Chapter 10

Physician Specialization

CONTENTS

Page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...........’”’” .“209Reliability of the Indicator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........”...2~2Validity of the Indicator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...........214

Is Physician Specialization a Reasonable Indicator of Quality~ . .............214Does the Board Certification Process Accurately Reflect

a Physician’s Competence? . . . . . . . . . . . . . . . . . . . . . . . . . “ “ “ “ “ . ‘ “ . “ “ “ “ . “ “ “ .214Does Physician Specialization Accurately Predict a Physician’s Quality? .. ...215Is the Use of Physician Specialization as a Quality Indicator Generalizable

Across Specialties? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... ... .”””221Feasibility of Using the Indicator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ....222Conclusions and Policy Implications . . . . . . . ..............................224

TablesTable Page

10-1. Specialty Boards Recognized by the American Boardof Medical Specialties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . “ . . . . . . . . “ ~ “ “ . “21O

10-2. Specialty Boards Recognized by the Advisory Board for OsteopathicSpecialists . . . . . . . . . . . . . . . . . . . . . . . .m c. .m . . . . . .< ’ . . . ” ’”2 1 1

103. Independent Boards That Certify Physicians . . . . . . . . . . . . . . . . . . . . . . . . . ..21210-4. Recertification y Specialty Boards Recognized by the American Board

of Medical Specialties: Current Status and Requirements . . . . . . . . . . . . . . . .21310-5. Studies on Physician Specialization Reviewed by TA . ................21610-6. Self-Designated Practice Specialties Recognized by the

American Medical Association . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Chapter 10

Physician Specialization

INTRODUCTION

The use of physician specialization to measurethe quality of care provided by individual physi-cians represents a structural approach to meas-uring quality. Like other structural indicators,physician specialization is often used to assessquality on the assumption that certain character-istics of physicians may lead to better perform-ance, which in turn may bring about better pa-tient outcomes.

A person who wants to practice medicine andsurgery legally in a State must obtain a licenseor certification of qualification from the StateBoard of Medical Examiners or other designatedagency (70 Corpus Juris Sec. 12). Although therequirements for medical licensure vary amongStates, in general, a person must be a graduateof a medical school accredited by the LiaisonCommittee on Medical Education, ] have com-pleted 1 year of residency training in a programapproved by the Accreditation Council for Grad-uate Medical Education, * and have passed the Fed-eration Licensing Examination sponsored by theFederation of State Medical Boards (470).3 Witha medical license from a given State, a physiciancan practice medicine in that State, in whateverspecialty area he or she chooses.

Some physicians, in addition to having generalmedical training, may have received training in

‘The Liaison Committee on Medical Education is the official ac-crediting body for educational programs leading to the M.D. de-gree and is listed for this purpose by the U.S. Secretary of Educa-tion and recognized by the Council of PostSecondary Accreditation.The committee consists primarily of members from the Council onMedical Education of the American Medical Association and theAssociation of American Medical Colleges (157).

‘The Accreditation Council for Graduate Medical Education iscomposed primarily of members from the American Board of Med-ical Specialties, the American Hospital Association, the AmericanMedical Association, the Association of American Medical Colleges,and the Council of Medical Specialty Societies. Louisiana, Missouri,Ohio, and Tennessee do not require any residency training for licen-sure. Connecticut, Guam, Maine, New Hampshire, and Washing-ton require completion of 2 years of residency training, and Ne-vada requires 3 years (47o).

3Most States also recognize certifying examinations of the Na-tional Board of Medical Examiners to license physicians (513).

a particular specialty area. Such training is notrequired for medical licensure, but physicians whohave specialty training may be eligible to becomecertified by a specialty board.4 Even if they havenot received specialty training or been board-certified, however, physicians may designatethemselves specialists.

Two major operational definitions of physicianspecialization have been used:

● certification by a specialty board, and● the fact that a physician is practicing in his

or her area of specialty training.

Many organizations certify physicians. TheAmerican Board of Medical Specialties and theAmerican Medical Association (AMA) officiallyrecognize the 23 specialty boards shown in table10-1. These boards certify 63.5 percent of the phy-sicians practicing in the United States (365). TheAdvisory Board for Osteopathic Specialists rec-ognizes the 17 osteopathic specialty boards shownin table IO-2. All of the 40 specialty boards rec-ognized either by the American Board of Medi-cal Specialties or by the Advisory Board forOsteopathic Specialists require physicians to com-plete a specified amount of training and a certainset of requirements and to pass an examination.In addition to these boards, there exist at least 69specialty boards not recognized by the AmericanBoard of Medical Specialties or the AdvisoryBoard for Osteopathic Specialists (see table 1o-3).

4Depending on the specialty, a physician may complete I to 5years of additional training in a specialty area. The American Boardof Orthopedic Surgery requires 5 years of additional specialty train-ing for a physician to become board certified, while the AmericanBoard of Colon and Rectal Surgery requires only 1 year of addi-tional training. The term “board eligible” is sometimes used to de-scribe a physician who has completed the necessary training andother predetermined requirements to become board certified, buthas not taken the formal examination offered by the board. Becauseof continuing confusion about the term board eligible, however,the American Board of Medical Specialties’ policy has disavowedthe use of the term. The American Board of Medical Specialties hasdeclared that the term has been given “such diverse meanings bydifferent agencies that it has lost its usefulness as an indicator ofa physician’s progress toward certification by a specialty board” (18).

209

210

Table IO-.– Specialty Boards Recognized by the American Board of Medical Specialties

Certificates in subspecialty areas Date initial

Certificates of Certificates of subspecialty

American Board Of: General certification special qualifications added qUaliflCatlOns offered

Allergy and lmmunO109Y. Allergy and immunology Diagnostic laboratory immunokJ9Y1986

AnesthesiO109Y . . . . . . . . . . .Anesthesi0i09Y Critical care medicine1986

Colon and Rectal Surgery. Colon and rectal surgeryDermatology . . . . . . . . . . . . . Dermato109Y

Emergency Medicine . . . . . . Emergency medicineFamily Practice . . . . . . . . . . Family practiceInternal Medicine ., ., Internal medlcme

Neurological Surgery . . . . . . Neurological surgeryNuclear Medicine ., Nuclear medicine

Obstetrics and Gynecology Obstetrics and gynecology

Ophthalmo109Y . . . . . . . . . . . Ophthalm0109Y()~hopaedic Surgery . Orthopedic sur9erYOtolaryng0109Y ., ., ., .0tolaryng0109YPathology . . . . . . . . . . . . . . . Anatomic and clin. path.

Anatomic pathologyClinical pathology

Pediatrics . . . . . . . . . pediatrics

Physical Medicine andRehabilitation . . . . . . . Physical medicine and rehabili”

tationPlastic Surgery . . . . . . Piastic surgeryPreventive Medicine ., ., . . . Aerospace medicine

Occupational medicinePublic health and generalpreventive medicine

Psychiatv and Neurology. PsychiatryNeurologyNeurology with specialqualifications in childneuroiogy

Radiology ., ., . . . . . . . . RadiologYDiagnostic radiologyRadiation Oncoiogy

Surgery . . . . . . . . . Surgery

DermatopathologyDermatO109ical immunOi09Y/diagnostic and laboratOrY immunology

Cardiovascular diseaseCritical care medicineDiagnostic laboratory immuno109YEndocrin0109Y and metabolismGastroenterology

HematologyInfectious diseaseMedical OnCO109YNephrologypulmonary diseaseRheumatoio9Y

Cooperates with American Board of Radi-ology and American Board of Pathology inradioisotopic pathology and nuclearradiology

Gynecologic oncoi09YMaternal and fetal medicineReproductive endocrinO109Y

Geriatric medicine

Geriatric medicine

Critlcd care m8dicine

Critical care

Hand surgery

Blood bankingChemical pathologyDermatopathologyForensic pathologyHematologyImmunopathologyMedical microbiologyNeuropathologyRadioisotopiC pathologyDiagnostic laboratory immunologyPediatric cardiologyPediatric critical care medicinePediatric endocrinO109YPediatric hematology -oncologyPediatric nephrologyPediatric puimonologyNeonatal-perinatal medicine

Hand surgery

Child psychiatry

1974

1985

1987194119871986197219411988197219721973197219411972

197419741974

19731950197419591952198319491947197419861961198719781974197419861975

1959

Nuciear radiology1957

Pediatric surgery

General vascular surgery

Hand surgery1975

Surgical critical care 19861982

General vascuiar surgery 1988

Thoracic Surgery Thoracic surgery

Urology . . . . . . . UrologySOURCE: American Board of Medical Specialties, Annual Report and Reference Handbook (Evanston, IL: 1987).

211

Table 10-2.—Specialty Boards Recognized by theAdvisory Board for Osteopathic Specialists

American Osteopathic Board of: Subspecialties

AnesthesiologyDermatologyEmergency MedicineGeneral PracticeInternal Medicine ., .Allergy/immunology

CardiologyEndocrinologyGastroenterologyHematologyHematology/oncologyInfectious diseasesMedical diseases of the chestNephrologyOncologyRheumatology

Neurology and Psychiatry ., Child psychiatryChild neurology

Nuclear MedicineObstet r ics and Gynecology . , Gynecolog ic onco logy

Maternal and fetal medicineReproductive endocrinology

Ophthalmology and Otorhinolaryngology, ,Oro-facial plastic surgeryOtorhinolaryngology and

oro-facial plastic surgeryOrthopedic Surgery. Hand surgeryP a t h o l o g y Laboratory medicine

Anatomic pathologyAnatomic pathology and

laboratory medicineForensic pathology

Pediatrics NeonatologyPediatric allergy/immunologyPediatric cardiologyPediatric hematology/

oncologyPediatric infectious diseasesPediatric intensive carePediatric nephrology

Preventive Medicine ... ., Preventive medicine/aerospace medicine

Preventive medicine/occupational-environ-mental medicine

Preventive medicine/publichealth

ProctologyRadiology ., ., Diagnostic radiology

Radiation oncologyRehabilitation MedicineSurgery .,, . . . . . . . . Surgery (general)

Neurological surgeryPlastic and reconstructive

surgeryThoracic cardiovascular

surgeryUrological surgeryGeneral vascular surgery

SOURCE: Advisory Board for Osteopathic Specialists, “Requirements for Cer.tification: Advisory Board of Osteopathic Specialists and Boards of Cer-tification, ” Chicago, IL, 1987

In addition to offering a general certification,several specialty boards offer certificates in sub-specialty areas. Altogether, there are 49 subspe-cialty areas of the 23 specialty boards recognizedby the American Board of Medical Specialties (seetable 10-1). Qualifications in these subspecialtyareas are recognized by certificates of special oradded qualifications. s Within the 17 specialtyboards recognized by the Advisory Board for Os-teopathic Specialists, there are 40 subspecialtyareas (see table 1o-2).

The 13 studies reviewed for this chapter per-tain to board certification by the 23 specialtyboards recognized by the American Board ofMedical Specialties.b This chapter evaluateswhether certification by these boards or practic-ing in one’s area of specialty training are valid in-dicators of the quality of a physician’s perform-ance. Although the literature and this chapterexamine physician specialization among allopathicphysicians [Doctors of Medicine (M. D.s)], the dis-cussion and conclusions drawn here are generally

applicable to both allopathic and osteopathic phy-sicians [Doctors of Osteopathy (D. O.S)].

The next two sections of this chapter evaluatethe reliability and validity of physician speciali-zation as a measure of the quality of care. Thethird section considers the feasibility of using phy-sician specialization as a quality indicator. Thefinal section of the chapter presents conclusionsabout physician specialization as an indicator ofquality. That section also discusses methods toimprove the reliability and validity of physicianspecialization as a quality indicator and considersalternatives to better assure consumers of theacceptability of a physician’s quality of care.

‘According to the American Board of Medical Specialties, “It isnot necessary for physicians in a recognized specialty to hold spe-cial certification in a subspecialty of that field in order to be con-sidered qualified to include aspects of that subspecialty within a spe-cialty practice. Such special certification is a recognition ofexceptional expertise and experience and has not been created tojustify a differential fee schedule or to confer other professional ad-vantages over other diplomats not so certified” (18).

bAdditional details on the studies revien’ed can be found in OTA’stechnical working paper “Physician Specialization as an Indicatorof Quality: An Evaluation of the Literature” (434).

212

Table 10-3.—lndependent Boards That Certify Physicians”

The following boards are called “American Board of “ unless otherwise designated, and each claims to certify physicians.

Abdominal SurgeonsAcupuncture MedicineAddictionologyAesthetic Plastic SurgeryAlcoholism and Other Drug DependenciesAlgology (Chronic Pain)Ambulatory AnesthesiaBariatric MedicineBloodless SurgeryChelation TherapyChemical DependenceClinical ChemistryClinical EcologyClinical NutritionClinical PharmacologyClinical ToxicologyCosmetic Plastic SurgeryCosmetic SurgeryCouncil of Non-Board-Certified PhysiciansDisability Evaluating PhysiciansElectroencephalographElectromyography and ElectrodiagnosisEpidemiology (College)Facial Cosmetic SurgeryFacial Plastic SurgeryForensic PsychiatryForensic ToxicologyHead, Facial & Neck Pain & TMJ OrthopedicsHealth PhysicsHomeotherapeuticsInsurance MedicineInterventional RadiologyLaser SurgewLaw in MedicineLegal Medicine

Malpractice PhysiciansMaxillofacial SurgeonsMedical Accreditation (American Federation for)Medical GeneticsMedical HypnosisMedical Laborato~ ImmunologyMedical Legal Analysis in Medicine & Surge~Medical Legal &Workers Compensation Medicine & SurgeryMedical Legal ConsultantsMedical MicrobiologyMedical Preventics (Academy of)Medical PsychotherapistsMedical ToxicologyMicrobiology (Medical Microbiology)Milita~ MedicineNeurological Orthopedic SurgeryNutritionOtorhinolaryngologyPlastic Esthetic SurgeonsPrison MedicinePsychiatric MedicinePsychiatry (American National Board of)Psychoanalysis (American Examining Board in)Psychological Medicine (International)Quality Assurance and Utilization ReviewRadiology and Medical ImagingRingside Physicians and SurgeonsSkin SpecialistsSpinal Cord InjuryToxicologyTrauma SurgeryTropical MedicineUltrasound TechnologyUrologic Allied Health Professionals

aThe b~ard~ listed below are not members of the American Board of Medical Specialties and are not recognized by the Advisow Board for Osteopathic Specialists.

SOURCE: American Board of Medical Specialties, “Self-Designated Boards,” Evanston, IL, June 16, 1967.

RELIABILITY OF THE INDICATOR

Advances in medical science and technologicalchanges are inherent in medical care. Unless aphysician’s knowledge and skills in a specialtyarea are periodically updated or assessed, physi-cian specialization as represented by board cer-tification or practicing in one’s area of specialtytraining may bean unreliable measure of the qual-ity of a physician’s performance over time.

In the past 10 years, there has been an increas-ing trend towards recertification by specialtyboards. The American Board of Medical Special-ties encourages the periodic reassessment of phy-sicians and has written guidelines on recertifica-

tion for the specialty boards to use (18). So far,15of the 23 specialty boards recognized by theAmerican Board of Medical Specialties haveadopted or decided to adopt time-limited certifi-cation, and 1 board offers voluntary recertifica-tion (see table 1o-4). Among these boards, the in-tervals between evaluations range from 6 to 10years.

Without a recertification process, there is noguarantee that physicians have maintained thesame level of skills and knowledge they demon-strated for their initial certification. Thus, boardcertification of a physician who was certified 20

213

iial.--

0a)

(%

1y“0

c0.—

c2. i

i!a

tiL

>

0

. .. .. .. .. .. .. .. .. .. .. .a)c.—

. .. .. .. .. .. .. .. .. ... ... .. .

. .. .. ... .. .. .. .. .. .

. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .:*:

. .

. .. .

.C0“ .-

:3.4-

.=

“z;4!:$:73.C.(Q

r=om==trbcoco c o a lUlmcnmul---77

a)a)a) a l a )

( n Co(nct)fn

>>>>>

( n U)ulcom

> > > > >

i’a)UI

. .

84-752 0 - 88 -- 8

214

years ago may not indicate the same level of com-petence as the same board certification of a phy-sician who has been assessed more recently.7 Thisvariability in the significance of board certifica-tion over time reduces the reliability of the useof board certification as an indicator of quality.Recertification requirements would increase thereliability of board certification as an indicatorof quality by making its significance more con-stant over time.

7Physicians who were certified before their respective boards ini-tiated recertification policies have not been subject to recertification.

VALIDITY OF THE INDICATOR

Is Physician Specialization aReasonable Indicator of Quality?

Intuitively, certification by a board recognizedby the American Board of Medical Specialties isa valid indicator of the quality of a physician’smedical performance. Board certification indicatesthat a physician has met a specified set of require-ments and has performed up to a certain level ona qualifying examination in the specialty area. Itmakes sense that physicians who have had a cer-tain amount of training in a specialty area wouldperform better than physicians who have had lessor no training in the field.

Examples of the current uses of board certifi-cation demonstrate a general acceptance of its useas an indicator of quality. Patient brochures andother articles prescribing how to choose amongphysicians (published by hospitals or consumerhealth information centers) encourage consumersto use board certification as a measure of qual-ity. Although the Joint Commission on Accredi-tation of Healthcare Organizations’ standards forhospitals do not require that board certificationbe used for granting hospital privileges to physi-cians, the standards do state that “specialty boardcertification is an excellent benchmark for thedelineation of clinical privileges” (330).

The fact that a physician is practicing medicinein the area in which that physician has beentrained is also intuitively valid as an indicator ofquality. It makes sense that a specialist practic-

A major issue in implementing recertificationprocedures is whether medical specialty boardscan truly measure physician competence. Muchof the opposition to recertification arises notbecause of recertification’s quality assessmentmechanism, but rather because of doubts aboutwhether current examination procedures are anaccurate measure of clinical skills. The AmericanBoard of Medical Specialties maintains that “recer-tification will be focused on performance assess-ment instead of the broad cognitive examinationsused for primary certification” (365).

ing in the area in which he or she has been trainedwould provide better quality care than, say, aboard-certified specialist who is not practicing inhis or her area of specialization.

Does the Board Certification ProcessAccurately Reflect a Physician’sCompetence?

Many specialty boards limit their evaluationsof a physician to evaluation of the physician’sknowledge of the pertinent subject matter; theyoften do not evaluate a physician’s interpersonalskills or skills used to technically apply theirknowledge. Part of the reason for this situationmay be that knowledge is fairly easy to test. Bycomparison, a physician’s interpersonal skills arerather difficult to measure. Judgment and clini-cal skills are other qualities that are difficult tomeasure, yet are of utmost importance for deter-mining the competence of a physician. The cer-tification process of the American Board of In-ternal Medicine does include a form that asksseveral questions about the interpersonal skills aphysician demonstrated during his or her periodof residency (16). This form (“Evaluation Formfor Clinical Competence”) is sent to the programdirector of each physician’s residency program.

At any rate, many of the aspects of a physi-cian’s practice mentioned in the last paragraph areonly proxy measures of physician competence.Burg and Lloyd emphasize the importance of the

215

specialty boards’ defining competence to ensurethe comprehensiveness of their evaluationmeasures:

Definitions of competence within a medicalspecialty discipline serve the purpose of provid-ing a first step toward the development of morevalid procedures for the certification of special-ists, This is because one form of validity, con-tent validity, calls for a comprehensive delinea-tion of the skills and abilities the board isattempting to measure. Ideally, measures of com-petence in a specialty should sample from thecomponents of competence identified as impor-tant by members of that specialty (110).

To evaluate the competence of a physician withpatients, it is important to use direct assessmentmethods. Evaluating a medical audit of pediatricperformance, a study by the National Board ofMedical Examiners demonstrated that for manycommon diagnoses, cognitive certifying exams donot test the same content area covered by directaudit assessments of clinical performance (612).Although several specialty boards utilize tech-niques in which a physician’s practice is evalu-ated more directly by requiring information forspecific cases treated, these methods have not beenadopted by a majority of the boards.

Further complicating the issue of validity andboard certification is that board-certified physi-cians’ practices are not necessarily limited to thearea in which they have been certified. In fact,statistics from the AMA’s Physician Character-istics and Distribution: 1986 demonstrate that 5percent of board-certified physicians are not cer-tified by the board corresponding to the primaryarea of their practice (35). This fact makes it im-possible for board certification, which assesses aphysician’s knowledge and skills in one specialtyarea, to be an accurate reflection of all specialtyareas in which a physician may practice. Physi-cians who are practicing in a specialty area inwhich they have not been board certified may not,outside of medical school or residency training,have had their skills assessed in that particulararea.

Does Physician SpecializationAccurately Predict a Physician%Quality?

A review of the 13 available studies on physi-cian specialization and quality (see table 105)gives one little confidence that board certificationaccurately predicts which physicians will providehigh-quality care and which will not (220,477,481,604).

One explanation for board certification’s lowpredictive power could be weaknesses in avail-able studies. The studies may have too manymethodological problems—small sample size, abiased physician sample (if it includes physicians’volunteering to participate), or no inclusion ofpatient-mix indices and severity-of-illness adjust-ments—to accurately assess the relationship be-tween board certification and physician per-formance.

Other issues may affect the accuracy of boardcertification in predicting that physicians will pro-vide high-quality care. If board certification werea mandatory process as opposed to a voluntaryone, it would be more likely that non-board-certified physicians had failed the certificationprocess and were substandard practitioners. Sinceboard certification is voluntary, however, somephysicians who are as well qualified as board-certified physicians may simply choose not to sub-stantiate their training through certification. Ofcourse, a percentage of non-board-certified phy-sicians are physicians who have attempted andfailed the certification process. Other non-board-certified physicians may not have met the board’smandated prerequisites to be eligible for certifi-cation. In the studies OTA reviewed, however,the percentage of unqualified non-board-certifiedphysicians was not high enough to affect the per-formance results in favor of board-certified phy-sicians.

Another possible explanation for board certifi-cation’s not being predictive of a physician’s per-formance in practice could be the imprecise evalu-ation procedures used by the specialty boards tocertify physicians, All of the 23 boards recognizedby the American Board of Medical Specialties use

Table 10.5.–Studies on Physician Specialization Reviewed by OTA

Physlclan Condltlons/sDeclaltles procedures Level of Performance Adjustments for Sample

Studya included studied aggregation measure patient characteristics size Results

By dlagnosls Performance on Amerl- No adjustments 185 Board-certlfledRamsey, et al General Internists1986 (510)

5 Condmons● Diabetes● Hypertension● Respwatory mfectlon● Urinary tract mfectlon● Ischem!c heart dwease

Kelly and Helhnger, Primary surgeons 4 Condmons1986 (347)

Strauss, et al1986 (604)

Goldberg andDletnch, 1985(257)

● Colon and rectal● Neurologlc● Orthopedic● Thoraclc● General

Pulmonary speclaltstsGeneral mtermstsFamily prachtloners

Famdy physiciansGeneral mterrustsMedical subspeclallstsg

● Stomach operation withcancer diagnosis

● Stomach operahon withulcer dlagnosls

● Intestinal operation withcancer of the colon orrectum

● Blood vessel surgerywith abdommal aneurysm

Chrome obstructwe pulmo-nary lung dtsease

General primary care vrats

By surgicalprocedure

By dlagnosls

By specialty status

can Board of InternalMedlcme ExamProcess measures

● Items relatlng todiagnosis

● Treatmentstrateglesc

● Monltorln9strategies

Outcome measures e

. Level of bloodpressure controlover a period oftime

● Glucose control● Exercise tolerance● Adverse outcomesf

Patient satisfactionEvaluahons by profes-sional associatesPostsurglcal mortaltty

Outcome measures● Pulmonary function● Functional ablhty. Instltutlonalized

days● Morfahty

Contmulty of careh

Adjustments made forseverity of illness, age,sex, and number of di-agnoses

Adjustments made forseverity and patientcharacteristics

Adjustments made forpatient age, sex, andyears with primaryphyslclan

75 Non-board-certlf led

1 241 Total surgeons

96 Total physicians

40 Total phystclans

●

●

●

Board-certlfled physiciansperformed slgmflcantly bet-ter on certlflcatlon examNo slgnlflcant differenceswere found between board.cert!fled and non-board-certrfled physicians forprocess or outcome meas-ures for any condltlonNo difference was foundbetween mean patientsatrsfachon score for certl-fled and non-board-cerlfledphysicians Board-cerhfledphysicians received slgnlfl-cantly higher ratings fromprofessional associates Inmost categories

Board-cerhfled surgeons werefound to be associated wthlower patient mortalty rates

No slgmflcant differenceswere found between thegroups of speclahsts for out-come measures

No slgmflcant differenceswere found In continuity ofcare provided by sub-speclahsts and generahstsSubspeclahsts providedhigher levels of contmuty topatients with a dlagnosls ly-ing wlthln their areas of ex.pertlse. but only at highuttllzatlon levels

217

0 O-J

al

In

WY

Table 10-5.-Studies on Physician Specialization Reviewed by OTA (Continued)

Physic]an Conditions/specialties procedures Level of Performance Adjustments for Sample

Study a included studied aqgreqatlon measure Datient characteristics s]ze Results

Rhee, 1977 (515) 18 Specialtlesn 20 Diagnoses

Hulka. et al , Family/general practitioners 4 Conditions.1976 (308) Intermsts c Adult-onset diabetes

Pedlatrlclans melhtusObstetricians ● Congestwe heart failure

● Normal pregnant woman. Normal newborn during

the first year of hfeRhee. 1976 (514) 18 Specialtiesn 20 Diagnoses

Compared across di-agnostic categories

By dlagnosls

Compared across di-agnostic categories

PP1/Physician Perfor-mance Score’

Mimmum explicit comsensus criteria formanagement protocol

PP1/Physician Perfor-mance Score’

No adjustments 321 Specialists133 General practi.tioners

No adjustments 34 Family physicians11 Intermsts8 Pediatricians8 Obstetricians

No adjustments 321 Specialists133 General practi-tioners

Board-certified/board-eligiblephysicians had higher perfor-mance scores than generalpractitioners or self-designated specialists.Pediatricians and obstetri-cians performed better for in-fancy and pregnancy. Therewere no differences amongphysicians in performance fordiabetes mellitus or conges-tive heart failure,Physician specialists werenot found to relate to qualityof t)hvsician performance.

aNu~&r~ in ~ arent he5e~ refe r to numbered entries in the list of references at the end of this report.bThi~ in~lude~ determination of underlying factors, determination of severity of the condition, and determination Of comorbid conditions.

cThis includes complexity of therapy and avoidance of potentially harmful therapy.d Frequency of foilowup office visits.e Measured for chronic diseases.fHypotension, hypoglycemia, hypokalemia.gMedical subspecialties included rheumatology, cardiology, hematologyloncology, and gastroenterology.hDefined as the p ropo~ion of visits that patients received from their prima~ physicians.

‘These diagnoses included chronic heart failure, acute myocardial infarction, chronic obstructive pulmonav diseases, diabetes mellitus, acute bacerial pneumonia.JThese diagnoses included hypetiension, diabetes mellitus, chronic heart failure, angina pectoris, chronic obstructive pulmonary disease, osteoarthritis.kperiodic adult medical examination, periodic gynecological examination, periodic pediatric medical examination, therapeutic use Of Chloramphenicol, keflex, digitalis preparation, and prednisone, aneMia,

essential hypertension, chronic heart disease (arteriosclerotic, hypertensive rheumatic), vulvovaginitis, acute urinary tract infection, chronic or recurrent urinary tract infection.IThe physician pe~ormance index (ppl) is a process measure Of Performance developed by payne and Lyons in lg~. Explicit process criteria for a variety Of diagnoses and exalllinations Were developed ifl

1974 by panels of practicing specialists. Physician performance was measured according to the level of physicians’ compliance with these explicit criteria. The criteria were weighted by the physician panelso that a single PPI score for each diagnosis or examination was generated. A Physician Performance Score represents a physician’s average PPI score over all of his or her treated cases.

mThe surgical Procedures included were gastric surge~ for ulcer, selected surgery of the biliary tract, surgery of large bowel, appendectomy, splendectomy, abdominal hysterectomy, va9iflal hysterectomy,

craniotomy, amputation of lower limb (ankle to hip), repair of fractured hip, arthroplasty of the hip, lumbar Iaminectomy (with and without fusion), pufmonaw resection, prostatectomy, and selected surgeryof abdominal aorta and/or iliac arteries.

nAnesthesiology dermatology, internal medicine, necrologic Surgev, Obstetrics/gynecology, ophthalmology, orthopedic surgery, otolaryngotogy, pathology, pediatrics, Plastic sur9erY, Preventive medicine!psychiatry and neurology, radiology, general surgery, thoracic surgery, urology, general practice.

SOURCE: Office of Technology Assessment, 1988.

I

I

219

written examinations with multiple-choice ques-tions to evaluate physicians, and 16 of the boardsrequire oral examinations. Nine of the boards re-quire physicians to submit a case list for recer-tification. 8 In the oral examination for recertifi-cation, these nine boards ask physicians abouttheir management of several cases. The particu-lar cases discussed during the examination arepicked by the specialty board from the case listsubmitted by the physician. Four boards requireinformation from the physician’s medical records(e.g., patient history, physician findings, andtreatment outcome) for a specified number ofcases (365).9

The relationship of the written and oral exam-ination used by the American Board of MedicalSpecialties boards to actual physician perform-ance is ambiguous. As noted earlier, test questionsare more likely to measure what a physicianknows about a certain field than the physician’sactual clinical performance. If tests do not includean assessment of a physician’s clinical competence,they may not provide an accurate prediction ofa physician’s performance. Studies have fre-quently demonstrated large discrepancies betweenlevels of knowledge and levels of clinical perform-ance (552).

Unfortunately, the assessment of a physician’sclinical performance is not as adaptable to the for-mat of a written examination as is an assessmentof a physician’s knowledge. Direct performanceassessment instruments may provide a more ac-curate reflection of a physician’s clinical compe-tence and may have more predictive validity thanwritten examinations. Although the 1975 Amer-ican Board of Medical Specialties guidelines sug-gested that practice audits and performance evalu-ations should be part of the certification process,only four member boards—the American Boardsof Family Practice, Obstetrics and Gynecology,Surgery, and Thoracic Surgery—have so far

8A case list specifies the number of diagnoses/procedures treatedby a physician. The nine boards requiring case lists for evaluationare the American Boards of Colon and Rectal Surgery, Neurologi-cal Surgery, Obstetrics and Gynecology, Orthopedic Surgery,Otolaryngology, Plastic Surgery, Surgery, Thoracic Surgery, andUrology.

‘The four boards requiring information on cases treated are theAmerican Boards of Obstetrics and Gynecology, Family Practice,Surgery, and Thoracic Surgery.

adopted such techniques. The American Board ofFamily Practice requires an office record reviewas part of its recertification process.10is board’sparticular methods of assessment increase the va-lidity of its certification process, but they may givetoo much control to the physicians being evalu-

lOEvery 6 t. 7 years, physicians certified by the American Boardof Family Practice are required to undergo an office record reviewas part of their recertification process. This process involves eachphysician’s choosing two individual patient records for each of threedifferent conditions. The three conditions are chosen by the physi-cian from a list of 20 possible conditions decided upon by the Boardof Family Practice. The board sends the physician an extensive ques-tionnaire and scansheet, to be filled out for each condition. Ques-tions pertain to patient history, physical exam, medicationsprescribed, and diagnostic procedures. After receiving the completedscansheet, the board analyzes and scores it by computer. The scoresare based on the physician’s compliance with explicit process cri-teria (determined by the board) for the diagnosis and treatment ofspecific conditions. If the scansheets reveal that the physicians arenot handling their patients as the board’s standards dictate, the Boardof Family Practice gives physicians the opportunity to send in ad-ditional patient records until their scansheets are approved. Theboard randomly selects physicians to be spot checked by requiringthem to send in a specific patient record and comparing this recordto the physician’s own scansheet on that same patient. Accordingto the Board of Family Practice, conflicts between a physician’s scan-sheet and the spot checked records rarely occur (490).

Photo credit” /nte///genf /mages

By taking physicians through multiple stages of cl inicalproblem-solving, computer-based methods forassessing a physician’s performance, such as theDxTer system shown above, may be more predictive

of a physician’s quality of care thanwritten examinations.

220

ated. Physicians may be biased in their selectionof records to send to the board for evaluation .11

Other approaches to office practice evaluationentail onsite reviews of actual patient records. TheCollege of Family Physicians of Canada, whichcombines the functions of a certifying board anda professional association, has developed a med-ical practice quality assessment model based onchart abstractions (see ch. 7 for a description). Thecollege plans to apply these techniques for usewithin certification examinations for “practice-eligible” candidates and for use in practice accred-itation (85).12 Similar onsite office evaluations arebeing performed by the College of Physicians andSurgeons of Ontario, the medical licensing bodyof Ontario, 13 and thepark Nicollet Medical Cen-ter in Minneapolis, Minnesota .14

Computer-based testing is another example ofa technique to assess physician performance.Computer-based assessment techniques that re-produce a physician’s clinical practice can providean interactive representation of patient/physicianencounters. ” Although such techniques may be

lllf ~hY~iCianS choose only their best records, they will not beproviding a representative sample of office records. Allowing a phy-sician to send in records until the records are approved may alsobias the sample.

lzpractice accreditation, as opposed to certification, involves therandom assessment of practicing physicians regardless of their spe-cialty status.

lJEach year since 1981, a total of 200 specialists and general Prac-titioners out of all Ontario’s physicians have been randomly selectedto undergo a mandatory office evaluation, entitled the Peer Assess-ment Program. These evaluations are fairly subjective and basic inscope and structure. They are performed mainly for finding physi-cians with significant deficiencies in their records or patient care.Certain physician groups assumed to be more at risk of providingpoor quality care, such as physicians over the age of 70 or thosein solo practice, are specially targeted for review (139). Quebecslicensing body is involved in similar office reviews of physicianswho are reported as needing special attention (54).

14A PrimaV Care practice Profile was developed to SUPPIY in-formation to the MedCenters Health Plan about the quality of careprovided by family practitioners in various settings. The PrimaryCare Practice Profile is used by an audit team of three nurses andone physician during an onsite visit and incorporates a diagnostic-specific chart to evaluate medical records in physicians’ offices (417).

lsT’he National Board of Medical Examiners has developed aComputer-Based Exam that provides electronic patient simulation(456). The computer presents X-rays and electrocardiograms for ex-ample, and allows the physician to order any test or procedure re-quired. The Computer-Based Exam keeps track of the results in termsof time, costs, and patient outcome. One physician score is produced.This exam has had the most experience, but other such patient simu-lation devices also exist. A DxTer system developed by David Al-len of Intelligent Images allows the viewer to see the image of a person

more predictive of a physician’s quality of carethan written examinations, their relative levels ofaccuracy remain to be validated through research.

The variations in the literature evaluating therelationship between board certification and qual-ity of care reviewed for this OTA report are areflection, in part, of the limited predictive powerof various methods of assessing physician per-formance. The studies vary in regard to the per-formance measure used to assess a physician’squality of care. A physician’s performance accord-ing to one technique may be different from hisor her performance as measured by anothermethod. In a study evaluating various proceduresused to assess quality of care, Brook and Appelfound the results of quality assessment to be de-termined by the method used (100). Between 1.4and 63.2 percent of patients were determined tohave received satisfactory care, depending onwhich method was used. Several studies reviewedfor this report use the Physician Performance In-dex (PPI),l’ a process measure of performance de-veloped by Payne and Lyons (242,481,514,515,516). Other studies use different process measuresor various outcome measures of performance(220,257,347,385,510,552,604) .

Although there are several inconsistent results,the literature suggests that specialists practicingin the area in which they have been trained pro-vide a higher quality of care than specialists prac-ticing outside their area of training. Restrictingtheir scope of practice presumably enables thesespecialists to treat patients in defined areas ger-

brought into an emergency room or clinical office (364). Althoughexperwve co produce, this system provides a very realistic simula-tion of clinical practice. Another uncued system developed by Hadessof the National Library of Medicine allows the physician to speakto the patient on the screen and receive a response from the patient(364).

16Explicit process criteria for a variety of diagnoses and exami-nations were developed in 1974 by panels of practicing specialists.Panels developing criteria for infectious disease, heart disease, hyper-tension, gynecology, and pediatrics were made up of specialists prac-ticing in the corresponding fields. Physician performance was meas-ured according to the level of physicians’ compliance with theseexplicit criteria. The criteria were weighted by the physician panelso that a single PPI score for each diagnosis or examination wasgenerated. Several studies are one step removed from comparingphysicians’ PPI scores, and go further to derive a performance scorefor each physician. A Physician Performance Score is calculated bytaking a mean of the standard scores for all the cases a physicianhas treated.

221

mane to their experience and training. One studyevaluating performance differences betweenboard-certified and non-board-certified physiciansand between physicians practicing in their areaof specialty training and physicians not practic-ing in their area of specialty training affirms thegreater predictive power of physicians practicingin their area of training as an indicator of quality(481).

Methodological weaknesses in available studiesevaluating physicians practicing in their area ofspecialty training may explain some of the varia-tions in the results of the studies (242,515,604).In some studies, the inclusion of self-designatedspecialists in the category of physicians practic-ing in their area of specialty training may haveconfounded the performance scores of thesespecialists, seemingly limiting the predictive powerof practicing in one’s area of training as an indi-cator of the quality of care.

The use of physician specialization as an indi-cator of quality is more valid when its meaningis clarified by either of the operational definitionsused in this chapter than when its meaning is leftundefined. As noted earlier in this chapter, phy-sicians may designate themselves specialistsregardless of the amount of training they have hadin the field or whether they are currently practic-ing in that specialty. There are no regulations orguidelines that limit who may call themselvesspecialists. Because specialists are reimbursed byMedicare at a higher rate than nonspecialists,there are financial incentives for physicians to la-bel themselves specialists even if they have notreceived specialty training. Consequently, phy-sician specialization (unless a physician’s trainingand qualifications can be identified with certainty)allows for a wide range of interpretations and isnot an accurate predictor of quality.

Is the Use of Physician Specializationas a Quality Indicator GeneralizableAcross Specialties?

Board certification by one specialty board sayslittle about a physician’s qualifications in anotherspecialty area. Owing to the variation in meth-ods of practice across medical fields and to thewide range of certification techniques utilized by

each of the 23 specialty boards that belong to theAmerican Board of Medical Specialties,17 boardcertification can be defined only as it applies toa specific board. Since many available studies as-sess only one type of board-certified specialist,their results address only that particular specialty.One study, for example, assessed only the per-formance of physicians certified by the AmericanBoard of Surgery as compared to surgeons notcertified by that board (347).

Along the same lines, a major limitation witha number of the studies OTA reviewed is that theyevaluate board-certified physicians or specialistspracticing in their area of training in the aggregaterather than by specialty category. Aggregating thescores of physicians certified by different specialtyboards and practicing in different specialty areasmay mask subtle differences among specialty cat-egories. Furthermore, specialists from each of thespecialty groups may not be equally representedin the sample. One study with these problemscompared board-certified and board-eligible phy-sicians to general practitioners and self-designatedspecialists. Performance scores for the physicianswere aggregated by board certification or eligi-bility status and by self-designation or generalpractice instead of being interpreted separately byspecialty category (515).

There is an inevitable trade-off in generaliza-bility between studies that assess physician per-formance with respect to many diagnoses andstudies that assess performance with respect toonly one diagnosis. Internal medicine, a specialtystudied by Sanazaro and Worth (552), is an espe-cially broad specialty field and includes many cat-egories of subspecialties. An evaluation of sub-specialists of internal medicine with respect totheir performance of a subspecialty procedure,may be a more accurate approach to measuringthe quality of a specialist’s performance in a par-ticular specialty, but would not necessarily begeneralizable to other subspecialties.

The generalizability of physician specializationalso has limitations within an area of practice. A

lprhe boar& may also differ in the required amount of time foreducation in a residency program and the specified stage for apply-ing for certification—whether during or shortly after residency orafter a specific amount of experience in the field.

222

physician’s performance with respect to onediagnosis or procedure is not necessarily gener-alizable to other conditions and procedures. A cer-tifying process that measures a physician’s per-formance by assessing his or her skills for one orseveral diagnoses may not adequately serve as anindicator of the physician’s performance overall(552). Thus, one cannot assume that a physicianwho performs poorly or well in the case of oneor several diagnoses will perform poorly or wellin all diagnoses.

In its broadest sense, certification by a boardrecognized by the American Board of MedicalSpecialties is an indication that a physician hasmet certain requirements and has passed certain

examinations to practice in a specialty area. Un-less board certification as an indicator of qualityis used in reference to a specific specialty area,the generalizability of board certification as an in-dicator is unclear.

If one defines a specialist as a physician whois practicing in the area in which he or she hasbeen trained, the link between previous trainingand specialty practice is set up. Unlike the board-certification literature, available studies on thequality of physicians practicing in their area ofspecialty training generalIy do classify speciali-zation at the level of the individual specialty(242,257,385,481,604),

FEASIBILITY OF USING THE INDICATOR

Information on physicians’ board-certificationstatus and specialty designation is already avail-able to individuals and organizations and easilyunderstood by the public. The American Medi-cal Directory, provided by the AMA, contains in-formation on the American Board of Medical Spe-cialties board certification, on self-designatedprimary and secondary practice specialties (see ta-ble 10-6), and on dates of recertification for allU.S. physicians alphabetically and geographicallyby city and county. This 4-volume directory ispublished every 2 years and is available in pub-lic libraries. The ABMS Compendium of Certi-fied Medical Specialists is a 7-volume publicationthat lists all of the specialists certified by the23 boards that belong to the American Board ofMedical Specialties. Although this compendiumis also published biennially, The ABMS Compen-dium Supplement is published in between pub-lication dates of the original volumes and bringsthe lists of certified specialists more up to date.

The Directory of Medical Specialists, publishedby Marquis Who’s Who and available in mostpublic, hospital, medical, and university libraries,also contains information on board certification.This publication may be incomplete, however, be-cause some boards do not supply informationabout new board-certified specialists to thissource.

Data on American Board of Medical Special-ties board-certification status and self-designatedpractice specialties are also available from theAMA Physician Masterfile. This computer database contains current and historical informationon all physicians practicing in the United Statesand on those U.S. physicians temporarily prac-ticing overseas. The data are provided to theMasterfile by primary sources. ’8 The AMA willprovide a computerized printout containing back-ground information on any U.S. physician toorganizations such as hospitals, State licensingboards, medical schools, and medical societies forthe purpose of credentials verification. Althoughinformation in the AMA Masterfile is primarilyintended to assist organizations in verifying thecredentials of physicians, the data are also avail-able to individuals (672).

The general availability of information onboard certification and specialty designationmakes it apparent that confidentiality of this in-formation is not an issue. Consumers may obtaininformation on board certification from county

18The American Board of Medical specialties provides informa-tion on board certification, and the Federation of State MedicalBoards (see ch. 6) provides information on final disciplinary actionsby State boards that affect medical licensure.

223

Table 10-6.—Self-Designated Practice Specialties Recognized by the American Medical Association

Adolescent MedicineAerospace MedicineAllergyAllergy and ImmunologyAnesthesiologyCardiovascular DiseasesCritical Care MedicineDermatologyDermatopathologyDiabetesDiagnostic Laboratory ImmunologyEmergency MedicineEndocrinologyFacial Plastic Surgery, OtolaryngologyFamily PracticeGast roenterologyGeneral PracticeGeneral Preventive MedicineGeriatricsGynecological OncologyGynecologyHematologyImmunologyImmunopathologyInfectious DiseasesInternal MedicineLegal MedicineMaternal and Fetal MedicineMedical Microbiology

Neonatal-Perinatal MedicineNephrologyNeurologyNeurology, ChildNeuropathologyNuclear MedicineNuclear RadiologyNutritionObstetricsObstetrics and GynecologyOccupational MedicineOncologyOphthalmology

Physical Medicine and RehabilitationPsychiatryPsychiatry, ChildPsychoanalysisPublic HealthPulmonary DiseasesRadiation OncologyRadiologyRadiology, DiagnosticRadiology, PediatricReproductive EndocrinologyRheumatologySurgery, Abdominal

Otolaryngology Surgery, CardiovascularPathology, Anatomic/Clinical Surgery, Colon and RectalPathology,Pathology,Pathology,Pathology,Pathology,Pathology,PediatricsPediatric Allergy Surgery, ThoracicPediatric Cardiology Surgery, TraumaticPediatric Endocrinology Surgery, UrologicalPediatric Hematology-Oncology Surgery, VascularPediatric NephrologyPediatric PulmonologyPharmacology, Clinical

Anatomic Surge~, GeneralBlood Banking Surgery, HandChemical Surgery, Head and NeckClinical Surgery, NeurologicalForensic Surgery, OrthopedicRadio isotopic Surgery, Pediatric

Surgery, Plastic

SOURCE American Medical Association, Division of Survey and Data Resources. “intended Use of AMA Physician Masterfile Codes for Self-Designation of PracticeSpecialty,” Chicago, IL, January 1987

medical societies or from the American Board ofMedical Specialties. Directly asking a physicianor requesting the information from the hospitalwhere the physician has staff privileges are fur-ther possibilities.

Determining whether physicians are practicingin the area of their training is not as easy for con-sumers as is determining a physician’s board-certification status, largely because of inconsisten-cies regarding the qualifications and range of prac-tice of a specialist. One study demonstrates asignificant disparity between the number of phy-sicians listed under the physician specialty head-ings in the Yellow Pages and the number of phy-sicians listed in the Directory of MedicalSpecialists as board certified (511). In the currentsystem, physicians can designate as their specialtyan area in which they have had no or little train-ing. Thus, consumers could be confused or misledby specialty designations. In most States exceptfor Maryland, there is no system in place to sub-

stantiate that specialists have been trained in theirpractice areas.19

As noted earlier, requirements may vary sub-stantially among the many boards that claim tocertify physicians. For some boards, a set fee maybe the only prerequisite for certification. A con-sumer uninformed about different types of cer-tification may falsely assume that certificationfrom a particular board is significant.

]gIn 19g5, the MaVland Genera] Assembly enacted legislation pro-hibiting physicians from presenting themselves to the public asspecialists unless identified by the Maryland Board of Medical Ex-aminers. Although any physician licensed in Maryland may applyfor specialty designation, only physicians certified by the AmericanBoard of Medical Specialties are allowed automatically to publiclydesignate themselves as specialists. Other physicians must completea form outlining specific training and experience in the requestedspecialty. The forms are reviewed for completeness and then referredto a multispecialty peer-review committee of the State medical so-ciety for evaluation. The Maryland Board of Medical Examinersmakes the final decision on the basis of recommendations from thecommittee. The identification is permanent, and the physician maypublicize himself or herself in that area of specialty (Annotated Codeof Maryland, 10.32.09).

224

CONCLUSIONS AND POLICY IMPLICATIONS

Certification by the American Board of Medi-cal Specialties or the Advisory Board for Os-teopathic Specialists enables consumers to iden-tify those physicians who meet a standard set ofqualifications, including specific training in a spe-cialty field and passing a certification examina-tion. Consumers should be made aware, however,that such certification is not a reflection of theamount of practical experience a physician has ina specialty area or an adequate measure of dem-onstrated proficiency in the field

The low predictive power of current methodsof assessing physician performance limits the va-lidity of board certification as an indicator of high-quality care. Furthermore, board certification isnot an indicator of quality that is generalizableacross specialties, diagnoses, or procedures. Anaccurate measurement of a physician’s perform-ance requires that many diagnoses in the physi-cian’s specialty area be evaluated and interpretedindividually.

Consumers should also be made aware that inmost States, the fact that physicians designatethemselves specialists does not necessarily meanthat they have had advanced training in the spe-cialty field corresponding to the area of their prac-tice. To establish whether physicians are practic-ing in an area that correlates to their training, aconsumer must verify that a physician’s desig-nated areas of practice match published listingsof board-certification status. If a physician “spe-cialist” is not board certified in any specialty field,there is no reliable method for a consumer to ver-ify that the physician has had advanced specialtytraining.

To strengthen the validity of the relationshipbetween board certification and clinical perform-ance would require improving the reliability andvalidity of the specialty board evaluation proce-

ZOAlthough Medicare’s conditions of participation for hospitalsonce used board certification as a requirement for participation, theconditions have since changed (FR 22021-22023, 1986). The currentconditions of participation with respect to a hospital’s medical staffrequire the hospital governing board to “ensure that under no cir-cumstances is the accordance of staff membership or professionalprivileges in the hospital dependent solely upon certification, fel-lowship, or membership in a specialty body or society” (641).

dures. The American Board of Medical Special-ties recognizes this need and is involved in vari-ous studies working towards improving thepredictive power of its specialty boards’ evalua-tion processes (21).

The recertification of physician specialistswould increase the reliability and the validity ofcertification. Recertification procedures couldmaintain the significance of board certificationover time and would encourage physicians to up-date their skills. Although 16 boards recognizedby the American Board of Medical Specialtieshave adopted some form of recertification, 7boards still have not. Directories that list certifi-cation and recertification dates for specialists arepublicly available, but consumers may not beaware of them (17,31).

Unfortunately, recertification would assure thepublic of the quality of care provided by onlythose physicians who hold a specialty certificate.Mandatory recredentialing (using valid perform-ance assessment methods) for all licensed physi-cians would establish a more comprehensive sys-tem and would reassure the public of a physician’sclinical competence21

Performance assessments through medical au-dits would increase the predictive power of thecertification/ recertification process. The shiftfrom a knowledge-based to a performance-basedassessment, by making the process more a reflec-tion of a physician’s actual practice, would in-crease the validity of board certification. Langs-ley of the American Board of Medical Specialtieswrites, “Competence represents the potential forperformance, but only performance assessmentdemonstrates that such potential is used in actualprofessional practice” (364).

ZIIn a response to New York State Governor Mario Cuomo’s pro-posal to require periodic reviews to measure physicians’ clinical com-petence, an advisory committee agreed upon a plan to use exami-nations, peer review, or audits of office practice (similar to thesystems used in Canada) to assess physicians who are not affiliatedwith a hospital (86). To determine the competence of physicians withhospital privileges, the committee is considering using hospital re-view systems that currently accredit physicians. A strong force be-hind this recredentialing initiative is a concern for the 10,000 to15,000 non-hospital-affiliated physicians in New York who are sub-ject to little peer review and critique (362).

225

.

n3s

226

requires that “professional criteria” be used forgranting clinical privileges. These criteria includecurrent licensure, relevant training or experience,current competence, and health status.

Perhaps a more rigorous system for evaluatingspecialized skills of physicians than the currentprocess used for board certification would pro-vide a more useful indication of the quality of aphysician’s care. The American College of Phy-sicians, as a part of its Clinical Privileges PilotProject, has developed guidelines defining mini-mum skills, education, and training that physi-cians need to perform competently eight specificmedical procedures.

22 Although intended to as-sist hospitals by furnishing objective standards toassess clinical competence and delineate hospitalprivileges, these guidelines may be useful in assess-ing the competence of physicians performing theseprocedures on an ambulatory basis, where clini-cal review is less common.

Guidelines for granting physician privileges toperform specific procedures are also being devel-oped by other organizations. The American Asso-ciation of Urology set up guidelines for the train-ing of physicians in the use of the extracorporealshock-wave lithotripsy.

23 In addition, in 1987, atask force from the American College of Cardi-ology proposed training standards for perform-ing coronary angioplasty.24

22s0 far guide]ine5 have been developed for renal biopsy, acutehemodialysis, acute peritoneal dialysis, continuous arteriovenoushemofiltration, flexible fiberoptic sigmoidoscopy, colonoscopy,esophagogastroduodenoscopy, and endoscopic retrograde cholan-giopancreatography. These guidelines, which have become officialpolicy of the American College of Physicians, were based on medi-cal literature and expert consensus. Approximately 4,200 generalinternists and subspecialists were surveyed as part of the pilot projectto obtain more objective data about the experience and training nec-essary for competence in the procedures. A project to develop guide-lines for a number of other procedures in all subspecialties of inter-nal medicine is planned for 1988 (24).

ZJTO obtain a certificate of training for this procedure, a physi-cian is required to have hads days of training and to have performed15 procedures. Most hospitals require that a physician be certifiedin the use of the lithotripter before he or she is allowed to performthe procedure. The Association also approves potential sites wherelithotripsy training can take place (380).

ZJThe task force calls for “three levels of training for three typesof cardiologists: those doing cardiac catheterization or angiogra-phy, those doing both, and those doing both plus angioplasty orother advanced procedures that may be developed” (22). The guide-lines state that the physician training to perform angioplasty mustcomplete a fourth year of residency training and a minimum of 125coronary angioplasty procedures, including 75 as primary opera-

To date, delineated qualifications for physiciansperforming specific procedures are mostly in theform of guidelines. In actuality, any licensed phy-sician could perform any number of complicatedsurgeries or medical procedures unrestrictedly.Many procedures are done at ambulatory clinicsor “surgicenters,” where there may not be a for-mal physician credentialing system in place. Inthese settings, there is a need to assure the publicthat the physicians performing these proceduresare competent and well qualified. Perhaps the cer-tificates of special competence currently offeredby specialty boards could be used for this pur-pose. The qualifications needed by a physician toacquire such a certificate could be made morerigorous—instead of representing only the pass-ing of an examination, perhaps representing adelineated amount of training, experience, andcompetence that the physician has acquired in per-forming the procedure. The qualifications re-quired for a physician to obtain a certificate ofspecial competence could be those credentialsdemonstrated to relate to better outcomes of pa-tient care.

To encourage physicians to acquire the specialtraining and experience to perform procedures,those physicians who hold such certificates couldbe reimbursed by Medicare at a higher rate. Amore stringent regulation might be to require un-der Medicare’s conditions of participation that aphysician hold a certificate of special competence.

Additional research is needed to explore thequalifications of physicians that relate to im-proved quality of care. Research on specific pro-cedures, similar to the studies of the AmericanCollege of Physicians on eight specific medicalprocedures (24), is needed to determine an ade-quate standard of training and experience for phy-sicians to perform the procedures.

Also needed is research that evaluates boardcertification and physicians practicing in their areaof training by type of specialty. The conditionsor procedures chosen to be evaluated within eachstudy should be conditions that are highly prev-alent among the types of patients a particular spe-

tor. The task force also calls for a “certifying process for a certifi-cate of added experience and qualification in advanced cardiaccatheterization procedures [such as angioplasty]. ”

227

cialist sees, so that they can serve as a valid rep-resentation of a specialist’s practice. Cautionwould still be warranted with respect to general-izing from a physician’s performance for one con-dition to performance for other diagnoses.

To increase the validity of various measures ofphysician performance, further research on vari-ous performance assessment techniques needs tobe conducted. Techniques with greater predictive

power would increase the significance of boardcertification as an indicator of the quality of aphysician’s performance. Validated methods to as-sess physician performance would also increasethe significance of other criteria used in determin-ing a physician’s competence, such as the certifi-cates of special competence. Consumers couldthen rely more heavily on these criteria as accept-able indicators of quality care.