Association of consultation between generalists and cardiologists with quality and outcomes of heart failure care

Congestive Heart Failure

Association of consultation between generalists and cardiologists with quality and outcomes of heart failure care Ali Ahmed, MD, MPH, FACP, Richard M. Allman, MD, Catarina I. Kiefe, PhD, MD, Sharina D. Person, PhD, Terrence M. Shaneyfelt, MD, MPH, Richard V. Sims, MD, FACP, George Howard, DrPH, and James F. DeLong, MD, FACP Birmingham, Ala

Background The appropriate roles for generalists and cardiologists in the care of patients with heart failure (HF) are unknown. The objective of this retrospective cohort study was to determine whether consultation between generalists and cardiologists was associated with better quality and outcomes of HF care. Methods We studied left ventricular function evaluation (LVFE) and angiotensin-converting enzyme inhibitor (ACEI) use and 90-day readmission and 90-day mortality rates in patients with HF who were hospitalized. Patient care was categorized into cardiologist (solo), generalist (solo), or consultative cares. The processes and outcomes of care were compared by care category using logistic regression analyses fit with generalized linear mixed models to adjust for hospital-related clustering. Results Of the 1075 patients studied, 13% received cardiologist care, 55% received generalist care, and 32% received consultative care. More patients who received consultative care (75%) received LVFE than patients who received generalist care (36%) and cardiologist care (53%; P ⬍ .001). Fewer patients who received solo care (54% each) received ACEI compared with 71% of patients who received consultative care (P ⬍ .001). After multivariable adjustment, consultative care was associated with higher odds of LVFE than generalist care (adjusted odds ratio [OR], 6.06; 95% CI, 3.97-9.26) or cardiologist care (adjusted OR, 2.96; 95% CI, 1.70-5.13) care. Consultation was also associated with higher odds of ACEI use compared with generalist (adjusted OR, 2.42; 95% CI, 1.42-4.12) or cardiologist (adjusted OR, 2.32; 95% CI, 1.14-4.72) care. Compared with patients who received generalist care, patients who received consultative care had lower odds of 90-day readmission (adjusted OR, 0.54; 95% CI, 0.34-0.86). Conclusion Collaboration between generalists and cardiologists, rather than solo care by either, was associated with better HF processes and outcomes of care. (Am Heart J 2003;145:1086-93.) Heart failure (HF) is the number 1 hospital discharge diagnosis for persons aged ⱖ65 years.1 Most patients with HF are older adults, who also have multiple comorbidities.2-5 Whether patients with HF should be From the Divisions of Gerontology and Geriatric Medicine, General Internal Medicine, and Preventive Medicine, Department of Medicine, School of Medicine, Departments of Epidemiology and International Health, and Biostatistics, School of Public Health, Center for Aging, Center for Outcomes and Effectiveness Education and Research, and Prime Care Geriatric Heart Failure Clinic, University of Alabama at Birmingham, the Geriatric Heart Failure Clinic, Sections of Geriatrics and Internal Medicine, and Geriatric Research, Education, and Clinical Center, VA Medical Center, and the Alabama Quality Assurance Foundation, Birmingham, Ala. Supported by career development awards from the Southeast Center of Excellence in Geriatric Medicine and the American Federation for Aging Research to Dr Ahmed. A paper based on preliminary results of this study was presented by Dr Ahmed at the 7th World Congress on Heart Failure, Vancouver, Canada, in July 2000. Reprints not available from the authors. © 2003, Mosby, Inc. All rights reserved. 0002-8703/2003/$30.00 ⫹ 0 doi:10.1016/S0002-8703(02)00070-4

treated by cardiologists or generalists or by consultation between the 2 is still being debated.6 Although generalists provide better overall care, they lack the specialized training and experience of cardiologists.7-10 Cardiologists, on the other hand, provide better HF care.11-16 However, little is known about the quality of primary care provided by cardiologists. Studies have demonstrated that consultation between generalists and specialists is associated with superior quality of care.17-19 To our knowledge, no study has examined the quality of HF care provided by solo attending physicians (generalists or cardiologists) compared with the care provided through consultation. We hypothesized that the processes and outcomes of care of patients with HF who were cared for collaboratively by generalists and cardiologists would be superior to those of patients with HF who were cared for by generalists or cardiologists alone.

American Heart Journal Volume 145, Number 6

Methods Subjects We studied Medicare beneficiaries, 65 years and older, discharged in 1994, with a principal discharge diagnosis of HF from 11 Alabama hospitals. This study is a secondary analysis of the baseline dataset of a HF quality improvement project conducted by the Alabama Quality Assurance Foundation (AQAF), the peer review organization for Alabama. The details of the study methodology and subject selection criteria have been described elsewhere.20 In brief, with Medicare claims data for index hospitalizations, patients were identified using the International Classification of Diseases, 9th Revision, Clinical Modification codes 428 and 402.91, and data were abstracted using chart reviews on demographics and other clinical characteristics, including indicators of processes of care, by specially trained research nurses. The reliability and validity of the abstraction process was checked with a random re-abstraction of 5% of the charts by physician members of the project, which showed ⬎95% concordance rates on key variables

Processes of care Left ventricular function (LVF) evaluation and discharge use of angiotensin-converting enzyme (ACE) inhibitors were considered to be indicators for processes of care. Patients were considered to have LVF evaluations when they had an echocardiography, multiple-gated acquisition scan, or contrast left ventriculography during the index hospitalization. Patients who had past LVF evaluation were not considered eligible for LVF evaluation during the current hospitalization and were excluded from the denominator. Patients who died in the hospital (n ⫽ 83), patients with contraindications to ACE inhibitors (n ⫽ 179), and patients with preserved left ventricular (LV) systolic function (n ⫽ 194) were considered to be ineligible for ACE inhibitor therapy. The presence of a systolic blood pressure ⬍90 mm Hg, a serum potassium level ⱖ5.5 mEq/L, a serum creatinine level ⱖ2.5 mg/dL on admission, severe aortic stenosis, or intolerance to ACE inhibitors (history of cough, hypotension, hyperkalemia, deteriorating renal function, or angioneurotic edema in response to previous use of ACE inhibitors) were considered contraindications.

Outcomes of care We examined 90-day readmission and 90-day mortality rates as outcomes of care. Readmission data were ascertained by AQAF using the claims history of the study subjects for all HF admissions to any acute care hospital in Alabama. Data on patients admitted to out-of-state hospitals or patients whose care was not paid for by Medicare were not included. About 98% of the subjects in the original project were Alabama beneficiaries, which is higher than the typical 92% rate for other in-state beneficiaries. Mortality data were obtained from the Centers for Medicare and Medicaid Services membership lists (Denominator File), which is created every year for each Medicare beneficiary in the state.

Physician specialty Physician specialty was ascertained from chart. Admitting attending physicians were at first classified as cardiologist or

Ahmed et al 1087

generalists. Generalist physician included general internists, family physicians, general practitioners, or others. The definition of physician specialty was made on the basis of accredited training, board certification, or both. Family physicians who were not residency trained were classified as general practitioners. When the attending physician was not a cardiologist and obtained a cardiology consultation, we recorded a consultation order. When the physician specialty was not clear at the time of data abstraction, it was later determined and verified with administrative data both at AQAF and in the respective hospital.

Statistical analysis The SPSS21 software program was used for overall data analysis. The SAS procedure PROC MIXED22 through the use of a macro called GLIMMIX23 was used to create generalized linear mixed models (GLMM) to account for potential cluster differences among hospitals.24-26 Demographic data, admission characteristics, and historical features of the patients were compared across categories of care. Performances on the quality indicators were also compared across the care categories. We also examined potential associations between performances on the quality indicators and the demographic variables, sources of admission, historical and admission characteristics of the patients, physician characteristics, length of stay (square root transformed), and hospital. Statistical significance of the comparisons for the categorical variables was tested with both the Cochran-Mantel Haenszel ␹2 test27 and one-way analysis of variance (ANOVA) test fit through GLMM. We calculated the odds ratios (OR) and 95% CIs of LVF evaluation, ACE inhibitor use, and 90-day mortality (including in-hospital mortality) and 90-day readmission (excluding inhospital mortality) rates for patients cared for by consultations between generalists and cardiologists compared with patients receiving solo generalist or cardiologist care. To control for confounding, we created separate multiple logistic regression models for each quality indicator. In each model, we at first assessed the overall significance and then evaluated pair-wise comparisons. Age (as a continuous variable), sex (female), and race (African American vs others) were forced into each of the models (step 1). Other clinically important covariates were then entered into the models in a forward stepwise fashion. In the model for LVF evaluation, past history of HF, admission from nursing home, prior use of ACE inhibitors, contraindication to ACE inhibitors, third heart sound, ⱖ3 comorbidities, length of stay (square root transformed), and hospital were entered in a forward stepwise fashion (step 2). In the model for ACE inhibitors use, history of HF, pre-admission use of ACE inhibitors, presence of hypertension or diabetes mellitus, ⱖ3 comorbidities, length of stay, and admission symptoms of dyspnea, orthopnea and paroxysmal nocturnal dyspnea were similarly entered. In the model for 90-day mortality and readmission rates, we entered history of HF, LVF evaluation, contraindication to ACE inhibitors, discharge use of ACE inhibitors, presence of ⱖ3 comorbidities, length of stay, and hospital. In the last step, we entered admission symptoms of dyspnea on exertion, orthopnea, paroxysmal nocturnal dyspnea, fatigue, leg swelling, and angina. Because of the potential for bias in documentation of symptoms,

American Heart Journal June 2003

1088 Ahmed et al

Table I. Demographic and historical patient characteristics by care category

Age category (y) (%) 65-74 75-84 ⱖ85 Female (%) African-American (%) Admission from nursing homes (%) Medical history (%) Heart failure Past LVF evaluation Comorbidities (%) CAD Hypertension Diabetes Cardiomyopathy Arrhythmias COPD Home medications (%) Diuretics ACE inhibitors Digoxin Hydralazine Isosorbide

Total, n ⴝ 1075 (100%)

Solo care by generalist, n ⴝ 588 (55%)

Solo care by cardiologist, n ⴝ 145 (13%)

Care by consultation, n ⴝ 342 (32%)

339 (31) 461 (43) 275 (26) 640 (60) 191 (18) 95 (9)

156 (26) 252 (43) 180 (31) 375 (64) 127 (22) 70 (12)

60 (41) 59 (41) 26 (18) 74 (51) 19 (13) 3 (2)

123 (36) 150 (44) 69 (20) 191 (56) 45 (13) 22 (6)

⬍.007

791 (74) 198 (18)

424 (72) 59 (10)

115 (79) 43 (30)

252 (74) 90 (26)

.29 ⬍.001

269 (25) 178 (17) 271 (25) 164 (15) 370 (34) 342 (32)

133 (23) 119 (20) 163 (28) 60 (10) 169 (29) 178 (30)

45 (31) 18 (12) 20 (14) 33 (23) 69 (48) 42 (29)

91 (27) 41 (12) 88 (26) 71 (21) 132 (39) 122 (36)

.06 .45 .014 .009 ⬍.001 .72

296 (27) 406 (38) 578 (54) 56 (5) 146 (14)

147 (25) 217 (37) 302 (51) 31 (5) 69 (12)

35 (24) 56 (39) 91 (63) 14 (10) 33 (23)

114 (33) 133 (39) 185 (54) 11 (3) 44 (13)

.62 .5 .011 .06 .07

P

.01 ⬍.001 .006

LVF, Left ventricular function, CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease.

these variables were entered in a separate step. These independent variables were selected because of their clinical significance or statistical significance in bivariate analyses. Because under-documentation of past LVF evaluation might result in underestimation of LVF evaluation during the index hospitalization, and because patients with a past history of HF were more likely to receive LVF evaluation in the past, we repeated the aforementioned analyses in a subgroup of patients who had no past history of HF. We similarly performed subgroup analyses for the use of ACE inhibitors among patients not using ACE inhibitors on admission.

Results Patient characteristics Subjects (N ⫽ 1075) had a mean (⫾ SD) age of 79 (⫾ 7.5) years, 60% were women, and 18% were African American. Solo generalists cared for 588 patients (55%), solo cardiologists cared for 145 patients (13%), and 342 patients (32%) received consultative care. A total of 883 patients were eligible for an LVF evaluation, and 433 patients (49%) received an LVF evaluation during the index hospitalization. A total of 619 patients were eligible for ACE inhibitor therapy, and 363 patients (59%) were discharged receiving ACE inhibitor therapy. Tables I and II describe the baseline demographic and other clinical characteristics of the patients by the 3 care categories.

LVF evaluation Patients who received consultative care were more likely to receive LVF evaluations (75%) than patients who received solo care from either a cardiologist (53%) or a generalist (36%; P ⬍ .001; Figure 1). Table III shows the crude and adjusted ORs of LVF evaluation for patients who received consultative care compared with patients who received generalist and cardiologist care. After multivariable adjustments, compared with generalist care, cardiology care (either solo or via consultation) was associated with higher odds of LVF evaluations. However, patients who received consultation had higher odds of LVF evaluation than patients who received solo cardiology care (adjusted OR, 2.96; 95% CI, 1.70-5.13; Table III). Among patients with new onset HF (N ⫽ 284), consultation was associated with higher odds of LVF evaluation than either generalist (adjusted OR, 6.89; 95% CI, 2.96-16.05) or cardiologist (adjusted OR, 4.40; 95% CI, 1.44-13.72) care.

ACE inhibitor use Figure 1 illustrates that, compared with either solo care category, more patients who received consultative care were discharged receiving ACE inhibitors (P ⬍ .001). Table IV shows that patients who received consultative care had higher odds of receiving ACE

American Heart Journal Volume 145, Number 6

Ahmed et al 1089

Table II. Admission characteristics by care category Total, n ⴝ 1075 (100%) Symptoms (%) Dyspnea Dyspnea on exertion Orthopnea PND Fatigue Angina Leg swelling Physical signs Pulse (per minute) Systolic BP (mm Hg) Diastolic BP (mm Hg) Third heart sound (%) JVD (%) Pulmonary ra ˆ les (%) PMI displaced (%) Laboratory values BUN (mg/dL) S. creatinine (mg/dL) S. sodium (mEq/L) S. potassium (mEq/L) Chest radiograph (%) Cardiomegaly Pulmonary edema Electrocardiogram (%) Normal sinus rhythm Atrial fibrillation LVH

Solo care by generalist, n ⴝ 588 (55%)

Solo care by cardiologist, n ⴝ 145 (13%)

Care by consultation, n ⴝ 342 (32%)

P

961 (89) 275 (26) 351 (33) 229 (21) 55 (5) 30 (3) 615 (57)

516 (88) 101 (17) 140 (24) 95 (16) 17 (3) 13 (2) 329 (56)

125 (86) 43 (30) 57 (39) 42 (29) 21 (15) 4 (3) 81 (56)

320 (94) 131 (38) 154 (45) 92 (27) 17 (5) 13 (4) 205 (60)

.04 ⬍.001 ⬍.001 ⬍.003 ⬍.001 .12 .051

92 (⫾ 22) 147 (⫾ 32) 80 (⫾ 20) 200 (19) 447 (42) 733 (68) 97 (9)

91 (⫾ 22) 149 (⫾ 34) 80 (⫾ 21) 106 (18) 228 (39) 391 (66) 46 (8)

91 (⫾ 21) 140 (⫾ 29) 78 (⫾ 15) 33 (23) 66 (45) 100 (69) 20 (14)

93 (⫾ 23) 145 (⫾ 29) 80 (⫾ 20) 61 (18) 153 (45) 242 (71) 31 (9)

.59 .03 .54 .68 .38 .45 .19

29 (⫾ 21) 1.6 (⫾ 1.3) 139 (⫾ 5.9) 4.3 (⫾ 0.7)

30 (⫾ 22) 1.7 (⫾ 1.5) 139 (⫾ 6.3) 4.3 (⫾ 0.7)

34 (⫾ 28) 1.8 (⫾ 1.2) 137 (⫾ 5.3) 4.4 (⫾ 0.7)

27 (⫾ 16) 1.4 (⫾ 0.9) 138 (⫾ 5.2) 4.3 (⫾ 0.6)

⬍.001 .04 .41 .32

699 (65) 332 (31)

357 (61) 166 (28)

97 (67) 45 (31)

245 (72) 121 (35)

.02 .11

580 (54) 248 (23) 171 (16)

325 (55) 128 (22) 99 (17)

69 (48) 33 (23) 15 (10)

186 (54) 87 (25) 57 (17)

.2 .64 .12

PND, Paroxysmal nocturnal dyspnea; BP, blood pressure; JVD, jugular venous distension; PMI, point of maximum impulse; BUN, blood urea nitrogen; S, serum; LVH, left ventricular hypertrophy.

inhibitors than patients who were cared for by either cardiologist (adjusted OR, 2.32; 95% CI, 1.14-4.72) or generalist (adjusted OR, 2.42; 95% CI, 1.42-4.12) alone. Consultation was also associated with higher odds of initiation of therapy with an ACE inhibitor (adjusted OR, 2.28; 95% CI, 0.99-5.26) than solo cardiologist care (adjusted OR, 2.26; 95% CI, 1.24-4.14) and solo generalist care. The superiority of consultative care in prescribing ACE inhibitor over either solo care was also noted among patients with incident HF.

Discussion

Mortality and readmission

Quality of care: consultation versus solo care

The overall 90-day mortality rate was 22%, and the 90-day readmission rate among patients who were discharged alive (n ⫽ 988) was 15%. The care category was not associated with mortality (Table V). However, compared with patients cared for by generalists, patients cared for by consultation had lower odds of having a HF-related readmission within 90 days of discharge (adjusted OR, 0.54; 95% CI, 0.34-0.86).

We do not know why a cardiologist acting in collaboration with a generalist would treat HF differently than when acting alone. However, we speculate that there are several possible explanations. First, patients who received consultations had more symptoms and were more likely to be taking diuretics before admission, which suggests a greater severity of symptoms. It is possible that patients were chosen for cardiology

The results of our study demonstrate that collaboration between generalists and cardiologists was associated with improved processes and outcomes of care for patients with HF who were hospitalized. These findings are important because more than half of all patients with HF in the United States receive solo care from generalists, and 10% to 30% of all patients with HF in the United States receive solo care from cardiologists,11,28,29 proportions that are consistent with those we observed.

1090 Ahmed et al

Figure 1

American Heart Journal June 2003

cardial ischemia. It was possible that some of these patients were referred for cardiology consultation for cardiovascular procedures.

Quality of care: generalists versus cardiologists Other investigators have also demonstrated that patients cared for by cardiologists were more likely to receive LVF evaluation,11,29 but not ACE inhibitors,15,16,19,31 than patients cared for by generalists. One possible explanation for this disparity might be under-documented past LVF evaluation by generalist physicians. This was apparent because the past LVF evaluation documentation rate for patients admitted by generalists was higher when they had received cardiology consultations (Table I). Under-documentation of past LVF evaluation by generalist physicians might have resulted in misclassification of their patients with past LVF evaluation as eligible for LVF evaluation (denominator), which in turn led to an underestimation of current LVF evaluation (numerator). Proportion of patients receiving LVF evaluation (LVFE) during the current hospitalization and ACE inhibitors (ACEI) at the time of discharge by physician care category.

consultation because they had more symptoms, which might also have compelled the consulting cardiologists to act. However, when we adjusted for number of symptoms, the associations remained statistically significant, suggesting that the positive confounding by symptoms was minimal. Second, patients who were referred for consultation were more likely to be white, younger, and male and were less likely to reside in long-term care facilities and, therefore, were more likely to receive LVF evaluation and ACE inhibitors. However, patients admitted by cardiologists also had comparable proportions of these characteristics. Therefore, superior care of the patients who received consultative care (vs that of patients who received solo cardiologist care) cannot be explained by these variables. Third, compared with admitting cardiologists, consulting cardiologists had a smaller window of opportunity and might have acted more aggressively. Other factors known to influence generalist physicians’ referral decisions, such as requests made by patients and family members, treatment of an underlying condition (viz coronary artery disease), performance of a specific procedure (viz multiple gated acquistion scan), and the availability of qualified experts30 might have also influenced referrals in our study. In 1 study, non-HF cardiac conditions, diagnostic and therapeutic dilemmas, or severity of symptoms and signs were motivating factors for consultation.29 Most of the patients in our study had a history of HF, and many of them might have had acute decompensation caused by myo-

Outcomes of care: consultation versus solo care Receipt of cardiologist care, either by admission or via consultation, apparently was not associated with improved survival rates, a finding that was also observed by other investigators.11,29,31 However, like us, other investigators also noted that consultative care was associated with lower HF-related readmissions11 and emergency department visits.29

Heart failure management: best model of care There is an ongoing debate about the most appropriate roles for generalists and cardiologists in the care of patients with HF.6 Some experts in this area have recommended a leading role for HF specialists over generalist physicians and community-based cardiologists,32 whereas others recommend a model with an initial evaluation by cardiologists with subsequent follow-up by generalists.33 Although the multidisciplinary HF program approach has been shown to improve quality and outcomes of care,34-36 such programs do not delineate specific roles of generalists and cardiologists in the care of patients with HF. However, other programs have emphasized the need for prospective longitudinal studies that would also involve outpatient care.37 The recently published American College of Cardiology/ American Heart Association HF guidelines also recommended a collaborative model in which generalists and cardiologists would work together to optimize the care of patients with HF.6 Collaborative care is likely to improve patient/family satisfaction and quality of life. Symptoms of dyspnea, especially in elderly patients with HF could well be caused by concomitant depression, and the early diagnosis and treatment of this will improve symptoms and obviate the need for unneces-

American Heart Journal Volume 145, Number 6

Ahmed et al 1091

Table III. Crude and adjusted odds ratio (95% CI) for LVF evaluation during the current hospitalization for pair-wise comparisons of 3 categories of care: Generalists alone, cardiologist alone, and consultative care

Unadjusted Adjusted for age-race-sex Adjusted for additional baseline characteristics and hospital* Adjusted for additional baseline characteristics, hospital, and admission symptoms†

Cardiologist alone vs generalist alone

Consultation vs generalist alone

Consultation vs cardiologist alone

Overall P

2.69 (1.70-4.28) 2.57 (1.60-4.11) 2.27 (1.38-3.74)

7.74 (5.30-7.74) 7.63 (5.21-11.19) 7.45 (4.95-11.20)

2.87 (1.75-4.73) 2.97 (1.80-4.92) 3.28 (1.94-5.55))

⬍.001 ⬍.001 ⬍.001

2.05 (1.21-3.46)

6.06 (3.97-9.26)

2.96 (1.70-5.13)

⬍.001

*History of heart failure, admission from nursing home, ACE inhibitor use before admission, contraindication to ACE inhibitors, presence of third hear sound, ⱖ3 comorbidities, and length of stay (square root transformed). †Dyspnea on exertion, orthopnea, paroxysmal nocturnal dyspnea, fatigue, and leg swelling.

Table IV. Crude and adjusted odds ratio (95% CI) for ACE inhibitors use for pair-wise comparisons of 3 categories of care: Generalists alone, cardiologist alone, and consultative care

Unadjusted Adjusted for age-race-sex Adjusted for additional baseline characteristics and hospital* Adjusted for additional baseline characteristics, hospital, and admission symptoms†

Cardiologist alone vs generalist alone

Consultation vs generalist alone

Consultation vs cardiologist alone

Overall P

1.16 (0.67-2.01) 1.21 (0.69-2.11) 1.11 (0.57-2.17)

2.57 (1.69-3.91) 2.69 (1.75-4.14) 2.77 (1.65-4.63)

2.21 (1.23-3.97) 2.23 (1.23-4.03) 2.48 (1.22-5.02)

⬍.001 ⬍.001 ⬍.001

1.04 (0.53-2.04)

2.42 (1.42-4.12)

2.32 (1.14-4.72)

.003

*History of heart failure, preadmission ACE inhibitor use, presence of hypertension or diabetes, ⱖ3 comorbidities, and length of stay (square root transformed). †Dyspnea on exertion, orthopnea, and paroxysmal nocturnal dyspnea.

Table V. Crude and adjusted odds ratio (95% CI) for readmission and mortality for pair-wise comparisons of 3 categories of care: Generalists alone, cardiologist alone, and consultative care

90-Day readmission Unadjusted Adjusted* 90-Day mortality Unadjusted Adjusted*

Cardiologist alone vs generalist alone

Consultation vs generalist alone

Consultation vs cardiologist alone

Overall P

1.17 (0.71-1.93) 0.79 (0.46-1.35)

0.70 (0.46-1.07) 0.54 (0.34-0.86)

0.60 (0.34-1.04) 0.69 (0.38-1.24)

.14 .06

0.87 (0.55-1.37) 0.70 (0.41-1.19)

0.94 (0.67-1.31) 1.02 (0.68-1.53)

1.08 (0.66-1.75) 1.45 (0.84-2.50)

.81 .36

*Adjusted for age, race, sex, history of heart failure, left ventricular function evaluation, admission from nursing home, contraindication to ACE inhibitors, discharge on ACE inhibitors, ⱖ3 comorbidities, length of stay (square root transformed), and hospital.

sary titration of HF medications. Similarly, a patient with HF who has severe osteoarthritis of hip will not benefit from the control of HF symptoms if the arthritis pain is not well controlled. Even with the large number of specialists that exist in this nation, there are not enough cardiologists to provide longitudinal care to every patient with HF.

Moreover, longitudinal primary cardiology care may be unsuitable for patients with HF, most of whom are older adults. Unlike younger patients, who more often receive longitudinal care from a single specialist (for example, a young woman with systemic lupus erythromatosus being cared for by a rheumatologist, or a young man with cardiomyopathy being cared for by a

American Heart Journal June 2003

1092 Ahmed et al

cardiologist), patients with HF who also suffer from multiple comorbidities do not receive optimum care from such a model of care. In a study of Medicare beneficiaries in Washington state, 26% of the “majority-ofcare” cardiologists’ patients received influenza vaccinations (vs 74% of generalists’ patients).38 The same study also found that most cardiologists’ patients received their vaccination from visits to a generalist’s office.

Limitations Because of the retrospective chart review nature of our study, we were not able to adjust for functional status, quality of life, patient, family, and physician preferences, and socioeconomic status. Although we were able to adjust for many obvious confounders in our multivariable models, residual confounding might be operative. It is possible that patients who were seen by consultation, in some unmeasured fashion, were considered to be more appropriate candidates for LVF evaluation or ACE inhibitor use. The generalizability of our findings is somewhat limited by the older age of the study subjects and the geographic limitation of the study to Alabama. Another limitation of the study is the age of the dataset. The patients were discharged from the hospital in 1994, the same year the first HF guideline was published.39 Although this partly explains the lower overall performance observed in our study, it is unlikely to explain the variations related to the care categories. A recent study has indicated similar overall rates of LVF evaluation and ACE inhibitors use in 1998 to1999 in Alabama (65% and 62%, respectively, for LVF evaluation and ACE inhibitor use).40 Articles on the benefits of ACE inhibitors in HF were published in major medical journals in the late 1980s,41-43 and physician practice patterns started to change by the early 1990s.44 It is possible that some of the patients who truly received consultative care were misclassified as solo-generalists’ or solo-cardiologists’ patients. For example, patients admitted by generalists who also received outpatient cardiology consultations might have been misclassified as receiving care from solo-generalists. Similarly, patients admitted by solo-cardiologists who also received outpatient care from generalists might have been misclassified as receiving care from solo-cardiologists. However, because only patients who received consultative care could be misclassified as patients receiving solo-care, not the other way around, we believe that these misclassifications would only have overestimated the performance of the patients who received solo-care.

Conclusions Our study demonstrates an association of collaboration between generalists and cardiologists with im-

proved processes of HF care. In addition, collaboration was associated with reduced HF readmission rates, without a significant difference in mortality rates. The results of this preliminary observational study, however, should be interpreted cautiously when making policy and practice recommendations. Future prospective longitudinal studies involving inpatient and outpatient care should explore the potential causal relationships suggested by these associations and should also determine the impact of collaborative care on the functional status and quality of life of older patients with HF.

Center for Medicare and Medicaid Services disclaimer The analyses on which this publication is based were performed under Contract Number 500-96-P60, titled “Utilization and Quality Control Peer Review Organization for the State of Alabama,” sponsored by the Center for Medicare and Medicaid Services (CMS, formerly HCFA), Department of Health and Human Services. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Service, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. The authors assume full responsibility for the accuracy and completeness of the ideas presented. This article is a direct result of the Health Care Quality Improvement Program initiated by the CMS, which has encouraged identification of quality improvement projects derived from analysis of patterns of care, and therefore required no special funding on the part of this contractor. Ideas and contributions to the author concerning experience in engaging with issues presented are welcomed. We thank Crystal Walls for her secretarial assistance and Robert Farmer of Alabama Quality Assurance Foundation and Eric Bodner of University of Alabama at Birmingham Center for Aging for their contribution in data management.

References 1. National hospital discharge survey: annual summary, 1995. Vitals Statistics Series 13; No. 133. DHHS Publication No. 98-1794. Hyattsville, Md: National Center for Health Statistics; 1998. 2. Congestive heart failure in the United States: a new epidemic. Data facts sheets. Bethesda, Md: National Heart, Lung, and Blood Institute, National Institutes of Health; 1996. 3. Havranek EP, Ware MG, Lowes BD. Prevalence of depression in congestive heart failure. Am J Cardiol 1999;84:348 –50. 4. Rich MW. Heart failure. Cardiol Clin 1999;17:123–35. 5. Rich MW. Epidemiology, pathophysiology, and etiology of congestive heart failure in older adults. J Am Geriatr Soc 1997;45: 968 –74.

American Heart Journal Volume 145, Number 6

6. Hunt SA, Baker DW, Chin MH, et al. ACC/AHA guidelines for the evaluation and management of chronic heart failure in the adult: executive summary. A report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines (Committee to revise the 1995 Guidelines for the Evaluation and Management of Heart Failure). J Am Coll Cardiol 2001;38:2101–13. 7. Ramsey PG, Carline JD, Inui TS, et al. Changes over time in the knowledge base of practicing internists. JAMA 1991;266:1103–7. 8. Carline JD, Inui TS, Larson EB, et al. The knowledge base of certified internists: relationships to training, practice type, and other physician characteristics. Arch Intern Med 1989;149:2311–3. 9. Petersdorf RG, Goitein L. The future of internal medicine. Ann Intern Med 1993;119:1130 –7. 10. Greenfield S, Nelson EC, Zubkoff M, et al. Variations in resource utilization among medical specialties and systems of care: results from the medical outcomes study. JAMA 1992;267:1624 –30. 11. Philbin EF, Weil HF, Erb TA, et al. Cardiology or primary care for heart failure in the community setting: process of care and clinical outcomes. Chest 1999;116:346 –54. 12. Davie AP, McMurray JJ. ACE inhibitors and heart failure in hospital: any difference between cardiologists and general physicians? Postgrad Med J 1999;75:219 –22. 13. Bello D, Shah NB, Edep ME, et al. Self-reported differences between cardiologists and heart failure specialists in the management of chronic heart failure. Am Heart J 1999;138:100 –7. 14. Baker DW, Hayes RP, Massie BM, et al. Variations in family physicians’ and cardiologists’ care for patients with heart failure. Am Heart J 1999;138:826 –34. 15. Reis SE, Holubkov R, Edmundowicz D, et al. Treatment of patients admitted to the hospital with congestive heart failure: specialtyrelated disparities in practice patterns and outcomes. J Am Coll Cardiol 1997;30:733– 8. 16. Chin MH, Wang JC, Zhang JX, et al. Utilization and dosing of angiotensin-converting enzyme inhibitors for heart failure: effect of physician specialty and patient characteristics. J Gen Intern Med 1997;12:563–6. 17. Vollmer WM, O’Hollaren M, Ettinger KM, et al. Specialty differences in the management of asthma: a cross-sectional assessment of allergists’ patients and generalists’ patients in a large HMO. Arch Intern Med 1997;157:1201– 8. 18. Willison DJ, Soumerai SB, McLaughlin TJ, et al. Consultation between cardiologists and generalists in the management of acute myocardial infarction: implications for quality of care. Arch Intern Med 1998;158:1778 – 83. 19. Chin MH, Wang JC, Zhang JX, et al. Differences among geriatricians, general internists, and cardiologists in the care of patients with heart failure: a cautionary tale of quality assessment. J Am Geriatr Soc 1998;46:1349 –54. 20. DeLong J, Allman R, Sherrill R, et al. A congestive heart failure project with measured improvement in care. Eval Health Profession 1998;21:472– 86. 21. SPSS for Windows, Rel. 10, 2000. 22. SAS/STAT software, version 8 of the SAS System for Unix. Copyright 2000. 23. Littell R, Milliken G, Stroup W, et al. SAS system for mixed models. Cary, NC: SAS Institute; 1996. 24. Hosmer D, Lemeshow S. Applied logistic regression, 2nd ed. New York: John Wile & Sons Inc; 2000. 25. Breslow N, Clayton D. Approximate inference in generalized linear models. J Am Stat Assoc 1993;88:9 –25. 26. McCullagh P, Nelder J. Generalized linear models. London: Chapman and Hall; 1989.

Ahmed et al 1093

27. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22: 719 – 48. 28. Chin MH, Friedmann PD, Cassel CK, et al. Differences in generalist and specialist physicians’ knowledge and use of angiotensinconverting enzyme inhibitors for congestive heart failure. J Gen Intern Med 1997;12:523–30. 29. Diller PM, Smucker DR, David B. Comanagement of patients with congestive heart failure by family physicians and cardiologists: frequency, timing, and patient characteristics. J Fam Pract 1999; 48:188 –95. 30. Nutting PA, Franks P, Clancy CM. Referral and consultation in primary care: do we understand what we’re doing? (editorial). J Fam Pract 1992;35:21–3. 31. Auerbach AD, Hamel MB, Davis RB, et al. Resource use and survival of patients hospitalized with congestive heart failure: differences in care by specialty of the attending physician. SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments. Ann Intern Med 2000;132:191–200. 32. Smith JJ, Konstam MA. Heart failure: a case for subspecialized care management (editorial). Am Heart J 1999;138:14 – 6. 33. Deedwania PC. Underutilization of evidence-based therapy in heart failure: an opportunity to deal a winning hand with ace up your sleeve (editorial). Arch Intern Med 1997;157:2409 –12. 34. Rich MW, Beckham V, Wittenberg C, et al. A multidisciplinary intervention to prevent the readmission of elderly patients with congestive heart failure. N Engl J Med 1995;333:1190 –5. 35. Gattis WA, Hasselblad V, Whellan DJ, et al. Reduction in heart failure events by the addition of a clinical pharmacist to the heart failure management team: results of the Pharmacist in Heart Failure Assessment Recommendation and Monitoring (PHARM) Study. Arch Intern Med 1999;159:1939 – 45. 36. Stewart S, Marley JE, Horowitz JD. Effects of a multidisciplinary, home-based intervention on unplanned readmissions and survival among patients with chronic congestive heart failure: a randomised controlled study. Lancet 1999;354:1077– 83. 37. Morris DL, Petruccelli DF. Congestive heart failure: who should provide care? Ann Intern Med 2000;132:238 –9. 38. Rosenblatt RA, Hart LG, Baldwin LM, et al. The generalist role of specialty physicians: is there a hidden system of primary care? JAMA 1998;279:1364 –70. 39. Konstam M, Dracup K, Bottoroff M, et al. Heart failure: evaluation and care of patients with left ventricular systolic dysfunction: clinical practice guideline. Agency for Health Care Policy and Research (AHCPR) 1994;DOHSS Publication 94-0612. 40. Jencks SF, Cuerdon T, Burwen DR, et al. Quality of medical care delivered to Medicare beneficiaries: a profile at state and national levels. JAMA 2000;284:1670 – 6. 41. Packer M, Lee WH, Yushak M, et al. Comparison of captopril and enalapril in patients with severe chronic heart failure. N Engl J Med 1986;315:847–53. 42. Anonymous. Angiotensin-converting-enzyme inhibitors in treatment of heart failure (editorial). Lancet 1985;2:811-2. 43. The Consensus Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure: results of the Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS). N Engl J Med 1987;316:1429 –35. 44. McDermott MM, Feinglass J, Lee P, et al. Heart failure between 1986 and 1994: temporal trends in drug-prescribing practices, hospital readmissions, and survival at an academic medical center. Am Heart J 1997;134:901–9.