Assessing prioritization of a good death across Brazil, Italy, Japan, and the United States: A psychometric evaluation

Abstract

Empirically vetted measures capturing individual priorities for a good death are lacking. This study assessed the measurement properties, measurement invariance, and internal consistency reliability of a prospective measure, Prioritization of a Good Death Index. We used data from a probability sample (N = 4,239) of the general public in Brazil (n = 1,233), Italy (n = 1,000), Japan (n = 1,000), and the United States (n = 1,006). We assessed aims through confirmatory factor analysis (CFA), multi-group CFA, and composite reliability, coefficient H, and average variance extracted. CFA results indicated adequate fit for the full sample, χ²(14) = 174.41, p<.001 (CFI = .96, TLI = .94, RMSEA = .05, SRMR = .93), and within each country subsample. Multi-group CFA results did not support scalar invariance, indicating differing end-of-life values across country subsamples. Sufficient internal consistency reliability was achieved within each country. These initial results support the use of the Prioritization of a Good Death Index as a valid and reliable measure to assess end-of-life preferences in practice and research.