- Email: [email protected]
Are things different in the light of day? A time study of internal medicine house staff days
SPECW ARTiCLE
/
Are Things Different in the Light of Day? A Time Study of Internal Medicine House Staff Days CONNIE PARENTI, M.D., NICOLE LURIE, M.D., M.s.P.H., Minneapo/is, Minnesota
To determine how internal medicine house staff spend their days and compare activities during the day with those previously observed during night calL SEITING: University-affiliated VeteranS Affairs Medical Center. DESIGN: Two internal medicine house staff teams (one PGY-1 [postgraduate year] and one PGY-2 or PGY-3) observed during 5 short call admitting days. m Time in each activity recorded by trained observers, computed, summed, and compared with that of similar activities of house staff on night call. RJBULTSZ House staff admitted an average of two patients each day. They spent, on average, 25 minutes per patient performing new patient histories and physical examinations, 29 minutes charting new patient information, and were interrupted after an average of 12 minutes during the new patient evaluation compared with 20,19, and 7 minutes at night (p >0.05). The average house officer spent 44 minutes in nonphysician duties and 11 minutes answering pages duriug the day. On average, house staff spent 3 minutes each day tall&g in person witi+patients’ families. CONCLUSIONs: A significant amount of time each day was spent performing nonphysician duties. Little time was spent evaluating each patient or in person with patients’ families and similar amounts of time were spent in charting and in patient evaluation. House staff appeared to spend more time with new patients during the day than they did at night, although this finding was not statistically significant. As noted during night call, evaluations were frequently interrupted. Future studies should examine why PURPOSE:
From the Departments of Medicine, Minneapolis Veterans Affairs Medical Center, and the Hennepin County Medical Center, University of Minnesota, Minneapolis. Minnesota. This work was presented, in part, at the National Meeting of the Association of Program Directors of Internal Medicine in April 1991, where the presentation received an abstract competition award. Requests for reprints should be addressed to Connie Parenti. M.D., Department of Medicine (111). Veterans Affairs Medical Center, One Veterans Drive, Minneapolis, Minnesota 55417. Manuscript submitted August 21. 1992. and accepted in revised form December 8.1992.
654
June 1993
The Amerlcen Journal of Medlclne
Volume 94
house staff choose to distribute their time in the manner described in this and similar studies.
I
n 1987, internal medicine departments began to critically evaluate their residency training programs in response to fewer students choosing careers in internal medicine [l] and to accommodate for the mandated changes in house staff work schedules [2]. Data on how internal medicine house staff spent their time on duty had been published almost 20 years earlier [3,4]. Therefore, in the spring of 1988, we observed 35 internal medicine house officers for 5 nights on call at 3 Minneapolis public teaching hospitals. That study [5] found that house staff at all three hospitals spent as much time charting as they did in direct patient contact, that patient evaluation’was frequently interrupted, and that almost no time was spent with patients’ families. In contrast, there was substantial time spent performing duties that potentially could have been done by a nonphysician. We hypothesized that the daytime experience would be different, specifically that there would be more patient and family contact when house officers were less tired and stressed. Therefore, we also observed the daytime activity of house staff at one of the hospitals that had participated in the study of night call. The results of those daytime observations are described and compared with our nighttime findings.
SUBJECTS AND METHODS Setting In June 1988, we studied the activity of 10 internal medicine house staff teams at the Minneapolis Veterans Affairs Medical Center (VA). The VA is a 650-bed primary- and tertiary-care hospital affiliated with the University of Minnesota Medical School. Each team consisted of one PGY-1 (postgraduate year) and one PGY-2 or PGY-3. The Department of Medicine at the VA had implemented a night float system 3 months prior to our observations. The detailed schedule has been previously published [6]. Ward teams were studied during five “short call” admitting days. During these days, each team was at risk for admitting a maximum of three new ward patients before 1 PM. None of the house staff teams was studied during the day pre-
TIME-MOWN STUDY OF HOUSE STAFF DAYS / PARENTI AND LURK
ceding or following long call. The average service size during the study was 12 patients, and teams admitted an average of 2 new patients per day. Study days were chosen to minimize duplication of resident teams and to include different days of the week but without regard to who the actual residents were. House Staff All house staff were University of Minnesota categorical internal medicine residents. PGY-1 residents are referred to as interns, and PGY -2 or PGY 3 residents are referred to as residents or senior residents. A total of 12 house staff participated in the study for 10 “intern days” and 10 “resident days” (some teams were observed for more than 1 day). All house staff observed were judged by the program director to be performing at a satisfactory level or better and had agreed to participate in the study. The mean age of the group was 29 years; 42% were women. Data Collection Data collection techniques have been previously described [5]. Each team member was followed from 8 AMuntil 4 PM by a lay observer with training in observational techniques. The observer recorded the type of activity, as well as the times at which it began and ended, with digital watches rounded to the nearest minute, recording consecutively. When simultaneous activities occurred (e.g., eating and charting), the activity most closely related to patient care was recorded. Activities Five major categories of activities were used: procedures, patient evaluation, communication, basics (eating, sleeping, hygiene), and miscellaneous. A modified list of the subcategories of activities for which data were collected appears in Table I. In the patient evaluation category, we distinguished patients newly admitted by the house officer the day he or she was on short call (new patients) from those who had been admitted previously and were evaluated as part of routine ward rounds (other patients). Because the intern and resident usually reviewed and discussed old records jointly, we grouped chart review and case discussion together. Time spent attending patient care conferences was recorded under other patient evaluation activities. We included in the communication category three types of chart entries-documentation of new patients’ histories and physical examination results, evaluation of the problems of patients who had been admitted previously, and writing of daily notes. Admission orders were considered part of the
TABLE I AverageTime Spentby Internsand Residentsin EachActMy Duringa ShortCallAdmittingDay ComparedWiih While on NightCall* Day (8 AM-~PM) Activity Procedures Patient Conducting evaluation new patients’ histories and physical examinations Evaluating other patients Reading charts or discussing patients Other Communication Charting Using the phone Receiving or answering pages Filling out forms Talking in person with patients families Other
Night(4 PM-~
AM)
Intern Resident (n = 10) (n = 10) (n = 10) (n = 10) Minutes(% oi total time) Intern
Resident
19 (4)
10 (2)
179 55 (35)
194 41 (38)
87 (12) l;;t(19)
21 (3) 1;; (18)
47
40
46
33
61
103
80
48
16
10
0
0
213 (42)
1;; (35)
1;;1(201
1;; (23)
z; 18
43 5
%
!
6 2
17 1
8 2
59
65
28
31
Basics
55 (11)
62 (12) 286 (40) 350 (49)
Miscellaneous Checking laboratory or radiographic results Looking for charts, radiographs, or equipment Transporting patients, laboratory specimens Other
40 (8) 20
63 (12) 24
L Total
!35 (g)
:z
58 (8)
11
12
9
12
19
2
0
15
6
6
30
21
22
506t
510t
720
720
‘his presentation of the night data is adapted with psrmfssron from 151 In four of the five mghts, the mtems ma not admrt new pattents. Ine trme shown ISme trme spent this activity on the remaining night. The total number of minutesfor these categories is based on is time averaged across all five nights. Percentagesmay not sum to 100 becauseof rounding. otal time represents slightfy more than 8 hours of duration because observers allowed house staff finish a given activity if it had been started prior to 4 PM.
documentation for new patients. Face-to-face communication with the nursing staff or other health care personnel and attending educational conferences were the principal entries under other types of communication. Only time spent in person with patients’ families was recorded in that subcategory. Data Analysis The time spent on each activity was computed and summed. The mean time spent on each activity was separated for interns and residents. The mean time and standard deviation spent per patient was computed for activities relating to obtaining histories, performing a physical examination, and charting information for newly admitted patients and for those patients previously admitted. Patients admitted the previous night by the “float team” were excluded from the separate analysis of new patients’ histories, physicals, and charting but were June 1993
The American Journal of Medlclne
Volume 94
655
TIME-MOTION
STUDY OF HOUSE STAFF DAYS / PARENTI AND LURIE
patients during the day than at night (an average of 23 minutes during the day versus 17 minutes at night for residents and 28 minutes during the day versus 24 minutes at night for interns); however, this trend did not reach statistical significance. During the day, house staff, like those at night, spent about as much time charting new patient information as they did actually performing new patient histories and physical examinations. Patient evaluation during the day was also frequently interrupted, although not as frequently as at night. For interns, the new patient history and physical was interrupted after an average of only 11 minutes, which represented two interruptions during each new evaluation. Once the initial evaluation was complete, only an average of 3.5 min/d was spent with each patient.
TABLE II AverageTime per New Patient Spent on History Taking, Physical Examination,and Charting andthe Amount ofTimeBefore These ActivitiesWere Interrupted* Day
Night Resident In = 30)
Intern (n = 16)
Resident Intern (n = 13) (n = 7) Minutes (SD)
History and physical examination of new patients
28 (12)
23 (14)
24 (8)
17 (10)
Charting new patient information
30 (17)
28 (22)
21 (10)
17 (11)
Time before interruption of the history and physical examination
11
14
Variable
7
7
included in the other measures of patient contact. The mean time spent on these activities for new patients was compared with that obtained observing internal medicine house staff on night call at the VA 1 month earlier [5]. The methods of data collection and analysis were the same for the day and night observations.
RESULTS Table I shows the average number of minutes and the percentage of total time spent during the day by interns and residents for each activity. The previously published data [5] from observations of house staff at the VA on night call are listed for comparison. As had been observed at night, surprisingly little time, an average of only 2.5 min/d, was spent with patients’ families. Time spent performing procedures during the day accounted for less time than at night but, as had been noted at night, most of these procedures did not require a physician (e.g., obtaining blood specimens and performing electrocardiograms). Other duties, such as transporting specimens or patients, or looking for charts, radiographs, or equipment, when combined with procedure time, accounted for more than an hour (an average of 76 minutes) each day spent on tasks that need not have been performed by a physician. Although house staff spent less time responding to pages during the day than they did at night, this activity, in addition to the time spent on the phone, accounted for about as much time as conducting new patient histories and physicals (on average, 47 minutes spent responding to pages and phone calls compared with 47.5 minutes in new patient evaluations). Table II shows the average time interns and residents spent on evaluation of new patients. There was a trend toward spending more time with new 666
June 1993
The American
Journal
of Medlclne
Volume
94
COMMENTS We have documented the activity of the house officers of a large categorical internal medicine residency program during short call admitting days. Two previously reported studies of the daytime activities of medicine house staff [3,4] had noted that little time was spent in patient contact and that a large proportion of time was spent charting and in communication with staff. However, those studies were reported almost 20 years earlier and had observed only interns on consecutive days when they had variable admitting responsibilities. Our study observed house staff during days when their admitting responsibilities did not vary, and senior residents as well as interns were observed. The expected new patient workload for the days we chose to observe should have allowed adequate time for house staff to evaluate patients and interact with family members. We found that the house staff we observed (interns and senior residents) spent surprisingly little time with patients even when they were rested with a manageable workload. Not only did we observe that house officers in our study spent little time with patients, we also observed that they spent little time with patients’ families. This was not an observation unique to the daytime observation or to the hospital site. We had observed similarly little time spent with patients’ families at the university and the county hospitals during our study of night call. It is possible that family members were not available for interaction with house staff and since we did not include time spent on the phone with patients’ families or when families were simultaneously seen with the patient, it is possible that we underestimated the total time spent in communication with the family. If we have not substantially underestimated the time spent with families, this is a particularly disturbing obser-
TIME-MOTION STUDY OF HOUSE STAFF DAYS / PARENTI AND LURIE
vation given a population of patients in whom the medical care of a significant number might be improved with family input. Once the initial evaluation was complete, house staff in our study spent an average of less than 5 minutes per patient each day, similar to what Wray et al [7] noted (an average of 4.6 minutes per patient each day) in their observations of internal medicine work rounds in 1986. Payson et al [3], following interns at Yale University in 1959, and Gillanders and Heiman [4], observing interns at the University of California, San Francisco, in 1969, had also observed that much less time was spent with patients as their hospital course progressed. Since there is no “gold standard” for the appropriate amount of time a physician should spend evaluating a given patient, we cannot conclude that the amount of time house staff in our study spent with patients was inadequate, only that it was less than we predicted and hoped it would be. Why house staff chose to spend so little time with patients or their families is not clear. We doubt that it was because of inadequate time available or because of a fear of a large number of admissions. Teams were at risk for a maximum of only three new patients and actually admitted an average of only two new patients each day. It is possible, however, that the unpredictable nature of the work associated with caring for acutely ill patients (a sense of “impending disaster around every corner”) creates an underlying pressure to gather data as quickly as possible even when the expected workload should be manageable. Nerenz et al [8] found that interns observed during the same academic year as this study spent a similar amount of time with new patient evaluations as those in our study. They also noted, as did we, a trend toward less time being spent with patient evaluation later in the day. We suspect that even less time is spent with patients at night because house staff feel even more pressure to be “efficient” at night to prepare themselves for unexpected work, to preserve time to spend in other activities such as sleep, or simply because they are adversely affected by fatigue as the workday progresses. Other explanations are also possible. For example, patients admitted at night may have characteristics different from those admitted during the day, which could explain this trend. In a study of work rounds at Baylor College of Medicine, Wray and colleagues [7] suggested that patient characteristics (such as do-not-resuscitate status, being nursing home candidates) may influence the amount of time house staff spend with patients. Finally, it is possible that house staff choose to avoid patients or their families, either because they are not comfortable spending time with them or
because it is not a personally satisfying activity. This may, in part, be due to the lack of emphasis given to teaching house officers about communication and the physician-patient relationship. It may also be that house staff are uncomfortable obtaining or do not value the clinical information that can be obtained from the patient. The availability of high-technology diagnostic tests may influence house staff to rely on means other than speaking with and examining their patients to formulate a plan of care. Two recent studies, one demonstrating poor physical diagnosis skills of house staff [9], and the other, poor clinical performance evaluation skills of faculty [lo], suggest that house staff may be deficient physical diagnosticians and that training programs may lack reliable methods for detecting these deficiencies. As suggested by a study designed to detect physical diagnosis errors and improve house staff clinical skills [ 111, poor physical diagnosis skills of house staff may not only have important implications for patient care but also for house staff satisfaction as well. That study noted improved house staff satisfaction following a program designed to enhance physical diagnosis skills. The house staff we observed spent a significant amount of time each day charting patient information, performing nonphysician duties (“scut” work), and responding to pages. This latter task was particularly frustrating because it represented a frequent source of interruption. Interns’ new patient evaluations were interrupted after an average of only 11 minutes, often by a page. Some of the time house staff spent in these activities could be eliminated by decreasing disruptions from paging as suggested by Katz and Schroeder [12], allowing dictation of patient information, and hiring personnel to perform nonphysician duties. Some limitations of the current study deserve mention. There is always a concern that observation alone may affect what house staff do-that is, the Hawthorne effect. We believe this is unlikely since residents frequently commented they were too busy to be affected by the observer, and we would have expected that, if they had been affected by the observer, more time would have been spent in patient evaluation or with patients’ families. Simultaneous activities were not recorded, and house staff were observed for only part of their workday (until 4 PM); therefore, time spent in some activities may have been underestimated. However, they had previously been observed from 4 PM until 8 AM during our study of night call [5] where the observations were similar. A larger sample size or observation of house staff at another time in the academic year may have resulted in different findings. Finally, this study of house staff days was performed June 1693
The American Journal of Medicine
Volume 94
657
TIME-MOTION
STUDY OF HOUSE STAFF DAYS / PARENTI AND LURIE
at only one hospital where a night float call system was in operation; therefore, the results may not be generalizable. Although these limitations should be considered when interpreting our results, many of the observations noted in this study are similar to those made by other investigators in different settings. In summary, during the day, patient evaluation was frequently interrupted, little time was spent with patients or patients’ families, and as much time was spent charting as in patient evaluation. Although the ideal distribution of house staff time is unknown, studies are now beginning to assess the effect of changing house staff work schedules on patient and resident outcome [6]. These studies have, however, primarily focused on the effects of decreasing sleep deprivation or altering workload. Few data are available that directly measure the effect of redistributing house staff activities within a given work schedule on patients or residents. Before it can be determined whether altering house staff activities would result in improved patient care or satisfaction, or in improved resident education or satisfaction, the motivation behind house staff choices regarding time distribution needs to be determined. For example, if house staff lack skill in physical diagnosis and therefore choose not to spend time with patients because they are more proficient at ordering and interpreting tests or documenting rather than obtaining clinical information, then, as suggested in a recent editorial [13] , clinical science needs to clearly define the complementary roles of clinical diagnosis and technologic assessment and encourage the development of outstanding clinical skills. However, if there are other reasons, such as lack of role modeling, that diminish the perceived importance of house staff-patient contact or if the current state of residency training creates an environment that does not encourage interaction with patients or their families, it is important that we determine if indeed “too little” time is being spent with patients and the effect of inade-
666
June 1993
The American
Journal
of Medlclne
Volume
94
quate personal contact on the patient or the physician in training. As residency programs attempt to balance service and education missions, it is critical to carefully evaluate why house staff choose to spend time as they do and the impact those choices have on patient care and house staff well-being.
ACKNOWLEDGMENT We are indebted to the house officers who made this study possible: to Thomas Ferris, Charles Moldow, and Robert Petzel for their support; and to Pamela Nesbitt for assistance in the preparation
of the manuscript.
REFERENCES 1.Graettinger JS. Internal medicine in the National Resident Matching Program 1987; the ides of March. Ann Intern Med 1988; 108: 101-15. 2. New York State Department Advisory Committee
of Health. Report of the New York State Ad Hoc
on Emergency
dent Working Conditions.
Services Regarding Supervision and Resi-
New York: New York State Department
of Health,
1987. 3. Payson HE, Gaenslen EC Jr, Stargardter
FL. Time study of an internship on a
university medical service. N Engi J Med 1961;264:439X3.
4. Gillanders W, Heiman M. Time study comparisons of 3 intern programs. J Med Educ 1971;46: 142-9. 5. Lurie N, Rank 8. Parenti C, Woolley T, Snoke W. How do house officers spend their nights? A time study of internal medicine house staff on call. N Engl J Med 1989; 320: 1673-7. 6. Gottlieb DJ, Parenti CM, Peterson CA, Lofgren RP. Effect of a change in house staff work schedule on resource utilization and patient care. Arch Intern Med
1991;151:2065-70. 7.Wray NP, Friedland JA, Ashton CM, Scheurich J. Zollo Al. Characteristics house staff work rounds on two academic Educ 1986,61:
of
general medicine services. J Med
893-900.
6. Nerenz D, Rosman H. Newcomb internal medicine.
C, et al. The on-call experience
of interns in
Arch intern Med 1990; 150: 2294-7.
9. St. Clair EW, Oddone
EZ, Waugh RA, Corey GR. Feussner JR. Assessing
housestaff diagnostic skills using a cardiology patient simulator. Ann Intern Med 1992; 117: 751-6.
10.Noel
GL, Herbers JE, Caplow MP, Cooper GS, Pangaro LN, Harvey J. How
well do internal medicine faculty members
evaluate
the clinical skills of resi-
dents? Ann Intern Med 1992: 117:757-65.
11.Wray NP, Friedland JA. Detection and correction
of house staff error in
physical diagnosis. JAMA 1983; 249: 1035-7. 12. Katz MH, Schroeder SA. The sounds of the hospital.
Paging patterns
in three
teaching hospitals. N Engl J Med 1988; 319: 1585-9. 13. Fletcher RH, Fletcher SW. Has medicine outgrown physical diagnosis? Ann Intern Med 1992;117:766-7.
/
Are Things Different in the Light of Day? A Time Study of Internal Medicine House Staff Days CONNIE PARENTI, M.D., NICOLE LURIE, M.D., M.s.P.H., Minneapo/is, Minnesota
To determine how internal medicine house staff spend their days and compare activities during the day with those previously observed during night calL SEITING: University-affiliated VeteranS Affairs Medical Center. DESIGN: Two internal medicine house staff teams (one PGY-1 [postgraduate year] and one PGY-2 or PGY-3) observed during 5 short call admitting days. m Time in each activity recorded by trained observers, computed, summed, and compared with that of similar activities of house staff on night call. RJBULTSZ House staff admitted an average of two patients each day. They spent, on average, 25 minutes per patient performing new patient histories and physical examinations, 29 minutes charting new patient information, and were interrupted after an average of 12 minutes during the new patient evaluation compared with 20,19, and 7 minutes at night (p >0.05). The average house officer spent 44 minutes in nonphysician duties and 11 minutes answering pages duriug the day. On average, house staff spent 3 minutes each day tall&g in person witi+patients’ families. CONCLUSIONs: A significant amount of time each day was spent performing nonphysician duties. Little time was spent evaluating each patient or in person with patients’ families and similar amounts of time were spent in charting and in patient evaluation. House staff appeared to spend more time with new patients during the day than they did at night, although this finding was not statistically significant. As noted during night call, evaluations were frequently interrupted. Future studies should examine why PURPOSE:
From the Departments of Medicine, Minneapolis Veterans Affairs Medical Center, and the Hennepin County Medical Center, University of Minnesota, Minneapolis. Minnesota. This work was presented, in part, at the National Meeting of the Association of Program Directors of Internal Medicine in April 1991, where the presentation received an abstract competition award. Requests for reprints should be addressed to Connie Parenti. M.D., Department of Medicine (111). Veterans Affairs Medical Center, One Veterans Drive, Minneapolis, Minnesota 55417. Manuscript submitted August 21. 1992. and accepted in revised form December 8.1992.
654
June 1993
The Amerlcen Journal of Medlclne
Volume 94
house staff choose to distribute their time in the manner described in this and similar studies.
I
n 1987, internal medicine departments began to critically evaluate their residency training programs in response to fewer students choosing careers in internal medicine [l] and to accommodate for the mandated changes in house staff work schedules [2]. Data on how internal medicine house staff spent their time on duty had been published almost 20 years earlier [3,4]. Therefore, in the spring of 1988, we observed 35 internal medicine house officers for 5 nights on call at 3 Minneapolis public teaching hospitals. That study [5] found that house staff at all three hospitals spent as much time charting as they did in direct patient contact, that patient evaluation’was frequently interrupted, and that almost no time was spent with patients’ families. In contrast, there was substantial time spent performing duties that potentially could have been done by a nonphysician. We hypothesized that the daytime experience would be different, specifically that there would be more patient and family contact when house officers were less tired and stressed. Therefore, we also observed the daytime activity of house staff at one of the hospitals that had participated in the study of night call. The results of those daytime observations are described and compared with our nighttime findings.
SUBJECTS AND METHODS Setting In June 1988, we studied the activity of 10 internal medicine house staff teams at the Minneapolis Veterans Affairs Medical Center (VA). The VA is a 650-bed primary- and tertiary-care hospital affiliated with the University of Minnesota Medical School. Each team consisted of one PGY-1 (postgraduate year) and one PGY-2 or PGY-3. The Department of Medicine at the VA had implemented a night float system 3 months prior to our observations. The detailed schedule has been previously published [6]. Ward teams were studied during five “short call” admitting days. During these days, each team was at risk for admitting a maximum of three new ward patients before 1 PM. None of the house staff teams was studied during the day pre-
TIME-MOWN STUDY OF HOUSE STAFF DAYS / PARENTI AND LURK
ceding or following long call. The average service size during the study was 12 patients, and teams admitted an average of 2 new patients per day. Study days were chosen to minimize duplication of resident teams and to include different days of the week but without regard to who the actual residents were. House Staff All house staff were University of Minnesota categorical internal medicine residents. PGY-1 residents are referred to as interns, and PGY -2 or PGY 3 residents are referred to as residents or senior residents. A total of 12 house staff participated in the study for 10 “intern days” and 10 “resident days” (some teams were observed for more than 1 day). All house staff observed were judged by the program director to be performing at a satisfactory level or better and had agreed to participate in the study. The mean age of the group was 29 years; 42% were women. Data Collection Data collection techniques have been previously described [5]. Each team member was followed from 8 AMuntil 4 PM by a lay observer with training in observational techniques. The observer recorded the type of activity, as well as the times at which it began and ended, with digital watches rounded to the nearest minute, recording consecutively. When simultaneous activities occurred (e.g., eating and charting), the activity most closely related to patient care was recorded. Activities Five major categories of activities were used: procedures, patient evaluation, communication, basics (eating, sleeping, hygiene), and miscellaneous. A modified list of the subcategories of activities for which data were collected appears in Table I. In the patient evaluation category, we distinguished patients newly admitted by the house officer the day he or she was on short call (new patients) from those who had been admitted previously and were evaluated as part of routine ward rounds (other patients). Because the intern and resident usually reviewed and discussed old records jointly, we grouped chart review and case discussion together. Time spent attending patient care conferences was recorded under other patient evaluation activities. We included in the communication category three types of chart entries-documentation of new patients’ histories and physical examination results, evaluation of the problems of patients who had been admitted previously, and writing of daily notes. Admission orders were considered part of the
TABLE I AverageTime Spentby Internsand Residentsin EachActMy Duringa ShortCallAdmittingDay ComparedWiih While on NightCall* Day (8 AM-~PM) Activity Procedures Patient Conducting evaluation new patients’ histories and physical examinations Evaluating other patients Reading charts or discussing patients Other Communication Charting Using the phone Receiving or answering pages Filling out forms Talking in person with patients families Other
Night(4 PM-~
AM)
Intern Resident (n = 10) (n = 10) (n = 10) (n = 10) Minutes(% oi total time) Intern
Resident
19 (4)
10 (2)
179 55 (35)
194 41 (38)
87 (12) l;;t(19)
21 (3) 1;; (18)
47
40
46
33
61
103
80
48
16
10
0
0
213 (42)
1;; (35)
1;;1(201
1;; (23)
z; 18
43 5
%
!
6 2
17 1
8 2
59
65
28
31
Basics
55 (11)
62 (12) 286 (40) 350 (49)
Miscellaneous Checking laboratory or radiographic results Looking for charts, radiographs, or equipment Transporting patients, laboratory specimens Other
40 (8) 20
63 (12) 24
L Total
!35 (g)
:z
58 (8)
11
12
9
12
19
2
0
15
6
6
30
21
22
506t
510t
720
720
‘his presentation of the night data is adapted with psrmfssron from 151 In four of the five mghts, the mtems ma not admrt new pattents. Ine trme shown ISme trme spent this activity on the remaining night. The total number of minutesfor these categories is based on is time averaged across all five nights. Percentagesmay not sum to 100 becauseof rounding. otal time represents slightfy more than 8 hours of duration because observers allowed house staff finish a given activity if it had been started prior to 4 PM.
documentation for new patients. Face-to-face communication with the nursing staff or other health care personnel and attending educational conferences were the principal entries under other types of communication. Only time spent in person with patients’ families was recorded in that subcategory. Data Analysis The time spent on each activity was computed and summed. The mean time spent on each activity was separated for interns and residents. The mean time and standard deviation spent per patient was computed for activities relating to obtaining histories, performing a physical examination, and charting information for newly admitted patients and for those patients previously admitted. Patients admitted the previous night by the “float team” were excluded from the separate analysis of new patients’ histories, physicals, and charting but were June 1993
The American Journal of Medlclne
Volume 94
655
TIME-MOTION
STUDY OF HOUSE STAFF DAYS / PARENTI AND LURIE
patients during the day than at night (an average of 23 minutes during the day versus 17 minutes at night for residents and 28 minutes during the day versus 24 minutes at night for interns); however, this trend did not reach statistical significance. During the day, house staff, like those at night, spent about as much time charting new patient information as they did actually performing new patient histories and physical examinations. Patient evaluation during the day was also frequently interrupted, although not as frequently as at night. For interns, the new patient history and physical was interrupted after an average of only 11 minutes, which represented two interruptions during each new evaluation. Once the initial evaluation was complete, only an average of 3.5 min/d was spent with each patient.
TABLE II AverageTime per New Patient Spent on History Taking, Physical Examination,and Charting andthe Amount ofTimeBefore These ActivitiesWere Interrupted* Day
Night Resident In = 30)
Intern (n = 16)
Resident Intern (n = 13) (n = 7) Minutes (SD)
History and physical examination of new patients
28 (12)
23 (14)
24 (8)
17 (10)
Charting new patient information
30 (17)
28 (22)
21 (10)
17 (11)
Time before interruption of the history and physical examination
11
14
Variable
7
7
included in the other measures of patient contact. The mean time spent on these activities for new patients was compared with that obtained observing internal medicine house staff on night call at the VA 1 month earlier [5]. The methods of data collection and analysis were the same for the day and night observations.
RESULTS Table I shows the average number of minutes and the percentage of total time spent during the day by interns and residents for each activity. The previously published data [5] from observations of house staff at the VA on night call are listed for comparison. As had been observed at night, surprisingly little time, an average of only 2.5 min/d, was spent with patients’ families. Time spent performing procedures during the day accounted for less time than at night but, as had been noted at night, most of these procedures did not require a physician (e.g., obtaining blood specimens and performing electrocardiograms). Other duties, such as transporting specimens or patients, or looking for charts, radiographs, or equipment, when combined with procedure time, accounted for more than an hour (an average of 76 minutes) each day spent on tasks that need not have been performed by a physician. Although house staff spent less time responding to pages during the day than they did at night, this activity, in addition to the time spent on the phone, accounted for about as much time as conducting new patient histories and physicals (on average, 47 minutes spent responding to pages and phone calls compared with 47.5 minutes in new patient evaluations). Table II shows the average time interns and residents spent on evaluation of new patients. There was a trend toward spending more time with new 666
June 1993
The American
Journal
of Medlclne
Volume
94
COMMENTS We have documented the activity of the house officers of a large categorical internal medicine residency program during short call admitting days. Two previously reported studies of the daytime activities of medicine house staff [3,4] had noted that little time was spent in patient contact and that a large proportion of time was spent charting and in communication with staff. However, those studies were reported almost 20 years earlier and had observed only interns on consecutive days when they had variable admitting responsibilities. Our study observed house staff during days when their admitting responsibilities did not vary, and senior residents as well as interns were observed. The expected new patient workload for the days we chose to observe should have allowed adequate time for house staff to evaluate patients and interact with family members. We found that the house staff we observed (interns and senior residents) spent surprisingly little time with patients even when they were rested with a manageable workload. Not only did we observe that house officers in our study spent little time with patients, we also observed that they spent little time with patients’ families. This was not an observation unique to the daytime observation or to the hospital site. We had observed similarly little time spent with patients’ families at the university and the county hospitals during our study of night call. It is possible that family members were not available for interaction with house staff and since we did not include time spent on the phone with patients’ families or when families were simultaneously seen with the patient, it is possible that we underestimated the total time spent in communication with the family. If we have not substantially underestimated the time spent with families, this is a particularly disturbing obser-
TIME-MOTION STUDY OF HOUSE STAFF DAYS / PARENTI AND LURIE
vation given a population of patients in whom the medical care of a significant number might be improved with family input. Once the initial evaluation was complete, house staff in our study spent an average of less than 5 minutes per patient each day, similar to what Wray et al [7] noted (an average of 4.6 minutes per patient each day) in their observations of internal medicine work rounds in 1986. Payson et al [3], following interns at Yale University in 1959, and Gillanders and Heiman [4], observing interns at the University of California, San Francisco, in 1969, had also observed that much less time was spent with patients as their hospital course progressed. Since there is no “gold standard” for the appropriate amount of time a physician should spend evaluating a given patient, we cannot conclude that the amount of time house staff in our study spent with patients was inadequate, only that it was less than we predicted and hoped it would be. Why house staff chose to spend so little time with patients or their families is not clear. We doubt that it was because of inadequate time available or because of a fear of a large number of admissions. Teams were at risk for a maximum of only three new patients and actually admitted an average of only two new patients each day. It is possible, however, that the unpredictable nature of the work associated with caring for acutely ill patients (a sense of “impending disaster around every corner”) creates an underlying pressure to gather data as quickly as possible even when the expected workload should be manageable. Nerenz et al [8] found that interns observed during the same academic year as this study spent a similar amount of time with new patient evaluations as those in our study. They also noted, as did we, a trend toward less time being spent with patient evaluation later in the day. We suspect that even less time is spent with patients at night because house staff feel even more pressure to be “efficient” at night to prepare themselves for unexpected work, to preserve time to spend in other activities such as sleep, or simply because they are adversely affected by fatigue as the workday progresses. Other explanations are also possible. For example, patients admitted at night may have characteristics different from those admitted during the day, which could explain this trend. In a study of work rounds at Baylor College of Medicine, Wray and colleagues [7] suggested that patient characteristics (such as do-not-resuscitate status, being nursing home candidates) may influence the amount of time house staff spend with patients. Finally, it is possible that house staff choose to avoid patients or their families, either because they are not comfortable spending time with them or
because it is not a personally satisfying activity. This may, in part, be due to the lack of emphasis given to teaching house officers about communication and the physician-patient relationship. It may also be that house staff are uncomfortable obtaining or do not value the clinical information that can be obtained from the patient. The availability of high-technology diagnostic tests may influence house staff to rely on means other than speaking with and examining their patients to formulate a plan of care. Two recent studies, one demonstrating poor physical diagnosis skills of house staff [9], and the other, poor clinical performance evaluation skills of faculty [lo], suggest that house staff may be deficient physical diagnosticians and that training programs may lack reliable methods for detecting these deficiencies. As suggested by a study designed to detect physical diagnosis errors and improve house staff clinical skills [ 111, poor physical diagnosis skills of house staff may not only have important implications for patient care but also for house staff satisfaction as well. That study noted improved house staff satisfaction following a program designed to enhance physical diagnosis skills. The house staff we observed spent a significant amount of time each day charting patient information, performing nonphysician duties (“scut” work), and responding to pages. This latter task was particularly frustrating because it represented a frequent source of interruption. Interns’ new patient evaluations were interrupted after an average of only 11 minutes, often by a page. Some of the time house staff spent in these activities could be eliminated by decreasing disruptions from paging as suggested by Katz and Schroeder [12], allowing dictation of patient information, and hiring personnel to perform nonphysician duties. Some limitations of the current study deserve mention. There is always a concern that observation alone may affect what house staff do-that is, the Hawthorne effect. We believe this is unlikely since residents frequently commented they were too busy to be affected by the observer, and we would have expected that, if they had been affected by the observer, more time would have been spent in patient evaluation or with patients’ families. Simultaneous activities were not recorded, and house staff were observed for only part of their workday (until 4 PM); therefore, time spent in some activities may have been underestimated. However, they had previously been observed from 4 PM until 8 AM during our study of night call [5] where the observations were similar. A larger sample size or observation of house staff at another time in the academic year may have resulted in different findings. Finally, this study of house staff days was performed June 1693
The American Journal of Medicine
Volume 94
657
TIME-MOTION
STUDY OF HOUSE STAFF DAYS / PARENTI AND LURIE
at only one hospital where a night float call system was in operation; therefore, the results may not be generalizable. Although these limitations should be considered when interpreting our results, many of the observations noted in this study are similar to those made by other investigators in different settings. In summary, during the day, patient evaluation was frequently interrupted, little time was spent with patients or patients’ families, and as much time was spent charting as in patient evaluation. Although the ideal distribution of house staff time is unknown, studies are now beginning to assess the effect of changing house staff work schedules on patient and resident outcome [6]. These studies have, however, primarily focused on the effects of decreasing sleep deprivation or altering workload. Few data are available that directly measure the effect of redistributing house staff activities within a given work schedule on patients or residents. Before it can be determined whether altering house staff activities would result in improved patient care or satisfaction, or in improved resident education or satisfaction, the motivation behind house staff choices regarding time distribution needs to be determined. For example, if house staff lack skill in physical diagnosis and therefore choose not to spend time with patients because they are more proficient at ordering and interpreting tests or documenting rather than obtaining clinical information, then, as suggested in a recent editorial [13] , clinical science needs to clearly define the complementary roles of clinical diagnosis and technologic assessment and encourage the development of outstanding clinical skills. However, if there are other reasons, such as lack of role modeling, that diminish the perceived importance of house staff-patient contact or if the current state of residency training creates an environment that does not encourage interaction with patients or their families, it is important that we determine if indeed “too little” time is being spent with patients and the effect of inade-
666
June 1993
The American
Journal
of Medlclne
Volume
94
quate personal contact on the patient or the physician in training. As residency programs attempt to balance service and education missions, it is critical to carefully evaluate why house staff choose to spend time as they do and the impact those choices have on patient care and house staff well-being.
ACKNOWLEDGMENT We are indebted to the house officers who made this study possible: to Thomas Ferris, Charles Moldow, and Robert Petzel for their support; and to Pamela Nesbitt for assistance in the preparation
of the manuscript.
REFERENCES 1.Graettinger JS. Internal medicine in the National Resident Matching Program 1987; the ides of March. Ann Intern Med 1988; 108: 101-15. 2. New York State Department Advisory Committee
of Health. Report of the New York State Ad Hoc
on Emergency
dent Working Conditions.
Services Regarding Supervision and Resi-
New York: New York State Department
of Health,
1987. 3. Payson HE, Gaenslen EC Jr, Stargardter
FL. Time study of an internship on a
university medical service. N Engi J Med 1961;264:439X3.
4. Gillanders W, Heiman M. Time study comparisons of 3 intern programs. J Med Educ 1971;46: 142-9. 5. Lurie N, Rank 8. Parenti C, Woolley T, Snoke W. How do house officers spend their nights? A time study of internal medicine house staff on call. N Engl J Med 1989; 320: 1673-7. 6. Gottlieb DJ, Parenti CM, Peterson CA, Lofgren RP. Effect of a change in house staff work schedule on resource utilization and patient care. Arch Intern Med
1991;151:2065-70. 7.Wray NP, Friedland JA, Ashton CM, Scheurich J. Zollo Al. Characteristics house staff work rounds on two academic Educ 1986,61:
of
general medicine services. J Med
893-900.
6. Nerenz D, Rosman H. Newcomb internal medicine.
C, et al. The on-call experience
of interns in
Arch intern Med 1990; 150: 2294-7.
9. St. Clair EW, Oddone
EZ, Waugh RA, Corey GR. Feussner JR. Assessing
housestaff diagnostic skills using a cardiology patient simulator. Ann Intern Med 1992; 117: 751-6.
10.Noel
GL, Herbers JE, Caplow MP, Cooper GS, Pangaro LN, Harvey J. How
well do internal medicine faculty members
evaluate
the clinical skills of resi-
dents? Ann Intern Med 1992: 117:757-65.
11.Wray NP, Friedland JA. Detection and correction
of house staff error in
physical diagnosis. JAMA 1983; 249: 1035-7. 12. Katz MH, Schroeder SA. The sounds of the hospital.
Paging patterns
in three
teaching hospitals. N Engl J Med 1988; 319: 1585-9. 13. Fletcher RH, Fletcher SW. Has medicine outgrown physical diagnosis? Ann Intern Med 1992;117:766-7.