Implementation of the Community Outpatient Practice Experience in a Large Pediatric Residency Program

(1)

Implementation

of the

Community

Outpatient

Practice

Experience

in a

Large

Pediatric

Residency

Program

Kimberlee C. Recchia, MD*; Teresa M. Petros, MD*4; S. Andrew Spooner, MDII; and

Janet L. Cranshaw, MD*

ABSTRACT. Objectives. To determine the feasibility

of implementing the Community Outpatient Practice

Experience (COPE), a community-based continuity

program, in a large, tertiary-care-oriented pediatric residency; to assess the impact of the continuity program on pediatric residents’ experience; and to compare the experience in a variety of community practice settings.

Settings. Continuity clinic settings included a hospi-tal-based residents’ group practice (RGP) clinic (1989 through 1991) and a community-based program in which each resident was paired with a practicing pediatrician in the community (1991 through 1993). Community practice types included publicly funded clinics (n = 9), private

practices (n = 38), and managed-care practices (n = 14). In

all settings, residents spent half a day per week in

continuity activity.

Methods. Measures of residents’ experience (patient encounters, patient age distribution, and diagnostic

mix) were compared in both settings and among

com-munity practice types. RGP data were derived from a

patient scheduling database, and COPE data were

ob-tamed from patient encounter records submitted by

each resident.

Results. Residents in RGP (108.5 resident years) had 5294 encounters with 1568 patients. In COPE (102.5 resi-dent years), 21 978 encounters with 19 235 patients

oc-curred. COPE residents saw significantly more patients

per session (6.2 vs 1.7) than residents in RGP. The mean patient age in COPE was significantly higher than RGP (5.3 vs 2.6 years). A greater proportion of encounters in

RGP were for health supervision (61% vs 30%), but a

greater number of health supervision encounters per

resident occurred in COPE. There was a higher propor-tion of patients with chronic disease in RGP (38% vs 7%), but a greater number of patients with chronic disease

was seen per resident in COPE. Analysis of COPE data

by practice type showed fewer patient encounters per

session and a younger patient age in publicly funded

sites than in private- or managed-care practices. The

proportion of health supervision encounters was greatest

in publicly funded sites, but the greatest number of

health supervision encounters per resident occurred in managed-care practices.

From the *Depament of Pediatrics and the §Division of Medical Informat-ics, Department of Internal Medicine, Washington University School of Medicine and St Louis Children’s Hospital; and $Egelston Children’s Hos-pital and Emory University School of Medicine, Atlanta, GA.

lIDr Spooner is an American Academy of Pediatrics Fellow of the Pediatric

Scientist Development Program of the Association of Medical School Pediatric Department Chairmen, Inc.

Received for publication Jul 8, 1994; accepted Sep 8, 1994.

Reprint requests to (K-CR-) Director, COPE Program, 3532, St Louis

Children’s Hospital, One Children’s Place, St Louis, MO 63110.

Acad-emy of Pediatrics.

Conclusions. We successfully integrated a large-scale

community-based continuity experience into a large,

tertiary-care- oriented pediatric residency program.

We present COPE as an alternative to the

hospital-based continuity clinic and suggest it as a model for

improving residents’ primary-care experience. Pediatrics

1995;96:90-98; continuity clinic, graduate medical

educa-tion, ambulatory care, primary health care, internship and residency, pediatrics.

ABBREVIATIONS. COPE, Community Outpatient Practice

Expe-rience; RGP, residents’ group practice; ANOVA, analysis of

variance.

During the past several decades, the proportion of

health care delivered in the outpatient setting has

increased. Until recently, medical education has been primarily oriented toward inpatient care and subspe-cialty medicine.1’2 As the complexity of the hospital-ized patient increases, and the average length of stay

decreases, the inpatient experience has become less

representative of primary-care practice. Shorter hos-pital admissions have diminished the opportunities

for residents to evaluate presenting complaints and

to assess therapeutic outcomes. To ensure a

compre-hensive educational experience, which allows

ade-quate exposure to primary-care areas, medical

edu-cators must make increasing use of ambulatory

teaching sites.5

Despite the perceived need for increased numbers of primary-care physicians,6 there has been a signif-icant decrease in the proportion of residents

choos-ing careers in primary-care areas such as general

pediatrics, general internal medicine, and family

practice.79 It has been suggested that the biased

ex-posure of trainees toward academic medicine

with-out comparable high-quality experiences in primary

care is an important factor.7’1#{176}The academic stature of the generalist is underemphasized in the tertiary-care center, where a significant proportion of activity centers around biomedical science.11

Community-based programs shift the balance of

residents’ education from the heavily emphasized

area of specialty-oriented “academic” medicine

to-ward primary care. In so doing, they also serve to

increase the stature of the generalist. Although many residencies offer elective rotations in the community,

few have a major, ongoing community-based

com-ponent.12’13 Prompted by the initial successes of

two pediatric residencies that shifted a portion of

their continuity sites to the community,12’14’15 we

at Viet Nam:AAP Sponsored on September 1, 2020

www.aappublications.org/news

(2)

developed and successfully integrated a community-based continuity experience for all residents into our

large, tertiary-care-oriented residency program. The

diversity of practice settings required to implement a

community-based continuity program in this large

residency has allowed us to extend the existing

anal-yses of resident experience in the community by

comparing practice settings. The Community

Outpa-tient Practice Experience (COPE) program at

Wash-ington University School of Medicine pairs residents

with practicing pediatricians in the community and

has entirely replaced the hospital-based continuity clinic as the setting for resident continuity activity.

As residency program directors and medical

school deans seek to develop community-based

ex-periences for residents and students, the COPE

pro-gram is a model for developing such programs in the

setting of large, tertiary-care-oriented medical

cen-ters. These settings present unique challenges for

primary-care education, because they often attract

students and residents with strong interests in sub-specialty medicine and research rather than primary-care practice. We provide a description of the first 30

months of COPE, compare COPE with the

previ-ously existing hospital-based continuity clinic, and

assess resident experience among various COPE

practice settings.

PROGRAM DESCRIPTION

Residency Program

The pediatric residency training program at

Wash-ington University School of Medicine is based at St

Louis Children’s Hospital, a 235-bed facility with

approximately 60 000 emergency department visits,

30 000 ambulatory subspecialty pediatric visits, and

17 000 outpatient pediatric surgery visits per year. The hospital serves a metropolitan area of 2.5 million population and is a referral center for communities

located within a 200-mile radius. Approximately 60%

of patients have private insurance, and 40% receive

public assistance. Currently, a total of 63 residents from 37 medical schools are enrolled in the

Accred-itation Council for Graduate Medical

Education-accredited residency. Twenty residents (32%) are

graduates of Washington University School of

Medicine. Thirteen residents (20%) have combined

MD-PhD degrees, and five (8%) have additional

degrees including MS. MPH, and MPP.

Residents’ Group Practice

Before 1991, the residents’ continuity experience occurred exclusively in the setting of the

hospital-based residents’ group practice (RGP) clinic.

Resi-dents were assigned one half-day per week to attend

the clinic. First-, second-, and third-year residents

who shared an RGP day also shared coverage

re-sponsibility after hours. This model was designed to

provide a continuity experience for all residents in

which they would have maximum responsibility for

primary patient care, as if they were in a community

group practice. Residents attended the RGP clinic

with the exception of vacations, elective months, and

rotations in the emergency department (up to 12

weeks per resident per year). Residents recruited

patients for the clinic from the newborn nurseries,

the emergency department, and the inpatient wards

to supplement the panel of existing patients assigned to the RGP clinic during internship.

To provide residents with exposure to

primary-care role models and to improve education in

prac-tice-related areas such as practice management,

bill-ing, and malpractice, the RGP clinic was staffed by

community pediatricians. Each pediatrician attended the clinic on a voluntary (unsalaried) basis one

half-day per week for 4 consecutive weeks at least once

per year. Each resident was supervised by as many

as 10 different attending physicians per year. Didac-tic sessions on general pediatric topics (eg, otitis media, toilet training, and breast feeding) were held at the beginning of each clinic session.

The evaluation process in the RGP clinic was

over-seen by a supervisory committee consisting of the

pediatric chief residents, resident representatives, two faculty members, and a hospital administrator.

This group met monthly to review activity in the

clinic, to address problems, to develop quality

assur-ance indicators, and to make recommendations to

the residency program director and department

chairman.

Transition to COPE

In 1989, prompted in part by the lack of

enthu-siasm that our residents demonstrated toward

pri-mary care, we reviewed our residents’ continuity

clinic experience. Despite efforts to improve the

quality of the continuity clinic by developing a

practice-like model, we found that the RGP was

failing to meet the educational needs of our

resi-dents. In addition, the clinic was also failing to

meet the needs of the community both for primary

care and subspecialty follow-up care. The

limita-tions of the RGP included small patient panels

with few adolescents, many chronically ill infants,

and the high resident-to-faculty ratio (6 to 10

res-idents to I faculty member) with little continuity of

faculty supervision.

During a period of 18 months, we shifted our

residents’ continuity experience from the

hospital-based RGP to COPE. To facilitate this transition, the

COPE program and the RGP existed concurrently for

18 months. The first group of residents entered the

COPE program in January 1991, during their

intern-ship (postgraduate year FPGY-1]) year. During that

year, the PGY-2 and PGY-3 residents continued to

attend the RGP. All successive resident groups

en-tered COPE at the beginning of internship, and

res-idents originally assigned to the RGP remained there until completion of their residencies. As residents in

the RGP graduated, their patients were reassigned to

pediatric practices or community-based clinics in the

COPE program. By July 1992, all residents were

assigned to community sites for their continuity

activity.

COPE (1991 to Present)

In COPE, residents spend one half-day per week

in a community pediatrician’s practice during the

(3)

entire 3 years of residency. This responsibility

con-tinues through all rotations with the exception of

out-of-town electives and vacation (4 to 8 weeks/

resident per year). In addition to the regularly

scheduled weekly COPE sessions, each resident is

relieved of inpatient responsibilities to spend I

week in the preceptor’s office during the first few

months of residency. The objective of this block

ambulatory rotation is to facilitate resident

inte-gration into the preceptor’s practice and to provide

an introduction to ambulatory primary care.

Dur-ing this time, residents are encouraged to join the

preceptor for morning rounds and to participate in

as many of the preceptor’s other activities as

pos-sible. During the second and third years, residents

may spend up to 8 weeks of elective rotations in

the pediatrician’s office.

Sixty-eight community pediatricians participate in

the program as COPE preceptors. Preceptor

candi-dates are selected on the basis of their qualifications as outstanding role models as well as their

enthusi-asm for teaching. All preceptors agree to a 3-year,

unsalaried commitment. Forty percent of preceptors

are alumni of the residency training program at

Washington University. Of these general

pediatri-cians, 25% percent are trained in a pediatric subspe-cialty area. Preceptors are expected to attend faculty

development programs, which focus on challenging

general pediatric topics and are designed to facilitate

teaching in the ambulatory setting. These half-day

seminars, conducted by nationally known educators

and physicians, have covered a range of topics,

in-cluding teaching in the ambulatory setting and

be-havioral pediatrics. Efforts to encourage preceptor participation in the seminars (offering continuing

medical education credit, multiple scheduled times

for each seminar, and convenient locations) have

resulted in a 60% to 70% attendance rate at each

program. The response from the preceptors

attend-ing the programs has been enthusiastic; because of

the positive feedback, residents and other faculty

members have begun to attend these programs.

Residents are paired with pediatrician preceptors

by a matching system. During orientation to the

internship, each resident visits four to five different offices, and each preceptor meets four to five interns.

After completion of interviews, residents and

pre-ceptors submit their preferences and are matched by

the COPE program directors.

COPE practices range in size from one to seven

pediatricians per office and are located 0.2 to 50 miles

from the hospital (median, 20 miles, 44% within 10

miles). The average travel time between the hospital

and COPE sites is 30 minutes (range, 5 to 60

mm-utes). COPE sites include 15% publicly funded

din-ics, 62% privately owned practices and 23%

man-aged-care practices in urban (30%), suburban (64%),

and rural (6%) settings. PEDnet, a

hospital-sup-ported computer system developed in 1989, electron-ically links 23 of the preceptors to the hospital and

allows electronic mail communication, on-site data

entry, access to laboratory and radiology reports,

access to Physician’s Desk Reference, and medical

literature searches.

The COPE staff includes the program director, the

assistant director, and the program coordinator. The

program director is a physician whose responsibili-ties include supervising all aspects of the COPE pro-gram, including preceptor recruitment, educational

support, program evaluation, and research. The

as-sistant director, also a physician, conducts office site

visits and assists with program development. Other

responsibilities of the program directors include

at-tending on the inpatient service, conducting

resi-dents’ report, and involvement in other residency

programs activities. The COPE coordinator, a

full-time staff member, provides administrative support, collects experience data, coordinates resident

sched-uling, and facilitates communication among house

staff, faculty, preceptors, and preceptors’ office staff.

The evaluation process in COPE was designed to

allow assessment of individual resident experience

as well as evaluation of the program as a whole.

Residents submit copies of the office billing sheets

from each patient encounter to the program

coordi-nator, who enters the date of each encounter, the

patient identifier, and the diagnosis code(s) into a

database. Preceptors and residents complete

semian-nual written evaluations. The COPE program

direc-tors make annual site visits to each office, at which

time the resident is observed; the interactions

be-tween the resident and the preceptor are assessed;

and individual summary data regarding the number

and type of patients seen by the resident are

re-viewed. Feedback regarding resident performance is

provided at that time. Resident and preceptor group

feedback sessions occur at regular intervals through-out the year. These sessions provide an opportunity

for residents and preceptors to assess how well the

program is meeting its objectives. Exit interviews are

conducted with each resident and preceptor on

corn-pletion of the 3-year program. These 30-minute

in-terviews address specific logistic issues as well as the

overall educational objectives and impact of the

program.

A steering committee composed of community

pe-diatricians, residents, chief residents, the COPE staff, the residency program director, and the department

chairman meets quarterly to oversee the activities of

the program. This committee was instrumental in

developing the program curriculum and is

responsi-ble for periodically reviewing and revising the

educational objectives set out in the curriculum. The

curriculum is organized in a progressive manner,

emphasizing clinical problems of increasing

com-plexity at each resident level. The curriculum also

addresses topics such as practice management,

tele-phone advice, and office-based research.

Several key elements that allowed for successful

implementation of the COPE program included:

strong administrative and financial support from the

department chairman and the hospital

administra-tion; recruitment of a full time COPE program

direc-tor who was empowered to implement innovative

strategies and provide leadership and direction for

the program; a long-standing, strong relationship

between the institution and the community

physi-cians, which facilitated preceptor recruitment; and

(4)

the development of a clearly defined set of program

goals and a plan for continued evaluation and

faculty development.

METHODS

To provide a detailed description of the COPE program, we

compared resident activity in COPE with existing data from the

RGP. In addition, we examined the effects of practice type on

resident experience in COPE.

Definitions

A patient encounter was defined as any distinct pairing of a

patient and a resident on a distinct date. Sessions were defined as

days for which a date existed in the database for a given resident

(days on which the resident attended clinic but had no patients

scheduled were not included in this analysis). Resident years

expresses the cumulative resident attendance in RGP or COPE

(eg, one resident attending COPE for 1 year = I resident year).

Publicly funded sites were distinguished from others on basis of

receiving federal or state support. Managed-care sites included

health-maintenance organizations and managed multispecialist

group practices. Continuity was determined by the number and

proportion of repeat visits a resident had with any given patient.

Health supervision visits (well child visits) were defined as

en-counters in which normal growth and development and

anticipa-tory guidance were the primary emphasis.

Chronic disease was defined as any congenital or acquired

condition that was expected to increase the frequency of physician contact (for example, asthma, prematurity, or seizure disorder).

Diagnoses were analyzed using a modification of a diagnosis clustering system from a previously published analysis of

pediat-nc continuity clinic diagnostic content)6 Multiple diagnoses for a

given patient encounter in RGP or COPE were counted as separate

diagnoses in the diagnostic clustering technique.

RGP Data Collection

Seventy-five residents participated in the RGP during the

2-year period fromJuly 1989 throughJune 1991. The data collected

represented a total of 108.5 resident-years’ experience. By resident

level, there were 29.5 PGY-1 years, 41 PGY-2 years, and 38 PGY-3

years represented during the study period. During the 2-year

study period there were 5294 encounter records for 1568 patients.

The computer-based patient scheduling system for RGP provided

information including date of visit, patient identifier, date of birth,

and International Classification of Diseases, Ninth

Revision-Clinical Modification diagnostic code for each encounter. Patient

visit data, mean patient age, patient age distribution, and

continuity were determined on the basis of these records.

A chart review was performed to augment and verify the

diagnostic data obtained through the scheduling database, which

allowed only one diagnostic code per encounter. Five hundred

twenty-one RGP encounters (152 patients) were reviewed. This

sample (approximately 10% of total RGP visits) was randomly

selected and stratified by the number of total visits per patient to

the clinic. Analyses of RGP diagnoses in this report were based on

the chart review. The concurrence rate between diagnoses

deter-mined by chart review and those obtained through the scheduling

database was 66%. Examples of nonconcurrence included patients

for whom more than one diagnosis existed (the database allowed

one diagnosis per visit) and patients with miscoded diagnoses

(eg, urinary tract infection coded as upper respiratory infection).

COPE Data Collection

Sixty-four residents participated in COPE between January

1991 and June 1993. A total of 21 978 encounters with 19 235

patients occurred during the study period. This analysis includes data from 61 residents. Two residents failed to submit an adequate

number of billing sheets, and one resident dropped out of the

residency program several months after entering COPE; these

three residents were excluded from the analysis. The data

dis-cussed here include 49.5 PGY-1 years, 37 PGY-2 years, and 16

PGY-3 years for a total of 102.5 resident-years’ experience.

Resi-dents submitted copies of the office billing sheets from each

patient encounter to the program coordinator, who entered the

date, patient name, patient age, and diagnosis codes into the

relational database.

To assess the validity of diagnostic information entered into the

COPE database, a chart review of a randomly selected sample of

COPE records stratified by office type and resident level was

conducted. One hundred fifty-two patient records were reviewed.

The concurrence rate between the diagnosis in the chart and the

diagnosis in the COPE database was 88%. Because of this high rate

of concurrence, the analysis of diagnostic information in COPE is

based on the entire database (21 978 encounters) rather than the

subset of reviewed charts.

COPE sites were characterized as either publicly funded (n = 9),

privately owned (n = 38), or managed-care (n = 14) practices.

Resident-years’ experience was distributed among practice types

as follows: publicly funded, 15 years; privately owned, 63 years;

and managed care, 23.5 years.

Measures of Resident Experience

Total patient encounters, total patients, encounters per resident

per year, encounters per resident per session, and repeat visits

were compared in the RGP and COPE. Mean patient age and

patient age distribution were compared between the RGP and

COPE. Diagnostic data, including health supervision visits and

incidence of chronic disease, were compared between the RGP

and COPE. Total patient encounters, total patients, patient

en-counters per resident, mean patient age, the proportion of health

supervision visits, and diagnostic data also were compared among

COPE practice types.

Statistical Analyses

Pooled, two-tailed t tests of means were used to compare

numbers of patient encounters, patient ages, and continuity

rates. Where applicable, means are expressed with SD as mean

± SD in the text. Chi-square analysis was used to compare types

of patient visits. One-way analysis of variance (ANOVA) was

used to compare encounters per session per resident, the

per-centage of health supervision encounters, and the mean patient

age across practice types.

RESULTS

Comparison of RGP and COPE Experience

Cumulative patient encounters, encounters per

resident per year, encounters per resident per

ses-sion, the number of repeat visits (continuity), the

mean patient age and patient age distribution, and

diagnoses were compared in the RGP and COPE.

Table I shows the total patient encounters that

occurred in the RGP and COPE (5294 vs 21 978). A

fourfold increase in patient contact occurred in

COPE compared with the RGP (214 encounters per

resident per year vs 51 encounters per resident per

year). Patient encounters per resident per session

were significantly higher in COPE than in the RGP

(6.2 vs 1.7; P < .01).

Total repeat encounters were similar in the RGP

and COPE (2984 vs 2374); because total encounters

were greater in COPE than the RGP, there was a

marked difference in the proportion of repeat visits

TABLE 1. Comparison of Patient Encounters in RGP and

COPE*

RGP COPE

Resident years in study 108.5 102.5

Total patient encounters 5294 21 978

Total patients 1568 19 235

Encounters per resident per year 49.0 ± 22.9+ 215.5 ± 98.2+

Encounters per resident per session I .7 ± 0.2+ 6.2 ± 1.8+

* Values are expressed as mean ± SD where applicable.

t P < .01 for pooled t test comparing RGP and COPE means.

(5)

M).0

40()

Age inyears (mean ±andard deviation)

I

RGP: 2.6 ± 3.2 y. N=5.294 UCOPE: 5.3 ± 5.0 y. N 21,975

I

I00

in COPE (I I %) versus RGP (56%). Repeat encounters per resident per year were slightly higher in the RGP than in COPE (27.5 vs 22.9).

The Figure illustrates the distribution of patient

encounters in the RGP and COPE by patient age in

years. Residents in COPE saw significantly greater

numbers of patients in each age group than did

residents in the RGP. The mean patient age in COPE

was significantly higher than in the RGP (5.3 vs 2.6

years; P < .01).

Table 2 summarizes the type of patients seen in the

RGP and COPE by diagnosis. Health supervision

was the most frequent primary reason for visits in

each setting, occurring in 61 % of RGP encounters

and 30% of encounters in COPE. The proportion of

patients in the RGP identified as having chronic

dis-ease was 38%, compared with 7% of patients in

COPE with identified chronic disease. Specific diag-noses or diagnostic categories are presented in COPE

from most common (RGP plus COPE) to least

corn-mon. Otitis media was the most frequently

diag-nosed acute problem in both the RGP and COPE

(approximately 25% of total encounters). A greater

proportion of encounters in COPE was for acute

infectious diseases (upper respiratory infection and

sinusitis), behavior and parenting problems, and

skin diseases. The proportion of encounters in the

RGP for asthma, respiratory disorders, neurologic

disorders, anemia, congenital disorders,

prematu-rity, developmental delay, and seizure disorders was

greater than in COPE. More importantly, the greater

total number of encounters in COPE resulted in more

encounters per resident for many of these problems.

Gynecologic problems and drug use were notably

infrequent in both groups.

Resident Experience in Various COPE Practice Settings Table 3 summarizes total encounters, encounters per resident per session, the mean patient age, health

supervision visits, and diagnoses according to COPE

practice type. Residents in public clinics saw signif-icantly fewer patients per session than did residents in private- or managed-care offices (4.7 vs 6.4 or 6.6;

P < .05). The mean patient age was significantly

lower in public than in private- or managed-care

settings (3.3 vs 5.3 or 6.0; P < .05).

The proportion of health supervision visits was

significantly higher in public clinics than in

pri-act . .

0 1 2 4 5 6 7 11 9 10 II 2 13 14 IS 16 17 15 19 20 Age.y,

Figure. Distribution of patient ages in RGP and COPE.

vate- or managed-care practices (47.4% vs 29.0% or

28.1%). The number of well child visits was 76 per

resident per year in managed-care vs 65 in publicly

funded and 59 in private practices. The range of

diagnoses seen in the three settings was similar. Of

note, the frequency of behavioral and

developmen-tal problems was greatest in managed-care

prac-tices (7.5% managed visits vs 2.6% public and 1.3%

private visits).

DISCUSSION

Traditionally, the continuity clinic at a children’s

hospital-based pediatric training program serves a

number of important functions. It provides a setting for the delivery of primary care as well as

subspe-cialty hospital follow-up care within the medical

center. It also serves as the primary focus of the

residents’ ambulatory primary-care experience and

is the setting in which residents are expected to

acquire many of the skills necessary for general

pe-diatric practice. Our experience indicates that it is

possible to implement a large-scale

community-based continuity experience for residents using

practicing pediatricians as preceptors. We have

successfully handled the logistical aspects of

inte-grating a community-based program into a large,

tertiary-care- oriented pediatric residency program.

Through comparison of the community-based

expe-rience (COPE) with the experience in the

hospital-based clinic (RGP), we observed a number of

advan-tages of using community sites for resident continuity experience.

Comparison of RGP and COPE Experience

Patient Encounters

Residents in COPE saw a fourfold greater number

of patients per session than did residents in the pre-existing hospital-based continuity clinic. Our

find-ings support those of Osborn et al,17 who reported

increased patient contact for residents in community sites compared with residents in university-based

practices. There are a number of factors that may

contribute to the increased patient contact observed

in the community. The COPE setting may allow

res-idents to see the preceptor’s patients when their own patients cancel and also may provide an opportunity

for the resident to observe the preceptor between

scheduled patient visits. Increased preceptor avail-ability because of the favorable faculty-to-resident

ratio (1:1 in COPE vs 1:8 in the RGP) may increase

patient contact in COPE by allowing the resident to

see patients more efficiently. The physical separation

of the COPE sites from the hospital also limits

un-necessary interruptions and distractions, allowing

the resident to learn more efficiently and to function in a more productive manner.

Although the number of patients seen per session

in the RGP seems very low in comparison with

COPE, data from other programs suggest that our

experience in the hospital-based setting is not

unusu-al.’8’19 Our observations suggest that through

in-creased patient contact, residents in COPE had

sig-nificantly greater opportunities for primary-care

education than did residents in the RGP.

(6)

RGP* COPE 521 152 320 (61%) 58 (38%) n 447 100.0 106 23.7 51 11.4 34 7.6 4 0.9 6 1.3 5 1.1 11 2.5 27 6.0 25 5.6 2 0.4 13 2.9 11 2.5 5 1.1 3 0.7 14 3.1 11 2.5 3 0.7 6 1.3 11 2.5 3 0.7 12 2.7 21 4.7 14 3.1 1 0.2 48 10.7 Continuiti,

Exposing residents to continuity of care is an

im-portant goal of primary care residency programs.2#{176} Insofar as repeat visits with the same patient

repre-sents continuity of care, COPE afforded residents

with about the same amount of continuity as the

RGP; there were similar numbers of repeat visits in

the RGP and COPE groups. Because there were

many more total patient encounters in the COPE

group, the proportion of repeat encounters in COPE

was much lower than in the RGP. This higher

pro-portion of repeat visits in the RGP may be accounted for by the younger mean patient age of the patients in the RGP, the higher incidence of chronic disease in

the patient population in the RGP, and the ability of

residents in the RGP to see their patients on a day

other than their assigned clinic day. Additionally, the

greater proportion of upper-level residents (PGY-2

and PGY-3) in the RGP who have had more

oppor-tunities to establish ongoing relationships with

pa-tients and families may have contributed to the

observed difference in continuity rates.

The simple count of repeat encounters is probably

an inadequate measure of continuity of care. The

measures of continuity of care developed in the

health services research literature21’22 are all based on counting repeat visits and ignore the effect of clinical context on interpretation of patterns of encounters. For instance, a pair of visits spaced 9 months apart may represent poor continuity if the clinical problem is otitis media but excellent continuity if the clinical

152 patients (10% of total patients) reviewed manually for diagnostic

problem is health supervision of a 2 year old. We are

developing a measure of continuity that takes this

clinical context into account and that should allow

better monitoring of improvements in continuity in

the COPE program.

Age Distribution

A greater number of patients of all ages were seen

in COPE than in the RGP. A large proportion of the

encounters in both COPE and the RGP occurred in

infants and children younger than 4 years of age. It is likely that this reflects adherence in both settings to

the American Academy of Pediatrics’ Guidelines for

Health Supervision 11,23 which recommends that

in-fants and young children be seen more frequently

than school-age children or adolescents.

Interest-ingly, the mean age of patients seen by residents in

COPE was significantly greater than that seen in the

RGP. The tendency for the patients in the RGP to be

clustered in younger age categories may have been

attributable to the large proportion of these patients

who were recruited from the neonatal intensive care

unit and the newborn nurseries. In COPE, residents

see patients from the preceptor’s established practice

and are likely to encounter greater diversity with

respect to patient age. It is important to note that in

COPE, residents saw a significantly greater number

of school-age and adolescent patients. On average,

residents in COPE each saw 60 to 70 school-age

patients (5 through I I years) per year and

approxi-mately 30 adolescents (age 12 years) per year.

TABLE 2. Summary Data and Diagnosis Clusters for RGP and COPE

Total encounters Total patients

Health supervision encounters Total patients with chronic disease

Total diagnoses (other than health sup.)

Acute otitis media and effusion

Upper respiratory infection

Other communicable diseases

Sinusitis

Behavior, development, and parenting

Skin disease

Vomiting, diarrhea, and gastroenteritis

Asthma

Respiratory disorders

Allergy

Eczema

Neurologic disorders Urinary tract infection Diaper dermatitis Candida infection

Anemia and other hematologic disorders

Headache

Poor feeding and failure to thrive Congenital and inherited disorders Gynecologic problems

Prematurity

Developmental delay

Seizure disorder Drug use, drug toxicity

Other

* Data for RGP are based on a sample of 521 encounters with

accuracy.

21 978

19 235

6551 (30%)

I 270 (7%) n 20007 100.0 5190 25.9 4224 21.1 2150 10.7 968 4.8 668 3.3 755 3.8 575 2.9 575 2.9 625 3.1 429 2.1 335 1.7 175 0.9 172 0.9 136 0.7 106 0.5 85 0.4 104 0.5 41 0.2 67 0.3 77 0.4 27 0.1 56 0.3 23 0.1 9 0.0 2435 12.2

(7)

TABLE 3. Summary Data and Diagnosis Cluster s for COPE A ccordin g to Practice Type (Val ues Exp ressed as Mean ± S D Where Applicable)

Total encounters

Public (n = 9) Private (n = 38) Managed (n = 14)

2 234 13 574 6 170

Total patients I 945 1 1 966 5 334

Resident years 15 63 23.5

Encounters per resident per session* 4.7 ± 0.4* 6.4 ± 0.3* 6.6 ± 0.4*

Patient age at encountert 3.3 ± 3.9t 5.3 ± 4.9t 6.0 ±5.3+

Health supervision encounters 998 3 756 1 797

Percentage health supervisions 47.4 ± 17.2* 29.0 ± 13.5 28.1 ± 10.6

n % n % n %

Total diagnoses (other than health sup.) I 583 100.0 12 069 100.0 6 355 100.0

Acute otitis media and effusion 438 27.7 3 180 26.3 1 572 24.7

Upper respiratory infection 209 13.2 2 776 23.0 1 239 19.5

Other communicable diseases 221 14.0 1 418 11.7 511 8.0

Sinusitis 36 2.3 704 5.8 228 3.6

Skin disease 72 4.5 387 3.2 296 4.7

Behavior, development, and parenting 41 2.6 152 1.3 475 7.5

Respiratory disorders 47 3.0 466 3.9 112 1.8

Vomiting, diarrhea, and gastroenteritis 72 4.5 346 2.9 157 2.5

Asthma 67 4.2 309 2.6 199 3.1

Allergy 19 1.2 287 2.4 123 1.9

Eczema 45 2.8 196 1.6 94 1.5

Neurologic disorders 16 1.0 50 0.4 109 1.7

Urinary tract infection 13 0.8 114 0.9 45 0.7

Diaper dermatitis 16 1.0 97 0.8 23 0.4

Headache 7 0.4 66 0.5 31 0.5

Candida infection 22 1.4 73 0.6 11 0.2

Gynecologic problems 5 0.3 47 0.4 25 0.4

Anemia and other hematologic 31 2.0 35 0.3 19 0.3

Developmental delay 10 0.6 15 0.1 31 0.5

Congenital and inherited disorders 6 0.4 26 0.2 35 0.6

Poor feeding and failure to thrive 4 0.3 25 0.2 12 0.2

Seizure disorder 5 0.3 8 0.1 10 0.2

Prematurity 5 0.3 11 0.1 11 0.2

Drug use, drug toxicity I 0.1 4 0.0 4 0.1

Other 1408 11.1 10792 10.6 5372 15.5

* ANOVA for mean sessions per resident produced, F = 3.91; P < .05.

t ANOVA for mean patient ages at time of encounter produced, F = 232.8; P < .05.

:1:ANOVA for mean percentage health supervision encounters per resident produced, F = 7.29; P < .05.

Diagnostic Clusters

A broad range of diagnoses was encountered in

both the RGP and COPE. One of the important

aspects of the residents’ continuity experience is

exposure to health supervision and anticipatory

guidance. From the subset of RGP encounters (10%

of total) that were reviewed for diagnostic

infor-mation, we estimate that more than twice as many

health supervision encounters occurred in COPE

per resident.

Providing ambulatory care for patients with

chronic or complex medical problems is another

im-portant component of residents’ continuity

experi-ence. Although a larger proportion of patients in the

RGP had chronic disease, we estimate that in COPE,

residents encountered at least twice as many patients

with chronic disease because of the greater number

of total patients seen in that setting. The types of

chronic disease differed in the RGP and COPE.

Pre-maturity, asthma, respiratory disorders, congenital

disorders, seizures, neurologic disorders, anemia,

and developmental delay constituted a large propor-tion of encounters in the RGP. In COPE, behavioral

and developmental problems (other than

develop-mental delay), asthma, respiratory disorders, and

al-lergy were the most common chronic disorders.

Be-cause patients often entered the RGP clinic after

hospitalization or discharge from the neonatal unit, it is not surprising that the incidence of prematurity,

congenital disorders, seizures, and other chronic

problems was more frequent in the RGP group. The

increased incidence of behavioral and

developmen-tal problems and skin disorders in COPE may reflect

the preceptors’ efforts to broaden the residents’

ex-perience with these problems, which are poorly

represented in the inpatient setting. We suspect that

the spectrum of disease seen in COPE is much more

representative of general pediatric practice; data re-garding pediatricians’ activity are limited24 and may be significantly affected by practice setting.

Effect of Practice Settings on Resident Experience in COPE

Through COPE, residents are placed in a variety of

practice settings. Prompted by reports of differences in resident experience in various continuity settings

at other residency programs,17’19 we examined

resi-dent experience in three types of community

prac-tices; publicly funded clinics (n = 9), private offices

(n = 38), and managed-care practices (n = 14).

Our observations support those of Osborn et al,17

who reported fewer patient visits per session for

residents assigned to community clinics for

conti-nuity experience compared with residents

(8)

signed to private practices. COPE residents in

pub-licly funded clinics saw significantly fewer

patients per session than did residents in

private-or managed-care offices. Our observations during

COPE site visits suggest that the efficiency with

which each of the various types of practices

func-tion in part may account for this difference.

Resi-dents in some publicly funded sites report long

waiting times for patients to be put into examining

rooms. In addition, we have observed that a large

proportion of the patients seen in publicly funded

settings have complicated social and family issues,

which may require residents to spend additional

time with each patient. The mean patient age was

lower in the publicly funded setting than in

pri-vate- or managed-care practices, as observed by

Osborn et al,’7 who noted a greater proportion of

younger patients in community clinic settings.

Thirty percent of all visits in COPE were for health

supervision. Although residents in publicly funded

clinics encountered the greatest proportion of visits for health supervision, residents in managed-care

sites provided the greatest number of health

super-vision visits per resident per year. Osborn et al17 also

reported a similar difference in the proportion of

health supervision encounters between community

clinics and private offices. One explanation for these differences may be that patients cared for in commu-nity clinic settings use alternative sites (emergency

departments and urgent-care centers) for acute

problems after hours.

A wide variety of diagnoses was seen in each

practice setting. There were no striking differences in

the distribution of diagnoses among the three site

types, with the possible exception of behavioral and

developmental problems, which were encountered

much more frequently in the managed-care setting

than in private or publicly funded practices.

CONCLUSIONS

Increasingly, a greater proportion of health care is

occurring in ambulatory sites in both tertiary and

community settings. To prepare physicians

ade-quately for the challenges of providing care on an

outpatient basis, medical education must move from

the hospital ward to ambulatory sites. The

commu-nity-based continuity programs developed at the

Universities of Utah and Massachusetts represent

important first steps toward shifting the emphasis of

resident education to ambulatory sites in the

com-munity. The differences between the settings in

which these initial community-based programs were

developed and the setting in which COPE was

estab-lished are significant. The pediatric residency pro-grams at the Universities of Utah and Massachusetts

are medium size (7 to 13 residents per level)12 and

have established reputations for excellence in the

field of primary care education. In contrast, the pe-diatric residency program at Washington University

is a large residency program (20 to 22 residents per

level), which has a long tradition of educating large numbers of academic specialists. A large proportion

of residents entering our residency program has

strong interests in research; more than 20% have

combined MD-PhD degrees. Our experience in

COPE during the past 3 years indicates that it is

possible to implement a large-scale

community-based continuity experience using practicing

pedia-tricians as preceptors.

Benefits of COPE include increased patient contact

across a wide range of patient ages, without the

relative overemphasis of chronic medical problems

that commonly predominate hospital-based

continu-ity clinics. The community-based setting offers

Un-paralleled exposure to practice-related issues such as

office management, patient scheduling, and

tele-phone contacts. COPE offers a unique opportunity

for mentorship-it is one of the only settings in

which the resident is offered an opportunity to form a long-term, professional relationship with a faculty

member. COPE has strengthened the relationship

between the community practitioners and the

tertia-ry-care center by bringing residents and

primary-care physicians together.

The major strength of this study was the large

number of patient encounters that occurred during

the 30-month period in COPE. Limitations included

the inaccuracy inherent in using a scheduling

data-base for diagnostic information, which necessitated a

manual review of RGP charts for diagnostic content.

That the programs did not exist concurrently must be

taken into consideration in interpreting the results. It

is possible that the differences observed between the

RGP and COPE programs were accentuated by the

increased emphasis placed on COPE during that

program’s development.

We present COPE as an alternative to the

hospital-based continuity clinic and suggest it as a model for

improving residents’ primary care experience. This

model has been successfully integrated into our

large, tertiary-care--oriented residency. Although it

is too early to assess the impact of this program on

resident attitudes toward primary care adequately,

in the first group of graduates of the program more

than 50% selected careers in primary care (a twofold

increase from previous years) and more than 70% of

graduates reported that COPE had a significant

in-fluence on their career choice. We think the best way

to allow residents to make informed career choices is

to provide high-quality ambulatory, primary-care

experience in the community. We anticipate that

through participation in COPE, all residents will

benefit, regardless of career path.

ACKNOWLEDGMENTS

We acknowledge Harvey R. Colten, MD, James P. Keating, MD,

and John B. Watkins, MD, for their support of program

develop-ment and direction, as well as their editorial assistance. We thank

the COPE preceptor faculty for their generous contributions to the

education of our residents. We thank Valerie Schulz (COPE

coor-dinator) for administrative support, Marion Johnson and Jim Rice

from the Pediatric Computing Facility of the Department of

Pedi-atrics for their assistance in establishing the COPE database, and

the Department of Pediatrics at Washington University School of

Medicine and St. Louis Children’s Hospital for their ongoing

support of the COPE program.

REFERENCES

1. Perkoff GT. Teaching clinical medicine in the ambulatory setting-an

idea whose time may have finally come. N EngI IMed. 1986;314:27-31

(9)

2. Greenberg LW, Getson P, Brasseux C, et al. How are pediatric training

programs preparing residents for practice? Am IDis Child. 1991;145: 1389-1392

3. Wartman SA, O’Sullivan PS, Cyr MG. Ambulatory-based residency

education: improving the congruence of teaching, learning, and patient care. Anti Intern Med. 1992;116:1071-1075

4. Reeb KG. Education of residents in the pediatric office. Pediafr Cliii North Am. 1981;28:601-615

5. Schroeder SA. Expanding the site of clinical education: moving beyond the hospital walls. / Geii Intern Med. 1988;3(suppl):S5-S14

6. Iglehart JK. The American health care system. Teaching hospitals. N Engl IMed. 1993;329:1052-1056

7. Colwill JM. Where have all the primary care applicants gone? N Eng!

IMed. 1992;326:387-393

8. American Medical Association Primary Care Task Force. Report of the

medical schools section primary care task force. JAMA. 1992;268:

1092-1094

9. Stimmel B. The crisis in primary care and the role of medical schools. Defining the issues. JAMA. 1992;268:2060-2065

10. Kassler WJ, Wartman SA, Silliman RA. Why medical students choose

primary care careers. Acad Med. 1991;66:41-43

I 1. Perkoff GT. Should there be a merger to a single primary care specialty

for the 21st century? IFan Pract. 1989;29:185-190

12. Sargent JR. Osborn LM, Roberts KB, DeWitt TG. Establishment of

primary care continuity experiences in community pediatricians’ offices: nuts and bolts. Pediatrics. 1993;91:1185-1189

13. Barratt MS. Tanz RR. A survey of the structure and function of pediatric continuity clinics. Am JDis Child. 1992;146:937-940

14. Osborn LM, Sargent JR. Pediatric continuity clinics in the private prac-tice setting. Utah State Med Assoc Bull. 1987;35:3-5

15. Roberts KB, DeWitt TG. The office experience: don’t leave residency without it! Contemp Pediatr. 1992;9(suppl):15-20

16. Wilton R, Pennisi AJ. Insurance coverage and residents’ experience in a

pediatric teaching clinic. Am IDis Child. 1993;147:284-289

17. Osborn LM, Sargent JR. Williams SD. Effects of time-in-clinic, clinic

setting, and faculty supervision on the continuity clinic experience.

Pediatrics. 1993;91 :1089-1093

18. Kittredge D, Shissler GE. Monitoring continuity clinic (CC) with com-puter-generated visit data 1984-1992. The Ambulatory Pediatric

Asso-ciation 33rd Annual Meeting, Washington, DC, May 3-6, 1993. Am JDis Child. 1993;147:432. Abstract

19. Rice TD, Holmes SE, Drutz JE. Effects of clinic setting and post-graduate level on continuity clinic experience. The Ambulatory Pediatric

Asso-ciation 34rd Annual Meeting, Seattle, WA, May 2-5, 1994. Arch Pediatr Adolesc Med. 1994;148:P47. Abstract

20. zones sz, Schroeder SA. Evolving residency requirements for ambula-tory care training for five medical specialties, 1961 to 1989. West IMed.

1989;151 :676-678

21. Shortell SM. Continuity of medical care: conceptualization and

mea-surement. Med Care. 1976;14:377-391

22. Smedby O, Eklund G, Eriksson EA, Smedby B. Measures of continuity of care. A register-based correlation study. Med Care. 1986;24:51 1-518

23. American Academy of Pediatrics Committee on Psychosocial Aspects of

Child and Family Health. 1986-1988 Guidelines for Health Supervision II.

Elk Grove Village, IL: The American Academy of Pediatrics; 1988. 24. American Academy of Pediatrics Task Force of Pediatric Education. The

Future of Pediatric Education: A Report by the Task Force on Pediatric

Education. Elk Grove Village, IL: The American Academy of Pediatrics; 1978

PRETENDING PHYSICIANS

Many clinicians resent medical decisions being made by theoretician staff

offic-ers who have never been in the trenches. They contrast the attending physician

with yet another new breed, the pretending physician. Such pretending physicians

only think they are practicing medicine. There being only so many hours in the

day, they feel they can better help humanity by tackling the problems of the many

rather than those of the few.

Pretending physicians include some chiefs of medicine who rarely see patients

with their students and residents, and most certainly never see them alone. They

work on the larger picture, on new cures (bench research), new systems of health

care (planning and politics), new breeds of doctors (curriculum reform, selection

and promotion committee), new modes of practice (clinical guidelines, chart

re-view, quality assurance, peer review), and new forms of medical governance

(bylaws, committees, contracts). How much more exciting and cost-effective than

dealing with the sufferings of merely one single sick person. No wonder

pretend-ing physicians are paid more than mere clinicians.

Dunnea G. Pretending physicians and other new breeds. Br Med I. 1995;24:65.

Submitted by Student

(10)

1995;96;90

Pediatrics

Kimberlee C. Recchia, Teresa M. Petros, S. Andrew Spooner and Janet L. Cranshaw

Pediatric Residency Program