-
[show abstract]
[hide abstract]
ABSTRACT: PURPOSE: To determine whether postsecondary research experiences (in college, medical school, and residency) and other variables mediate racial/ethnic disparities in U.S. medical school graduates' full-time faculty appointments in academic medicine. METHOD: Individualized, deidentified records for 1994-2000U.S. medical school matriculants who graduated with MDs before 2005, completed graduate medical education before 2009, and had data for all variables were examined for potential mediators of racial/ethnic disparities in full-time faculty appointments using the SAS macro "MEDIATE" for estimation and statistical inference. Controlling for gender, parents' occupation, and graduation year, the authors estimatedthe effects of potential mediators in separate models comparing Asian/Pacific Islander (PI) versus underrepresented minority (URM; including African American, Hispanic, and Native American/Alaska Native) graduates and white versus URM graduates. RESULTS: Of 82,758 eligible graduates, 62,749 (75.8%) had complete data; of these, 11,234 (17.9%) had full-time faculty appointments, including 18.4% (7,848/42,733) of white, 18.8% (2,125/11,297) of Asian/PI, and 14.5% (1,261/8,719) of URM graduates. Proportion of total race/ethnicity effect on full-time faculty appointment explained by all mediators was 66.0% (95% confidence interval [CI], 44.7%-87.4%) in a model comparing Asians/PIs with URMs and 64.8% (95% CI, 52.2%-77.4%) in one comparing whites with URMs. Participation in postsecondary research activities, authorship during medical school, academic achievement, and faculty career intentions at graduation were among the significant mediators explaining the effect of race/ethnicity onfull-time faculty appointment. CONCLUSIONS: Postsecondary research experiences for URM students are among the mediators of racial/ethnic disparities in full-time faculty appointments and, therefore, may increase academic medicine faculty diversity.
Academic medicine: journal of the Association of American Medical Colleges 09/2012; · 2.34 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Randomized start and withdrawal designs have been recently proposed to test the disease-modifying agents on Alzheimer's disease (AD). This article provides methods to determine the optimum parameters for these designs. A general linear mixed effects model is proposed. This model employs a piecewise linear growth pattern for those in the delayed treatment or early withdrawal arm, and incorporates a potential correlation on the rates of change on efficacy outcome before and after the treatment switch. Based on this model, we formulate the disease-modifying hypothesis by comparing the rate of change on efficacy outcome between treatment arms with and without a treatment switch, and develop a methodology to optimally determine the sample size allocations to different treatment arms as well as the time of treatment switch for subjects whose treatment is changed. We then propose an intersection-union test (IUT) to assess the disease-modifying efficacy, and study the size and the power of the IUT. Finally, we employ two recently published symptomatic trials on AD to obtain pilot estimates to model parameters, and provide the optimum design parameters including total and individual sample size to different arms as well as the time of treatment switch for future disease-modifying trials on AD.
Statistics in Biopharmaceutical Research 08/2012; 4(3):216-227. · 0.54 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Income disparities in mortality are profound in the United States, but reasons for this remain largely unexplained. The objective of this study was to assess the effects of health behaviors, and other mediating pathways, separately and simultaneously, including health insurance, health status, and inflammation, in the association between income and mortality.
This study used data from 9925 individuals aged 20 years or older who participated in the 1999-2004 National Health and Nutrition Examination Survey (NHANES) and were followed up through December 31, 2006 for mortality. The outcome measures were all-cause and CVD/diabetes mortality. During follow-up 505 persons died, including 196 deaths due to CVD or diabetes.
After adjusting for age, sex, education, and race/ethnicity, risk of death was higher in low-income than high-income group for both all-cause mortality (Hazard ratio [HR], 1.98; 95% confidence interval [CI]: 1.37, 2.85) and cardiovascular disease (CVD)/diabetes mortality (HR, 3.68; 95% CI: 1.64, 8.27). The combination of the four pathways attenuated 58% of the association between income and all-cause mortality and 35% of that of CVD/diabetes mortality. Health behaviors attenuated the risk of all-cause and CVD/diabetes mortality by 30% and 21%, respectively, in the low-income group. Health status attenuated 39% of all-cause mortality and 18% of CVD/diabetes mortality, whereas, health insurance and inflammation accounted for only a small portion of the income-associated mortality (≤6%).
Excess mortality associated with lower income can be largely accounted for by poor health status and unhealthy behaviors. Future studies should address behavioral modification, as well as possible strategies to improve health status in low-income people.
PLoS ONE 01/2012; 7(11):e49929. · 4.09 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: To develop and validate a risk prediction model that could identify patients at high risk for Clostridium difficile infection (CDI) before they develop disease.
Retrospective cohort study in a tertiary care medical center.
Patients admitted to the hospital for at least 48 hours during the calendar year 2003.
Data were collected electronically from the hospital's Medical Informatics database and analyzed with logistic regression to determine variables that best predicted patients' risk for development of CDI. Model discrimination and calibration were calculated. The model was bootstrapped 500 times to validate the predictive accuracy. A receiver operating characteristic curve was calculated to evaluate potential risk cutoffs.
A total of 35,350 admitted patients, including 329 with CDI, were studied. Variables in the risk prediction model were age, CDI pressure, times admitted to hospital in the previous 60 days, modified Acute Physiology Score, days of treatment with high-risk antibiotics, whether albumin level was low, admission to an intensive care unit, and receipt of laxatives, gastric acid suppressors, or antimotility drugs. The calibration and discrimination of the model were very good to excellent (C index, 0.88; Brier score, 0.009).
The CDI risk prediction model performed well. Further study is needed to determine whether it could be used in a clinical setting to prevent CDI-associated outcomes and reduce costs.
Infection Control and Hospital Epidemiology 04/2011; 32(4):360-6. · 3.67 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: We examined the impact of surgical treatments (breast-conserving surgery [BCS], mastectomy alone, mastectomy with reconstruction) and surgical side-effects severity on early stage (0-IIA) breast cancer patients' body image over time. We interviewed patients at 4-6 weeks (T1), six (T2), 12 (T3), and 24 months (T4) following definitive surgical treatment. We examined longitudinal relationships among body image problems, surgery type, and surgical side-effects severity using the Generalized Estimating Equation approach, controlling for demographic, clinical, and psychosocial factors. We compared regression coefficients of surgery type from two models, one with and one without surgical side-effects severity. Of 549 patients enrolled (mean age 58; 75% White; 65% BCS, 12% mastectomy, 23% mastectomy with reconstruction), 514 (94%) completed all four interviews. In the model without surgical side-effects severity, patients who underwent mastectomy with reconstruction reported poorer body image than patients who underwent BCS at T1-T3 (each P < 0.02), but not at T4. At T2, patients who underwent mastectomy with reconstruction also reported poorer body image than patients who underwent mastectomy alone (P = 0.0106). Adjusting for surgical side-effects severity, body image scores did not differ significantly between patients with BCS and mastectomy with reconstruction at any interview; however, patients who underwent mastectomy alone had better body image at T2 than patients who underwent mastectomy with reconstruction (P = 0.011). The impact of surgery type on body image within the first year of definitive surgical treatment was explained by surgical side-effects severity. After 2 years, body image problems did not differ significantly by surgery type.
Breast Cancer Research and Treatment 02/2011; 126(1):167-76. · 4.43 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Limited psychometric research has examined the reliability of self-reported measures of neighbourhood conditions, the effect of measurement error on associations between neighbourhood conditions and health, and potential differences in the reliabilities between neighbourhood strata (urban vs rural and low vs high poverty). We assessed overall and stratified reliability of self-reported perceived neighbourhood conditions using five scales (social and physical disorder, social control, social cohesion, fear) and four single items (multidimensional neighbouring). We also assessed measurement error-corrected associations of these conditions with self-rated health.
Using random-digit dialling, 367 women without breast cancer (matched controls from a larger study) were interviewed twice, 2-3 weeks apart. Test-retest (intraclass correlation coefficients (ICC)/weighted κ) and internal consistency reliability (Cronbach's α) were assessed. Differences in reliability across neighbourhood strata were tested using bootstrap methods. Regression calibration corrected estimates for measurement error.
All measures demonstrated satisfactory internal consistency (α ≥ 0.70) and either moderate (ICC/κ=0.41-0.60) or substantial (ICC/κ=0.61-0.80) test-retest reliability in the full sample. Internal consistency did not differ by neighbourhood strata. Test-retest reliability was significantly lower among rural (vs urban) residents for two scales (social control, physical disorder) and two multidimensional neighbouring items; test-retest reliability was higher for physical disorder and lower for one multidimensional neighbouring item among the high (vs low) poverty strata. After measurement error correction, the magnitude of associations between neighbourhood conditions and self-rated health were larger, particularly in the rural population.
Research is needed to develop and test reliable measures of perceived neighbourhood conditions relevant to the health of rural populations.
Journal of epidemiology and community health 11/2010; 66(4):342-51. · 3.04 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: The relationship between presence of diabetes and adverse neighborhood and housing conditions and their effect on functional decline is unclear. We examined the association of adverse neighborhood (block face) and housing conditions with incidence of lower-body functional limitations among persons with and those without diabetes using a prospective population-based cohort study of 563 African Americans 49-65 years of age at their 2000-2001 baseline interviews.
Participants were randomly sampled African Americans living in the St. Louis area (response rate: 76%). Physician-diagnosed diabetes was self reported at baseline interview. Lower-body functional limitations were self reported based on the Nagi physical performance scale at baseline and the three-year follow-up interviews. The external appearance of the block the respondent lived on and five housing conditions were rated by study interviewers. All analyses were done using propensity score methods to control for confounders.
109 (19.4%) of subjects experienced incident lower-body functional limitations at three-year follow-up. In adjusted analysis, persons with diabetes who lived on block faces rated as fair-poor on each of the five conditions had higher odds (7.79 [95% confidence interval: 1.36-37.55] to 144.6 [95% confidence interval: 4.45-775.53]) of developing lower-body functional limitations than the referent group of persons without diabetes who lived on block faces rated as good-excellent. At least 80 percent of incident lower-body functional limitations was attributable to the interaction between block face conditions and diabetes status.
Adverse neighborhood conditions appear to exacerbate the detrimental effects on lower-body functioning associated with diabetes.
BMC Public Health 01/2010; 10:283. · 2.00 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Medical education topics might be locally prioritized using public health data on health outcomes and risk factors unrelated to quality of care.
The Missouri Information for Community Assessment (MICA) supplied preventable hospitalization rates (PHRs) for asthma, chronic obstructive pulmonary disease (COPD), diabetes, heart failure, and hypertension in 114 counties from 1998 to 2002. For each disease, a linear regression model predicted PHR from behavior, access, and disease prevalence data from MICA and other public data sources. For each disease in each county, the residual, unexplained PHR should include effects of local medical practices. Variation in relative priority of diseases between counties was estimated from raw PHR and unexplained PHR.
The raw values of the five PHRs varied geographically in different patterns. Regression models explained between 46% and 83% of the variability. The medical education priorities implied by unexplained PHR values differ from priorities inferred from unadjusted PHR or disease prevalence.
Patient behavior and poor health care access contribute to PHR but do not fully explain variation in PHR. If county-level unexplained PHR values identify high priority medical education topics, then other measures of importance, notably disease prevalence and PHR, are poor identifiers of high value topics. Although available predictor and outcome variables constrain the current analysis, unexplained variation in health outcome measures might identify educational opportunities. These observations suggest strategies for balancing and evaluating controlled trials of knowledge dissemination efforts and eventually for deploying educational activities.
Journal of Continuing Education in the Health Professions 01/2009; 28(4):197-204. · 1.52 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: To compare the efficacy of pain reduction of triage oxycodone (O) versus codeine (C) to children with suspected forearm fractures.
Children, aged 4 to 17 years, were randomized to receive O (0.2 mg/kg; maximum, 15 mg) or C (2 mg/kg; maximum, 120 mg) if isolated forearm fracture was suspected by the emergency department (ED) triage nurse. All other ED staff were blinded to the assignment. The primary outcome measure was a 5-point facial scale (0 = no pain, 4 = severe) completed by subjects to assess pain at baseline then at 30-minute intervals until ED discharge or procedural sedation for fractures requiring reduction. Ten adverse effects were assessed at baseline and the succeeding intervals. Identification of the most painful part of the visit was assessed at discharge. Efficacy and adverse effects of O versus C were compared using generalized estimate equation modeling.
One hundred seven subjects (mean age, 10.4 years; African American, 55%; males, 56%) were randomized to O (n = 51) or C (n= 56). Subjects taking O reported a pain score significantly lower than subjects taking C (0.4 faces, P = 0.01). Minor adverse effects occurred in both groups, but itching occurred less in O subjects (odds ratio, 0.37; 95% confidence interval, 0.14-0.99). The most painful part of the visit was radiography (O = 41%, C = 38%) followed by extremity examination (O = 16%, C = 13%) then casting (O = 8%, C = 13%).
Triage-administered O tended toward greater pain reduction compared with C in children with suspected forearm fractures. Although minor adverse effects occurred in both groups, itching occurred more in C. Identification of radiography as the most painful part of fracture evaluation underscores the need for early triage administration of analgesia for suspected fractures.
Pediatric emergency care 10/2008; 24(9):595-600. · 0.92 Impact Factor
-
Clinical Infectious Diseases 09/2008; 47(3):430-1. · 9.15 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Mortality attributable to bloodstream infection (BSI) is still controversial. We studied the impact of BSI on mortality after coronary artery bypass surgery, including the specific impact of different etiologic organisms.
Our cohort consisted of 4515 patients who underwent coronary artery bypass procedures at a university hospital from 1996 through 2004. We used Society of Thoracic Surgery data supplemented with laboratory and infection control data. Mortality dates were identified using Society of Thoracic Surgery data and the Social Security Death Index. BSI within 90 days after surgery was defined by a positive blood culture result. Cox proportional hazards and propensity score models were used to analyze the association between BSI and mortality.
Patients with BSI had a 4.2-fold increased risk of death (95% confidence interval [CI], 3.0-5.9) 2-90 days after coronary artery bypass surgery, compared with uninfected patients. The risk of death was higher among patients with BSI due to gram-negative bacteria (hazard ratio [HR], 6.8; 95% CI, 3.9-12.0) and BSI due to Staphylococcus aureus (HR, 7.2; 95% CI, 3.3-15.7) and lowest among patients with BSI caused by gram-positive bacteria other than S. aureus (HR, 2.2; 95% CI, 1.1-4.6). The risk of death was highest among patients who developed BSI but had the lowest likelihood of infection (HR, 10.0; 95% CI, 3.5-28.8) and was lowest among patients who developed BSI but had the highest likelihood of infection (HR, 2.3; 95% CI, 1.2-4.6).
BSIs due to gram-negative bacteria and BSIs due to S. aureus contributed significantly to mortality. Mortality attributable to BSI was highest among patients predicted to be least likely to develop infection and was lowest among severely ill patients who were most likely to develop infection. BSI appears to be an important contributor to death after coronary artery bypass surgery, particularly among the healthiest patients.
Clinical Infectious Diseases 06/2008; 46(10):1537-46. · 9.15 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: The relationship between US Medical Licensing Examination Step 1 scores and core surgical-specialty match outcomes has not been well defined.
With IRB approval, we measured associations between aggregate Step 1 scores and other specialty-specific, match-process variables for 3 surgical-specialty matches. Chi-square tests measured differences between proportions of US students and independent applicants (ie, all non-US allopathic student applicants) who matched. Independent samples t-tests compared differences in Step 1 scores between matched- and unmatched-applicant groups. Pearson correlations measured the magnitude and direction of associations between matched-applicants' Step 1 scores and other variables of interest and between Step 1 scores for all match participants and percentage of positions filled by US students (two-tailed p values).
Step 1 scores were lower for unmatched- than matched-applicant groups for each specialty examined (each p < 0.0001). Matched-applicant groups' Step 1 scores positively correlated with each unmatched-applicant groups' Step 1 scores (r =.82, p < 0.0001), Step 1 gap between matched- and unmatched-applicant groups' scores (r = .40, p = 0.035), percentage of positions filled by US students (r = .62, p < 0.0001), and mean number of applications filed/applicant (r = .50, p < 0.0001). Step 1 scores for all match participants correlated with percentage of positions filled by US students (r = .61, p = 0.0006).
Step 1 scores were closely related to match process outcomes and match participation itself, with increasing Step 1 scores among both matched- and unmatched-applicant groups as specialty selectivity increased.
Journal of the American College of Surgeons 03/2008; 206(3):533-9. · 4.55 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Receiver operating characteristic (ROC) methodology is widely used to evaluate diagnostic tests. It is not uncommon in medical practice that multiple diagnostic tests are applied to the same study sample. A va-riety of methods have been proposed to combine such potentially correlated tests to increase the diagnostic accuracy. Usually the optimum combina-tion is searched based on the area under a ROC curve (AUC), an overall summary statistics that measures the distance between the distributions of diseased and non-diseased populations. For many clinical practitioners, however, a more relevant question of interest may be "what the sensitivity would be for a given specificity (say, 90%) or what the specificity would be for a given sensitivity?". Generally there is no unique linear combination superior to all others over the entire range of specificities or sensitivities. Under the framework of a ROC curve, in this paper we presented a method to estimate an optimum linear combination maximizing sensitivity at a fixed specificity while assuming a multivariate normal distribution in diagnostic tests. The method was applied to a real-world study where the accuracy of two biomarkers was evaluated in the diagnosis of pancreatic cancer. The performance of the method was also evaluated by simulation studies.
Journal of Data Science. 01/2008; 6:1-13.
-
[show abstract]
[hide abstract]
ABSTRACT: Previous studies of risk factors for Clostridium difficile-associated disease (CDAD) have been limited by small sample sizes and case-control study designs. Many of these studies were performed during outbreaks of CDAD. Colonization pressure and use of fluoroquinolones, vancomycin, and gastric acid suppressors have not been fully evaluated as risk factors for CDAD. The purpose of this study was to determine risk factors for endemic CDAD, including CDAD pressure, a modified version of colonization pressure.
We performed a retrospective cohort study of 36,086 patients admitted to Barnes-Jewish Hospital (St. Louis, MO) during the period from 1 January 2003 through 31 December 2003. Administrative, laboratory, and pharmacy data were collected from electronic hospital databases. Colonization pressure was measured through a surrogate variable (i.e., CDAD pressure). Multivariable pooled logistic regression models were used to evaluate independent risk factors for CDAD.
The analysis included 382 CDAD case patient admissions and 35,704 non-case patient admissions. Significant independent risk factors for CDAD included increasing age, admission(s) in the previous 60 days, hypoalbuminemia, leukemia and/or lymphoma, mechanical ventilation, and receipt of antimotility drugs, histamine-2 blockers, proton pump inhibitors, intravenous vancomycin, fluoroquinolones, and first-, third-, or fourth-generation cephalosporins. Increasing CDAD pressure was a strong risk factor for CDAD (for a CDAD pressure >1.4, the odds ratio was 4.0; 95% confidence interval, 2.9-5.6). Receipt of metronidazole was protective against CDAD (odds ratio, 0.5; 95% confidence interval, 0.3-0.6).
This study identified the previously underrecognized CDAD risk factors of CDAD pressure and vancomycin. More studies are needed to evaluate the relationship between CDAD, these risk factors, and use of gastric acid suppressors and fluoroquinolones.
Clinical Infectious Diseases 12/2007; 45(12):1543-9. · 9.15 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: In an effort to examine a cluster of late-stage breast cancer, this study reports (1) the extent of geographic variation in adequacy of diagnostic follow-up (ADFU) after abnormal breast cancer screening results across Missouri's counties and census tracts, (2) whether various personal characteristics or area poverty account for any geographic clustering observed, and (3) the association between area poverty rate and ADFU.
We used 1998-2002 Missouri Show Me Healthy Women breast and cervical cancer program data from 2580 low-income women aged 50-64 who had abnormal breast cancer screening results. ADFU was based on established guidelines. Poverty rate was from the 2000 census data. We used 3 complementary statistical approaches.
Overall, 26.9% of screening results were inadequately followed up. County-level geographic variation accounted for 6.7% of the total variance in ADFU, while the census-tract-level variation was negligible. Women's sociodemographic characteristics, symptoms reported at time of screening, and screening results accounted for 25% of the county-level variation in ADFU. Statistically significant geographic variation in ADFU remained that could not be explained. Beyond 70 miles from the women's residence, the likelihood of receiving ADFU was geographically uncorrelated. We identified one large geographic cluster extending beyond the borders of counties and census tracts where women were less likely to receive ADFU (relative risk = 0.64; p = 0.01).
Efforts to improve the likelihood of ADFU should be directed at examining the relative contributions of the healthcare and social environments and characteristics of the women in the area where women were less likely to receive ADFU especially in the cluster area of late-stage breast cancer rather than targeting efforts at the county or census-tract level.
Annals of Epidemiology 10/2007; 17(9):704-12. · 3.21 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Many medical diagnostic studies involve three ordinal diagnostic groups in which the diagnostic accuracy can be summarized by the volume or partial volume under a Receiver Operating Characteristic (ROC) surface. We study in this paper the statistical comparison of diagnostic accuracy from multiple diagnostic tests when three ordinal diagnostic groups are involved. Under the assumption that the multiple diagnostic tests follow a multivariate normal distribution within each diagnostic group, we provide the asymptotic variance and covariance for the maximum likelihood estimates of the volumes under the ROC surfaces from multiple diagnostic tests and propose statistical tests to test whether the diagnostic accuracy as measured by the volume under the ROC surface is the same for multiple diagnostic tests. We also propose a confidence interval estimate to the difference of two volumes under two ROC surfaces. Our approach depends crucially on the assumptions of normal distributions on diagnostic tests, which might not be robust when such assumptions are violated. Finally, we apply our proposed methodology to a real data set of 118 subjects to compare the diagnostic accuracy of early stage Alzheimer's disease (AD) from multiple neuropsychological tests.
Biometrical Journal 09/2007; 49(5):682-93. · 1.25 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: The authors examined the associations of observed neighborhood (block face) and housing conditions with the incidence of diabetes by using data from 644 subjects in the African-American Health Study (St. Louis area, Missouri). They also investigated five mediating pathways (health behavior, psychosocial, health status, access to medical care, and sociodemographic characteristics) if significant associations were identified. The external appearance of the block the subjects lived on and housing conditions were rated as excellent, good, fair, or poor. Subjects reported about neighborhood desirability. Self-reported diabetes was obtained at baseline and 3 years later. Of 644 subjects without self-reported diabetes, 10.3% reported having diabetes at the 3-year follow-up. Every housing condition rated as fair-poor was associated with an increased risk of diabetes, with odds ratios ranging from 2.53 (95% confidence interval: 1.47, 4.34 for physical condition inside the building) to 1.78 (95% confidence interval: 1.03, 3.07 for cleanliness inside the building) in unadjusted analyses. No association was found between any of the block face conditions or perceived neighborhood conditions and incident diabetes. The odds ratios for the five housing conditions were unaffected when adjusted for the mediating pathways. Poor housing conditions appear to be an independent contributor to the risk of incident diabetes in urban, middle-aged African Americans.
American Journal of Epidemiology 09/2007; 166(4):379-87. · 5.22 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Radiotherapy following breast-conserving surgery for the treatment of first primary breast cancer is the standard of care and is widely used despite its small survival benefit. The effects of radiotherapy in metachronous contralateral breast cancer are unknown. We examined the use of radiotherapy and its effect on cause-specific and all-cause mortality among women with metachronous contralateral breast cancer treated with breast-conserving surgery in community settings. Using data from the 1985-2000 Surveillance, Epidemiology, and End Results program, we identified women with stage 0-III metachronous contralateral breast cancer that occurred at least six months after stage 0-III first primary breast cancer. Cause-specific and all-cause mortality of women age 40-69 who did and who did not receive radiotherapy following breast-conserving surgery for metachronous contralateral breast cancer were compared in proportional hazard models using propensity scores to balance covariates by radiotherapy use. We adjusted for misclassification of radiotherapy use. Based on misclassification-corrected analyses, 43.2 percent of 1,083 women with metachronous contralateral breast cancer did not receive radiotherapy after BCS. After adjustment for propensity scores and radiotherapy misclassification, women who did not receive radiotherapy had 2.2 times greater risk of cause-specific and 1.7 times greater risk of all-cause mortality. In community settings, a high percentage of women with stage 0-III metachronous contralateral breast cancer did not receive radiotherapy following breast-conserving surgery. Unlike the small survival benefit of radiotherapy after first primary breast cancer, omission of radiotherapy after metachronous contralateral breast cancer significantly increased the risk of cause-specific and all-cause mortality.
Breast Cancer Research and Treatment 06/2007; 103(1):77-83. · 4.43 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: We examined the geographic bias of four methods of geocoding addresses using ArcGIS, commercial firm, SAS/GIS, and aerial photography. We compared "point-in-polygon" (ArcGIS, commercial firm, and aerial photography) and the "look-up table" method (SAS/GIS) to allocate addresses to census geography, particularly as it relates to census-based poverty rates.
We randomly selected 299 addresses of children treated for asthma at an urban emergency department (1999-2001). The coordinates of the building address side door were obtained by constant offset based on ArcGIS and a commercial firm and true ground location based on aerial photography.
Coordinates were available for 261 addresses across all methods. For 24% to 30% of geocoded road/door coordinates the positional error was 51 meters or greater, which was similar across geocoding methods. The mean bearing was -26.8 degrees for the vector of coordinates based on aerial photography and ArcGIS and 8.5 degrees for the vector based on aerial photography and the commercial firm (p < 0.0001). ArcGIS and the commercial firm performed very well relative to SAS/GIS in terms of allocation to census geography. For 20%, the door location based on aerial photography was assigned to a different block group compared to SAS/GIS. The block group poverty rate varied at least two standard deviations for 6% to 7% of addresses.
We found important differences in distance and bearing between geocoding relative to aerial photography. Allocation of locations based on aerial photography to census-based geographic areas could lead to substantial errors.
Annals of Epidemiology 06/2007; 17(6):464-70. · 3.21 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: When the efficacy of a treatment in a randomized controlled trial is required for multiple primary endpoints, trial design and analysis differ from trial requiring efficacy in only one of the multiple endpoints.
We consider a two-arm clinical trial requiring efficacy analysis for multiple primary endpoints, formulating the appropriate null and alternative hypotheses for the test of treatment efficacy. We study the significance level/statistical power of an intersection-union test (IUT) in this situation. We compare IUT with the intuitive approach (selecting the maximum sample size over those obtained from testing individual primary endpoints one by one) for determination of sample size.
The proposed IUT reserves the same Type I error rate as shared by all endpoint-specific tests. The statistical power of the proposed IUT is no more than the minimum from the individual tests. The maximum sample size from multiple endpoint-specific tests is often inadequate for the test of treatment efficacy, especially when the standardized effect sizes are similar. Finally, the IUT can be applied to Alzheimer's disease treatment trials in which two primary endpoints are typically used.
The IUT is a valid method for use in the design and analysis of clinical trials requiring efficacy at multiple primary endpoints.
Clinical Trials 02/2005; 2(5):387-93. · 1.92 Impact Factor
