Air quality and attributable mortality among city dwellers in Kampala, Uganda: results from 4 years of continuous PM_2.5 concentration monitoring using BAM 1022 reference instrument

Ambient air pollution from human activities is a major causal risk factor for non-communicable diseases and premature death, contributing significantly to the global disease burden, with the largest impacts in low- and middle-income countries. Africa’s rapid development and urbanization have increased emissions from household biomass fuel use, windblown dust, agriculture, traffic, and industrialization. Monitoring data in Sub-Saharan Africa are limited but indicate particulate matter levels exceeding WHO Air Quality Guidelines. To inform policy and control efforts in Uganda, this study reports four years (2018–2021) of continuous PM2.5 monitoring in Kampala and estimates the associated mortality burden, providing the first evidence-based estimates of air pollution–associated mortality for the city.

The paper references global and African evidence linking PM2.5 to mortality and disease burden, including estimates of millions of premature deaths attributable to fossil fuel–derived PM2.5 and more than one million deaths in Africa in 2019. It highlights the scarcity of high-quality monitoring and health risk data in Sub-Saharan Africa and notes that existing monitoring indicates PM levels above WHO guidelines. The work complements prior GEOHealth Hub efforts, including a three-year monitoring and health impact assessment in Addis Ababa, and situates Kampala’s findings within broader literature on seasonal and meteorology-driven PM2.5 variation and the contribution of traffic, dust, and biomass burning to urban air pollution.

Design: Time-series analysis using prospectively collected PM2.5 and mortality data for Kampala, Uganda (2018–2021). PM2.5 monitoring: Hourly PM2.5 measured with a U.S. EPA reference Beta Attenuation Monitor (BAM-1022) installed on the Makerere University School of Public Health rooftop within Mulago National Referral Hospital complex. Instrument operation maintained specified ranges for flow rate (15.8–17.5 L/min), relative humidity (10–99%), temperature (1–40 °C), and barometric pressure (650–670 mm Hg). Data were averaged to 1-hour values (24 per day). Quality control included automated error flags, routine maintenance, and semiannual background concentration offset calibration with a BX-302 Zero Filter Calibration Kit, with additional verification at setup. Sampling period: January 1, 2018 to December 31, 2021. Mortality data: Annual all-cause mortality data for Kampala were obtained from the Uganda Bureau of Statistics and the Ministry of Health (DHIS2, Health Information Management Division). Deaths due to accidents (external causes) were excluded. Population inputs: Kampala area 189.3 km², density 9352 persons/km², total population by year (2018–2021), and proportions aged ≥30 years were compiled; accidental deaths were subtracted before final mortality inputs. Data management and completeness: Data were downloaded weekly and evaluated under standard operating procedures. For analysis, days with ≥18 of 24 hourly values were retained. Missing data, primarily from power outages and maintenance interruptions (including COVID-19 lockdown–related service disruptions), were imputed using a before-after mean method: missing values replaced by the mean of the adjacent observations; if adjacent data were unavailable, a 2-day window was used. This process recovered 120 days (8.2%) of expected data. Exposure analysis: Hourly BAM data aggregated to daily averages; descriptive statistics and time-series plots explored diurnal, day-of-week, and seasonal patterns using R (v3.6.2). Health impact assessment: WHO AirQ+ tool used to estimate deaths attributable to long-term exposure to PM2.5. Annual mean PM2.5 concentrations from the BAM-1022 and annual non-accidental deaths were inputs. Health impact estimates were benchmarked against WHO annual PM2.5 guideline (5 µg/m3) and interim targets (IT-1: 35 µg/m3, IT-2: 25 µg/m3, IT-3: 15 µg/m3, IT-4: 10 µg/m3).

- Data completeness: 1069 days over four years (73.2% completeness); missingness driven by power outages and pump failure, including service lapses during COVID-19 lockdown (June–July 2020). - PM2.5 levels: Four-year annual average 39 µg/m3; overall mean across 2018–2021 was 38.8 µg/m3 (SD 18.6; range 1.2–162.9). Levels were roughly eight times the WHO annual guideline (5 µg/m3) and about twice the WHO daily guideline. - Seasonal pattern: Lower concentrations during rainy seasons (March–June and October–December); highest levels typically in December–February, with the highest recorded concentration in February 2021 (dry season). - Diurnal and weekly patterns: Peaks around 09:00 and 21:00, likely reflecting traffic and meteorological influences. Saturdays had the highest daily average PM2.5 (41.2 µg/m3) over the four years. - Attributable mortality (WHO annual guideline 5 µg/m3 as reference): Estimated non-accidental deaths attributable to long-term PM2.5 exposure were 2777 (19.3%) in 2018, 2136 (17.9%) in 2019, 1281 (17.9%) in 2020, and 1063 (19.8%) in 2021; total 7257 deaths over 2018–2021. - Public health impact: Indicates substantial premature mortality burden associated with observed PM2.5 concentrations in Kampala.

The study addressed the research objective by linking continuously monitored PM2.5 levels to mortality estimates using WHO’s AirQ+ tool. Findings show PM2.5 levels far above WHO guidelines with clear temporal patterns: lower concentrations in rainy seasons and higher values during morning and evening traffic peaks and dry seasons. These patterns implicate traffic emissions, non-exhaust sources (brake and tire wear, road dust), biomass and solid waste burning, dust from unpaved roads, and potential nighttime industrial activity. The high fraction of deaths attributable to long-term PM2.5 exposure (about 17–20% yearly) underscores a significant public health burden, especially for vulnerable groups (elderly and those with cardiopulmonary diseases). The results are policy-relevant, highlighting the need for strengthened emission controls, improved fuel and vehicle standards, enhanced public transport, electrification, and enforcement mechanisms. Protection strategies for vulnerable populations during high-exposure periods are warranted.

Ambient PM2.5 levels in Kampala greatly exceed WHO Air Quality Guidelines, contributing to substantial adverse health effects and premature mortality. This four-year, reference-grade monitoring effort provides the first evidence-based estimates of air pollution–attributable mortality for Kampala. Immediate, multipronged policy actions are needed, including curbing household and institutional solid waste burning; promoting reduction, reuse, and recycling of plastics; facilitating clean cooking through subsidies for electricity and LPG and reduced taxes on electric/gas cookers; expanding clean mass transit; enforcing vehicle emission standards and regular maintenance; and consistent traffic rule enforcement. Continued monitoring and research will support effective air quality management and further quantify health benefits of interventions.

- Data gaps due to power outages and instrument pump failure, including missed regular maintenance during the COVID-19 lockdown (June–July 2020). - Single-city, single-site monitoring may limit spatial representativeness within Kampala. - Use of imputation for missing hourly data (before-after mean) introduces uncertainty. - Health impact estimates rely on model assumptions and available mortality data after excluding accidental deaths, which may be subject to reporting limitations.