Supportive Care Needs Survey: A reliability generalization meta-analysis.

Abstract

The purpose of this study is to investigate the reliability generalization of 2 forms of the Supportive Care Needs Survey (SCNS), the questionnaires commonly used to assess the unmet needs of cancer patients. Reviewed articles were retrieved through databases including PubMed, Ovid, Embase, CINAHL (Cumulative Index to Nursing and Allied Health Literature), Web of Science, Scopus, and ProQuest. The inclusion criteria were quantitative studies that assessed the unmet needs of cancer patients using the SCNS and presented reliability coefficients with sample size. Two independent reviewers examined the studies according to inclusion criteria and quality. The final studies included in the meta-analysis were determined by consensus. A random effects model was adopted for the analysis. To estimate reliability coefficients, the alpha coefficients for each study were transformed into the Z statistic for normalization and back to alpha. The values were weighted by the inverse of the studies' variance. The Higgins I2 statistic was used to test for heterogeneity, and the Egger's test and funnel plot were performed to evaluate publication bias. Out of 12,522 studies, 26 studies were included in the meta-analysis. The overall mean weighted effect size of the SCNS long-form (LF) was 0.90 and the subdomains ranged from 0.90 to 0.97. The overall alpha for the SCNS short-form (SF) was 0.92, and the alphas for the subdomains were between 0.81 and 0.92. The estimated reliability coefficients in both LF and SF were highest in psychological and health information needs and lowest in sexuality. No publication bias was indicated in this study. In this study, the overall reliability of SCNS was presented and the factors affecting the reliability of SCNS were identified. The results of this study may help clinicians or researchers make decisions about selecting tools to measure unmet needs of cancer patients.