Poor-quality data show music interventions may improve post-op anxiety and pain in adults

Reference

Kuhlmann AYR, de Rooij A, Kroese LF, van Dijk M, Hunink MGM, Jeekel J. Meta-analysis evaluating music interventions for anxiety and pain in surgery. Br J Surg 2018;105(7):773-783.

Synopsis

These authors systematically searched multiple databases and trial registries to identify randomized trials of the use of music to decrease anxiety and pain in adults undergoing invasive surgical procedures. Four authors independently determined which papers to include and asked a fifth author to resolve disagreements. The authors don't describe how they assessed each study's risk of bias but report that the included studies were of moderate to high risk of bias. Of the 92 small studies they identified (total number of patients = 7385), 81 provided enough data to conduct a meta-analysis. From the reported data, the authors imputed visual analog equivalent scores for anxiety and for pain. Because the studies assessed many different kinds of surgeries, types of patients, and nature of the music interventions (timing, frequency, and so forth) the authors used a random effects model, a more conservative approach, to estimating their treatment effects. The studies also varied as to the nature of the controls: 4 of the studies used a music therapist while the rest used recorded music; 8 used sham sounds; 6 provided head phones with no music. Forty-seven studies reported on anxiety and 42 reported on pain. Overall, the included studies demonstrated that patients exposed to music had less anxiety (21 mm on a 100-mm scale) and less pain (10 mm on a 100-mm scale). Generally 15 mm to 20 mm is considered the minimal clinically important difference. In addition to methodologic heterogeneity, the authors also identified statistical heterogeneity in the data. Finally, the authors report the potential for publication bias, especially in the studies that assessed anxiety. Since the existing data are based on studies that are at risk of bias and have a tendency toward publication bias that potentially negates the anxiety data, an appropriate conclusion would be that the jury is still out. Some might argue that this intervention is low cost and likely to cause no harm. This latter argument is to be countered with a recent report of singing and dancing in the operating room causing other mayhem (https://www.cnn.com/2018/05/25/health/dancing-doctor-malpractice-suits/index.html)!