Written and maintained by CASRAI Editorial Board
Published
On this page
The Zarit Burden Interview (ZBI) is a standardized psychometric instrument designed to measure subjective caregiver burden—the emotional, physical, social, and financial strain experienced by informal caregivers. Developed by Steven H. Zarit and colleagues in 1980, the questionnaire operationalizes perceived burden as an independent construct distinct from the care recipient’s objective functional impairment. The standard 22-item instrument (ZBI-22) uses a 5-point Likert scale (0 to 4), producing a composite sum from 0 to 88 where higher scores indicate greater personal and role strain. Widely used in gerontology, dementia trials, and nursing research, the ZBI also exists in validated 12-item (ZBI-12) and 4-item short forms, distributed internationally under formal licensing by Mapi Research Trust through ePROVIDE.
Formula with every symbol explained
The standard Zarit Burden Interview score is the unweighted sum of all administered items:
ZBI_total = ∑ x_i (for i = 1 to k)
Where the mathematical components represent:
- k: The total number of items administered (k = 22 for full ZBI-22; k = 12 for the Bédard et al. short form).
- x_i: The ordinal score for item i, where x_i ∈ {0, 1, 2, 3, 4}. Response anchors are: 0 = Never, 1 = Rarely, 2 = Sometimes, 3 = Quite Frequently, and 4 = Nearly Always.
- ZBI_total: The composite score, ranging from 0 to 88 for ZBI-22, and 0 to 48 for ZBI-12.
When items are missing, treating blanks as zero underestimates burden. If at least 80% of items are completed (at most 4 missing on ZBI-22, or 2 on ZBI-12), analysts calculate a prorated score:
ZBI_prorated = (∑ x_j / m_obs) × k
Where x_j represents observed item scores, m_obs is completed item count (m_obs ≥ 18 for ZBI-22), and k is 22. If fewer than 18 items are answered, code the score as missing.
Worked example with real numbers and the arithmetic shown
Consider a family caregiver in a clinical trial completing all 22 items of the ZBI-22. The 22 recorded responses are:
Items 1 to 10: 2, 3, 1, 2, 3, 2, 1, 3, 2, 1
Items 11 to 20: 0, 2, 3, 1, 2, 2, 3, 1, 1, 0
Items 21 and 22: 2, 3
The total score is computed by direct summation:
ZBI_total = 2 + 3 + 1 + 2 + 3 + 2 + 1 + 3 + 2 + 1 + 0 + 2 + 3 + 1 + 2 + 2 + 3 + 1 + 1 + 0 + 2 + 3 = 40
If the caregiver omits Items 11, 19, and 20, 19 items remain valid (m_obs = 19 ≥ 18). The observed sum is 39 (40 − [0 + 1 + 0] = 39). Applying the prorating formula:
ZBI_prorated = (39 / 19) × 22 = 2.0526 × 22 = 45.16 ≈ 45
Prorating projects the observed response average across the full scale, preventing missing items from artificially reducing the burden estimate.
Interpretation (what the result means, with a small table of reading guidance)
Higher scores indicate greater perceived caregiver burden. Clinical researchers divide ZBI-22 composite scores into four standard severity tiers:
| Score Interval | Burden Classification | Interpretation and Research Action |
|---|---|---|
| 0 to 20 | Little or no burden | Manageable strain; routine observation without targeted psychoeducational intervention. |
| 21 to 40 | Mild to moderate burden | Emerging stress and schedule disruption; support resources indicated. (Worked example score 40 falls at upper boundary). |
| 41 to 60 | Moderate to severe burden | Substantial personal strain; structured respite services or counseling recommended. |
| 61 to 88 | Severe burden | Critical caregiver exhaustion; elevated risk of clinical depression and physical health decline. |
In intervention studies, researchers frequently apply a binary screening cut-off. Summaries in the Shirley Ryan AbilityLab Rehabilitation Measures Database establish that a score of 24 or higher indicates high caregiver burden warranting targeted clinical evaluation. When evaluating physical and psychiatric symptoms, comparing ZBI outcomes against clinical tools such as the PHQ-15 somatic symptom scale helps differentiate somatic manifestations of caregiver stress from situational exhaustion. For the 12-item short form (ZBI-12), a score of 17 or higher identifies clinically significant burden.
Assumptions and common mistakes
When incorporating the Zarit Burden Interview into research protocols, investigators must account for several methodological assumptions:
- Assuming unidimensionality without validation: Although total scores are routinely reported as a composite metric, factor analyses demonstrate that the ZBI captures at least two distinct sub-dimensions: Personal Strain and Role Strain. Reporting subscale scores can reveal intervention targets obscured by an aggregate sum.
- Confusing caregiver burden with care recipient impairment: Objective functional impairment does not correlate perfectly with subjective burden, which is heavily mediated by coping resources, social support, and neuropsychiatric behavioral symptoms of the patient.
- Ignoring ceiling and floor effects: In non-clinical cohorts, the ZBI can exhibit floor effects where participants score under 15. In palliative care, score compression near the ceiling can obscure treatment effects. Researchers should review protocols for managing floor and ceiling effects when designing statistical power analyses.
- Conflating reliability metrics: While the ZBI-22 demonstrates high Cronbach’s alpha (0.88 to 0.93), test-retest reliability fluctuates if patient symptoms change between visits. Reviewing the difference between reliability vs. validity and evaluating test-retest vs. inter-rater reliability ensures appropriate interval selection in longitudinal trials.
- Erroneous reverse coding: None of the items on the standard ZBI-22 are reverse-scored; all items are positively keyed toward greater burden. Analysts occasionally mistake Item 22 as reverse-coded, introducing calculation errors.
- Using unverified translations: Because the ZBI is copyrighted by Dr. Steven H. Zarit and distributed through Mapi Research Trust / ePROVIDE, research teams must obtain certified linguistic validations rather than ad-hoc open-web translations.
How to report it (APA example)
When reporting Zarit Burden Interview results in APA 7th edition style, researchers should document the scale version, licensing authorization, item scoring anchors, missing data handling, descriptive statistics, sample internal consistency, and construct validation evaluating convergent and discriminant validity:
“Subjective caregiver burden was evaluated using the 22-item Zarit Burden Interview (ZBI-22; Zarit et al., 1980), administered under institutional user licensing from Mapi Research Trust. Each item was scored on a 5-point Likert scale ranging from 0 (Never) to 4 (Nearly Always), yielding a total possible score between 0 and 88. Missing items (observed in ≤ 3% of responses) were prorated when participants completed at least 18 items. Across the sample (N = 142), total scores were normally distributed with a mean of 34.20 (SD = 11.45, range = 8–66). The scale exhibited excellent internal consistency in the current cohort (α = .91, McDonald’s ω = .92). Using the established clinical threshold of 24, 61.3% of caregivers (n = 87) met criteria for high burden. Bivariate Pearson correlations supported convergent validity, showing robust positive associations with caregiver somatic distress measured via the PHQ-15 (r = .48, p < .001) and depressive symptoms on the CES-D (r = .56, p < .001). Discriminant validity was supported by a non-significant association between caregiver burden and caregiver chronological age (r = −.06, p = .47).”
How to calculate it in Excel, R or Python
Below are reproducible code implementations for calculating composite ZBI scores and implementing missing data prorating rules across spreadsheet software, R, and Python.
1. Microsoft Excel
Assuming raw item responses for Items 1 through 22 are entered in columns B through W for row 2:
=IF(COUNT(B2:W2)>=18, ROUND((SUM(B2:W2)/COUNT(B2:W2))*22, 0), "Incomplete")
This formula checks whether at least 18 of the 22 items are completed. If valid, it computes the mean of the available items, scales it by 22, and rounds to the nearest whole integer.
2. R (Base R)
The following R function processes a data frame of item responses, enforcing the 80% completion rule (at least 18 completed items for ZBI-22):
compute_zbi22 <- function(data, item_cols) {
items <- data[, item_cols]
n_valid <- rowSums(!is.na(items))
raw_sum <- rowSums(items, na.rm = TRUE)
zbi_score = 18,
round((raw_sum / n_valid) * 22),
NA_real_
)
return(zbi_score)
}
3. Python (pandas and numpy)
In Python data science workflows, vectorized pandas operations ensure rapid processing across large cohorts:
import pandas as pd
import numpy as np
def calculate_zbi22(df: pd.DataFrame, item_columns: list[str]) -> pd.Series:
items = df[item_columns]
valid_count = items.notna().sum(axis=1)
observed_sum = items.sum(axis=1, skipna=True)
prorated = (observed_sum / valid_count) * 22
final_score = np.where(valid_count >= 18, prorated.round(), np.nan)
return pd.Series(final_score, index=df.index, name="zbi_total")
Frequently asked questions
What is the clinical cut-off score for the Zarit Burden Interview?
On the full 22-item scale (ZBI-22), a score of 24 or greater indicates high caregiver burden and heightened risk of secondary anxiety and depression. Scores of 0 to 20 reflect little or no burden, 21 to 40 mild-to-moderate burden, 41 to 60 moderate-to-severe burden, and 61 to 88 severe burden.
How many items are in the Zarit Burden Interview?
The standard full scale has 22 items (ZBI-22). Validated short forms include a 12-item version (ZBI-12) by Bédard et al., an 8-item version, a 4-item screening version, and a 1-item global screener for rapid clinical triage.
Who holds the copyright to the Zarit Burden Interview and how is it licensed?
The copyright is held by Dr. Steven H. Zarit. The instrument is officially distributed and managed by Mapi Research Trust through its ePROVIDE platform (eprovide.mapi-trust.org), where investigators obtain authorized copies, manuals, and standardized translations.
Can the Zarit Burden Interview be used for non-dementia conditions?
Yes. Although developed for dementia caregivers, the ZBI is psychometrically validated in caregivers of individuals recovering from stroke, acquired brain injury (ABI), cancer, COPD, and palliative illnesses.
How should missing items be handled when scoring the ZBI?
Standard guidelines permit prorating if at least 80% of items are answered (≥ 18 on ZBI-22, or ≥ 10 on ZBI-12). The sum of answered items is divided by the number of answered items and multiplied by 22 (or 12). If more than 20% are missing, code the score as missing.
Sources
- Mapi Research Trust / ePROVIDE. (2026). Zarit Burden Interview (ZBI) Catalogue. https://eprovide.mapi-trust.org/instruments/zarit-burden-interview. Checked November 8, 2026.
- Shirley Ryan AbilityLab Rehabilitation Measures Database. (2026). Zarit Burden Interview. https://www.sralab.org/rehabilitation-measures/zarit-burden-interview. Checked November 8, 2026.
- Zarit, S. H., Reever, K. E., and Bach-Peterson, J. (1980). Relatives of the impaired elderly: Correlates of feelings of burden. The Gerontologist, 20(6), 649–655. https://doi.org/10.1093/geront/20.6.649. Checked November 8, 2026.
- Bédard, M., Molloy, D. W., Squire, L., Dubois, S., Lever, J. A., and O’Donnell, M. (2001). The Zarit Burden Interview: A new short version and screening version. The Gerontologist, 41(5), 652–657. https://doi.org/10.1093/geront/41.5.652. Checked November 8, 2026.
- Hébert, R., Bravo, G., and Préville, M. (2000). Reliability, validity and development of short forms of the Zarit Burden Interview in Ontario and Quebec. Canadian Journal on Aging, 19(4), 494–507. https://doi.org/10.1017/S0714980800012480. Checked November 8, 2026.







