Comparing psychotic experiences in low-and-middle-income-countries and high-income-countries with a focus on measurement invariance. Issue 8 (7th June 2022)

• Record Type: Journal Article
• Authors: Jaya, Edo S.; Wüsten, Caroline; Alizadeh, Behrooz Z.; van Amelsvoort, Therese; Bartels-Velthuis, Agna A.; van Beveren, Nico J.; Bruggeman, Richard; Cahn, Wiepke; de Haan, Lieuwe; Delespaul, Philippe; Luykx, Jurjen J.; Myin-Germeys, Inez; Kahn, Rene S.; Schirmbeck, Frederike; Simons, Claudia J. P.; van Haren, Neeltje E.; van Os, Jim; van Winkel, Ruud; Fonseca-Pedrero, Eduardo; Peters, Emmanuelle; Verdoux, Hélène; Woodward, Todd S.; Ziermans, Tim B.; Lincoln, Tania M.
• Is Part Of: Psychological medicine. Volume 52:Issue 8(2022)
• Page Start: 1509
• Page End: 1516
• Publication Date: 2022-06-07
• Subjects: Cross-cultural -- cross-culture -- cross-national -- psychosis -- schizophrenia -- validity; Psychiatry -- Periodicals; Medicine and psychology -- Periodicals; Clinical psychology -- Periodicals; 616.89
• DOI: 10.1017/S0033291720003323
• Languages: English
• ISSNs: 0033-2917

Abstract

Abstract: Background: The prevalence of psychotic experiences (PEs) is higher in low-and-middle-income-countries (LAMIC) than in high-income countries (HIC). Here, we examine whether this effect is explicable by measurement bias. Methods: A community sample from 13 countries ( N = 7141) was used to examine the measurement invariance (MI) of a frequently used self-report measure of PEs, the Community Assessment of Psychic Experiences (CAPE), in LAMIC ( n = 2472) and HIC ( n = 4669). The CAPE measures positive (e.g. hallucinations), negative (e.g. avolition) and depressive symptoms. MI analyses were conducted with multiple-group confirmatory factor analyses. Results: MI analyses showed similarities in the structure and understanding of the CAPE factors between LAMIC and HIC. Partial scalar invariance was found, allowing for latent score comparisons. Residual invariance was not found, indicating that sum score comparisons are biased. A comparison of latent scores before and after MI adjustment showed both overestimation (e.g. avolition, d = 0.03 into d = −0.42) and underestimation (e.g. magical thinking, d = −0.03 into d = 0.33) of PE in LAMIC relative to HIC. After adjusting the CAPE for MI, participants from LAMIC reported significantly higher levels on most CAPE factors but a significantly lower level of avolition. Conclusion: Previous studies using sum scores to compare differences across countries are likely to be biased. The direction of the bias involves both over- and underestimation of PEs in LAMIC compared to HIC. Nevertheless, the study confirms the basic finding that PEs are more frequent in LAMIC than in HIC.