Biomedical subjects

E F Cook

Publications and source records attributed to E F Cook.

At least 37 records · Page 2Linked to original sources

PubMed1989

Performance of tests of significance based on stratification by a multivariate confounder score or by a propensity score.

Stratification by a multivariate confounder score or by a propensity score has been proposed for the multi-confounder situations that are commonly encountered in epidemiologic evaluations of the effects of a treatment or a triage decision. However, the use of these scores in clinical research has been limited, perhaps in part because of the concern that the stated level of statistical significance may be exaggerated when there is a high degree of correlation between the exposure and the set of confounders. We present a specific example and computer simulations to suggest that exaggeration of statistical significance occurs only under unusual circumstances when the correlation between the exposure and the confounders is extreme. Our simulations also suggest that an analysis based on stratification by a propensity score is less affected by high correlation between the exposure and the confounders than is an analysis based on a multivariate confounder score.

Computer Simulation↗

PubMed1989

Allopurinol use and the risk of cataract formation.

Several reports have suggested an association between chronic allopurinol ingestion and cortical and subcapsular cataract formation. To examine this possibility we identified 51 allopurinol users and compared their lenses with those of 76 patients who did not use allopurinol. The existence of lens opacities and the level of visual acuity were assessed by review of medical records or by prospective ophthalmic examinations; in both phases of the study the examiners were blinded as to the patient's use or non-use of allopurinol. Three different outcomes were considered: formation of any cataract, formation of a posterior subcapsular cataract, and formation of a cataract contributing to a corrected visual acuity of 20/30 or worse. The risk ratio for the formation of any cataract was 1.3 (95% confidence interval: 0.8, 2.0), the risk ratio for the formation of a posterior subcapsular cataract was 0.9 (0.3, 2.0), and the risk ratio for the formation of a cataract contributing to a loss of visual acuity was 1.3 (0.6, 2.9). None of these risk ratios was changed appreciably after controlling for age, sex, hypertension, or diabetes. Thus, after a mean of 6.9 years of allopurinol use, we found no evidence to confirm that allopurinol users were at higher risk of acquiring cataracts.

Allopurinol↗

PubMed1989

Indirect costs for medical education. Is there a July phenomenon?

Medicare currently pays for "indirect costs" of medical education to support the higher costs of care in teaching hospitals. To investigate whether indirect costs are higher earlier in the training year when house officers might be less efficient--the "July phenomenon"--we compared utilization by 1251 patients hospitalized during July and August with 1338 patients hospitalized during April and May from 1982 through 1984 at Brigham and Women's Hospital, Boston, Mass. These included all patients in the 10 most prevalent medical and surgical diagnosis related groups. Using analysis of covariance to correct for age, sex, diagnosis related group, urgency of admission, temporal change, and mortality, we found no differences in length of stay, total charges, or categories of ancillary charges. These results suggest that there is no substantial increase in the cost of care early in the training year; there was no evidence of a "July phenomenon."

Ancillary Services, Hospital↗

PubMed1988

Costs and effectiveness of routine therapy with long-term beta-adrenergic antagonists after acute myocardial infarction.

We analyzed the costs and effectiveness of routine therapy with beta-adrenergic antagonists in patients who survived an acute myocardial infarction. On the basis of data pooled from the literature, this form of therapy resulted in a 25 percent relative reduction annually in the mortality rate for years 1 to 3 and a 7 percent relative reduction for years 4 to 6 after a myocardial infarction. The estimated cost of six years of routine beta-adrenergic-antagonist therapy to save an additional year of life was $23,400 in low-risk patients, $5,900 in medium-risk patients, and $3,600 in high-risk patients, assuming that the entire benefit of earlier treatment is lost immediately after six years. Under a more likely assumption--that the benefit of six years of treatment wears off gradually over the subsequent nine years--the estimated cost of therapy per year of life saved would be $13,000 in low-risk patients, $3,600 in medium-risk patients, and $2,400 in high-risk patients. As compared with coronary-artery bypass grafting and the medical treatment of hypertension, routine beta-adrenergic-antagonist therapy has a relatively favorable cost-effectiveness ratio.

Adrenergic beta-Agonists↗

PubMed1988

A computer protocol to predict myocardial infarction in emergency department patients with chest pain.

To achieve more appropriate triage to the coronary care unit of patients presenting with acute chest pain, we used clinical data on 1379 patients at two hospitals to construct a simple computer protocol to predict the presence of myocardial infarction. When we tested this protocol prospectively in 4770 patients at two university hospitals and four community hospitals, the computer-derived protocol had a significantly higher specificity (74 vs. 71 percent) in predicting the absence of infarction than physicians deciding whether to admit patients to the coronary care unit, and it had a similar sensitivity in detecting the presence of infarction (88.0 vs. 87.8 percent). Decisions based solely on the computer protocol would have reduced the admission of patients without infarction to the coronary care unit by 11.5 percent without adversely affecting the admission of patients in whom emergent complications developed that required intensive care. Although this protocol should not be used to override careful clinical judgment in individual cases, the computer protocol for the most part yields accurate estimates of the probability of myocardial infarction. Decisions about admission to the coronary care unit based on the protocol would have been as effective as those actually made by the unaided physicians who cared for the patients, and less costly. Whether physicians who are aided by the protocol perform better than unaided physicians cannot be determined without further study.

Angina Pectoris↗

PubMed1988

Body mass index and mortality among nonsmoking older persons. The Framingham Heart Study.

The relationship of weight at age 65 years and subsequent mortality was examined in a population of 1723 nonsmokers who were followed up from one to 23 years (mean, 9.5 years) during the Framingham Heart Study. In sex-specific proportional hazards analyses, risks of mortality were increased for men and women at the high and low extremes of body mass index, even when accounting for potential effects of excess weight on serum cholesterol level, blood glucose level, and systolic blood pressure. For those at the lower extreme of body mass index, the relative risk of death was almost twice as high in the years immediately after age 65 years as in later follow-up, suggesting that the increased early death rate was due to disease that was already present. At the upper extreme, risk of death was twofold over the entire follow-up period for persons with body mass indexes at or above the 70th percentile at both 55 and 65 years of age. We conclude that, even when accounting for cardiovascular risk factors, being overweight is a serious health problem for older people, especially for those with long-standing weight problems.

Aged↗

PubMed1988

Electrocardiographic diagnosis of myocardial infarction in the presence of complete left bundle branch block.

Whether myocardial infarction (MI) can be diagnosed electrocardiographically in the presence of left bundle branch block (LBBB) is controversial. Our study sought to test the reliability of ECG criteria for diagnosing MI in patients with LBBB. Among 985 patients presenting within 18 hours of the onset of ischemic chest pain, 35 had complete LBBB. Acute MI was identified by serial MB-creatine kinase (CK) elevations and prior MI was determined by previously documented ECG and/or enzyme changes. Among those with LBBB, 24 patients had acute and/or prior MI, while 11 had neither. Eleven ECG criteria previously proposed for detecting MI in the presence of LBBB were evaluated. In patients presenting with ischemic chest pain and complete LBBB, presence of any one of the following ECG criteria was highly specific (90% to 100%) and predictive (85% to 100%) for acute or prior MI: Q waves in at least two of leads I, aVL, V5, or V6; R wave regression from V1 to V4; notching of the upstroke of the S wave in at least two of leads V3, V4, or V5, and primary ST-T wave changes in two or more adjacent leads.

Bundle-Branch Block↗

PubMed1988

Asymmetric stratification. An outline for an efficient method for controlling confounding in cohort studies.

Confounding is usually controlled by either cross-stratification or multivariate modeling. The first approach is simple and intuitive, but it is not practical for controlling many factors. The second approach, although less intuitive, may provide a more efficient means for controlling many confounders, but its ability to control confounding depends on the appropriateness of the chosen model. Hybrid methods based on a multivariate confounder score or a propensity score combine the favorable characteristics of both methods and may be better suited for controlling many confounders. However, the resulting strata are defined by subranges of a multivariate model, and, therefore, may possess little intrinsic meaning. The authors propose the principle of asymmetric stratification to control efficiently a number of confounders in cohort studies while retaining the intuitive appeal and general framework of cross-stratification. The proposed method resembles a propensity score analysis but does not use a multivariate model to define the strata. Instead, strata are defined by the categories of only a subset of the original potential confounders. The authors also demonstrate how our proposed method can be implemented by an application of classification and regression trees (CART) (recursive partitioning), as outlined by Breiman et al. (Classification and Regression Trees. Belmont, CA: Wadsworth, 1984). Computer simulations and an actual example suggest that the proposed method is a potentially simpler alternative to the standard propensity score analysis. Specific recommendations on how the proposed method can be improved are also presented.

Aged↗

PubMed1988

Proportional hazards analysis of risk factors for coronary heart disease in individuals aged 65 or older. The Framingham Heart Study.

Risk factors for coronary heart disease were examined in 2,501 individuals in the Framingham Heart Study who survived to the age of 65 without evidence of coronary artery disease. We used a proportional hazards (Cox) analysis that examined risk factors over time and included events through the 16th biennial examination. The independently significant multivariate correlates of the development of coronary heart disease after the age of 65 were sex (incidence rate ratio [RR] for males of 1.7 [95% confidence interval of 1.4,2.0]), left ventricular hypertrophy (RR = 2.4 [1.7,3.5]), systolic blood pressure (RR = 2.2 [1.4,3.3] for systolic blood pressure of 160 mmHg or higher as compared with less than 120 mmHg), casual blood glucose (RR = 2.2 [1.5,3.4] for 175 mg/dL or more as compared with less than 90 mg/dL), Metropolitan relative weight from examination 1 (RR = 1.3 [1.0,1.6] for those 130% or more of ideal weight compared with those less than 110% of ideal), and total serum cholesterol (RR = 1.8 [1.3,2.5] for cholesterol in the 90th percentile or higher compared with less than 200 mg/dL). Controlling for all these risk factors, those smoking 20 cigarettes a day or more were at slightly increased risk of coronary heart disease (RR = 1.2 [.9,1.6]) compared with nonsmokers. These analyses suggest that older persons share the same risk factors for coronary artery disease that are significant in younger populations.

Aged↗

PubMed1988

The impact of Medicare on early cancer detection in the elderly.

In an analysis of 23 sex-specific tumors in the Connecticut Tumor Registry, only seven tumors showed more than a 20 per cent increase in the ratio of localized tumors detected post-Medicare (1967-75) versus ante-Medicare (1960-65) in the 65-75 year age group compared to the 60-64 year age group. Of these seven, the information was more convincing for lung cancer in males. Overall, however, Medicare had little if any impact on early tumor diagnosis in the elderly.

Age Factors↗

PubMed1987

Diagnostic implications for myocardial ischemia of the circadian variation of the onset of chest pain.

To determine whether the occurrence of chest pain is randomly distributed during the day and to study whether the time of onset is useful in discriminating among causes of chest pain, patients older than 30 years who presented to 7 emergency departments with a chief complaint of chest pain unexplained by trauma or chest x-ray abnormalities were studied. A total of 7,759 patients presented during the study period; of these, 3,990 presented within 6 hours of the onset of pain and were included in the primary analysis. Chest pain caused by acute myocardial infarction, unstable angina pectoris and stable angina pectoris was more likely to begin during the period from 6 AM to noon than would be expected if the onset were uniformly distributed during the day (relative risks 1.15, 1.29 and 1.32, respectively), but chest pain that was caused by nonischemic cardiac causes and by noncardiac causes was also more likely to begin during the same time period (relative risks 1.28 and 1.17). Although chest pain from coronary arterial causes had a distinct circadian variation, the time of onset of pain was not a helpful criterion for determining the cause of chest pain.

Adult↗

PubMed1987

Identification and preliminary validation of predictors of major bleeding in hospitalized patients starting anticoagulant therapy.

Among 617 hospitalized patients who started long-term anticoagulant therapy, major bleeding developed before discharge in 28 (5 percent) and minor bleeding in another 38 (6 percent), with daily incidence rates of 0.4 and 0.5 percent, respectively. The most common site of bleeding was gastrointestinal, and one patient died from bleeding. Four independent risk factors for major in-hospital bleeding were identified and weighted using multivariate discriminant analysis in a randomly chosen group of 411 patients: co-morbid conditions other than the indication for anticoagulant therapy (specific signs of heart, liver, or kidney dysfunction, cancer, and severe anemia); the use of heparin to begin therapy in patients age 60 years or older; the intensity of therapy (measured by the maximal prothrombin time or partial thromboplastin time); and liver dysfunction that worsened during treatment. These findings were validated in an independent testing group of 206 patients; the risk factors identified 151 patients at low (1 percent) risk of major bleeding, 33 at moderate (6 percent) risk, and 22 at high (23 percent) risk. The accuracy and clinical impact of this prediction rule should be evaluated further in other hospitals.

Adolescent↗

PubMed1987

Correlates of cardiac and sudden death after ambulatory monitoring in a community hospital. Importance of clinical characteristics, congestive heart failure and tachyarrhythmias.

To analyze the prognostic importance of arrhythmias on routine 24-hour ambulatory monitoring, we prospectively followed 755 consecutive patients undergoing monitoring at a community hospital in the context of usual medical care. Of the 755 patients, 114 (15%) had ventricular tachycardia on monitoring. At a mean follow-up of 38 months, multivariate survival analysis indicated that congestive heart failure was the strongest correlate of death from all causes (relative risk (RR) = 2.6), cardiac death (RR = 3.5), and the sudden cardiac death (RR = 5.6); ventricular tachycardia was significantly correlated with death from all causes and with cardiac death, but had only a borderline association with sudden cardiac death (RR = 1.9, p = 0.08). While ventricular tachycardia on ambulatory monitoring is of prognostic importance, congestive heart failure, as determined by routine clinical examination, was a more powerful prognostic correlate of deaths from all causes, death from cardiac causes, and sudden cardiac death.

Adolescent↗

PubMed1987

Circadian variation of transient myocardial ischemia in patients with coronary artery disease.

To examine whether a significant circadian variation of transient myocardial ischemia exists and to better understand the character of such variation, 32 patients with chronic stable symptoms of coronary artery disease underwent one or more days of ambulatory monitoring of ischemic ST segment changes during daily life. A total of 251 episodes of ischemic ST segment depression occurred in 24 (75%) of the 32 patients with a median duration of 5 min (range 1 to 253). A significant circadian increase in ischemic activity was found with 39% of episodes and 46% of total ischemic time occurring between 6 A.M. and 12 P.M. (p less than .05 and p = .02, respectively). In 21 patients with ST segment depression during the 6 hr after waking and the 6 hr before sleep, 68% of episodes occurred in the morning compared with 32% in the evening. There were no significant differences in heart rate at onset, heart rate at 1 min before onset, and activity score associated with ST segment depression. The proportion of minutes showing ST segment depression when the heart rate was above the lowest rate associated with ST segment depression was significantly greater in the morning compared with the evening (26% vs 15%; p = .03). Thus the early morning increase in ST segment depression does not appear to be explained by differences in extrinsic activity and/or stress measured by physical activity score and heart rate response. More importantly, this phenomenon is often ignored by the usual patterns of drug administration for angina.(ABSTRACT TRUNCATED AT 250 WORDS)

Adult↗

PubMed1987

Sensitivity of routine clinical criteria for diagnosing myocardial infarction within 24 hours of hospitalization.

Myocardial infarction was diagnosed in 431 (30%) of 1460 patients with acute chest pain who had serial enzyme testing after admission to intensive or intermediate care units at three teaching and three community hospitals. The diagnosis was made within 12 hours of admission in 331 (77%) patients and within 24 hours in 415 (96%). Of the 16 patients with myocardial infarction who did not have enzyme abnormalities within 24 hours, 9 (56%) had recurrent ischemic pain during this 24-hour period. Of 451 patients who had neither enzyme abnormalities nor recurrent ischemic pain in the first 24 hours, only 7 (2%) ultimately met diagnostic criteria for myocardial infarction. These findings were prospectively validated in an independent testing set of 275 patients with myocardial infarction, 271 (99%) of whom either met diagnostic criteria for myocardial infarction or had recurrent ischemic pain within 24 hours of admission. These data suggest that 24 hours is nearly always a sufficient period to exclude myocardial infarction in patients without recurrent chest pain.

Aged↗

PubMed1987

Evaluation of creatine kinase and creatine kinase-MB for diagnosing myocardial infarction. Clinical impact in the emergency room.

We prospectively studied the performance of emergency room strategies using a single sampling of total creatine kinase (CK) only and total CK with, if total CK levels were elevated, CK-MB levels in 639 patients with acute chest pain, including 386 patients who were admitted and 253 patients who were discharged. Acute myocardial infarction was diagnosed in 104 patients and excluded in 535. An elevated total CK level had a sensitivity of only 38% and specificity of only 80%, whereas a CK-MB level over 5% of an elevated total CK level had a sensitivity of only 34% and specificity of 88%. The sensitivities of both CK and CK-MB were higher in patients who arrived more than four hours after the onset of symptoms, and, in this population, the strategy using CK-MB performed significantly better than the strategy using total CK alone. Since a very positive CK-MB in a low-risk patient can greatly raise the probability of myocardial infarction, future strategies using CK-MB may have a role in selected subsets in determining which patients should not be sent home. However, the sensitivity of a single sampling of CK and CK-MB is too low for these assays to be used to exclude myocardial infarction in the emergency room or to be used as the rationale for deciding not to admit a patient.

Adult↗

PubMed1986

Long-term survival of more than 2,000 patients after coronary artery bypass grafting.

Among 2,004 patients who underwent their first coronary artery bypass graft operation between January 1970 and December 1980 without concomitant valve replacement or aneurysmectomy, life-table survival was 89% at 5 years and 80% at 8 years after surgery. In a multivariate Cox model analysis, the independent correlates of long-term survival were emergent operation with cardiogenic shock (multivariate mortality rate ratio [RR] = 14.0), use of a postoperative intraaortic balloon pump (RR = 3.9), ejection fraction less than 50% (RR = 2.4), preoperative history of congestive heart failure (RR = 2.2), cardiopulmonary bypass time (RR = 1.4 for each 30-minute increment), uncorrected mitral regurgitation (RR = 1.5 for each increment of angiographic gradation), left main coronary artery narrowing (RR = 1.7) and diabetes (RR = 1.6). After controlling for these factors, age, sex and the percentage of narrowings that were bypassed were not independent correlates of long-term survival.

Adult↗

PubMed1986

Value of noninvasive testing in adults with suspected aortic stenosis.

To determine the predictors of surgically correctable aortic stenosis in patients with systolic murmurs, 231 patients were evaluated. Five variables (carotid upstroke timing, carotid upstroke volume, aortic valve calcification on chest radiography, single or absent second heart sound, and a murmur with its maximal intensity at the right upper sternal border) were significant multivariate correlates. Two echocardiographic factors (a maximal aortic valve leaflet separation of 7 mm or less and hypertrophy of the posterior wall of the left ventricle to 12 mm or more) and one systolic time interval factor (a rate-corrected ejection time of more than 340 msec) added significant incremental information. When prospectively tested on an independent set of 86 patients with suspected aortic outflow obstruction, the combined clinical and noninvasive information correctly placed 10 patients (12 percent) into a low-risk group in which catheterization may not be indicated and 15 patients (17 percent) into a high-risk group in which it might be avoided or limited to coronary arteriography. This approach to predicting aortic stenosis deserves wider prospective testing.

Angina Pectoris↗