Consultation Between Cardiologists and Generalists in the Management of Acute Myocardial Infarction Implications for Quality of Care

ORIGINAL INVESTIGATION Consultation Between Cardiologists and Generalists in the Management of Acute Myocardial Infarction Implications for Quality of Care

Donald J. Willison, ScD; Stephen B. Soumerai, ScD; Thomas J. McLaughlin, ScD; Jerry H. Gurwitz, MD; Xiaoming Gao, MA; Edward Guadagnoli, PhD; Steven Pearson, MD; Paul Hauptman, MD; Barbara McLaughlin, BAN

Background: The rapid expansion of managed care in was no variation in the use of effective agents between the United States has increased debate regarding the ap- patients cared for by a cardiologist attending physician propriate mix of generalist and specialist involvement in and a generalist with a consultation by a cardiologist. How- medical care. ever, there was a consistent trend toward increased use of aspirin, thrombolytics, and ␤-blockers in these pa- Objective: To compare the quality of medical care when tients compared with those with a generalist attending generalists and cardiologists work separately or to- physician only (PϽ.05 for ␤-blockers only). Differences gether in the management of patients with acute myo- between groups in the use of lidocaine were not statis- cardial infarction (AMI). tically significant. The adjusted probabilities of use of thrombolytics for consultative care and cardiologist at- Methods: We reviewed the charts of 1716 patients tending physicians were 0.73 for both. Corresponding with AMI treated at 22 Minnesota hospitals between probabilities were 0.86 and 0.85 for aspirin and 0.59 and 1992 and 1993. Patients eligible for thrombolytic aspi- 0.57 for ␤-blockers, respectively. rin, ␤-blockers, and lidocaine therapy were identified using criteria from the 1991 American College of Car- Conclusions: For patients with AMI, consultation be- diology guidelines for the management of AMI. We tween generalists and specialists may improve the quality compared the use of these drugs among eligible patients of care. Recent policy debates that have focused solely on whose attending physician was a generalist with no car- access to specialists have ignored the important issue of diologist input, a generalist with a cardiologist consul- coordination of care between generalist and specialist phy- tation, and a cardiologist alone. sicians. In hospitals where cardiology services are avail- able, generalists may be caring for patients with AMI who Results: Patients cared for by a cardiologist alone were are older and more frail. Future research and policy analy- younger, presented earlier to the hospital, were more likely ses should examine whether this pattern of selective re- to be male, had less severe comorbidity, and were more ferral is true for other medical conditions. likely to have an ST elevation of 1 mm or more than generalists’ patients. Controlling for these differences, there Arch Intern Med. 1998;158:1778-1783

HE RAPID expansion of man- The focus on comparative perfor- aged care in the United mance of generalists and specialists ig- States has increased pub- nores opportunities for sharing of knowl- lic and scientific debate re- edge and experience through formal and garding the appropriate mix informal consultation. In 1 study,4 the ofT generalist and specialist involvement in quality of psychoactive drug prescribing medical care. Previous research suggests in nursing homes was highest among gen- that cardiologists have better knowledge eralists who reported frequent consulta- than generalists concerning efficacious tions with psychiatrists. A recent study5 of therapies in the treatment of acute myo- a US health maintenance organization cardial infarction (AMI).1 However, pre- showed that an intervention fostering col- vious studies2 also suggest that there is laborative care between generalists and little relationship between knowledge or psychiatrists improved adherence to an- self-reported practice and actual behav- tidepressant regimens, patient satisfac- The affiliations of the ior. Despite intense controversy, few tion, and other outcomes in patients with 3 authors are listed in the studies exist comparing the care pro- major depression. Acknowledgments section at vided to comparable patients by general- In contrast to previous studies1,6,7 that the end of the article. ists and specialists. used survey data to measure differences be-

ARCH INTERN MED/ VOL 158, SEP 14, 1998 1778

Downloaded From: https://jamanetwork.com/ on 09/27/2021 PATIENTS AND METHODS cardiologist consultation. We defined 3 categories of generalist and cardiologist involvement in patient care: (1) generalist only, if the attending physician was an inter- PATIENT SAMPLE nist, family physician, or general practitioner, and there was no cardiologist consultation during the hospitaliza- The patient sample was drawn from a larger study of 37 tion; (2) consultative care, generalist attending with car- hospitals participating in a randomized controlled trial to diologist consultation; and (3) cardiologist attending phy- improve quality of care in the treatment of AMI in com- sician alone. In 6.7% of cases (n = 123), the attending munity hospitals.13 Data reported herein were collected prior physician was some other specialist and was excluded from to the intervention. In this study we included 20 hospitals the analysis. in Minneapolis, St Paul, Duluth, and St Cloud, Minn, which had both cardiologists and generalists on staff. We ex- DATA QUALITY cluded the 17 rural hospitals that did not have a cardiologist on staff because of confounding of the characteristics A complete description of the overall quality control meth- of rural practice and cardiologist availability. The 20 study ods have been reported elsewhere.13 Interrater agreement hospitals represented 91% of the number of community hos- concerning the identity of the attending physician was 86%. pital beds (adults) in the 4 urban areas. The medical records of 1839 patients admitted to the DATA ANALYSIS 20 hospitals for AMI between September 1992 and Au- gust 1993 were abstracted by trained nurse-abstractors. We To determine whether patient characteristics were simi- identified patients with an admission diagnosis of AMI or lar, we compared several demographic and clinical char- suspected AMI who met 2 or more of the following 3 cri- acteristics of patients across the 3 categories of generalist/ teria14: (1) clinical signs and symptoms consistent with AMI cardiologist involvement (Table 2). (arm or shoulder pain, chest pain, diaphoresis, dyspnea, Next, in separate models, we regressed the use of each nausea or vomiting, or neck/jaw pain); (2) electrocardio- study drug on the category of generalist/specialist involve- graphic evidence of AMI; or (3) laboratory evidence of AMI ment using binary logistic regression analysis,16 modeling (elevated levels of serum creatine kinase and its isoen- the 3 groups with 2 indicator variables, and controlling for zyme MB subfraction). We excluded patients who had sus- potentially confounding covariates. Covariates included pa- tained a myocardial infarction in the 2-week period prior tient age, sex, presence of severe comorbidity (Greenfield to the present hospital admission, were dead on hospital and coworkers’ Index of Coexistent Disease17), time from arrival, or were transferred from a nonstudy hospital. onset of symptoms to presentation, use of lipid-lowering For each drug, a subsample of eligible subjects was iden- drugs prior to admission, method of presentation (emer- tified, based on criteria developed from the 1990 American gency medical services transport, emergency department, College of Cardiology/American Heart Association (ACC/ or other), location of AMI (anterior, inferior, or other), and AHA) guidelines for the management of AMI, which were presence of ST elevation higher than 1 mm. Patient age and in effect in the study period.15 A detailed description of the sex were included in all statistical models. Other control eligibility criteria for each target therapy is listed in Table 1. variables were retained if for the Wald statistic PՅ.10 in all models. Age of the attending physician was initially in- PHYSICIAN SAMPLE cluded in the model, but dropped because it was not a significant predictor of use of the study drugs. Finally, we ad- We recorded the unique identification code and specialty justed for the hospital-level nesting of the binary dependent of the attending physician, along with evidence of any variable using generalized estimating equations.18

tween generalist and specialist care, we investigated dif- use of cardiologist consultations and had no patients ferences in quality of care actually provided to patients with treated by a cardiologist attending physician. At 1 hos- AMI when generalists and cardiologists work separately pital, almost all patients with AMI were under the care and in consultation. Quality of care was defined as care of a cardiologist attending physician. consistent with nationally recognized evidence-based prac- A description of patient characteristics, stratified tice guidelines. Specifically, we examined the use of drugs by level of generalist and specialist involvement in the known to reduce morbidity and mortality in eligible pa- care of that patient, is provided in Table 2. Patients tients (aspirin, thrombolytics, and ␤-blockers),8-11 and non- cared for by a cardiologist were younger and had less se- indicated use of lidocaine, which may cause increased mor- vere comorbidity. They were more likely to be male, to tality.12 We compared use of these drugs among patients be taking lipid-lowering medications prior to admis- whose attending physician was (1) a generalist with no sion, to have presented to the hospital within 6 hours of cardiologist input, (2) a generalist with cardiologist con- onset of symptoms, and to have electrocardiographic sultation, and (3) a cardiologist. findings clearly indicating AMI (ie, ST elevation Ն1 mm) (Table 2). RESULTS Controlling for differences in patient characteristics, cardiologist involvement in care, whether as a con- Seventeen (85%) of the 20 hospitals had similar distri- sultant or as the attending physician, was associated with butions of involvement of generalists and specialists in a statistically significant increase in the use of ␤- the management of patients. Two hospitals made little blockers only, with a consistent trend toward increased

ARCH INTERN MED/ VOL 158, SEP 14, 1998 1779

Downloaded From: https://jamanetwork.com/ on 09/27/2021 There was a trend toward increased use of lido- Table 1. Eligibility and Contraindications for Study Drugs* caine with increasing cardiologist involvement, but this was not statistically significant (Table 3; 95% CI, 0.6- Oral aspirin 2.6 and 0.6-4.7 for cardiologist consultation and cardi- Eligible population: all patients with AMI and no contraindications ologist attending physician, respectively). Because this Absolute contraindications: history of allergy to aspirin; serious gastrointestinal bleeding trend was unexpected, we conducted further analyses of Relative contraindications: asthma; nasal polyps (aspirin could lead specialty-associated variations in the use of study drugs to anaphylaxis); history of bleeding and/or significant risk of at the hospital level. bleeding; history of peptic ulcer disease When we plotted levels of use of aspirin, thrombo- ␤ -Blockers lytic agents, and ␤-blockers by hospital, the pattern of Eligible population: all patients with AMI and no contraindications Absolute contraindications: heart rate too low (Ͻ60/min); low increased use of these effective agents among patients in systolic blood pressure (Ͻ100 mm Hg); severe left ventricular the consultative care and cardiologist-treated groups was failure (rales Ͼ10 cm from base of lungs [10 cm = 1⁄3 from consistent across all hospitals. On the other hand, most base]); severe bronchospastic lung disease; signs of peripheral of the increase in use of lidocaine was concentrated among hypoperfusion; atrial ventricular conduction abnormalities; 14 cardiologists in 2 hospitals (Figure 2). In these 2 hos- history of adverse reaction to ␤-blockers Relative contraindications: systolic blood pressure Ͻ110 mm Hg; pitals, lidocaine was given along with thrombolytic history of asthma; severe peripheral vascular disease; therapy in 92% of cases and without thrombolytic difficult-to-control, severe, type 1 diabetes agents in 29% of cases. Corresponding figures for other Thrombolytics hospitals were 60% and 20%. Thus, the questionable Eligible population: all patients with AMI or suspected AMI use of lidocaine in these hospitals may have been due to presenting within 12 h of onset of symptoms; ST-segment elevation Ն1 mm; no medical contraindications prophylactic use of lidocaine for arrhythmias associated Absolute contraindications: active internal bleeding; suspected aortic with reperfusion. dissection; prolonged or traumatic cardiopulmonary resuscitation; recent head trauma (Յ2 wk); intracranial neoplasm; hemorrhagic ophthalmic conditions; pregnancy; COMMENT previous allergic reaction to the thrombolytic agent; sustained systolic blood pressure Ͼ180 mm Hg or diastolic blood pressure Overall, our results reinforce previous evidence suggest- Ͼ110 mm Hg; any recorded blood pressure Ͼ200/120 mm Hg ing that consultation between generalists and special- on admission; trauma or surgery Յ2 wk; AMI onset Ͼ24 h ists results in improved quality of care. In our study, the Relative contraindications: Major bleeding; recent trauma or surgery quality of consultative care for patients with AMI, mea- Ͼ2wkandϽ2 mo; history of chronic severe hypertension with or without drug therapy; history of CVA; current use of warfarin sured as the proportion of eligible patients who re- anticoagulants; prior use of streptokinase or APSAC (if they are ceived highly effective drugs, was almost identical to the the agents of choice); significant liver dysfunction; active peptic quality of care under a cardiologist attending physician, ulcer; AMI onset Ͼ12 h and somewhat higher than the quality of care provided Lidocaine by generalists without cardiologist consultation. Eligible population: patients with AMI (American College of Cardiology/American Heart Association [ACC/AHA] class I) Our data are generally consistent with the survey 1 or suspected AMI (ACC/AHA class IIa) with frequent data of Ayanian and colleagues, who found that cardi- (Ͼ6/min) ventricular premature beats; nonsustained or ologists were more willing to prescribe thrombolytic sustained (Ͼ30 s) ventricular tachycardia at a rate Ͼ100/min; agents, aspirin, and ␤-blockers for patients with AMI when ventricular fibrillation compared with internists or family physicians. How- Absolute contraindications: allergy to lidocaine ever, our data on actual practice in the sample of physi- *AMI indicates acute myocardial infarction; CVA, cerebrovascular cians in urban and suburban hospitals in Minnesota sug- accident; and APSAC, asinoylated plasminogen-streptokinase activator gest smaller differences between generalists and complex (anistreplase). Guidelines were previously published (JAMA. cardiologists in the use of these drugs. Also, our esti- 1998;279:1358-1363). mates of drug use are somewhat lower than the self- reported behavior,1 after adjusting for important clini- use of other effective medications (Table 3 and cal variables. Differences between these studies suggest Figure 1). Compared with generalist attending physi- a shortfall in the translation of knowledge into action. cians with no cardiologist input, odds ratios for use of To some extent, the differences may be due to social de- aspirin for consultative care and cardiologist attending sirability bias on the part of respondents to surveys. How- physicians alone were 1.5 (95% confidence interval [CI], ever, we may also have failed to capture some aspects of 1.0-2.2) and 1.4 (95% CI, 0.8-2.6), respectively. Odds the patients’ presentation that caused physicians to re- ratios were 1.9 (95% CI, 1.0-3.9) and 2.0 (95% CI, 0.8- frain from using these drugs. 4.8), respectively, for thrombolytic agents; 2.3 (95% CI, Our findings regarding use of lidocaine differ quali- 1.3-3.9) and 2.1 (95% CI, 1.1-3.9), respectively, for ␤- tatively from those of Ayanian and colleagues,1 who re- blockers. To provide a better representation of the ac- ported that cardiologists were less likely to prescribe pro- tual differences in prescribing of study medications, we phylactic lidocaine. We found a trend toward increased converted odds ratios into adjusted probabilities, con- use of lidocaine for nonindicated purposes among car- trolling for patient level covariates (Figure 1). The ad- diologists when compared with generalists. The trend justed probability of use of thrombolytics for consulta- seemed to be explained by a higher rate of inappropriate tive care and cardiologist attending physicians was 0.73 use among cardiologists in only 2 hospitals. The marked for both. Corresponding probabilities were 0.86 and 0.85 difference in use of lidocaine at these 2 sites presumably for aspirin; 0.59 and 0.57 for ␤-blockers, respectively. reflected the common practice, among cardiologists at

ARCH INTERN MED/ VOL 158, SEP 14, 1998 1780

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 2. Characteristics of Study Patients by Category of Involvement of Generalists and Cardiologists*

Generalist Cardiologist Cardiologist Variable Overall Only Consult Attending Physician P † Sample size 1716 (100) 273 (16) 932 (54) 511 (30) . . . Age, y Ͻ65 711 (41) 62 (23) 384 (41) 265 (52) 65-74 489 (28) 70 (26) 278 (30) 141 (28) Ͻ.01 Ն75 516 (30) 141 (52) 270 (29) 105 (21) Male sex 1061 (62) 129 (47) 586 (63) 346 (68) Ͻ.01 Insurance Medicare/HMO/other‡ 1612 (95) 260 (95) 879 (95) 473 (94) Uninsured 41 (2) 7 (3) 20 (2) 14 (3) .82 Medicaid 46 (3) 5 (2) 27 (3) 14 (3) History of AMI 447 (26) 75 (27) 246 (26) 126 (25) .65 Smoking status Never smoked 375 (23) 63 (25) 219 (24) 93 (19) Past smoker 775 (47) 126 (50) 428 (48) 221 (46) Ͻ.05 Current smoker 487 (30) 63 (25) 254 (28) 170 (35) Diabetes mellitus 369 (22) 66 (24) 215 (23) 88 (17) Ͻ.05 Location of AMI Anterior 440 (26) 63 (23) 239 (26) 138 (27) Ͻ.01 Inferior/posterior 702 (41) 89 (33) 380 (41) 233 (46) ST elevation Ն1 mm 937 (55) 120 (44) 480 (52) 337 (66) Ͻ.01 Presented with heart failure 337 (20) 74 (27) 191 (22) 72 (14) Ͻ.01 Severe comorbidity§ 390 (23) 93 (34) 203 (22) 94 (18) Ͻ.01 Preadmission medications Aspirin 442 (26) 68 (25) 248 (27) 126 (29) .68 ␤-blockers 310 (18) 39 (14) 192 (21) 79 (15) Ͻ.05 Nitrates 375 (22) 66 (24) 211 (23) 98 (19) .19 Lipid-lowering agents 603 (35) 54 (20) 345 (37) 204 (40) Ͻ.01 Digoxin 175 (10) 46 (17) 81 (9) 48 (9) Ͻ.01 Calcium channel blockers 428 (25) 73 (27) 228 (24) 127 (25) .75 Any contraindications to Aspirin 474 (28) 77 (28) 267 (29) 130 (25) .42 Thrombolytics࿣ 383 (22) 82 (30) 206 (22) 95 (19) Ͻ.01 ␤-Blockers 1409 (82) 222 (81) 746 (80) 441 (86) Ͻ.05 Symptom onset Ͻ6 h 1014 (60) 136 (50) 533 (57) 345 (68) Ͻ.01 First contact Through ED 704 (41) 86 (32) 390 (42) 228 (45) Ͻ.01 EMS transport 750 (44) 139 (51) 387 (42) 224 (44) Ͻ.05 Admitted to ICU/CCU 1543 (91) 229 (86) 830 (90) 484 (95) Ͻ.01 Median LOS, d Hospital . . . 5 6 6 . . . ICU/CCU . . . 2 2 2 . . . Procedures PTCA 135 (8) 1 (Ͻ1) 57 (6) 77 (15) Ͻ.01 CABG 22 (1.3) 0 (0) 11 (1.2) 11 (2.1) Ͻ.05

*The study patients were eligible for aspirin, thrombolytic, or ␤-blocker therapy, or had no indication for lidocaine. All values are number (percentage). Ellipses indicate not applicable; HMO, health maintenance organization; AMI, acute myocardial infarction; NS, not significant; ED, emergency department; EMS, emergency medical services; ICU, intensive care unit; CCU, coronary care unit; LOS, length of stay; PTCA, percutaneous transluminal coronary angioplasty; and CABG, coronary artery bypass graft. †Based on ␹2 test. ‡Other indicates indemnity/commercial insurance. §Based on Greenfield and colleagues’ Index of Coexistent Disease.17 ࿣Does not include patients without indications (ST elevation Ն1 mm or symptom onset Ͻ12 hours).

these institutions, of routinely administering prophylac- tients with the best prognosis for survival are more likely tic lidocaine for the possible development of arrhyth- to be referred to specialists.19 In another report,21 we found mias associated with thrombolytic infusion. that otherwise eligible patients with more severe comor- We observed important patient differences across bidity were less likely to receive aspirin and thrombo- the 3 categories of generalist and cardiologist involve- lytic therapy. Alternatively, this observed difference may ment in patient care, consistent with previous stud- be because younger, healthier patients are less likely to ies.19,20 Overall, generalists had older, more frail pa- have an established relationship with a generalist phy- tients. These results question the assumption that sician and are, therefore, assigned to a cardiologist. Fu- cardiologists care for sicker and more complex patients, ture studies should examine whether this pattern of at least in the setting of AMI. One hypothesis is that pa- selective referral exists for other medical conditions.

ARCH INTERN MED/ VOL 158, SEP 14, 1998 1781

Downloaded From: https://jamanetwork.com/ on 09/27/2021 60 Table 3. Use of Study Drugs by Level Generalist Alone of Generalist/Cardiologist Involvement* Consultative Care Cardiologist Attending Physician 50 OR (95% CI Using GEE)

Cardiologist Consultation Cardiologist vs 40 vs Generalist Generalist Drugs Attending Physician Attending Physician

Aspirin† 1.5 (1.0-2.2) 1.4 (0.8-2.6) 30 Thrombolytics‡ 1.9 (1.0-3.9) 2.0 (0.8-4.8) ␤-Blockers§ 2.3 (1.3-3.9) 2.1 (1.1-3.9) Lidocaine࿣ 1.3 (0.6-2.6) 1.7 (0.6-4.7) 20

*GEE indicates generalized estimating equations, as described in the “Results” section in the text. 10 †Controlling for age, sex, ST elevation 1 mm or more, and use of lipid-lowering agents before hospitalization (N = 1242). Patients Without Indication Who Received Lidocaine, % ‡Controlling for age, sex, ST elevation 1 mm or more, severe comorbidity, 0 acute myocardial infarction (AMI) location, use of lipid-lowering agents Other Hospitals Two Outlier Hospitals before hospitalization, and location of first contact with hospital (N = 578). §Controlling for age, sex, severe comorbidity, and AMI location (N = 307). Figure 2. Nonindicated use of lidocaine by specialty of the attending ࿣Controlling for age, sex, ST elevation 1 mm or more, severe comorbidity, physician. Comparison of 2 outlier hospitals with all others. AMI location, presence of heart failure, use of lipid-lowering agents before hospitalization, presentation to hospital within 6 hours of symptom onset, and location of first contact with hospital (N = 1136). The presence of a cardiologist may influence quality of care in several ways other than through direct con-

1.00 Generalist Alone sultation. For example, cardiologists may have been in- Consultative Care strumental in developing evidence-based protocols for 0.90 Cardiologist Attending Physician the management of AMI at individual hospitals. Also, a 0.80 generalist who had previously consulted with a cardiologist for a similar case would not necessarily consult 0.70 formally with the cardiologist over the same matter in 0.60 the next case. We have not included these consider- ations in our model. 0.50 An additional limitation is that our study findings 0.40 are derived from a retrospective chart audit of the medical record. While richer in detail than hospital dis-

Predicted Probability of Use 0.30 charge and claims data,20 it is possible that important fac- 0.20 tors relating to the use of these drugs may not have been

0.10 recorded in the charts. If this measurement error oc- curred at random, then we have underestimated the ac- 0.00 Thrombolytic Aspirin β-blocker Lidocaine tual difference in the use of study therapies. We do not Study Drug expect that there would be any differences in recording of important clinical contraindications by specialty, par- Figure 1. Predicted probability (SE) of use of thrombolytic agents, aspirin, ticularly because data on medical history and clinical con- and ␤-blockers among eligible patients, and lidocaine among ineligible patients, by level of generalist and cardiologist involvement. traindications were gathered from all sources in the chart and not solely from physicians’ notes. Finally, this study was conducted in 1 state and may differ from other states LIMITATIONS AND ALTERNATE EXPLANATIONS in ways that affect the generalizability of our findings.

As in any observational study, our results may be biased POLICY IMPLICATIONS by variables that are either difficult to measure or are un- measured. For example, in our study generalists cared The changes that are currently occurring in the organi- for patients with more severe comorbidity, which is in- zation and financing of health care in the United States dependently associated with reduced use of study medi- are influencing how specialists’ services are used in the cations and increased mortality risk.21 If we adjusted in- care of large patient populations. Instead of the open, pa- completely for comorbidity, our results would be biased tient- and physician-driven system of traditional fee-for- toward greater specialty-associated differences in use of service practice, many of the conditions under which gen- the study drugs. This confounding may be even more se- eralists now collaborate with specialist colleagues are rious in studies comparing postmyocardial infarction mor- largely determined at the organizational level, where fi- tality rates that did not collect information on severe co- nancial incentives, prior authorization requirements, and morbidity.20 Only a randomized controlled trial could other managed care policies are developed. These changes adequately address these limitations. However, random have the potential to affect the quality of patient care. How- allocation of patients to different levels of involvement ever, evidence to evaluate the added value of specialist of generalists and cardiologists would not be feasible. care within these various arrangements is not yet avail-

ARCH INTERN MED/ VOL 158, SEP 14, 1998 1782

Downloaded From: https://jamanetwork.com/ on 09/27/2021 able to help inform organizational policy decisions that 2. Lomas J, Anderson GM, Domnic-Pierre K, Vayda E, Enkin MW, Hannah WJ. Do impact on the access of patients to specialists. practice guidelines guide practice? the effect of a consensus statement on the practice of physicians. N Engl J Med. 1989;321:1306-1311. Our findings suggest that, when a generalist physi- 3. Greenfield S, Rogers W, Mangotich M, Carney MF, Tarlov AR. Outcomes of pa- cian is managing the care of a patient with AMI, the in- tients with hypertension and non–insulin-dependent diabetes mellitus treated by volvement of a cardiologist consultant offers an oppor- different systems and specialties: results from the Medical Outcomes Study. JAMA. tunity to enhance the quality of care. Although most of 1995;274:1436-1444. the recent health policy debate has focused solely on ques- 4. Beers MH, Finegold SF, Ouslander JG, Reuben DB, Morgenstern H, Beck JC. Char- 22 acteristics and quality of prescribing by doctors practicing in nursing homes. tions of patient access to specialty care, a narrow focus J Am Geriatr Soc. 1993;41:802-807. on an either/or dichotomy between specialist and gen- 5. Katon W, von Korff M, Lin E, et al. Collaborative management to achieve treat- eralist care ignores the important issue of coordination ment guidelines: impact on depression in primary care. JAMA. 1995;273:1026- of care between generalist and specialist physicians. We 1031. believe that our data should encourage further research 6. Hutchison SJ, Cobbe SM. Management of myocardial infarction in Scotland: have clinical trials changed practice? BMJ. 1987;294:1261. to delineate the optimum strategy for ensuring effective 7. Hlatky MA, Cotugno H, O’Connor C, Mark DB, Pryor DB, Califf RM. Adoption of coordination between generalists and cardiologists for the thrombolytic therapy in the management of acute myocardial infarction. Am J care of patients with AMI. Cardiol. 1988;61:510-514. 8. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Ran- domised trial of intravenous streptokinase, oral aspirin, both, or neither among Accepted for publication January 15, 1998. 17,187 cases of suspected acute myocardial infarction: ISIS-2. Lancet. 1988;2: From the Department of Ambulatory Care and Preven- 349-360. tion,HarvardMedicalSchoolandHarvardPilgrimHealthCare, 9. GISSI (Gruppo Italiano per lo Studio della Streptochinasi nell’Infarto Mio- Boston, Mass (Drs Willison, Soumerai, McLaughlin, and Pear- cardico). Effectiveness of intravenous thrombolytic treatment in acute myocar- son and Ms Gao); Centre for the Evaluation of Medicines, St dial infarction. Lancet. 1986;1:397-402. 10. ISIS-1 (First International Study of Infarct Survival) Collaborative Group. Ran- Joseph’sHospitalandMcMasterUniversity,Hamilton,Ontario domised trial of intravenous atenolol among 16,027 cases of suspected acute (Dr Willison); the Meyers Primary Care Institute, University myocardial infarction: ISIS-1. Lancet. 1986;2:57-66. of Massachusetts Medical Center and the Fallon Healthcare 11. MIAMI Trial Research Group. Metoprolol in acute myocardial infarction System, Worcester (Dr Gurwitz); the Department of Health (MIAMI): a randomised placebo-controlled trial. Eur Heart J. 1985;6:199-226. Care Policy, Harvard Medical School (Dr Guadagnoli); Car- 12. MacMahon S, Collins R, Peto R, Koster RW, Yusuf S. Effects of prophylactic lidocaine in suspected acute myocardial infarction: an overview of results from diovascular Division, Brigham and Women’s Hospital, Bos- the randomized, controlled trials. JAMA. 1988;260:1910-1916. ton (Dr Hauptman); and Healthcare Education and Research 13. McLaughlin TJ, Soumerai SB, Willison DJ, et al. Adherence to national guide- Foundation, St Paul, Minn (Ms McLaughlin). lines for drug treatment of suspected acute myocardial infarction: evidence for This study was supported by grant HSO 7357 from the undertreatment in women and the elderly. Arch Intern Med. 1996;156:799-805. Agency for Health Care Policy and Research, Rockville, Md, 14. Beaglehole R, Stewart AW, Butler M. Comparability of old and new World Health Organization criteria for definite myocardial infarction. Int J Epidemiol. 1987;16: and a grant from the Harvard Pilgrim Health Care Foun- 373-376. dation, Boston, Mass. Dr Gurwitz is supported by a Clini- 15. ACC/AHA Task Force. Guidelines for the early management of patients with acute cal Investigator Award K08 AG00510 from the National In- myocardial infarction: a report of the American College of Cardiology/American stitute on Aging, Bethesda, Md. At the time of writing the Heart Association Task Force on Assessment of Diagnostic and Therapeutic Car- diovascular Procedures (Subcommittee to Develop Guidelines for the Early Man- manuscript Dr Willison was supported by a Post-Doctoral agement of Patients with Acute Myocardial Infarction). J Am Coll Cardiol. 1990; Research Fellowship from the Medical Research Council of 16:249-292. Canada. 16. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York, NY: John Presented in abstract form at the Annual Meeting of Wiley & Sons Inc; 1989. the Association for Health Services Research, Atlanta, Ga, 17. Greenfield S, Apolone G, McNeil BJ, Cleary PD. The importance of co-existent disease in the occurrence of postoperative complications and one-year recov- June 10, 1996. ery in patients undergoing total hip replacement: comorbidity and outcomes af- Reprints: Stephen B. Soumerai, ScD, Department of ter hip replacement. Med Care. 1993;31:141-154. Ambulatory Care and Prevention, Harvard Medical School 18. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. and Harvard Pilgrim Health Care, 126 Brookline Ave, Suite Biometrika. 1986;73:13-22. 200, Boston, MA 02215 (e-mail: oumerai@warren 19. Horner RD, Matchar DB, Divine GW, Feussner JR. Relationship between physician specialty and the selection and outcome of ischemic stroke patients. Health .med.harvard.edu). Serv Res. 1995;30:275-287. 20. Jollis JG, DeLong ER, Peterson ED, et al. Outcome of acute myocardial infarc- REFERENCES tion according to the specialty of the admitting physician. N Engl J Med. 1996; 335:1880-1919. 21. McLaughlin TJ, Soumerai SB, Willison DJ, et al. The impact of comorbidity on 1. Ayanian JZ, Hauptman PJ, Guadagnoli E, Antman EM, Pashos CL, McNeil BJ. use of thrombolysis or aspirin in acute myocardial infarction patients eligible for Knowledge and practices of generalist and specialist physicians regarding drug treatment. J Gen Intern Med. 1997;12:1-6. therapy for acute myocardial infarction. N Engl J Med. 1994;331:1136-1142. 22. Kassirer JP. Access to specialty care. N Engl J Med. 1994;331:1151-1153.

ARCH INTERN MED/ VOL 158, SEP 14, 1998 1783

Downloaded From: https://jamanetwork.com/ on 09/27/2021