Moral Dissonance Is Greater When Clinicians Act Contrary to Evidence, Experience, and Values Than When They Disappoint the Patient

Most musculoskeletal visits, tests, and treatments are optional. Specialists may experience moral dissonance when patient requests are difficult to reconcile with evidence, experience, and professional values. This dissonance may arise when clinicians feel torn between acquiescing to a request they view as unsupported and declining the request while worrying about disappointing the patient. Understanding the specific factors associated with greater moral dissonance, either in the form of acquiescence or disappointment, may inform efforts to improve both patient and clinician experiences when guiding patients toward effective health strategies.

(1) Is the severity of moral dissonance experienced by musculoskeletal specialists greater when acquiescing to patient requests that are contrary to evidence, experience, and values compared with disappointing the patient by declining a request? (2) What scenario factors are associated with the severity of moral dissonance?

We invited members of the Science of Variation Group (SOVG), an international collaborative of musculoskeletal surgeons who participate in monthly survey-based experiments addressing variation in care, to complete an online survey- and scenario-based experiment between April 12 and May 4, 2025. The SOVG is open to all musculoskeletal surgeons, but its active participants are mostly White academic surgeons (54%) in the US and Europe (26%), which is not representative of the average musculoskeletal surgeon. Our experiments are therefore designed to measure associations rather than to estimate rates or prevalence. The analysis depends on variation in how surgeons judge and respond to the same clinical scenarios and should be reproducible in any sample of surgeons with sufficient variation in opinions. Of approximately 200 SOVG members who typically participate in at least one survey annually, 68% (135) of surgeons completed at least one scenario. Participants reviewed 11 clinical scenarios with randomized elements, including patient age, patient gender, and combinations of clinical presentation and patient request. Participants were then randomly assigned to rate moral dissonance as either pressure to act contrary to evidence, experience, and values ("Imagine you give the patient what they want. To what degree do you feel like you are acting contrary to your values and best evidence?") or as a feeling of disappointing the patient ("To what degree do you feel like you're letting the patient down if you can't offer a diagnosis and treatment?"). Both items were rated on a scale from 0 ("I don't feel it at all") to 100 ("I feel it strongly"). Multilevel mixed-effects linear regression models were used to identify factors associated with levels of distress, accounting for surgeon-level nesting. Regression coefficients represent the estimated difference in moral dissonance on the 0-to-100 rating scale associated with each factor compared with the reference category. Positive coefficients indicate higher moral dissonance and negative coefficients indicate lower moral dissonance. For categorical variables, the coefficient represents the average difference in moral dissonance between that category and the reference category, adjusted for the other variables in the model.

Levels of moral dissonance were higher when rated as acting contrary to evidence, experience, and values (median [IQR] 70 [40 to 90]) than when rated as disappointing the patient (40 [20 to 70]); the group medians differed by 30 points on the 0-to-100 scale. In the multilevel mixed-effects regression model, disappointing the patient was associated with a 21-point lower moral dissonance score than acting contrary to evidence, experience, and values (regression coefficient -21 [95% confidence interval (CI) -27 to -15]). Relative to requests for electrodiagnostic testing for pain with typing, moral dissonance ratings were higher for requests for carpal tunnel release despite normal electrodiagnostic testing results (regression coefficient 6.7 [95% CI 0.9 to 13]) and surgical excision of a wrist ganglion for symptoms consistent with a wrist ganglion (regression coefficient 7.1 [95% CI 1.3 to 13]). These coefficients indicate approximately 7-point higher moral dissonance ratings on the 0-to-100 scale compared with the reference scenario. Moral dissonance ratings were lower than in the reference scenario for requests for proximal median nerve release for nonspecific forearm pain (regression coefficient -14 [95% CI -20 to -8.2]) and surgery for distal biceps tendinopathy (regression coefficient -14 [95% CI -20 to -7.9]), indicating approximately 14-point lower ratings than the reference scenario.

The finding that moral dissonance was greater when rated as acquiescing to care that conflicts with evidence, experience, and values than when rated as disappointment suggests that the source of distress may be the surgeon's sense of responsibility for potentially reinforcing unhelpful beliefs, exposing patients to low-value care, or contributing to overuse. The variation by clinical presentation and patient request suggests that distress may be greatest when requested care is discretionary but still clinically plausible, making it harder to decline.

These findings suggest that surgeons may reduce moral dissonance by strategizing and practicing specific communication steps when patients request discretionary but plausible care, such as curiosity about the concern and goal driving the request, validation of that concern, and introducing options that might better achieve the patient's goals while introducing less potential harm and false hope.