Uganda Liver Disease study – Dataset for analysis of liver, metabolic and cardiovascular parameters associated with mortality in the General Population Cohort – User Guide

Persistent identifier

10.17037/DATA.00004497

Description

This dataset contains information collected through the Uganda General Population Cohort (GPC) that was used by the Uganda Liver Disease study to assess how liver, metabolic and cardiovascular parameters are associated with all-cause mortality. This includes data on liver function tests (serum AST, ALT, ALP, GGT and BR) obtained in survey rounds 22 (years 2010-11) and 24 (years 2013-14); metabolic biomarkers (hemoglobin A1c, high density lipoprotein, low density lipoprotein, and triglycerides) collected in round 22 (2010-11); screening results for Human Immunodeficiency Virus (HIV) and Hepatitis B Virus (HBV) infection that were obtained in survey round 22, and follow-up data for mortality analyses collected up to study round 34 (2022-23). Death was ascertained through mortality registers and verbal autopsy.

Data collection methods

Primary data were collected using hard copy standardized questionnaires (standard GPC data collection templates). The data were collected from the GPC, located in the Ugandan Kyamulibwa sub-county of the Kalungu district, a prospective population cohort under longitudinal follow up.

Overall, 7896 individuals are represented:

• Liver function tests (serum AST, ALT, ALP, GGT and BR) are included from survey rounds 22 (years 2010-11) and 24 (years 2013-14).
• We also report round 22 data for metabolic biomarkers: hemoglobin A1c (HbA1c), high density lipoprotein (HDL), low density lipoprotein (LDL) and triglycerides.
• Screening results for Human Immunodeficiency Virus (HIV) and Hepatitis B Virus (HBV) infection are included from survey round 22.
• Follow-up data for mortality analyses were collected up to study round 34 (2022-23).
• Death was ascertained by from mortality registers and verbal autopsy.

Data analysis and preparation

The data were checked by the ULIDS analysis team for inconsistencies, out of range values, duplication and other anomalies. The cleaned data were analysed using descriptive statistics, and univariable and multivariable Cox proportional hazards models to investigate factors associated with hazards of death.

Geographic regions

Kyamulibwa sub-county of the Kalungu district, in rural South-Western Uganda.

Key dates

Data were collected from GPC survey rounds 22 (years 2010-11) and 24 (years 2013-14). Follow-up data for mortality analyses were collected up to study round 34 (2022-23). Analysis was undertaken between 2021-2023.

Quality controls

Data were checked for completeness by the ULIDS study team, and cross-checked with the UVRI data management team in the case of missingness or discrepancies to minimise errors and optimise completeness in the dataset.

Population

This study focuses on the human population of participants enrolled in the GPC. We used laboratory data from survey rounds 22 (years 2010-11) and 24 (years 2013-14), and mortality data up to round 34 (2022-23). In most survey rounds, all individuals aged ≥13 years planning to reside ≥3 months in a household located within the study region were eligible for inclusion, with age restrictions relaxed to include younger participants in certain survey rounds. GPC participants have been prospectively followed-up to detect relevant endpoints, and participant data from the survey are linked to data collected at the Medical Research Council (MRC) / Uganda Virus Research Institute (UVRI) clinic via unique individual identifiers.

Privacy

Data were stored in pseudonymised format in accordance with GPC protocols. No identifiable data (e.g. name, date of birth) were shared with the ULIDS study team.

The following masking was performed when preparing the data for submission to the LSHTM repository.

• Age has been categorized into age group
• Dates have been rounded. Day has been floored to 01 (i.e. the first of the month) in all instances. Month-year remain valid.
• The persona identifier primary key has been replaced with a random 8-digit integer.
• The original unique lab identifier primary key “labno” has been replaced with a random 6-digit integer (named “lab_identifier”).

Ethics

The study obtained ethical approvals from the Science and Ethics Committee of the Uganda Virus Research Institute (GC/127/12/11/06), the Ugandan National Council for Science and Technology (HS870) and the East of England-Cambridge South (formerly Cambridgeshire 4) NHS Research Ethics Committee UK (11/H0305/5).

Keywords

Liver function tests, hypertension, diabetes, MASLD, mortality, HIV, HBV, Uganda, epidemiology

Language of written material

The study analysis and reporting has been conducted in English. Study paperwork was also available in Luganda for participants.

Project information

Creators

File description