估計中風病患姿勢評估量表之最小重要差異值及可觀實質進步量

Abstract

背景與目的：中風病人姿勢控制量表 (Postural Assessment Scale for Stroke Patients, PASS) 為常用之平衡功能評估工具，其多數心理計量特性驗證完備。然其最小重要差異值 (minimal important difference, MID) 與可觀實質進步量 (substantial clinical benefit, SCB) 皆未知。MID為判斷患者功能變化是否具備最小且有意義之改變量，SCB為判斷患者功能變化是否達到明顯大幅且有意義之改變量。因此MID與SCB為判斷與解釋治療成效所需指標之一。故本研究目的為估計PASS之MID與SCB。 方法：本研究為次級資料分析。使用20點計分巴氏量表 (Barthel index, BI) 之行走功能項目作為PASS之外在效標。於估計MID方面：研究者共納入275筆資料，依據行走項目（0、1、2、3計分）之是否有進步（進步或1分以上）將患者分為二組：進步組與無進步組；於估計SCB方面：研究者共納入243筆資料，依據行走項目是否有可觀進步（進步2分或以上）將患者分為二組：可觀進步組與無可觀進步組。研究者依據不同分組，個別使用接受者操作特徵曲線估計PASS改變分數之最佳切截點以作為MID及SCB之估計值。再以曲線下面積 (area under the curve, AUC) 檢驗以「PASS改變分數」分類「BI行走項目是否（可觀）進步」之整體準確性。 結果：本研究顯示PASS之MID為3分，AUC為0.87。其中進步組共213人 (77%)，未進步組共62人 (23%)。PASS之SCB為8分，AUC為0.86，其中可觀進步組共179人 (73.7%)，無可觀進步組共64人 (26.3%)。MID組與SCB組之AUC結果顯示以BI行走功能改變作為PASS改變之外在效標，整體分類效果為可接受。 結論：本研究顯示患者於PASS上進步3分時，可呈現患者平衡功能之最小且有意義之進步；患者若於PASS上進步8分，可呈現患者平衡功能有明顯大幅且有意義之進步。其結果可提供研究人員療效驗證之新指標，使研究人員可判斷不同介入方式之成本，並同時促進精準醫療；本結果亦可輔助臨床人員判斷平衡功能之治療成效，並協助治療計畫擬定，然應用上有諸多限制，使用者須多加注意。

Background and purpose: Post-stroke patients have balance dysfunctions, which have a great influence on their lives. The Postural Assessment Scale for Stroke Patients (PASS) is commonly used to measure patients’ balance function. Most psychometric properties of the PASS have been examined and are better than those of other balance measures. However, the minimal important difference (MID) and the substantial clinical benefit (SCB) of the PASS remain unknown. MID can be defined as the smallest difference in a patient's functional change score that is considered as important. SCB can be defined as the substantial improvement in a patient's functional change score that is considered as important. Thus, MID and SCB are indices that clinicians or researchers use to evaluate and interpret the treatment effects. This thesis has 2 main purposes. The first purpose is to estimate the MID of the PASS. The second purpose is to estimate the SCB of the PASS. Methods: Secondary data was used in this study. The author used the mobility item of 20-point Barthel index (BI) as the external criterion of the PASS. The author divided the secondary data into different groups depending on the purposes. To estimate the MID, the author recruited 275 patients and divided them into two groups, the improved group and non-improved group, depending on whether the score increased by at least 1 point on the mobility item (0, 1, 2, 3 scoring). To estimate the SCB, the author recruited 243 patients and divided them into two groups, the substantially improved group and non-substantially improved group, depending on whether the score increased by at least 2 points on the mobility item. The author used the receiver operating characteristic curve to estimate the best cut-off point for the MID and SCB. The author then used the area under the curve (AUC) to examine the classification accuracy of the change scores of the PASS compared with the change scores of the mobility item in the BI. Results: Results showed that the MID of the PASS was 3 points and the AUC was 0.87. For estimating the MID, 213 patients (77%) were allocated to the improved group and 62 patients (23%) were allocated to the non-improved group. The SCB of the PASS was 8 points and the AUC was 0.86. For estimating the SCB, 179 patients (73.7%) were allocated to the improved group and 64 patients (26.3%) were allocated to the non-improved group. The AUCs of the MID and SCB groups showed that the classification accuracies were acceptable. Conclusion: The MID of 3 points indicates that a 3 point improvement on the PASS signifies a minimal important and meaningful improvement in balance function. The SCB of 8 points indicates that an 8 point improvement on the PASS represents a substantial benefit and meaningful improvement in balance function. The MID and SCB provide new indices for researchers to evaluate the treatment effect in future studies which use the PASS as an assessment, while comparing the cost of different interventions, and enhancing the precision of treatment. The MID and SCB also help clinicians judge the treatment effect on individual patients, but clinicians must be aware of the limitations when applying the MID and SCB in clinical settings.