the emerging role of “hospitalists” in the american health care system

8/2/2019 THE EMERGING ROLE OF HOSPITALISTS IN THE AMERICAN HEALTH CARE SYSTEM

1/11

EDITORIALS

Vo lu me 3 35 N um be r 7 507

Editorials

R

EDESIGNING

G

RADUATE

M

EDICAL

E

DUCATION

L

OCATIONAND

C

ONTENT

EN years ago several educators proposed thatthe teaching of clinical medicine should shift

from inside to outside the hospital. They observedthat there was a discrepancy between the kinds ofpatients seen on the inpatient services and thoseseen in physicians offices, and suggested that resi-dent training should be better tailored to match therequirements of clinical practice.

1-3

For many yearsfew heeded their call for such a dramatic change inresidency training. Inpatient teaching of residents

was efficient in terms of faculty time, it was funded

by Medicare and other third-party payers, and a gen-eration of faculty members was accustomed to theconvenience of making teaching rounds on patientsconfined to their hospital rooms. Not only was out-patient teaching inadequately funded, but most fac-ulty members were also uncomfortable with the ideaof office-based teaching. At present, a large fractionof resident training in all but a few specialties is stillfocused on hospitalized patients.

Today we are confronted with a new, rather urgenteducational dilemma. The inpatient service has be-come an even more anachronistic site for learningclinical medicine, and at the same time the imped-iments to shifting training away from hospitalized

patients are even greater. As the cost of care hasincreased faster than our willingness to pay for it,hospital stays have shortened, many patients who areadmitted stay only long enough to have a cardiac, ra-diographic, or endoscopic procedure, and most of theimportant diagnostic problems are solved outside thehospital. Moreover, the spectrum of diseases seen onthe inpatient services is narrowing as hospitals havebecome havens for patients with complex or multiplecoexisting diseases, those who have failed to respondto office treatment, and those who need the ad-

vanced technology of intensive care units.There are also increasing restrictions on ambula-

tory teaching. Clinicians whose main location is ahospital-based faculty ambulatory practice have be-come busier, encumbered by more administrative andquality-improvement tasks and by higher standardsfor clinical productivity that force them to devotemore time to hands-on patient care. Given thesefactors and the increasingly persuasive view thattraining needs to be better matched to practice re-quirements, program directors have begun to seektraining sites other than those directly connectedto the mother institution. Training at free-standing

ambulatory care centers, community health centers,ambulatory care sites of managed-care organiza-tions, and private doctors offices are all being con-sidered.

4-6

There has been substantial experience with resi-dent education in some of these remote sites, but not

much in others. Family medicine has a well-devel-oped ambulatory care teaching program in nearly450 free-standing family practice centers. Residentsspend a minimum of half their total training withambulatory patients under supervision by staff physi-cians in these centers and in other ambulatory carefacilities.

7

Although a few managed-care organiza-tions have had accredited residency programs (espe-cially in internal medicine) for decades, they providemostly inpatient experience. In addition, at the timeof the last survey the aggregate amount of residenteducation in health maintenance organizations wasminimal,

8

and it has probably changed little since.There are no reliable data on ambulatory teaching

in other sites not directly connected with academicmedical centers, but the amount appears to be small.Some program directors do send their residents toprivate practitioners offices for short periods to ex-pose them to practice management, but this ap-proach is not widespread.

To switch the locale of most clinical teaching,we must solve another formidable problem name-ly, educational credentialing and evaluation. Meth-ods for selecting practitioners, monitoring their per-formance, helping them become more effectiveteachers, and evaluating the experience of residentsassigned to office practices are seriously underdevel-oped.

5,6

Without an educationally sound program,

which would entail time-consuming reporting ar-rangements, we could be heading for a retrogressivesplitting off of clinical education from the healthyacademic environment that has sheltered it. It willbe difficult to capture the same quality of education-al experience in the ambulatory setting that has ex-isted for decades in the hospital. Over the years aremarkable variety of educational conferences wassqueezed into residents schedules, and the high con-centration of faculty, subspecialty fellows, and seniorresidents optimized the opportunity to learn.

If inpatient care is relegated to a small cadre ofhospital-based specialists in the future, as suggestedby Wachter and Goldman in this issue of the Jour-nal,

9

it will still be necessary for future outpatientdoctors to have substantial experience with sick in-patients. Understanding the continuum of care fromacute illness through a diseases chronic phase is fun-damental to a physicians education. If we erred inthe past, it was in overemphasizing acute care. Asmore clinical education moves into ambulatory set-tings, we should take care not to make the inversemistake.

How much does the content of our educational

The New England Journal of MedicineDownloaded from nejm.org at LANE MEDICAL LIBRARY on February 13, 2012. For personal use only. No other uses without permission.

Copyright 1996 Massachusetts Medical Society. All rights reserved.


2/11

508

August 15, 1996

The New England Journal o

f Medicine

programs need to be revised? Is there a serious dis-junction between current training and current prac-tice? A group of practitioners only a few years out oftheir training recently reported that they felt inade-quately trained in practice management and cost-effective practice,

10,11

and some managed-care leaders

have complained that graduates of generalist trainingprograms require up to 18 months of retraining toenable them to practice effectively in their environ-ment.

12,13

A few managed-care organizations havedeveloped curricula to help their newly recruited pri-mary care physicians adjust to the companys prac-tices. These programs feature new disciplines suchas population management, team building, informa-tion-systems management, and clinical-resource man-agement. Some even attempt to expand physiciansnotion of duty beyond patient advocacy to includepayers, associated providers and other patients underpopulation-based medical practice.

14

I believe that medical centers do need to improve

the way they teach students and residents how to usediagnostic tests and that they must focus more oncost-conscious decision making.

15

It is not clear, how-ever, how much the training of primary care phy-sicians should be influenced by the special needsof managed-care plans. What part of the benefit ofteaching these new disciplines accrues to patients and

what part to the bottom line of managed-care organ-izations should be an important consideration.

The financial implications of moving more resi-dent education into ambulatory settings are critical.Hospitals are reimbursed by Medicare for its share ofthe costs of resident training on the basis of thenumber of residents on duty in hospital settings (in-

cluding the outpatient clinics) but not for those as-signed to other ambulatory sites. Because physicianspracticing in ambulatory settings will lose income ifthey take on more responsibility for teaching, they

will want to be paid, and there is no source of fundsfor this. It will also be expensive to hire full-time phy-sicians or other health care personnel to care for thesick inpatients now cared for by residents. Given thelikelihood of decreases in income from faculty-prac-tice plans,

16

academic departments will be unable tosubsidize ambulatory teaching with funds from thissource. Although for-profit managed care, the fast-est-growing segment of the health care system, hasbeen urged to support education and training asother industries do it has resisted doing so.

17

In my opinion, the most rational way to financegraduate medical education is through a govern-ment-mediated program supported by all payers,

with specific provisions for dealing with the financingproblems that will result from relocating training.

18-20

Legislative changes to deal with these funding issueshave been proposed for several years, and a bill intro-duced recently in Congress addresses the problem.

21

It mandates funding of graduate medical education

from sources that broadly represent the entire healthcare system (Medicare, Medicaid, and the private sec-tor) and recommends that policies be formulated toexpand eligibility for graduate medical education toinstitutions other than teaching hospitals. Unfortu-nately, there is little hope that legislation of this kind

will be enacted in the near future.While changes in the health care system are in high

gear, the location and content of graduate medicaleducation (and undergraduate medical education)sit on idle. Even without federal intervention, wecan do a great deal. As we seek new training sitesand teachers outside the hospital, we must developexplicit criteria for accepting new faculty and new

venues and implement monitoring methods to en-sure that educational goals are achieved. We mustdecide how much of the subject matter of the re-training done by managed-care plans is appropriatefor the residency curriculum. In doing so, depart-ment chairs and training-program directors must

continue to insist on a rigorous education in theknowledge base and cognitive skills of clinical med-icine, assert their responsibility to decide what is bestto teach, and avoid trendy educational fads.

Rather than wait until the public insists on it, weshould make graduate education more accountablein terms of cost. To do so, we must identify the fairand actual costs of direct medical education and theincreased cost attributable to residents at all sites.

We should anticipate that the federal governmentwill supply even fewer funds in the future, but weshould continue to press for federal reform of grad-uate medical education. As resident and studenteducation moves to ambulatory sites, federal funds

should move with it. Hospitals should probably nolonger control these funds; turning them over to de-partments that administer the training programsseems more rational. Training programs must alsoforge alliances with managed-care organizations, in-sisting that they sign on as partners and take real,not just token, responsibility for education. If theyuse the products of our programs, they must con-tribute to the funding of training.

Lastly, as we change teaching sites and modifyour curricula to keep pace with the rapidly changinghealth care system, we must be unwavering in our at-tention to the quality of the educational experienceand the benefits of training that accrue to patients.

We stand to lose a lot if we lose sight of this goal.

J

EROME

P. K

ASSIRER

, M.D.

REFERENCES

1.

Schroeder SA, Showstack JA, Gerbert B. Residency training in internalmedicine: time for a change? Ann Intern Med 1986;104:55-61.

2.

Shine KI. Innovations in ambulatory-care education. N Engl J Med1986;314:52-3.

3.

Perkoff GT. Teaching clinical medicine in the ambulatory setting: anidea whose time may have finally come. N Engl J Med 1986;314:27-31.




3/11

EDITORIALS

Vo lu me 3 35 N um be r 7

509

4.

Wartman SA. Managed care and its effect on residency training in inter-nal medicine. Arch Intern Med 1994;154:2539-44.

5.

Greenberg LW. Managed care, re-engineering and downsizing: willmedical education survive change? Pediatrics 1995;96:1146-7.

6.

Lesky LG, Hershman WY. Practical approaches to a major educationalchallenge: training students in the ambulatory setting. Arch Intern Med1995;155:897-904.

7.

Colwill JM. Financing graduate medical education in family medicine.

Acad Med 1989;64:154-8.

8.

Corrigan JM, Thompson LM. Involvement of health maintenance or-ganizations in graduate medical education. Acad Med 1991;66:656-61.

9.

Wachter RM, Goldman L. The emerging role of hospitalists in theAmerican health care system. N Engl J Med 1996;335:514-7.

10.

AAMC Executive Council. Roles for medical education in health carereform. Acad Med 1994;69:512-5.

11.

Cantor JC, Baker LC, Hughes RG. Preparedness for practice: youngphysicians views of their professional education. JAMA 1993;270:1035-40.

12.

Shine KI. The future of academic health centers. Physiologist 1995;38(2):51-5.

13.

Nordgren R, Hantman JA. The effect of managed care on undergrad-uate medical education. JAMA 1996;275:1053, 1058.

14.

Kertesz L. Payers start to worry about impact of HMOs bad press.Modern Healthcare. February 12, 1996:112, 114, 125.

15.

Kassirer JP. Academic medical centers under siege. N Engl J Med1994;331:1370-1.

16.

Fogelman AM. Impact of managed care on departments of internalmedicine. Am J Med 1994;96:I-V.

17.

Hasan MM. Lets end the nonprofit charade. N Engl J Med 1996;334:1055-7.

18.

Petersdorf RG. A proposal for financing graduate medical education.N Engl J Med 1985;312:1322-4.

19.

Eisenberg JM. How can we pay for graduate medical education in am-bulatory care? N Engl J Med 1989;320:1525-31.

20.

Idem.

Financing ambulatory care education in internal medicine.J Gen Intern Med 1990;5:Suppl:S70-S80.

21.

Moynihan DP. S.1870: The Medical Education Trust Fund Act of1996. Congressional Record Senate. Washington, D.C.: GovernmentPrinting Office, June 13, 1996.

1996, Massachusetts Medical Society.

M

ULTIMODAL

T

HERAPY

FOR

A

DENOCARCINOMA

OF

THE

E

SOPHAGUS

AND

E

SOPHAGOGASTRIC

J

UNCTION

N the Western world the incidence of adenocar-cinoma of the esophagus has risen markedly in

the past three decades. Barretts specialized colum-nar-cell metaplasia (Barretts esophagus), which isthe most important predisposing factor, is found inmore than 80 percent of patients with adenocarci-noma of the distal esophagus.

1

Although recent stud-ies indicate that microscopic areas of metaplasia mayalso occur in true carcinoma of the cardia and sub-cardial carcinoma,

2

the etiology, epidemiology, andpathology of these tumors appear to differ fromthose of adenocarcinoma of the esophagus.

3

Ade-nocarcinoma of the distal esophagus often growssubmucosally toward the proximal esophagus andspreads through the lymphatic system toward theposterior mediastinum, whereas true carcinoma ofthe cardia and subcardial carcinoma more closely re-semble gastric cancer in their spread.

3

I

Surgical resection has been the mainstay of thera-py for localized adenocarcinoma of the distal esoph-agus. In recently published series, the five-year sur-

vival rate after resection of tumor in early stages(stages 1 and 2a) is about 80 percent.

4-6

Even amongpatients with more advanced penetration of the

esophageal wall or early lymphatic spread, long-termsurvival is possible with surgery alone, provided thatthe resection is complete, as judged by macroscop-ical and microscopical criteria.

4-6

However, since mostpatients present with late-stage adenocarcinoma ofthe esophagus, the overall prognosis after surgicalresection alone is dismal. Multimodal therapy, espe-cially an approach in which chemotherapy, radio-therapy, or both are given before surgery (neoadju-

vant treatment), has therefore received increasingattention.

7

In this issue of the Journal,

Walsh et al. presentthe results of an interim analysis (after three years offollow-up) of a prospective, randomized, controlled

trial comparing resection alone with resection pre-ceded by radiotherapy and chemotherapy in patients

with adenocarcinoma of the esophagus or esophago-gastric junction.

8

They found a statistically signifi-cant survival benefit for patients who received mul-timodal therapy. However, a critical evaluation ofthe data raises several points of concern that must beaddressed before we can draw general conclusions.

The study included patients with different typesof tumor (i.e., adenocarcinoma of the esophagus orthe cardia). This is reflected in the variety of surgi-cal approaches and the finding of Barretts mucosain less than 40 percent of the patients. The studyalso included patients with early and advanced tu-

mors. Although the number of early-stage tumorswas low in the group treated with surgery alone, theproportion of patients with early disease in thegroup assigned to multimodal therapy is unknown,because we consider the staging methods used inthe trial to be insufficient. An imbalance in the ini-tial distribution of tumor stages might have influ-enced the results. In our experience, endoscopic ul-trasonography accurately assesses the tumor stage

with respect to size, location, and local invasion,

7,9

but Walsh et al. did not use this method. Endo-scopic ultrasonography could have allowed precisestratification of the patients according to tumorstage and the identification of subgroups that mightbenefit most from multimodal therapy or primaryresection. Interpretation of the results of this studyalso requires a clearer account of the kind of surgicalresection (e.g., complete resection on the basis ofboth microscopical and macroscopical inspection)

10

and the extent of lymphadenectomy than the au-thors give. Complete resection is the single mostimportant prognostic factor in patients with adeno-carcinoma of the esophagus and esophagogastric

junction.

5,6

Adequate lymphadenectomy is essential




4/11

510

August 15, 1996

The New England Journal o

f Medicine

for accurate postoperative staging and may influ-ence the prognosis.

5,6,10

Despite these shortcomings, the survival benefitof multimodal therapy in this study is impressive.The results should encourage intensified efforts toidentify subgroups of patients who might benefit

from this expensive, complex, and potentially dan-gerous treatment.

Most studies indicate that we can expect a multi-modal therapy to be beneficial only in patients withlocally advanced tumors who have a substantial re-sponse to preoperative therapy and who have neithermicroscopical nor macroscopical evidence of residualtumor after resection. Since the benefit of multimod-al therapy in patients with early tumors is question-able,

7,11

accurate staging methods and ways of pre-dicting therapeutic responses are essential. In ouropinion, staging before treatment should includeendoscopic ultrasonography and diagnostic laparos-copy with ultrasonography to search for peritoneal

and liver metastases.

9,12

No method available todaycan reliably predict the clinical or histopathologicalresponse to neoadjuvant therapy. In patients withcomplete clinical responses, the resected specimenoften contains residual tumor, and in many patients

with only partial responses the specimen has no via-ble tumor cells. Research on this problem has begunto focus on the endoscopic appearance of the growthpattern of the tumor and on molecular markers suchas thymidylate synthase.

13

Positron-emission tomog-raphy is a promising method for assessing the re-sponse to neoadjuvant therapy.

Combined preoperative chemotherapy and radio-therapy appear to induce histopathologically com-

plete remissions more effectively than preoperativechemotherapy alone.

7,11

However, neoadjuvant che-motherapy and radiotherapy for squamous-cell carci-noma of the esophagus increased postoperative mor-bidity and mortality.

7

These risks must be taken intoaccount in the planning of multimodal treatment, es-pecially because the control of distant disease is a ma-

jor problem, even in patients with histopathologicallycomplete remissions. We need a detailed analysis ofthe pattern of recurrence and long-term follow-upbefore deciding on the optimal mode of neoadjuvanttherapy.

In our centers, primary resection remains thetreatment of choice in patients with early (stages 1and 2a) adenocarcinoma of the distal esophagus. Weconsider multimodal therapy to be experimental inpatients with such early tumors. In contrast, patients

with locally advanced tumors in whom a successfulcomplete resection of the tumor and regional lymphnodes appears doubtful (stages 2b, 3, and 4) shouldreceive neoadjuvant therapy to reduce the size of thetumor and increase the chance of a successful com-plete resection. Because of the associated morbidityand mortality, multimodal therapy should be restrict-

ed to patients with sufficient physiologic reserve to withstand not only its complications but also thesurgical resection. Randomized trials that investigatethe optimal type of preoperative therapy for partic-ular subgroups of patients are needed before multi-modal therapy for adenocarcinoma of the esophagus

and esophagogastric junction can be widely recom-mended.

H

ANSJOCHEN

W

ILKE

, M.D.

Essen University Medical School45122 Essen, Germany

U

LRICH

F

INK

, M.D.

Technische Universitt Mnchen81675 Munich, Germany

REFERENCES

1.

Stein HJ, Siewert JR. Barretts esophagus: pathogenesis, epidemiology,functional abnormalities, malignant degeneration and surgical manage-ment. Dysphagia 1993;8:276-88.

2.

Cameron AJ, Lomboy CT, Pera M, Carpenter HA. Adenocarcinoma ofthe esophagogastric junction and Barretts esophagus. Gastroenterology1995;109:1541-6.

3.

Siewert JR, Hlscher AH, Becker K, Gssner W. Kardiakarzinom: Ver-such einer therapeutisch relevanten Klassifikation. Chirurg 1987;58:25-34.

4.

Clark GW, Peters JH, Ireland AP, et al. Nodal metastasis and sites ofrecurrence after en bloc esophagectomy for adenocarcinoma. Ann ThoracSurg 1994;58:646-54.

5.

Hlscher AH, Bollschweiler E, Bumm R, Bartels H, Hofler H, SiewertJR. Prognostic factors of resected adenocarcinoma of the esophagus. Sur-gery 1995;118:845-55.

6.

Steup WH, DeLeyn P, Deneffe I, Van Raemdonck D, Coosemans W,Lerut T. Tumors of the esophagogastric junction: long-term survival in re-lation to the pattern of lymph node metastasis and a critical analysis of theaccuracy or inaccuracy of pTNM classification. J Thorac Cardiovasc Surg1996;111:85-95.

7.

Fink U, Stein HJ, Bochtler H, Roder JD, Wilke HJ, Siewert JR. Neo-adjuvant therapy for squamous cell esophageal carcinoma. Ann Oncol1994;5:Suppl 3:S17-S26.

8.

Walsh TN, Noonan N, Hollywood D, Kelly A, Keeling N, HennessyTPJ. A comparison of multimodal therapy and surgery for esophageal ad-enocarcinoma. N Engl J Med 1996;335:462-7.

9.

Dittler HJ, Siewert JR. Role of endoscopic ultrasonography in esopha-geal cancer. Endoscopy 1993;25:156-61.

10.

Hermanek P, Sobin LH. UICC TNM classification of malignant tu-mors. 5th rev. ed. Berlin, Germany: Springer-Verlag, 1992.

11.

Stahl M, Wilke H, Fink U, et al. Combined preoperative chemother-apy and radiotherapy in patients with locally advanced esophageal cancer:interim analysis of a phase II trial. J Clin Oncol 1996;14:829-37.

12.

Molloy RG, McCourtney JS, Anderson JR. Laparoscopy in the man-agement of patients with cancer of the gastric cardia and oesophagus. Br JSurg 1995;82:352-4.

13.

Lenz HJ, Leichman CG, Danenberg KD, et al. Thymidylate synthasemRNA level in adenocarcinoma of the stomach: a predictor for primary tu-mor response and overall survival. J Clin Oncol 1996;14:176-82.





5/11

EDITORIALS

Vo lu me 3 35 N um be r 7

511

S

EX

AND

U

RINARY

T

RACT

I

NFECTIONS

HE association between sexual intercourse andacute cystitis in women has been recognized for

generations in folk wisdom (e.g., honeymoon cys-

titis), but researchers have been less certain of thisassociation.

1

In fact, a causal relation between twovery common events, intercourse and acute cystitis,is difficult to establish.

Several studies have linked urinary tract infectionsto heterosexual intercourse. Kunin and McCormacknoted that nuns had a lower prevalence of bacteriuriathan other populations of women during early adultlife.

2

Buckley et al.

3

reported an increase in bacterialcounts in the urine after intercourse in 30 percent of

women, and Nicolle et al.

4

found that both asymp-tomatic and symptomatic bacteriuria was more com-mon on the day after coitus. Remis et al.

5

and Stromet al.,

6

in cross-sectional casecontrol studies, found

that coitus is a major risk factor for symptomatic uri-nary tract infection in young women.

In this issue of theJournal, Hooton et al.7 reporton a prospective, population-based study of two co-horts of otherwise healthy, sexually active, nonpreg-nant premenopausal women who were selected onthe basis of their willingness to participate and theirintention to start a new method of contraception.The annual incidence of acute cystitis was higherthan anticipated during 323 person-years of follow-up, with rates of 0.7 episode per person-year amonguniversity women and 0.5 per person-year among

women enrolled in a health maintenance organiza-tion (HMO). Using these data, the investigators cre-

ated a statistical model to identify the risk factors foracute cystitis and the doseresponse relation betweenacute cystitis and coitus.

After correction for other risk factors, the relativerisk of urinary tract infection among unmarried uni-

versity women increased dramatically from 1.0 forwomen who had not been sexually active during thepreceding week to 9.0 for women who had had in-tercourse seven times during that period. Coital fre-quency in this population was in the expected range,

with a median of 6.4 episodes per month in the university cohort and 5.0 episodes per month in theHMO cohort. These numbers imply that at leasttwo thirds of the acute episodes of urinary tract in-fection in this population are attributable to coitusand that probably twice as many episodes occuramong sexually active women as are reported on thebasis of data obtained from office visits.7

Most issues surrounding the putative link betweensexual activity and acute cystitis remain unresolved.Hooton et al. have substantiated the relation butprovide no insights into its pathogenesis. We knowthat the establishment of pathogenic microbial florausually precedes episodes of acute cystitis8 and that

the use of spermicides with or without a diaphragmalters normal flora and facilitates colonization withpathogens.9 We also know that bacterial vaginosishas been linked to acute cystitis.10 What we do notknow is whether a particular coital position or sexualpractice is more likely to lead to acute cystitis.

Postmenopausal women remain sexually active,and some also have frequent episodes of acute cysti-tis. Raz and Stamm11 showed that local estrogen re-placement with topical vaginal cream can preventboth symptomatic and asymptomatic infections. Nostudies have addressed the relation of coitus and uri-nary tract infections in postmenopausal women or in

women who have undergone hysterectomy.What interventions can reduce the burden of dis-

comfort and cost caused by urinary tract infections?Often-prescribed methods such as postcoital mictu-rition do not prevent acute cystitis.7 Contraceptivemethods other than a diaphragm with spermicide canbe tried. Other unproved interventions include the

ingestion of cranberry or blueberry juice. These juic-es contain a substance with biologic activity that canreduce bacterial adhesion to uroepithelial cells.12 Al-though the daily ingestion of 300 ml of cranberry

juice reduced the incidence of bacteriuria and pyuriain elderly women,13 this approach has not been stud-ied in young women. Regular or intermittent antimi-crobial prophylaxis after intercourse will prevent mostrecurrences of urinary tract infections, but concernabout the emergence of resistance limits the useful-ness of this approach. Couples need to understandthat sexually associated infections are not necessarilysexually transmitted and that neither partner is re-sponsible for their occurrence. Guilt and fear must

not be allowed to interfere with sexual fulfillment.The direct health costs of an episode of acute cys-

titis range from $40 to $80, and the patient losesabout one day of productive activity. Unfortunately,less attention has been focused on urinary tract in-fection in the recent past: fewer scientists are work-ing in the field, and fewer studies have been pub-lished.14 Womens health, reproductive health, andsexually transmitted infections are high-priority ar-eas for further research. We need to find better waysof preventing urinary tract infections in sexually ac-tive women.

ALLAN RONALD, M.D.

St. Boniface HospitalWinnipeg, MB R2H 2A6, Canada

REFERENCES

1. Kunin CM. Sexual intercourse and urinary infections. N Engl J Med1978;298:336-7.2. Kunin CM, McCormack RC. An epidemiologic study of bacteriuria andblood pressure among nuns and working women. N Engl J Med 1968;278:635-42.3. Buckley RM Jr, McGuckin M, MacGregor RR. Urine bacterial countsafter sexual intercourse. N Engl J Med 1978;298:321-4.




6/11

512 August 15, 1996

The New England Journal of Medicine

4. Nicolle LE, Harding GK, Preiksaitis J, Ronald AR. The association ofurinary tract infection with sexual intercourse. J Infect Dis 1982;146:579-83.5. Remis RS, Gurwith MJ, Gurwith D, Hargrett-Bean NT, LaydePM. Risk factors for urinary tract infection. Am J Epidemiol 1987;126:685-94.6. Strom BL, Collins M, West SL, Kreisberg J, Weller S. Sexual activity,contraceptive use, and other risk factors for symptomatic and asympto-

matic bacteriuria: a case control study. Ann Intern Med 1987;107:816-23.7. Hooton TM, Scholes D, Hughes JP, et al. A prospective study of riskfactors for symptomatic urinary tract infection in young women. N Engl JMed 1996;335:468-74.8. Kunin CM, Polyak F, Postel E. Periurethral bacterial flora in women:prolonged intermittent colonization with Escherichia coli. JAMA 1980;243:134-9.9. Hooton TM, Roberts PL, Stamm WE. Effects of recent sexual activityand use of a diaphragm on the vaginal microflora. Clin Infect Dis 1994;19:274-8.10. Hooton TM, Fihn SD, Johnson C, Roberts PL, Stamm WE. Associa-tion between bacterial vaginosis and acute cystitis in women using dia-phragms. Arch Intern Med 1989;149:1932-6.11. Raz R, Stamm WE. A controlled trial of intravaginal estriol in post-menopausal women with recurrent urinary tract infections. N Engl J Med1993;329:753-6.12. Ofek I, Goldhar J, Zafriri D, Lis H, Adar R, Sharon N. Anti Escheri-chia coliadhesin activity of cranberry and blueberry juices. N Engl J Med

1991;324:1599.13. Avorn J, Monane M, Gurwitz JH, Glynn RJ, Choodnovskiy I, LipsitzLA. Reduction of bacteriuria and pyuria after ingestion of cranberry juice.JAMA 1994;271:751-4.14. Harding GKM, Ronald AR. The management of urinary infections:

what have we learned in the past decade? Int J Antimicrob Agents 1994;4:83-8.


LAPAROSCOPIC HYSTERECTOMY

IS THEREA BENEFIT?

XCEPT for abortion, laparoscopically assisted vaginal hysterectomy has generated more con-

troversy and discussion than any other type of gyne-cologic surgery in recent times. Is it because hys-terectomy is itself controversial? Could it be thatthe benefits of laparoscopy remain uncertain? A cyn-ic would suggest that the issues surrounding laparos-copy and laparoscopically assisted vaginal hysterecto-my in particular continue to be debated because thetechnique has been promoted by those closely tiedto the industries that gain financially from its useand because it has been tested with study designs bi-ased toward achieving favorable results. In somestudies the outcome evaluated was success in remov-ing the uterus, rather than success in providing atrue benefit to the patient.

Initially, investigators published small case seriesand concluded that laparoscopically assisted vaginalhysterectomy offered advantages over traditional vag-inal hysterectomy in terms of decreased pain, shorterhospital stays and recovery times, and decreased com-plication rates. Some authors suggested that laparo-

E

scopically assisted vaginal hysterectomy should re-place the traditional vaginal approach.1,2 In contrast,our prospective, randomized trial3 found no differ-ences between laparoscopically assisted vaginal hys-terectomy and the traditional approach with regardto estimated blood loss, operative and postoperative

complications, postoperative hospital stays, and peri-ods of convalescence. In addition, the mean directmedical charge for laparoscopically assisted vaginalhysterectomy ($7,905) was higher than for vaginalhysterectomy ($4,891).

Some have also suggested that laparoscopic assist-ance has benefits if an oophorectomy is planned inconjunction with the hysterectomy,4 because theybelieve that transvaginal oophorectomy is often notfeasible and that it increases the risk to the patient.

Although adding a procedure increases the risk ofany operation somewhat, in most instances the ova-ries can be removed at the time of a vaginal hyster-ectomy without an associated increase in risk.3,5 Un-

fortunately, instrument manufacturers and surgicalenthusiasts have both continued to advocate thebenefits of laparoscopically assisted vaginal hysterec-tomy despite data that should have led to more tem-pered enthusiasm.

In this issue of the Journal, Weber and Lee6 ad-dress the possible overuse of laparoscopically assisted

vaginal hysterectomy. The number of hysterectomiesdecreased during the seven-year study period, prob-ably because more conservative surgical procedures

were used and because of the influence of managedcare. They found that laparoscopic assistance wasmost likely to be used when the patient was coveredby health insurance. Does this finding reflect a true

difference in the presence of disease states betweenthe insured and the uninsured, or is it that newforms of technology are more readily available to pa-tients in certain socioeconomic classes? Althoughthe trend away from the abdominal approach to hys-terectomy was a positive one, it is impossible to de-termine its importance, because we do not know

whether the patients who underwent abdominal orlaparoscopically assisted vaginal hysterectomy couldhave successfully undergone vaginal hysterectomy

without laparoscopy. If 53 percent of the patients inthe vaginal-hysterectomy group were operated onfor prolapse, for example, why did only 18 percentundergo oophorectomy, since these patients wereprobably of postmenopausal age? How did the twogroups compare with regard to uterine weight orstage of endometriosis? Such data are generally notavailable in data bases such as the one used in thisstudy, but they are needed before one can criticallyevaluate the type of hysterectomy chosen.

In another article in this issue, Dorsey et al. exam-ine the costs and charges associated with three typesof hysterectomy.7 As with the patients reported on by

Weber and Lee, we cannot determine which of the




7/11

EDITORIALS


patients who had laparoscopically assisted procedureswould have been candidates for a vaginal procedure that is, why was the laparoscopic technique used?In addition, in some of the patients undergoing lap-aroscopically assisted vaginal hysterectomy, the indi-cation for surgery was pelvic relaxation. Common

wisdom suggests that these patients would have beenbetter served by a traditional vaginal approach. Be-cause the patients were not cared for in a standard-ized way, caution must be used in evaluating the dataon factors such as the length of the hospital stay. Fi-nally, information about complications and treat-ments after discharge may be incomplete.

Despite their limitations, both studies confirmwhat others have found that laparoscopic assist-ance offers the patient no immediate economic ad-

vantage and uses more hospital resources than eithervaginal or abdominal hysterectomy.8,9 In one study10

the hospital charges associated with the laparoscopictechnique were lower than those for an abdominal

approach. The charges were $5,869 for vaginal hys-terectomy, $6,552 for abdominal hysterectomy, and$6,431 for laparoscopically assisted vaginal hysterec-tomy. These authors achieved this reduction by elim-inating all use of disposable laparoscopic trocars andequipment from their procedures and using a special-ized operative team.

What can we conclude about laparoscopically as-sisted vaginal hysterectomy? We know that when pos-sible, vaginal hysterectomy is the procedure of choice

when hysterectomy is required. We suspect that lap-aroscopic assistance may allow some patients toundergo a vaginal hysterectomy when they wouldotherwise have required laparotomy. For laparoscop-

ically assisted vaginal hysterectomy to be cost effec-tive, the expensive disposable instrumentation willhave to be eliminated, and patients undergoing theprocedure will need to have a shorter convalescencethan those undergoing abdominal hysterectomy. Todetermine the appropriate role for the new proce-dure, gynecologic surgeons must begin to conduct

well-designed clinical trials that examine not onlythe short-term surgical outcomes, but also the over-all economics of the procedure and the quality of lifethat results.

THOMAS G. STOVALL, M.D.

Bowman Gray School of Medicine

Winston-Salem, NC 27157-1066

ROBERT L. SUMMITT, JR., M.D.

University of TennesseeMemphis, TN 38103

REFERENCES

1. Padial JG, Sotolongo J, Casey MJ, Johnson C, Osborne NG. Laparos-copy-assisted vaginal hysterectomy: report of seventy-five consecutive cases.J Gynecol Surg 1992;8:81-5.2. Liu CY. Laparoscopic hysterectomy: a review of 72 cases. J Reprod Med1992;37:351-4.3. Summitt RL Jr, Stovall TG, Lipscomb GH, Ling FW. Randomized

comparison of laparoscopic-assisted vaginal hysterectomy with standardvaginal hysterectomy in an outpatient setting. Obstet Gynecol 1992;80:895-901.4. Raju KS, Auld BJ. A randomised prospective study of laparoscopic vag-inal hysterectomy versus abdominal hysterectomy each with bilateral sal-pingo-oophorectomy. Br J Obstet Gynaecol 1994;101:1068-71.5. Sheth SS. The place of oophorectomy at vaginal hysterectomy. Br J Ob-stet Gynaecol 1991;98:662-6.6. Weber AM, Lee J-C. Use of alternative techniques of hysterectomy inOhio, 19881994. N Engl J Med 1996;335:483-9.7. Dorsey JH, Holtz PM, Griffiths RI, McGrath MM, Steinberg EP. Costsand charges associated with three alternative techniques of hysterectomy.N Engl J Med 1996;335:476-82.8. Nezhat C, Bess O, Admon D, Nezhat CH, Nezhat F. Hospital costcomparison between abdominal, vaginal, and laparoscopy-assisted vaginalhysterectomies. Obstet Gynecol 1994;83:713-6.9. Boike GM, Elfstrand EP, DelPriore G, Schumock D, Holley HS, LurainJR. Laparoscopically assisted vaginal hysterectomy in a university hospital:report of 82 cases and comparison with abdominal and vaginal hysterecto-

my. Am J Obstet Gynecol 1993;168:1690-701.10. Johns DA, Carrera B, Jones J, DeLeon F, Vincent R, Safely C. Themedical and economic impact of laparoscopically assisted vaginal hysterec-tomy in a large, metropolitan, not-for-profit hospital. Am J Obstet Gynecol1995;172:1709-19.





8/11

514 August 15, 1996


Sounding Board

THE EMERGING ROLEOF

HOSPITALISTS IN THE AMERICANHEALTH CARE SYSTEM

HE explosive growth of managed care has led toan increased role for general internists and other

primary care physicians in the American health caresystem. This change is welcome in many respects,since generalists have perennially been undervaluedby health care institutions, payers, and even pa-tients.1-3 The greater prominence of generalism hasled to an increase in the number of medical students

who choose careers in primary care,4 expanded jobopportunities for generalists,5 and a modest increasein the incomes of primary care physicians.6

Two of the principles underlying generalism,whether in the form of internal medicine, pediatrics,or family medicine, have been comprehensivenessand continuity.7,8 Ideally, the primary care physician

would provide all aspects of care, ranging from pre- ventive care to the care of critically ill hospitalizedpatients. This approach, argued the purists, wouldresult in medical care that was more holistic, less frag-mented, and less expensive.9 To its proponents, thenotion was so attractive the general internist ad-mits the patient to the hospital, directs the inpatient

workup, and arranges for a seamless transition backto the outpatient setting that questioning it wouldhave seemed sacrilegious merely a few years ago.

Unfortunately, this approach collides with the re-alities of managed care and its emphasis on efficien-cy. As a result, we anticipate the rapid growth of anew breed of physicians we call hospitalists spe-cialists in inpatient medicine who will be respon-sible for managing the care of hospitalized patientsin the same way that primary care physicians are re-sponsible for managing the care of outpatients. Spe-cialists in inpatient care have long had a central rolein urban hospitals in Canada and Great Britain, butuntil recently, such specialists have been scarce in theUnited States. However, a role for this specialty isnow being developed both in and outside academia,especially in areas where managed care predominates,such as San Francisco, and we expect this growth toaccelerate soon.

We believe the hospitalist specialty will burgeonfor several reasons. First, because of cost pressures,managed-care organizations will reward profession-als who can provide efficient care. In the outpatientsetting, the premium on efficiency requires that thephysician provide care for a large panel of patientsand be available in the office to see them promptlyas required. There is no greater barrier to efficiency

in outpatient care than the need to go across thestreet (or even worse, across town) to the hospital tosee an unpredictable number of inpatients, some-times several times a day. There are parallel pressuresfor efficiency in the hospital. Since the inpatient set-ting involves the most intensive use of resources, it

is the place where the ability to respond quickly tochanges in a patients condition and to use resources

judiciously will be most highly valued. This shouldprove to be the hospitalists forte.

Equally pressing is the question of value, definedas the quality of care divided by its cost.10 The sur-

vival of all health care systems is becoming increas-ingly dependent on the delivery of high-value care.(For academic health centers, this means that moreexpensive care must be justified by better outcomes.)Many physicians, though primarily serving outpa-tients, have exceptional skills in providing inpatientcare. It seems unlikely, however, that high-value carecan be delivered in the hospital by physicians who

spend only a small fraction of their time in this set-ting. As hospital stays become shorter and inpatientcare becomes more intensive, a greater premium willbe placed on the skill, experience, and availability ofphysicians caring for inpatients.

The debate over the role of hospitalists is takingplace against the backdrop of the larger controversyover whether generalists or specialists should pro-

vide care for relatively ill patients.11 The first decadeof managed care has been dominated by a gatekeep-er model, in which care is managed by a primarycare physician. There is some evidence that this mod-el saves money,12,13 and for common diseases, thequality of care provided by generalists and specialists

appears to be similar.14 Building on a considerablebody of data demonstrating a positive relation be-tween procedural experience and outcomes,15-18 anumber of recent studies have examined whether asimilar relation exists for nonprocedural care of pa-tients with complex medical illnesses. Those who fa-

vor the use of inpatient specialists for hospital carepoint to the strong correlation of experience withthe quality of care provided for patients in an inten-sive care unit,19,20 as well as for those with AIDS,21-25

asthma,26,27 rheumatoid arthritis,28 or acute coronarysyndromes.29-31

If our prediction of an increased role for hospital-ists is borne out, the effects on academic medicalcenters will be profound. The triple threat leader skilled clinician, researcher, and educator wasthe paradigm of exceptional faculty achievement (orfantasy) for more than a generation. Balancing aproductive research career with teaching and clinicalcare was easier when academic health centers wereless accountable for the quality and cost of clinicalcare than they are now. Given the parallel pressurefor funding research,32 one can envision fewer triplethreats in the future, with researchers concentrating




9/11

SOUNDING BOARD


on research and clinician-educators concentrating onclinical work and teaching. And the clinician-educa-tors may branch again, with some focusing on out-patients and others on inpatients. We also believethat the relation between quality and volume inthe performance of procedures may lead to another

schism between medical specialists who primarily per-form procedures and those who do not.

What will hospitalist jobs in academia look like? Inthe light of the increasing intensity of inpatient care,

we believe that 12 months as an attending physicianis a formula for burnout; 3 to 6 months a year seemsmore sustainable. The experience of critical care spe-cialists (intensivists) is a close parallel.19,20,33 In ac-ademic settings, these specialists typically limit their

yearly service on the intensive care unit to three tosix months in order to prevent burnout and to haveopportunities for academic productivity (Cohen N,Luce J: personal communication). As with intensiv-ists, a major challenge is to link the hospitalist role

successfully with other activities. The outpatient en-terprise, which is subject to the same pressures forefficiency, high quality, and low cost, may have littleuse for a physician who is otherwise occupied 80percent of the time during half the year, except per-haps in drop-in settings that do not require continu-ity of care. For some physicians who are trained in aspecialty, work as a hospitalist may be complement-ed by an inpatient or outpatient consulting practicein that specialty, and for generalists, inpatient con-sultation in general medicine will have a similar role.In the academic sett ing especially, a premium will beplaced on clinical quality improvement, the develop-ment of practice guidelines, and outcomes research,

not only to provide the physician with a creative out-let and a potential source of funding during thenonclinical months but also to give the academiccenter a practical research-and-development arm. Oneof the advantages of the hospitalist model is that itcreates a core group of faculty members whose in-patient work is more than a marginal activity and

who are thus committed to quality improvement inthe hospital.

For house staff in internal medicine, the introduc-tion of hospitalists may mean a greater likelihood ofbeing supervised by attending physicians who arehighly skilled and experienced in providing inpatientcare. House staff have long enjoyed a certain amountof autonomy, because many of their faculty supervi-sors have been relatively unfamiliar with moderninpatient care. Such autonomy may be diminished

with the new approach to inpatient care. Althoughthere is bound to be transitional pain, we believethat the potential for improved inpatient teaching

will more than compensate for it. Moreover, thischange will help answer public calls for closer andmore effective faculty oversight of house staff andstudents.34

Training programs in internal medicine have em-phasized flexibility. Many traditional programs, al-though based in inpatient settings, pride themselveson providing training that is flexible enough to al-low graduates to practice primary care competently.Pressures from residency-accreditation agencies have

also resulted in a broader curriculum. Over the pastfew years, many traditional programs have augment-ed and improved training in ambulatory care so thattheir graduates will have the necessary flexibility. Andmost primary care programs, while training residentsfor careers as outpatient generalists, have includedenough inpatient and intensive care medicine in thecurriculum to ensure that their graduates are com-petent in these settings. However, if the medicalmarketplace creates jobs that are based in either in-patient or outpatient settings (but not both), theprimary care program of the future may need to pro-

vide only enough inpatient training so that its grad-uates will know how to care for sick outpatients.

Conversely, some traditional programs may develophospitalist tracks that emphasize acquisition of theskills most relevant to inpatient practice. If suchtracks are developed, it will be important not to re-duce training in ambulatory care too aggressively,since the competent hospitalist will need a full un-derstanding of what can and cannot be donein the outpatient setting.

The hospitalist trend is already visible at both teach-ing and nonteaching hospitals in areas where man-aged care has taken root. Some medical groups, suchas the Scripps Clinic in La Jolla, California, use arotating schedule of primary care physicians, eachof whom is the dedicated admitting physician for

week-long tours of duty.35 The Park Nicollet MedicalGroup, a large multispecialty practice in Minneapolis,uses a hybrid model with two full-time hospitalistscomplemented by rotating general internists and fam-ily physicians.36 Other groups, such as San FranciscosCalifornia Pacific Medical Group, employ full-timehospitalists to handle inpatient care for a large groupof patients receiving care on a capitated basis (Aron-owitz P: personal communication). Similarly, the Kai-ser Permanente system now uses full-time hospitalistsin 3 of its 15 hospitals in northern California and isplanning to introduce this model in most of its otherfacilities in the region over the next few years (Likosky

W: personal communication). Anecdotal reports sug-gest that the use of each of these models has resultedin substantial decreases in lengths of stay, hospitalcosts, and specialty consultation.35-38

In both academic and nonacademic settings, themost effective way to organize hospitalists may be asa multispecialty group. Envision a model for a largeintegrated health care system in which a team ofhospitalists some trained as generalists, others asspecialists shares responsibility for the manage-ment of inpatient care. Consultation is provided by




10/11

516 August 15, 1996


members of the group who have the appropriate ex-pertise; specialty consultation is obtained outsidethe group when the required knowledge is unavail-able from the members. Although not everyone inthe group possesses the skills to oversee the de-

velopment of practice guidelines and to study the

outcomes of care, all the members participate indeveloping and disseminating guidelines, as well asteaching clinical medicine. Group members developstrong relationships with hospital staff, dischargeplanners, specialty consultants, and outpatient phy-sicians. The outpatient physicians, in turn, have moretime in the office to see patients and less hospitalduty. Potential problems with the transfer of carebetween the outpatient and inpatient settings areprevented by meticulous communication betweenoffice and hospital.

Despite this optimistic vision, objections to thehospitalist model, not surprisingly, come from bothgeneralists and specialists. To preserve continuity and

maintain their acute care skills, some primary carephysicians clearly prefer to manage their patientscare during hospitalization, even when there is noeconomic incentive to do so (Aronowitz P: personalcommunication). Specialists fear that skilled hospital-ists may order fewer consultations than primary carephysicians. But flexibility in the implementation ofthe hospitalist model may allay some of these fears.To date, most systems employing hospitalists havenot required that inpatient physicians manage thecare of all hospitalized patients. This flexibility hassatisfied primary care physicians who worried that ahospitalist model would block their access to inpa-tient medicine.36 The fears of specialists fewer

consultations and lower income are more difficultto address. In fact, objections raised by specialistshave impeded the implementation of the model insome nonteaching hospitals in southern California(Chandler W: personal communication).

The hospitalist model of inpatient care challengesmany of the basic assumptions of generalists, special-ists, academic medical centers, and managed-careorganizations. Nevertheless, we believe that the forc-es promoting the use of the model are sufficientlycompelling that it will continue to be adopted inboth teaching and nonteaching settings. As with anymajor transition, the medical community must con-tinually reevaluate the new approach to ensure thatany possible discontinuity in care is outweighed byimproved clinical outcomes, lower costs, better edu-cation for physicians, and greater satisfaction on thepart of patients.

ROBERT M. WACHTER, M.D.

LEE GOLDMAN, M.D.

University of California, San FranciscoSan Francisco, CA 94143-0120

We are indebted to Paul Aronowitz, M.D., Andrew Bindman,M.D., Weston Chandler, M.D., William Likosky, M.D., Amy Mark-owitz, and Ed ONeil, Ph.D., for their helpful comments on themanuscript.

REFERENCES

1. Kassirer JP. Primary care and the affliction of internal medicine. N Engl

J Med 1993;328:648-51.2. Levinsky NG. Recruiting for primary care. N Engl J Med 1993;328:656-60.3. Weeks WB, Wallace AE, Wallace MM, Welch HG. A comparison of theeducational costs and incomes of physicians and other professionals.N Engl J Med 1994;330:1280-6.4. Kassebaum DG, Szenas PL. Specialty intentions of 1995 U.S. medicalschool graduates and patterns of generalist career choice and decision mak-ing. Acad Med 1995;70:1152-7.5. Weiner JP. Forecasting the effects of health reform on US physician

workforce requirement: evidence from HMO staffing patterns. JAMA1994;272:222-30.6. Jaklevic MC. Primary-care docs see incomes rise. Modern Healthcare.September 26, 1994:64.7. Donaldson M, Yordy K, Vanselow N. Institute of Medicine. Definingprimary care: an interim report. Washington D.C.: National AcademyPress, 1994.8. Starfield B. Primary care: concept, evaluation, and policy. Oxford, Eng-land: Oxford University Press, 1992.

9. Starfield B, Simpson L. Primary care as part of US health services re-form. JAMA 1993;269:3136-9.10. Eddy DM. Clinical decision making: from theory to practice: connect-ing value and costs: whom do we ask, and what do we ask them? JAMA1990;264:1737-9.11. Kassirer JP. Access to special ty care. N Engl J Med 1994;331:1151-3.12. Greenfield S, Nelson EC, Zubkoff M, et al. Variations in resource uti-lization among medical specialties and systems of care: results from theMedical Outcomes Study. JAMA 1992;267:1624-30.13. Martin DP, Diehr P, Price KF, Richardson WC. Effect of a gatekeeperplan on health services use and charges: a randomized trial. Am J PublicHealth 1989;79:1628-32.14. Greenfield S, Rogers W, Mangotich M, Carney MF, Tarlov AR. Out-comes of patients with hypertension and non-insulin dependent diabetesmellitus treated by different systems and specialties: results from the Med-ical Outcomes Study. JAMA 1995;274:1436-44.15. Moore MJ, Bennett CL. The learning curve for laparoscopic cholecys-tectomy. Am J Surg 1995;170:55-9.16. Hughes RG, Hunt SS, Luft HS. Effects of surgeon volume and hospi-

tal volume on quality of care in hospitals. Med Care 1987;25:489-503.17. Hannan EL, ODonnell JF, Kilburn H Jr, Bernard HR, Yazici A. In-

vestigation of the relationship between volume and mortality for surgicalprocedures performed in New York State hospitals. JAMA 1989;262:503-10.18. Kimmel SE, Berlin JA, Laskey WK. The relationship between coronaryangioplasty procedure volume and major complications. JAMA 1995;274:1137-42.19. Pollack MM, Katz RW, Ruttimann UE, Getson PR. Improving theoutcome and efficiency of intensive care: the impact of an intensivist. CritCare Med 1988;16:11-7.20. Li TC, Phillips MC, Shaw L, Cook EF, Natanson C, Goldman L. On-site physician staffing in a community hospital intensive care unit: impacton test and procedure use and patient outcome. JAMA 1984;252:2023-7.21. Kitahata MM, Koepsell TD, Deyo RA, Maxwell CL, Dodge WT, Wag-ner EH. Physicians experience with the acquired immunodeficiency syn-drome as a factor in patients survival. N Engl J Med 1996;334:701-6.22. Stone VE, Seage GR III, Hertz T, Epstein AM. The relation betweenhospital experience and mortality for patients with AIDS. JAMA 1992;268:2655-61.

23. Turner BJ, McKee L, Fanning T, Markson LE. AIDS specialist versusgeneralist ambulatory care for advanced HIV infection and impact on hos-pital use. Med Care 1994;32:90216.24. Paauw DS, Wenrich MD, Curtis JR, Carline JD, Ramsey PG. Abilityof primary care physicians to recognize physical findings associated withHIV infection. JAMA 1995;274:1380-2.25. Bennett CL, Adams J, Gertler P, et al. Relation between hospital ex-perience and in-hospital mortality for patients with AIDS-related Pneu-mocystis cariniipneumonia: experience from 3,126 cases in New York Cityin 1987. J Acquir Immune Defic Syndr 1992;5:856-64.26. Mayo PH, Richman J, Harris HW. Results of a program to reduce ad-missions for adult asthma. Ann Intern Med 1990;112:864-71.27. Zeiger RS, Heller S, Mellon MH, Wald J, Falkoff R, Schatz M. Facili-tated referral to asthma specialist reduces relapses in asthma emergency




11/11

SOUNDING BOARD


room visits. J Allergy Clin Immunol 1991;87:1160-8. [Erratum, J AllergyClin Immunol 1992;90:278.]28. Ward MM, Leigh JP, Fried JF. Progression of functional disability inpatients with rheumatoid arthritis: associations with rheumatology subspe-cialty care. Arch Intern Med 1993;153:2229-37.29. Schreiber TL, Elkhatib A, Gr ines CL, ONeill WW. Cardiologist versusinternist management of patients with unstable angina: treatment patternsand outcomes. J Am Coll Cardiol 1995:26:577-82.30.

Ayanian JZ, Hauptman PJ, Guadagnoli E, Antman EM, Pashos CL,McNeil BJ. Knowledge and practices of generalist and specialist physiciansregarding drug therapy for acute myocardial infarction. N Engl J Med1994;331:1136-42.31. Mark DB, Naylor CD, Hlatky MA, et al. Use of medical resources andquality of life after acute myocardial infarction in Canada and the UnitedStates. N Engl J Med 1994;331:1130-5.32. Movsesian MA. Effect on physician-scientists of the low funding rateof NIH grant applications. N Engl J Med 1990;322:1602-4.33. Guidelines Committee, Society of Critical Care Medicine. Guidelines

for the definition of an intensivist and the practice of critical care medicine.Crit Care Med 1992;20:540-2.34. Robins N. The girl who died twice: every patients nightmare: the Lib-by Zion case and the hidden hazards of hospitals. New York: DelacortePress, 1995.35. Brandner J. Will hospital rounds go the way of the house call? Man-aged Care, July 1995:19-28.36. Kilgore C. Some internists bid farewell to rounds. Internal Medicine

News. March 1, 1995: 1,33.37. Gipe B. A Pennsylvania model for in-house acute care physician serv-ices. Improving inpatient performance and relieving outpatient stress. Cost& Quality 1996;2:6.38. Winslow R. Employer group rethinks commitment to big HMOs. WallStreet Journal. July 21, 1995:B1, B4.


Gulf of Mexico JOHN N. WHITAKER, M.D.

the emerging role of “hospitalists” in the american health care system

Documents