Allocation of Internal Medicine Resident Time in a Swiss Hospital: A Time and Motion Study of Day and Evening Shifts

The practice of hospital medicine is constantly evolving with the increasing complexity of patients (14). During training, residents in internal medicine are educated to manage this complexity and to summarize a wide array of medical information. The structure of a resident's workday has also changed dramatically in recent decades, with limitations on hours worked per week, wide implementation of electronic medical records (EMRs), and a growing volume of clinical data and administrative tasks (5). Many new scheduling paradigms have been proposed and implemented to address duty hours; these include the separation of inpatient and outpatient rotations (mandatory clinics) (6) and the use of advanced practitioners (nurse practitioners and physician assistants) (7). Information technologies have increasingly been used in health care because they allow better sharing and availability of medical data. However, negative effects of EMRs have also been described, as they increase the time physicians spend performing administrative tasks and writing notes (8). Similarly, Alkureishi (9), Ratanawongsa (10), and Asaro (11) and their respective colleagues reported that EMRs reduce communication between patients and physicians and the time they spend together. Recently, Sinsky and associates (12) described the allocation of physicians' time in ambulatory practice and confirmed that they spend up to 50% of their time using EMRs. In hospital practice, only 9% to 22% of residents' time is spent with patients, and more than half is dedicated to activities indirectly related to patients (1317). This trend is worrying because less time spent with patients decreases physician satisfaction (18) and patient education and health promotion (19) and increases inappropriate prescribing and medical malpractice (20). Although information technologies are improving, EMRs have thus far failed to fulfill their promise in hospital practice (21). Few time and motion studies have focused on how computer use affects residents' time allocation. We therefore aimed to objectively assess the type and duration of activities performed by internal medicine residents throughout the day. The primary focus was to estimate the time spent with patients and using a computer. The secondary focus was to identify individual factors influencing the residents' allocation of time to different activities and contexts. Methods Study Design, Setting, and Participants We conducted an observational study between May and July 2015 in the Department of Internal Medicine of Lausanne University Hospital, 1 of 5 teaching hospitals in Switzerland, with more than 1400 beds and 47300 patients hospitalized in 2015 (www.chuv.ch). The hospital implemented its current EMR (Soarian, Cerner Corporation) in 2009; it compiles medical documentation, laboratory results, radiographs and providers' views, an electronic prescription system, and scanned medical archives and is available on desktop and laptop computers. The Department of Internal Medicine receives approximately 6200 patients per year and has 203 beds organized in 8 wards. Each ward is staffed with 1 senior physician and 1 chief resident supervising 3 residents. The minimum postgraduate training for internal medicine is planned over 5 years (3 years of basic training, mostly in an internal medicine department, and 2 additional years more freely organized in specialties). At least 6 months have to be certified in an ambulatory practice. Chief residents usually have at least 4 years of postgraduate training. Within each ward, a resident is responsible for 6 to 10 patients. There are day (8:00 a.m. to 6:00 p.m.), evening (4:30 p.m. to 11:30 p.m.), and night (10:30 p.m. to 8:30 a.m.) shifts. Scheduling of the residents' daily activities is similar to that in U.S. hospitals (Figure 1), with daily patient rounds, supervision, training, and new patient admissions. The staff is reduced to 2 residents and 1 chief resident (for 203 beds) during evening and night shifts. Evening shifts mostly involve late patient admissions, unstable patient care, and emergency situations. In this study, only day and evening shifts were considered; night shifts were not considered because activities are mostly limited to emergency situations. Figure 1. Official schedule of day and evening shifts in the Department of Internal Medicine at Lausanne University Hospital. During evening shifts, staff is reduced to 2 residents in charge of all wards. All residents working in internal medicine inpatient wardswhich do not cover any critical care or specialty unitduring the study period were eligible for inclusion. There were no exclusion criteria. The Human Research Ethics Committee of Canton de Vaud certified that the study was exempt from human subjects ethics review. All residents were informed of the study and provided written consent. No patient identifier or health information was recorded. Data Collection Procedures Data were collected by undergraduate medical students, who were extensively trained to collect data without interfering with residents' work. The training consisted of a dedicated e-learning program on how to categorize the residents' various activities, a teaching session that focused on the definition of activities and the use of recording devices, a 2-hour practice session based on a 1-hour video of residents engaging in typical medical activities, 8 hours of observation and recording of a resident's activity in the wards (the resulting data were not included in the study), and a final session to resolve any remaining issues. The reproducibility of the observers was assessed during the practice session; overall, observers recorded activities and contexts similarly (Appendix Table 1). Appendix Table 1. Results of the Reproducibility Study Between Observers (n= 6) Recording began when the resident arrived at his workplace and lasted until he left. Residents were randomly assigned to an observer. To decrease observation bias, observers had to avoid communicating with the residents and were allowed to ask for clarification about an activity or context only. To optimize observation accuracy, day shifts were sequentially covered by 2 observers, with handoff after the first 6 hours. Evening shifts were covered by only 1 observer because these shifts were shorter. Each activity was recorded in real time using a tablet (Appendix Figure). The application was designed by the investigators and developed with the Information Technology Department of the hospital. To promote similar studies, the source code is available as is on the GitHub Web site (https://github.com). The observer selected an activity, a context, or both. After the observer pressed a confirmation icon, the application automatically recorded the start time of the activity or context being performed. We defined 22 activities in 6 categories (Table 1), based on expertise and earlier studies (1317, 22). Our department is keen on quality development and is particularly interested in assessing residents' communication skills; hence, we created a category labeled communication to find out how much time was dedicated to news delivery (for example, bad news or therapeutic orientation) and meeting with family members. Other types of communication were collected in the category directly related to the patient, within such activities as admission, patient rounds, and patient discharge. Similarly, we created 4 contexts in which the activity was being performed: in the presence of 1 or more colleagues (any professional), in the presence of the patient, with a computer, and with a telephone. Context could change regardless of the activity being performed. Appendix Figure. Dedicated tablet application to record observations. The screen is split into 3 areas. 1) Observer sets the next activity and/or context. 2) After the observer presses the Confirm icon, preset activity becomes the current activity, exposed in detail in the green area. 3) The log allows editing of past activities. Observers and residents are identified by a number. Once the observation is finished, results are sent by secured e-mail to a designated investigator. Table 1. Distribution of Activities According to Shift in the Department of Internal Medicine of Lausanne University Hospital For each resident, sex, age, country of medical school graduation, postgraduate training (in months), and distance between home and the hospital were collected. The number of patients the resident was in charge of during the observed shifts was also collected. Statistical Analysis Based on similar studies (1317), a pragmatic sample size of 64, corresponding to 2 shifts per resident, was chosen. Statistical analyses were performed using Stata 14.0 (StataCorp). Descriptive results for residents' characteristics are presented as means and SDs for continuous data or number and percentage of participants for categorical data. Because residents could be assessed several times on the same or different shifts, we used a linear mixed model with clustering by resident to compute the means and corresponding 95% CIs of the times dedicated to each activity. We calculated the percentage of a resident's shift time devoted to a specific activity by dividing the time for that activity by the total shift duration. Role of the Funding Source The study was sponsored by the Information Technology Department and the Department of Internal Medicine of Lausanne University Hospital. The funding sources had no involvement in the study design; collection, analysis, or interpretation of the data; writing of the manuscript; or the decision to submit the manuscript for publication. Results Thirty-six residents were included; 23 were women, and the mean age was 29.4 years (SD, 2.5; range, 25.7 to 39.4 years). Thirteen (36%) residents held a foreign medical diploma. The average amount of postgraduate training in internal medicine was 29 months (SD, 11; range, 0 to 50 months), and the