Development And Application Of A Syndrome Element Scale For Diabetic Retinopathy

Objective:To develop a Syndrome Element Scale for Diabetic Retinopathy (DR), evaluate it's acceptability, reliability, validity and responsiveness, and establish diagnostic criteria of every Syndrome Element. Finally implement quantitative diagnosis of syndrome through exploring their combination laws in patient with DR.Method:①Extract syndrome element through systematic reviewing the neoteric literature of DR, then propose theory scheme according to the extracted Syndrome Element. The items were developed on the bases of pilot study which have summarized 106 symptoms and signs that concerned with DR. And select the items from it which can be used in diagnosis and differential diagnosis among syndrome elements of DR. Translate these symptoms and signs to items by referring to generally accepted scales. The incipient scale in theory was formed, which domains and items were regarded as syndrome element and symptoms or signs.②Responsiveness analysis,t test, correlation analysis and factor analysis were used in item selection among 130 DR recruited patients. If more than 2 methods indicate that the items would be deleted, this item will be deemed as useless firstly. But the final test scale was will be formed according to the outcome of statistic analysis, theory of TCM and some experts' opinion.③Four indexes such as average time for completing the scale, the rate of reception, completion, recovery were used in evaluating the feasibility. Three methods such as test-test reliability, split-half reliability, and Chronbach'sαCoefficient were used to evaluate the reliability of the scale. Experts assessment, correlation analysis between items and domains were used to examine the content validity, factor analysis was used in examining the construct validity. Paired t-test will be used to evaluate the responsiveness between target population and healthy population.④Maximum likelihood discriminatory analysis method was used to assure the weight of second order which belongs to the items of syndrome element scale. Firstly calculate the conditional probability of second order for each item both diagnosis as "yes" and "no". Secondly, convert conditional probability to index number and establish the index number scale for each syndrome element to realize quantitative diagnosis. At last, substituting the patients'selection to index number scale, index number should be added lengthways, and diagnosis as "yes" or "no" according to the maximum. Implement quantitative diagnosis of syndrome through their combination which Syndrome Elements had been diagnosed.⑤The standard of quantitative diagnosis about each Syndrome Element should be checked by Prospective study through evaluating the indexes such as sensitivity, specificity, accuracy, positive predictive value, and negative predictive value. Results:①The contents included 14 domains (Syndrome Element):Qi deficiency, Yin deficiency, Yang deficiency, Fire (feat), Blood stasis, Qi stagnation, Phlegm-Dampness, Blood deficiency, Liver, Kidney, Spleen, Lung, Stomach, Heart. 49 items were remained which has been selected by different methods to constitute the Syndrome Element Scale for DR.②The result of feasibility show that average time for completing the scale is 18 minutes, the rate of reception is 96.92%, the rate of completing the scale is 99.21%, the rate of completing the items is 99.41%, and the rate of recovery is 100%.③The result of test-retest reliability show that Spearman coefficient correlation of the scale is 0.686 and 9 domains are ranged from 0.535 to 0.713. The result of split-half reliability show that the coefficients of the scale is 0.926 and 11 domains are ranged from 0.664 to 0.895. And the Chronbach's a coefficient of the scale is 0.883, and most domains were above 0.60, which showed the scale had better internal consistency.④Content validity was consented firstly by experts assessment. And the result of correlation analysis show that almost every affiliated item has high coefficient correlation (≥0.4) with it's domain, and has low coefficient correlation (<0.4) with other domains, which show that the scale had better content validity again.15 factors were developed from 49 items by principal component analysis, with explaination of 65.397% of total variance. After rotation of variances, each factor values exceeded 0.4. The result consistent with the theory scheme mostly.⑤The outcomes of responsiveness analysis show that the scores has significant difference (P<0.05) between target population and healthy population not scale but also domains, which showed that the scale had satisfactory item discrimination.⑥The prospective study showed that these new standards had high sensitivity and specificity, and accuracy, positive predictive value, and negative predictive value were satisfied.Conclusion:①The Scale had better feasibility, reliability, validity and responsiveness, and can reflect the condition of illness of DR. So the scale is so reliable and sensitive that can be prepared to establish the standard of quantitative diagnosis about each Syndrome Element.②The prospective study indicated that these new standards had better diagnosis efficacy, and finally realize syndrome element combination laws in patient with DR.