Clinical characteristics and factors associated with COVID-19-related mortality and hospital admission during the first two epidemic waves in 5 rural provinces in Indonesia: A retrospective cohort study

Background Data on coronavirus disease 2019 (COVID-19) clinical characteristics and severity from resource-limited settings are limited. This study examined clinical characteristics and factors associated with COVID-19 mortality and hospitalisation in rural settings of Indonesia, from 1 January to 31 July, 2021. Methods This retrospective cohort included individuals diagnosed with COVID-19 based on polymerase chain reaction or rapid antigen diagnostic test, from five rural provinces in Indonesia. We extracted demographic and clinical data, including hospitalisation and mortality from a new piloted COVID-19 information system named Sistem Informasi Surveilans Epidemiologi (SISUGI). We used mixed-effect logistic regression to examine factors associated with COVID-19-related mortality and hospitalisation. Results Of 6,583 confirmed cases, 205 (3.1%) died and 1,727 (26.2%) were hospitalised. The median age was 37 years (Interquartile range 26–51), with 825 (12.6%) under 20 years, and 3,371 (51.2%) females. Most cases were symptomatic (4,533; 68.9%); 319 (4.9%) had a clinical diagnosis of pneumonia and 945 (14.3%) presented with at least one pre-existing comorbidity. Age-specific mortality rates were 0.9% (2/215) for 0–4 years; 0% (0/112) for 5–9 years; 0% (1/498) for 10–19 years; 0.8% (11/1,385) for 20–29 years; 0.9% (12/1,382) for 30–39 years; 2.1% (23/1,095) for 40–49 years; 5.4% (57/1,064) for 50–59 years; 10.8% (62/576) for 60–69 years; 15.9% (37/232) for ≥70 years. Older age, pre-existing diabetes, chronic kidney disease, liver diseases, malignancy, and pneumonia were associated with higher risk of mortality and hospitalisation. Pre-existing hypertension, cardiac diseases, COPD, and immunocompromised condition were associated with risk of hospitalisation but not with mortality. There was no association between province-level density of healthcare workers with mortality and hospitalisation. Conclusion The risk of COVID-19-related mortality and hospitalisation was associated with higher age, pre-existing chronic comorbidities, and clinical pneumonia. The findings highlight the need for prioritising enhanced context-specific public health action to reduce mortality and hospitalisation risk among older and comorbid rural populations.


Unfunded studies
Enter: The author(s) received no specific funding for this work.

59
Conclusion Clinical characteristics and risk factors of severe COVID-19 outcomes in rural 60 provinces were broadly similar to those in urban settings. The risk of COVID-19-related 61 mortality and hospitalisation was associated with higher age, pre-existing chronic 62 comorbidities, and clinical presentation of pneumonia.

101
The majority of cases and deaths in Indonesia were reported on Java Island, a more 102 developed setting populated by 152 million individuals (56% of Indonesia's total population).

103
Recent studies from Indonesia's capital city of Jakarta suggested that COVID-19 104 disproportionately affected individuals with older age and pre-existing chronic comorbidities, 105 as well as those areas with lower vaccine coverage and higher poverty and population density 106 [8,15]. However, data on the impact of COVID-19 in less developed areas of Indonesia with 107 greater geographic, cultural and socio-economic diversity, outside of the island of Java, are 108 scarce.

110
The limited evidence regarding COVID-19 epidemiology and severity from low resource 111 settings in Indonesiaespecially in areas outside of Java -has been particularly associated 112 with the lack of epidemiological data recorded by the surveillance systems in these areas. To

169
We used bivariable and multivariable mixed effects logistic regression models to determine 170 the risk of death and hospitalisation, expressed as odds ratios with 95% confidence intervals.

252
·· the Mann-Whitney U test was not performed due to identical median and IQR between group.

258
Details on age-specific mortality and hospitalisation rate can be seen in Table 2.

267
In bivariable mixed effects logistic regression analysis (Supplementary Table 3

282
·· the variable did not enter the multivariable model therefore statistics were not estimated.

284
Based on the first multivariable mixed effects logistic regression model (

301
In the second multivariable model (Supplementary immunocompromised conditions were associated with higher risk of hospitalisation, but not 324 associated with mortality. Gender and number of health care workers at province-level were 325 not associated with COVID-19-related mortality and hospitalisation. The overall mortality rate in these five rural provinces was higher than the mortality rate 327 reported from Jakarta (3% vs 1.5%), potentially due to a higher prevalence of at least one 328 comorbidity in these settings (14% vs 0.7%). However, our findings suggest that the 329 hospitalisation rate was slightly lower than in Jakarta (26% vs 33%), possibly due to a lower