Multiple imputation for missing continuous and dichotomous data: The effects of patterns of missingness and variable correlations on type I error with small samples

The purpose of this study is to investigate the effects of predictor variable correlations and patterns of missingness with dichotomous and/or continuous data in small samples when missing data is multiply imputed. Missing data of predictor variables is multiply imputed under three different multivariate models: the multivariate normal model for continuous data, the multinomial model for dichotomous data and the general location model for mixed dichotomous and continuous data. Subsequent to the multiple imputation process, Type I error rates of the regression coefficients obtained with logistic regression analysis are estimated under various conditions of correlation structure, sample size, type of data and patterns of missing data. The distributional properties of average mean, variance and correlations among the predictor variables are assessed after the multiple imputation process. ^ For continuous predictor data under the multivariate normal model, Type I error rates are generally within the nominal values with samples of size n = 100. Smaller samples of size n = 50 resulted in more conservative estimates (i.e., lower than the nominal value). Correlation and variance estimates of the original data are retained after multiple imputation with less than 50% missing continuous predictor data. For dichotomous predictor data under the multinomial model, Type I error rates are generally conservative, which in part is due to the sparseness of the data. The correlation structure for the predictor variables is not well retained on multiply-imputed data from small samples with more than 50% missing data with this model. For mixed continuous and dichotomous predictor data, the results are similar to those found under the multivariate normal model for continuous data and under the multinomial model for dichotomous data. With all data types, a fully-observed variable included with variables subject to missingness in the multiple imputation process and subsequent statistical analysis provided liberal (larger than nominal values) Type I error rates under a specific pattern of missing data. It is suggested that future studies focus on the effects of multiple imputation in multivariate settings with more realistic data characteristics and a variety of multivariate analyses, assessing both Type I error and power. ^

Toward a cognitive taxonomy of medical errors.

One critical step in addressing and resolving the problems associated with human errors is the development of a cognitive taxonomy of such errors. In the case of errors, such a taxonomy may be developed (1) to categorize all types of errors along cognitive dimensions, (2) to associate each type of error with a specific underlying cognitive mechanism, (3) to explain why, and even predict when and where, a specific error will occur, and (4) to generate intervention strategies for each type of error. Based on Reason's (1992) definition of human errors and Norman's (1986) cognitive theory of human action, we have developed a preliminary action-based cognitive taxonomy of errors that largely satisfies these four criteria in the domain of medicine. We discuss initial steps for applying this taxonomy to develop an online medical error reporting system that not only categorizes errors but also identifies problems and generates solutions.

Influence of temperature of incubation and type of growth medium on pigmentation in Serratia marcescens.

Maximal amounts of prodigiosin were synthesized in either minimal or complete medium after incubation of cultures at 27 C for 7 days. Biosynthesis of prodigiosin began earlier and the range of temperature for formation was greater in complete medium. No prodigiosin was formed in either medium when cultures were incubated at 38 C; however, after a shift to 27 C, pigmentation ensued, provided the period of incubation at 38 C was not longer than 36 hr for minimal medium or 48 hr for complete medium. Washed, nonpigmented cells grown in either medium at 38 C for 72 hr could synthesize prodigiosin when suspended in saline at 27 C when casein hydrolysate was added. These suspensions produced less prodigiosin at a slower rate than did cultures growing in casein hydrolysate at 27 C without prior incubation at 38 C. Optimal concentration of casein hydrolysate for pigment formation by suspensions was 0.4%; optimal temperature was 27 C. Anaerobic incubation, shift back to 38 C, killing cells by heating, or chloramphenicol (25 mug/ml) inhibited pigmentation. Suspensions of washed cells forming pigment reached pH 8.0 to 8.3 rapidly and maintained this pH throughout incubation for 7 days. Measurements of viable count and of protein, plus other data, indicated that cellular multiplication did not occur in suspensions of washed cells during pigment formation. By this procedure utilizing a shift down in temperature, biosynthesis of prodigiosin by washed cells could be separated from multiplication of bacteria.

Penetrating injuries of the thorax and abdomen: Type of medical-care institution and case fatality

The purpose of this study was to determine, for penetrating injuries (gunshot, stab) of the chest/abdomen, the impact on fatality of treatment in trauma centers and shock trauma units compared with general hospitals. Medical records of all cases of penetrating injury limited to chest/abdomen and admitted to and discharged from 7 study facilities in Baltimore city 1979-1980 (n = 581) were studied: 4 general hospitals (n = 241), 2 area-wide trauma centers (n = 298), and a shock trauma unit (n = 42). Emergency center and transferred cases were not studied. Anatomical injury severity, measured by modified Injury Severity Score (mISS), was a significant prognostic factor for death, as were cardiovascular shock (SBP $\le$ 70), injury type (gunshot vs stab), and ambulance/helicopter (vs other) transport. All deaths occurred in cases with two or more prognostic factors. Unadjusted relative risks of death compared with general hospitals were 4.3 (95% confidence interval = 2.2, 8.4) for shock trauma and 0.8 (0.4, 1.7) for trauma centers. Controlling for prognostic factors by logistic regression resulted in these relative risks: shock trauma 4.0 (0.7, 22.2), and trauma centers 0.8 (0.2, 3.2). Factors significantly associated with increased risk had the following relative risks by multiple logistic regression: SBP $\le$ 70 (RR = 40.7 (11.0, 148.7)), highest mISS (42 (7.7, 227)), gunshot (8.4 (2.1, 32.6)), and ambulance/helicopter transport (17.2 (1.3, 228.1)). Controlling for age, race, and gender did not alter results significantly. Actual deaths compared with deaths predicted from a multivariable model of general-hospital cases showed 3.7 more than predicted deaths in shock trauma (SMR = 1.6 (0.8, 2.9)) and 0.7 more than predicted deaths in area-wide trauma centers (SMR = 1.05 (0.6, 1.7)). Selection bias due to exclusion of transfers and emergency center cases, and residual confounding due to insufficient injury information, may account for persistence of adjusted high case fatality in shock trauma. Studying all cases prospectively, including emergency center and transferred cases, is needed. ^

RELATION OF TYPE OF REPORTED ILLNESSES TO PATTERN OF USE OF THE STUDENT HEALTH CENTER BY THE INTERNATIONAL STUDENTS AT TEXAS SOUTHERN UNIVERSITY, HOUSTON, TEXAS

The state of knowledge on the relation of stress factors, health problems and health service utilization among university students is limited. Special problems of stress exist for the international students due to their having to adjust to a new environment. It is this latter problem area that provides the focus for this study. Recognizing there are special stress factors affecting the international students, it is first necessary to see if the problems of cultural adaptation affect them to any greater degree than American students attending the same university.^ To make the comparison, the study identified a number of health problems of both American and international students and related their frequency to the use of the Student Health Center. The expectation was that there would be an association between the number of health problems and the number of life change events experienced by these students and between the number of health problems and stresses from social factors. It was also expected that the number of health problems would decline with the amount of social support.^ The population chosen were students newly enrolled in Texas Southern University, Houston, Texas in the Fall Semester of 1979. Two groups were selected at random: 126 international and 126 American students. The survey instrument was a self-administered questionnaire. The response rate was 90% (114) for the international and 94% (118) for the American students.^ Data analyses consisted of both descriptive and inferential statistics. Chi-squares and correlation coefficients were the statistics used in comparing the international students and the American students.^ There was a weak association between the number of health problems and the number of life change events, as reported by both the international and the American students. The study failed to show any statistically significant association between the number of stress from social factors and the number of health problems. It also failed to show an association between the number of health problems and the amount of social support. These findings applied to both the international and the American students.^ One unexpected finding was that certain health problems were reported by more American than international students. There were: cough, diarrhea, and trouble in sleeping. Another finding was that those students with health insurance had a higher level of utilization of the Health Center than those without health insurance. More international than American students utilized the Student Health Center.^ In comparing the women students, there was no statistical significant difference in their reported fertility related health problems.^ The investigator recommends that in follow-up studies, instead of grouping all international students together, that they be divided by major nationalities represented in the student body; that is, Iranians, Nigerians and others. ^

Natural history study of arthrogryposis multiplex congenita, amyoplasia type

Arthrogryposis or Arthrogrypsosis Multiplex Congenita (AMC) are terms used to describe the clinical finding of multiple congenital contractures. There are more than 300 distinct disorders associated with arthrogryposis. Amyoplasia is the most common type of arthrogryposis and is often referred to as the “classic” type. There is no known cause of amyoplasia and no risk factors have been identified. Moreover, there is no established diagnostic criteria, which has led to inconsistency and confusion in the medical literature. The purpose of this study was to describe the natural history of amyoplasia, to determine if there are any identifiable risk factors and develop a list of diagnostic criteria. A retrospective chart review of 59 children with arthrogryposis ascertained at the Shriners Hospitals for Children in Houston, Texas was performed and included the following information: prenatal, birth, and family histories, and phenotypic descriptions. Forty-four children were identified with amyoplasia and 15 children with other multiple congenital contractures and other anomalies (MCC) were used as a comparison group. With the exception of abnormal amniotic fluid levels during pregnancy, there were no significant demographic or prenatal risk factors identified. However, we found common features that discriminate amyoplasia from other types of arthrogryposis and developed a diagnostic checklist. This checklist can be used as diagnostic criteria for discriminating amyoplasia from isolated and multiple contracture conditions.

Estimating the mean and standard deviation of concentration distributions using censored samples

Environmental data sets of pollutant concentrations in air, water, and soil frequently include unquantified sample values reported only as being below the analytical method detection limit. These values, referred to as censored values, should be considered in the estimation of distribution parameters as each represents some value of pollutant concentration between zero and the detection limit. Most of the currently accepted methods for estimating the population parameters of environmental data sets containing censored values rely upon the assumption of an underlying normal (or transformed normal) distribution. This assumption can result in unacceptable levels of error in parameter estimation due to the unbounded left tail of the normal distribution. With the beta distribution, which is bounded by the same range of a distribution of concentrations, $\rm\lbrack0\le x\le1\rbrack,$ parameter estimation errors resulting from improper distribution bounds are avoided. This work developed a method that uses the beta distribution to estimate population parameters from censored environmental data sets and evaluated its performance in comparison to currently accepted methods that rely upon an underlying normal (or transformed normal) distribution. Data sets were generated assuming typical values encountered in environmental pollutant evaluation for mean, standard deviation, and number of variates. For each set of model values, data sets were generated assuming that the data was distributed either normally, lognormally, or according to a beta distribution. For varying levels of censoring, two established methods of parameter estimation, regression on normal ordered statistics, and regression on lognormal ordered statistics, were used to estimate the known mean and standard deviation of each data set. The method developed for this study, employing a beta distribution assumption, was also used to estimate parameters and the relative accuracy of all three methods were compared. For data sets of all three distribution types, and for censoring levels up to 50%, the performance of the new method equaled, if not exceeded, the performance of the two established methods. Because of its robustness in parameter estimation regardless of distribution type or censoring level, the method employing the beta distribution should be considered for full development in estimating parameters for censored environmental data sets. ^

Effect of nonnormal random components on level and power in group-randomized trials

The use of group-randomized trials is particularly widespread in the evaluation of health care, educational, and screening strategies. Group-randomized trials represent a subset of a larger class of designs often labeled nested, hierarchical, or multilevel and are characterized by the randomization of intact social units or groups, rather than individuals. The application of random effects models to group-randomized trials requires the specification of fixed and random components of the model. The underlying assumption is usually that these random components are normally distributed. This research is intended to determine if the Type I error rate and power are affected when the assumption of normality for the random component representing the group effect is violated. ^ In this study, simulated data are used to examine the Type I error rate, power, bias and mean squared error of the estimates of the fixed effect and the observed intraclass correlation coefficient (ICC) when the random component representing the group effect possess distributions with non-normal characteristics, such as heavy tails or severe skewness. The simulated data are generated with various characteristics (e.g. number of schools per condition, number of students per school, and several within school ICCs) observed in most small, school-based, group-randomized trials. The analysis is carried out using SAS PROC MIXED, Version 6.12, with random effects specified in a random statement and restricted maximum likelihood (REML) estimation specified. The results from the non-normally distributed data are compared to the results obtained from the analysis of data with similar design characteristics but normally distributed random effects. ^ The results suggest that the violation of the normality assumption for the group component by a skewed or heavy-tailed distribution does not appear to influence the estimation of the fixed effect, Type I error, and power. Negative biases were detected when estimating the sample ICC and dramatically increased in magnitude as the true ICC increased. These biases were not as pronounced when the true ICC was within the range observed in most group-randomized trials (i.e. 0.00 to 0.05). The normally distributed group effect also resulted in bias ICC estimates when the true ICC was greater than 0.05. However, this may be a result of higher correlation within the data. ^

A full pedigree based method for the statistical assessment of genetic anticipation

Genetic anticipation is defined as a decrease in age of onset or increase in severity as the disorder is transmitted through subsequent generations. Anticipation has been noted in the literature for over a century. Recently, anticipation in several diseases including Huntington's Disease, Myotonic Dystrophy and Fragile X Syndrome were shown to be caused by expansion of triplet repeats. Anticipation effects have also been observed in numerous mental disorders (e.g. Schizophrenia, Bipolar Disorder), cancers (Li-Fraumeni Syndrome, Leukemia) and other complex diseases. ^ Several statistical methods have been applied to determine whether anticipation is a true phenomenon in a particular disorder, including standard statistical tests and newly developed affected parent/affected child pair methods. These methods have been shown to be inappropriate for assessing anticipation for a variety of reasons, including familial correlation and low power. Therefore, we have developed family-based likelihood modeling approaches to model the underlying transmission of the disease gene and penetrance function and hence detect anticipation. These methods can be applied in extended families, thus improving the power to detect anticipation compared with existing methods based only upon parents and children. The first method we have proposed is based on the regressive logistic hazard model. This approach models anticipation by a generational covariate. The second method allows alleles to mutate as they are transmitted from parents to offspring and is appropriate for modeling the known triplet repeat diseases in which the disease alleles can become more deleterious as they are transmitted across generations. ^ To evaluate the new methods, we performed extensive simulation studies for data simulated under different conditions to evaluate the effectiveness of the algorithms to detect genetic anticipation. Results from analysis by the first method yielded empirical power greater than 87% based on the 5% type I error critical value identified in each simulation depending on the method of data generation and current age criteria. Analysis by the second method was not possible due to the current formulation of the software. The application of this method to Huntington's Disease and Li-Fraumeni Syndrome data sets revealed evidence for a generation effect in both cases. ^

Nonlinear tests for genetic studies of complex disease

Linkage and association studies are major analytical tools to search for susceptibility genes for complex diseases. With the availability of large collection of single nucleotide polymorphisms (SNPs) and the rapid progresses for high throughput genotyping technologies, together with the ambitious goals of the International HapMap Project, genetic markers covering the whole genome will be available for genome-wide linkage and association studies. In order not to inflate the type I error rate in performing genome-wide linkage and association studies, multiple adjustment for the significant level for each independent linkage and/or association test is required, and this has led to the suggestion of genome-wide significant cut-off as low as 5 × 10 −7. Almost no linkage and/or association study can meet such a stringent threshold by the standard statistical methods. Developing new statistics with high power is urgently needed to tackle this problem. This dissertation proposes and explores a class of novel test statistics that can be used in both population-based and family-based genetic data by employing a completely new strategy, which uses nonlinear transformation of the sample means to construct test statistics for linkage and association studies. Extensive simulation studies are used to illustrate the properties of the nonlinear test statistics. Power calculations are performed using both analytical and empirical methods. Finally, real data sets are analyzed with the nonlinear test statistics. Results show that the nonlinear test statistics have correct type I error rates, and most of the studied nonlinear test statistics have higher power than the standard chi-square test. This dissertation introduces a new idea to design novel test statistics with high power and might open new ways to mapping susceptibility genes for complex diseases. ^

Roles of mismatch repair proteins in the recognition and processing of DNA interstrand crosslinks in human cells

DNA interstrand crosslinks (ICLs) are among the most toxic type of damage to a cell. Many ICL-inducing agents are widely used as therapeutic agents, e.g. cisplatin, psoralen. A bettor understanding of the cellular mechanism that eliminates ICLs is important for the improvement of human health. However, ICL repair is still poorly understood in mammals. Using a triplex-directed site-specific ICL model, we studied the roles of mismatch repair (MMR) proteins in ICL repair in human cells. We are also interested in using psoralen-conjugated triplex-forming oligonucleotides (TFOs) to direct ICLs to a specific site in targeted DNA and in the mammalian genomes. ^ MSH2 protein is the common subunit of two MMR recognition complexes, and MutSα and MutSβ. We showed that MSH2 deficiency renders human cell hypersensitive to psoralen ICLs. MMR recognition complexes bind specifically to triplex-directed psoralen ICLs in vitro. Together with the fact that psoralen ICL-induced repair synthesis is dramatically decreased in MSH2 deficient cell extracts, we demonstrated that MSH2 function is critical for the recognition and processing of psoralen ICLs in human cells. Interestingly, lack of MSH2 does not reduce the level of psoralen ICL-induced mutagenesis in human cells, suggesting that MSH2 does not contribute to error-generating repair of psoralen ICLs, and therefore, may represent a novel error-free mechanism for repairing ICLs. We also studied the role of MLH1, anther key protein in MMR, in the processing of psoralen ICLs. MLH1-deficient human cells are more resistant to psoralen plus UVA treatment. Importantly, MLH1 function is not required for the mutagenic repair of psoralen ICLs, suggesting that it is not involved in the error-generating repair of this type of DNA damage in human cells. ^ These are the first data indicating mismatch repair proteins may participate in a relatively error-free mechanism for processing psoralen ICL in human cells. Enhancement of MMR protein function relative to nucleotide excision repair proteins may reduce the mutagenesis caused by DNA ICLs in humans. ^ In order to specifically target ICLs to mammalian genes, we identified novel TFO target sequences in mouse and human genomes. Using this information, many critical mammalian genes can now be targeted by TFOs.^

Nurse likelihood to report a pediatric medication error: Examination of the "pre-reporting" period

Each year, hospitalized patients experience 1.5 million preventable injuries from medication errors and hospitals incur an additional $3.5 billion in cost (Aspden, Wolcott, Bootman, & Cronenwatt; (2007). It is believed that error reporting is one way to learn about factors contributing to medication errors. And yet, an estimated 50% of medication errors go unreported. This period of medication error pre-reporting, with few exceptions, is underexplored. The literature focuses on error prevention and management, but lacks a description of the period of introspection and inner struggle over whether to report an error and resulting likelihood to report. Reporting makes a nurse vulnerable to reprimand, legal liability, and even threat to licensure. For some nurses this state may invoke a disparity between a person‘s belief about him or herself as a healer and the undeniable fact of the error.^ This study explored the medication error reporting experience. Its purpose was to inform nurses, educators, organizational leaders, and policy-makers about the medication error pre-reporting period, and to contribute to a framework for further investigation. From a better understanding of factors that contribute to or detract from the likelihood of an individual to report an error, interventions can be identified to help the nurse come to a psychologically healthy resolution and help increase reporting of error in order to learn from error and reduce the possibility of future similar error.^ The research question was: "What factors contribute to a nurse's likelihood to report an error?" The specific aims of the study were to: (1) describe participant nurses' perceptions of medication error reporting; (2) describe participant explanations of the emotional, cognitive, and physical reactions to making a medication error; (3) identify pre-reporting conditions that make it less likely for a nurse to report a medication error; and (4) identify pre-reporting conditions that make it more likely for a nurse to report a medication error.^ A qualitative research study was conducted to explore the medication error experience and in particular the pre-reporting period from the perspective of the nurse. A total of 54 registered nurses from a large private free-standing not-for-profit children's hospital in the southwestern United States participated in group interviews. The results describe the experience of the nurse as well as the physical, emotional, and cognitive responses to the realization of the commission of a medication error. The results also reveal factors that make it more and less likely to report a medication error.^ It is clear from this study that upon realization that he or she has made a medication error, a nurse's foremost concern is for the safety of the patient. Fear was also described by each group of nurses. The nurses described a fear of several things including physician reaction, manager reaction, peer reaction, as well as family reaction and possible lack of trust as a result. Another universal response was the description of a struggle with guilt, shame, imperfection, blaming oneself, and questioning one's competence.^

Bayesian estimation of multilevel item response model with missing data

This study proposed a novel statistical method that modeled the multiple outcomes and missing data process jointly using item response theory. This method follows the "intent-to-treat" principle in clinical trials and accounts for the correlation between outcomes and missing data process. This method may provide a good solution to chronic mental disorder study. ^ The simulation study demonstrated that if the true model is the proposed model with moderate or strong correlation, ignoring the within correlation may lead to overestimate of the treatment effect and result in more type I error than specified level. Even if the within correlation is small, the performance of proposed model is as good as naïve response model. Thus, the proposed model is robust for different correlation settings if the data is generated by the proposed model.^

A simulation study of the effects of small center enrollment in multi -center clinical trials

Multi-center clinical trials are very common in the development of new drugs and devices. One concern in such trials, is the effect of individual investigational sites enrolling small numbers of patients on the overall result. Can the presence of small centers cause an ineffective treatment to appear effective when treatment-by-center interaction is not statistically significant?^ In this research, simulations are used to study the effect that centers enrolling few patients may have on the analysis of clinical trial data. A multi-center clinical trial with 20 sites is simulated to investigate the effect of a new treatment in comparison to a placebo treatment. Twelve of these 20 investigational sites are considered small, each enrolling less than four patients per treatment group. Three clinical trials are simulated with sample sizes of 100, 170 and 300. The simulated data is generated with various characteristics, one in which treatment should be considered effective and another where treatment is not effective. Qualitative interactions are also produced within the small sites to further investigate the effect of small centers under various conditions.^ Standard analysis of variance methods and the "sometimes-pool" testing procedure are applied to the simulated data. One model investigates treatment and center effect and treatment-by-center interaction. Another model investigates treatment effect alone. These analyses are used to determine the power to detect treatment-by-center interactions, and the probability of type I error.^ We find it is difficult to detect treatment-by-center interactions when only a few investigational sites enrolling a limited number of patients participate in the interaction. However, we find no increased risk of type I error in these situations. In a pooled analysis, when the treatment is not effective, the probability of finding a significant treatment effect in the absence of significant treatment-by-center interaction is well within standard limits of type I error. ^