Assessing the cost-effectiveness of interventions for Hypertensive Disorder of Pregnancy and Diabetes Mellitus in Pregnancy in Bangladesh

Progress in reducing maternal mortality have stalled during the Sustainable Development Goals era both globally and in Bangladesh. Maternal mortality is often called the tip of the iceberg. The morbidities leading to maternal deaths often remain ignored. Long-term impact of these morbidities are rarely discussed. While hypertensive disorder is one of the leading causes of maternal deaths, there is a growing burden of diabetes in pregnancy. Economic evaluation models addressing either of the two Non-Communicable Diseases are scarce, especially in the context of low and middle income countries. Preventive interventions are highly recommended for addressing the burden of NCDs and are expected to be easier to implement given the capacity problems evidence in the Bangladesh health system. The aim of this thesis was to develop a cost-effectiveness model for interventions addressing hypertensive disorder and diabetes mellitus in pregnancy among Bangladeshi women. This was done in several steps including multiple reviews, stakeholder consultations and finally developing the model.
First, a review of World Health Organisation and Bangladesh clinical guideline review was undertaken to understand the two diseases and their recommended treatment pathways. The clinical guideline review led to the identification of the risk factors of the two conditions, their symptoms or diagnostic criteria and the treatment to be undertaken once they are detected. This review also identified the individual level risk factors of the two diseases and revealed that diabetes mellitus, both pre-existing and gestational are risk factors for hypertensive disorder in pregnancy. The review recommendations suggested antenatal care to be the key stage of the pregnancy continuum of care to address the two disease conditions. It also indicated antenatal period to be the most crucial stage for delivering preventive interventions.
Next, a systematic review of economic evaluation models assessing interventions related to hypertensive disorder and diabetes mellitus in pregnancy identified those methods that had been used previously. While decision trees represented the dominant model structures, markov state transition and microsimulation models had also been used. Interventions could be divided into several types; screening and diagnosis, treatment, diet and lifestyle, labour induction and others. Most models were built from a health system’s perspective while some were focused on the payer’s or societal perspective. Model time horizons ranged from pre-conception to the lifetime of the women and their offspring. A varied set of model outcomes were identified including immediate outcomes like development of hypertensive disorder to mortality and development of chronic conditions in the long-term. This review highlighted the lack of a consistent approach to modelling of these two disorders and the lack of a model that was suitable for decision making in Bangladesh. Both reviews also revealed the need for preventive care measures, specifically during the antenatal period.
The reviews led to the development of a detailed list of interventions relevant for Bangladesh and a draft conceptual framework. These were presented to stakeholders and amended via interviews. Both the reviews and interviews with stakeholders identified the antenatal care related interventions as the key to address the two diseases conditions. The antenatal care package was deemed to be the most important intervention in addressing the two conditions which contains multiple preventive interventions. Rather than modelling the whole package of intervention, a single preventive intervention was selected as an exemplar. The final intervention selected for the model was calcium supplementation among pregnant women in Bangladesh scaled up from 18% to reach 80% coverage level. Calcium supplementation is a recommended preventive intervention that is supposed to be delivered as part of the antenatal care package through the Bangladeshi health system. Methods for incorporating multiple intervention effect were explored and documented. The model programming ensured flexibility in incorporating additional interventions for future work.
The model took health system’s perspective as suggested by the stakeholders. Model outcomes covered the development of hypertensive disorders and gestational diabetes mellitus in pregnancy, c-section, preterm births, stillbirths, newborn and maternal deaths. Long-term outcomes needed to account for development of chronic conditions among women for a lifetime and developmental delay among children for 5 years. The model took into account two pre-existing chronic conditions as risk factors; pre-existing chronic hypertension and diabetes mellitus.
An individual based markov microsimulation model was developed. Increased risks of events and outcomes were assigned from existing literature through ad hoc reviews. Costs of intervention and all downstream care-seeking were estimated based on Bangladesh national data. Disability Adjusted Life Years (DALYs) were estimated based on disability weights from Global Burden of Disease studies and country-specific life tables for women. DALYs were estimated both for pregnancy and long-term health conditions. Finally, a cost effectiveness analysis was undertaken and a probabilistic sensitivity analysis (PSA) was incorporated accounting for parameter uncertainty. The model was validated for the base case scenario against national prevalence levels and verified using a prescribed technical verification checklist. The model produced summary outputs in the form of Net Monetary Benefit, the Incremental Cost Effectiveness Ratio, the cost-effectiveness plane, cost-effectiveness acceptability curve. Selected scenario and sub-group analysis and deterministic sensitivity analyses were also conducted. The willingness to pay threshold was set at 1x and 3x Gross Domestic Product (GDP) per capita values based on stakeholder requirements.
The intervention reduced adverse pregnancy and birth outcomes for mothers and their babies. The direct impact of the intervention led to a reduction in the number of women developing gestational hypertension, pre-eclampsia and eclampsia (23%, 40% and 36% reduction respectively). Caesarean-sections reduced by 7% through an indirect effect of the intervention. Preterm births and stillbirths reduced by 7% each and maternal deaths saw a 45% reduction through both direct and indirect impact. Newborn deaths were reduced by 8% through reduction in preterm births. Among the long-term outcomes, there was an 8% reduction in chronic hypertension while diabetes mellitus remained almost unchanged. The incremental cost per woman was estimated at BDT -5122 which indicates a cost saving for the scaled-up provision of care. The probability of the scaled-up provision to be cost-effective was 1 at both 1xGDP and 3xGDP per capita threshold level. The incremental net monetary benefit was positive at both threshold level and for all sub-group level analyses, which covered pre-existing conditions, age, education and wealth quintiles.
The model is novel and added to the existing evidence base in several ways. It is the first model that took into account two cardio-metabolic diseases; hypertensive disorder and diabetes mellitus in pregnancy together. Pre-existing conditions and risks related to them were incorporated as risk factors in the model while interaction between the two pregnancy related conditions were also considered. The model was strengthened in terms of its validity and relevance by use of a large number of country specific data and involving stakeholders at an early stage. Findings from this thesis can add value to the existing evidence base of economic evaluation of calcium in the context of low and middle income countries. The thesis reemphasised the need for ensuring access to calcium supplementation during pregnancy in Bangladesh to prevent hypertensive disorders in pregnancy. It also highlighted that this simple intervention can release resources, which can then be redirected to other areas of need. In order to ensure successful implementation of this intervention, there is value in exploring the most effective channel for delivering the intervention. Further research should also consider ways to improve compliance in calcium intake, alternative sources of calcium like food fortification during pregnancy and pre-conception and explore the optimum dose of calcium for Bangladeshi women.