Evaluating the impact of e-registration and mHealth on institutional delivery in hazard-prone areas of Bangladesh: A protocol for a non-randomized controlled cluster trial

BackgroundDespite substantial progress, Bangladesh still has a high rate of maternal deaths owing to difficulties during pregnancy, delivery, and the postpartum period. Increasing facility delivery is mandatory to reach the goal of bringing down the MMR to <70 deaths/100,000 live births by 2030. In the era of digitalization, the introduction of e-registration and mHealth may aid the government in reaching this target. The southern part of Bangladesh is a hazard-prone area, where service uptake from institutions is low. This study aimed to determine the effect of an e-registration tracking system and mHealth counseling on institutional deliveries to pregnant mothers in hazard-prone areas of southern Bangladesh.

MethodsWe will conduct an open-label, two-arm, non-randomized controlled cluster trial for six months and use three hazard-prone areas for intervention and another three hazard-prone areas for control. We will collect data at baseline and end-line of the study period using a structured questionnaire. We will enroll at least 268 pregnant mothers from the intervention and 268 pregnant mothers from the control areas after screening based on the inclusion and exclusion criteria. Pregnancy information will be obtained from the Family Welfare Assistant register. We will follow the participants until their delivery and exclude those respondents from the study who will have post-dated delivery, migrate out, lost to follow-up, or die during the study period. Random-intercept mixed-effect logistic regression will be performed to explain the relationship of e-registration and mHealth package with institutional delivery.

DiscussionInstitutional delivery is still uncommon in Southern Bangladesh despite several interventions. Innovative approaches like e-registration and mHealth counseling may be helpful to bring women to health facilities.

Trial registrationThis experiment is registered in the open science framework (registration number: DOI 10.17605/OSF.IO/YZE5C) and https://www.clinicaltrials.gov/ (registration number: NCT05398978).


23
Despite substantial progress, Bangladesh still has a high rate of maternal deaths owing to 24 difficulties during pregnancy, delivery, and the postpartum period. Increasing facility 25 delivery is mandatory to reach the goal of bringing down the MMR to <70 deaths/100,000 26 live births by 2030. In the era of digitalization, the introduction of e-registration and 27 mHealth may aid the government in reaching this target. The southern part of Bangladesh 28 is a hazard-prone area, where service uptake from institutions is low. This study aimed to 29 determine the effect of an e-registration tracking system and mHealth counseling on 30 institutional deliveries to pregnant mothers in hazard-prone areas of southern Bangladesh. 31 Methods 32 We will conduct an open-label, two-arm, non-randomized controlled cluster trial for six 33 months and use three hazard-prone areas for intervention and another three hazard-prone 34 areas for control. We will collect data at baseline and end-line of the study period using a 35 structured questionnaire. We will enroll at least 268 pregnant mothers from the intervention 36 and 268 pregnant mothers from the control areas after screening based on the inclusion and 37 exclusion criteria. Pregnancy information will be obtained from the Family Welfare 38 Assistant register. We will follow the participants until their delivery and exclude those 39 respondents from the study who will have post-dated delivery, migrate out, lost to follow-40 up, or die during the study period. Random-intercept mixed-effect logistic regression will 41 be performed to explain the relationship of e-registration and mHealth package with 42 institutional delivery.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 30, 2022. However, the effect of e-registration tracking and mHealth counseling in both voice calls 98 and text messages as a package on institutional delivery in Bangladesh country context is 99 yet unknown. Although the study conducted in Ethiopia showed positive results in 100 increasing the facility delivery using both text messages and voice calls, the cultural, social, 101 and geographical variations present between the Ethiopian and the Bangladeshi population 102 beg the importance of a separate study to be conducted in the Bangladesh country context. 103 Additionally, the Southern part of Bangladesh is a hazard-prone area, and the service uptake 104 tendency from the institution is much lower here. Thus, achieving the targeted institutional 105 delivery rate is a challenge. 106 In Bangladesh, paper-based registers are used in the health system to keep a record of Family Welfare Assistant (FWA) registers to keep track of pregnant mothers at the 110 community level. Each FWA is assigned to one or more than one adjacent union that visits 111 houses to conduct population registration, eligible couple registration, and pregnant mother . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 30, 2022. ; https://doi.org/10.1101/2022.06.30.22277099 doi: medRxiv preprint 6 112 registration quarterly. The data are stored in the paper-based registers, and union-level data 113 are compiled monthly to produce sub-district level reports. The field activity and data Strengthening pregnancy registration is one of the key mandates of this project, and 128 assisting DGFP in the roll-out of the eMIS registration system in Chandpur district was 129 done to achieve the mandate. The project has employed midwives in the Chandpur District 130 Hospital, Kachua Upazila Health Complex, and Faridganj Upazila Health Complex who 131 will send text messages and make voice calls to counsel the registered pregnant mothers 132 for institutional delivery. Therefore, we plan to conduct a two-arm, non-randomized 133 controlled cluster trial to evaluate the joint effort of the DGFP eMIS tracking system and 134 mHealth counseling done by the midwives in overcoming the challenge and improving 135 institutional deliveries in the hazard-prone areas of Southern Bangladesh. We expect that 136 e-Registration and mHealth intervention will provide policy-relevant lessons on scaling up . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  Pregnant mothers registered in the FWA register, . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 30, 2022. decided on this proportion we need to determine the number of respondents for our study . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 30, 2022.  Figure 3 provides the conceptual framework of data collection process. We will use a 242 structured questionnaire developed in Bangla for data collection at the baseline and end 243 line of the study period. Two local recruits at each of the intervention and control areas will 244 be employed for data collection to avoid any issues with the local dialect. Among the two . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 30, 2022. ;https://doi.org/10.1101https://doi.org/10. /2022 12 245 data collectors, one will be a female who will collect data from the pregnant mothers 246 directly for the cultural issue and the nature of the questions asked, and another will be a 247 male who will support the female data collector during community visits. We will train the 248 data collectors on the questionnaire, data collection method, ethics, and infection The questionnaire is subdivided into three main sections -screening measure, baseline 254 survey, and end-line survey. During the baseline period, we will screen pregnant mothers 255 using the screening tools. The screening tools will include mothers who will give consent 256 and exclude pregnant mothers with severe pregnancy-related complications, such as 257 eclampsia, antepartum hemorrhage, sepsis, and/or COVID-19 infection. The baseline 258 section is designed to collect data of the respondents about the socio-demographic 259 information, pregnancy-related information, such as parity, gravida, months of pregnancy 260 at the first antenatal visit, counseling on facility visit, birth preparedness plan, knowledge 261 on danger signs, location, and outcome of previous pregnancy and existing comorbidities, 262 and other information such as personal belief, social norm, choice of delivery place, health-263 seeking behavior in general, and the distance between the respondent's home, and the 264 nearest healthcare facility. It will take approximately 20 minutes to complete the baseline 265 section of the data collection tool. The pregnant mothers will then be followed-up until 266 their delivery. Then their delivery-related data, such as place of delivery, mode of delivery, 267 and the delivery outcome, will be collected at the end line of the study period using the 268 relevant section of the questionnaire, which will take about 10 minutes to fill up.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 30, 2022. ; https://doi.org/10. 1101/2022 Since the baseline, the end line data collection will be done at the end of the fourth month 270 of the study period. The baseline and end-line data collections will be done by face-to-face 271 interviews. The questions will be read out to the participants and the data collectors will fill 272 up the questionnaire with the information given by the mothers. The mothers who will have 273 post-dated delivery, develop severe pregnancy-related complications, will be lost to follow-274 up, or die during the study period will be excluded from the study. We will exclude post- online. The FPIs perform data quality checks, approval, and monthly aggregation of union-286 level data to the sub-district levels. The aggregated list will be sent to the midwives. Then 287 the midwives will call and send text messages to the pregnant mothers whose expected 288 delivery dates are nearby and counsel them to conduct their deliveries at the healthcare 289 facility. The midwives will call three times in case of a failure of the first two attempts to 290 reach each participant of the intervention arm. The text messages will be tailored 291 beforehand in Bangla and the midwives will be trained on how to counsel the pregnant 292 mothers on facility delivery. Following e-Registration, the mHealth intervention will be . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 30, 2022. Data management and analysis plan 318 We will analyze the data in R 4.1.2. We will do a frequency distribution and proportion 319 analysis for categorical variables, and for continuous variables, we will perform a mean 320 and standard deviation analysis. Finally, a random-intercept mixed-effect logistic 321 regression model will be fitted to explain the relationship of e-Registration and mHealth 322 package with institutional delivery, where the union will be included as the random 323 intercept. An adjusted odds ratio with p-value less than 0.05 will be considered statistically 324 significant for the variable. Quality control and quality assurance 327 The data monitoring committee will regularly check data and monitor to maintain the data 328 quality. An expert will double-check the validation of data collected by the data collectors. 329 The data entry team will link records that have been collected with the help of individual 330 identifiers assigned. We will remove this identifying information once the records are done 331 matching. The anonymized data will be uploaded to a central web-based platform regularly, 332 after following all the appropriate confidentiality requirements. We will use an encrypted, 333 password-protected centralized database system to store data and only authorized members 334 of the research team will be allowed access.

335
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2022. ;https://doi.org/10.1101https://doi.org/10. /2022 Patient and public involvement statement 337 There were no funds or time allocated for the patient and public involvement, so we will 338 not be able to involve patients. We will ask pregnant women to assist us in evaluating an 339 effective digital health strategy to reduce adverse events. This two-arm non-randomized controlled cluster trial of six months will determine the 343 effect of e-registration tracking and mHealth counseling as a complete package on 344 institutional deliveries in the hazard-prone areas of Southern Bangladesh. A minimum of 345 268 pregnant mothers of 28-36 weeks will be enrolled and followed up until their deliveries 346 in each intervention and control area. Baseline and end-line data will be used to synthesize 347 the study result. Since the end-line data will be collected four months later than baseline 348 data collection, the study might have a recall bias. 349 However, to our best knowledge, this two-arm non-randomized controlled cluster trial will 350 be the first study that will evaluate the effect of e-Registration and mHealth as a complete 351 package on institutional deliveries in the hazard-prone areas of Southern Bangladesh. e-352 registration and mHealth intervention package may increase facility delivery, thus, reduce 353 delivery-related complications and maternal mortality. Thus, our study will facilitate the 354 evidence gap related to the effectiveness of e-registration and mHealth on facility delivery 355 and provide substantial evidence which may help the government to take the necessary 356 steps to roll out e-registration and mHealth counseling as a complete package to increase 357 institutional deliveries and reduce the burden of maternal mortality ratio of Bangladesh by 358 2030.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2022. ;https://doi.org/10.1101https://doi.org/10. /2022