A framework to lower maternal mortality and morbidity rates in Kenya using mobile technology

Background. Maternal health represents an area of significant concern in the world. With various innovations in healthcare, maternal mortality rates are decreasing exponentially. However, this is not the situation in developing countries, specifically Kenya. Several causes of maternal mortality exist; however, it is noted that one of the key causes is due to a lack of information by pregnant women. Traditional strategies such as free maternal health care at public hospitals have been in place to improve overall pregnancy outcomes. While this is aimed at a reduction in maternal mortalities, it has not been as effective in Kenya. Non-conventional strategies are needed to improve maternal health outcomes and reduce maternal mortality. Information Communication Technology (ICT) is one of the areas that has been proven successful in reducing maternal mortality. Aim. The aim of the research was to create an ICT framework that aided in educating pregnant women using an mHealth dissemination tool and thus reduce complications that led to mortalities within Kajiado North Constituency. Methods. The study utilized a sequential mixed-method design. Phase one consisted of a retrospective chart review and a cross-sectional survey on nurses and pregnant women. The first phase focused on understanding the maternal mortality rates within Kajiado North and to determine procedures pregnant women and nurses took during pregnancy. The retrospective chart review was conducted for a period of six months at two health facilities, the Ongata Health Centre (OHC) and Ngong Sub District hospital (NSD). The cross-sectional survey interrogated the mitigation strategies with a focus on information and communication technologies (ICT). Phase two was a prospective multi-location randomized clinical trial (RCT). A two-arm, two-site RCT was conducted using an intervention in the form of an ICT prototype with messages aimed at educating the pregnant participants. The trial was conducted at two public health facilities namely the Ongata Health Center and the Ngong Sub District. A total of 211 pregnant women were recruited from both locations after they had met the inclusion and exclusion criteria and after providing consent to participate in the study. During the RCT, an intervention was developed using the Design Science Research Methodology (DSRM) and was used to send messages to participants within the intervention arm. The DSRM approach allowed for two iterations to be created, with one iteration being tested during the pilot test and the other during the RCT. Pregnant participants within the intervention groups received messages on their mobile phones about well-being during pregnancy. Women in the control group continued to receive their established standard of care. Both groups completed a post-test survey at the end of the trial. Data were analysed using ANOVA with the probability set at p≤0.05%. The relationship between the number of antenatal visits and the place of delivery on the complication rate was shown using the correlation coefficient. Additionally, a multiple regression model was generated based on the antenatal visits, place of delivery and the study arms and their impact on the complications. Results. Data from phase one of the study showed a need for a messaging system to send messages to pregnant women. The retrospective data showed no maternal mortalities, however, the nurse survey highlighted possible explanations for the lack of mortalities. From the RCT, there were no known maternal mortalities. There were three neonatal mortalities (p=0.154), one from the OHC intervention group and two from the OHC control group. The ANC visits relationship towards the complication rate was calculated. At the NSD site, the effect size of the ANC visits based on the participants' study arm toward the complication rate was low (0.027) and statistically insignificant (p=0.15). At the OHC site, the effect size was moderate (0.405) and statistically significant (p=0.003) for the ANC visits variable. The place of delivery relationship towards the complication rate was calculated. At the NSD site, the effect size of the place of delivery based on the participants' study arm toward the complication rate was moderate (0.366) but statistically insignificant (p=0.479). At the OHC site, the effect size of the variables was low (0.237) and statistically insignificant (p=0.789). The stepwise regression model at the OHC site showed significance when ANC visits (p=0.007), place of delivery (p=0.003) and participants study arm (p=0.008) were sequentially entered. The multiple variables (R=0.516) Only had a medium effect size (0.266) toward the complication rate. The stepwise regression model at the NSD site was statistically insignificant when the place of delivery (p=0.283), participants study arm (p=0.445) and ANC visits (p=0.655) were sequentially entered. The multiple variables (R=0.217) had a small effect size (0.047) toward the complication rate. Conclusion: Qualitative findings revealed that maternal health was affected adversely by several lengthy health worker strikes. Negligence on part of the health worker was a lead contributor to neonatal deaths. The study also found that accountability systems for referrals were lacking within the county and measures needed to be put in place to mitigate the consequences. In addition, feedback from the study participants indicated that the messages had aided in helping them to take necessary action based on complications and warning symptoms. Based on the data, the study finally proposed a framework that would allow for a reduction of maternal and neonatal mortality rates using ICT technologies. The study equally contributed to knowledge when using technology to empower women on maternal health matters as well as areas of maternal health practice that need improvement