Proving the worth of neurologists?

Neurology 1996;46:276-277

The role of specialist care in the emerging managed health delivery system is changing. Most health maintenance organizations (HMOs) are limiting access to specialists because of concerns about cost. Despite the need to control spending, there is optimism that specialists can "prove their worth," presumably by providing a much better outcome even if it means a higher price. To convince payers not to skimp on specialty care, however, we need more data on qualitative improvements in care provided by specialist consultants. ^[1] Specialists will need to prove their usefulness to generalists acting as gatekeepers of the health delivery system. Can this be done?

To evaluate the impact of specialist contributions to inpatient management, we assessed the usefulness of neurologic consultations in a university setting for 5 months in 1994. We arbitrarily defined a consultation as useful if it met one of the following criteria: (1) The neurologist's differential diagnosis contained a diagnosis (whether ultimately correct or not) that had not been included by the referring physician. (2) The neurologist suggested even a single procedure not previously ordered by the primary physician that was ultimately useful in making the correct diagnosis or eliminating other alternative diagnoses. The procedure must have been done. (3) The neurologist suggested a treatment regimen that was carried out by the referring physician and was different from the regimen proposed by the primary physician prior to the consultation. If no treatment regimen had been proposed by the referring physician, the consultation was considered useful. (4) We elected to consider a consultation as not useful if it only confirmed the primary physician's differential diagnosis or management. We did not consider a neurologist's role in interpreting or performing tests (electroencephalogram, electromyogram, nerve conduction velocity, etc), because these functions were performed by subspecialists other than the consulting neurologist. We are aware that this is not the case in many settings but is usual in most university hospitals.

Seventy-nine inpatients were seen by five neurologists from the university inpatient consultation service over a 5-month period. We were able to obtain adequate information to evaluate 72 of these patients. Emergency room consultations were not included in the analysis. Disease categories and the number of patients in each category are listed in Table 1. Thirty-two consultations were requested for diagnostic purposes, 21 for management, and 19 for both.

Only 31 consultations (43%) were deemed useful (see *table 1* for a breakdown by diagnostic category). This percentage held across the board for patients seen for diagnosis (9 of 21 patients) or management (14 of 32 patients), or both (8 of 19 patients). We then asked if the consultants were more able than referring physicians to pinpoint a few key tests that were helpful to the decision-making process. To this end, we counted the number of "non-routine" tests that had a positive effect in making the correct diagnosis or management decision. Primary doctors ordered 111 "nonroutine" tests. We considered 59 of these to be useful (53%). Consultant neurologists suggested an additional 102 "nonroutine" tests, 36 of which (35%) we deemed to be useful. Twenty-two tests suggested by the consultants were not even ordered by the primary physician.

The case for the neurologist is a little less grim, but by no means bright, if one looks at some of the situations in which a consultation was useful, or not useful. Four of six consultations that called for syncope were correctly diagnosed and managed by the referring doctor (not useful category). In two cases, however, the initial diagnosis was incorrect and the consultations led to the correct diagnoses (useful category) (seizures and vestibular disease). Even the most hardened HMO manager should agree that specialty consultation for all six syncope patients was worthwhile and cost-effective to rectify an initially incorrect diagnosis in two cases. Consultants were useful in the workup of new onset seizures and in the management of poorly controlled epilepsy and, surprisingly, ethanol withdrawal seizures. Patients admitted with a stroke benefited when there was a question about anticoagulation, if the initial diagnosis was wrong, when the event occurred in the context of a procedure (e.g., angiography), and when the referring physician had not initiated any plan of diagnosis and management. Most consultations for stroke and seizures that fell into the not useful category were called to confirm decisions that had already been made. This "hand-holding" was also common for patients with primary neurologic diseases (Parkinson's disease, motor neuron disease, multiple sclerosis) when the admission was based on clearcut medical or surgical reasons. It was particularly notable in patients with hypoxic-ischemic encephalopathy after cardiac arrest where consultations were called because the primary physician seemingly did not wish to or feel competent to talk with the family about the neurologic state or prognosis.

We decided to look at inpatient consultations because this segment of health care delivery is of particular importance to neurologic teaching and practice. Neurologic inpatient wards are shrinking not only because more patient problems are now handled outside the hospital, but also because many patients with neurologic disease are admitted to general medicine services. Will HMOs be willing to pay for consultations in situations that mostly confirm what others strongly suspect? Will the fear of malpractice or the unwillingness to relay bad news be enough to sustain neurologic consultation services in the face of rising costs? Will analysis of outpatient consultative practice be able to show its worth? If not, what role should neurologists play in this new world?

We suspect that neurologists will be able to prove their worth, but not by continuing their present practices. Neurologists may need to become the primary care physicians for patients with neurologic diseases. One recent study has shown that patients with stroke are discharged from the hospital more rapidly when admitted to a neurologic service rather than to a general medical ward, where the patient is better off and it costs less. ^[2] Managed care may actually provide an opportunity to test this approach.

This kind of analysis is not easy. Deciding on ultimate usefulness is as much a judgment call based on experience and knowledge as it is a numbers game based on cost and definitions. We tried our best to be fair by using a set of defined criteria, even though we started out with the bias that specialty consultation is useful, more so than we found. We are not sure, however, if our tentative conclusions are at all close to the mark. To what degree, for example, is patient satisfaction tied to the blessing of a specialist? Do generalists really have enough neurologic knowledge to forego hand-holding? Future studies of this kind will need to be carefully crafted by more experienced experts than us to reach conclusions that stand a chance of being correct. Such studies, however, must not be left only to epidemiologists and accountants. Neurologists cannot be excluded from the design and implementation of such studies, or the information gained may be as stillborn as the late presidential health plan designed by experts with little or no "hands-on" experience. We welcome advice and criticism from our neurologic colleagues. We hope our comments stir more debate on how best to use our nation's specialists in the interests of the people.

Finally, the training of future generations of medical students may be at risk. Despite a shift of medical student teaching to outpatient settings, we believe, as do most of the medical students on our neurology rotation, that there is no substitute for inpatient experience. There is simply not enough time for students to evaluate patients on their own, go to the library and read about the disease, and think things over and re-examine the same patient, all in an hour-long clinic visit. The situation will become worse as we shift to more managed care. If neurologists cannot prove their worth and if inpatient consultations dry up as a result, where will students get this learning experience? Even if a consultation is not useful for the patient, it may still be useful for the student!