Interitem Psychometric Validation of the Stanford Integrated Assessment for Transplant Scale Among Thoracic Transplant Candidates

Assessing reliability and validity of SIPAT and opportunities for improvement: A single-center cohort study

“…[9] It also challenges the interpretation of internal consistency for the composite and domain scores, which requires domain unidimensionality. [27] Consistent with recent work, [14] our findings also demonstrate that the SIPAT may bias scores across groups. SIPAT did not demonstrate subgroup validity across insurance type, race/ethnicity, and education: evaluators responded differently to questions based on patient group, controlling for mean group differences.…”

“…[28] For example, recent item-response theory work suggests reducing the number of response categories on some SIPAT questions may improve measurement accuracy across different levels of psychosocial risk. [14] Here, we found 2 pairs of questions were collinear and did not improve model fit: (a1/a2) knowledge of medical illness and transplant process, and (d16/d17) history of substance use and risk of relapse. Our analysis suggests that within each pair, each question provided the same degree of predictive information despite contributing equally to the total score.…”

“…Structural Validity: The SIPAT is multidimensional. [9,14] We evaluated how well the SIPAT's 4-domain design reflected the constructs they were developed to assess. Using a confirmatory factor analysis, we assessed how well, on average, patients' responses to SIPAT questions aligned, or "fit" with the predetermined SIPAT domains.…”

“…Three studies found high inter-rater reliability, [9,12,13] and 2 found good internal consistency reliability of the composite score and poor to good reliability of the domains. [13,14] One study [9] found strong correlations between SIPAT and The Psychosocial Assessment of Candidates for Transplant, [15] demonstrating concurrent criterion validity. Validation studies for translated versions (107 ≤ N ≤ 150) found high inter-rater reliability, [16][17][18][19] moderate concurrent validity, [19] and good internal consistency.…”

“…Validation studies for translated versions (107 ≤ N ≤ 150) found high inter-rater reliability, [16][17][18][19] moderate concurrent validity, [19] and good internal consistency. [16,18] However, a recent prospective single-center study (N = 173) [14] found bias across groups and mixed construct validity. Evidence that SIPAT can predict posttransplant adherence or outcomes is mixed; some report small to moderate associations between SIPAT and posttransplant adherence and rejection, [12,20,21] while others found no associations.…”