发功表现的独立预测力:控制联盟、期望与安慰剂后,外气疗效还剩多少治疗师效应?

Written by

in

Language

Evidence level: B

传统说”功夫在功外”——但外气研究要回答的是:把患者觉得”这人靠谱”(联盟)、患者本来就觉得会好(期望)、以及纯粹的心理安慰(安慰剂)全部剥离之后,治疗师本人的”发功表现”还能不能独立预测疗效?

正文

外气研究有一个二十年前就暴露的老问题:2008 年一项 112 名膝骨关节炎患者的双盲 RCT 发现,两位治疗师的效果完全不同——治疗师 2 的患者疼痛改善显著优于假治疗对照组,而治疗师 1 的患者改善程度与对照无异(PMID 18654733)。研究者当时的结论很谨慎:”外气疗效似乎依赖于治疗师的某种品质。” 但这个”品质”到底是什么,此后二十年一直没人把它测出来。本文要讨论的,正是把”发功表现”变成一把经过检验的尺子之后,它的预测力能不能在剔除各种”非特异因素”之后依然存活。

先看心理治疗领域怎么解决类似问题。Facilitative Interpersonal Skills(FIS,促进性人际技能)是一套标准化行为任务:给受训治疗师播放一系列”难搞患者”的录像(怀疑、抗拒、情绪崩溃等情境),让他们像真做治疗一样口头回应,再由受过训练的评分者按八个维度打分(语言流畅性、共情、温暖接纳、联盟维系、破裂修复等)。2016 年的一项前瞻性研究发现,44 名受训治疗师在培训初期完成的 FIS 得分,能显著预测一年多以后他们真实接诊患者的症状变化——高 FIS 治疗师的疗效在短程治疗(≤8 次)中显著优于低 FIS 者(PMID 26594945)。更早的 RCT 也显示,FIS 挑选出的高人际技能治疗师,无论联盟质量还是治疗结局都优于低技能者,且这一效应与受训状态无关(PMID 26344392)。换言之,心理治疗已经把”治疗师个体差异”从玄学名词变成了可测量的工程参数。

关键在于:FIS 的预测力必须经过”增量检验”(incremental validity)才算数。因为联盟(患者对治疗关系的评价)本身就是疗效的强预测因子——一项 14,000 余例治疗的元分析显示联盟-结局相关稳定存在(PMID 22922705 证实治疗师层面的联盟变异比患者层面的更影响结局);期望与安慰剂更是公认的”非特异因素”,慢性疼痛患者的积极治疗期望与临床疼痛缓解的相关可高达 r=0.72(PMID 31302644),而 2020 年一项 200 人 RCT(SOPPRANO)显示高期望患者的安慰剂反应会强到把真实药物与安慰剂的差异完全抹平(NNT 从 3.2 恶化到 29.6,PMID 32297945 为同一试验的期望分层分析)。如果 FIS 的预测力只是”高 FIS 治疗师更会让人感觉联盟好、更容易营造期望”的副产品,那它就没有独立价值。

把这条逻辑搬到外气场景,就构成了一个干净的三步检验设计。第一步,用外气版 FIS 任务(按情境视频 + 行为评分改编,见姊妹篇)给 ≥10 名治疗师测”发功表现”分。第二步,做一个多治疗师 × 多患者的盲法 RCT:治疗师与患者都纳入联盟问卷(患者版 + 治疗师版)、期望问卷(治疗前对”气感/疗效”的预期)、以及标准假治疗对照臂来控制安慰剂。第三步,用多层模型做增量检验:先在疗效上拟合”联盟 + 期望 + 治疗臂”的基线模型,再加入治疗师层 FIS 得分,看 Δ方差/ΔAIC 是否显著。若显著,还要做”特异性”检验——把 FIS 得分与通用人际质量(如自我报告社交技能、一般共情量表)同时放入模型,看”发功表现”是否比”通用人际质量”多解释一部分治疗师效应。

这个设计的精妙之处在于它能同时回答两个相反方向的假说。如果 FIS 增量显著,说明外气的确有某种独立于联盟/期望/安慰剂之外的”治疗师特异性”——这正是气功传统的”功夫在功外”(发功功力本身可测且有效);如果 FIS 增量不显著、全部被联盟/期望吃光,那外气的治疗师效应就只是”看起来更专业的治疗师更让人信”的人际现象,功力维度可以放弃。两种结局都收窄争论,没有第三种模棱两可的中间态。

样本量与统计把握度也值得预先想清楚。参考 2008 年 RCT 中两位治疗师即出现显著差异的先例(PMID 18654733),治疗师效应量可能不小,但治疗师层样本才是瓶颈——心理治疗文献中治疗师效应通常只解释结局方差的 5-8%(PMID 22922705),要稳定检出 FIS 的增量贡献,推荐 ≥10 名治疗师 × 每名 ≥10 名患者(总 ≥100 对)的分层设计,用多层模型(患者嵌套于治疗师)配合 ΔAIC 与置信区间报告,并预先注册效应量阈值,避免”显著但不稳定”的假阳性。若样本无法一次到位,可先做”联盟/期望控制后 FIS-疗效相关”的预注册观察性分析作为中间步骤,再升级为完整 RCT。

对易学桥接而言,这个设计把”发功”从不可言说的内证体验,推向与心理治疗 FIS 同等地位的”可训练、可评分、可检验增量”的行为变量。一旦通过,气功教学中的”带功状态””发功功力”就有了第一个现代操作化定义——不是疗效本身,而是”剥离一切人际与心理因素之后依然存在的治疗师层预测力”。

证据与局限

已验证: FIS 行为任务信效度及前瞻预测力(PMID 26594945、PMID 26344392);联盟-结局关联与治疗师层联盟变异(PMID 22922705);期望对安慰剂反应的放大效应(PMID 31302644);外气疗效依赖治疗师个体(PMID 18654733)。

部分验证: 联盟中介治疗师能力-结局路径(PMID 34210234 显示早期治疗师能力经联盟中介影响抑郁症状变化,但样本仅 50 对);高期望抹平药物-安慰剂差异(同一 SOPPRANO 试验的期望分层分析,需与主试验结果一并解读)。

待验证: 外气版 FIS 得分在控制联盟/期望/安慰剂后的独立增量(尚无任何直接数据);”发功表现”与”通用人际质量”的特异性分离(心理治疗领域也尚未在 FIS 上做完整的特异性对照);治疗师层样本量对增量检验把握度的影响。

相关阅读

证据等级: B(FIS 增量预测有 PubMed 直接实证,但外气场景的独立增量检验尚无直接数据,属跨领域类比 + 实验设计提案)
参考文献: PMID 18654733, PMID 26594945, PMID 26344392, PMID 22922705, PMID 31302644, PMID 32297945, PMID 34210234


免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。

Predictive Validity of Qi Manipulation Performance: Remaining Therapist Effect After Controlling for Alliance, Expectation, and Placebo in External Qi Therapy

Evidence level: B

Traditional discourse posits that “skill lies beyond the technique” — yet external qi research must address whether the therapist’s own “performance manifestation” can independently predict efficacy once all non-specific factors are stripped away: the alliance (patient trust in the practitioner), pre-existing expectation that the patient feels will improve, and pure psychological placebo.

Main Text

External qi research has faced a longstanding issue exposed two decades ago: a 2008 double-blind randomized controlled trial (RCT) involving 112 patients with knee osteoarthritis revealed starkly different outcomes between two therapists. Patients treated by Therapist 2 showed significantly greater pain improvement compared to the sham treatment control, whereas patients treated by Therapist 1 showed no difference from the control group (PMID 18654733). Researchers’ conclusion at the time was cautious: “External qi efficacy appears to depend on some quality of the therapist.” However, what exactly constitutes this “quality” has remained unmeasured for the subsequent twenty years. This article discusses whether transforming “performance manifestation” into a validated metric allows it to retain predictive power after eliminating various non-specific factors.

First, consider how similar issues are addressed in psychotherapy. Facilitative Interpersonal Skills (FIS), or Promoting Interpersonal Skills, constitute a standardized set of behavioral tasks: trained therapists watch a series of videos depicting “difficult patients” (scenarios involving suspicion, resistance, emotional collapse, etc.) and respond verbally as if conducting real therapy. Trained raters then score responses across eight dimensions (linguistic fluency, empathy, warmth and acceptance, alliance maintenance, repair of ruptures, etc.). A 2016 prospective study found that FIS scores completed by 44 trained therapists in early training significantly predicted symptom changes in their actual patients more than a year later—therapists with high FIS scores demonstrated significantly better outcomes in short-term treatment (≤8 sessions) compared to those with low FIS scores (PMID 26594945). Earlier RCTs also indicated that FIS-selected therapists with high interpersonal skills yielded superior alliance quality and treatment outcomes compared to low-skill counterparts, regardless of training status (PMID 26344392). In essence, psychotherapy has transformed “individual therapist differences” from esoteric concepts into measurable engineering parameters.

Crucially, the predictive validity of FIS must undergo “incremental validation” to be considered meaningful. Because alliance (patient evaluation of the therapeutic relationship) itself is a strong predictor of outcome—a meta-analysis of over 14,000 treatments confirms a stable alliance-outcome correlation (PMID 22922705 confirms therapist-level alliance variance impacts outcomes more than patient-level variance); expectation and placebo are recognized “non-specific factors,” where positive treatment expectations in chronic pain patients correlate with clinical pain relief as high as r=0.72 (PMID 31302644), and a 2020 RCT with 200 participants (SOPPRANO) showed that high-expectation patients exhibited placebo responses strong enough to completely erase the difference between active drug and placebo (NNT deteriorated from 3.2 to 29.6, PMID 32297945 represents the expectation-stratified analysis of the same trial). If FIS predictive validity is merely a byproduct of “high FIS therapists fostering better alliance and easier expectation cultivation,” it lacks independent value.

Applying this logic to the external qi scenario constructs a clean three-step validation design. First, measure “performance manifestation” scores for ≥10 therapists using an external qi version of the FIS task (adapted via situational videos and behavioral scoring; see sister article). Second, conduct a multi-therapist × multi-patient blinded RCT: include both therapists and patients in alliance questionnaires (patient version + therapist version), expectation questionnaires (pre-treatment expectations regarding “qi sensation/efficacy”), and a standard sham treatment control arm to manage placebo. Third, perform incremental validation using multilevel modeling: first fit a baseline model including “alliance + expectation + treatment arm” on efficacy, then add therapist-level FIS scores to assess whether Δ variance/Δ AIC is significant. If significant, conduct a “specificity” test—simultaneously include FIS scores and general interpersonal quality (e.g., self-reported social skills, general empathy scales) in the model to determine if “performance manifestation” explains an additional portion of the therapist effect beyond “general interpersonal quality.”

The sophistication of this design lies in its ability to simultaneously address two opposing hypotheses. If FIS increment is significant, it indicates external qi possesses a certain “therapist-specific” quality independent of alliance/expectation/placebo—precisely the “skill lies beyond the technique” concept from qigong tradition (where performance capability itself is measurable and effective). If FIS increment is not significant and fully absorbed by alliance/expectation, then the therapist effect in external qi is merely an interpersonal phenomenon where “appearing more professional therapists are simply believed more,” and the capability dimension can be discarded. Both outcomes narrow the debate, eliminating a third ambiguous middle ground.

Sample size and statistical power also require pre-meditation. Referencing the precedent from the 2008 RCT where significant differences emerged between just two therapists (PMID 18654733), the therapist effect size may be substantial; however, the therapist-level sample remains the bottleneck. In psychotherapy literature, therapist effects typically explain only 5-8% of outcome variance (PMID 22922705). To stably detect the incremental contribution of FIS, a stratified design is recommended with ≥10 therapists × ≥10 patients per therapist (total ≥100 pairs), utilizing multilevel modeling (patients nested within therapists) with Δ AIC and confidence interval reporting, along with pre-registered effect size thresholds to avoid false positives that are “significant but unstable.” If sample size cannot be achieved in a single study, a pre-registered observational analysis of “FIS-outcome correlation after controlling for alliance/expectation” can serve as an intermediate step before upgrading to a full RCT.

For bridging Yijing Studies, this design elevates “performance manifestation” from an ineffable introspective experience to a behavioral variable on par with psychotherapy FIS—one that is trainable, scoreable, and capable of incremental validation. Once validated, “performance state” and “performance capability” in qigong pedagogy acquire their first modern operational definitions—not as efficacy itself, but as “therapist-level predictive power persisting after stripping away all interpersonal and psychological factors.”

Evidence and Limitations

Verified: FIS behavioral task reliability, validity, and prospective predictive power (PMID 26594945, PMID 26344392); alliance-outcome correlation and therapist-level alliance variance (PMID 22922705); expectation’s amplification effect on placebo response (PMID 31302644); external qi efficacy dependence on therapist individuality (PMID 18654733).

Partially Verified: Alliance mediating the therapist capability-outcome pathway (PMID 34210234 shows early therapist ability affects depression symptom changes via alliance mediation, but sample size was only 50 pairs); high expectation erasing drug-placebo differences (expectation stratified analysis of the same SOPPRANO trial, requiring interpretation alongside main trial results).

To Be Verified: Independent incremental contribution of external qi version FIS scores after controlling for alliance/expectation/placebo (no direct data exists yet); specificity separation between “performance manifestation” and “general interpersonal quality” (even in psychotherapy, a complete specificity control has not yet been conducted within the FIS framework); impact of therapist-level sample size on power for incremental validation.

Related Reading

Evidence Level: B (FIS incremental prediction has direct PubMed empirical evidence, but independent incremental validation in the external qi scenario lacks direct data, representing cross-domain analogy + experimental design proposal)
References: PMID 18654733, PMID 26594945, PMID 26344392, PMID 22922705, PMID 31302644, PMID 32297945, PMID 34210234


Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *