- Email: [email protected]
Coronary arteriography performed by a physician assistant
CoronaryArteriographyPerformed by a PhysicianAssistant HENRY DeMOTS, MD, BARBARA COOMBS, PA, EDWARD MURPHY, MD, and ROBERT PALAC, MD
Funding constraints and an oversupply of cardiologists mitigate against continued training of increasing numbers of cardiology fellows. In some institutions, the workload of the catheterization laboratory is an overriding factor. The ability of a physician assistant to perform some of this work was tested to determine if the~number of fellows and the content of the fellowship training program could be uncoupled from the catheterization laboratory workload. Among the first 150 patients in whom coronary arteriography was performed by a physician assistant, no patient died or had a myocardial infarction or stroke. Two patients (1.3% ) had minor complications: a retinal embolus and an infected puncture site. The complication rate in 150 consecutive
cases performed by fellows was also 1.3 % , a small myocardial infarction and a transient ischemic attack. Procedure times for the physician assistant and for the fellows were 41 f 13 and 44 ,f 18 minutes for preoperative patients and 62 f 24 and 70 f 20 minutes for postoperative patients. Corresponding fluoroscopy times were 11 f 5 and 12 f 7 minutes for the preoperative and 22 f 12 and 20 f 6 for postoperative patients. Only preoperative fluoroscopy times were statistically different (p = 0.02). Thus, substituting a physician assistant for a fellow to perform coronary arteriography ts an option in institutions at which the number of studies exceeds the training needs of fellows. (Am J Cardiol 1987;60:784-787)
T
he number of cardiologists in the U.S. is increasing rapid1y.l Projections made by the Graduate Medical Education National Advisory Committee in 1980 suggested that the supply is not related to need. The Committee projected that the supply of cardiologists would be nearly twice the number required to meet the nation’s need in 19!Xk2 There may be disagreement regarding the precise number of cardiologists needed. It is clear, however, that the Committee’s projections of a growing supply of cardiologists are accurate3 and that an excess has occurred or will soon occur. Among the reasons given for continued growth in the numbers is a need for training institutions to manage their patient care workload. This workload not only plays a powerful role in establishing the size of the training program, but also strongly influences its
content because time for other academic and research pursuits often compete poorly with the need to provide timely patient care services. We initiated a pilot project to determine the feasibility of reducing the clinical workload of cardiology fellows in our training program by substituting a physician assistant for a cardiology fellow to perform coronary arteriography. We reasoned that a successful outcome of the project would permit us greater freedom to design our present training program based on educational and research needs rather than the patient care workload.
Methods Coronary arteriography in the context of this trial included the procedure itself and a number of activities before and after the procedure that relate to it. Having 1 person perform all of these activities enhances continuity of care. The Subcommittee on Human Studies of the Portland Veterans Administration Medical Center approved the protocol. The Board of Medical Examiners of the State of Oregon approved a practice description for the physician assistant that included performing coronary arteriography.
From the Division of Cardiology of the Portland Veterans Administration Medical Center and Oregon Health Sciences University, Portland, Oregon. Manuscript received March 2, 1987; revised manuscript received and accepted June 2,1987. Address for reprints: Henry DeMots, MD, Portland Veterans Administration Medical Center, 3719 S.W. Veterans Road, Portland, Oregon 97201. 784
OCtObW
The physician assistant in this trial was initially trained in a program approved by the American Medical Association Council on Allied Health Education and Accreditation and had 5 years of experience caring for patients with heart disease. Her experience did not include invasive procedures. Patients were evaluated as inpatients in response to a request from ward physicians who considered the patient a candidate for coronary arteriography. The results of the history, physical examination and noninvasive data were reviewed with 1 of 3 staff cardiologists and the patient was scheduled for arteriography. The physician assistant educated the patient regarding the risks and benefits of the procedure and obtained informed consent for the procedure. We used the Judkins transfemoral technique4 and a General Electric Fluoricon 300 biplane x-ray system. Physiologic information was acquired and processed using a Honeywell Meddars 300 cardiac catheterization laboratory computer system with a left ventricular analysis subsystem. Our standard technique includes biplane left ventricular cineangiography, 3 to 6 views of the left coronary artery system and 2 to 4 views of the right coronary artery system. A videotape recording was reviewed before withdrawal of catheters to assure the diagnostic adequacy of the study. Patients who required evaluations that included elements other than angiographic visualization of the left ventricle, coronary arteries and bypass grafts were excluded. Exceptions included hemodynamic evaluations for valvular heart disease, electrophysiologic studies and percutaneous transluminal angioplasty. These procedures were considered to be important training opportunities for fellows and were performed exclusively by them. After the procedure the physician assistant verified the accuracy of the computer analysis of the pressure data and determined the ventricular volumes and ejection fraction. A staff cardiologist reviewed these data and the physician assistant’s interpretation of the coronary artery anatomy. The physician assistant communicated the results to the patient and the physicians caring for the patient and, when appropriate, presented the data to the cardiac surgery team for consideration of bypass surgery. We recorded the frequency and type of complications, amount of fluoroscopy time and total procedure time. We observed for the following complications: death, myocardial infarction, stroke, embolus, infection, vascular obstruction, and bleeding that resulted in a need for surgical intervention or prolonged hospitalization. We also noted the frequency of incomplete studies in which catheterization and angiography of the left ventricle and both coronary arteries and bypass grafts, if present, could not be accomplished. Fluoroscopy time, which served as an index of the ability to manipulate catheters within the vascular system, was measured by a timing device attached to the fluoroscopy switch. The procedure time, which served as an index of proficiency in all aspects of the procedure, was measured from the time when the local anesthetic was injected to the time when the last catheter
7, 1987
WE
AXRiCAR;
Compiications
JOL‘RRiAL
OF CARDIOLOGY
and t~c~~p~eie Fellows (n =
Death Myocardial infarction Transient ischemic attack Retinal embolus Serious arrhythmia Septicemia incomplete study
150) 0 1 1 0 0 0 3
Volume
60
785
Studies Physician
Assistant
(n = 150) 0 0 0 1 0 1 1
was withdrawn. Because the complexity of the angiographic procedure is greater in patients with previous coronary bypass surgery, we segregated the data into preoperative and postoperative groups. We analyzed the first 150 patients in the physician assistant group. Three patients declined to participate in the trial. We assessed the same variables in 150 consecutive studies performed by cardiology fellows, beginning at the onset of an academic year. The same exclusion criteria were applied. Three fellows in their second clinical year and one in the first clinical year participated. We cornpared the performance of the physician assistant with that of the fellows rather than the performance of staff cardiologists because the fellow model is the model used in most academic centers and is probably used in developing the standards used in the literature. A staff cardiologist, who was present in the procedure room, supervised all procedures, whether performed by a fellow or the physician assistant. Patients were not randomly allocated to 1 of the 2 groups, nor were they purposely selected for 1 group or the other on the basis of anticipated difficulty or ease in performing the procedure. Availability of the fellow and physician assistant at a time when the laboratory was available or prior knowledge of the patient usually dictated who performed the procedure. The frequency of complications in the groups was compared using a chi-square test. Fluoroscopy times and procedure times were compared using a nonpaired t test.
Results No patient in either group died or had arrhythmias that required specific therapy. There were 2 complications among 150 patients (1.3%) in the physician assistant group and 2 complications among 150 patients in the fellow group [Table I). These results compare favorably with published results of the incidence of complications for the procedure.5-7 One myocardial infarction, diagnosed by cardiac enzyme levels and electrocardiogram, occurred among the patients studied by a fellow. Chest pain occurred several hours after an uneventful study. Another patient in the fellow group had slurred speech and transient left hemiparesis during the procedure. Three studies were incomplete. One patient studied by the physician assistant reported a visual field defect that began during the procedure, but this was not reported until the next day.
788
NONPHYSICIAN
TABLE
II
Fluoroscopy
CORONARY
and
ARTERIOGRAPHY
Procedure
Times
Preoperative Fluoroscopy time Physician assistant Fellow Procedure time Physician assistant Fellow
(minutes) Postoperative
11 f 12 f
5’ 7”
22 f 21 f
12 6
41 f 63 f
18 25
44f 70 f
18 20
“p <0.02.
Another patient in whom catheterization was performed using an artery that had been catheterized 1 week earlier had a fever and pectechiae of the extremity the day after the procedure. Rapid resolution occurred with antibiotic therapy. One study in the physician assistant group was incomplete. The fluoroscopy and procedure times in the preoperative and postoperative groups are shown in Table II. Only the difference in fluoroscopy times in the preoperative patients was statistically significant.
Discussion The results of this investigation suggest that coronary arteriography, a procedure generally considered to require skills found only in physicians, can be safely performed by a physician assistant. The complication rates and the time required to perform the procedures were similar. The shorter time required for fluoroscopy by the physician assistant in the preoperative patients may be related to an unidentified bias in case selection, since patients were not randomized. The number of studies performed without a death by the physician assistant is small and the mortality rate for widespread use of the model is still unknown. However, these data suggest that there is nothing inherent in the procedure that mandates that it be performed only by physicians. One might choose to develop a model in which some of these duties are performed by technical personnel. We believe there may be advantages to using a specifically trained professional such as a physician assistant, whose involvement with the patient begins with initial evaluation and may continue, uninterrupted, into long-term outpatient care. We assumed that continuity of care not only enhances patient satisfaction, but also quality of care. Delivery of care and trainee numbers: The academic training model in which residents and subspecialty fellows implement a plan in concert with a faculty member provides effective training and highquality patient care. There is concern, however, that the fellow model is producing an oversupply of specialists. The oversupply of specialists in turn may lead to an increase in the aggregate health care costs of the nation and threatens survival of the teaching hospitals that produce these trainees.1~3~8-10 Physician assistants have successfully substituted for residents, enabling teaching hospitals to function
with fewer residents than would otherwise be possible.11J2 This study shows a role for the physician assistant in support of the technical duties of the medical subspecialist in addition to primary care duties. Patient care revenues and general appropriations provide over 60% of the funds required for fellowship stipends.13 The fraction of subspecialty stipends that is paid by hospitals is increasing.3J3 As long as hospitals are reimbursed for the clinical care provided by trainees there appears to be insufficient incentive to reduce their number even though they will eventually ‘compete with the training institution for patients. Reimbursement for clinical training will be decreased when proposed changes in Medicare reimbursement policy are implemented, however.13 Since the report of the Graduate Medical Education National Advisory Committee in 1980 the number of cardiology trainees has not diminished, even though cardiology was cited as one of the most oversupplied of the subspecialties. A 4.9% increase in the number of cardiology trainees occurred between 1983 and 1984.3 Introducing this model may be perceived as a way of increasing the aggregate capability for cardiology services at a time when there is an excess of cardiologists. The model, however, by providing a mechanism to cope with fewer trainees, will have the opposite effect. O’Rourke described several policies that could be implemented on a national scale to reduce the number of trainees.lO Whether these methods are used may be determined by changes in funding mechanisms beyond the control of the training director.14 Structure of training programs: Dependence on the trainee for clinical productivity may compromise academic and research efforts.1°J5 This is particularly true of the trainee who anticipates a career in research. The allocation of time of these young physicians is often determined primarily by the patient mix and style of practice of the institution. However, when 1 aspect of training displaces opportunities in other aspects of clinical training or in research, a training program director may need to uncouple the allocation of trainee time from the clinical workload. Scheduling of catheterization procedures to be performed by a physician assistant while the fellows were attending educational conferences increases the productivity of the catheterization laboratory and allows all fellows to participate in the conferences. Allocation of the cardiologist’s time: Widespread use of the physician assistant model will not be necessary while the number of clinical cardiology trainees is increasing. If a decrease occurs, the cardiology faculty could perform the work. In some instances this may be appropriate, but in others it will be important to preserve the time of the faculty member for teaching and research. Large institutions, hospital groups or corporations formed for the purpose of providing comprehensive, prepaid health care may develop angiography teams that provide the service for their constituency. We have shown a role for the physician assistant in that setting, but further study is required to determine cost-effectiveness.
October
The time of the staff cardiologist is savedby using all of the advantagesof the fellow model. We closely supervise all critical activities of both fellows and physician assistants,Although the requirement for over-the-shouldersupervision is the same as the fellow model, it representsa time saving for the staff cardiologistcomparedwith performing the role of primary operator. Additional time is saved becausethe physician assistant,unlike the cardiology fellow, does not leave the program after 3 years to be replaced by another untrained person. A changein the method of reimbursement of physicians to one of several types of prospective payment systems is expected.16,~If this occurs, cardiologists may reexamine the way they spend their time. The cognitive elements of their activities may receive greater emphasis and the technical elements less. Acknowledgment: We are indebted to Dr. J. David Bristow and Dr. JohnA. Bensonfor their review of the manuscript, encouragementand suggestions.
References 1. Graduate Medical Education Advisory tary, Department of Health and Human October 1981.
Committee, Report Services. Publ. No.
to the Secre-
(HRA) 82-617,
1, 1987
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume
60
783
2. Graduate
Medical Education National Advisory Committee. Report to the Department of Health and Human Services. PubJ. No. (HRA) 81651. April 1981. 3. Schleiter MK, Tarlov AR. National Study of Internal Medicine Manpower. IX. Internai medicine residency and fellowship training: 1984 Update. Ann Intern Med 1985;102:681-685, 4. Judkins MP. Selective coronary arteriography: a percutaneous tronsfemoral technic. Radiology 3967;80:815-824. 5. Bourassa MC, Noble J. Complication rate of coronary arteriography: a review of528 cases studied by a percutaneous femoral technique. Circulation 1976;53:106-114. 6. Davis K, Kennedy ]W, Kemp HG, Judkins MP, Gosselin AJ, Killip T. Complications of coronary arteriogrophy from the collaborative study of coronary artery surgery (CASS]. Circulation 2979;59:1105-1111. - ’ ‘7. Abrams HL. Complications of coronary arteriography. In: Abrams HI,, ed. Coronary Arteriography: A Practical Approach. Boston: Little. Brown, 1988: 71-86. 8. Petersdorf RG. Is the establishment defensible? N EngJ J Med 1988; 309:1053-1057. 6. January C. Manpower changes in the future of cardiology. Chest 1981; Secretary,
81:138-139.
10. O’Rourke RA. Cardiology at a precipice. Circulation 1985;72:258-261. 11. Silver HK, McAtee PA. On the use of nonphysician “associate residents” in overcrowded specia!ty training programs N EngJ J Med 1984;311:326-328. 12. Perry HB, Detmer DE, Redmond EL. The current and future role of surgical physician assistants: report of national surveyofsurgical chairmen in Jorge U.S. hospitals. Ann Surg 1981;193:132-137, 13. Pfordresher K. House staff expenditures and funding in COTH member hospitals. J Med Educ 1985;60:422-424. 14. Iglehart JK. Federal support of graduate medical education. N Engl J Med 1985;312:1000-1004, 15. Murphy E. Cardiology fellowship training programs. Chest 1982;82:139140. 16. Petersdorf RG. Current and future directions for hospital and physician reimbursement. JAMA 1985;253:2543-2548. 17. Mitchell JB. Physician DRGs. N EngJ J Med 1985;313:670.
Funding constraints and an oversupply of cardiologists mitigate against continued training of increasing numbers of cardiology fellows. In some institutions, the workload of the catheterization laboratory is an overriding factor. The ability of a physician assistant to perform some of this work was tested to determine if the~number of fellows and the content of the fellowship training program could be uncoupled from the catheterization laboratory workload. Among the first 150 patients in whom coronary arteriography was performed by a physician assistant, no patient died or had a myocardial infarction or stroke. Two patients (1.3% ) had minor complications: a retinal embolus and an infected puncture site. The complication rate in 150 consecutive
cases performed by fellows was also 1.3 % , a small myocardial infarction and a transient ischemic attack. Procedure times for the physician assistant and for the fellows were 41 f 13 and 44 ,f 18 minutes for preoperative patients and 62 f 24 and 70 f 20 minutes for postoperative patients. Corresponding fluoroscopy times were 11 f 5 and 12 f 7 minutes for the preoperative and 22 f 12 and 20 f 6 for postoperative patients. Only preoperative fluoroscopy times were statistically different (p = 0.02). Thus, substituting a physician assistant for a fellow to perform coronary arteriography ts an option in institutions at which the number of studies exceeds the training needs of fellows. (Am J Cardiol 1987;60:784-787)
T
he number of cardiologists in the U.S. is increasing rapid1y.l Projections made by the Graduate Medical Education National Advisory Committee in 1980 suggested that the supply is not related to need. The Committee projected that the supply of cardiologists would be nearly twice the number required to meet the nation’s need in 19!Xk2 There may be disagreement regarding the precise number of cardiologists needed. It is clear, however, that the Committee’s projections of a growing supply of cardiologists are accurate3 and that an excess has occurred or will soon occur. Among the reasons given for continued growth in the numbers is a need for training institutions to manage their patient care workload. This workload not only plays a powerful role in establishing the size of the training program, but also strongly influences its
content because time for other academic and research pursuits often compete poorly with the need to provide timely patient care services. We initiated a pilot project to determine the feasibility of reducing the clinical workload of cardiology fellows in our training program by substituting a physician assistant for a cardiology fellow to perform coronary arteriography. We reasoned that a successful outcome of the project would permit us greater freedom to design our present training program based on educational and research needs rather than the patient care workload.
Methods Coronary arteriography in the context of this trial included the procedure itself and a number of activities before and after the procedure that relate to it. Having 1 person perform all of these activities enhances continuity of care. The Subcommittee on Human Studies of the Portland Veterans Administration Medical Center approved the protocol. The Board of Medical Examiners of the State of Oregon approved a practice description for the physician assistant that included performing coronary arteriography.
From the Division of Cardiology of the Portland Veterans Administration Medical Center and Oregon Health Sciences University, Portland, Oregon. Manuscript received March 2, 1987; revised manuscript received and accepted June 2,1987. Address for reprints: Henry DeMots, MD, Portland Veterans Administration Medical Center, 3719 S.W. Veterans Road, Portland, Oregon 97201. 784
OCtObW
The physician assistant in this trial was initially trained in a program approved by the American Medical Association Council on Allied Health Education and Accreditation and had 5 years of experience caring for patients with heart disease. Her experience did not include invasive procedures. Patients were evaluated as inpatients in response to a request from ward physicians who considered the patient a candidate for coronary arteriography. The results of the history, physical examination and noninvasive data were reviewed with 1 of 3 staff cardiologists and the patient was scheduled for arteriography. The physician assistant educated the patient regarding the risks and benefits of the procedure and obtained informed consent for the procedure. We used the Judkins transfemoral technique4 and a General Electric Fluoricon 300 biplane x-ray system. Physiologic information was acquired and processed using a Honeywell Meddars 300 cardiac catheterization laboratory computer system with a left ventricular analysis subsystem. Our standard technique includes biplane left ventricular cineangiography, 3 to 6 views of the left coronary artery system and 2 to 4 views of the right coronary artery system. A videotape recording was reviewed before withdrawal of catheters to assure the diagnostic adequacy of the study. Patients who required evaluations that included elements other than angiographic visualization of the left ventricle, coronary arteries and bypass grafts were excluded. Exceptions included hemodynamic evaluations for valvular heart disease, electrophysiologic studies and percutaneous transluminal angioplasty. These procedures were considered to be important training opportunities for fellows and were performed exclusively by them. After the procedure the physician assistant verified the accuracy of the computer analysis of the pressure data and determined the ventricular volumes and ejection fraction. A staff cardiologist reviewed these data and the physician assistant’s interpretation of the coronary artery anatomy. The physician assistant communicated the results to the patient and the physicians caring for the patient and, when appropriate, presented the data to the cardiac surgery team for consideration of bypass surgery. We recorded the frequency and type of complications, amount of fluoroscopy time and total procedure time. We observed for the following complications: death, myocardial infarction, stroke, embolus, infection, vascular obstruction, and bleeding that resulted in a need for surgical intervention or prolonged hospitalization. We also noted the frequency of incomplete studies in which catheterization and angiography of the left ventricle and both coronary arteries and bypass grafts, if present, could not be accomplished. Fluoroscopy time, which served as an index of the ability to manipulate catheters within the vascular system, was measured by a timing device attached to the fluoroscopy switch. The procedure time, which served as an index of proficiency in all aspects of the procedure, was measured from the time when the local anesthetic was injected to the time when the last catheter
7, 1987
WE
AXRiCAR;
Compiications
JOL‘RRiAL
OF CARDIOLOGY
and t~c~~p~eie Fellows (n =
Death Myocardial infarction Transient ischemic attack Retinal embolus Serious arrhythmia Septicemia incomplete study
150) 0 1 1 0 0 0 3
Volume
60
785
Studies Physician
Assistant
(n = 150) 0 0 0 1 0 1 1
was withdrawn. Because the complexity of the angiographic procedure is greater in patients with previous coronary bypass surgery, we segregated the data into preoperative and postoperative groups. We analyzed the first 150 patients in the physician assistant group. Three patients declined to participate in the trial. We assessed the same variables in 150 consecutive studies performed by cardiology fellows, beginning at the onset of an academic year. The same exclusion criteria were applied. Three fellows in their second clinical year and one in the first clinical year participated. We cornpared the performance of the physician assistant with that of the fellows rather than the performance of staff cardiologists because the fellow model is the model used in most academic centers and is probably used in developing the standards used in the literature. A staff cardiologist, who was present in the procedure room, supervised all procedures, whether performed by a fellow or the physician assistant. Patients were not randomly allocated to 1 of the 2 groups, nor were they purposely selected for 1 group or the other on the basis of anticipated difficulty or ease in performing the procedure. Availability of the fellow and physician assistant at a time when the laboratory was available or prior knowledge of the patient usually dictated who performed the procedure. The frequency of complications in the groups was compared using a chi-square test. Fluoroscopy times and procedure times were compared using a nonpaired t test.
Results No patient in either group died or had arrhythmias that required specific therapy. There were 2 complications among 150 patients (1.3%) in the physician assistant group and 2 complications among 150 patients in the fellow group [Table I). These results compare favorably with published results of the incidence of complications for the procedure.5-7 One myocardial infarction, diagnosed by cardiac enzyme levels and electrocardiogram, occurred among the patients studied by a fellow. Chest pain occurred several hours after an uneventful study. Another patient in the fellow group had slurred speech and transient left hemiparesis during the procedure. Three studies were incomplete. One patient studied by the physician assistant reported a visual field defect that began during the procedure, but this was not reported until the next day.
788
NONPHYSICIAN
TABLE
II
Fluoroscopy
CORONARY
and
ARTERIOGRAPHY
Procedure
Times
Preoperative Fluoroscopy time Physician assistant Fellow Procedure time Physician assistant Fellow
(minutes) Postoperative
11 f 12 f
5’ 7”
22 f 21 f
12 6
41 f 63 f
18 25
44f 70 f
18 20
“p <0.02.
Another patient in whom catheterization was performed using an artery that had been catheterized 1 week earlier had a fever and pectechiae of the extremity the day after the procedure. Rapid resolution occurred with antibiotic therapy. One study in the physician assistant group was incomplete. The fluoroscopy and procedure times in the preoperative and postoperative groups are shown in Table II. Only the difference in fluoroscopy times in the preoperative patients was statistically significant.
Discussion The results of this investigation suggest that coronary arteriography, a procedure generally considered to require skills found only in physicians, can be safely performed by a physician assistant. The complication rates and the time required to perform the procedures were similar. The shorter time required for fluoroscopy by the physician assistant in the preoperative patients may be related to an unidentified bias in case selection, since patients were not randomized. The number of studies performed without a death by the physician assistant is small and the mortality rate for widespread use of the model is still unknown. However, these data suggest that there is nothing inherent in the procedure that mandates that it be performed only by physicians. One might choose to develop a model in which some of these duties are performed by technical personnel. We believe there may be advantages to using a specifically trained professional such as a physician assistant, whose involvement with the patient begins with initial evaluation and may continue, uninterrupted, into long-term outpatient care. We assumed that continuity of care not only enhances patient satisfaction, but also quality of care. Delivery of care and trainee numbers: The academic training model in which residents and subspecialty fellows implement a plan in concert with a faculty member provides effective training and highquality patient care. There is concern, however, that the fellow model is producing an oversupply of specialists. The oversupply of specialists in turn may lead to an increase in the aggregate health care costs of the nation and threatens survival of the teaching hospitals that produce these trainees.1~3~8-10 Physician assistants have successfully substituted for residents, enabling teaching hospitals to function
with fewer residents than would otherwise be possible.11J2 This study shows a role for the physician assistant in support of the technical duties of the medical subspecialist in addition to primary care duties. Patient care revenues and general appropriations provide over 60% of the funds required for fellowship stipends.13 The fraction of subspecialty stipends that is paid by hospitals is increasing.3J3 As long as hospitals are reimbursed for the clinical care provided by trainees there appears to be insufficient incentive to reduce their number even though they will eventually ‘compete with the training institution for patients. Reimbursement for clinical training will be decreased when proposed changes in Medicare reimbursement policy are implemented, however.13 Since the report of the Graduate Medical Education National Advisory Committee in 1980 the number of cardiology trainees has not diminished, even though cardiology was cited as one of the most oversupplied of the subspecialties. A 4.9% increase in the number of cardiology trainees occurred between 1983 and 1984.3 Introducing this model may be perceived as a way of increasing the aggregate capability for cardiology services at a time when there is an excess of cardiologists. The model, however, by providing a mechanism to cope with fewer trainees, will have the opposite effect. O’Rourke described several policies that could be implemented on a national scale to reduce the number of trainees.lO Whether these methods are used may be determined by changes in funding mechanisms beyond the control of the training director.14 Structure of training programs: Dependence on the trainee for clinical productivity may compromise academic and research efforts.1°J5 This is particularly true of the trainee who anticipates a career in research. The allocation of time of these young physicians is often determined primarily by the patient mix and style of practice of the institution. However, when 1 aspect of training displaces opportunities in other aspects of clinical training or in research, a training program director may need to uncouple the allocation of trainee time from the clinical workload. Scheduling of catheterization procedures to be performed by a physician assistant while the fellows were attending educational conferences increases the productivity of the catheterization laboratory and allows all fellows to participate in the conferences. Allocation of the cardiologist’s time: Widespread use of the physician assistant model will not be necessary while the number of clinical cardiology trainees is increasing. If a decrease occurs, the cardiology faculty could perform the work. In some instances this may be appropriate, but in others it will be important to preserve the time of the faculty member for teaching and research. Large institutions, hospital groups or corporations formed for the purpose of providing comprehensive, prepaid health care may develop angiography teams that provide the service for their constituency. We have shown a role for the physician assistant in that setting, but further study is required to determine cost-effectiveness.
October
The time of the staff cardiologist is savedby using all of the advantagesof the fellow model. We closely supervise all critical activities of both fellows and physician assistants,Although the requirement for over-the-shouldersupervision is the same as the fellow model, it representsa time saving for the staff cardiologistcomparedwith performing the role of primary operator. Additional time is saved becausethe physician assistant,unlike the cardiology fellow, does not leave the program after 3 years to be replaced by another untrained person. A changein the method of reimbursement of physicians to one of several types of prospective payment systems is expected.16,~If this occurs, cardiologists may reexamine the way they spend their time. The cognitive elements of their activities may receive greater emphasis and the technical elements less. Acknowledgment: We are indebted to Dr. J. David Bristow and Dr. JohnA. Bensonfor their review of the manuscript, encouragementand suggestions.
References 1. Graduate Medical Education Advisory tary, Department of Health and Human October 1981.
Committee, Report Services. Publ. No.
to the Secre-
(HRA) 82-617,
1, 1987
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume
60
783
2. Graduate
Medical Education National Advisory Committee. Report to the Department of Health and Human Services. PubJ. No. (HRA) 81651. April 1981. 3. Schleiter MK, Tarlov AR. National Study of Internal Medicine Manpower. IX. Internai medicine residency and fellowship training: 1984 Update. Ann Intern Med 1985;102:681-685, 4. Judkins MP. Selective coronary arteriography: a percutaneous tronsfemoral technic. Radiology 3967;80:815-824. 5. Bourassa MC, Noble J. Complication rate of coronary arteriography: a review of528 cases studied by a percutaneous femoral technique. Circulation 1976;53:106-114. 6. Davis K, Kennedy ]W, Kemp HG, Judkins MP, Gosselin AJ, Killip T. Complications of coronary arteriogrophy from the collaborative study of coronary artery surgery (CASS]. Circulation 2979;59:1105-1111. - ’ ‘7. Abrams HL. Complications of coronary arteriography. In: Abrams HI,, ed. Coronary Arteriography: A Practical Approach. Boston: Little. Brown, 1988: 71-86. 8. Petersdorf RG. Is the establishment defensible? N EngJ J Med 1988; 309:1053-1057. 6. January C. Manpower changes in the future of cardiology. Chest 1981; Secretary,
81:138-139.
10. O’Rourke RA. Cardiology at a precipice. Circulation 1985;72:258-261. 11. Silver HK, McAtee PA. On the use of nonphysician “associate residents” in overcrowded specia!ty training programs N EngJ J Med 1984;311:326-328. 12. Perry HB, Detmer DE, Redmond EL. The current and future role of surgical physician assistants: report of national surveyofsurgical chairmen in Jorge U.S. hospitals. Ann Surg 1981;193:132-137, 13. Pfordresher K. House staff expenditures and funding in COTH member hospitals. J Med Educ 1985;60:422-424. 14. Iglehart JK. Federal support of graduate medical education. N Engl J Med 1985;312:1000-1004, 15. Murphy E. Cardiology fellowship training programs. Chest 1982;82:139140. 16. Petersdorf RG. Current and future directions for hospital and physician reimbursement. JAMA 1985;253:2543-2548. 17. Mitchell JB. Physician DRGs. N EngJ J Med 1985;313:670.