Abstract
Background
Acute kidney injury in children is a serious but often overlooked condition in low-resource settings. In Rwanda, although referral hospitals provide advanced care, most children are managed at district hospitals, where limited diagnostic services and low provider awareness may delay diagnosis. We assessed healthcare providers’ knowledge of pediatric AKI, audited real-world case management, and evaluated whether a brief training intervention could improve early detection and care.
Methods
We conducted a mixed-method study in six Rwandan hospitals affiliated with the University Teaching Hospital of Kigali from 2024 to 2025. A cross-sectional survey assessed provider knowledge, followed by retrospective and prospective case audits of pediatric AKI management. Cases aged 1 month to 14.9 years were screened using serum creatinine ≥ 1.0 mg/dL, and 155/156 met KDIGO Serum Criteria for AKI. A KDIGO-based educational workshop was delivered on April 25, 2025; patients admitted before formed pre-intervention cohort (n = 138) and post-intervention cohort (n = 18). Care quality was assessed using nine indicators adapted from the Recognition-Action-Results framework. Multivariable and stratified analyses were performed.
Results
Among 166 providers (65.7% female; 51.2% nurses), the mean knowledge score was 2.1/5.0 (42%), with only 3.6% achieving > = 80%. Knowledge gaps were consistent across professional categories and hospitals, although 89.8% expressed willingness to adopt AKI guidelines. In the 156 pediatric cases (mean age was 6.6 +/- 5.3 years; 54.5% male), 76.3% presented with KDIGO Stage 3 AKI, most commonly associated with acute gastroenteritis (33.3%). Baseline care quality was poor with low documentation of staging (5.8%), urine output monitoring (3.6%), and follow-up creatinine testing (27.5%); however, laboratory investigations were performed in 87.7% of cases. After training, follow-up creatinine monitoring increased modestly to 33.3%. Composite care quality showed non-significant improvement (40.6% vs. 44.4%; OR 1.17, 95% CI 0.44–3.15; p = 0.802). Overall mortality was 9.6% (15/156) with hypovolemic shock (OR 8.73, p = 0.008), severe dehydration (OR 5.29, p = 0.009), and hypernatremia (OR 4.18, p = 0.045) as independent predictors.
Conclusion
Pediatric AKI remains underrecognized in Rwandan district hospitals, with critical gaps in staging and monitoring. Training improved selected practices but did not translate into improved outcomes. Sustained improvements require mentorship, standardized protocols, and system-level support. Larger prospective studies are needed to confirm impact.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12887-026-06946-9.
Keywords: Acute kidney injury, AKI, Healthcare provider knowledge, Clinical training intervention, Rwanda, Resource-limited settings
Background
Acute kidney injury (AKI) is a serious complication in pediatric patients worldwide, with particularly devastating consequences in low-resource settings where access to specialized care and renal replacement therapy is limited [1]. The global incidence of pediatric AKI ranges from 8.4% to 17.6% in hospitalized patients, with mortality rates reaching 73% when dialysis is needed but unavailable [2, 3]. In sub-Saharan Africa, pediatric AKI is often associated with poverty-related conditions, including severe dehydration from diarrheal diseases, sepsis, malaria, and exposure to nephrotoxic traditional medicines [4].
Early recognition of pediatric AKI is essential because once injury occurs, effective therapies to reverse or accelerate recovery remain limited. Timely identification of at-risk children allows clinicians to prevent progression, mitigate complications, and optimise supportive care [5]. The Kidney Disease: Improving Global Outcomes (KDIGO) guidelines provide evidence-based Recommendations for AKI diagnosis, classification, and management [6]. However, the implementation of these guidelines in resource-limited settings faces numerous challenges, including inadequate training, a shortage of specialized healthcare workers, limited diagnostic capabilities, and a lack of awareness of AKI among primary healthcare providers [7, 8].
Healthcare provider knowledge and practices play pivotal roles in AKI outcomes [9]. Studies from high-income countries have demonstrated that educational interventions and the implementation of standardized protocols can significantly improve AKI recognition, reduce the time to diagnosis, and enhance patient outcomes [10, 11]. However, limited data exist regarding healthcare provider knowledge about pediatric AKI management in sub-Saharan Africa, particularly in Rwanda’s evolving healthcare system [12].
Rwanda has made remarkable progress in healthcare delivery since 1994, with significant investments in health system strengthening and human resource development [13]. The country’s healthcare system is organized in a three-tiered pyramid structure: national referral hospitals at the apex, intermediate-level hospitals (provincial and district referral hospitals) in the middle, and health centers forming the base [14]. Within this structure, most patients receive care at district hospitals, which serve as the primary inpatient facilities for their catchment populations. These hospitals are technically supported by national referral centers, including the University Teaching Hospital of Kigali (CHUK), which receives patients who need advanced consultation services and specialized care. However, district hospitals and or primary healthcare facilities face inherent limitations in diagnostic capacity, specialized personnel, and advanced treatment options that may impact early AKI recognition and management [15, 16].
Research in Rwanda identified significant knowledge gaps among healthcare providers regarding adult AKI management, with limited diagnostic capacity and treatment options at the district hospital level [17]. A previous study conducted at the University Teaching Hospital of Kigali (CHUK) evaluated AKI among children aged 1 month to 14 years admitted with septic and non-septic illness in the pediatric emergency unit, high-dependency unit (HDU), and pediatric intensive care settings, where the most critically ill children in Rwanda receive care. Many required advanced organ-support measures, including intravenous fluid resuscitation, vasopressor support for septic shock, oxygen therapy, blood transfusion, and close hemodynamic monitoring. The study reported an overall AKI incidence of 6.8%, with markedly higher mortality among children presenting with shock, severe sepsis, and multiorgan dysfunction and Mortality was 36.2% in AKI Children compared to the overall mortality of 10.1% [15].
Understanding knowledge levels and attitudes toward guideline implementation is essential for designing effective educational interventions and measuring their impact on clinical practice. This study aimed to assess healthcare providers’ knowledge of pediatric AKI management and attitudes toward clinical practice guidelines across six hospitals in Rwanda’s healthcare system, providing crucial baseline data for future quality improvement initiatives. It was conducted at six hospitals within the CHUK referral network, comprising four district hospitals (Kibagabaga, Muhima, Kacyiru, and Byumba) and two Level II/Referral hospitals ( Ruhengeri Level II Teaching Hospital and Kibuye Referral Hospital), selected based on patient volume and varying levels of expertise as noted in Fig. 1.
Fig. 1.
Geographical location of hospitals participating in this study on the Rwandan map
Methods
Study design and setting
We carried out a mixed-methods before-and-after intervention study from July 1, 2024, to June 30, 2025, in 6 of the 25 hospitals in the CHUK referral network: Kibagabaga, Muhima, Kacyiru, Byumba, Ruhengeri, and Kibuye (Fig. 1). The study had three main parts. First, a cross-sectional survey of healthcare providers was conducted to gather information on their demographics, clinical experience, knowledge of pediatric AKI, and attitudes toward guideline-based care. Second, on April 25, 2025, we provided a standardized KDIGO-based educational session at all participating hospitals to improve how pediatric AKI is recognized, staged, and managed. Third, we conducted case file reviews of children aged 1 month to 14.9 years with serum creatinine ≥ 1.0 mg/dL (88.4 µmol/L) at any point during hospitalization between July 1, 2024, and June 30, 2025. Of the 156 children meeting the screening threshold, 155 met KDIGO criteria for AKI; the remaining 1 was retained in the analytic cohort as they met the primary inclusion criterion. Patients admitted before April 25, 2025, formed the pre-intervention group (n = 138), and those admitted after formed the post-intervention group (n = 18). This study follows the SQUIRE 2.0 (Standards for Quality Improvement Reporting Excellence) guidelines.
Participants and recruitment
Healthcare providers
All healthcare providers involved in pediatric care including pediatricians, general practitioners, registered nurses, and medical students/residents rotating through pediatric departments were eligible. Inclusion criteria were: (1) currently working in pediatric, neonatal, emergency, or intensive care units; (2) direct involvement in patient care; and (3) provision of informed consent. Providers not involved in direct patient care and those who declined were excluded. Participants were recruited through onsite training meetings and written informed consent was obtained before data collection.
A structured, pretested questionnaire covered four domains: (1) demographic and professional characteristics (age, sex, education level, years of practice, department, prior AKI training); (2) AKI clinical experience (self-reported frequency of AKI encounters, missed diagnoses, and comfort managing pediatric AKI); (3) knowledge assessment — five multiple-choice questions based on KDIGO guidelines, covering creatinine-based AKI diagnosis, AKI staging, intrinsic AKI causes, management principles, and absolute indications for dialysis; and (4) guideline acceptability six statements rated on a 5-point Likert scale assessing perceived need, expected benefits, and willingness to implement guidelines. The questionnaire was developed in English and pretested among 15 healthcare providers at CHUK; minor modifications were made based on feedback. The full questionnaire, guideline materials, training slides, and patient data collection instrument are provided in the Supplementary File.
Pediatric case reviews
Pediatric files were identified at each participating hospital using either: (i) a documented diagnosis of AKI in the medical record or admission register, or (ii) serum creatinine ≥ 1.0 mg/dL (88.4 µmol/L) at any point during hospitalization. This threshold was selected based on prior pediatric cohort studies and published reference intervals, where the upper limit of normal in older children approaches 0.79 mg/dL [18]. Children with a prior diagnosis of chronic kidney disease, those lacking medical records, and infants below 1 month of age were excluded. All cases meeting the screening criterion were staged using KDIGO serum creatinine criteria [6]. Where a prior measured baseline creatinine was available from medical records, this value was used. Where baseline creatinine was unavailable common in district hospital settings [19]; It was estimated using the Pottel height-independent equation assuming a normal eGFR of 120 mL/min/1.73 m², as described and validated by Pottel et al. and Blufpand et al. in children and adolescents [20, 21]. This approach is supported by studies in African pediatric populations [22, 23]. Urine output criteria were not included in AKI staging due to inconsistent documentation across study sites; this limitation is discussed below.
Care quality was assessed using a nine-item composite indicator adapted from the Recognition-Action-Results (RAR) framework described by Kashani et al. (2019) [24]. The framework used in this study is annexed in the Supplementary file.
Data collection
Provider data were collected via structured questionnaire administered in person and captured digitally through KoBoCollect. At each hospital, two research assistants with medical backgrounds were deployed after a one-day training led by the principal investigator. For the knowledge assessment, the primary outcome was the overall knowledge score (sum of correct responses, range 0–5, expressed as a percentage). Secondary outcomes included performance on individual questions, guideline acceptability (composite score range 6–30), and factors associated with knowledge performance. Knowledge levels were categorized as excellent (≥ 80%), good (60–79%), fair (40–59%), or poor (< 40%). For the clinical case review, patient-level data were extracted from medical records via a standardized 85-variable form. Clinician recognition of AKI was coded as ‘yes’ if AKI, acute renal failure, rising creatinine, oliguria, or any synonymous term was explicitly documented in clinical notes, the problem list, or assessment plan.
Statistical analysis
Survey responses were exported from KoBoCollect into Microsoft Excel and analyzed in Jamovi V2.6.44 Descriptive statistics summarize provider characteristics, knowledge scores, and guideline acceptability. Knowledge performance was reported as the means with standard deviations, medians with interquartile ranges, and categorical distributions. Differences across professional cadres and hospitals were assessed via chi-square or Fisher’s exact tests for categorical variables and ANOVA for continuous variables.
For the case audit, continuous variables are presented as means ± SD or medians with IQR; categorical variables as proportions. Differences between pre- and post-intervention groups were assessed using Fisher’s exact test for categorical variables and the Mann-Whitney U test for continuous variables. All tests were two-tailed; p < 0.05 was considered statistically significant.
Each of the nine RAR quality indicators was compared between intervention periods using Fisher’s exact test, reported as the proportion meeting the criterion with the absolute change in percentage points. Domain sub-scores and the total composite score were compared using the Mann-Whitney U test. The proportion meeting the composite threshold of ≥ 3 indicators was compared using Fisher’s exact test, with effect size expressed as an odds ratio (OR) and 95% confidence interval (CI). All integer cutoffs from ≥ 1 to ≥ 9 were evaluated as sensitivity analyses.
In-hospital mortality was the primary clinical outcome. Each candidate predictor was tested individually using Fisher’s exact test; unadjusted ORs with 95% CIs were calculated from 2 × 2 contingency tables. Given marked heterogeneity in mortality rates across hospital sites (range 0%–24%), Mantel-Haenszel stratified analysis was performed for the two strongest predictors (severe dehydration and hypernatremia), stratifying by hospital. Full multivariate logistic regression was not performed because the 15 mortality events precluded stable estimation with multiple covariates.
Results
Healthcare provider knowledge assessment
One hundred sixty-six healthcare providers participated in this baseline assessment across six hospitals in Rwanda. Only four paediatricians were included because the participating hospitals had a total of eight; one declined, and three were absent on the assessment day. Paediatricians were analysed separately because they constitute a specialised clinical cadre with advanced pediatric training and responsibilities distinct from general medical doctors/medical officers. Therefore, they were not included in the “medical doctor” group. The demographic and professional characteristics of the participants are summarized in Table 1.
Table 1.
Demographic characteristics of health care providers trained by hospital
| Characteristic | Overall (n = 166) | Byumba (n = 48) | Kacyiru (n = 12) | Kibagabaga (n = 17) | Kibuye (n = 23) | Muhima (n = 35) | Ruhengeri (n = 31) | P value |
|---|---|---|---|---|---|---|---|---|
| Gender, n (%) | ||||||||
| Female | 109 (65.7) | 32 (66.7) | 9 (75.0) | 11 (64.7) | 12 (52.2) | 24 (68.6) | 21 (67.7) | 0.678 |
| Male | 57 (34.3) | 16 (33.3) | 3 (25.0) | 6 (35.3) | 11 (47.8) | 11 (31.4) | 10 (32.3) | |
| Education Level, n (%) | ||||||||
| Registered Nurse | 85 (51.2) | 25 (52.1) | 6 (50.0) | 8 (47.1) | 8 (34.8) | 12 (34.3) | 16 (51.6) | 0.045 |
| Medical Doctors (General Practitioners) | 24 (14.5) | 6 (12.5) | 3 (25.0) | 4 (23.5) | 4 (17.4) | 5 (14.3) | 2 (6.5) | |
| Pediatrician | 4 (2.4) | 0 (0.0) | 1 (8.3) | 2 (11.8) | 0 (0.0) | 2 (5.7) | 0 (0.0) | |
| Other/Student | 53 (31.9) | 17 (35.4) | 2 (16.7) | 3 (17.6) | 11 (47.8) | 16 (45.7) | 13 (41.9) | |
| Years of Practice | ||||||||
| Mean (SD) | 6.8 (7.2) | 6.4 (7.8) | 4.2 (4.1) | 8.9 (9.8) | 5.3 (6.1) | 7.8 (7.4) | 7.1 (6.9) | 0.312 |
| Median (IQR) | 3.0 (1.0–10.0) | 2.0 (1.0–11.0) | 3.0 (1.0-6.8) | 3.0 (1.0-15.5) | 2.0 (1.0-8.5) | 5.0 (1.0–13.0) | 4.0 (2.0–10.0) | |
| Previous AKI Training, n (%) | ||||||||
| Graduate training | 89 (53.6) | 27 (56.3) | 7 (58.3) | 8 (47.1) | 11 (47.8) | 18 (51.4) | 18 (58.1) | 0.832 |
| CPD training | 12 (7.2) | 4 (8.3) | 0 (0.0) | 0 (0.0) | 2 (8.7) | 1 (2.9) | 5 (16.1) | |
| Never | 58 (34.9) | 15 (31.3) | 5 (41.7) | 8 (47.1) | 9 (39.1) | 14 (40.0) | 7 (22.6) | |
| Other | 7 (4.2) | 2 (4.2) | 0 (0.0) | 1 (5.9) | 1 (4.3) | 2 (5.7) | 1 (3.2) | |
AKI Acute kidney injury, CPD Continuous professional development, IQR Interquartile range
p values from the chi-square test for categorical variables and ANOVA for continuous variables
Knowledge scores were uniformly low across all professional categories (overall mean 2.1/5.0, 42%), with only 3.6% achieving adequate performance (≥ 80%) and fewer than one-fifth scoring ≥ 60% (Table 2). Differences across professional categories (p = 0.312) and hospital sites (p = 0.756) were not statistically significant, indicating system-wide rather than cadre-specific deficits. Performance was lowest on identification of intrinsic AKI causes (33.7% correct) and highest on AKI management principles (65.1%) (Table 3).
Table 2.
Knowledge assessment scores by professional category
| Group | n | Mean Score (SD) | Median (IQR) | Range | % Scoring ≥ 80% | % Scoring ≥ 60% |
|---|---|---|---|---|---|---|
| Overall | 166 | 2.1 (1.2) | 2.0 (1.0–3.0) | 0–5 | 3.60% | 18.10% |
| By Professional Category | ||||||
| Registered Nurses | 85 | 2.0 (1.1) | 2.0 (1.0–3.0) | 0–5 | 2.40% | 15.30% |
| Medical Doctors | 24 | 2.4 (1.3) | 2.0 (1.0–3.0) | 0–5 | 8.30% | 25.00% |
| Pediatricians | 4 | 2.8 (1.5) | 3.0 (1.5-4.0) | 2–4 | 20.00% | 60.00% |
| Other/Students | 53 | 2.0 (1.2) | 2.0 (1.0–3.0) | 0–4 | 3.80% | 15.10% |
| pvalue* | 0.312 | |||||
*Knowledge scores out of 5 points. p-values from ANOVA
Table 3.
Performance by individual knowledge questions
| Question | Topic | Correct Answer | n Correct (%) |
|---|---|---|---|
| Q1 | AKI Diagnosis (KDIGO criteria) | A | 89 (53.6%) |
| Q2 | AKI Classification/Staging | B | 98 (59.0%) |
| Q3 | Intrinsic AKI Etiology | C | 56 (33.7%) |
| Q4 | AKI Management | B | 108 (65.1%) |
| Q5 | Dialysis Indications | A | 95 (57.2%) |
Question topics: Q1-AKI diagnosis using creatinine criteria; Q2-AKI staging; Q3-Identifying intrinsic causes (HUS); Q4-Fluid management; Q5-Absolute dialysis indications
Despite these knowledge gaps, providers were highly receptive to guideline implementation. Overall acceptability scores averaged 29.1/30, with 89.8% expressing strong willingness to follow guidelines if available (Table 4 and Fig. 2). Multivariable linear regression identified no significant predictors of higher knowledge scores across hospital, professional category, years of practice, or prior training (R²=0.087, overall p = 0.523), though providers reporting high comfort managing AKI showed a non-significant trend toward better performance (p = 0.061) (Table 5).
Table 4.
Acceptability of the AKI clinical practice guidelines
| Guideline Aspect | Strongly Agree | Agree | Neutral | Disagree | Strongly Disagree | Mean Score (SD) |
|---|---|---|---|---|---|---|
| Guidelines are needed in the hospital | 139 (83.7%) | 24 (14.5%) | 2 (1.2%) | 0 (0.0%) | 1 (0.6%) | 4.81 (0.51) |
| Will improve provider knowledge | 147 (88.6%) | 17 (10.2%) | 1 (0.6%) | 0 (0.0%) | 1 (0.6%) | 4.86 (0.45) |
| Will improve patient outcomes | 144 (86.7%) | 20 (12.0%) | 2 (1.2%) | 0 (0.0%) | 0 (0.0%) | 4.85 (0.38) |
| Will improve healthcare service | 145 (87.3%) | 19 (11.4%) | 2 (1.2%) | 0 (0.0%) | 0 (0.0%) | 4.86 (0.37) |
| Will reduce unnecessary transfers | 142 (85.5%) | 21 (12.7%) | 2 (1.2%) | 0 (0.0%) | 1 (0.6%) | 4.82 (0.48) |
| Will follow the guidelines if available | 149 (89.8%) | 14 (8.4%) | 1 (0.6%) | 1 (0.6%) | 1 (0.6%) | 4.86 (0.49) |
| Overall Acceptability Score | 29.1 (2.1) |
Likert scale: 5 = Strongly Agree, 4 = Agree, 3 = Neutral, 2 = Disagree, 1 = Strongly Disagree overall acceptability score range: 6—30
Fig. 2.
Figure showing acceptability and willingness to use the Aki guideline
Table 5.
Factors associated factors with higher knowledge scores (multivariable linear regression)
| Variable | Coefficient | 95% CI | p value |
|---|---|---|---|
| Hospital (ref: Byumba) | |||
| Kacyiru | 0.28 | −0.52, 1.08 | 0.485 |
| Kibagabaga | 0.41 | −0.26, 1.08 | 0.228 |
| Kibuye | 0.02 | −0.53, 0.57 | 0.943 |
| Muhima | 0.11 | −0.38, 0.60 | 0.658 |
| Ruhengeri | −0.02 | −0.54, 0.50 | 0.936 |
| Professional Category (ref: Nurses) | |||
| Medical Doctors | 0.35 | −0.22, 0.92 | 0.225 |
| Pediatricians | 0.63 | −0.48, 1.74 | 0.263 |
| Other/Students | 0.08 | −0.36, 0.52 | 0.720 |
| Years of Practice | 0.01 | −0.02, 0.04 | 0.612 |
| Previous AKI Training (ref: Never) | |||
| Graduate training | 0.21 | −0.24, 0.66 | 0.358 |
| CPD training | 0.47 | −0.26, 1.20 | 0.203 |
| Comfort managing AKI (ref: Not comfortable) | |||
| Very comfortable | 0.84 | −0.04, 1.72 | 0.061 |
| Comfortable | 0.31 | −0.26, 0.88 | 0.282 |
| Less comfortable | 0.26 | −0.32, 0.84 | 0.377 |
R² = 0.087; overall model p value = 0.523
Case audit: study population and baseline characteristics
A total of 156 children aged 1 month to 14.9 years with serum creatinine ≥ 1.0 mg/dL were enrolled across six hospitals between July 2024 and June 2025 (pre-intervention n = 138; post-intervention n = 18). Mean age was 6.6 ± 5.3 years; 54.5% were male. The majority (76.3%) presented with KDIGO Stage 3 AKI. Acute gastroenteritis was the leading precipitating cause (33.3%), followed by pneumonia and sepsis (10.3% each). Baseline characteristics and KDIGO staging distribution are presented in Tables 6 and 7.
Table 6.
Baseline demographics characteristics of the patient enrolled in the case file review
| Characteristic | Overall (N = 156) | Pre-intervention (N = 138) | Post-intervention (N = 18) | P-value |
|---|---|---|---|---|
| Demographics | ||||
| Age (years), mean ± SD | 6.62 ± 5.25 | 6.26 ± 5.20 | 9.36 ± 4.97 | |
| Age category, n (%) | ||||
| Infant (1 < 12mo) | 28 (17.9%) | 28 (20.3%) | 0 (0.0%) | |
| Young child (1 < 5y) | 45 (28.8%) | 40 (29.0%) | 5 (27.8%) | |
| Older child (5 < 12y) | 48 (30.8%) | 42 (30.4%) | 6 (33.3%) | |
| Adolescent (≥ 12y) | 35 (22.4%) | 28 (20.3%) | 7 (38.9%) | |
| Gender, n (%) | ||||
| Male | 85 (54.5%) | 73 (52.9%) | 12 (66.7%) | 0.270 |
| Female | 71 (45.5%) | 65 (47.1%) | 6 (33.3%) | |
| Hospital Enrolled, n (%) | ||||
| Byumba District Hospital | 17 (10.9%) | 10 (7.2%) | 7 (38.9%) | |
| Kacyiru District Hospital | 10 (6.4%) | 9 (6.5%) | 1 (5.6%) | |
| Kibagabaga District Hospital | 27 (17.3%) | 27 (19.6%) | 0 (0.0%) | |
| Kibuye Referral Hospital | 36 (23.1%) | 35 (25.4%) | 1 (5.6%) | |
| Muhima District Hospital | 21 (13.5%) | 21 (15.2%) | 0(0.0%) | |
| Ruhengeri Level II Teaching Hospital | 45 (28.8%) | 36 (26.1%) | 9 (50.0%) | |
| KDIGO Stage (SCr criteria), n (%) | ||||
| no AKI | 1 (0.6%) | 0(0.0%) | 1 (5.6%) | < 0.001 |
| Stage 1 | 7 (4.5%) | 3 (2.2%) | 4 (22.2%) | |
| Stage 2 | 29 (18.6%) | 25 (18.1%) | 4 (22.2%) | |
| Stage 3 | 119 (76.3%) | 110 (79.7%) | 9 (50.0%) | |
| Associated Conditions, n (%) | ||||
| Acute Gastroenteritis | 52 (33.3%) | 46 (33.3%) | 6 (33.3%) | 0.631 |
| Pneumonia | 16 (10.3%) | 15 (10.9%) | 1 (5.6%) | 0.376 |
| Sepsis | 16 (10.3) | 16 (11.6%) | 0 (0.0%) | 0.094 |
| Malaria | 14 (9.0%) | 13 (9.4%) | 1 (5.6%) | 0.473 |
| Others | 36 (23.1%) | 28 (20.3%) | 8 (44.4%) | 0.062 |
| Comorbidities, n (%) | ||||
| Severe Malnutrition | 7 (5.1%) | 7 (5.1%) | 0(0.0%) | 0.268 |
| Heart Failure | 2 (3.7%) | 2 (3.7%) | 0(0.0%) | 0.588 |
| Liver Failure | 5 (3.6%) | 5 (3.6%) | 0(0.0%) | 0.367 |
Table 7.
Clinical profile of cases and associated presenting symptoms, conditions and outcomes
| Characteristic | Overall (N = 156) |
Pre-intervention (N = 138) |
Post-intervention (N = 18) |
P-value |
|---|---|---|---|---|
| PRESENTING COMPLAINTS | ||||
| Fever | 84/156 (53.8%) | 77/138 (55.8%) | 7/18 (38.9%) | 0.213 |
| Vomiting | 82/156 (52.6%) | 68/138 (49.3%) | 14/18 (77.8%) | 0.025 |
| Diarrhea | 63/156 (40.4%) | 55/138 (39.9%) | 8/18 (44.4%) | 0.800 |
| Others | 49/156 (31.4%) | 45/138 (32.6%) | 4/18 (22.2%) | 0.432 |
| Abdominal pain | 34/156 (21.8%) | 28/138 (20.3%) | 6/18 (33.3%) | 0.228 |
| Difficulty breathing | 22/156 (14.1%) | 22/138 (15.9%) | 0/18 (0.0%) | 0.078 |
| Generalised body weakness | 21/156 (13.5%) | 20/138 (14.5%) | 1/18 (5.6%) | 0.470 |
| Cough | 16/156 (10.3%) | 14/138 (10.1%) | 2/18 (11.1%) | 1.000 |
| ASSOCIATED CONDITIONS | ||||
| Acute gastroenteritis | 52/156 (33.3%) | 46/138 (33.3%) | 6/18 (33.3%) | 1.000 |
| Others | 36/156 (23.1%) | 28/138 (20.3%) | 8/18 (44.4%) | 0.035 |
| Pneumonia | 16/156 (10.3%) | 15/138 (10.9%) | 1/18 (5.6%) | 0.696 |
| Sepsis | 16/156 (10.3%) | 16/138 (11.6%) | 0/18 (0.0%) | 0.219 |
| Malaria | 14/156 (9.0%) | 13/138 (9.4%) | 1/18 (5.6%) | 1.000 |
| Traditional medication use | 11/156 (7.1%) | 9/138 (6.5%) | 2/18 (11.1%) | 0.617 |
| Shock Hypovolemic | 10/156 (6.4%) | 9/138 (6.5%) | 1/18 (5.6%) | 1.000 |
| COMORBIDITIES | ||||
| Others | 20/156 (12.8%) | 16/138 (11.6%) | 4/18 (22.2%) | 0.253 |
| Severe malnutrition | 7/156 (4.5%) | 7/138 (5.1%) | 0/18 (0.0%) | 1.000 |
| Liver failure | 5/156 (3.2%) | 5/138 (3.6%) | 0/18 (0.0%) | 1.000 |
| KDIGO STAGE | ||||
| Stage 1 | 6/154 (3.9%) | 3/138 (2.2%) | 3/16 (18.8%) | 0.015 |
| Stage 2 | 29/154 (18.8%) | 25/138 (18.1%) | 4/16 (25.0%) | 0.505 |
| Stage 3 | 119/154 (77.3%) | 110/138 (79.7%) | 9/16 (56.2%) | 0.054 |
| LABORATORY INVESTIGATIONS | ||||
| Baseline creatinine (mg/dL) | 0.3 ± 0.1 | 0.3 ± 0.1 | 0.5 ± 0.2 | |
| Peak creatinine (mg/dL) | 1.6 ± 0.9 | 1.6 ± 0.9 | 1.7 ± 0.9 | |
| Hyperkalemia (K + > 5.0 mEq/L) | 58/156 (37.2%) | 50/138 (36.2%) | 8/18 (44.4%) | 0.605 |
| Hypernatremia (Na + > 145 mEq/L) | 15/156 (9.6%) | 14/138 (10.1%) | 1/18 (5.6%) | 1.000 |
| Hyponatremia (Na + < 135 mEq/L) | 53/156 (34.0%) | 42/138 (30.4%) | 11/18 (61.1%) | 0.016 |
| OUTCOMES | ||||
| Alive | 125/156 (80.1%) | 112/138 (81.2%) | 13/18 (72.2%) | 0.358 |
| Deceased | 15/156 (9.6%) | 12/138 (8.7%) | 3/18 (16.7%) | 0.385 |
| Renal recovery status: | ||||
| Complete recovery | 27/156 (17.3%) | 23/138 (16.7%) | 4/18 (22.2%) | 0.519 |
| Partial recovery | 2/156 (1.3%) | 1/138 (0.7%) | 1/18 (5.6%) | 0.218 |
| Not assessed | 116/156 (74.4%) | 106/138 (76.8%) | 10/18 (55.6%) | 0.081 |
| LENGTH OF HOSPITALIZATION (days) | ||||
| Overall | 6.6 ± 10.4 | 7.0 ± 10.9 | 3.7 ± 4.2 | 0.249 |
| Stage 1 | 5.2 ± 5.2 | 5.3 ± 2.3 | 5.0 ± 7.8 | 0.947 |
| Stage 2 | 6.8 ± 10.0 | 7.0 ± 10.4 | 4.5 ± 3.5 | 0.745 |
| Stage 3 | 6.6 ± 10.9 | 7.0 ± 11.2 | 1.8 ± 1.7 | 0.191 |
Values are N (%) for categorical variables and mean ± SD for continuous variables
Others: refers to variables that were recorded in categories but were not prevalent above 10%; they were grouped
Significant differences between intervention periods included KDIGO staging distribution (p < 0.001), with post-intervention patients showing more Stage 1 cases (22.2% vs. 2.2%, p = 0.015) and fewer Stage 3 cases (50.0% vs. 79.7%, p = 0.054). Post-intervention patients were also older (9.4 ± 5.0 vs. 6.3 ± 5.2 years), had higher baseline creatinine (0.5 ± 0.2 vs. 0.3 ± 0.1 mg/dL, p < 0.001), and had more hyponatremia (61.1% vs. 30.4%, p = 0.016).
Quality of care assessment
We assessed nine evidence-based quality indicators across Recognition (R), Action (A), and Results Evaluation (RE) domains (Tables 8 and 9). Pre-intervention, the mean composite score was 3.0 ± 1.8 out of 9 possible points. Individual quality indicator performance varied widely: essential laboratory investigations were performed in 87.7% of cases, while KDIGO staging was documented in only 5.8% and urine output monitoring in only 3.6%.
Table 8.
Quality of care using RAR framework scores within intervention categories
| Based on Kashani et al. 2019 Recognition-Action-Results framework [24] | |||
|---|---|---|---|
| RAR Domain/Metric | Pre-intervention (N = 138) | Post-intervention (N = 18) | P-value |
| RECOGNITION SCORE (0–3) | |||
| Mean ± SD | 0.54 ± 0.80 | 0.50 ± 0.71 | 0.8270 |
| R1. AKI documented | 39 (28.3%) | 4 (22.2%) | |
| R2. AKI staged (KDIGO) | 8 (5.8%) | 1 (5.6%) | |
| R3. Baseline creatinine | 28 (20.3%) | 4 (22.2%) | |
| ACTION SCORE (0–3) | |||
| Mean ± SD | 1.07 ± 0.61 | 1.11 ± 0.47 | 0.7580 |
| A1. Urine monitoring | 5 (3.6%) | 1 (5.6%) | |
| A2. Supportive management | 21 (15.2%) | 3 (16.7%) | |
| A3. Lab investigations | 121 (87.7%) | 16 (88.9%) | |
| RESULTS SCORE (0–3) | |||
| Mean ± SD | 1.39 ± 0.84 | 1.28 ± 0.75 | 0.5867 |
| RE1. Final creatinine | 38 (27.5%) | 6 (33.3%) | |
| RE2. Survival | 126 (91.3%) | 15 (83.3%) | |
| RE3. Recovery status | 28 (20.3%) | 2 (11.1%) | |
| TOTAL RAR SCORE (0–9) | |||
| Mean ± SD | 3.00 ± 1.84 | 2.89 ± 1.49 | 0.8063 |
Table 9.
Detailed breakdown of quality indicators assessed
| Pre-intervention (n = 138) vs. Post-intervention (n = 18) | ||||||
|---|---|---|---|---|---|---|
| Quality Indicator | Pre n (%) | Post n (%) | Change | OR | 95% CI | P-value |
| INDIVIDUAL QUALITY INDICATORS (9 total) | ||||||
| R1: AKI documented in the medical record | 39/138 (28.3%) | 4/18 (22.2%) | -6.0pp | 0.73 | [0.22–2.34] | 0.7809 |
| R2: AKI staging Documented | 8/138 (5.8%) | 1/18 (5.6%) | -0.2pp | 0.96 | [0.11–8.12] | 1.0000 |
| R3: Baseline creatinine available. | 28/138 (20.3%) | 4/18 (22.2%) | + 1.9pp | 1.12 | [0.34–3.68] | 0.7654 |
| A1: Urine output monitoring measured | 5/138 (3.6%) | 1/18 (5.6%) | + 1.9pp | 1.56 | [0.17–14.20] | 0.5269 |
| A2: AKI supportive management provided | 21/138 (15.2%) | 3/18 (16.7%) | + 1.4pp | 1.11 | [0.30–4.19] | 1.0000 |
| A3: Essential laboratory investigations | 121/138 (87.7%) | 16/18 (88.9%) | + 1.2pp | 1.12 | [0.24–5.32] | 1.0000 |
| RE1: Final creatinine before discharge | 38/138 (27.5%) | 6/18 (33.3%) | + 5.8pp | 1.32 | [0.46–3.76] | 0.5879 |
| RE2: Survival to discharge | 126/138 (91.3%) | 15/18 (83.3%) | -8.0pp | 0.48 | [0.12–1.88] | 0.3850 |
| RE3: Renal recovery status documented | 28/138 (20.3%) | 2/18 (11.1%) | -9.2pp | 0.49 | [0.11–2.26] | 0.5284 |
| COMPOSITE SCORE DISTRIBUTION (0–9 points) | ||||||
| Score | Pre n (%) | Post n (%) | ||||
| 0/9 | 0 (0.0%) | 0 (0.0%) | ||||
| 1/9 | 19 (13.8%) | 2 (11.1%) | ||||
| 2/9 | 63 (45.7%) | 8 (44.4%) | ||||
| 3/9 | 18 (13.0%) | 2 (11.1%) | ||||
| 4/9 | 6 (4.3%) | 4 (22.2%) | ||||
| 5/9 | 15 (10.9%) | 0 (0.0%) | ||||
| 6/9 | 6 (4.3%) | 2 (11.1%) | ||||
| 7/9 | 9 (6.5%) | 0 (0.0%) | ||||
| 8/9 | 1 (0.7%) | 0 (0.0%) | ||||
| 9/9 | 1 (0.7%) | 0 (0.0%) | ||||
| PRIMARY COMPOSITE OUTCOME | ||||||
| COMPOSITE: Adequate quality (≥ 3/9 indicators) | 56/138 (40.6%) | 8/18 (44.4%) | + 3.9pp * | 1.17 | [0.44–3.15] | 0.8020 |
pp*: percentage point(s)
To evaluate the adequacy of laboratory monitoring, a laboratory composite score was constructed based on five AKI-relevant biochemical parameters: serum creatinine, potassium, sodium, Full Blood count and Malaria Testing. Each parameter documented during hospitalisation was assigned one point. Adequate laboratory monitoring was defined as completion of at least three of the five parameters (≥ 3/5). Post-intervention, modest improvements were observed in several indicators: final creatinine monitoring increased from 27.5% to 33.3% (+ 5.8% points), urine output monitoring from 3.6% to 5.6% (+ 1.9pp), and baseline creatinine documentation from 20.3% to 22.2% (+ 1.9pp). However, none of these individual improvements reached statistical significance (all p > 0.05).
Using a composite quality indicator defined as achievement of ≥ 3 of 9 quality indicators, adequate quality care improved from 40.6% (56/138) pre-intervention to 44.4% (8/18) post-intervention, representing a 3.9% point increase (OR 1.17, 95% CI 0.44–3.15, p = 0.802). When examining the distribution of composite scores, a notable shift occurred at the 4-point threshold, with 22.2% of post-intervention patients achieving 4/9 indicators compared to only 4.3% pre-intervention, suggesting a trend toward better performance despite a lack of statistical significance.
Renal recovery status documentation was defined according to predefined clinical and biochemical criteria. Recovery was considered present if there was a documented physician note indicating renal recovery, and/or return of the final measured serum creatinine to the patient’s known baseline value, or normalisation of serum creatinine to age-appropriate reference ranges when baseline values were unavailable. Patients who did not meet these criteria were categorised as having no documented recovery at discharge.
Partial recovery was defined as serum creatinine remaining above baseline at discharge.
AKI documented in medical records refers to cases where the treating clinician explicitly recognises (and notes) AKI in the patient notes and initiates at least one AKI-directed intervention (ideally items explicitly mentioned in the KDIGO guideline and/or taught in the AKI CPD event). Final creatinine control was defined as documentation of a repeat serum creatinine measurement obtained before discharge to assess improvement in kidney function. All recovery assessments were based on the serum creatinine trajectory and clinician notes recorded during the hospitalisation period up to the day of discharge.
Mortality and predictors
Overall mortality was 9.6% (15/156), with 12 deaths (8.7%) in the pre-intervention period and 3 deaths (16.7%) post-intervention (p = 0.385). Among the 15 deaths, 73.3% were male, 46.7% were young children aged 1–5 years, and 80.0% had KDIGO Stage 3 AKI at presentation. Significant independent predictors of mortality included hypovolemic shock (OR 8.73, 95% CI 2.13–35.76, p = 0.008), severe dehydration (OR 5.29, 95% CI 1.57–17.86, p = 0.009), and hypernatremia (OR 4.18, 95% CI 1.02–17.06, p = 0.045).
Protective factors included older age (5–12 years: OR 0.14, p = 0.038), some dehydration status (OR 0.13, p = 0.021), and care at Kibuye Referral Hospital (OR 0.00, p = 0.022). In Mantel-Haenszel stratified analysis adjusting for hospital, both severe dehydration (adjusted OR 4.14) and hypernatremia (adjusted OR 3.99) remained independent predictors, demonstrating effects beyond facility-level confounding.
Complete renal recovery was documented in only 17.3% of survivors (27/156), with partial recovery in 1.3%; recovery status was not assessed in 74.4% of cases, representing a major gap in discharge care (Table 10).
Table 10.
Bivariate analysis of factors associated with in-hospital mortality among children with AKI
| Variable Category | Deceased n/15 (%) |
95% CI | Alive n/141 (%) |
Odds Ratio | 95% CI | P-value |
|---|---|---|---|---|---|---|
| Sex | ||||||
| Male | 11/15 (73.3%) | [0.0%-20.4%] | 65/141 (52.0%) | 2.54 | [0.77–8.40] | 0.1703 |
| Female | 4/15 (26.7%) | [10.9%-52.0%] | 59/141 (47.2%) | 0.41 | [0.12–1.35] | 0.1727 |
| Age Category | ||||||
| Infant (1–12 mo) | 5/15 (33.3%) | [15.2%-58.3%] | 20/141 (16.0%) | 2.62 | [0.81–8.50] | 0.1452 |
| Young child (1–5 y) | 7/15 (46.7%) | [24.8%-69.9%] | 35/141 (28.0%) | 2.25 | [0.76–6.67] | 0.1466 |
| Older child (5–12 y) | 1/15 (6.7%) | [1.2%-29.8%] | 42/141 (33.6%) | 0.14 | [0.02–1.11] | 0.0377 |
| Adolescent (≥ 12 y) | 2/15 (13.3%) | [3.7%-37.9%] | 28/141 (22.4%) | 0.53 | [0.11–2.50] | 0.5252 |
| KDIGO Stage | ||||||
| Stage 1 | 1/15 (6.7%) | [1.2%-29.8%] | 4/141 (3.2%) | 2.16 | [0.23–20.71] | 0.4375 |
| Stage 2 | 2/15 (13.3%) | [3.7%-37.9%] | 26/141 (20.8%) | 0.59 | [0.12–2.76] | 0.7354 |
| Stage 3 | 12/15 (80.0%) | [54.8%-93.0%] | 93/141 (74.4%) | 1.38 | [0.36–5.19] | 0.7614 |
| Hydration Status | ||||||
| Well hydrated | 3/15 (20.0%) | [7.0%-45.2%] | 29/141 (23.2%) | 0.83 | [0.22–3.13] | 1.0000 |
| Mild- Moderate dehydration | 1/15 (6.7%) | [1.2%-29.8%] | 45/141 (36.0%) | 0.13 | [0.02-1.00] | 0.0213 |
| Severe dehydration | 6/15 (40.0%) | [19.8%-64.3%] | 14/141 (11.2%) | 5.29 | [1.64–17.08] | 0.0088 |
| Shock | ||||||
| Any type of shock | 4/15 (26.7%) | [10.9%-52.0%] | 6/141 (4.8%) | 7.21 | [1.76–29.48] | 0.0125 |
| Hypovolemic shock | 4/15 (26.7%) | [10.9%-52.0%] | 5/141 (4.0%) | 8.73 | [2.04–37.30] | 0.0080 |
| Septic/Cardiogenic shock | 1/15 (6.7%) | [1.2%-29.8%] | 1/141 (0.8%) | 8.86 | [0.52-149.55] | 0.2035 |
| Sepsis | ||||||
| Sepsis | 1/15 (6.7%) | [1.2%-29.8%] | 14/141 (11.2%) | 0.57 | [0.07–4.64] | 1.0000 |
| Electrolyte Abnormalities | ||||||
| Hyperkalemia (K + > 5.0 mEq/L) | 7/15 (46.7%) | [24.8%-69.9%] | 45/141 (36.0%) | 1.56 | [0.53–4.57] | 0.4143 |
| Hypernatremia (Na + > 145 mEq/L) | 4/15 (26.7%) | [10.9%-52.0%] | 10/141 (8.0%) | 4.18 | [1.12–15.56] | 0.0452 |
| Hyponatremia (Na + < 135 mEq/L) | 4/15 (26.7%) | [10.9%-52.0%] | 43/141 (34.4%) | 0.69 | [0.21–2.31] | 0.7734 |
| Intervention Period | ||||||
| Pre intervention | 12/15 (80.0%) | [54.8%-93.0%] | 112/141 (89.6%) | 0.46 | [0.12–1.86] | 0.3806 |
| Post intervention | 3/15 (20.0%) | [7.0%-45.2%] | 13/141 (10.4%) | 2.15 | [0.54–8.64] | 0.3806 |
| ADJUSTED ANALYSIS (Mantel-Haenszel, stratified by hospital) | ||||||
|---|---|---|---|---|---|---|
| Variable | Crude OR | Crude 95% CI | Adjusted OR | Interpretation | ||
| Severe dehydration | 4.56 | [1.32–15.78] | 4.14 | Independent predictor | ||
| Hypernatremia (Na + > 145) | 4.30 | [1.02–18.06] | 3.99 | Independent predictor | ||
Discussion
This study provides the first comprehensive assessment of pediatric AKI care quality in Rwanda using a structured quality indicator framework, revealing substantial gaps across all domains of recognition, action, and results. KDIGO staging was performed in only 5.8% of pre-intervention cases and urine output monitoring in 3.6%, despite essential laboratory investigations being completed in 87.7%; a fundamental investigation-action disconnect. The composite adequate-care rate rose from 40.6% to 44.4% post-intervention (OR 1.17, 95% CI 0.44–3.15, p = 0.802), a null result attributable to critical under-powering with only 18 post-intervention cases (estimated power < 8%). Nonetheless, a five-fold increase in patients achieving 4/9 quality indicators (4.3% to 22.2%) is consistent with the ‘early adoption’ phase described in implementation science, where single-session training produces initial behaviour change in a minority of providers before broader dissemination [11, 25]. Achieving comprehensive practice change will require repeated reinforcement, audit-and-feedback cycles, and integration into routine clinical supervision [25, 26].
The overall mortality of 9.6% compares favourably with the pooled international rate of 18.27% (95% CI 14.89–21.65) among hospitalized children with AKI [27]. Mortality was almost entirely driven by modifiable clinical factors. Hypovolemic shock was the strongest predictor (OR 8.73, 95% CI 2.04–37.30, p = 0.008), consistent with the dominance of dehydration-associated pre-renal injury in African children [1]. The similar shock prevalence pre- and post-intervention (6.5% vs. 5.6%) confirms that a recognition-focused training programme alone cannot address delays in volume resuscitation. Severe dehydration (OR 5.29) and hypernatremia (OR 4.18) remained independent mortality predictors after Mantel-Haenszel adjustment for hospital site (adjusted ORs 4.14 and 3.99), confirming true patient-level risk beyond facility confounding. Hypernatremic dehydration carries distinct risk through cellular dehydration, cerebral vascular injury, and the complexity of safe electrolyte correction, particularly in settings without neurological monitoring capacity [28]. The protective association of mild-to-moderate dehydration (OR 0.13, p = 0.021) likely reflects a critical intervention window: children with partial volume depletion still respond to rehydration, whereas those presenting with severe dehydration or shock may arrive too late for recovery [29]. Promoting earlier care-seeking and primary-level triage of dehydrated children with elevated creatinine may therefore be the highest-yield intervention in this setting.
The overwhelming majority (76.3%) presented with KDIGO Stage 3 AKI, a pattern consistently reported across sub-Saharan Africa where Stage 3 disease predominates (e.g., 68% in a Nigerian series), contrasting sharply with high-income settings where Stage 1 typically predominates [1, 30]. This reflects limited community awareness, out-of-pocket costs, and recourse to traditional medicine, used by 7.1% of our cohort prior to admission [31]. The post-intervention shift toward earlier KDIGO staging (Stage 1: 22.2% vs. 2.2%, p = 0.015) is encouraging but must be interpreted cautiously given the older age and higher baseline creatinine of post-intervention patients, which preclude confident causal attribution to the educational intervention.
The overconfidence gap 45.2% of providers felt comfortable managing pediatric AKI, yet the mean knowledge was only 42% carries direct patient safety implications, potentially delaying recognition of knowledge limitations and discouraging consultation [17]. Knowledge deficits were uniform across all professional categories and uncorrelated with years of practice (R²=0.087), confirming that experience alone does not drive AKI competency. Poor identification of intrinsic AKI causes (33.7%), particularly HUS a leading cause in sub-Saharan Africa risks inappropriate management and delayed referral [17]. The exceptionally high guideline acceptability (mean 4.8/5.0) creates an optimal foundation for sustained improvement, but attitudinal readiness must be matched by system-level capacity and repeated structured training.
Finally, the 74.4% rate of undocumented renal recovery at discharge represents a critical continuity-of-care gap, particularly given emerging evidence that AKI survivors face elevated long-term risks of CKD, hypertension, and cardiovascular complications, with a post-discharge mortality of 6.84% reported in recent meta-analysis [27].
Limitations
Several limitations should be considered when interpreting these findings. First, case identification relied on an operational creatinine threshold of ≥ 1.0 mg/dL, which maximized specificity for clinically relevant AKI but reduces sensitivity in younger children. Infants with a baseline creatinine of ~ 0.3 mg/dL could meet KDIGO criteria for AKI with a rise to 0.6 mg/dL yet would not have been captured by our screen, likely underestimating the true AKI burden in the youngest age groups [32].
Second, baseline serum creatinine was unavailable for most patients, necessitating estimation using the Pottel height-independent equation. Although validated in pediatric populations and applied uniformly across all cases to minimize systematic bias, the equation was developed in European cohorts and no Rwanda-specific reference values currently exist. Differences in nutritional status and body composition may have introduced some degree of misclassification in AKI staging.
Third, urine output data were inconsistently documented across sites and could not be incorporated into KDIGO staging. Reliance on serum creatinine criteria alone may have led to underrecognition of AKI severity in some patients, though the ongoing revision of KDIGO guidelines is expected to reassess the practical relevance of urine output criteria in resource-limited settings. (https://kdigo.org/wp-content/uploads/2023/10/KDIGO-AKI-Guideline_Scope-of-Work_25Oct2023_Final.pdf).
Fourth, documentation quality and laboratory availability varied across district hospitals. Management actions such as nephrotoxin discontinuation or fluid optimization may have been performed but not recorded, potentially underestimating true adherence to recommended care.
Fifth, the knowledge level thresholds used (≥ 80% excellent, 60–79% good, 40–59% fair, < 40% poor) are commonly applied in competency surveys but are inherently arbitrary. Small score variations could shift respondents between categories without reflecting meaningful differences in clinical competence.
Finally, with only 15 in-hospital deaths, multivariable mortality analyses are limited in precision. The small post-intervention sample (n = 18) and short follow-up (~ 3 months) constrain statistical power and may not capture sustained behavioural change. Retrospective case identification and the possibility that post-training awareness improved documentation independent of actual care quality are additional sources of potential bias.
Recommendations
On the basis of these findings, we recommend: implementing risk-stratified surveillance systems targeting patients with high-risk presentations; establishing institutional requirements for core quality processes (AKI staging > 80%, urine output monitoring > 90%, serial creatinine assessment > 80%) with performance monitoring; developing structured discharge and follow-up protocols to address the 74% gap in recovery status documentation; moving beyond one-time training to incorporate quarterly refresher sessions, case-based learning, simulation, and mentorship; integrating comprehensive pediatric AKI curricula into preservice training; and establishing continuous quality improvement infrastructure with 3–6-monthly case reviews and institutional dashboards. A systematic programme of creatinine measurement and nephrology follow-up at 3 months post-discharge should be established to capture the AKI-to-CKD transition.
Conclusion
Paediatric AKI in Rwanda carries a substantial burden characterised by late presentation, profound care quality deficits, and mortality concentrated in high-risk clinical phenotypes and a single high-mortality facility. A brief educational intervention showed encouraging distributional shifts but was underpowered to demonstrate statistically significant quality improvement. Mortality is driven primarily by hypovolemic shock and severe dehydration — factors that are clinically modifiable with prompt, protocolised resuscitation. Persistent staging, monitoring, and follow-up gaps underscore that training must extend beyond initial recognition to encompass the complete care continuum. The exceptional baseline guideline acceptability provides a promising foundation for sustained improvement, suggesting Rwanda’s healthcare system is well positioned for progress if training and system-strengthening initiatives receive sustained institutional support.
Supplementary Information
Acknowledgements
We thank the CHUK Directorate of Research and Education for administrative support, the staff of participating district hospitals for their collaboration, and the healthcare providers who participated in the knowledge assessment. Special thanks to the data collectors for their dedicated field work. Our gratitude goes also to Prof Martin Bitzan for his contribution to clinical pediatric nephrology in Rwanda through International society of Nephrology (ISN) sister renal center initiative.
Abbreviations
- AKI
Acute Kidney Injury
- CHUK
University Teaching Hospital of Kigali
- CPD
Continuous Professional Development
- DH
District Hospital
- IQR
Interquartile Range
- KDIGO
Kidney Disease: Improving Global Outcomes
- L2TH
Level Two Teaching Hospital
- OR
Odds Ratio
- pRIFLE
Pediatric Risk, Injury, Failure, Loss, and End–stage Renal Disease
- RH
Referral Hospital
- SD
Standard Deviation
- SE
Standard Error (if present in tables)
- CI
Confidence Interval
- HUS
Hemolytic Uremic Syndrome
- AKIN
Acute Kidney Injury Network
Authors’ contributions
- **Vainqueur Ineza Habyarimana: ** Conceptualization, study design, data collection supervision, data analysis, manuscript drafting.- **Gilbert Rugamba: **Principal Investigator, Conceptualization, study oversight, methodology refinement, manuscript review.- **Co-authors: Jean Claude Ntiyamira, Febronie Mushimiyimana, Juliette Unyuzumutima, Faustine Agaba, Janvier Hitayezu, Oswald Habyarimana: **Conceptualization Data interpretation, critical revisions of the manuscript.- **Martin Bitzan: **Supervision, Conceptualisation, Manuscript review. All authors read and approved the final manuscript.
Funding
This project was supported by a small grant from the University Teaching Hospital of Kigali (CHUK) The funders had no role in study design, data collection, analysis, interpretation, or manuscript preparation.
Data availability
The datasets generated and analyzed during this study are included within the published article and its additional files. De-identified data may also be made available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
This study was reviewed and approved by the CHUK Institutional Review Board (approval number: CHUKIRB/2024/015) and Administrative authorization was also granted from Rwanda Biomedical Centre (ref no 2464/RBC/2025) and all participating district hospitals Ethics committee.
All procedures involving human participants were conducted in accordance with the ethical standards of the institutional and national research committees and in line with the Declaration of Helsinki and its subsequent revisions. Informed consent was obtained from all healthcare providers who participated in the knowledge and attitudes survey.
Given the retrospective nature of the case record review and the use of de-identified data, the requirement for informed consent to participate was waived by the CHUK Institutional Review Board and Participating hospitals ethics committee. All the data were anonymized and stored securely, with access limited to authorized research personnel.
Consent for publication
Not applicable. No identifiable individual data are presented in this manuscript.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Olowu WA, Niang A, Osafo C, Ashuntantang G, Arogundade FA, Porter J, et al. Outcomes of acute kidney injury in children and adults in sub-Saharan Africa: a systematic review. Lancet Glob Heal. 2016;4(4):e242–50. Available from: 10.1016/S2214-109X(15)00322-8. [DOI] [PubMed]
- 2.Kaddourah A, Basu RK, Bagshaw SM, Goldstein SL. Epidemiology of Acute Kidney Injury in Critically Ill Children and Young Adults. N Engl J Med. 2017;376(1):11–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Vasudevan A, Phadke K, Yap HK. Peritoneal dialysis for the management of pediatric patients with acute kidney injury. Pediatr Nephrol. 2017;32(7):1145–56. [DOI] [PubMed] [Google Scholar]
- 4.Cleto-Yamane TL, Gomes CLR, Suassuna JHR, Nogueira PK. Acute Kidney Injury Epidemiology in pediatrics. J Bras Nefrol. 2019;41(2):275–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Robinson CH, Iyengar A, Zappitelli M. Early recognition and prevention of acute kidney injury in hospitalised children. Lancet Child Adolesc Heal. 2023;7(9):657–70. Available from: 10.1016/S2352-4642(23)00105-0. [DOI] [PubMed]
- 6.Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120(4):c179–84. https://kdigo.org/wp-content/uploads/2016/10/KDIGO-2012-AKI-Guideline-English.pdf. [DOI] [PubMed] [Google Scholar]
- 7.Lunyera J, Kilonzo K, Lewington A, Yeates K, Finkelstein FO. Acute Kidney Injury in Low-Resource Settings: Barriers to Diagnosis, Awareness, and Treatment and Strategies to Overcome These Barriers. Am J kidney Dis Off J Natl Kidney Found. 2016;67(6):834–40. [DOI] [PubMed] [Google Scholar]
- 8.Crews DC, Bello AK, Saadi G. Burden, access, and disparities in kidney disease. J Nephrol Italy. 2019;32:1–8. [DOI] [PubMed]
- 9.Xu G, Baines R, Westacott R, Selby N, Carr S. An educational approach to improve outcomes in acute kidney injury (AKI): report of a quality improvement project. BMJ Open. 2014;4(3):e004388. 10.1136/bmjopen-2013-004388. PMID: 24650804; PMCID: PMC3963099. [DOI] [PMC free article] [PubMed]
- 10.Roy JP, Devarajan P. Acute Kidney Injury: Diagnosis and Management. Indian J Pediatr. 2020;87(8):600–7. [DOI] [PubMed] [Google Scholar]
- 11.Molitoris BA, Levin A, Warnock DG, Joannidis M, Mehta RL, Kellum JA, et al. Improving outcomes from acute kidney injury. J Am Soc Nephrol. 2007;18:1992–4. JASN. United States. [DOI] [PubMed]
- 12.Cabana MD, Rand CS, Powe NR, Wu AW, Wilson MH, Abboud PA, et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999;282(15):1458–65. [DOI] [PubMed] [Google Scholar]
- 13.Binagwaho A, Farmer PE, Nsanzimana S, Karema C, Gasana M, de Dieu Ngirabega J, et al. Rwanda 20 years on: investing in life. Lancet (London England). 2014;384(9940):371–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Rwanda Ministry of Health. Health Sector Policy. 2015;41. Available from: http://www.moh.gov.rw/fileadmin/templates/policies/Health_Sector_Policy___19th_January_2015.pdf.
- 15.Odiit A. Acute Kidney Injury; its mortality rate in septic and critically ill children at Kigali University Teaching Hospital, Rwanda. Afr J Paediatr Nephrol. 2016;3:2016. [Google Scholar]
- 16.Hoste EAJ, Kellum JA, Selby NM, Zarbock A, Palevsky PM, Bagshaw SM, et al. Global epidemiology and outcomes of acute kidney injury. Nat Rev Nephrol. 2018;14(10):607–25. [DOI] [PubMed] [Google Scholar]
- 17.Igiraneza G, Dusabejambo V, Finklestein FO, Rastegar A. Challenges in the Recognition and Management of Acute Kidney Injury by Hospitals in Resource-Limited Settings. Kidney Int Rep. 2020;5(7):991–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chuang GT, Tsai IJ, Tsau YK. Serum Creatinine Reference Limits in Pediatric Population-A Single Center Electronic Health Record-Based Database in Taiwan. Front Pediatr. 2021;9:793446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bouchard J. Estimating Baseline Serum Creatinine for Assessing Acute Kidney Injury: Not a One Size Fits All Approach. Kidney Int Reports. 2021;6(3):562–4. Available from: 10.1016/j.ekir.2021.01.030. [DOI] [PMC free article] [PubMed]
- 20.Pottel H, Hoste L, Martens F. A simple height-independent equation for estimating glomerular filtration rate in children. Pediatr Nephrol. 2012;27(6):973–9. [DOI] [PubMed] [Google Scholar]
- 21.Blufpand HN, Westland R, van Wijk JAE, Roelandse-Koop EA, Kaspers GJL, Bökenkamp A. Height-Independent Estimation of Glomerular Filtration Rate in Children: An Alternative to the Schwartz Equation. J Pediatr. 2013;163(6):1722–7. Available from: https://www.sciencedirect.com/science/article/pii/S0022347613010032. [DOI] [PubMed]
- 22.Batte A, Starr MC, Schwaderer AL, Opoka RO, Namazzi R, Phelps Nishiguchi ES, et al. Methods to estimate baseline creatinine and define acute kidney injury in lean Ugandan children with severe malaria: a prospective cohort study. BMC Nephrol. 2020;21(1):417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ibrahim OR, Afolayan FM, Alao MA, Mohammed B, Suleiman BM, Adedoyin OT. Impact of methods of estimating baseline Serum Creatinine (bSCr) on the incidence and outcomes of acute kidney injury in childhood severe malaria. Egypt Pediatr Assoc Gaz [Internet]. 2023;71(1):36. Available from: 10.1186/s43054-023-00186-7.
- 24.Kashani K, Rosner MH, Haase M, Lewington AJP, O’Donoghue DJ, Wilson FP, et al. Quality Improvement Goals for Acute Kidney Injury. Clin J Am Soc Nephrol. 2019;14(6):941–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Rowe AK, de Savigny D, Lanata CF, Victora CG. How can we achieve and maintain high-quality performance of health workers in low-resource settings? Lancet (London, England). 2005;366(9490):1026–35. [DOI] [PubMed]
- 26.Ameh CA, Kerr R, Madaj B, Mdegela M, Kana T, Jones S, et al. Knowledge and Skills of Healthcare Providers in Sub-Saharan Africa and Asia before and after Competency-Based Training in Emergency Obstetric and Early Newborn Care. PLoS ONE. 2016;11(12):e0167270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zarei H, Azimi A, Ansarian A, Raad A, Tabatabaei H, Roshdi Dizaji S, et al. Incidence of acute kidney injury-associated mortality in hospitalized children: a systematic review and meta-analysis. BMC Nephrol. 2025;26(1):117. Available from: 10.1186/s12882-025-04033-2. [DOI] [PMC free article] [PubMed]
- 28.Moritz ML, Ayus JC. Management of hyponatremia in various clinical situations. Curr Treat Options Neurol. 2014;16(9):310. [DOI] [PubMed] [Google Scholar]
- 29.Zieg J, Narla D, Gonsorcikova L, Raina R. Fluid management in children with volume depletion. Pediatr Nephrol. 2024;39(2):423-34. 10.1007/s00467-023-06080-z. Epub 2023 Jul 14. PMID: 37452205. [DOI] [PubMed]
- 30.Jetton JG, Boohaker LJ, Sethi SK, Wazir S, Rohatgi S, Soranno DE, et al. Incidence and outcomes of neonatal acute kidney injury (AWAKEN): a multicentre, multinational, observational cohort study. Lancet Child Adolesc Heal. 2017;1(3):184–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Naicker S, Plange-Rhule J, Tutt RC, Eastwood JB. Shortage of healthcare workers in developing countries–Africa. Ethn Dis. 2009;19(1 Suppl 1):S1–60. [PubMed] [Google Scholar]
- 32.Koeze J, Keus F, Dieperink W, van der Horst ICC, Zijlstra JG, van Meurs M. Incidence, timing and outcome of AKI in critically ill patients varies with the definition used and the addition of urine output criteria. BMC Nephrol. 2017;18(1):70. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and analyzed during this study are included within the published article and its additional files. De-identified data may also be made available from the corresponding author upon reasonable request.