Application of Delphi Technique in Identification of Appropriate Screening Questions for Chronic Low Back Pain from Traditional Chinese Medicine Experts' Opinions

Abstract

The goal of the present study was to obtain a standard list of traditional Chinese medicine (TCM) symptoms and signs for screening chronic low back pain (cLBP) from a group of experts and to assess agreement and consistency among their opinions on the items of a questionnaire. The study design involved three rounds of modified Delphi technique, and it was carried out by 13 experts in orthopedics, massage, and acupuncture working in four hospitals affiliated with Yunnan University of Traditional Chinese Medicine, China. The outcome was measured on the 5-score Likert-scale self-administered checklists. A review of eight textbooks identified 12 pain characteristics, 11 associated factors, and 25 physical and tongue diagnostic expressions as important factors in the TCM diagnosis of cLBP. These 48 diagnostic characteristics were rated by 13 experts as "not important" to "very important" on a scale of 1-5. After three rounds of rating, 13 characteristics were eliminated from the list, with the final numbers for each group being 8, 11, and 16, respectively. Seven items based on Western medicine were also added by the experts. The intra-class correlation (ICC) coefficient for agreement among the experts was 0.2 at the end. Intra-rater, between rounds, consecutive pair-wise median kappa values were 0.53 and 0.66. Analysis of variance using items appearing in all three rounds revealed significant effects of expert and group of symptoms and signs (p < 0.001) and nonsignificant differences among scores of the same expert in the three rounds (p = 0.97). Mean score of physical and tongue expressions was significantly (p < 0.001) lower than that of all other groups of symptoms and signs. Modern TCM experts have de-emphasized the items on physical and tongue expressions and have adopted instead those from Western medicine. Intra-expert agreement across items was low, and each expert tended to stick to her/his original opinions.