Validation of the Zarit Burden Interview for Polish caregivers of individuals with rare diseases: a multidimensional approach to assessing caregiver burden

Abstract

Background

The study aimed to validate the Zarit Burden Interview (ZBI) in a sample of 942 caregivers of patients with rare diseases (RDs) in Poland.

Methods

The validation process included exploratory and confirmatory factor analyses, as well as an assessment of internal consistency and interpretability.

Result

Principal Component Analysis was conducted to determine the underlying structure of the scale, identifying a three-factor solution: Personal Sacrifice, Emotional Strain, and Caregiver Guilt and Future Worries. This model was then tested using Confirmatory Factor Analysis, which demonstrated superior model fit compared to a unidimensional approach (CFI = 0.942, TLI = 0.931, SRMR = 0.052, RMSEA = 0.061). Reliability analysis showed strong internal consistency, with Cronbach’s alpha values of 0.861 for Personal Sacrifice, 0.848 for Emotional Strain, and 0.722 for Caregiver Guilt and Future Worries. Additionally, weighted scores were introduced for better interpretability, and stratified comparisons by age and sex provided normative references.

Conclusions

Overall, the results confirm that the three-factor model of the ZBI effectively captures caregiver burden and can serve as a reliable tool for assessing the impact of caregiving in diverse populations. The final section provides normative data for the three ZBI subscales among Polish caregivers of individuals with RDs Stratified by sex and age, these norms aid clinicians and researchers in contextualizing individual scores, identifying high-risk groups, and guiding targeted support efforts.

Introduction

Caregivers of patients with rare diseases (RDs) face unique challenges that can significantly impact their well-being. The burden associated with caregiving extends beyond physical and financial strain, encompassing emotional distress, social isolation, and psychological exhaustion. Assessing this burden accurately is essential for developing appropriate interventions and support systems tailored to caregivers’ needs [1–4]. One of the most widely used instruments for evaluating caregiver burden (CB) is the Zarit Burden Interview (ZBI), originally designed to measure the impact of caregiving on individuals providing care to patients with dementia [5, 6]. However, its applicability to caregivers of patients with RDs remains underexplored.

Rare diseases, by definition, affect a small percentage of the population, yet they often require long-term, complex, and highly specialized care. Due to the rarity of these conditions, caregivers frequently encounter difficulties in accessing medical resources, receiving adequate social support, and maintaining their own health [7–10]. These challenges necessitate the use of a validated tool to assess CB reliably within this population. The ZBI, with its well-established psychometric properties, has been validated in various caregiving contexts, but its performance in the specific setting of RDs caregiving requires further examination. Although the ZBI has been validated in numerous studies [11–13], its application in the context of RDs remains underexplored. Most prior validations have focused on caregivers of individuals with neurodegenerative, such as dementia, Parkinson’s disease, multiple sclerosis, or chronic conditions, while the specific burden experienced by caregivers of people with RDs has received limited empirical attention, particularly in Poland [11, 14–17].

Existing literature highlights a growing awareness of the multidimensional nature of caregiving in rare disease contexts, but significant gaps remain. While recent studies have documented high emotional burden, anticipatory grief, and caregiver guilt—especially in pediatric and neurologically complex conditions—much less is known about long-term psychological outcomes, coping mechanisms, or resilience factors among RD caregivers [18–20]. Additionally, there is a lack of consensus regarding which aspects of burden are most salient in this population, and existing tools may insufficiently capture components such as chronic uncertainty, future-related anxiety, or system-level barriers [21–24]. These limitations underscore the need to validate widely used instruments like the ZBI in rare disease caregiving to ensure they reflect the full scope of caregiver experiences.

RDs include a diverse array of conditions, most commonly neurological, metabolic, and genetic disorders, each posing unique challenges for caregivers. Neurological conditions like Angelman or Dravet syndromes often involve seizures, cognitive impairments, and motor dysfunction, necessitating continuous supervision, specialized therapy and rehabilitation [25, 26]. Metabolic disorders, such as Gaucher disease, maple syrup urine disease or phenylketonuria, disrupt the body’s ability to metabolize nutrients and typically require lifelong dietary restrictions and careful biochemical monitoring [27]. Genetic disorders, including muscular dystrophy and cystic fibrosis, often impact multiple organ systems and involve progressive disability, demanding coordinated, multidisciplinary care [28]. The broad variability of these diseases places a considerable emotional, physical, and organizational burden on family caregivers, who must constantly adapt to complex care needs and navigate often-disjointed healthcare systems.

The purpose of this study is to validate the ZBI in a population of caregivers for individuals with RDs in Poland. We aim to assess the psychometric properties of the instrument, including its reliability, validity, and factor structure within this unique caregiver group. This represents the first validation of the ZBI in a large and diagnostically diverse sample of Polish caregivers of individuals with RDs and includes both an evaluation of the scale’s factor structure and the development of normative data. In doing so, the study addresses a notable gap in the literature by providing age- and sex-stratified norms to support clinical interpretation and cross-study comparability. Additionally, in light of the ongoing debate regarding the unidimensional versus multidimensional structure of the ZBI, this study also seeks to clarify its dimensionality in the context of rare disease caregiving [29–31]. By doing so, we hope to provide evidence for the suitability of the ZBI as a standardized tool for evaluating CB in RDs settings. Moreover, our findings will contribute to the broader understanding of the caregiving experience in RDs and support the development of targeted interventions aimed at improving caregiver well-being.

This study is particularly important in the Polish healthcare context, where RDs caregiving is mainly informal and lacks adequate institutional support [10]. By validating the ZBI in this population, we seek to offer a reliable assessment tool that can be used in clinical and research settings to better address the needs of caregivers and ultimately improve the quality of life for both caregivers and patients. Our research will also help guide policymakers and healthcare providers in recognizing the specific burdens faced by caregivers of RDs patients and implementing necessary support mechanisms.

Methods

Study design

This study utilized a self-administered, anonymous, computer-assisted online survey to examine the relationship between parenting a person with a RDs and CB.

Participants and setting

Family caregivers of individuals with RDs were recruited through a convenience sampling method, as no Polish RDs patient registry exists. Although definitions of RD vary globally, the EU Orphan Regulation defines a rare (or orphan) disease as one affecting no more than 5 in 10,000 people in the European Economic Area [32]. Based on this definition, Orphanet – Europe’s official database on rare diseases – currently lists over 10,000 such conditions, with new ones continuing to be identified. Importantly, eligibility for orphan designation under EU law depends not only on prevalence, but also on how a “condition” is defined, which may vary based on the specificity of the disease’s definition regional epidemiology, diagnostic criteria, and the availability of diagnostic technologies and research [33, 34].

Patient associations, foundations, and organizations shared the survey link via their websites and Facebook pages. Eligible participants were Polish-speaking adults aged 18 years or older who were parents or family caregivers of a person with a confirmed RD diagnosis, had internet access, and provided informed consent.

In this study, the term family caregiver refers to a relative, most often a parent, guardian, or close family member, who provides regular, unpaid support to a person with a chronic illness, disability, or other long-term health or care need with whom they share a personal relationship [35, 36] This role may involve either primary or secondary caregiving, regardless of whether the caregiver resides with the care recipient. In addition to assisting with daily living activities such as household tasks, personal care, and mobility, family caregivers often provide emotional support, manage medical responsibilities, coordinate care, navigate complex healthcare and social service systems, and act as advocates or surrogate decision-makers.

The sample population comprised 942 caregivers (see Table 1), with 130 (13.8%) identified as male and 812 (86.2%) as female. The age distribution of the sample population was categorized into four groups: 16–29, 30–39, 40–49, and 50 years or older. The largest proportion of respondents, constituting 43.9% of the entire sample, falls within the 40–49 age range (n = 414). This is followed by caregivers aged 30–39 (n = 381; 40.4%). The smallest segment is that of younger caregivers (16–29 years) (n = 34; 3.6%), while those aged 50 and above comprise 12.0% of the total sample (n = 113).

Table 1.

Characteristics of the sample by sex and age of caregivers

	Male	Female	Total
Age (in years)
18–29	2 (1.5%)	32 (3.9%)	34 (3.6%)
30–39	44 (33.8%)	337 (41.5%)	381 (40.4%)
40–49	64 (49.2%)	350 (43.1%)	414 (43.9%)
50 and more	20 (15.4%)	93 (11.5%)	113 (12.0%)
Total	130 (13.8%)	812 (86.2%)	942 (100%)

When disaggregated by sex, a similar pattern emerges. Among the male caregivers (n = 130), those aged 40–49 constitute the predominant demographic (49.2%), followed by those in the 30–39 age range (33.8%), while only 1.5% are aged 16–29, and 15.4% are 50 or older. A similar distribution is observed among women, with 43.1% of the sample falling into the 40–49 age group, 41.5% in the 30–39 age group, 3.9% in the 16–29 age group, and 11.5% aged 50 or older. These figures suggest that middle-aged adults account for the bulk of caregiving responsibilities, with women representing a particularly large share of the sample.

To reflect the diagnostic diversity within the sample, respondents were asked to indicate the rare disease diagnosis of the individual under their care. The reported conditions encompassed a wide spectrum of neurological, metabolic, and genetic disorders, each associated with distinct caregiving challenges. Neurological disorders were most frequently represented and included Duchenne/Becker muscular dystrophy (22.4%), Dravet syndrome (11.3%), Angelman syndrome (12.7%), 22q11.2 deletion syndrome (13.9%), tuberous sclerosis complex (1.5%), neurofibromatosis (3.5%), and PANS/PANDAS (1.5%). Metabolic disorders, such as glycogen storage disease (1.1%) and Niemann-Pick disease (1.1%), were also present in the sample. Genetic syndromes included Williams syndrome (5.1%), Turner syndrome (1.5%), Kabuki syndrome (1.2%), Prader-Willi syndrome (1.1%), and Ehlers–Danlos syndrome (2.0%). Notably, 20.3% of the sample reported caring for individuals diagnosed with one of 144 other rare or ultra-rare conditions, further underscoring the heterogeneity of rare diseases and the complexity and unpredictability of caregiving roles in this population.

Ethical considerations

The study followed the ethical principles of the Declaration of Helsinki [37] and was approved by the Bioethics Committee of Poznan University of Medical Sciences (KB – 228/24).

Research questionnaire

The survey included sociodemographic questions on caregivers and individuals with RDs, followed by the ZBI. The ZBI total score (0–88) sums responses on a 5-point Likert scale, with higher scores indicating greater caregiver burden. In this study, we used the Polish version of the Zarit Burden Interview developed and validated by Szatkowska and Basińska [38]. This version underwent full linguistic and cultural adaptation, including forward-backward translation, pilot testing, and psychometric validation. It was specifically tailored for use among Polish family caregivers and demonstrated robust reliability and construct validity, making it suitable for the RD caregiving context.

Data collection

The survey was conducted from March to August 2024. The research coordinator collaborated with RD patient organizations, which distributed the survey link. Three follow-up reminders were sent in January, March, and June. The estimated completion time was 25 min.

Statistical analysis

This document provides a comprehensive examination of the ZBI using both exploratory (Principal Component Analysis-PCA) and confirmatory (Confirmatory Factor Analysis-CFA) approaches. This dual-analytic framework enables a robust assessment of the scale’s factor structure, offering valuable insights into how different facets of caregiver burden cluster within a given population. All statistical analyses were conducted using R Statistical Software (v4.3.1) [39].

Results

Evaluation of the latent structure of the Zarit burden interview

The ZBI is a caregiver self-report measure used by numerous agencies. We started with an in-depth exploratory examination of the ZBI through PCA. This step was essential for identifying the underlying dimensional structure of the scale’s items before proceeding to confirmatory techniques. The PCA approach seeks to reduce the total number of variables—in this case, the 22 ZBI items—to a smaller set of core latent factors that best explain the shared variance among those items. This approach enables a preliminary evaluation of how well each item contributes to the overall construct of CB. It also informs subsequent decisions about retaining or discarding certain items based on factor loadings and interpretability.

As illustrated in Fig. 1, the eigenvalues corresponding to each principal component are presented in descending order, with the characteristic “elbow” or break in the scree plot becoming particularly evident after the third component. This observation indicates that factors beyond the third component contribute only a marginal incremental variance. Consequently, the first three components collectively capture a substantial portion of the variance in the data, thereby serving as the most meaningful underlying dimensions. This visual evidence supports the decision to focus on a three-factor solution (see Table 2), aligning with the theoretical and empirical rationale for multidimensional CB.

Fig. 1.

Scree plot for 22-items of ZBI

Table 2.

PCA results for three components: number of items reduced based on factor loadings values and complexity measure

Item	Subscale 1: Personal Sacrifice	Subscale 2: Emotional Strain	Subscale 3: Caregiver Guilt and Future Worries	Complexity
Q2	0.691			1.26
Q3	0.630			1.66
Q8	0.716			1.10
Q10	0.593			1.84
Q11	0.738			1.46
Q12	0.682			1.63
Q14	0.584			1.50
Q4		0.654		1.10
Q5		0.721		1.21
Q6		0.705		1.46
Q9		0.712		1.41
Q13		0.661		1.57
Q16		0.566		1.61
Q18		0.621		1.17
Q7			0.477	1.97
Q20			0.859	1.10
Q21			0.848	1.13

Note that for further analysis, we selected items with a strong loading (≥ 0.45) on at least one component (RC1, RC2, or RC3) while ensuring they have a manageable complexity (≤ 2). This decision helps retain items associated with a single or limited number of components, improving interpretability and reducing cross-loading issues.

To test the latent structure of the ZBI, we performed CFA. The analysis evaluated a baseline unidimensional solution and a 3-dimensional configuration based on the explanatory part of our study. The baseline model demonstrated poor fit indices, indicating that a single-factor solution was inadequate for capturing the complexity of CB. In contrast, the 3-dimensional model, encompassing Subscale 1: Personal Sacrifice, Subscale 2: Emotional Strain, and Subscale 3: Caregiver Guilt and Future Worries, demonstrated significantly superior fit indices (CFI = 0.942, TLI = 0.931, SRMR = 0.052, and RMSEA = 0.061). The results suggest that the proposed 3-dimensional model effectively captures the nuanced experiences of CB, supporting its validity and utility in diverse caregiving contexts.

In addition, a reliability assessment of the three ZBI subscales indicates that each demonstrates acceptable internal consistency. Specifically, Cronbach’s alpha is 0.861 for Personal Sacrifice (7 items), 0.848 for Emotional Strain (7 items), and 0.722 for Caregiver Guilt and Future Worries (3 items).

Normalized (weighted) scores for 3-dimensional model

Subscale scores for Personal Sacrifice, Emotional Strain, and Caregiver Guilt and Future Worries were calculated by weighting each item’s response with its standardized factor loading derived from the CFA model. This approach ensures that each item’s empirical contribution is accurately reflected in the final score.

For the subscales Personal Sacrifice and Emotional Strain, which range from 0 to 28, scores below 7 indicate no to mild burden, scores between 7 and 14 represent mild to moderate burden, scores between 14 and 21 signify moderate to severe burden, and scores above 21 reflect severe burden. The Caregiver Guilt and Future Worries subscale ranges from 0 to 12 and is similarly structured. Scores below three on this subscale indicate no to mild burden, scores between 3 and 6 represent mild to moderate burden, scores between 6 and 9 signify moderate to severe burden, and scores above 9 reflect severe burden. The Total Score (17 items) ranges from 0 to 68, with scores below 17 indicating no to mild burden, scores between 17 and 34 indicating mild to moderate burden, scores between 34 and 51 indicating moderate to severe burden, and scores above 51 indicating severe burden. These standardized intervals provide a structured framework for interpreting and comparing caregiving burdens across diverse caregiver populations.

Figure 2 presents a histogram and density plot for the Total Score of the ZBI. At the same time, Table 3 provides descriptive statistics for the three latent subscales—Personal Sacrifice, Emotional Strain, and Caregiver Guilt and Future Worries—following the normalization or weighting of the raw ZBI item scores. The table provides key distributional indices, including the mean, standard error, skewness, kurtosis, and quartiles. It also includes floor (percentage scoring the minimum) and ceiling (percentage scoring the maximum) effects.

Fig. 2.

Histogram and density plot for the total score of the Zarit burden interview

Table 3.

Descriptive characteristics of latent scores of three burden subscales and the total score of the Zarit burden interview

Statistic	Subscale 1: Personal Sacrifice [range: 0–28]	Subscale 2: Emotional Strain [range: 0–28]	Subscale 3: Caregiver Guilt and Future Worries [range: 0–12]	Total score [range 0–68]
Mean	14.8	7.3	6.9	29.0
Standard error of mean	0.20	0.18	0.09	0.39
Skewness	-0.231	0.647	-0.112	0.067
Kurtosis	2.374	2.724	2.598	2.485
Shapiro-Wilk statistic	0.986	0.946	0.981	0.995
Shapiro-Wilk p-value	< 0.001	< 0.001	< 0.001	0.002
Minimum	0.0	0.0	0.0	0.0
Q1	10.2	2.8	5.3	20.3
Median	15.4	6.3	7.0	29.0
Q3	19.5	11.2	9.0	37.5
Maximum	28.0	27.2	12.0	64.7
Floor effect (%)	0.8	7.2	0.2	0.1
Ceiling effect (%)	0.6	0.0	5.5	0.0

In the domain of Personal Sacrifice, which ranges from 0 to 28, the mean (14.8) and median (15.4) both fall within the upper portion of the “mild to moderate” burden range (7–14), approaching the threshold of “moderate to severe” (14–21). The skewness of − 0.231 indicates a marginally left-skewed distribution, suggesting a slight clustering of higher scores. Floor and ceiling effects are minimal, with each accounting for less than 1% of respondents.

The mean score of 7.3 for Emotional Strain (range 0–28) places the typical caregiver near the boundary between “no to mild” (0–7) and “mild to moderate” (7–14), indicating that approximately half of the sample remains in the lower range while the other half extends into higher levels of strain. The median score of 6.3 is marginally lower than the mean, suggesting that some participants report relatively high emotional strain, although most remain in the lower-to-moderate zones. A floor effect of 7.2% is more pronounced than in Personal Sacrifice, reflecting a small but distinct group with minimal emotional burden.

The Total Score, with a range of 0 to 68, exhibited a mean and median of 29.0, suggesting that most caregivers in the sample were classified as “mild to moderate” regarding their caregiving responsibilities. The skewness value (0.067) was close to zero, indicating an almost symmetrical distribution. Additionally, both floor and ceiling effects were negligible, with a percentage of 0.1%. The interquartile range (Q1 = 20.3, Q3 = 37.5) indicates that half of the respondents are between these values, suggesting a broad but not sharply polarized distribution of total burden. Although the Shapiro-Wilk test (p = 0.002) suggests statistically significant non-normality, the departure from normality appears minor in practical terms.

The study found that Personal Sacrifice, Caregiver Guilt and Future Worries exhibited higher levels of burden in the cohort, while Emotional Strain, though slightly lower, remained near the threshold between “no to mild” and “mild to moderate.”

Normative data

The final section of this study presents normative data for the three subscales of the ZBI among Polish caregivers of individuals with RDs. By stratifying the sample by sex and age group, these norms enable clinicians and researchers to interpret individual ZBI scores in a broader, population-specific context. This approach enhances the practical utility of the scale, helping stakeholders identify which demographic subgroups may be at heightened risk of CB and, thus, more in need of targeted support.

As indicated by Table 4, mean scores for Personal Sacrifice, Emotional Strain, and Caregiver Guilt and Future Worries vary by sex and age group. However, considering the subscales’ possible score ranges, the observed differences may be modest. For instance, Personal Sacrifice averages cluster around 14–16 points (on a 0–28 scale), suggesting relatively similar levels of perceived burden across most demographic segments. While slight elevations appear in some subgroups, such as women in the 50 + category, the absolute gaps remain within a few points. A comparable pattern emerges for Emotional Strain, with most scores hovering near the lower-to-mid single digits, implying that large fractions of caregivers experience only mild to moderate emotional burden. Finally, Caregiver Guilt and Future Worries, scaled from 0 to 12, show group means concentrated around 6–8, highlighting a moderate but not extreme degree of worry or self-criticism across the board. Similarly, Total Scores on the ZBI show only modest variation across demographic subgroups, hovering in the mid-to-high 20s out of 68 points. This result places the typical caregiver in the “mild to moderate” burden range, with slight increases in specific segments such as older cohorts or women, but differences remain within a few points. The floor effect remains minimal (with no ceiling effect detected), suggesting that extremes of burden are relatively rare in this population. Our results suggest that while there are discernible variations by sex and age, the overall burden levels remain within a relatively constrained range, underscoring the multifaceted but moderately distributed nature of CB in this population.

Table 4.

Burden scale norms by sex and age group for the Polish population of caregivers of persons with rare disease

	Male			Female
	16–39	40–49	50+		16–29	30–39	40–49	50+
Subscale 1: Personal Sacrifice
Mean	13.8	14.1	14.2		14.6	15.0	14.7	15.7
Standard error of mean	0.87	0.75	1.45		1.22	0.33	0.34	0.68
Minimum	3.4	1.9	2.7		0.0	0.0	0.0	0.8
Q1	8.9	10.7	10.1		8.7	10.6	9.7	11.8
Median	13.2	13.7	14.0		15.2	15.6	15.6	16.8
Q3	18.0	18.1	18.0		19.2	19.7	19.4	20.8
Maximum	27.1	25.4	25.1		27.0	28.0	28.0	28.0
Floor effect (%)	2.2	1.6	5.0		3.1	0.6	1.4	1.1
Ceiling effect (%)	0.0	0.0	0.0		0.0	0.9	0.3	2.2
Subscale 2: Emotional Strain
Mean	6.9	5.9	6.2		7.7	7.3	7.2	8.6
Standard error of mean	0.71	0.60	1.06		1.10	0.29	0.29	0.61
Minimum	0.0	0.0	0.8		0.0	0.0	0.0	0.0
Q1	3.4	2.1	2.0		2.9	2.7	2.8	3.1
Median	6.4	4.1	5.6		5.7	6.4	6.2	8.2
Q3	10.0	9.5	10.3		12.4	11.3	10.8	12.9
Maximum	21.0	19.1	16.9		22.2	22.6	27.2	24.3
Floor effect (%)	6.5	6.2	10.0		9.4	7.4	7.4	7.5
Ceiling effect (%)	0.0	0.0	0.0		0.0	0.0	0.0	0.0
Subscale 3: Caregiver Guilt and Future Worries
Mean	7.0	6.3	6.5		7.9	7.2	6.7	6.6
Standard error of mean	0.41	0.31	0.59		0.48	0.14	0.14	0.27
Minimum	0.5	1.0	1.7		0.0	0.5	0.0	1.0
Q1	5.0	4.4	5.1		6.5	5.8	4.9	4.5
Median	7.0	6.5	6.3		8.0	7.0	6.5	7.0
Q3	9.3	7.7	8.1		9.5	9.5	9.0	8.3
Maximum	12.0	12.0	12.0		12.0	12.0	12.0	12.0
Floor effect (%)	2.2	1.6	5.0		3.1	0.6	0.3	2.2
Ceiling effect (%)	6.5	4.7	10.0		12.5	6.2	4.3	4.3
Total score: ZBI
Mean	27.7	26.3	26.9		30.3	29.6	28.6	30.8
Standard error of mean	1.59	1.35	2.55		2.33	0.63	0.65	1.33
Minimum	4.4	3.4	7.3		0.0	1.5	2.7	1.8
Q1	20.8	20.1	19.8		20.0	20.4	19.6	21.3
Median	27.3	24.7	26.6		29.7	29.8	28.1	32.3
Q3	36.0	34.6	31.9		37.8	38.2	36.6	40.3
Maximum	52.6	51.0	53.2		55.8	55.7	64.7	57.8
Floor effect (%)	2.2	1.6	5.0		3.1	0.3	0.3	1.1
Ceiling effect (%)	0.0	0.0	0.0		0.0	0.0	0.0	0.0

Discussion

This study validated the Polish version of the Zarit Burden Interview in a large and diverse sample of caregivers for individuals with RDs. The results confirmed the scale’s strong psychometric properties, including internal consistency and structural validity. A three-factor solution: Personal Sacrifice, Emotional Strain, and Caregiver Guilt and Future Worries emerged as the most appropriate model for capturing the complexity of caregiver burden in this population. The revised Total Score, based on these three dimensions, offered greater precision than the unidimensional sum of the ZBI-22 items. The findings also revealed sociodemographic differences, with age and gender showing variable associations with burden intensity. These results highlight the multidimensional nature of caregiving experiences and the importance of context-specific tools for their assessment. These findings align with previous validations of the ZBI in diverse caregiving contexts and contribute new evidence supporting its multidimensional structure in the rare disease setting [14, 22, 40].

The primary objective of this study was to adapt the ZBI for use in Poland to assess CB among individuals caring for patients with RDs. Originally developed by Zarit, this tool evaluates the perceived burden experienced by family caregivers [5, 6]. The adaptation and validation of the Polish version confirm its value as a reliable tool for assessing CB in this population. Evidence suggests that a Polish version of the ZBI was in use prior to 2020, as demonstrated by its application in the study by Jabłoński et al. [41], which evaluated the burden experienced by male partners of women with cancer. Similarly, Folwarski et al. used the ZBI to assess the burden among caregivers of patients receiving home enteral nutrition in Poland, identifying significant levels of burden within this population [42]. However, it remains unclear whether these studies used validated Polish versions of the instrument. More recently, a 2024 study by Szatkowska and Basińska further confirmed the good reliability and three-factor structure of the Polish ZBI, reinforcing its validity for measuring CB across diverse research contexts in Poland and supporting its continued use in this field [38]. However, in this validation study, over half of the sample consisted of children and spouses caring for a sick relative, which raises concerns regarding its applicability for parents, who are often the primary caregivers of children with RDs. From their perspective, the current version may not fully capture the specific challenges they face.

While the ZBI has been widely validated across diverse caregiving contexts, it is important to recognize that some dimensions of burden – particularly feelings of guilt or anxiety about the child’s future – may be underrepresented in instruments originally designed for geriatric caregiving. This highlights the need for adapted or supplementary models when assessing caregiver burden in pediatric or rare disease settings, where emotional and anticipatory stressors often differ significantly from those in adult care contexts.

In our study, the ZBI was found to have a three-factor structure utilizing 17 items from the original Zarit scale [5, 6]. The baseline model demonstrated poor fit indices, indicating that a single-factor solution was inadequate for capturing the complexity of CB. The factor structure of the 22-item ZBI has been explored in numerous studies, revealing substantial variability in the number and nature of the factors identified. While the ZBI-22 is frequently used as a unidimensional measure [43–46], many studies suggest that it is a multidimensional construct. Whitlatch, Zarit, and von Eye proposed a two-factor structure, distinguishing between personal strain, such as feeling overwhelmed when around the care recipient, and role strain, such as the negative impact on social life [47]. Siegert et al. also identified two factors—personal strain and role strain—among caregivers of adults with acquired brain injuries, with Rasch analysis confirming their reliability as separate subscales [11]. A three-factor structure has also been observed in multiple studies. The Spanish version of the ZBI-22, according to Martín-Carrasco et al., showed a factorial structure of three dimensions: overload, competence, and dependence, with adequate sensitivity and specificity with mental health [48]. Knight, Fox, and Chou proposed a three-factor model that included embarrassment and anger, self-criticism, and patient dependency [29]. Similarly, caregivers of ALS patients reported a three-factor structure consisting of social restrictions, self-criticism, and anger and frustration [12].

Other studies have identified four-factor structures. Ruiz-Lozano et al. found four dimensions in informal caregivers of patients with severe visual impairment: embarrassment and frustration, self-efficacy expectations, interference in personal relationships, and personal strain [49]. In the Croatian version of the ZBI, Lucijanić et al. identified four factors: personal strain, negative feelings toward the care recipient, perceived negative impact on social life and financial strain, and guilt. While the first three factors were highly correlated with the overall burden score, guilt appeared to be a separate and distinct component [14].

Five-factor structures have also been proposed [13], particularly in research on caregivers of patients with multiple sclerosis and dementia. A study on the Persian version of the ZBI supported a five-factor structure, confirming its validity for assessing CB in this population [15]. Similarly, research on Chinese informal caregivers for dementia patients identified five dimensions: role strain, personal strain, incompetency, dependency, and guilt [50].

This variability in factor structures may be attributed to differences in the populations studied, including caregivers of individuals with conditions such as dementia, Parkinson’s disease, ALS, intellectual disabilities, cancer, acquired brain injuries, visual impairment, or multiple sclerosis. Cultural differences in how CB is perceived and experienced may also play a role, as well as methodological variations in statistical analysis [14, 51]. Some studies rely on exploratory factor analysis to uncover potential latent structures, while others use confirmatory factor analysis to validate predefined models [50, 52–56]. Rasch analysis has further contributed to understanding the dimensionality and psychometric properties of the ZBI [11, 51]. Despite this evidence supporting its multidimensionality, the ZBI-22 is still frequently used as a unidimensional measure to obtain a global score of CB [30, 57–59]. The strong correlations found between the identified factors and the total ZBI score support this practice to some extent. However, recognizing distinct dimensions of burden provides a more nuanced understanding of the caregiver experience and could inform more targeted interventions tailored to specific aspects of caregiver strain [60].

The three-dimensional structure identified in the present study – Personal Sacrifice, Emotional Strain, and Caregiver Guilt and Future Worries – aligns with previous findings emphasizing the multifaceted nature of caregiver burden. Recent studies on caregivers of children with rare epileptic syndromes, such as Dravet syndrome [61], have similarly revealed a distinct pattern of emotional burden arising not only from clinical responsibilities but also from broader challenges including social isolation, occupational disruption, and strained family dynamics. These insights underscore the relevance and necessity of multidimensional models in fully capturing the complexity of caregiver experiences, particularly in the context of rare pediatric diseases [62].

The three-dimensional structure obtained in our validation, appears to be a useful tool for capturing the experiences of caregivers of individuals with RDs. Similarly, the Total Score effectively reflects the overall burden experienced by caregivers. Notably, the low Floor Effect and particularly low Ceiling Effect suggest that the instrument is capable of capturing a wide range of burden levels among caregivers. In our study sample, an uncritical adoption of the Total Score as a simple sum of all ZBI-22 items does not seem appropriate. Instead, the proposed Total Score, calculated as the sum of the three validated dimensions, appears to provide a more accurate representation of CB in this population. The multidimensional nature of the ZBI can be valuable both in clinical practice and research, as it allows clinicians to target specific aspects of CB that might otherwise be overlooked in interventions [63]. Although the ZBI was originally designed as a unidimensional scale by its developer, later studies—including those co-authored by Zarit—have suggested that CB is a multidimensional construct [47, 64]. These studies argue that the Total Score may not fully capture the complexity of the burden experienced by caregivers. Certain factors, such as the sense of fulfillment derived from caregiving, may help moderate the level of CB [63]. An additional issue is the arbitrariness surrounding what constitutes a suitable cutoff score, as well as the classification of CB into different groups based on its severity [30].

Our study also reveals certain differences in CB based on gender and age. Given the context outlined above, this finding is not surprising, although previous research using the ZBI has not always been consistent in confirming such associations. Regarding age, some studies have identified a positive correlation between caregiver age and burden, suggesting that older caregivers may experience greater caregiving overload and more symptoms of irritability and apathy [54, 58, 65]. However, other studies, including those on caregivers of individuals with dementia [66] and Parkinson’s disease [16], found no significant relationship between age and ZBI scores. Studies on caregivers of individuals with schizophrenia [67], Duchenne muscular dystrophy [68], and mechanically ventilated patients [69] have also failed to establish a clear link between age and CB.

Similarly, the relationship between gender and burden remains inconclusive [14]. Some studies have reported higher ZBI scores among female caregivers across multiple burden dimensions, suggesting that women may experience a greater caregiving strain [70]. However, other studies, particularly those on dementia and Parkinson’s disease caregivers, found no significant differences between men and women [66]. Interestingly, research on pediatric patients with neurofibromatosis type 1 found that male caregivers reported higher burden levels [59]. Studies on caregivers of individuals with schizophrenia, atypical parkinsonian syndromes, and severe visual impairment have also failed to establish a significant association between gender and ZBI scores [49, 67, 71]. These inconsistencies suggest that while age and gender may influence CB, their impact likely varies depending on the specific caregiving context and population under study.

Limitations

Despite nearly a thousand caregivers participating, making this the largest study of its kind in Poland, the actual number of family caregivers is much higher. However, without a national RD patient registry, their total remains unknown. Additionally, not all caregivers are affiliated with online patient associations or support groups that facilitated recruitment. The study also has limitations, including gender bias, as only 130 fathers participated, and potential selection bias due to its online format, which may have excluded caregivers without internet access or those hesitant to use social media. Moreover, while caregivers of individuals with various RDs were included, differences in symptoms and prognoses could have influenced their experiences. However, validating the scale for each condition separately is impractical given the rarity of these diseases. Finally, the study might have benefited from a comparison group of caregivers of healthy individuals or those caring for persons with more common chronic illnesses.

Conclusions

This study validates the three-factor structure of the ZBI in a large sample of Polish caregivers of individuals with RDs. The identified dimensions: Personal Sacrifice, Emotional Strain, and Caregiver Guilt and Future Worries, demonstrate strong internal consistency and offer a more nuanced assessment of CB than a unidimensional approach. The introduction of weighted scores and normative data stratified by age and sex enhances the interpretability and clinical utility of the scale. These findings support the use of the ZBI as a reliable tool for assessing CB in diverse populations, facilitating more targeted interventions and support strategies.

Beyond its methodological contribution, this study has important policy and practical implications. First, the ZBI can now be reliably used in both clinical and research settings to identify RD caregivers at heightened risk of psychological distress or burnout. This enables healthcare professionals to design tailored psychological and social support interventions. Second, the findings underscore the need for systemic recognition of caregiving as a public health priority, especially in the context of RD, where support structures are often fragmented or insufficient. Third, normative data can inform national caregiver screening protocols, helping allocate resources more effectively and equitably. Lastly, our results advocate for integrating caregiver needs into broader health and social policy frameworks, including respite care services, financial support, and caregiver training programs.

These directions highlight the necessity of shifting from individualized coping to institutional responsibility for caregiver well-being. Future research should explore how caregiver burden changes over time and how system-level interventions might mitigate long-term strain.

Abbreviations

CB

Caregiver burden

CFA

Confirmatory Factor Analysis

PCA

Principal Component Analysis

RDs

Rare diseases

ZBI

Zarit Burden Interview

Author contributions

JD and DW conceptualised and designed the study. JD collected the data, coordinated and supervised the study. PJ analysed the data and prepared the tables and figures. All authors conducted the literature search, wrote and revised various versions of the main manuscript text. All authors contributed equally to this paper.

Funding

This research received no funding.

Data availability

The data supporting this study’s findings are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was performed in line with the principles of the Declaration of Helsinki. Ethics and research governance approval were obtained from the PUMS Bioethics Committee (KB – 228/24). Informed written consent was obtained from all individual participants in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.