A Pilot Study of Patient Photographs and Handoff Communication

We conducted a pilot mixed-methods study comprised of a 2 group, 2 period randomized crossover trial with simulated handoff sessions and posttrial semi-structured interviews. Resident participants were in a training program in the mid-Atlantic region serving a 200-bed pediatric hospital. Our institution does not routinely incorporate patient photographs as identifiers in our electronic health record (EHR). Simulation was chosen for this study because it allowed for experimental control of patient illness severity and complexity, did not interrupt clinical workflows, and did not risk harm to patients. Figure 1 describes the study design and methods. Convenience sampling of residents was employed with the goal of recruiting the first 40 residents across all training years, representing 40% of the eligible pediatric residents. The study was advertised by e-mail and at noon conference sessions and interested participants contacted 1 of the research staff directly. The research protocol was approved by our institutional review board and all participants provided written consent. The study was conducted between October 2017 and June 2018.

Residents participated in 2 in person simulated handoff sessions that were audio-recorded. One session included stock photographs of fictional patient faces that were provided for each patient, along with written information; the other with written information alone. Handoffs were delivered in I-PASS format²  by 1 investigator (H.H.) acting as the giver (delivered handoff information) with residents in the receiver role (received handoff information). Residents were provided written handoff documents that mimicked ones used in our institution’s EHR.

Resident participants were randomized 1 to 1 (blocks of 4) to whether patient photographs accompanied either the first or second handoff simulation session (Fig 1). Photographs were included on hardcopies of handoff documents for relevant sessions as well as in the corresponding postsession online questionnaire. Each of the 2 sequentially conducted handoff sessions were composed of 8 fictional vignettes (16 cases total) and delivered in the same order to every participant. Half of participants were randomized to have photographs accompany the first 8 scenarios, whereas the second group was randomized to have photographs accompany the second 8 scenarios. For example, a participant that received photographs with the first handoff session was then “crossed-over” to receive no photographs with the second simulation session. Vignettes included within the 2, 8- handoff sessions (with and without photos) reflected similar levels of illness severity, medical complexity, and age distribution based on the review of 2 study investigators (H.H. and J.S.). Because the 2, 8-case handoff sessions were matched by age, the same 8 photographs were used to match with either the first group of 8 cases or the second group of 8 cases.

Immediately after each of the handoff sessions, residents completed a questionnaire, developed de novo by the research team via Qualtrics Online Survey Software (Supplemental Fig 2). The survey instrument was developed by the study authors and iterated through feedback from institutional medical education leaders to assess the intended primary outcomes. The resident then participated in the next handoff session followed by completing the questionnaire regarding their second handoff experience. Patient photographs were included in the questionnaire corresponding to the handoff session that was accompanied by photographs. Residents were prompted to answer 10 questions related to information recall and response to clinical scenarios (eg, “Who is the sickest patient?”) and 4 questions related to their perception of the handoff and familiarity with patients (eg, “How well do you feel you know the patients on this census?”). Printed handoff documents were removed while participants completed questionnaires so that answers reflected retention from the preceding session.

Information retention and efficiency of response to clinical scenarios were based on the percent of correct responses and cumulative response time (automatically captured in Qualtrics) for the 10 clinical questions asked after each of the 2 handoff simulations. The 4-items assessing participant perceptions of the preceding handoff each used a 5-item Likert scale (eg, “How likely would it be for you to get confused between patients on this census?”) with scores ranging from 1 (extremely likely) to 5 (extremely unlikely). See Supplemental Fig 2 for full list of Likert response options.

Demographic characteristics were compared between the 2 study arms using Mann-Whitney U and χ² or Fisher’s exact tests for continuous and categorical variables, respectively. Analyses followed the intention to treat principle. Mean differences in overall percent correct answers and cumulative response time for the 10-item post-simulation questionnaire were compared between handoffs with and without photographs using paired t tests. Analysis of variance (ANOVA) was used to estimate the photographs, period, and sequence effects for questionnaire answers and response times. For the 5-point item Likert questions assessing perception of the handoff session, Fisher’s exact test was used to compare responses after photograph sessions versus responses after nonphotograph sessions. A P value <.05 was considered statistically significant. All analyses were conducted in Stata 14.1 (StataCorp, College Station, TX).

Immediately after completion of the 2 simulated handoff sessions, 1 investigator (H.H.) conducted semi-structured interviews with individual participants to assess the benefits and challenges of incorporating patient photographs in handoffs. Interviews were audio-recorded and transcribed.

Domains assessed in the semi-structured interview included: (1) perceptions of benefits and challenges of photograph inclusion in handoffs; (2) perceived impact of photographs on retention of medical information, response to clinical scenarios, and familiarity with patients; (3) experiences with photographs in patients’ charts in real life clinical care; and (4) reflection about memory strategies during handoffs.

We used content analysis to identify patterns and themes in audio-recorded, transcribed interviews. All coding was conducted in Atlas.ti 8 (ATLAS.ti Scientific Software Development GmbH) by 2 investigators (H.H. and L.Z.), 1 of whom completed pediatric residency training at our institution and 1 of whom was a research assistant without clinical experience. Disagreement in coding was resolved by consensus.

A total of 40 pediatric residents participated; 20 were randomized to receive patient photographs in the first handoff session and 20 were randomized to receive patient photographs in the second handoff session. Demographic characteristics of participant groups are listed in Table 1. There were no statistically significant differences between groups by participant age, gender, or postgraduate year.

When comparing responses to post-simulation questionnaires immediately following the simulation (Table 2), there were no significant differences in the overall percentage of correct responses between handoffs with photographs to handoffs without the photos (79% vs 76%; mean difference = 4% [95% confidence interval (95% CI): −3% to 11%]; P = .259). Additionally, there were no significant differences in the overall cumulative response time to the 10 clinical questions (168 vs 160 seconds; mean difference = 7 second [95% CI: −6 to 21]; P = .267). There was no evidence of a photograph or sequence effect when comparing either outcome through ANOVA. When looking to time to response stratified by handoff session, both groups demonstrated slower performance in the second simulation (173 seconds vs 178 seconds for the group that received photographs with the first session and 142 vs 163 seconds for the group that received photographs with the second session).

Out of the 40 total participants, 27 (68%) felt that it was at least somewhat likely they would be confused between patients on the census after a handoff session without photographs, in comparison with only 12 (30%) participants after handoff sessions with photographs (P = .003). All 40 participants (100%) indicated that that they would feel at least somewhat comfortable caring for the patients after the photograph handoff sessions overnight, in comparison with 32 (80%) of participants after nonphotograph handoffs (P = .020). Participants also more frequently rated the photograph handoff sessions to be of high quality; 37 (93%) indicated the photograph handoff sessions where at least somewhat high quality versus 34 (85%) for the nonphotograph sessions (P = .007). There was no significant difference in how well participants felt they knew patients between the photograph and nonphotograph sessions when assessed immediately after the handoff session.

In the interviews following handoff sessions, all (40 of 40) participants identified benefits to using patient photos during handoffs, whereas 75% (30 of 40) of participants identified challenges (Table 3). Key benefit themes included assisting memory, enhancing humanism and engagement with patients, visualizing patient physical traits, and improving patient safety. Multiple participants commented that receiving a new list of patients can be overwhelming, and it can be hard to appreciate each person’s individuality. One stated, “It’s like they’re all just names on a sheet. It’s not until I see the patient that I feel like I can associate all of the things that we read to a specific person.” Another participant commented, “I think a lot of us are people-people. You’re more driven by a human connection. Definitely, it makes it a lot more real when you see the patient’s face as opposed to read different names of people you’ve never met.”

Potential challenges of including photographs that respondents identified included visual distraction (eg, could not focus on listening while looking at the photograph) and the potential for bias and discrimination (eg, because of race or disability). Others mentioned challenges that they might predict, for example, misleading photographs (eg, out of date or does not represent patient’s current clinical status) and privacy considerations (eg, that a photograph could be shared or distributed). Almost all participants (39 of 40, 98%) described having real life experiences related to incorporating patient photos into charts.

In this pilot study, we found that the inclusion of patient photographs in handoffs did not result in overall improved immediate information retention or efficiency of response to clinical scenarios compared with handoffs without photographs. Participants did, however, report higher levels of comfort with handoffs and felt they were less likely to confuse patients when photographs were used. In interviews, residents suggested that handoff photographs could be helpful for assisting memory, enhancing humanism, differentiating between patients, and improving patient safety. These potential benefits may explain why residents indicated higher perceptions of comfort and quality with photograph handoffs. Reported drawbacks included distracting or misleading handoff receivers, introducing bias, and potentially infringing upon patient privacy. This is the first pilot study to our knowledge to report the potential implications of incorporating patient photographs into handoffs.

Ultimately, we did not find any statistically significant differences in our primary outcomes (information retention and efficiency of response to clinical scenarios). This may be because of a true lack of effect or may have been related to an inadequately powered sample size given this is the first study of its kind. It may also be because of the measurement of these outcomes immediately after the handoff was completed, instead of sometime later (eg, 12 hours delay), which may be more representative of what would happen in a real-world shift. We also do not have validity evidence to indicate that a fast response time is an accurate proxy for retention of clinical information after a handoff session.

The nonstatistically significant differences for our pilot study's primary outcomes are supplemented with the nuanced perspectives of resident participants in the qualitative interview comments. The unintended consequences of duty hours regulations (eg, work compression and more frequent handoffs) have led to higher levels of stress and depersonalization, as well as lower work satisfaction among residents.^16–18  Residency programs must develop innovative methods of helping residents to thrive in this changing learning environment. We are encouraged by participant responses regarding humanism and engagement with patients and feel that photographs could be a relatively simple intervention for programs to promote meaningful resident-patient connections.

Although there is limited literature addressing patient photographs in EHRs, evidence suggests they do have use beyond their potential in handoffs. In response to increased feelings of personal detachment because of limited patient contact in radiology,^19,20  a recent study reported that radiologists include more thorough, descriptive reporting for studies with patient photographs than without.²¹  Further, other studies have shown that the use of photographs reduced the risk of erroneous order placement.^12,13 

Although this study was conducted before the coronavirus disease 2019 pandemic, we feel our findings may have more relevance in the pandemic era. The accelerated use of telemedicine prompted by the coronavirus disease 2019 pandemic will no doubt change the nature of provider-patient interactions. There are no clear national guidelines for whether and how photographs of patient faces should be incorporated into EHRs or handoffs. Some of our participant comments indicate there may be benefits in humanizing patients to providers when face-to-face contact is limited.

There are several limitations that should inform interpretation of the findings. Most notably, given that this was a small, single institution pilot study with a de novo survey instrument, results may not be generalizable to other programs or to real life clinical practice. Additionally, participants may have been affected by their awareness of the simulated environment and fictional patient cases. We did not employ a washout period, which may partially explain the observed period effect in which both arms answered questions more slowly in the second period, irrespective of whether photographs were in the first or second handoff session. Unblinded to the patient photograph as intervention in this crossover design, participants may have approached how they answered questions differently in the second session. Despite this limitation, exposing each participant to both photograph-handoff and nonphotograph-handoff led to more fruitful qualitative interviews.

Despite these limitations, we feel that our results provide some early lessons learned that could help inform future intervention related to incorporating photographs into handoffs.

A future study in the real-world setting would have multiple benefits over this pilot simulation study. To assess information recall and efficiency of response to clinical scenarios, participants were not able to reference a paper handoff document. Although this allowed for scientific measurement of recall, it is not representative of real-life handoff situations. More importantly, a follow-up real-world study would allow for more robust assessment of the impact of photograph incorporation on clinical and quality outcomes, patient experience, and patient-provider relationship and communication measures.

In simulated handoff sessions in this small pilot study, the use of patient photographs was associated with resident perceptions of increased comfort and decreased likelihood of confusion between patients. Despite these perceptions, photographs were not associated with improved performance on post-simulation questionnaires when assessed immediately after the handoff. Participants articulated potential benefits to using photographs, but also some challenges and barriers. Future research should expand upon this pilot work and examine how photographs impact handoffs in real-world settings, including how to address logistical challenges (eg, how are photos taken and stored), whether photographs impact information retention over a longer time frame, and determine whether their utilization results in improved patient outcomes.

We thank members of the I-PASS study group, including Dr. Alisa Khan and the I-PASS Study Group Executive Council and members of the Johns Hopkins Institute for Excellence in Education led by Dr. Joseph Cofrancesco.