Skip to main content

Emergency department physicians spend only 25% of their working time on direct patient care

Laila Maria Füchtbauer¹, Birgitte Nørgaard² & Christian Backer Mogensen³

1. jan. 2013
13 min.



Modern health care is increasingly complex with the growing demand of multidisciplinary care and the demand for high efficiency, but also for high quality and timely patient treatment. At the same time, numbers of malpractice claims are rising and the costs of malpractice trials are increasing [1], which leads to a growing need of detailed documentation and administrative tasks for physicians. This means less time for direct patient care and it is one of the major causes of physician discontent [2].

So far only a few studies have evaluated how physicians spend their time and no meta-analyses have been performed in this field. In a recent review [3], Tipping et al stated that there is a lack of standardization of methods and detailed definitions of work categories, which would enable comparison of different studies. In their review of the existing literature, they concluded that physicians spent an average of 23% of their time on direct patient care.

While the time spent on different tasks partly depends on the level of training of the physician [4], no significant difference in time spent during day versus night shifts was reported [5]. To our knowledge, no direct comparison has ever been made of the time distribution in different clinical functions, i.e. admitting physicians versus physicians working in wards or in fast tracks.

The aim of this study was to investigate how physicians in the emergency department of a public non-university hospital in Denmark spent their time during day shifts. We were also interested in whether there was a difference in the distribution of time used for different tasks when working in the admission area compared with the emergency room.


The study was conducted at the Emergency Department (ED) of the community hospital in Kolding, Denmark. This hospital is a non-university hospital that trains physicians at various stages of their postgraduate career and frequently has medical students as interns. It has a total of 360 beds, 52 in the ED.

Structurally, the ED is made up of an admission area and an emergency room (ER). Patients are referred to the admission area from the primary healthcare sector or other hospitals with medical, surgical or vascular surgical conditions. In the ER patients are self-referred without any pre-hospital visitation, primarily with orthopaedic complaints and trauma, but medical and other surgical emergencies are also seen. Patients who require admission are transferred from the admission area or ER after initial examination to the bed area of the ED for further investigation and treatment.

On average, 30 patients are admitted every day, 2% of whom require intensive care and 70% are discharged to home within 48 hours. In the ER, about 80 patients are seen every day by a physician and/or a nurse and are treated as out-patients.

The ED is staffed with 14 young and two to three senior physicians and supported by a range of more experienced physicians and specialists from the other departments of the hospital. In the Danish hospital system, the young physicians are primarily residents and are in charge of triage, admissions and the primary care. After admission, the details of the patient’s condition are communicated by the admitting physician to a more experienced physician, often a specialist and always a doctor with at least one year’s training within the specialty in question. The more experienced physician is obliged to come to the ED and review the patient and to initiate further investigations or treatment if necessary.

To evaluate the time physicians spent on different tasks during a dayshift, we chose a work sampling (WS) approach. Three physicians were observed every three minutes and their activity was registered in a predefined classification system (Table 1). Classification was tested in a pilot study and followed the categories suggested by Tipping et al with subcategories adjusted to our ED [3]. One exception was that communication with patients and relatives was considered part of direct patient care in this study. The observers generally attempted to minimize interruption of the physicians’ work by not entering the exam room if they were able to hear what task the physician was performing through the door. For this reason, the physicians were asked to leave the door ajar when possible during the course of this study.

Data collection

Three trained observers (two secretaries and the 1st author) observed nine residents during dayshifts from 20 to 24 February 2012. Dayshifts for admitting doctors were from 8 a.m. to 6 p.m., while the physician primarily responsible for minor orthopaedic traumas in the ER worked from 8 a.m. to 4 p.m. (fast track). The three residents on duty each day were observed simultaneously at an interval of three minutes. This interval had previously been determined by a pilot study conducted to determine the number of observations needed to obtain the desired accuracy [6]. We targeted an accuracy of 2σ ≤ 2% for the overall results and a 2σ ≤ 3% for the stratified results.

For each category, the number of observations and the relative frequency expressed in percentage and 95% confidence interval (95% CI) was calculated. Statistical differences in the distribution of observations between the admission group and the emergency room group were tested by χ2-test for the major categories.

Since it is well known that the distribution of work changes around shift changes [7], the physicians were observed for a full workday each day including any additional hours used to finish the work.

Trial registration: NCT01722721.


We observed nine physicians at three-minute intervals for a total of 137 hours during dayshifts of a normal week at the ED with 2,748 observation points distributed over six predefined main categories and 27 subcategories.

The overall results are summarized in Table 2 and Figure 1: physicians spent 24.8% of their time with direct patient contact, 5.8% with indirect patient care, 23.8% communicating with other staff, 31.4% documenting their work and 5.9% on transport. Personal time accounted for 4.8% and other activities for the remaining 3.5%.

The data were analyzed separately for the ER physicians (fast track) and the physicians of the admitting area (admission track). Even though there were differences in the subcategories, no significant differences were found between the observation frequencies in the primary category groups (p = 0.26, χ2-test).


In this study, we showed that physicians working with the primary contact with patients at a middle-sized public hospital in Denmark spent 25% of their time during day shifts with direct patient contact. An almost equal amount of time was spent on communicating with other staff, while indirect care took a lesser share of their time. Most time was spent on documentation (31%).

Our data confirm previous findings [3] which showed that physicians do not spend more than about a fourth of their time in direct contact with their patients. We stratified our observations to investigate whether the overall distribution of time differed with different tasks. In this respect, physicians working on the fast track have substantially different tasks as they rarely admit patients and often treat and discharge them immediately. We therefore hypothesized that time distribution would differ between the two groups. To prove this hypothesis, we calculated 95% CI for the main categories and for subcategories. This hypothesis was, however, rejected as the CI of all main categories overlapped, while there were significant differences between subcategories. χ2-tests performed on the main categories confirmed this finding.

The overlap of the main categories was independent of the task and specialty being performed. This could indicate that the structure of the IT and administrative systems which the physicians are working under may have more pronounced influence on the way they work than the work tasks per se. The fact that the distribution of time spent on the subcategories defined in Table 1 was significantly different between the two groups of physicians illustrates that the tasks being performed were, indeed, different: The physicians in the ER spent twice the time on procedures (3.7% versus 1.7% of their total work day) and only half the time on taking admission histories (6.4% versus 11.4%) compared with their colleagues who worked primarily with admitting patients, although all physicians in total spent the same amount of time on direct patient care.

Strengths and weaknesses

There are different methods to approach the question of how physicians distribute their time. We chose a WS method as described by Sittig in 1992 [8] as opposed to the equally used time motion studies described by Tippett in the 1930s [9]. This was because time-motion studies involve a 1:1 ratio between observer and physician, which has the disadvantage of potentially introducing stronger observer bias and also biasing the physicians work. On the other hand, very accurate timing figures are obtained by this technique. WS has the advantage that one observer can observe several physicians, in our case three at a time. It is less invasive and is therefore expected to introduce less bias. The number of observations in our study was sufficient to achieve the targeted accuracy calculated from our pilot study.

We observed three physicians at a time during day shifts of a normal week (Monday-Friday) at our hospital. We did not observe night shifts or weekend shifts, which may introduce a bias although a previous study has shown that work distribution does not change significantly during night time [5]. Another weakness of our study is that three different observers collected the data, which, in theory, could introduce inter-observer bias. However, all observers were acquainted with the IT systems and the facilities of the hospital and had received training prior to this study. To further minimize the bias, precise definitions of the observation categories were noted on the observation scheme.

It is a subject of debate how the optimal time distribution for physicians’ work should be. Physicians often themselves think that they spend too little time on direct patient care because of documentation requirements. However, it remains a question whether this is a new development. There is a need for further studies to investigate this and to establish how the time spent changes. To the best of our knowledge, all time studies evaluating physicians in hospitals have been conducted in Europe, Australia or the US. It would be interesting to see a time study from a part of the world with a different healthcare system with less dependency on IT systems or documentation requirements.

We were surprised at how much time physicians in our study spent on discussing their patients with other staff and on documentation. Since this was independent of whether the physician was working in the fast track or admission track, we speculate that this may be due to the organization of the Danish hospital system, where the same physician rarely visits a patient twice. Physicians may therefore feel the need to document and communicate every detail with other staff such as the nurses in the acute ward or the physician taking over charge. In a system with more continuity between treating staff and patient, less time would probably be spent on these categories. These thoughts lead us to hypothesize that the strongest determinants for time distribution may be the administrative and IT systems of the hospital, and that variations in time of day or weekday are of minor importance. In a time where treatment and diagnostics become more time-consuming and specialized and with a relative lack of physicians, authorities might consider rethinking their approach towards improving safety and efficacy as this may have a large impact on workflow and on the time available for direct patient care.


We found that physicians spent 25% of their time on direct patient care, including communicating with the patients. More than twice this time (55%) was spent on communicating with other staff and documentation of their work. No significant differences in these proportions were found between physicians admitting or treating in the fast track. We therefore hypothesize that the administrative system including IT solutions may have much influence on how physicians spend their time, which indicates that this is a vital area on which to focus when implementing measures to improve the efficiency of hospital medicine.

The lack of physicians with advanced experience in many industrialized countries stresses the need for solutions that simplify documentation and allow more time for direct patient care.

Correspondence: Laila Maria Füchtbauer, Medicinsk Afdeling, Kolding Hospital, 6000 Kolding, Denmark. E-mail:

Accepted: 24 October 2012

Conflicts of interest:Disclosure forms provided by the authors are available with the full text of this article at




  1. Carroll AE, Parikh PD, Buddenbaum JL. The impact of defense expenses in medical malpractice claims. J Law Med Ethics 2012;40;135-42.

  2. Mechanic D. Physician discontent: challenges and opportunities. JAMA 2003e:290;941-6.

  3. Tipping MD, Forth VE, Magill DB et al. Systematic review of time studies evaluating physicians in the hospital setting. J Hosp Med 2010;5;353-9.

  4. Westbrook JI, Ampt A, Kearney L et al. All in a day’s work: an observational study to quantify how and with whom doctors on hospital wards spend their time. Med J Aust 2008e:188;506-9.

  5. Parenti C, Lurie N. Are things different in the light of day? A time study of internal medicine house staff days. Am J Med 1993:94;654-8.

  6. Sittig DF. Work-sampling: a statistical approach to evaluation of the effect of computers on work patterns in healthcare. Methods Inf Med 1993:32;167-74.

  7. Kim CS, Lovejoy W, Paulsen M et al. Hospitalist time usage and cyclicality: opportunities to improve efficiency. J Hosp Med 2010 5;329-34.

  8. Sittig DF. Work-sampling: a statistical approach to evaluation of the effect of computers on work patterns in the healthcare industry. Proc Annu Symp Comput Appl Med Care 1992;537-41.

  9. Tippett LHC. Statistical methods in textile research. Uses of the binominal and Poisson distributions: A snap reading method of making time studies of machines and operatives in factory surveys. J Textile Inst Trans 1935:26;51-5.