University of South Florida
Clinical Features, Laboratory Characteristics, and Outcomes of COVID-19 Patients in a Tertiary Care Reference Hospital in Panama City, Panama. A Retrospective Observational Study
Abstract
dc:descriptionLatin America experienced some of the world’s highest COVID-19 mortality rates, yet data on in-hospital outcomes and their social and geographic determinants remain limited. This dissertation investigated multilevel baseline factors associated with in-hospital mortality among adults with COVID-19 at Panama’s largest public tertiary hospital (Hospital Santo Tomás), integrating survival, mediation, and spatial-temporal modeling within a causal-inference framework. A retrospective cohort of 1,526 RT-PCR–confirmed COVID-19 hospitalizations (May 2020–November 2021) was analyzed using multivariable Cox proportional-hazards, Fine–Gray competing-risk, mediation, and Bayesian spatial models. Directed acyclic graphs (DAGs) pre-specified causal relationships and minimally sufficient adjustment sets. DAG-guided models quantified total effects without conditioning on mediators. Temporal non-proportionality and volatility in access to care were examined using time-varying-coefficient and GARCH models. Overall, in-hospital mortality was 27.4 %, with most deaths occurring within 30 days of admission. Age was the dominant risk factor (HR 1.04 per year; p < 0.001). Acute complications (such as pneumonia (HR 3.05), ARDS (1.68), septic shock (2.18), and cerebrovascular (2.71) or cardiovascular (1.64) events) were the strongest factors associated with time to in-hospital death. Laboratory abnormalities reflecting inflammation and organ failure (elevated lactate, BUN, LDH, CRP, INR, bilirubin, and D-dimer) were associated with in-hospital mortality. Diabetes increased the unadjusted risk but attenuated after adjustment, consistent with confounding and/or pathways through downstream complications. Invasive mechanical ventilation showed a time-increasing hazard (HR ≈ 2.1 → 4.5 over 21 days), indicating non-proportional risk and potential reverse-causation dynamics. Fine–Gray competing-risk models confirmed a 6–7 percentage-point excess 30-day mortality among diabetic patients. Mediation analysis detected no significant racial or ethnic disparities in in-hospital mortality, implying that inequities are likely to occur upstream of admission. Spatial and GARCH analyses of travel to the hospital revealed persistent geographic inequities and long-memory volatility in accessibility but no evidence of an association between travel burden and survival once admitted. COVID-19 mortality at Hospital Santo Tomás was shaped by an interaction of age, comorbidity, acute complications, and physiologic derangement. Integrating DAG-based causal modeling with time-series and spatial analytics provided a coherent, multilevel view of disease severity and health-system performance. Policy recommendations include strengthening chronic-disease prevention, integrating laboratory-based triage protocols, harmonizing hospital data collection teams, and investing in equitable referral and transport infrastructure to enhance national preparedness for future pandemics.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy (Ph.D.)
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Public Health
- Grantor
- University of South Florida
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Hernandez Bojorge, Santiago E.
- Contributors dc:contributor
-
- Ismael Hoare, Ph.D., MPH
- Benjamin Jacob, Ph.D.
- Jason Salemi, Ph.D., MPH, FACE
- Matthew Valente, Ph.D.
- Ricardo Izurieta, MD, DrPH, MPH
Subjects
dc:subject × 6Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.usf.edu/etd/11302