Resource availability for the management of maternal sepsis in Malawi, other low‐income countries, and lower‐middle‐income countries

To assess the availability of key resources for the management of maternal sepsis and evaluate the feasibility of implementing the Surviving Sepsis Campaign (SSC) recommendations in Malawi and other low‐resource settings.

in some low-income countries, accounting for between 16.3% 4 and 29.4% 5 of maternal mortality in Malawi. To meet the maternal mortality target set out in the Sustainable Development Goals, considerable progress is needed.
In high-income settings, there has been widespread implementation of the Surviving Sepsis Campaign (SSC) guidelines. 6 These guidelines recommend a group of evidence-based interventions aimed at improving the diagnosis and management of sepsis. The interventions include 3-hour and 6-hour "bundles," which represent selected elements of care that-when implemented together-improve outcomes and reduce mortality (odds ratio 0.66; 95% confidence interval 0.61-0.72), 7 although data specific to the maternal population are unavailable. Although effective, doubts exist regarding their applicability in low-resource settings. 8 The feasibility of implementing the SSC guidelines has not previously been examined in low-resource settings and would be critical in future attempts to produce guidance suitable for such settings.
There are several key elements that make the SSC guidelines clinically effective. Alongside the 3 and 6-hour bundles, the "key resources" needed for the initial management of maternal sepsis are fluids, antibiotics, basic monitoring (blood pressure, pulse, temperature), and basic infrastructure (water, sanitation, hygiene). 9 These have been highlighted as the cornerstones of effective sepsis management in any setting. 10 Better understanding of their availability is critical in planning programmatic interventions to improve sepsis outcomes in low-income and lower-middle-income settings.
Malawi was the focus of the current project; it is an example of a country with a high sepsis morbidity and mortality. 4,5 It has been selected as the location for a feasibility study examining the introduction of a maternal sepsis bundle designed specifically for use in the maternity population in low-income countries. This will form part of a wider WHO maternal sepsis program. 11 The present study aimed to investigate the perceived availability of key resources for maternal sepsis management and SSC guideline/ bundle implementation in health centers and hospitals in Malawi, and to compare resource availability across hospitals in Malawi, other lowincome countries, and lower-middle-income countries.

| MATERIALS AND METHODS
Data were collected using a self-reported, cross-sectional survey of healthcare professionals in maternal healthcare facilities, conducted between January 31 and March 31, 2016. The study questionnaire (Appendix S1) was distributed to healthcare professionals from 21 healthcare facilities in Blantyre, Malawi, and in an online survey targeting all 82 low-income and lower-middle-income countries as classified by the World Bank. In Malawi, all facilities providing maternity care in Blantyre were sampled. Any healthcare professional involved in maternal health care at the time of the questionnaire distribution was invited to participate using a paper-based, structured questionnaire to assess maternal sepsis management resources. For the online survey, snowball sampling was used and internet searches were conducted to identify relevant international obstetrics and gynecology organizations, healthcare facilities, and professionals active in the field through published literature in the relevant countries. Snowball sampling is a nonprobability sampling technique whereby existing study participants recruit further study participants; therefore, response rate calculations are inappropriate.
Participants were informed that completion of the questionnaire or online survey constituted consent. Ethics approval was obtained from the University of Birmingham (Birmingham, UK) and from the College of Medicine in Blantyre, Malawi.
Healthcare professionals included doctors, nurses, midwives, and administrators among others. All healthcare professionals were required to read and write English to complete the questionnaire (or online survey). Any healthcare professional who was a US citizen was excluded from participating because the University of Birmingham was unable to cover potential liability for their involvement.
The questionnaire and online survey (conducted using SmartSurvey were identified by consensus among the research team. Resources were classified as "always," "sometimes," or "never" available, or as "don't know." Only resources considered to be "always" available were deemed sufficient to consistently implement the SSC guidelines. The "key resources" investigated in the current analysis (fluids, antibiotics, basic monitoring, and basic infrastructure) have been highlighted as essential in the initial management of maternal sepsis. 9,10 "Basic monitoring" includes measurement of the blood pressure, pulse, respiratory rate, and temperature. "Basic infrastructure" includes the availability of water, sanitation, and hygiene services, 13 specifically the availability of clean gloves, running water, skin cleaning preparation, soap, and alcohol gel.
Moreover, several "additional resources" central to other internationally recommended sepsis bundles 9,14 were investigated, including oxygen therapy and measurement of oxygen saturation and lactate. The questionnaire and survey responses were categorized into responses from facilities in Malawi, other low-income countries, and lower-middle-income countries. The planned comparisons were between facility types within Malawi, between hospitals in Malawi and other low-income countries, and between hospitals in low-income and lower-middle-income countries.
The statistical analysis was conducted using SPSS version 22.0 (IBM, Amonk, NY, USA). The respondents from Malawi were grouped on the basis of their type of facility, and the mean percentage of implementable SSC recommendations per facility type was calculated. If there were two respondents from the same facility, the more conservative response was used. If there were more than two respondents from a facility, the mode of their responses regarding the reported availability of that resource was used to represent that facility. Tests of normality were conducted using the Shapiro-Wilk test with a priori comparisons conducted between nonparametric continuous data using the Mann-Whitney U test, and between categorical data using the Fisher exact test. P<0.05 was considered statistically significant.  Despite an overall poor availability of resources, there was great variation between the facilities within Malawi. Facilities in the largest districts (Lilongwe, Blantyre, and Mzuzu [central hospitals]) had a greater availability of key resources than did hospitals in smaller, more rural districts (Fig. 1). In Blantyre, facilities closer to the center of town had greater resource availability than did more-peripheral facilities (Fig. 1).

Study invitations were
The percentage of hospitals with all key resources available was greatest in lower-middle-income countries, followed by low-income counties and Malawi (Table 2). Although hospitals in lower-middleincome countries and low-income countries had similar availability of the key resources for basic monitoring and infrastructure, fluids and antibiotics were significantly more often available in hospitals in lower-middle-income countries (Fig. 2).
In terms of the availability of additional resources within Malawian healthcare facilities, oxygen therapy was as often available as some of the key resources (15 [ Hospitals in lower-middle-income countries were able to implement a higher percentage of SSC recommendations than hospitals in Malawi or in low-income countries ( Table 2, Fig. 2). Compared with hospitals in low-income countries, hospitals in lower-middle-income countries also had a greater ability to implement elements of the SSC 3-hour and 6-hour sepsis bundles, including antibiotics and vasopressors (Table 2).

| DISCUSSION
To our knowledge, the present study is the first to evaluate the avail- Regarding elements that make the SSC recommendations clinically effective, the timely administration of antibiotics is vital and improves mortality outcomes. 8,15 However, the availability of antibiotics varied between hospitals in lower-middle income countries (95.1%), hospitals in low-income countries (58.8%), and health centers in Malawi (39.1%). Efforts to increase the availability of antibiotics in low-income settings and to allow the rapid initiation of effective treatment should be a priority. More promisingly, fluid resuscitation-another mainstay of sepsis management 8 -was more widely available in hospitals T A B L E 1 Availability of key resources for the management of maternal sepsis and percentage of implementable SSC guideline recommendations in Malawi. a T A B L E 1 (Continued)

Variable Public vs private health centers Health centers vs hospitals
F I G U R E 1 Availability of key resources for the management of maternal sepsis and percentage of SSC guideline recommendations that could be implemented in Malawi. Abbreviation: SSC, Surviving Sepsis Campaign.
T A B L E 2 Comparison of the availability of key resources for the management of maternal sepsis and percentage of implementable SSC guideline recommendations between hospitals in Malawi, other low-income countries, and lower-middle-income countries. a

Malawi vs other low-income countries Other low-income vs lower-middleincome countries
Availability of key resources All key resources available 4 (30. 8 To improve the outcomes of maternal sepsis, prompt identification of this condition and immediate initiation of care are required. A bundle-based approach has proven effective in other settings, 7 but an essential principle of the success of this approach is that all elements must be available and used in every eligible patient. 20 With the current SSC bundles, this is clearly not feasible in low-income and T A B L E 2 (Continued) F I G U R E 2 Percentage of SSC guideline recommendations that could be implemented at hospitals in Malawi, other low-income countries, and lower-middle-income countries. Abbreviation: SSC, Surviving Sepsis Campaign.

Percentage of recommendations that could be implemented
Component of maternal sepsis management bundle P = 0.005* P = 0.019* P = 0.002* lower-middle-income settings. Previous research 8 has indicated that these bundles need to be adapted for use in low-resource settings.
To be able to initiate care early, the adapted bundles should ideally be deliverable not only in hospital settings but also within communitybased healthcare centers where patients present initially. 21 To improve sustainability and maximize coverage, the development of new lowresource maternal sepsis bundles should be rigorous, have wide involvement from practitioners delivering care in this setting, and be grounded in a realistic assessment of the available resources. 22 The limited and inconsistent supply of resources demonstrated by the present study is probably a reflection of weak systems, with widespread limitations across all aspects of resource provision rather than those related to maternal sepsis only. 23 Quality improvement approaches that aim to address the limitation in resources for the management of maternal sepsis will therefore need to be integrated into, and matched by, wider health system strengthening initiatives and international advocacy efforts. 24,25 The present study has several limitations. The availability of Therefore, the present results cannot be generalized to all sites within these countries, given that there will be variability between the facilities and over time.
In conclusion, implementation of the existing SSC guidelines or bundles is unrealistic in most facilities in Malawi and other lowincome countries because of severe resource limitations; facilities in lower-middle-income countries are more adequately resourced. The present work highlights the urgent need to improve the provision of broad-spectrum antibiotics, the availability of basic monitoring equipment, and the basic hygiene infrastructure in facilities in low-income countries. New maternal sepsis bundles should be developed that take account of the actual availability resources in low-income countries, to make it feasible to implement the bundles in these countries' health centers and hospitals.

AUTHOR CONTRIBUTIONS
MA conducted the primary data collection, analyzed the data, and drafted the manuscript. CM assisted with the data collection and managed the study in Malawi. DL conceived the study. AW, AC, FT, and DL planned and supervised the study. All authors reviewed and edited the manuscript.

SUPPORTING INFORMATION
Additional Supporting Information may be found online in the supporting information tab for this article. Table S1. Characteristics of the healthcare facilities included in the present survey. Table S2. List of responding countries.