Browsing by Subject "Latent class analysis"

Background: The Kilimanjaro region has one of the highest rates of reported alcohol use per capita in Tanzania. Alcohol-related risky behaviors pose substantial threats to the health and well-being of alcohol users and people around them. This study sought to provide a better understanding of how risky behaviors are associated with alcohol use perceptions. Methods: This mixed method study took place in the Kilimanjaro Christian Medical Center. Quantitative data on alcohol use, the alcohol use disorder identification tool, alcohol-related consequences, and qualitative data on alcohol use perception and risky behaviors were collected from a hospital- and non-hospital-based sample in Moshi, Tanzania. Latent class analysis was applied to examine alcohol-related risky behaviors. Results: Three classes of risky behavior patterns were identified: “no risky behavior”, “moderate risky behaviors” and “high risky behaviors”. Membership of classes 3 was associated with the most alcohol use quantity and frequency. No association between classes and alcohol-stigma was found. Our qualitative results explored alcohol perceptions and risky behaviors and illustrated their possible associations. Conclusions: Although alcohol stigma may not associate the number of risky behaviors directly, our qualitative result helped us to understand how stigma associates with risky behaviors. This study may serve as a reference for designing and adjusting interventions for alcohol-related injury patients' needs; we can improve interventions by using our knowledge about misconception and stigma and the identified risky behaviors classes as a form of classification system.

This study examines whether trajectories of multiple social role occupancy, measured by level and dynamics of spouse, parent, and worker roles, are associated with mortality and concurrent trajectories of depressive symptoms and self-rated health. I frame hypotheses with role strain, role enhancement, role context, stress process, and life course theories to examine both within-person changes over age and between-person predictors of health status.

I use data from the Americans' Changing Lives Study, a nationally representative accelerated cohort panel study of U.S. adults interviewed in 1986, 1989, 1994, and 2001/2 with mortality tracking through 2006. I use latent class analysis to estimate disaggregated trajectories of role occupancy, role strain, role satisfaction, depressive symptoms, and self-rated health across the adult life course. I then use multinomial and logistic regression analyses to examine associations between role trajectories and health outcomes.

I find that (1) there is significant heterogeneity in trajectories of role occupancy and characteristics across the adult life course; (2) higher levels of social role occupancy are associated with better depressive symptom and mortality outcomes; (3) lower levels of role strain and higher levels of role satisfaction are associated with better depressive symptom outcomes, and (4); the association between role occupancy and health is robust to the inclusion of role characteristics. Thus, I find support for the role enhancement hypothesis in that higher levels of role occupancy are associated with better health outcomes irrespective of reward and strain associated with those roles.

Background: It is unclear if strategies such as using different interferon-gamma release assay (IGRA) cutoffs or performing serial testing with a tuberculin skin test (TST) and an IGRA, or two different IGRAs could improve test accuracy in those from countries with high TB prevalence. Methods: We used Bayesian latent class analysis to construct receiving operating characteristic (ROC) curves and calculate their area under the curve (AUC) based on our cohort of HIV-negative, non-US-born individuals 5 years and older residing in the USA with valid results for all three LTBI tests (i.e TST, Quantiferon Gold-In-Tube (QFT) and T-SPOT.TB (TSPOT)). We compared the sensitivity, specificity and predictive values of 1) standard U.S. cutoffs for test positivity and higher or lower cutoffs, and 2) a single LTBI test and two tests: TST-QFT, TST-TSPOT, QFT-TSPOT. Results: A total of 14,167 participants were included in the analysis. The ROC curves had an AUC of 0.81 (95% Credible Interval (CrI) 0.78 - 0.86), 0.89 (95% CrI 0.86-0.93), and 0.92 (95% CrI 0.88-0.96) for TST, QFT and TSPOT respectively. QFT at the 0.35 IU/ml cutoff and TSPOT at a cutoff of 5 spots made the least number of diagnostic errors. The TST-QFT combination had similar sensitivity, specificity, PPV and NPV to a TST at a cutoff of 8 spots and the TST-TSPOT performed similarly to a QFT at a cutoff of 0.7 IU/mL. Conclusions: Using the lowest IGRA cutoffs performed better in our study population by increasing test sensitivity and NPV, therefore minimizing the total number of diagnostic errors. Two tests performed similarly to single testing with an IGRA at higher cutoffs.