September 2008

Objective To compare two methods of developing short forms of the Malaysian Oral Health Impact Profile (OHIP-M) measure. Method Cross sectional data obtained using the long form of the OHIP-M was used to produce two types of OHIP-M short forms, derived using two different methods; namely regression and item frequency methods. The short version derived using a regression method is known as Reg-SOHIP(M) and that derived using a frequency method is known as Freq-SOHIP(M). Both short forms contained 14 items. These two forms were then compared in terms of their content, scores, reliability, validity and the ability to distinguish between groups. Results Out of 14 items, only four were in common. The form derived from the frequency method contained more high prevalence items and higher scores than the form derived from the regression method. Both methods produced a reliable and valid measure. However, the frequency method produced a measure, which was slightly better in terms of distinguishing between groups. Conclusion Regardless of the method used to produce the measures, both forms performed equally well when tested for their cross-sectional psychometric properties. Key words: Oral health-related quality of life, psychometric methods, reliability, shortened version, validity.