Shreyash Deshmukh

As part of this role, I developed clinical risk prediction methods using longitudinal and survival data, contributing to more robust analytical approaches in clinical research.

I actively assisted in the development of the statistical analysis plan for the HATHI study, working in close collaboration with the Principal Investigator (PI) and study coordinator.

This collaboration also involved designing and finalizing paper-based Case Report Forms (CRFs) to ensure alignment with the study's data collection objectives.

Additionally, I supervised and trained clinicians, counselors, and field staff on data entry protocols, including the use of REDCap for electronic data capture and integration with Electronic Health Records (EHRs), while standardizing procedures through well-defined Standard Operating Procedures (SOPs).

To uphold data integrity, I conducted comprehensive quality assurance and quality control (QA/QC) processes to identify and rectify discrepancies in CRF and electronic data, ensuring the delivery of high-quality, reliable datasets for analysis.

Conducted a meta-analysis to estimate survival rates from HIV diagnosis to AIDS onset and from AIDS onset to death.

Key Contributions:

• Reviewed 20 cohort studies with 25000 participants from PubMed, Web of Science, and Scopus databases.
• Analyzed survival probabilities for HIV-to-AIDS progression and AIDS-related mortality with and without HAART (Highly Active Antiretroviral Therapy).
• Applied random-effects models to estimate survival rates with 95% confidence intervals.
• Identified significant heterogeneity and evaluated studies for risk of bias.
• Concluded that HAART significantly improves survival, with most untreated patients dying within two years of AIDS onset.

This study is exploratory study, which examines the dynamic relationships between air pollutants (PM_2.5, PM₁₀, O₃, NO₂, SO₂, CO) and Air Quality Index (AQI) using a Vector Autoregressive (VAR) model.

The Augmented Dickey-Fuller (ADF) test confirms stationarity, and optimal lag selection criteria (AIC, BIC, HQIC) suggest a lag length of 3 or 4.

Granger causality tests indicate that PM_2.5, PM₁₀, NO₂, and CO significantly influence AQI, while impulse response analysis shows that shocks in PM₁₀ and CO have persistent effects on AQI.

These findings highlight the need for targeted policy interventions to control key pollutants, particularly vehicular emissions and industrial pollution.