Using a prospective study design, we assessed the influence of residents’ workloads arising from patient care on supervision and students’ activity profiles (medical versus non-medical activities). A medical activity was defined as an activity that could not be delegated to nurses and non-medical staff, while non-medical activities do not necessarily have to be performed by a physician. While the non-medical activities are also important and sometimes difficult and demanding, they can be performed by staff other than physicians and can thus also be learned and supervised by staff other than physicians. Our study concentrated mainly on the activities that are usually exclusively performed by physicians and thus can only be learned from and supervised by physicians.
Final-year students recorded their daily activities in a diary. The recorded activity profile and students’ supervision by the ward residents on each day were subsequently matched with a corresponding ward-specific workload index calculated from the electronic hospital information system.
In our hospital the wards are run by one to three residents with one attending doctor supervising the ward round once or twice a week and discussing problems on the ward once a day. Usually, training and supervision of final year student is almost exclusively performed by the residents. Every student is obliged to attend a 16-week internship in Internal Medicine, Surgery and an elective.
On the first day of their internship in Internal Medicine, 54 final-year students were asked to complete protocols of their daily activities for a period of three weeks (15 working days) using a diary (see Table ). Students from each medical ward at our University Hospital participated. Diaries were completed between March and December 2009. We sent the students weekly reminders to fill in the diary per email. For completed diaries the students Workload Index received a voucher for a bookshop.
Informal interviews with the residents on ward (n
18) revealed that residents considered their workload to primarily correlate with the number of patients seen, the complexity of their comorbidities, and the numbers of admissions and discharges per day. For each day recorded in the diary, we calculated a workload index (Figure ) based on our electronic hospital information system. The index comprised the following elements:
n: The number of patients on the respective ward on the day in question.
: A correction factor for the patient comorbidity complexity level (PCCL). The PCCL modifies reimbursement for each patient in the national DRG (Diagnosis-Related Group) system according to the individual comorbidities and clinical complications arising during the corresponding hospital stay. After students diaries had been handed back, we asked residents on the wards (n
18) to state how much more care a patient with a PCCL factor of 4 (maximum) requires compared to a patient with a PCCL factor of 0 (minimum, no relevant comorbidities). The mean result was a factor of 4.8
SD). We therefore applied a correction factor (CFPCCL
) of 1 for PCCL 0, 1.95 for PCCL 1, 2.9 for PCCL 2, 3.85 for PCCL 3, and 4.8 for PCCL 4.
CFAd/Dis: A correction factor for admission or discharge of the corresponding patient. According to the mean estimate of interviewed ward residents, admitting a patient is approximately 4.1 (± .9) times more time consuming than simply caring for the patient on ward. Discharges were estimated to be approximately 2.7 (± 1.1) times more time consuming. Therefore the correction factor CF was set to 4.1 for a patient admitted and 2.7 for a patient discharged from the ward at the corresponding day.
The overall resulting workload was divided by the number of residents present on the ward on each day examined (see Figure ).
Workload values showed different ranges (11 to 109 points) on each of the wards under examination. For each individual ward, we therefore categorized each day according to the tertile distribution of values into high, average, or low workload.
Student supervision and activity survey
Students were asked to complete diaries for a three-week period and to reflect on each day whether they had
(1) only watched,
(2) assisted the ward resident,
(3) performed the activity themselves under supervision of the ward resident, or
(4) performed the activity without supervision. (see Table ).
The diary was constructed after an in-depth discussion with former final-year-students according to the tasks frequently performed on our medical wards. At the end there was space to fill in additional activities.
We regarded the following activities as medical (at our hospital not delegable to nurses or non-medical staff): arterial puncture, informative/explanatory consultations, documentation in patient files, ECG interpretation, prescriptions, presenting patients on ward rounds, punctures (joint, pleura, paracentesis, etc.), ward rounds, and writing discharge letters. The administration of infusions, arranging/cancelling patient appointments, taking blood samples, enquiring about test results, running errands, the administration of intravenous catheters, sorting/copying files, and patient transportation were all considered to represent non-medical activities since in our hospital they can be delegated to nurses and other staff. This distinction (medical, non-medical) was made since we wanted to assess the influence of the residents’ workload on students’ medical tasks supervision. Activities that can be delegated to nurses or other staff can be learned from and supervised by the staff in question.
History taking and physical examinations were analyzed separately, since these tasks are usually seen as core activities for medical students and thus the probability that the students practice these activities was considered to be rather high.
Every medical activity that was noted as (3) (performed under supervision) was counted as supervised activity. The students were asked to count constructive feedback as supervision, delivered either during or after the activity. We then recorded, how many medical and non-medical activities were reported as actively performed [(3) or (4)] and counted them as actively performed medical versus non-medical activities, respectively.
In addition, the students were asked to judge the motivation for teaching of the ward.
The study protocol was approved by the local ethics committee. Since it was not possible to ensure complete anonymity of the diaries (we needed the ward and the time period of diary completion for calculation of the workload index, see above), we were obliged to ensure that diaries were handed in on a strictly voluntary basis.
A linear mixed model was used to evaluate the influence of time and workload on supervision or the amount of medical activities per day and the influence of time and supervision on the amount of medical activities per day, respectively. Due to the fact that workload and supervision varied each day, these were modeled separately as time-varying factors. Days 1–8 and Days 9–15 were analyzed individually, since the number of medical activities performed by students was only found to increase during the first eight days, with levels remaining constant across the rest of recorded ward time (see the Results section). As the residuals violated the assumption of a normal distribution, results were recalculated using the sandwich estimator [21
]. A comparison of the results demonstrated the robustness of the applied model against deviation from a normal distribution.
The significance level was corrected for three different tests according to Bonferroni. Therefore a p-value <0.017 was regarded as being significant. The software used was JMP 9.1.3 (SAS Inc, Cary, NC, USA).