Study area
The study area was Maiduguri, the Borno state capital which is located in North-eastern Nigeria. Borno state lies between latitudes 10° 30' and 13° 50' north and longitudes 11.00° and 13° 45' east, with a total land area of 69,435 km2 [32]. The climate varies according to the time of the year, with temperatures ranging from 25oC to 44oC, and an average annual rainfall of 613 mm [32]. Its population is reported to be 540,016 consisting of 282,409 males and 257,607 females [33]. The study location which was the State Specialist Hospital, is one of the three secondary-level hospitals in Maiduguri, Bono State. It was chosen because it happens to be the biggest of the three state hospitals in terms of size, man-power and patient load. Its location is also at the centre of the city, making it the most geographically accessible, to the Maiduguri populace. The ante-natal care clinic is run as a unit under the Department of Obstetrics and Gynaecology, and is run from Mondays to Fridays. Mondays are reserved for booking, while follow-up visits are conducted on the other days. Each booking clinic session has an average load of approximately a hundred clients. Booking at the hospital’s clinic is open to all pregnant women, irrespective of their gestational ages, and with no requirement for referral letters from other centres. Routine antenatal health education is given by the mid-wives at the waiting hall, before antenatal care consultations start. The talks given are not structured, and cut across issues of hygiene, proper nutrition, breastfeeding, malaria prevention and any health issue deemed relevant by the staff.
Study design and study population
The study was a double-blind parallel-group randomized controlled trial guided by the CONSORT Statement [34]. Participants were randomly selected from the ante-natal care attendees, and then randomly assigned to either the intervention or control group (with equal number of participants in each group). The intervention group received the IMB-based health educational intervention on malaria, while the control group received health education on breast feeding. The study population was pregnant women attending the State Specialist Hospital, Maiduguri for their ante-natal care.
Those who do not understand Hausa language were excluded because the intervention was prepared and delivered in Hausa language. Also excluded were participants who had history of medical conditions which could interfere with the study outcomes like: hypertension, pre-eclampsia or eclampsia [35-37], diabetes [38,39] and history of per vaginal bleeding in index pregnancy [40]. The minimum sample size required to test the study hypotheses was calculated using the two proportions formula [41]. In a previous study which had included ITN use and IPTp [42], substituting the results of ITN use (P = 0.295; P1 = 0.38; P2 = 0.21) gave a higher sample size. Subsequent adjustments were made for an estimated 30% drop out, and an estimated eligibility of 70%, to give a final sample size of 352 (176 participants in each group).
Randomization
Random assignment of the participants to either the intervention or control group, was done on the same day the participants were selected into the study. It was however performed after baseline data had been collected.
Sequence generation
The permuted block randomization technique was used. Permuted blocks of four (each containing two interventions and two controls), containing all the six possible combinations (AABB, BBAA, ABBA, BAAB, ABAB, BABA) were generated using the random function in Microsoft Excel 2013.
Allocation concealment mechanism
The sequences generated were then placed inside opaque envelopes and sealed, by the medical records’ staff who generated them. The envelopes were also serially numbered from the outside, to guide those doing the allocation.
Implementation
The sequence generation was performed by a trained staff of the hospital’s Medical Records Department, who was not part of any of the other research processes. Two other staff of the antenatal clinic worked independently to assign groups to the participants. The first staff serially handed over the envelopes to the participants without opening, then directed them to the second staff. The second staff then opened the envelopes, informed them of the day to come for their health education session (based on the group to which they belonged), and then documented their respective groups on a sheet provided. Each participant was then given a hand card, which carried their bio-data and serial numbers. The hand card also contained a short note indicating that they had been enrolled into a follow-up study, and requesting of the attending physician or nurse to kindly complete the post-natal findings as indicated on the card.
Blinding
The study was double-blinded, as participants and assessors (enumerators) were blinded to the assigned interventions. The list of assignments was then kept confidential by the staff who documented the participants’ allocations. This staff was only aware of the group coding (A and B), but unaware of the interventions assigned to them. To ensure blinding of the participants, the groups they belonged to, were not indicated on their cards. The subsequent clinic appointment dates for the two groups were set on different days, to minimize contamination. The enumerators were blinded, as they were not involved in any of the processes of group allocation, or delivery of the intervention.
Development of the health educational intervention modules
For the malaria health educational intervention, the information sources for developing the modules were: the National Guidelines and Strategies for Malaria Prevention and Control During Pregnancy, a publication of the Nigerian Federal Ministry of Health [43]; WHO published materials [11,44]; publication of the Global Health Learning Centre on Malaria in Pregnancy [45]; and other publications from studies carried out in Nigeria. It was developed based on the Information-Motivation-Behavioural Skills theory [29], using local scenarios for demonstration purposes. The breastfeeding modules were developed from the following sources: National Policy on Infant and Young Child Feeding in Nigeria [46]; Lactation Management Self-Study Modules, Level I [47] and other studies conducted in Nigeria on breastfeeding practices. The items extracted from these sources were compiled to follow a similar pattern with that for the malaria health education intervention. This topic was chosen because of its relevance to their pregnancy state (that is, being prospective mothers).
Structure of the malaria health educational intervention modules
The intervention consisted of four modules, covering each of the constructs of the IMB model as presented in Table 1. The first and second modules covered the information construct, the third module covered the motivation construct, while the fourth module covered the behavioural skills construct.
Table 1. Tabular illustration of the malaria intervention modules by construct and contents
Theory
Construct
|
Module &
Strategy
|
Contents
|
Estimated time
|
Information
|
Modules 1
&
Module 2
(Lectures)
|
Transmission;
Clinical features;
Complications;
Prevention measures
|
30 minutes
&
30 minutes
|
Motivation
|
Module 3
(Interactive discussion;
Brainstorming)
|
Participants’ experiences and those of other pregnant women
|
1 hour 30 minutes
|
Behavioural skills
|
Module 4
(Lectures;
Videos;
Demonstrations)
|
Evaluation of self-efficacy for taking IPTp-SP and using ITN; Goal setting
|
1 hour 30 minutes
|
The first module was titled ‘Understanding Malaria in Pregnancy’. This module gave a general discussion on malaria in pregnancy, at the end of which the participants were expected to be able to identify what causes malaria; its mode of transmission; its signs and symptoms; its complications during pregnancy; and how it can be prevented. This section was delivered basically through lectures. The second module was titled, ‘The Main Preventive Measures for Malaria in Pregnancy’ which introduced the two most important preventive measures for malaria during pregnancy (TN use and IPTp uptake). This module discussed the importance and efficacy of these preventive measures, as well as success stories of these measures as reported in some studies. Participants were expected to be able to highlight the protective importance of IPTp and ITN at the end of this section. This section was also delivered through lectures.
The third module was titled, ‘Motivation for Malaria Prevention during Pregnancy’, and it dwelt on motivating the participants to adopt these protective measures. This module was designed to guide an interactive session, for misconceptions and other possible deterrent factors to be addressed interactively. The facilitator together with the participants were to brainstorm and proffer possible ways of overcoming those deterrent factors. At the end of this session, participants were expected to be apparently convinced of the efficacy and safety of these preventive measures, and also motivated to carry them out. The fourth module titled, ‘Insecticide Treated Net and Fansidar’, focused on empowering the participants. They were to be practically shown how to hang, and care for their insecticidal nets, and how to look out for defects in their nets, and how to repair them. They were also taught about the timings, indications and contra-indications for IPTp, and also educated on how to take their drugs. Their level of self-efficacy was to be assessed at a group level, and collectively, goals were set to achieve better adherence to these malaria preventive measures. There were evaluation questions at the end of Modules 1, 2 and 4.
Structure of the breastfeeding health educational intervention modules
The intervention consisted of three modules, covering each of the constructs of the IMB model. As presented in Table 2, the first module covered the information construct, the second module covered the motivation construct, while the fourth module covered the behavioural skills construct. The first module titled, ‘An Overview of Breastfeeding’ had two sections: the first section was on the benefits of breastfeeding and the consequences of not breastfeeding; while the second section discussed the practices to avoid during breastfeeding, alternatives to breastfeeding, and challenges associated with breastfeeding such as sore nipples, low milk supply and breast engorgement. The second module titled, ‘Motivation on Breastfeeding’ contained a list of misconceptions and other deterrent factors for exclusive breastfeeding as identified in previous studies. These were to be discussed and brainstormed by the participants under the facilitator’s guidance to come up with solutions. The third module titled, ‘Practical Session’ was for breastfeeding procedures to be demonstrated practically using models as well as video demonstrations to be shown to the participants.
Table 2. Tabular illustration of the breastfeeding intervention modules by construct and contents
Theory
Construct
|
Module &
Strategy
|
Contents
|
Estimated time
|
Information
|
Module 1
(Lectures)
|
Benefits of breastfeeding,
Consequences of not breast feeding.
Problems associated with breastfeeding.
|
30 minutes
1 hour
|
Motivation
|
Module 2
(Interactive discussion;
Brainstorming)
|
Participants’ experiences and those of other pregnant women
|
1 hour 30 minutes
|
Behavioural skills
|
Module 3
(Lectures;
Videos;
Demonstrations)
|
Practical demonstrations of breastfeeding techniques; expression of breast milk
|
|
Quality control of the health educational interventions
This entailed appraising the modules for accuracy, relevance and comprehensibility, as well as giving the facilitator adequate training on the modules.
Quality control of the module contents
The modules for the intervention group as well as the control group were appraised for their contents and relevance by expert teams. For the malaria health educational intervention modules, the team comprised of three Public Health specialists, an Obstetrics and Gynaecology specialist, a Health Educator, and a Mid-wife; while the appraisal team for the modules on breastfeeding comprised of three Public Health specialists, a Health Educator, and a Paediatric nurse. Both interventions were pre-tested with a sample of twenty five pregnant women from the same hospital, who were not part of the study. Appropriate feedback was then obtained, from both the experts as well as the participants, and appropriate modifications and corrections were effected.
Training of the module facilitator
One facilitator who was a midwife was chosen to deliver the modules at all the health education sessions for both groups. For each of the two interventions, the facilitator received two sessions of one-on-one training by the researcher. The first session lasted about four hours long, during which the contents of the modules, and how to deliver them within their limits were treated. A revision session was then held again in a similar way after the module had been revised (after the pre-testing).
Delivery of the malaria IMB-based health educational intervention
The interventions were given a week after baseline data had been collected. As suggested by McMillan,McMillan et al. delivery of educational interventions should be in modes that are consistent with what is used in practice [48], and as such, they were given at a group level, and in a single-session, with each session having around fifty participants. The module was delivered by the facilitator (a mid-wife nurse), so as to eliminate experimenter effects. The facilitator was blinded to the objectives of the study, and delivery of the module was also closely monitored and supervised by the researcher, while giving necessary feed-back to the facilitator to ensure that the module was strictly followed without straying. The facilitator went through all the four modules, while actively engaging the participants. The durations were approximately 30 minutes, 30 minutes, 1 hour 30 minutes and 1 hour 30 minutes, for modules 1, 2, 3 and 4 respectively, with breaks of about 10 to 15 minutes between modules. Each participant was given the chance to demonstrate what she had learnt, and corrective feed-back was given.
Delivery of the breastfeeding health educational intervention
The health education modules on breastfeeding were delivered by the same midwife who gave the malaria health education. The total duration of the session was also around four hours, and the delivery was of similar methodology to that of the malaria health education modules, comprising of lectures with power-point presentations, videos, discussions, and practical sessions. A paediatric nurse helped out with the demonstrations using models. The delivery was also strictly monitored to ensure strict adherence to the module and its contents, without straying.
Follow-up of Participants
After collecting baseline data, randomization, and delivery of the intervention, participants were then followed up. Follow-up data was collected using the same questionnaire, at 2-months post-intervention and at 4-months post intervention. Participants were also requested to always carry their hand cards with them each time they visited the hospital, especially when they were going for delivery, or in the case of a home delivery, to present to the hospital immediately. Data on babies’ birth weights and pregnancy outcomes were filled on the hand card by the attending physician or nurse, which participants submitted back after their delivery. For each visit, compensation for transport fare was given to each participant.
Variables
The independent variable in this study was group, which was either the intervention or control group. The dependent variables comprised of primary outcomes (ITN use and IPTp uptake) and secondary outcomes (malaria diagnosis, haematocrit, pregnancy outcome, and babies’ birth weights). Other variables which were considered potential confounding factors in this study, included ethnicity, family type, type of residence, educational status, occupational status, income level, age at first marriage, gravidity; parity, period of amenorrhoea, history of previous preterm delivery, and history of previous miscarriage.
A monogamous family was defined as one with one husband and one wife, while a polygamous family was defined as one in which the husband had more than one wife [49]. An internally-displaced person (IDP) was defined as a Nigerian who had been displaced from her town of residence, but not outside the country, as a result of armed conflict situation [50]. The Nigerian minimum wage of N18,000 was used to categorize monthly income. A live birth was defined as one in which there was any sign of life after expulsion from the mother, irrespective of the gestational age [51], while a stillbirth was defined as a baby born after 28 weeks of gestation with no signs of life [52]. A baby was defined as low birth weight, if its birth weight was below 2.5 kg [53].
Instruments and Data Collection Methods
The instruments used for data collection in this study were the questionnaire and study hand card. Due to low literacy rates, the questionnaires were administered as face-to-face interviews in Hausa language, by trained enumerators. The questionnaire had four sections: socio-demographic characteristics; obstetric and gynaecological history; preventive measures; malaria infection and haematocrit. The section on preventive measure asked about how frequently they slept under an ITN, and whether or not they had taken IPTp, and the number of doses. Frequency of ITN use was categorized as: Never, Seldom (1-2 times a week), Sometimes (3-4 times a week), Often (5-6 times a week) and Almost always. These categories were scored: 1, 2, 3, 4 and 5 respectively. History of malaria was based on self-reporting, and a client was said to have had malaria if she had been told so in a health centre, after a blood test. Their haematocrit measurements were obtained from their ante-natal cards as it is routinely done in the ante-natal clinic.
The study hand cards were designed by the researchers to collect information about post-pregnancy events. They were given the cards immediately after recruitment, and were to be submitted to the researcher at the end of the pregnancy. It contained the participants’ serial numbers, basic bio-data, and a short appeal to the attending physician or nurse, informing them that the patient had been enrolled into a follow-up study, and his/her help was required to kindly complete the blanks provided on the card. The items it contained were: outcome of pregnancy (live birth, stillbirth or miscarriage), place of delivery (hospital/health centre or home) and baby’s birth weight. For those who gave birth at home, only information on pregnancy outcome and place of delivery were collected when they came to submit the study cards.
Statistical analyses
Data collected from the questionnaires and study hand cards (Additional file 1) were entered into the IBM Statistical Package for Social Sciences (SPSS) version 22, followed by data cleaning. The data set for this study are available as supplementary material. Histogram and absolute values of skewness and kurtosis of not more than 2 and 7 respectively, were used to check for normality of the haematocrit and birth weights [54]. The outcome showed that there was no substantial non-normality in these continuous variables, and as such the data could be treated as normally distributed. Mean and standard deviation were used to summarize continuous variables, while the categorical variables were summarized as frequency and percentage. Pearson’s Chi-squared test was used to compare the baseline and follow-up categorical variables of the two groups. Fisher’s Exact test was performed for variables where any cell had less than five expected observations. Independent t-test was performed on the haematocrit and birth weights to determine the between-group differences in these variables.
The intention-to-treat (ITT) analysis was performed after replacing missing values [55], for which the multiple imputations method was used. This method was chosen since some correlations were expected between the missing observations, such that non-missing values on one outcome can inform the imputation of missing values on another outcome [56]. An exploration of the missing data showed that 11.09% and 13.96% of the data of the intervention and control groups were missing respectively. The Little’s Missing Completely at Random (MCAR) test was not significant (χ2=140, df=137, p=0.400) for both groups, making the multiple imputations method suitable [57]. The baseline variables were used only for prediction, while the outcome variables were imputed and used for prediction as well. The output of the multiple imputations analysis generated five different data sets which is the default in SPSS, each with missing values replaced. These were then pooled into a single data set for subsequent analyses.
The Generalized Linear Mixed Models (GLMM) analysis was performed to determine the overall effectiveness of the intervention [58]. Twelve potential confounding factors (ethnicity; family type; type of residence, educational status, occupational status, income level; age at first marriage; gravidity; parity level; period of amenorrhoea; history of previous preterm delivery, and history of previous miscarriage) plus the baseline data for the outcome variable in question were included in the GLMM analysis. The combination of these variables gave the best model, with the lowest Akaike Corrected Information Criterion (ACIC) and Bayesian Information Criterion (BIC). The small intervention groups were entered as the third cluster level, after the identifier (serial number) and intervention group. For sensitivity analysis, the GLMM analysis was repeated without the replacements (per-protocol), to determine the effect of drop out and non-response on the findings of the study. For all analyses, Confidence Interval (CI) of 95% was used, and level of significance was set at 0.05.