The study will be conducted at Liaquat University of Medical Health Sciences (LUMHS), which is a public sector tertiary hospital located in Hyderabad district of Pakistan. The hospital has a total of 3000 beds and 35 departments which serves a large number of mostly underprivileged populations. The hospital provides 24 hours’ emergency cover to patients coming from nearby urban and rural areas. LUMHS has three Obstetrics and Gynecology units.
The current data from the facility shows that a total of approximately 11205 patients were admitted in OBGYN units from the period of January to August 2021; and the maternal mortality rate was recorded as 159/11205 (1.4%). Out of these 159 deaths, 45 were due to confirmed maternal sepsis (28.3%). These indicators direct that there is a need of a robust system to early detect and manage maternal sepsis cases in the hospital.
The study will use a mix-method design and will be conducted in two phases.
Phase 1-Adaptation of FAST-M intervention (Qualitative)
For a FAST-M bundle to be effective in Pakistan, it is necessary to identify how best to implement the FAST-M bundle in the context of local settings. In order to adapt this intervention, a systematic method will be taken to understand the nature of existing practices and an appropriate system for characterising the intervention and its components that can make use of this understanding. This constitute phase 1 of the study.
This formative research (phase 1) will adopt a qualitative research design involving focus group discussion (FGD) and key-informant interviews (KIIs) and a purposive sampling approach. The aim of group discussion and interviews will be to engage health practitioners, government officials and other key stakeholders to understand the behavior of existing practices in the study setting for maternal sepsis care, to finalize the FAST-M tools for the context of Pakistan, and to identify various facilitators and barriers that may influence implementation of the FAST-M intervention. The FGD and KIIs will be conducted using interview guides developed through the use of the Consolidated Framework for Implementation Research (CFIR) .
Consolidated Framework for Implementation Research (CFIR)
The CFIR is a commonly used framework to facilitate implementation research design, evaluate and implement evidence-based interventions, and comprises five major domains: 1) Intervention characteristics, 2) Outer setting, 3) Inner setting, 4) Characteristics of individuals, and 5) Process of implementation. It is categorized as a determinant framework with the objective to understand and explain factors (individual or organization) which influence implementation outcomes . CFIR has been used in a wide range of studies because this flexible framework can be tailored to different settings across multiple contexts . We aim to use the tailored CFIR framework to assess critical barriers and facilitators to implementation that need to be addressed at multiple levels if the FAST-M bundle is to be successfully optimized, and adopted in health care practices in Pakistan (Additional file 1).
The interview guides (see supplementary file 1) for KIIs and the FGD have been developed using five major domains of CFIR to identify existing practices for sepsis management. These guides will also identify the facilitators and barriers to implementation of FAST-M intervention in the study setting. The identification of existing practices for maternal sepsis care and facilitators and barriers in phase 1 will then form the basis of feasibility testing of FAST-M intervention in phase 2.
Inclusion criteria for KIIs and the FGD
- HCPs including physicians, nursing staff, healthcare administrators who are associated with maternal sepsis care and management
- HCPs who have worked at the study site for last six months
15 to 20 semi-structured key informant interviews are planned in the qualitative phase of the study until data saturation is reached. One focus group will be conducted before initiation of the study to adapt the tools and identify implementation approaches; and a second will be conducted at the end of the study as a summative evaluation of the study to identify perceptions about success of implementation. Therefore, two focus group discussions (before and after implementation) will be conducted with 8-10 health care providers in each discussion.
Data collection and management
A semi-structured interview guide has been developed to explore healthcare professionals’ views and attitudes towards FAST-M intervention and its implementation at their facility. Before beginning the interview, the qualitative researchers will describe the FAST-M bundle components and the patient referral pathway demonstrating the algorithm and summary for utilization of FAST-M bundle care tool (Additional file 2).
A free flow of discussion among participants will be encouraged, using probes from these discussions to obtain healthcare professionals’ perceptions about the feasibility of the FAST-M intervention. Interviews will be conducted face-to-face in Urdu and English according to the participants’ preference, and will be audio recorded following consent from study participants. Interviews and focus group discussion will be conducted by experienced study team members who are also trained qualitative researchers. Detailed field notes will be also taken during each interview to capture non-verbal language and cues.
All data will be kept confidential for seven years on password-protected computers and/or locked filing cabinets only accessible to members of the research team. During transcription, audio-recordings will be referenced only with an identification number for anonymity of participants, with all identifying information removed before using the software analysis tool.
COVID-19- Standard Operating Procedures (SOPs)
In view of current of current COVID-19 pandemic situation, all project related activities will comply with standard operating procedures (SOPs). The following measures will be taken related to this study: 1) All research staff will be provided with appropriate masks, sanitizers, and/or other applicable Personal protective equipment (PPE) to the field staff; 2) Daily mandatory screening for COVID-19 symptoms of all project staff; 3) KIIs and FGDs will be conducted with social distancing (6 feet) with all vaccinated participants wearing face coverings.
Qualitative data gained through individual interviews and FGDs will be audio recorded, transcribed and analysed using an inductive approach to determine the facilitators and barriers for implementation of the intervention and will be summarized according to CFIR domains. This will help to understand the important contextual features that are helping or hindering the operationalization of the FAST-M intervention.
The analysis will be an ongoing iterative process during phase 1 of this study. The research team will conduct multiple reviews of the transcripts and tapes to familiarize themselves with the data and identify initial themes that will be reflexive and interactive. Analysis will begin as soon as the first interview is completed in phase 1 and will be continued concurrently with data collection to help determine when new information is no longer being generated from interviews.
Although, we identified the CFIR as the appropriate framework, additional codes may emerge during the familiarization process to develop a thematic framework from experiences of participants. The codes, categories and themes will be developed using NVivo version 10 (QSR International, Pty Ltd) software.
An audit trail will be used to document our decision-making process. Sections of the transcripts will be charted, organized by CFIR domains, and then re-framed to better reflect descriptions from participants. The primary team will review the codes and associated themes multiple times to check for potential biases, to ensure they are reflecting participants’ words and meanings, and improve the credibility of their interpretation of the interviews. Initial findings will be shared with a group of participants to help with interpretation and generate meaning from the data.
The facilitators and hindering factors will be identified through phase 1 of the work. The FAST-M bundle care tools (MEOWS chart, decision tool and treatment bundle) will be modified through construal gained from interviews and discussion with health care providers.
Phase 2- Intervention phase
Following phase 1, intervention phase will be implemented for the feasibility testing.
During the intervention phase, patients will be assessed by a healthcare practitioner on decision to initiate screening for potential maternal sepsis that will be based on the following inclusion criteria:
- Women who are pregnant or within 6 weeks of miscarriage, termination of pregnancy or delivery
- Abnormal maternal observations triggered on the inpatient MEOWS chart
- Healthcare practitioner concern regarding potential maternal infection
- Fetal tachycardia greater than or equal to160 beats per minute
For enrollment of sepsis cases, we will power to a primary process outcome of "sepsis management compliance”. This is defined as "the proportion of patients admitted with features of sepsis who receive appropriate monitoring (full set of vital sign measurements on admission) and antibiotics within 1 hour (if required)." This means the notes of all patients with suspected or confirmed sepsis will be reviewed and their data would be collected using study Case Report Forms (CRFs).
Assuming baseline compliance is less than 10%, grounded on observations from FAST-M study in Malawi, to detect an increase in compliance to 20%, with an alpha of 0.05, we will require the observation of 199 participants in each phase to achieve a power of 80%. This is adequate precision to allow important increases to be estimated. Allowing for loss to follow-up and missing / laboratory results, we consider an initial sample size of 400 as appropriate to allow the study to have adequate power to detect an increase in compliance. This number of cases will be feasible to collect within 6 months, based on current rate of sepsis from hospital records of anticipated site. The flow of participants through the study is presented in Fig. 1.
This feasibility study is anticipated to run for seven months. This includes a baseline assessment period of two months, and training programme planned to schedule at completion of baseline phase before commencing intervention phase of four months.
The intervention phase will be introduced after training all health care provides involved in management of maternal sepsis at the study site. At the start of the intervention phase, FAST-M bundle care tools (Additional file 2) will be introduced including MEOWS chart, FAST-M decision tool, and FAST-M treatment bundle. Fig.2. provide the summary of enrollment, intervention and assessment
Modified early obstetric warning score
MEOWS stands for modified early obstetric warning score (MEOWS) to identify suspected maternal sepsis patients. This tool helps in identifying any early warning scores used to track the physiological parameters of an individual over time onto a chart, with guidance thresholds to trigger clinical action of they become abnormal . The MEOWS chart used during implementation of the FAST-M intervention in the districts of Malawi will be adapted in context of Pakistan for the purpose of this feasibility study .
The use of obstetric early warning systems (OEWS) in UK maternity units was recommended in the 2007 Confidential Enquiry into Maternal and Child Health (CEMACH) report as an adjunct to reducing maternal morbidity and mortality.  MEOWS consisted of scores of respiratory rate, oxygen saturation, temperature, heart rate, blood pressure, assessment of urine, including for proteinuria, color of amniotic fluid, neurological response, pain score, assessment of lochia, and an overall assessment of whether the woman appears well . Clinical action is triggered by a single parameter exceeding a red threshold or any two parameters exceeding a yellow threshold. MEOWs chart have been widely adopted in the UK and internationally .
To complete the MEOWS chart, the healthcare providers involved in the study will be trained to record patient observations (heart rate, respiratory rate, blood pressure, conscious level, urine output and temperature) and fetal heart rate (if applicable) from medical records. These observations will be charted on a MEOWS chart in the inpatient setting.
Abnormal observations (indicated by a single red or two yellow thresholds) will trigger a review by an attending doctor or nurse. This will be agreed locally prior to study commencement. These patients will then be screened for potential sepsis using the FAST-M decision tool. In addition to abnormal maternal observations, cases of suspected sepsis will also be identified using the FAST-M patient pathway when prompted by attending clinician concern regarding potential maternal sepsis or an increased fetal heart rate greater than or equal to160 beats per minute.
Patients will be defined as having or are at a higher risk of having sepsis, who will trigger a red flag on the decision tool and will be commenced immediately on the FAST-M treatment bundle pathway. These patients will receive a review from a doctor/nurse as soon as possible, with the bundle initiated within one hour. Those patients who trigger two yellow flags on the decision tool and have or at a higher risk of having sepsis require a review from a doctor/nurse within three hours. All suspected cases will remain in observation for possible development of red flags. Half-hourly (if red trigger) or hourly (if two yellow triggers) observations will be made in the first instance, until otherwise specified by an attending clinical decision maker. Those patients without at least one red or two yellow flags will be considered to have a low risk of sepsis and will be managed according to local guidelines by the screening healthcare practitioner.
FAST-M treatment bundle
Patients managed with the FAST-M treatment bundle will have their treatment recorded on the FAST-M treatment bundle form including documentation of actions completed and any reasons for not completing certain component of the bundle.
The FAST-M treatment bundle consists of the timely consideration of all the following:
· Source identification and control
· Assessment of the need to Transport / Transfer to a high level of care
· Ongoing Monitoring (of the mother and neonate)
Co-interventions for implementation of intervention
Multiple full day training sessions by the study team will be delivered to healthcare practitioners working for maternal care and sepsis management at the study site. The interactive sessions will be offered in English and Urdu languages for each healthcare practitioner to understand the processes completely. Any requirement for supplementary educational material such as posters and a study booklet will be determined during the implementation programme via feedback from front line clinical staff and stakeholders on facilitators and barriers to use of the tools. This will be done using qualitative interviews and focus groups discussion.
The training and implementation programme is likely to consist of:
· Background information on maternal sepsis, including risk factors, signs and symptoms and the potential consequences if untreated
· Use of the MEOWS chart to track and trigger the recognition of deteriorating patients
· Use of the FAST-M decision tool to recognise and screen for potential study participants at risk of maternal sepsis
· Use of the FAST-M treatment tool to initiate the bundle components
· Guidance around implementing the individual components of the FAST-M bundle
· Use of feedback tools (run chart and dashboard) and approaches the team can use to work together to improve compliance and outcomes
Post training, an impact survey will be made to measure the extent to which skills and knowledge learned in the program have translated into improved behavior among participants who attended the training program.
The local clinical champions and team leaders will be identified and trained to take a lead at study sites from different units where study will be implemented, and will remain engaged throughout the implementation process. The overarching goal of each champion will be to encourage engagement and compliance with the FAST-M bundle. To achieve this goal, champions at each site will be engaged in a number of key activities: disseminating knowledge, advocating, navigating boundaries, facilitating consensus, arranging meetings with stakeholders, tracking quality indicators and developing organizational communication strategies and relationships.
Ongoing improvement approaches
Ongoing improvement practices at different units of the study site will be carried out by clinical champions of the respective units. The improvement strategies include: 1) weekly/biweekly training of health care providers on FAST-M tools, 2) display of run charts, dashboards in units to demonstrate rate of maternal sepsis and outcomes of maternal sepsis cases over-time, and 3) meeting with stakeholders for communicating needs and requirements for implementation of the FAST-M intervention. Table 1. shows the summary of ongoing improvement approaches planned to implement for FAST-M implementation
An overview of the implementation of the complex intervention is illustrated in Figure 3 below;
Data collection and management
During the intervention phase, data will be collected by a member of the research team who will not be part of the clinical team. Data will be collected using CRFs on various outcomes; structural, clinical, organizational and any adverse events.
If the patient requires a transfer as part of the FAST-M treatment bundle to any other health facility due to shortage of beds or other resources, the data collector will continue to follow up the patient’s clinical outcomes. The data collection team will keep their study site updated on their performance using this data, and will visually display it on run charts and dashboards and work on strategies to improve performance. The data will be maintained in an investigator file to be secured in a locked cabinet. Information recorded on the data collection sheet will be recorded in a database located on a secure server.
Quantitative analyses will be done to assess numerous outcomes; process, organizational, clinical, structural and adverse events with quantitative comparisons made between before and after implementation of the bundle. Quantitative data will be analysed using percentages, means, medians interquartile ranges and 95% confidence intervals and the change identified over time. Binary outcomes will be analysed using logistic regression and continuous measures by linear regression.
A mixed methods approach will be used to explore the implementation of the intervention. In this approach both quantitative and qualitative data collection methods will be used, and then integrated to draw conclusions. A sequential exploratory design will be used to collect qualitative data for adaption of the FAST-M bundle care tools and will be applied to make these tools contextual based. This will be then followed by the implementation of contextual based modified FAST-M tools at the study setting. This mixed-methods study will help in exploring the perspectives and adaptation of FAST-M intervention in phase 1 and evidence of its feasibility in phase 2 of the study. This will allow us to assess practicality of implementation in order to build a robust and successful full-scale trial for future.
Main outcome measurements
We will explore a range of outcomes measurement for maternal sepsis care. Primary process include 1) the proportion of patients admitted with features of sepsis who received appropriate monitoring (full set of vital sign measurements on admission recorded on MEOWS chart) 2) the proportion of women with suspected maternal sepsis received antibiotics within 1 hour (if required), 3) the proportion of women with suspected maternal sepsis receiving the FAST-M treatment bundle (including each bundle component) within 1 hour of identification of sepsis. Secondary outcomes will include: 1) the proportion of women with suspected maternal sepsis referred to clinical decision maker on the basis of abnormal vital signs records; and 2) the proportion of women with suspected maternal sepsis receiving a clinical review by a senior clinical decision maker following their diagnosis.
Fluid resuscitation in patients with sepsis if not managed appropriately can precipitate volume overload and subsequent pulmonary edema. This is a particular concern in patients with preeclampsia. Clear teaching and guidance regarding fluid resuscitation will be provided during the training programme. When fluid resuscitating patients with suspected maternal sepsis, the decision regarding the rate of fluid administration will be made by the responsible clinician based on clinical examination findings and ongoing monitoring.