Development and validation of the Rapid Positive Mental Health Instrument (R-PMHI) for measuring mental health outcomes in the population

• Main Authors: Vaingankar, Janhavi Ajit, Abdin, Edimansyah, van Dam, Robertus Martinus, Chong, Siow Ann, Tan, Linda Wei Lin, Sambasivam, Rajeswari, Seow, Esmond, Chua, Boon Yiang, Wee, Hwee Lin, Lim, Wei Yen, Subramaniam, Mythily
• Other Authors: Lee Kong Chian School of Medicine (LKCMedicine)
• Format: Article
• Language: English
• Published: 2021
• Subjects: Science::Medicine; Partial Credit Rasch Model; Split-half Sample
• Online Access: https://hdl.handle.net/10356/146304
• Institution: Nanyang Technological University

Background: The multidimensional Positive Mental Health Instrument (PMHI) has 47 items and six subscales. This study aimed to develop and validate a short unidimensional version of the PMHI among Singapore’s adult resident population. Methods: Using pooled data from three earlier studies (n = 1050), PMHI items were reduced by Partial Credit Rasch Model (PCRM) runs in a random split-half sample, while psychometric properties of the resulting measure were tested through confirmatory factor analysis (CFA), item response theory-graded response model and internal consistency reliability in the other half. Its reliability, construct and concurrent validity, agreement with the original scale, floor and ceiling effect, and scale estimates were further investigated in an external representative general population sample (n = 1925). Results: The average age of the participants was around 41 years. Four PCRM re-runs for item selection resulted in a 6-item unidimensional Rapid PMHI (R-PMHI). CFA confirmed the unidimensional structure of the R-PMHI in the internal (RMSEA = 0.075, CFI = 0.985, TLI = 0.974) and external (RMSEA = 0.051, CFI = 0.992, TLI = 0.987) validation samples. In the external validation sample, the R-PMHI met concurrent validity criteria, showing high agreement with the 47-item version with intra-class correlation coefficient of 0.872 (95% CI: 0.861 to 0.882) and low floor and ceiling effects. Weight-adjusted mean (SE, 95% CI) R-PMHI score in the population was 4.86 (0.2, 4.82–4.90). Conclusion: The unidimensional 6-item R-PMHI offers brevity over the original multidimensional measure while appropriately representing the positive mental health construct. Prospective studies are needed to assess its responsiveness and test-retest reliability.