The Paradox of Primary Care

Stange's masterful series of three articles on the fundamentals of generalist care superbly articulate the complexity of what generalists do. He shows us why the integrative generalist function is so essential, what gets in its way in the United States, and why health care looks different when viewed through generalist versus specialist lenses. In doing so, he distills the frustrations of practicing generalist care in a system designed to reward other types of care into a theoretical foundation we can begin to act on.

Because he has done so well thus far, I would like to challenge him to take on one more layer of paradox. That is--the paradox of patients choosing primary care as their home base, yet not valuing it.

Let me illustrate what I mean through the case of a close relative. At 90, she has had her share of medical issues, yet is basically healthy and extremely sharp. I attribute her continued health in significant part to her access to primary care. She is fortunate, in the super-specialized town of Los Angeles, to have an experienced primary care clinician. She relies on her primary care doctor and values him highly as a person. Yet she takes his knowledge, skill, and availability for granted. Of course he will always be there. Of course he's paid enough. What was so complicated about seeing her at 5 pm one day for a severe bronchitis, deciding she didn't need hospitalization, starting antibiotics, following up on the succeeding days? Of course he cares about her, she thinks; he's a friend and supporter. She is politically active, but how can defending her relationship to her doctor require letters to Congress? Ultimately, she might ask, why does a generalist require high level scientific training? Dr. Stange's articles provide the beginnings of an answer.

Decades ago, we provided evidence on a nationally representative sample of patients (1) that the large majority of individuals turn to generalists when asked to identify a doctor, despite the great preponderance of specialists in US communities. Yet few citizen's groups actively monitored the availability and vibrancy of US primary care during the succeeding years. In the VA system, as another example, more than twice as many veterans chose VA care after the system transformed to a primary care-based design. Prior to transformation, veterans rated the quality of their specialty care highly, yet overall rated VA quality of care as low. After the transformation, VA patient satisfaction has consistently exceeded non-VA satisfaction with care. Yet protests from veteran groups during the shift toward primary care focused on loss of special technical services or programs; few perceived the fundamental benefits of the shift to primary care as the root or foundation for all other care.

We blame insurance companies and others for the way our payment system is structured. But perhaps for all of us, it is easier to pay for something that is a thing than for something that is embedded in a relationship. Can we get patients to fight for something that seems vague and unformed to them? Can we get them to focus on what they gain from scientifically-based generalist care, rather than on potential losses? I don't think we've managed to distill what we do well enough to accomplish these goals yet. I'm hopeful that Dr. Stange's work can be the basis for this next step.

(1) Spiegel JS, Rubenstein LV, Scott B, et al: "Who is the Primary Care Physician?" New England Journal of Medicine 1983;308:1208-12.

Competing interests:   None declared

Any general practitioner working in a busy practice will conduct about 7000 consultations a year with a range of patients. Despite this large number, the GP will always find that patients actually bring small numbers of ANY disease or illness to the practice. For example the GP might be interested in doing an audit of people who had presented with myocardial infarction in the previous year. Surely there would be large numbers of people to study because we know this is a "common problem" people suffer in primary care. Unfortunately no. It is always the case that the numbers are actually very small - often less than 10.

GPs are generalist because they encounter a range of people clustered in small numbers of specific diseases and illnesses. GPs develop skills in dealing with the uncertainties that surround the paradox of the small number.

The paradox is created - in part - by the WEIGHT of the small numbers upon the individual GP. It is likely that each of the 10 people with their infarction were known personally to the GP, and in great detail. This knowledge alters the GP's perception of the smallness of the numbers.

Unlike a hospital colleague who will encounter many more more people with infarction in the same year (the difference is always one of many zeros of order of magnitude), but with less personal impact. Such that the perception of numbers is not affected.

The other part of the paradox is created by having small clusters of disease or illness remain constant in size despite a steady stream of people moving in and out of the locality. GPs realise over time that although they might care for clusters of "only" 10 patients with specific disease or illness a year, it turns out that the clusters change composition from year to year. Therefore, from an experienced GP's perspective the small cluster is bigger than it looks.

Competing interests:   None declared

Primary care, and not the availability of (sub)specialist care is a determining factor of populations’ health. Although specialists, in comparison to generalist, might perform better in the treatment of individual diseases, this does not affect the health of the people and populations with these diseases. Stange and Ferrer present this as ‘the paradox of primary care’.[1] In our view, the paradox is much more in specialist care than in primary care. The way specialization has been incorporated in health care is of major importance. The ‘paradox of specialization’ may help gain further insight in exploring the effectiveness of health care.

Stange and Ferrer correctly explain the paradox of primary care in the discrepancy between ‘disease’ and ‘people or populations with the disease’. Their description of complexity at the level of single diseases, individual patients and populations is helpful to come to a better understanding of the contribution of generalists and specialists. In the (primary care) population co-morbidity and multi-morbidity have become a rule rather than an exception.[2] This explains why sheer ‘disease specific’ interventions have so little impact on people’s and populations’ health. Most individuals with health problems in the community do not visit health care.[3, 4] However, when individuals are visiting health care a substantial part of the interventions performed are diagnostic in nature (approximately 65 per cent). Therefore, comparisons on (specialist versus generalist) performance should not ignore the diagnostic phase of health care. Furthermore, from a population perspective, performance at the disease and procedure specific level is of limited relevance, as many presented symptoms never evolve into diseases or disorders.[5] The contribution of specialists to populations’ health, in other words, depends on the way their expertise is called-in for the care of individual patients.

This brings us to explore Stange and Ferrer’s paradox from the angle of specialization in medical care: ‘the paradox of specialization’. There can be no doubt that specialization on defined domains of disease and procedure has contributed substantially to the progress of health care.[1] One would expect that this contribution would result in a role distribution in health care between generalists and specialists, in which specialists’ optimize their performance on the defined and restricted domain of their specialization. This is in fact what Stange and Ferrer recommend for the management of ‘complex diseases’.[1] Implicit, though in this recommendation, is the importance of prior patient selection, to allow specialists to concentrate on their clinical core competencies. Patient selection, however, calls for an expert in undifferentiated health problems, the generalist.[6] In this context another paradox is important: the current position of the specialist in health care and in society is not based on the needs of people’s and populations’ health, but on a societal status (including political, financial and income consequences) beyond the limits of their domain of specialization. This is reflected in their dominance of medical education, teaching, training and research. ‘Specialist’, in other words, has become a generic title, outside the specialty context. This results in problems with professional responsibilities, in which specialists either act outside their (sub)specialty area or redefine and reduce the patients’ health needs to just the disease specific interventions in which they are specialized.

The three complexities identified by Stange and Ferrer all depend on interactions and collaboration between different experts:

(a) where the professionally diagnosed health problem is complex or rare, this falls in the domain of the (sub)specialist, but requires as well the diagnostic or selective expertise of the generalist;
(b) where the patient (co-morbidity) is the complex, this is the generalist’s domain with eventual subspecialist support for defined questions;
(c) where the population is complex, it is public health leadership to involve generalist and specialist support for defined questions.

Given the fact that health care is to be performed in ever more complex health care systems, we propose to add a fourth ‘complexity’ item: where the health care system is complex, there is the need of a(n FP)- generalist with knowledge about the local situation and organization of health care, who can guide (‘navigate’) the patient to get the right care at the right moment. These four interactions between different experts will only work when all professionals involved acknowledge the importance of the expertise of the other professionals and take full responsibility to jointly and in collaboration address the patients’ and populations’ needs. Therefore, the specialists should optimize the quality of their specialty interventions and transfer health problems that fall outside this domain (back) to the generalists.

The ‘leadership to steer’ care and a timely transfer of expertise between specialists and generalists asks for a sound balance in which Kerr White’s dictum that ‘knowing the patient who has the disease is as important as knowing the disease the patient has’ comes true.

References 1. Stange KC, Ferrer RL. The Paradox of Primary Care. Ann Fam Med 2009; 7:293-299.
2. Weel C van, Schellevis FG. Comorbidity and guidelines: conflicting interests. Lancet 2006;367:550-1.
3. Green LA, Fryer, GE, Yawn, BP, Lanier D, Dovey, SM. The ecology of medical care revisited. N Engl J Med 2001; 344: 2021-25
4. Lisdonk EH van de. Perceived and presented morbidity in general practice. Scand J Prim Health Care 1989; 7: 73-8.
5. Okkes IM, Oskam SK, Van Boven K, Lamberts H. EFP. Episodes of Care in Family Practice. Epidemiological data based on the routine use of the International Classification of Primary Care (ICPC) in the Transition Project of the Academic Medical Center/University of Amsterdam. (1985–2003). In: ICPC in the Amsterdam Transition Project. CD-Rom—Okkes IM, Oskam SK, Lamberts H, eds. (2005) Amsterdam: Academic Medical Center/University of Amsterdam, Department of Family Medicine.
6. Wonca Europe. The European definition of general practice/family medicine. http:/www.Wonca Europe 2002. The European Definition of General Practice/Family Medicine (accessed July 27th, 2009).

Competing interests:   None declared

General Practice/Family Medicine/Primary care are as much about a value as approach – improving patient health (the value) through an ongoing personal healing relationship between a patient and doctor (the approach) [1]. It ‘patient health’ – is denominator – that should solve the apparent paradox so brilliantly outlined by Stange and Ferrer [2].

Patient health is a subjective state, a state achieved when the patient perceives his biological, psychological, social and sensemaking dimensions to be in balance. Health, illness and “dis-ease” are states on the same subjective scale and need to be distinguished from the objective scale of identifiable pathologies (also known in colloquial terms as disease).

A holistic perspective focuses on the denominator – the patient's experience of health. Achieving a good patient health experience depends as much on the 'presenting problem' as on coordinating the 'right health care team' for this patient's problem. Problem is used here in a generic way, as we know from community epidemiology only very few require 'specialist intervention', and most can be managed by the local health care team. Restoration of health is an emergent process in the presence as well as the absence of identifiable pathologies and must be distinguished from the instrumental nature inherent in the 'potentially curative interventions' of reductionist specialism.

Trying to disentangle the contributions of specific agents on the outcome in complex systems like health care is dangerous, as many of the failed health care reforms are based in the prevailing reductionist framework, and are excused as “unforseen unindented consequences”. The key question must always be – for the patient with the common cold as much as for the patient with many discrete health problems – 'How is this intervention going to impact on this patient's health'.

1. Sturmberg, J. P. with a contribution by Martin, C. M. (2007) The foundations of primary care. Daring to be different. Oxford San Francisco: Radcliffe Medical Press.

2. Stange, K. C. & Ferrer, R. L. (2009) The paradox of primary care. Annals of Family Medicine, 7 (4), 293-299.

Competing interests:   None declared

Congratulations on this very throughtful and timely article in the interesting times of major US health care reform. The article stimulates many questions including the following, What is health and are our measures appropriate? What are the priorities for health in health systems? How is primary care generalism also a specialism dealing with practice populations as well as individuals? Does equity mean equality of treatment for all on selected components of primary care - for example an outcome of the UK QOF endeavour? Should there be an ecological approach to individual and community needs?

Ultimately the answers to most of these questions rest not on evidence but on societal values and politics in order to balance competing and even contradictory demands in complex health and political economic systems. For example, there will never be an evidence based answer to determine a hierarchy and balance among the priorities of whole population health, local community health and individual needs.

We, as family physicians, by the nature of our chosen discipline, probably have values that make us strong advocates for the 'most health (as a public good) for the most people', while strongly advocating for our individual patients. This requires a conscious shift from simple thinking that relies on linear evidence to complexity principles in order to accomodate the non-linear holistic dynamic systems components of primary care.

Often our discipline is not easliy understood by policy makers and funders. Here the human mind is required to rely not only on 'facts' but dynamic knowledge and wisdom and overtly recognize values. Also research approaches to address many primary care questions range from philosophy, mathematical modelling and trials to simple observations and with feedback loops over time. Thus synthesis using mechanistic approaches can never provide answers to many big questions.

Primary care will always be on the edge of chaos, as we need to reflect and adapt our practice and research in response to the shifting dynamics of complex systems, and articluate both adaptive and changing and fairly stable knowledge and wisdom at different levels.

Competing interests:   None declared

Stange and Ferrer do a masterful job of summarizing the current literature regarding the perceived competencies of generalists and specialists as well as the inherent limitations of a disease-based model of care. Whether the described paradox is real is unclear, as the authors indicate, particularly since a generalist should bring a level of expertise in their own domain that exceeds that of a specialist in the same domain. In cases where the domains of the generalist and specialist overlap, the former should have an advantage, if for no other reason than the former is more likely to consult the latter rather than vice versa and should have the advantage of expertise from both domains.

These issues are not adequately addressed in research that uses a disease-based model—putatively the domain of the specialist—for defining its methodology. If the metrics are correct, a specialist in primary care should be no less effective at the level of the individual patient as that of the population. In the last article in this series (Stange KC. The generalist approach. Ann Fam Med. 2009;7(3):198-203), Stange has already provided an excellent example of one of his own patients who benefited from his expertise as a family physician. Many such examples exist.

A few days ago my family medicine residents and I had the privilege of delivering our 30th consecutive obstetrical patient vaginally in a teaching hospital with a 41% caesarean section rate. Ironically, four hours before the arrival of the healthy infant, the obstetrical attending observing the fetal monitoring tracing on the screen at the nursing station had recommended a caesarean section. While an objective family physician would have concurred with this assessment based on the information available from the telemetry unit (3 successive deep variable decelerations, mild fetal tachycardia and cervical dilatation still at 4 cm), additional data was available at the bedside. As a result, a vaginal delivery was achieved by using a primary care heuristic that included contextualization (slow but consistent progress of labor over time that involved cephalic molding), breadth of data (absence of late decelerations), personalization (partnership and teamwork of those at the bedside—patient, physicians, nurses, family) and integration (carefully observing the physiologic response to the stress of each contraction as well as to positioning, oxygen and other interventions).

The authors cite 31 articles (#37-67) that show a greater quality of care by specialists. Yet, there is not a single family physician represented among the 134 individuals who conducted these research studies. The research and the metrics need to be better aligned.

Competing interests:   None declared

As I read the Paradox of Primary Care I kept making connections with a sobering encounter I had this week. This encounter made me realise that it is not the geographic fact of primary care that causes cost-efficient care, but the mindset of generalists and their ability to team-work.

I was facilitating a multidisciplinary workshop on Case Management for patients who have serious conditions such as Chronic Obstructive Airways Disease (COPD). In the UK community matrons case manage such patients to avoid expensive hospital admissions. There were about 30 people in the room – a quarter general practitioners (GPs), a quarter community matrons, and the rest a mixture of other primary care practitioners, managers and academics.

At my table were a mother and daughter whose husband/father had died from COPD the previous week. They insisted on coming despite their recent bereavement, because they had something urgent to say. They wanted to complain about their treatment in hospital where the clinicians ignored their emotional needs and were unavailable even to answer simple practical questions about what might happen. I have heard this complaint many times before. I reason that when doctors are overly concerned with treating diseases and too little concerned with healing people they become unable to respond to aspects of health other than their own specialty.

It fell to me to interview her, in the hearing of the whole room. She extolled the virtues of the community matron, Dom (imaginary name) who was unobtrusively sitting in the shadows. “You always knew that Dom would solve problems, get information we needed… he was like a friend, he bantered with my husband, who was not an easy man….”

“And what about your GP?” I asked.

“Well he is a nice man. We have known him for years. He is our family doctor. But he didn’t really know how to help. He came and told my husband to stop smoking and prescribed some pills. But he didn’t really understand.”

This wasn’t what I wanted to hear. And I had a flash of insight into why generalism causes cost-efficient care. At that moment her GP was not fulfilling the potential of a generalist. He is ‘primary care’ because he is geographically located outside of hospital. He is a generalist in the sense that he treats a wide range of diseases. But he was unable to respond to their emotional and social needs.

One reason for the cost-effectiveness of generalist primary care practitioners is a personal ability to identify and address social and emotional needs when others might medicalise them and start expensive counter-productive investigations. Another reason is team-working that allows quick access to advice and help from other practitioners – medical and non-medical – that avoid costly referrals. In my anecdote the GP was medicalising social/emotional needs and not working close enough with the community matron to reinforce their messages.

Stange and Ferrer are right when the say that horizontal integration is essential and neglected. This is the plane that helps me to span the domains of physical, emotional and social health, and work with non- medical colleagues. The vertical plane is equally important, but is much easier because we all speak the same medical language. The paradox is that the secret about ‘better quality, better equity, lower cost’ is the plane that is most neglected – the horizontal plane.

Competing interests:   None declared

Excellent article-I tend towards the article's latter explanation that emphasizes multi-level measurement of performance. Another factor to consider is that disease- or specialty- specific performance measures tend to be "blind" to undercertainty-they assume that a diagnosis is clearly established, and the technical performance measures are thus clearly defined. By contrast, primary care's superior performance from a ecologic/population perspective may be derived from superior management of uncertaintly via primary care core attributes of continuity, comprehensiveness, patient-centeredness, etc. These attributes are likely to be particularly important in management of ill-defined symptom complexes, and in the care of complex patients with multiple chronic conditions.

Competing interests:   Employee/Stock Ownership, UnitedHealth Group

Not quite gate keeper and wizard, I think, in that "gatekeeper" sounds too passive. The right brain sees patterns and relationships and sees health in context, and both initiates and manages a pattern of co-ordinated responses to a pattern of illness. Within that pattern of responses, the left brain gets specific details in focus and zaps them. Both are necessary - but different - skills. But complementary. Both-and, not either-or. The problem is to get proper understanding and respect between two very different skills which need to be measured in very different ways and value themselves in very different ways and indeed are developed in very different ways.

Competing interests:   None declared

Competing interests:   None declared

Stange and Ferrer, albeit with an erudite discussion of levels of analysis, have basically rediscovered the truth neatly encapsulated in the classic paper, "The Gatekeeper and the Wizard: A Fairy Tale" (Mathers N, Hodgkin P. BMJ 298:172-174, 1989). The Gatekeeper (primary care physician) and the Wizard (specialist) need each other; both suffer when a dysfunctional health care system tries to reassign one's tasks to the other.

Competing interests:   None declared