COCHRANE ROB 2 · 偏誤風險評讀
FFI 總分(6 個月組間)
← 回到「足底筋膜炎」的圖表
總體判斷 OVERALL-High(高風險)
Domain 與 overall 判斷由程式依 Cochrane RoB 2 官方演算法計算;AI 只負責回答下方各訊號問題。符號:+ 低風險、? 有些疑慮、- 高風險(不只靠顏色)。
風險偏誤紅綠燈
- ?Domain 1 隨機化過程:Some concerns(有些疑慮)
- ?Domain 2 偏離預定介入:Some concerns(有些疑慮)
- ?Domain 3 缺失結局資料:Some concerns(有些疑慮)
- -Domain 4 結局測量:High(高風險)
- ?Domain 5 選擇性報告:Some concerns(有些疑慮)
- -總體(Overall):High(高風險)
中文摘要(給一般讀者)
這篇論文的偏誤風險與 limitation
本篇為平行組隨機對照試驗(乾針 vs 單次皮質類固醇注射治療足底筋膜病),主要結局取病人自陳之 FFI 總分於 6 個月的組間比較。隨機化:使用線上亂數產生器(1.1 Y),但完全未描述分配隱匿機制(1.2 NI),基線大致平衡、僅治療前 FFI 總分微小顯著差異但幅度不足以致偏(1.3 PN)。偏離介入(ITT 效果):病人與施治者皆知分組、無假對照(2.1/2.2 Y),未見試驗情境造成之方案偏離;但 2 名合格病人於隨機化後被排除於分析、非完整 ITT(2.6 PN),惟排除量僅 2% 影響有限(2.7 PN)。缺失資料是最大隱憂:含 LOCF 補值在內,6 個月真正有資料者約 91.8%(低於 95%),且 8 例缺失/補值全部集中在療效流失的皮質類固醇組(乾針組 0 例),LOCF 依指引不能視為校正偏誤、亦無敏感度分析(3.1/3.2 PN、3.3 PY、3.4 PN)。結局測量:FFI 量表適當且組間一致(4.1/4.2 N),但屬病人自陳主觀結局、病人未盲化即為評估者,加上研究者方向性期待,知情很可能影響評估(4.3 Y、4.4 Y、4.5 PY),為本結局主要偏誤來源之一。選擇性報告:無事前計畫可查(5.1 NI),但所有次量表/時間點/分析皆完整呈報,無擇優跡象(5.2/5.3 PN)。整體而言,本結局偏誤傾向偏高,主要受無盲化的主觀自陳結局與單側集中之缺失資料驅動;domain 與 overall 判斷交由官方演算法計算。
詳細評讀(給專業讀者)
Domain 1:隨機化過程 — 判斷:Some concerns
偏誤來源:隨機分組的過程——分配序列是否隨機產生、是否隱蔽到分組為止、基線(第一期開始時)是否平衡。
1.1 Was the allocation sequence random?
分配序列是否隨機產生?
- 回答:是(Yes — 有明確證據支持)
- 原文引述:One week later ... we randomly divided patients into 2 groups using an online random number generator (16).一週後……我們用一個線上亂數產生器將病人隨機分成 2 組。
- 理由:明確使用線上亂數產生器產生分配序列,屬隨機化方法,判 Y。
1.2 Was the allocation sequence concealed until participants were enrolled and assigned to interventions?
分配是否隱蔽到分組為止?
- 回答:資訊不足(No information — 原文未提供足以判斷的資訊)
- 原文引述:The numbers, obtained from online software, were matched to the patients on the order of the application.由線上軟體取得的號碼,依申請(就診)順序配對給病人。
- 理由:只描述亂數如何依就診順序配給病人,完全未提及任何分配隱匿機制(不透明信封、集中式/藥局分配、遠端服務等),無法判斷收案者或病人是否能事先得知下一個分配,故資訊不足判 NI。
1.3 Did baseline differences between intervention groups suggest a problem with the randomization process?
(第一期開始時)基線差異是否暗示隨機化有問題?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:There were no differences between groups in terms of sex, age, or FFI score before treatment. (Table 1: FFI total pretreatment Group 1 56.2 ± 8.2 vs Group 2 54.3 ± 5.8, p = .006)兩組在性別、年齡與治療前 FFI 分數上無差異。(表 1:治療前 FFI 總分 第1組 56.2±8.2 對 第2組 54.3±5.8,p=.006)
- 理由:多數基線特徵(性別、年齡、各次量表)皆平衡;雖表 1 顯示治療前 FFI 總分達統計顯著(p=.006),但兩組差距僅約 1.9 分、數值極接近,且乾針組起始分數略高(較差),此微小差異幅度不足以造成介入效果估計的偏誤,屬可與隨機化相容的少數顯著差異。官方指引明言少數 0.05 顯著差異通常視為機遇所致,故判 PN。
Domain 2:偏離預定介入 — 判斷:Some concerns
偏誤來源:實際執行時偏離預定的介入——盲性如何、是否依「隨機分到的組別」分析(ITT)。
2.1 Were participants aware of their assigned intervention during the trial?
參與者是否(在每一期)知道自己的分組?
- 回答:是(Yes — 有明確證據支持)
- 原文引述:Group 1 underwent dry needling, whereas group 2 underwent corticosteroid injection. ... Applications were repeated twice a week for 5 sessions.第1組接受乾針、第2組接受皮質類固醇注射……乾針每週 2 次、共 5 次療程。
- 理由:乾針(每週2次共5次、俯臥插15針)與單次皮質類固醇注射兩種介入形式與療程截然不同,無假針/假注射對照,病人必然知道自己被分到哪一組,判 Y。
2.2 Were carers and people delivering the interventions aware of participants' assigned intervention during the trial?
照護/執行者是否(在每一期)知道分組?
- 回答:是(Yes — 有明確證據支持)
- 原文引述:All dry needling was performed by a single physiotherapist, and all corticosteroid injections were performed by the second author.所有乾針由單一物理治療師施作,所有皮質類固醇注射由第二作者施作。
- 理由:施治者依組別分工(物理治療師施乾針、第二作者施注射),施治者必然知道分組,判 Y。
2.3 Were there deviations from the intended intervention that arose because of the trial context?
是否因試驗情境造成偏離?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:Patients were prohibited from taking any medications during the trial.試驗期間禁止病人服用任何藥物。
- 理由:報告未提及任何因試驗情境(trial context)而生、且與試驗方案不一致的介入偏離;規定禁藥屬方案本身。無證據顯示有試驗情境造成的方案偏離,判 PN。
2.4 Were these deviations likely to have affected the outcome?
這些偏離是否可能影響結局?
- 回答:不需評估(NA)
- 原因:條件題:依 2.3 的答案,本題依 RoB 2 官方路由不需評估(NA)。
2.5 Were these deviations from intended intervention balanced between groups?
這些偏離在兩組間是否平衡?
- 回答:不需評估(NA)
- 原因:條件題:依 2.4 的答案,本題依 RoB 2 官方路由不需評估(NA)。
2.6 Was an appropriate analysis used to estimate the effect of assignment to intervention?
是否用了適當分析(ITT)?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:Two (2%) patients were excluded from the analysis, whereas data from 6 (6.1%) patients were handled by the LOCF method. (Fig. 1: Group I Analysed n=49, Excluded n=0; Group II Analysed n=47, Excluded n=2)2 名(2%)病人被排除於分析之外,另 6 名(6.1%)病人以 LOCF 法處理。(圖1:第1組分析 n=49、排除 0;第2組分析 n=47、排除 2)
- 理由:6 名以 LOCF 補值者仍納入分析(屬 mITT 概念可接受),但另有 2 名已隨機化且合格的病人(搬遷至他城)被直接排除於分析之外。官方指引明言將合格受試者於隨機化後排除屬不適當分析,此非完整 ITT,判 PN。
2.7 Was there potential for a substantial impact (on the result) of the failure to analyse participants in the group to which they were randomized?
未依隨機分組分析,是否可能對結果造成重大影響?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:Two (2%) patients were excluded from the analysis ... 2 of the patients moved to another city.2 名(2%)病人被排除於分析……其中 2 名病人搬到他城。
- 理由:被排除者僅 2 名(2%),結局為連續變項(非罕見事件),排除原因(搬遷)與預後/療效無關,且雙方隨機化人數原為 49:49。此小量排除不足以對結果造成實質影響,判 PN。
Domain 3:缺失結局資料 — 判斷:Some concerns
偏誤來源:結局資料缺失——是否幾乎所有人都有資料、缺失是否可能與病情(結局真實值)相關。
3.1 Were data for this outcome available for all, or nearly all, participants randomized?
是否幾乎所有受試者都有此結局的資料?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:Lost to follow-up: none (Group I) ... Moving to another city (n=2), Lost at random (n=6) (Group II). Missing data at the last control was handled by the carrying forward of the last observation (LOCF) method.第1組失訪:無;第2組:搬到他城 2 名、隨機失訪 6 名。最後一次追蹤的缺失資料以末次觀察值前推(LOCF)法處理。
- 理由:官方指引規定補值(imputed)資料應視為缺失、不算真實結局資料。故 6 個月真正有結局資料者為 98−2(排除)−6(LOCF 補值)=90/98 ≈ 91.8%,低於連續變項常用的 95% 門檻;且全部 8 例缺失/補值皆集中在皮質類固醇組(第1組失訪為 0)。未達「幾乎所有人皆有資料」,判 PN。
3.2 Is there evidence that the result was not biased by missing outcome data?
是否有證據顯示結果未受缺失資料偏誤?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:Another limitation is the LOCF method, which was used for handling missing data from 6 patients (12.2%). Although it may impart a bias, we believed that this method would be less problematic than simply omitting the data.另一限制是用於處理 6 名病人(12.2%)缺失資料的 LOCF 法;作者亦承認此法可能帶入偏誤,但認為比直接刪除資料問題較小。
- 理由:官方指引明言 LOCF 補值不能視為校正缺失資料偏誤,作者自己也承認 LOCF「可能帶入偏誤」;且未進行任何在不同缺失假設下的敏感度分析。缺乏結果未受缺失偏誤的證據,判 PN。
3.3 Could missingness in the outcome depend on its true value?
缺失是否可能取決於結局真實值?
- 回答:可能是(Probably yes — 綜合判斷後傾向「是」)
- 原文引述:Lost to follow-up: none (Group I) ... Moving to another city (n=2), Lost at random (n=6) (Group II); the corticosteroid injection group showed a loss of efficacy at 6 months.第1組無失訪;第2組搬遷 2 名、隨機失訪 6 名;皮質類固醇組在 6 個月時療效流失。
- 理由:所有缺失/補值皆發生在療效於 6 個月流失的皮質類固醇組(乾針組 0 例),兩組缺失比例懸殊(0/49 對 8/49);『隨機失訪 6 名』理由含糊,療效回退的病人可能較易失訪,且 LOCF 把較早(約 3.6 個月)較佳的值前推。缺失有可能取決於結局真值,判 PY。
3.4 Is it likely that missingness in the outcome depended on its true value?
缺失是否很可能取決於結局真實值?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:The average follow-up duration of the patients, whose outcomes were carried forward, was 3.6 (range 3 to 4) months. *Moving to another city (n=2), *Lost at random (n=6).被前推結局之病人平均追蹤時間為 3.6(3 至 4)個月;搬到他城 2 名、隨機失訪 6 名。
- 理由:缺失全落在單一組確實令人擔憂(→3.3 判 PY,屬「有疑慮」),但作者將原因記為『搬到他城』與『隨機失訪』,表面上與結局無明確關聯,尚無足夠證據斷定缺失『很可能』取決於結局真值。故停在有疑慮層級、不升至高風險,判 PN。
Domain 4:結局測量 — 判斷:High
偏誤來源:結局的測量——測量工具是否恰當、評估者是否盲化、評估是否可能受分組知情影響。
4.1 Was the method of measuring the outcome inappropriate?
結局測量方法是否不適當?
- 回答:否(No — 有明確證據支持)
- 原文引述:The FFI score was used for clinical evaluations. The FFI measures pain, disability, and activity limitation. ... The use of the FFI score is a validated and useful method for evaluating PF.臨床評估採用 FFI 分數;FFI 量測疼痛、失能與活動受限……FFI 分數是評估足底筋膜病之經驗證且有用的方法。
- 理由:FFI 為已驗證、適用於足底筋膜病的量表,量測方法本身適當,判 N。
4.2 Could measurement or ascertainment of the outcome have differed between intervention groups?
(每一序列內)結局測量是否可能因介入而異?
- 回答:否(No — 有明確證據支持)
- 原文引述:Patients were assessed at 3 weeks and 6 months after the interventions by a single investigator (E.U.) using the criteria described previously.兩組病人皆由同一位研究者(E.U.)於介入後 3 週與 6 個月、以相同準則評估。
- 理由:兩組皆以同一 FFI 量表、同一位研究者、相同時間點與準則評估,測量方法組間一致,判 N。
4.3 Were outcome assessors aware of the intervention received by study participants?
結局評估者是否知道受試者接受的介入?
- 回答:是(Yes — 有明確證據支持)
- 原文引述:Patients were assessed ... by a single investigator (E.U.) ... in a single-blind manner. The FFI score was used for clinical evaluations.病人由單一研究者(E.U.)以單盲方式評估;臨床評估採用 FFI 分數。
- 理由:本結局為病人自陳的 FFI(疼痛/失能/活動受限自填量表),依官方指引,病人自陳結局的『結局評估者即為病人本人』。病人未被盲化(無假針/假注射、兩種療程明顯不同,見 2.1),故評估者知道分組,判 Y。(論文所稱『單盲』指施評研究者對分組設盲,但無法改變病人本人知情此一事實。)
4.4 Could assessment of the outcome have been influenced by knowledge of intervention received?
結局評估是否可能受介入知情影響?
- 回答:是(Yes — 有明確證據支持)
- 原文引述:The FFI measures pain, disability, and activity limitation.FFI 量測疼痛、失能與活動受限。
- 理由:FFI 為主觀之病人自陳疼痛/失能量表,官方指引明列此類病人自陳結局可能受知道分組所影響,判 Y。
4.5 Is it likely that assessment of the outcome was influenced by knowledge of intervention received?
結局評估是否很可能受介入知情影響?
- 回答:可能是(Probably yes — 綜合判斷後傾向「是」)
- 原文引述:We hypothesized that dry needling would be as effective as corticosteroid injections ... dry needling seems to be a reliable procedure for treating PF, with better outcomes than corticosteroid injection.我們假設乾針與皮質類固醇注射同樣有效……乾針似乎是可靠的療法,結果優於皮質類固醇注射。
- 理由:病人完全未盲化、結局為主觀自陳疼痛/功能,且兩組療程體驗差異極大(5 次親手施針的照護接觸 vs 單次注射),研究者對乾針療效有明確方向性期待。官方指引以『同類療法試驗中病人自陳症狀』為知情很可能影響評估的示例,本例情境相符,判 PY。
Domain 5:選擇性報告 — 判斷:Some concerns
偏誤來源:選擇性報告結果——是否依事前訂定的分析計畫分析、有無從多個結果中挑選有利者報告(交叉另含:是否因偵測到殘留而只報第一期)。
5.1 Were the data that produced this result analysed in accordance with a pre-specified analysis plan that was finalized before unblinded outcome data were available for analysis?
是否依事前訂定(取得未盲化資料前定版)的分析計畫分析?
- 回答:資訊不足(No information — 原文未提供足以判斷的資訊)
- 原文引述:Statistical analyses were performed with SPSS version 21.0 ... In all analyses, p < .05 indicated statistical significance.以 SPSS 21.0 進行統計分析;所有分析以 p<.05 為統計顯著。
- 理由:報告未提及任何試驗註冊號、事前方案或統計分析計畫,無法確認分析是否在取得解盲結局資料前即事前定案,判 NI。
5.2 Is the numerical result being assessed likely to have been selected, on the basis of the results, from multiple eligible outcome measurements (e.g. scales, definitions, time points) within the outcome domain?
此數值結果是否可能依結果,從多個量表/定義/時間點中選出?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:The findings were the same for pain, disability, and activity limitation scores (Tables 1 and 2). ... when the third week and sixth month total FFI scores were calculated, the outcomes in the dry needling group were better ...疼痛、失能與活動受限分數的結果一致(表1、表2)……計算第3週與第6月的 FFI 總分後,乾針組結果較佳……
- 理由:論文將 FFI 全部三個次量表與總分、於 3 週與 6 個月兩個時間點的組內與組間結果都完整列於表 1、表 2,未見從多個量測中擇優呈報的跡象,判 PN。
5.3 Is the numerical result being assessed likely to have been selected, on the basis of the results, from multiple eligible analyses of the data?
此數值結果是否可能依結果,從多種分析中選出?
- 回答:可能否(Probably no — 綜合判斷後傾向「否」)
- 原文引述:Normally distributed data were analyzed by paired samples t testing and independent samples t testing. ... not distributed normally were analyzed by the Wilcoxon signed-rank test ... the Mann-Whitney U test.常態分布資料以配對與獨立樣本 t 檢定分析;非常態者以 Wilcoxon 符號等級檢定與 Mann-Whitney U 檢定分析。
- 理由:分析方法依常態性事先擇定並一致套用於所有量表與時間點,各表結果一致呈報,未見從多種分析中擇優呈報的跡象,判 PN。