Search tips
Search criteria

Results 1-6 (6)

Clipboard (0)

Select a Filter Below

Year of Publication
Document Types
1.  Cumulative Risk, Cumulative Outcome: A 20-Year Longitudinal Study 
PLoS ONE  2015;10(6):e0127650.
Cumulative risk (CR) models provide some of the most robust findings in the developmental literature, predicting numerous and varied outcomes. Typically, however, these outcomes are predicted one at a time, across different samples, using concurrent designs, longitudinal designs of short duration, or retrospective designs. We predicted that a single CR index, applied within a single sample, would prospectively predict diverse outcomes, i.e., depression, intelligence, school dropout, arrest, smoking, and physical disease from childhood to adulthood. Further, we predicted that number of risk factors would predict number of adverse outcomes (cumulative outcome; CO). We also predicted that early CR (assessed at age 5/6) explains variance in CO above and beyond that explained by subsequent risk (assessed at ages 12/13 and 19/20). The sample consisted of 284 individuals, 48% of whom were diagnosed with a speech/language disorder. Cumulative risk, assessed at 5/6-, 12/13-, and 19/20-years-old, predicted aforementioned outcomes at age 25/26 in every instance. Furthermore, number of risk factors was positively associated with number of negative outcomes. Finally, early risk accounted for variance beyond that explained by later risk in the prediction of CO. We discuss these findings in terms of five criteria posed by these data, positing a “mediated net of adversity” model, suggesting that CR may increase some central integrative factor, simultaneously augmenting risk across cognitive, quality of life, psychiatric and physical health outcomes.
PMCID: PMC4452593  PMID: 26030616
2.  Follow-Up Visit Patterns in an Antiretroviral Therapy (ART) Programme in Zomba, Malawi 
PLoS ONE  2014;9(7):e101875.
Identifying follow-up (FU) visit patterns, and exploring which factors influence them are likely to be useful in determining which patients on antiretroviral therapy (ART) may become Lost to Follow-Up (LTFU). Using an operation and implementation research approach, we sought 1) to describe the timing of FU visits amongst patients who have been on ART for shorter and longer periods of time; and 2) to determine the median time to late visits, and 3) to identify specific factors that may be associated with these patterns in Zomba, Malawi.
Methods and Findings
Using routinely collected programme monitoring data from Zomba District, we performed descriptive analyses on all ART visits among patients who initiated ART between Jan. 1, 2007–June 30, 2010. Based on an expected FU date, each FU visit was classified as early (≥4 day before an expected FU date), on time (3 days before an expected FU date/up to 6 days after an expected FU date), or late (≥7 days after an expected FU date). In total, 7,815 patients with 76417 FU visits were included. Ninety-two percent of patients had ≥2 FU visits. At the majority of visits, patients were either on time or late. The median time to a first late visit among those with 2 or more visits was 216 days (IQR: 128–359). Various patient- and visit-level factors differed significantly across Early, On Time, and Late visit groups including ART adherence and frequency of, and type of side effects.
The majority of patients do not demonstrate consistent FU visit patterns. Individuals were generally on ART for at least 6 months before experiencing their first late visit. Our findings have implications for the development of effective interventions that meet patient needs when they present early and can reduce patient losses to follow-up when they are late. In particular, time-varying visit characteristics need further research.
PMCID: PMC4102478  PMID: 25033285
3.  Prediction of Drosophila melanogaster gene function using Support Vector Machines 
BioData Mining  2013;6:8.
While the genomes of hundreds of organisms have been sequenced and good approaches exist for finding protein encoding genes, an important remaining challenge is predicting the functions of the large fraction of genes for which there is no annotation. Large gene expression datasets from microarray experiments already exist and many of these can be used to help assign potential functions to these genes. We have applied Support Vector Machines (SVM), a sigmoid fitting function and a stratified cross‐validation approach to analyze a large microarray experiment dataset from Drosophila melanogaster in order to predict possible functions for previously un‐annotated genes. A total of approximately 5043 different genes, or about one‐third of the predicted genes in the D. melanogaster genome, are represented in the dataset and 1854 (or 37%) of these genes are un‐annotated.
39 Gene Ontology Biological Process (GO‐BP) categories were found with precision value equal or larger than 0.75, when recall was fixed at the 0.4 level. For two of those categories, we have provided additional support for assigning given genes to the category by showing that the majority of transcripts for the genes belonging in a given category have a similar localization pattern during embryogenesis. Additionally, by assessing the predictions using a confidence score, we have been able to provide a putative GO‐BP term for 1422 previously un‐annotated genes or about 77% of the un‐annotated genes represented on the microarray and about 19% of all of the un‐annotated genes in the D. melanogaster genome.
Our study successfully employs a number of SVM classifiers, accompanied by detailed calibration and validation techniques, to generate a number of predictions for new annotations for D. melanogaster genes. The applied probabilistic analysis to SVM output improves the interpretability of the prediction results and the objectivity of the validation procedure.
PMCID: PMC3669044  PMID: 23547736
Gene ontology; Support Vector Machines; Drosophila melanogaster; Gene expression data; Gene function prediction
4.  Using geographical information systems mapping to identify areas presenting high risk for traumatic brain injury 
The aim of this study is to show how geographical information systems (GIS) can be used to track and compare hospitalization rates for traumatic brain injury (TBI) over time and across a large geographical area using population based data.
Results & Discussion
Data on TBI hospitalizations, and geographic and demographic variables, came from the Ontario Trauma Registry Minimum Data Set for the fiscal years 1993-1994 and 2001-2002. Various visualization techniques, exploratory data analysis and spatial analysis were employed to map and analyze these data. Both the raw and standardized rates by age/gender of the geographical unit were studied. Data analyses revealed persistent high rates of hospitalization for TBI resulting from any injury mechanism between two time periods in specific geographic locations.
This study shows how geographic information systems can be successfully used to investigate hospitalizaton rates for traumatic brain injury using a range of tools and techniques; findings can be used for local planning of both injury prevention and post discharge services, including rehabilitation.
PMCID: PMC3260231  PMID: 22054220
traumatic brain injury; geographic information systems; geographic visualization; spatial analysis
5.  Incidence and estimated rates of residual risk for HIV, hepatitis C, hepatitis B and human T-cell lymphotropic viruses in blood donors in Canada, 1990–2000 
Since 1990, the Canadian Red Cross Society and Canadian Blood Services have been testing blood donors for hepatitis C virus (HCV) antibody and HCV nucleic acids and have supplemented HIV antibody testing with p24 antigen testing. We report trends in the incidence of blood-transmissible viral markers and estimates of the risk of undetected infection in donors over the last decade.
We extracted anonymous donor and blood-transmissible disease information from the Canadian Blood Services National Epidemiology Donor Database for 8.9 million donations from 2.1 million donors between June 1990 and December 2000. The risk of transfusion-transmitted infection (or “residual risk”) refers to the chance that an infected donation escapes detection because of a laboratory test's window period (i.e., the time between infection and detection of the virus by that test). We determined the probability of residual contamination of a unit of blood after testing by using the incidence/window period model, which is based on the incidence of infection in repeat donors and the window period for each laboratory test. The viral markers evaluated in the study were HIV, HCV, hepatitis B virus (HBV) and human T-cell lymphotropic virus (HTLV).
Except for HBV, the transmissible-disease rates of the other evaluated viruses decreased over the study period, with less of a decrease for HTLV. In 2000, the transmissible-disease–positive rate per 100 000 donations was 0.38 for HIV, 16.83 for HCV, 12.40 for HBV and 1.77 for HTLV. The residual risk of HIV, HCV and HTLV decreased over the study period; the residual risk of HBV fluctuated throughout the decade. The current residual risk per million donations is 0.10 for HIV, 0.35 for HCV, 13.88 for HBV and 0.95 for HTLV.
Except for HBV, the estimated risk of undetected infection (residual risk) has decreased over time. The rates of transmissible disease and the probability of undetected transmission of infection are at par with, if not lower than, those reported for other industrialized countries.
PMCID: PMC203278  PMID: 14557314
6.  Effect of motor vehicle emissions on respiratory health in an urban area. 
Environmental Health Perspectives  2002;110(3):293-300.
Motor vehicles emit particulate matter < 2.5 microm in diameter (PM(2.5)), and as a result, PM(2.5) concentrations tend to be elevated near busy streets. Studies of the relationship between motor vehicle emissions and respiratory health are generally limited by difficulties in exposure assessment. We developed a refined exposure model and implemented it using a geographic information system to estimate the average daily census enumeration area (EA) exposure to PM(2.5). Southeast Toronto, the study area, includes 334 EAs and covers 16 km(2) of urban area. We used hospital admission diagnostic codes from 1990 to 1992 to measure respiratory and genitourinary conditions. We assessed the effect of EA exposure on hospital admissions using a Poisson mixed-effects model and examined the spatial distributions of variables. Exposure to PM(2.5) has a significant effect on admission rates for a subset of respiratory diagnoses (asthma, bronchitis, chronic obstructive pulmonary disease, pneumonia, upper respiratory tract infection), with a relative risk of 1.24 (95% confidence interval, 1.05-1.45) for a log(10) increase in exposure. We noted a weaker effect of exposure on hospitalization for all respiratory conditions, and no effect on hospitalization for nonrespiratory conditions.
PMCID: PMC1240770  PMID: 11882481

Results 1-6 (6)