Peer Pressure, Role modeling, and the Social Dynamics of Medical Training: What Unconscious Factors Influence the Hand Hygiene Behavior of Medicine Teams?

Background: The process of becoming a competent physician is complex, involving academic, experiential and social learning.  Residents and medical students are educated about the importance of hand hygiene (HH) but compliance remains low. The impact of attending physicians in role modeling appropriate professional behaviors including HH is incompletely understood 

Objective: Does the attending’s compliance with HH influence the behavior of residents, medical and pharmacy students, and vice versa?

Methods: A research assistant, already embedded into the internal medicine teams at a 750 bed academic hospital under a different pretense, was recruited to secretly observe HH compliance during rounds. Teams included 1 attending physician, 1 senior resident, 2 interns, 1 medical and 1 pharmacy student. All subjects were completely unaware that HH was being observed. Multiple different teams were observed daily over 2 months. The observer covertly noted the order each team member entered and left the patient encounter and whether they performed HH.

Results: 103 patient encounters, resulting in 437 HH opportunities prior to contact and 489 after contact were observed.  Overall compliance was 47% prior to and 68% after contact. Compliance by training level ranged from 30-70% prior to and 48-93% after contact (p<0.001) and was highest among medical and pharmacy students. The order in which team members entered a room did not affect likelihood of HH (p=0.184).  However, if the first person entering the room performed HH, then 72% of others also performed HH, whereas first person failure of HH resulted in only 38% of others performing HH (p<0.001).  If the attending performed HH entering, then 65% of others did, but attending failure resulted in only 35% of others performing HH (p<0.001).  Performance of HH by the first person exiting the room had no impact on likelihood of others performing HH (70% vs. 61%, p=0.28).  However, if the attending performed HH upon exiting then 70% of others also did, vs. 43% if the attending failed HH upon exiting (p=0.006).

Conclusions: There was a strong follower effect entering patient encounters, but not leaving them.  If the first person entering a room (regardless of their training level) performs HH, then others are more likely to perform it too.  This implies peer pressure may be playing a role in HH compliance.  Similarly, there was a strong attending effect both entering and exiting rooms, regardless of whether the attending went in first.  If the attending performs HH, then others are more likely to do so too, implying that role modeling by attending physicians impacts the unconscious behavior of learners.  HH compliance was greater exiting patient encounters than entering them, implying that self protection may be a stronger driver of behavior than patient protection.  These results may be helpful in designing HH programs targeting physicians.