Introduction

The overuse of medical interventions, including pharmaceuticals, testing, and screening, has come under scrutiny for its economic costs and potential to harm patients [1]–[5]. In the United States, unnecessary services cost private and public payers $210 billion in 2009, approximately 30% of all healthcare costs [6]. In Medicare, the US program that provides health insurance for all Americans over age 65, the number of diagnostic tests per beneficiary grew approximately 85% from 2002 to 2012 [7]. Drivers of overuse in the United States include fee-for-service payment structures, physicians’ fear of malpractice lawsuits, industry marketing of pharmaceuticals and devices to medical professionals, and physicians’ culture of thoroughness [8]. To address overuse, the American Board of Internal Medicine (ABIM) Foundation and a coalition of professional medical associations have begun to engage physicians in efforts to “choose wisely” among treatments and screening and testing procedures [9].

However, physicians are not solely responsible for the types and volume of clinical care that patients receive. Two generations ago, patients were expected to defer to physicians’ authority and accept the passivity inherent in what Talcott Parsons called the sick role [10]. Patients were meant to seek technically competent help and follow physicians’ orders. But patients’ and physicians’ roles have changed. Starting in the 1960s, bioethical, legal, and administrative principles promoted patient autonomy and circumscribed physicians’ discretion [11], [12]. The trend toward patient autonomy accelerated in the 1970s and 1980s, in part because of activism by feminists and people with AIDS who were determined to control their medical care [13]. Patient access to medical information through the internet has reduced their dependence on physicians for medical information [14], [15]. Although patient choice remains constrained by physicians, payers, regulators, and industry, patient participation has become an institutionalized part of the healthcare system [16].

Given patients’ participation in decision-making and care, their potential contributions to overuse have begun to draw attention. Carman et al found that patients strongly believe that more testing and more treatment lead to better outcomes and, to a lesser extent, that newer treatments are more effective [17]. We use the term “medical maximalism” to describe this reported preference for more medical intervention. Patients are especially enthusiastic about routine cancer screenings, regardless of false positives or limited effects on mortality [18]. Recent United States Preventive Services Task Force (USPSTF) recommendations to limit screening for breast and prostate cancer, for example, have provoked considerable public resistance [19]–[21]. Patients have also been found to be hostile to research-based treatment guidelines, which they believe could lead to rationing and “one size fits all” medicine [22].

Emanuel has attributed these maximalist attitudes to a general American enthusiasm for technology [8]. But Americans are not in fact enthusiastic about all technologies, nor do all Americans adopt technologies at the same rate [23]. Meanwhile, medical overuse is not solely an American problem [24]–[28]. In many economically more developed countries, the increased availability of screening and diagnostic technologies has made it possible for people to closely monitor their bodies for biomarkers associated with disease [29], [30]. However, specific features of the American healthcare system likely predispose patients and physicians towards maximalism. Third-party payers shield insured patients and their physicians from many of the costs of pharmaceuticals, testing, and screening. Pharmaceutical and device manufacturers now use direct-to-consumer advertising (DTC) to increase patient demand for and physician awareness of expensive new pharmaceuticals [31].

We sought to explore the logic underlying patients’ apparently maximalist attitudes, not to determine the quantitative distribution of their attitudes. We therefore conducted focus groups to better understand patient preferences for more pharmaceuticals, testing, and screening. We considered whether attitudes towards testing and screening differ from attitudes toward pharmaceutical use. We also sought to understand how problems of overuse and non-adherence could be related to patients’ efforts to take responsibility for their health.

Methods

Participants (n = 43) were recruited from a database of individuals who had expressed interest in taking part in focus groups but who had not done so within the previous six months. Participants were selected based on self-reported answers to a screener developed by the authors (Appendix S1). Following Carman et al’s approach, we selected privately insured participants because previous research indicates that such individuals feel relatively empowered to make healthcare choices [17], [32]. We chose individuals who reported themselves and their immediate families as in “excellent” or “good” health because we assumed that people with terminal or chronic illnesses merit separate inquiry. We selected individuals in regular contact with the healthcare system, defined as having visited a physician 2–6 times and having filled 1–5 prescriptions for themselves in the past 12 months, excluding contraception. Individuals who reported more than one hospitalization in the past two years besides childbirth were excluded. Participants were middle-aged, employed full-time or part-time, spoke English fluently, had at least some college education and had household incomes above $50,000, the approximate 2011 median [33]. Anyone employed in healthcare was excluded, including medical professionals and employees of medical insurance companies, device manufacturers, or the pharmaceutical industry.

Recruiters asked potential participants to rate their agreement on a Likert scale with the statement: “More, more expensive, newer medical tests and procedures are usually better.” The screener also asked potential participants to rate on a Likert scale their satisfaction with the healthcare that they and their family receive. These questions were not used to qualify or disqualify potential participants, but served as prompts during the focus group discussions.

In March and April 2012, a trained moderator led five 60-minute focus groups using discussion guides developed by the authors (Appendix S2). Three groups were conducted in suburban locations – one in Tarrytown, NY and two in a suburban area of Atlanta, GA – and two in an urban location, Chicago, IL. However, we draw no conclusions about geographical variation. Focus group participants received $100 to compensate them for their time. The Institutional Review Board at Columbia University Medical Center approved the study and participants gave written informed consent (Appendix S3).

The focus group discussions were organized into three parts. First, the moderator briefly discussed participants’ degrees of satisfaction with their healthcare, including their attitudes about medical professionals, insurance companies, hospitals, and government entities roles in healthcare. Next, she initiated a longer discussion of maximalist and minimalist attitudes. This included discussing their responses to the screener question “More, more expensive, newer medical tests and procedures are usually better” and asking about instances in which they have felt over-treated and under- treated. Lastly, in order learn how participants respond to messages that challenge maximalist attitudes, the moderator solicited responses to popular media articles and editorials reporting recommendations to screen less, as well as a Journal of the American Medical Association editorial about the potential harms of cancer screening [34]–[43]. These conversational prompts mentioned recommendations to selectively limit prostate-specific antigen (PSA) testing, mammograms, pap smears, and annual physical exams; over-diagnosis of breast cancer and over-treatment of prostate cancer. We assumed that participants would be familiar with these procedures, thereby minimizing the need for explanation from the moderator.

The focus group discussions were observed by the authors. Audiotapes were professionally transcribed. The authors reviewed the transcripts independently and analyzed them jointly, in conjunction with the moderator. Ellipses in quotations from participants included below indicate that the authors have edited for expediency and clarity. Because focus group techniques are appropriate for assessing similarities among participants rather than differences between them, our analysis includes only themes that arose consistently across participants in all groups [44].

Results

Discussion

We found that participants placed enormous value on testing and screening. They reacted with hostility to messages recommending fewer procedures, using highly charged words like “stupid” and “dumb.” The rate of false positives was not a deterrent to testing. However, rather than wanting all medical interventions, participants were suspicious of pharmaceuticals. Several reported allowing physicians to write prescriptions that they did not fill or take. DTC advertisements actually contributed to their wariness. Although previous research has shown that DTC advertisements prompt requests for prescriptions, this wariness and reported non-adherence are consistent with other findings that patients choose not to follow pharmaceutical regimens in order to protect their health [45]–[48].

We acknowledge several limitations to our research. Qualitative methods do not indicate the proportion of individuals holding a given attitude. Nonetheless, the attitudes we found occurred consistently in all focus groups, indicating that they may reflect attitudes held more broadly by privately insured, middle-aged, financially stable Americans who are in contact with the healthcare system but are not seriously ill. We specifically sought to investigate attitudes among those types of people, and we cannot draw conclusions about other demographic groups or insurance cohorts. As discussed below, our findings raise timely questions about how other groups’ attitudes may differ. Furthermore, our findings are based on opinions stated by individuals in group settings. We did not observe clinical encounters or health-seeking activities. We assume that individuals’ stated opinions differ from their actual behaviors, but we cannot estimate the extent of those differences. In addition, focus groups provide only a snapshot of participants’ opinions. We cannot draw conclusions about how opinions change in response to political and economic events or in response to personal events such as illness, aging, and deaths of family members or financial setbacks. Finally, we did not collect data on race in our first focus group in Tarrytown NY and therefore cannot provide complete racial data for the sample. However, as stated in the methodology section above, focus groups are appropriate for assessing similarities among participants rather than differences between them. We therefore do not discuss race, gender, age, or geographic differences in our analysis and believe that the missing information about race does not significantly weaken our findings.

The limitations of our study suggest questions for future research, particularly about uninsured people’s attitudes about medical intervention. Uninsured and intermittently insured Americans use less healthcare than insured people do, regardless of income [49]–[51]. These lower rates of utilization are almost certainly associated with insurance status, rather than indicating any preference for less medical intervention. In fact, when previously uninsured people enroll in Medicare at age 65, they begin to use more healthcare than their previously insured peers [52]. Those higher rates of use by the newly insured likely reflect the poorer health status of uninsured people [53], [54]. But just as our focus group participants expressed opinions that reflected their experience as insured, financially secure participants in the healthcare system, future research should investigate how lack of insurance shapes others’ attitudes about overuse and appropriate care. Do uninsured people fear scarcity and therefore develop more maximalist attitudes about medical interventions? Or does lack of insurance make them more circumspect about the healthcare system in general and therefore less maximalist in their attitudes? These questions will become particularly important as millions of previously uninsured Americans enter the healthcare system given that the United States Supreme Court upheld the Patient Protection and Affordable Care Act and Barack Obama was reelected president in 2012.

Our discussions revealed how our participants’ attitudes towards pharmaceuticals differ from attitudes towards testing and screening. Future research on the insured, the uninsured, and the newly insured should identify and explore differences among attitudes towards other medical interventions. The analysis should include distinguishing between diagnostic testing and routine screening, between different modes of drug delivery, and between out-patient and in-patient procedures.

Our findings highlight the complexity of addressing the overuse of testing and screening. While patients are skeptical about pharmaceuticals, this perspective cannot necessarily be extended to testing and screening. Pharmaceuticals can give patients quick negative feedback in the form of side-effects, whereas the side-effects of testing and screening are not necessarily immediately apparent. Moreover, DTC pharmaceutical advertisements often list long rosters of side-effects, something not required for testing and screening procedures. Recommendations to test less effectively tell patients that they should have less information, which conflicts with their active, engaged, information-seeking roles. Despite the limited numbers of clinically relevant findings from screening procedures, no one wants to be among the tiny percentage of people who could have “caught it early.” The anxiety associated with false positives pales in comparison to the fear of uncertainty, disease and death. Furthermore, system-wide healthcare costs do not matter to individuals concerned about their own and their families’ health. In fact, discussions of costs may only reinforce suspicions that recommendations to test and screen less are designed to protect payers rather than patients.

The suspicion that guidelines are actually efforts by the federal government to cut costs suggests that the writing and distribution of guidelines needs fundamental alteration. First, both the professional medical associations that produce most guidelines and the USPSTF must clearly and continually communicate their independence from government payers. While that independence may be obvious to medical professionals and policymakers, it is not obvious to the public. Second, the guidelines writing process must engage healthy people who are at risk of excessive testing and screening. The Institute of Medicine’s (IOM) standards for improving guidelines recommend engaging disease-specific organizations in guideline development [55]. But disease-specific organizations can be expected to advocate for more screening, just as they advocate for more research funding [56]. Healthy people seeking preventive services, by contrast, are not represented by organizations and their interests are harder to define. Guidelines writing committees must therefore consider diseases-specific organizations’ biases and develop methods to include members of the public who are not patients. The USPSTF has invited public comments on its draft recommendations only since 2011. Whether those public engagement efforts include non-patients remains to be seen. Finally, the IOM does not discuss strategies for communicating guidelines beyond the medical profession. But given patients’ information-seeking activities, accessibly written guidelines should be published and publicized in the places online where people already seek information, rather than only in medical journals or on the USPSTF website.

Patients are not solely responsible for overuse. Even affluent, well-educated patients tend to see physicians as paternalistic and intimidating [57]. But we cannot rely solely on physicians to police overuse, just as we do not rely on them to police their own conflicts of interest [58]. Efforts to address overuse must involve professional medical associations, hospital systems, payers, and medical schools in modifying fee-for-service payment systems, enabling better coordination of care, and integrating lessons about overuse into training and continuing education. But the preferences of active patients nonetheless merit attention. Both the mistrust of pharmaceuticals and the enthusiasm for testing and screening reflect individuals’ efforts to take care of their health. The challenge is to engage patients in understanding the connection between over-testing and over-treatment, to see both as detrimental to their health, and to actively choose to do less.

Supporting Information

Acknowledgments

The authors thank Susan Chimonas for her comments on an earlier draft.