電腦化進行醫院臨床失智評量表之可行性分析－以南部某醫學中心檢查紀錄為例

目的：比較電腦化及傳統手工失智症的評估效益及其可行性。方法：以366病人所做臨床失智評估量表（CDR）資料（含病人及家屬雙方），以組內相關係數（ICC）、Kappa　係數、卡方檢定（1）電腦化6向度（即記憶、方向感、問題解決能力、社區活動、居家習慣、及自我照顧能力）與（2）分數加總法，對照（3）傳統手工6向度失智分數程度（由0至5分）的一致性及其構面結構的相同性。尤登圖作為醫師對病人家屬解說失智症程度的對照視覺化圖示報表。結果：電腦化法比分數加總法有較高的（1）Kappa係數（0.87，95%CI=0.66,1.00及0.66，95%CI=0.46，0.87）、（2）ICC（0.96，95%CI=0.91，1.0及0.88，95%CI=0.85, 0.90），及（3）較小的卡方值（0.423及17.84）。電腦化法比傳統手工法每100人的失智症評估，可節省專業技術員時間達6.5小時。利用尤登圖可以協助醫師對病人家屬解說失智症程度，提供更多有用的訊息。結論：作者設計的電腦化失智症評估法，結合尤登圖來評估病人的失智症，推薦未來研究者進行更多的臨床檢測。

關鍵字

尤登圖；失智量表；分數加總法；電腦化法；傳統手工法

並列摘要

Objective: To compare the feasibility and efficiency of using computer-aided digital (CAD) and traditional hand-plus-human-brain (THB) methods to score the Clinical Dementia Rating (CDR) Scale in patients with dementia. Methods: The CDR scale was used with 366 outpatients in a hospital using (1) CAD and (2) THB methods to score six dimensions (memory, directional sense, problem-solving skills, community activity, habits at home, and self-care ability), as well as (3) the summing-all-item (SAI) scores in comparison with each other on the CDR scales responded by patients and family members. The multiple-regression method with scores of patients and family members was used to predict the CDR scores derived by the traditional method. Kappa, ICC (intra-class coefficient) and Chisquare tests were done to compare consistency and structural similarity between the traditional and computer-aided methods. A Youden plot was used to create a visual presentation of the CDR results for patients and physicians. Results: In comparing CAD and THB in terms of SAI scores, CAD was more consistent and reached more similar conclusions about the extent of patient dementia as shown by: (1) Kappa coefficient (0.87, 95%CI=0.66,1.00 vs. 0.66,95%CI=0.46,0.87), (2)ICC (0.96, 95%CI=0.91, 1.0 vs. 0.88, 95%CI=0.85, 0.90) and (3) a smaller Chi-square value (0.423 vs. 17.84). The time saving in finishing CDR evaluations per 100 patients using the CAD method could be as much as a 6.5 hour reduction in a professional technician's workload. Using a Youden plot to characterize the patient with dementia offers much more useful information to clinical consultants. Conclusion: The author-designed CAD method along with a Youden plot to assess patients with dementia is recommended to future researchers for use in clinical settings.