Measurement properties of translated versions of the Shoulder Pain and Disability Index: A systematic review

Abstract

Purpose: To summarise measurement properties of translated versions of the Shoulder Pain and Disability Index (SPADI) and to assess their methodological quality. Methods: Relevant studies testing measurement properties of translated versions of the SPADI in non-specific shoulder pain participants were included from 11 databases (August 2020). Two reviewers independently screened articles and assessed individual measurement property risk of bias using the COSMIN checklist as very good, adequate, doubtful or inadequate. For each measurement property results were pooled and rated sufficient, insufficient, or inconsistent. Synthesised evidence was graded as high, moderate, low or very low (GRADE approach). Results: Thirty-four studies (21 languages and 26 different versions) were included from 4402 articles. A total of 141 measurement properties were reported with 60 rated as very good or adequate. These included; internal consistency (19), test-retest reliability (4), construct validity (6), structural validity (10), measurement error (5), responsiveness (9), and cross-cultural validity (2). Comprehensibility was adequate in the Chinese, German, Nepali, Spanish and Urdu versions. Only the Danish, Dutch and Nepali versions confirmed all, or all but one, of their measurement properties with sound methodology. Pooled results of all measurement properties except structural validity were rated as sufficient. Quality of evidence was graded moderate to high with downgrading due to inconsistent results. Conclusion: Overall evidence suggests the SPADI is valid, reliable and responsive in translated form but less than half the measurement properties tested were of adequate quality. Further testing is required in many languages particularly in; test-retest reliability, measurement error and construct validity.