Scales to evaluate pain in general practice : only the BPI and the SFMPQ are stable and efficient.

Jean Yves Le Reste, Sophie Roignan, Jeanlin Viala, Jerome Fonseca, Sophie Lalande, Delphine Le Goff, Patrice Nabbe

Keywords: pain scale, general practice, validity


Pain is a multidimensional phenomenon that is subjective with physiological and psychological components. It seems necessary to have valid and reliable pain measures. A previous study (presented at EGPRN in 2019) has identified Brief Pain Inventory (BPI), SF MPQ-1 and 2 (Short Form McGill Pain Questionnaire) as reliable. Visual Analog Pain Scale (VAS) and Pain drawing were less reliable but as they are commonly used the research team decided to collect their reliability data. The present study aimed to collect reliability data of these five scales.

Research questions:

What are the reliability data of the BPI, SFMPQ 1 and 2, VAS and pain drawing scales?


A systematic review of literature via PubMed was undertaken with no restriction on the publication date. Only English and French language papers were examined. screening, selection and Inclusion were conducted until June 2019 in accordance with the PRISMA statement. Data of validity (convergent, discriminant, construct, face and responsiveness) were extracted for each four scales.


17 relevant articles were included for BPI, 8 for SF-MPQ-1, 5 for SF-MPQ-2, none for Pain drawing and 2 for VAS. Convergent validity was studied in several articles included for BPI, SFMPQ 1 and 2, VAS. Construct validity was studied for BPI and SF-MPQ. Discriminant validity was studied for BPI in one article. Face validity was studied in one article for both SF-MPQ-1 and 2. Responsiveness was studied in one article for BPI, 2 articles for SF-MPQ-1 and 3 articles for SF-MPQ-2.


Satisfactory validity data were found for BPI and SF-MPQ. Data were inadequate for Pain drawing and insufficient for VAS. The five relevant validities were not all studied for BPI and SF-MPQ. Further studies could supplement these data. It is disturbing that validity of VAS, which is very used in practice is so poorly studied.

Points for discussion:

do you use BPI and SFMPQ?

do you use VAS?

what do you think about reliability and external validity of VAS?