Skip to main content

Evidence > Predictive Model for Postpartrum Haemorrhage and Sepsis


There is an urgent need for low-cost, easy-to-use tools that can accurately predict adverse outcomes associated with maternal morbidities. A prediction model could assist in the management of adverse outcomes by allowing early identification of high-risk women based on early manifestations of diseases (signs and symptoms). This can guide the decision-making for transfer and care in health facilities. Low- and middle-income countries with high burden of maternal mortality and morbidities may especially benefit from these simple technologies.


Through funding from Grand Challenges Canada, we have initiated a predictive modeling project focussed on hypotension. A prospective cohort study will be conducted at Jinnah Post Graduate Medical Center (JPMC), Karachi, Pakistan. We are recruiting women who have admitted for delivery or are postpartum within forty-two days of delivery and aim to reach a sample size of one hundred adverse maternal outcomes. With this project, we are ultimately aiming to incorporate a hypotension risk model into the PIERS platform. Once data collection is completed a prediction model will be developed by Samina Bakhtawar, Masters Student at Aga Khan University, Karachi, Pakistan.

Current status

We successfully obtained ethics at the two participating sites in Pakistan: Jinnah Post Graduate Medical Centre (JPMC) and Aga Khan University (AKU), Karachi as well as at UBC. We have finalised the data collection tools and processes, and also created an electronic data capture application. Doctors and nurses at JPMC have been hired and trained for data collection, which began on 25 July 2016. See below for a live recruitment graph of the study.

Total Enrolment:
*Data updated on (2017-05-22)