Incidence of postpartum and neonatal illnesses and utilization of healthcare services in rural communities in southern Ethiopia: A prospective cohort study

Although improving postpartum and neonatal health is a key element of the Ethiopian health extension program, the burdens of postpartum and neonatal illnesses and healthcare-seeking in rural communities in Ethiopia are poorly characterized. Therefore, we aimed to assess the incidence and risk factors for these illnesses and measure the utilization of healthcare services. We conducted a prospective cohort study of 784 postpartum women and their 772 neonates in three randomly selected kebeles in rural southern Ethiopia. Eight home follow-up visits were conducted during the first 42 postpartum days, and six neonate follow-ups were conducted at the same home over the first 28 days of life. The Prentice, Williams, and Peterson’s total time Cox-type survival model was used for analysis. We recorded 31 episodes of postpartum illness per 100 women-weeks (95% confidence interval [CI]: 30%, 32%) and 48 episodes of neonatal illness per 100 neonate-weeks (95% CI: 46%, 50%). Anemia occurred in 19% of women (95% CI: 17%, 22%) and low birth weight (<2,500g) in 15% of neonates (95% CI: 13%, 18%). However, only 5% of postpartum women (95% CI: 4%, 7%) and 4% of neonate (95% CI: 3%, 5%) reported utilizing healthcare services. Walking over 60 minutes to access healthcare was a factor of both postpartum illnesses (AHR = 2.61; 95% CI: 1.98, 3.43) and neonatal illnesses (AHR = 2.66; 95% CI: 2.12, 3.35)). Birth weight *2500g was identified factor of neonatal illnesses (AHR = 0.39; 95% CI: 0.33, 0.46). Compared with younger mothers, older mothers with sick newborns (AHR = 1.22; 95% CI: 1.00, 1.50) or postpartum illnesses (AHR = 1.40; 95% CI: 1.03, 1.89) were more likely to seek healthcare. Reasons for not utilizing healthcare services included a belief that the illnesses were not serious or would resolve on their own, little confidence in the healthcare institutions, and the inability to afford the cost. The burden of postpartum and neonatal illnesses in rural communities of southern Ethiopia remains high. Unfortunately, few participants utilized healthcare services. We recommend strengthening the health system that enables identifying, managing, treating, and referring maternal and neonatal illnesses and provide reasonable healthcare at the community level. A prospective cohort study of 784 women who had recently given birth and their 772 neonates was conducted from May 2017 to July 2018 in three randomly selected kebeles (i.e., Mekonisa, Hase-Haro, and Tumata-Chiricha) in the Wonago district of southern Ethiopia, which is located 420 km from the city of Addis Ababa. Each woman was followed up eight times at home during her first 42 postpartum days, and each neonate was followed up six times at the same home over the first 28 days of life. The details of this cohort are described in our previous study about pregnancy-related illnesses [22]. During our study, some women and neonates could not be contacted because of social unrest in a small part of our study area and were therefore excluded. We defined postpartum illnesses as any disorder after childbirth that hurt a woman’s health [23]. Neonatal illnesses were similarly defined as any disturbance of the normal state of the body and organs of the neonate [16]. We defined utilization of healthcare services as any use of healthcare services or any visit to a healthcare facility to get outpatient and inpatient healthcare services to treat postpartum or neonatal illnesses [24]. We defined healthcare services according to the Encyclopedia of Behavioral Medicine as “an array of medical care or services that are recognized under state law and are performed by healthcare professionals or under their direction, for promoting, maintaining, or restoring health to those in need (i.e., patients, families, and communities) in either of all settings of care (i.e., health posts, health centers, hospitals, and homes)” [24]. In Ethiopia, the health service is restructured into a three-tier system: primary, secondary, and tertiary level of healthcare, and the healthcare system is organized based on the type of care provided [25]. In the primary tier system; the primary healthcare unit consists of a health center and five satellite health posts. One health center is for 15,000–25,000 people in rural areas, and 40,000 people in urban areas and each health post is for 3,000–5,000 people in rural areas. The primary hospital serves 60,000–100,000 people. The secondary tier system includes general hospital which is for 1–1.5 million people, and the tertiary tier system also includes a specialized hospital that is for 3.5–5 million people. The primary tier system embraces all the healthcare services provided at all levels, and at health posts, most primary healthcare services are provided by health extension workers. These lay individuals are not nurses or trained healthcare professionals, although they have received one-year training in primary healthcare. Health extension workers can address issues related to infectious diseases (e.g., tuberculosis, malaria), communicable and sexually transmitted infections (e.g., HIV/AIDS), maternal and child health, common nutritional disorders, hygiene, and environmental health, immunization and family planning, and reproductive health. Most pregnant women in Ethiopia who seek healthcare use their local health posts. If the case is serious, the Health Extension Worker may refer them to a health center [25]. In our study, women with postpartum illnesses were identified based on symptoms and clinical measurements, such as hemoglobin and blood pressure levels [23], recorded by trained data collectors. Hemoglobin was measured at the end of the 6-week postnatal period using the HemoCue analyzer Hb 301 System (www.hemocue.com). We defined anemia as a hemoglobin value of <12 g/dL at six weeks after childbirth [26]. Blood pressure was measured during each visit using a Riester ri-champion N digital apparatus (www.riester.de). High blood pressure was defined as >140/90 mm Hg [27]. Neonates with neonatal illnesses were identified based on reported symptoms by the mother and some measurements, such as neonates’ weight and length were measured using the Health O Meter® Portable Home Care Baby Scale (Pelstar, LLC; www.chichestershomecare.net). Neonatal mortality was defined as the death of the neonate within the first 28 days of life. Low birth weight was defined as <2,500g, measured within the first 48 hours of life. As birth weight is closely associated with neonatal mortality and morbidity and is used as a public health indicator [28], we recorded birth weight as an exposure variable together with other neonatal illnesses [28]. We assessed two primary outcomes for mothers and their neonates: illnesses and utilization of healthcare services. We measured these outcomes in two ways: as a count and as a dichotomous value (0 = no, 1 = yes). We assessed 10 symptoms and signs of postpartum illnesses [29,30]: hemorrhage, high fever, foul-smelling vaginal discharge, blurred vision with severe headache, severe abdominal pain, urinary incontinence, breast pain and engorgement, severe tiredness, anemia, and hypertension. S1 Table summarizes the symptoms of postpartum illnesses. We also assessed 12 symptoms and signs of neonatal illnesses [30,31]: not sucking properly, high fever (≥37.5°C), diarrhea, cord stump with redness or pus, persistent vomiting, hypothermia (≤35.5°C), fast breathing (≥60 breaths per minute), severe chest in-drawing, no spontaneous movement, jaundice, red or discharging eye, and lethargy. S2 Table summarizes the symptoms of neonatal illnesses. The exposure variables for illnesses included participant characteristics and community-level variables. The women’s basic characteristics included her age, age at first marriage, age at first birth, birth interval (≥2 years or <2 years between births), occupation (domestic service or other), household wealth index, total monthly household expenditure (more or less than $30), gravidity, parity, history of abortion, history of stillbirth, marital status, and educational status. The neonates’ exposure characteristic was the birth weight (≥2500g or <2,500g). Low birth weight <2,500g was regarded as an exposure variable. Community-level variables included the type of road (asphalt or other) and walking distance to the nearest health facility. The exposure variables for the utilization of healthcare services were postpartum and neonatal illnesses and all previously listed exposure variables. Continuous variables were assessed for symmetry, and parametric tests were used for normally distributed variables. To assess the sample, we made some assumptions based on our previous study [22]. We assumed a 15.5% incidence of illnesses among pregnant women and a 1.65 relative risk among poor women, compared with rich women (95% confidence level [CI], 80% power, and 1:1 ratio of unexposed to exposed) [32]. The sample size was estimated at 898 after adding 10% for non-responses. S1 Fig summarizes the flowchart of the recruitment of participants. During the initial recruitment, 898 pregnant women were included in the study. Of those, 86 were excluded due to incomplete data, 14 had abortions, two were not pregnant, and instead had ovarian cysts, one died in an accident, and one refused to participate. Thus, 794 pregnant women were included in the analysis of pregnancy-related illnesses. During the follow-up visits, 10 women with incomplete data were excluded. The final sample thus included 784 postpartum women (S1 Fig). Neonates born from 794 pregnant women were eligible for inclusion in the study. Out of 808 births, 782 were live births and 26 were stillbirths. Thirteen women had multiple births, and 781 women had singleton births. Of the 782 live births, we excluded 10 due to incomplete data, leaving a final sample of 772 neonates (S1 Fig). Baseline socioeconomic and demographic data and follow-up data were collected via a questionnaire during visits to participants’ homes. Data from postpartum women were collected eight times: within 24 hours and within 24–72 hours after birth and at the end of the first, second, third, fourth, fifth, and sixth weeks. The questionnaire was adapted from the WHO Maternal Morbidity Measurement Tool [33]. Data from neonates were collected six times at the same home visits: within 24 hours and within 24–72 hours of birth and at the end of the 7th, 14th, 21st, and 28th days. The questionnaire was adapted from the Johns Hopkins University Tools and Indicators for Maternal and Newborn Health [34]. Postpartum and neonatal questionnaires were prepared in English, translated into the local Gedeo (see S1 File Gedeo questionnaire for postpartum women and neonates.rar) and Amharic (see S2 File Amharic questionnaire for postpartum women.rar and S4 File Amharic questionnaire for neonates.rar) languages, and then translated back into the English language (see S3 File English questionnaire for postpartum women.rar and S5 File English questionnaire for neonates.rar). The trained data collectors read the symptoms aloud and then asked the women to indicate whether they or their neonates had any of the symptoms, whether they utilized healthcare services, and reasons why they did not seek healthcare during illnesses. To assess the need for the utilization of healthcare services during postpartum and neonatal illnesses, a community-based approach was used rather than a facility-based approach [35]. The data were entered in EpiData version 3.1 software (EpiData Association, Odense, Denmark). Descriptive statistical analysis was used to determine the distribution of illness incidences and the utilization of healthcare services. In this paper, one statistical model was used to analyze recurrent event data. The interpretation of results and analyses of risk factors were based on Prentice, Williams, and Peterson’s total time Cox-type survival model [36,37]. This model is a robust option for recurrent events of illnesses and the utilization of healthcare services. To control for the effect of missing values, the analysis was restricted to women and neonates with complete data [38]. Correlations among variables during pregnancy and postpartum periods or during neonatal and pregnancy periods also were assessed. STATA software version 15 was used for analysis (Stata Corp LLC, College Station, TX, USA). Detailed information on the methods, study design, procedures, sample size, statistical methods, and major findings of illnesses during pregnancy are presented in our previous study [22]. This study was approved by the institutional ethical review board at Hawassa University, College of Medicine and Health Sciences (IRB/100/08), and by the Regional Committees for Medical and Health Research Ethics of Western Norway (2016/1626/REK vest). Written permission was obtained from the Gedeo Zone health department and the Wonago district health office. Written informed consent was obtained from each woman after she received an explanation of the purpose of the study. The privacy, anonymity, and confidentiality of all participants were maintained. If a data collector observed any illness among participants during the study period, they tried to link the mother or child with health extension workers in the kebele.