The cost of training third-year primary care medical students in an ambulatory setting.

The role of the primary care physician (PCP) has undergone a significant change. These physicians are being called upon to play an enlarged role in the coordination of medical services provided to patients, resulting in a tremendous increase in the demand for PCPs (Seifer 1996). Attempts to address the imbalance between specialists and general practitioners have also contributed to this increased demand (Huang 1999). The proper venue for the training of these physicians has been extensively debated. Classically, medical education has been carried out in the teaching hospital ward and outpatient clinic. However, there has been a substantial increase in the training of medical students in ambulatory settings. This increase has been driven by several factors.

Changes in the health care environment have led to a narrower, more chronic, case mix in teaching hospitals. This skewed patient population is not representative of the patient problems that the majority of physicians will treat after they graduate. Additionally, with the increasingly sophisticated diagnosis and treatment methodologies that are now available on an outpatient basis, the ambulatory care setting has become a major site for making important patient care decisions. Finally, there is increased emphasis on preventive care and the care of chronic diseases, which may be learned best by working with ambulatory patients (Delbanco & Calkins 1988).

In order to encourage students to choose careers in primary care areas and to provide the skills they need, medical schools have developed programs which include training at alternative ambulatory care sites (Bazell & Kahn 2001). These include private practices, health maintenance organizations (HMOs), community health centers, and reorganized teaching hospital ambulatory practices.

Few academic health centers make payments for clinical education rotations. (Holder 1988; Kumar et al. 1999). Furthermore, the presence of medical students can lengthen the workday of the teaching physician and reduce the number of patients seen per day (Vinson et al., 1997). These facts have been cited as the reason for the limited participation of alternative ambulatory care sites in teaching programs (Dorsey 1973). While preceptoring physicians are largely motivated to teaching due to intangible rewards such as the love of teaching, a desire to "give back" to the profession, or increased enjoyment of the practice of medicine (Grayson et al. 1998; Kollisch et al., 1997), it has been noted that neither an HMO with fixed income and fixed salaries, nor a private group can tolerate inefficiency in patient care and physician usage (Olness 1975).

The increase in ambulatory medical education requires that a number of issues be addressed. These include a determination of the cost of ambulatory teaching, an examination of how this teaching can be conducted more efficiently, and exploration of the implications for financing medical education if more teaching occurs in ambulatory settings (Delbanco & Calkins 1988). The objective of this study was to examine this first issue, and provide a basis for future exploration of the other issues.

Previous Studies

A substantial amount of research has been conducted examining the impact of training medical students in ambulatory care settings. Employing inconsistent methodologies, they report contradictory findings. Some studies (Delbanco & Calkins 1988; Garg et al. 1991; Gavett & Mushlin 1986; Kirz & Larsen 1986; Kosecoff et al. 1987; Lindenmuth et al. 1972; Paulson et al. 1979; Pawlson et al. 1980; Usatine et al. 2000; Vinson & Paden 1994; Wollstadt et al. 1979) indicate a reduction in physician productivity when undergraduate medical students are present, while others indicate no effect (Frank et al. 1997; Pawlson et al. 1980; Kearl & Mainous 1993), or even an increase (Lindenmuth et al. 1972) in physician productivity (Isaacs & Madoff 1984). The ability to generalize from these studies is also limited by differences in the objectives of the studies and, thus, differences in relevant costs (Garg et al. 1991). Some studies have focused solely on the cost resulting from the decline in the preceptor physician's efficiency. Others have included the estimated impact on indirect costs, such as occupancy costs. Finally, some studies have included programmatic costs and an allocation of university and school overhead. Additional limitations of the existing literature include its emphasis on graduate medical training and its emphasis on training in a traditional hospital setting. It has been noted that the bulk of the literature on ambulatory care education focuses on graduate education, with graduate and undergraduate education often enmeshed as if they were identical activities (Yonke & Foley 1991). A description of the findings of the literature most relevant to the current study follows.

Based on interviews with faculty physicians at a university-operated HMO, Pawlson et al. (1979) reported that the physicians assigned an average of 22 percent of their time in clinical sessions to teaching third-year medical students, or an average of 31 minutes per student per 3.5-hour session. The lack of independent corroboration of this survey data leads one to question the accuracy of this finding.

Lindenmuth et al. (1972) examined the effect of third-year medical students on the number of patients examined per 3.5-hour session by two physicians in an academic primary care clinic. The students' role varied from merely observing the physician to seeing patients by themselves and presenting their findings with an assessment and plan to the physician. Lindenmuth et al. report that the presence of students significantly increased the number of patients seen, especially at the end of the third year when the students had some clinical experience. However, as Pawlson et al. (1980) point out, the Lindenmuth study did not control for changes in patient mix (the preceptoring physician tended to see walk-ins). Lindenmuth et al. (1972) note in their findings that the presence of students did not significantly increase the length of the physicians' day or add to patient dissatisfaction.

Pawlson et al. (1980) studied the effect of the same third-year primary care clerkship as Lindenmuth et al. (1972). They found no effect on productivity due to the presence of the students. However, they note that in this university HMO students were present for nearly every session for 48 of the 52 weeks of the year. They hypothesize that the baseline productivity for this practice may have been at a level that "assumed" the presence of students, and note the lack of financial incentives to increase productivity during the brief periods when students were not present.

Pawlson et al. (1980) also studied the effect of the presence of third-year students in a private office setting. The student role was limited to observation in some offices. A reduction of 1.2 visits per day occurred when students were present. No added space or material costs were identified. Delbanco & Calkins (1988) note that in both settings, however, it is not clear whether faculty spent less time with patients when teaching. If so, no adjustment was made for this factor in calculating teaching costs.

Kirz & Larsen (1986) report a decline in productivity of 1.0 patient visit per half-day session when medical students were present, based on an analysis of visit logs. In addition, the primary care physicians reported that they spent, on average, 43.8 minutes per half day in direct teaching activities. The interpretation of these results is limited by the failure to take into account the level of the students, which ranged from first to fourth year.

Kearl & Mainous (1993) defined physicians' productivity as the number of patients seen per half-day session. They report no impact on the productivity of family practice faculty at an academic ambulatory-care center from the presence of third-year medical students. Given that the study period encompassed the period March through June, their results are consistent with the idea that by the end of their third year, medical students are able to perform enough services to offset the reduction in the preceptor physician's productivity due to teaching. Kearl and Mainous note that the generalizability of their findings is limited to academic health care facilities. They also note that the staff that schedule patients' appointments were unaware of the daily teaching status of the PCPs. A strength of the Kearl and Mainous study is their use of a paired-sample t-test. Most other studies employed independent sample designs, which could obscure differences in productivity due to individual practice style.

Garg et al. (1991) conclude that the productivity of the teaching faculty in their study was 30-40 percent lower than the productivity of nonteaching physicians nationally and regionally. However, the design of this study is flawed in that it compares physician productivity to published averages. The reported results could thus be attributable to individual practice style or inefficiencies inherent in the subject practice.

AMC PROGRAM

As part of its undergraduate curriculum, Albany Medical College requires its students to participate in ambulatory-based (third year) and hospital-based (fourth year) clerkships. Such clerkships have been shown to be effective in increasing the proportion of students choosing family practice as a specialty (Campos-Outcalt & Senf 1999; Musham & Chessman 1994). The third-year clerkships take place in the practices of a network of affiliated health care organizations. This study examined the impact of the third-year students on the practices at four (two HMO and two non-HMO) health centers.

No change was made in the scheduling of patients when students were present at any of the health centers involved in the study. A common comment from the faculty preceptors at all of the sites was that a change in scheduling should be made in order to allow additional time for teaching activities. This lack of additional time appears not to be unique to this program, as other research studies have found that teaching by community-based preceptors tended to decrease with increased patient care workload (Masters & Nester 2001).

METHOD

Physician productivity was measured in three ways. In order to facilitate comparability with prior studies, the number of patients seen per session (and per hour) was measured, as well as the length of patient visits. The effect of the presence of students and whether a site was affiliates with an HMO on the number of encounters per hour and per session was analyzed using an ANOVA model.

Additionally, physician productivity was measured utilizing an activity-based approach. This methodology involves looking at a patient visit as a process that consists of a sequence of tasks. These tasks and their relationship to the overall process were identified at one of the HMO sites, based on a genetic patient visit. These tasks were aggregated into the activities listed in Figure 1. The length of time required to perform each of these activities within a patient visit was measured. These times were then analyzed using multiple regression.

FIGURE 1

LIST OF ACTIVITIES PERFORMED BY PCPs.

Office Encounter
  1. Brief student on patient
  2. Review chart for visit
  3. In-room encounter
  4. Initiate request for ancillary testing; complete referral form
  5. Write prescription
  6. Document visit in chart
  7. Discuss case with student
  8. Wait for nursing assistance

Other Activities
  1. Doctor to doctor consults
  2. Urgent referrals / appointment scheduling
  3. Requesting chart
  4. Reviewing diagnostic test results and consults
  5. Review medical references / literature
  6. Triage advice for nurses
  7. Ordering referrals / diagnostic tests
  8. Processing no-shows
  9. Complete insurance & disability forms
 10. Complete administrative forms
 11. Consultation/assisting NPs, PAs
 12. Telephone calls
 13. Breaks

Breaking down the patient visit process into its constituent activities has several advantages. First, it enables us to more accurately measure the time actually spent by a physician with patients and with students. Secondly it makes it possible to identify which activities of the practice are affected by the presence of medical students, enabling us to better understand the effect of students on physician productivity.

RESULTS

The cost of training primary care physicians was measured at the four health centers indicated above. Data was collected for 1,069 patient encounters conducted by 16 physicians in 121 half-day sessions (see Table 1).

The most aggregated indicator of physician productivity used was the number of encounters per hour, equal to the number of patients seen in a session divided by the length, in hours, of the session. Based on this measure, the presence of students did not have a significant (F =1.121, p < 0.292) impact on productivity. However, there was a significant (F = 4.025, p < 0.047) difference between the number of encounters per hour at the HMO (2.26) and non-HMO (2.54) sites.

There was a significant difference (F = 6.785, p < 0.01) between the average number of patients seen per session with (7.88) and without (9.66) students. This was due to the longer average length of the sessions without students.

A second measure of productivity was based on the length of patient encounters, calculated as the difference between when a patient first and last saw a physician. Overall, the presence of students significantly (F = 15.438, p < 0.001) increased the average length of a patient encounter from 18.6 minutes to 24.0 minutes. (Differences between these figures and those implied by the previous measures of productivity are due to factors such as physician breaks and the overlapping of patient visits.)

In order to make some adjustment for case severity, encounters were classified as being for a new or an established patient, and as a routine or urgent visit. For all encounters, no significant difference existed between routine and urgent care visits (F = 0.466, p < 0.495). Physicians spent significantly (F = 5.775, p < 0.016) more time on a new patient encounter (29.12 min.) than for an established patient (19.83 min.). Also, as might be expected, the length of office visit was significantly longer (F = 38.301, p < 0.001) when tests were prescribed (30.28) than when they were not (17.89). A significant (F = 9.030, p < 0.003) interaction was observed between the testing variable and routine-urgent care variables, with urgent care visit taking longer (18.75) than routine visits (12.93) when no tests were performed, but taking shorter (29.39) than routine visits (34.50) when tests were performed.

Patient visits were significantly shorter (F = 7.456, p < 0.006) for HMO visits (16.13) than for non-HMO visits (22.53). A significant interaction (F = 13.436, p < 0.001) between the type of site and the presence of students were observed, with the presence of students resulting in an increase in the average visit length at a non-HMO site from 19.98 to 32.30 minutes, while increasing the average visit length at the HMO sites only from 15.55 to 16.96 minutes. As indicated in Table 2, this effect varied considerably among physicians, with no significant difference observed for any of the HMO physicians and significant differences being observed for most of the non-HMO physicians.

The presence of students did not affect the rate at which ancillary tests were prescribed (F = 0.199, p < 0.655). However, there was a significant interaction (F = 5.038, p < 0.025) between the presence of students and ancillary testing on the length of patient visits. For visits without a student present, performing an ancillary test resulted, on average, in an incremental increase in encounter length of 9.65 minutes, versus an increase of 18.67 minutes with a student.

The third measure of productivity focused on the activities performed by the physicians and the time needed to perform those activities. This data was collected by fourth year medical students at the HMO sites and by nursing staff at the non-HMO sites. One problem encountered, especially with the data collected at the non-HMO sites, was that of "joint product costs" (Gavett & Mushlin 1986). For example, a physician might document a visit in the chart while performing the in-room encounter. In order to avoid arbitrary time (and cost) allocations, all of the activities involved in the office encounter were combined for data analysis purposes, with the exception of those related to direct student interaction (numbers 1 and 7) and the one related to waiting for nursing assistance (number 8). The presence of students did not significantly (F = 2.307, p < 0.135) affect the length of this aggregated in-office activity.

Further analysis of the data was hampered by the frequent failure of the data collectors at the non-HMO sites to indicate on the data collection forms the time physicians ended their day. Use of the data from these sessions could possibly have failed to accurately capture any "after hours" paperwork that needed to be done by preceptoring physicians due to the presence of students. In order to eliminate this potential source of bias, these sessions were eliminated from further consideration.

Table 3 presents a comparison of the average time spent per session on each activity with and without students. Contrary to expectation, the sessions with students were shorter on average than those without students, although the difference is statistically insignificant. Statistical significance was determined using a multiple regression model where the dependent variable was the length of time spend on a given activity and the dependent variables were the presence of a student and the percentage of visits in a session that were routine. The latter variable was included as a proxy for case mix/acuity. The t statistics and levels of significance presented in Table 3 are for the "presence of student" independent variable. Except for the two activities directly involving students, there were no significant differences (at p < 0.05) for any of the remaining activities.

Another possible confounding variable in the above analysis is the number of patients seen per session. In order to adjust for this factor, the time to perform each activity was adjusted for the number of patients seen for each session. Table 4 summarizes the average time, per patient, required to perform each activity. Again, with the same two exceptions regarding student-related activities, there were no significant differences. Examination of the difference in mean time between sessions with and without students indicates that the presence of students does not decrease physician productivity, and that teaching time spent with students is more than offset by a reduction in the actual in-office encounter and in time spent waiting for nursing assistance. An examination of the raw data indicated that physicians frequently used this waiting time to discuss cases with students. This more efficient use of preceptor physician's time was observed in another study (Usatine et al., 2000), which found that time savings from student charting allowed preceptors to teach and care for patients without losing valuable practice time.

COST IMPLICATIONS

The objective of this study was to determine the incremental cost to the study sites resulting from the presence of the medical students. This cost includes the cost of establishing the programs at the various sites. Inasmuch as established programs were used for this study, this data was not available.

Additional costs include the cost of administering the program and the cost of conferences with the students. The non-HMO sites estimated that the administrative duties amounted to approximately 6 hours per rotation, or approximately $120. Student conferences require approximately 4 hours per week, including preparation time. The cost of these conferences over the four-week rotation, at an estimated cost of $80 per hour, is $1,080. Additionally, student evaluations during the rotation require 2 hours per student, at a cost of $160. The total of these costs for a three-student program is $1,680, or $560 per student.

The presence of students decreased the productivity of the preceptor physicians only at the non-HMO sites. Productivity at these sites was reduced by 0.31 patients per hour. The cost for the additional physician time is estimated to be $26 per session (0.31 patients/hour x 4 hours/day x $50/hour / 2.4 patients/hour), or $517 per student per rotation.

The final element of cost is the impact of the student on other costs at the health center. This study found no significant difference in the rate of ancillary testing. (This is consistent with the results of other studies [Health & Beatty 1998].) In addition, the presence of the students did not result in the need for additional examination space. Aside from insignificant use of supplies, the presence of students did not increase non-physician related health center costs.

CONCLUSION AND DISCUSSION

The results of this study yielded mixed results. At the HMO sites, the presence of students did not affect the length of patient visits, nor did it have a significant impact on preceptor physician productivity. Activity analysis indicated that the physicians were able to efficiently use the third year students in the patient visit process, and were able to be more efficient by utilizing what would be unproductive time to teach students.

At the non-HMO sites, patient visits were significantly longer when students were present, and physician productivity decreased by 0.31 patients per hour. This resulted in an increase in the physicians' workday of 1.0 hour (0.31 patients/hour x 8 hours/day / 2.4 patients/hour). This result is consistent with statements by physicians participating in the study, who indicated that the presence of students resulted in their staying late to complete necessary paperwork. It is also consistent with non-HMO health centers administrators' estimates that physicians participating in the training program spent 0.5 to 2 hours longer per session when students were present, depending on the amount of teaching taking place. While physicians participating in the teaching program were not compensated for this time, there is an associated "opportunity cost" of $517 per student per rotation.

It should be noted that there is a tremendous difference between the ability level of students starting their third year and those at the end of that year. Inasmuch as much of the data collected for this study was collected near the end of the students' third year, the results may not be representative of those that would be achieved with students just beginning their third year.

A second caveat is that no site adjusted its patient scheduling for the presence of teaching. The preceptor physicians simply did not have the luxury of spending unlimited time teaching students during the day. An oft-heard comment from the physicians was that some allowance needed to be made in the scheduling when the students were present. While they were able to fit in some teaching during what would otherwise be unproductive time, the adequacy of this time is an issue that may need to be explored further.

A final caveat is that for the program studied here, only one or two students were present at a given site at any given time. A significant increase in the scope of the training activities could lead to significant additional costs being incurred.

The productivity of the physicians in this study was a function of three factors: the practice site, the individual physician, and the presence of students. The presence of students affected productivity only at the non-HMO sites. This appears to be due to the fact that the non-HMO sites, being in a relatively rural area that was experiencing a shortage of physicians, regarded their participation in the training program as having potential physician recruitment benefits. They thus designed a program they regarded as being of premier quality. The HMO sites, on the other hand, did not face a similar situation and did not offer a program with similar features.

Previous studies have found that physicians' practice patterns, rather than clinic or patient characteristics, may account for most of the variation in physician productivity (Smith et al. 1995). In a similar way, the data from this study indicates that physicians' practice patterns affected their efficiency more than the presence of students. The data in Table 2 indicate that the average encounter time varied among physicians by a factor of 5.1 with students present and by a factor of 4.1 without students, while the presence of students increased the average encounter time by only 29 percent. Given that differences among physicians contributed more to variability in efficiency than the presence of students, a fruitful avenue for future research would be the exploration of the individual physician characteristics that affect productivity, possibly employing the methodology used in this study.

The effect of the third factor affecting physician productivity, the presence of students, appears dependent upon the other two factors. First the health care site must have structured the training program and its objectives in such a way as to enable the program to affect productivity, and secondly the individual physicians must be willing to accept the cost of reduced productivity, increase in session length. It is clear, then, that when evaluating whether to participate in such training programs, both the institution and its physicians need to have a clear understanding of the time commitment expected of the participating physicians and its resulting cost.

TABLE 1

NUMBER OF ENCOUNTERS AND PARTICIPATING PHYSICIANS BY HEALTH CENTER

                                   No. Physicians
    Site         No. Encounters    Participating in
                                        Study

HMO No. 1               36                1
HMO No. 2              362                7
NON-HMO No. 1          304                4
NON-HMO No. 2          367                4
Total                1,069               16
TABLE 2

EFFECT OF PRESENCE OF STUDENTS ON AVERAGE ENCOUNTER TIME

                                   Encounters
                                  With Student

                              Avg. Time      No.
    Site         Physician     (min.)     Encounters

HMO No. 1            1         27.133         20
HMO No. 2            1         20.134         35
                     2         14.547         24
                     3         14.769         22
                     4          8.282         33
                     5         21.317         10
                     6         12.056          3
                     7         19.663         18
NON-HMO No. 1        1         30.560         14
                     2         21.599         13
                     3         27.428         13
                     4           n.a.          0
NON-HMO No. 2        1         38.276         31
                     2         42.315         20
                     3         26.097         13
                     4         30.068         35
All sites                      23.973        304

                                  Encounters
                                Without Student

                              Avg. Time       No.
    Site         Physician     (min.)      Encounters

HMO No. 1            1         24.734          16
HMO No. 2            1         14.757          41
                     2         13.580          67
                     3         16.439          61
                     4         12.336          20
                     5         17.668          18
                     6         15.444           9
                     7          6.000           1
NON-HMO No. 1        1         19.120         120
                     2         19.558          49
                     3         17.094          61
                     4         19.280          34
NON-HMO No. 2        1         22.925          58
                     2         17.919          88
                     3         24.303          59
                     4         21.254          63
All sites                      18.632         765

                              Difference

                              In Avg.       Sig.
    Site         Physician     Time         p <

HMO No. 1            1         2.399        0.804
HMO No. 2            1         5.377        0.212
                     2         0.967        0.728
                     3        -1.670        0.592
                     4        -4.054        0.227
                     5         3.649        0.592
                     6        -3.389        0.575
                     7        13.663        0.151
NON-HMO No. 1        1        11.439        0.001
                     2         2.041        0.520
                     3        10.335        0.001
                     4          n.a.         n.a.
NON-HMO No. 2        1        15.351        0.002
                     2        24.396        0.001
                     3         1.794        0.755
                     4         8.814        0.007
All sites                      5.340        0.001
TABLE 3

COMPARISON OF AVERAGE TIME (PER SESSION) TO PERFORM ACTIVITIES

                                              Mean Time (min.)

                                        Without      With     Diffe-
       Acitivity                        Student    Student    rence

All activities                          249.184    221.145    -28.039
Brief student on patient                  n.a.       1.908      1.908
Review chart for visit
In-room encounter
Initiate request for ancillary          155.655    129.740    -25.915
  testing; complete referral form
Write prescription
Document visit in chart
Discuss case with student                 n.a.      13.244     13.244
Wait for nursing assistance              23.413      3.559    -19.854
Doctor to doctor consults                 5.779      7.677      1.898
Urgent referrals/appointment              0.069      0.517      0.448
  scheduling
Requesting chart                          0.321      0.190     -0.131
Reviewing diagnostic test results         6.542      5.729     -0.813
  and consults
Review medical references/literature      2.388      0.357     -2.031
Triage advice for nurses                  0.164      0.030     -0.134
Ordering referrals/diagnostic tests       0.343      0.273     -0.070
Complete administrative forms            22.688     23.134      0.446
Consultation/assisting NPs, PAs           8.432      8.153     -0.279
Telephone calls                          11.320     12.251      0.931
Breaks                                   12.070     14.384      2.314

                                                    Sig.
       Activity                          t-Stat.    p <

All activities                           -0.422     0.256
Brief student on patient                 -1.813     0.013
Review chart for visit
In-room encounter
Initiate request for ancillary           -0.503     0.228
  testing; complete referral form
Write prescription
Document visit in chart
Discuss case with student                -1.254     0.001
Wait for nursing assistance              -0.003     0.427
Doctor to doctor consults                 0.189     0.696
Urgent referrals/appointment             -2.973     0.231
  scheduling
Requesting chart                         .0.367     0.810
Reviewing diagnostic test results        -3.911     0.774
  and consults
Review medical references/literature      0.027     0.226
Triage advice for nurses                 -2.944     0.516
Ordering referrals/diagnostic tests        n.a.      n.a.
Complete administrative forms             0.602     0.997
Consultation/assisting NPs, PAs           1.752     0.730
Telephone calls                           0.066     0.747
Breaks                                    0.425     0.807
TABLE 4

COMPARISON OF AVERAGE TIME (PER PATIENT) TO PERFORM ACTVITIES

                                             Mean Time (min.)

                                       Without    With
       Activity                        Student   Student   Difference

All activities                         34.520    27.767      -6.753
Brief student on patient                n.a.      0.288       0.288
Review chart for visit
In-room encounter
Initiate request for ancillary         20.095    16.085      -4.010
  testing; complete referral form
Write prescription
Document visit in chart
Discuss case with student               n.a.      1.994       1.994
Wait for nursing assistance             4.642     0.466      -4.176
Doctor to doctor consults               0.819     0.960       0.141
Urgent referrals/appointment            0.008     0.069       0.061
  scheduling
Requesting chart                        0.029     0.032       0.003
Reviewing diagnostic test results       1.221     0.831      -0.390
  and consults
Review medical references/literature    0.273     0.042      -0.231
Triage advice for nurses                0.016     0.010      -0.006
Ordering referrals/diagnostic tests     0.092     0.068      -0.024
Complete administrative forms           3.255     2.580      -0.675
Consultation/assisting NPs, PAs         1.191     1.041      -0.150
Telephone calls                         1.457     1.658       0.201
Breaks                                  1.423     1.643       0.220

                                                   Sig.
       Activity                         t-Stat.    p <

All activities                          -1.042     0.305
Brief student on patient                 2.910     0.006
Review chart for visit
In-room encounter
Initiate request for ancillary          -1.189     0.243
  testing; complete referral form
Write prescription
Document visit in chart
Discuss case with student                5.605     0.001
Wait for nursing assistance             -0.846     0.404
Doctor to doctor consults               -0.269     0.789
Urgent referrals/appointment             1.303     0.202
  scheduling
Requesting chart                         0.482     0.633
Reviewing diagnostic test results        0.516     0.609
  and consults
Review medical references/literature    -1.278     0.210
Triage advice for nurses                -0.657     0.516
Ordering referrals/diagnostic tests       n.a.      n.a.
Complete administrative forms           -0.603     0.550
Consultation/assisting NPs, PAs         -0.640     0.526
Telephone calls                          0.274     0.786
Breaks                                  -0.006     0.995

REFERENCES

Bazell, C. & R. Kahn (2001) "From the Primary Care Organizations Consortium's Proposal to the Interdiscipinary Generalist Curriculum Project," Academic Medicine (76):S13-18.

Campos-Outcalt, D. & J. Senf (1999) "A Longitudinal, National Study of the Effect of Implementing a Required Third-Year Family Practice Clerkship or a Department of Family Medicine on the Selection of Family Medicine by Medical Students," Academic Medicine (74):1016-20.

Delbanco, T.L. & D.R. Calkins (1988) "The Costs and Financing of Ambulatory Medical Education," Journal of General Internal Medicine (3):S34-S43.

Dorsey, J.L. (1973) "The Prepaid Group Practice Plan in the Education of Future Physicians: Initial Efforts at the Harvard Community Health Plan," Medical Care (XI):2-19.

Frank, S.H., K.C. Stange, D. Langa & M. Workings (1997) "Direct Observation of Community-Based Ambulatory Encounters Involving Medical Students," Journal of the American Medical Association (278):712-716.

Garg, M., J. Boero, R.G. Christiansen & C.G. Booher (1991) "Primary Care Teaching Physicians' Losses of Productivity and Revenue at Three Ambulatory-Care Centers," Academic Medicine (66):348-353.

Gavett, J. & A. Mushlin (1986) "Calculating the Costs of Training in Primary Care," Medical Care (24):301-312.

Grayson, M.S., M. Klein, J. Lugo & P. Visintainer (1998) "Benefits and Costs to Community-Based Physicians Teaching Primary Care to Medical Students," Journal of General Internal Medicine (13):485-8.

Heath, J.M. & P.G. Beatty (1998) "Does Teaching Medical Students in the Office Affect the Way Physicians Complete Patient-Encounter Forms?" Academic Medicine (73):439-41.

Holder, L. (1988) "Paying for Clinical Education: Fact of Fiction?" Journal of Allied Health (17):221-229.

Huang, K. (1999) "Graduate Medical Education: The Federal Government's Opportunity to Share the Nation's Physician Workforce," Yale Journal on Regulation (16):175-210.

Isaacs, J.C. & M.A. Madoff (1984) "Undergraduate Medical Education in Prepaid Health Care Plan Settings," Journal of Medical Education (59):615-24.

Keari, G.W. & A.G. Mainous, II (1993) "Physicians' Productivity and Teaching Responsibilities," Academic Medicine (68):166-7.

Kirz, H. & C. Larsen (1986) "Costs and Benefits of Medical Student Training to a Health Maintenance Organization," Journal of the American Medical Association (2):734-739.

Kollisch, D.O., P.Y. Frasier, L. Slatt & M. Storaasli (1997) "Community preceptors' views of a required third-year family medicine clerkship," Archives of Family Medicine (6):25-8.

Kosecoff, J., R. Brook, A. Fink, C. Kamberg, C. Roth, G. Goldberg, L. Linn, V. Clark, J. Newhouse & T. Delbanco (1987) "Providing Primary General Medical Care in University Hospitals: Efficiency and Cost," Annals of Internal Medicine (107):399-405.

Kumar A., D. Loomba, R.Y. Rahangdale & D.J. Kallen (1999) Journal of General Internal Medicine (14):370-2.

Lindenmuth, N., A. Stone & M. Donaldson (1972) "The Effect of Third-Year Clinical Clerks on Physician Productivity in a Primary Care Practice," Journal of Medical Education (53):357-359.

Masters, P.A. & C. Nester (2001) "A Study of Primary Care Teaching Comparing Academic and Community-Based Settings," Journal of General Internal Medicine (16):9-13.

Musham, C. & A. Chessman (1994) "Changes in Medical Students' Perceptions of Family Practice Resulting from a Required Clerkship," Family Medicine (26):500-503.

Olness, K. (1975) "Teaching Ambulatory Pediatrics in a University HMO," Journal of Medical Education (50):687-689.

Pawlson, L., S. Schroeder & M. Donaldson (1979) "Medical Student Instructional Costs in a Primary Care Clerkship," Journal of Medical Education (54):551-555.

Pawlson, L., R. Watkins & M. Donaldson (1980) "The Cost of Medical Student Instruction in the Practice Setting," Journal of Family Practice (10):847-852.

Seifer, S.D. (1996) "Changes in Marketplace Demand for Physicians: A Study of Medical Journal Recruitment Advertising," Journal of the American Medical Association (276):696-97.

Smith, D., D. Martin, C. Langefeld, M. Miller & J. Freedman (1995) "Primary Care Physician Productivity: The Physician Factor," Journal of General Internal Medicine (10):495-503.

Usatine, R.P., P.T. Tremoulet & D. Irby (2000) "Time-Efficient Preceptors in Ambulatory Care Settings." Academic Medicine (75):639-642.

Vinson, D.C. & C. Paden (1994) "The Effect of Teaching Medical Students on Private Practitioners' Workloads," Academic Medicine (3):237-8.

Vinson, D.C., C. Paden, A. Devera-Sales, B. Marshall & E.C. Waters (1997) "Teaching Medical Students in Community-Based Practices: A National Survey of Generalist Physicians," Journal of Family Practice (45):487-94.

Wollstadt, L., G. Boyd, G. Gaumer & C.H. Bazuin (1979) "Resource Requirements for Teaching Continuity in Primary Care: Contrasts Between a Graduate and an Undergraduate Program, Journal of Family Practice (9):1065-1071.

Yonke, A.M. & R.P. Foley (1991) "Overview of Recent Literature on Undergraduate Ambulatory Care Education and a Framework for Future Planning," Academic Medicine (66):750-5.

Address for correspondence: Raef A. Lawson, The Institute for the Advancement of Healthcare Management, School of Business, The University at Albany, State University of New York, Albany, NY 12222, USA, [e-mail].