Why strengthen clinical assessments during pregnancy?

Our implementation research on maternal safety in Koppala district, Karnataka,15–17 identified four interconnected contributors to the problem of poor clinical care for rural women during the pregnancy.

First, a tendency among healthcare providers, as well as women and their families, to ‘normalise’ commonly encountered risks. For example, in poor communities where anaemia is highly prevalent, doctors tend not to actively manage diminished haemoglobin levels until they fall perilously.

Second, time constrained doctor–patient interactions due to large inflows of women for antepartum care and delivery, especially after the roll-out of the National Rural Health Mission. In such a scenario, women often fail to receive individual clinical attention. Harried doctors tend to fall back on a checklist of discrete symptoms to quickly screen women instead of using clusters of symptoms, signs and test results to guide their diagnoses of maternal risks and comorbidities. The latter is essential if clinical diagnoses are to be rigorous. For example, we found that border-line blood pressure did not prompt doctors to look for other indications of pregnancy-induced hypertension, such as swelling of the legs and albumin in urine.

Third, inconsistent clinical capacity among doctors (some being non-allopaths), due to variations in the content and quality of their basic training, and of their clinical exposure during internships. Inconsistencies also arise due to lax implementation of standard protocols.

Fourth, task shifting from higher to lower levels of care, often with inadequate supportive supervision. For example, although ANMs are formally responsible for tracking ‘risk cases’ identified by the MO, the ASHAs are often the ones who actually follow up. With insufficient training and only sporadic supervision by the ANMs, ASHAs provide little more than nutrition-linked information to women. They are also expected to collect data about women’s antenatal well-being. Yet, without sufficient knowledge, the ASHAs tend to ask non-specific questions (‘do you have any problems?’), and report symptoms that are neither graded for severity nor linked to action. Such task shifting is counterproductive. ASHAs invest considerable time and effort in collecting information that is not particularly useful for the MO, who requires reliable, actionable information.

M-CAT: a digital tool for effective risk identification and clinical management

As an information technology-enabled tool, M-CAT was designed to enable ASHAs to collect clinical data from pregnant and postpartum women and assist MOs to identify maternal risks. It expanded the current focus on age, height, parity and spacing by including information about a woman’s obstetric and medical history, as well as commonly occurring comorbidities. More importantly, it emphasised a cluster-based approach to diagnosing and grading the severity of both maternal risks and comorbidities. It oriented ASHAs to the idea that maternal risks require prompt action and provided doctors with case-sheets functioning like protocols to prompt systematic clinical assessments. In its design, the tool was an individualised response mechanism to clinical needs rather than a way of aggregating population level data to guide health programme planning.

M-CAT has four major components (table 1).

Drawing from a handbook on maternal risks,18 19 the forms included all potentially life-threatening obstetric conditions and comorbidities (acute and chronic) that can be algorithmically deduced from clusters of symptoms/signs and simple tests. Our medical team together with a group of experienced, practising obstetricians finalised this list of conditions (table 2).

The tool excluded health problems that can be diagnosed only after detailed investigations (eg, domestic violence, postpartum depression) or sophisticated tests (eg, cortical venous thrombosis). It also excluded diseases that must be kept confidential (eg, HIV).

Questions in the tool were arranged in a systematic way. Algorithms to provide preliminary clinical assessments were created using the clinical guidelines in our handbook. The software was developed as an offline application written in Java, running on a Tomcat server with a MySQL database. The software had a simple user interface and could generate reports for each woman, identifying any risks and their severity, in a format similar to a short clinical case sheet.

The implementation and feasibility study

The feasibility study was conducted in all villages under one PHC in Koppala district, Karnataka. M-CAT implementation involved 14 ASHAs, 2 ANMs and the PHC MO, as well as two project functionaries (data entry operator (DEO) and link worker (LW)). A total of 349 pregnant and postpartum women were registered and followed up over four successive 30-day cycles between September 2016 and January 2017. Our team monitored and evaluated this implementation closely, via direct observation of ASHA’s home visits, participant observation of debrief meetings, focus group discussions with ASHAs, in-depth interviews with randomly selected women and their families, and review of the clinical database generated through M-CAT work cycles. While monitoring the roll out, the research and implementation teams met regularly to share and analyse field experiences.

A detailed discussion on the methods and findings from the feasibility study are beyond the scope of this paper. Our focus is to describe our experience of developing and implementing the M-CAT by working closely with the ASHAs. We reflect on the strategies that we employed in working with the ASHAs and how the ASHAs felt empowered as a result of those strategies.

Refining the forms

Our first interaction with the ASHAs focused on refining the clinical questions that were already phrased using simple yet accurate language. These were tweaked further through dialogue between the ASHAs and our medical team: the former suggesting modifications to the questions in the form, the latter vetting these for accuracy. The aim was to have questions that can be administered easily while keeping the women comfortable (eg, avoiding sensitive questions requiring privacy).

Building knowledge and skills

We then held a 3-day training workshop for 14 ASHAs in the PHC along with two ANMs. Employing interactive methodologies, the training aimed to enthuse and orient each functionary to the forms, processes and their roles. Half a day was also spent on orienting the MO, LW and DEO to their respective roles.

Following this initial training, the ASHAs, LW and DEO were provided practical field-based training and support. The research team and LW accompanied the ASHAs over 2 days for hands-on training in form filling. This activity provided a baseline picture of the ASHAs’ capabilities and indicated the type of support they would need. It also initiated the LW into her role of providing hands-on support.

Six of the 14 ASHAs who were less literate and communicative, needed more support than the others. Continuous handholding was crucial for the first couple of cycles as these six ASHAs were failing to adhere to instructions that were critical for eliciting symptoms. The other eight ASHAs required much less oversight to ensure that form-filling proceeded as mandated.

Over time, even the six ASHAs who required extensive support became adept at form-filling. Sheer practice also helped them elicit symptoms expertly. They were given an incentive of Rs50 per form to fill it without errors, which may have sustained their motivation to learn and to be accurate.

By the end of the fourth cycle, all ASHAs had become very familiar with the questions, which they articulated clearly, and rarely made errors in filling the forms.

Complementing the individual support mechanisms were monthly debrief meetings at the PHC. The ASHAs and project staff (LW, DEO) as well as representatives of our implementation and monitoring & evaluation teams attended these meetings that aimed to: (1) give ASHAs feedback on their performance; (2) allow them to talk about their experiences, challenges and concerns; (3) troubleshoot, clarify queries and retrain them (as required) and (4) develop action plans for women at risk (identified through M-CAT or otherwise). The quality of the interactions at the debrief meetings was usually very good, with the ASHAs participating actively and demonstrating their openness to learning

Dovetailing with the existing structure

We learnt from our field interactions that only two cadres of health workers effectively engage with pregnant and postpartum women: the ASHAs and the doctors. The role of the ANMs had significantly reduced since the introduction of the ASHAs, and the staff nurses merely supported the doctors. Given the serious time constraints faced by the doctors, the M-CAT had to be designed with the ASHAs in mind.

Since the ASHAs were already saddled with multiple responsibilities, the M-CAT had to dovetail with their existing responsibilities. Moreover, it was important for this activity to improve the efficiency of the doctor without becoming onerous.

Keeping this in mind, we integrated data collection into the ASHAs’ schedule of activities during regular home visits. We also ensured that M-CAT forms systematised and improved the questions that the ASHAs would ask anyway. In addition, we asked the ASHAs to plan their home visits before the scheduled monthly ANC clinics at the PHC, leaving enough time for the data to be processed and for reports to be generated for the MO.