Knowledge,attitudes and practices on antimicrobial resistance among healthcare workers in sub-Saharan Africa: a systematic review and meta-analysis

Abstract

Background:

Antimicrobial resistance (AMR) poses a significant threat to public health in sub-Saharan Africa (SSA). Healthcare workers (HCWs) are critical to preventing and controlling AMR.

Objectives:

This systematic review and meta-analysis aimed to synthesise evidence on the knowledge, attitude and practices (KAP) of HCWs regarding AMR in SSA.

Design:

Systematic review and meta-analysis.

Data sources and methods:

We conducted a comprehensive search of multiple databases (PubMed, Web of Science, APA PschINFO and CINAHL via EBSCOhost) for studies published between 2014 and 2024. Studies reporting KAP of HCWs regarding AMR in SSA were eligible. MeSH search terms and related keywords were used. Data extraction was conducted individually by each team member and then cross-verified to ensure accuracy. Quality appraisal was conducted using the Critical Appraisal Skills Programme Checklist to evaluate the methodological rigour of the included studies. Meta-analysis was performed using a random-effects model. All statistical data analyses were performed in the R software (version 4.5.0) using both the meta and metafor packages.

Results:

Thirty-one studies involving 8873 HCWs were included. The pooled proportions of AMR knowledge, attitude and practices were 65.4% (95% CI: 55.7–74.5), 69.2% (95% CI: 58.1–79.3) and 54.1% (95% CI: 44.1–63.9), respectively. Doctors had the highest knowledge (65.1%) yet poor practices (42.8%), while pharmacists and nurses showed lower knowledge and inconsistent practices. Despite favourable attitudes, especially among nurses, behavioural implementation remained weak, underscoring substantial knowledge–practice gaps across professional groups.

Conclusion:

HCWs demonstrated moderate knowledge and attitudes but poor practices, revealing a persistent knowledge–practice gap. Targeted, context-specific interventions and strengthened professional training are urgently needed to enhance evidence-based practice and improve healthcare quality and patient outcomes.

Trial registration:

The review protocol was prospectively registered with PROSPERO, CRD42024579356.

Keywords

antimicrobial resistance attitude healthcare workers knowledge practices sub-Saharan Africa

Introduction

Antimicrobial resistance (AMR) continues to pose a critical global health threat that undermines decades of progress in infectious disease management. The burden of AMR is particularly acute in sub-Saharan Africa (SSA),¹ where infectious diseases such as malaria, HIV, tuberculosis, diarrheal diseases and respiratory tract infections remain leading causes of morbidity and mortality.^2,3 Weak health systems, poor laboratory capacity, unregulated access to antimicrobials and limited infection prevention and control (IPC) infrastructures accelerate the emergence and spread of resistant pathogens in the region.^4,5 Compounding this issue is the high prevalence of Hospital-Acquired Infections (HAIs) in SSA, estimated at 12.9% (95% CI: 8.9–17.4).⁶ The implications are severe: higher treatment failures, increased mortality, prolonged hospitalisations and escalating healthcare costs, which collectively strain already fragile health systems and compromise progress towards universal health coverage.⁷

Healthcare workers (HCWs) are at the frontline of AMR containment. Their knowledge, attitudes and practices (KAP) directly influence antibiotic prescribing, patient education and adherence to IPC measures.⁸ Rational antimicrobial use and effective stewardship programs depend heavily on HCWs’ awareness of AMR risks, access to guidelines and ability to implement best practices.^9,10 In SSA, HCWs frequently work under demanding circumstances characterised by inadequate training, excessive workloads and limited resources,¹¹ which collectively undermine compliance with antimicrobial stewardship protocols.

Over the past decade, several studies across SSA have investigated HCWs’ KAP regarding AMR. However, findings remain fragmented and heterogeneous. For instance, while some studies demonstrate high levels of awareness about the global threat of AMR, they also reveal gaps in knowledge of local policies or treatment guidelines.^10,12 Similarly, while HCWs may express positive attitudes towards prudent antibiotic use, these attitudes often fail to translate into practice due to factors such as diagnostic uncertainty, patient demand, or systemic barriers.^13,14 Comparative evidence across settings further highlights inconsistencies in KAP, with substantial variation by profession, level of training and institutional support.¹⁰

In this review, knowledge refers to HCWs’ understanding of AMR, including its causes, consequences and prevention strategies. Attitude denotes their beliefs, perceptions and willingness to engage in antimicrobial stewardship and infection control measures. Practice represents the actual application of this knowledge and attitude in clinical settings—such as appropriate antibiotic prescribing, adherence to infection prevention protocols and rational antimicrobial use.¹⁵

Understanding HCWs’ KAP in SSA is therefore fundamental for informing interventions to strengthen antimicrobial stewardship and IPC in the region. Against this backdrop, this study aimed to systematically review and quantitatively synthesise existing evidence on the KAP of HCWs regarding AMR in SSA. By quantitatively pooling available data, this study provides a more robust and generalisable estimate of KAP levels among HCWs. The innovation of this study lies in its ability to transform disparate and context-specific findings into consolidated evidence that can inform regional AMR action plans, strengthen HCW training and contribute to global AMR containment strategies.

Methods

Search strategy

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.¹⁶ A comprehensive literature search of PubMed, APA PsycINFO and CINAHL via EBSCOhost was performed using the MeSH terms ‘antimicrobial resistance’ AND ‘healthcare workers’ AND ‘knowledge’ AND ‘attitude’ AND ‘practice’ AND ‘Sub-Saharan Africa’ from 2014 to 2024. Our first database search was undertaken on February 20, 2025, with a subsequent update on July 30, 2025.

Eligibility criteria

Studies were eligible for inclusion if they assessed HCWs’ KAP regarding AMR, antibiotic use, or antimicrobial stewardship and were conducted in SSA. Eligible participants included licensed healthcare professionals such as physicians, nurses, pharmacists, laboratory scientists, dentists and allied health staff. Both observational (cross-sectional, cohort) and interventional (quasi-experimental or randomised) studies reporting quantitative or qualitative KAP outcomes were included. No language restrictions were applied.

Studies were excluded if they focused on non-healthcare populations (e.g. patients, students, community members), lacked explicit assessment of AMR-related KAP, or did not report extractable quantitative or qualitative outcomes. Editorials, commentaries, reviews and case reports were excluded, as were studies with inadequate methodological clarity or insufficient data for meta-analysis. Studies conducted outside of the SSA were excluded.

In all included studies, the levels of KAP towards AMR among HCWs were assessed based on participants’ responses to structured questionnaires or surveys.

The review protocol was prospectively registered with PROSPERO (registration number: CRD42024579356).

Study selection and data extraction

A standardised and pilot-tested data extraction form was developed to ensure systematic and consistent collection of information from all included studies. The form captured comprehensive details including bibliographic information (first author, year of publication and country), study design and setting, participant characteristics (eligibility criteria, sample size, age, sex distribution and baseline features), and all reported outcome measures.

Data extraction was performed independently by two reviewers using the predesigned form, and all extracted entries were cross-checked for accuracy and completeness. Any discrepancies between reviewers were resolved through discussion, and unresolved disagreements were adjudicated by a third senior reviewer.

For item-level missing data, that is, individual KAP items (e.g. missing responses within a questionnaire), the handling followed the approach reported by the original investigators. If the study provided details on how nonresponses or ‘don’t know’ answers were treated (e.g. excluded from denominators, coded as neutral), these conventions were adopted. When no clarification was available, missing responses were assumed to be excluded from denominators, and results were recalculated based on reported valid responses. If only proportions or percentages were provided without absolute numbers, missing numerators or denominators were back-calculated where possible. When data for one or more KAP domains (e.g. attitudes) were missing but others were available, analyses were conducted using the available domains, and the missing component was excluded from the pooled synthesis. The implications of this selective reporting were explored through sensitivity analyses. To further enhance the accuracy and reliability of our findings, we contacted seven authors to seek clarification on their results. Feedback was received from four authors, which was incorporated into our analysis.

Outcome measures

The primary outcomes were the pooled proportions of HCWs with good knowledge, positive attitude and good practice towards AMR. Secondary outcomes included subgroup analyses by profession (e.g. doctors, nurses, pharmacists).

Data analysis

To quantify HCWs’ KAP regarding AMR, we calculated the proportion of KAP cases for each study by dividing the number of KAP cases by the total number of participants. To account for heterogeneity across studies due to varying study designs and HCW characteristics, we employed random-effects models. Additionally, for the meta-analysis, we calculated individual sample proportion estimates with 95% confidence intervals (CIs) using stabilised variances via the Freeman–Tukey double arcsine transformation.¹⁷ This transformation enabled approximation of normal distribution, facilitating robust meta-analytic estimates.

Random-effects models were fitted using the restricted maximum-likelihood estimation, to quantify variance heterogeneity. The 95% CIs of summary measures were calculated using the Knapp–Hartung variance estimator. Heterogeneity was assessed using the I² statistic, with I² ⩾ 75% indicating substantial heterogeneity.¹⁸ To evaluate publication bias, we used funnel plots, Kendall’s tau (τ) rank correlation test and Egger’s test, where applicable (i.e. outcomes reported by ⩾10 studies).¹⁹

In our review, the team systematically assessed studies reporting on KAP related to AMR. Where AMR-specific KAP measures were explicitly stated, these data were directly extracted. In studies where the term ‘KAP’ was not used, we identified the variables that most closely aligned with the conceptual domains of knowledge, attitudes, or practices on AMR and extracted these accordingly. The meta-analysis was conducted using these harmonised data.

Numerical values represent the proportion of respondents who provided correct answers (knowledge), held appropriate or favourable views (attitudes) or reported recommended behaviours (practices) related to AMR. For studies that reported composite KAP scores, we extracted or calculated percentage scores based on the number of correct or favourable responses relative to the total items assessed in each domain. Where only raw counts were provided, these were converted into percentage values to ensure comparability across studies. After pooling data, we applied the modified Bloom’s cut-off criteria to classify AMR-related KAP levels, categorising scores as good (80%–100%), moderate (50%–79%) or poor (<50%).²⁰ To illustrate KAP scoring approaches, the questionnaire used by Chukwu et al.21 was consulted as a reference, alongside instruments from other included studies where applicable. All statistical analyses were performed using R software (version 4.5.0) with the meta and metafor packages.

Quality appraisal

To evaluate the methodological quality of the included studies, we employed the Critical Appraisal Skills Programme (CASP) checklist (Supplemental File 1). This involved assessing each study’s rigour in areas such as research design, participant recruitment, data analysis and reporting. Studies were categorised as high, moderate, or low quality based on the proportion of ‘Yes’ responses to the CASP criteria. A score of ⩾75% denoted high quality, indicating strong methodological rigour and minimal risk of bias. Scores between 50% and 74% were classified as moderate quality, reflecting some methodological limitations but overall acceptable validity. Studies scoring <50% were considered low quality, suggesting major methodological flaws and a high risk of bias.

Results

Characteristics of the included studies

The literature search yielded 1102 records, with 771 remaining after duplicates were removed. After screening titles and abstracts, 55 full-text articles were reviewed, and 31 studies^{10,12,13,21,48} were included in the systematic review and meta-analysis. Studies from 13 SSA countries were eligible, including Benin, Cameroon, Ethiopia, Ghana, Nigeria, Sierra Leone, Somalia, South Africa, Sudan, Togo (in conjunction with Ivory Coast), Uganda and Zambia, with the following number of studies per country: Benin (n = 1), Cameroon (n = 1), Ethiopia (n = 5), Ghana (n = 4), Nigeria (n = 5), Sierra Leone (n = 2), Somalia (n = 1), South Africa (n = 5), Sudan (n = 1), Togo (n = 1, jointly with Ivory Coast), Uganda (n = 2) and Zambia (n = 3). A PRISMA flow diagram illustrating the study selection process, from initial screening to final inclusion based on the predefined eligibility criteria, is presented in Figure 1.

Figure 1.

PRISMA flow diagram of the systematic search and selection process.

The characteristics of the included studies conducted between 2014 and 2024 are summarised in Supplemental File 2, Table 1. The majority of studies focused on HCWs in three professional categories: doctors (n = 11), pharmacists (n = 10) and nurses (n = 6). A detailed breakdown of the KAP of HCWs regarding AMR is presented in Supplemental File 2, Table 2.

Of the 31 studies appraised using the CASP cohort checklist, 22 were rated as high quality, 7 as moderate quality and 2 as low quality. A recurring methodological limitation was the inadequate identification and adjustment for confounding factors during analysis.

A comprehensive analysis of 8873 HCWs revealed that 49.2% (4364) demonstrated knowledge of AMR, while 43.5% (3857) and 18.6% (1647) exhibited positive attitudes and practices towards AMR, respectively (Table 1).

Table 1.

Summary proportion of characteristic dimensions of AMR KAP in SSA from 2014 to 2024.

Professional group	N	n	p (%)	EP	95% CI
Overall
Knowledge	8873	4364	49.2	65.4	55.7–74.5
Attitudes	8873	3857	43.5	69.2	58.1–79.3
Practices	8873	1647	18.6	54.1	44.1–63.9
Doctors
Knowledge	2239	1292	57.7	65.1	46.4–81.7
Attitudes	2239	136	6.1	79.8	32.1–99.9
Practices	2239	65	2.9	42.8	7.9–82.4
Pharmacists
Knowledge	1631	540	33.1	56.3	33.9–77.5
Attitudes	1631	154	9.4	76.6	51.8–94.6
Practices	1631	464	28.4	38.9	16.1–64.5
Nurses
Knowledge	1148	341	29.7	28.9	16.4–43.2
Attitudes	1148	308	26.8	88.4	82.6–93.2
Practices	1148	211	18.4	27.8	11.7–47.6

AMR, antimicrobial resistance; CI, confidence interval; EP, estimated proportion in percentages from meta-analysis; KAP, knowledge, attitude and practices; n, number of cases; N, sample size; p (%), proportion in percentages (n/N); SSA, sub-Saharan Africa.

Temporal trends indicated that HCWs’ knowledge of AMR peaked in 2014, 2017, 2020 and 2022, with proportion rates ranging from 72% to 77%, and a maximum of 77% observed in 2020 (Figure 2). Conversely, practice towards AMR was highest in 2020 (50%) and lowest in 2021 (9%) (Supplemental File 2, Table 3).

Figure 2.

Trends in overall KAP of AMR in SSA between 2014 and 2024. Blank spaces show absent data of required variable in a particular year.

Profession-specific disparities emerged, with doctors exhibiting higher knowledge of AMR (58%), whereas pharmacists and nurses demonstrated more favourable attitudes (28%) and practices (27%), respectively (Figure 3). Among the doctors, higher knowledge of AMR (82%) was exhibited in 2014, while more attitude (100%) was observed in 2020 (Supplemental File 2, Figure 1). Amidst the HCWs who were pharmacists, higher knowledge of AMR (91%) was exhibited in 2020, while more attitude (100%) and practice (60%) were observed in 2020 and 2024, respectively (Supplemental File 2, Figure 2). However, more attitude (91%) and practice (31%) were exhibited in 2020 and 2023, respectively, among the nurses, while higher knowledge (48%) of AMR was observed in 2020 (Supplemental File 2, Figure 3). Overall, the doctors had a good knowledge of AMR (63%) but a moderate attitude (50%) and poor practice (23%). On the other hand, the pharmacists had a moderate knowledge of AMR (54%) and a poor attitude (33%) and practice (38%). The nurses had a poor knowledge of AMR (31%) and practice (25%), but a moderate attitude (62%).

Figure 3.

Comparison of HCWs’ KAP of AMR in SSA between 2014 and 2024.

Trends in HCWs’ KAP towards AMR (2014–2024)

Over the decade under consideration, HCWs demonstrated fluctuating levels of KAP regarding AMR. Baseline knowledge in 2014 was relatively high (72.2%), likely reflecting ongoing awareness campaigns in urban tertiary facilities, though attitudes and practices data were not reported. By 2017, knowledge remained stable (72.8%), yet a comprehensive measurement of attitudes and practices remained absent.

The first attitude data appeared in 2018 (67.2%), while practice data remained largely unavailable. Knowledge peaked in 2020 (76.8%), coinciding with intensified implementation of the World Health Organization’s (WHO’s) AMR Global Action Plan and hospital-level stewardship programs. This year also provided the first concurrent measurements of attitude (65.2%) and practice (49.8%), indicating moderate translation of awareness into behaviour.

The COVID-19 pandemic had a profound impact on the KAP of HCWs towards AMR. In 2021, knowledge decreased sharply 41.8% and practice dropped substantially 9%, whereas attitudes remained relatively high (65.2%). Post-pandemic recovery in 2022 saw knowledge rebound (74.9%), but attitudes (29.6%) and practices (31.5%) lagged behind. Subsequent years (2023–2024) showed variability: knowledge fluctuated (29.3%–64.4%), attitudes modestly improved (21.5%–34.1%) and practices remained low (19.5%–36.8%).

Overall KAP of AMR

The pooled current proportion of AMR knowledge, attitude and practice was estimated at 65.4% (95% CI: 55.7–74.5; n = 4364; number of included estimates (k) = 25), 69.2% (95% CI: 58.1–79.3; n = 3857; k = 17) and 54.1% (95% CI: 44.1–63.9; n = 1841; k = 16) of the HCW populations, respectively.

The random effect models (Figure 4(a)–(c)) indicated that the effect of proportions was significant (the diamond did not cross the origin). In addition, from the random effect models, the tests of heterogeneity (all p < 0.001) suggested the presence of heterogeneous results, with the heterogeneity statistics (I² = 98.5%, 97.8%, 96.0%) indicating high between-sample heterogeneity and their 95% CI (98.3%–98.7%, 97.2%–98.2%, 94.7%–97.0%) indicative of a potentially important to substantial heterogeneity in the KAP effect sizes.

Figure 4.

Current proportion of overall (a) AMR knowledge in SSA, (b) AMR attitudes in SSA and (c) AMR practices in SSA.

Knowledge of AMR

The pooled current proportions of AMR knowledge among doctors, pharmacists and nurses were estimated at 65.1% (95% CI: 46.4–81.7; n = 1292; k = 10), 56.3% (95% CI: 33.9–77.5; n = 540; k = 9) and 28.9% (95% CI: 16.4–43.2; n = 341; k = 6) of the HCW populations, respectively.

The random effect models (Supplemental File 2, Figures 4–6) indicate that the effect of proportions was significant. From the random effect models, the tests of heterogeneity (all p < 0.001) suggested the presence of heterogeneous results, with the heterogeneity statistics (I² = 97.3%, 97.9%, 95.1%) indicating high between-sample heterogeneity and their 95% CI (96.3%–98.1%, 97.1%–98.5%, 91.7%–97.1%) indicative of a potentially important to substantial heterogeneity in the knowledge effect sizes.

Attitudes towards AMR

The pooled current proportion of AMR attitudes among doctors, pharmacists and nurses was estimated at 79.8% (95% CI: 32.1–99.9; n = 136; k = 3), 76.6% (95% CI: 51.8–94.6; n = 154; k = 5) and 88.4% (95% CI: 82.6–93.2; n = 308; k = 3) of the HCW populations, respectively.

From the random effect models (Supplemental File 2, Figures 7 and 8), the tests of heterogeneity (all p < 0.001 except for nurses) suggested the presence of heterogeneous results, with the heterogeneity statistics (I² = 98.2%, 92.2%), indicating high between-sample heterogeneity and their 95% CI (96.7%–99.0%, 84.7%–96.0%) indicative of a potentially important to substantial heterogeneity in the attitudes effect sizes. However, for the nurses, the heterogeneity test (p = 0.1354) indicated that there were few heterogeneous results, with a low between-sample heterogeneity (I² = 50.0%).

Practice towards AMR

The pooled current proportion of AMR practice among doctors, pharmacists and nurses was estimated at 42.8% (95% CI: 7.9–82.4; n = 65; k = 3), 38.9% (95% CI: 16.1–64.5; n = 464; k = 6) and 27.8% (95% CI: 11.7–47.6; n = 211; k = 2) of the HCW populations, respectively.

The random effect models (Supplemental File 2, Figures 9–11) indicate that the effect of proportions was significant. From the random effect models, the tests of heterogeneity (all p < 0.001) suggested the presence of heterogeneous results, with the heterogeneity statistics (I² = 95.7%, 99.5%, 96.6%), indicating high between-sample heterogeneity and their 95% CI (90.6%–98.0%, 99.3%–99.6%, 90.8%–98.7%) indicative of a potentially important to substantial heterogeneity in the practice effect sizes.

Publication bias

Funnel plots for overall KAP of AMR and knowledge of AMR among doctors exhibit asymmetrical patterns (Figure 5(a) and (b)). This asymmetry suggests publication bias and small-study effects, although the latter effect is subtle.

Figure 5.

(a) Funnel plot for overall knowledge, attitudes and practice of antimicrobial resistance in sub-Saharan Africa and (b) funnel plot for antimicrobial resistance knowledge among doctors in sub-Saharan Africa.

Rank correlation tests (Supplemental File 2, Table 4) revealed a significant relationship between sample size and observed effect size for KAP of AMR (all p-values < 0.05). This finding was corroborated by Egger’s regression tests (Supplemental File 2, Table 4), which detected significant small-study effects (all p-values < 0.05). These results statistically confirm the visual evidence of funnel plot asymmetry, highlighting the influence of small-study effects in the meta-analysis of overall KAP of AMR.

Due to the limited number of studies examining doctors’ attitudes (n = 3) and practices (n = 3) regarding AMR, as well as pharmacists’ knowledge (n = 9), attitudes (n = 5) and practices (n = 6), publication bias tests could not be reliably conducted for these HCW subgroups. Similarly, the limited number of studies on nurses’ knowledge (n = 6), attitudes (n = 3) and practices (n = 2) precluded meaningful publication bias assessments for this subgroup.

Sensitivity analysis

Application of the trim-and-fill method to account for asymmetry in funnel plots revealed the presence of potential publication bias. This method imputed 9 and 2 missing studies to the meta-analyses of overall knowledge and attitude, respectively, towards AMR. The adjusted random-effects proportions were 51.0% (95% CI: 39.7–62.3) for knowledge, characterised by substantial heterogeneity (I² = 98.8%, 95% CI: 98.6–99.0; k = 34) and 65.0% (95% CI: 53.3–75.8) for attitude, also exhibiting considerable heterogeneity (I² = 98.2%, 95% CI: 97.8–98.5; k = 19). Notably, the trim-and-fill method did not impute any missing studies for the overall practice meta-analysis.

Application of the trim-and-fill method to adjust for funnel plot asymmetry in the meta-analysis of AMR knowledge among doctors resulted in the imputation of two additional studies. This adjustment yielded a revised random-effects proportion of 57.3% (95% CI: 38.4–75.1), accompanied by substantial heterogeneity (I² = 97.4%, 95% CI: 96.6–98.1; k = 12). Notably, the trim-and-fill method did not significantly alter the heterogeneity statistics for either the overall KAP of AMR or AMR knowledge among doctors, suggesting that the estimated summary effect sizes are robust. Furthermore, the heterogeneity results remained consistent before and after applying the trim-and-fill method, lending support to the reliability of the findings.

Influence analysis

Overall KAP of AMR

To identify influential studies with high effect sizes and their impact on the overall heterogeneity, a leave-one-out sensitivity analysis was conducted. This involved iteratively removing each study from the meta-analysis and re-running the analysis to assess the changes in the overall effect estimate. By systematically excluding each study, we evaluated the robustness of the findings and identified studies with disproportionate influence on the overall effect size and heterogeneity.

Influence analysis identified one study³¹ as a significant outlier, substantially contributing to the heterogeneity in the meta-analyses of overall knowledge and attitude. Upon exclusion of this study, the revised meta-analyses yielded a pooled proportion of 66.4% (I² = 98.2%, n = 3338, k = 24) for overall knowledge and 69.3% (I² = 97.9%, n = 2168, k = 16) for overall attitude. Despite the removal of this outlier, the remaining studies continued to exhibit substantial heterogeneity.

Application of the trim-and-fill method to address funnel plot asymmetry resulted in the imputation of seven and three additional studies to the meta-analyses of overall knowledge and attitude, respectively. This adjustment yielded revised random-effects proportions of 56.7% (I² = 98.4%, k = 31) for knowledge and 62.6% (I² = 98.3%, k = 19) for attitude. Notably, the trim-and-fill method had a minimal impact on heterogeneity, with a slight increase in I² heterogeneity from 98.2% to 98.4% for knowledge and from 97.9% to 98.3% for attitude. In contrast, no outliers were detected in the meta-analysis of overall practice.

Knowledge of AMR

An influence analysis identified one study³¹ as an outlier, influencing the heterogeneity of the meta-analyses for knowledge among doctors, pharmacists and nurses. Excluding this study, the revised meta-analyses yielded pooled proportions of 66.6% (I² = 97.2%, n = 710, k = 9) for doctors, 60.2% (I² = 96.5%, n = 327, k = 8) for pharmacists and 27.7% (I² = 94.5%, n = 110, k = 5) for nurses. Subsequent application of the trim-and-fill method resulted in the imputation of 0, 1 and 2 additional studies for doctors, pharmacists and nurses, respectively. This adjustment yielded revised random-effects proportions of 66.6% (I² = 97.2%, k = 9), 66.6% (I² = 96.5%, k = 9) and 14.6% (I² = 96.3%, k = 7), with minimal to no change in I² heterogeneity.

Attitudes towards AMR

Influence analyses identified multiple outliers influencing the meta-analyses of attitude among healthcare professionals. Specifically, two studies^40,49 were detected as outliers for doctors’ attitude, while two studies^37,49 and two studies^40,43 were identified as outliers for pharmacists’ and nurses’ attitudes, respectively. Excluding these studies from their respective meta-analyses yielded revised proportions of 35.0% (I² = N/A, n = 62, k = 1) for doctors, 63.0% (I² = 91.0%, n = 96, k = 3) for pharmacists and 90.7% (I² = N/A, n = 78, k = 1) for nurses. Subsequent application of the trim-and-fill method was not feasible for doctors’ and nurses’ attitudes due to the limited number of remaining studies (k = 1) left after the influential ones were excluded. However, the trim-and-fill method did not impute any additional studies for pharmacists’ attitudes.

Practice towards AMR

Influence analyses identified outliers contributing to heterogeneity in the meta-analyses of practice among healthcare professionals. Specifically, one study³⁸ was detected as an outlier in the pharmacists’ practice studies, while two studies^38,50 were identified as outliers in the doctors’ practice studies. Excluding these studies from their respective meta-analyses resulted in revised proportions of 33.0% (I² = 99.2%, n = 209, k = 5) for pharmacists and 55.6% (I² = N/A, n = 25, k = 1) for doctors. Application of the trim-and-fill method to the pharmacists’ practice studies imputed three additional studies, yielding an adjusted random-effects proportion of 3.8% (I² = 99.5%, k = 8), with a minimal increase in I² heterogeneity from 99.2% to 99.5%. However, the trim-and-fill method could not be applied to doctors’ practice due to an insufficient number of remaining studies (k = 1) after excluding influential ones. Furthermore, influential analysis was not feasible for nurses’ practice studies due to the limited sample of included studies (n = 2).

Discussion

This systematic review and meta-analysis revealed that while HCWs demonstrate moderate to high knowledge of antimicrobial resistance, substantial gaps persist in attitudes and practices in SSA. The overall proportion of HCWs with adequate knowledge indicates moderate to good knowledge of AMR, mirroring findings from other studies in the region.^51,52 The persistence of moderate knowledge levels despite global awareness efforts may suggest that pre-service and in-service training on AMR remains insufficient across the subcontinent. In the current study, overall attitudes among HCWs were moderate, comparable to findings from Tanzania.⁵³ Moderate level of positive attitudes implies recognition of AMR as a public health concern but also reflects uncertainty about personal and institutional capacity to enact change. Limited stewardship infrastructure, insufficient administrative support and the absence of feedback mechanisms are often cited as barriers to attitudinal commitment towards AMR prevention and control.⁵⁴ Even more concerning is the low level of appropriate antimicrobial practices observed. Comparable findings have been reported in Uganda, South Africa and Cameroon.^31,55,56

The study’s results are consistent with those reported in other regions of the world, highlighting the global nature of the challenge posed by AMR. The observed prevalence of good knowledge among HCWs in SSA (65.4%) is remarkably high, surpassing the Western Pacific region (45.9%) and closely trailing Europe (70.3%).¹¹ This notable finding warrants recognition, as it suggests that sub-Saharan African HCWs demonstrate a solid understanding of AMR and its implications. The variations in AMR knowledge among HCWs across Europe, the Western Pacific and SSA can be attributed to a complex interplay of factors. It begins with divergent healthcare systems, where differences in infrastructure, funding and policies impact the quality of HCW training and education. In some regions, like Europe, robust healthcare systems⁵⁷ provide HCWs with ample opportunities for AMR training and updates. In contrast, regions with resource-constrained healthcare systems, like the Western Pacific, might struggle to provide HCWs with adequate AMR education. Furthermore, the prevalence of AMR in a region plays a significant role. In areas where AMR is rampant, such as the Western Pacific, HCWs may become desensitised, leading to complacency. Conversely, regions with lower AMR prevalence may foster a greater sense of urgency among HCWs to learn about AMR.

The longitudinal trends (Figure 2) highlight several key insights. First, improvements in knowledge generally precede enhancements in attitudes and practices, consistent with the KAP theoretical framework. Second, external disruptions such as the COVID-19 pandemic can sharply undermine both knowledge retention and practical implementation, even when positive attitudes persist. Third, despite post-pandemic recovery of knowledge, persistent gaps between attitudes and practice indicate that awareness alone does not ensure optimal antimicrobial stewardship. This emphasises the need for continuous reinforcement, training and system-level support to translate knowledge into behaviour.

Similar to findings from other studies, the KAP of HCWs towards AMR appeared to be compromised during the COVID-19 pandemic. This decline has been attributed to insufficient training, limited access to accurate and timely information and inadequate institutional preparedness.⁵⁸ Moreover, attitudinal barriers and systemic constraints—such as shortages of personal protective equipment (PPE) and increased clinical workload—further impeded effective antimicrobial practices. Evidence suggests that enhanced training, targeted education programs and improved access to reliable information are critical to strengthening HCWs’ KAP on AMR, particularly in low- and middle-income settings.^59,60

This study has several strengths. Firstly, it provides a comprehensive overview of the KAP of HCWs regarding AMR in SSA, a region with limited data on this topic. Secondly, the study’s methodology was rigorous, with a systematic search strategy and a thorough quality assessment of the included studies. Finally, the study’s findings have significant public health relevance, highlighting the need for targeted interventions to improve AMR KAP among HCWs in SSA.

Despite its valuable insights, this study is subject to several limitations that may influence the reliability and validity of its findings. Variability in the methodological quality and sample sizes of the included studies introduces heterogeneity that can compromise the consistency and precision of pooled estimates, thereby reducing the reliability of the results. The reliance on self-reported data further poses a risk of recall and social desirability bias, which may inflate reported knowledge or positive practices, ultimately threatening internal validity. Incomplete or inconsistent data from earlier years (2014 and 2017) weaken the temporal comparability of findings and constrain robust longitudinal inferences, limiting external validity over time. Moreover, while a notable gap between knowledge and practice was identified, the absence of analysis on contextual or behavioural determinants underlying this discrepancy restricts the explanatory depth of the study and impedes a nuanced understanding of causal mechanisms. Together, these limitations suggest that while the observed patterns are informative, they should be interpreted with caution, and future research employing standardised methodologies and mixed-methods approaches is warranted to strengthen causal and contextual validity.

The public health relevance of this study’s findings cannot be overstated. AMR is a growing global health threat, and HCWs play a critical role in preventing and controlling its spread. The findings of this study suggest that there are significant gaps in AMR KAP among HCWs in SSA, which may be contributing to the spread of AMR in the region. Addressing these gaps through targeted interventions, such as education and training programs, is critical to mitigating the spread of AMR.⁶¹

Future research should prioritise the development and validation of standardised, context-sensitive instruments for assessing HCWs’ knowledge, attitudes and practices towards antimicrobial resistance. The current heterogeneity in measurement tools impedes cross-study comparability and undermines the precision of pooled estimates in meta-analyses. Establishing unified metrics and validated KAP scales—adapted across healthcare cadres and settings—would enhance methodological rigour and enable more accurate global benchmarking. Additionally, future studies should move beyond descriptive assessments to rigorously evaluate the impact of targeted interventions, such as antimicrobial stewardship training, audit-and-feedback mechanisms and digital decision-support tools, using robust experimental or quasi-experimental designs. Longitudinal and mixed-methods approaches are also warranted to capture behavioural change over time and elucidate contextual drivers of KAP gaps, particularly in resource-limited settings. Strengthening research in these directions will not only improve the reliability and generalisability of evidence but also provide actionable insights to inform policy and practice in combating AMR at both institutional and system levels.

Conclusion

This systematic review and meta-analysis demonstrate considerable heterogeneity and overall suboptimal levels of knowledge, attitudes and practices among HCWs towards AMR. Although pooled estimates indicate moderate knowledge (65.4%) and attitude (69.2%) levels, translation into effective practices remains limited (54.1%). Stratified analyses revealed that doctors exhibited relatively higher knowledge but poor practice levels, while pharmacists and nurses showed both inadequate knowledge and inconsistent practices. Despite generally favourable attitudes, particularly among nurses, these did not consistently result in improved behavioural outcomes.

Collectively, these findings underscore a persistent knowledge–practice gap that threatens the effectiveness of evidence-based healthcare delivery. Strengthening continuous professional development, integrating stewardship principles into pre- and in-service training and implementing institutional support mechanisms are essential to improve HCWs’ competencies and practice behaviours towards AMR. Future interventions should be context-sensitive, profession-specific and rigorously evaluated to achieve sustainable improvements in healthcare quality and patient outcomes.

Supplemental Material

sj-docx-1-tai-10.1177_20499361251410677 – Supplemental material for Knowledge, attitudes and practices on antimicrobial resistance among healthcare workers in sub-Saharan Africa: a systematic review and meta-analysis

Supplemental material, sj-docx-1-tai-10.1177_20499361251410677 for Knowledge, attitudes and practices on antimicrobial resistance among healthcare workers in sub-Saharan Africa: a systematic review and meta-analysis by Herbert Melariri, Prince Campbell, Rosa Du Randt, Ilse Truter, Simo Zulu, Beth Rose Wiehahn, Alive Ntunja, Marilyn Hutton-Willemse, Zukiswa Ntlokonkulu, Ntokozo Mzimela, Opeoluwa Oyedele and Paula Melariri in Therapeutic Advances in Infectious Disease

Supplemental Material

sj-docx-2-tai-10.1177_20499361251410677 – Supplemental material for Knowledge, attitudes and practices on antimicrobial resistance among healthcare workers in sub-Saharan Africa: a systematic review and meta-analysis

Supplemental material, sj-docx-2-tai-10.1177_20499361251410677 for Knowledge, attitudes and practices on antimicrobial resistance among healthcare workers in sub-Saharan Africa: a systematic review and meta-analysis by Herbert Melariri, Prince Campbell, Rosa Du Randt, Ilse Truter, Simo Zulu, Beth Rose Wiehahn, Alive Ntunja, Marilyn Hutton-Willemse, Zukiswa Ntlokonkulu, Ntokozo Mzimela, Opeoluwa Oyedele and Paula Melariri in Therapeutic Advances in Infectious Disease

Footnotes

Acknowledgements

None.

Declarations

ORCID iD

Herbert Melariri

Supplemental material

Supplemental material for this article is available online.

References

Totaro

, et al. Antimicrobial resistance in Sub-Saharan Africa: a comprehensive landscape review. Am J Trop Med Hyg 2025; 113(2): 253–263.

Murray

, et al. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet 2022; 399: 629–655.

O’Neill

Tackling drug-resistant infections globally: final report and recommendations. AMR Review, 2016.

Okeke

, et al. The scope of the antimicrobial resistance challenge. Lancet 2024; 403: 2426–2438.

Tadesse

, et al. Antimicrobial resistance in Africa: a systematic review. BMC Infect Dis 2017; 17: 616.

Melariri

, et al. The Burden of Hospital-Acquired Infections (HAI) in Sub-Saharan Africa: a systematic review and meta-analysis. The Lancet 2024.

WHO. No time to wait: securing the future from drug-resistant infections. World Health Organization, 2019.

Rana

, et al. Knowledge, attitudes and practices of antimicrobial resistance awareness among healthcare workers in India: a systematic review. Front Public Health 2024; 12: 1433430.

WHO. WHO library cataloguing-in-publication data global action plan on antimicrobial resistance. World Health Organization, 2015.

10.

Reddy

, et al. Antimicrobial resistance and antimicrobial stewardship in South Africa: a survey of healthcare workers in academic and nonacademic hospitals. Antimicrobial Steward Healthcare Epidemiol 2023; 3: e202, 1–7.

11.

Jahromi

, et al. Global knowledge, attitudes, and practices towards antimicrobial resistance among healthcare workers: a systematic review and meta-analysis. Antimicrob Resist Infect Control 2025; 14(1): 47.

12.

Mudenda

, et al. Knowledge, awareness and practices of healthcare workers regarding antimicrobial use, resistance and stewardship in Zambia: a multi-facility cross-sectional study. JAC Antimicrob Resist 2024; 6: dlae076.

13.

Koroma

, et al. Behavioral practices towards antibiotic use among health care workers-Sierra Leone, 2021: a facility-based cross-sectional study. Pan African Medical J 2024; 47(63).

14.

Kpokiri

, et al. Knowledge, awareness and practice with antimicrobial stewardship programmes among healthcare providers in a Ghanaian Tertiary Hospital. Antibiotics 2022; 11(1): 6.

15.

Ogoina

, et al. Predictors of antibiotic prescriptions: a knowledge, attitude and practice survey among physicians in tertiary hospitals in Nigeria. Antimicrob Resist Infect Control 2021; 10: 73.

16.

Page

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Syst Rev 2021; 372: n71.

17.

Chen

Wang

, et al. Using Freeman-Tukey double arcsine transformation in meta-analysis of single proportions. Aesthetic Plastic Surg 2023; 47: 83–84.

18.

Thorlund

, et al. Evolution of heterogeneity (I2) estimates and their 95% confidence intervals in large meta-analyses. PLoS One 2012; 7(7): e39471.

19.

Terrin

Schmid

Lau

, et al. Adjusting for publication bias in the presence of heterogeneity. Stat Med 2003; 22: 2113–2126.

20.

Bloom

Engelhart

Furst

, et al. Taxonomy of educational objectives. 1964.

21.

Chukwu

, et al. Antimicrobial resistance awareness and antibiotic prescribing behavior among healthcare workers in Nigeria: a national survey. BMC Infect Dis 2021; 21(1): 22.

22.

Abera

Kibret

Mulu

Knowledge and beliefs on antimicrobial resistance among physicians and nurses in hospitals in Amhara Region, Ethiopia. BMC Pharmacol Toxicol 2014; 15: 26.

23.

Asante

, et al. Knowledge of antibiotic resistance and antibiotic prescription practices among prescribers in the Brong Ahafo Region of Ghana; a cross-sectional study. BMC Health Serv Res 2017; 17: 422.

24.

Gebretekle

, et al. Opportunities and barriers to implementing antibiotic stewardship in low and middle-income countries: lessons from a mixed-methods study in a tertiary care hospital in Ethiopia. PLoS One 2018; 13(12): e0208447.

25.

Labi

, et al. Physicians’ knowledge, attitudes, and perceptions concerning antibiotic resistance: a survey in a Ghanaian tertiary care hospital. BMC Health Serv Res 2018; 18.

26.

Assefa

Diress

Adane

Infection prevention knowledge, practice, and its associated factors among healthcare providers in primary healthcare unit of Wogdie District, Northeast Ethiopia, 2019: a cross-sectional study. Antimicrob Resist Infect Control 2020; 9: 136.

27.

Dougnon

, et al. Prescription of antibiotics as a source of emerging antibiotic resistance: knowledge, attitudes, and practices of medical staff in the Dassa-Glazoué and Savalou-Bantè’s health zones (Benin, West Africa). Int J One Health 2020; 6: 34–40.

28.

Engler

, et al. Compliance with South Africa’s antimicrobial resistance national strategy framework: are we there yet? J Chemother 2021; 33(1): 21–31.

29.

Musoke

, et al. A one health approach to strengthening antimicrobial stewardship in Wakiso District, Uganda. Antibiotics 2020; 9: 1–16.

30.

Akpan

Udoh

Akpan

, et al. Community pharmacists’ management of self-limiting infections: a simulation study in Akwa Ibom State, South-South Nigeria. Afr Health Sci 2021; 21: 576–584.

31.

Balliram

Sibanda

Essack

SY.

The knowledge, attitudes and practices of doctors, pharmacists and nurses on antimicrobials, antimicrobial resistance and antimicrobial stewardship in South Africa. S Afr J Infect Dis 2021; 36: a262.

32.

Domche Ngongang

Basera

Mendelson

Tertiary hospitals physician’s knowledge and perceptions towards antibiotic use and antibiotic resistance in Cameroon. BMC Infect Dis 2021; 21: 1116.

33.

Bedekelabou

Oyetola

Coulibaly

, et al. First assessment of the knowledge, attitudes, and practices of health actors in Togo and Ivory Coast in regard to antibiotic resistance. Int J One Health 2022; 8: 108–123.

34.

Magadze

Nkhwashu

Moloko

, et al. The impediments of implementing infection prevention control in public hospitals: Nurses’ perspectives. Health SA Gesondheid 2022; 27(0): a2033.

35.

Tembo

Mudenda

Banda

, et al. Knowledge, attitudes and practices on antimicrobial resistance among pharmacy personnel and nurses at a tertiary hospital in Ndola, Zambia: implications for antimicrobial stewardship programmes. JAC Antimicrob Resist 2022; 4(5): dlac107.

36.

Akande-Sholabi

Oyesiji

Adebisi

YA.

Antimicrobial stewardship: community pharmacists’ antibiotic dispensing practices, knowledge, and perception regarding antibiotics and antibiotic resistance. J Pharmaceutical Health Services Res 2023; 14: 383–391.

37.

Chetty

Leigh-De Rapper

Exploring the discord between pharmacy education and practice in antimicrobial stewardship. Health SA Gesondheid 2023; 28(0): a2114.

38.

Koroma

, et al. An assessment of knowledge, attitude, and practice of medical professionals on factors related to antimicrobial resistance in three (3) selected University Teaching Hospital Complexes in Sierra Leone: a cross-sectional analytic study. Texila Int J Public Health 2023; 11.

39.

Nabidda

, et al. Antimicrobial resistance and rational prescription practices: knowledge, perceptions and confidence of health profession interns in Uganda. JAC Antimicrob Resist 2023; 5(5): dlad105.

40.

Sefah

, et al. A multicenter cross-sectional survey of knowledge, attitude, and practices of healthcare professionals towards antimicrobial stewardship in Ghana: findings and implications. Antibiotics 2023; 12: 1497.

41.

Abejew

Wubetu

Fenta

TG.

Assessment of challenges and opportunities in antibiotic stewardship program implementation in Northwest Ethiopia. Heliyon 2024; 10: e32663.

42.

Amorha

Chima

Ugochukwu

, et al. Assessment of community pharmacists’ knowledge, perception and practice regarding antimicrobial stewardship in Ebonyi State, Southeastern Nigeria. African J Biomed Res 2024; 27: 65–72.

43.

Djochie

RDA

, et al. Assessment of knowledge and perception of prescribers towards rational medicine use in the Ashanti Region of Ghana. PLoS One 2024; 19(10): e0308406.

44.

Kıratlı

, et al. Evaluation of doctors’ knowledge, attitudes, behaviors, awareness and practices on rational antimicrobial stewardship in a training and research Hospital in Mogadishu-Somalia. Infect Drug Resist 2024; 17: 2759–2771.

45.

Mufwambi

, et al. Community pharmacists’ knowledge and practices towards antimicrobial stewardship: findings and implications. JAC Antimicrob Resist 2024; 6(5): dlae176.

46.

Ojo

Ipinnimo

Osho

, et al. Antibiotics use, resistance and self-medication practices among healthcare workers in a Federal Teaching Hospital in Southwest, Nigeria. J Public Health Develop 2024; 22: 336–346.

47.

Osman

, et al. Antibiotic resistance in Sudan: assessing the knowledge and practices of healthcare workers in Khartoum. JAC Antimicrob Resist 2024; 6(2): dlae049.

48.

McKnight

, et al. Evaluating hospital performance in antibiotic stewardship to guide action at national and local levels in a lower-middle income setting. Glob Health Action 2019; 12: 1761657.

49.

Gebrehiwot

Tadiwos

Knowledge and beliefs of health care professionals towards antimicrobial resistance in Hiwot Fana specialized University Hospital, in Harar, Ethiopia. Infect Drug Resist 2020; 13: 2027–2035.

50.

Menkem

, et al. Attitudes and practices of the use of third-generation cephalosporins among medical doctors practicing in cameroon. Int J Clin Pract 2023; 2023: 8074413.

51.

Getu

, et al. Knowledge, attitudes, and practices regarding antimicrobial resistance and associated factors among healthcare workers in Addis Ababa, Ethiopia: a cross-sectional study. Can J Infect Dis Med Microbiol 2025; 2025: 5354292.

52.

Davwar

, et al. Knowledge, attitudes, and practice of doctors in Nigeria regarding antimicrobial resistance. Niger Med J 2023; 64: 492–502.

53.

Sindato

, et al. Knowledge, attitudes and practices regarding antimicrobial use and resistance among communities of Ilala, Kilosa and Kibaha districts of Tanzania. Antimicrob Resist Infect Control 2020; 9: 194.

54.

Pulcini

, et al. Developing core elements and checklist items for global hospital antimicrobial stewardship programmes: a consensus approach. Clin Microbiol Infect 2019; 25(1): 20–25.

55.

Kizito

, et al. Antibiotic consumption and utilization at a large tertiary care level hospital in Uganda: a point prevalence survey. PLoS One 2025; 20: e0313587.

56.

Mbapah

, et al. Compliance with transmission-based precautions, and associated factors among healthcare providers in Cameroon: a cross-sectional study. Antimicrob Resist Infect Control 2025; 14: 21.

57.

ECDC 2024. ECDC/EC/EFSA country visit to Montenegro to advance One Health responses against antimicrobial resistance. European Centre for Disease Prevention and Control, 2024.

58.

Nemr

, et al. Knowledge, attitude, and practice (KAP) of antimicrobial prescription and its resistance among health care providers in the COVID-19 era: a cross sectional study. PLoS One 2023; 18(8): e0289711.

59.

Al Sulayyim

Ismail

Hamid

, et al. Knowledge, attitude and practice of healthcare workers towards antibiotic resistance during the COVID-19 pandemic. JAC Antimicrob Resist 2023; 5: dlad068.

60.

Gulleen

, et al. Knowledge and perceptions of antimicrobial resistance and antimicrobial stewardship among staff at a national cancer referral center in Uganda. Antimicrob Steward Healthc Epidemiol 2022; 2: e54.

61.

Ahmed

, et al. Antimicrobial resistance: impacts, challenges, and future prospects. J Med Surg Public Health 2024; 2: 100081.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.27 MB

0.93 MB

0.00 MB