Background Preconception care (PCC) is an evidence-based health promotion intervention to prevent adverse pregnancy outcomes. Nevertheless, it is one of the missing elements within the continuum of maternal and child healthcare. Despite the WHO’s recommendation, most of the developing countries have not yet started implementing preconception care. Objective To determine the knowledge level of healthcare providers about PCCand to identify predictors of effective knowledge of preconception care. Method This is a cross-sectional study conducted among 634 healthcare providers (HCP) working in public health institutions of Hawassa. A pilot-tested and validated self-administered survey tool was used to collect data from individual healthcare providers who were selected randomly using a multistage sampling technique. The data entry and analysis were conducted using SPSS version 20 software. Frequency, proportions, means and standard deviations were used to describe the data. Bivariate and multivariate logistic regression models were implemented to determine the predictors of HCP’s PCC knowledge. Results Only a few (31%) of the healthcare providers demonstrated a good level of knowledge on preconception care. The odds of having good PCC knowledge was high among HCPs working in hospitals (AOR = 1.8, 95% C.I. 1.3-2.6), HCPs using their smart phone to access clinical resources (AOR = 1.4, 95% C.I. 1.1-2.0), among those HCPs ever have read PCC guideline prepared by organization outside of Ethiopia (AOR = 1.9, 95% C.I. 1.4-2.7), among those who claimed practicing PCC (AOR = 3.4, 95% C.I. 2.0-5.9), and among those who earn salary of ≥ 146.0 $(AOR = 1.5, 95% C.I. 1.1-2.1). Conclusion There is an unacceptably low level of knowledge about PCC among most of the healthcare providers in public health facilities in Ethiopia. The predictors identified in this study can be used to enhance the knowledge of healthcare providers about preconception care.
The study was a cross-sectional quantitative study conducted from May to June 2017 among healthcare providers working in public health institutions (PHI) within the jurisdiction of Hawassa, 275km south of the capital (Addis Ababa) of Ethiopia. The public health institutions consist of nine healthcentres and two hospitals of which one was a secondary level public hospital and the other a tertiary level comprehensive specialized hospital. Under the public health centres are seventeen health posts where the health extension workers are working. During the study period, healthcare workers consisted of 106 doctors, 826 nurses, 60 health officers, 95 midwifes, and 142 health extension workers who were employed in the institutions and formed the target population for the study. The health extension workers are primarily nurses, specifically trained to provide community health service in line with the country’s primary healthcare package. The maternal healthcare includes antenatal care, postnatal care and institutional delivery services. These services are provided in every health facility by all healthcare providers but mainly by midwives and gynaecologists. Preconception care is not a specified area of care in any of these facilities. The authors of this study purposively selected Hawassa City Administration as the study area. Selection of the study area considered the goals of the study, feasibility issues, and the availability of all healthcare providers working at all levels of the referral system located at both rural and urban areas. The study sample of healthcare providers in public health institutions in Hawassa was randomly selected by using the employers register as a sampling frame. The sampled healthcare professionals were all taken proportional to their profession, their number, and the type of health facility where they are working. Healthcare workers who were employed for less than six months were excluded from the study. Multistage sampling technique was applied to draw a total of 647 HCPs. The minimum sample size required for the study was determined by using a single population proportion formula. While computing the minimum sample size, the following parameters were considered: a 0.05 margin of error (α), a 95% Confidence Interval (CI), a 50% estimated proportion of healthcare providers’ knowledge about preconception care, 10% non-response rate and a design effect of 2. The design effect (DEEF) was calculated with the formula DEEF = 1+ δ (n-1). The “δ” or the interclass correlation coefficient (ICC) was calculated from the cluster data by using SPSS and it was found 0.169. Since the average size of clusters (n) was 11/2 = 5.5, the final DEEF was determinedas 1.79 ≈ 2. Given that the total number of HCPs working in PHIs of the study area was 1239, a population correction factor was considered. Due to the absence of a similar study or comparative study in the country we preferred taking a 50% proportion which is a proportion to yield adequate sample size. Concerning sampling procedure, first, five PHIs out of the 11 PHIs found within the city administration were randomly selected. By using simple random sampling technique, 3 out of the 9 health centres were included in the study. Since the remaining two public hospitals were quite different in their level and type, both were selected. The study population was also stratified in terms of profession. In the second stage, HCPs were selected by the systematic random sampling method using employer’s employee registry document as a sampling frame. The study participants were all taken from each strata using probability proportional to size method. All HCPs were selected and consented to participate in the study without any coercion. A data collection tool, namely ‘Andarg-Ethio PCC-KAP-Questionnaire for HCP’, based on literature and evidence-based guidelines on PCC was developed and validated by the principal investigator to conduct this research project. The instrument was tested for content and face-validity by a panel of experts and was scored with a content validity index (CVI) of 92.4%. The reliability of the instrument was checked for its internal consistency with a Cronbach’s α test and demonstrated a score of 0.945[19]. The questionnaire which was originally prepared in English was translated to local language Amharic and then translated back to English. The survey was administered using the Amharic version. The instrument was designed to assess socio-demographic characteristics of HCPs, their knowledge on PCC, their attitude towards PCC, their practice on PCC,issues on training of PCC,and in-service training opportunities on PCC. The HCPs’ knowledge of PCC was measured through 18 questions, each containing only one correct answer. A further 36 items measured various elements of PCC practice, including reproductive life planning,screening practices, access to resources to practice PCC. The attitude of healthcare workers on PCC was assessed by using 10 items each with five-point Likert scale responses. The calculated single knowledge factor was then categorized into three ordinal categories. Respondents who scored less than the 50th percentile or below the mean score were categorized as HCPs with ‘poor/low PCC knowledge’. Whereas, HCPs who scored ≥ 50th percentile to 75th percentile and those who scored > 75th percentile were categorized as HCPs with ‘medium’ and ‘high’ PCC knowledge respectively. For analytical purpose, those HCPs who scored ‘high’ and ‘medium’ PCC knowledge were merged all together into another category called ‘HCPs with good PCC knowledge’. The instrument was piloted on 65 (10% of the minimum sample) healthcare practitioners in a different town as Hawassa, after which minor revision to improve on the clarity of questions were done. The questionnaire was administered by two nurses, one health officer, one 2nd year Master of public health student, and one pharmacist after being trained by the primary investigator. There were two field supervisors. The principal investigator was the main supervisor throughout the study. The data collected was entered to the SPSS version 20 software by an experienced statistician. The analysis used descriptive statistics such as frequency, proportion, standard deviation, mean, mode and range to describe the variables of the study. In addition, the inferential statistics applied a binary and multiple logistic regression analytical models to determine the crude (COR) and adjusted odds ratios (AOR) respectively. The analysis was all made fixing CI at 95%. The variables with their P-Value of less than 0.20 were all considered in the second or multivariate logistic regression model. The second analytical step used a stepwise backward model to determine factors associated with ‘HCP’s good PCC knowledge’. The goodness of fit of the models was tested by using the Hosmer-Lemeshow test. Thus, the model which was found to be greater than the significance level (P-value = 0.05) was accepted. The project proposal of the study was approved by IRBs of Hawassa University and the University of South Africa. Ethical principles such as confidentiality, beneficence, respect and human rights were maintained throughout the study.