Primary Care for Elderly People: Why Do Doctors Find It So Hard?

Abstract

Purpose:Many primary care physicians find caring for elderly patients difficult. The goal of this study was to develop a detailed understanding of why physicians find primary care with elderly patients difficult. Design and Methods: We conducted in-depth interviews with 20 primary care physicians. Using an iterative approach based on grounded theory techniques, a multidisciplinary team analyzed the content of the interviews and developed a conceptual model of the difficulty. Results:Three major domains of difficulty emerged: (i) medical complexity and chronicity, (ii) personal and interpersonal challenges, and (iii) administrative burden. The greatest challenge occurred when difficulty in more than one area was present. Contextual conditions, such as the practice environment and the physician's training and personal values, shaped the experience of providing care and how difficult it seemed. Implications:Much of the difficulty participants experienced could be facilitated by changes in the health care delivery system and in medical education. The voices of these physicians and the model resulting from our analysis can inform such change.

Decision Editor: Laurence G. Branch

Book Reviews

Practice Concepts

The Forum

America is in the midst of a major demographic shift that will have repercussions for health care for some time to come (Manton and Vaupel 1995). Currently, people aged 65 and older account for 30–40% of primary care physician visits (Schappert 1999; Stafford et al. 1999; U.S. Bureau of the Census 1996). As the rapid aging of the population continues toward its projected midcentury plateau, general internists and family physicians will be called upon to provide primary care to an increasing volume of elderly patients. At present, many of these physicians are unwilling or unable to do so. Surveys of primary care physicians show that between 30% and 50% limit the number of elderly patients they admit to their practices (AARP 1995; Cykert, Kissling, Layson, and Hansen 1995; Damiano, Momany, Willard, and Jogerst 1997; Geiger and Krol 1991; Lee and Gillis 1993; Lee and Gillis 1994). To meet the primary care needs of the aging population, researchers and policy makers must understand and respond to this phenomenon.

There have been surprisingly few attempts to determine the reasons physicians limit the number of elderly patients in their practices, and results have been inconsistent. Studies have focused on concerns with Medicare fees and documentation requirements, which clearly are sources of frustration for physicians (Cykert et al. 1995; Geiger and Krol 1991). However, frustration with Medicare seems to explain only a small part of physicians' willingness to provide care to elderly patients. In one survey of primary care physicians, 65% reported that low Medicare fees were a very important problem in their practices, but this did not predict whether or not they limited the number of Medicare patients they accepted (Damiano et al. 1997). Some demographic variables are associated with practice limitation, including primary care specialty (Lee and Gillis 1993; Lee and Gillis 1994), urban location (Cykert et al. 1995), and type of practice (solo, single specialty, or multispecialty; Cykert et al. 1995). Studies have generally not measured psychosocial or practice level variables that might contribute to physicians' perceived need to limit geriatric practice and no previous qualitative studies have addressed these issues.

Though data are sparse, there are suggestions in the literature that primary care physicians find elderly patients more difficult to treat (Damiano et al. 1997). This may have to do with medical training. In a national survey, only 60% of general and/or family practice physicians and 50% of general internists felt that their formal medical training did a good or excellent job of preparing them to manage care needs for frail elders (Cantor, Baker, and Hughes 1993). Another survey of primary care physicians in Virginia found that fewer than half thought their current geriatric knowledge was adequate (Perez, Mulligan, and Myers 1991). Characteristics of the health care system may also contribute to physicians' willingness to provide care to elders. In a survey of Canadian family physicians, respondents endorsed poor reimbursement, time pressure, and inadequate community resources all as sources of frustration in caring for older patients (Pereles and Russell 1996). Although these studies suggest potential contributors to physicians' limitations on practice with elderly patients, a detailed understanding of the problems physicians encounter in geriatric primary care and a clear direction for change are sorely needed.

Given the paucity of data in this area, a research approach that allows in-depth examination of physicians' perspectives is needed. To gain a deeper, more detailed understanding of the key issues, we conducted a qualitative study that explored how physicians view providing primary care to elderly people. This article focuses on the theme that most consistently pervaded the interviews: the increased difficulty of primary care with elderly patients. We present a conceptual model, developed from these data, which suggests vital areas to be addressed to ensure that primary care of elderly people meets current and future needs.

Methods

Design and Participants

We conducted a qualitative in-depth interview study with a diverse sample of 20 practicing general internists and family physicians. The first two respondents were physicians known by one of the authors to have busy internal medicine practices with a relatively high proportion of elderly patients. Subsequently, we selected physicians practicing in the vicinity of Omaha, Nebraska, from a database maintained in the Chancellor's office at the University of Nebraska Medical Center comprising demographic information about all physicians practicing in the state. We used a maximum variation sampling strategy (Kuzel 1999), in which we selected physicians from the list by gender, age, and specialty to compile a sample representing both men and women, internists and family practitioners, and a wide age range. We approached physicians by an introductory letter followed up by a telephone call. In all, we contacted 141 physicians to recruit the 20 participants.

Demographic and practice information about participants is shown in Table 1 . Of the 20 participants recruited, 19 were White and 1 was Hispanic. Eight were women. Ages ranged from 32 to 70 years. Three respondents limited the number of elderly patients they accept into their practices; all three were busy internists with a high volume of elderly patients. In this article, a code letter has been randomly assigned to identify participants.

Procedure

Two of the authors (W. A. and H. M.), both physicians, conducted in-depth interviews (Crabtree and Miller 1999) with the participants. The average interview lasted 50 min, with a range from 30 to 120 min. Participants appeared to respond in an equally open and forthcoming way to both interviewers. We examined interview content for systematic differences in responses to the different interviewers and were unable to detect any. We were also unable to detect any systematic differences between the responses of the two participants who were previously acquainted with the interviewer and the others, who were not. The interview questions were broad and open ended. We invited the participants to relate personal narratives regarding experiences with geriatric primary care with the initial "grand tour" question: "Please tell me about some of your experiences taking care of elderly people." We then asked them to relate both satisfying and frustrating experiences. The existing literature suggests certain topics important for physician satisfaction that may relate to their views on care of elderly patients. If these did not come up spontaneously, we asked participants to comment on them. We used such prompts for reimbursement issues (Cykert et al. 1995; Damiano et al. 1997; Lee and Gillis 1993; Lee and Gillis 1994), time pressure (Burdi and Baker 1999; Lewis, Prout, Chalmers, and Leake 1991; Linn, Yager, Cope, and Leake 1985; Linzer et al. 2000; Mawardi 1979), confidence in addressing geriatric syndromes (Cantor, Baker, and Hughes 1993; Perez, Mulligan and Myers 1991), community resources for elderly patients (Pereles and Russell 1996; Siu and Beck 1990), the doctor–older patient relationship (Adelman, Greene, and Ory 2000; Bates, Harris, Tierney, and Wolinsky 1998; Greene 1993; McMurray et al. 1997; Roter 1991), and frailty and death (Krakowski 1982; Morrison, Morrison, and Glickman 1994). We asked the physicians to describe how the doctor– patient relationship was different with older and younger patients. Other questions did not ask physicians to compare and contrast experiences with older and younger patients, but they frequently made such comparisons when discussing their experiences.

Analysis

We audiotaped and transcribed interviews verbatim. A multidisciplinary team including 2 physicians, a nurse practitioner, a medical anthropologist, a medical sociologist, and a psychologist then analyzed these data. We used a three-stage coding process derived from the sociologic tradition of grounded theory (Strauss and Corbin 1998). In the initial open coding stage, each team member independently read each transcript several times and marked key phrases, terms, or sentences. We then met and discussed the interviews in detail, sharing insights from our various disciplines and assigning topical codes to the text of the interviews. We grouped these codes into categories as it became evident which concepts were emerging as keys to understanding physicians' perspectives on primary care with elderly patients. As the analysis proceeded, we compared the content of each new interview to the existing categories and the coding modified accordingly. In the axial coding phase, we developed the categories further and began to define the relationships among them and their possible implications. In the final selective coding process, we developed the conceptual model that is presented here.

We used several techniques common to qualitative research to ensure that standards of rigor were met. To maximize the trustworthiness of our data collection and analysis, we continued recruiting participants until no new major themes were emerging (Patton 1990). In the process of developing codes and interpreting the data, the diversity of the team kept one point of view from dominating and biasing the results (Creswell 1998; Lincoln and Guba 1985). We also routinely searched for disconfirming evidence in the interviews (Patton 1990). We conducted follow-up interviews, also known as member checking (Lincoln and Guba 1985), with 5 of our participants. In these interviews, we gave participants written descriptions of the categories of difficulty and contextual conditions we had developed in the analysis process. In the last member checking interview, we presented the evolving conceptual model, similar to Fig. 1 in this article. We then asked for discussion and feedback. Although not every point of difficulty was important to every physician, all strongly confirmed the importance of the increased difficulty and the appropriateness of the categories of difficulty presented here.

Results

Overview

Most participants enjoyed their interactions with older patients and emphasized that advanced patient age alone was not problematic. All, however, related experiencing increased difficulty in caring for elderly patients, which fell into three major domains: (i) medical complexity and chronicity, especially patients' vulnerability to adverse events; (ii) personal and interpersonal challenges, including time pressure, communication problems, and ethical dilemmas; and (iii) administrative burden, including more telephone calls and paperwork as well as Medicare's documentation requirements. As illustrated in Fig. 1, these categories overlap and interact. For example, a medically complex situation may lead to nursing home placement, which challenges the doctor–patient–family relationship and increases administrative burden. Fig. 1 also illustrates that the difficulty was experienced in the context of the practice environment and seen in the light of the personal characteristics of the physician. Although the nature of the difficulty was similar for physicians with small and large volumes of elderly patients, it had more impact on physicians with a high volume.

The Nature of the Difficulty

Medical Complexity and Vulnerability to Adverse Events

Elderly patients were seen as medically more difficult to care for than younger people. They had more medical conditions, and their illnesses often presented atypically. They were more likely to become seriously ill and were vulnerable to rapid declines in their condition. Multiple medications and the risk of adverse medication reactions also contributed to the difficulty. Participants described diagnostic and therapeutic uncertainty as well as anxiety about causing unintended harm to patients. An internist who had inherited a large volume of elderly patients from a retiring colleague remarked, "Their problems were kind of special compared with the general medical population… The thing that impressed me the most is, their homeostatic mechanisms didn't leave much room for goof-ups" (Dr. C). Another internist described a patient's adverse drug reaction: "I thought, here is something I have done to hurt this patient by giving him this medicine… What other things can I do to hurt people? … In a lot of ways you've got to be very careful with things" (Dr. I).

Elderly patients often have chronic conditions with symptoms that are difficult to control. "In general, they have more things wrong with them and in general, they're on way more medication and in general, they don't feel good most of the time and they don't sleep at night and they are deteriorating …" (Dr. P). This can lead to a disinclination to see these patients:

Every time they come in something's aching or hurting or … "My back's a little sore" or "I'm a little stiff, I don't have the energy I used to," "Well, maybe I'm a little depressed." Sometimes they get to be those people that you look at the list and go, "Ah-h-h-h, doggone, that name again" (Dr. E).

Many participants described frustration at their perceived inability to help with older patients' chronic conditions. One family physician related,

No matter what you do, they hurt. No matter what you do, they get agitated. And no drug exists to stop a cognitively impaired patient from falling. You know, yeah, that's frustrating. You bet it is. But hey, somebody needs to take care of these folks (Dr. L).

An internist reported,

You know, there are some patients that they're always going to have the same problems year after year after year. They're not going to be fixed. You know, it's their back pain from their osteoporosis and scoliosis and you can't do anything about it, or they may be a little depressed, but they won't take any medicine, and they're chronically constipated and, you know, sometimes those are the most frustrating (Dr. O).

Medical complexity also had a positive side. Several participants enthusiastically told of satisfying experiences in which they had made a difficult diagnosis and helped patients substantially. Regarding a 96-year-old woman with an atypical presentation of ischemic heart disease, a family physician remarked,

I was able to stabilize her in the hospital, get her feeling good and actually took care of her for another two years or so… She was so grateful that I had been able to find what was wrong with her, and she became a very dear patient to me … so that was a really good experience (Dr. J).

Adjusting to the increased prevalence of chronic illness and the relative infrequency of cures requires a change of outlook on the physician's part. One young internist seemed to be in the midst of this process when she related,

But then I was thinking, I need to think of it in a different frame of mind. More of maybe getting them to understand that this is a chronic problem and what can we do to make them feel better as opposed to fix them. (Dr. O).

This may be an adjustment that not all physicians are able to make. Regarding caring for cognitively impaired patients, Dr. L said,

I mean in general, there's not a lot that medicine can do about that. Our interventions are somewhat limited, so this just adds to this area of medicine. It takes a special kind of mind set, a special kind of provider to grapple with those on a day-to-day basis.

Personal and Interpersonal Challenges

Communication barriers, especially those resulting from hearing problems or cognitive impairment, contributed to difficulties with history taking, treatment, and the quality of the relationship. One physician remarked, "[There are] lots of various obstacles to getting the whole story, getting the truth out and sometimes 'cause they don't remember and sometimes they just don't think it's important and sometimes they're just in denial of what's really wrong" (Dr. P). Another commented, "It's sometimes frustrating when you've got an older person who can't hear and won't wear hearing aids and you know, you have to shout so loud that everyone else in the building hears everything you say to them" (Dr. J).

Families often became involved in the care of frail elders. For the physicians we interviewed, this had both positive and negative implications. Involved family members increased the safety of medication use and the home environment. However, their participation increased the length of office visits, the complexity of the doctor–patient relationship, and the difficulty of decision making. Friction with families sometimes arose when it was unclear whose responsibility it was to provide personal care:

You know, I don't mind dealing with it as long as the family is going to deal with it too. If they act like it's all my problem to deal with Mom or Dad and figure out, you know, a solution at home for care and you know, that's what's irritating because that's not my responsibility (Dr. G).

When older patients were unsafe driving or living alone but wished to continue to do so, the need to balance safety and autonomy was sometimes difficult.

It's usually a struggle between the family wanting them to move to a more supervised level of care or out of their home and the … parents not wanting to do that, so it's usually a negotiating process, usually a slow process (Dr. Q).

On the whole, physicians found caring for elderly patients who are dying one of the most important and meaningful aspects of practice. Most, however, had experienced serious conflicts with family members in this area. One related,

The most difficult thing … is just the actual end of life issue when the patient is in the hospital and you have a family there, and the family doesn't get along, and then trying to be a mediator within the family to get some kind of good consensus (Dr. K).

Physicians were challenged to examine their values and balance them with the family's.

When you internally feel like a family member is making decisions on behalf of the patient that are maybe prolonging the patient's misery … then we are kind of put into the awkward position of having to carry out what they want (Dr. L).

These decisions are frequently emotionally charged. "Our culture is so afraid of death, that usually it isn't that peaceful. It's just wrought with being torn apart by just an incredible amount of argument and bickering between family members. It's terrible" (Dr. P).

Time pressure was a major issue for participants with a large volume of elderly patients. "That's probably the biggest problem I have right now, is managing my time with the older individuals" (Dr. A). Medical complexity, family involvement, ethical decision making, and communication barriers all made caring for frail elders more time consuming. History taking was slower, physical examinations took longer, and mobility impairment slowed down the flow of office activities. Medicare's extensive documentation requirements and lengthy claims processing also make heavy demands on physicians' time, as do paperwork and phone calls from home health agencies and nursing homes. "[If you see 15 elderly people] it takes time. You feel like you've done a big day's work. You can see 15 young people with sore throats and be done in an hour" (Dr. M). In the current health care environment, where efficiency is highly valued, this presents a major difficulty for physicians. "You have to have sheer volume with Medicare patients but Medicare patients also require most of your time because they need so much, so it's a hard situation out there" (Dr. I).

Administrative Burden

Nearly all of the physicians felt they spent too much time, effort, and worry on Medicare regulations. Claims were often denied for apparently trivial reasons and resubmitting them required substantial personnel time. In some situations, "The amount of return is less than the effort made in acquiring the reimbursement" (Dr. Q). Medicare regulations seemed particularly frustrating because they did not seem to relate to the quality of care.

It has nothing to do with the care the patient got… You go through a whole long physical exam of stuff that is irrelevant really to the problem at hand, … and spend more time on the paperwork than you do taking care of the patient. And so that's extremely frustrating as well as stupid (Dr. M).

The threat of legal action from Medicare adds additional anxiety to geriatric primary care:

You wake up in the middle of the night in a cold sweat thinking, "Oh my God! The Office of Inspector General showed up at my office today and wants to go through every file in my charts!" So it's sobering to know what Medicare could do to you and your practice if they chose to. And I'm of the opinion they could probably find improper documentation/coding/billing in every office in this country (Dr. L).

In general, Medicare was seen in an adversarial light, increasing the burden of providing primary care to older patients.

Multifaceted Complexity

In the initial coding process, complexity and difficulty were noted in all 20 interviews. As we returned to the data for the axial coding process, it was evident that participants rarely felt overwhelmed by difficulty in one area alone. In every interview, however, there was a discussion of at least one situation where an elderly patient's medical needs overlapped with psychosocial and/or administrative difficulty. These were the situations in which caring for older patients became seriously problematic. "It's just that you have a number of these things happening all at the same time. Physicians are human. It wears on you" (Dr. N). When considering Fig. 1 in a member-checking interview, one participant remarked,

This helps me understand why these patients are so hard. It's OK if they just have difficulty in one of these areas, but when there's more than one, and especially in that area (pointing to the model) where they have all three, the difficulty is exponential, or logarithmic or something (Dr. B).

Contextual Conditions

The three-faceted difficulty presented above occurred within the context of the practice environment. There was also a context of personal and interpersonal factors. For instance, a complex medical situation that occurred within the context of a long-term doctor–patient relationship was perceived differently from a complex medical situation in the context of a new relationship. Various constellations of these contextual conditions shaped the experience of providing care and how difficult it seemed. In Fig. 1, the larger circle represents this context.

Personal and Interpersonal Factors

All participants found elderly patients more grateful and appreciative than younger patients. Some also enjoyed hearing their stories and experiencing their wisdom. For many, this mitigated the difficulty of their care.

I enjoy taking care of elderly patients, mostly for the personal interaction with them as opposed to their medical problems. I would look at the medical care of those individuals to be a little more cumbersome than younger people from an operational standpoint. It's harder to do things, more difficult. But the interaction with the individuals is more rewarding I would say (Dr. Q).

When patients are severely cognitively impaired, on the other hand, the limited relationship often made the care seem meaningless. One internist related,

The very severe cognitively impaired people, … I don't find any particular satisfaction in taking care of them. Whatever was … the essence of their humanity is long since gone and I'm tending to a body, which has no hope of recovery and it's hard for me to get real excited and enthusiastic in that setting (Dr. B).

A family physician said, "You have to tell them the same thing every visit. And they don't remember you. It eliminates some of the camaraderie, if you will, with the patient. That's inevitable" (Dr. L).

Physicians' personal characteristics, values, and training also affected how they viewed geriatric primary care. For instance, older physicians felt closer to their elderly patients:

I'm not exactly young myself anymore and so I guess I have a fair amount of good feeling towards the elderly. It's easier for me to identify with somebody who is 75 and has lived through some of the things they lived through or the depression, World War II, raising children, than a very young person with an earring in their nose and ear and their lip and I'm not sure I have much in common with that person (Dr. H).

Some participants felt a social obligation to care for nursing home patients, whereas others did not.

It's not that much fun but I just feel like it's something that I have to do for society, part of my job. I could never do that as a full-time job or even have a larger practice in a nursing home (Dr. G).

The Practice Environment

Certain aspects of the practice environment facilitated or hindered caring for elderly patients. The volume of older patients in the practice had a major impact on how the difficulty was experienced and whether the participant was limiting or was planning to limit the number of new elderly patients. Physicians with a high volume of older patients found it more difficult to incorporate their complex care into the usual flow of work. One internist who had recently cut back her practice related,

The patients are so complex and they take so much time sometimes and they have side effects from medications and phone calls, that yeah, you get overwhelmed. It's just not physically, humanly possible. It just isn't. You would need to have a smaller patient population to do a good job (Dr. P).

The roles of office staff members and their relationships with the physician and each other also affected how well they were able to cope with a high volume of elderly patients. One geriatrician remarked, "[Nurses] can make you or break you. I mean, if they left, I'd have to leave" (Dr. A).

Community resources were generally perceived as inadequate. None of our participants had ready access to social workers in the office, so arranging home health care, adult daycare, and other community services added to the difficulty of primary care.

You know, there's no one place, no one clearing house that you can go for those kind of services. You just have to kind of make a patchwork quilt almost of that. It'd be nice to have someplace where you can have one phone call … and say, here's my patient's needs, what can you provide for us? (Dr. C).

Caring for patients in nursing homes was generally regarded as difficult and unpleasant. Prominent difficulties with nursing home care included the logistics of providing care, communication with nursing home staff, and dysfunctional regulations.

Their regulations are ridiculous, you know, especially the one where they have to call you if somebody scrapes their elbow. Nursing home visits usually aren't the most stimulating … and you have to sift through charts that you're not familiar with and where anything is and I don't know (Dr. G).

Although caring for frail elders is difficult and time consuming, Medicare reimbursement is lower than private insurance. Low fees did not contribute to the difficulty of geriatric primary care, but clearly influenced how physicians responded to it.

If you told me that I had to run this place on the basis of what I get from Medicare, I would have to tell you I couldn't do it, which is kind of sad, because they claim that they're bankrupt and everything. Where in the hell are they spending their money? They sure ain't giving it to me (Dr. F).

The mismatch between patient needs and the level of reimbursement generates a conflict between the physician's role as healer and his or her role as business person or employee.

You owe it to your employer to be as productive as you can but you also owe it to your patient to be as helpful as you can and sometimes the two masters can't be served at the same time (Dr. C).

The imbalance between the time required and reimbursement sometimes leads to physicians limiting geriatric practice even if they enjoy it.

In the real world, communication takes time, whether you're communicating with an elderly person who has a delay between the time that you give them a question and the time they give you an answer, or those that can't understand or deal with complex questions… It takes longer to take care of patients like that. You superimpose upon this slow reacting patient a worried … family member who has a number of questions… It adds more time to the office visit and the way Medicare is paying us for office visits. From an economic standpoint it just does not make sense to take care of old people (Dr. C).

Discussion

This study, using face-to-face interviews with practicing physicians, gives an in-depth look at the difficulty involved in providing primary care to elderly patients. The voices of these physicians and the framework we propose for understanding the difficulty they described can inform future efforts to meet the health care needs of our aging population. On the whole, participants enjoyed interactions with their elderly patients, but the high prevalence of multiple medical problems and declining physical and cognitive function among these patients gave rise to interacting medical, interpersonal, and administrative difficulty. Physicians struggled to deal with the difficulty in a practice environment that was not set up to provide the support and resources these patients needed.

We are by no means the first to recognize the mismatch between the chronic care needs of our aging population and the acute orientation of our health care system (Kottke, Brekke, and Solberg 1993; Wagner, Austin, and Von Korff 1996). This study vividly demonstrates the real impact of this mismatch on the daily practice of medicine. In so doing, it strongly supports the need for health system change. The recent Institute of Medicine report, Crossing the Quality Chasm, calls for efforts to improve health care by approaching it as a "complex adaptive system" (Institute of Medicine 2001). To effect positive change in such a system, it is essential to recognize which elements can change and which cannot. The three-faceted difficulty at the center of our model must be regarded as a fixed element of the system. Caring for chronically ill elders is and will remain complex and time consuming. There is great potential for positive change in the context in which care is delivered, however.

Our results suggest potential for change in practice organization, health care policy and medical education. In the area of practice organization, a number of interventions to facilitate primary care of chronically ill elders have been proposed and a few have been studied (Boult, Boult, Morishita, Smith, and Kane 1998; Leveille et al. 1998; Schraeder, Shelton, and Sager 2001; Netting and Williams 2000; Wagner et al. 1996). The participation of nurse case managers in primary care practices, for instance, has shown benefits in elderly patient mortality and physician satisfaction (Schraeder et al. 2001). As yet, however, such interventions have met with very little acceptance by health care organizations or third party payers (Boult, Kane, Pacala, and Wagner 1999; Wagner, Davis, Schaefer, Von Korff, and Austin 1999). None of our participants had access to such personnel. Perhaps the greatest interpersonal challenge our participants experienced was the expansion of the doctor–patient relationship to include family members and other caregivers. Programs that facilitate communication between families and staff in the nursing home setting have shown great promise (Pillemer, Hegeman, Albright, and Henderson 1998; Specht, Kelley, Manion, Maas, Reed, and Rantz 2000). A similar intervention to enhance doctor–patient–family communication could be extremely helpful in the primary care setting.

Regarding health care policy, participants confirmed that Medicare documentation requirements are onerous and fees too low. Simplification of documentation requirements and increased reimbursement for complex nonprocedural care would clearly facilitate caring for elders. Participants also found the infrastructure of support services inadequate and difficult to access. Policy directed at improving community resources to meet the needs of chronically ill elders would also be extremely beneficial.

Changes in medical education could have important impact on physicians, who are themselves modifiable elements of the health care system. On the whole, participants felt confident managing specific illnesses, but lacked confidence in dealing with geriatric issues, such as vulnerability to adverse medical events and cognitive impairment. They experienced the greatest difficulty when the medical problems overlapped with interpersonal challenges and administrative burden. Despite the long recognition of the demographic imperative, few medical schools have mandatory geriatrics rotations and residencies devote minimal time to geriatrics training (Association of Professors of Medicine 2001). With additional training, physicians could become more skilled and comfortable with the special needs of elderly patients.

This report has both strengths and limitations to consider. The qualitative format allowed participants' views to be explored in depth, adding important information to our understanding of primary care for elderly patients. Because of its intensive nature, however, a qualitative study can include only a small number of participants. Although we found striking consistency in the main themes, it is possible that our participants were systematically different from nonparticipants or physicians in other locales. Larger quantitative studies will determine the generalizability of our findings.

Although primary care for elderly people is rewarding and enjoyable, it is also complex, difficult, and time-consuming. Physicians alone cannot meet the wide range of needs these people have in the current practice environment. Our findings suggest that changes in practice organization, health policy, and medical education will be needed if primary care physicians are to care for a larger volume of elderly patients effectively.

We are most grateful to all the physicians who donated their time for the interviews. We also thank Jeff Susman, MD, Kurt Stange, MD, and Lynn Meadows, PhD, for their helpful critiques of earlier drafts of this article; John Creswell, PhD, for methodologic advice; and Linda Ferring for manuscript preparation. This study was approved by the Institutional Review Board at the University of Nebraska Medical Center.

References