Illusory predictors: Generalizability of findings in cocaine treatment retention research.

Treatment retention is of paramount importance in cocaine treatment research as treatment completion rates are often 50% or less. Failure to retain cocaine patients in treatment has both significant research and clinical implications. In this paper we qualitatively and quantitatively demonstrate the inconsistency found across analyses of retention predictors in order to highlight the problem. First, a qualitative review of the published literature was undertaken to identify the frequency of predictors studied and their relations to treatment retention. Second, an empirical demonstration of predictor stability was conducted by testing a common set of variables across three similar 12-week cocaine clinical trials conducted by the same investigators in the same research clinic within a five-year period. Results of the literature review indicated inconsistently selected variables of convenience, widely varying statistical procedures, and discrepant findings of significance. Further, quantitative analyses resulted in discrepancies in variables identified as significant predictors of retention among the three studies. Potential sources of heterogeneity affecting the consistency of findings across studies and recommendations to improve the validity and generalizability of predictor findings in future studies are proposed.

An exploratory investigation of America's most costly delinquents: Juvenile firesetters

This study focused on the instruments that are currently being used by fire department personnel to identify and classify juvenile firesetters, these instruments, as published by the Federal Emergency Management Agency (F.E.M.A.) have never been empirically validated as to their ability to discriminate between first time and multiple firesetters and to predict the degree of risk for future firesetting by juveniles that come to the attention of authorities for firesetting behaviors. The study was descriptive in nature and not designed to test the validity of these instruments. The study was designed to test the ability of the instruments to discriminate between first time and multiple firesetters and to categorize known firesetters, based on the motive for firesetting, as to their degree or risk for future firesetting.^ The results suggest that the F.E.M.A. instruments are of little use in discriminating between first time and multiple firesetters. The F.E.M.A. instruments were not able to categorize juvenile firesetters as to their potential risk for future firesetting. A subset of variables from the F.E.M.A. instruments was identified that may be useful in discriminating between youth that are troubled firesetters and those that are not. ^

Anthropometric measures as a predictor of cephalopelvic disproportion

This study compared three body measurements, height, hip width (bitrochanteric) and foot length, in 120 Hispanic women who had their first birth by cesarean section (N = 60) or by spontaneous vaginal delivery (N = 60). The objective of the study was to see if there were differences in these measurements that could be useful in predicting cephalopelvic disproportion. Data were collected from two public hospitals in Houston Texas over a 10 month period from December 1994 to October 1995. The statistical technique used to evaluate the measures was discriminant analysis.^ Women who delivered by cesarean section were older, shorter, had shorter feet and delivered heavier infants. There were no differences in the bitrochanteric widths of the women or in the mean gestational age or Apgar scores of the infants.^ Significantly more of the mothers and infants were ill following cesarean section delivery. Maternal illness was usually infection; infant illness was primarily infection or respiratory difficulties.^ Discriminant analysis is a technique which allows for classification and prediction to which group a particular entity will belong given a certain set of variables. Using discriminant analysis, with a probability of cesarean section 50 percent, the best combination to classify who would have a cesarean section was height and hip width, correctly classifying 74.2 percent of those who needed surgery. When the probability of cesarean section was 10 percent and probability of vaginal delivery was 90 percent, the best predictor of who would need operative delivery was height, hip width and age, correctly classifying 56.2 percent. In the population from which the study participants were selected the incidence of cephalopelvic disproportion was low, approximately 1 percent.^ With the technologic assistance available in most of the developed world, it is likely that the further pursuit of different measures and their use would not be of much benefit in attempting to predict and diagnose disproportion. However, in areas of the world where much of obstetrics is "hands on", the availability of technology extremely limited, and the incidence of disproportion larger, the use of anthropometric measures might be useful and of some potential benefit. ^

An empirical evaluation of the Random Forests classifier models for variable selection in a large-scale lung cancer case-control study

Random Forests™ is reported to be one of the most accurate classification algorithms in complex data analysis. It shows excellent performance even when most predictors are noisy and the number of variables is much larger than the number of observations. In this thesis Random Forests was applied to a large-scale lung cancer case-control study. A novel way of automatically selecting prognostic factors was proposed. Also, synthetic positive control was used to validate Random Forests method. Throughout this study we showed that Random Forests can deal with large number of weak input variables without overfitting. It can account for non-additive interactions between these input variables. Random Forests can also be used for variable selection without being adversely affected by collinearities. ^ Random Forests can deal with the large-scale data sets without rigorous data preprocessing. It has robust variable importance ranking measure. Proposed is a novel variable selection method in context of Random Forests that uses the data noise level as the cut-off value to determine the subset of the important predictors. This new approach enhanced the ability of the Random Forests algorithm to automatically identify important predictors for complex data. The cut-off value can also be adjusted based on the results of the synthetic positive control experiments. ^ When the data set had high variables to observations ratio, Random Forests complemented the established logistic regression. This study suggested that Random Forests is recommended for such high dimensionality data. One can use Random Forests to select the important variables and then use logistic regression or Random Forests itself to estimate the effect size of the predictors and to classify new observations. ^ We also found that the mean decrease of accuracy is a more reliable variable ranking measurement than mean decrease of Gini. ^

Household health assessment and associated water sample analysis in the Miches watershed of the Dominican Republic

Background. Diarrhea and malnutrition are the leading causes of mortality for children age one to four in the Dominican Republic. Communities within the Miches watershed lack sanitation infrastructure and water purification systems, which increases the risk of exposure to water-borne pathogens. The purpose of this cross-sectional study was to analyze health information gathered through household interviews and to test water samples for the presence of diarrheagenic pathogens and antibiotic-resistant bacteria within the Miches watershed. Methods. Frequency counts and thematic analysis were used to investigate Human Health Survey responses and Fisher's exact test was used to determine correlation between water source and reported illness. Bacteria cultured from water samples were analyzed by Gram stain, real-time PCR, API® 20E biochemical identification, and for antibiotic resistance. Results. Community members reported concerns about water sources with respect to water quality, availability, and environmental contamination. Pathogenic strains of E. coli were present in the water samples. Drinking aquifer water was positively-correlated with reported stomach aches (p=0.04) while drinking from rivers or creeks was associated with the reported absence of “gripe” (cold or flu) (p=0.01). The lack of association between reported illnesses and water source for the majority of variables suggested that there were multiple vehicles of disease transmission. Antibiotic resistant bacteria were isolated from the water samples tested. Conclusions. The presence of pathogenic E. coli in water samples suggested that water is at least one route of transmission for diarrheagenic pathogens in the Miches watershed. The presence of antibiotic-resistant bacteria in the water samples may indicate the proliferation of resistance plasmids in the environment as a result of antibiotic overuse in human and animal populations and a lack of sanitation infrastructure. An intervention that targets areas of hygiene, sanitation, and water purification is recommended to limit human exposure to diarrheagenic pathogens and antibiotic-resistant organisms. ^

A retrospective analysis of patients referred for tuberculosis testing at Parkland Hospital, Dallas, TX

Tuberculosis (TB) is an infectious disease of great public health importance, particularly to institutions that provide health care to large numbers of TB patients such as Parkland Hospital in Dallas, TX. The purpose of this retrospective chart review was to analyze differences in TB positive and TB negative patients to better understand whether or not there were variables that could be utilized to develop a predictive model for use in the emergency department to reduce the overall number of suspected TB patients being sent to respiratory isolation for TB testing. This study included patients who presented to the Parkland Hospital emergency department between November 2006 and December 2007 and were isolated and tested for TB. Outcome of TB was defined as a positive sputum AFB test or a positive M. tuberculosis culture result. Data were collected utilizing the UT Southwestern Medical Center computerized database OACIS and included demographic information, TB risk factors, physical symptoms, and clinical results. Only two variables were significantly (P<0.05) related to TB outcome: dyspnea (shortness of breath) (P<0.001) and abnormal x-ray (P<0.001). Marginally significant variables included hemoptysis (P=0.06), weight loss (P=0.11), night sweats (P=0.20), history of homelessness or incarceration (P=0.15), and history of positive skin PPD (P=0.19). Using a combination of significant and marginally significant variables, a predictive model was designed which demonstrated a specificity of 24% and a sensitivity of 70%. In conclusion, a predictive model for TB outcome based on patients who presented to the Parkland Hospital emergency department between November 2006 and December 2007 was unsuccessful given the limited number of variables that differed significantly between TB positive and TB negative patients. It is suggested that a future prospective cohort study should be implemented to collect data on TB positive and TB negative patients. It may be possible that a more thorough prospective collection of data may lead to clearer comparisons between TB positive and TB negative patients and ultimately to the design of a more sensitive predictive model for TB outcome. ^

A study of nurses' perceptions of and participation in the resolution of treatment dilemmas for critically ill newborns

This study was an exploratory investigation of variables which are associated with neonatal intensive care nurses' perceptions of and participation in life-sustaining treatment decisions for critically ill newborns. The primary purpose of the research was to examine the extent to which assessment of infants' physical and mental prognoses, parents' preferences regarding treatment, and legal consequences of non-treatment influence nurses' recommendations about life-saving treatment decisions for handicapped newborns. Secondly, the research explored the extent and nature of nurses' reported participation in the resolution of treatment dilemmas for these critically ill newborns. The framework of the study draws upon the work of Crane (1977), Blum (1980), and Pearlman (1982) who have explored the sociological context of decision-making with critical care patients.^ Participants in the study were a volunteer sample of eighty-three registered nurses who were currently working in neonatal intensive care units in five large urban hospitals in Texas. Data were collected through the use of intensive interviews and case study questionnaires. Results from the study indicate that physical and mental prognoses as well as parent preferences and concerns about legal liability are related to nurses' treatment recommendations, but their levels of significance vary according to the type of handicapping condition and whether the treatment questions are posed in terms of initiating aggressive therapy or withdrawing aggressive therapy.^ The majority of nurses reported that the extent of their participation in formal decision-making regarding handicapped newborns was fairly minimal although they provide much of the definitive data used to make decisions by physicians and parents. There was substantial evidence that nurse respondents perceive their primary role as advocates for critically ill newborns, and believe that their involvement in the resolution of treatment dilemmas should be increased. ^

Selection bias and the cross-validation of regression models for prediction

Strategies are compared for the development of a linear regression model with stochastic (multivariate normal) regressor variables and the subsequent assessment of its predictive ability. Bias and mean squared error of four estimators of predictive performance are evaluated in simulated samples of 32 population correlation matrices. Models including all of the available predictors are compared with those obtained using selected subsets. The subset selection procedures investigated include two stopping rules, C$\sb{\rm p}$ and S$\sb{\rm p}$, each combined with an 'all possible subsets' or 'forward selection' of variables. The estimators of performance utilized include parametric (MSEP$\sb{\rm m}$) and non-parametric (PRESS) assessments in the entire sample, and two data splitting estimates restricted to a random or balanced (Snee's DUPLEX) 'validation' half sample. The simulations were performed as a designed experiment, with population correlation matrices representing a broad range of data structures.^ The techniques examined for subset selection do not generally result in improved predictions relative to the full model. Approaches using 'forward selection' result in slightly smaller prediction errors and less biased estimators of predictive accuracy than 'all possible subsets' approaches but no differences are detected between the performances of C$\sb{\rm p}$ and S$\sb{\rm p}$. In every case, prediction errors of models obtained by subset selection in either of the half splits exceed those obtained using all predictors and the entire sample.^ Only the random split estimator is conditionally (on $\\beta$) unbiased, however MSEP$\sb{\rm m}$ is unbiased on average and PRESS is nearly so in unselected (fixed form) models. When subset selection techniques are used, MSEP$\sb{\rm m}$ and PRESS always underestimate prediction errors, by as much as 27 percent (on average) in small samples. Despite their bias, the mean squared errors (MSE) of these estimators are at least 30 percent less than that of the unbiased random split estimator. The DUPLEX split estimator suffers from large MSE as well as bias, and seems of little value within the context of stochastic regressor variables.^ To maximize predictive accuracy while retaining a reliable estimate of that accuracy, it is recommended that the entire sample be used for model development, and a leave-one-out statistic (e.g. PRESS) be used for assessment. ^

THE RELATIONSHIP BETWEEN DEGREE OF BLOOD PRESSURE REDUCTION AND MORTALITY AMONG HYPERTENSIVES IN THE HYPERTENSION DETECTION AND FOLLOW-UP PROGRAM

The relationship between degree of diastolic blood pressure (DBP) reduction and mortality was examined among hypertensives, ages 30-69, in the Hypertension Detection and Follow-up Program (HDFP). The HDFP was a multi-center community-based trial, which followed 10,940 hypertensive participants for five years. One-year survival was required for inclusion in this investigation since the one-year annual visit was the first occasion where change in blood pressure could be measured on all participants. During the subsequent four years of follow-up on 10,052 participants, 568 deaths occurred. For levels of change in DBP and for categories of variables related to mortality, the crude mortality rate was calculated. Time-dependent life tables were also calculated so as to utilize available blood pressure data over time. In addition, the Cox life table regression model, extended to take into account both time-constant and time-dependent covariates, was used to examine the relationship change in blood pressure over time and mortality.^ The results of the time-dependent life table and time-dependent Cox life table regression analyses supported the existence of a quadratic function which modeled the relationship between DBP reduction and mortality, even after adjusting for other risk factors. The minimum mortality hazard ratio, based on a particular model, occurred at a DBP reduction of 22.6 mm Hg (standard error = 10.6) in the whole population and 8.5 mm Hg (standard error = 4.6) in the baseline DBP stratum 90-104. After this reduction, there was a small increase in the risk of death. There was not evidence of the quadratic function after fitting the same model using systolic blood pressure. Methodologic issues involved in studying a particular degree of blood pressure reduction were considered. The confidence interval around the change corresponding to the minimum hazard ratio was wide and the obtained blood pressure level should not be interpreted as a goal for treatment. Blood pressure reduction was attributed, not only to pharmacologic therapy, but also to regression to the mean, and to other unknown factors unrelated to treatment. Therefore, the surprising results of this study do not provide direct implications for treatment, but strongly suggest replication in other populations. ^

Age -period -cohort analysis: An extension of Cox's lifetable regression with time -dependent covariates

It is well known that an identification problem exists in the analysis of age-period-cohort data because of the relationship among the three factors (date of birth + age at death = date of death). There are numerous suggestions about how to analyze the data. No one solution has been satisfactory. The purpose of this study is to provide another analytic method by extending the Cox's lifetable regression model with time-dependent covariates. The new approach contains the following features: (1) It is based on the conditional maximum likelihood procedure using a proportional hazard function described by Cox (1972), treating the age factor as the underlying hazard to estimate the parameters for the cohort and period factors. (2) The model is flexible so that both the cohort and period factors can be treated as dummy or continuous variables, and the parameter estimations can be obtained for numerous combinations of variables as in a regression analysis. (3) The model is applicable even when the time period is unequally spaced.^ Two specific models are considered to illustrate the new approach and applied to the U.S. prostate cancer data. We find that there are significant differences between all cohorts and there is a significant period effect for both whites and nonwhites. The underlying hazard increases exponentially with age indicating that old people have much higher risk than young people. A log transformation of relative risk shows that the prostate cancer risk declined in recent cohorts for both models. However, prostate cancer risk declined 5 cohorts (25 years) earlier for whites than for nonwhites under the period factor model (0 0 0 1 1 1 1). These latter results are similar to the previous study by Holford (1983).^ The new approach offers a general method to analyze the age-period-cohort data without using any arbitrary constraint in the model. ^

Critical indicators associated with negative performance in a selectively manned unit

Introduction. Selectively manned units have a long, international history, both military and civilian. Some examples include SWAT teams, firefighters, the FBI, the DEA, the CIA, and military Special Operations. These special duty operators are individuals who perform a highly skilled and dangerous job in a unique environment. A significant amount of money is spent by the Department of Defense (DoD) and other federal agencies to recruit, select, train, equip and support these operators. When a critical incident or significant life event occurs, that jeopardizes an operator's performance; there can be heavy losses in terms of training, time, money, and potentially, lives. In order to limit the number of critical incidents, selection processes have been developed over time to “select out” those individuals most likely to perform below desired performance standards under pressure or stress and to "select in" those with the "right stuff". This study is part of a larger program evaluation to assess markers that identify whether a person will fail under the stresses in a selectively manned unit. The primary question of the study is whether there are indicators in the selection process that signify potential negative performance at a later date. ^ Methods. The population being studied included applicants to a selectively manned DoD organization between 1993 and 2001 as part of a unit assessment and selection process (A&S). Approximately 1900 A&S records were included in the analysis. Over this nine year period, seventy-two individuals were determined to have had a critical incident. A critical incident can come in the form of problems with the law, personal, behavioral or family problems, integrity issues, and skills deficit. Of the seventy-two individuals, fifty-four of these had full assessment data and subsequent supervisor performance ratings which assessed how an individual performed while on the job. This group was compared across a variety of variables including demographics and psychometric testing with a group of 178 individuals who did not have a critical incident and had been determined to be good performers with positive ratings by their supervisors.^ Results. In approximately 2004, an online pre-screen survey was developed in the hopes of preselecting out those individuals with items that would potentially make them ineligible for selection to this organization. This survey has aided the organization to increase its selection rates and save resources in the process. (Patterson, Howard Smith, & Fisher, Unit Assessment and Selection Project, 2008) When the same prescreen was used on the critical incident individuals, it was found that over 60% of the individuals would have been flagged as unacceptable. This would have saved the organization valuable resources and heartache.^ There were some subtle demographic differences between the two groups (i.e. those with critical incidents were almost twice as likely to be divorced compared with the positive performers). Upon comparison of Psychometric testing several items were noted to be different. The two groups were similar when their IQ levels were compared using the Multidimensional Aptitude Battery (MAB). When looking at the Minnesota Multiphasic Personality Inventory (MMPI), there appeared to be a difference on the MMPI Social Introversion; the Critical Incidence group scored somewhat higher. When analysis was done, the number of MMPI Critical Items between the two groups was similar as well. When scores on the NEO Personality Inventory (NEO) were compared, the critical incident individuals tended to score higher on Openness and on its subscales (Ideas, Actions, and Feelings). There was a positive correlation between Total Neuroticism T Score and number of MMPI critical items.^ Conclusions. This study shows that the current pre-screening process is working and would have saved the organization significant resources. ^ If one was to develop a profile of a candidate who potentially could suffer a critical incident and subsequently jeopardize the unit, mission and the safety of the public they would look like the following: either divorced or never married, score high on the MMPI in Social Introversion, score low on MMPI with an "excessive" amount of MMPI critical items; and finally scores high on the NEO Openness and subscales Ideas, Feelings, and Actions.^ Based on the results gleaned from the analysis in this study there seems to be several factors, within psychometric testing, that when taken together, will aid the evaluators in selecting only the highest quality operators in order to save resources and to help protect the public from unfortunate critical incidents which may adversely affect our health and safety.^

HEALTH CARE COST INCREASES IN METROPOLITAN AREAS

The research project is an extension of the economic theory to the health care field and health care research projects evaluating the influence of demand and supply variables upon medical care inflation. The research tests a model linking the demographic and socioeconomic characteristics of the population, its community case mix, and technology, the prices of goods and services other than medical care, the way its medical services are delivered and the health care resources available to its population to different utilization patterns which, consequently, lead to variations in health care prices among metropolitan areas. The research considers the relationship of changes in community characteristics and resources and medical care inflation.^ The rapidly increasing costs of medical care have been of great concern to the general public, medical profession, and political bodies. Research and analysis of the main factors responsible for the rate of increase of medical care prices is necessary in order to devise appropriate solutions to cope with the problem. An understanding of the community characteristics and resources-medical care costs relationships in the metropolitan areas potentially offers guidance in individual plan and national policy development.^ The research considers 145 factors measuring community milieu (demographic, social, educational, economic, illness level, prices of goods and services other than medical care, hospital supply, physicians resources and techological factors). Through bivariate correlation analysis, the number of variables was reduced to a set of 1 to 4 variables for each cost equation. Two approaches were identified to track inflation in the health care industry. One approach measures costs of production which accounts for price and volume increases. The other approach measures price increases. One general and four specific measures were developed to represent each of the major approaches. The general measure considers the increase on medical care prices as a whole and the specific measures deal with hospital costs and physician's fees. The relationships among changes in community characteristics and resources and health care inflation were analyzed using bivariate correlation and regression analysis methods. It has been concluded that changes in community characteristics and resources are predictive of hospital costs and physician's fees inflation, but are not predictive of increases in medical care prices. These findings provide guidance in the formulation of public policy which could alter the trend of medical care inflation and in the allocation of limited Federal funds.^

INFECTION: ITS ECOLOGICAL DETERMINANTS AMONG CHILDREN IN NORTHERN CHILE

In developing countries, infection and malnutrition, and their interaction effects, account for the majority of childhood deaths and chronic deficits in growth and development. To promote child health, the causal determinants of infection and malnutrition and cost-effective interventions must be identified. To this end, medical examinations of 988 children (age two weeks to 14 years) living at three altitudes (coastal < 300m; sierra (TURN) 3,000m; and altiplano > 4,000m) in Chile's northermost Department of Arica revealed that 393 (40%) of the youngsters harbored one or more infections. When sorted by region and ethnicity, indigenous children of the highlands had infection rates 50% higher than children of Spanish descent living near the coast.^ An ecological model was developed and used to examine the causal path of infection and measure the effect of single and combined environmental variables. Family variables significantly linked to child health included maternal health, age and education. Significant child determinants of infection included the child's nutrient intake and medical history. When compared to children well and free of disease, infected youngsters reported a higher incidence of recent illness and a lower intake of basic foodstuffs. Traditional measures of child health, e.g. birth condition, weaning history, maternal fertility, and family wealth, did not differentiate between well and infected children.^ When height, weight, arm circumference, and subcapular skinfold measurements were compared, infected children, regardless of age, had smaller arm circumferences, the statistical difference being the greatest for males, age nine to eleven. Height and weight, the traditional growth indices, did not differentiate between well and infected groups.^ Infection is not determined by a single environmental factor or even a series of variables. Child health is ecological in nature and cannot be improved independent of changes in the environment that surrounds the child. To focus on selected child health needs, such as feeding programs or immunization campaigns, without simultaneously attending to the environment from which the needs arose is an inappropriate use of time, personnel, and money. ^

Ethnic differences in food availability at home among 3--5 year old preschoolers attending Head Start centers in 2008

Background. The childhood obesity epidemic has disproportionately impacted the lives of low-income, minority preschoolers and their families. Research shows that parents play a major role as "gatekeepers" who control what food is brought into the home and as role models for dietary practices. Currently, there is limited research regarding ethnic differences in families of low-income preschoolers. ^ Objective. The objective of this study was to look at ethnic differences in food availability at home among the low-income families of Hispanic and African American preschoolers attending Head Start centers in Harris County, Texas. ^ Design/Subjects. Descriptive data on food availability at home between Hispanic and African American families were used and analyzed for this study. Parents or primary caregivers (n = 718) of children enrolled at Head Start Centers in Houston, Texas completed the Healthy Home Survey. ^ Methods. In the Healthy Home Survey, participants were asked to answer open-ended questions regarding various types of foods currently available at home, such as fresh, canned or jarred, dried and frozen fruits; fresh, canned or jarred, and frozen vegetables; salty snacks, sweet snacks, candy, and soda. Descriptive analyses were conducted to identify significant differences between Hispanics and African Americans via a paired t-test to compare the means of variables between the study groups and a Pearson's chi-square or Fischer's exact (if cell size was <5) test calculated for food availability (food types) between ethnicities to determine differences in distributions. ^ Results. Although both Hispanics and African Americans reported having all categories of food types at home, there were statistically significant differences between ethnic groups. Hispanics were more likely to have fresh fruits and vegetables at home than African Americans. At the same time, more African American families reported having canned or jarred fruits and canned green/leafy vegetables than Hispanics. More Hispanic families reported having diet, regular, and both diet and regular sodas available compared to African American families. However, high percentages of unhealthy foods (including snacks and candy) were reported by both ethnicities. ^ Conclusions. The findings presented in this study indicate the implicit ethnic differences that exist in the food availability among low-income families of Hispanic and African American preschoolers. Future research should investigate the associations between food availability and children's weight status by ethnicity to identify additional differences that may exist.^