Abstract
Background
Studies on how psychologic factors influence the placebo effect have shown conflicting results in an experimental setting. Pessimists are more likely to experience a nocebo effect (feel worse after an inert intervention), whereas other studies suggest that patients with more symptoms of depression or anxiety or greater neuroticism have a greater response to a placebo. This is important because treatment benefits are potentiated by placebo effects, and optimal utilization of this phenomenon may improve clinical outcomes.
Questions/purposes
(1) What psychologic factors are associated with a decrease in magnitude of limitations (Disabilities of the Arm, Shoulder and Hand [DASH] score) and pain intensity (visual analog scale [VAS] for pain) after placebo injections for the treatment of painful nontraumatic upper extremity conditions? (2) What psychologic factors are associated with achieving a minimum clinically important difference (MCID) in disability and pain intensity?
Methods
We performed a secondary analysis of data acquired in two prospective, double-blind, randomized controlled trials of patients with lateral elbow pain, trapeziometacarpal arthrosis, and de Quervain tendinopathy who received a single injection of dexamethasone and lidocaine or lidocaine alone (placebo). One hundred six patients were included between June 2003 and February 2008. Sixty-three patients (59%) received dexamethasone and lidocaine, and we analyzed the subset of 43 patients (41%) who received lidocaine alone. The primary outcomes of interest were the DASH questionnaire and the VAS for pain measured three times: when they received the injection, between 1 and 3 months after the injection, and between 5 and 8 months after the injection. Seven patients missed the first followup visit and 14 patients missed the second visit. Based on previous research, we chose a MCID threshold of 10 for the DASH and a threshold of 1.0 for the VAS score. In bivariate analysis, we accounted for sex, race, marital status, degree, education, work status, pretreatment pain, diagnosis, symptoms of depression (Center of Epidemiologic Studies–Depression Scale), coping strategies in response to nociception (Pain Catastrophizing Scale), and personality traits (measured with the Multidimensional Health Locus of Control scale and the Eysenck Personality Questionnaire-Revised score). Variables with p values < 0.10 in bivariate analysis were included in the multivariable regression models. An a priori power analysis showed that a sample of 43 participants provides 80% statistical power, with α set at 0.05, for a regression with five predictors if the depression score would account for 15% or more of the variability in pain score. We used multiple imputations (imputations = 50) for a total of 66 (8.5%) missing or incomplete questionnaires.
Results
In the final multivariable models, no psychologic factors were associated with a change in DASH score between injection and followup, and no factors were associated with greater decrease in pain intensity. After injection, no psychologic factors were independently associated with achieving a MCID in the DASH and VAS.
Conclusions
Our study confirms that patient factors are less important mediators of the placebo effect than clinician factors. In other words, clinician warmth and competence can help diminish symptoms and limitations of people in various states of mind, even when using inert or ineffective treatments.
Level of Evidence
Level II, therapeutic study.
Introduction
The apparent effectiveness of a given treatment is determined by several factors. First, there may be improvement in symptoms resulting from external pharmacologic or structural changes to physiology (eg, penicillin for a strep throat). Second, the body naturally changes over time, and many conditions or symptoms resolve without treatment (they are self-limited). Patients may visit a doctor when symptoms are peaking, and symptoms reduce closer to an average symptom level without any intervention, creating a false sense of response to treatment—a phenomenon referred to as “regression to the mean.” Treatment can also activate our intrinsic physiology for well-being such as enkephalins and endorphins—a phenomenon referred to as the placebo effect. Evidence points to the warmth and competence of the clinician as key factors influencing the magnitude of placebo effects [5, 7, 11]. Consequently, an effective clinician-patient relationship increases the effectiveness of both effective and inert or ineffective (eg, sugar pill) medications.
One might expect that psychologic factors such as effective coping strategies and symptoms of depression would alter the magnitude of the placebo effect, because there is mounting evidence that these psychologic factors have important associations with symptom intensity and magnitude of limitations [6, 10, 12, 23, 25] and may affect treatment decisions [4]. The same may be argued for certain personality traits, which may affect patient trust in a treatment or clinician. In a survey of hand surgeons, many participants indicated that they believed the placebo effect reflects the strong influence of psychologic factors on illness [18]. However, studies on this topic show conflicting results. For instance, one study showed that in an experimental setting, pessimists were more likely to experience a negative placebo effect (they felt worse after receiving an inert intervention), also known as the nocebo effect [9]. In contrast, in another study, patients with lower back pain who had more symptoms of depression or anxiety and higher levels of neuroticism had a greater response to placebo [26]. In a trial of patients with persistent upper arm pain, expectations and symptoms of depression and anxiety were not associated with a response to a placebo [2]. Because of these conflicting results, it is unclear whether the placebo effect of treatment is mediated by psychologic factors.
Therefore, we performed a secondary analysis of data from two placebo injection controlled randomized trials to ask: (1) What psychologic factors are associated with decreases in disability (Disabilities of the Arm, Shoulder and Hand [DASH] score) and pain intensity (visual analog scale [VAS] for pain) after placebo injections for the treatment of painful nontraumatic upper extremity conditions? (2) What psychologic factors are associated with achieving a minimum clinically important difference (MCID) in disability and pain intensity?
Patients and Methods
In this institutional review board-approved study, we performed a secondary analysis of data acquired in two prospective, double-blind, randomized controlled trials comparing the effect of a single injection of dexamethasone and lidocaine versus a single injection of lidocaine (placebo) on upper extremity disability in patients with lateral elbow pain, trapeziometacarpal arthrosis, and de Quervain tendinopathy [1, 15, 17]. Between June 2003 and February 2008, 106 patients were enrolled. Forty-three patients (41%) received a placebo injection, and 63 patients (59%) received a corticosteroid injection because in the later months of the trial, randomization was four to one in favor of a corticosteroid injection to induce participation (Fig. 1). We analyzed the 43 patients who received a placebo injection (Table 1). Thirty-three of 43 patients (77%) had lateral elbow pain, six (14%) had de Quervain tendinopathy, and four (9.3%) had trapeziometacarpal arthrosis. A research assistant called patients to remind them to return for a second appointment between 1 and 3 months and for a third appointment between 5 and 8 months after their initial appointment to complete additional questionnaires. Seven patients missed the first followup visit and 14 patients missed the second visit, leaving 36 and 29 patients for analysis, respectively.
Fig. 1.
This Consolidated Standards of Reporting Trials (CONSORT) flow diagram displays the progress of all participants who received placebo injection through the phases of the randomized controlled trials.
Table 1.
Patient demographics and questionnaire scores
Arm-specific limitations were measured with the DASH questionnaire (n = 38) and the VAS for pain intensity (n = 42). Additionally, the general Multidimensional Health Locus of Control (MHLC) scale (n = 38) was administered to measure patients’ health-related beliefs. The subscales of the MHLC were the Internal Health Locus of Control (IHLC), the Powerful others Health Locus of Control (PHLC), and the Chance of Health Locus of Control (CHLC). A measure of personality traits, the Eysenck Personality Questionnaire (EPQ-R; n = 37), was administered until the lateral elbow pain enrollment was complete and then was discontinued. This questionnaire measured the patient’s degree of extraversion, neuroticism, dishonesty, and psychoticism. After the first 28 patients, the protocol was amended to include the Center of Epidemiologic Studies–Depression Scale (CES-D) (n = 26) and the Pain Catastrophizing Scale (PCS) (n = 26). Seventeen CES-D and PCS questionnaires were missing in the group that received a placebo injection. We did not account for these missing data in bivariate analysis, but intended to use multiple imputations [22, 27] in all multivariable models.
In this study, we accounted for psychologic factors previously identified as mediators of pain intensity and magnitude of limitations measured with patient-reported outcomes. Psychologic factors account for a greater percentage of variation in outcomes than differences in objectifiable pathology: greater symptoms of depression have been associated with greater symptoms and limitations after surgery; and less effective coping strategies in response to nociception (measured with the PCS) are correlated with pain interference (a greater tendency to limit activity owing to pain) [13]. The MHLC assesses beliefs about control of one’s health status and illness, which is known to mediate recovery from injury and surgery [3, 24]. The EPQ-R is a scale that is used by psychologists and psychiatrists that measures four major dimensions of the personality: extraversion, neuroticism, dishonesty, and psychoticism. Extroverted people tend to be more sociable and outgoing and might be more likely to have a larger social support system and more likely to seek help for their health problems. People with greater neuroticism tend to have more worries and anxiety, and this trait is associated with more pain for a given pathology. Psychoticism measures the tendency of people to develop psychotic disorders; and people with more symptoms of this trait tend to be more aggressive and less empathetic. This trait is correlated with greater negative emotion associated with pain.
Statistical Analysis
To identify factors associated with a change in DASH and VAS scores as well as a MCID in the DASH and the VAS [19, 21], we performed bivariate analyses using chi-square and one-way analysis of variance tests for categorical variables, and we conducted Pearson correlation and Student’s t-tests for all continuous variables. Subsequently, variables with p values < 0.10 were included in the multivariable linear regression model. We subtracted the DASH and VAS scores at baseline from the first followup and the score at baseline from the second followup. Therefore, a positive number indicates an increase in these respective scores, whereas a negative number indicates a decrease. The MCID is the smallest incremental change in treatment outcome that an individual patient would classify as “important” and that would warrant a change in management [20, 28]. Based on previous research, we chose an MCID threshold of 10 for the DASH [21] and a threshold of 1.0 for the VAS score [19]. We accounted for the following variables: sex, race, marital status, degree, education, work status, pretreatment pain, diagnosis, MHLC, CES-D, PCS, and EPQ-R score. Regression coefficients, standard errors, 95% confidence intervals, and p values are reported. All p values < 0.05 were deemed statistically significant.
An a priori power analysis showed that a sample of 43 participants provides 80% statistical power, with α set at 0.05, for a regression with five predictors if depression score would account for 15% or more of the variability in pain intensity, and our complete model would account for 25% of the overall variability. A post hoc power analysis confirmed that the data met the assumptions for the a priori calculation, yielding 94% statistical power to detect a 0.97-point difference in pain intensity (MCID = 1.0) with two predictor variables.
We used multiple imputations [22, 27] for the following missing or incomplete predictor variables (imputations = 50) in all multivariable analyses: 14 PCS questionnaires, 14 CES-D questionnaires, six EPQ-R psychoticism questionnaires, seven DASH questionnaires at time 2, nine VAS questionnaires at time 2, and 16 VAS questionnaires at time 3. After multiple imputations, all 43 patients were available for analysis.
Results
After controlling for potential confounders such as pain catastrophizing, symptoms of depression, and demographic factors, no psychologic factors were associated with a change in DASH score (Table 2) between injection and followup visits, and no factors were associated with a greater decrease in pain intensity (VAS; Table 3). The bivariate analyses for factors associated with a change in DASH and VAS were included as an addendum (Table 1, Supplemental Digital Content 1, and Table 2, Supplemental Digital Content 2, respectively).
Table 2.
Multivariable analysis of factors associated with change in Disabilities of the Arm, Shoulder and Hand (DASH) score
Table 3.
Multivariable analysis of factors associated with change in visual analog scale (VAS) for pain score
No multivariable analysis was performed for factors associated with achieving a MCID in DASH and VAS, because only PCS was below the threshold of p < 0.10 in bivariate analysis between baseline and the first followup for the DASH score (Table 4); only female sex was associated with achieving a MCID in VAS to the first followup (Table 5); and only the work status “full-time or part-time” was associated with achieving a MCID in VAS to the second followup (Table 5).
Table 4.
Bivariate analysis of factors associated with a minimum clinically important difference (MCID) in the Disabilities of the Arm, Shoulder and Hand (DASH) score
Table 5.
Bivariate analysis of factors associated with a minimal clinically important difference (MCID) in VAS score
Discussion
Conflicting results have been found on the influence of psychologic factors on the placebo effect [2, 8, 9, 26]. This is important because treatment benefits are increased by placebo effects, and optimal utilization of this phenomenon can help patients become and remain healthy. This study sought to identify the factors associated with a change in the magnitude of limitations and pain intensity as well as achieving a MCID after placebo treatment for atraumatic painful upper extremity conditions. We found that no factors were associated with a greater decrease in pain intensity or upper extremity disability after placebo injection. No factors were independently associated with achieving a MCID in limitations or pain intensity after a placebo injection.
Readers should keep several limitations in mind when considering our work. First, this study represents secondary use of data from studies comparing placebo and corticosteroid injections, which introduces the risk of assessment bias that is inherent to a retrospective study design. We mitigated this risk by determining our research questions a priori, and we did not perform tertiary analyses. Additional prospective studies are needed to confirm our results, which may include several psychologic factors that have not been accounted for in this study (such as general mental health questionnaires, beliefs in effectiveness of conventional and alternative medicine, and the presence of other chronic painful comorbidities). Second, the MCID in pain and disability probably varies by diagnosis but could not be analyzed as such because we did not have a sufficient sample size to address each diagnosis separately. Third, because the population was smaller than the original studies and we only analyzed the subset of patients who received a placebo injection, we had to pool categories into “other,” making these findings less specific. Therefore, the “other’’ category is heterogeneous and is less likely to provide meaningful clinical implications. Fourth, changes in the trial protocol resulted in a moderate amount of missing data, meaning we had to use multiple imputation. This may lead to underestimation of correlations in the multivariable models if the imputed variables do not accurately describe these data, a risk we tried to reduce by using auxiliary variables in the imputation. However, none of the variables in these models were close to reaching statistical significance. Finally, data of this study may not be generalizable to all patients visiting orthopaedic clinics, because all patients who were enrolled were English-speaking adults, and patients with systemic inflammatory disease (eg, rheumatoid arthritis), prior elbow surgery, and prior treatment with corticosteroids were excluded.
After accounting for potential confounders in multivariable analysis, we found that no factors were associated with a change in pain intensity and magnitude of physical limitations in the first 6 months after placebo injection. This is consistent with a prior study that tested the influence of psychologic factors (such as depression and anxiety) on the placebo effect comparing the efficacy of a sham device with a sham pill on 119 patients with persistent distal arm pain that found no correlation between pain at the end of treatment and psychologic factors [2]. In another randomized, double-blind, placebo-controlled, crossover study of the anticonvulsant oxcarbazepine for patients with neuropathic pain, the variables anxiety, depression, neuroticism, and catastrophizing were not associated with a greater response to a placebo [16]. On the other hand, a previous study that combined pain, anxiety, depression, and neuroticism scores, and stratified patients into three psychopathology groups (low, moderate, and high) based on these scores, found that the group with the highest degree of psychologic symptomatology experienced the most pain relief from placebo treatment [26]. These conflicting findings may reflect methodological differences because our study analyzed psychologic factors on a continuum rather than creating categories. We discourage categorization of psychologic factors and personality traits because they reinforce a false mind/body dichotomy and reinforce the stigma and shame often associated with the mental aspects of health. As for higher order personality traits such as psychoticism and neuroticism, a previous study of 154 patients with trigger finger found no correlation between pain and psychoticism [12]. Another study of 66 patients subjected to the cold pressor test showed that higher order personality traits (such as psychoticism, neuroticism, negative/positive affectivity, extraversion) do not correlate as well with pain as lower order personality traits (pain catastrophizing and fear of pain) [14]. These lower order personality traits represent pain coping strategies and reflect cognitive fusion (thoughts are facts) and negative affectivity (a tendency toward negative interpretations and narratives).
No factors were independently associated with a MCID in magnitude of limitations and pain intensity, which is consistent with our finding that psychologic factors did not influence the decrease in pain and limitations after placebo injection on a continuous scale. If psychologic factors would be associated with a placebo response in a larger study design, this is unlikely to have a clinically important impact on pain and limitations. Our findings may indicate that psychologic factors have a limited impact on the placebo effect or that the effect of the placebo treatment did not alter the natural course of these painful illnesses. The conditions studied here are generally self-limited (with the exception of trapeziometacarpal arthrosis), and pain intensity fluctuates. Therefore, the observed decrease in pain intensity and disability may be the result of nonspecific effects rather than placebo effects.
Stress, distress, and less effective coping strategies affect symptom intensity and the magnitude of limitations, but they do not seem to influence the magnitude of the placebo effect. Current best evidence supports that clinicians can enhance placebo effects through conveyance of warmth and competence [5, 7, 11]. In other words, there is an opportunity for clinicians to bolster a more effective internal physiology through effective communication, irrespective of individual patient factors. Clinicians can enhance the health of patients and make treatment more effective by identifying and taking advantage of opportunities to improve mental health as well as by cultivating highly effective communication strategies to optimize placebo effects. Given that patients who stand to benefit more from improved mental health may tend to see less appeal in a comprehensive approach, effective communication strategies are paramount to both potentiating the placebo effect and addressing less effective coping strategies.
Acknowledgments
We thank A. Lindenhovius, M. Henket, D. J. S. Makarawung, S. J. E. Becker, B. P. Gilligan, S. Lozano-Calderon, J. B. Jupiter, and S. Bekkers for collecting the data and allowing us access to the data to perform this secondary analysis.
Footnotes
One of the authors (DR) received royalties from Tornier (formerly Wright Medical; Memphis, TN, USA) for elbow plates in the amount of less than USD 10,000 per year and from Skeletal Dynamics (Miami, FL, USA) for an internal joint stabilizer elbow in the amount of less than USD 10,000 per year. One of the authors certifies that he (DR) is a Deputy Editor for Hand and Wrist, Journal of Orthopaedic Trauma, and Clinical Orthopaedics and Related Research® and has received or may receive payments or benefits in the amount of USD 5000 per year. One of the authors certifies that he (DR) received honoraria from meetings of AO North America (Wayne, PA, USA), AO International (Davos, Switzerland), and various hospitals and universities. One of the authors certifies that he (TT) received payments in the amount of less than 10,000 USD per year from AO Trauma (Dubendorf, Switzerland), DePuy Synthes (West Chester, PA, USA), and PATIENT+ (The Hague, The Netherlands).
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research® editors and board members are on file with the publication and can be viewed on request.
Clinical Orthopaedics and Related Research® neither advocates nor endorses the use of any treatment, drug, or device. Readers are encouraged to always seek additional information, including FDA approval status, of any drug or device before clinical use.
Each author certifies that his institution approved the human protocol for this investigation and that all investigations were conducted in conformity with ethical principles of research.
References
- 1.Becker SJE, Guitton TG, Ring D. Predictors of missed research appointments in a randomized placebo-controlled trial. Arch Bone Jt Surg. 2014;2:192–198. [PMC free article] [PubMed] [Google Scholar]
- 2.Bertisch SM, Legedza ART, Phillips RS, Davis RB, Stason WB, Goldman RH, Kaptchuk TJ. The impact of psychological factors on placebo responses in a randomized controlled trial comparing sham device to dummy pill. J Eval Clin Pract. 2009;15:14–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Christino MA, Fleming BC, Machan JT, Shalvoy RM. Psychological factors associated with anterior cruciate ligament reconstruction recovery. Orthop J Sports Med. 2016;4:2325967116638341. 10.1177/2325967116638341. eCollection 2016 Mar. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Crijns TJ, Bernstein DN, Ring D, Gonzalez R, Wilbur D, Hammert WC. Factors associated with a discretionary upper-extremity surgery. J Hand Surg Am. 2018. Jun 13. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 5.Crum AJ, Phillips DJ, Goyer JP, Akinola M, Higgins ET. Transforming water: social influence moderates psychological, physiological, and functional response to a placebo product. PLoS One. 2016;11:e0167121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Das De S, Vranceanu A-M, Ring DC. Contribution of kinesophobia and catastrophic thinking to upper-extremity-specific disability. J Bone Joint Surg Am. 2013;95:76–81. [DOI] [PubMed] [Google Scholar]
- 7.Enck P, Bingel U, Schedlowski M, Rief W. The placebo response in medicine: minimize, maximize or personalize? Nat Rev Drug Discov. 2013;12:191–204. [DOI] [PubMed] [Google Scholar]
- 8.Ernst E, Resch KL. Concept of true and perceived placebo effects. BMJ. 1995;311:551–553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Geers AL, Helfer SG, Kosbab K, Weiland PE, Landry SJ. Reconsidering the role of personality in placebo effects: dispositional optimism, situational expectations, and the placebo response. J Psychosom Res. 2005;58:121–127. [DOI] [PubMed] [Google Scholar]
- 10.Hermanussen HH, Menendez ME, Chen NC, Ring D, Vranceanu A-M. Predictors of upper-extremity physical function in older adults. Arch Bone Jt Surg. 2016;4:359–365. [PMC free article] [PubMed] [Google Scholar]
- 11.Howe LC, Goyer JP, Crum AJ. Harnessing the placebo effect: exploring the influence of physician characteristics on placebo response. Health Psychol. 2017;36:1074–1082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kennedy SA, Vranceanu A-M, Nunez F, Ring D. Association between psychosocial factors and pain in patients with trigger finger. J Hand Microsurg. 2010;2:18–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kortlever JTP, Janssen SJ, van Berckel MMG, Ring D, Vranceanu AM. What is the most useful questionnaire for measurement of coping strategies in response to nociception? Clin Orthop Relat Res. 2015;473:3511–3518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lee JE, Watson D, Frey Law LA. Lower-order pain-related constructs are more predictive of cold pressor pain ratings than higher-order personality traits. J Pain. 2010;11:681–691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lindenhovius A, Henket M, Gilligan BP, Lozano-Calderon S, Jupiter JB, Ring D. Injection of dexamethasone versus placebo for lateral elbow pain: a prospective, double-blind, randomized clinical trial. J Hand Surg Am. 2008;33:909–919. [DOI] [PubMed] [Google Scholar]
- 16.Lund K, Demant DT, Vase L, Sindrup SH, Jensen TS, Finnerup NB. Placebo responses in patients with peripheral neuropathic pain. Scand J Pain. 2017;5:210. [Google Scholar]
- 17.Makarawung DJS, Becker SJE, Bekkers S, Ring D. Disability and pain after cortisone versus placebo injection for trapeziometacarpal arthrosis and de Quervain syndrome. Hand (NY). 2013;8:375–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Mallette P, Ring D. Attitudes of hand surgeons, hand surgery patients, and the general public regarding psychologic influences on illness. J Hand Surg Am. 2006;31:1362–1366. [DOI] [PubMed] [Google Scholar]
- 19.Salaffi F, Stancati A, Silvestri CA, Ciapetti A, Grassi W. Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain. 2004;8:283–291. [DOI] [PubMed] [Google Scholar]
- 20.Schünemann HJ, Guyatt GH. Commentary—goodbye M(C)ID! Hello MID, where do you come from? Health Serv Res. 2005;40:593–597. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sorensen AA, Howard D, Tan WH, Ketchersid J, Calfee RP. Minimal clinically important differences of 3 patient-rated outcomes instruments. J Hand Surg Am. 2013;38:641–649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM, Carpenter JR. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338:b2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Talaei-Khoei M, Fischerauer SF, Lee S-G, Ring D, Vranceanu A-M. Pain catastrophizing mediates the effect of psychological inflexibility on pain intensity and upper extremity physical function in patients with upper extremity illness. Pain Pract. 2017;17:129–140. [DOI] [PubMed] [Google Scholar]
- 24.Thomee P, Wahrborg P, Borjesson M, Thomee R, Eriksson BI, Karlsson J. Determinants of self-efficacy in the rehabilitation of patients with anterior cruciate ligament injury. J Rehabil Med. 2007;39:486–492. [DOI] [PubMed] [Google Scholar]
- 25.Vranceanu AM, Jupiter JB, Mudgal CS, Ring D. Predictors of pain intensity and disability after minor hand surgery. J Hand Surg Am. 2010;35:956–960. [DOI] [PubMed] [Google Scholar]
- 26.Wasan AD, Kaptchuk TJ, Davar G, Jamison RN. The association between psychopathology and placebo analgesia in patients with discogenic low back pain. Pain Med. 2006;7:217–228. [DOI] [PubMed] [Google Scholar]
- 27.Wood AM, Royston P, White IR. The estimation and use of predictions for the assessment of model performance using large samples with multiply imputed data. Biom J. 2015;57:614–632. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Wright A, Hannon J, Hegedus EJ, Kavchak AE. Clinimetrics corner: a closer look at the minimal clinically important difference (MCID). J Man Manip Ther. 2012;20:160–166. [DOI] [PMC free article] [PubMed] [Google Scholar]