The reliability, validity and screening effect of the happiness index scale among inpatients in a general hospital

BACKGROUND: The Happiness Index Scale (HIS) is a newly developed scale by our group to screen for common psychological illnesses among general hospital inpatients. This study aimed to analyze the reliability, validity and screening effect of the HIS and to explore its clinical application. METHODS: From April 1, 2021, to December 31, 2021, a total of 8405 continuous inpatients were enrolled from different departments of a large tertiary general hospital with 1385 inpatient beds in Guangzhou, Guangdong Province, China. Using a cross-sectional survey design, each participant was assessed with the Patient Health Questionnaire 9(PHQ-9), Generalized Anxiety Disorder 7 items(GAD-7), Athens Insomnia Scale (AIS), Columbia Suicide Severity Rating Scale (C-SSRS) and HIS within 24 h of admission. McDonald's ω coefficient, the Guttman split-half coefficient and the test–retest reliability coefficient were used to evaluate the reliability of the HIS and the construct validity and criterion validity of the validity tests. Scores on the PHQ-9, GAD-7, AIS, and C-SSRS were used as the gold standard tools to analyze the screening effect of the HIS. RESULTS: The HIS exhibited very good reliability, with a McDonald's ω coefficient of 0.825, a Guttman split-half coefficient of 0.920 and a test–retest reliability coefficient of 0.745 (P < 0.05). Confirmatory factor analysis showed a satisfactory model fitting index with a χ(2)/df = 2.602, a root mean squared error of approximation (RMSEA) of 0.014, a standardized root mean square residual (SRMR) of 0.010, a comparative fit index (CFI) of 0.992, and a Tucker–Lewis index (TLI) of 0.983. The correlation coefficient between the total score of each dimension of the scale and the corresponding criterion was 0.854 ~ 0.949 (P < 0.001). The HIS showed a very good distinguishing effect. The average HIS score of inpatients who screened positive for psychological problems was significantly higher than that of inpatients who screened negative for psychological problems (t = 3790.619, P < 0.001). The effect size was very large (Cohens d = 2.695, 95% CI = 2.630 ~ 2.761). Approximately 90.2% of the positive and negative screening results of the HIS were matched with the gold standard tools, with a kappa value of 0.747 (P < 0.001). The screening effect test showed a sensitivity (true positive rate) of 92.9% and a specificity (true negative rate) of 89.5%. CONCLUSION: The HIS exhibited satisfactory reliability and validity and a clinically meaningful screening effect with a much shorter version compared to the commonly used screening scales. Thus, it could potentially be useful as the first screening step to rule out psychological conditions for inpatients in general hospitals or to remind medical teams of further psychological concerns. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12888-022-04219-0.