Electronic Partograph: A Way of Improving Partograph Use During Labour Monitoring Process in Selected District Hospitals in Bangladesh

Study design, site and duration: A prospective, follow-up crossover study design was used. Two secondary level referral hospitals in Jessore and Kushtia districts were selected as study sites. The selected facilities were similar in terms of infrastructure, human resources, and service delivery scenarios to typical district hospitals and were thus representative of public hospitals in Bangladesh. However, there were a few differences in the number of c-section and stillbirth rates in between two study hospitals. The crossover study design adopted under this study was designed to minimize the impact of this non-similarity on the study outcomes. The same facility was served as its own control at different time point in time. The study was conducted in two phases with an intervening pause period. During the first phase, an electronic partograph was used in Kushtia district hospital and a paper based partograph was used in Jessore district hospital. Both of these hospitals were being crossed to alternate partograph tool after the pause period. The nurse-midwives from each hospital posted in obstetric wards were the user who observed the calculated deliveries (N=506 for each facility) during the study period. Each participant attended a two-day long training to become acquainted with the respective partograph that they would be utilizing in the first phase of the study. Another training session was also arranged during the pause period. Clinical parameters recorded in the partograph were validated by the project appointed research assistants with midwifery qualifications who observed labor progression and deliveries. Sample selection: Sample size calculations were completed on the basis of existing prevalence of partograph use (17.1 %), and delivery outcome data related to paragraph use such as incidence of birth asphyxia (22%) and prolonged labour (7%) . Sample size calculations also took into account the refusal rate (10%) and design effects. Thus to observe the 50% improvement for all three mentioned clinical parameters (increased use of partograph, decreased level of birth asphyxia, and prolonged labour), the number of deliveries that needed to be observed were as follows, 105, 216 and 506 respectively at 80% power and 5% significance level. Thus the final sample size in each district hospital was taken as 506; half of the sample (253) before and the other half (253) after applying the crossover design. Field data collection continued for 12 months and the data collectors tried to capture all deliveries until the required minimum sample size was achieved. Inclusion criteria Inclusion criteria for the study were women with spontaneous labour in the first stage of labour with cervical dilatation in between 4-7 centimeters, singleton pregnancy, gestation of at least 37 completed weeks, cephalic presentations, and no additional complications. Exclusion criteria Exclusion criteria for the study were women with ante partum hemorrhage, breech presentation, multiple pregnancies, premature labour (before 37 weeks), eclampsia, elective caesarean section, and induced labour. Study procedure and data collection: An expert obstetrician, who was also a study investigator, provided training to the nurses and midwives. The basics for the paper and electronic partographs were essentially the same, so the training methods were not vastly different. For the e-partograph, a junior programmer was available by mobile phone, around-the-clock, to provide any instant trouble shooting or feedback. In addition, the Android Tablet was used for implementing the digital partograph study. Participants were also introduced to varying inconsistencies and errors they might face when utilizing the e-partograph. The research assistants became expert users of the e-partograph and facilitated the usage for the weak users. e-partograph: The electronic version of the partograph was a state-of-the-art application that is accessed through smart phone or tablet pc or computer device. The application's user interface (UI) is segmented; users will have to concentrate only on a single portion at a time that would lessen the existing complexity of using paper-based partograph. e-partograph application's user interface in Android programming language for smart tabs, and in ASP.net with C# language for personal computers. The application has options to save the data in local storage and in a remote central database storage concurrently. Local storage contains data for temporarility; the remote server contains the data permanently which makes the partograph information searchable at any time and place. This application allows partograph data to be monitored remotely. The existing knowledge on how to use the paper and e-partograph was assessed for each participant before the start of the training to understand the skill status of the individuals. The training consisted of two days of intensive hands on training as well as a refresher course that was organized six months after the first training to ensure that participants would not forget the information. In order to measure user rates, rates of birth asphyxia, and prolonged labour, record reviews generated by the partographs were collected and analyzed. Data Analysis: Quantitative data was analyzed using SPSS 23 statistical software. The outcome variables were use of the partograph and delivery outcomes (prolonged labour and birth asphyxia). The independent variables covered maternal demographic and obstetric variables along with fetal characteristics as well as partograph type and facility. Partograph was considered done when it was correctly used for labour monitoring among the delivered women who fall under inclusion criteria. Prolonged labour was defined as labour extending more than 12 hours and birth asphyxia was defined as an APGAR score less than 7 on the 5th minute after delivery. Discrete variables were expressed as percentages and presented as frequency tables and cross tabulations. Chi square (χ2) tests were employed to test the association between proportions of respondents and type of partograph. Statistical significance was defined as p-values of \<0.05. The data collected between the two groups was analyzed in such a way that a clear comparison could be made between the user rate of different periods (before and after applying the crossover design). Potential confounding factors were adjusted using binary logistic regression. Ethical Assurance: The Research Review Committee (RRC) of the International Centre for Diarrheal Disease Research, Bangladesh (icddr, b) approved the technical part of the proposed study. Then the Ethical Review Committee (ERC) of icddr, b who oversees the protection of human rights, approved the study. In addition, informed consent was obtained from the hospital administration, health workers, and the laboring mothers.