Screening for infectious maternal morbidity - knowledge, attitudes and perceptions among healthcare providers and managers in Malawi: a qualitative study

Key informant interviews and focus group discussions were conducted with healthcare providers working in the obstetric department the Queen Elizabeth Hospital, in Blantyre, Malawi, in June 2019 (prior to Covid-19 pandemic caused by SARS-Co-V). This hospital is the largest teaching hospital in the south of Malawi providing routine and specialised antenatal, intrapartum, and postnatal care, in addition to receiving high risk referrals from surrounding healthcare facilities. There are on average 1000 deliveries per month, with a Caesarean rate of 30%. All interviews were held a quiet office in the healthcare facility away from the clinic rooms to ensure privacy.

Healthcare providers (doctors, nurses, midwives) were included if they provided routine maternity care (including antenatal and postnatal care) at the chosen study site. Healthcare managers (Ward matron, Head of Department, Head of facility) were included to enable the triangulation of the data and broadened the scope of the topic. Snowballing and opportunistic sampling techniques were employed to identify the participants. Participants were chosen purposively, based on their ability to speak English, and were recruited sequentially until data saturation was met.

A topic guide was developed and piloted at the study site in Malawi. The topic guide was a flexible tool that enabled the interviewer to capture the healthcare providers’ responses as well as acting as a cue to probe further to understand the participants’ perceptions and beliefs (Supplementary File 1). In addition to sociodemographic questions, the topic guide included five main subject areas: (1) overall understanding of screening for infectious morbidity (2) knowledge and perception of early warning scores; (3) experience and views on use of early warning scores; (4) approaches to management of women with infectious morbidity; and (5) suggestions on how to identify or screen women with possible infectious maternal morbidity.

Key informant interviews and focus group discussions were conducted face-to-face in English, lasted on average 30–45 min, were recorded on a digital recording device, and transcribed on completion. Anonymity and confidentiality with regards to data reporting were emphasised to reassure participants’ confidence in providing honest answers. All participants approached agreed to participate in the study and completed the interviews.

The interviews and focus group discussions were transcribed verbatim by the first author (ES). The first author (ES) and a second reviewer (HWU) independently coded all transcripts. The identified codes were grouped into categories and reviewed by three researchers (ES, HU, MMC) to ensure consistency. This enabled the first extraction of data [36]. Key themes were then discussed and checked by all researchers together to reach consensus. We used the Standards for Reporting Qualitative Research guidelines in reporting the analysis [37].

Ethical approval was granted by the Liverpool School of Tropical Medicine, UK (LSTM14.025) and by the University of Malawi College of Medicine Research and Ethics Committee (COMREC 2724). This research was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants of the study.

Twenty healthcare providers and four healthcare managers participated in the study (10 doctors and 14 nurse-midwives; 16 females and 8 males). Two Consultant obstetrician-gynaecologists, four nurse-midwives and four healthcare managers (Head of facility, Head of Department, Matrons) provided key informant interviews and six doctors with varied levels of experience (junior doctor, specialist registrar, and consultant) and eight midwives participated in two separate focus group discussions. The age of participants ranged between 25 and 65 years, with the majority between 25 and 35 years. Most participants were female and had between one to 5 years of experience of providing maternity care.

The main emerging themes are presented below with illustrative quotes provided in Tables 1, 2 and 3.

Healthcare providers were aware that women were at high risk of infective morbidity both during and after pregnancy (Table 1, Q1–2) and were aware that infective maternal morbidity includes a spectrum of disorders which vary in severity (Table 1, Q3–4). Healthcare providers had knowledge and experience of infectious morbidity in women not being recognized and leading to severe and life-threatening infections and sepsis especially where women had been discharged after birth and had to be readmitted in the postnatal period (Table 1, Q5). Healthcare providers reported that current approaches to screening for infection included: physical examination, checking vital signs as a form of triage, the ChEWA (the Chantinka Early Warning Alarm) screening tool in the high dependency unit and one participant mentioned knowledge of the SOFA (Sequential Organ Failure Assessment) score for detection of sepsis (Table 1, Q6–10). Many healthcare providers were aware of the benefits of using an early warning score as an approach to screening for infection (Table 1, Q11–14). There was an underlying willingness of the healthcare providers to provide better care. However, there were significant barriers in place.

Reported challenges to the comprehensive assessment of infectious maternal morbidity included lack of simple equipment such as a thermometer, a lack of time with high patient volume and a small staff complement, staff not trained to use early warning tools or scores other clinical emergencies taking priority (Table 2, Q14–20). Many healthcare providers reported that a lack of time due to the large number of women attending for maternity care and the lack of simple equipment were challenges that made it difficult for them to screen and manage women who might have infectious morbidity but were not (yet) seriously sick (Table 2, Q21–23). Participants commented on the challenges to train and orientate staff in the interpretation of using early warning scores (Table 2, Q24–25); and that some patients may not be able to buy the treatment recommended if infection is detected (Table 2, Q26).

Healthcare providers were keen to discuss possible solutions and had recommendations regarding how to introduce screening as part of routine maternity care (including antenatal and postnatal care) with various approaches to implementation suggested. Healthcare providers were keen for the introduction of routine infection prevention health guidelines and a standardised questionnaire or triage system or an early warning score to help guide the assessment of women during antenatal and postnatal care, both in outpatient and inpatient settings. The use of visual aids such as posters and the use of colour to highlight severity (for example using a traffic light system: red, amber, green) were suggested as useful aids to implementation (Table 3 Q27). Many healthcare providers reported that training and orientation for new (and existing staff) would help (Table 3 Q28–30) with the implementation of an early warning score, and that screening for infection should be used in all departments of the hospital (Table 3 Q28–33).

Some participants reported that the adaptation of any triage system or early warning score to their own local setting was important, especially if the scoring system was originally designed and validated in high resource settings, as the resources in place in their setting (for example laboratory support) were much more limited. Participants emphasised that any additional screening approach would need to be perceived as a useful additional aide to help them care for their patients, instead of an additional cumbersome and ineffective ‘paper task’ to complete. They also felt a multidisciplinary team approach, collaborating with laboratory technicians and microbiologists as part of routine maternity care was important and encouraged a more comprehensive and holistic approach to maternity care. Local clinical leadership would benefit implementation and the use of clinical audit would demonstrate the effectiveness of such implementation to other areas of the hospital, including antenatal and postnatal departments (Table 3, Q34–36).

Healthcare providers are aware of the problem and impact of maternal morbidity and mortality due to different types of infection during pregnancy and after childbirth. Healthcare providers understand that maternal morbidity related to infection should be specifically assessed and can present at any time from early pregnancy through to the late postnatal period. Early identification of infectious morbidity can be difficult if healthcare providers can only rely on the overall clinical condition of women and/or on women’s self-reporting of symptoms. However, ‘routine’ screening of women for infectious maternal morbidity during and after pregnancy as part of routine maternity care is not in place due to a lack of training, lack of time, and lack of free and easily available diagnostic tools (including thermometers). Screening using an early warning score that was easy to use, colour coded, adapted to the local settings, and properly implemented was welcomed. Healthcare providers also reported needing support from the wider multidisciplinary team including laboratory technician as well as a re-focus on prevention and early identification of underlying maternal morbidity in addition to life-threatening conditions.

Our study adds to existing evidence exploring the knowledge and attitude of healthcare providers regarding the potential usefulness and effectiveness of approaches to early detection of infectious maternal morbidity as part of routine maternity care in low resource settings where the burden of disease is high and unrecognized infection is known to be one of the causes of maternal and neonatal mortality. This study highlights barriers as well as solutions to inform programmes that seek to introduce and establish routine screening for infectious maternal morbidity during and after pregnancy in low resource settings. Practical recommendations were provided by a range of different cadres of healthcare providers (both female and male), who worked in different departments within a busy maternity setting and had varied level of experiences. All healthcare providers interviewed recognised this as an important area and were keen to embrace change in practice that would help reduce maternal and neonatal morbidity and mortality. This study population comprised mainly doctors and nurse-midwives providing routine maternity care in one large teaching hospital and the findings cannot be assumed to be the same in other hospitals and other settings. However, it may be assumed that resource limitations for community-based healthcare providers in different settings are similar or worse. This study used a qualitative study design and limitations are that findings cannot be generalised because of smaller sample sizes and context specificity; data analysis can be influenced by researchers’ personal perceptions and unconscious biases; and rigor is more difficult to maintain, assess, and demonstrate [38]. However, for this study we chose to use qualitative research methodology to understand the local context and influences; to listen to healthcare provider’s voices, experiences, and opinions; to explore in-depth knowledge, attitudes; and to understand perceptions of a proposed clinical intervention.

In our study, many healthcare providers had knowledge of and were aware of the benefits of early warning scores, like other studies [39‐41]. Healthcare providers were familiar with the use of an early warning score in the maternity high dependency unit as part of critical care, the ChEWA score which uses standard clinical observations such as pulse rate, respiratory rate, temperature, blood pressure, in combination with the mental state score (AVPU – alert, responds to visual stimuli, responds to painful stimuli, unresponsive) to detect the deteriorating unwell mother after childbirth [23]. The benefits of the use of early warning scores in our study was like a study in Ethiopia where healthcare providers were positive about the implementation of an early warning score, reported that this was easy to use and improved communication between nurses and doctors [40]. However, like our study and other studies conducted in low resource settings, healthcare providers reported challenges to implementation including regarding lack of staff, lack of time, lack of resources and lack of training on how to interpret early warning scores [41‐44]. One study conducted in Malawi highlighted that healthcare providers reported that the time required to measure vital signs required to obtain an early warning score is too much and that healthcare providers would prefer to continue with a quick subjective clinical judgement only to identify if a patient has an infection or not [41]. Similar to findings in our study, local ownership and clinical leadership, good orientation, staff training and integration of clinical change into respective care bundles and local management guidelines are important for successful implementation [23, 40‐45]. In our study, many healthcare providers did appreciate the benefits of screening for infection in women as part of routine antenatal and postnatal care, not just during labour and childbirth as part of critical care. However, similar to another study conducted in Uganda, in our study healthcare providers did emphasis the importance of ensuring any proposed change in clinical practice was adapted to the local setting and resources made available for any change to be adopted and effective, specifically highlighting the difficulty in using a blood result as a parameter for early warning score due to lack of rapid results from the laboratory [46]. There are portable rapid diagnostic testing devices for white cell count, lactate and C-Reactive protein available but assessment of their feasibility and cost effectiveness as screening for infection in routine maternity care in low resource settings is awaited [47‐49]. The importance of the provision of free targeted treatment and further management if infection is detectedmust be available alongside the implementation of routine screening for infectious maternal morbidity. Any proposed single implementation must be contextually appropriate and complement current clinical care to support the provision of high quality comprehensive holistic maternity care [50].

There are currently several international clinical and policy guidelines on who should enquire, screen, and manage infectious maternal morbidity during and after pregnancy as part of routine antenatal and postnatal care, and how this should be conducted, including identification, counselling, documentation, first line medication and provision of higher referral pathways [51, 52]. However, the practicalities associated with the introduction, implementation, sustainability of change and acceptability of these guidelines as part of routine maternity care in countries such as Malawi are currently uncertain. There is a need to better understand how infectious maternal morbidity can be prevented, detected using early warning scores, investigated, and treated across different healthcare levels in low resource settings. There is also debate as to who is most suitable to enquire, screen for and manage infectious maternal morbidity and at what level of the health system (community, primary or secondary health care level) across LMIC. In many high-income countries, specially trained nurses or midwives routinely assess, support, and provide further referral between different levels of care. This approach in a low resource setting requires further research and evaluation.