Conditional Probabilities of AIDS Disease Transitions Using Semi-Markov Models

Analyzing progression of diseases is vital to monitor patient's traversal over time through a disease. Clinical study settings present modeling challenges, as patients' disease trajectories are only partially observed, and patients' disease statuses are only assessed at clinic visit times. HIV disease is a continuum of progressive damage to the immune system from the time of infection to the manifestation of severe immunologic damage. We proposed a semi-Markov model and collected data at Yirgalem General Hospital. Our study found that for an HIV/AIDS patient the transition probability from a given state to the next worse state increases within the good states as time gets optimum and then decreases with increasing time during a follow up. In a specific state of the disease a patient will stay in that state with a non- zero probability in good states and a patient will transit to the next state either to the worst or to the good state with a non-zero probability. The probability of being in same state decreases over time. With the good or alive states, the probability of being in a better state is non-zero, but less than the probability of being in worst states. The survival probabilities are decreasing with increasing time. Therefore, we recommend that increased clinical care for patients on ART services should be strengthen and patients need to regularly check their CD4 T cell count in the appropriate day based on physician order to timely know and monitor their disease status to improve the survival probability and to reduce mortality.

Bayesian Joint Modelling of Disease Progression Marker and Time to Death Event of HIV/AIDS Patients under ART Follow-up.

Objectives: To develop separate and joint statistical models in the Bayesian framework for longitudinal measurements and time to death event data of HIV/AIDS patients. Study design: Longitudinal study. Place and Duration of Study: The population of study includes all HIV/AIDS patients who had been under follow up of Antiretroviral Therapy (ART) from January 2006 to December 2012 at Shashemene Referral Hospital in Ethiopia. Methodology: The posterior model was analyzed using Gibbs sampler by sampling from the distributions of the parameters given the data. Convergence of each sample was maintained. Results: The results indicated that the joint model was not significant indicating that the CD4 count did not have significant effect on the patient’s survival time. The results of both the separate and joint analyses were consistent. The separate model was better interims of goodness of fitness than the joint model, while the final joint model was found to be simpler (less complex) model than the separate models. In the longitudinal sub-model, the predictors: linear time, squared time, sex, and tobacco addiction were statistically significant at 0.05 level of significance. For the survival submodel, knowledge of ART and condom use were significantly related with time to death. Conclusion: The Bayesian Joint model provides results consistent with that of the separate models.

Factors Influencing the Risk of Death among Patients with Heart Failure.

The purpose of this study was to investigate the impact of risk factors on the death of patients with heart failure in a cohort of patients hospitalized with heart failure disease. In this paper we used chisquare tests with the aim of studying the relationship of each factor with survival. Generalized Additive Models (GAM), particularly Generalized Additive Logistic Regression Model, was used to examine the impact of risk factors on the death of patients with heart failure out of 263 patients considered in the analysis, 18.6% patients died of heart failure. A death proportion for female was 19.6% and that of male patients was 17.5%. From the GAM analysis the predictors: age, anemia, Tuberculosis, HIV status, renal inefficiency, diabetes, hypertension and sinus were found to significantly affect the death status of a patient. Being older age, anemic, renal inefficient, TB positive, HIV positive, diabetic, hypertensive and sinus positive increase the risk of death of a heart failure patient.

Learning Bayesian Networks Using Heart Failure Data.

Background: Several factors may affect heart failure status of patients. It is important to investigate whether or not the effects are direct. The purpose of this study was learning Bayesian networks that encode the joint probability distribution for a set of random variables. Methods: The design was a retrospective cohort study. The target population for this study was heart failure patients who were under follow- up at Asella referral teaching Hospital from February, 2009 to March, 2012. Bayesian Network is used in this paper to examine causal relationships between variables via Directed Acyclic Graph (DAG). Results: Death of patients can be determined using HIV, hypertension, diabetes, anemia, renal inefficiency and sinus. Hypertension and sinus were found to have direct effects while TB had only indirect effect. Age did not have an effect. Conclusion: Anemia, HIV, diabetes mellitus renal inefficiency and sinus directly affect the death of heart failure patient. Death is conditionally independent on TB and age, given all other variables.