594 TYPES OF SPECIALTY IN RELATION TO LEVEL AT WHICH PROVIDED
HOW MANY SPECIALISTS?* BRYAN JENNETT
University Department of Neurosurgery, Institute of Neurological Sciences, Glasgow G51 4TF The need for specialists cannot be adequately assessed on the basis of current practice and an estimated case-load (based on prevalence/population ratio). Specialists should declare which patients can or cannot be expected to benefit from their treatment, and they should consider the consequences of underprovision of specialist services. Specialists should also decide how much of what they normally do might equally well be carried out by generalists using the knowledge of specialists. Only then can the appropriate level of provision for various specialties be assessed, and effective deployment of the resources provided ensured.
other surgeons the care of patients with head injury or lumbar-disc disease). The characteristic of the work of relative specialists is that all of it may be undertaken by generalists; indeed, it usually is, when or where a specialist is r available. The consequences of restricting the level of provision of an absolute specialist service is clearly different from that of restricting a relative specialist service. By definition a substitute cannot carry out an absolute specialist activity-there is no alternative but the waiting list or, if the condition is one demanding acute management, denial of the service.
THE growth of specialisation, especially over the past 20 years, has been stimulated mostly by specialist doctors themselves, and it has seldom occurred as a planned response to the identified needs of patients or the health service. Since the reorganisation of the National Health Service in 1974, and since the subsequent declaration of the policy of restricting the growth of acute services, it seems likely that further development of specialist services will depend on the outcome of debate about the relative merits of various needs. It is therefore timely to consider how decisions might be reached on the scale of provision of specialties, and on where each should be available-at regional, area, or district level, or in every
major hospital. TYPES OF SPECIALISATION
Traditionally, it is doctors who specialise, but there are also specialist hospitals (or departments or beds within a hospital). But few patients fall into any one specialty; many require the services at different times of a variety of specialists, and one of the dangers of increased specialisation is that a patient may not have access to the appropriate specialist when required. Specialties can be classified simply (see accompanying table) into traditional specialties, which are usually available in every general hospital, and into two other kinds, known as relative and absolute specialties. The availability of relative and absolute specialties, many of which relate to expertise in particular procedures rather than specialties as normally recognised, is usually centralised. This classification into relative and absolute specialties recognises that even "superspecialists", such as neurosurgeons or cardiothoracic surgeons, also undertake some work which might be done by others not in that specialty (e.g., neurosurgeons often share with *Based on talks given to a Department of Health and Social Security symposium on Clinicians and Management, and to a Faculty of Community Medicine symposium on the Acute Hospital Services.
PROVISION OF SPECIALIST SERVICES
The chance a patient has of access to the services of specialist depends on the availability of facilities, on the willingness of other doctors to refer patients, and on the acceptance policies of the specialist unit. There is considerable unevenness in the level of provision of beds, equipment, and staff in different parts of the U.K.; these differences derive from a complex interaction between patient-need, the demands of non-specialist doctors, the response to these of health authorities (at national and local level), and also the attitude of the specialists concerned. The attitude of specialists tends to be dominated by the enthusiasm and interests of local specialists, while the health-service reponse is usually to regard the status
Seldom have attempts been made to calculate the needs of the population for a particular service, or to appraise what the contribution of that service to the general standard of medicine might be. An exception is the assessment of the potential demand for treatment of renal failure, from which an estimate has been made of the resources required to provide renal dialysis or renal transplantation for the patient population at risk. There have also been assessments of the consequence of failure to meet the demand for treatment-not only the mortality and morbidity, but also the costs of caring for the untreated patients. In North America the prevalence (per 100 000 population) per year of certain major conditions and the frequency with which neurological procedures are performed have been related to the distribution of neurosurgeons and to their workload, as a basis for future planning.I SPECIALISTS’
VIEW OF THEIR ROLE
Without a definition of their roles, "acceptance policies" of various specialties will depend a great deal on the view taken by local specialist groups about their role. Although these policies may depend partly on the facilities available, the extent of these facilities usually reflect the
demand made by local specialists. There is a variation of Tudor Hart’s inverse-care law,2 which applies particuand which may be termed the states that in tend to show the interest the clinical least specialists conditions (within their areas of expertise) which cause the biggest burdens in the health service. Many neurosurgeons show little interest in head injuries, rheumatologists in backache, gastroenterologists in nervous dyspepsia, and urologists in incontinence in the elderly. These specialists may have carefully considered what benefit would accrue from their paying more attention
"inverse-burden law of specialists". It
patients with these
conditions, and perhaps
have wisely (and modestly) concluded that they have little more to offer than the generalist. However, a more likely explanation is that other problems present more interesting challenges to the specialist. In the other approach taken by specialists an inclusive rather than an exclusive policy dominates the scenefor example, the belief by paediatricians that no one else can cope with children, or the claim by radiologists of a monopoly on reading films. Where market forces considerably influence medical practice it is usual to find specialists willing to take on as many patients as possible
who could be
regarded as justifying their special skills. They encouraged in this if fear of litigation makes generalists unwilling to deal with patients who might be thought the proper concern of the specialist. Under an inclusive system, the specialist legitimately calculates how small a population could provide enough patients for his economic survival. Where an exclusive policy dominates, the question is how large a population could reasonably be served by one specialist (or unit). The contrast can be exemplified by considering neurosurgical practice in the U.S. and U.K. The U.S.A. has are
times as many neurosurgeons as does Britain in relation to the population served, and most American neurosurgeons accept responsibility for a much wider range of conditions than do most of their British colleagues. The American neurosurgeon deals with a large proportion of the head injuries admitted to hospital (even those with mild concussion), with spinal trauma, and with peripheral-nerve injuries. In most parts of Britain the neurosurgeon accepts less than 5% of patients admitted to hospital with head injuries, whilst spinal and peripheral-nerve trauma are dealt with almost seven
exclusively by orthopxdic surgeons. TYPE OF SPECIALIST UNIT
Almost all British neurosurgeons work in regional centres, whereas in the U.S.A. many neurosurgeons work in community hospitals, where the extent of supporting services, in particular specially trained staff in neuroradiology, neuroanoesthesia, and neuropathology are very limited’-by British standards. The cost of capital equipment is not the only reason why in most of Europe neurosurgery is organised in a small number of large centres. The scale of the enterprise influences the degree of competence, only if there is a sufficient patient throughput can competence be gained and maintained. The dilution of experience which follows the dispersion and multiplication of specialists is a disadvantage to be weighed against the advantages of providing a service within smaller communities.
Not only do large regional centres allow a high standard of equipment, staffing, and competence, but they also make subspecialisation possible-in neurosurgery, one surgeon may deal mainly with paediatrics, stereotaxis, transsphenoidal surgery, or the treatment of pain. They also provide the opportunity for training of staff (medical, nursing, and technical), for exploration of new technologies, and for research. HOW MANY SPECIALISTS ARE
This begs the question, "To do what"? The cases made out for the expansion of specialist services (for extra staff or additional units) usually depend on the estimated case-load in the community (based on prevalence/population ratio) or on the demands made on present services. One important factor is usually not taken into account-namely, what proportion (or which) of the patients whose conditions might be regarded as of "interest" to a specialty will "benefit" from specialist referral? Other crucial questions to be considered include-what are the alternatives to specialist care? What difference would there be between specialist and non-specialist care? Can the demand for specialist-care be regulated? These questions cannot be answered unless specialists are required to declare their policies and to defend their position with some facts. The need is for a statement of what a specialist can do (according to the present state of the art), what is provided (as a local resource), and what is offered (allowing for the bias of interest of the
local staff). A declaration by staff of the balance of their interest between major categories of activity-diagnosis, rescue, cure, rehabilitation, or care-may help to clarify the role of a specialist service. All doctors might claim to undertake all of these activities, but in practice many specialists, and certainly some special units, clearly accept that their predominant concern is only with some of these roles. Clearly the resources needed by a neurologist who regards himself as providing a diagnostic service, and who sees patients with conditions such as muscular dystrophy or multiple sclerosis once or twice only, are quite different from those of a colleague who sees his role as taking over the life-time support and care of patients with progressive conditions. Similarly the neurosurgeon who sends back to the primary surgical or the medical wards patients whose in-
requiring special neurosurgical
facilities has been completed, will need fewer beds than if he takes continuing responsibility for all patients referred. Once a specialty has declared its objectives, it is possible to consider the likely consequences of overprovision or underprovision of services:
always be easy to compare the outcome of with that of non-specialist care-mortality specialist and morbidity provide obvious indicators, but there is It may
596 need also to consider the relative costs of management by one system or the other. It is easy to assert that everyone with a stroke or a fit should see a neurologist, or that everyone with a head injury or sciatica should see a neurosurgeon; but it is fair to ask what benefit is supposed to come from this. American neurosurgeons are- in no doubt that their greater numbers enables them to provide a "kind of care which is truly not available in other countries."3 What is needed before a decision is taken on an appropriate level of specialist provision are comparative studies of outcome in large numbers of patients treated under different systems, rather than what a prominent member of the American College of Surgeons has called "The Rhetoric of Specialzation".44 .
DEMAND FOR SPECIALIST SERVICES
This comes almost wholly from other doctors, at least with the closed system of referral practised in Britain. It might therefore be considered relatively easy to regulate demand by applying the declared policies of acceptance agreed between specialists and their hospital colleagues, provided that referral is restricted to hospital doctors. What happens if this is not done can be exemplified by the experience of neurologists who accept referrals from family doctors. They usually find that they are seeing a large proportion of patients with conditions which do not justify special investigation; indeed many of these patients have complaints that are not organic. But if neurologists are willing to accept referrals from family doctors, why should they not? Because in a regionalised system it is inevitable that the majority of referrals from general practitioners will come from the area in which the regional unit happens to be sited. When resources are limited (and that should be regarded as the normal situation) similar patients living a distance away from a regional unit will not have equal access to its facilities. But more important is the possibility that those patients throughout the region who really do require the facilities of a special unit will have to compete with those who do not, and will have to join the waiting list for clinic appointments, for investigations, and perhaps for admission. There is another danger in specialists seeing uncomplicated patients. Because superspecialists regard themselves as a court of last appeal, they often tend to overinvestigate patients with
trivial complaints, merely because they
are expected to give a once-and-for-all diagnosis. Such a practice is costly to the health service and may be hazardous to the patient. Everyone would benefit from the interposition of a well-trained general physician between the family doctor and the neurologist.
SUBSTITUTES FOR SPECIALISTS
It is often presumed that the only alternative to a specialist is the untutored generalist. But that need not be so. An important contribution which a specialty should make to overall medical care is to define by trial within a specialist unit the usefulness of various techniques, and then to decide which can safely be undertaken by others elsewhere. The specialist must then un- ’ dertake to inform and instruct the "others elsewhere" in what to do, how to do it, and most important, who to do it to.
encouraging evidence of the willingness of specialists to do this. For example it has been suggested that radiologists might train other doctors to read certain types of film.5 It has been further suggested that radiology is an artificial specialty which should be disbanded, and that the work should be dispersed among clinicians,6 a small number of specialised radiology departments being maintained for research and development and for the training of other doctors. Also, a team of junior doctors and nurses in medical wards, who had been briefly trained for the task, did as well as psychiatrists in assessing suicidal risk in patients recovering from deliberate self-poisoning, and in identifying which patients required psychiatric treatment or help from social workers.’ Another study showed little difference in outcome between children treated for leukaemia in special centres and in district hospitals; the few patients treated in district hospitals who had done less well had not received the full regimen of treatment recommended by the specialist centre.8 The conclusion was that district hospitals should continue to treat patients, but that the regional centre would monitor facilities, regimens, and results. A final example of specialists teaching others their tricks comes from the development of the Glasgow coma scale in a large neurosurgical unit, where it was subjected to an observer-error trial with junior doctors, There is
non-neurosurgeons, and nurses.9 This scale is now used worldwide, and in the U.S.A. ambulancemen use it, so that patients arrive at hospital with the Glasgow comascore recorded from the time of pick-up. Coma is liable
affect patients with a wide variety of conditions -head injury, hypoxia, stroke, poisoning, hepato-renal failure, and so on-and the state of coma often influences decisions which can affect outcome critically. This transference of skills from high-technology specialists to paramedical technicians is another example of the educational and training role of specialists. to
SELECTION OF PATIENTS FOR TERTIARY CARE
The effective use of specialist services depends on the of a good system of selecting the right patients for referral, in accordance with the declared role of each specialist service. Such a policy will be effective only if it is generally recognised. A deliberate educational effort is needed to inform generalists, and to train them to do some specialist work and to become experts at selecting the patients who need to be referred to a specialist. The misuse of specialist services is not limited to overburdening them with patients whose referral is unjustified because their condition is not sufficiently serious. Certain specialties (e.g., intensive care, neonatal surgery, radiotherapy, and medical oncology) often have to deploy an undue proportion of their resources in attempts to rescue patients who can neither be cured nor restored temporarily to an acceptable quality of life. The concept of "rescue" has been discussed elsewhere,’O as has the considerable (? undue) proportion of the expenditure of some expensive specialist units which goes on rescue or on other "unsuccessful" treatment." However, when activities are appropriately deployed,
medicine may be very cost-effective,
allowing for the large capital outlay on equipment.
favourably with some activities which are accepted without question as part of the acute hospital services-for example, the treatment of some forms of malignant disease by general surgeons, radiotherapists, and oncologists, which at best may result only in short-term palliation, which might be achieved more simply, and which at worst amounts to an expensive placebo. In contrast, the success of renal dialysis and transplantation, in terms of initial mortality, actuariallycorrected survival, and degree of rehabilitation achieved, is striking.12 The same is true for much open-
It compares very
and hence to more effective use of specialists. They should concentrate on activities which only they can do; and they should be more willing to consider ways and means whereby some of the work currently regarded as their responsibility might be carried out by generalists. In the work which they retain within their units they should ensure that an appropriate proportion of their resources are reserved for effective procedures. This demands careful selection of patients, in accordance with agreed and declared policies. REFERENCES 1. Drake, C. G. J. Neurosurg 1978, 2. Hart, J. T. Lancet, 1971,i, 405.
(personal communication), cited by C. G. Drake in J. Neuro1978, 49, 483. 4. Hanlon, C. R. J. Neurosurg. 1978, 49, 785. 5. Editorial. Br. J. Radiol. 1975, 48, 517. 6. Thomas, M. L. Br. med. J. 1978,ii, 706 & 1225. 7. Gardner, R., Hanka, R., Evison, B., Mountford, P. M., O’Brien, V. C., Roberts, S. J. ibid. 1392. 8. McCarthy, M. Lancet, 1975,i, 1128. 9. Teasdale, G., Knill-Jones, R., Van der Sande, J. J. Neurol. Neurosurg. Psychiat. 1978, 41, 603 10. Jennett, B. Lancet, 1976, ii, 1235. 11. Jennett, B. in Clinical Practice and Economics (edited by C. I. Phillips and J. N. Wolfe); p. 46. London, 1977. 12. Morris, P. J., Bishop, M., Fellows, G., et al. Lancet, 1978,ii, 1353. 13. Ross, J. K., Monro, J. L., Manners, J. M., Edwards, J. C., Lewis, B., Hyde, I., Conway, N., Johnson, A. M. Br. med. J. 1976,ii, 1485. 14. Monro, J. L., Mollo, S., Brookbanks, S., Conway, N., Ross, J. K. ibid. 1978, 3. Ransohoff, J. surg.
CONCLUSIONS a fact of life in modern medicine in countries. It allows the development of effective methods of investigation and treatment, but the benefits of these techniques are seldom as widely available to appropriate patients as they might be. This is not only, or indeed mainly, due to lack of resources. A more serious hindrance is the reluctance of specialists to define their role, and to declare what they can and cannot do. Agreement between them and their colleagues in other disciplines could lead to more effective referral patterns,
Hospital Practice CAN PATIENTS KEEP THEIR OWN PEAK-FLOW RECORDS RELIABLY? I. P. WILLIAMS M. R. HETZEL* R. M. SHAKESPEARE
Department of Thoracic Medicine, St. James’ Hospital, London SW12
Fifty patients recorded their peak expiratory-flow rate (P.E.F.R.) in hospital, unaided by nursing staff, five times a day for 5 days. Each patient’s readings were randomly and independently checked on two occasions during this period. 69% of checked readings were accurate. Most patients kept satisfactory records as a table, but were less efficient in recording their results on a P.E.F.R. chart. Recording of P.E.F.R. by patients with respiratory disease saves nursing time and provides valuable clinical information.
Regular monitoring of peak expiratory-flow rate (P.E.F.R.)’ in respiratory disease is valuableespecially in severe asthma3where different patterns of response to treatment,4and high-risk groups,5 can be identified. In some general medical wards, however, routine P.E.F.R. monitoring by nursing staff is not feasible. We therefore investigated the reliability of P.E.F.R. records kept by patients themselves.
disease, and ability to blow a P.E.F.R. of > 100 litres min"’on admission. Fifty suitable patients were recruited from consecutive admissions to our wards; although most had respiratory disease, this was not a prerequisite for inclusion. There was no age limit. Each patient used the same miniature Wright peak-flow meter6 throughout the study and made recordings at 4-hourly intervals between 6 A.M. and 10 P.M. for the first 5 consecutive days of admission. Those patients who were treated with bronchodilator drugs took them 30 min after the times appointed for P.E.F.R. measurement. On each occasion, 3 readings were taken and recorded on a simple table with the date and time. Patients were asked to plot the highest of these 3 readings on a standard P.E.F.R. chart printed with gradations for the day and time on the x axis and a P.E.F.R. scale in 20 litres min-l increments on the y axis. On admission these procedures and the use of the meter were demonstrated by the, first observer who then supervised the first 2 sets of readings. Patients were then left to keep their own tables and charts. On two randomly selected occasions during the 4th and 5th days the second observer, who was not involved in patients’ management or allowed to see their records, visited patients 10 min after they should have made a recording and asked them to blow a further three peak flows into their own meters. Patients were not warned that they would be checked and the observer attempted to reduce their awareness of the purpose of his visit by memorising his check readings and recording them after leaving the patient. Means of the three readings for each set of doctor and patient results were compared. The total number of readings which patients recorded in their tables and the number and accuracy of readings which patients transferred from their tables to the P.E.F.R. charts were assessed; a reading was regarded as being accurately recorded if it was plotted within +20 litres min-of its correct position on the chart. RESULTS
PATIENTS AND METHODS
Criteria for inclusion in the study were literacy, eyesight good enough to read the meter scale, absence of psychiatric *Present address:
Road., London SW3.
Patients’ mean age was 54.8years (range 13-89 There were thirty-five men and fifteen women. Clinical diagnoses on admission were asthma (ten), chronic bronchitis (fourteen), other respiratory disease