Search PubMed⌕ Search

Biomedical subjects

E F Cook

Publications and source records attributed to E F Cook.

At least 19 recordsLinked to original sources
PubMed1992

Detection and significance of intraoperative and postoperative myocardial ischemia in peripheral vascular surgery.

OBJECTIVE: To determine the incidence and significance of intraoperative and postoperative myocardial ischemia and their relationship to preoperative ischemia and postoperative cardiac events in patients undergoing peripheral arterial surgery. DESIGN: Prospective cohort trial. PATIENTS: One hundred fifteen patients undergoing elective vascular surgery who met predefined eligibility criteria and were thought to have acceptable cardiac risk as assessed by independent cardiologists. INTERVENTIONS: Ambulatory electrocardiographic monitoring preoperatively, intraoperatively, and up to 72 hours postoperatively. MEASUREMENTS: Preoperative clinical characteristics and laboratory data were collected. Predefined adverse cardiac events were identified by an investigator who was "blinded" to monitoring results. Monitor recordings were interpreted for ST-segment depression by investigators blinded to patient information. MAIN RESULTS: Intraoperative ischemia was present in 21 patients (18%), and postoperative ischemia was present in 35 (30%). There were 16 postoperative cardiac events. The relative risk of suffering a cardiac event was 2.7 in patients with intraoperative ischemia and was 16 in patients with postoperative ischemia. Preoperative ischemia closely correlated with intraoperative and postoperative ischemia. Preoperative and postoperative ischemia preceded cardiac events in 14 of 16 patients. CONCLUSIONS: Preoperative ischemia appears to identify high-risk patients, and subsequent perioperative monitoring detects silent ischemia that commonly precedes clinical events and that may be treatable with anti-ischemia therapy.

Adult↗
PubMed1992

Development of the 'Activities of Daily Vision Scale'. A measure of visual functional status.

To develop a method for the evaluation of visual function in subjects with cataracts, the authors identified 20 visual activities and categorized them into five subscales (distance vision, near vision, glare disability, night driving, and daytime driving) that comprised the Activities of Daily Vision Scale (ADVS). Each subscale in the ADVS was scored between 100 (no visual difficulty) and 0 (inability to perform the activity because of visual difficulty). In 334 subjects scheduled for cataract extraction (mean age 75 +/- 9 years, 67% women), ADVS scores (mean +/- standard deviation) for each subscale ranged from 44 +/- 31 for night driving to 72 +/- 24 for near vision activities. When administered by telephone, inter-rater reliability coefficients (r) were 0.82 to 0.97 (P < 0.001) for each of the subscales, and test-retest reliability was 0.87 for the scale overall. Cronbach's coefficient alpha was very high for both the in-person (alpha = 0.94) and telephone (alpha = 0.91) formats. Criterion validity, the correlation between visual loss and ADVS score, was -0.37 (P < 0.001) when the ADVS was administered in person and -0.39 (P < 0.001) when it was administered by telephone. Content validity as assessed with factor analysis showed that 88% of the variance of the principal components weighted on one factor. The authors conclude that substantial visual disability is not captured by routine visual testing and that the ADVS is a reliable and valid measure of patient's perception of visual functional impairment.

Activities of Daily Living↗
PubMed1991

Outcome of patients who were admitted to a new short-stay unit to "rule-out" myocardial infarction.

For emergency room patients with a low probability of acute myocardial infarction, we established a new short-stay coronary observation unit, a 2-bed nonintensive care unit with telemetry monitoring adjacent to the emergency room. Of 512 consecutive admissions to the coronary observation unit, 425 (83%) were discharged home without evidence of acute myocardial infarction or serious complications (mean length of stay, 1.2 days; median length of stay, 1 day); 87 (17%) were transferred to other hospital beds. The rate of acute myocardial infarction was 3%. No deaths and only 1 serious complication occurred in the coronary observation unit. At 6 month follow-up, the cardiac survival rate was 99% for patients sent home directly from this unit. It is concluded that the coronary observation unit is safe and adequate for ruling out acute myocardial infarction in a defined subset of patients. Short-stay units, however, encourage early discharges which, when premature, may miss patients who are at risk of having complications shortly thereafter. Strategies such as mandatory but expeditious predischarge stress testing to encourage early but not premature discharge may augment the efficiency of coronary observation units.

Coronary Care Units↗
PubMed1991

Ruling out acute myocardial infarction. A prospective multicenter validation of a 12-hour strategy for patients at low risk.

BACKGROUND: Although previous investigations have suggested that 24 hours is required to exclude acute myocardial infarction in patients who are admitted to a coronary care unit for the evaluation of acute chest pain, we hypothesized that a 12-hour period might be adequate for patients with a low probability of infarction at the time of admission. METHODS: Using a Bayesian model, we developed a strategy to identify candidates for a shorter period of observation from an analysis of a derivation set of 976 patients with acute chest pain who were admitted to three teaching and four community hospitals. In the derivation set, patients whose clinical characteristics in the emergency room predicted a low (less than or equal to 7 percent) probability of myocardial infarction had only a 0.4 percent risk of infarction if they had neither abnormal levels of cardiac enzymes nor recurrent ischemic pain during the first 12 hours of hospitalization. In an independent testing set of 2684 patients from the seven hospitals, 957 admitted patients (36 percent) were classified as candidates for this 12-hour period of observation according to a previously published multivariate algorithm. Few of these patients were actually transferred from a monitored setting at 12 hours. RESULTS: Of the 771 candidates for a 12-hour period of observation who did not have enzyme abnormalities or recurrent pain during the first 12 hours, 4 (0.5 percent) were subsequently found to have acute myocardial infarction, and only 3 (0.4 percent) died after primary cardiac arrests, all of which occurred three to five days after admission. Rates of other major cardiovascular complications were low in the patients who might have been transferred from the coronary care unit after 12 hours with this strategy. In patients with a higher initial risk of infarction, the standard strategy of 24-hour observation identified all but 11 of 739 acute myocardial infarctions (1 percent). CONCLUSIONS: Emergency room clinical data can be used to identify a large subgroup of patients for whom a 12-hour period of observation is normally sufficient to exclude acute myocardial infarction. Patient-specific evaluation and treatment can then proceed without the restrictions imposed by "rule-out" protocols for myocardial infarction.

Adult↗
PubMed1991

Impact of physician experience on triage of emergency room patients with acute chest pain at three teaching hospitals.

PURPOSE: To determine whether the experience of the physician (as measured by postgraduate training level or time during the academic year) who performs the initial evaluation affects the triage of patients with acute chest pain. PATIENTS AND METHODS: Prospective data on the presenting clinical features, initial triage, final diagnosis, and complications were collected for 7,857 patients who presented to the emergency rooms of three teaching hospitals, including 1,118 (14%) with acute myocardial infarction (AMI), 2,477 (32%) with acute ischemic heart disease (AIHD) (i.e., AMI or unstable angina), and 335 (4%) with major complications. The experience of the evaluating physicians, who were in their first three postgraduate years in 93% of cases, was measured in three ways: (1) postgraduate training level, (2) month during the academic year, and (3) number of patients with acute chest pain previously evaluated. Multivariate logistic regression analyses that adjusted for hospital site and 20 clinical variables estimated the odds ratios for admission to the coronary care unit (CCU) and hospital associated with each incremental increase in physician experience. RESULTS: With more experience (as measured by postgraduate training level or time during the academic year), the sensitivity of physicians for admitting patients with AMI, AIHD, or major complications to the hospital increased. For example, each incremental increase in postgraduate training level carried a 1.4 increase in the adjusted odds ratio for admission of a patient with AIHD to the hospital (p less than 0.05), corresponding to an increase in the probability of admission from 93% to 97%. However, increasing physician experience was also associated with an elevated false-positive rate in admitting patients without these diagnoses to the CCU and hospital. Thus, each incremental increase in postgraduate training level carried a 1.2 increase in the adjusted odds ratio for admission of a patient without AIHD to the CCU and hospital (p less than 0.005), corresponding to an increase in the probability of admission from 34% to 47%. By receiver operating characteristic curve (ROC) regression analyses, these changes in triage patterns were consistent with movement along a single ROC curve, rather than a shift to a new or better ROC curve. CONCLUSIONS: As the experience of the physician who performed the initial evaluation increased, there was a lower threshold for admitting all patients with and without AMI, AIHD, or major complications to the CCU and hospital without a detectable improvement in diagnostic accuracy.

Adult↗
PubMed1991

Cyclophosphamide cardiotoxicity in bone marrow transplantation: a prospective evaluation of new dosing regimens.

Cyclophosphamide (CY) cardiotoxicity may be a lethal complication of bone marrow transplantation. Previous echocardiographic studies have reported that left ventricular dysfunction due to CY occurs in over 50% of patients undergoing transplantation. To evaluate the cardiotoxicity of new dosing protocols that included twice-daily rather than once-daily CY, 44 bone marrow transplantation patients were prospectively evaluated with serial ECGs and echocardiograms. Twenty-six patients received a once-daily lower-dose protocol (mean total 87 +/- 11 mg/kg), and 18 patients received a twice-daily higher-dose (mean total 174 +/- 34 mg/kg) CY regimen. In the higher-dose CY group, significant reductions in summed ECG voltage (-20%) (P less than .01) and increases in left ventricular mass index (LVMI) (+10%) (P less than .05) were detected in the first week following therapy. These changes resolved by the third week following CY and were significantly greater than the changes noted in the lower-dose group. However, left ventricular ejection fraction (EF) did not change significantly in either group. Five patients developed clinical cardiotoxicity (four, pericarditis; one, congestive heart failure); four of the five patients were in the higher-dose group (P = .14). Only a prior history of congestive heart failure or a baseline EF less than 50% was an independent correlate of clinical cardiotoxicity (P less than .05). Thus, dose-dependent cardiotoxicity following the use of CY for bone marrow transplantation is evident as reversible decreases in ECG voltage and increases in left ventricular mass, possibly reflecting myocardial edema or hemorrhage. However, systolic dysfunction is much less common with these new twice-daily dosing regimens when compared with earlier studies of high-dose once-daily CY.

Adult↗
PubMed1990

Long-term prognosis of myocardial ischemia detected by Holter monitoring in peripheral vascular disease.

To assess the long-term prognostic significance of myocardial ischemia, as measured by ambulatory electrocardiographic monitoring, in patients with occlusive peripheral arterial disease, 176 eligible patients scheduled for elective peripheral arterial surgery at Brigham and Women's Hospital were prospectively studied. All patients were monitored preoperatively without alterations to baseline medications. Prospective follow-up was obtained during routine medical care as provided by blinded, independent physicians and by subsequent telephone contact with the patients. Thirty-two patients (18%) had a total of 75 episodes of myocardial ischemia, 73 (97%) of which were asymptomatic. During a mean follow-up period of 615 days, there were 9 cardiac deaths, 1 occurring in-hospital after peripheral vascular surgery, and 13 nonfatal myocardial infarctions, 4 occurring in-hospital after peripheral vascular surgery. Cardiac events occurred in 12 of 32 patients with ischemia (38%), including 6 cardiac deaths, and in 10 of 144 patients without ischemia (7%), including 3 cardiac deaths (risk ratio 5.4, 95% confidence interval 2.6 to 11.4). The sensitivity of ischemia was 55%, the specificity was 87%, the positive predictive value was 38%, and the negative predictive value was 93%. In a multivariate Cox proportional-hazards model controlling for age, gender, coronary risk factors, history of angina, myocardial infarction, coronary artery disease and antianginal medications, the presence of ischemia was the only independent predictor of outcome. In patients with peripheral arterial disease, who often are unable to perform adequate exercise testing, ambulatory monitoring for myocardial ischemia is a significant independent predictor of 1- to 2-year prognosis.

Adult↗
PubMed1990

Predicting bacteremia in hospitalized patients. A prospectively validated model.

OBJECTIVE: To develop and validate a model for the prediction of bacteremia in hospitalized patients, and to identify subgroups of patients with a very low likelihood of bacteremia in whom a positive blood culture has a low positive predictive value. DESIGN: Prospective cohort study with clinical data on 1516 episodes collected from a random sample of all patients who had blood cultures done at one institution. SETTING: Urban, tertiary care hospital. PATIENTS: Derivation set: 1007 blood culture episodes sampled from all blood cultures done on patients at Brigham and Women's Hospital between October 1988 and February 1989. Validation set: 509 episodes, May 1989 to June 1989. The unit of evaluation was the episode, defined as a 48-hour period beginning after a blood culture was drawn. MEASUREMENTS AND MAIN RESULTS: True- and false-positive rates of blood cultures in the derivation set as assessed by independent reviewers were 7% (74 of 1007) and 8% (81 of 1007), respectively. Independent multivariate predictors of true bacteremia were temperature of 38.3 degrees C or higher, presence of a rapidly (less than 1 month) or ultimately (less than 5 years) fatal disease; shaking chills; intravenous drug abuse; acute abdomen on examination; and major comorbidity. In the low-risk group, defined by absence of these predictors, the misclassification rate of the model in the derivation set was 1% (4 of 303), and a positive blood culture had a positive predictive value of only 14% for true bacteremia. The model also identified a high-risk subset in which 16% (41 of 264) of episodes represented true bacteremia. The model was prospectively validated in 509 additional episodes, and the misclassification rate in the low-risk group was 2% (3 of 155). INTERVENTIONS: None. CONCLUSION: These findings provide a means of stratifying hospitalized patients according to their risk for bacteremia. If prospectively validated in other settings, this model may be helpful when deciding whether or not to do blood cultures or start antibiotic therapy and, when evaluating a positive blood culture, to determine whether or not it is a true-positive.

Adult↗
PubMed1990

Outcomes in patients with myocardial infarction who are initially admitted to stepdown units: data from the Multicenter Chest Pain Study.

PURPOSE: To assess whether the admission of patients with chest pain to a stepdown unit would jeopardize the outcome of those patients who ultimately "ruled in" for a myocardial infarction. PATIENTS AND METHODS: We compared the risk of an adverse outcome in initially uncomplicated, "rule-out myocardial infarction" patients who were admitted directly to a stepdown unit (n = 58) or to a coronary care unit (n = 409) at 6 hospitals and who then ultimately "ruled in" for a myocardial infarction. RESULTS: An adverse outcome (death, serious complication, or invasive intervention) occurred in 16 (28%) stepdown unit patients compared with 159 (39%) coronary care unit patients. Among patients eligible for initial care in either location, the risk of an adverse outcome after controlling for clinical characteristics was similar in the two groups using each of two different multivariate approaches. CONCLUSION: Although our study was not of sufficient size to exclude the possibility of a small benefit from initial triage to a coronary care unit, our data suggest that (1) admission of initially uncomplicated chest pain patients with a relatively low probability of acute myocardial infarction to a stepdown unit does not seriously jeopardize those who eventually "rule in" for myocardial infarction; and (2) larger observational or randomized studies, which could reduce the residual possibility of somewhat higher risk in the stepdown unit, would be ethical to perform.

Aged↗
PubMed1990

Comparison of the natural history of new onset and exacerbated chronic ischemic heart disease. The Chest Pain Study Group.

To compare the natural history of patients with new onset ischemic heart disease with that of patients with exacerbations of chronic ischemic heart disease, short- and long-term outcomes of 3,465 emergency room patients with acute ischemic heart disease at four community and three university hospitals were evaluated. Acute myocardial infarction was diagnosed in 598 (33%) of the 1,835 patients with a prior history of infarction or angina and 934 (57%) of the 1,630 without such a history (p less than 0.001). Patients with new onset ischemic heart disease with acute myocardial infarction were more likely than patients with infarction and exacerbated chronic ischemic heart disease to have Q wave infarction (57% versus 36%) and to receive thrombolytic therapy (11% versus 5%); they also had higher maximal creatine kinase levels (1,088 +/- 1,299 versus 733 +/- 906 U/liter) (p less than 0.0001 for all three). After adjustment for differences in clinical presentation and initial triage, patients with new onset ischemic heart disease with acute myocardial infarction were less likely than the comparison group to have congestive complications (odds ratio 0.63, 95% confidence interval 0.47 to 0.84, p less than 0.01) but not less likely to have arrhythmic, ischemic or overall complications. Among patients with angina without acute myocardial infarction, patients with new onset ischemic heart disease were less likely to have recurrent ischemic pain and congestive heart failure. In multivariate analysis of long-term follow-up data on 457 patients from one hospital, patients with new onset ischemic heart disease had better cardiovascular survival rates.(ABSTRACT TRUNCATED AT 250 WORDS)

Acute Disease↗
PubMed1990

Statistical methods in SUPPORT.

The analysis and interpretation of the data collected in SUPPORT provide great potential for understanding the relationships among treatment choices, patient and physician values and preferences, perceptions about the risks and benefits of treatments, institutional characteristics, and outcomes (as measured by quality of life, survival, and satisfaction). The complicated analyses required to elucidate these relationships will pose many technical challenges in dealing with longitudinal observational data collected from seriously ill patients at multiple sites. Major challenges include the handling of incomplete data, proper parameterization of treatment effects, strategies to avoid various potential biases, validating predictive models, and constructing endpoints that combine survival with quality of life. Within the structure of the SUPPORT study, mechanisms have been established to guide the analyses and to ensure their quality and validity.

Confounding Factors, Epidemiologic↗
PubMed1990

Impact of initial triage decisions on nursing intensity for patients with acute chest pain.

The results of a prospective evaluation of the patient-specific correlates of nursing intensity for 183 consecutive emergency room patients admitted for evaluation of acute chest pain, including 33 (18%) with acute myocardial infarction (AMI), are reported. These correlates were measured with a previously-validated, commercially-available patient classification tool (Medicus). In multiple linear regression analysis that adjusted for the effects of 31 clinical variables from the initial presentation and subsequent course, initial triage to the coronary care unit had a significant independent correlation with initial Medicus score (P less than 0.0001) and mean Medicus score from the first three days of hospitalization (P less than 0.0001). In a subset of 74 uncomplicated non-AMI patients, coronary care unit admissions were more likely to have vital signs taken every 2 hours, to receive oxygen therapy and assistance with feeding, and to be transferred to another unit within three days. Findings suggest that, after adjusting for severity of illness, initial triage of patients with acute chest pain to the coronary care unit is associated with increased nursing utilization because of 1) the routine application of standard coronary care unit protocols that were developed for high-risk patients, and 2) the nursing efforts required by early transfer of patients out of the coronary care unit.

Adult↗
PubMed1990

Role of lung biopsy in patients undergoing a modified Fontan procedure.

Fifty-six patients undergoing a modified Fontan procedure from 1977 to 1985 had lung tissue available from the perioperative period (n = 51) or recent palliative procedure (n = 5) (37 biopsy, 12 autopsy, 7 biopsy and autopsy). Twenty-six patients had tricuspid atresia, and 30 had other causes of functional univentricular heart. Pulmonary arterial structure was classified by both the Heath-Edwards and morphometric scales. Biopsy and autopsy grades were identical except for a minor difference in Heath-Edwards scale in one case. Elevated preoperative mean pulmonary artery pressure (greater than 15 mm Hg) was associated with medial hypertrophy (p = 0.04), although six patients with normal pulmonary artery pressure had medial hypertrophy and six patients with mean pulmonary artery pressure greater than 15 mm Hg did not have medial hypertrophy. Medial hypertrophy was a possible risk factor in patients with univentricular heart: Patients with medial hypertrophy had a ninefold greater risk of death (p = 0.08). Controlling for confounding variables did not alter this result (0.06 less than or equal to p less than or equal to 0.10). Medial hypertrophy was not a risk factor for death in patients with tricuspid atresia: Four had this feature and each survived. Four patients with tricuspid atresia died and none had medial hypertrophy. A mean transpulmonary pressure gradient was measured on the first postoperative day in 45 patients. In patients with medial hypertrophy there was a trend toward a higher gradient (p = 0.13). All patients with medial hypertrophy who had this measurement (n = 9) had a transpulmonary pressure gradient of 7 mm Hg or greater, whereas this was true in only 18 of 36 without medial hypertrophy (p = 0.007). On the basis of an evaluation of medial hypertrophy, we conclude that biopsy does not aid in selection of patients with tricuspid atresia for a Fontan procedure; its role in patients with univentricular heart is uncertain.

Adolescent↗
PubMed1989

Clinical characteristics and outcome of acute myocardial infarction in patients with initially normal or nonspecific electrocardiograms (a report from the Multicenter Chest Pain Study).

To determine the prevalence and characteristics of acute myocardial infarction (AMI) patients who present to emergency departments with normal or nonspecific electrocardiograms (ECGs), data were analyzed from 7,115 consecutive patients in the Multicenter Chest Pain Study. AMI patients with normal or nonspecific initial ECGs (n = 107) were less likely to have a past history of coronary artery disease or to be diaphoretic on presentation (p less than 0.01) than AMI patients with initial ECGs highly suggestive of AMI (n = 811). The overall probability of AMI among patients with chest pain and initially normal or nonspecific ECGs was 3%, but ranged from less than 1 to 17% depending on the patient's age and sex and whether the patient had pressure-type pain or pain radiating to the shoulder, neck or arms. Among initially admitted patients, the time elapsed between onset of pain and presentation was similar in both groups. However, the time between onset of pain and definitive diagnosis of AMI by enzymes or clinical course was longer in patients with initially normal or nonspecific electrocardiograms (8.3 vs 7.5 hours, p less than 0.05), their peak creatine kinase levels were lower (mean 643 vs 1,032 mg/dl, p less than 0.001) and their mortality was slightly lower (6 vs 12%, p = 0.10). These findings suggest that AMI patients with initially normal or nonspecific ECGs may have a less severe short-term clinical outcome.

Adult↗
PubMed1989

Correlation between preoperative ischemia and major cardiac events after peripheral vascular surgery.

Patients who undergo peripheral vascular surgery are at increased risk for postoperative cardiac events and are difficult to assess preoperatively because of limitations on their activity. We prospectively studied 176 consecutive eligible patients undergoing elective vascular surgery to determine the value in predicting a postoperative cardiac event of preoperative electrocardiographic monitoring to detect myocardial ischemia. Of the 176 patients, 32 (18 percent) had 75 episodes of monitored ischemic ST-segment depression preoperatively (of which 73 were asymptomatic), and 13 (7 percent) met strict criteria for major postoperative cardiac events, including 1 with a fatal myocardial infarction, 3 with nonfatal infarctions, 4 with unstable angina, and 5 with ischemic pulmonary edema. Of the 32 patients with ischemia before their operations, 12 had postoperative events (univariate relative risk, 54; 95 percent confidence interval, 7.2 to 400). Only 1 postoperative event occurred among 144 patients who did not have preoperative ischemia. The sensitivity of preoperative ischemia was 92 percent, the specificity 88 percent, the predictive value of a positive result 38 percent, and the predictive value of a negative result 99 percent. In multivariate analyses, preoperative ischemia was the most significant correlate of postoperative cardiac events and remained a statistically significant independent correlate even after we had controlled for all other preoperative factors (multivariate relative risk, 24.4; 95 percent confidence interval, 6.8 to 88). These preliminary data suggest that preoperative electrocardiographic monitoring to detect episodes of myocardial ischemia is a useful method for assessing cardiac risk in patients who undergo elective vascular surgery. In particular, the absence of ischemia during monitoring indicates a very low risk.

Aged↗
PubMed1989

Comparison of clinical presentation of acute myocardial infarction in patients older than 65 years of age to younger patients: the Multicenter Chest Pain Study experience.

To assess whether the manifestations of acute ischemic heart disease in the elderly are less typical than in younger patients, the presenting clinical features and their associated relative risks for acute myocardial infarction (AMI) were compared in 2,625 patients greater than or equal to 65 years of age and 5,109 patients less than 65 years of age. These patients were evaluated for acute chest pain in the emergency departments of 7 hospitals. The same features were associated with increased relative risks for AMI in younger and elderly patients. The relative risks among older patients, however, were consistently closer to 1.0 for classic features, including male gender, pressure-like quality of pain, substernal location, typical pattern of pain radiation and electrocardiographic evidence of ischemia or AMI. Analyses for the endpoint "acute ischemic heart disease" (i.e., AMI or unstable angina) yielded similar findings. Elderly patients were more likely to be admitted to the hospital (56 vs 35%; p less than 0.0001) and to the coronary care unit (37 vs 23%; p less than 0.0001) in the absence of either AMI or unstable angina. These data support the hypothesis that diagnosis of acute chest pain is especially difficult in elderly patients.

Adult↗
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↗