Integrated studies addressing the incidence of severe maternal morbidity (SMM) in Kilombero district, Tanzania
Mgonja, Miriam Lucy
Severe maternal morbidities (SMM ) are a diverse range o f serious complications which occur during pregnancy, childbirth and the post-natal period (six weeks after childbirth). They can affect any of the body organ systems, and their severity means that if untreated, or wrongly treated, maternal death is likely to result. Therefore prevention of maternal death requires better understanding of the epidemiology and outcome of the various forms of SMM. Tanzania is among developing countries in the world having a high maternal mortality ratio (454 per 100,000 live births), but there are few published evidence addressing all major forms of SMM in Tanzania. Not all pregnant Tanzanian women use health facilities for antenatal care or delivery, not even for treatment in the event of a complication. Therefore ascertaining the level of SMM in health facilities will not give a true estimation of the incidence among all pregnant women. While ascertaining SM M in the community would therefore seem to be a preferable approach, there are considerable practical barriers to executing such research in rural areas of developing countries. Furthermore previous studies have raised concerns about the reliability and validity of women’s self report of obstetric complications, especially when such report is retrospective.This PhD research has therefore taken two complementary approaches: (1) At St Francis Referral Hospital, Kilombero, Tanzania, using hospital routine data with case note review of all admissions experiencing complications, it has ascertained over the course of a full calendar year, the incidence and outcome of SMM , and associated risk factors (N= 5500 total admissions, n=1723 with case note review); (2) By undertaking a community interview survey in that hospital’ s rural catchment area, to determine the incidence of SMM via women’s self reporting of symptoms. This survey attempted to interview all women in selected villages that are subject to the Ifakara Centre Demographic Surveillance System (IC-DSS), and who were recorded by IC-DSS as having given birth over the same time-frame (690 were interviewed, while 663 had given birth during study period). Validation of self-reporting of SMM was undertaken using interview response data of the subset of 173 interviewees who had attended St Francis hospital for obstetric delivery/treatment during the study interval, and for whom hospital diagnosis made at that time were therefore also available.In the hospital study, the incidence of both intra partum (IP) and postpartum (PP) SMM (among all women delivering there or admitted postpartum) was 19.2%, with a case fatality rate of 1.7%. The incidence of solely intra partum SM M was 17.2%, the rate being higher among primigrávida than among multigravid (28% v 8%, p<0.001), among teenagers than older women (23% v 14%, p<0.001) and single women than among other women (68.1% v 48.7% , p<0.001). Among 909 intra partum SMM , the three commonest conditions were obstructed labour (69% ), eclampsia (14% ) and severe preeclampsia (9% ) while in the 258 postpartum SM M ; the three commonest conditions were postpartum eclampsia (26% ), puerperal sepsis (24% ) and postpartum haemorrhage (23% ). Neonatal death was more com m on among women with intra partum SM M than among other deliveries (4.7% v 1.1%, p<0.001). There were no intra partum maternal deaths, but across all antenatal, intra partum and postpartum admissions, maternal mortality was four-fold greater among women admitted in critical condition than among other women (2.1% v0.5%, p=0.006). The majority of the women who became maternal deaths were admitted in critical condition (11/17).Among Ifakara IC-DSS women interviewed, the majority had used health facilities at some level for obstetric care (64.7% ), and affirmed prior birth and emergency complication readiness plans (97.6% & 81.5% ). The incidence o f any SMM (antenatal to postnatal, derived from self reported symptoms by interviewees) was 16.7%; while for both intra partum and postpartum SMM , the incidence was 15.4%, lower than that of the hospital study (19.2% ). Considering solely intra partum (interview-ascertained) SMM, the rate was 8.1% (n=54), lower than the corresponding estimate of 17.2% from the hospital study. Validation interview-ascertained intra partum SMM classification could be undertaken in the subset of 173 who had delivered in St Francis hospital. Of these, 48 could be matched with the list of names of hospital admissions subjected to case note-review (i.e. all deliveries with any evidence of complications), while the remainder had no hospital-ascertained complication even (and hence must be presumed to be ‘normal deliveries’). In this group, interview-ascertained IP SM M rate was 21.4% , and the sensitivity and specificity for interview-ascertained IP SM M were 59.3% (16/27) and 85.6% (125/146) respectively.This is the first study in Tanzania to address the incidence of all intra partum and postpartum severe morbidities, using both retrospective review of referral hospital data/records and an interview survey in the surrounding rural community. The research findings have described the relative frequency of various subtypes of SMM , and the relationship of SMM to sociodemographic and obstetric factors and to maternal-foetal outcome. Such information will be helpful for clinicians and maternity staff to understand the pattern of SMM and how the hospital is performing in treating SMM cases. The finding that maternal death is so strongly associated with admission in critical condition emphasises the importance of addressing factors that delay admission of pregnancies that develop complications.In the thesis the findings from the research studies will be elaborated and the challenges of such research will be discussed. While the sample size for validation of intra partum SMM was too small to be able to make conclusive remarks, the finding of borderline sensitivity will be discussed in relation to specific morbidities and the reliability o f the hospital data ‘ gold standard’ comparator.With a pragmatic approach one can undertake research on SMM that enhances understanding of the complexities of SMM and its measurement, while also imparting knowledge on the epidemiology of the condition and potential actions that could be considered to improve outcome for SMM.