Translating and Validating a Hindi Version of the Spine Functional Index: Cross-cultural Adaptation and Psychometric Testing in an Indian Orthopedic Population

Study design and setting

This prospective, single-center, cross-sectional validation study was conducted in the orthopedics outpatient department of a tertiary-care postgraduate teaching institute in northern India between June 2025 and April 2026. The center functions as a high-volume tertiary referral site serving patients from several neighboring states in northern India, in addition to the local population. The study was approved by the Institutional Ethics Committee (IRB No. DRB/Ortho/2025/56). Written informed consent was obtained from all participants before enrolment. This research was conducted in accordance with the World Medical Association Declaration of Helsinki.

Translation and cross-cultural adaptation

The Hindi SFI (SFI-Hi) was developed using a double forward-backward translation procedure based on Beaton et al. ^[10] and the COSMIN framework ^[11,12], to achieve semantic, idiomatic, experiential, and conceptual equivalence. Two bilingual, native Hindi-speaking translators, independently prepared forward translations; an expert committee comprising an orthopedic spine surgeon, a physiotherapist, a clinical psychologist, and both translators developed a consensus version, which two further bilingual translators back-translated independently. The committee resolved discrepancies by consensus and adapted units for the Indian context (e.g., kilograms rather than pounds) ^[6]. The near-final questionnaire was pilot tested in 22 patients, and three items underwent minor wording changes based on their feedback to produce the final SFI-Hi.

Participants

Eligible participants were aged 18 years or older, literate in Hindi, symptomatic for at least 12 weeks, and diagnosed with a cervical, lumbar, or combined spinal disorder. Exclusion criteria were spinal tumor, infection, or fracture; pregnancy; systemic rheumatological or inflammatory disease; previous spinal surgery; or neurological or psychiatric conditions likely to interfere with self-reported assessment. Of 204 eligible patients approached, 186 (91.2%) consented and completed baseline assessment; 90 clinically stable participants repeated the SFI-Hi after 5–7 days for test-retest analysis.

Outcome instruments

All participants completed the 25-item SFI-Hi (response options “Yes,” “Partly,” and “No”), the ten-item Hindi Functional Rating Index (FRI-Hi) Feise and Michael ^[13], Hindi SF-12, and a 100-mm pain VAS. Cervical patients also completed the Hindi NDI (NDI-Hi) ^[8,9], and lumbar patients completed the ODI-Hi ^[7]. SFI-Hi total scores were expressed on a 0–100 scale for comparison with the disability-oriented comparators, with higher scores indicating greater functional limitation. Baseline variables, included age, sex, body mass index, residence, education, occupation, smoking status, comorbidities, symptom duration, prior treatment, radiating pain, and magnetic resonance imaging (MRI), confirmed degenerative changes where available.

A priori hypotheses for construct validity

Six hypotheses were prespecified for construct validity: High correlation (r ≥ 0.60) with the FRI-Hi, with the ODI-Hi in the lumbar subgroup, and with the NDI-Hi in the cervical subgroup; high correlation (|r| ≥ 0.60) with the SF-12 physical functioning domain; weak correlation (|r| ≤ 0.30) with the SF-12 mental health domain; and moderate correlation (r = 0.31–0.60) with VAS pain. Construct validity was considered supported if at least 75% of hypotheses were confirmed ^[14].

Responsiveness and MCID

A subgroup of 92 participants who received a defined treatment intervention (physiotherapy, pharmacological management, or a combination) during the study period returned for a scheduled clinical follow-up visit 6 weeks after baseline and completed a repeat SFI-Hi assessment together with a patient-reported global rating of change (GRC) anchor question. Responsiveness was evaluated by comparing baseline and 6-week SFI-Hi scores and calculating the standardized response mean (mean change divided by the standard deviation of the change score). The MCID was estimated using an anchor-based approach, comparing the observed SFI-Hi change among patients reporting meaningful improvement on the GRC anchor with that among patients reporting no meaningful improvement; receiver operating characteristic (ROC) curve analysis was used to identify the SFI-Hi change score that best discriminated between these groups ^[12,14].

Statistical analysis

Item distributions were assessed for skewness (|Z| > 1.96 indicating non-normality), and item-total correlations below 0.30 were flagged for review. Internal consistency was assessed using Cronbach’s alpha (0.70–0.95 interpreted as acceptable to excellent) ^[15]. Test-retest reliability was assessed using the two-way random-effects, single-measure intraclass correlation coefficient (ICC2,1) with 95% confidence intervals, and Bland–Altman analysis was used to assess agreement. Standard error of measurement (SEM) was calculated as SD√(1−ICC), and minimal detectable change at 90% confidence (MDC90) as SEM × √2 × 1.65 ^[5]. Construct validity was assessed using Pearson correlation coefficients, interpreted as low (0.00–0.30), moderate (0.31–0.60), or high (≥0.60) ^[6]. Principal component analysis (PCA) explored factor structure using eigenvalues >1.0 and the scree plot as extraction criteria. Responsiveness was interpreted using conventional standardized-response-mean benchmarks (<0.2 trivial, 0.2–0.5 small, 0.5–0.8 moderate, and >0.8 large) ^[14], and the anchor-based MCID analysis used ROC-derived sensitivity, specificity, and area under the curve (AUC). Floor and ceiling effects were considered present when more than 15% of participants scored at the lowest or highest possible score ^[1,16]. All tests were two-sided, with P < 0.05 considered statistically significant. Analyses were performed using IBM Statistical Package for the Social Sciences Statistics version 31 (IBM Corp., Armonk, NY, USA).

Content validity and item analysis

All 25 items met the prespecified item-level criteria; no item exceeded the skewness threshold or had an item-total correlation below 0.30, so none was removed. Missing responses occurred in fewer than 2% of administrations, most often for items concerning lifting/carrying and assistance with personal care.

Internal consistency and test-retest reliability

Overall Cronbach’s alpha was 0.92 (item-deleted range 0.90–0.93; Table 2). In the ninety-patient retest sample, test-retest reliability was excellent (ICC2,1 = 0.96, 95% CI 0.94–0.98), with an SEM of 2.8% and an MDC90 of 6.5%. Bland–Altman analysis demonstrated a mean test-retest bias of 0.4%, with 95% limits of agreement from approximately −7.5% to +8.2% (Fig. 3).

Structural (factor) validity

PCA identified a single dominant component, with an eigenvalue of 8.74 explaining 35.0% of the total variance; the scree plot showed a clear inflection after the first component (Fig. 4). Item loadings ranged from 0.46 to 0.75, with the highest loadings for difficulty with walking/recreation/sport, transport independence, and getting in and out of chairs, and the lowest for appetite and concentration (Table 3).

Construct validity: Hypothesis testing

Three of the six prespecified hypotheses were confirmed – those concerning the ODI-Hi, the SF-12 mental health domain, and VAS pain (Table 4 and Fig. 5). Correlations with the FRI-Hi (r = 0.56) and NDI-Hi (r = 0.53) were moderate rather than high, and the correlation with the SF-12 physical functioning domain (r = −0.54) did not reach the prespecified threshold. The observed results therefore supported several expected construct relationships but did not meet the predefined criterion of 75% of hypotheses.

Floor and ceiling effects

No participant (0/186) scored at the floor, and 1 (0.5%) scored at the ceiling; both values were well below the prespecified 15% threshold for a clinically important floor or ceiling effect (Fig. 6).

Responsiveness

Ninety-two treated participants completed follow-up assessment at 6 weeks. Mean SFI-Hi score improved from 51.3 ± 15.8 at baseline to 38.1 ± 16.4 at follow-up, a mean improvement of 13.2 points (95% CI, 10.8–15.6; P < 0.001), with a standardized response mean of 0.84, indicating good responsiveness (Table 5). Patients reporting meaningful improvement on the GRC anchor showed a significantly greater SFI-Hi change than those reporting no meaningful improvement.

MCID

Anchor-based analysis identified an estimated MCID of 8 points on the 0–100 SFI-Hi scale. ROC analysis identified an optimal discriminating threshold of 7.8 points, with an AUC of 0.81, sensitivity of 78%, and specificity of 76% (Table 5).

Conflict of Interest:

Nil

Source of Support:

Nil

Consent:

The authors confirm that informed consent was obtained from the patient for publication of this article

How to Cite this Article

Raj S, Kumar D, Kumar V, Jose A, Kumar SR, Goni VG. Translating and Validating a Hindi Version of the Spine Functional Index: Cross-cultural Adaptation and Psychometric Testing in an Indian Orthopaedic Population. Journal of Orthopaedic Case Reports 2026 October;16(10): 415-422.