Predicting childhood sexual or physical abuse: a logistic regression geo-mapping approach to prevention.

This study investigates the degree to which gender, ethnicity, relationship to perpetrator, and geomapped socio-economic factors significantly predict the incidence of childhood sexual abuse, physical abuse and non- abuse. These variables are then linked to geographic identifiers using geographic information system (GIS) technology to develop a geo-mapping framework for child sexual and physical abuse prevention.

Logistic regression models for ordinal response: A study of self-efficacy in colorectal cancer screening

The ordinal logistic regression models are used to analyze the dependant variable with multiple outcomes that can be ranked, but have been underutilized. In this study, we describe four logistic regression models for analyzing the ordinal response variable. ^ In this methodological study, the four regression models are proposed. The first model uses the multinomial logistic model. The second is adjacent-category logit model. The third is the proportional odds model and the fourth model is the continuation-ratio model. We illustrate and compare the fit of these models using data from the survey designed by the University of Texas, School of Public Health research project PCCaSO (Promoting Colon Cancer Screening in people 50 and Over), to study the patient’s confidence in the completion colorectal cancer screening (CRCS). ^ The purpose of this study is two fold: first, to provide a synthesized review of models for analyzing data with ordinal response, and second, to evaluate their usefulness in epidemiological research, with particular emphasis on model formulation, interpretation of model coefficients, and their implications. Four ordinal logistic models that are used in this study include (1) Multinomial logistic model, (2) Adjacent-category logistic model [9], (3) Continuation-ratio logistic model [10], (4) Proportional logistic model [11]. We recommend that the analyst performs (1) goodness-of-fit tests, (2) sensitivity analysis by fitting and comparing different models.^

An ordinal logistic regression model with misclassification of the outcome variable and categorical covariate

Ordinal outcomes are frequently employed in diagnosis and clinical trials. Clinical trials of Alzheimer's disease (AD) treatments are a case in point using the status of mild, moderate or severe disease as outcome measures. As in many other outcome oriented studies, the disease status may be misclassified. This study estimates the extent of misclassification in an ordinal outcome such as disease status. Also, this study estimates the extent of misclassification of a predictor variable such as genotype status. An ordinal logistic regression model is commonly used to model the relationship between disease status, the effect of treatment, and other predictive factors. A simulation study was done. First, data based on a set of hypothetical parameters and hypothetical rates of misclassification was created. Next, the maximum likelihood method was employed to generate likelihood equations accounting for misclassification. The Nelder-Mead Simplex method was used to solve for the misclassification and model parameters. Finally, this method was applied to an AD dataset to detect the amount of misclassification present. The estimates of the ordinal regression model parameters were close to the hypothetical parameters. β1 was hypothesized at 0.50 and the mean estimate was 0.488, β2 was hypothesized at 0.04 and the mean of the estimates was 0.04. Although the estimates for the rates of misclassification of X1 were not as close as β1 and β2, they validate this method. X 1 0-1 misclassification was hypothesized as 2.98% and the mean of the simulated estimates was 1.54% and, in the best case, the misclassification of k from high to medium was hypothesized at 4.87% and had a sample mean of 3.62%. In the AD dataset, the estimate for the odds ratio of X 1 of having both copies of the APOE 4 allele changed from an estimate of 1.377 to an estimate 1.418, demonstrating that the estimates of the odds ratio changed when the analysis includes adjustment for misclassification. ^

Logistic regression with incompletely observed binary

Logistic regression is one of the most important tools in the analysis of epidemiological and clinical data. Such data often contain missing values for one or more variables. Common practice is to eliminate all individuals for whom any information is missing. This deletion approach does not make efficient use of available information and often introduces bias.^ Two methods were developed to estimate logistic regression coefficients for mixed dichotomous and continuous covariates including partially observed binary covariates. The data were assumed missing at random (MAR). One method (PD) used predictive distribution as weight to calculate the average of the logistic regressions performing on all possible values of missing observations, and the second method (RS) used a variant of resampling technique. Additional seven methods were compared with these two approaches in a simulation study. They are: (1) Analysis based on only the complete cases, (2) Substituting the mean of the observed values for the missing value, (3) An imputation technique based on the proportions of observed data, (4) Regressing the partially observed covariates on the remaining continuous covariates, (5) Regressing the partially observed covariates on the remaining continuous covariates conditional on response variable, (6) Regressing the partially observed covariates on the remaining continuous covariates and response variable, and (7) EM algorithm. Both proposed methods showed smaller standard errors (s.e.) for the coefficient involving the partially observed covariate and for the other coefficients as well. However, both methods, especially PD, are computationally demanding; thus for analysis of large data sets with partially observed covariates, further refinement of these approaches is needed. ^

POLYCHOTOMOUS LOGISTIC REGRESSION ANALYSIS (HYPERTENSION, CORONARY HEART DISEASE)

The history of the logistic function since its introduction in 1838 is reviewed, and the logistic model for a polychotomous response variable is presented with a discussion of the assumptions involved in its derivation and use. Following this, the maximum likelihood estimators for the model parameters are derived along with a Newton-Raphson iterative procedure for evaluation. A rigorous mathematical derivation of the limiting distribution of the maximum likelihood estimators is then presented using a characteristic function approach. An appendix with theorems on the asymptotic normality of sample sums when the observations are not identically distributed, with proofs, supports the presentation on asymptotic properties of the maximum likelihood estimators. Finally, two applications of the model are presented using data from the Hypertension Detection and Follow-up Program, a prospective, population-based, randomized trial of treatment for hypertension. The first application compares the risk of five-year mortality from cardiovascular causes with that from noncardiovascular causes; the second application compares risk factors for fatal or nonfatal coronary heart disease with those for fatal or nonfatal stroke. ^

Logistic regression with Markov chains as covariates

The tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) is an obvious carcinogen for lung cancer. Since CBMN (Cytokinesis-blocked micronucleus) has been found to be extremely sensitive to NNK-induced genetic damage, it is a potential important factor to predict the lung cancer risk. However, the association between lung cancer and NNK-induced genetic damage measured by CBMN assay has not been rigorously examined. ^ This research develops a methodology to model the chromosomal changes under NNK-induced genetic damage in a logistic regression framework in order to predict the occurrence of lung cancer. Since these chromosomal changes were usually not observed very long due to laboratory cost and time, a resampling technique was applied to generate the Markov chain of the normal and the damaged cell for each individual. A joint likelihood between the resampled Markov chains and the logistic regression model including transition probabilities of this chain as covariates was established. The Maximum likelihood estimation was applied to carry on the statistical test for comparison. The ability of this approach to increase discriminating power to predict lung cancer was compared to a baseline "non-genetic" model. ^ Our method offered an option to understand the association between the dynamic cell information and lung cancer. Our study indicated the extent of DNA damage/non-damage using the CBMN assay provides critical information that impacts public health studies of lung cancer risk. This novel statistical method could simultaneously estimate the process of DNA damage/non-damage and its relationship with lung cancer for each individual.^

Evaluation of the Hosmer -Lemeshow global goodness -of-fit statistic for mixed variable logistic regression models

The performance of the Hosmer-Lemeshow global goodness-of-fit statistic for logistic regression models was explored in a wide variety of conditions not previously fully investigated. Computer simulations, each consisting of 500 regression models, were run to assess the statistic in 23 different situations. The items which varied among the situations included the number of observations used in each regression, the number of covariates, the degree of dependence among the covariates, the combinations of continuous and discrete variables, and the generation of the values of the dependent variable for model fit or lack of fit.^ The study found that the $\rm\ C$g* statistic was adequate in tests of significance for most situations. However, when testing data which deviate from a logistic model, the statistic has low power to detect such deviation. Although grouping of the estimated probabilities into quantiles from 8 to 30 was studied, the deciles of risk approach was generally sufficient. Subdividing the estimated probabilities into more than 10 quantiles when there are many covariates in the model is not necessary, despite theoretical reasons which suggest otherwise. Because it does not follow a X$\sp2$ distribution, the statistic is not recommended for use in models containing only categorical variables with a limited number of covariate patterns.^ The statistic performed adequately when there were at least 10 observations per quantile. Large numbers of observations per quantile did not lead to incorrect conclusions that the model did not fit the data when it actually did. However, the statistic failed to detect lack of fit when it existed and should be supplemented with further tests for the influence of individual observations. Careful examination of the parameter estimates is also essential since the statistic did not perform as desired when there was moderate to severe collinearity among covariates.^ Two methods studied for handling tied values of the estimated probabilities made only a slight difference in conclusions about model fit. Neither method split observations with identical probabilities into different quantiles. Approaches which create equal size groups by separating ties should be avoided. ^

GSTM1 and GSTT1 null polymorphisms and risk of salivary gland carcinoma.

Glutathione S-transferase (GST) genes detoxify and metabolize carcinogens, including oxygen free radicals which may contribute to salivary gland carcinogenesis. This cancer center-based case-control association study included 166 patients with incident salivary gland carcinoma (SGC) and 511 cancer-free controls. We performed multiplex polymerase chain reaction-based polymorphism genotyping assays for GSTM1 and GSTT1 null genotypes. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated with multivariable logistic regression analyses adjusted for age, sex, ethnicity, tobacco use, family history of cancer, alcohol use and radiation exposure. In our results, 27.7% of the SGC cases and 20.6% of the controls were null for the GSTT1 (P = 0.054), and 53.0% of the SGC cases and 50.9% of the controls were null for the GSTM1 (P = 0.633). The results of the adjusted multivariale regression analysis suggested that having GSTT1 null genotype was associated with a significantly increased risk for SGC (odds ratio 1.5, 95% confidence interval 1.0-2.3). Additionally, 13.9% of the SGC cases but only 8.4% of the controls were null for both genes and the results of the adjusted multivariable regression analysis suggested that having both null genotypes was significantly associated with an approximately 2-fold increased risk for SGC (odds ratio 1.9, 95% confidence interval 1.0-3.5). The presence of GSTT1 null genotype and the simultaneous presence of GSTM1 and GSTT1 null genotypes appear associated with significantly increased SGC risk. These findings warrant further study with larger sample sizes.

Time to renal disease and end-stage renal disease in PROFILE: a multiethnic lupus cohort.

BACKGROUND: Renal involvement is a serious manifestation of systemic lupus erythematosus (SLE); it may portend a poor prognosis as it may lead to end-stage renal disease (ESRD). The purpose of this study was to determine the factors predicting the development of renal involvement and its progression to ESRD in a multi-ethnic SLE cohort (PROFILE). METHODS AND FINDINGS: PROFILE includes SLE patients from five different United States institutions. We examined at baseline the socioeconomic-demographic, clinical, and genetic variables associated with the development of renal involvement and its progression to ESRD by univariable and multivariable Cox proportional hazards regression analyses. Analyses of onset of renal involvement included only patients with renal involvement after SLE diagnosis (n = 229). Analyses of ESRD included all patients, regardless of whether renal involvement occurred before, at, or after SLE diagnosis (34 of 438 patients). In addition, we performed a multivariable logistic regression analysis of the variables associated with the development of renal involvement at any time during the course of SLE.In the time-dependent multivariable analysis, patients developing renal involvement were more likely to have more American College of Rheumatology criteria for SLE, and to be younger, hypertensive, and of African-American or Hispanic (from Texas) ethnicity. Alternative regression models were consistent with these results. In addition to greater accrued disease damage (renal damage excluded), younger age, and Hispanic ethnicity (from Texas), homozygosity for the valine allele of FcgammaRIIIa (FCGR3A*GG) was a significant predictor of ESRD. Results from the multivariable logistic regression model that included all cases of renal involvement were consistent with those from the Cox model. CONCLUSIONS: Fcgamma receptor genotype is a risk factor for progression of renal disease to ESRD. Since the frequency distribution of FCGR3A alleles does not vary significantly among the ethnic groups studied, the additional factors underlying the ethnic disparities in renal disease progression remain to be elucidated.

Notification of abnormal lab test results in an electronic medical record: do any safety concerns remain?

BACKGROUND: Follow-up of abnormal outpatient laboratory test results is a major patient safety concern. Electronic medical records can potentially address this concern through automated notification. We examined whether automated notifications of abnormal laboratory results (alerts) in an integrated electronic medical record resulted in timely follow-up actions. METHODS: We studied 4 alerts: hemoglobin A1c > or =15%, positive hepatitis C antibody, prostate-specific antigen > or =15 ng/mL, and thyroid-stimulating hormone > or =15 mIU/L. An alert tracking system determined whether the alert was acknowledged (ie, provider clicked on and opened the message) within 2 weeks of transmission; acknowledged alerts were considered read. Within 30 days of result transmission, record review and provider contact determined follow-up actions (eg, patient contact, treatment). Multivariable logistic regression models analyzed predictors for lack of timely follow-up. RESULTS: Between May and December 2008, 78,158 tests (hemoglobin A1c, hepatitis C antibody, thyroid-stimulating hormone, and prostate-specific antigen) were performed, of which 1163 (1.48%) were transmitted as alerts; 10.2% of these (119/1163) were unacknowledged. Timely follow-up was lacking in 79 (6.8%), and was statistically not different for acknowledged and unacknowledged alerts (6.4% vs 10.1%; P =.13). Of 1163 alerts, 202 (17.4%) arose from unnecessarily ordered (redundant) tests. Alerts for a new versus known diagnosis were more likely to lack timely follow-up (odds ratio 7.35; 95% confidence interval, 4.16-12.97), whereas alerts related to redundant tests were less likely to lack timely follow-up (odds ratio 0.24; 95% confidence interval, 0.07-0.84). CONCLUSIONS: Safety concerns related to timely patient follow-up remain despite automated notification of non-life-threatening abnormal laboratory results in the outpatient setting.

Timely follow-up of abnormal diagnostic imaging test results in an outpatient setting: are electronic medical records achieving their potential?

BACKGROUND: Given the fragmentation of outpatient care, timely follow-up of abnormal diagnostic imaging results remains a challenge. We hypothesized that an electronic medical record (EMR) that facilitates the transmission and availability of critical imaging results through either automated notification (alerting) or direct access to the primary report would eliminate this problem. METHODS: We studied critical imaging alert notifications in the outpatient setting of a tertiary care Department of Veterans Affairs facility from November 2007 to June 2008. Tracking software determined whether the alert was acknowledged (ie, health care practitioner/provider [HCP] opened the message for viewing) within 2 weeks of transmission; acknowledged alerts were considered read. We reviewed medical records and contacted HCPs to determine timely follow-up actions (eg, ordering a follow-up test or consultation) within 4 weeks of transmission. Multivariable logistic regression models accounting for clustering effect by HCPs analyzed predictors for 2 outcomes: lack of acknowledgment and lack of timely follow-up. RESULTS: Of 123 638 studies (including radiographs, computed tomographic scans, ultrasonograms, magnetic resonance images, and mammograms), 1196 images (0.97%) generated alerts; 217 (18.1%) of these were unacknowledged. Alerts had a higher risk of being unacknowledged when the ordering HCPs were trainees (odds ratio [OR], 5.58; 95% confidence interval [CI], 2.86-10.89) and when dual-alert (>1 HCP alerted) as opposed to single-alert communication was used (OR, 2.02; 95% CI, 1.22-3.36). Timely follow-up was lacking in 92 (7.7% of all alerts) and was similar for acknowledged and unacknowledged alerts (7.3% vs 9.7%; P = .22). Risk for lack of timely follow-up was higher with dual-alert communication (OR, 1.99; 95% CI, 1.06-3.48) but lower when additional verbal communication was used by the radiologist (OR, 0.12; 95% CI, 0.04-0.38). Nearly all abnormal results lacking timely follow-up at 4 weeks were eventually found to have measurable clinical impact in terms of further diagnostic testing or treatment. CONCLUSIONS: Critical imaging results may not receive timely follow-up actions even when HCPs receive and read results in an advanced, integrated electronic medical record system. A multidisciplinary approach is needed to improve patient safety in this area.

Ethnic differences in do-not-resuscitate orders after intracerebral hemorrhage.

OBJECTIVE: To explore ethnic differences in do-not-resuscitate orders after intracerebral hemorrhage. DESIGN: Population-based surveillance. SETTING: Corpus Christi, Texas. PATIENTS: All cases of intracerebral hemorrhage in the community of Corpus Christi, TX were ascertained as part of the Brain Attack Surveillance in Corpus Christi (BASIC) project. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Medical records were reviewed for do-not-resuscitate orders. Unadjusted and multivariable logistic regression were used to test for associations between ethnicity and do-not-resuscitate orders, both overall ("any do-not-resuscitate") and within 24 hrs of presentation ("early do-not-resuscitate"), adjusted for age, gender, Glasgow Coma Scale, intracerebral hemorrhage volume, intraventricular hemorrhage, infratentorial hemorrhage, modified Charlson Index, and admission from a nursing home. A total of 270 cases of intracerebral hemorrhage from 2000-2003 were analyzed. Mexican-Americans were younger and had a higher Glasgow Coma Scale than non-Hispanic whites. Mexican-Americans were half as likely as non-Hispanic whites to have early do-not-resuscitate orders in unadjusted analysis (odds ratio 0.45, 95% confidence interval 0.27, 0.75), although this association was not significant when adjusted for age (odds ratio 0.61, 95% confidence interval 0.35, 1.06) and in the fully adjusted model (odds ratio 0.75, 95% confidence interval 0.39, 1.46). Mexican-Americans were less likely than non-Hispanic whites to have do-not-resuscitate orders written at any time point (odds ratio 0.37, 95% confidence interval 0.23, 0.61). Adjustment for age alone attenuated this relationship although it retained significance (odds ratio 0.49, 95% confidence interval 0.29, 0.82). In the fully adjusted model, Mexican-Americans were less likely than non-Hispanic whites to use do-not-resuscitate orders at any time point, although the 95% confidence interval included one (odds ratio 0.52, 95% confidence interval 0.27, 1.00). CONCLUSIONS: Mexican-Americans were less likely than non-Hispanic whites to have do-not-resuscitate orders after intracerebral hemorrhage although the association was attenuated after adjustment for age and other confounders. The persistent trend toward less frequent use of do-not-resuscitate orders in Mexican-Americans suggests that further study is warranted.

Back pain in migrant farmworker families and farmworker high school students from Starr County, Texas

Back symptoms are a major global public health problem with the lifetime prevalence ranging between 50-80%. Research suggests that work-related factors contribute to the occurrence of back pain in various industries. Despite the hazardous nature, strenuous tasks, and awkward postures associated with farm work, little is known about back injury and symptoms in farmworker adults and children. Research in the United States is particularly limited. This is a concern given the large proportion of migrant farmworkers in the United States without adequate access to healthcare as well as a substantial number of youth working in agriculture. The present study describes back symptoms and identifies work-related factors associated with back pain in migrant farmworker families and farmworker high school students from Starr County, TX. Two separate datasets were used from two cohort studies "Injury and Illness Surveillance in Migrant Farmworkers (MANOS)" (study A: n=267 families) and "South Texas Adolescent Rural Research Study (STARRS)" (study B: n=345). Descriptive and inferential statistics including multivariable logistic regression were used to identify work-related factors associated with back pain in each study. In migrant farmworker families, the prevalence of chronic back pain during the last migration season ranged from 9.5% among youngest children to 33.3% among mothers. Chronic back pain was significantly associated with increasing age; fairly bad/very bad quality of sleep while migrating; fewer than eight hours of sleep at home in Starr County, TX; depressive symptoms while migrating; self-provided water for washing hands/drinking; weeding at work; and exposure to pesticide drift/direct spray. Among farmworker adolescents, the prevalence of severe back symptoms was 15.7%. Severe back symptoms were significantly associated with being female; history of a prior accident/back injury; feeling tense, stressed, or anxious sometimes/often; lifting/carrying heavy objects not at work; current tobacco use; increasing lifetime number of migrant farmworker years; working with/around knives; and working on corn crops. Overall, results support that associations between work-related exposures and chronic back pain and severe back symptoms remain after controlling for the effect of non-work exposures in farmworker populations. ^

Relationship between Clostridium difficile toxin type and clinical features, severity and outcome in patients with C. difficile diarrhea (CDAD)

Background. Clostridium difficile infection is one of the major causes of antibiotic associated diarrhea and colitis in the United States. Currently, there is a dearth of literature on the risk factors and outcomes differences between the patients with infection due to the hypervirulent strain vs. the non-hypervirulent strains. The objective of this study was to determine the relationship between C. difficile toxin type and clinical features, severity and outcome in patients with C. difficile diarrhea. ^ Methods. The case group included 37 patients who had infections due to hypervirulent strain (tcdC deletion) and the control group included 55 patients with other toxin types (toxin A, B, binary toxin). A univariate analysis was performed followed by a multivariable logistic regression analysis to assess the differences between cases and controls. ^ Results. In the multivariate analyses, we found out that being a male was a protective factor for developing the infection due to the hypervirulent strain [OR 0.33; 95% CI 0.12-0.90]. Also, the hypervirulent group has worse clinical and economic outcomes, although the differences were small and nonsignificant. ^ Conclusions. There may likely be no predictive risk factor for acquiring infection due to the hypervirulent strain and the acquisition may be more linked to the infection control practices of the individual hospitals or location of patients. Hence, better infection control practices may prove helpful in decreasing the overall disease burden and thus improve patient outcomes. ^

Racial/ethnic differences in factors influencing screening and treatment among prostate cancer patients

Background. Racial/ethnic differences have been found in various aspects of cancer care. But a limited number of studies have examined the racial/ethnic differences in predictors of prostate-specific antigen (PSA) screening in a group of prostate cancer patients and have attempted to identify the racial/ethnic differences in treatment discussions, treatment choice and treatment received for organ-confined localized prostate cancer (PCa) among three major racial/ethnic groups of the USA. This study was conducted to redress this lack of information. ^ Methods. This study was conducted on a group of 935 prostate cancer patients representing all three major race/ethnic groups (Whites, African Americans and Hispanics) who were treated at various medical institutes of the Texas Medical Center, Houston between 1996 and 2004 to identify the racial/ethnic differences in predictors of PSA screening. A subset of 640 patients who had organ-confined localized prostate cancer was selected to examine the racial/ethnic differences in treatment discussions, treatment choice and treatment received for their localized prostate cancer. They were interviewed by trained research interviewers of MD Anderson Cancer Center using a validated structured questionnaire. ^ Results. The results showed that African American (54.4%) and Hispanic patients (42.3%) were significantly less likely (p=0.004 and p<.001, respectively) than White patients (63.2%) to report having had PSA screening before their prostate-cancer diagnosis. Among Whites, only education and annual check-ups predicted the use of PSA screening, whereas in African Americans two more additional factors, marital status and bode-mass index (BMI), significantly predicted PSA screening. Among Hispanics, like two other groups, education and annual check-ups also appeared as a significant predictor of PSA screening. ^ Results from multivariable logistic regression showed that African American patients were 15% less likely (OR=0.85, 95% CI=0.61-1.17, p=0.32) and Hispanics patients were 40% less likely (OR=0.60, 95% CI=0.41-0.87, p=0.008) to undergo PSA screening than Whites after adjusting for education and age at diagnosis for African Americans, and for education, annual check-ups and age at diagnosis for Hispanics. ^ This study revealed that health professionals were less likely to discuss surgery (79.9% vs. 93.2%) and watchful waiting (27.9% vs. 43.9%) with Hispanics compared to Whites. African Americans were more likely to choose (35.1% vs. 27.7%) and receive radiation therapy (38.3% vs.31.4%) than Whites. A comparison of concordance between treatment choice and treatment received showed that the highest concordance was found for watchful waiting and radiation therapy among African Americans (100% and 85.9%, respectively) whereas the highest concordance (96.9%) was found for surgery among Hispanics. ^ Conclusions. In this multiethnic study, the rates of PSA screening and its potential predictors varied by racial/ethnic groups. Substantial racial/ethnic variations were also found in treatment discussion, but the differences were not evident for treatment choice and treatment received. Health-education programs and culturally appropriate educational outreach efforts, especially targeted for high-risk groups, are needed to reduce these disparities. In the current climate of uncertainty about the benefits of PSA screening, or the benefit of one treatment over others, men should have access to information and services regardless of race/ethnicity so that they can make informed decisions. Further in-depth studies are needed in other settings to confirm these findings with the goal of developing an intervention to address these concerns. ^