Study Design, Period and Setting
This study employed a facility-based unmatched case-control design to investigate the determinants of puerperal sepsis among mothers who delivered at Jigjiga public hospitals between January 2022 and December 2022. The Somali Regional State is a regional state located in eastern Ethiopia, which holds the second-largest territory in the country after the Oromia Region. It is bordered by the Ethiopian regions of Afar and Oromia to the west, Djibouti to the north, Somalia to the northeast, east, and south, and Kenya to the southwest. Jigjiga is the capital city, located 621 kilometers from Addis Ababa, Ethiopia's capital city. The city has a population of approximately 304,400 people, comprising 151,422 males and 153,578 females [11]. The study was conducted at two public hospitals in Jigjiga, specifically Sultan Sheikh Hassan Yabare Comprehensive Specialized Hospital and Karamara General Hospital. These hospitals provide health services to the local community as well as serving an estimated population of more than 7 million people living in the region, including neighboring districts of the Oromia region, region in northern Somalia (approximately 3.5-5 million according to the standard). The hospitals are responsible for training, teaching, and research for Jigjiga University's College of Medicine and Health Sciences, as well as internship training for students from other health science colleges in the region.
Study Population
The study population comprised all postpartum mothers who delivered at Jigjiga Public hospitals over a 12-month period. The case group consisted of mothers who were admitted with puerperal sepsis, while the control group comprised mothers who did not develop puerperal sepsis and remained hospitalized for more than 24 hours.
Eligibility Criteria
Cases
Postpartum mothers who developed puerperal sepsis within 24 hours after delivery or within 42 days after delivery admitted to Jigjiga Public hospitals during the data collection period were included in the study. Postpartum mothers who had a history of previous puerperal sepsis, who had pre-existing medical conditions, and who had incomplete or inaccurate medical records admitted to Jigjiga Public hospitals during the data collection period were excluded from the study.
Controls
Postpartum mothers who did not develop puerperal sepsis within 24 hours after delivery or within 42 days after delivery, gave birth normally, had a normal pregnancy outcome admitted to Jigjiga Public hospitals during the data collection period were included in the study.
Sample Size Determination
The sample size was determined using the double population proportion formula, with Epi Info 7, taking into account the prevalence of puerperal sepsis based on a previous study in West Ethiopia [12], the desired level of precision (5%), a power of 80%, and an expected odds ratio of 3.4. Therefore, the calculated sample size was calculated to be 216 (72 cases and 144 controls), and considering a 5% non-response rate, the final sample size for this study was 227 (76 cases and 151 controls).
Sampling Procedure
Consecutive sampling was used to select cases from postpartum mothers with puerperal sepsis admitted to the maternity wards of the two public hospitals Sultan Sheikh Hassan Yabare Comprehensive Specialized Hospital and Karamara General Hospital. The total sample size was allocated proportionally to each hospital based on the number of registered puerperal sepsis cases in each hospital. Cases were selected from the medical records of mothers who had developed puerperal sepsis during their postpartum stay. Controls were selected from the same population, but were mothers who did not develop puerperal sepsis during their postpartum stay. Every third case of puerperal sepsis and every third mother without puerperal sepsis were selected from the hospital records until the desired sample size was reached.
Data Collection Tool and Procedure
A standardized questionnaire was utilized to collect data for this study, which was adapted from a previous study conducted in Ethiopia [4, 8, 13]. The questionnaire was initially developed in English and subsequently translated into Somali language, followed by a back-translation into English by independent language experts to ensure consistency between the two versions. The questionnaire comprehensively covers all relevant information necessary to meet the objectives of the study, categorized into the following sections: socio-demographic characteristics, obstetric factors, and maternal and co-morbidities. All collected data were recorded on a specially designed data collection sheet for the study. A team of trained data collectors and supervisors was recruited for this study. Two midwives and two nurses served as data collectors, while a health officer acted as a supervisor. Prior to data collection, a comprehensive one-day training session was conducted by the principal investigator, emphasizing the study's objectives, data recording methods, data quality control measures, and confidentiality protocols. The training also covered the contents of the questionnaire and the approach to be taken when interacting with study participants (cases and controls), with the aim of minimizing information bias and ensuring high-quality data collection.
Data Quality Control
To assure the quality of data, adequate training was given to data collectors and supervisor for one-day duration, the objective of the training is to clearly explain data collectors and supervisor the objective of the study to strengthen their skills about how to interview and approach with the study participants during the actual data collection. The questioners were pre-tested on 5% (4 Cases and 8 Controls) of the total sample size at Jigjiga public hospitals. The Questionnaires were checked for their completeness, validity and consistency by the principal investigator.
Data Analysis
The collected and cleaned data were subsequently entered into Epi data version 3.1 and exported to SPSS version 26 for analysis. Descriptive statistics were performed to examine the distribution of variables, and the results were presented in text, frequency tables, and proportion formats. The Pearson Chi-square test was employed to identify the determinants of puerperal sepsis among mothers who gave birth in Jigjiga public hospitals. Binary logistic regression analysis was conducted to investigate the relationship between puerperal sepsis and the explanatory variables. Variables with a p-value less than 0.25 in the bivariate analysis were selected as candidate variables for multivariate logistic regression modeling to account for confounding variables. The model fitness was subsequently assessed using the Hosmer-Lemeshow test. The strength of the associations between independent and dependent variables was evaluated using odds ratios with 95% confidence intervals. Finally, a statistically significant level was declared at a p-value less than 0.05.