Evaluation of diagnostic interval in children with newly diagnosed paediatric cancers: A Cross-sectional study

Abstract

Background

Diagnostic interval/delay (DI) denotes the duration between first presentations to healthcare provider to the definitive diagnosis of cancer. Prolonged DI is a major contributor to deaths among children with malignancies in low-middle-income countries (LMICs). The data on DI and factors affecting it are limited from LMIC.

Methods

This cross-sectional study enrolled patients/children<12 years with proven/suspected malignancies. Children already on definitive chemotherapy, who died before a definitive diagnosis and with benign tumours were excluded. The parents or caregivers were interviewed individually using a customized questionnaire. Various intervals in referral chain like patient interval/delay (PI), DI and treatment interval/delay (TI) and factors affecting DI were evaluated.

Results

Out of 120 eligible children; 79 with median age 43 months (interquartile range [IQR]: 28–81) were analyzed. Haematological malignancies (n = 40) and non-haematological malignancies (n = 39) were equally distributed. Median total delay was 74 days (14–88 days). Median DI was 17 days (IQR: 8–54). The main contributor to delay was due to referral delay by primary and secondary physician (p < 0.05). The median PI was 2 day (IQR: 0–7.5) and median TI was 1 day (IQR: 0–4). Prolonged DI (DI > 30 days) was seen in 33 (41.8%) children. Trial of alternative medicine was the only significant factor associated with prolonged DI in univariate and multivariate analysis with odds ratio of 6.24. Other demographic, socioeconomic, health-seeking journey and disease-related factors were not found to be associated with prolonged DI.

Conclusions

Significant delay exists in paediatric cancer management in LMIC. Augmentation of physician and parental awareness is the key to decrease these delays.

Introduction

Childhood cancers constitute around 4% of the total cancer cases.¹ The annual prevalence of childhood malignancies in India varies from 38 to 140 cases per million children.² Despite significant advances in the field of paediatric oncology over the past twenty years, cancer continues to be a significant reason of mortality in children. Any delays in diagnosis of cancers might adversely affect the prognosis and psychological well-being.^{3, 4, 5, 6} Even a four-week delay of cancer treatment is associated with increased mortality.⁷

Three important intervals describe the time patients spend in the cancer treatment pathway: patient interval which is from start of symptom to first visit to a healthcare professional; diagnostic interval (DI) denotes the duration between first visit to healthcare provider to the definitive confirmation of the cancer diagnosis and treatment interval (TI) which is interval from final diagnosis to initiation of definitive therapy. DI reflects the efficiency of chain of referral to tertiary-care centre while TI reflects competence of tertiary healthcare system. Prolonged DI is a major public health concern worldwide and remains the prime cause of death among children with malignancies in low-middle-income countries (LMICs).⁸ Delay in cancer diagnosis in children is typically classified into three categories: delays caused by parents or caregivers (patient delay), delays caused by the disease itself, and delays caused by issues in healthcare delivery (practitioner/system delay).⁹

Various social, educational, cultural, healthcare-related and disease-related factors are recognized as variables for DI.¹⁰ Delayed diagnosis of cancer in children is prevalent in LMIC. There is a paucity of research on DI and factors contributing to delayed diagnosis in LMIC.¹¹ Therefore, there was a felt need on the estimation of the DI and factors contributing to DI in our patients throughout the referral chain leading to our tertiary care centre.

Materials and methods

Participants

This cross-sectional study was conducted at a tertiary care referral paediatric haemato-oncology centre in North India between Aug 2022 and Dec 2023. All children less than 12 years of age who were newly referred to our hospital with proven or suspected malignancies were consecutively enrolled and included in the study. Children who were already on definitive therapy for malignancies, who died or left before a definitive diagnosis and children whose final histopathology of tumour was benign were excluded. Informed consent was taken from parents/caregivers. The study was approved by Institutional Ethics Committee. All good clinical practice guidelines as per declaration of Helsinki were followed.

Sample size calculation

We calculated a convenient sample size from our finite study population. Considering the strength of our study population to be approximately 25,00,000 with existing fertility rate of 2.18% and percentage population of children<12 years by national statistics to be 20–25%; total population of children<12 years in our study population was 11,50,000.¹² With annual incidence rate of paediatric cancers being 140/10,00,000 in India,¹³ and with assumption that our centre accounts for 50% workload for childhood malignancies amongst our study patients; the calculated convenient sample size was 120 for study duration.

Data collection

The parents/caregivers were interviewed individually using a customized questionnaire. All relevant information from the participant was cross-verified from previous medical records and documented on master sheet. Various intervals in the referral path were defined in line with the Aarhus Statement 2012 to promote greater consistency in defining and a greater precision in reporting various points and intervals.¹⁴

Another customized questionnaire was administered to the parents/caretakers of every enrolled patient and their response was accumulated and collaborative data was calculated to study factors affecting DI. The collected data was broadly categorized into three categories.

Category 1: Demographic and socioeconomic.

Category 2: Healthcare related.

Category 3: Disease related.

Definitions

The various time intervals and delays used in the chain of referral in the study are depicted in Fig. 1.

• 1.Patient Interval (PI) or patient delay: First symptom to first visit to a healthcare provider.
• 2.Primary physician: the first healthcare provider the patient visited.
• 3.Secondary physician; the first doctor to whom the primary physician has referred the patient.
• 4.Primary physician delay: Time interval taken by primary physician to refer to secondary physician.
• 5.Secondary physician delay: Time interval taken by secondary physician (or subsequent physicians) to refer to tertiary-care referral centre.
• 6.Diagnostic Interval (DI): Interval between first visit to a healthcare provider to definitive diagnosis.
• 7.Treatment Interval (TI) or Treatment delay: Interval between definitive diagnosis to initiation of definitive treatment.
• 8.Prolonged DI: DI >30 days.
• 9.Alarming symptoms: Bluish discolouration or patches on the body, bleeding from any site, throbbing headache, swelling of any part, severe pain abdomen and seizures.
• 10.Vague symptoms: Fever, cold, cough, dull-aching abdominal pain and lethargy.
• 11.Advanced stage: ≥ Stage III for solid malignancies and all haematological malignancies.

Fig. 1.

Various intervals/delays in referral pathway for childhood cancers.

Statistical analysis

For comparison, we had divided our cohort into two groups: DI ≤ 30 days and >30 days.⁷ All categorical data was expressed as frequency (%). Categorical data was analysed by Chi-square/Fisher exact test and continuous variables were analysed by Student t test. All categorical variables were subjected to multiple logistic regression analysis. For comparison of >2 median values Kruskal–Wallis test was used. A P-value>0.05 was taken as significant. All data was analysed using SPSS software (version 25.0).

Results

A total of 120 patients with proven or suspected malignancies were eligible and 79 with median age 43 months (interquartile range [IQR]: 28–81) were analysed. The analysed patients were mostly male 44/79 (55.7%), with rural background 43/79 (54.4%) and resided at plains 62/79 (78.5%). Primary physician was a paediatrician in majority 44/79 (55.7%) of the cases, while only one-third of the total patients took first consultation at military hospitals. Haematological malignancies (n = 40) and non-haematological malignancies (n = 39) were equally distributed (Fig. 2). Haematological malignancies comprised of acute lymphoblastic leukaemia (33), acute myeloid leukaemia (3), lymphoma (4) and non-haematological malignancies encompassed neuroblastoma (7), brain tumour (6), bone tumour (4), Wilms’ tumour (4), rhabdomyosarcoma (3) and others (15). The median total delay was 74 days (14–88 days) due to various reasons as shown in Supplementary table 1. Median DI was 17 days (IQR: 8–54). On comparison of various delays in study, referral delay by primary and secondary physician significantly contributed to total delay (p < 0.05) (Fig. 3). Median patient delay was 2 days (IQR: 0–7.5) and median treatment delay was 1 day (IQR: 0–4) (Fig. 3). Prolonged DI was seen in 33 (41.8%) patients. Trial of alternative medicine was the only factor found as a major contributor to prolonged DI in univariate and multivariate analysis with odds ratio of 6.24 (Table 1, Table 2). Other demographic, socioeconomic, health-seeking journey and disease-related factors were not found to be associated with prolonged DI.

Fig. 2.

Consort diagram depicting flow of enrolled children in the study.

Fig. 3.

Plot and whisker graph depicting various delays and their comparison in the study population.

Table 1.

Univariate analysis of factors affecting Diagnostic Interval.

	Diagnostic Interval			p-value
	≤30 (n = 46)	>30 (n = 33)	Total (n = 79)
Demographic/Socioeconomic factors
Age of child at first presentation of symptoms >5 Year	16 (34.8%)	14 (42.4%)	30 (38%)	0.228
Male Gender	30 (65.2%)	14 (42.4%)	44 (57%)	0.06
Number of Siblings >1	34 (73.9%)	26 (78.8%)	60 (75.9%)	0.931
Birth order >1	23 (50%)	12 (36.4%)	35 (44.4%)	0.256
Rural location	24 (52.2%)	19 (57.6%)	43 (54.4%)	0.529
Hilly terrain of residence	11 (23.9%)	6 (18.2%)	17 (21.5%)	0.373
Father's education: Graduate and above	23 (50%)	17 (51.5%)	40 (50.6%)	0.889
Mother's education: Graduate and above	31 (67.4%)	21 (63.6%)	52 (65.4%)	0.391
Trial of alternative medication	3 (6.5%)	7 (21.2%)	10 (12.7%)	0.046∗
Healthcare-related factors
Type of primary healthcare centre patient first consulted: Military Hospital	16 (34.8%)	12 (36.4%)	28 (35.4%)	0.97
Qualification of the primary healthcare professional: Paediatrician	26 (56.5%)	18 (54.5%)	44 (55.7%)	0.767
Disease-related factors
Parents assessment of initial symptom: Vague	38 (82.6%)	25 (75.7%)	63 (79.7%)	0.413
Occult site of tumour/malignancy	43 (93.5%)	27 (81.8%)	70 (88.6%)	0.192
Haematological malignancy	27 (58.7%)	13 (39.4%)	40 (50.6%)	0.116
Advanced stage at diagnosis	33 (71.7%)	19 (57.6%)	52 (65.8%)	0.255
Non-metastatic disease at diagnosis	40 (87%)	25 (75.7%)	65 (82.3%)	0.176

∗Significant p values are mentioned in bold.

Table 2.

Multivariate regression analysis of factors affecting Diagnostic Interval.

	Coefficient of beta (β)	Odds Ratio (OR)	95% C.I. for Odds Ratio
			Lower	Upper
Demographic/Socioeconomic Factors
Age of child at first presentation of symptoms (RC: >5 years)	0.90	2.47	0.79	7.68
Gender (RC: Male)	0.73	2.07	0.57	7.51
Number of siblings (RC: >1)	−0.90	0.41	0.04	4.44
Birth order (RC:>1)	−0.79	0.46	0.12	1.70
Rural/Urban location (RC: Rural)	−0.19	0.83	0.24	2.84
Terrain of residence (RC: Hilly Region)	0.65	1.91	0.40	9.13
Highest level of education the child's father has attained (RC: Graduate)	−0.52	0.59	0.18	1.98
Highest level of education the child's mother has attained (RC: Graduate)	−0.66	0.52	0.13	2.03
Trial of alternative medication (RC: Yes)	1.83	6.24∗	1.19	32.77
Healthcare-related factors
Type of primary healthcare centre (RC: Military Hospital)	−0.37	0.69	0.20	2.38
Qualification of the primary healthcare professional/expert (RC: Paediatrician)	−0.37	0.69	0.21	2.32
Disease-related factors
Parents assessment of initial symptom (Vague/Alarming) (RC: Vague)	−0.10	0.91	0.21	3.87
Site of tumour (RC: Occult)	0.86	2.36	0.31	17.82
Type of tumour (RC: Haematological)	−1.44	0.24	0.02	3.57
Stage at diagnosis (RC: Advanced)	1.98	7.25	0.52	101.16
Metastatic or non-metastatic (RC: Non-metastatic)	−2.39	0.09	0.00	2.17

∗Significant p-value (p = 0.03) are mentioned in bold.

Discussion

The patients often exhibit distinct healthcare demographics, influenced by factors such as deployment, geographic location, and access to specialized medical resources. Our study population consisted predominantly of lower-middle socioeconomic class as per modified Kuppuswamy scale and was not native of a particular region. Consequently, our study unravelled the intricacies within this specific demographic and socioeconomic cohort and enabled us to comprehend the present scenario of childhood cancer detection in our study patients. The study has highlighted potential challenges and tailored insights for public health initiatives or significant policy changes to enhance the efficacy of childhood cancer care within such a unique healthcare domain.

In our settings, median DI was of 17 days (IQR: 8–54), with a significant part of the delay occurring after their initial health interaction (primary and secondary physician delay). These findings were highly consistent with the existing literature.^{15, 16, 17} In our study, the median patient interval of 2 days (IQR: 0–7.5) did not contribute much to total delay, as our patients were predominantly from lower-middle class and not from lower class and majority of mothers were graduate (64.5%), which led to early health-seeking journey. Median TI of 1 day (IQR: 0–4) in our study was also minimal and comparatively lesser to published studies from India confirming the availability of rapid diagnostic facilities with minimal waiting period, leading to early initiation of definitive therapy at our tertiary-care referral centre. Comparison of DI, PI and TI in our study with other published studies from India is shown in Table 3.

Table 3.

Comparison of various previously published Indian studies on diagnostic delays in Paediatric malignancies.

	Verma et al.17 2020	Singh et al.16 2020	Lashkari et al.15 2021	Present Study
Type of study	Prospective observational	Cross-sectional	Retrospective	Cross-sectional
Age at enrollment	Not specified	<15 Years	<18 years	<12 years
Geographical location	North India	South-west	South	Pan-India
Patient enrolled	111	570	115	120
Patient analysed	111	75	111	79
Patient Interval (Median)	5	7	14	2
Diagnosis Interval (Median)	4	76	21	17
Treatment Interval (Median)	2	NA	5	1
Factors associated with delays	Gender and use of alternative medicine	Maternal education, state of residence, site of lesion, disease stage, and tumour type	Age, maternal education, and type of cancers	Use of alternative medicine

NA- Not available.

We divided our cohort in two groups with variable DI (≤30 days and >30 days), as a recent meta-analysis concluded that the mortality related to cancer increases by 10% for a DI of 28–30 days.⁷

In our study 10/79 (12.7%) patients received alternative medication, which was the primary factor contributing to the prolonged DI in our cohort. In India, there is a widespread conviction among the general people in the efficacy of alternative forms of medicine, which are perceived as being devoid of negative side effects, are cost-effective, and are not subjected to rigorous examinations as compared to allopathic therapies. Parents explore alternative therapies such as holy water, religious rites, visits to herbalists, and the use of over-the-counter pharmaceuticals.¹⁰ A recent study found that approximately one-fifth of caregivers of patients chose an alternative medication, resulting in a notably increased delay.¹⁷

Previous studies have documented the presence of gender bias favouring male in early healthcare-seeking behaviour and cancer treatment in LMIC, particularly in India.¹⁷ Higher age significantly influences the initial seeking of medical help and the period spent for diagnosis.^9,10,15,18 Also, aggressive disease is likely to have lesser delay due to alarming symptoms and rapid deterioration. Place of residence of patients and their socioeconomical background and education have affected the diagnostic delays.^17,19 In our study, these factors were not found to affect DI. Our patients predominantly consisted of lower-middle class, which was higher socioeconomic strata compared to other studies with higher literacy rates and have better access to healthcare facilities and resources, which may explain the above discrepancy. Better access to healthcare and increased awareness might have mitigated the gender, regional, geographical, and socioeconomic disparities and thereby delays in diagnosis. Moreover, since the above factors are interdependent, the impact of isolated factors may not hold true in regression analysis. Similar findings were observed by other authors as well from LMIC.^{15,16,20, 21, 22}

Our study did not find any association between DI and other variables such as the type and site of cancer, disease stage and different types of healthcare facilities, which have been identified as determinants for delay in cancer diagnosis in previous studies. Hence, certain correlated factors identified as predictors of delay in diagnosis in one study may not necessarily hold true in another study as depicted in Table 3. This supports the notion that potential determinants of delay vary depending on the geographic distribution, socioeconomic disparities, cultural beliefs and health-seeking behaviour of the particular individual.

There were a few limitations in this study. It was a single-centre study with small sample size. Moreover, most of our patients were from lower-middle socioeconomic group thereby compromising the generalizability of this study.

To conclude, a significant delay exists in paediatric cancer management in LMIC. We also found delayed referral by primary and secondary physician as a major contributor to total delay. Trial of alternative medicine was single-most significant contributor to prolonged DI in our study. Therefore, it is crucial to exert all possible endeavours to augment physician and parental awareness regarding childhood cancers. We also feel that there is an urgent need to analyse referral routes in order to mitigate problems encountered in the chain of referral, facilitating timely referral to suitable paediatric tertiary cancer care centres. A nationwide study on large number of patients is required to exactly delineate various delays in referral routes and various factors affecting DI.

Patients/ Guardians/ Participants consent

Parents/guardians informed consent was obtained.

Ethical clearance

Institute/hospital ethical clearance certificate was obtained.

Source of support

Nil.

Disclosure of competing interest

The authors have none to declare.

Appendix A. Supplementary data

The following is the Supplementary data to this article: