- Email: [email protected]
Diagnostic Management by Gatekeepers Is Not Cost Effective for Neuro-ophthalmology
Diagnostic Management by Gatekeepers Is Not Cost Effective for Neuro--ophthalmology Edward C. Dillon, MD, Robert C. Sergott, MD, Peter]. Savino, MD, Thomas M. Bosley, MD Purpose: To determine whether the "gatekeeper physician system" for evaluating neuro-ophthalmologic problems is cost effective. Methods: The authors retrospectively reviewed the records of 588 patients referred for neuro-ophthalmologic evaluation between July and December 1989 to determine the frequency and cost of unnecessary diagnostic testing ordered by "gatekeeper physicians." Pre-referral diagnostic testing costs were compared with the cost of neuroophthalmologic consultation for four common problems: (1) optic neuropathy; (2) diplopia; (3) ptosis; and (4) proptosis. Results: Between 16% and 26% of patients in the first three diagnostic categories were subjected to overtesting, resulting in $57,900 of excessive costs, a 724% overcharge. Although the evaluation of proptosis was performed correctly, the quality of 10 of the 18 neuro-imaging procedures was substandard. Conclusions: The gatekeeper system managed by primary care physicians for these four neuro-ophthalmologic problems not only did not conserve healthcare dollars but also had a negative impact on cost control. For neuro-ophthalmologic disorders, prompt subspecialty evaluation and examination appear to be a cost-effective strategy. Ophthalmology 1994;101:1627-1630
Escalating, virtually uncontrolled, healthcare costs have evolved into one of the United States's most pressing domestic issues, thereby providing the stimulus for both new healthcare policies and delivery systems. 1 Many new policies are based on a resource-based relative value scale (RBRVS) for physician services. The Harvard University RBRVS, designed by William Hsiao, PhD, has recommended that primary care physicians, and even some nonphysicians such as optometrists, should receive increased financial reimbursement?-4 In contrast, physicians with subspecialty training such as ophthalmologists, radiologists, thoracic surgeons, pathologists, dermatologists, and psychiatrists have been labeled as overpaid. This reimbursement plan paid for and endorsed by the Health Originally received: December 17, 1993. Revision accepted: April II, 1994. From the Wills Eye Hospital, Neuro-Ophthalmology Service, Philadelphia. Reprint requests to Robert C. Sergott, MD, Wills Eye Hospital. NeuroOphthalmology Service, 900 Walnut St, Philadelphia, PA 19107.
Care Financing Administration, the overseer of Medicare, has been enacted and may be extended to the private sector.5 The advocates of RBRVS envision reduced healthcare expenditures by empowering primary care physicians, instead of subspecialists, with the control over costly diagnostic and therapeutic procedures. Using a similar philosophy, healthcare delivery systems such as health maintenance organizations and preferred provider organizations attempt to limit expenditures by placing primary care physicians ("gatekeepers") in command of diagnostic and therapeutic efforts. 6 - 9 The cost effectiveness of this gatekeeper system has not been assessed for highly specialized, complex diagnostic situations such as neuro-ophthalmologic disorders. As a hybrid discipline concentrating on unusual problems of the afferent and efferent visual system, neuro-ophthalmology remains highly dependent on astute medical history assessments and precise physical examinations, rather than a strict reliance on expensive procedures. To determine whether primary care physicians and specialists without neuro-ophthalmologic post-residency
1627
Ophthalmology
Volume 101, Number 9, September 1994
fellowship training, including family practitioners, general neurologists, ophthalmologists, and optometrists, provide cost-effective diagnostic strategies, we compared the costs generated before neuro-ophthalmologic referral with the costs that would have been incurred with a referral requested sooner during the patients' illnesses.
Patients and Methods The records of all new patients evaluated during the 5 months between July and December 1989 at the NeuroOphthalmology Service, Wills Eye Hospital, were reviewed by one of us (ECD) who is a board-certified internist and emergency room physician. All the patients had been evaluated by one of us, all of whom are neuroophthalmologists. The 1989 timeframe was selected to ensure that long-term patient follow-up care was available, and that accurate diagnoses were obtained in all cases. The records were reviewed for four referring diagnoses: (l) optic neuropathy; (2) diplopia; (3) ptosis, and (4) proptosis. These four categories were selected because the proper diagnostic conclusion usually is possible based on an accurate neuro-ophthalmologic history and physical examination without relying on extensive ancillary testing. Pre-referral diagnostic testing and costs were compared with the evaluations the patient would have received had they been referred at the onset of their problems. Estimated cost savings were obtained by a summation of the pre-referral costs compared with the diagnostic costs at Wills Eye Hospital and its affiliated hospitals in the Philadelphia area. "Overtesting" was defined as obtaining neuroimaging studies when the diagnosis could be established on the patient's history and results of physical examination. For example, a patient with ptosis and double vision underwent computed tomographic (CT) scanning and magnetic resonance imaging (MRI) as well as an arteriogram when the diagnosis of ocular myasthenia was clinically apparent and confirmed by intravenous Tensilon testing. The determination of overtesting was made by one of us (EDC). The records of the patients considered to have had excessive testing then were reviewed by the other three authors, and a consensus was reached. For a patient to be considered "overtested," all three participating neuroophthalmologists (RCS, PJS, TMB) had to come to an agreement. For each diagnostic category, the ancillary studies and clinical management considered to be appropriate were those recommended by an acknowledged, "standard," neuro-ophthalmologic textbook. 10 Neuroimaging studies were judged to be of insufficient quality when only one, two, or no views of the orbits or base of the skull were included. Patients with scans of insufficient quality were deemed to be "overtested."
Table 1. Overtesting for Optic Neuropathies Test MRI scan ($900/scan) CT scan ($500/scan) Inappropriate serologic studies (antinuclear antibody, FTA-ABS) ($100/both tests) Cerebral angiography (4-vessel study) ($1800/study) MRI
=
magnetic resonance image; CT
=
No. of Patients
Cost
25 20 26
$22;500 $10,000 $ 2600
5
$ 9000
computed tomography.
Optic Neuropathies Eighty-eight patients with optic neuropathies were found, and 21 (26%) had undergone excessive diagnostic testing usually in the form of neuro-radiologic studies, including CT scanning and MRI. Table 1 illustrates the type and cost of excessive testing for optic neuropathies. The most common diagnoses generating overtesting were ( 1) nonarteritic ischemic optic neuropathy, (2) congenitally anomalous optic discs, "pseudopapilledema," and (3) primary chronic open-angle glaucoma.
Diplopia Of 53 patients with diplopia, 11 (20%) were overtested. Table 2 indicates the scope of the overtesting and cost. As with optic neuropathies, neuroimaging studies were the procedures most frequently abused. The most common diagnoses that produced excessive testing were ( 1) ocular myasthenia gravis; (2) microvascular ischemic mononeuropathy involving the third, fourth, and sixth cranial nerves; and (3) thyroid eye disease (Graves ophthalmopathy).
Proptosis None of the 18 patients with proptosis were subjected to excessive diagnostic procedures. All procedures were evaluated appropriately with CT scans, although for ten patients the scans were not of sufficient quality to evaluate the orbits properly.
Ptosis Of the 19 patients with ptosis, 3 ( 16%) were overtested as shown in Table 3. One patient had ocular myasthenia gravis, and two had dehiscence of the levator palpebral muscle (age-related ptosis). Again, neuro-radiologic imaging procedures were the inappropriate studies.
Results
Cost Comparison
Five hundred eighty-eight new patient charts were reviewed.
The total cost of unnecessary diagnostic studies for these 35 patients was $65,900 in 1989 dollars. The cost assess-
1628
Dillon et al · Cost Effectiveness of Diagnostic Management by Gatekeepers Table 2. Overtesting for Diplopia Test
No. of Patients
MRI scan CT scan
7
Arteriography
11 3
MRI
=
=
magnetic resonance image; CT
Cost
$6300
$5500 $5400 computed tomography.
ment does not consider hospitalization costs for angiography, potential morbidity of the procedures, and subsequent costs due to complications of the invasive procedures. In comparison, if the patients had been referred at the onset of their problems for a neuro-ophthalmologic consultation ($200 per patient), the net cost would have been $8000. Therefore, the gatekeeper system generated $57,900 in unnecessary costs, a 724% overcharge, without including the additional costs of hospitalization in some patients. The 724% figure may be somewhat of an overestimation because some diagnostic testing may have been ordered if the patients had been referred directly to a neuroophthalmologist. However, the diagnostic categories of optic neuropathy, diplopia, proptosis, and ptosis rarely require extensive neuroimaging studies in standard neuroophthalmologic algorithmic diagnostic evaluations. 10 Another method of cost comparison would be to tabulate the total testing costs of the 588 patients in the four categories surveyed. The excess testing charges then could be viewed as a percentage of charges, perhaps resulting in a less dramatic surcharge amount. Attempts to perform such an analysis for this cohort of patients was unsuccessful because of incomplete information about the costs, both in terms of testing, physicians' time, and adminis~ trative expenses, before being referred for neuro-ophthalmologic evaluation. Patient outcomes. The 35 patients subjected to excessive testing did not incur additional loss of vision or neurologic sequelae due to diagnostic delay. In addition, the patients did not have any significant complications from the testing procedures.
Discussion Our data indicate that between 16% and 26% of patients in three of four neuro-ophthalmologic diagnostic categories were subjected to unnecessary and expensive diagnostic procedures under the direction of "gatekeeper physicians." In the one category in which the diagnostic approach was appropriate, the quality of the neuroimaging studies often was found to be substandard. Compared with a prompt referral for neuro-ophthalmologic consultation, the gatekeeper system generated $57,900 in excessive costs, a 724% overcharge. In addition, many patients may have undergone the same unnecessary procedures and not been referred for consultation.
Neuro-radiologic imaging studies, including CT and MRI scans as well as cerebral angiography, encompassed the majority of unnecessary testing. The excess neuroradiologic procedures imply that the overtested patients have so-called "CT and MRI negative" disease, and these disorders frequently are diagnosed based on medical history and physical findings. One explanation for the results is that gatekeeper physicians lack education and experience in the basic fundamentals of medical history taking and physical examination assessment involved in complex neuro-ophthalmic problems. Residency and post-residency fellowship training would appear to be a prerequisite to elicit precise medical history data and physical findings, and thereby save healthcare dollars. An unavoidable shortcoming of this study is an inability to determine how many patients with these diagnostic problems were appropriately managed by the referring physicians. At this time, we do not have access to such records. The current study hopefully will serve as a stimulus to begin a prospective investigation to examine this issue. It is also possible that if the patients were seen first by neuro-ophthalmologists that some overtesting may have been incurred. We tried to minimize this problem by choosing relatively common neuro-ophthalmologic disorders in which the diagnostic criteria and management approach are in general agreement. 10 Because we did not use an outside arbitrator to judge "overtesting," this study may have incurred some additional bias. An attempt was made to overcome this potential error by requiring that all three neuro-ophthalmologists agree when a patient was included as "overtested." In addition, it is difficult to attach a price to the cost of delayed diagnosis engendered by "false negative CT or MRI scans" obtained by gatekeeper physicians. In this context, we refer to patients who had structural, potentially treatable lesions and then incurred diagnostic delays. Although the current study did not have a means to examine this aspect of the gatekeeper system, future studies will need to address the issue of delays in diagnosis to the point where therapy is inadequate to reverse the visual loss while the patient was presumed to have another, nontreatable diagnosis. Fortunately, none of the patients in this study had loss of vision or permanent neurologic disability because of a lengthy diagnostic delay. An alternative method ofhealthcare delivery would be the "centers of excellence" concept into which difficult diagnostic problems are triaged immediately. In this approach, the gatekeeper would direct the patient to the
Table 3. Overtesting for Ptosis Test
No. of Patients
Cost
2 2
$1800 $1000 $1800
MRI scan CT scan Arteriography MRI
=
magnetic resonance image; CT
=
computed tomography.
1629
Ophthalmology
Volume 101, Number 9, September 1994
proper tertiary referral center and perhaps a "tertiary care gatekeeper," instead of ordering a standard battery of diagnostic procedures. These centers of excellence would be staffed by physicians with post-residency fellowship training and certified by peer-review groups. In summary, the gatekeeper system was found not to be cost effective for the evaluation offour common neuroophthalmologic problems. Although the Hsiao RBRVS system and the health maintenance organization network have been valuable in providing stimuli to save healthcare dollars in some respects, it appears that the system may break down for highly specialized problems. In actuality, post-graduate subspecialty medical education appears to provide patients with neuro-ophthalmic problems with more cost-effective and safer healthcare. Based on the data in this study, the gatekeeper method for neuro-ophthalmology consumed rather than conserved a significant amount of health care dollars within a very limited time.
References I. Williams AP. Memorandum to the President-elect. Parameters for health system reforms [editorial]. JAMA 1992;268: 2699-2700.
1630
2. Hsiao WC, Braun P, Yntema D, Becker ER. Estimating physicians' work for a resource-based relative-value scale. N Eng! J Med 1988;319:835-41. 3. Hsiao WC, Braun P, Dunn D, et a!. Results and policy implications of the resource-based relative-value study. N Eng! J Med 1988;319:881-8. 4. Hsiao WC, Dunn DL, Verrilli DK. Assessing the implementation of physician-payment reform. N Eng! J Med 1993;328:928j33. 5. Health Care Firancing Administration. Medicare program: fee schedule fo~ physicians services: final notice. Fed Regist 1991 ;56:(227):59577-87, 59635-784. 6. Franks P, Clanby CM, Nutting PA. Gatekeeping revisitedprotecting patients from overtreatment. N Eng! J Med 1992;327:424-9. 7. Greenfield S, Nelson EC, Zubkoff M, eta!. Variations in resource utilization among medical specialties and systems of care. Results from the Medical Outcomes Study. JAMA 1992;267: 1624-30. 8. Hillman AL. Health maintenance organizations, financial incentives, and physicians' judgements. Ann Intern Med 1990; 112:891-3. 9. Hiduchenko K. Do health maintenance organizations control costs or shift costs [letter]? N Eng! J Med 1993;328: 971. 10. Burde RM, Savino PJ, Trobe JD. Clinical Decisions in Neuro-Ophthalmology, 2nd ed. St. Louis: CV Mosby, 1992.
Escalating, virtually uncontrolled, healthcare costs have evolved into one of the United States's most pressing domestic issues, thereby providing the stimulus for both new healthcare policies and delivery systems. 1 Many new policies are based on a resource-based relative value scale (RBRVS) for physician services. The Harvard University RBRVS, designed by William Hsiao, PhD, has recommended that primary care physicians, and even some nonphysicians such as optometrists, should receive increased financial reimbursement?-4 In contrast, physicians with subspecialty training such as ophthalmologists, radiologists, thoracic surgeons, pathologists, dermatologists, and psychiatrists have been labeled as overpaid. This reimbursement plan paid for and endorsed by the Health Originally received: December 17, 1993. Revision accepted: April II, 1994. From the Wills Eye Hospital, Neuro-Ophthalmology Service, Philadelphia. Reprint requests to Robert C. Sergott, MD, Wills Eye Hospital. NeuroOphthalmology Service, 900 Walnut St, Philadelphia, PA 19107.
Care Financing Administration, the overseer of Medicare, has been enacted and may be extended to the private sector.5 The advocates of RBRVS envision reduced healthcare expenditures by empowering primary care physicians, instead of subspecialists, with the control over costly diagnostic and therapeutic procedures. Using a similar philosophy, healthcare delivery systems such as health maintenance organizations and preferred provider organizations attempt to limit expenditures by placing primary care physicians ("gatekeepers") in command of diagnostic and therapeutic efforts. 6 - 9 The cost effectiveness of this gatekeeper system has not been assessed for highly specialized, complex diagnostic situations such as neuro-ophthalmologic disorders. As a hybrid discipline concentrating on unusual problems of the afferent and efferent visual system, neuro-ophthalmology remains highly dependent on astute medical history assessments and precise physical examinations, rather than a strict reliance on expensive procedures. To determine whether primary care physicians and specialists without neuro-ophthalmologic post-residency
1627
Ophthalmology
Volume 101, Number 9, September 1994
fellowship training, including family practitioners, general neurologists, ophthalmologists, and optometrists, provide cost-effective diagnostic strategies, we compared the costs generated before neuro-ophthalmologic referral with the costs that would have been incurred with a referral requested sooner during the patients' illnesses.
Patients and Methods The records of all new patients evaluated during the 5 months between July and December 1989 at the NeuroOphthalmology Service, Wills Eye Hospital, were reviewed by one of us (ECD) who is a board-certified internist and emergency room physician. All the patients had been evaluated by one of us, all of whom are neuroophthalmologists. The 1989 timeframe was selected to ensure that long-term patient follow-up care was available, and that accurate diagnoses were obtained in all cases. The records were reviewed for four referring diagnoses: (l) optic neuropathy; (2) diplopia; (3) ptosis, and (4) proptosis. These four categories were selected because the proper diagnostic conclusion usually is possible based on an accurate neuro-ophthalmologic history and physical examination without relying on extensive ancillary testing. Pre-referral diagnostic testing and costs were compared with the evaluations the patient would have received had they been referred at the onset of their problems. Estimated cost savings were obtained by a summation of the pre-referral costs compared with the diagnostic costs at Wills Eye Hospital and its affiliated hospitals in the Philadelphia area. "Overtesting" was defined as obtaining neuroimaging studies when the diagnosis could be established on the patient's history and results of physical examination. For example, a patient with ptosis and double vision underwent computed tomographic (CT) scanning and magnetic resonance imaging (MRI) as well as an arteriogram when the diagnosis of ocular myasthenia was clinically apparent and confirmed by intravenous Tensilon testing. The determination of overtesting was made by one of us (EDC). The records of the patients considered to have had excessive testing then were reviewed by the other three authors, and a consensus was reached. For a patient to be considered "overtested," all three participating neuroophthalmologists (RCS, PJS, TMB) had to come to an agreement. For each diagnostic category, the ancillary studies and clinical management considered to be appropriate were those recommended by an acknowledged, "standard," neuro-ophthalmologic textbook. 10 Neuroimaging studies were judged to be of insufficient quality when only one, two, or no views of the orbits or base of the skull were included. Patients with scans of insufficient quality were deemed to be "overtested."
Table 1. Overtesting for Optic Neuropathies Test MRI scan ($900/scan) CT scan ($500/scan) Inappropriate serologic studies (antinuclear antibody, FTA-ABS) ($100/both tests) Cerebral angiography (4-vessel study) ($1800/study) MRI
=
magnetic resonance image; CT
=
No. of Patients
Cost
25 20 26
$22;500 $10,000 $ 2600
5
$ 9000
computed tomography.
Optic Neuropathies Eighty-eight patients with optic neuropathies were found, and 21 (26%) had undergone excessive diagnostic testing usually in the form of neuro-radiologic studies, including CT scanning and MRI. Table 1 illustrates the type and cost of excessive testing for optic neuropathies. The most common diagnoses generating overtesting were ( 1) nonarteritic ischemic optic neuropathy, (2) congenitally anomalous optic discs, "pseudopapilledema," and (3) primary chronic open-angle glaucoma.
Diplopia Of 53 patients with diplopia, 11 (20%) were overtested. Table 2 indicates the scope of the overtesting and cost. As with optic neuropathies, neuroimaging studies were the procedures most frequently abused. The most common diagnoses that produced excessive testing were ( 1) ocular myasthenia gravis; (2) microvascular ischemic mononeuropathy involving the third, fourth, and sixth cranial nerves; and (3) thyroid eye disease (Graves ophthalmopathy).
Proptosis None of the 18 patients with proptosis were subjected to excessive diagnostic procedures. All procedures were evaluated appropriately with CT scans, although for ten patients the scans were not of sufficient quality to evaluate the orbits properly.
Ptosis Of the 19 patients with ptosis, 3 ( 16%) were overtested as shown in Table 3. One patient had ocular myasthenia gravis, and two had dehiscence of the levator palpebral muscle (age-related ptosis). Again, neuro-radiologic imaging procedures were the inappropriate studies.
Results
Cost Comparison
Five hundred eighty-eight new patient charts were reviewed.
The total cost of unnecessary diagnostic studies for these 35 patients was $65,900 in 1989 dollars. The cost assess-
1628
Dillon et al · Cost Effectiveness of Diagnostic Management by Gatekeepers Table 2. Overtesting for Diplopia Test
No. of Patients
MRI scan CT scan
7
Arteriography
11 3
MRI
=
=
magnetic resonance image; CT
Cost
$6300
$5500 $5400 computed tomography.
ment does not consider hospitalization costs for angiography, potential morbidity of the procedures, and subsequent costs due to complications of the invasive procedures. In comparison, if the patients had been referred at the onset of their problems for a neuro-ophthalmologic consultation ($200 per patient), the net cost would have been $8000. Therefore, the gatekeeper system generated $57,900 in unnecessary costs, a 724% overcharge, without including the additional costs of hospitalization in some patients. The 724% figure may be somewhat of an overestimation because some diagnostic testing may have been ordered if the patients had been referred directly to a neuroophthalmologist. However, the diagnostic categories of optic neuropathy, diplopia, proptosis, and ptosis rarely require extensive neuroimaging studies in standard neuroophthalmologic algorithmic diagnostic evaluations. 10 Another method of cost comparison would be to tabulate the total testing costs of the 588 patients in the four categories surveyed. The excess testing charges then could be viewed as a percentage of charges, perhaps resulting in a less dramatic surcharge amount. Attempts to perform such an analysis for this cohort of patients was unsuccessful because of incomplete information about the costs, both in terms of testing, physicians' time, and adminis~ trative expenses, before being referred for neuro-ophthalmologic evaluation. Patient outcomes. The 35 patients subjected to excessive testing did not incur additional loss of vision or neurologic sequelae due to diagnostic delay. In addition, the patients did not have any significant complications from the testing procedures.
Discussion Our data indicate that between 16% and 26% of patients in three of four neuro-ophthalmologic diagnostic categories were subjected to unnecessary and expensive diagnostic procedures under the direction of "gatekeeper physicians." In the one category in which the diagnostic approach was appropriate, the quality of the neuroimaging studies often was found to be substandard. Compared with a prompt referral for neuro-ophthalmologic consultation, the gatekeeper system generated $57,900 in excessive costs, a 724% overcharge. In addition, many patients may have undergone the same unnecessary procedures and not been referred for consultation.
Neuro-radiologic imaging studies, including CT and MRI scans as well as cerebral angiography, encompassed the majority of unnecessary testing. The excess neuroradiologic procedures imply that the overtested patients have so-called "CT and MRI negative" disease, and these disorders frequently are diagnosed based on medical history and physical findings. One explanation for the results is that gatekeeper physicians lack education and experience in the basic fundamentals of medical history taking and physical examination assessment involved in complex neuro-ophthalmic problems. Residency and post-residency fellowship training would appear to be a prerequisite to elicit precise medical history data and physical findings, and thereby save healthcare dollars. An unavoidable shortcoming of this study is an inability to determine how many patients with these diagnostic problems were appropriately managed by the referring physicians. At this time, we do not have access to such records. The current study hopefully will serve as a stimulus to begin a prospective investigation to examine this issue. It is also possible that if the patients were seen first by neuro-ophthalmologists that some overtesting may have been incurred. We tried to minimize this problem by choosing relatively common neuro-ophthalmologic disorders in which the diagnostic criteria and management approach are in general agreement. 10 Because we did not use an outside arbitrator to judge "overtesting," this study may have incurred some additional bias. An attempt was made to overcome this potential error by requiring that all three neuro-ophthalmologists agree when a patient was included as "overtested." In addition, it is difficult to attach a price to the cost of delayed diagnosis engendered by "false negative CT or MRI scans" obtained by gatekeeper physicians. In this context, we refer to patients who had structural, potentially treatable lesions and then incurred diagnostic delays. Although the current study did not have a means to examine this aspect of the gatekeeper system, future studies will need to address the issue of delays in diagnosis to the point where therapy is inadequate to reverse the visual loss while the patient was presumed to have another, nontreatable diagnosis. Fortunately, none of the patients in this study had loss of vision or permanent neurologic disability because of a lengthy diagnostic delay. An alternative method ofhealthcare delivery would be the "centers of excellence" concept into which difficult diagnostic problems are triaged immediately. In this approach, the gatekeeper would direct the patient to the
Table 3. Overtesting for Ptosis Test
No. of Patients
Cost
2 2
$1800 $1000 $1800
MRI scan CT scan Arteriography MRI
=
magnetic resonance image; CT
=
computed tomography.
1629
Ophthalmology
Volume 101, Number 9, September 1994
proper tertiary referral center and perhaps a "tertiary care gatekeeper," instead of ordering a standard battery of diagnostic procedures. These centers of excellence would be staffed by physicians with post-residency fellowship training and certified by peer-review groups. In summary, the gatekeeper system was found not to be cost effective for the evaluation offour common neuroophthalmologic problems. Although the Hsiao RBRVS system and the health maintenance organization network have been valuable in providing stimuli to save healthcare dollars in some respects, it appears that the system may break down for highly specialized problems. In actuality, post-graduate subspecialty medical education appears to provide patients with neuro-ophthalmic problems with more cost-effective and safer healthcare. Based on the data in this study, the gatekeeper method for neuro-ophthalmology consumed rather than conserved a significant amount of health care dollars within a very limited time.
References I. Williams AP. Memorandum to the President-elect. Parameters for health system reforms [editorial]. JAMA 1992;268: 2699-2700.
1630
2. Hsiao WC, Braun P, Yntema D, Becker ER. Estimating physicians' work for a resource-based relative-value scale. N Eng! J Med 1988;319:835-41. 3. Hsiao WC, Braun P, Dunn D, et a!. Results and policy implications of the resource-based relative-value study. N Eng! J Med 1988;319:881-8. 4. Hsiao WC, Dunn DL, Verrilli DK. Assessing the implementation of physician-payment reform. N Eng! J Med 1993;328:928j33. 5. Health Care Firancing Administration. Medicare program: fee schedule fo~ physicians services: final notice. Fed Regist 1991 ;56:(227):59577-87, 59635-784. 6. Franks P, Clanby CM, Nutting PA. Gatekeeping revisitedprotecting patients from overtreatment. N Eng! J Med 1992;327:424-9. 7. Greenfield S, Nelson EC, Zubkoff M, eta!. Variations in resource utilization among medical specialties and systems of care. Results from the Medical Outcomes Study. JAMA 1992;267: 1624-30. 8. Hillman AL. Health maintenance organizations, financial incentives, and physicians' judgements. Ann Intern Med 1990; 112:891-3. 9. Hiduchenko K. Do health maintenance organizations control costs or shift costs [letter]? N Eng! J Med 1993;328: 971. 10. Burde RM, Savino PJ, Trobe JD. Clinical Decisions in Neuro-Ophthalmology, 2nd ed. St. Louis: CV Mosby, 1992.