Validation of prostate cancer index and SF-12 short forms

Assessment of prostate cancer (PCa) specific and generic health-related quality-of-life (HRQOL) is frequently omitted due to several obstacles, such as respondent burden and infrastructure-related limitations. We attempted to reduce the number of items of two commonly used HRQOL assessment tools, namely the UCLA PCa Index (PCI) and the RAND SF-12, with the intent of generating the most parsimonious, yet psychometrically valid and reliable HRQOL assessment tool.

The PCI and SF-12 were administered to 2415 radical prostatectomy patients, and re-tested in a convenience sample of 35 men with PCa. Multivariate linear regression models defined the most predictive and item-reduced SF-12 and PCI item combinations. These were subjected to standard psychometric reliability and validity tests.

The 8-item PCI sexual function (SF) scale was reduced to three items. The 5-item PCI urinary function (UF) scale was reduced to three items. The 6-item SF-12 mental health scale was reduced to three items, and the 6-item SF-12 physical scale was also reduced to three items. The total number of items was reduced from 27 to 12 (44%). The item-reduced scales accounted for over 85% of full-scale variance. All reliability and validity tests yielded highly satisfactory results.

We developed SF-12 and PCI short-forms, which consist of 12 of 27 (44%) original items and can be completed by most men within 2 minutes. The short-forms represent a valid substitute for the full scales, as they provide over 85% of full-scale information and demonstrate excellent reliability statistics. The short forms have the potential for decreasing respondent burden and infrastructure-related requirements, which may in turn promote HRQOL assessment after radical prostatectomy.