Association between maternal age and adverse perinatal outcomes in Arba Minch zuria, and Gacho Baba district, southern Ethiopia: A prospective cohort study

Background: Globally, delayed childbearing to the advanced age is a growing option. It is an emerging public health issue in developing countries. Currently, adverse perinatal outcomes significantly increased. A few studies showed the effect of advanced maternal age on adverse perinatal outcomes. However, most used secondary data or chart reviews, and this increases the risk of biases. Besides, there are limited studies in-country Ethiopia as advanced maternal age steadily increased. Therefore, this study aimed to assess the status of advanced maternal age and its effect on perinatal outcomes in the study setting. Methods: A community-based prospective cohort study was conducted among 709 study participants from October 15, 2018, to September 30, 2019, in Arba Minch zuria, and Gacho Baba district, southern Ethiopia. The data were collected by a pretested interviewer-Administered structured Open Data Kit survey tool and analyzed by SPSS version 25. The log-linear regression model was used to compare perinatal outcomes among women aged 20-34 years and ≥ 35 years. The log-likelihood ratio tested for the goodness of fit. In this study, P-value < 0.05 was considered to declare a result as a statistically significant association. Results: In this study, 209(29.5%) of the women were age group ≥35 years old, and 500(70.5%) were age group from20-34 years old. Stillbirth (β = 0.29, 95%CI: 0.05, 0.52), and neonatal mortality (β = 0.11, 95%CI: 0.01, 0.21) were significantly associated with the advanced maternal age. Conclusions: Perinatal outcomes such as stillbirth and neonatal mortality were independently associated with advanced maternal age after controlling for possible cofounders. Therefore, different strategies should design for the women who planned to bear child, and information should provide for women who are advanced age or delayed childbearing to alert them. In this study, women’s in Arba Minch zuria, and Gacho Baba district, Arba Minch-Health, and Demographic Surveillance System sites (AM-HDSS), southern Ethiopia involved, from October 15, 2018, to September 30, 2019. Arba Minch-Health and Demographic Surveillance System sites were established in collaboration between Arba Minch University and Ethiopian Public Health Association with the support of the Centers for Disease Control and Prevention (CDC) Ethiopia in 2009 to track demographic changes. The surveillance site included nine kebeles from the 29 kebeles located in Arba Minch zuria, and Gacho Baba district, Gamo zone, southern Ethiopia [28]. Arba Minch is an administrative town in the Gamo zone, located 505 km south of Addis Ababa and 275 km southwest of Hawassa. Based on the 2007 Census conducted by the Central Statistical Agency (CSA), these districts have a total population of 164,529, of whom 82,199 are men and 82,330 women. According to the HDSS report, there is a total population of 74,157 in the surveillance site. A community-based prospective cohort study design was employed to meet study objectives. The source population for this study was all women who were pregnant in Arba Minch zuria, and Gacho Baba district, AM-HDSS site, southern Ethiopia. Those women who were pregnant during the study period (2018–2019) were study population for this study. At enrollment for this study, all women who were pregnant and inhabitants to a minimum of six months in the study area were eligible for this study. The eligibility defined by the pregnancy screening checklist, which was developed by Whiteman et al. [29]. During recruitment, all women whose ages less than twenty years old and known to be preexisting illnesses excluded from the study. Epi info7 software Stat Calc used to estimate the sample sizes. For the first objective, a single population proportion was used by considering the following assumptions: P = 0.334 from the study conducted in Norway [9], 95% level of confidence, and 5% margin of error used. Based on this, the estimated sample size was 342. A two-sample comparison proportion used to estimate the sample size for the second objective. The assumption was P1 (age group 20–34) = 0.207 and P2 (age group ≥35) = 0.124 in the study conducted in Malaysia [10], 95%CI, ratio 1:1, and Power = 80% and the sample size estimated by this assumption was 676. The sample size for this study estimated by adding a non-response rate of 10% to the larger sample size. Therefore, the calculated sample size for this study was 744. The data were collected using a pretested interviewer-administered structured Open Data Kit (ODK) survey tool. The tools were developed by reviewing different works of literature. The wealth index assessment questionnaire adapted from the questionnaire used in the Ethiopian Demographic Health Survey (EDHS) 2016 [30]. The household food insecurity level measured with Household Food Insecurity Access Scale (HFIAS), a structured, standardized, and validated tool that developed mainly by Food and Nutrition Technical Assistance (FANTA) [31]. They have three main parts for the questionnaire: Part I (pregnancy screening checklist), Part II (baseline information), and Part III (follow-up survey tool) (Additional file 1). The tools pretested in the Chencha district, which was out of the study area to verify the appropriateness, and modifications and amendments were taken accordingly before actual data collection. The well-trained nine data collectors and three field supervisors were prospectively identified perinatal outcomes among pregnant women during the study period. Intensive three days training gave for data collectors and supervisors separately regarding objectives of the study and data collection ways. Data collectors discussed the information about the ODK survey tool and pregnancy screening checklists to identify pregnant women. The data collected in different phases, as this was a community-based prospective follow-up study. In the first phase: all the baseline information about the women obtained and pregnancy status was checked by using a pregnancy-screening checklist. After identified whether women were advanced age or not, and the data collectors have recruited the women into the cohort. In the second phase: the women were followed started from the time pregnancy confirmed up to the immediate postpartum period to identify some of the perinatal outcomes. The follow-up terminated at the end of the neonatal period that the neonates reassessed with a similar fashion in the above mechanism. In the community setting, the data collectors frequently contacted women or any household members, surround health care institutions, and health extension workers during the follow-up period. The description and measurements for some of the outcome and explanatory variables were stated in detail below (Table 1). Measurements to assess the status of advanced maternal age and their effect on perinatal outcomes in Arba Minch zuria, and Gacho Baba district, southern Ethiopia, 2018/9 Classified into underweight (< 18.5 kg/m2), normal (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), obese (30–34.9 kg/m2), and morbidly obese (≥35 kg/m2). To ensure quality, experts translated questionnaires into the local language. A standard tool, which was commented by many experts, was used to collect the information. The data collectors and supervisors trained to standardize and ensure consistency of data collection. The principal investigator and supervisors critically checked the data for completeness before uploaded to the ODK cloud server. Multiple imputation techniques used for the missed data that were not more than 20% of the needed information. The inconsistent data excluded from the final analysis. The data coded correctly and categorized to maintain quality. The collected data were downloaded from ODK aggregate and exported to SPSS version 25 for analysis. Proportions and summary statistics computed with maternal age. The wealth quintiles determined by the Principal Component Analysis (PCA). A crude and adjusted log-linear regression analysis was done for each outcome variable with maternal age to estimate the beta coefficient (β). The assumptions for log-linear regression checked, and the log-likelihood ratio tested for the goodness of fit. To control the confounding effect, all the variables with P ≤ 0.25 in the bivariate analysis included in the final model. The model adjusted for educational and occupational status, parity, wealth index, body mass index (BMI), HFIAS, lifestyle factors, distance to health care institution, and sex of the neonate, antenatal care, postnatal care, and place of delivery. A standard error of > 2 considered as suggestive of the existence of multi co-linearity. In this study, P < 0.05, considered to declare a result as a statistically significant association. Then the information, presented in simple frequencies, summary measures, tables, and figures.