Study settings and design
Hospital based cross-sectional study was conducted at Debre tabor general Hospital Obstetrics and Gynecology department, high risk and labor ward from March 14 - June 20, 2019 GC. This hospital is found in South Gondar zone, Amhara Regional state of Ethiopia. It is located 103 km from Bahir Dar, the capital city of Amhara regional state and 667km from Addis Ababa, the capital city of Ethiopia.
According to 2015 Debre Tabor town health statics report, the total population in the town was around 83081and from those, 43301 (52.12%) were Females and 39780 (47.88%) were males. The town has one general hospital, three health center and four private clinics.
Debre tabor general Hospital is having around 5 major wards and gives service for more than 5 million people within its catchment area. The labor ward provides services for approximately 380 deliveries per month. The Department of Obstetrics and Gynecology has maternity ward, high risk ward and labor ward. High risk ward has around 20 beds and approximately 80 pregnant women were admitted in this ward per month. Maternity, labor ward and high risk ward are staffed with four obstetrics and gynecology specialists, twenty three midwives, five emergency surgeons and twenty four interns.
All pregnant women who were admitted in obstetric wards of Debre tabor general hospital between 28 to 36+6weeks of gestation were the source population. This study included all pregnant women who were admitted in obstetric wards of Debre tabor general hospital during data collection period between 28 to 36+6weeks of gestation.
Sample size determination
The required sample size was determined using a single population proportion formula [n = [(Za/2)2*P (1-P)]/d2] by assuming a 95% confidence level of Z a/2 = 1.96, marginal of error 5%, and prevalence of preterm PROM 50% since there is no previous similar studies. Accordingly, the sample size was 385. By adding 10% non- response rate, the final sample size was 424.
Socio-demographic variables include MUAC of the mother, age, ethnicity, residency, religion, educational status, marital status, average monthly income and occupational status. Obstetric variables include gravidity, parity, gestational age, ANC follow up, vaginal bleeding in current pregnancy, history of premature rupture of membrane, history of preterm labor, history of abortion, type of pregnancy, polyhydramnios in current pregnancy, presentation and labor pain. Medical and behavioral variables include abnormal vaginal discharge, gestational diabetes mellitus, anemia, lifting heavy objects, falling in accident, UTI, Cigarette smoking, Cocaine use, alcohol addicted and chat chewing
Sampling and sampling procedure
Study participants were selected from high risk and labor ward by using systematic random sampling technique. The average numbers of women who were admitted in Debre tabor general hospital during the data collection periods were estimated based on the previous admissions which were found by reviewing a six months registration book in high risk and labor wards prior to data collection. Totally 480 pregnant women were admitted in high risk wards in six months. We planned to collect the data within six months. So to find the sampling interval; the total number of women who were admitted in six months (480) was divided by the total number of sample size (424) and it was approximately 2. The first woman was selected by lottery method then every other woman who was admitted in the high risk ward was recruited for the study.
Data collection tools and procedures
Interview, chart review and measurements were used to collect the data. Three BSc midwives and one supervisor were used for data collection and supervisory activity respectively after training was given for them. Structured interviewer-administered data collection formats were adapted and modified from different kinds of literature. Questionnaires were structured into four logical sections (socio-demographic characteristics, obstetric related factors, medical history and behavioral related factors). Data on respondent’s specific socio-demographic, obstetric, medical and behavioral characteristics were collected by reviewing her medical records and through interview of the respondents.
In addition, physical measurements were used to obtain data on mid upper arm circumstance (MUAC) of pregnant women. In this condition, MUAC of each woman was measured at the midpoint between the tips of the shoulder and elbow of the left arm using non-elastic, non-stretchable MUAC tapes. In this study, a poor nutritional status of the mother was defined as MUAC < 23 cm(1).
Gestational age was estimated using from 1st trimester or 2nd trimester ultrasound (up to 24weeks) and last normal menstrual period.
Data management and analysis
Data were entered into EPI data version 4.2 then exported to SPSS version 23 for analysis. Descriptive statistics such as cross tabulations and frequencies were computed. Binary logistic regressions were used to select variables associated with preterm PROM. In binary logistic regression, both bivariate and multivariable logistic regression was computed. In bivariate analysis, independent variables with p value less than 0.2 were selected as a candidate for multiple logistic regression. In multivariable logistic regression; Statistical significant was considered at P<0.05. Adjusted Odds Ratio (AOR) and their 95% Confidence Interval (CI) were used to measure strength of association. Back ward stepwise logistic regression was applied. The Hosmer -Lemeshow goodness-of-fit statistic was used to check if the necessary assumptions for multiple logistic regressions were fulfilled and the model had a p-value >0.05 which proved the model was good.