted among Dutch consumers eliciting their willingness to switch general practitioners (GP) or pharmacies in response to different levels and presentations of copayments and very succinct information on quality. Results suggest that negative incentives (e.g., higher copayments) are far more effective than positive, and that respondents were far more likely to switch pharmacies than GPs, a finding attributed to status quo bias. In addition, respondents were aware of and more willing to pay for GP's with higher quality ratings, a dichotomous summary rating, than for pharmacies. Individuals register with a GP in the Dutch system, who serves as a “gatekeeper” to all hospital and specialized services and longstanding relationships are the norm. It is plausible that this context underlies the finding that status quo bias was more predominant for choice of GP than pharmacy. The authors note that insurers may be more successful in channeling all patients to selected pharmacies including those with previous experience, while efforts regarding the choice of GP would be better focused on educating patients who have not yet selected one. Of note, such cost sharing strategies are currently nonexistent in the Netherlands, so the context is quite different than the United States.

In contrast, Sinaiko's experimental study was designed to assess how quality information from multiple sources and financial incentives affect consumer choice of physician in Massachusetts. State employees enrolled in Group Insurance Commission (GIC) tiered plans were randomly assigned to one of six sites of hypothetical-tiered networks of specialist physicians, with different scenarios involving choice of either a cardiologist or a dermatologist, with varying levels of copayments. Participants were also asked about choices of a tier 1 or tier 2 specialist according to GIC quality rating, either with no additional information or with positive information about lower-rated (tier 2) physicians derived from personal experiences or information from family or friends. Starting with a base case of 84 percent of respondents choosing a tier 1 specialist, the results found that almost one half of respondents would switch to a tier 2 specialist if recommended by a friend or family member, and two-thirds would switch to a tier 2 specialist if recommended by their own physician. Copayments between U.S.$10.00 and U.S.$35.00 increased the probability of selecting a tier 1 specialist from 3.5 percent to 11.7 percent. Simulations suggested that copayments would need to exceed U.S.$300.00 to counteract positive recommendations of tier 2 specialists from trusted sources, with some differences by specialty. The authors note that previous analyses suggested that one-third of enrollees do not have full confidence in GIC quality ratings. Whether the results would have been different if respondents had more confidence in the quality ratings in this study, or if more detail regarding quality had been provided, presents fertile ground for future study.

Once decisions regarding when and where to seek care have been made, individuals can seek information and make decisions about involvement in their health care. Skolasky assessed the psychometric properties of Hibbard's Patient Activation Measure (PAM), central to emerging models of chronic illness care, in older adults with multiple comorbid conditions (Hibbard et al. 2004). The findings support the construct validity of the PAM in this population, with significant associations between some health-related behaviors and functional status. Patient activation was not related to the number of comorbid conditions. Patients with higher PAM stages reported better quality of care, suggesting that activated patients may go to extra efforts to seek and obtain better care, raising intriguing questions regarding the extent to which interventions to effect health behavior changes are mediated by patient activation, itself a multidimensional construct (Skolasky et al. 2011). The quality assessments used in this study are both patient-reported surveys. The authors acknowledge that their design cannot demonstrate causality, leaving open the possibility that perceived quality is confounded by the fact that older patients tend to be far less skeptical about medical care than younger patients (Fiscella, Franks, and Clancy 1998). Future studies should address the relationship between patient activation and clinical quality assessments, and whether the results observed here are generalizable to younger people.

Patients' reported experience of care has become an important component of quality assessment of health plans, hospitals, and other care delivery settings since the late 1990s. Most recently, a tool to assess patient experience with clinicians and groups, Clinician-Group CAHPS (CG-CAHPS), has been endorsed and used by some medical groups. Several studies have demonstrated that individuals from different racial and ethnic minority groups frequently report very different summary ratings than Caucasians. Weinick's innovative study was designed to assess the extent to which racial/ethnic differences in ratings of patient experience represent true differences or perceptions (Weinick et al. 2011). Taking advantage of a nationally representative online panel, the investigators developed a video that simulated CG-CAHPS items with varying degrees of physician responsiveness for a patient with a headache. African Americans and whites had similar perceptions of the quality of the physician–patient interaction when presented with the same behaviors, underscoring that reported differences by race are not merely due to differences in how they judge effective interaction. Based on these findings, the authors make specific policy recommendations regarding the use of CAHPS report items rather than summary ratings to stratify findings by race and ethnicity.

Conway's paper presents a patient-centered approach to assessing national health expenditures (Conway et al. 2011). The investigators used the Medical Expenditure Panel Survey (MEPS) to categorize expenditures into seven patient-centered categories: chronic illness (47 percent), acute illness (25 percent), trauma and poisoning (8 percent), dental (7 percent), routine preventive care (6 percent), pregnancy (4 percent), and other (3 percent). The authors appropriately note that the MEPS does not include people who are institutionalized or receive care from the military of Veterans Health Administration. However, as public and private policy makers struggle to identify and communicate approaches that support high quality, affordable care, the potential for presenting information in ways that patients and families experience care is quite intriguing. In particular, this approach may be particularly useful to the newly created Patient Centered Outcomes Research Institute, a private organization supported by a combination of public and private sector revenues, with the unique feature of a 21 member multistakeholder board.

Together, these papers push well beyond the boundaries of earlier work examining how individuals use information on quality or benefit design and illustrate multidimensional challenges ahead. For example, the paper by Sinaiko suggests that the perceived credibility and trustworthiness of quality information and the ability to evaluate information from multiple sources are essential components of effective engagement, albeit far from straightforward. As efforts to reform the quality and efficiency of care in the United States accelerate, it is clear that a focus on the interactions between new models of care and patients' engagement will be an indispensable component of research illuminating which models are most effective.