The Time Needed for Clinical Documentation versus Direct Patient Care

Page created by Warren Tucker
 
CONTINUE READING
84   © Schattauer 2009                                                                                                                 Original Articles

     The Time Needed for Clinical
     Documentation versus Direct
     Patient Care
     A Work-sampling Analysis of Physicians’ Activities
     E. Ammenwerth; H.-P. Spötl
     Institute for Health Information Systems, UMIT – University for Health Sciences, Medical
     Informatics, and Technology Tyrol, Hall in Tyrol, Austria

                                                                                                         (physicians, nurses, etc.), clinical areas (radi-
      Keywords                                        ing physicians on two internal medicine
                                                                                                         ology, surgery, etc.), and health care organi-
      Documentation, time and motion studies,         wards of a 200-bed hospital in Austria. A
                                                                                                         zations (primary care, hospitals, nursing
      workload, physicians, medical record systems    37-item classification system was applied to
                                                                                                         homes, etc.), producing a high demand for
                                                      categorize tasks into five categories (direct
      Summary                                                                                            the documentation and communication of
                                                      patient care, communication, clinical docu-
      Objectives: Health care professionals seem                                                         patient-related data. This is aggravated by
                                                      mentation, administrative documentation,
      to be confronted with an increasing need for                                                       rising economic pressure, decreasing lengths
                                                      other).
      high-quality, timely, patient-oriented docu-                                                       of stay [2], and legal regulations all requiring
                                                      Results: From the 5555 observation points,
      mentation. However, a steady increase in                                                           additional documentation under great time
                                                      physicians spent 26.6% of their daily working
      documentation tasks has been shown to be                                                           pressure. Overall, health care professionals
                                                      time for documentation tasks, 27.5% for di-
      associated with increased time pressure and                                                        seem to be confronted with an increasing
                                                      rect patient care, 36.2% for communication
      low physician job satisfaction. Our objective                                                      need for high-quality, timely, patient-
                                                      tasks, and 9.7% for other tasks. The documen-
      was to examine the time physicians spend on                                                        oriented documentation.
                                                      tation that is typically seen as administrative
      clinical and administrative documentation                                                              An increase in administrative tasks has
                                                      takes only approx. 16% of the total documen-
      tasks. We analyzed the time needed for clini-                                                      been shown to be associated with increasing
                                                      tation time.
      cal and administrative documentation, and                                                          time pressure and low physician job satis-
                                                      Conclusions: Nearly as much time is being
      compared it to other tasks, such as direct                                                         faction [3], whereas adequate time for phy-
                                                      spent for documentation as is spent on direct
      patient care.                                                                                      sician-patient interaction seems to be associ-
                                                      patient care. Computer-based tools and, in
      Methods: During a 2-month period (De-                                                              ated with higher physician satisfaction [4].
                                                      some areas, documentation assistants may
      cember 2006 to January 2007) a trained in-                                                         Furthermore, in Austria, this rising need for
                                                      help to reduce the clinical and administrative
      vestigator completed 40 hours of 2-minute                                                          documentation is criticized by clinicians and
                                                      documentation efforts.
      work-sampling analysis from eight participat-                                                      regarded as a danger for the quality of patient
                                                                                                         care. The Austrian Medical Association states
                                                                                                         that clinicians spend too much time at the
                                                      Methods Inf Med 2009; 48: 84–91                    computer, and that the administrative and
      Correspondence to:
                                                      doi: 10.3414/ME0569                                documentation tasks (“paper chaos”) are
      Elske Ammenwerth
                                                      received: May 3, 2008                              taking too much time away from patient care
      Institute for Health Information Systems
                                                      accepted: August 15, 2008                          [5]. A recent survey of 2000 Austrian hospital
      UMIT – University for Health Sciences,
                                                      prepublished:                                      physicians showed decreasing job satisfaction
      Medical Informatics,
                                                                                                         compared to earlier years, with 82% of the
      and Technology Tyrol
                                                                                                         physicians stating that they feel stressed
      Eduard Wallnöfer Zentrum 1
                                                                                                         partly or heavily due to administration and
      6060 Hall in Tyrol
                                                                                                         documentation tasks [6] – this representing
      Austria
                                                                                                         the category with the highest stress level,
      E-mail: elske.ammenwerth@umit.at
                                                                                                         higher than, for example, stress from a high
                                                                                                         personal workload or from night shifts. In
                                                                                                         this survey, 53% of the physicians stated that,
     Introduction                                     other things, increase the quality and efficien-   in recent years, work has become more
                                                      cy of patient care, and to support health care     unpleasant, with increasing documentation
     Health care is increasingly influenced by the    professionals in their daily tasks. Modern         and administration efforts being the fre-
     use of modern information technologies (IT)      health care is characterized by the distribu-      quently mentioned reasons (52%) for this
     [1]. IT systems are introduced to, among         tion of tasks between professional groups          feeling [6].

     Methods Inf Med 1/2009
E. Ammenwerth; H.-P. Spötl: The Time Needed for Clinical Documentation versus Direct Patient Care                     85

    Researchers, therefore, have attempted to       Objectives of this Paper                           advantage is that work sampling just provides
quantify the actual time needed for docu-                                                              an estimate of the real-time distribution [18].
mentation, especially compared to the time          The objectives of this paper were, therefore, to   In addition, work sampling is only feasible
available for direct patient care. For example,     objectively measure the time physicians            when the clinicians remain in a defined area,
in an outpatient oncology clinic, Fontaine et       spend on clinical and administrative docu-         where they can easily be located by the ob-
al. found that U.S. physicians spend 29% of         mentation tasks, and to compare it with the        server. Both time-motion studies as well as
their time entering and retrieving informa-         time needed for other activities.                  work sampling have been conducted in clini-
tion from paper-based medical records, and                                                             cal areas for many years [19]. In the 23 studies
43% on direct patient care [7]. In another                                                             reviewed by Poissant et al., 58% used time
U.S. study, Gottschalk et al. found that family     Methods                                            motion, 33% work sampling, and 8% a self-
physicians spend 55% of their time with face-                                                          report survey approach [9].
to-face patient care, while other activities pri-   The traditional methods for time measure-              For our present study, we selected work
marily involved reviewing medical records,          ment comprise either the subjective esti-          sampling, as it allows for only one observer
writing notes, and writing prescriptions [8].       mation by the actors themselves in a survey,       documenting the activities of several clini-
For Austria, the Austrian Medical Association       or the objective measurement by a trained          cians. We followed the steps of work sampling
estimates that physicians in hospitals spend        observer. The second method is typically pre-      as described by Sittig [20]: First, the identifi-
no more than 63% of their time for direct           ferred, as the first one only provides an impre-   cation of working categories; then, the con-
patient care, without providing the source of       cise, and potentially biased, measure of ac-       duction of a pilot study for a sample size cal-
these data [5].                                     tivity [11]. For objective time measurements,      culation; finally, the conduction and analysis
    The electronic patient record (EPR) and         the two most widely-used approaches are            of the main study.
other more specialized computer-based               time-motion studies, as introduced by F. W.            Our analysis was conducted at a 200-bed
documentation systems promise to support            Tayler (in the 1880s) [12], and work sam-          hospital in Tyrol between November 2006
documentation and to reduce documen-                pling, as introduced by L. H. C. Tippett (in the   and January 2007. The study was conducted
tation efforts. Several evaluation studies have     1930s) [13].                                       in two wards of the inpatient area of the de-
investigated the relationship between intro-            In time-motion studies, trained observers      partment of internal medicine. Both wards
duction of an EPR system and time efficiency.       measure the duration of activities by docu-        admit around 520 patients each year. The
In a recent review, Poissant et al. [9] analyzed    menting their beginning and end, using a           mean patient length of stay in this depart-
seven studies evaluating the effects of an EPR      predefined classification of activities. This      ment is around 18 days, with mostly post-
on the time efficiency of physicians. Four of       method has been applied, for example, to           surgical patients treated. All of the eight
those studies reported an increase in the           measure the impact of an EPR system on the         physicians (one doctor-in-training, four resi-
documentation time (by 11-41%), whereas             time use of oncologists [14], or to analyze the    dent physicians, three senior physicians)
three studies reported a reduction (by              time needed after the introduction of com-         working in the observed wards during the
13–46%). Poissant et al. [9] found com-             puter-based physician order entry [15].            study period agreed to participate and were
parable varying results when analyzing                  In a work-sampling analysis, a trained ob-     included in the study.
studies on the time efficiency of nurses. Rea-      server documents which activity is just being          This hospital is equipped with a clinical
sons for the observed differences among the         executed at predefined (for example, every five    information system (Cerner Millenium,
reviewed studies may comprise differences in        minutes) or randomly selected moments in           [21]) that supports several clinical activities,
the amount of documented information, in            time, also by using a predefined classification    such as order entry and result reporting for
hardware equipment (for example mobile              of activities. By counting the number of ob-       lab and x-ray, report writing, and patient-
tools), clinical workflow, and usability and        served activities in each category, the overall    related scheduling. A paper-based record is
quality of the IT systems in the evaluated set-     distribution and thus the duration of each         still maintained for the documentation of
tings [7, 9]. An increase in workload for phys-     task can be estimated. The larger the number of    clinical admissions, vital signs, prescriptions,
icians can lead to low user satisfaction and        observations, the more precise this estimation     ongoing status documentation, and nursing
even user boycott [10].                             can be. Work sampling was used, for example,       care planning.
    Overall, a rising demand for clinical and       to study the work distribution of physicians           We developed the initial classification of
administrative documentation may lead to a          in a general medical service unit [16].            activities that are needed for the work-sam-
decrease in the direct time available for pa-           The most important advantage of work           pling analysis, based on an earlier work of
tient care and reduced job satisfaction for         sampling is that the data for several clinicians   Blum et al. who investigated the documen-
physicians. This problem is currently being         can be obtained by only one observer, which        tation efforts in German hospitals [22].
actively discussed in Austria. However, there       makes it rather efficient compared to con-         Castelein later adapted Blum’s classification
seems to be no objective data on the overall        tinuous time-motion studies where, typically,      for an Austrian hospitals’ setting [23]. We
time for documentation compared to the              one observer shadows one clinician [17]. In        used his classification as the basis for our
overall time for patient care.                      addition, work sampling minimizes the risk         study. We also reviewed the international
                                                    that clinicians’ behavior will be affected by      literature to check the completeness of our
                                                    being observed permanently [7]. The dis-           classification system.

© Schattauer 2009                                                                                                            Methods Inf Med 1/2009
86   E. Ammenwerth; H.-P. Spötl: The Time Needed for Clinical Documentation versus Direct Patient Care

     Table 1 Distribution of the most important activities of the observed physicians for the overall study          The resulting list of activities was refined
     period. Mean and standard deviation of the percentage of the overall working time is indicated. Only        by a pilot study in the hospital, which was
     those categories higher than 1.5% (for categories I to III) resp. higher than 0.5% (for categories IV and   conducted in November 2006. This pilot
     V) are indicated. For a complete list and definition of the categories, see Appendix.                       study comprised both direct observations of
      Category                                                             Mean (Standard Deviation)
                                                                                                                 clinical workflow as well as interviews with
                                                                                                                 the physicians. The interviews that were con-
      I. Direct patient care                                               027.5% (10.5%)                        ducted with two physicians were used to dis-
         Communication with patients                                       00 9 .3% (5.5%)
                                                                                                                 cuss face validity of our instrument, and to
         Other patient care                                                00 7 .2% (6.0%)
                                                                                                                 check the definitions of each category. The di-
         Medical activities                                                00 5 .9% (5.3%)
         Read in patient record                                            00 5 .1% (3.2%)                       rect observations within the pilot study lasted
         Waiting for patient                                               00 0 .0% (0.1%)                       eight hours using a one-minute work sam-
                                                                                                                 pling interval. The observations were used to
      II. Communication                                                    036.2% (10.5%)                        train the observer, to test the prepared docu-
          Personal communication with physicians                           0 12.9% (6.8%)                        mentation form and to assess the complete-
          Regular meetings                                                 0 10.0% (8.4%)                        ness and clarity of each category. Overall, only
          Phone calls                                                      00 4.6% (3.2%)                        slight modifications mostly in wording of in-
          Personal communication with non-physicians                       00 4.0% (2.5%)
                                                                                                                 dividual categories were done as a result of
          Other communication                                              00 1.6% (2.9%)
                                                                                                                 the pilot study. The pilot study was conducted
      III. Clinical documentation                                          022.4% (10.7%)                        by that observer who also conducted the final
          Writing of a preliminary discharge letter                        00 4.5% (3.9%)                        study. No further formal reliability testing of
          Ongoing clinical documentation                                   00 3.2% (3.0%)                        the instrument was conducted.
          Writing of a final discharge letter                              00 3.6% (2.8%)                            The findings from the pilot observations
          Documentation of an initial examination                          00 2.2% (3.1%)                        were used to calculate the needed number of
          Prepare documentation forms                                      00 3.4% (2.6%)                        observations, using the formula provided by
          Documentation of findings                                        00 1.9% (2.6%)                        Sittig (n = p (1 – p)/σ2, with n = total number
          Prepare forms for order entry                                    00 1.9% (1.1%)
                                                                                                                 of observations, p = expected percentage of
          Documentation of medication                                      00 1.5% (1.4%)
                                                                                                                 time required by the most important cat-
      IV. Administrative documentation                                     004.2% (4.6%)                         egory of study (estimated from pilot), and σ
         Generation of duty rosters                                        00 2.1% (4.6%)                        standard deviation of percentage) [20]. Based
         Writing of discharge documents                                    00 0.7% (1.8%)                        on this formula, we calculated n = 2244 for
         Completing of transportation orders                               00 0.5% (0.9%)                        our study. Estimating a planned duration of
         Other administrative documentation                                00 0.5% (0.9%)                        observation of 5 days à 8 hours, this n would
                                                                                                                 be reached by 449 observations per day resp.
      V. Other activities                                                  009.7% (7.4%)
                                                                                                                 28 observations per hour. This means one
         Walking times                                                     00 2.6% (1.4%)
                                                                                                                 observation every two minutes.
         Breaks                                                            00 5.1% (3.6%)
         Other                                                             00 2.1% (7.1%)                            The final classification system comprised
                                                                                                                 37 categories, 21 describing documentation
      Sum                                                                  100%                                  activities, with 11 related to clinical docu-
                                                                                                                 mentation and 10 to administrative docu-
                                                                                                                 mentation. 씰Appendix 1 shows the classifi-
                                                                                                                 cation system.
                                                                                                                     The main work-sampling study was con-
                                                                                                                 ducted in December 2006 and January 2007,
                                                                                                                 and comprised 40 hours of observations dur-
                                                                                                                 ing the day shifts, with each day of the week
                                                                                                                 covered equally. Based on the results from the
                                                                                                                 pilot study, the chosen sampling period was
                                                                                                                 two minutes. The observer (HS) used a pro-
                                                                                                                 grammable watch that beeped every two
                                                                                       Fig. 1
                                                                                                                 minutes. Typically, three to four physicians
                                                                                       Distribution of the
                                                                                       activities of the ob-     were observed in parallel by the observer. A
                                                                                       served physicians for     typical observation session lasted eight hours.
                                                                                       the overall study         If necessary, the observer looked into the pa-
                                                                                       period. The details       tients’ rooms in case one of the physicians was
                                                                                       are shown in Table 1.     there at the moment of observation, to

     Methods Inf Med 1/2009                                                                                                                  © Schattauer 2009
E. Ammenwerth; H.-P. Spötl: The Time Needed for Clinical Documentation versus Direct Patient Care                      87

capture patient-related activities. Overall, 30   documentation effort of an individual phy-          none of the studies reviewed by Poissant et al.
so-called physician-days were observed (a         sician was 21.6% of the overall working time,       [9] were comparable to our study. Most
physician-day reflecting a full-day observa-      and the highest was 36.2%. In 12 of 30 ob-          studies focus either on other professional
tion of one physician).                           served individual physician-days, the daily         groups (for example nurses), on outpatient
    The observation form that we used docu-       documentation load for a physician was near         areas, on specialized inpatient clinical settings
mented the following information for each         30% or above.                                       (such as intensive care or emergency care
observation point: the actor (name of phy-                                                            units), or only on certain activities (such as
sician); the performed activity (see Appen-                                                           order entry). For example, the study of Gott-
dix); and the tool used (computer-based,          Discussion                                          schalk et al. [8] analyzing activities of general
paper-based, or other). Overall, 5500 obser-                                                          physicians found that they spent around 20%
vations were documented. MS Excel 2003 was        Meaning and Generalizability                        of their time documenting – this lower
used to analyze the respective data. First, for   of the Results                                      number may reflect, however, the lower docu-
each individual physician, the number of ac-                                                          mentation requirements in outpatient care.
tivities documented in each category was          Our most interesting finding was the sub-           Oddone et al. [24] analyzed the work dis-
translated into an individual percentage,         stantial proportion of 27% of the working           tribution of 36 physicians at a university
using the overall number of documented            time dedicated to documentation, compris-           medical center and found 43.6% for “patient
activities of this physician as denominator.      ing both clinical and administrative tasks.         evaluation” (comprising direct patient care
Then, based on those numbers, the mean and        We conducted our work-sampling analysis             and discussing patient care), 18.9% for edu-
standard deviation of the categories of all       during the main working hours, i.e. 8 a.m. to       cational activities, and 13.9% for adminis-
physicians were calculated.                       4.30 p.m. Physicians later stated that they         tration (for example charting, dictating,
                                                  often work overtime (i.e. after 4.30 p.m.), to      label/forms). Here it is unclear as to whether
                                                  finalize documentation tasks. If we estimate        the activities noted for patient evaluation
Results                                           that this overtime was around 40 minutes per        (such as physical exam, patient history, and
                                                  physician per day during the study period (as       ward rounds) may have also included related
씰Table 1 and 씰Figure 1 show the overall re-       estimated from the administrative working           documentation activities. Educational activ-
sults of the work-sampling study. 27.5% of all    time documentation of the department), the          ities were not relevant in our study, as the hos-
activities were related to direct patient care,   overall daily documentation workload would          pital is not an academic hospital. Hollings-
36.2% to communication activities, 26.6% to       increase from 26.6% to 32.4%.                       worth et al. [25] used a time-motion study to
documentation activities, and 9.7% to other           Before the study, the hospital manage-          analyze the time distribution in an emergency
activities. The clinical documentation activ-     ment had stated that the documentation              unit. They found that the observed ten faculty
ities accounted for 22.4%. Documentation          should not exceed 30% of the working time.          physicians spent 32% of their time on direct
activities typically defined as “adminis-         While the mean (without overtime) is just           patient care, 22% on communication, and
trative” (i.e. coding for billing purposes,       below this threshold, each of the eight ob-         around 18.5% on charting and other paper-
documentation for quality management) ac-         served physicians spend at least one day            work. Mamlin et al. [19] conducted a com-
counted for 4.2% overall, that is 15.7% of all    (from five) with more than 30% of their time        bined time-and-motion and work-sampling
documentation time.                               needed for documentation activities.                study in a general medicine clinic and found
    We also documented which tools were               In a survey-based self-assessment study of      that physicians spent 37.8% of their time
used for the documentation activities (cat-       1010 German physicians conducted by Blum            charting, this reflecting the purely paper-
egories III and IV). Here, we found that for      et al., they found a documentation effort of        based documentation at the time of the study.
49.3 ± 19.7% of the documentation tasks,          40.6% [22]. The Austrian Medical Associ-                A recent study by Westbrook et al. [26] is
paper-based tools (for example paper-based        ation has stated that physicians in hospitals       better comparable to our study; they used a
patient record, paper-based forms for order       spend up to 63% of their time on documen-           time-motion approach to quantify work
entry or duty rostering) were used. Com-          tation [5]. Those subjective estimations may        activities of doctors in a 400-bed teaching
puter-based tools (for example electronic pa-     be biased [24] – a rising dissatisfaction of        hospital where also a mix of computer-based
tient record, office, and statistic tools) were   physicians in Germany and Austria with what         and paper-based tools was used. They found
used for 49.3 ± 19.7% of the documentation        they call “bureaucracy” may have led to those       that 33% of the time was spent on professional
tasks.                                            rather high subjective estimates. Our objec-        communication, 32% on (direct and indirect)
    We also analyzed the activity distribution    tive measurements confirm that the docu-            patient care, and 12% for documentation (ex-
for each individual physician to calculate the    mentation efforts in the inpatient area are         cluding medication documentation).
individual daily and weekly documentation         quite high with 27%. However, only one-sixth            A high documentation effort is often not
effort. The lowest daily documentation effort     of this time is clearly devoted to adminis-         well accepted, and physicians argue that
of an individual physician at a given day was     trative documentation.                              documentation takes away time from direct
8.5% (for a resident), the highest 55.5% (for a       Studies that analyzed the distribution of       patient care, and thus endangering the quality
senior physician). When analyzing activity        physicians’ activities in a clinical setting com-   of care. Especially documentation tasks that
distribution over one week, the lowest weekly     parable to our study are rare. For example,         are seen as not directly related to patient care

© Schattauer 2009                                                                                                           Methods Inf Med 1/2009
88   E. Ammenwerth; H.-P. Spötl: The Time Needed for Clinical Documentation versus Direct Patient Care

     (such as documentation for quality manage-         physician (this is comparable to the 5% found      who used 29 categories. Other authors have
     ment or for legal reasons) are often not well      by Westbrook et al. [26]), which means ap-         used much fewer detailed categories for
     accepted. In our study, we found 4.2% of the       prox. 30% of the overall documentation time.       documentation. For example, Bürkle et al.
     working time (i.e. 15.7% of all the documen-       Better computer-based support (including           [33] used 23 categories to document nursing
     tation time) is spent exclusively on “adminis-     speech recognition and computer-supported          activities, only one of which was clearly re-
     trative” documentation tasks, arguably re-         correction workflow) may help to reduce            lated to documentation activities. Westbrook
     flecting a moderate effort. In addition, both      documentation efforts, and reduce the turn-        et al. [26] used 22 categories, two of them
     clinical and administrative documentation is       around time of discharge letters.                  devoted to documentation.
     vital to provide good-quality, affordable and          In Austria, physicians’ organizations are          We decided to execute the work-sampling
     coordinated care to patients [8], especially in    calling for the introduction of so-called          observations at fixed intervals. Sittig [20] rec-
     a health care setting that is characterized by a   documentation assistants [5], who would            ommends fixed intervals for observation of
     large number of different professional groups      take over certain documentation tasks in           random work activities such as the hospital-
     and institutions and by rising economic pres-      order to reduce the workload of the phy-           related activities in our study. For fixed obser-
     sure. Therefore, the question should not just      sicians. However, if we look at the documen-       vations, Nickman [11] recommends a mini-
     be how to reduce the documentation efforts,        tation activities, only a few of them (such as     mum of eight observations per hour. With 30
     but rather how to plan and organize it to best     coding of diagnosis and services, completion       observations per hour, we were well over this
     support patient care [27].                         of transportation orders, documentation for        limit. The 2-minute interval observations
         Computer-based documentation systems           quality management, preparation of docu-           placed a high workload on the observer, but
     may be helpful in streamlining documen-            mentation forms) seem to be appropriate for        guaranteed that the activities of a short du-
     tation tasks, integrating data, and avoiding       delegation to documentation assistants.            ration were captured, which increased the
     unnecessary double data entry. In our hospi-       Other major documentation activities such          precision of our measurements. Instead of
     tal, already around half of the documentation      as discharge letter writing, ongoing clinical      using a paper-based work-sampling form, in
     tasks are supported by computer-based tools,       documentation, documentation of initial            order to reduce the time needed for data
     and this percentage is expected to increase in     examinations, and documentation of medi-           analysis as well as to eliminate transcription
     the coming years. For example, the documen-        cation, could not be delegated to non-             errors, the use of a PDA for activity docu-
     tation of the initial examination and of the       medical professionals. In our opinion, it is,      mentation might have been helpful, as, for
     medical history is still performed on paper-       therefore, questionable as to whether docu-        example, proposed by [34] and used in
     based forms. By using bedside terminals or         mentation assistants can help to reduce            studies by [11, 26].
     mobile tools, such as laptops or PDAs, docu-       documentation tasks of physicians. Earlier             Our results may be subject to certain er-
     mentation may be facilitated and might in-         studies showed that a medical assistant can        rors. For example, a misinterpretation of the
     crease the data quality [28, 29]. In addition,     help support physicians in certain areas of        work category definitions is one possible
     the use of mobile tools could also help to re-     general information logistics, such as looking     source of error. We attempted to limit the im-
     duce the time needed for the paper-based           for records or test results [32]. However, these   pact of this threat by having just one well-
     documentation of medication and other              were results from mostly settings using            trained observer, by defining each category,
     clinical notes as well as for general order        paper-based records and may not be repro-          and by conducting a pilot study to validate
     entry, which are at the moment performed in        ducible in settings with already high levels of    the working categories. Overall, each of our
     the paper-based records. Using a CPOE sys-         computer support.                                  physicians was observed over approximately
     tem (computerized physician order entry)                                                              four full days. We cannot be certain that the
     may even help to increase patient safety by of-                                                       observed days were representative, even when
     fering checks and alerts [30, 31]. Altogether,     Strengths and Weaknesses                           we attempted to guarantee representative
     we estimate that another approx. 25% of the        of the Study                                       days by distributing the observation days
     documentation time could be supported by                                                              equally over the whole week and over two
     computer-based point-of-care tools.                Using work sampling it was possible for us to      months. On all the observation days, the bed
         In addition, workflow reorganization may       observe four physicians at the same time be-       occupancy was high, which seems to be rep-
     help to reduce any unnecessary documen-            cause physicians mostly stayed in the defined      resentative for the overall situation of the
     tation tasks. For example, the writing of dis-     area of a ward. The observer only had to take      observed department.
     charge letters is – at the moment – a rather       a short look at what a physician was doing at a        The present study was conducted in a de-
     complicated process, with a physician dictat-      certain moment, thereby minimizing the             partment of internal medicine, characterized
     ing both a preliminary and later a final letter,   danger of a Hawthorne effect and avoiding          by a mean length of stay of approx. 18 days,
     and the final letter is written by secretaries     any disturbance of the clinical workflow.          and with mostly post-surgical patients
     with a long paper-based correction process             We developed a classification system with      treated. This duration of stay is much higher
     involving the author as well as the senior and     37 activity categories, 21 of which were re-       than the overall mean duration of stay in Aus-
     head physician. Overall, discharge letter pro-     lated to documentation activities, as this was     trian hospitals, which was 5.7 days in 2006 [2].
     duction (both preliminary and final ones)          the major focus of our study. Our classifi-        Our results may, therefore, not be generaliz-
     sums to 8% of the daily working time of a          cation was based on the work of Blum [22]          able to other inpatient settings in Austria.

     Methods Inf Med 1/2009                                                                                                            © Schattauer 2009
E. Ammenwerth; H.-P. Spötl: The Time Needed for Clinical Documentation versus Direct Patient Care                                             89

                                                                 http://www.ifes.at/upload/1156086926_folie.pdf.               aufnahme und Verbesserungsvorschläge – Unter-
Conclusion                                                       Last accessed: August 1, 2008.                                suchung im Auftrag der Deutschen Kranken-
                                                             7. Fontaine BR, Speedie S, Abelson D, Wold C. A work-             hausgesellschaft [Clinical documentation effort
We found that a substantial proportion of                        sampling tool to measure the effect of electronic             in hospitals]. Düsseldorf: Wissenschaft und Praxis
26.6% of working time was dedicated to                           medical record implementation on health care                  der Krankenhausökonomie (Band 11). Deutsche
documentation in a department of internal                        workers. J Ambul Care Manage 2000; 23 (1): 71–85.             Krankenhaus           Verlagsgesellschaft.       ISBN
                                                             8. Gottschalk A, Flocke SA. Time spent in face-to-face            3-935762.22-4; 2003. http://www.dkvg.de/index.
medicine, 22.4% of which was for clinical                        patient care and work outside the examination                 php?cat=c23_Wissenschaft-und-Praxis-der-
documentation, and 4.2% for administrative                       room. Ann Fam Med 2005; 3 (6): 488–493.                       Krankenhausoekonomie.html. Last accessed: Au-
documentation. The time for direct patient                   9. Poissant L, Pereira J, Tamblyn R, Kawasumi Y. The              gust 1, 2008.
care was 27.5% and, therefore, was only                          impact of electronic health records on time efficien-   23.   Castelein J. Analyse des Dokumentationsaufwands
                                                                 cy of physicians and nurses: a systematic review. J           im ärztlichen Bereich und seine Auswirkung auf die
slightly higher than the time spent for docu-                    Am Med Inform Assoc 2005; 12 (5): 505–516.                    Arbeitszeitbelastung von Krankenhaus-Ärzten am
mentation. The further introduction of com-                  10. Ornstein C. California; Hospital Heeds Doctors,               Beispiel des Landeskrankenhauses Innsbruck (Uni-
puter-based tools and the reorganization of                      Suspends Use of Software: Cedars-Sinai physicians             versitätskliniken). PhD thesis, UMIT – University
the working processes (such as discharge                         entered prescriptions and other orders in it, but             for Health Sciences, Medical Informatics and Tech-
                                                                 called it unsafe. In: The Los Angeles Times, Jan 22;          nology, Hall in Tyrol, Austria. In preparation; 2008.
letter writing) may help to reduce the docu-                     2003.                                                   24.   Oddone E, Guarisco S, Simel D. Comparison of
mentation efforts that physicians often state                11. Nickman NA, Guerrero RM, Bair JN. Self-reported               housestaff ’s estimates of their workday activities
are excessively high. Further research is                        work-sampling methods for evaluating phar-                    with results of a random work-sampling study.
                                                                 maceutical services. Am J Hosp Pharm 1990; 47 (7):            Acad Med 1993; 68 (11): 859–861.
needed to see whether similar results can be
                                                                 1611–1617.                                              25.   Hollingsworth JC, Chisholm CD, Giles BK, Cordell
found in other inpatient settings with a short-              12. Barnes R. Motion and Time Study Design and                    WH, Nelson DR. How do physicians and nurses
er mean duration of stay. In addition, evalu-                    Measurement of Work. 6th ed. New York: John                   spend their time in the emergency department?
ation studies may help to show how docu-                         Wiley and Sons; 1968.                                         Ann Emerg Med 1998; 31 (1): 87–91.
                                                             13. Tippett L. Statistical Methods in Textile Research.     26.   Westbrook JI, Ampt A, Kearney L, Rob MI. All in a
mentation efforts will develop after the intro-                  Uses of the Binomial and Poissant Distributions. J            day’s work: an observational study to quantify how
duction of computer-based tools and/or the                       Textile Inst Trans 1935 (26): 51–55.                          and with whom doctors on hospital wards spend
employment of documentation assistants.                      14. Pizziferri L, Kittler AF, Volk LA, Shulman LN,                their time. Med J Aust 2008; 188 (9): 506–509.
                                                                 Kessler J, Carlson G, et al. Impact of an Electronic    27.   Leiner F, Haux R. Systematic Planning of Clinical
                                                                 Health Record on oncologists’ clinic time. AMIA               Documentation. Methods Inf Med 1996; 35 (1):
Acknowledgments                                                  Annu Symp Proc 2005. p 1083.                                  25–34.
We thank the management and staff of the                     15. Overhage JM, Perkins S, Tierney WM, McDonald            28.   Lu YC, Xiao Y, Sears A, Jacko JA. A review and a
analysed Tyrolean hospital for their co-oper-                    CJ. Controlled trial of direct physician order entry:         framework of handheld computer adoption in
                                                                 effects on physicians’ time utilization in ambulatory         healthcare. Int J Med Inform 2005; 74 (5): 409–422.
ation and support.                                               primary care internal medicine practices. J Am Med      29.   Wu RC, Straus SE. Evidence for handheld electronic
                                                                 Inform Assoc 2001; 8 (4): 361–371.                            medical records in improving care: a systematic re-
                                                             16. Guarisco S, Oddone E, Simel D. Time analysis of a             view. BMC Med Inform Decis Mak 2006; 6: 26.
                                                                 general medicine service: results from a random         30.   Kaushal R, Shojania KG, Bates DW. Effects of com-
References                                                       work sampling study. J Gen Intern Med 1994; 9 (5):            puterized physician order entry and clinical deci-
                                                                 272–277.                                                      sion support systems on medication safety: a sys-
1. Haux R, Ammenwerth E, Herzog W, Knaup P.                  17. Wirth P, Kahn L, Perkoff GT. Comparability of two             tematic review. Arch Intern Med 2003; 163 (12):
   Health Care in the Information Society: A Progno-             methods of time and motion study used in a clinical           1409–1416.
   sis for the Year 2013. Int J Med Inform 2002; 66              setting: work sampling and continuous observa-          31.   Rothschild J. Computerized physician order entry
   (1–3): 3–21.                                                  tion. Med Care 1977; 15 (11): 953–960.                        in the critical care and general inpatient setting: a
2. BMGF. Krankenanstalten in Zahlen. 2006; Available         18. Finkler SA, Knickman JR, Hendrickson G, Lipkin                narrative review. J Crit Care 2004; 19 (4): 271–278.
   from: http://www.kaz.bmgf.gv.at/. Last accessed:              M, Jr, Thompson WG. A comparison of work-sam-           32.   Wipf JE, Fihn SD, Callahan CM, Phillips CM. How
   August 1, 2008.                                               pling and time-and-motion techniques for studies              residents spend their time in clinic and the effects of
3. Mechanic D. Physician discontent: challenges and              in health services research. Health Serv Res 1993; 28         clerical support. J Gen Intern Med 1994; 9 (12):
   opportunities. JAMA 2003; 290 (7): 941–946.                   (5): 577–597.                                                 694–696.
4. Probst JC, Greenhouse DL, Selassie AW. Patient and        19. Mamlin JJ, Baker DH. Combined time-motion and           33.   Bürkle T, Kuch R, Prokosch HU, Dudeck J. Stepwise
   physician satisfaction with an outpatient care visit. J       work sampling study in a general medicine clinic.             Evaluation of Information Systems in an University
   Fam Pract 1997; 45 (5): 418–425.                              Med Care 1973; 11 (5): 449–456.                               Hospital. Methods Inf Med 1999; 38 (1): 9–15.
5. Brettenthaler R, Mayer H. Stopp der Zettel-               20. Sittig DF. Work Sampling: A Statistical Approach to     34.   Westbrook JI, Ampt A, Williamson M, Nguyen K,
   wirtschaft – Bürokratie macht krank [stop paper               Evaluation of the Effect of Computers on Work Pat-            Kearney L. Methods for measuring the impact of
   chaos – bureaucracy makes sick]. 2007. Available              terns in The Healthcare Industry. Methods Inf Med             health information technologies on clinicans’
   from: http://www.buerokratiestopp.at/statements.              1993; 32 (2): 167–174.                                        patterns of works and communication. In: Kuhn K,
   html. Last accessed: August 1, 2008.                      21. Lechleitner G, Pfeiffer K-P, Wilhelmy I, Ball M.              Warren J, Leong T-Y, editors. Medinfo 2007 – Pro-
6. Mayer H. Spitalsärztinnen in Österreich 2006.                 Cerner Millenium: The Innsbruck Experience.                   ceedings of the 12th World Congress on Health
   Ergebnis der Befragung für die Bundeskurie                    Methods Inf Med 2003; 42 (1): 8–15.                           (Medical) Informatics, Brisbane, August 20–24,
   Angestellte Ärzte der österreichischen Ärzte-             22. Blum K, Müller U. Dokumentationsaufwand im                    2007. Amsterdam: IOS Press; 2007. pp 1083–1087.
   kammer. Österreichische Ärztekammer; 2006.                    Ärztlichen Dienst der Krankenhäuser – Bestands-

© Schattauer 2009                                                                                                                                     Methods Inf Med 1/2009
90   E. Ammenwerth; H.-P. Spötl: The Time Needed for Clinical Documentation versus Direct Patient Care

     Appendix
     Classification System Used for Analysis of Physician´s Activities

     No.     Name of Category                             Definition of Category
     I       Direct patient care
     I.1     Medical activities                           Any diagnostic and therapeutic activity of the physician, related to the care of a patient
     I.2     Communication with patients                  Direct conversation between the physician and patient
     I.3     Waiting for a patient                        Physician is waiting for the next patient to arrive
     I.4     Read in patient record                       Get information on the patient from the patient record
     I.5     Other direct patient care                    Other activities of direct patient care
     II      Communication
     II.1    Personal communication with physicians       Direct conversation with other physicians
     II.2    Personal communication with non-physicians   Direct conversation with health care professionals other than physicians (for example,
                                                          nurses, co-therapists)
     II.3    Personal communication with relatives        Direct conversation with family members of a patient
     II.4    Phone calls                                  Phone calls with other health care providers (excluding phone calls for patient-related
                                                          scheduling)
     II.5    Communication for scheduling                 Organization (mostly by phone calls) of patient-related appointments (such as diagnostic
                                                          or therapeutic examinations, next inpatient admission)
     II.6    Electronic communication                     Use of e-mail, Intranet, and Internet
     II.7    Regular meetings                             Any meetings that take place at a predefined time
     II.8    Other communications                         Any other communication activities
     III     Clinical documentation
     III.1   Documentation of the initial examination     Documentation of the initial examination of a patient after his admission to the hospital
     III.2   Ongoing clinical documentation               Any written entries in the patient record, such as clinical notes (for example, during ward
                                                          rounds)
     III.3   Documentation of findings                    Filing or copying of recent findings (such as lab or x-ray reports) into the patient record
     III.4   Writing of preliminary discharge letter      Writing of the preliminary discharge letter upon the discharge of the patient from the hospital
     III.5   Writing of final discharge letter            Writing of the final discharge letter, including the correction process and transport time
     III.6   Writing of consultation letters              Writing of a consultation letter for other departments
     III.7   Documentation of medication                  Documentation of the prescribed drugs of a patient and of any changes to prescriptions
     III.8   Preparation of documentation forms           Prepare weekly documentation forms for a patient (for example, for care planning and care
                                                          documentation)
     III.9   Preparation of the forms for order entry     Order diagnostic or therapeutic procedures using predefined forms
     III 10 Writing of prescriptions                      Filling-out paper-based prescription forms for a patient that is going to be discharged
     III.11 Other clinical documentation                  Any other documentation related to a patient
     IV      Administrative documentation
     IV.1    Coding of the diagnosis and services         Documentation and coding of the diagnoses and services for accounting and legal reasons
     IV.2    Completing of transportation orders          Completing of a transportation order form for a patient
     IV.3    Writing of doctors’ certificates             Writing of any patient-related certificates (for example, inability to work)
     IV.4    Documentation for external quality manage-   Documentation of data for any quality reports
             ment
     IV.5    Documentation of the working time            Personal documentation of the daily hours of work

     Methods Inf Med 1/2009                                                                                                               © Schattauer 2009
E. Ammenwerth; H.-P. Spötl: The Time Needed for Clinical Documentation versus Direct Patient Care                    91

Appendix
Classification System Used for Analysis of Physician´s Activities (continued)

No.    Name of Category                             Definition of Category
IV.6   Generation of departmental statistics        Development and update of departmental-oriented statistics related to patient care
IV.7   Generation of duty rosters                   Development and update of departmental duty rosters for the clinical staff
IV.8   Writing of discharge documents               Finalization of the administrative discharge documents of a patient
IV.9   Writing of requests                          Prepare patient-oriented applications for example, for rehabilitation, aftercare, nursing care, or
                                                    further hospitalization
IV.10 Other administrative documentation            Any other administrative documentation
V      Other activities
V.1    Walking times                                Physician walking between rooms, departments, etc.
V.2    Breaks                                       Any breaks
V.3    Other                                        Any other activities (for example, private phone calls)

© Schattauer 2009                                                                                                           Methods Inf Med 1/2009
You can also read