To the Editor
We read with great interest the article by Mitrev et al concluding that higher anesthesiologist compassion scores were associated with a reduction in postoperative pain scores. The results presented in the study warrant further discussion due to limitations with the study design specific to the assessment of compassion and anxiety levels and lack of adjustment for variables affecting preoperative anxiety.
The 5-item compassion measure (CM) tool used in the study to assess the compassion of anesthesiologist may not accurately assess patients’ perceived anesthesiologist compassion. The original survey was validated for outpatient clinic visits, conducted via mail, and reliable only when conducted on a large scale. Although the authors have established the validity of this survey as a stand-alone distinct domain from the Clinician and Group Survey of the Consumer Assessment of Health care Providers and Systems, they failed to address the bias introduced by Hawthorne effect stemming from consenting and administering the survey before and after the preanesthesia visit. Administering the survey to a small group of patients will not address response homogeneity, which could be better managed by conducting the survey on a larger scale. In fact, the same author who validated the outpatient CM tool has also validated a separate 5-item, 2-part CM tool specifically for inpatient settings. A recent critical comparative review found the Sinclair Compassion Questionnaire to be the most reliable and valid measure of compassion compensating for the observed homogeneity associated with limited number of survey questions.
The study predominantly includes female patients (87.1%). Male and female sex hormones have been known to alter pain perception and affect total opioid consumption in a dose-dependent manner. Higher levels of testosterone show a protective effect against pain perception, whereas postmenopausal estrogen fluctuations are associated with a lower pain threshold. The linear mixed model may have minimized the bias due to disproportionate sample size but cannot address the effect of sex hormones on postoperative pain. This study does not discuss the preexisting use of psychiatric medications or preoperative anxiolytic medications before the surgery. Any anxiety-reducing measures or major depressive disorder can affect the postoperative pain scores and opioid consumption. Additionally, the effect of patient interactions with other providers and physicians on preoperative anxiety level was not assessed. The degree of compassion and competence of the surgeon, as perceived by the patient, could have a significant impact on SA levels, perhaps even a greater impact than that of an anesthesiologist given the relatively longer duration of familiarity with the surgeon. Since there are different combinations of anesthesiologists and surgeons providing care for each patient, it is not possible to isolate the effect of the interactions between the patients and these providers to SA levels. The original CM tool is not validated for interaction with multiple physicians.
Finally, the study does not provide details on the total number of unplanned admissions exceeding 23 hours, nor does it provide details on the reason for the hospital admission. Assuming these patients were admitted due to complications or delayed postoperative recovery, they would have different postoperative opioid requirements and pain scores within the first 24 hours compared to those who are discharged home the same day. Similarly, pain intensity can vary among different surgical procedures. Failure to make necessary adjustment in pain scores for unplanned admission and surgical procedures can produce unreliable results. The study measures State Anxiety (SA) using the STAI Y1 form which consists of 20 statements each scored from 1 to 4, yielding a total score range of 20 to 80. However, the median and quartile range for SA on postoperative day 0 (Table 2)1 falls outside of the mentioned range. It is not clear how these values were calculated. This calculation alone could have significantly affected the data analysis and the results presented in the article.We commend the authors’ efforts in conducting a challenging study exploring the complex relationship between compassion, patients’ psychology, and its effect on postoperative pain management. However, the study did not account for the confounding effect of anxiety and cannot establish it as a mediator in the compassion-postoperative pain pathway.