Prevalence and Sources of Patients' Unmet Expectations for Care

Methods

Sampling of Practices and Patients

Our study was done in three general internal-medicine practices in a mid-sized city in northern California. Two practices were branch offices of a large group-model health maintenance organization. Within these two practices, all patients were members of the health maintenance organization and received care on a capitated basis; the 10 physicians and three nurse practitioners were each paid a salary. The third practice consisted of 8 general internists (2 of whom had additional subspecialty training) and one nurse practitioner, all of whom worked in a single private office and received a combination of salary and productivity bonuses. Patients in this practice were insured under various prepaid and fee-for-service plans. All three practices attempted to schedule patients with their own regular practitioners when possible.

We selected these practices to represent the dominant models of health care in northern California, an area heavily penetrated by managed care. Because managed care providers may have financial incentives to restrict medical care, managed care settings are potential sites of conflict between patients' expectations and the ability of clinicians to meet those expectations.

During 36 half-day sessions (12 sessions per practice) scheduled between October and December 1994, trained research assistants approached patients in waiting rooms and encouraged them to complete a brief form that was designed to identify and thus exclude patients presenting for routine checkups. [In preliminary research, we had found that such patients were relatively unlikely to have unmet expectations.] The form asked two questions: 1) Do you have a new or worsening problem that you wish to discuss with the doctor today [yes or no]? and 2) How concerned are you that you might have a serious disease or condition that has not yet been diagnosed (not-at-all to extremely)? Of the 1221 patients who completed the form, 804 had a new or worsening problem or were at least moderately concerned about an undiagnosed condition and thus were eligible for further study. Of the 804 eligible patients, 688 (86%) agreed to participate in the study and completed a post-visit questionnaire that asked about demographic characteristics, recent health care utilization, health status, satisfaction with the visit, and perceived omissions of care.

The core of the written survey was eight questions about perceived omissions of care. Patients were given the following instructions:

When people go to the doctor, they usually bring some thoughts about how he or she can be of the most help. Sometimes, however, the doctor may not be able or willing to do exactly what the patient wants … These questions are about things you felt were necessary for the doctor to do today but which (for whatever reason) didn't happen …

These instructions were followed by questions about sets of possible omissions involving physician preparation for the visit; history taking; physical examination; laboratory testing or diagnostic imaging; prescription of medication; referral to specialists; information, counseling, or personal help; and “anything else you felt was necessary or might be necessary for the doctor to do today but which didn't happen.” The wording of the instructions reflected semantic concerns reviewed elsewhere [5] and was intended to convey a broad definition of “patients' expectations,” emphasizing expectations as things that patients value. Of the 688 respondents, 125 (18%) reported one or more omissions and 108 (86%) consented in writing to participate in a telephone interview within several days of the visit.

Telephone Interviews

Within 1 to 7 days of the visit, one investigator conducted telephone interviews with 90 of the 108 patients who had reported one or more omissions on their questionnaire and who had provided their home telephone numbers as part of the consent process. The interviewer identified himself as a “researcher from UC Davis” but not as a physician (he did say that he was a physician if specifically asked). The interviewing approach was modeled on the critical incident technique [6], which focuses on patients' accounts of events that have actually happened rather than on generalizations or opinions [7]. In this case, the critical incident was the perceived omission of care. Thus, the opening question was, “You mentioned on your questionnaire that you were hoping that the doctor would [perform a particular intervention], but that didn't happen. Can you tell me more about that?” Patient interviews averaged 15 minutes in length.

Data Transformation

Of the 90 interviews, 88 were successfully tape recorded and transcribed. To locate key themes related to the genesis of patients' expectations for care, we used an iterative process, which began with a careful reading by three investigators of approximately 20 randomly selected transcripts. In a series of meetings, the research team developed by consensus a preliminary coding scheme. Codes were developed inductively, consistent with the grounded theory approach used in sociology [8, 9]. We focused on the sources and the process of development of patient's expectations. Examples of codes and their definitions are shown in Table 1.

In the second stage of the analysis, the lead investigator applied the coding scheme to 50% of the transcripts (n = 44) by using a qualitative research tool, the Ethnograph [10]. A health psychologist independently coded 20% of these transcripts. The κ coefficient for inter-rater reliability at the coded-segment level was 0.39, indicating moderate agreement beyond chance. The lead investigator then coded the remaining transcripts, and the study team as a whole reviewed the results and reached a consensus on major concepts and themes.

Results

Patient Characteristics

The patients who completed the post-visit survey (n = 688) had a mean age of 51 years and a mean education level of 14.2 years; 63% were female and 83% were white. Six percent were visiting a physician for an initial, comprehensive evaluation; 49% were making follow-up visits with their own physicians; 38% were being seen on an urgent basis by someone other than their own physicians; and 7% were making other types of visits.

In terms of sex, education level, and the percentage of patients who were making follow-up visits to their own physicians, the group of patients who reported at least one omission of care (n = 125) was similar to the group of patients who reported no omissions of care. However, patients with unmet expectations were younger (mean age, 47 years compared with 52 years; P = 0.001). The likelihood of reporting one or more omissions was lowest among the 188 patients aged 65 years and older (11% reported one or more omissions), highest for the 127 patients younger than 35 years of age (23% reported one or more omissions), and intermediate for the 365 middle-aged patients (20% reported one or more omissions) (P = 0.007). Patients who reported at least one omission were less likely than patients who did not report an omission to provide an “excellent” rating for the visit overall (35% compared with 75%; P < 0.001).

Prevalence of Unmet Expectations

A total of 125 patients (18.2%) reported at least one unfulfilled expectation within eight broad categories of care (Table 2). The prevalence of specific perceived omissions ranged from 2.8% (for failure to provide needed information, counseling, or personal help) to 5.5% (for omission of a component of the physical examination) (Table 2). Among patients who reported at least one perceived omission, the mean number of perceived omissions ±SD was 1.9 ± 1.17.

Sources of Expectations

In the telephone interviews, patients identified four major sources of their unmet expectations: somatic symptoms, perceived vulnerability to illness, previous experience, and transmitted knowledge. Ninety-five percent of patients cited at least one of these sources, and 73% cited two or more.

Somatic Symptoms

Somatic symptoms influenced the expectations of 74% of the interviewed patients, who described their symptoms in terms of four dimensions: intensity of symptoms, functional impairment, duration of symptoms, and perceived seriousness of symptoms (Table 3). Patients often linked salient symptom dimensions to specific goals of care-seeking. For example, patients who had severe or disabling symptoms frequently sought empathy or relief, whereas those with frightening symptoms sought reassurance that they did not have a serious disorder.

Forty-two percent of patients commented on the intensity or severity of their symptoms. As justification for patients' expectations, symptoms often spoke for themselves (“There is a red hot poker that is drilled through my leg and it's real painful at times.”) Other patients explicitly contrasted the severity of their symptoms with the clinician's apparent nonchalance. Distress about symptoms was compounded by concern that the physician was not taking the problem seriously: “Yes, it's my entire body. My extremities, connective tissue, everything is sore and painful. … It warranted more than, well, let's just talk about it in 2 weeks.”

About one sixth of patients emphasized the functional consequences of their symptoms (Table 3). These patients focused on resuming short-term responsibilities (“I'm having a really big problem with my back right now and I'm missing work and would like to go ahead and get a refill on my Flexeril”) or achieving long-term goals, such as finishing school or having a baby. Implicit in these comments was the belief that the physician had underappreciated the extent of the patient's distress.

Almost half of the patients focused on “how long” instead of “how much” (Table 3). Prolonged symptoms not only taxed patients' ability to cope (“My leg is not getting better. … I can't stand it”); they also raised questions about therapy (“If something doesn't work, why keep using it?”). These patients viewed refractory symptoms as a signal that the clinician was off-track.

More than one third of patients expressed concern about the implications of their symptoms (Table 3). The prospect of serious disease triggered anxieties that were not assuaged by attempts to relieve symptoms; these attempts were often dismissed as blind empiricism. “I want to know,” said one patient, “‘is this exactly what is wrong with me, or is there something else?’ before I start taking medication that might mask what really is the problem.” In this case, the physician misread the patient's primary goal, which was not relief from hand cramps but reassurance that a serious disease was not lurking behind them.

Perceived Vulnerability to Illness

The degree to which patients attended to sensations, interpreted them as abnormal, and developed concerns about them was related to each patient's unique vulnerabilities. Forty-four patients (50%) mentioned at least one specific vulnerability related to one of five factors: aging, a previously diagnosed condition, a family history of illness, personal lifestyle factors, and the utterances of medical office staff (Table 3).

Patients' recognition that they were growing older affected their responses to specific symptoms (“… at my age I'm just kind of getting worried [about my backache]”) as well as their perceived susceptibility to illness in general (“… I'm in my mid-life and I'm thinking a little bit about my age and health. … I thought maybe I should have some kind of … heart test and a cholesterol test, blood test, urinalysis”). Age-related concerns were reported by 5% of interviewed patients younger than 35 years of age, 15% of those 35 to 64 years of age, and 14% of those older than 65 years of age. Thus, attaining a particular chronological age was neither necessary nor sufficient for the development of age-related expectations.

About one fourth of study participants had a previously diagnosed medical condition that appeared to influence current expectations by magnifying perceived vulnerability to illness (Table 3). Symptomatic and asymptomatic patients reacted differently. Having had a medical problem in the past could cast a morbid shadow over otherwise minor symptoms, as it did for one patient with a sty who was unhappy when he was advised to use warm compresses rather than to see an ophthalmologist (“I have extremely poor eyes. … Therefore I am always concerned if I have a problem with my eyes …”).

However, previous or underlying illnesses influenced expectations even among persons who currently felt well. As one 67-year-old patient told us, “So I feel like, having a heart condition where I take two drugs for it, that he automatically should listen to the heart and lungs when I go in. I don't feel like I should have to ask.” As a “heart patient,” this patient expected that her special medical vulnerabilities would be recognized and acted on by her physician.

For about 9% of patients, personal vulnerability was derived from a family history of illness, often heart disease or cancer (Table 3). For these patients, diseases that had killed their relatives colored even the most routine medical encounter. One young woman made such concerns explicit: “I went in for some concerns about ovaries … I was having some lower abdominal pain which was obviously not appendix, and my mother, grandmother, and great-grandmother all died of cancer. I brought it up and nothing was really said about it.” Here the patient reflects on family history and feels vulnerable, the physician overlooks or downplays the connection, and the patient feels slighted or endangered.

Even when patients were not concerned about genetic predispositions to disease, recalling the difficult experiences of family members could inspire specific demands for care. As one 46-year old man put it, “Walking pneumonia killed my father, and I know the best way to tell if you have pneumonia is with a chest x-ray, so that is why I'm going back to get this to make sure I haven't caught it.”

Another 9% of patients felt vulnerable because of their own health habits (Table 3). These patients recognized the threat inherent in ongoing substance abuse or stress, and their concerns were often grounded in epidemiologic reality. However, some patients, such as a 38-year-old man who was experiencing work-related stress and wanted the physician “to take a look at my eyes [to make sure] I wasn't having a stroke” seemed to overestimate their absolute risk.

Expectations sometimes came from odd quarters: “I called and gave them my symptoms on the phone [and] they … told me you have to come in because of our meningitis scare. I guess your mind starts to work overtime then.” Nurses, technicians, and medical clerical personnel could wield considerable influence. Casual asides led patients to downgrade their perceived health, leading to changed expectations for care.

Previous Experience

For 42% of interviewed patients, unmet expectations for care were shaped by past experiences with similar symptoms or illnesses or by experiences acquired while caring for others (Table 3).

Direct personal experience influenced expectations for both diagnosis and therapy. Reasoning by analogy, patients interpreted their current symptoms in light of the past. They also used past experiences to develop definitions of an appropriate diagnostic process. As one woman told us, “Well, I think … if you're as tired as I am and everything, that a basic blood test wouldn't hurt. … That's how they found out [the last time I was sick]—my blood [platelet] count was down to 5000 on the night they put me in the hospital for my spleen. It was just a simple blood test.” We do not know enough about the case to judge whether a blood count might have been diagnostically helpful at this visit, but the patient's commitment to this strategy, derived from her own experience, seems plain enough.

Patients also acquired experience while accompanying relatives or friends to the physician. For example, a 68-year-old man said: “The doctor [my wife] was going to before would do … a thorough run up of tests. All that ceased when we went to [the current health maintenance organization].”

Transmitted Knowledge

A final category of influences on the development of patients' expectations included knowledge transmitted from sources other than personal experience. A total of 54% of patients reported having expectations that had been acquired through personal education and training; through conversations with friends, relatives, and physicians and other health care professionals; or through instruments of popular culture (Table 3).

Approximately 16% of our largely middle-class study sample had had formal training in the health professions or were closely related to health care professionals. These patients used their specialized medical knowledge to anticipate or influence their physicians' plans for diagnosis and therapy (“I work in veterinary medicine and we would have routinely done a blood count”). They also tended to react harshly to perceived medical omissions (“I'm an RN, and when someone presents with these symptoms I think I would be inclined to at least examine them.”) To these medically sophisticated persons, failed expectations were not merely a source of disappointment but hinted at medical incompetence.

Similarly, patients whose close friends or family members were health care professionals drew on the expertise—real or perceived—of others. Friends, family, and acquaintances were important sources of medical information when they had specialized medical training (“… my mother who is a nurse said that it sounded like an infection because of all the other things that were going on”) and even when they didn't (“… my girlfriend recommended [that I see] a dermatologist since she had seen one herself who she liked a lot”). Patients did not generally disclose their sources of information to the clinician. Only when providers appreciated the influence of informal counselors could they address patients' expectations directly.

Patients also drew on the authority of health care professionals other than the physicians they were currently seeing. More than 40% of patients reported on the salience of pronouncements, admonitions, and asides delivered by previously encountered physicians and nurses (Table 3). When the advice of trusted health professionals (often a physician the patient had seen in the past) was challenged, patients became suspicious. One elderly woman stated, “Dr. X thought it was very important once a year that I have that test where they put the dye and you ride the treadmill and then you go into nuclear medicine … [But now] suddenly I don't need it. Now was it because they wanted the money from the insurance or was it because I really did need it? You know, which was it?”

To our surprise, references to the media were relatively uncommon (7%; Table 3). Those patients who did mention the media tended to discount its influence. As one patient pointed out, however, the media's influence may be so pervasive that it is often difficult to identify it as a separate factor (“I didn't necessarily read [about needing regular blood tests] in the paper but just because of cholesterol; we live in a high fat society”). In this statement, the patient articulates how preventive health ideology has become part of the social canon, a media-dependent process.

Discussion

Our study both documents the prevalence of patient's unmet expectations for care in office practice and elucidates the multifactorial ontogenesis of those expectations. The complexity of the process by which patients develop expectations may lessen the allure of “demand management,” an increasingly popular concept that seeks to discourage health care utilization by helping patients make rational decisions about medical services [12, 13]. However, rationality has its limits. Approaches that do not consider patients' underlying vulnerabilities and past experiences will merely encourage patients to express their symptoms in more urgent terms [14]. A more realistic goal would be to support clinicians' efforts to understand patients' expectations and, when necessary, initiate negotiations around them.

Our data allow us to construct a preliminary model of how patients' expectations develop in the context of symptom-driven clinical encounters (Figure 1). Each patient comes to the physician's office with a unique set of perceived vulnerabilities, past experiences, and stores of knowledge. These antecedents influence the interpretation of symptoms and lead to the formulation of a response that is expected for the practitioner. The practitioner's actual response is then evaluated in light of the patient's expectations.

For many patients, three symptom dimensions (duration of and lack of alleviation of symptoms, functional impairment, and perceived seriousness of symptoms) were especially salient. Inattention to these dimensions may or may not have affected the ability of physicians to reach a correct biomedical diagnosis, but it surely restricted their ability to impart a sense of concern and understanding. Our data highlight the previously recognized gap between the clinician's focus on objective disease and the patient's subjective experience of illness [15]. A rich and multifaceted understanding of patients' interpretations of symptoms, perceptions of vulnerability, and ways of knowing can enable clinicians to meet patients “where they are” and can lead to more productive clinical negotiation [16].

We found that for at least half of the patients interviewed in our study, symptoms and acquired knowledge were filtered through a mesh of perceived vulnerabilities related to age, previously diagnosed conditions, family history, lifestyle, or the casual utterances of health care professionals or medical office staff. Patients who considered themselves to be at risk for specific diseases tended to have broader expectations about history taking, physical examination, diagnostic testing, and therapy. Because patients often do not voice such concerns unless asked about them, physicians may find themselves arguing with patients about a particular clinical strategy (such as whether antibiotics should be used to treat an upper respiratory tract infection) when the fundamental difference turns on risk assessment—for example, the physician may see a “cold,” but the patient thinks of Uncle Charlie, whose fatal case of pneumonia started with the sniffles. Our results argue for the medical relevance of the social and family history. Clinicians hoping to understand their patients' expectations should inquire about a family history of serious illness, anniversaries of significant life events, and previous diagnoses and their meaning to the patient.

Patients' increasing expectations for the application of expensive medical technologies have been cited as an important contributor to increased health care costs [17]. Physicians may feel uncomfortable when patients request costly interventions of equivocal benefit. Yet our data suggest that one of the most powerful influences on patients' expectations is physicians themselves. Physicians can promote inappropriate expectations by prescribing marginally beneficial tests and therapies, couching clinical beliefs as medically authoritative dicta, and giving in to inappropriate patient requests without discussion. In contrast, they may be able to interrupt the development of unrealistic expectations by reducing unjustified practice variation, learning about and sharing with patients some of medicine's inherent uncertainties, and engaging patients as partners in a clinical negotiation [16, 18].

Several patients in our study reported that they received conflicting information from their health care providers. The amount of contradictory advice that patients receive is likely to increase as managed care (and societal mobility) continues to disrupt long-term patient–physician relationships [4]. Policies that encourage continuity of care might not only improve patient satisfaction [19] but also reduce unfulfilled expectations and the clinical conflict that they can cause.

Our data and conclusions should be evaluated in light of our study's limitations. The study was done in three practices (two health care systems) in one geographic region. The sample of patients interviewed was large by qualitative standards, but it comprised only 88 patients. Moreover, we intentionally excluded patients who were presenting for routine check-ups. This strategy probably increased the proportion of patients who had unmet expectations for diagnosis and treatment but eliminated many whose expectations may have centered on prevention.

In summary, we have identified some of the major sources of patients' expectations within three highly managed general internal medicine practices in northern California. By stripping away some of the mystique surrounding patients' expectations, we hope our study will enable physicians to refine their history-taking skills and priorities, to be sensitive to patients' expectations even when these are not made explicit, and to reduce needless clinical conflict. Further study is needed to confirm our findings in other settings, help clinicians elicit patients' expectations efficiently, and evaluate strategies for reshaping expectations when they are unreasonable and in need of change.

From the University of California, Davis, and Kaiser Permanente Medical Center, Sacramento, California; and the University of California, Los Angeles, California.

Dr. Callahan: University of California, Davis, Department of Family Practice, Primary Care Center, Room 2118, 2221 Stockton Boulevard, Sacramento, CA 95817.

Dr. Paterniti: Yale University School of Medicine, Department of Epidemiology and Public Health, 60 College Street, Box 20834, New Haven, CT 06520.

Dr. Dunham: Kaiser Permanente, 2345 Fair Oaks Boulevard, Med 6, Sacramento, CA 95825.

Dr. Lewis: University of California, Los Angeles, Center for the Health Sciences, Room 61-236, Los Angeles, CA 90024.