Maternal and perinatal Health Research Collaboration, India (MaatHRI): methodology for establishing a hospital-based research platform in a low and middle income country setting

Establishing a network of hospitals and clinical collaborators

Successful completion of the pilot work in two hospitals in Assam allowed us to expand the network from two to nine government hospitals within Assam: six teaching hospitals and four district hospitals. In each hospital, we identified a lead collaborator who were obstetricians. Through their professional networks, we were able to reach out to other hospitals. A hospital was included in the network based on two criteria: (i) willingness of the hospital to participate in a large research collaboration and (ii) a high burden of maternal and perinatal deaths in the population covered by the hospital. Similar to the process used in the pilot work⁶, we mapped the hospitals to assess the spread and coverage of the population in each state.

Setting up a high-quality secure system for data collection, storage and transfer

One of the major reasons for success of the pilot work was having dedicated research staff for data collection and data entry. A pragmatic approach was adopted to develop a high quality secure electronic system to overcome the challenges of human resource constraints, lack of dedicated secure computer servers for data storage in the hospitals, and securely sharing data. The following methods were employed:

Standardised laboratory infrastructure

A laboratory infrastructure was created through a partnership with a private laboratory in India, Dr Lal PathLabs (LPL). LPL has a pan-India presence with a network of sample collection centres, regional laboratories and a national reference laboratory in New Delhi, India. Their existing service delivery model was adapted to the requirements of the MaatHRI platform through extensive consultations between the Indian clinical collaborators, and experts at the University of Oxford and LPL. The following services were agreed and are being currently used to standardise the laboratory infrastructure:

We tested the model in a run-in phase before full implementation.

Regulatory systems

The steering committee constituted for the pilot work (IndOSS-Assam) was expanded to form the MaatHRI steering committee. The committee includes representatives from all the collaborating hospitals, the University of Oxford, Indian policy advocates and experts in statistics and ethics. As MaatHRI is a research platform set up to conduct studies on a long-term basis, an independent ‘Data safety and monitoring board’ (DSMB) was set up, including members from India and the UK who are not associated with the MaatHRI platform. A DSMB charter was drafted outlining the roles and responsibilities of the members and how the board will function to provide independent safety review of participants and data, and guidance for observational studies during the course of the ongoing projects. Since the studies currently undertaken through the platform are observational studies, review of adverse event data and reports of serious adverse events (SAEs) are not currently applicable to MaatHRI DSMB. However, should randomised controlled trials be conducted through MaatHRI in the future we would expect the DSMB to be involved in reviewing this type of information.

Academic collaboration agreements were signed between the University of Oxford and all collaborating institutions in India. These were discussed extensively with each institution’s legal team to agree on the terms and conditions including data security and sharing, confidentiality, and intellectual property (IP) rights. To ensure equitable partnership, IP is jointly owned by the collaborating institutions in proportion to the respective contribution of each institution.

Ethics approvals

Ethics approval was obtained for setting up the platform and to undertake the observational studies from the institutional review boards (IRB) of each coordinating Indian institution, namely: Srimanta Sankaradeva University of Health Sciences, Guwahati, Assam (No.MC/190/2007/Pt-1/126); Nazareth hospital, Shillong, Meghalaya (Ref No. NH/CMO/IEC/COMMUNICATIONS/18-01); Emmanuel Hospital Association, New Delhi (Ref. Protocol No.167); Mahatma Gandhi Institute of Medical Sciences, Sevagram, Maharashtra (Ref No. MGIMS/IEC/OBGY/118/2017); and the Institute of Medical Sciences, Banaras Hindu University, Varanasi, Uttar Pradesh (No.Dean/2018/EC/290). The platform and the studies have also been approved by the Government of India’s Health Ministry’s Screening Committee, the Indian Council of Medical Research, New Delhi (ID number 2018-0152) and by the Oxford Tropical Research Ethics Committee (OxTREC), University of Oxford, UK (OxTREC Ref: 7-18).

MaatHRI network of hospitals

We were able to establish a network of 14 hospitals by September 2018 across four states in India – Assam, Meghalaya, Uttar Pradesh, and Maharashtra. After establishing the network, two more hospitals joined MaatHRI, but two government district hospitals left the collaboration. A lack of interest in research and high patient load were the main reasons given by the lead collaborators of the departing hospitals. The MaatHRI platform currently includes a network of 14 hospitals (11 Government and 3 private) who collaborate based on the shared vision of generating scientific evidence to reduce maternal and perinatal mortality and morbidity in India and other LMICs. Figure 1 shows the distribution of the network across India and within the state of Assam. A list of the Indian collaborating institutions are available on the MaatHRI website (https://www.npeu.ox.ac.uk/maathri). None of the site-collaborators or any clinician involved in the project charge their time contribution to the project. This is another testament of their passion towards the shared vision. The 14 hospitals together conduct about 100,000 deliveries per year. The network includes an Indian research team of 20 nurses, a project manager, 16 obstetricians, two pathologists, a public health specialist, a general physician and a paediatrician.

Figure 1. Distribution of the MaatHRI collaborating hospitals and regions covered by the hospitals across India.

Setting up a high-quality and secure system for data collection and storage

Data and biological sample collection: Depending on the patient load and related participant recruitment rates, one or two research nurses have been appointed in each collaborating hospital for the MaatHRI work. The nurses are responsible for recruiting study participants, providing participant information and obtaining informed consent, collecting data and blood samples, and following up participants. The research nurses were specifically trained to undertake these activities. In addition, a project manager has been appointed to manage the research nurses and provide supportive supervision.

Data entry and storage: Our original plan was for research nurses to collect data in online electronic forms using tablet PCs enabling automatic collation in the Microsoft Azure cloud computing platform (Microsoft Corporation) with servers located in India; there is no provision for storing data on the tablets. However, after an initial trial we found that direct data entry in an online form was not possible due to problems with internet connections in several hospitals and the sensitivity associated with a nurse standing with a tablet PC next to a very sick woman. It was therefore decided that paper forms would be used to collect data in these hospitals and the nurse would enter the data immediately afterwards into the online data portal and then destroy the paper forms. Each hospital has a unique Login ID and password to access the data collection forms and their collated data on the online portal.

Quality assurance and data security: The electronic data collection forms have checks and validations to flag logical errors. The project manager is responsible for monitoring data entry on a day-to-day basis. Red flags are raised for errors and incomplete forms immediately so that the research nurse can rectify the errors before the participant is discharged from the hospital. Data stored in the cloud server are encrypted and password protected. Each collaborating hospital can only view and download its own data. Identifiable information are collected for follow-up of participants, but these can only be viewed by the authorised hospital staff and cannot be downloaded by anyone. Once the data collection is complete, in preparation for analysis, all identifiable information is completely delinked from the clinical data to generate pseudonymised analysis files. We have developed secure mechanisms for transferring data within India and between India and the UK with recommended level of end-end-encryption.

Laboratory infrastructure

Dr Lal Pathlabs (LPL) provides the laboratory infrastructure for MaatHRI. The following services were tested in a trial run before being fully incorporated into the platform.

Service 1: Blood collection kits with instructions to all study hospitals. LPL provides the required blood collection kits with specific written guidance to all study hospitals for collecting, processing and packing the blood samples.

Service 2: Train MaatHRI research nurses to collect and prepare blood samples. Technical experts from the laboratory trained the MaatHRI staff (project manager and research nurses) to collect, centrifuge and pack samples before the start of studies. When required, a phlebotomist from their collection centre provided supportive supervision to the research nurses during the initial few weeks to correct or prevent any errors.

The MaatHRI research nurses collect, centrifuge and pack blood samples as per instructions in transportation boxes ready for collection by LPL. A standard test requisition form for each participant is filled in by the obstetrician caring for the participant. This form only includes the participant ID, age and a barcode to ensure participant confidentiality and blinding to minimise reporting bias. The test results are only used for research purposes and not for the provision of clinical care.

Service 3: Transport samples at ambient conditions from the hospitals to the laboratory. A designated person from the LPL collection centre collects the boxes from the hospital. These are transported via road to the nearest regional laboratory where they are checked and then shipped via air to the national laboratory in New Delhi. A flow-chart describing the transportation process from the hospitals to the LPL National Reference Laboratory is shown in Figure 2 and the network is presented in a map in Figure 3. Time in transit is regularly monitored by LPL and reported for each participant along with their test results.

Figure 2. A flow-chart showing the transportation of samples from the hospital to the Dr Lal PathLabs National Reference Laboratory for processing and analysis.

Figure 3. Laboratory network for the MaatHRI platform.

Service 4: Standardising blood assays. All samples are processed and analysed in the LPL National Reference Laboratory based at New Delhi. The assay methods, traceability and performance characteristics are discussed by experts from the University of Oxford’s Wolfson laboratory and LPL before including a test in the study. Table 1 shows the traceability and Table 2 shows the performance characteristics for assays that are commonly used for the epidemiological studies undertaken using the MaatHRI platform. The details of specific tests will be presented in subsequent publications. Traceability and assay performance monitoring are important for standardisation of laboratory procedures and quality control. If the quality of a blood sample is compromised in transit, it is not processed, and the site-collaborator and research nurse are advised to collect a fresh sample. The laboratory runs quality control checks daily for each assay (twice a day for some) and monitors their mean coefficient of variation and standard deviation. The results are shared as part of a performance monitoring plan during monitoring and feedback meetings. In addition, LPL also runs a quarterly Quality Improvement Programme.

The LPL National Reference Laboratory is accredited by the following bodies – College of American Pathologists (CAP); National Accreditation Board for Testing and Calibration (NABL); British Standards Institution (Quality Management System ISO 9001: 2015, FS 60411).

Service – 5: Test reports. Test reports are securely made available to the site-collaborator in each hospital through their usual communication channel. Data from the reports are entered in the electronic forms by the research nurse.

Regulatory systems

MaatHRI steering committee has met biannually since the platform was established in September 2018. The role of the steering committee is to guide the platform in terms of vision, scope, equitable partnership, and research and training priorities. It is also responsible for communicating the results of the studies undertaken through MaatHRI to the Ministry of Health and Family Welfare (MoHFW), Government of India.

The DSMB periodically reviews participant recruitment, data safety and confidentiality, ethical issues and data quality, and examines whether the overall safety and feasibility of the MaatHRI project is acceptable. Although conventionally DSMB is set up for individual studies, we found that setting up a DSMB for the research platform that has oversight of all studies undertaken through the platform could be an effective way to ensure data safety. If the DSMB estimates a potential risk to participant or data security, it will be communicated to the MaatHRI steering committee who has the responsibility to mitigate the problem as soon as possible. The risk and measures taken to mitigate it will also be communicated to all collaborating hospitals and obstetricians as lessons learnt. If the risk or compromise has the potential to cause harm to any participant, the information will be communicated with the participant (at risk) by the site investigator.

The DSMB has met twice since MaatHRI was established in September 2018 and membership includes two obstetricians (one from the UK and one from India), one paediatrician (from India), one biostatistician (from the UK), and one expert in bioethics (from India), all with prior experience and expertise in observational epidemiological studies. They were nominated by the study investigators. So far, no risk or compromise to participant or data security has been identified.

Studies currently being undertaken through the MaatHRI platform

One survey and three observation studies are currently being undertaken through the platform. A monthly survey of nine life-threatening complications of pregnancy has been in progress since July 2018. The complications are defined using standard definitions and include eclampsia, pre-eclampsia, postpartum haemorrhage, maternal peripartum infection, septic abortion, uterine rupture, heart failure during pregnancy and postpartum, transient peripheral neuropathy, and Japanese encephalitis complications.

The epidemiological studies undertaken are informed by the knowledge and hypothesis generated during the pilot work for IndOSS-Assam. They include: (i) an unmatched case-control study examining the risk factors, clinical characteristics, and outcomes of heart failure in pregnant and postpartum women; (ii) a prospective cohort study investigating the safety of induction and augmentation of labour in pregnant women with anaemia; and (iii) a nested study within the prospective study comparing the coagulation parameters in pregnant women with and without anaemia. Of these, the nested coagulation study is complete, and the other two studies will be completed by June 2022. The monthly survey will continue as long as the collaborative platform exists.

Indian collaborating obstetricians suggest topics for research based on the needs of the local population. For example, based on the observation of high number of cases of health failure during pregnancy, this was proposed as a research topic by one of the collaborating obstetricians. In addition, the conditions included in the monthly survey of pregnancy complications were proposed by the obstetricians. Thus, all research projects undertaken through the platform are co-developed by the Indian and the UK collaborators. Aim of the MaatHRI collaboration is to undertaken research on conditions that adversely affect pregnant women in India so that the results from the studies benefit the future mothers in India and in other LMICs.

Strengths and challenges

MaatHRI is a collaboration of hospitals that covers deprived populations, some of which are located in remote rural areas of India. While this provides the opportunity to conduct research to improve the health of mothers and babies in areas of the country that have the highest burden of maternal and perinatal deaths, it also poses challenges related to resources and capacity. Appointing new research nurses to collect data and blood samples ensured that the MaatHRI platform was not depriving the hospitals of their scarce human resource. This has created an employment opportunity for nurses in the field of research, which is not a usual job for trained nurses in India. However, the challenge associated with this was the need for extensive training and constant supervision of the nurses. Furthermore, most of the collaborating hospitals had not been involved in a project of this scale and intensity encompassing not just implementation, but designing, standardising and developing the project as equal partners. Therefore, it took more than 20 months of continuous engagement with staff and collaborators to achieve the desired level of quality and standardisation for the MaatHRI platform.

Within the resource constraints, a further challenge is achieving a balance between an ideal collaborative research platform and a pragmatic solution. For example, the ideal platform would have collected data electronically on tablets using online forms, but this was not feasible due to a lack of good internet connectivity in the remote hospitals and cultural sensitivities. Therefore, paper forms are used in some hospitals. However, to mitigate risks and as advised by the DSMB, we have developed a documented process of securely storing and destroying the paper forms within an agreed timeline for each hospital.

Costs related to research staff, standardised laboratory parameters, programming data collection forms, and storing data on Microsoft Azure make studies undertaken through the MaatHRI platform more expensive compared with existing similar systems in the UK⁸, Europe and Australia⁹. It is our belief, however, that the benefits of generating high quality scientific evidence to answer important and urgent clinical research questions that will save the lives of thousands of future mothers and babies, outweigh these additional costs.