Effect of guided counseling on nutritional status of pregnant women in West Gojjam zone, Ethiopia: A cluster-randomized controlled trial

Background: Undernutrition during pregnancy affects birth outcomes adversely. In Ethiopia, despite nutrition counseling on the maternal diet being given by the health workers during pregnancy, maternal undernutrition is still high in the country. Hence, this study aimed to assess the effect of guided counseling using the health belief model and the theory of planned behavior on the nutritional status of pregnant women. Methods: A two-arm parallel cluster randomized controlled community trial was conducted in West Gojjam Zone, Ethiopia, from May 1, 2018, to April 30, 2019. The nutritional status of the women was assessed using mid-upper arm circumference. A total of 694 pregnant women were recruited from the intervention (n=346) and control (n=348) clusters. Of which endline data were collected from 313 and 332 pregnant women in the intervention and control clusters, respectively. The intervention was started before 16 weeks of gestation and pregnant women in the intervention group attended 4 counseling sessions. Counseling was given at the participants’ homes using a counseling guide with the core contents of the intervention. Leaflets with appropriate pictures and the core messages were given for women in the intervention arm. Women in the control group got the routine nutrition education given by the health care system. Data were collected using interviewer administered structured questionnaires and mid-upper arm circumference was measured using an adult MUAC tape. Descriptive statistics and linear mixed-effects model were used to assess the intervention effect after adjusting for potential confounders. Results: After the intervention, the prevalence of undernutrition was 16.7% lower in the intervention group compared with the control arm (30.6% Vs 47.3%, P = < 0.001). Women in the intervention group showed significant improvement in nutritional status at the end of the trial than the control group (β = 0.615, p = < 0.001). Conclusion: This study demonstrated that guided counseling using the health belief model and the theory of planned behavior was effective in improving the nutritional status of pregnant women. The results imply the need for the design of model and theory based nutrition counseling guidelines. The trial was registered in Clinical Trials.gov (NCT03627156). This study was a 1 year two-arm parallel design cluster randomized controlled community trial. Clusters (kebeles or the smallest administrative units in Ethiopia) were taken as a unit of randomization. The study was conducted in West Gojjam Zone from May 1, 2018, to April 30, 2019. It is one of the 11 zones in Amhara Region comprised of 15 woredas/districts/ with a total population of 2,641,240, half of which (50.7%) were females [25]. The number of estimated pregnant women was 61,072. The study was conducted in accordance with the principles of Helsinki Declaration and the requirements of Good Clinical Practice [26]. The research protocol was approved by the Institutional Review Board of Bahir Dar University (protocol number: 092/18–04). Written informed consent (fingerprint for women who could not read and write) was secured from each participant prior to starting the trial. The trial was registered in the Clinical Trials.gov ({"type":"clinical-trial","attrs":{"text":"NCT03627156","term_id":"NCT03627156"}}NCT03627156). Consolidated Standards of Reporting Trials (CONSORT) guideline was used for reporting the results (Fig. 1 and Additional file 1) [27]. This figure shows the flow of the study participants through the trial according to the criteria recommended in the CONSORT guideline The study targeted pregnant women before 16 weeks of gestation. Women who had no intention of leaving the study area until delivery were included in this study. Women with confirmed or diagnosed hypertension and/or diabetes mellitus were excluded from participating in the study [28]. The sample size was calculated using G power 3.1.9.2 program with a power of 85% for Fisher’s exact test and precision of 5%. According to Kedir H et al., (2016) the prevalence of undernutrition among pregnant women (p1) was 24% [7] and P2 was 9% by assuming a 15% difference between p1 and P2 [29]. The calculated sample size was multiplied by design effect of two due to cluster sampling. Considering a 10% loss to follow up, the final sample size was 214 pregnant women in each arm. Since cluster randomization was used and pregnant women who fulfilled the requirement were included, 346 women in the intervention group and 348 women in the control group were enrolled in this study. From 15 woredas in the zone, eight had nutrition education intervention on complementary feeding practice. These eight woredas were excluded from the study. From the seven eligible woredas, three woredas namely: Bahir Dar Zuria Woreda, South Achefer Woreda, and Burie Zuria Woreda were selected using simple random sampling (SRS) technique. Then, samples of non-adjacent clusters were selected from the three woredas using SRS (lottery) method. The following formula was used to determine the number of clusters [30]. Where c is the number of required clusters, Po was undernourished pregnant women (24%) [7], p1 was the expected number of undernourished women after intervention (9%) by assuming 15% difference between p0 and p1, n was the number of households that had pregnant women in each cluster (assuming an average of 35 women). K was a coefficient of variation of undernourished women between clusters within each arm. Since there was no study to estimate K, it was taken as 0.5. Therefore, 11 clusters per arm were included in this study. Based on proportional to size allocation, ten clusters from Bahir Dar Zuria Woreda, six clusters from South Achefer Woreda and another six clusters from Burie Zuria Woreda were selected, randomly. Finally, SRS (lottery) method and a 1:1 ratio were used to allocate intervention and control clusters (Additional file 2). Cluster randomization was used to prevent message contamination because women in the same cluster had a high probability of communicating and discussing the intervention messages. To avoid this, all eligible pregnant women in one cluster were enrolled in the same arm (either in the intervention or control arms). Moreover, buffer zones (non-selected clusters) were also left between the intervention and control clusters to prevent information contamination [31]. Eligible pregnant women were screened through the house to house survey by inquiring about the first date of the last menstrual period and confirming pregnancy with a pregnancy test. All eligible pregnant women were included in the study. Nurses working in selected woredas randomized the cluster, screened and enrolled the study participants from May to August 2018. Community-based guided counseling using the HBM and the TPB was the intervention package for this study (Additional file 3). It was adapted from the recommendations by World Health Organization and Ministry of Health of Ethiopia [32, 33]. The core contents of the counseling guide were increasing meal frequency and portion size with increasing gestational age. Message on taking diversified meals by giving emphasis to iron-rich foods, animal products, fruits, and vegetables was also one component of the counseling guide. Messages on the consumption of iron/folic acid supplement and iodized salt were also included in the core contents of the counseling guide. Additional messages of the core contents were reducing heavy workload, taking day rest, impregnated bed net use and utilization of health care services. Moreover, the consequences of taking inadequate nutrient, susceptibility to and severity of the consequences of insufficient nutrient intakes were also discussed during counseling. The benefits of taking an adequate amount of diversified meals and barriers that interfere with taking a balanced diet were also included in the counseling guide. Attitude, subjective norms, self-efficacy, perceived control, intention, knowledge and dietary practice were assessed during each counseling session. Then, counseling was given based on the identified gaps and household income. Each pregnant woman attended four counseling sessions throughout her entire pregnancy. Individual nutrition counseling was given through a home visit on non-working days (religious holidays and weekends). During counseling, counselors used a client-centered approach to identify women’s dietary practices and their specific needs in terms of nutrition. Counselors considered women’s needs, household income and identified gaps and allowed the women to choose recommendations that were locally available, acceptable and affordable. Counseling was delivered monthly using a counseling guide with the core contents and each counseling session lasted for 40 to 60 minutes. The first counseling was given before 16 weeks of gestation, focused on basic nutrition, food groups, food selection, preparation, meal frequency, portion size, and iodized salt utilization. The second and third sessions of the counseling were given during the second trimester of pregnancy and covered the whole contents of the counseling guide. The last counseling was given based on the identified gaps during the early third trimester of pregnancy. Leaflets with the core messages in Amharic (local language) and appropriate pictures were prepared and delivered to each pregnant woman in the intervention arm. For women who couldn’t read, anyone at home or in the neighborhood who could read was requested to read the leaflet to the woman and other family members. Women in the control arm received nutrition education given by the health care system. Pregnant women from both the control and intervention arms had access to ANC services. Six BSc nurses and three MSc nutritionists were recruited as counselors and supervisors of the counseling process, respectively. Counselors were selected based on their previous experience in giving counseling services. A three-day intensive training with role-playing and fieldwork were given to the counselors and supervisors using the training manual. Moreover, a one-day additional training was given for the counselors and supervisors after two months of intervention implementation to keep providers sticking to the standardized procedures over time. Criteria were established to assess fidelity of the intervention, based on the National Institutes of Health Behavioural Change Consortium developed best practice recommendations [34]. The criteria included checklists to assess intervention design, training of counselors, counseling process, receipt of intervention and enactment of skills gained from the intervention [35]. The intervention design had theoretical ground. Non-adjacent clusters were selected to prevent information contamination. Equal numbers of clusters for the intervention and control groups were taken from each woreda to balance variations. The trial used a control group and counseling guide. The intervention process was pretested before the implementation of the trial. Besides, each woman received equal numbers and frequencies of counseling, and the lengths of contacts within an intervention group were similar to make the process standardized. Counselor training was given in a group using a training manual, role-playing, and mock counseling practice. Counselors’ knowledge and skill were assessed by pre and post-training tests and practical evaluation. Counseling sessions were randomly selected for process evaluation and all selected sessions were evaluated by one process evaluator. The process observer rated the educator using a ‘yes/no’ rating system on items such as using a counseling guide, provision of the whole content, duration and frequency of counseling, preparedness, accuracy, and ability to properly respond to questions. Intervention receipt was assessed using checklists on knowledge of the women on diet during pregnancy through interviewing about their understanding of the core contents of the intervention. Intervention enactment was also assessed using the checklist on a demonstration of food preparation and consumption. Even if, participant allocation concealment was not possible due to the nature of the intervention, participants, counselors, and data collectors were blinded to the study hypotheses. Additionally, the data entry clerk was blinded by labeling the groups with a non-identifiable unique number until analysis was finalized. The counseling process was supervised by the counseling supervisors and principal investigator. Six nurses collected data using structured questionnaires through one-to-one interview of the participants at their homes. The questionnaire included socio-demographic variables, obstetric history, HBM, and TPB constructs. Data on socio-demographic and obstetric characteristics were collected at the baseline. Whereas, data on food security, MUAC, HBM, and TPB constructs were taken before and after implementation of the intervention. Data collectors and supervisors were trained for 3 days using a training manual focused on the data collection tools, procedures, and ethical issues. To prevent the breaching of privacy of the women, no one was allowed to have free access to the place where the interview was conducted. The supervisors and the principal investigator overhauled the data collection procedure. The data collection team held a daily meeting to discuss challenges encountered during the day. Besides, nutrition counselors and their supervisors also held a monthly meeting to discuss difficulties during nutrition counseling and feedback was given to the counselors. The secondary outcome of this trial was the nutritional status of the pregnant women that was assessed by measuring MUAC. Post-intervention data were measured from 36 to 37 weeks of pregnancy. Women who didnot attend all counseling sessions were considered as ‘did not adhere- to the guideline’ and those withdraw from participating in the study were taken as ‘lost to follow up’. There is minimal change in MUAC during pregnancy, accordingly, MUAC is a better indicator of pre-pregnancy body fat and the nutritional status of pregnant women than body mass index [12, 36, 37]. Therefore, in this study, MUAC was used to assess the nutritional status of pregnant women and was measured on the upper left arm. During the procedure, the midpoint of the upper arm was located by flexing the women’s elbows to 900 with the palm facing upwards. Then the distance from the acromion to olecranon processes was measured and the midpoint was marked. Finally, measuring tape was placed snugly around the arm at the midpoint mark while hanging arm freely, palm facing towards the thigh. Two measurements were taken and read the measurement to the nearest 0.1 cm. Women with MUAC > = 23 cm were considered normal nourished whereas participants with MUAC < 23 cm were labeled as undernourished [12, 13, 38]. The wealth index of the household was determined using principal component analysis (PCA) by considering latrine, water source, household assets, livestock, and agricultural land ownership. The responses of all non-dummy variables were classified into three parts, and the highest one was coded as 1 and the two lower values were given code 0. Factor scores were produced using variables having a commonality value of greater than 0.5 in PCA. Quintiles of the wealth score were created using the first principal component. Food security status was assessed using 27 previously validated questions [39]. A household that experienced less than the first 2, 2–10, 11–17 and > 17 food insecurity indicators were considered as food secure, mildly, moderately and severely food insecure households, respectively. The attitude, knowledge, subjective norms, intention, perceived susceptibility, severity, benefit, and barriers were assessed using the sum of their respective composite questions. The full description of data collation, measurements, the study area and participants described elsewhere [28]. Descriptive statistics were used to summarize the baseline socio-demographic characteristics of the women by group status. A chi-square test was performed to compare the baseline characteristics of the intervention and control groups. Comparisons of MUAC between and within the intervention and control groups were done using independent samples and paired sample t-tests, respectively. A per-protocol analysis was performed in this study. The per-protocol analysis includes all the study participants who adhered to the predetermined guideline. Therefore, in this study, women who attended four education sessions and gave endline data were included in the analysis. A linear mixed-effects model was used to determine the impacts of the intervention on changes in the nutritional status of pregnant women over time. This model enables to accommodate the correlation of observations due to the repeated measures (pre- and post-intervention) and the clustering of individuals within the 22 randomly selected clusters. During fitting the model, participants and clusters were analyzed as random effects. This model also enables to control the effects of potential confounding factors (food security, latrine utilization, education, family size, source of drinking water and age). The intercept-only model estimates the variance of the cluster-level residual errors as 0.0035 (variability of the average nutritional status across all clusters was 0.0035 and which wasn’t statistically significant (p = 0.90). The intra-cluster correlation coefficient was closer to zero (0.001) which showed that no need for fitting a third-level model. Therefore, the two-level model was fitted to account for time-invariant variables at the individual level. The effect of the intervention was evaluated by testing the interaction term between time and treatment allocation. All statistical analyses were performed using the SPSS package version 23.