| INTRODUCTION: Thermal pattern analysis is used to assess the neurological component of the vertebral subluxation. Interpretation of thermal scans is often a subjective visual process. The present study investigates the reliability of three methods of using the more objective thermal pattern calculator software. METHODS:This study was reviewed and approved by the Institutional Review Board of the Sherman College of Straight Chiropractic. Three independent examiners each compared two thermal scans from the same 30 subjects. Each examiner twice performed analysis by three different methods. Methods 1 and 2 involved vertically aligning scans for analysis. Method 1 also included the addition of usable data points at the top and bottom of the scans. Method 3 left the graphs in situ. The results were evaluated by intraclass correlation coefficients (ICC) and t-tests. RESULTS: Method 1 intra-examiner reliability ICC scores were acceptable (mean of 0.932, ranging from 0.859 to 0.987). Inter-examiner ICC scores for this method demonstrated a mean of 0.876 (range of 0.798 to 0.926). T-test intra-examiner alpha scores for Method 1 revealed one significant difference for examiner A. Method 2 intra-examiner reliability ICC scores were acceptable (mean of 0.923, ranging of 0.790 to 0.981) and inter-examiner reliability ICC scores for this method demonstrated a mean of 0.854 (range of 0.805 to 0.941). T-test alpha scores revealed no significant intra-examiner differences for Method 2. Method 3 intra-examiner reliability ICC scores were high (mean of 0.979, range of 0.936 to 1.0) and inter-examiner reliability ICC scores for this method demonstrated a mean of 0.978 (range of 0.956 to 0.999). T-test alpha scores did not reveal any significant intra-examiner differences for Method 3. The three methods were compared for each examiner. Comparisons of Methods 1 and 2 resulted in ICC scores showing strong agreements (mean of 0.962, range of 0.870 to 0.991). T-test alpha scores revealed no significant differences between these two methods for any of the examiners. Comparison of Methods 1 and 3 resulted in lower ICC scores (mean of 0.639, range of 0.385 to 0.892). T-test alpha scores revealed significant differences between these two methods in 44% of the thermal pattern calculator calculations. Comparisons of Methods 2 and 3 also resulted in lower ICC scores (mean of 0.657, range of 0.437 to 0.891). T-test alpha scores revealed significant differences between these two methods in 61% of the thermal pattern calculator calculations. ICC scores for comparisons of Methods 1 and 2 were, on average, higher than any other method comparison. DISCUSSION: Intra- and inter-examiner reliability is acceptable for thermal pattern calculator procedures. Vertically aligning the graphs increases inter and intra-examiner agreement regarding percent similarity. Leaving the graphs in situ significantly lowers the percent similarity. Including extra data points made no significant difference in the percent of similarity. This abstract is reproduced with the permission of the publisher. |
