Prevalence and factors associated with delayed surgery among children with Wilms tumor at Mulago Hospital: a mixed-method study

Abstract

Introduction

Wilms tumor (WT) is a highly treatable form of childhood cancer, with five-year survival rates greater than 90% in high-income countries. In low- and middle-income countries, survival is still low. Delayed surgery is one of the leading factors for poor survival, as it leads to tumor complications, making subsequent treatment more complex. However, there is a paucity of information in Uganda on the prevalence and factors associated with delayed surgery. This study aimed to determine the prevalence and factors associated with delayed surgery, and to explore the barriers and facilitators of timely surgery among children with WT at Mulago Hospital.

Methods

The study employed a convergent concurrent mixed-methods design. The quantitative component involved a retrospective cross-sectional design using 261 patient records selected through consecutive sampling and reviewed via a data abstraction tool. Modified Poisson regression was used to assess associated factors. The qualitative component included in-depth interviews with 10 healthcare workers and 10 caregivers selected purposively and was analyzed via inductive thematic analysis.

Results

The prevalence of delayed surgery was 63.6% (95% confidence interval (CI) 57.0-68.7). The year of diagnosis (2021: adjusted prevalence ratio (aPR) 2.26, 95% CI 1.40–3.65, p value 0.001; 2022: aPR 1.78, 95% CI 1.07–2.99, p value 0.026; and 2023: aPR 2.01, 95% CI 1.24–3.25, p value 0.004), tumor laterality (aPR 1.41, 95% CI 1.07–1.85, p value 0.014), hemoglobin level after preoperative chemotherapy (POPC; aPR 1.22, 95% CI 1.05–1.51, p value 0.032), and chemotherapy regimen (VAD: aPR 1.32, 95% CI 1.11–1.57, p value 0.02; and AV/CE: aPR 1.46, 95% CI 1.07–1.99) were associated with delayed surgery. Qualitative findings revealed systemic, patient-level, chemotherapy, and tumor-related barriers to and facilitators of timely surgery.

Conclusion

The prevalence of delayed surgery was high among children with WT. Delayed surgery results from a complex interplay of clinical, systemic, and patient-related factors. Addressing barriers at both the institutional and patient levels may help reduce surgical delays and improve outcomes.

Introduction

Wilms tumor (WT) is the most common childhood renal cancer, accounting for 5.9% of childhood cancers [1]. The age-standardized rate worldwide is 7.5 per million person-years for children aged 0–14 and 0.3 per million person-years for those aged 15–19 [2]. Hospital-based reports from tertiary referral centers in Africa show that 4 to 10 new WT cases are diagnosed annually [3]. However, hospital-based data may underestimate the true incidence in Africa, as some affected children may not receive hospital care; thus, the actual incidence remains unclear [4].

Advances in WT management from surgery alone to multimodal treatment have improved overall survival over the years [5]. The sequence of multimodal therapies depends on the specific protocol used. The Société Internationale d’Oncologie Pédiatrique (SIOP) protocol begins with neoadjuvant chemotherapy, followed by surgery, postoperative chemotherapy, and radiotherapy [6]. In contrast, the Children’s Oncology Group (COG) recommends upfront surgery [7]. This treatment progress has achieved survival rates > 90% for WT in high-income countries, whereas rates in low- and middle-income countries (LMICs) range from 20% to 50% [8, 9]. The disparity is largely due to late presentation, treatment abandonment, and surgical delays [8].

Despite multimodal treatment strategies, surgery remains a critical component for local control of the primary tumor, preventing metastatic spread, and tumor staging [10]. However, in LMICs, pediatric patients often experience delays in receiving surgery [11]. In Tanzania, only 30% of patients underwent surgery within one week of completing preoperative chemotherapy (POPC) [12]. In Uganda, 37.5% of pediatric surgical admissions are postponed, and more than three-quarters of these patients are asked to return after more than a month [11]. Studies in Uganda and Somalia using the three delays framework revealed that limited theatre space, workforce shortages, equipment gaps, and residence significantly impacted access to surgical services [11, 13].

Delayed surgery substantially worsens treatment outcomes for children with WT, leading to lower survival rates [14]. Specifically, in Uganda, WT survival rates range from 44% to 59%, with delayed surgery identified as a primary factor in these low rates [15, 16]. However, the prevalence of delayed surgery and its associated factors in pediatric WT patients have not yet been studied. This study aimed to determine the prevalence of delayed surgery, identify related factors, and explore the barriers and facilitators of timely surgery among pediatric WT patients at the largest tertiary hospital in Uganda.

Methods

Study design

This study used a convergent concurrent mixed methods approach, where both quantitative and qualitative data were collected simultaneously, analysed separately, and then the results were merged. A retrospective cross-sectional design was used for the quantitative component, and an exploratory descriptive qualitative study design was used for the qualitative component.

Study setting

The study was conducted at the Pediatric Haematology and Oncology Unit of Mulago Hospital in Kampala, Uganda. This is the largest tertiary hospital in Uganda and a national referral hospital, with an official bed capacity of 1790. The hospital’s surgical department has a total of 22 operating theatres distributed among different surgical departments. Approximately 20–30 surgical operations are conducted from the main operating theatre daily. The Pediatric Oncology unit offers both outpatient and inpatient services with a bed capacity of 30. Patients receive oncology services at no cost. This has been achieved by the collaborative effort of the Global HOPE program with Mulago National Referral Hospital (MNRH). On average, 50 children with Wilms tumor are diagnosed and treated every year.

Study population

The quantitative component included medical records of patients aged ≤ 18 years who were diagnosed with WT between January 2019 and December 2024. Records of patients aged ≤ 18 years who died before starting POPC, and those with missing data on the date of surgery, were excluded. The qualitative component included healthcare workers (HCWs) involved in the care of children with WT, pediatric WT patients, and caregivers of pediatric WT patients whose scheduled surgery was delayed and who were undergoing postoperative chemotherapy and/or radiotherapy or who were scheduled for medical reviews. Healthcare workers and caregivers who were not available at the pediatric oncology unit during the data collection period were excluded from the interviews.

Sample size and sampling strategy

The sample size for the objective of determining the prevalence of delayed surgery was calculated via the Kish–Leslie formula. Considering a study by Concepcion and authors that reported a prevalence of delayed surgery of 83% [13], along with a 5% sampling error and a 20% adjustment for missing data, a sample size of 261 was determined. To identify factors associated with delayed surgery, the formula for comparing two proportions was employed. Proportions for children under 10 years of age and those aged 10–19 years were obtained from a study by Adugna and authors [17], adjusted for 20% missing data, resulting in a sample size of 90. Using the largest sample size, the overall sample size for the quantitative component was set at 261 participants. Consecutive sampling was used to select the 261 files. Additionally, 10 healthcare workers and 10 caregivers were purposively selected with maximum variation for interviews, and data saturation was achieved. Maximum variation considered unit, cadre, and years of experience for health care workers, and sex, relationship with the child, residence, level of education, and occupation for caregivers.

Operational definitions

Delayed surgery was defined as surgery occurring ≥ 6 weeks and ≥ 9 weeks from the start of neoadjuvant chemotherapy for localized and metastatic disease, respectively, for pediatric Wilms tumor patients aged 1–18 years. For pediatric Wilms tumor patients younger than 12 months, delayed surgery was defined as surgery occurring 2 weeks or more after the diagnosis of Wilms tumor (resource-adapted International Society of Pediatric Oncology protocol at the pediatric oncology unit, Mulago hospital). Wilms tumor was defined as a confirmed diagnosis on the basis of abdominal CT scan findings and, where available, histologically following biopsy.

Study procedure

The quantitative data were collected via a piloted paper data abstraction tool. The variables included sociodemographics, clinical characteristics, and chemotherapy-related characteristics. The abstracted data were double-entered into EpiData 3. Qualitative data were collected via in-depth interviews through a semistructured interview guide. The interviews were conducted in English and Luganda, audio-recorded with consent, and transcribed verbatim. The Luganda audio was translated to English at the time of transcription. To mitigate potential loss, the data were backed up on an external hard drive and a Google drive.

Data analysis

The quantitative data were analysed via STATA version 17 (STATA Corp, College Station, TX). Continuous variables are expressed as medians and interquartile ranges. Categorical data are summarized as percentages and frequencies. The prevalence of delayed surgery was calculated as the proportion of participants who experienced delay out of the total number, along with its 95% confidence interval. Bivariable and multivariable analyses were performed via modified Poisson regression with robust standard errors. All variables with a p value < 0.2 in the bivariable analysis were included in the multivariable analysis.

The manual stepwise backwards method was used to remove the least significant variables. Interaction was evaluated by adding product terms of the remaining significant variables and applying the log-likelihood ratio test to compare the reduced and full models for any significant interaction. Confounding was assessed by at least a 10% change in the crude and adjusted prevalence ratios. Statistical significance for the final model was set at a p value < 0.05. Qualitative analysis was conducted via Open Code version 4.03, following the six steps outlined by Braun and Clarke in 2013.

Results

Study profile and sociodemographic characteristics of the participants

A total of 266 patient records were reviewed, 5 records were excluded. Among the excluded records, 3 patients had died before starting POPC, and 2 had missing data on the date of surgery. The median age of the children was 3 years (IQR 2–5), with 50.2% of the children being male. Patients travelled a median distance of 66 km to reach the health facility, and over half of the patients resided in rural areas (57.5%, n = 150). Most of the children (64.4%, n = 168) were from central Uganda. Approximately one-third of the children were Catholic (33.2%, n = 85), and 60.6% (n = 80) of their caregivers were employed in informal jobs (Table 1).

Table 1.

Sociodemographic characteristics of 261 children diagnosed with Wilms tumor at Mulago Hospital between Jan 2019 and December 2024

Variable	Category	n	%	Median(IQR)
Age in Years				3 (2–5)
Sex of the Child	Male	131	50.2
	Female	130	49.8
Residence	Urban	111	42.5
	Rural	150	57.5
Distance from health facility (km)				66 (14–188)
Region of origin	Central	168	64.4
	Western	29	11.1
	Eastern	47	18.0
	Northern	17	6.5
Religion (n = 256)	Anglican	71	27.7
	Catholic	85	33.2
	Muslim	60	23.4
	Pentecostal	29	11.3
	Others#	11	4.3
Caregiver’s occupation (n = 132)	Unemployed	31	23.5
	Informal Job	80	60.6
	Formal Job	21	15.9

Abbreviation: IQR Interquartile range

# Seventh Day Adventist 9 (3.5%) and Jehovah witness 2 (0.7%)

Clinical characteristics

The highest and lowest numbers of cases were recorded in 2024 (21.5%, n = 56) and 2019 (11.5%, n = 30), respectively. The most common presenting symptom was abdominal distension (64.4%, n = 168). More than half of the participants (59.0%, n = 154) had comorbidities. Among those with comorbidities, the most common comorbidity was secondary hypertension (39.0%, n = 60). The median Hb, white blood cell (WBC), and platelet counts after preoperative chemotherapy were 9.44 g/dl, 6.88 × 10^9 l and 437 × 10^9 l, respectively (Table 2).

Table 2.

Clinical characteristics of 261 children diagnosed with Wilms tumor at Mulago Hospital between Jan 2019 and December 2024

Variable	Category	n	%	Median(IQR)
Year of Diagnosis	2019	30	11.5
	2020	52	19.9
	2021	35	13.4
	2022	39	14.9
	2023	49	18.7
	2024	56	21.5
Presenting symptoms	Abdominal distension	168	64.4
	Abdominal pain	4	1.5
	Hematuria	8	3.1
	Others*	81	31.0
Presence of comorbidities	No	107	41.0
	Yes	154	59.0
Comorbidity type (n = 154)	Hypertension	60	39.0
	Malnutrition	45	29.2
	Hypertension and malnutrition	47	30.5
	Others#	2	1.3
Hb level after POPC(g/dl)				9.44(8.3–10.2)
WBC after POPC(x10^9l)				6.87(4.71-10)
Platelet after POPC(x10^9l)				437(311–550)

Abbreviations: IQR Interquartile range, Hb Hemoglobin, POPC Preoperative chemotherapy, WBC White blood cell

*- Abdominal distension and pain (n = 64), abdominal distension and hematuria (n = 7), abdominal pain and hematuria (n = 10), and #-HIV (n = 1), pneumonia (n = 1)

Tumor and chemotherapy-related characteristics

Most tumors (94%, n = 245) were unilateral. Among the unilateral tumors, 46.1% (n = 113) and 53.9% (n = 132) were located in the right and left kidneys, respectively. The majority of the tumors (70.1%, n = 183) were localized, whereas 29.9% (n = 78) were metastatic. All patients with localized disease received the AV regimen. Among the metastatic cases, the lungs were the most common site of metastasis (70.0%, n = 53). Only 14.6% (38) of the patients had inferior vena cava (IVC) thrombosis, with most cases (44.7%, n = 17) reaching the ipsilateral renal vein level. Histopathological findings revealed that the majority of the tumors (68.4%, n = 121) had no anaplasia. The majority of the tumors (87.2%, n = 227) were large (> 10 cm) at baseline (Table 3).

Table 3.

Tumour- and chemotherapy-related characteristics of 261 children diagnosed with Wilms tumour at Mulago Hospital between Jan 2019 and December 2024

Variable	Category	n	%
Tumor Laterality	Unilateral	245	94.0
	Bilateral	16	6.0
Tumor position for the unilateral(n = 245)	Right	113	46.1
	Left	132	53.9
Tumor status	Localized	183	70.1
	Metastatic	78	29.9
Area of Metastasis (n = 78)	Lungs	53	70.0
	Liver	8	10.3
	Lungs and Liver	17	21.7
Presence of IVC thrombus	No	223	85.4
	Yes	38	14.6
Level of IVC thrombus (n = 38)	Ipsilateral renal vein	17	44.7
	Up to the hepatic vein	4	10.5
	Below the diaphragm	2	5.3
	Above the diaphragm	15	39.5
Chemotherapy regimen received	AV	183	70.1
	VAD	65	24.9
	AV/CE	13	5.0
Histopathology findings (n = 177)	No anaplasia	121	68.4
	Focal anaplasia	16	9.0
	Diffuse anaplasia	40	22.6
Baseline tumor size	Small (≤ 10 cm)	34	13.0
	Large (> 10 cm)	227	87.0

Abbreviations: VAD Vincristine, Actinomycin D, and Doxorubicin, AV/CE Actinomycin D, Vincristine / Cyclophosphamide, and Etoposide, IVC Inferior Vena Cava

Prevalence, bivariable and multivariable analysis of the factors associated with delayed surgery

The prevalence of delayed surgery was 63.6% (166/261), with a 95% CI of 57.6%-69.2%. At bivariable analysis, distance from healthy facility, region, religion, Hb level after POPC, WBC count, platelet count after POPC, year of diagnosis, presence of comorbidity, tumor laterality, tumor status, presence of IVC-thrombus, and chemotherapy regimen had p values < 0.2 (Tables 4 and 5, and 6). Factors significantly associated with delayed surgery in the multivariable analysis were year of diagnosis (202, 2022, and 2023), Hb level after POPC, tumor laterality, and chemotherapy regimen received (VAD and AV/CE). There was neither interaction nor confounding (Table 7).

Table 4.

Bivariable analysis of the sociodemographic characteristics of 261 children with Wilms tumor at Mulago Hospital

Variable	Delayed surgery Yes n(%) No n(%)	95% CI	P value
Sex
Male	85(64.9) 46(35.1)
Female	81(62.3) 49(37.7)	0.79–1.15	0.67
Age in years
< 5	129(62.0) 79(38.0)
≥ 5	37(69.8) 16(30.2)	0.92–1.38	0.262
Residence
Urban	69(62.2) 42(37.8)
Rural	97(64.7) 53(35.3)	0.86–1.26	0.68
Distance from health facility
< 10 km	23(51.1) 22(48.9)
10–50 km	47(67.1) 23(32.9)	0.94–1.83	0.11
> 50 km	96(65.8) 50(34.3)	0.94–1.75	0.11
Region
Central	107(63.1) 62(36.9)
Western	22(75.9) 7(24.1)	0.95–1.52	0.126
Eastern	27(57.5) 20(42.6)	0.69–1.19	0.500
Northern	11(68.8) 6(31.3)	0.77–1.55	0.631
Caregiver’s occupation(n = 136)
Unemployed	24(77.4) 7(22.6)
Informal job	53(66.3) 27(33.7)	0.67–1.09	0.217
Formal job	12(57.1) 9(42.9)	0.49–1.12	0.275
Religion
Catholic	56(65.9) 29(34.1)
Anglican	47(66.2) 24(33.8)	0.80–1.26	0.967
Muslim	33(55.0) 27(45.0)	0.63–1.09	0.200
Pentecostal	19(65.5) 10(34.5)	0.73–1.35	0.972
Others	9(81.8) 2(18.2)	0.90–1.7	0.182

Abbreviation: CI Confidence Interval

Statistical note: Bivariable analysis was performed via modified Poisson regression with robust standard errors. All variables with a p value < 0.2 in the bivariable analysis were included in the multivariable analysis

Table 5.

Bivariable analysis of the clinical characteristics of 261 children with Wilms tumor at Mulago Hospital between Jan 2019 and December 2024

Variable	Delayed surgery Yes n(%) No n(%)	95% CI	P value
Hb level after POPC
Normal	72(55.4) 58(44.6)
Low	94(71.8) 37(28.2)	1.07–1.56	0.007
WBC after POPC
Low	42(56.8) 32(43.2)
Normal	114(67.1) 56(32.9)	0.94–1.48	0.147
High	10(58.8) 7(41.2)	0.66–1.62	0.875
Platelet count after POPC
High	78(78.0) 49(38.6)
Normal	13(50.0) 13(50.0)	0.54–1.23	0.324
Low	75(69.4) 33(30.6)	0.94–1.36	0.197
Year of diagnosis
2019	11(36.7) 19(63.3)
2020	31(59.6) 21(40.4)	0.96–2.74	0.068
2021	30(85.7) 5(14.3)	1.43–3.82	0.001
2022	26(66.7) 13(33.3)	1.08–3.06	0.025
2023	38(77.6) 11(22.4)	1.29–3.47	0.003
2024	30(53.6) 26(46.4)	0.86–2.48	0.161
Presenting symptoms
Abdominal distension	111(66.1) 57(33.9)
Abdominal pain	2(50.0) 2(50.0)	0.28–2.03	0.580
Hematuria	4(50.0) 4(50.0)	0.37–1.53	0.437
Others	49(60.5) 32(39.5)	0.74–1.12	0.404
Presence of comorbidity
No	61(57.0) 46(43.0)
Yes	105(68.2) 49(31.8)	0.98–1.46	0.075
Comorbidity type
Hypertension	38(62.3) 23(37.7)
Malnutrition	30(66.7) 15(33.3)	0.79–1.39	0.64
Hypertension and malnutrition	35(74.5) 13(25.5)	0.91–1.52	0.24
Others	1(50.0) 1(50.0)	0.19–3.21	0.76

Abbreviations: CI Confidence Interval, Hb: Haemoglobin, WBC White Blood cell, POPC Preoperative Chemotherapy

Statistical note: Bivariable analysis was performed via modified Poisson regression with robust standard errors. All variables with a p value < 0.2 in the bivariable analysis were included in the multivariable analysis

Table 6.

Bivariable analysis of tumor- and chemotherapy-related factors in 261 children with Wilms tumor at Mulago Hospital

Variable	Delayed surgery Yes n(%) No n(%)	95% CI	P value
Tumor laterality
Unilateral	153(62.5) 92(37.5)
Bilateral	13(81.3) 3(18.7)	1.008–1.68	0.043
Tumor Status
Localized	106(57.9) 77(42.1)
Metastatic	60(76.9) 18(23.1)	1.12–1.58	0.001
Area of metastasis
Lungs	43(81.1) 10(18.9)
Liver	6(75.0) 2(25.0)	0.61–1.41	0.716
Both	11(64.7) 6(35.3)	0.55–1.16	0.239
Presence of IVC thrombus
No	138(61.9) 85(38.1)
Yes	28(73.7) 10(26.3)	0.96–1.48	0.11
Baseline tumor size
Small	24(70.6) 10(29.4)
Large	142(62.6) 85(37.4)	0.69–1.13	0.323
Chemotherapy regimen
AV	106(57.9) 77(42.1)
VAD	50(76.9) 15(23.1)	1.11–1.59	0.002
AV/CE	10(76.9) 3(23.1)	0.96–1.83	0.085
Histopathology
No anaplasia	72(59.5) 49(40.5)
Focal anaplasia	9(56.2) 7(43.8)	0.59–1.49	0.810
Diffuse anaplasia	21(52.5) 20(47.5)	0.63–1.23	0.457

Abbreviations: CI Confidence Interval, VAD Vincristine, Actinomycin D, and Doxorubicin, AV/CE Actinomycin D, Vincristine / Cyclophosphamide, and Etoposide, IVC Inferior vena cava

Statistical note: Bivariable analysis was performed via modified Poisson regression with robust standard errors. All variables with a p value < 0.2 in the bivariable analysis were included in the multivariable analysis

Table 7.

Multivariable analysis of the factors associated with delayed surgery among 261 children with Wilms tumor at Mulago National Referral Hospital

Variable	aPR (95% CI)	P value
Year of diagnosis
2019	1
2020	1.53 (0.92–2.56)	0.098
2021	2.26 (1.40–3.65)	0.001
2022	1.78 (1.07–2.99)	0.026
2023	2.01 (1.24–3.25)	0.004
2024	1.41 (0.85–2.37)	0.187
Tumor laterality
Unilateral	1
Bilateral	1.41 (1.07–1.85)	0.014
Hb level after POPC
Normal	1
Low	1.22 (1.05–1.51)	0.032
Chemotherapy regimen received
AV	1
VAD	1.32(1.11–1.57)	0.002
AV/CE	1.46(1.07–1.99)	0.016

Abbreviations: aPR Adjusted Prevalence Ratio, CI Confidence Interval, Hb Haemoglobin, POPC Preoperative chemotherapy, VAD Vincristine, Actinomycin D, and Doxorubicin, AV/CE Actinomycin D, Vincristine / Cyclophosphamide, and Etoposide

Statistical note: Multivariable analysis was performed via modified Poisson regression with robust standard errors. significance was defined as p < 0.05

Barriers and facilitators to timely surgery

In-depth interviews were conducted with 10 HCWs and 10 caregivers to explore the barriers to and facilitators of timely surgery. Six themes emerged inductively, namely, (1) systemic and institutional-level barriers, (2) patient-related barriers, (3) tumor and chemotherapy-related barriers, (4) systemic-level facilitators, (5) patient-level facilitators, and (6) community-level facilitators.

Theme 1: institutional and systemic barriers

High patient load and surgical backlog

This issue stems from a significant imbalance between the high number of patients requiring care, the existing queue, and the available medical personnel and infrastructure. “…part of the barriers leading to that; we have a large patient-to-doctor ratio.” (HCW07). A caregiver also noted that “…. We arrived, and there were people on the list who were going to have surgery as well….” (Caregiver06).

Inadequate equipment and consumables

The scarcity of essential medical equipment and consumables poses another significant institutional barrier. A health care worker reported “… then also some consumables like they don’t have what to use; the things of central line, some drugs that they are using in theatre, anaesthetic drugs, you know the government doesn’t supply enough…” (HCW04).

Limited surgical workforce and theatre space

The qualitative data revealed that a constrained surgical workforce significantly contributed to surgical delays. A healthcare worker said, “… this could be because there are few surgeons and sometimes when those doing fellowships in pediatric surgeries are in their exams, they tend to push the patients…” (HCW03).

Theme 2 patient-level barriers

Treatment abandonment

Treatment abandonment has emerged as a significant patient-level barrier to timely surgery. One healthcare worker recalled, “…unfortunately, the mother didn’t turn up and later came back when the tumor had progressed in size and thus the plans for surgery had changed…” (HCW04). One caregiver similarly noted, “… I had a friend who was told that the child was to be operated on and she ran away from the hospital but later came back when the child was in a worse state….” (Caregiver08).

Financial constraints

Financial constraints are frequently cited as major patient-level barriers that impact various stages of care, including diagnostics and treatment. One HCW explained, “…the image-guided biopsies that we take off for histology, they are quite expensive nowadays. The cost of 550,000 for an ordinary Ugandan is something that is not easily achievable…” (HCW07). One caregiver concisely summarized their struggle, “…because the money I have been using, I have to pay it back. And I’m not working now…… that’s the problem, the problem is money….” (Caregiver02).

Lack of awareness and misconceptions

A critical patient-level barrier was the ongoing lack of awareness about WT and, in some instances, harmful misconceptions regarding its cause and treatment. One HCW noted that “…And the other point would be ignorance…” (HCW07). A caregiver also shared, “…. They take away children from the hospital thinking it is witchcraft, and by the time they return, there is nothing more that can be done….” (Caregiver05).

Late presentations

Healthcare workers and caregivers consistently noted that a significant proportion of patients arrive late with advanced disease. One participant explained, “…. Many of our patients come late, so there is late presentation, and they therefore come when the tumor is too huge, deemed inoperable…” (HCW08). Another participant said, “…In my own view as to why they don’t get timely surgery is because of the delay to come to the hospital to check that this is the particular illness they are suffering…” (caregiver09).

Transport difficulties

Transport-related challenges presented a significant logistical barrier, particularly for patients residing far from MNRH. As one healthcare worker explained, “…and sometimes when we call them back for surgery, they are not able to get that transport, to meet the transport costs in time…” (HCW08). One caregiver lamented, “…I would leave my home to this side and then go back, making the transport costs be too much….” (Caregiver06).

Theme 3: tumour and chemotherapy-related barriers

Tumor-related barriers

Tumor characteristics, particularly tumor size and invasiveness at various stages of treatment, were identified as direct contributors to surgical delays. A healthcare worker noted that “…sometimes during interim assessment, the tumor of the patient is still big, and more chemotherapy is needed so that they can have surgery…”.

Chemotherapy-related barriers

In addition to tumor characteristics, side effects arising from chemotherapy regimens have also emerged as a cause of surgical delays. A HCW highlighted that “…occasionally some patients get delayed count recovery from the myelosuppressive effect of the chemotherapy…” (HCW07).

Theme 4: system-level facilitators

Adequate supply of resources

A key system-level facilitator identified was the consistent and adequate supply of essential resources. One participant highlighted, “…. but the factors that would facilitate timely surgery would be…. funding to have materials available, utilities, gowns, available water supply in theatre, surgical supplies being available…” (HCW01).

Increased surgical workforce

An increase in the surgical workforce was consistently identified as a critical systemic facilitator. The participants suggested, “…I think maybe they can increase human resource. May be you have more than two anaesthetists, more surgeons. But in case one has a problem, another will be around to cover… (HCW06).

Multidisciplinary coordination and communication

Effective multidisciplinary coordination and open communication are recognized as pivotal in streamlining the surgical pathway. One participant explained that “…informing surgeons from the start, which is a multidisciplinary approach. And then working hand in hand with them, communicating each step of the treatments…(HCW01).

Theme 5: patient-level facilitators

Early presentation

Early presentation to the hospital emerged as a crucial patient-level facilitator for timely surgery. A healthcare worker noted that “…when a patient has come in early, the tumor is not very large enough; they can have their surgery in time after receiving neo-adjuvant chemotherapy…” (HCW03). One caregiver explained, “… I immediately came to Mulago because if I had delayed, it would have been worse….” (Caregiver01).

Patient education

Comprehensive patient education was highlighted as a significant facilitator. A participant noted that “… you just empower them with the knowledge they need and you try to demystify all their myths…” (HCW03). Caregivers themselves affirmed the positive impact of clear communication; one stated that “…. the doctors explaining to me helped to be strong knowing exactly the child was going to be operated….” (Caregiver07).

Proactive patient follow-up

Proactive patient follow-up by the healthcare team was identified as a key facilitator in ensuring treatment continuity and preventing abandonment. A healthcare worker emphasized, “… proper follow-up to pick the ones probably who have absconded and then try to convince them and call them back in time….” (HCW07). Caregivers also appreciated these efforts, with one confirming, “…. Even when I could miss, they would call me on the phone…” (Caregiver02).

Caregiver adherence to treatment plan

Caregiver adherence to the prescribed treatment plan was recognized as a direct facilitator of timely surgery. One caregiver affirmed this commitment, stating, “…we were able to follow how the treatment went as instructed by the healthcare workers….” (Caregiver04).

Community-level facilitators

Enhanced community awareness

Enhanced community awareness was proposed as a vital community-level facilitator to promote early presentation and reduce treatment delays. A participant suggested, “…. spreading awareness out there in the communities. I feel like in this country, awareness and advocacy for cancer is still a low level…” (HCW10).

Decentralized health system strengthening and surgical outreach

Strengthening the health system at lower-level facilities and extending specialized care beyond the central referral hospital were identified as crucial community-level facilitators. Healthcare workers suggested, “…. Health system strengthening for lower-level facilities whereby we have probably more pediatric surgeons at regional hospitals. This can enable things like organizing surgical camps…….” (HCW01). Caregivers strongly echoed the need for decentralized services, with one explicitly requesting that “… I request the government; this cancer treatment should not just be in Mulago; it should be even extended to the district level because there is a parent that can’t make it to Mulago….” (Caregiver04).

Discussion

The prevalence of delayed surgery in our study was 63.6% (95% CI 57.02–68.65). This means that approximately six of every ten children with WT had delayed surgeries. This is similar to the 67% reported by a global study on surgery delay in children [18]. However, this percentage was higher than the 56% reported in Uganda [11] and lower than the 83% reported in Somalia [13]. The difference in the prevalence arises from differences in the study population, study designs, and settings. Although it is the same study setting, Kakembo et al. used a prospective design and recruited children with any surgical conditions in MNRH. The Somalia study used household surveys and included any disease, illness, or injury in which surgical care improves outcome. The high prevalence of delayed surgery among children with WT at Mulago Hospital may contribute significantly to low survival rates.

Delayed surgery was associated with the year of diagnosis, haemoglobin level after POPC, tumour laterality, and chemotherapy regimen received. Patients diagnosed in 2021, 2022, and 2023 were approximately two times more likely to experience surgical delays than those diagnosed in 2019. The year 2021 marked the peak of the COVID-19 pandemic, during which healthcare systems worldwide were severely disrupted. Factors such as restricted hospital admissions, reduced staffing levels, and the reallocation of operating theatres for COVID-19 contributed to these delays in surgical care [19]. These findings align with a multicenter study in Sub-Saharan Africa, which revealed substantial surgery delays for children in 2021 due to the impacts of the pandemic [20]. The persistence of surgery delays in 2022 and 2023 may reflect the prolonged impact of the COVID-19 pandemic-related backlogs, increased referrals after the COVID-19 pandemic, and ongoing systemic challenges. These findings are consistent with findings from a global study that predicted that many operations would be cancelled or postponed due to disruption caused by COVID-19 [21]. The nonsignificance of the year 2020 likely reflects a complex, transitional year where the full effects of the pandemic on surgical systems were only beginning to unfold, and the nonsignificance of the year 2024 may reflect health system recovery and adaptation after the pandemic.

Children with bilateral WT were 41% more likely to experience surgical delays than those with unilateral tumors. This is consistent with prior research, as bilateral WT requires a more complex, staged surgical approach to preserve renal function. These cases often necessitate extended preoperative chemotherapy, advanced imaging, and specialized surgical teams—all factors that can contribute to delays in a resource-constrained environment [22, 23]. In this study, children who received the VAD and AV/CE regimens were 32% and 46% more likely to experience delayed surgery, respectively, than those who received the AV regimen. These findings are consistent with findings from Malawi, where 27% of the children who received the VAD regimen had delayed surgery due to unresectable disease [24]. This could be attributed to the inclusion of doxorubicin, cyclophosphamide and etoposide, which are often used in more advanced (metastatic) or aggressive disease. In cases of advanced and aggressive disease, there might be a need for more treatment cycles, an increased likelihood of toxicity occurrence, and longer recovery periods. Israels and colleagues reported that grade 4 neutropenia occurred significantly more frequently in children receiving the VAD regimen than in those receiving the VAD regimen [24].

Children with low Hb levels after POPC were 25% more likely to experience surgical delay than those with normal Hb levels after POPC. These findings align with a study in Nigeria, where a significant number of children had their surgeries delayed due to anaemia [25]. Additionally, a systematic review in LMICs revealed that low haemoglobin levels can lead to surgeries being cancelled or postponed [26]. Hemoglobin is a critical indicator of a child’s hematologic stability and surgical fitness. Anaemia can increase perioperative risk, including complications such as anaesthesia, poor wound healing, and infection.

Preoperative chemotherapy, particularly multiagent regimens such as VAD, can suppress bone marrow function, leading to chemotherapy-induced anaemia, especially in children who are already malnourished or have advanced disease. As a result, surgical teams may defer surgery in children with low haemoglobin levels until stabilization is achieved. This commonly involves blood transfusion or correction of nutritional deficiencies, which can introduce delays depending on resource availability. However, a previous study in MNRH reported that major elective surgeries were cancelled or postponed due to blood shortages [27]. Findings from in-depth interviews with HCWs and caregivers highlighted several interrelated barriers spanning the institutional, patient, tumor, and chemotherapy domains. Institutional and systemic barriers included high patient load and surgical backlog, inadequate equipment and consumables, and limited surgical workforce and theatre space. These findings align with a study by Israel and authors, who reported that children with WT in LMICs experience surgical delay due to a lack of staff or theatre equipment [28].

Patient-level barriers included treatment abandonment, financial constraints, lack of awareness and misconceptions, late presentations, and transport difficulties. These issues align with prior findings that link financial constraints and transport difficulties to treatment delays [29]. Additionally, findings from a study in Malawi reported that treatment abandonment and late presentation are associated with preoperative treatment failure, which affects the timing of surgery [30]. The tumor and chemotherapy-related barriers align with the quantitative findings, where tumor size after preoperative chemotherapy and the chemotherapy regimen were associated with delayed surgery.

The study identified several enablers that contributed to timely surgical access. At the institutional level, increasing the surgical workforce and effective multidisciplinary coordination and communication are among the facilitators of timely surgery. Such integrated team-based approaches have been recommended in studies on pediatric cancer in LMICs to improve cancer care efficiency [31]. At the patient level, early presentation, patient education, proactive patient follow-up, and caregiver adherence to the treatment plan facilitated timely surgery. These factors are supported by previous research showing that well-informed caregivers are more likely to comply with care plans [32] and that continued efforts are needed to encourage early presentation and prevent abandonment [30]. At the community level, enhanced community awareness and decentralized health system strengthening and surgical outreach could facilitate timely surgery. These findings are consistent with those of studies in Uganda and Kenya, which suggested that increasing awareness of childhood cancers among the general public via effective means of communication, continuing education programs for health providers on signs and symptoms of common childhood cancers, including providers at lower-level facilities and improving referral linkages between community health providers and cancer referral centers, could address late-stage presentations [32].

Study limitations

The study relied on a retrospective chart review, which is inherently susceptible to incomplete or missing data. This was evident in the high percentage of missing data (greater than 15%) for certain variables, which could have led to a reduction in the statistical power to detect meaningful associations. Caregivers interviewed may have been affected by the tendency to provide responses perceived as socially acceptable, especially concerning institutional challenges, leading to social desirability bias. The study was conducted at a single tertiary referral hospital, which may limit the generalizability of the findings to other settings, especially lower-level health facilities or rural areas with different health system capacities.

Conclusion

The prevalence of delayed surgery among children with WT in this study was high (6 in 10). Delayed surgery was associated with the year of diagnosis, haemoglobin level after POPC, tumour laterality, and chemotherapy regimen received. Several barriers to timely surgeries have been reported, notably systemic and institutional barriers such as high patient loads and surgical backlogs, inadequate equipment and supplies, and limited surgical capacity. Patient-level barriers included treatment abandonment, financial constraints, a lack of awareness and misconceptions, late presentations, transportation difficulties, and other barriers related to chemotherapy and tumor characteristics. Facilitators of timely surgery were identified at the systemic, patient, and community levels. These included increasing the surgical workforce, multidisciplinary coordination and communication, early presentation, providing patient education, proactive patient follow-up, caregiver adherence to treatment plans, enhanced community awareness, decentralized health system strengthening, and surgical outreach.

Recommendations

We recommend carrying out targeted interventions addressing both systemic constraints and clinical complexities. Policies should focus on expanding surgical capacity, ensuring a consistent supply of equipment, and strengthening referral pathways. Further research should examine the impact of surgical delays on outcomes and interventions to reduce surgical delays among children with Wilms tumor.

Abbreviations

aPR

Adjusted Prevalence Ratio

AV

Vincristine and Actinomycin D

CE

Cyclophosphamide and etoposide

CI

Confidence Interval

COG

Children Oncology Group

Global HOPE

Global Haematology Oncology Pediatric Excellence

Hb

Haemoglobin

HCWs

Health Care Workers

HICS

High-income countries

IVC

Inferior Vena Cava

KIIs

Key informant interviews

LMICs

Low- and middle-income countries

MNRH

Mulago National Referral Hospital

NCDs

Non-Communicable Diseases

NWSTG

National Wilms’ Tumor Study Group

PHO

Pediatric Hematology and Oncology

PODC

Pediatric oncology in developing countries

POPC

Preoperative chemotherapy

RFS

Relapse-free survival

SIOP

Société Internationale d’Oncologie Pédiatrique

USA

United States of America

VAD

Vincristine, Actinomycin D and Doxorubicin

WBC

White blood cells

WHO

World Health Organization

WT

Wilms tumor

Authors’ contributions

D.L. conceived the research idea, wrote the research protocol, designed the study database and data abstraction tool, led the data collection, performed the data analysis, and wrote the first manuscript draft.S.N. polished the research idea and was key in the protocol and manuscript writing.A.M. polished the research idea and was key in the protocol and manuscript writing.R.N. polished the research idea and was key in the protocol and manuscript writing.A.A. refined the research idea, participated in protocol writing and designing the data abstraction tool, provided supervision during study implementation, and participated in manuscript writing.E.K.K. polished the research idea and was key in the qualitative component of the study.R.N. refined the research idea, participated in protocol writing and designing the data abstraction tool, provided supervision during study implementation, and participated in manuscript writing.K.N. refined the research idea, participated in protocol writing and designing the data abstraction tool, provided supervision during study implementation, and participated in manuscript writing.J.K. polished the research idea and provided expatriate guidance in the protocol and manuscript writing.

Funding

The author did not receive funding to do this work.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was performed in accordance with the principles of the Declaration of Helsinki. Approval was granted by the Makerere University School of Medicine Research and Ethics Committee (Mak-SOMREC-2024-1212), which also granted a waiver of informed consent. The authors confirm that the ethical policies of the journal, as noted on the journal’s author guidelines page, have been adhered to.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.