Supporting global antimicrobial stewardship: antibiotic prophylaxis for the prevention of surgical site infection in low- and middle-income countries (LMICs): a scoping review and meta-analysis.

Antimicrobial stewardship programmes (ASPs) commonly include the use of antibiotic prophylaxis prior to surgical procedures as a key area for improvement. 1 , 2 In addition to reducing the incidence of infection, appropriate prophylaxis contributes to the control of antimicrobial resistance (AMR). In the UK and other high-income countries, guidance on use of surgical antibiotic prophylaxis (SAP) to reduce surgical site infections (SSIs) has been in place for many years and the recommendation to use a single dose within 1 h of surgery 3 has become routine practice. The ‘Start smart—then focus’ resource 4 provides a useful algorithm for SAP together with a summary of components for best practice. In Scotland, SAP has been the focus of improvement work for the Scottish Antimicrobial Prescribing Group (SAPG) 5 and associated quality indicators have been used to drive improvements in practice. 6 Through research collaborations with teams in African countries and through reviews of global point prevalence studies, we were aware that practice varies considerably from that recommended by WHO. 3 , 7 Recent audits and point prevalence studies in Benin, Botswana, Ghana, Kenya, India, Nigeria, Pakistan, Saudi Arabia and Turkey confirm continued high occurrence of the use of multiple-dose SAP administered for >1 day, which is likely to exacerbate AMR. 8–16 In 2019, SAPG received a Commonwealth Partnerships for Antimicrobial Stewardship grant from the Fleming Fund to support partners in two hospitals in Ghana to develop ASPs. This work included data gathering via point prevalence surveys, 17 development and delivery of multi-professional education aimed at increasing knowledge and optimizing behaviours and supporting quality improvement interventions. Initial intelligence gathering suggested that antibiotic prophylaxis to prevent SSI was suboptimal, with extended durations of antibiotics common. 17 The objective of this scoping review was to inform our approach to supporting change in this area. We reviewed the evidence on the duration of SAP and related incidence of SSI in low- and middle-income countries (LMICs) including staff behaviours and attitudes to provide guidance. We are aware that cultural and contextual issues are important for influencing antimicrobial prescribing in LMICs. 18

This scoping review and meta-analysis is reported in accordance with the Joanna Briggs Institute (JBI) Reviewer’s Manual. 19 Studies that met the following criteria were included: all study designs that reported SAP in LMICs or developing countries as measured by SSI in patients undergoing any surgical procedure and studies that discussed clinicians’ perceptions and practices regarding the use of SAP. The search strategy aimed to find both published and unpublished studies. A systematic search of published literature in MEDLINE, Embase, Cochrane, CINAHL and Google Scholar from inception to 17 February 2020 was undertaken for trials, audits, guidelines and systematic reviews. The search was limited to studies published in English and involving human subjects. Grey literature, websites and reference lists of included studies were also searched for completeness. Key words and index terms were: antibiotics OR antibiotic prophylaxis; surgical site infection; low-income countries OR middle-income countries; Africa OR (individually named countries). The Medline search strategy is provided in Appendix S1 (available as Supplementary data at JAC-AMR Online). Two authors (L.C. and J.S.) screened all titles and abstracts identified. Full text was obtained for all relevant abstracts and those not satisfying the inclusion criteria were removed. All remaining studies were included in this review in accordance with the aim of scoping reviews to determine the extent of literature on the topic and are reported in accordance with JBI scoping review guidelines. 19 Publications selected for critical appraisal were assessed independently by two reviewers (L.C. and J.S.) for methodological validity using the standardized critical appraisal instrument from the JBI Meta-analysis of Statistics Assessment and Review Instrument ( https://joannabriggs.org/sites/default/files/2019-05/JBI_RCTs_Appraisal_tool2017_0.pdf ). Disagreements were resolved by consensus or a third reviewer (R.A.S.). Study data were extracted onto a customized form that included study characteristics, interventions, outcomes and recommendations. Two meta-analyses were conducted using Cochrane Review Manager (RevMan) version 5.3. Primary statistics were extracted from studies to calculate risk ratios for the association of (i) timing or (ii) treatment duration of SAP and incidence of SSI.

The detailed search results are reported in Figure  1 . Of the initial 423 publications identified, full text was retrieved for 68 potentially relevant studies and 51 are included in this review. Flow diagram for study selection. Key characteristics and findings of the studies are summarized in Table  1 . Publications in this review include 3 reviews, 15 randomized clinical trials (RCTs), 25 prospective studies, 5 retrospective studies, 1 cross-sectional study, 1 survey and 1 qualitative study. Included studies WHO 3 2016 (updated 2018) Guideline Global 1. When SAP is indicated, recommend administration prior to surgical incision. Recommend the administration of SAP within 120 min before incision while considering the half-life of the antibiotic. 2. Strong recommendation against the prolongation of SAP administration after completion of the operation for the purpose of preventing SSI. 1. Recommendation was based on systematic review of timing of antibiotic administration prior to incision. 2. Recommendation based on meta-analysis of 44 RCT. Systematic review and meta-analysis International SSI was the leading infection in hospitals in developing countries: 5.6 per 100 surgical procedures. Much higher than in developed countries (USA 2.6%; Europe 2.9%). Call for improvements in surveillance and infection control practices. WHO-funded review. 220 studies included (57 focused on SSI). HAI is poorly recorded in Africa. Systematic review Sub-Saharan Africa Review and synthesis of interventions that had been tested in sub-Saharan Africa to reduce the risk of SSI. 24 studies included from Nigeria ( n  = 8), South Africa ( n  = 5), Côte d’Ivoire ( n  = 2), Kenya ( n  = 2), Tanzania ( n  = 2) Uganda ( n  = 2). One study included from each of Ethiopia, Ghana and Mozambique. Antibiotic prophylaxis was the most common intervention ( n  = 10). Systematic review International RCT Ghana 320 women admitted for Caesarean section were randomized 1:1 to receive either triple therapy of ampicillin 1 g + metronidazole 500 mg + gentamicin 80 mg all given IV after cord clamping and repeated 12 h after first dose (Group 1) or co-amoxiclav 1.2 g given IV after cord clamping and repeated 12 h after first dose (Group 2). SSI were 13.1% in Group 1 and 3.7% in Group 2. RCT Uganda Conference abstract of RCT Rwanda RCT Uganda Common practice at the time of the study was to give multiple doses of antibiotics for 3–7 days post-op. Study concluded that single-dose therapy of metronidazole and ceftriaxone was as effective as multiple doses in the prevention of post-op wound infection and recommends use of single-dose prophylaxis. Cost–benefit RCT Mozambique Women ( n  = 288) with indication for emergency Caesarean section were randomly allotted to two groups. Group 1 ( n  = 143) received the single, combined dose of prophylactic antibiotics (160 mg gentamicin and metronidazole IV given before incision) and Group 2 ( n  = 145) received, over 7 days, the post-op standard scheme of antibiotics (crystalline penicillin 4 000 000 UI IV q6h and metronidazole 500 mg IV q8h during the first 24 h; erythromycin 500 mg q6h orally and metronidazole 500 mg q8h orally for 6 days). Outcome measures: post-op infection, measured at Day 7, mean hospital stay and costs of antibiotics used. Post-op infection rate was 5.2% in Group 1 and 6.4% in Group 2. Single-dose cost US$0.78; post-op scheme cost US$8.37. RCT Sudan RCT Benin RCT Tanzania RCT Sudan RCT Uganda RCT Nigeria RCT Tanzania RCT Nigeria RCT International RCT Nigeria Segmented time-series analysis of effect of improvements Columbia Time-series design Quality improvement intervention Kenya 18 month period of SSI surveillance and introduction of quality improvement intervention. Hospital staff conducted the surveillance daily—diagnosis of SSI within 30 days, antibiotics prescribed. Assessed costs. n  = 3343 patients were followed up (Caesarean section, laparotomy, hysterectomy and hernia repair). Before the intervention 99% were prescribed post-op (penicillin, gentamicin and metronidazole IV for 3–5 days followed by oral antibiotics). Locally developed policy: 2 g ampicillin IV + 500 mg metronidazole IV given pre-op; no antibiotics prescribed post-op. Results: at Week 1, 60% and at Week 6, 98% of patients had pre-op prophylaxis and post-op prophylaxis fell to 40% (Week 1) and 10% (Week 6): highly significant changes. Surveillance showed a modest reduction in the risk of superficial SSI (deep SSI results were non-significant). There is no evidence to support the use of post-op prophylactic antibiotics. Education was conducted at multidisciplinary seminars to develop an antibiotic prophylaxis policy—seminars were based on review of African and international research papers and national policy documents. Financial and time impacts: cost was reduced by US $2.50 per operation and 450 nurse-hours saved per month. This study emphasizes the importance of local engagement and patient education to facilitate change. Prospective pre- and post-intervention study International Prospective Rwanda The clinical guideline was based on WHO recommendations from 2015 that pre-op antibiotics be administered 30–60 min before incision and no post-op antibiotics be given. However, 66.7% of women were given 1 g ceftriaxone within 1 h before incision, almost all received post-op antibiotics. SSI was 10.9% at Day 10 and this was attributed to type of skin cleaning. No association was found with either pre- or post-op antibiotics. Prospective audit Nigeria Study objectives were to evaluate compliance with SAP measures (selection of antibiotic, timing and duration) and to determine the DDD of antibiotic per procedure in obs/gyn surgeries. Data were collected from patients’ notes in three hospitals N  = 248. Results Antibiotic selection: nitroimidazole (86.2%),  β-lactam/β-lactamase inhibitor (73.4%), third-generation (30.2%) and second-generation (22.6%) cephalosporins were the most frequent antibiotics prescribed for surgical prophylaxis. Timing : pre-op antibiotic was administered within 60 min before incision in 41 (16.5%) procedures (usually at induction of anaesthesia). Timing of antibiotic prophylaxis was inappropriate in >80% of the procedures. Duration : duration of antibiotic prophylaxis was prolonged in all the procedures (range 5–12 days). In ∼99% of cases, antibiotics were prescribed for ≥7 days. The mean duration of antibiotic prophylaxis was 8.7 ± 1.0 days. DDD : SAP was 1675.3 DDD per 100 procedures (16.75 DDD per procedure). 2/3 hospitals in this study did have infection control teams, but none had a pharmacist to review orders for SAP. The findings of this study highlighted the need for antimicrobial stewardship to reduce the excessive use of antibiotics; inappropriate prescribing of antibiotic prophylaxis may be due to poor knowledge regarding the spectrum of antibiotic activity; non-compliance with timing may be due to lack of a protocol or lack of knowledge regarding optimal timing of antibiotic prophylaxis; extended duration of antibiotic prophylaxis may have occurred as obstetricians and gynaecologists demonstrated lack of belief in single-dose antibiotic prophylaxis, citing poor infection control. There was a misconception that long duration of antibiotic prophylaxis would reduce risk of SSI. Prospective pre- and post-intervention study Nigeria Prevalence of SSI in Nigeria 9.1%–30.1%. SSI affects morbidity, mortality and cost. Cost for patients with SSI is double. SAP practice of obstetricians and gynaecologists in Nigeria is not compliant with guidelines. This study evaluated the impact of antibiotic stewardship interventions on prescribing (compliance, choice, timing and duration of antibiotic prophylaxis and antibiotic utilization), clinical (SSI rate) and economic (costs of antibiotic prophylaxis) outcomes. Compliance with timing of antibiotic prophylaxis increased from 14.2% to 43.3%. Compliance with duration increased from 0% to 29.2%. Prescription of third-generation cephalosporin was reduced from 29.2% to 20.6%. Antibiotic utilization significantly decreased and mean cost of SAP was reduced by US $4.20. SSI rates were recorded as 4% pre-intervention and 3.4% post-intervention; however, these measures were limited to period of hospitalization only. Antibiotic stewardship interventions in this study included: development and dissemination of a departmental protocol for SAP; educational meeting with the obstetricians and gynaecologists; audit and feedback using baseline data and reminder in the form of wall-mounted posters. In both hospitals, the protocol was developed by a team that comprised four to five consultant obstetricians and gynaecologists and a clinical pharmacist. The protocol presented by the clinical pharmacist recommended: antibiotic prophylaxis administered 60 min before incision and discontinued within 24 h after surgery and type of antibiotic to be used. The educational session focused on the principles of SAP for obs/gyn surgery and the data collected during pre-intervention period highlighting areas where practice did not align with guidelines. Prospective audit and feedback South Africa Study included 34 urban and rural South African hospitals. The aim of the study was to promote multidisciplinary, collaborative action of pharmacist-driven audit and feedback improvement model and to achieve a sustainable reduction in SSI. Started by multidisciplinary team working to achieve consensus on a PAP guideline. Defined four process measure and indicators: antibiotic choice, dose, timing of administration and duration. Workshops to introduce model to surgeons, anaesthetists and nurses. Learning cycles at 8–10 week intervals. 4 week survey of compliance preintervention. Results Pre-intervention : an antimicrobial was administered  to 34.7% (95% CI 31.7%–37.7%) of patients within 1 h before incision. Antimicrobial agents consistent with the guideline were administered to 81.2% (95% CI 78.5%–83.8%) of the patients and at the recommended dose in 70.5% (95% CI 67.1%–73.9%). Antimicrobial prophylaxis was limited to one dose or discontinued within 24 h of the end of surgery in 80.8% (95% CI 79.0%–82.5%). The mean SSI rate for the hospital group was 2.46 (95% CI 2.18–2.73). Post-intervention : Timely administration occurred in 56.4% (95% CI 53.1%–59.6%) ( P  < 0.0001), antibiotic choice consistent with the guideline was administered in 95.9% (95% CI 89.9%–100%) ( P  = 0.0004), at the recommended dose in 87.0% (95% CI 81.3%–92.8%) ( P  = 0.0002) and the duration of PAP was appropriate for 93.9% (95% CI 88.1–99.6) ( P  = 0.0005). SSI rate 1.97 (95% CI 1.79–2.15) ( P  = 0.0029). These authors question the effectiveness of previous interventions aimed at improving adherence to antimicrobial prophylaxis guidelines. Emphasized the need to develop effective teams and coordinate processes to institutionalize new approaches in order to make them sustainable. Before and after cohort study 5 hospitals in Kenya, Uganda, Zambia and Zimbabwe Before and after study Egypt Study included five tertiary acute care surgical hospitals (herniorrhaphy, colectomy, joint replacement, spinal fusion, obs/gyn). 6 month intervention aiming to improve timing of the first dose before surgery (at least one dose administered within 60 min before incision) and the duration (no longer than 24 h) of therapy. Surgeons and anaesthetists were targeted and programme was established within the hospitals’ infection control teams. International guidelines formed basis of education activities. Posters were used to remind prescribers of the optimal timing and duration for antibiotic prophylaxis. Results: post-intervention, 41.6% of patients received optimal antimicrobial prophylaxis and there was a significant reduction in amount of antimicrobials used. Few hospitals in Egypt have policies or guidelines on antimicrobial use either in general or specifically for surgical prophylaxis. After the intervention, surgeons were less compliant with duration than timing. Authors attribute this finding to poor awareness of AMR, surgeons’ resistance to changing routine practices and strong belief that hospitals in Egypt are different in terms of increased contamination. Beliefs should be explored and discussed. Before and after study Tanzania Routine ampicillin (88% post incision + continued for 5 days) pre-intervention (commonly chloramphenicol 60%, aminopenicillins 23% and benzylpenicillins 15%; 118/527 patients received more than one agent). Intervention: 2.2 g amoxicillin/clavulanate given IV 10–30 min before incision (target was 30 min before incision but not always achieved). N  = 527 pre-intervention group, n  = 114 (21.6%) developed SSI. N  = 276 intervention group, n  = 11 (4%) developed SSI. Authors concluded timing was less crucial as long as antibiotic was administered before incision. Poster presentation Before and after study Kenya Prospective cross-sectional study Sudan Patients in obs/gyn SSI detected by surveillance during admission and by structured telephone call post-discharge until Day 28. SAP given to 98.8% of patients in the operating room. Time of first dose was proper (30–60 min before incision) in 11.6%, late (1–29 min before incision) in 58.8% and too late (after incision) in 24.4%. All patients had post-op prophylaxis prescribed (average 8 days). SSI rate was 7.8% overall. Prospective cohort study Ethiopia Prospective observational study Tanzania Prospective observational survey Tanzania Prospective observational study Ethiopia Prospective cohort study Tanzania Prospective study Botswana Aim was to describe current SSI burden and antibiotic surgical patterns. Emergency and elective surgery included (laparotomy, appendectomy, excisions and mastectomy). 73.3% of patients prescribed antibiotics: 15% pre-op (majority continued these post-op), 58.3% post-op and 26.8% no antibiotics. Post-op antibiotics were started for suspected infections in patients with peritonitis ( n  = 8), abscesses ( n  = 4) and appendicitis ( n  = 50). Mean duration of post-op antibiotics was 5 ± 2 days. Drugs used: cefotaxime (80.7%), metronidazole (63.5%), cefradine (13.6%) and amoxicillin/clavulanate (11.6%). Co-administration of cefotaxime and metronidazole was found in 134 (48.6%) patients post-op. Oral antibiotics cloxacillin, amoxicillin, doxycycline and metronidazole were also prescribed. SSI rate = 8.5% in emergency and 9.9% in elective. 88.2% of patient with culture-positive SSI had not received pre-op antibiotics. Prospective study Ethiopia Prospective study Zimbabwe Prospective study Sub-Saharan Africa Cross-sectional comparative study Nigeria Aim of the study was to compare adherence to WHO guidelines in two hospitals. Hospital A followed government directive to follow WHO guidelines and had a multidisciplinary responsibility for education and adherence to the policy. Hospital B did not have an infection control committee or a policy. Data were collected from staff caring for women having Caesarean section using a semi-structured questionnaire and observations. The appropriate timing of the administration of prophylactic antibiotics (intra-op administration) was observed by 57.58% of the respondents in A, compared with 22.86% in B ( P  = 0.00). Emphasize the importance of a committee responsible for surveillance and education as well as regular supply of necessary antiseptics and consumables. Also need for a dedicated infection control team. Clinicians in Hospital A attributed their non-compliance to poor supervision by the infection control team and lack of training. Hospital B: lack of training and absence of a hospital policy on infection control. Observational study Ethiopia Observational study Kenya Primary focus was timing of PAP. Outcome measure was SSI. Compared rates of SSI in two hospitals in Nairobi. Hospital A provided antimicrobial prophylaxis prior to incision for all patients and Hospital B provided only post-op prophylaxis to all patients. Results: the SSI rate was 4.0% (12/299; 11 superficial SSI, 1 deep SSI) at Hospital A and 9.3% (28/301; 18 superficial SSI, 7 deep SSI, 3 organ/space SSI) at Hospital B. Multicentre prospective cohort study International Primary outcome measure was the 30 day SSI incidence comparing high-, middle- and low-HDI countries. Data were collected on the incidence and length of antimicrobial treatment surgery. Results: the unadjusted SSI incidence high HDI ( n  = 691/7339 [9.4%] patients), middle HDI ( n  = 549/3918 [14.0%] patients) and low HDI ( n  = 298/1282 [23.2%] patients). Administration of pre-op or prophylactic antibiotics, or both, was higher in groups with low HDI ( n  = 1224/1282 [95.5%]) than in countries with middle HDI ( n  = 3392/3918 [86.6%] patients) and high HDI ( n =6446/7339 [87.8%] patients). Patients in LMICs were more likely to receive post-op antibiotics than those in high-HDI countries (3376/7339 [46.0%] patients in high-HDI countries versus 3135/3918 [80.0%] patients in middle-HDI countries versus 1098/1282 [85.6%] patients in low-HDI countries; P <0.001). Courses of post-op antibiotics were longer in patients in LMICs than in high-income countries, with the number of patients receiving antibiotics for 5 days or more increasing from countries with high HDI to low HDI [1830/7339 (24.9%) patients in high-HDI countries versus 1837/3918 (46.9%) patients in middle-HDI countries versus 650/1282 (50.7%) patients in low-HDI countries; P <0.001]. Intervention India Point prevalence survey Botswana Point prevalence survey Kenya Point prevalence survey Benin 39/45 hospitals participated: n  = 3130 patients. SSI prevalence was 24.7%. 64.6% of patients surveyed were treated with antibiotics: 30% were non-infected. 40.8% self-medications. Point prevalence survey Ghana Retrospective chart review South Africa Chart review in a teaching hospital n  = 112 and a private hospital n  = 112. Prescription of antimicrobials compared with current SAP guidelines. Criteria: appropriate antimicrobial selection, dosing, timing of administration, re-dosing and duration of treatment. Teaching hospital: 77.3% received SAP when indicated, 21.1% when not indicated. Met 3/5 criteria on 58.8% of situations. Private: 100% received SAP when indicated, 45.9% when not indicated. Retrospective chart review Zimbabwe Retrospective case series Gambia Retrospective cross-sectional study Addis Ababa Retrospective study Ethiopia Survey Ghana Survey to assess the perceptions and practices of surgeons in Ghana regarding the use of antibiotics for groin hernia repair in relation to evidence-based guidelines. Structured questionnaire: 117/146 (80%) response rate. 62% no antibiotics, 10% pre-op, 15% pre- and post-op antibiotics when no mesh used. 53% pre- and post-op antibiotics when mesh was used. Previous published rates of SSI for groin hernia surgery in Ghana 1%–3% thus low risk—recommendations state no need for antibiotic prophylaxis. 25% practice inconsistent with evidence-based guidelines—55% in mesh repairs. Most common reason was concern for SSI. Authors advise other measures of infection control such as bathing and strict aseptic technique. Qualitative study Sub-Saharan Africa (Kenya, Uganda, Zambia, Zimbabwe) A central facilitator to implementation was the establishment of local multidisciplinary teams: they actively included other stakeholders to promote wider ownership of change and became a driving force and helped overcome barriers. Results related to optimization of antibiotic prophylaxis: regular feedback of data on SSI surveillance and compliance with SSI preventative measures was important to create a tension for change and to sustain engagement with the project: it was important to show staff that SSI rate was reducing at the same time as the number of patients receiving post-op antibiotics was reducing. HDI, human development index; IM, intramuscular; intra-op, intra-operative; obs/gyn, obstetrics/gynaecology; PAP, perioperative antimicrobial prophylaxis; post-op, post-operative; pre-op, pre-operative; RR, risk ratio; SUSP, Surgical Unit-based Safety Programme. Ten studies were critically appraised prior to inclusion in the meta-analysis. Five studies 20–24 compared pre- and post-operative SAP. Five studies 25–29 compared short- and long-term SAP. Overall quality scores ranged from 8 to 12 out of 13; the risk of selection bias was deemed high in 4 studies 21 , 24–26 and the risk of detection bias was deemed high in 5 studies. 23–26 , 28 The results of the critical appraisal process are presented in Table  2 . Critical appraisal scores of selected studies Publications were assessed using the JBI Meta-analysis of Statistics Assessment and Review Instrument. Y, yes; N, no; UC, unclear. Several studies in this review, including a recent international cohort study, confirmed that the incidence of SSI is significantly higher in LMICs than in higher-income countries. 30–37 SSIs have been reported as the leading hospital-acquired infection (HAI) in LMICs, 38 , 39 with highly variable reported rates, e.g. 8.5%–23.2% following gastrointestinal surgery and one study reported an incidence of 85.1% for SSI following Caesarean section. 20 However, surveillance data regarding HAI are generally poorly recorded across LMICs. 32 , 38 , 40–43 Patients in LMICs are more likely to receive post-operative antibiotics with treatment lasting for more than 5 days than patients in high-income countries. 21 , 22 The WHO guidelines on SAP are not always adhered to and local guidelines are not always available or followed. 34 , 35 , 40 , 44–52 Studies reported low compliance with recommended timing of SAP and patients being given antibiotics for long periods post-operatively or when they were not indicated. 12 , 32 , 42–44 , 46 , 48 , 51 , 53–55 Despite a lack of evidence to support the use of post-operative antibiotic prophylaxis, many surgeons in developing countries continue to use post-operative antibiotic prophylaxis. 12 , 30 , 32 , 51 , 53 , 56 , 57 Non-compliance with WHO guidelines may be due to poor knowledge of spectrum of antibiotic activity and of optimal timing of antibiotic prophylaxis. 12 , 45 Clinicians cite lack of an infection control committee and lack of policy and education 40 , 58 as well as concerns about SSI 59 , 60 as factors in non-compliance. A recent systematic review advocated the use of multifaceted strategies including staff engagement and education, as well as evaluation of the success of interventions in reducing SSI. 61 Several individual projects in LMICs that aim to reduce the incidence of SSI have included compliance with WHO guidelines on SAP as part of a wider programme, e.g. WHO surgical safety checklist, pre-operative bathing, avoiding hair removal, optimal surgical hand and skin preparation using locally produced alcohol-based products and improving operating room discipline. 31 , 62–64 Most of the individual projects demonstrated success in reducing the incidence of SSI, with the longest follow-up being 18 months. 64 Interventions to promote the correct use of SAP have been shown to be successful in reducing incidence of SSI, even in low-resource settings. 30 , 31 , 50 , 65 Several clinical trials assessed different SAP regimens; however, even in recent studies not all of the treatment arms complied with WHO guidelines regarding single dose and timing of administration. Details for individual studies are reported in Table 1 . The majority of RCTs were conducted in women having elective and/or emergency Caesarean section; 20–29 other studies investigated SAP in patients having surgical treatment for hernias, 24 , 66 transrectal prostate biopsy 67 and clean day-case surgery for children. 68 Five studies investigated the efficacy of single-dose pre-incision SAP versus single-dose or prolonged post-incision treatment in patients undergoing emergency or elective Caesarean section 20–24 and were included in a meta-analysis. Six studies compared a short duration of SAP to a long duration of SAP with patients in both groups receiving a first dose pre-incision and five were conducted in patients undergoing emergency or elective Caesarean section 25–29 and were included in a meta-analysis. One study that assessed SAP in patients undergoing transrectal biopsy was not included in the meta-analysis due to heterogeneity of the study population. 67 Other studies compared different types of antibiotics 69 and single-dose SAP with placebo. 66 , 68 , 70 Studies consistently reported either a reduction or no change in the incidence of SSI between groups, thereby supporting the use of single-dose pre-incision or short-duration SAP in LMICs. The random-effects pooled results across five studies ( n  =   1268 participants) showed the risk ratio for the development of SSI was 0.77 (95% CI: 0.51–1.17) ( P  =   0.23) for the single-dose pre-incision group versus the single- or multiple-dose post-incision group ( Figure 2 ). Heterogeneity was moderate (Tau2   =   0.09, I 2   =   43%). The random-effects pooled results across five studies ( n  =   1377 participants) showed the risk ratio for the development of SSI was 0.89 (95% CI: 0.55–1.14) ( P  =   0.63) for the short-duration group versus the long-duration group ( Figure 3 ). Heterogeneity was low (Tau2   =   0.00, I 2   =   0%). Single-dose pre-incision versus single/multiple post-incision SAP. Short-duration versus long-duration SAP. Several studies reported on the difference in the cost between SAP regimens, with mixed findings depending on type of antibiotic given. Single-dose pre-operative prophylaxis was shown to cost a tenth of post-operative prophylaxis, 21 reducing costs by US $2.50 56 or $4.20 per procedure. 34 There was overall agreement that correct use of SAP would reduce the costs for patients or institutions. 26 , 27 , 29 , 30 , 57 , 70 Studies also mentioned greater compliance with regimen and nurse-time saving as a result of a reduced need to administer IV antibiotics for lengthy post-operative periods. 25 , 26 It was proposed that emphasizing the potential cost savings would promote the engagement of hospital managers who have the power to drive changes such as production of local evidence-based guidelines and interventions to reduce SSI. 56 Studies reported that non-compliance with guidelines occurred for a variety of reasons. Several studies found that clinicians believed that extended duration of post-operative antibiotic prophylaxis is needed due to poor hygiene, poor infection control and overcrowding in hospitals. 12 , 30 , 51 There was also widespread belief that a long duration of antibiotic prophylaxis would reduce the risk of SSI and lack of belief in the efficacy of single-dose prophylaxis. 12 Patient expectation of longer duration of SAP may also be an important factor. 56 A qualitative study showed that at the same time as the number of patients on post-operative antibiotics was reducing, the SSI rate was also reducing, which provided proof of impact and encouraged compliance with new approaches. 71 In order to be successful, education interventions must be relevant to local circumstances and promote the development of effective local teams. 31 , 34 , 50 The process should encourage clinicians to discuss their beliefs, emphasize the importance of surveillance and involve them in the development of local guidelines that are consistent with evidence and WHO guidelines. 56 Education of patients may also reduce their expectation of extended use of SAP.

The aim of this review was to consider the effectiveness of SAP in the prevention of SSI in LMICs to support an education intervention in Ghana specifically but acknowledge there would be implications for practice in other LMICs. Clinicians in LMICs undoubtedly face environmental challenges that are exacerbated by scarce resources; however, studies in this review have demonstrated that improvements in practice can be successful in low-resource environments. Even though the sizes of absolute effect were small (risk ratio of 0.77 and 0.89) and not statistically significant, single-dose and short-duration SAP in Caesarean section was associated with lower risk of SSI. Although this cannot be directly extrapolated to other procedures it provides support for adherence to local, national and international guidelines on SAP, including single-dose and short-duration SAP. This can contribute significantly to reduction of SSI and control of AMR, as well as conserving resources and potentially reducing duration of hospitalization. Conversely, inappropriate post-operative antibiotic prophylaxis has been shown to increase costs and promote AMR. Additionally, this review supports the need for improvement in LMICs across all core components of infection control identified by the WHO. Staff education was highlighted in several studies as a key element to build infection surveillance and control skills in LMICs. An audit of SAP in Nigeria suggested that poor knowledge regarding the spectrum of antibiotic activity may lead to inappropriate prescribing 12 and an implementation study in Kenya found limited awareness of national policy documents, poor access to appropriate medicines and lack of awareness of the potential for cost savings. 56 Successful programmes to improve antimicrobial stewardship have identified the establishment of local multidisciplinary and cross-departmental teams to promote ownership of change. All ASPs require high-level support including resources, which can be a concern in LMICs. 72 Consequently, hospital managers must be made aware of the potential for cost and time saving with appropriate SAP as well as a potential reduction in AMR and its future consequences on costs as well as on morbidity and mortality. 73 , 74 Other essential factors include the development of a local SAP policy informed by review of research and guidelines, as local cultural and contextual issues are important for successful programmes to improve future antibiotic use, 18 and the establishment of a robust method of collection of process and outcome measurement data such as SSI surveillance and monitoring of compliance with local policy. Regularly sharing local data with clinical teams is essential to engage the workforce as well as identifying local champions to become the driving force to overcome barriers to change and sustain new ways of working. In order to address the common misconceptions, education programmes should include interactive workshops to explore clinicians’ beliefs and perceptions of factors that contribute to increased risk of SSI. Education sessions should also be used to raise awareness of national policy guidelines and relevant international research findings. Education of patients in terms of their expectations regarding SAP is also important. Patients in LMICs may believe that receiving more medicines is better for outcomes or for achieving value from national insurance schemes. Studies indicated that posters were a useful resource for the education of both clinicians and patients, and also served to reinforce key messages. 56 This literature review provided useful evidence to support compliance with WHO recommendations around using a single dose of surgical prophylaxis to prevent SSI in LMICs. We have utilized the findings to develop and deliver multi-professional interactive learning sessions in two hospitals in Ghana to share the evidence and promote compliance with WHO and national guidelines. Both hospitals, with the help of their AMS teams, are developing evidence-based local SAP policies for different surgical subspecialties and this review will help to guide this work. Local policies are being promoted to clinical staff through posters in wards and theatres as well as being included in ongoing training as part of their ASPs. We hope this work will be of interest as other countries seek to improve their use of antimicrobials to reduce high AMR rates and implement ASPs to optimize SAP.

Click here for additional data file.