Study design and setting

A quantitative, cross-sectional study was conducted among women with children less than two years in the Volta Region (Volta and Oti) of Ghana. The Volta Region was one of Ghana’s ten (10) administrative regions at the beginning of the study which was later split into sixteen (16) administrative regions. It is situated in the eastern part of the country sharing boundaries in the north with the Northern region, the south with the Gulf of Guinea, in the west with the Volta Lake, and in the east with the Republic of Togo. From the routine service, the region had IPTp3 Coverage of 50.6% in 2019, with Nkwanta South having the lowest coverage of 29% and North Tongu District having the highest coverage of 76.1%. However, comparing the average coverage from 2015 to 2019, Nkwanta North had the lowest coverage of 23.1% with North Tongu having the highest coverage of 59.5%, and both districts were selected for the study. The Volta Region was selected because it presents a peculiar and widespread case of the problem under investigation. Based on data from the Policy, Planning, Monitoring and Evaluation Directorate (PPMED) of the Ghana Health Service (GHS), the Volta Region was ranked 10^th (last) on the “Performance Gauge” (PG) scale in 2016 and 2017. The Holistic Assessment of the Health Sector Programme of Work 2014 and the DHS report of 2014 mentioned the Volta Region as an area that needs special attention. The region is, therefore, seen as a classical area for determining the demand-side predictors of intermittent preventive treatment of malaria in Ghana [25].

Study population

The study population includes all women who gave birth twenty-four months preceding the survey from the communities within the region expected to have used antenatal services. The survey date was 1^st June 2020 to 10^th July 2020.

Inclusion and exclusion criteria

Variables

Based on the information reviewed from earlier studies done elsewhere. This study included the following variables that have been linked in theory and empirically to malaria in pregnancy and the use of IPTp-SP. The outcome/dependent variable was the uptake of 3+ doses of IPTp-SP by women exposed to the policy and independent/explanatory variables included both individual-level demographic characteristics of the respondents and IPTp-SP implementation-related characteristics. This includes age, marital status, educational level of woman, occupation of the woman, religion, ethnic group, partner education, partner occupation, partner age, duration of the marriage, monthly household income, household assert, means of transportation to the nearest health facility, estimated traveled time to the nearest health facility, estimated distance to the nearest health facility, health insurance enrollment, and health insurance validity status, gestational age booked for antenatal, gestational age for IPTp-SP, gravida, parity, number of live births, and number of antenatal visits.

Sample size determination and sampling

The formula for sample size determination based on the Yamane (1967) approach is stated as follows [26] Eq 1

Where n is the sample size, N is the population size, e is the level of precision. Yamane’s (1967) approach to sample size determination assumes that the population of the study is known and is finite. The sample frame (target) population is estimated to be 76,341 for the entire Volta Region. The appropriate sample size is determined based on Yamane’s (1967) approach at the precision of 5% as presented below [26]: Eq 2

The required sample size that was selected is therefore 398. Given the possibility that not all the target respondents were reached (non-response), 10% of the obtained sample size, four hundred and thirty-eight (438), were added to compensate for persons that the researcher was unable to reach.

The sampling procedure was adapted from the 2014 Ghana Demographic and Health Survey (GDHS) and followed a two-stage sample design intended to allow estimates of key indicators for urban, and rural areas in the two selected districts. The first stage using probability proportional to the size and then thirty households per cluster scenario [27], the researcher selected eight (8) clusters for North Tongu and six (6) clusters for Nkwanta North, giving a total of fourteen(14) clusters for the study area with five (5) urban and nine (9) rural. The second stage of sampling involved the selection of the sampling unit for enumeration after listing all members of each household in the enumeration area. All the eligible women were sampled based on the population proportion to the size of the eligible population contribution after the listing of all the members of the households in the clusters and then a random selection of the eligible population was done using Stata to select the respondents. In all, 438 women were selected at random for the interview. However, those that were not available at the time of the enumeration were replaced by the next in the randomized order.

Data collection and quality management

The study used a structured questionnaire for data collection. The questionnaire solicited data on the demographics and use of IPTp-SP. Data on the number of antenatal visits, number of doses of IPTp-SP, parity, and gravida was obtained from the questionnaire and triangulated by their maternal health records. Data was captured using Kobo collect app, data extracted as a Microsoft Excel file, and exported to STATA 14.1 for analysis. Data cleaning was done at every stage to ensure that we had good-quality data for analysis.

Data analysis

In the analysis, the continuous independent variables were categorized into intervals range as required whilst the categorical variables such as sex, educational level, ethnicity, etc. retained their categorization during analysis. To ascertain the wealth index, the researcher had to first conduct a principal component analysis of the household assets and possessions to establish which variable should be included in the model to determine the wealth of the household. In the study, data was collected on household assets and possessions which may be correlated in an unknown and complex way. Thus, when a set of variables are correlated in a complex and unknown way along several dimensions, Principal Component Analysis (PCA) is employed to reduce these variables by assessing which variables behave similarly. Based on the variables and their relationship to each other, PCA creates a new set of variables called the principal component. The household assets and possessions collected during the study were assigned weights based on the PCA and the resulting scores were standardized to a standard normal distribution. Based on these aggregate scores, individuals and their household wealth were put into five quintiles.

The chi-square test of associations was used to check for dependence between the sociodemographic variables and the outcome variables. All independent variables with p<0.05 in the bivariate analysis were included in the multivariable logistic regression to further examine the association between the outcome and each independent variable while controlling the effect of other explanatory variables by calculating adjusted odds ratios (AOR). The level of significance used was 5% (0.05), two-tailed at 95% confidence interval (CI).

The goodness of fit of the model was tested by using the criterion-based method, usually known as the “Leaps-and-Bounds algorithm”, which was used in the variable selection because it involved a wider search and compared models in a preferable manner. It also identified the subset of variables with the most discriminatory power (predictive sense) is proposed. It minimized the parametric estimator of the error rate among all possible variable subsets, evaluating only a fraction of the total number of subsets. These criteria took into consideration the nature of the problem and the objective of the analysis.

To investigate the effect of the predictor variable on the outcome variables, binary logistic regression analysis was done at two levels. The first level was to establish the crude level of association between the independent variable and the dependent variables for the study. For this study, all independent variables were added to the model to establish the crude relationship with the dependent variables at various levels of measurement in accordance to the study objectives. Afterward, variables that were significant in the first model at the bivariate analysis level were moved to another model to adjust for the effects of those variables. A p-value < 0.05 was considered statistically significant.

Ethical approval and consent to participate

Ethical clearance was obtained from Ghana Health Service’s Ethical Review Committee (ERC) with protocol ID No: GHS-ERC 005/12/19 through the University of Ghana, Business School. The researcher obtained written informed consent and assent were obtained from each study participant whose age was 18 years and above and below 18 years respectively. Willingness to participate in the study and parental permission was confirmed by signing or fingerprinting on the informed consent form.

Sociodemographic characteristics of the service users

The average age of the respondents was 29 years (SD:6.1) with 211 (54.8%) of them falling within the 20–29 age group, 237(54.1%) of the respondents, had lower than Junior High School/Junior Secondary School/Middle education. Christians, 353 (80.6%), formed the majority of the religious groupings. Seventy-eight percent (77.6%), were married. With 155(35.39%) of the respondents were Farmers. The most dominant ethnic group was Ewes, who constituted 54.79% of the respondents. Most of the partners of the women, 196 (44.7%), were between the ages of 30 and 39 with a mean age of 35 years (SD:7.6). One hundred and eighty-nine (43.2%) of the partners of the respondents had less than JHS/JSS/Middle educational and 187(42.69%) of them were farmers. Most of the respondents, 242 (55.25%), had been married for 3 to 9 years with an average duration of 7 years (SD:4.9). An average household income for the respondents was GH ¢349($34.9) per month with a majority of them, 243 (55.48%), earning between 100($10) to 999($99) Ghana cedis per month. The highest wealth index quintile was the second with 90 (20.55%) of the women in this category. This was computed for the list of assets like owning a mobile phone, a television set, radio, bed, farmland, means of transport, a house, etc. The majority of the respondents, 245 (55.9%), resided in the rural areas of the region with only 193 (44.1%) in the urban areas. 210 (47.95%) of the respondents, walked to the nearest health facility during antenatal care with the average estimated travel time and distance being 32 minutes (SD:21.0) and 8km (SD:9.1) respectively. A high number of the respondents, 282 (64.68%), booked for antenatal care in the first trimester. An overwhelming majority of them, 408 (93.15%), had enrolled in the National Health Insurance Scheme. However, only 221 (50.46%) of them were active to access free services. The majority of the women, 237 (54.1%), had a parity of 0 to 2 pregnancies, the average gravida was 3 (SD:1.9) pregnancies and the average number of live births was 3 (SD:1.8) children as shown in Table 1 below.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Background characteristics of IPTp-SP service users.

https://doi.org/10.1371/journal.pone.0308321.t001

Uptake of 3 plus doses of IPTp-SP

The mean number of antenatal care (ANC) attendance was 5 (SD:2.6) visits per client, with 262 (59.82%) of them getting the 3 plus doses of IPTp-SP while 44 (10.1%) of the respondent did not receive any dose of IPTp-SP. Also, 221 (50.46%) were booked for IPTp-SP before the 17^th week of pregnancy as shown in the Table 1 below.

Crosstabulation of background variables and uptake of 3+ doses IPTp-SP

For the IPTp-SP uptake for prevention of MiP against sociodemographic variables, only one variable out of 25 was significant at (p<0.05) and thus had an association with the uptake of 3 plus doses of IPTp-SP. Some variables like age of women, marital status, ethnic group, partner’s age, district of residence, gestational age at booking for ANC, National Health Insurance Scheme (NHIS) card availability, NHIS card active, gravidity, parity, number of livebirths, number of antenatal visits and gestational age at booking for IPTp-SP, which are known in the literature to influence uptake, had no association with the uptake. Women who attend antenatal in the first trimester are more likely to get 3+ doses of IPTp-SP as shown in the Table 2 below:

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Association between sociodemographic variables and uptake of 3+ doses of IPTp-SP.

https://doi.org/10.1371/journal.pone.0308321.t002

Logistic regression for uptake of optimal doses of IPTp-SP

From the crosstabulation (Table 2) above, only the number of antenatal attendances had a connection with the IPTp-SP uptake (p<0.05). Thus to select the model with goodness of fit, the criterion-based method, usually known as the “Leaps-and-Bounds algorithm”, was used. Education level of the women was included based on information from the literature indicating that higher women’s education is associated with uptake of IPTp-SP. Thus, controlling for confounding variables, the optimal model selected among the 4 nested models was model 2 (AIC = 532.2588). The predictors selected are the number of antenatal care visits and maternal education. The number of antenatal attendances was connected with the uptake of 3 plus doses (optimal doses) of IPTp-SP. Overall, the respondents who attended the antenatal clinic 4–7 times had 7 (CI:3.9–12.3) times higher uptake of 3+ doses of IPTp-SP as compared to others who attended less than 4 visits. Similarly, women who had 8 or more visits had a 16.1 (CI: 5.9–43.6) times higher chance of getting more than 2 doses of IPTp-SP as compared with others who had fewer than 4 attendances as shown in Table 3 below:

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Crude and adjusted associations between variables and uptake of 3+ doses of IPTp-SP.

https://doi.org/10.1371/journal.pone.0308321.t003

Discussion of the predictors of IPTp-SP uptake

One of the major objectives of this study is to examine the magnitude and the reasons related to the uptake of 3+ doses of IPTp-SP by women in the Volta Region after the policy update by WHO in 2012. The research has revealed that generally, the percentage of women who took 3+ doses of IPTp-SP was 59.8%, which is nevertheless extremely lower in the surveyed populace as compared to the Roll Back Malaria (RBM) aim of a minimum of 80% for all pregnant women living in regions with high malaria spread [28]. Among the twenty-five (25) covariants that were examined to describe the of 3+ doses of IPTp-SP uptake among pregnant women in the Volta Region in Ghana, only the number of antenatal attendance was discovered to influence the uptake of IPTp-SP after modifying other determinants. The explanatory examination of this enquire reveals that the percentage of the women that receive 3+ doses of IPTp-SP is the highest within women making 4 or more antenatal care attendance than those making less attendance. Likewise, the corrected odds ratio scores indicate that pregnant women who visit antenatal care less than 3 times during gestational time had a lesser chance of taking maximum doses of IPTp-SP likened to pregnant women who had completed 4+ attendance. Thus, with the new requirements of making 8 visits or more, pregnant women who attend ANC 8 times or more have higher odds of getting the maximum 3+ dose as likened to pregnant women who had just 4 attendances. For pregnant women who booked for IPTp-SP later after the 16^th week of gestation, the adjusted odds ratio shows that these women have lesser odds of getting the maximum doses of IPTp-SP as compared to individuals who book for antenatal care earlier. This result is coherent with research in Malawi [29, 30], Uganda [21], Tanzania [23, 31], Ghana [32] and Burkina Faso [33]. Antenatal care clinics offer a stage for vital health packages and services such as health promotion, prevention, testing, and identification of illness, all intended at enhancing the health of pregnant women and their unborn children must be intensified [34]. Using IPTp-SP is one of the prevention strategies being executed at the antenatal care in the Volta Region. The 2012 update policy recommended 3+ doses of IPTp-SP and the first dose should be received as timely as feasible through the second trimester of gestation [35]. This suggests that timely ANC commencement would boost the likelihood of getting the maximum IPTp-SP doses as there is a probability of more antenatal clinic attendance to be made. In this study, 64.8% of the women commence antenatal clinic attendance during their first trimester, which agrees with the WHO recommendations that pregnant women must begin antenatal clinic attendance during the first trimester of pregnancy [36]. The results from the research also reported that less than half of the study respondents visit the antenatal clinics at least 8 times, implying the slow-paced of use of focused antenatal care (FANC) strategy was utilized at the period of the research as a guide. FANC strategy proposed at least eight (8) antenatal care attendance and examination by a skilled health provider before delivery [37]. The reviewed WHO policy was designed to match with the FANC Model in order to boost the maximum doses of IPTp-SP uptake. However, only 11.7% of pregnant women who made eight-plus attendance get the maximum doses of IPTp-SP. This is an overlooked chance to give the proposed maximum doses of IPTp-SP to women who attend antenatal at least 8 times before delivery [38].

The difference concerning the proportion of 8+ antenatal visits and the 3+ doses of IPTp-SP uptake could occur partially because of recurrent unavailability of SP in the antenatal care clinics. Additional clarifications might be that certain antenatal clinic attendance was not planned because of unforeseen problems; thus IPTp-SP doses were not given.

Limitations

Geographically, the scope of the study covers only the Volta Region (Volta and Oti) of Ghana, specifically the two purposively selected districts (North Tongu and Nkwanta North). In terms of the study unit, the scope was also narrowed to women who had given birth in the past 24 months. The focus of the study was on IPTp-SP policy and not on any other malaria intervention. Another possible limitation came from the use of questionnaires. The reliability of the data from the questionnaire depended on the skills and experience of field enumerators Research Assistants (RAs) that conducted the interviews as well as the extent to which the respondents could remember facts and events. Also, only well-trained graduates were used in the data collection to reduce errors. Furthermore, the study had financial challenges, hence the choice of only two districts for the study in the Volta Region. What this means is that the findings of the study should be interpreted with caution because the two participating districts were purposively selected based on low and high coverage of IPTp-SP [39].