Back

The medical signature of musicians: A Phenome-wide association study using an Electronic Health Record database.

Niarchou, M.; Lin, G.; Lense, M. D.; Gordon, R. L.; Davis, L. K.

2020-09-26 epidemiology
10.1101/2020.08.14.20175109 medRxiv
Show abstract

BackgroundBeing a musician is a physically and psychologically demanding undertaking. A limited set of previous studies suggest musicians may be at higher risk for a unique set of medical and mental health problems. To address the limitations of previous studies, we examined trends in the medical care of musicians in Vanderbilts Electronic Health Record research database. MethodsWe mined clinical notes for a principled collection of keywords and regular expressions which indicated that the patient was a musician. We identified 9,803 "musicians" that we matched for sex, median age (across the medical record), ethnicity, race, length of record (in days) and number of visits with 49,015 controls. We fitted 1,263 logistic regression models (one for each phenotype classification). Results257 medical diagnoses were significantly more prevalent in musicians than controls after strict Bonferroni adjustment for the total number of phenotypes tested (p-value<7.6 x 10-6). Diseases of the larynx and the vocal cords (OR=2.32, p<2.84 x 10-191), and hearing loss (OR=1.36, p=5.98 x 10-97) were among the top associations. Anxiety disorder (OR=1.25, p=7.67 x 10-71), and Major depressive disorder (OR=1.21, p=4.88 x 10-36) were also strongly associated with musician status. However, 15 phenotypes were significantly more prevalent in non-musicians than musicians, including Coronary atherosclerosis (OR=0.91, p=1.77 x 10-10), and Ischemic Heart Disease (OR=0.92, p=1.65 x 10-09). ConclusionsAlthough being a musician was related to a large number of occupational health problems, we also identified protective effects of musicianship in which certain diagnoses were less common in musicians than in non-musicians, indicating that active musical engagement could have similar health benefits to athletic engagement. Key messagesO_LIIn a sample of 14,379 musician cases and 71,895 matched controls identified in an Electronic Health Record database, we conducted the first and largest study to date to identify the medical diagnoses associated with musician patients in an EHR context. C_LIO_LIWe replicated previous associations of musician status with medical problems, including respiratory and hearing problems as well as mental health problems. C_LIO_LIWe identified a number of protective effects by observing diagnoses such as cardiovascular disease, respiratory failure, and renal failure, that were less common in musicians than in controls, in line with literature indicating that active musical engagement has similar health benefits to athletic engagement. C_LI

Matching journals

The top 17 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.