CONCORDANCE BETWEEN ANALOGUE AND VERBAL RATING PAIN SCALES: ARE THEY INTERCHANGEABLE?

Abstract

Introduction: Pain and disability are two parameters used to indicate and evaluate treatment in lumbar degenerative spine (LDS). Visual Analogue Scales (VAS) and Verbal Rating Scales (VRS) are frequently used to assess pain intensity. Both scales are validated and they have good psychometric properties.

Aim: To evaluate concordance and grade of interchangeability between the two scales.

Design: Retrospective and observational study with dates collected prospectively.

Outcome measures: Pain intensity was evaluated by VAS. As VRS we used pain categoric question of SF 36. Kappa coefficient was used to measure agreement.

Material and Methods: We analysed 50 patients operated by LDS. 28 women and 22 men, mean age was 50 y (17–81 y). All patients filled preoperatively a set of questionnaires (SF 36, Oswestry Disability Index -ODI-, Core set of Deyo and VAS). We evaluated whether VAS follows or not a normal distribution, correlation and concordance between VAS and VRS. Even we study if there are different related to sex, age or study level. Statistically studies were done using SPSS. We considered p=0’05 as statistically significant

Results: VAS not have a normal distribution (Kolmogorov-Smirnov p=0’000). VAS and VRS have a low, but statistically significant, correlation (Rho-Spearman r=0’408 p=0’004), whatever the two scales have low concordance, with large overlap of responses (Kappa=0’345 IC 95% 0’174–0’543). This disagreement dates are aleatory (RV=0’174) and non systematic (RP=−0’093. RC=−0’00005). Making a stratified study, using analogue scale we not found differences by sex (p=0’283) but using VRS we shown more pain intensity in women (p=0’042) and this dates are associated with significative differences in women’s disability (ODI 46’74 vs 57’79 p=0’05)

Conclusions:

• VAS data should be analyzed using non paramentrics methods because vas have non-linear properties

• VAS and VRS are not interchangeable and they have a low percentage of intra-scale agreement. Disagreement are aleatory and non systematic

• The two scales have different interpretation

• Probably, due to great correlation with disability measured by odi, it is recommended to use vrs

Study limitations: Using pain question of SF 36 we are asking about bodily pain and not specifically lumbar pain, but in a patient who are going to be operated by DLS we thing is a good approach.

Number of patients, although we find clinically and statistically significant differences