Time to non-adherence to iron and folic acid supplementation and associated factors among pregnant women in Hosanna town, South Ethiopia: Cox-proportional hazard model

PLoS ONE, Volume 17, No. 9 September, Year 2022

Interpretation

Backgrounds Micronutrient deficits in women of reproductive age have been linked to poor pregnancy outcomes. The most common micronutrient deficits in women are iron and folate. The World Health Organization recommends daily oral iron and folic acid supplementation (IFAS) as part of routine antenatal care to lower the risk of maternal anemia and adverse pregnancy outcomes. However, the effectiveness of the supplementation relies on client’s strict adherence. The aim of this study was to determine time- to- non-adherence to IFAS and associated factors among pregnant women in Hosanna Town, South Ethiopia. Methods A community based cross sectional study design was employed from May 15-June11, 2021. Data were entered into Epi-Data version 3.1 and exported to SPSS version 23 for analysis. The Cox regression hazard model was applied. The threshold of statistical significance was declared at a p-value <0.05 and adjusted hazard ratios (AHRs) with corresponding 95% confidence intervals were used to report. Result The median time-to-non-adherence was 74 days (95 percent CI: 65.33-82.67). After adjusting for the confounders, age (AHR = 1.05, 95% CI: 1.01-1.09), education status (AHR = 2.43 95%CI 1.34-4.40, AHR 3.00, 95% CI: 2.09-4.31, AHR 1.91, 95% CI: 1.32-2.77), household's wealth index (AHR = 1.73, 95% CI: 1.19-2.51, AHR = 1.64, 95% CI:1.15- 2.35), and counseling at service delivery (AHR = 2.53, 95% CI: 1.88-3.41) were independent predictors of time to non-adherence to IFAS among pregnant women. Conclusion The median time to non-adherence was short and women became non-adherent before the recommended duration. Improving women's education and counseling pregnant women on IFAS during pregnancy would make a change. The research was carried out in Hosanna Town, Southern Nations Nationalities and People Regional State (SNNPR) of Ethiopia. The town is situated in 230 kilometers to the south of national capital, Addis Ababa. According to 2007 census [16], total population of the town was 69,957; 35, 503 were males and 34, 454 were females. In the same census, the population growth rate in the region was 2.9% per year. Based on this, the projected total population of the Town for 2021 was 104,387. In the region, 23.3% and 3.5% population are women in their reproductive age and expected to be pregnant respectively [17]. Based on these evidences, there were 24,323 women of reproductive age (15–49 years) and 3,654 estimated pregnancies for year 2021. The research involving community based cross sectional study design was employed between May 15 and June 11, 2021. The source population consisted of all pregnant women in the research area and all pregnant women in selected sub-cities were sample population. The inclusion criteria considered all pregnant women in the study cluster who were booked for ANC one-week preceding the study. Pregnant women who were booked for ANC but whose registration for follow-up was less than one week prior the survey date were excluded. Also women who unable to recall their last normal menstrual cycle or gestational age (GA) at the time of booking were not allowed participating. The sample size was calculated with Epi Info version 7 using the double population proportion formula to detect a non-adherence rate of at least 25.1%, [14], 95% significance level and 5% margin of error; a sample size of 289 was obtained. The Cox proportional hazards model (power cox) was used to determine sample size for factors associated with non-adherence to IFAS using Stata version 15.0 considering the presence of censoring and adjusting for others. Non- adherence was considered a failure (outcome). Factors obtained from literatures having a significant association with adherence to IFAS were considered for sample size calculation; knowledge about IFAS [6], counseling on IFAS [12], Partner support [18], and Educational status [19]. After computing various factors, sample size calculated for educational status was 139; the largest sample size computed for factors associated with IFAS adherence. Therefore, the minimum sample size required for the non-adherence estimation would be 289. Considering 5% non-response rate [14, 20, 21] and design effect of 1.5, the final sample size was (289* +5%) *1.5 = 456. Where n is required sample size, zα/2 is 95% CI, p is population proportion and d is margin of error. Study clusters were identified using a two-stage cluster sampling procedure. After randomly selecting five Kebeles (the smallest administrative structure), total number of predefined distinct clusters (Mender) (smallest cluster within an administrative Kebele) were obtained, after which we obtained the size of the pregnant women for each cluster. The required number of clusters from each Kebele were assigned using probability proportional to population size approach, in which larger settlements have a higher chance of being selected as clusters. Reserve clusters were used until we obtain required sample size (Fig 1). Ethiopian Demographic and Health Survey 2016 (EDHS,2016) [22] and relevant literature [6, 11, 14, 18, 23] were used to adapt data collection tool. A household’s wealth status was computed based on 23 household assets and housing quality variables which were adapted from EDHS2016 [22], given that the study setup is urban. First, all the study participants were asked about the ownership of assets by their respective households. Those who owned the asset received a score of "1," while those who did not received a score of "0". A structured questionnaire was prepared in English, translated to Amharic, and then back translated into English to ensure consistency in order to measure the required parameters. The Amharic version tool was then employed. When a pregnant woman visiting an antenatal clinic took IFA tablets for less than four days per week for the week prior to the survey or for less than 90 days during the third trimester of pregnancy [12, 13]. Pregnant women who were adherent at the time of data collection were censored observations. When a pregnant lady took IFAS tablet for less than four days one week preceding the survey date, the event occurred. Is the amount of days a pregnant woman contributed while on IFAS until she experienced an event of interest (non-adherent) or censorship. The dependent variable was time to non- adherence of IFA supplementation measured in days. Censored observations were denoted by 0, whereas events were indicated by 1. The period of time a pregnant woman spent on IFAS in days was determined as the difference between the entire GA (from LNMP to the day of data collection) and the calculated and/or reported GA at initial booking for ANC. Maternal age and educational status, household’s wealth index, counseling status of health institution on IFAS, knowledge on IFAS, waiting time to receive care, frequency of ANC visit, history of adverse fetal outcomes and history of anemia were variables hypothesized to be independent predictors of time to non-adherence which is a primary outcome in this study. To ensure data quality, ten data collectors and one supervisor received two days training on tool clarity and overall data collection processes. The training was emphasized on sociodemographic information’s, household’s wealth status, obstetric factors, Personal exposure to media, health facility related factors, and Knowledge on IFAS and anemia. Structured questionnaire was prepared in English, translated into Amharic, and then back translated into English to ensure consistency. A pre-test was conducted on 5% of the sample size in a nearby town. Cronbach’s alpha was done to assess internal consistency (alpha coefficient for household wealth status (23 items) = 0.85, media access (3 items) = 0.71, counseling at health facility (7 items) = 0.76, Knowledge on IFAS (15 items) = 0.79). A public health officer supervised the data collection process, while the primary investigators (PIs) supervised the whole technique. All collected data were handled to PIs and checked and cleaned for consistency and completeness; daily discussion was held in case of inconsistencies. Epi-Data version 3.1 for data entry and Statistical Package for Social Science (SPSS) version 23 for analysis were use. Before analysis missing value, new categories and normality for continuous variables were checked. Households wealth index was computed by principal component analysis (PCA) based on household assets and housing quality variables which adapted from EDHS 2016 [22]. Pregnant women’s knowledge on IFAS was computed after performing PCA based on 15 items. Problematic variables were removed step by step, eleven items having four component factors that explains a total variance of 64.3% were retained; whose alpha coefficient was 0.78, all having acceptable correlation matrix (KMO = 0.78, x2 = 1048, P <0.001), sampling adequacy of each item was > 0.5. The value of retained variables was aggregated and used median as a cut off to declare knowledge status of study population. The difference between total GA (spanning from last normal menstrual cycle to date of data collection) and the calculated and or reported GA at first booking for ANC was taken as total time contributed in days during which a pregnant woman was on IFAS. Survival curve was used to display the survival status (time to non-adherence) among different characteristics. For survival analysis, the outcome variable was dichotomized to event and censored. The assumptions of proportional hazard were tested statistically and graphically. Against each categorical variable, we performed the log-og survival plot and the Kaplan-Meier survival plot. Both log-log survival plot and Kaplan-Meier survival and predicted plot revealed that the plots were parallel to each other. We have also conducted Schoenfeld test with the corresponding p-value for all variables. The Kaplan-Meier test was used to assess the median survival time between groups. The multivariate Cox Proportional Hazard model was used to examine the factors associated with time to non-adherence. The crude and adjusted hazards ratios with a 95% confidence interval (CI) were used as a measure of effect size. The Cox proportional hazard model assumption was tested graphically using log-minus-log survival plots against time for predictors. Multivariable Cox proportional hazard regression model was used to control the confounding effect of variables. In bivariate analysis, variables having a p-value < 0.25 were selected as potential predictors and used in multivariable analysis. A p-value < 0.05 with a corresponding 95% CI was declared statistically significant. The study was approved by the Institutional Review Board (IRB) of Hosanna Health Science College. In addition, permission was obtained from health department of the local government offices. Informed written consent was obtained from all participants. Respondents were informed that they had the right to refuse or discontinue the interview. The information provided by each respondent was kept confidential. Women who were non-adherent at the time of data collection were successfully counseled on the benefits of IFAS.

AI Digest

Study Justification:

This study aimed to investigate the time to non-adherence to iron and folic acid supplementation (IFAS) among pregnant women in Hosanna Town, South Ethiopia. The justification for this study is based on the importance of micronutrient supplementation, specifically iron and folate, in improving pregnancy outcomes. Micronutrient deficits in women of reproductive age have been linked to poor pregnancy outcomes, and the World Health Organization recommends daily IFAS as part of routine antenatal care to lower the risk of maternal anemia and adverse pregnancy outcomes. However, the effectiveness of the supplementation relies on the client’s strict adherence. Therefore, understanding the factors associated with non-adherence to IFAS is crucial for improving maternal and child health outcomes.

Study Highlights:

– The median time-to-non-adherence to IFAS among pregnant women in Hosanna Town was 74 days.
– Age, education status, household’s wealth index, and counseling at service delivery were identified as independent predictors of time to non-adherence to IFAS.
– The findings suggest that women became non-adherent to IFAS before the recommended duration.
– Improving women’s education and counseling pregnant women on IFAS during pregnancy are recommended interventions to improve adherence.

Recommendations for Lay Reader:

– Pregnant women should be educated about the importance of iron and folic acid supplementation and the potential risks of non-adherence.
– Health care providers should provide counseling and support to pregnant women regarding IFAS during antenatal care visits.
– Policy makers should prioritize improving access to education for women and implementing strategies to enhance counseling services for pregnant women.

Recommendations for Policy Maker:

– Develop and implement educational programs to raise awareness among pregnant women about the benefits of iron and folic acid supplementation and the risks of non-adherence.
– Strengthen the training and capacity of health care providers to provide effective counseling on IFAS during antenatal care visits.
– Allocate resources to improve access to education for women, particularly in rural areas.
– Invest in the development and implementation of interventions to enhance counseling services for pregnant women, including the provision of adequate time and resources for counseling.

Key Role Players:

– Ministry of Health: Responsible for developing and implementing policies and programs related to maternal and child health, including iron and folic acid supplementation.
– Health care providers: Responsible for providing counseling and support to pregnant women during antenatal care visits.
– Community health workers: Play a crucial role in educating and raising awareness among pregnant women about the importance of IFAS.
– Non-governmental organizations (NGOs): Can support the implementation of educational programs and provide resources to improve access to education and counseling services.

Cost Items for Planning Recommendations:

– Development and printing of educational materials: Cost of designing and printing educational materials to raise awareness among pregnant women about IFAS.
– Training and capacity building: Cost of training health care providers and community health workers on counseling skills and the importance of IFAS.
– Infrastructure and equipment: Cost of improving infrastructure and providing necessary equipment in health facilities to support counseling services.
– Monitoring and evaluation: Cost of monitoring and evaluating the implementation of interventions to assess their effectiveness and make necessary adjustments.

Strength of Evidence

The strength of evidence for this abstract is 7 out of 10.
The evidence in the abstract is moderately strong, but there are some areas for improvement. The study design is a community-based cross-sectional study, which provides valuable information but does not establish causality. The sample size calculation is appropriate and the statistical analysis is well-described. However, the abstract could benefit from providing more details on the methods used for data collection and the inclusion/exclusion criteria. Additionally, it would be helpful to include information on the response rate and any potential limitations of the study. To improve the evidence, future studies could consider using a longitudinal design to establish causality and include a larger and more diverse sample to enhance generalizability. Furthermore, providing more information on the data collection process, such as the training of data collectors and measures taken to ensure data quality, would strengthen the evidence.

Abstract

The research was carried out in Hosanna Town, Southern Nations Nationalities and People Regional State (SNNPR) of Ethiopia. The town is situated in 230 kilometers to the south of national capital, Addis Ababa. According to 2007 census [16], total population of the town was 69,957; 35, 503 were males and 34, 454 were females. In the same census, the population growth rate in the region was 2.9% per year. Based on this, the projected total population of the Town for 2021 was 104,387. In the region, 23.3% and 3.5% population are women in their reproductive age and expected to be pregnant respectively [17]. Based on these evidences, there were 24,323 women of reproductive age (15–49 years) and 3,654 estimated pregnancies for year 2021. The research involving community based cross sectional study design was employed between May 15 and June 11, 2021. The source population consisted of all pregnant women in the research area and all pregnant women in selected sub-cities were sample population. The inclusion criteria considered all pregnant women in the study cluster who were booked for ANC one-week preceding the study. Pregnant women who were booked for ANC but whose registration for follow-up was less than one week prior the survey date were excluded. Also women who unable to recall their last normal menstrual cycle or gestational age (GA) at the time of booking were not allowed participating. The sample size was calculated with Epi Info version 7 using the double population proportion formula to detect a non-adherence rate of at least 25.1%, [14], 95% significance level and 5% margin of error; a sample size of 289 was obtained. The Cox proportional hazards model (power cox) was used to determine sample size for factors associated with non-adherence to IFAS using Stata version 15.0 considering the presence of censoring and adjusting for others. Non- adherence was considered a failure (outcome). Factors obtained from literatures having a significant association with adherence to IFAS were considered for sample size calculation; knowledge about IFAS [6], counseling on IFAS [12], Partner support [18], and Educational status [19]. After computing various factors, sample size calculated for educational status was 139; the largest sample size computed for factors associated with IFAS adherence. Therefore, the minimum sample size required for the non-adherence estimation would be 289. Considering 5% non-response rate [14, 20, 21] and design effect of 1.5, the final sample size was (289* +5%) *1.5 = 456. Where n is required sample size, zα/2 is 95% CI, p is population proportion and d is margin of error. Study clusters were identified using a two-stage cluster sampling procedure. After randomly selecting five Kebeles (the smallest administrative structure), total number of predefined distinct clusters (Mender) (smallest cluster within an administrative Kebele) were obtained, after which we obtained the size of the pregnant women for each cluster. The required number of clusters from each Kebele were assigned using probability proportional to population size approach, in which larger settlements have a higher chance of being selected as clusters. Reserve clusters were used until we obtain required sample size (Fig 1). Ethiopian Demographic and Health Survey 2016 (EDHS,2016) [22] and relevant literature [6, 11, 14, 18, 23] were used to adapt data collection tool. A household’s wealth status was computed based on 23 household assets and housing quality variables which were adapted from EDHS2016 [22], given that the study setup is urban. First, all the study participants were asked about the ownership of assets by their respective households. Those who owned the asset received a score of "1," while those who did not received a score of "0". A structured questionnaire was prepared in English, translated to Amharic, and then back translated into English to ensure consistency in order to measure the required parameters. The Amharic version tool was then employed. When a pregnant woman visiting an antenatal clinic took IFA tablets for less than four days per week for the week prior to the survey or for less than 90 days during the third trimester of pregnancy [12, 13]. Pregnant women who were adherent at the time of data collection were censored observations. When a pregnant lady took IFAS tablet for less than four days one week preceding the survey date, the event occurred. Is the amount of days a pregnant woman contributed while on IFAS until she experienced an event of interest (non-adherent) or censorship. The dependent variable was time to non- adherence of IFA supplementation measured in days. Censored observations were denoted by 0, whereas events were indicated by 1. The period of time a pregnant woman spent on IFAS in days was determined as the difference between the entire GA (from LNMP to the day of data collection) and the calculated and/or reported GA at initial booking for ANC. Maternal age and educational status, household’s wealth index, counseling status of health institution on IFAS, knowledge on IFAS, waiting time to receive care, frequency of ANC visit, history of adverse fetal outcomes and history of anemia were variables hypothesized to be independent predictors of time to non-adherence which is a primary outcome in this study. To ensure data quality, ten data collectors and one supervisor received two days training on tool clarity and overall data collection processes. The training was emphasized on sociodemographic information’s, household’s wealth status, obstetric factors, Personal exposure to media, health facility related factors, and Knowledge on IFAS and anemia. Structured questionnaire was prepared in English, translated into Amharic, and then back translated into English to ensure consistency. A pre-test was conducted on 5% of the sample size in a nearby town. Cronbach’s alpha was done to assess internal consistency (alpha coefficient for household wealth status (23 items) = 0.85, media access (3 items) = 0.71, counseling at health facility (7 items) = 0.76, Knowledge on IFAS (15 items) = 0.79). A public health officer supervised the data collection process, while the primary investigators (PIs) supervised the whole technique. All collected data were handled to PIs and checked and cleaned for consistency and completeness; daily discussion was held in case of inconsistencies. Epi-Data version 3.1 for data entry and Statistical Package for Social Science (SPSS) version 23 for analysis were use. Before analysis missing value, new categories and normality for continuous variables were checked. Households wealth index was computed by principal component analysis (PCA) based on household assets and housing quality variables which adapted from EDHS 2016 [22]. Pregnant women’s knowledge on IFAS was computed after performing PCA based on 15 items. Problematic variables were removed step by step, eleven items having four component factors that explains a total variance of 64.3% were retained; whose alpha coefficient was 0.78, all having acceptable correlation matrix (KMO = 0.78, x2 = 1048, P 0.5. The value of retained variables was aggregated and used median as a cut off to declare knowledge status of study population. The difference between total GA (spanning from last normal menstrual cycle to date of data collection) and the calculated and or reported GA at first booking for ANC was taken as total time contributed in days during which a pregnant woman was on IFAS. Survival curve was used to display the survival status (time to non-adherence) among different characteristics. For survival analysis, the outcome variable was dichotomized to event and censored. The assumptions of proportional hazard were tested statistically and graphically. Against each categorical variable, we performed the log-og survival plot and the Kaplan-Meier survival plot. Both log-log survival plot and Kaplan-Meier survival and predicted plot revealed that the plots were parallel to each other. We have also conducted Schoenfeld test with the corresponding p-value for all variables. The Kaplan-Meier test was used to assess the median survival time between groups. The multivariate Cox Proportional Hazard model was used to examine the factors associated with time to non-adherence. The crude and adjusted hazards ratios with a 95% confidence interval (CI) were used as a measure of effect size. The Cox proportional hazard model assumption was tested graphically using log-minus-log survival plots against time for predictors. Multivariable Cox proportional hazard regression model was used to control the confounding effect of variables. In bivariate analysis, variables having a p-value < 0.25 were selected as potential predictors and used in multivariable analysis. A p-value < 0.05 with a corresponding 95% CI was declared statistically significant. The study was approved by the Institutional Review Board (IRB) of Hosanna Health Science College. In addition, permission was obtained from health department of the local government offices. Informed written consent was obtained from all participants. Respondents were informed that they had the right to refuse or discontinue the interview. The information provided by each respondent was kept confidential. Women who were non-adherent at the time of data collection were successfully counseled on the benefits of IFAS.

Materials

Innovations

Based on the research study titled “Time to non-adherence to iron and folic acid supplementation and associated factors among pregnant women in Hosanna town, South Ethiopia: Cox-proportional hazard model,” the following recommendations can be made to improve access to maternal health and increase adherence to iron and folic acid supplementation (IFAS) among pregnant women:

1. Strengthen education and awareness: Implement educational programs that emphasize the benefits of IFAS during pregnancy and the potential risks of non-adherence. Focus on improving women’s education and knowledge about the importance of IFAS.

2. Enhance counseling services: Provide comprehensive counseling on IFAS during antenatal care visits. Train healthcare providers to effectively communicate the importance of IFAS, address concerns or misconceptions, and provide guidance on proper adherence.

3. Address socioeconomic factors: Develop interventions that address socioeconomic barriers such as poverty and limited resources. Provide financial support or subsidies for pregnant women to afford IFAS.

4. Community engagement and support: Engage the community in promoting maternal health and IFAS adherence. Utilize community health workers or volunteers to raise awareness, provide support, and reinforce the importance of IFAS through home visits, group discussions, and community events.

5. Integration of IFAS into existing health services: Integrate IFAS into routine antenatal care services. Incorporate IFAS provision and counseling into existing maternal health programs and strengthen health systems to support the delivery of IFAS.

These recommendations, if implemented, can help improve access to maternal health and increase adherence to iron and folic acid supplementation among pregnant women, leading to better pregnancy outcomes and reduced maternal anemia.

AI Innovations Description

The research study titled “Time to non-adherence to iron and folic acid supplementation and associated factors among pregnant women in Hosanna town, South Ethiopia: Cox-proportional hazard model” provides valuable insights into the factors influencing non-adherence to iron and folic acid supplementation (IFAS) among pregnant women. This information can be used to develop recommendations and innovations to improve access to maternal health.

Based on the study findings, the following recommendations can be made:

1. Strengthen education and awareness: Improving women’s education and knowledge about the importance of IFAS during pregnancy is crucial. Implementing educational programs that emphasize the benefits of IFAS and the potential risks of non-adherence can help increase awareness among pregnant women.

2. Enhance counseling services: Providing comprehensive counseling on IFAS during antenatal care visits is essential. Health care providers should be trained to effectively communicate the importance of IFAS, address any concerns or misconceptions, and provide guidance on proper adherence.

3. Address socioeconomic factors: The study identified household wealth index as a predictor of non-adherence. To improve access to IFAS, interventions should focus on addressing socioeconomic barriers such as poverty and limited resources. This could include providing financial support or subsidies for pregnant women to afford IFAS.

4. Community engagement and support: Engaging the community in promoting maternal health and IFAS adherence can be beneficial. Community health workers or volunteers can play a crucial role in raising awareness, providing support, and reinforcing the importance of IFAS through home visits, group discussions, and community events.

5. Integration of IFAS into existing health services: Integrating IFAS into routine antenatal care services can help ensure consistent access and adherence. This can be achieved by incorporating IFAS provision and counseling into existing maternal health programs and strengthening health systems to support the delivery of IFAS.

By implementing these recommendations, it is possible to improve access to maternal health and increase adherence to iron and folic acid supplementation among pregnant women, ultimately leading to better pregnancy outcomes and reduced maternal anemia.

AI Innovations Methodology

To simulate the impact of the main recommendations on improving access to maternal health, the following methodology can be used:

1. Strengthen education and awareness: Implement educational programs in the community to increase awareness about the importance of iron and folic acid supplementation (IFAS) during pregnancy. Measure the impact by conducting pre- and post-intervention surveys to assess changes in knowledge and awareness among pregnant women.

2. Enhance counseling services: Provide comprehensive counseling on IFAS during antenatal care visits. Track the number of pregnant women receiving counseling and assess their adherence to IFAS through regular follow-up visits. Compare the adherence rates before and after implementing enhanced counseling services.

3. Address socioeconomic factors: Implement interventions to address socioeconomic barriers such as poverty and limited resources. Provide financial support or subsidies for pregnant women to afford IFAS and track the uptake of these interventions. Compare the adherence rates among women who receive financial support versus those who do not.

4. Community engagement and support: Engage community health workers or volunteers to promote maternal health and reinforce the importance of IFAS. Monitor the activities of community health workers, such as home visits and group discussions, and assess their impact on adherence rates.

5. Integration of IFAS into existing health services: Integrate IFAS provision and counseling into routine antenatal care services. Monitor the implementation of this integration and track the adherence rates among pregnant women receiving antenatal care.

To evaluate the impact of these recommendations, collect data on adherence rates to IFAS before and after implementing the interventions. Compare the adherence rates and assess the statistical significance of any changes using appropriate statistical tests. Additionally, conduct qualitative interviews or focus group discussions with pregnant women to gather their perspectives on the interventions and their experiences with IFAS adherence.

By analyzing the data collected and considering the perspectives of pregnant women, it will be possible to assess the impact of the recommendations on improving access to maternal health and increasing adherence to IFAS. This information can inform future interventions and policies to further enhance maternal health outcomes.

Author

Dima Health

Identifiers:

DOI: 10.1371/journal.pone.0275086

Research Areas:

Community Interventions, Environmental, Health System and Policy, Maternal Access, Maternal and Child Health, Sexual and Reproductive Health

Study Countries:

Ethiopia

Study Design:

Cohort Study, Cross Sectional Study, Grounded Theory

Study Approach:

Quantitative

Participants Gender:

Female