Application of the general linear model to assess the effect of missing data on bone marrow mononuclear cell therapy with infusion timing and follow-up MRI

With most clinical trials, missing data presents a statistical problem in evaluating a treatment's efficacy. There are many methods commonly used to assess missing data; however, these methods leave room for bias to enter the study. This thesis was a secondary analysis on data taken from TIME, a phase 2 randomized clinical trial conducted to evaluate the safety and effect of the administration timing of bone marrow mononuclear cells (BMMNC) for subjects with acute myocardial infarction (AMI).^ We evaluated the effect of missing data by comparing the variance inflation factor (VIF) of the effect of therapy between all subjects and only subjects with complete data. Through the general linear model, an unbiased solution was made for the VIF of the treatment's efficacy using the weighted least squares method to incorporate missing data. Two groups were identified from the TIME data: 1) all subjects and 2) subjects with complete data (baseline and follow-up measurements). After the general solution was found for the VIF, it was migrated Excel 2010 to evaluate data from TIME. The resulting numerical value from the two groups was compared to assess the effect of missing data.^ The VIF values from the TIME study were considerably less in the group with missing data. By design, we varied the correlation factor in order to evaluate the VIFs of both groups. As the correlation factor increased, the VIF values increased at a faster rate in the group with only complete data. Furthermore, while varying the correlation factor, the number of subjects with missing data was also varied to see how missing data affects the VIF. When subjects with only baseline data was increased, we saw a significant rate increase in VIF values in the group with only complete data while the group with missing data saw a steady and consistent increase in the VIF. The same was seen when we varied the group with follow-up only data. This essentially showed that the VIFs steadily increased when missing data is not ignored. When missing data is ignored as with our comparison group, the VIF values sharply increase as correlation increases.^

Relationships of impulsivity and physical fighting history to alcohol -induced aggression in women

Research suggests women respond to the aggression-inducing effects of alcohol in a manner similar to men. Highly aggressive men are more prone to alcohol-induced aggression, but this relationship is less clear for women. This study examined whether alcohol consumption would differentially affect laboratory-measured aggression in a sample of aggressive and non-aggressive women and how those differences might be related to components of impulsive behavior. In 39 women recruited from the community (two groups: with and without histories of physical fighting) ages 21–40, laboratory aggressive behavior was assessed following placebo and 0.80 g/kg alcohol consumption (all women experienced both conditions). Baseline laboratory impulsive behavior of three impulsivity models was later assessed in the same women. In the aggression model (PSAP), participants were provoked by periodic subtractions of money, which were blamed on a fictitious partner. Aggression was operationalized as the responses the participant made to subtract money from that partner. The three components of impulsivity that were tested included: (1) response initiation (IMT/DMT), premature responses made prior to the completion of stimulus processing, (2) response inhibition (GoStop), a failure to inhibit an already initiated response, and (3) consequence sensitivity (SKIP and TCIP), the choice for a smaller-sooner reward over a larger-later reward. I hypothesized that, compared to women with no history of physical fighting, women with a history of physical fighting would exhibit higher rates of alcohol-induced laboratory aggression and higher rates of baseline impulsive responding (particularly for the IMT/DMT), which would also be related to the alcohol-induced increases aggression. Consistent with studies in men, the aggressive women showed strong associations between laboratory aggression and self-report measures, while the non-aggressive women did not. However, unlike men, following alcohol consumption it was the non-aggressive women's laboratory aggression that was related to their self-reports of aggression and impulsivity. Additionally, response initiation measures of impulsivity distinguished the two groups, while response inhibition and consequence sensitivity measures did not; commission error rates on the IMT/DMT were higher in the aggressive women compared to the non-aggressive women. Regression analyses of the behavioral measures showed no relationship between the aggression and impulsivity performance of the two groups. These results suggest that the behavioral (and potentially biological) mechanism underlying aggressive behavior of women is different than that of men. ^

Behavioral effects of plasma tryptophan depletion in aggressive and nonaggressive men

Serotonin (5-HT) neurotransmission deficits have been implicated in impulsive aggression. A Trp-free beverage of amino acids competitively inhibits Trp uptake into the brain for 5-HT synthesis and also lowers endogenous plasma Trp for several hours. This has worsened mood and/or increased aggressive behavior, especially in hostile persons or those with histories of depression. In 24 community-recruited men (12 each with and without significant aggression histories), aggressive and impulsive behavior in the laboratory was assessed before and after plasma Trp depletion and Trp loading. In the aggression model, subjects were provoked by periodic subtractions of participation earnings, and these subtractions were blamed on a ficitious other participant. Aggression was measured as the responses the subject made to subtract money from his antagonist. Impulsiveness was operationalized as: (1) the choice of smaller reward after a shorter delay over having to wait longer to receive a larger reward, and (2) “false alarm” commission errors in a modified Continuous Performance Task, which represent a failure to inhibit responding to stimuli similar (but not identical) to target stimuli. Finally, plasma cortisol and Trp were measured under each condition immediately following a aggression testing session when subjects were highly provoked. I hypothesized that 5-HT may tonically modulate (inhibit) the hypothalmnic-pituitary-adrenal stress response, such that Trp depletion may enhance the cortisol response to high provocation in aggressive men. ^ Trp depletion had no effect in the laboratory tasks purported to measure impulsive behavior, and failed to cause increases in aggressive behavior under low provocation conditions. Under higher provocation, however, aggressive responses we re elevated under Trp-depleted conditions relative to Trp-loaded conditions in aggressive men, whereas the reverse was true in nonaggressive men. Cortisol levels nonsignificantly paralled the group differences in aggression under Trp-depleted and Trp-loaded conditions. Aggressive men achieved lower plasma Trp levels after Trp loading than did nonaggressive men, possibly due to heavy alcohol use histories. The high post-loading plasma Trp levels in nonaggressive men tended also to correlate with their aggressive responding rates, due perhaps to increases in other psychoactive Trp metabolites. ^

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. ^

A space-time analysis of reported shigellosis and salmonellosis cases in Texas from 2000 to 2004

Background. The purpose of this study was to describe the risk factors and demographics of persons with salmonellosis and shigellosis and to investigate both seasonal and spatial variations in the occurrence of these infections in Texas from 2000 to 2004, utilizing time series analyses and the geographic information system digital mapping methods. ^ Methods. Spatial Analysis: MapInfo software was used to map the distribution of age-adjusted rates of reported shigellosis and salmonellosis in Texas from 2000–2004 by zip codes. Census data on above or below poverty level, household income, highest level of educational attainment, race, ethnicity, and urban/rural community status was obtained from the 2000 Decennial Census for each zip code. The zip codes with the upper 10% and lower 10% were compared using t-tests and logistic regression to determine whether there were any potential risk factors. ^ Temporal analysis. Seasonal patterns in the prevalence of infections in Texas from 2000 to 2003 were determined by performing time-series analysis on the numbers of cases of salmonellosis and shigellosis. A linear regression was also performed to assess for trends in the incidence of each disease, along with auto-correlation and multi-component cosinor analysis. ^ Results. Spatial analysis: Analysis by general linear model showed a significant association between infection rates and age, with young children aged less than 5 and those aged 5–9 years having increased risk of infection for both disease conditions. The data demonstrated that those populations with high percentages of people who attained a higher than high school education were less likely to be represented in zip codes with high rates of shigellosis. However, for salmonellosis, logistic regression models indicated that when compared to populations with high percentages of non-high school graduates, having a high school diploma or equivalent increased the odds of having a high rate of infection. ^ Temporal analysis. For shigellosis, multi-component cosinor analyses were used to determine the approximated cosine curve which represented a statistically significant representation of the time series data for all age groups by sex. The shigellosis results show 2 peaks, with a major peak occurring in June and a secondary peak appearing around October. Salmonellosis results showed a single peak and trough in all age groups with the peak occurring in August and the trough occurring in February. ^ Conclusion. The results from this study can be used by public health agencies to determine the timing of public health awareness programs and interventions in order to prevent salmonellosis and shigellosis from occurring. Because young children depend on adults for their meals, it is important to increase the awareness of day-care workers and new parents about modes of transmission and hygienic methods of food preparation and storage. ^

Regression with autocorrelated data: A study of time trend of global infant mortality rate

The infant mortality rate (IMR) is considered to be one of the most important indices of a country's well-being. Countries around the world and other health organizations like the World Health Organization are dedicating their resources, knowledge and energy to reduce the infant mortality rates. The well-known Millennium Development Goal 4 (MDG 4), whose aim is to archive a two thirds reduction of the under-five mortality rate between 1990 and 2015, is an example of the commitment. ^ In this study our goal is to model the trends of IMR between the 1950s to 2010s for selected countries. We would like to know how the IMR is changing overtime and how it differs across countries. ^ IMR data collected over time forms a time series. The repeated observations of IMR time series are not statistically independent. So in modeling the trend of IMR, it is necessary to account for these correlations. We proposed to use the generalized least squares method in general linear models setting to deal with the variance-covariance structure in our model. In order to estimate the variance-covariance matrix, we referred to the time-series models, especially the autoregressive and moving average models. Furthermore, we will compared results from general linear model with correlation structure to that from ordinary least squares method without taking into account the correlation structure to check how significantly the estimates change.^