Development and validation of the short version of the Psychological General Well-Being Index (PGWB-S)

Background: The PGWBI is a 22-item health-related Quality of Life (HRQoL) questionnaire developed in US which produces a self-perceived evaluation of psychological well-being expressed by a summary score. The PGWBI has been validated and used in many countries on large samples of the general population and on specific patient groups. Recently a study was carried out in Italy to reduce the number of items of the original questionnaire, yielding the creation of a shorter validated version of the questionnaire (PGWB-S). The purpose of the present paper is to describe the methods adopted and to report and discuss the relevance of results. Methods: Data for this study were collected from 4 different population samples: two general population samples a student and a patient sample. On the basis of the results of the first (development) sample population, six relevant items were identified statistically from the original questionnaire and grouped to assemble a new summary scale. Following the newly created 6-item questionnaire was administered in three independent population samples. Descriptive statistics, correlation coefficients, univariate and multivariate regression analyses were used to compare the performance of the long and short questionnaire, within and between population samples and across relevant subgroups. A further independent sample extracted by an ongoing cancer clinical trial served as final validation step. Results: Overall, the questionnaires were administered to 1443 subjects. Six items were selected by a step-wise approach to explain 90% of the variance of the summary measure of the original questionnaire. Response rates reached 100%, while missing items were not observed. University students ( n = 400) showed the highest mean value of the summary measure (75.3); while the patient sample ( n = 28) had the lowest score (71.5). The correlation coefficients between the summary measures and the single items according to the different studies were satisfactory, reaching the highest estimates in the student sample. The internal consistency showed high values of the Cronbach's alpha coefficient ( range 0.80 - 0.92) for all three study samples, coming close to the value of the coefficient established for the original questionnaire ( 0.94). A cross-validation in an independent sample of 755 cancer patients confirmed the item selection procedure and amount of variance explained by the new shorter questionnaire ( ranging from 90.2 to 95.1%, across age and sex strata). Conclusion: The newly identified PGWB-S showed good acceptability and validity for the use in various settings in Italy. The translation of the PGWB-S into different languages, and its use in other linguistic settings will add evidence about its cross-cultural validity.