Birth preparedness, complication readiness and other determinants of place of delivery among mothers in Goba District, Bale Zone, South East Ethiopia

Background: Ethiopia is one of the countries with the highest maternal mortality ratio 676/100,000 LB and the lowest skilled delivery at birth (10 %) in 2011. Skilled delivery care and provision of emergency obstetric care prevents many of these deaths. Despite implementation of birth preparedness and complication readiness packages to antenatal care users since 2007 in the study area, yet an overwhelming proportion of births take place at home. The effect of birth preparedness and complication readiness on place of delivery is not well known and studied in this context. Methods: A community based case control study preceded by initial census was conducted on a total of 358 sampled respondents (119 cases and 239 controls) who were selected using stratified two stage sampling technique. A pre-tested and standardized questionnaire with a face-to-face interview was used to collect the data, and then data was cleaned, coded and entered in to SPSS version-21 for analysis. Binary logistic regression models were run to identify predictors of place of delivery and Odds ratio with 95 % CI was used to assess presence of associations at a 0.05 level of significance. Results: The mean (± Standard Deviation) age of respondents was; 27.41(±5.8) and 28.84(±5.7) years for the cases and the controls respectively. Two third (67.1 %) of the childbirths took place in the respondents house while only (32.9 %) gave birth in health facilities. Great proportion (79.7 %) of the cases and two third (34.0 %) of the controls were well-prepared for birth and complication. Maternal education, religion, distance from health facility, knowledge of availability of ambulance transport and history of obstetric complication were significantly associated with place of delivery (P-value <0.01). Birth preparedness and complication readiness practice had an independent effect on place of delivery (AOR =2.55, 95 % CI: 1.12, 5.84). Conclusion: The study identified better institutional delivery service utilization by mothers who were well-prepared for birth and complication. Strategies that increase the preparedness of mothers for birth and complication ahead of childbirth are recommended to improve institutional delivery service utilization. The study was conducted from February 10th to 5th of March, 2014 in Goba District, which is one of the 18 Districts in Bale Zone, Oromia regional state, located at 444 Km in the South-East direction from the capital city Addis Ababa. Goba District is administratively structured into: Goba rural District and Goba administrative town, with 15 rural and 2 urban Kebeles (the smallest administrative unit with minimum of 1000 households) respectively. The estimated total population of the study district was 89,859 (projected from the 2007 census) [11]. There were an estimated 20,668 women of child bearing age (15-49 years) and 3774 pregnant women in the district during the study period. The public health infrastructure of the district involves; one referral hospital, four health centres and 15 functional health posts. The referral hospital in the town provides comprehensive emergency obstetric care while the health centres provide basic emergency obstetric care. The Health Posts are the lowest health facilities in Ethiopia used to improve equitable access to basic health services, mainly disease prevention and health promotion through a community (Kebele) based health extension program. Each health post is staffed by two female health extension workers (HEWs) who are assigned after completing one year of training. And the HEWs provide family planning, immunization, ANC, clean delivery service and postnatal care. Furthermore, they are responsible for referring women with obstetric complications to health centers and hospitals where basic and comprehensive emergency obstetric care is available. A community based unmatched case control study design; preceded by initial census i.e. to enumerate and prepare list of all households with eligible women in the sampled kebeles was used. The cases were mothers who gave childbirth in health institution and the controls were mothers who gave childbirth at home. Childbearing women who gave childbirth in the last 12 months (February 10/2013 to February 09/ 2014) in the District, regardless of the birth outcome were included in the sample. The required sample size of eligible mothers for the study was determined with a formula to estimate two population proportion, using “Epi-info7 Stat calc sample size for unmatched case control study” with the assumptions of; a 95 % confidence level, 80 % power, level of exposure to BPACR package in the control group of 30.0 % [8], an expected odds ratio of 2.0, the ratio of controls to cases of 2:1, and an additional 5 % non-response rate. The calculated total sample size was 358 mothers who gave childbirth in the 12 months before the study period, with the number of sampled cases (n = 119) and sampled controls (n = 239). Initially the study area was stratified in to rural and urban, and there were fifteen rural and two urban kebeles during the study time. Five rural and one urban Kebele were selected using lottery method. A house to house census of the sampled Kebeles was done and a total of 864 eligible households with women who gave childbirth from 10th of February 2013 to 09th of February 2014 (regardless of birth outcome) were identified and sample frames with a list of 285 eligible households with a mother who gave childbirth in a health facility and 579 eligible households with a mother who gave childbirth at home were prepared. From the selected six Kebeles, 119 households with women who gave childbirth in a health facility (cases) and 239 households with women who gave childbirth at home (controls) were selected by simple random sampling technique. If the households were locked or the mothers were not available at the time of data collection, frequent visits were made until the data collectors could communicate with them during the data collection period. A lottery method was used in cases where there was more than one eligible woman in a single household. Structured and pre-tested questionnaire was prepared first in English and then translated into Afan Oromo and Amaharic, local languages. Eight nurses had conducted the face to face interview and a health extension worker in each Kebele was used as a local guider during the preliminary census and the data collection periods. The principal investigator and two public health officers supervised the whole data collection process. Training was given to the data collectors and supervisors before the actual data collection regarding the aim of the study, data collection tool (going through each question), data handling, sampling procedure and interview techniques. The questionnaires were reviewed daily by the supervisors and the principal investigator to check for completeness and early corrections and cleaning of the data were made. Data on place of delivery; on mothers’ age, religion, ethnicity, marital status, educational status, decision making of women on obstetric health care seeking, income, family size; on husbands age, education and occupation; availability of Television, Radio & Telephone in the household; on gravidity, parity, age at first delivery, number of abortions, still birth and live birth, obstetric complications experienced, antenatal care use, gestational age at the first ANC, the number (frequency) of ANC; awareness on availability and accessibility of health workers and skilled delivery care, service fee, knowledge of availability of free ambulance transportation service, average time of travel to the nearest health facility with emergency obstetric care; and on birth preparedness and complication readiness knowledge and practice of the mother were collected. The data was checked visually for completeness, coded and entered into SPSS version 21 soft-ware package. Frequency distribution of the variables was used to check data entry errors and consistency was checked by doing double data entry on 10 % of the questionnaire. The results were presented in the form of tables, figures and texts. Frequencies and summary statistics such as mean, standard deviation and percentages used to describe the study population (cases and controls) in relation to relevant variables. Binary logistic regressions models were used to determine the effects of BPACR (the main exposure variable) and other covariates on the outcome variable (place of delivery). Variables were recruited for multivariable analysis based on findings from the bivariate analysis. To be a candidate for multiple logistic regressions; variables whose p-value < 0.25 along with the variables of known clinical importance were considered and we compared the coefficients of each variable with the coefficient from the model containing only that variable and tried to verify the importance of each variable in the multiple model using Wald statistic and variable that doesn’t appear to be important were eliminated, and new model were refitted until the important covariates were included. Furthermore lists of possible pairs of variables in the main effects model that have some scientific basis to interact with each other were formed. Then we added the interaction terms, one at a time, in the model containing all the main effects and then assessed its significance using the likelihood ratio test and dropped any non-significant interaction. Finally the overall goodness-of-fit of the Model was assessed using the Hosmer and lemeshow’s test. Adjusted Odds ratios and 95 % confidence intervals were computed for each explanatory variable to determine the strength of association with place of delivery while controlling the effect of potential confounders, at a p-value ≤ 0.05. Ethical approval was secured from designated institutional review committee of the Addis Ababa University, School of Public Health. A formal letter in request of cooperation was written to Goba District Health Offices. Information on the purpose of the study and the right not to participate were given to the participants and they were also informed that all data obtained from them would be kept confidential by using codes instead of any personal identifiers as presented in Additional file 1. Oral consent was obtained from each study participant. Information on the importance of getting services from skilled attendants during pregnancy, delivery and postnatal period were provided by the data collectors to the participants who delivered outside health institutions at the end of data collection. The study was adherent to the STROBE criteria as outlined in Additional file 2.