Performance Evaluation And Item Optimization Of Patient-Reported Outcome Scale For Tension-Type Headache

ObjectiveThe performance of the first draft of the patient-reported outcome(PRO)scale for tension-type headache developed in previous work was evaluated.The items in the first draft of the scale were quantitatively optimized following the classical test theory to establish a basis for further formation of a mature PRO scale for tension-type headache that could be widely used in clinical practice.MethodsUsing the tension-type headache PRO scale as a measuring tool,the patients with tension-type headache who were treated at the Guangdong Provincial Hospital of Traditional Chinese medicine(TCM)were investigated on-site.Based on the survey data,the feasibility,re-confidence,viability and response of TCM were evaluated.Based on the survey data,the feasibility,reliability,validity,and responsiveness of tension-type headache PRO scale were evaluated.Following the classical test theory,the items were quantitatively selected by discrete trend method,Cronbach’s a coefficient method,distribution of item method,correlation coefficient method,re-test reliability method,and factor analysis method.Based on the common factor results extracted by exploratory factor analysis,the items remaining after screening were dimensionally divided to form the final scale.Results1.Clinical investigation.A total of 288 respondents were included,and 323 valid questionnaires were returned,including 17 patients receiving TCM treatment and 18 untreated patients who completed the questionnaire twice;the average completion time of each questionnaire was 332±112 seconds;the completion of each questionnaire and item was basically good,and the item 35 was missing up to 41.5%,which was directly removed.2.Performance evaluation.Reliability analysis,the Cronbach’s a coefficient and splithalf reliability a coefficient of the scale were greater than 0.7.Re-test reliability,the Spearman correlation coefficient of the data from the two measurements reached more than 0.8,with a P value<0.05.Validity analysis,KMO reached 0.887,Bartlett spherical test tip 4566.028,degree of freedom reached 703,and the P value<0.05;exploratory factor analysis extracted a total of nine factors with eigenvalues greater than 1 to form a factor matrix,the cumulative variance contribution rate was 60.508%.Responsiveness,Mann-Whitney U test was used to analyze the differences in the two measurement data before and after TCM treatment in the scores of various domains of the scale and the whole scale to evaluate the responsiveness of TCM treatment.Measurement results showed that the score of psychological domain,functional domain and the whole scale were P>0.05,without statistical difference;the score of physical domain was P<0.05,with statistical difference.Mann-Whitney U test was used to analyze the differences in the scores of each item,each domain and the overall scale between patients with qi stagnation and blood stasis syndrome and patients with other TCM syndrome to evaluate the discrimination of the scale for different TCM syndrome,and the measurement results suggested P>0.05,with no statistical difference;comparing the two groups of qi stagnation and blood stasis syndrome and qi deficiency and blood stasis syndrome,the results suggested P>0.05 in the scores of other items,each domain and the overall scale except for items 14 and 29,with no statistical significance.3.Item optimization.It was recommended to delete seven items for discrete trend method,five items for Cronbach’s a coefficient method,eight items for distribution of item method,24 items for correlation coefficient method,four items for re-test reliability method,and eight items for factor analysis method.The combined results of the above six methods finally removed eight items,and the remaining 30 items constituted the item pool.Thirty items were further divided into dimensions based on the common factors extracted by exploratory factor analysis,and finally a new scale was collated,containing one disease concept,three major domains,six dimensions,and 30 items.ConclusionAfter clinical investigation,it is verified that the developed scale has good feasibility,with good internal consistency reliability,split-half reliability,and Re-test reliability,but poor construct validity.Furthermore,the scale needs further validation in terms of TCM treatment responsiveness and discrimination of different TCM syndrome.