The Hungarian cross-cultural adaptation of the MMQL-AF for measuring quality of life in adolescents with cancer

Abstract

With improvements in childhood cancer survival, attention has shifted toward health-related quality of life (HRQoL) and long-term treatment effects. This study aimed to validate the Hungarian version of the Minneapolis-Manchester Quality of Life Instrument – Adolescent Form (MMQL-AF) and explore factors associated with HRQoL among adolescents and young people with cancer. Ninety-two patients with cancer (46 receiving active treatment and 46 ≥ 1 year post-treatment) and 46 healthy controls completed the MMQL-AF. The MMQL-AF was translated into Hungarian in accordance with Beaton et al. and TARES (2024) guidelines. Reliability was assessed using Cronbach’s alpha and intra-class correlation coefficients (ICCs). Validity was evaluated by correlations with the Hungarian Pediatric Quality of Life Inventory (PedsQL 4.0) and by comparing scores across groups. Five items were removed during confirmatory factor analysis (CFA) to improve model fit. The MMQL-AF demonstrated acceptable reliability (Cronbach’s alpha 0.73–0.90; ICC 0.78–0.96; RMSEA = 0.062; 90% CI: 0.055–0.069; CFI = 0.828; TLI = 0.814; SRMR = 0.079; GFI = 0.962) and strong concurrent validity with PedsQL scores. Cancer patients reported significantly lower HRQoL than healthy peers (mean 3.58 vs. 3.81; mean difference 0.23; 95% CI: 0.05–0.41; p = 0.012). No significant difference emerged between patients on active treatment and patients off treatment. Patients with bone, soft-tissue, or central nervous system tumors reported worse outcomes (p = 0.015). Positive parental relationships were strongly associated with a better outlook (p < 0.001). The Hungarian MMQL-AF demonstrated acceptable reliability and validity for assessing HRQoL among adolescents and young people with cancer. Tumor type and family environment were associated with differences in the well-being of the patients. This instrument enables comprehensive HRQoL assessment in Hungarian pediatric oncology, supporting identification of at-risk groups and informing survivorship care strategies tailored to medical and psychosocial needs.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-33243-9.

Introduction

In recent decades, advances in the treatment of childhood cancer have led to a significant increase in survival rates to over 70–80%¹. However, many survivors face long-term side effects associated with their treatment, which may include chemotherapy, radiation therapy, surgery, and complications arising from the cancer itself². These late effects can manifest in various ways, such as growth and developmental issues, endocrine disorders including gonadal dysfunction and hypothalamic-pituitary disorders, cardiovascular and kidney problems, secondary cancers, cognitive impairments, and psychological difficulties^3–5. Even peers who recover without noticeable physical late effects may face challenges in terms of psychological well-being and social integration. For example, they may struggle with the psychological burden or social difficulties associated with milestones such as starting school, entering the workforce, or forming relationships^6–8. These challenges can negatively affect the quality of life (QoL) of child⁹. The World Health Organization defines QoL as “individuals’ perceptions of their position in life, in the context of the culture and value systems in which they live, and in relation to their goals, expectations, standards, and concerns.”¹⁰. To address these issues, it is crucial to establish a comprehensive follow-up care system that focuses not only on maintaining physical health, but also on enhancing long-term QoL throughout the recovery and growth phases. Therefore, research on health-related quality of life (HRQoL) among childhood cancer patients and survivors is vital for optimizing their care and improving outcomes¹¹. HRQoL is a multidimensional construct that reflects an individual’s subjective perception of how their health status—including physical, emotional, cognitive, and social functioning—is affected by a medical condition or its treatment. Unlike general quality of life, HRQoL focuses specifically on health-related aspects of well-being, making it a key outcome in clinical research and healthcare evaluation¹². Currently, tools to assess HRQoL in childhood cancer patients and survivors are limited and there are no specific instruments for this population in Hungary. The Minneapolis-Manchester Quality of Life Survey of Health (MMQL) is a well-established, multidimensional self-report instrument that has been used to assess HRQoL among adolescent cancer patients in various countries, including the United States, Western Europe, and Asia^13–16. The original MMQL-Adolescent Form (MMQL-AF), developed by Bhatia et al. (2002)¹³, demonstrated strong psychometric properties in a sample of 499 adolescent cancer patients and survivors. Internal consistency was high across all seven domains, with Cronbach’s alpha coefficients ranging from 0.75 (cognitive functioning) to 0.90 (psychological functioning). Construct validity was supported through exploratory factor analysis, and convergent validity was demonstrated by moderate to strong correlations with the Child Health Questionnaire (CHQ). Discriminant validity was established by significant differences in HRQoL scores between cancer survivors and healthy controls across multiple domains (e.g., psychological functioning: p < 0.001). Subsequent validation studies in other countries have supported the robustness of the instrument. The UK version¹⁴, reported good internal consistency (Cronbach’s α = 0.72 to 0.90) and confirmed known-group validity by differentiating between patients on and off treatment. The Swedish version¹⁶, also demonstrated acceptable reliability (α = 0.69 to 0.91), and ICC ranged between 0.248 and 0.763. These findings further support the MMQL-AF as a reliable and valid tool for assessing HRQoL in adolescent cancer populations across diverse cultural contexts. The MMQL was chosen because its items are broad and can be administered to both healthy adolescents and those affected by or recovering from disease. Given its applicability across different conditions, the MMQL is a valuable tool for evaluating general health aspects. The aim of this study was to validate a Hungarian version of the MMQL, contributing to the establishment of a robust long-term follow-up care system for children fighting cancer and the survivors in Hungary. A secondary objective was to explore clinical and psychosocial factors affecting HRQoL. Specifically, the analysis aimed to determine whether the MMQL-AF could distinguish between patients with cancer and healthy peers, and whether factors such as parental relationship status or cancer type (e.g., CNS or bone tumors) are associated with HRQoL variation.

Methods

Adaptation process

We received approval from Dr. Smita Bhatia, the original author and creator of the questionnaire to translate the Minneapolis-Manchester Quality of Life Survey of Health Adolescent Form (MMQL-AF) into Hungarian, following a standardized adaptation procedure based on the TARES guideline¹⁷. This study was approved by the Scientific and Research Ethics Committee (TUKEB) of the Health Science Council (ETT) (BM/16408-1/2023). Informed written consent was obtained from all participating patients and their parents for participation in the study and anonymous analysis of the data.

Linguistic adaptation

The linguistic adaptation of the MMQL-AF was conducted using a forward-backward translation process, in accordance with the guidelines established by Beaton et al.¹⁸, and was performed by two independent investigators (M.Sz., K.T.). Nine discrepancies were identified between the translations. In collaboration with a third investigator (A.F.), the translators resolved these discrepancies and determined which translations should be included in the final synthesis. Subsequently, a professional bilingual translator performed the back-translation of the MMQL-AF into Hungarian. A provisional version of the Hungarian MMQL-AF was pretested on 14 patients during their follow-up visits to the Neuro-oncological Outpatient Clinic, Pediatric Center, Semmelweis University. Participants were asked to provide feedback on clarity, comprehensiveness and difficulty of the items in the Hungarian MMQL-AF. None of the participants reported any issues completing the questionnaire.

Patient selection

Study participants were recruited from the Pediatric Center, Semmelweis University, Budapest, Hungary between November 2023 and January 2024. The study interviewers enrolled patients with solid tumors from either the oncological outpatient clinic or inpatient wards. Healthy controls were selected from Babits Mihály Grammar School, Budapest, with participants confirming the absence of any chronic disease. Participants were asked to complete a self-administered questionnaire.

Eligibility criteria: Participants aged between 13 and 20 years – adolescents and young people, based on WHO definition.

Patients diagnosed with cancer who are currently undergoing active treatment (Group A) or have been off treatment for more than one year, and are in complete remission or have stable disease (Group O).

Healthy controls: confirmation of absence of known chronic disease.

Signed informed consent obtained from all parents or legal guardians of minor participants.

Exclusion criteria included cognitive or developmental impairment preventing questionnaire completion, severe clinical instability or hospitalization in critical condition, and refusal or withdrawal of consent.

Data collection

Once the adolescents and young people consented to participate, they completed the survey while waiting for the results in the outpatient clinic. Healthy controls from high schools received an informational booklet and a questionnaire package. Upon agreement from both the patients and their parents/guardians to participate, they completed the questionnaire.

All participants were provided with a second questionnaire and were instructed to return the completed questionnaire within two weeks to evaluate test-retest reliability. This timeframe is widely used in psychometric studies of HRQoL instruments as it balances the need to minimize recall bias while ensuring that the underlying construct is unlikely to have substantially changed¹⁹.

The Minneapolis-Manchester quality of life instrument

The original version of the MMQL-AF was developed for participants between 13 and 20 years, who recovering from serious illnesses, including cancer, but designed to be applicable to both clinical and healthy population. The instrument includes 46 items covering 7 HRQoL domains: (1) physical functioning (9 items), (2) psychological functioning (9 items), (3) social functioning (6 items), (4) cognitive functioning (9 items), (5) body image (6 items), (6) outlook on life (4 items), and (7) intimate relations (3 items). The MMQL-AF is scored on a 4- or 5-point Likert scale from 1 to 4 and 1–5 with higher scores indicating greater HRQoL. The questionnaire was originally developed to assess the well-being of adolescents fighting with or recovering from serious illnesses such as cancer and has also been applied more widely in various health contexts. The questionnaire helps clinicians and researchers to evaluate the long-term impact of illness and treatment, monitor recovery, and guide supportive care interventions¹³.

PedsQL SF15 questionnaire

The Hungarian version of the Pediatric Quality of Life Inventory (PedsQL) Version 4.0 SF15 is a validated tool used to assess the QoL of children and adolescents in generic population. In the Hungarian adaptation²¹the instrument showed good feasibility, with low missing-item rates (1.8–2.3). Most scales achieved Cronbach’s α > 0.70, indicating acceptable internal consistency, and intraclass correlation coefficients (ICCs) between child and parent reports ranged from 0.52 to 0.77, reflecting moderate to good agreement. The study also supported construct validity. This questionnaire measures various domains to evaluate physical, emotional, social, and school functioning, reflecting the overall impact of cancer and its treatment on a patient’s daily life. The measured domains were physical health (8 items), emotions (5 items), social relations (5 items), and school (5 items). Each item is scored on a 5-point Likert scale from 0 to 4, with higher scores indicating a better quality of life. The total QoL score was calculated by averaging the domain scores, providing an overall view of the patient’s well-being. Given its broad conceptual overlap with the MMQL-Adolescent Form, particularly in areas such as emotional well-being, social functioning, and physical health, the PedsQL serves as an appropriate external instrument for assessing concurrent convergent validity^20,21.

Additional subgroup comparisons

The secondary objective of our study was to assess whether the diagnosis, age groups, and socioeconomical factors, such as parental education and parental relationships can have an impact on HRQoL. Participants were categorized into two age groups: 13–16 years and > 16 years. We hypothetized that adolescents aged 13–16 and 16–18 differ in cognitive and emotional maturity. Two groups were assigned based on parental education level, if the patients’ mother and father have graduated or not. The questionnaire provided five response options on parental relationship status. For the analysis, we grouped participants according to whether their parents lived together or separately. To examine the impact of disease type on HRQoL, we categorized cancer patients into diagnostic groups, including bone- and soft-tissue tumors, central nervous system (CNS) tumors, neurofibromatosis, thyroid cancer, and other malignancies (Table 1). Due to small sample sizes and noting that there were no difference between the HRQoL of patients on and off therapy, we decided to group them up to assess the impact of their diagnosis. For patients who had completed treatment, we assessed whether the duration since treatment completion influenced their HRQoL. Participants were categorized into two groups: short-term survivors (≤ 5 years post-treatment) and long-term survivors (> 5 years post-treatment).

Table 1.

Demographic characteristics and diagnoses of the study population.

	Patients on active treatment (Group A)	Patients off treatment (Group O)	Healthy controls (Group C)
Sex
Female (%)	22 (47.83)	27 (58.70)	24 (52.17)
Male (%)	24 (52.17)	19 (41.30)	22 (47.83)
Age
Mean (SD)	15.33 (1.85)	17.14 (2.20)	16.87 (1.03)
Diagnoses
Central nervous system (CNS) tumors (%)	14 (30.43)	19 (41.30)
Pilocytic astrocytoma	5 (10.87)	2 (4.35)
Ganglioglioma	5 (10.87)	3 (6.52)
Medulloblastoma	2 (4.35)	8 (17.39)
Low-grade glioma	1 (2.17)	3 (6.52)
High-grade glioma	1 (2.17)	-
Craniopharyngioma	-	1 (2.17)
Germinoma	-	2 (4.35)
Bone- and soft-tissue tumors (%)	11 (23.91)	3 (6.52)
Ewing’s sarcoma	4 (8.7)	1 (2.17)
Rhabdomyosarcoma	5 (10.87)	-
Osteosarcoma	2 (4.35)	-
Giant-cell tumor	-	1 (2.17)
Teratoma	-	1 (2.17)
Neurofibromatosis (%)	13 (28.26)	6 (13.04)
Thyroid cancer (%)	6 (13.04)	11 (23.91)
Non-Hodgkin lymphoma (%)	2 (4.35)	-
Neuroblastoma (%)	-	7 (15.22)

Statistical analysis

The cross-cultural adaptation process followed internationally accepted guidelines [17; 18]. Descriptive statistics were used to summarize demographic and clinical characteristics.

To assess structural validity, we performed a confirmatory factor analysis (CFA), applying maximum likelihood estimation to test the hypothesized factor structure. Model fit was evaluated using several indices, including the Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR). Goodness-of-Fit Index (GFI) and Tucker-Lewis Index (TLI) were also examined. Model fit was considered acceptable if CFI and NFI were ≥ 0.90, RMSEA ≤ 0.08, and SRMR ≤ 0.10; and good if CFI and TLI were ≥ 0.95, RMSEA ≤ 0.06, and SRMR ≤ 0.08. For GFI, values ≥ 0.90 were considered acceptable^22,23.

Modification indices (MIs) indicate how much the model fit would improve if a fixed parameter (e.g., a path or correlation) were freely estimated. Items with high MIs (10 or higher) were reviewed to identify potential sources of model misfit. Five items (C1d, C1e, D1d, E6, and G2) showing consistently large MIs were removed to enhance model parsimony and fit. After removing the items, we examined the specific reasons why these items may have contributed to model misfit²⁴.

Internal consistency was assessed by calculating Cronbach’s alpha coefficients, and the Intraclass Correlation Coefficient (ICC) was applied for the test-retest reliability. Exploratory comparisons across the different parameters were made by using the Mann-Whitney U-test or the Kruskal-Wallis rank sum test. The significance level was set at p < 0.05. Post hoc multiple comparisons were adjusted using the Tukey Honestly Significant Difference (HSD) test. All statistical analyses were conducted using R (version 4.3.2, R Core Team, 2023. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. URL: https://www.R-project.org/)²⁵.

Results

Study population

The current study included 46 adolescents and young people with cancer who were actively undergoing treatment, 46 patients who had completed cancer treatment at least one year earlier and now have stable disease with watch-and-wait therapy, or are in complete remission, and 46 healthy controls with no history of cancer or other chronic diseases. Purpose of the three-group comparison: Participants were divided into three groups to examine known-group validity and assess how treatment status and medical history influence perceived HRQoL. This design allowed for comparison across different health states and recovery phases, which is a common approach in HRQoL instrument validation studies [12;13;15].

We distributed the questionnaire to 107 patients in our Clinic; 15 refused to participate. The majority of the cancer patients were diagnosed with central nervous system tumors, bone- and soft-tissue tumors, or patients with neurofibromatosis who had been treated for malignancies. In the present study, we included 138 participants for 46 items, corresponding to a patient-to-item ratio of 3:1. Table 1 provides a detailed overview of the study population characteristics.

Internal consistency

Cronbach’s alpha was calculated for each domain to confirm the scale structure of the Hungarian version of the MMQL-AF. Internal consistency was excellent for the overall score of the questionnaire (0.92; 95% CI 0.86–0.94). The results showed sufficient Cronbach’s alpha coefficients for each of the seven factors individually (0.73; 95% CI 0.65–0.91 to 0.90; 95% CI 0.87–0.92) (Supplementary Material 1).

Questionnaire adequacy

Initial results indicated poor model fit (RMSEA = 0.074; 90% CI: 0.070–0.080; CFI = 0.748; TLI = 0.731; SRMR = 0.086; GFI = 0.94; TLI = 0.765). After the removal of items with high modification indices (C1d, C1e, D1d, E6, and G2), the model demonstrated improved fit (RMSEA = 0.062; 90% CI: 0.055–0.069; CFI = 0.828; TLI = 0.814; SRMR = 0.079; GFI = 0.962). Modification indices are shown in Supplementary material 2. While the CFI and TLI remained below the conventional threshold, RMSEA, SRMR and GFI were within the acceptable ranges. Given the complexity of the model and the cross-cultural adaptation process, the overall model fit was considered adequate for exploratory validation. We evaluated factor loadings before and after item removal. In line with common psychometric criteria, factor loadings above 0.70 were interpreted as excellent, values between 0.50 and 0.69 as acceptable, and values below 0.49 as weak indicators of the corresponding latent construct. (Supplementary material 3 and 4).

Test-retest reliability

The retest achieved a notably high response rate of 90% (124 out of 138). Table 2 shows the test-retest reliability, which was assessed using intraclass correlation coefficients (ICCs). The analysis demonstrated excellent reliability, with ICC values exceeding 0.90 across most domains. However, the body image domain exhibited a lower ICC of 0.78, indicating good to moderate reliability.

Table 2.

Test-retest reliability of the Hungarian version of the MMQL-AF questionnaire.

MMQL domains		First measurement	Second measurement	ICC	CI (95%)
		Mean (SD)	Mean (SD)
Total score		3.66 (0.53)	3.74 (0.52)	0.96	0.94, 0.97
Physical functioning		3.14 (0.69)	3.22 (0.68)	0.90	0.86, 0.93
Psychological functioning		3.80 (0.68)	3.93 (0.63)	0.90	0.86, 0.93
Social functioning		3.86 (0.75)	3.93 (0.73)	0.90	0.86, 0.93
Cognitive functioning		3.72 (0.85)	3.80 (0.8)	0.94	0.91, 0.96
Body image		3.84 (0.74)	3.85 (0.73)	0.78	0.7, 0.84
Outlook on life		3.79 (0.95)	3.99 (0.86)	0.90	0.86, 0.93
Intimate relations		3.69 (0.92)	3.71 (0.88)	0.92	0.88, 0.94

Table legend - Test-retest reliability results for each domain of the questionnaire, assessed using Intraclass Correlation Coefficients (ICCs).

Validity assessment

The content validity assessment indicated that the respondents comprehended the questionnaire items and found them clear and readable. During pilot testing, the MMQL-AF was completed in an average of 11 min. Reliability analyses were conducted also on the final version after item removal. The pilot tested comprehensibility with adolescents, and full content validity assessment was done by the onco-psychologist working in our neuro-oncological ward.

To evaluate the concurrent convergent validity of the Hungarian version of the MMQL-AF, we calculated Spearman’s correlation coefficients between subscale scores and those of the PedsQL 4.0 SF questionnaire. Physical functioning domains were paired, psychological functioning was aligned with the emotional functioning domain, social functioning domains were matched, and the cognitive functioning domain was compared to the school functioning domain. Strong correlations were observed between the corresponding domains of the two instruments (Spearman’s coefficient 0.76–0.85; p < 0.001). Additionally, the intimate relations domain demonstrated a strong association with social functioning, while the body image and outlook on life domains exhibited moderate correlations with the corresponding PedsQL domains (Supplementary Material 5).

To assess the discriminant validity of the Hungarian version of the MMQL-AF, we compared the scores for each subscale of the MMQL-AF between adolescents with cancer and healthy controls, results are shown in Table 3. Results indicated that adolescents who completed had significantly lower scores on the physical functioning scale (mean score 3.01 (SD 0.79) 95% CI 2.78, 3.24 vs. 3.32 (SD 0.54) 95% CI 3.16, 3.48; p = 0.025) and social functioning (mean score 3.71 (SD 0.81) 95% CI 3.47, 3.95) vs. 4.05 (SD 0.53) 95% CI 3.89, 4.21, p = 0.048) compared to healthy controls. Additionally, adolescents receiving ongoing treatment had significantly lower total questionnaire scores than their healthy counterparts did (mean score 3.58 (SD 0.46) 95% CI 3.44, 3.72) vs. 3.81 (SD 0.45) 95% CI 3.68, 3.94; p = 0.012). No differences were observed between the groups in other domains.

Table 3.

Discriminant validity of the Hungarian version of the MMQL-AF questionnaire (mean + SD).

	Group A	Group O	Group C	A vs. C (p)	O vs. C (p)	A vs. O (p)
Total score	3.58 (0.46)	3.59 (0.46)	3.81 (0.45)	0.012	0.083	0.7
Physical functioning	3.08 (0.71)	3.01 (0.79)	3.32 (0.54)	0.083	0.025	0.7
Psychological functioning	3.79 (0.64)	3.80 (0.76)	3.80 (0.64)	0.9	0.8	0.8
Social functioning	3.81 (0.83)	3.71 (0.81)	4.05 (0.53)	0.2	0.048	0.6
Cognitive functioning	3.64 (0.80)	3.57 (1.05)	3.95 (0.61)	0.053	0.2	0.9
Body image	3.69 (0.80)	3.86 (0.81)	3.98 (0.58)	0.15	0.6	0.4
Outlook on life	3.63 (0.97)	3.78 (0.96)	3.96 (0.91)	0.085	0.4	0.4
Intimate relations	3.47 (1.03)	3.72 (0.96)	3.87 (0.71)	0.073	0.6	0.2

Table legend - Comparison of MMQL-AF subscale scores among adolescents with cancer - both actively treated (group A) and post-treatment (group O) - and healthy controls (group C) to assess discriminant validity.

The results of the PedsQL 4.0 SF questionnaire showed that healthy controls scored significantly higher in the total score, as well as in the physical functioning and school functioning domains. (Supplementary Material 6).

Comparative analysis

We analyzed the results based on socioeconomic factors and treatment-related characteristics of the study population.

No significant differences in questionnaire performance were observed between the age groups. (Supplementary Material 7).

Parental education level, assessed separately for mothers and fathers, had no significant impact on reported quality of life (Supplementary Material 8).

Table 4 shows that patients whose parents lived together had significantly higher scores in the outlook on life domain (mean score (SD): 3.97 (0.86) 95% CI 3.79, 4.15 vs. 3.39 (1.01) 95% CI 3.08, 3.70; p < 0.001) and the physical functioning domain (mean score (SD): 3.22 (0.66) 95% CI 3.09, 3.36 vs. 2.97 (0.73) 95% CI 2.75, 3.19; p = 0.044) compared to those patients whose parents lived separately.

Table 4.

Difference between participants with different parental relationship status.

	Together (n = 94)	Separate (n = 44)	p-value
	Mean (SD)
Total score	3.72 (0.52)	3.54 (0.55)	0.1
Physical functioning	3.22 (0.66)	2.97 (0.73)	0.044
Psychological functioning	3.85 (0.66)	3.68 (0.70)	0.2
Social functioning	3.92 (0.71)	3.73 (0.82)	0.3
Cognitive functioning	3.72 (0.81)	3.71 (0.93)	0.8
Body image	3.91 (0.72)	3.69 (0.77)	0.080
Outlook on life	3.97 (0.86)	3.39 (1.01)	< 0.001
Intimate relations	3.72 (0.91)	3.61 (0.95)	0.5

Table legend – Exploratory comparison of MMQL-AF scores based on parental relationship status. Patients whose parents lived together reported significantly higher on the Outlook on Life and Physical Functioning domains compared to those whose parents lived separately, indicating a potential impact of family structure on perceived quality of life.

Table 5 shows how differences were observed between the groups based on diagnoses. In the total mean score (3.45 (SD 0.47) vs. 3.48 (SD 0.57) vs. 3.48 (SD 0.56) vs. 3.77 (0.47) vs. 4.08 (0.50), p = 0.015), outlook on life (2.95 (SD 0.91) vs. 3.71 (SD 0.91) vs. 3.82 (SD 0.85) vs. 3.88 (SD 1.03) vs. 4.19 (0.94), p = 0.021), and social functioning (3.69 (SD 0.74) vs. 3.43 (SD 0.94) vs. 3.86 (SD 0.73) vs. 4.11 (SD 0.51) vs. 4.24 (SD 0.64), p = 0.013) domains. Adolescents and young adults diagnosed with bone- and soft-tissue tumors or CNS tumors scored worse than those in other diagnostic groups. Paired comparisons of the different disease types are presented in Supplementary Material 9.

Table 5.

Performance in MMQL-AF questionnaire of patients with different types of cancer.

	BST (n = 13)	CNS (n = 34)	NF1 (n = 19)	TC (n = 17)	Other (n = 9)	p-value
Total score	3.45 (0.47)	3.48 (0.57)	3.48 (0.56)	3.77 (0.47)	4.08 (0.50)	0.015
Physical functioning	3.04 (0.74)	2.94 (0.71)	2.9 (0.81)	3.14 (0.72)	3.58 (0.67)	0.200
Psychological functioning	3.59 (0.73)	3.8 (0.64)	3.63 (0.69)	3.88 (0.78)	4.28 (0.58)	0.150
Social functioning	3.69 (0.74)	3.43 (0.94)	3.86 (0.73)	4.11 (0.51)	4.24 (0.64)	0.013
Cognitive functioning	3.54 (0.75)	3.49 (0.96)	3.41 (1.02)	3.88 (0.76)	4.03 (1.02)	0.300
Body image	3.78 (0.81)	3.77 (0.72)	3.49 (0.899	3.8 (0.92)	4.39 (0.55)	0.300
Outlook on life	2.95 (0.91)	3.71 (0.91)	3.82 (0.85)	3.88 (1.03)	4.19 (0.94)	0.021
Intimate relations	3.39 (0.80)	3.35 (1.18)	3.63 (0.89)	3.96 (0.78)	4.11 (0.90)	0.120

Table legend – Exploratory comparison of MMQL-AF scores across different cancer diagnostic groups (Bone- and soft-tissue tumors (BST), central nervous system tumors (CNS), neurofibromatosis 1 (NF1), thyroid cancer (TC), others) to assess the impact of disease type on QoL. Comparison of MMQL-AF scores across different cancer diagnostic groups to assess the impact of disease type on HRQoL. Patients with bone- and soft-tissue tumors or central nervous system tumors reported lower scores in these domains compared to those with other malignancies.

The analysis based on how long patients are off treatment revealed no significant association between time since treatment and HRQoL outcomes (Supplementary Material 10).

Discussion

In Hungary, approximately 250 new cases of childhood cancer are diagnosed annually²⁶. Owing to significant advancements in pediatric oncology, the majority of these children and adolescents now have a good chance of recovery²⁷. With significant improvements in survival rates, the long-term side effects of cancer treatment have become increasingly important. In addition to physical endocrinological complications, psychological factors, including anxiety, depression, and social adjustment difficulties, also play a critical role in the well-being of survivors^1–5. HRQoL has emerged as a key outcome measure, encompassing physical, emotional, social, and cognitive functioning. To provide long-term support for childhood cancer survivors in their daily lives and school attendance or employment after completion of treatment, healthcare providers should ensure that long-term follow-up involves HRQoL assessments and appropriate measures from a multidimensional perspective. These should include not only physical, but also psychosocial aspects according to the developmental process²⁸. Although both generic and disease-specific multidimensional measures for pediatric patients have been developed in Western countries, only a limited number of HRQoL measures with established reliability and validity are available for children and adolescents in Hungary.

The MMQL-AF, an internationally used, well-established multidimensional self-report tool, was selected for adaptation. In this cohort, the Hungarian version demonstrated acceptable internal consistency and validity. Content validity was assessed by a clinical psychologist. Construct validity was evaluated using confirmatory factor analysis (CFA) and through known-group comparisons between cancer patients and healthy controls. CFA revealed only modest model fit. While RMSEA and SRMR reached acceptable levels, CFI and TLI remained below conventional tresholds, suggesting limited structural validty. After five items were removed with high modification indices, the fit improved, but it has reduced cross-cultural comparability. Although the CFI remained below the ideal threshold, overall model fit was acceptable for exploratory validation, considering the complexity of the model and the cross-cultural adaptation context. Although the CFA supported the proposed seven-factor structure, the sample size relative to the number of estimated parameters was somewhat limited, which may affect the stability and generalizability of the factor solution. The generally recommended participant-to-item ratio for questionnaire validation and cross-cultural adaptation ranges from 5:1 to 10:1²⁹. In the present study, we included 138 participants for 46 items, corresponding to a ratio of approximately 3:1. Although this is below conventional recommendations, the sample size was determined by the rarity of the cancer diagnosis within the Hungarian population. Therefore, we consider the number of participants acceptable and sufficient to provide meaningful psychometric information within this clinical context. Previous international adaptations of the MMQL-AF, including versions from the United Kingdom, South Korea and Sweden^14–16, have reported varying psychometric performance, particularly in confirmatory factor analysis results. For example, the Korean and UK adaptations demonstrated acceptable model fit only after item-level modifications or exploratory factor restructuring, mirroring our findings regarding the need for item removal to improve model parsimony. Test-retest reliability was strong.

Criterion validity was assessed by calculating Spearman’s correlations between corresponding subscales of the MMQL-AF and the PedsQL 4.0 SF15. Strong correlations were found between corresponding domains of the MMQL-AF and PedsQL questionnaires. Discriminant validity was also established, as adolescents with cancer scored significantly lower on physical functioning compared to healthy controls, especially in physical and social functioning. However, no significant difference was observed between those on active treatment and off treatment, possibly due to persistent long term psychological effects which can blur the differences between the groups, adaptive coping mechanism and resilience in the actively treated group can cause better outcomes than expected³⁰. The limited sample size can also be a reason why no difference presented itself between the two groups.

This study also examined the impact of various factors such as diagnosis and parental relationships on HRQoL. Exploratory analysis suggested that tumor type and family structure were associated with differences in HRQoL. Cancer patients with bone- and soft-tissue tumors and central nervous system tumors had worse outcomes, particularly in the outlook on life domain, than the other groups. This disparity can be attributed to the aggressive nature of the treatment protocols, which may include high-dose chemotherapy, cranial irradiation, and radical surgical interventions, frequently resulting in long-term physical disabilities and functional impairments. Central nervous system tumors are associated with neurological sequelae, including cognitive deficits, motor dysfunction, and behavioral disturbances, which further compromise psychosocial development and academic performance. In contrast, thyroid cancer and gliomas associated with neurofibromatosis typically require less intensive treatment and are associated with fewer debilitating physical or neurocognitive consequences. Together, these factors contribute to the more pronounced decline in health-related quality of life observed in pediatric patients with bone- and CNS tumors^31–34.

Consistent with previous findings in psychosocial oncology, our results suggest that stable family structure and parental cohabitation may be associated with higher HRQoL among pediatric and adolescent patients. Cohabiting parents are more likely to provide a consistent source of emotional support, daily stability, and practical assistance, which can buffer the psychological burden of illness and enhance overall adjustment. A supportive family environment also facilitates effective coping and communication, contributing to better emotional and social functioning during and after treatment³⁵. Parental education level, which is often considered a proxy for socioeconomic status, showed no significant association with QoL outcomes in this population. These results suggest that family structure may play a more significant role than educational background alone in shaping health-related quality of life in pediatric oncology patients^35–38. However, unmeasured confounders (e.g.: household income, psychosocial resources) may influence these findings.

The study employed a rigorous cross-cultural adaptation process, including forward-backward translation and pilot testing, to ensure linguistic and conceptual equivalence of the Hungarian MMQL-AF. The study achieved a test-retest response rate of 90%, indicating strong participant engagement and reinforcing the reliability of the temporal stability findings.

Justification for item removal

To improve overall model fit and enhance the validity of the factor structure, we removed several items due to their high modification indices and psychometric inconsistencies. The rationale for the exclusion of each item is described below.

C1d: “I cannot do many activities because of my health”.

C1e: “I cannot do many activities because of my arms or legs”.

In our analysis, these items scored consistently higher than the other items assessing physical health. We hypothesize that the negatively phrased items may have posed comprehension difficulties for the adolescents, leading to potential underestimation of the intended construct. Physical education in Hungarian schools improved in the last years, focusing on integration of children into classes, even for children with difficulties.

D1d: “I feel lonely.”

This item showed unexpectedly high modification indices, suggesting poor fit within the factor structure, with patients scoring higher than on other aspects of emotional life. The experience of loneliness is multifaceted and may be difficult for children and adolescents to accurately self-assess. With widespread access to online communities, young people are better equipped to cope with social isolation.

E6: “I am uncomfortable with the way my body is developing.”

This item, which addresses body image in the context of bodily development, appear to produce different response patterns compared to other body image items. The phrasing refers to a complex and gradual developmental process, the full understanding of which may be challenging for younger respondents. The negative wording and conceptual difficulty likely contributed to measurement inconsistency.

G2: “Do you have difficulty concentrating at other times (e.g., playing cards, computer games, or reading)?”

Unlike other items assessing cognitive function, this question specifically targets concentration during leisure activities. In the rigid Hungarian school system, children often have little time for free-time activities during the school day, so they may perceive these moments as something they have been looking forward to all day—making it less likely for them to report difficulties concentrating. This emphasis on enjoyable activities may have made it difficult for peers to accurately assess impairments in this area, leading to elevated modification indices and a poor model fit.

Limitations

This study has several limitations that should be considered when interpreting the findings. First, the sample size was relatively small, and single-center recruitment may limit the generalizability of the results, although this is a common challenge in pediatric oncology research given the low incidence of childhood cancer. Future validation efforts will include a multidisciplinary expert panel - comprising psychometricians, pediatric oncologists, and psychologists - to comprehensively evaluate content validity. The cultural and linguistic adaptation process required the removal of several items from the original MMQL-AF to improve model fit; while necessary for psychometric reasons, this may reduce the strict comparability with other international adaptations. The CFA model fit remained suboptimal despite item removal, and the sample-to-parameter ratio was likely insufficient for robust structural validation, future multicenter studies with larger and more diverse samples are needed to ensure robust construct validity of the Hungarian MMQL-AF. The reliance on self-reported data may introduce recall or social desirability bias. The heterogeneity of cancer diagnoses and treatment regimens within the study population may have influenced the outcomes and limits the precision of subgroup comparisons, and since we only included patients with solid tumors, results suitable for all childhood cancer patients. Because healthy controls were recruited from a single grammar school, the sample may not fully represent the socioeconomic diversity of the general adolescent population. Higher parental education levels or family resources could have contributed to slightly elevated HRQoL scores in the control group, potentially amplifying differences observed between patients and controls, which may introduce selection bias. Future studies with larger, multicenter cohorts and longitudinal designs are needed to confirm and extend these findings.

Conclusion

The Hungarian version of the MMQL-AF demonstrated acceptable internal consistency and validity in this cohort of adolescents and young adults with cancer. The instrument effectively captured multiple domains of health-related quality of life and may serve as a valuable tool for clinical and research use in Hungary. The removal of five items may limit direct cross-cultural comparability with other MMQL-AF versions. Our findings suggest that tumor type and family structure are associated with differences in HRQoL. The Hungarian MMQL-AF provides clinicians with a practical and psychometrically sound instrument for routine assessment of health-related quality of life in adolescent cancer patients. Its use in clinical follow-up may facilitate early identification of psychosocial challenges and inform tailored supportive interventions.

Implications and contributions statement

This study presents the first validated Hungarian version of the MMQL-AF, which fills a critical gap in HRQoL assessment for adolescent cancer patients and survivors. Its proven reliability and validity support clinical and research use. The findings highlight how cancer type and family structure influence HRQoL, informing personalized survivorship care and support strategies.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CFA

Confirmatory Factor Analysis

CFI

Comparative Fit Index

GFI

Goodness-of-Fit Index

HRQoL

Health-Related Quality of Life

ICC

Intraclass Correlation Coefficient

MI

Modification Index

MMQL

Minneapolis-Manchester Quality of Life

MMQL-AF

Minneapolis-Manchester Quality of Life Instrument – Adolescent Form

PedsQL

Pediatric Quality of Life Inventory

PedsQL-J

Pediatric Quality of Life Inventory – Japanese Version

QoL

Quality of Life

R

(Statistical software) R Language and Environment for Statistical Computing

RMSEA

Root Mean Square Error of Approximation

SF15

Short Form 15 (of the PedsQL)

SRMR

Standardized Root Mean Square Residual

TLI

Tucker-Lewis Index

TUKEB

Tudományos és Kutatásetikai Bizottság (Scientific and Research Ethics Committee)

Author contributions

M.S.: Conceptualization, study design, data analysis, data curation, manuscript drafting, interpretation of findings, manuscript writing. N.F.: Conceptualization, study design, data analysis, statistical analysis, critical manuscript revision. A.F., K.T., O.J.: Data collection, data curation, manuscript review. M.H., B.T., P.H.: Study design, manuscript revision. M.Cs.: Recruitment of participants, clinical input, critical manuscript revision. M.G.: Project supervision, funding acquisition, final approval of the manuscript.

Funding statement

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

Due to ethical restrictions, the data supporting the findings of this study are not publicly available. However, they may be obtained from the corresponding author upon reasonable request and with appropriate institutional approvals.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

The study was approved by the Scientific and Research Ethics Committee (TUKEB) of the Health Science Council (ETT), under reference number BM/16408-1/2023. All procedures involving human participants in the study were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its subsequent amendments. Written informed consent was obtained from all participants, as well as from the parents or legal guardians of minors, prior to enrollment in the study.

Data access statement

Due to ethical restrictions, the data supporting the findings of this study are not publicly available. However, they may be obtained from the corresponding author upon reasonable request and with appropriate institutional approvals.

The authors declare no conflicts of interest related to this study.