Study design and setting
A facility-based cross-sectional study was conducted from January 10, 2018, to February 10, 2018, in Mekelle, Ethiopia. Mekelle is a city located around 780 km north of the Ethiopian capital, Addis Ababa, at an elevation of 2084 m above sea level, in the region of Tigrai. The Tigrai regional state has five zones. Mekelle city being designated as the regional capital city. Mekelle is administratively divided into seven sub-cities in which the entire Tigrai population obtains their specific healthcare services. Mekelle has three public hospitals, nine health centers, four private hospitals, more than 30 private clinics, and one ortho-physiotherapy center. Of these, 12 public and 7 private health facilities were providing antenatal care services.
Mekelle has a total population of 358,529 of whom 176,986 are females [32]. The Ethiopian Demographic and Health Survey (EDHS 2016) reported a countrywide fertility rate of 4.6, 42, and 33% employment status for women. However, in Tigrai, this national report revealed a similar fertility rate 4.7% and a slightly higher, literacy rate 51%, and employment status 37.4% of women. In addition, in Tigrai, health facility delivery rate is 56.9% (716 births per 1129 interviewed women) with the same proportion (56.5%) gave births in public health facilities followed by at home (41.0%). Furthermore, 59.3% of births were delivered by a skilled provider, and 49.6% of these births were attended by nurse or midwife followed by traditional birth attendants (23.2%) [33].
Study population
Pregnant women during any trimester of pregnancy who were attending antenatal care during the data collection period at randomly selected public (12) and private health facilities (7).
Inclusion criteria
Study participants who were having antenatal care during the data collection period in the selected public and private health facilities.
Exclusion criteria
Pregnant women who presented with medical or obstetric complications and serious psychological conditions that could have an impact on the reliability of data/information were excluded.
Sample size, sampling techniques, and procedures
The required sample size was determined using a single population proportion formula based on the assumption of a 95% confidence interval with a margin of error of 5%. A simple random sampling technique was used to select the eight public and private health facilities out of 19 utilized for data collection. These health facilities were chosen because they provide antenatal services and report to the Tigrai Regional Health Bureau (TRHB). According to the Federal Democratic Republic of Ethiopia Ministry of Health; health and health-related indicators in 2016/2017 report indicated that in Tigrai, 125,373 (69.5%) of women had at least four antenatal visits during their last pregnancy, and 118,219 (65.5%) of births was delivered by skilled attendants [34].
The number recruited from each health facility was determined based on a population-based proportion that was developed considering the TRHB 2009 Ethiopian Fiscal Year (EFY) annual report and recent data reported by the health institutions for the third quarter of the previous year. All eligible pregnant women attending their antenatal care visits at selected health facilities during the study period were randomly approached for inclusion until the total sample size was attained. A total of 305 participants was targeted based on these calculations (Fig. 1).
Data collection instrument and methods
Data were collected using a structured questionnaire developed from reviewing relevant literature. The questionnaire contains socio-demographic, obstetric and health, physical activity characteristics, and Pregnancy Physical Activity Questionnaire (PPAQ) which was developed by Chasan-Taber et al. [35]. A slight modification was made on PPAQ tool; for instance, two items like “playing with pets, and “mowing the lawn by riding a mower”, and by using a walking mower, raking, and gardening” were omitted in the present study’s tool due to cultural and feasibility issues in the Ethiopian context. Four experts (one obstetric and gynecologist, two chief physiotherapy specialists and one chief midwifery) with good knowledge of the subject matter of the study’s theme and both language versions had contextually constructed the contents of the original and local language versions of the PPAQ.
A questionnaire was translated into Tigrigna which is entirely an official language spoken in the Tigrai region, Ethiopia. This questionnaire was pretested on 5% (15) of pregnant women who were not part of the study participants from similar health settings and based on pretesting feedback modifications were made to the survey prior to administration.
For each activity, respondents were asked to select the category that best approximates the amount of time spent on that activity per day or week during the current trimester. At the time of recruitment, the women were informed about the purposes and procedures of this study. After highlighting the introduction and receiving oral and written consent, the participants completed the questionnaires for over 15 min in a quiet place in the clinic.
The PPAQ measures the frequency and duration of activities and gives an intensity value to each activity developed by Chasan-Taber et al. (31). The calculation was total activity = sum of (duration * intensity) for each question. Each activity was classified according to intensity in Metabolic Equivalent Task (MET)—sedentary (< 1.5 MET), low (1.5 to < 3.0 MET), moderate (3.0 to 6.0 MET), and vigorous-intensity (> 6.0 MET)—and type—labor, domestic (e.g., caring for a person), and sports/exercise [35].
Data processing and analysis
After the collection of data, all collected questionnaires were checked for completeness, correctness, and internal consistency to exclude missing or inconsistent data. Data were entered, cleaned and analyzed using IBM SPSS Statistics version 23. Descriptive statistics were computed using frequencies with percentages for categorical variables and median and interquartile range (IQR) for continuous variables. The binary logistic regression model was used to model the association between outcome and independent variables. The final model was assessed for multicollinearity using Variance Inflation Factor (VIF) and goodness of fit using Hosmer and Lemeshow test for independents which make a variable selection decision at each step of the modeling process [36]. Any variable having a significant univariate test with cut-off point p, 0.25 was a candidate for the multivariate analysis [36]. Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were estimated. A p-value of < 0.05 was considered statistically significant. During the analysis phase, each categorical variable having five or more cells were not merged in the categorization of variables.
Ethical considerations
This research was conducted after obtaining ethical clearance from the Health Research Ethics Review Committee (HRERC) of the College of Health Sciences, Mekelle University. The study was conducted with written and oral consent that assures the willingness of each participant to participate in the study. The pregnant women who were unwilling to participate in the study were respected and only those who were willing to participate in the study were recruited. Each participant signed an informed consent form. Confidentiality and privacy of the pregnant woman were also kept protected.