the costs and effectiveness of nurse practitioners (part 3

Case Study #16:

The Costs and Effectivenessof Nurse Practitioners

Lauren LeRoy, C. Phil.Senior Research Associate

with the assistance of:

Sharon Solkowitz, B.S.Health Policy Intern

Health Policy ProgramUniversity of California, San Francisco

INTRODUCTION

Definition of Terms

The concept of using nonphysician healthprofessionals to perform basic medical servicestraditionally provided by physicians emerged inthe mid-1960’s amidst widespread concern overa perceived physician shortage. Variously re-ferred to as new health practitioners, mid-levelpractitioners, or physician extenders, they wereseen as a way to increase the availability ofhealth care services, particularly in primarycare. Although no single term adequately rep-resents the categories of professionals who com-prise this group, the term physician extender isused in this case study when general reference isbeing made to them. This term encompassesnurse practitioners (NPs), physicians’ assistants(PAs), and Medex, a group of former militarycorpsmen who apply their skills in civilian life.

The present analysis highlights data on NPs.However, the data that exist make it difficult tofocus on NPs exclusively. For that reason, dataon the various types of physician extenders areincorporated into the analysis where such dataare relevant for comparative purposes or incases where similarities in experience merit abroader discussion. The distinctions amongNPs, PAs, and Medex derive from differences in

legal requirements, training, and functions (seeappendix). In the wide variation of experiencedocumented for different practice settings, how-ever, these distinctions often break down. As aresult, attempts to distinguish between NPs andPAs understandably lead to broad definitions ofroles such as those quoted by the GraduateMedical Education National Advisory Com-mittee (32):

[A physician’s assistant is] a skilled personqualified by academic and practical on-the-jobtraining to provide patient services under thesupervision and direction of a licensed physi-cian, who is responsible for the performance ofthat assistant.

Today’s nurse, operating in an expanding roleas a professional nurse practitioner, provides di-rect patient care to individuals, families, andother groups in a variety of settings . . . . Thenurse practitioner engages in independent deci-sionmaking about the nursing needs of clients,and collaborates with other health professionals,such as the physician, social worker, and nu-tritionist in making decisions about other healthcare needs. The nurse working in an expandedrole practices in primary, acute, and chronichealth care settings. As a member of the health

3

4 ● Backround Paper #.?: Case Studies of Medical Technologies

care team, the nurse practitioner plans and in-stitutes health care programs.

The actual roles of PAs and NPs depend ontheir work settings. In some cases, the functionsthe two perform are virtually identical; inothers, they are very different. PAs are intendedto operate essentially as physician substitutesfor routine primary care. NPs, while trained toassume medical tasks, come from a traditionbased on fundamental nursing concepts stress-ing aspects of patient care not usually providedby physicians.

The scope of this analysis is limited by in-adequacies in the available data. Although thereis a rather extensive body of literature on physi-cian extender practice in a variety of differentpractice settings—private physician practice,prepaid group practice (PPGP), hospitals, andother organized settings that operate on either afixed budget or fee-for-service basis) —somegeneral problems, as noted by Schweitzer (87),appear consistently in the available studies: nar-rowness of site coverage, incomprehensivenessof variables considered, and weakness of the re-search design. Studies often focus on a singlesite or small nonranciom groups of sites. A num-ber of studies were conducted shortly after thephysician extender entered practice and there-fore leave issues regarding maturity unresolved.Many studies do not adequately identify po-tential biases influencing the research findingsor later interpretation of those findings: In someof them, the impact on the research results offactors unique to the type of practice settingbeing examined is not specified; in others, the

COST= EFFECTIVENESS ANALYSIS

Limitations of a CEA ofPhysician Extenders

Encouragement of NPs and PAs as an innova-tion in the delivery of primary care services isbased on their potential to improve access andto lower costs without compromising quality.This promise derives from several basic assump-tions (18):

● physician extenders can perform basic and

researchers conducting the study may be ad-vocates for the physician extender concept.Finally, from the perspective of cost-effec-tiveness analysis (CEA), perhaps the most seri-ous problem is a dearth of information specifi-cally defining what medical care tasks physi-cian extenders are qualified to perform. Withoutthis information, comparative analysis betweenphysician extenders and physicians is limited.

supply

Before 1970, there were fewer than 2,OOO for-mally trained physician extenders. Currently,there are roughly 22,OOO physician extenders inactive practice: 13,000 NPs and 9,000 PAs (18).The Congressional Budget Office (CBO) esti-mates that 2,000 NPs and 1,500 PAs and Medexgraduate annually. Assuming continuation ofFederal funding at the present level, CBO es-timates the supply of physician extenders by1990 will exceed 56,000 (18). While budget-ary constraints might preclude a decision tofurther increase training opportunities, the de-mand for training positions, as reflected cur-rently by a high ratio of applicants to availablepositions, would not be an obstacle (25,101).Assuming continuation of current trends, how-ever, physician extenders will continue to rep-resent a very small group of health profes-sionals. For purposes of comparison, it shouldbe noted that the number of physicians is ex-pected to rise from the present 400,000 to594,000 by 1990 (18). This increase aloneeral times more than the total number ofcian extenders.

is sev-physi-

routine medical care tasks traditionally per-formed by physicians;

● physicians working in concert with physi-cian extenders will thus be free to focus onmore serious and more complex medicalcare problems;

● training costs for physician extenders arecheaper than training costs for physicians;

● lower costs associated with physician ex-

Case Study #16 The Costs and Effectiveness of Nurse Practitioners 5

tender services will result in lower pricesfor the services provided; and

. improved access resulting from the addi-tion of physician extenders to the healthcare team will increase the frequency ofearly detection of disease and thus reducemedical care expenditures.

A number of issues concerning appropriatetraining, task delegation, performance quality,physician and consumer acceptance, costs, pro-ductivity, and barriers to practice have con-strained the realization of physician extenders’potential. The importance of these concerns isillustrated by their dominance in published re-search. Only recently, with several exceptions,has cost effectiveness provided the frameworkfor analysis of physician extender practice(69,71). The most recent contributions to CEAof physician extenders are a synthesis of relatedresearch and its application to cost-effectivenessquestions published by CBO (18), and an ex-haustive review of literature on task delegation,productivity, and cost by Jane Cassels Record(70).

The focus on literature review and synthesisin these studies reflects, in part, the data andmethodological problems associated with con-ducting a pure CEA in this area. Data from ex-isting studies are insufficient to meet the re-quirements of a thorough CEA. Moreover,while the findings of these more narrow studiesmay contribute to CEAS, the evidence they haveprovided to date is considered “limited but sug-gestive” and allowing for only “tentative” con-clusions (18). CEA seeks to determine which ap-proach accomplishes a given objective at mini-mum cost. Such a comparison between physi-cians and physician extenders is difficult, be-cause the approaches being compared, whileoverlapping, cannot be substituted for eachother in all instances. Moreover, there does notexist the same standardization with physicianextenders as with more traditional technologicalinnovations. Physician extenders differ in back-ground, temperament, training, attitude, abilityto make independent judgments, and desire forindependence. They cannot be considered aneutral “technology” to be utilized and acted on,

because they themselves exert an influence ontheir practice. Moreover, it is not knownwhether differences in productivity, quality, in-dependence, cost, and provider acceptance existamong NPs in different types of specialty prac-tice. These factors raise both data and method-ological questions which have yet to be an-swered in published research.

Physician extenders have been found to becapable of providing high percentages ofprimary care services traditionally provided byphysicians—but it is unclear which services areincluded in these percentages; which services areleft out; whether those left out have more im-pact on the delivery of care, physician attitudes,productivity, and costs than those provided;and so on. One way to focus an analysis wouldbe to select for comparison a set of tasks thathave been noted in the literature to be per-formed by both physicians and physician ex-tenders (e.g., well-baby care, history and phys-ical, hypertension monitoring). Even then, how-ever, the results of the analysis would be 1im-ited, because they would not provide a way todetermine what medical care tasks physician ex-tenders do not perform and the value of thoseservices. It would be difficult to determinewhether a “patient visit” (the usual standard ofmeasurement) were the same in both instancesin terms of content and outcome. The resultswould not provide adequate information for de-veloping staffing configurations that are costeffective in terms of services for which physicianextenders can substitute for physicians, servicesfor which physicians are the only providers, andservices provided by physician extenders thatphysicians traditionally have not provided.

The objectives of a CEA of physician ex-tenders will depend on the perspective of thosefor whom the study is undertaken (e.g., a physi-cian in fee-for-service private practice, a healthmaintenance organization (HMO) with a fixedbudget, or Federal policy makers concerned withreimbursement under Federal health insuranceprograms). Consideration of the type of practicesetting, with its different budget and staffingconstraints, will alter the study design. The con-text within which the analysis is conducted is

6 . Backround Paper #2: Case Studies of Medical Technologies

crucial. Different factors are taken as given; dif-ferent assumptions underlie the analysis.

Moreover, structural characteristics of thehealth care system have a profound impact onthe way an innovation is used. This is clearlyevident in the case of physician extenders andraises questions as to whether a CEA should re-strict its focus to current conditions or whetherits assessment should consider changes in rela-tion to a variety of policy alternatives to modifythe existing structure. This is a key question inthe case of physician extenders. Unlike the in-troduction of most new technologies, the in-troduction of physician extenders into healthservices delivery was not accompanied by re-imbursement. Unlike many new technologieswhich enhance the position of the physician, thephysician extender is potential competition.Physicians legally maintain a substantialamount of control over physician extender prac-tice; however, the structure of reimbursementreinforces that control by making it virtuallyimpossible for a physician extender to practiceindependently, Any analysis that is based onwhat physician extenders can do by virtue oftheir training, rather than what they actually doby virtue of the structural characteristics of thehealth care system, therefore has serious meth-odological limitations.

Yet developing a broad study based on actualexperience to make up for those limitations ismuch more difficult. The data base specifyingwhat physician extenders do and costs asso-ciated with their practice is incomplete. Mostexisting studies confine themselves to very smallsamples. The emphasis has been on ambulatorycare, leaving a dearth of information on theroles of physician extenders in hospital settings,including their potential substitution for housestaff. Expanding the sample size and composi-tion would entail the identification and surveyof physician extenders in a variety of practicesettings, itself a lengthy and expensive task.Moreover, it is difficult to disentangle physicianextender performance and cost characteristicsfrom the characteristics of the practice setting.In a CEA, which should be the focus? Is it possi-ble to control for those factors associated with

the practice setting that determine utilization ofhealth personnel?

Given the data that are available, it is notpossible to conduct a CEA of sufficient precisionto calculate cost-effectiveness ratios. To reiter-ate key constraints on such an analysis, it is notknown exactly what medical tasks physician ex-tenders perform, nor what tasks they cannotperform, nor the importance of either to thoseemploying physician extenders (i. e., do em-ployers seek to cover the average case or thevariance: routine versus emergency care?). It isunclear what occurs during the “patient visits”reported throughout the literature. Are therequalitative differences in the services providedby a physician extender and those provided by aphysician? The relationship between practicesetting characteristics and physician extendercost and performance factors is not fullyunderstood or described in available research.Cost information, when available, is not suffi-ciently broken down to compare the full costs ofa given service provided by a physician extenderto the service provided by a physician. Costs ofphysician extenders are generally presented asan add-on to an existing physician practice.While calculations often account for overhead,they also reflect a fully operational practice,thus minimizing startup costs and assuming dif-ferent degrees of practice independence forphysicians and physician extenders. Moreover,charges cannot be used to determine costs, be-cause existing evidence indicates little relationbetween the two.

Because the methodological problems andlack of data preclude a full CEA of physician ex-tenders at this time, no effort is made in theanalysis presented in this case study to developcost-effectiveness ratios. The purpose of theanalysis is to elaborate on the factors essentialto determining cost effectiveness and to testingassumptions regarding the cost-effectiveness po-tential of NPs against existing data and researchfindings. Even without data of sufficient preci-sion, quality, and quantity for a full CEA, onecan see trends and draw conclusions. From apolicy perspective, there emerge in this caseclear indications of areas in which alternative

Case Study #16: The Costs and Effectiveness of Nurse Practitioners ● 7

poIicies would have significant impact. Giventhe expense in time and dollars of going beyondsynthesis and analysis of available informationwithin a cost-effectiveness framework, it isquestionable how much more could be gainedthat would significantly alter the findings andconclusions suggested by existing studiesthrough fulfilling the requirements of a moreelaborate cost-effectiveness methodology.

Studies focusing on the removal of seriousdeficiencies in the data base most likely wouldbe more useful in clarifying still outstandingissues. Assessment of the cost effectiveness ofNPs must at a minimum consider the specificservices NPs are qualified to provide, perform-ance quality, productivity, task delegation ex-perience, changes in physician practice behaviorafter the introduction of NPs, employmentcosts, impact on average expenses per patientvisit, training costs, price effects, and revenuegeneration ability. The discussion that followsexamines each of these factors. The difficulty inreaching definitive conclusions in this area re-sults in part from the sensitivity of the analysisto modest changes in many of the variables andfrom the need to consider the combinations inwhich these variables are found as additionalfactors influencing the outcome.

On the basis of available data, it appears thatNPs do alter the production of medical servicesin a manner that can improve access to suchservices and reduce production costs. That theintroduction of NPs will result in a reduction inthe price of medical care services or in overallmedical expenditures appears less likely. With-out a reduction in price that reflects lower costs,the financial benefits derived from the cost-effective attributes of NPs accrue primarily tothe physician or to the employing institution.This situation with NPs is similar to the expe-rience with many new medical technologies thatare cost saving. Benefits to consumers comewhen the introduction of NPs results in im-proved access.

Services Provided by PhysicianExtenders

In order to determine the cost effectiveness of

physician extenders, it is necessary to knowwhat services they are qualified to provide andwhether those services are substitutive or com-plementary to those provided by physicians.This key question is one on which available dataare clearly inadequate. Most studies refer toservices provided by or delegated to physicianextenders in terms of office visits rather thandefinitive tasks. They describe services physi-cian extenders are producing rather than thosethey are qualified to produce (70). Instead ofcategorizing services by specific tasks, studiesare more likely to categorize services generallyinto those physician services that physician ex-tenders either can or cannot safely provide.

There are some studies that have attempted todefine areas of medical practice or diagnosesmanaged by physician extenders. Although thestudy samples are often very small, the findingsof these studies, accompanied by more generalconclusions drawn from the bulk of availableresearch, suggest several patterns:

●

●

●

physicians and physician extenders haveboth a complementary and substitutive re-lationship. NPs provide additive servicesand PAs serve as an extension of the physi-cian (70);physician extenders are capable of safelyproviding a high percentage of primarymedical care services (70); andstudies that document current performancereveal that the practice setting is the majordeterminant of services provided by physi-cian extenders (29,50).

In general, PAs work more closely with physi-cians than do NPs and also provide care whichis more oriented toward acute or emergency sit-uations. NPs often assume a large degree of in-dependence and responsibility and tend to be in-volved in a broader spectrum of patient care, in-cluding counseling, education, and general con-sultation on a continuing basis.

Among the studies that begin to define serv-ices performed by NPs is that of Coulehan andSheedy (20). The medical practice of an NPtrained in diagnosis and treatment of generalmedical conditions included the following:wellness care; stable chronic disease (hyperten-

8 . Background Paper #2: Case Studies of Medical Technologies

sion, diabetes, obesity, arteriosclerotic heartdisease, arthritis, chronic depression, psy-chophysiologic reactions); and acute self-limitedconditions (colds, sore throats, acute viral syn-drome, minor trauma, rashes, skin infection).Of the 15 most common diagnoses for theCoulehan and Sheedy study sample, the NPhandled so percent or more of the followingconditions: upper respiratory infections, otitismedia, otitis externa, soft tissue trauma, andgonorrhea. The NP managed one-third to one-half of patients presenting the following: muscleor back strain, dermatitis or eczema, hyperten-sion, diabetes, obesity, and urinary tract infec-tion. Again, it must be noted that these servicesreflect services the NP provided within the con-straints of the practice setting, not necessarilythe range of services the NP was qualified toprovide. Moreover, it should be noted that theCoulehan and Sheedy study was conducted in1973, and therefore reflects early experiencewith NPs, Subsequent studies reveal even higherpercentages of patients presenting the same con-ditions being treated by NPs.

In terms of specific tasks, the limited data thatare available indicate that physician extenderscan perform medical functions basic to primarycare such as taking medical histories, per-forming routine physical examinations, carry-ing out simple diagnostic procedures, orderingroutine lab tests and interpreting their results.Physician extenders commonly administer injec-tions, apply dressings, casts and splints, and canperform life-preserving measures in emergencysituations. Some are qualified to perform minorsurgical procedures such as removing a foreignobject from the eye 01 routine suturing (10).

Physician extenders generally are restrictedfrom prescribing drugs except under certainconditions (e. g., having prescriptions counter-signed by a supervising physician or prescribingwithin a limited “scope of practice”). EightStates prohibit physician extenders from writingdrug prescriptions. The issue of whether or notphysician extenders are qualified to prescribedrugs is one that a number of States are current-ly reviewing. Some States are experimentingwith the extension of prescribing privileges(e.g., California has a project which allows pre-

scribing by NPs, PAs, and pharmacists in fivegeographic areas of the State). The constraintson drug prescription represent the most sen-sitive unresolved issue in terms of tasks allowedto be performed by physician extenders, bothbecause of the integral role of prescribing inmedical care and because of the implications ofsuch constraints for professional independence(18).

Performance Quality

The quality of services provided by NPs iscrucial to their acceptance by both physiciansand patients. Indeed, this issue has been studiedmore than any other. Like most research in thisarea, the studies on quality generally use smallsamples, assess quality from a variety of dif-ferent perspectives and focus more on short-term results than on long-term outcomes of pa-tient care. These evaluations of physician ex-tender services repeatedly confirm their highquality (10,14,24,41,42,44,51,52,79). The qual-ity of medical care services provided by physi-cian extenders is at least comparable to the qual-ity of services provided by physicians them-selves. Furthermore, in some cases, physicianextenders have shown performance superior tophysicians in symptom relief, diagnostic accu-racy, and patient satisfaction (33,70). Sum-marizing the findings of the numerous studies ofphysician extender performance in a variety ofpractice settings, CBO concludes: “Physicianextenders have performed as well as physicians,with respect to patient outcomes, proper diag-noses, management of ‘indicator’l medical con-ditions, frequency of patient hospitalization,manner of drug prescription, documentation ofmedical findings and patient satisfaction” (18).

In its study of the Federal physician extenderreimbursement experiment (102), System Sci-ences, Inc., used nationally recognized diseasetreatment protocols to evaluate the quality ofcare provided to patients by physician/physi-cian extender teams and by physicians only.The results favored physician/physician ex-

Case Study #16. The Costs and Effectiveness of Nurse Practitioners ● 9

tender teams and revealedfor physician/NP teamsgroup.

Productivity

higher quality ratingsthan for any other

It is difficult to measure productivity in stricteconomic terms when applied to health man-power. The inputs and outputs of the medicalcare production process are difficult to defineand measure. Some people define the output ofthe medical care industry as an intermediategood to be combined with other inputs in theproduction of good health. However, most fo-cus on this intermediate product and try to de-velop proxy measures for what actually occurswhen patient and health professional come to-gether.

The output most commonly associated withhealth professional services is defined in termsof patient visits. Productivity of physicians andphysician extenders usually is measured by thenumber of patient visits per unit of time.Holmes, et al. (39) noted the inability of thismeasure to reflect either the complexity or thevolume of services provided during a patientvisit. These investigators also noted the difficul-ty of determining the relative contributionsmade to patient care when more than one pro-fessional is involved. To overcome these inade-quacies, they suggested the use of a relativevalue scale to assign values to the specific ac-tivities performed by physicians, NPs, andnurses. This and other attempts are being madeto develop more refined measures of pro-ductivity, but the use of patient visits is mostprevalent in the literature.

Although there is little doubt that the efficientuse of NPs can improve the productivity of thedelivery of medical services, it is essential to dis-tinguish between potential impact and actual ex-perience. As Reinhardt (74) points out, deter-mination of physician productivity must ac-count for both technical feasibility in the pro-duction process and the probable economic be-havior of the physician. One needs to know notonly what is technically feasible but also whatconfiguration of inputs physicians are likely tochoose and to what extent physicians will at-

tempt to maximize the output that is technicallyattainable with that combination of inputs (75).The physician, perhaps in collaboration withthe administrator in an organized setting, deter-mines how the physician extender will be usedas an input in the production of medical serv-ices. While the debate continues regarding theindependence of NPs, reality shows them to befunctionally dependent on the physician. Berki(37) defines this relationship as one of “con-strained substitutability” with the physiciandetermining most of the constraints.

The extent to which tasks are delegated fromphysician to NP, the amount of time it takes aphysician or NP to perform the same task, andthe impact of the introduction of NPs on physi-cian behavior are key productivity-related vari-ables in the cost-effectiveness calculation. In areview of 15 studies that used physician officevisits as a measure of delegability, Record (70)concluded that between 75 and 80 percent ofadult primary care services and up to 90 percentof pediatric primary care services could be del-egated to physician extenders. The purpose ofthe Record study was to estimate different com-binations of physicians and physician extendersthat could produce given levels of primary careservices. Cost estimates associated with the var-ious configurations revealed potential cost sav-ings of $0.5 billion to over $1 billion in caseswith higher physician extender participation.This amounted to 19 to 49 percent of total pri-mary care provider costs.

Steinwachs, et al. (94) reported on the ex-perience of the Columbia Medical Plan, anHMO, in expanding the use of physician ex-tenders over a 3-year period. As the involve-ment of physician extenders increased, dramaticchanges occurred in the distribution of patientencounters between physician extenders andphysicians. In 1971-72, 79 percent of patient en-counters in adult medicine were managed ini-tially by physicians. By 1973 -74, that figure haddropped to 38 percent. In 1971-72, physician ex-tenders managed 10 percent of initial encountersfor illness and injury and conducted no healthreviews. By 1973-74, physician extenders man-aged 50 percent of illness and 75 percent of in-

10 ● Backround Paper #2: Case Studies of Medical Technologies

jury initial encounters and conducted 50 percentof adult health reviews. The delegation of re-sponsibility for initial encounters in pediatricswas even higher, with all but 16 percent of pedi-atric health reviews being conducted by physi-cian extenders.

The changes in distribution of patient en-counters among physician extenders and physi-cians resulted in a major change in staffing pat-terns as the Columbia Medical Plan evolved(94). During the study period, enrollment nearlydoubled. In response to this growth, the numberof full-time equivalent (FTE) internists increasedless than 10 percent, while the number of physi-cian extender FTEs increased 260 percent. Theresult was a change in staffing patterns from theearly study period when physicians represented60 percent of the total number of FTE providersto the final months of the study when their rep-resentation decreased to 38 percent of total FTEproviders. Pediatrics experienced a similar butless pronounced staffing change, with physicianFTEs increasing by one-third and physician ex-tender FTEs nearly doubling.

In the Northern California Kaiser-Per-manente Medical Care Program, also an HMO,NPs conducted a Health Evaluation Service(H E S) consisting of automated multiphasichealth testing followed by a physical examina-tion and health appraisal (22). Of the patientswho entered the Kaiser system through HES, 74percent were managed without physician refer-ral. Of those referred to a physician, two-thirdswent to a specialty clinic, thus having the NPs’HES visit substitute for an initial primary carephysician visit. Moreover, pelvic exams con-ducted through HES replaced 5,207 visits to thegynecology clinic during the study period.

Similar experience is reported for two NPsworking in the offices of two family physiciansin the Burlington Randomized Trial (79,93). Pa-tients were divided randomly into two groups:one receiving first-contact, primary care serv-ices from a family physician working with anurse (“conventional” care); and the other re-ceiving such care services from NPs. The studyfound that NPs were able to provide primarycare services as safely and effectively as physi-

cians. In 67 percent of patient visits, care wasprovided with no physician consultation.

More specifically, in settings with both NPsand physicians, NPs assume primary responsi-bility for the diagnosis and treatment of acuteself-limited conditions and acute conditionswith limited uncertainty in the diagnosis and re-sponsiveness to a defined therapy. Adult healthreviews are shared by physicians and NPs,while the majority of well-baby care can be pro-vided by pediatric NPs, Physicians retain pri-mary responsibility for diagnosis and treatmentof more complex and serious acute conditionsand for chronic conditions (89,94,109).

It is obvious from the aforementioned andother studies (19,85, 105) that NPs can assume ahigh proportion of primary medical care tasks.Existing studies also reveal substantial variationamong practices, making more difficult thetranslation of specific expectations for taskdelegation to widespread experience. Record(70) outlines a number of factors accounting forthat variation, including type of practice set-ting, structure and age of practice, provider rolestrain, legal and reimbursement constraints, andlevel of demand.

The time spent by NPs in managing a patientvisit is significantly higher than that spent byphysicians. Table 1 shows the experience ofphysician extenders in the Southern CaliforniaKaiser-Permanente Health Facility.

For given presenting complaints, physicianextenders averaged 4 to 9 minutes longer thanphysicians. Recent research has shown that NPsspend up to 65 percent more time per patientvisit and see 60 percent as many patients perhour as do physicians (18,103). NPs see fewerpatients per day because of their longer time pervisit, a shorter workday, and more time de-voted to patient telephone consultation and ad-ministrative activities. The number of patientvisits reported for NPs ranges from 5 to 14 perday (101). Consideration of the number andduration of patient visits to NPs must accountfor the possibility that the content of the NP pa-tient visit differs from that of a physician visit.If NPs do, in fact, provide more patient edu-cation and counseling than physicians, they


298307, 1978

may be improving access and the patient’s ex-perience during the encounter. Since such serv-ices generally are not reimbursable, and reim-bursement is used as a measure of value, it is dif-ficult to determine the value of these services.

An NP who substitutes for a physician in pro-viding specified medical services allows the phy-sician to increase productivity. As Reinhardtpoints out, this is true even if the NP spendsmore time providing the same service (75s):

Because of the need for supervision of physi-cian extenders, the delegation of a task normallyrequiring 10 minutes of physician time to a phy-sician extender may actually free only 8 minutesof physician time (and may require 20 minutesof the physician extender’s time). Even so, aslong as some physician time is freed at all, taskdelegation will enable the physician to treatmore cases per unit of time and hence in-crease . . . hourly productivity.

Physicians working with NPs noted an addi-tional increase in their efficiency due to the needto be more rigorous and clear in communicatingtheir thoughts to the NPs (93).

Estimates of increases in the productivity ofphysician practices that include NPs range from20 to 90 percent (31,36,38,39,43,70,82,104). Insome cases, these estimates reflect actual ex-periences; in others, they are the resuIt of com-puter simulation models that determine produc-tivity increases based on optimal staffing config-urations for performing medical care tasks.Given that the computer s imulat ion models

measure potential rather than actual experience,these generally yield higher estimates. Thegreatest productivity increases come when theNP has primary responsibility for a subset ofpatients and when triage is performed by theNPs’ referring complicated cases “up” to thephysician rather than by the physician’s dele-gating routine medical problems “down” to theNP (90).

The amount of physician time freed by NPs isreflected in physician/NP substitution rates. Ina mathematical model for HMO staffing pat-terns, Schneider and Foley (82) estimated thatthe substitution of one physician extenderwould decrease physician requirements by 53 to60 percent depending on departments. After theexpanded use of physician extenders in the Col-umbia Medical Plan, the workload (number ofencounters per FTE) remained constant for thephysician extenders but declined for physicians,in part because physicians were freed from rou-tine medical procedures to concentrate on pa-tients with more time-consuming complex prob-lems (94). Although physicians still must beavailable for consultation with the NP, Green-field, et al. (33) found that the physician time re-quired for consultation was 92 percent less thanthe time that physicians would spend treatingthe same clinical problem.

In addition to the numerous studies that de-fine productivity in terms of patient visits,Holmes, et al. (39) used a relative value scale todetermine productivity differences between tra-

12 ● Backround Paper #2: Case Studies of Medical Technolologies

ditional physician/nurse practices and physi-cian/NP practices. They found productivitylevels in practices incorporating NPs to be anaverage of 25.8 percent higher (39).

Productivity increases that result from the useof NPs vary widely, depending on the practicesetting, the responsibilities delegated to the NP,the severity and stability of illness in the pa-tients served by the practice, and how the physi-cian chooses to use the free time resulting fromtask delegation. As will be discussed below, thepotential for productivity increases is not nec-essarily realized. Moreover, Hershey and Kropp(36) point out the negative impact that max-imum task delegation and resulting productivityincreases can have on the practice environment.Considering such factors as office hours, wait-ing room congestion, and supervision time, theyconclude that an operating environment re-sembling that of a conventional physician prac-tice cannot be achieved with more than a 50-per-cent increase in procductivity. If physician hoursremain constant, supervision time has a markedimpact on how much of an increase in pro-ductivity a practice environment can absorb.Tables 2 and 3 compare practice environmentswith and without PAs under varying supervi-sion times.

Employment Costs

To determine the cost effectiveness of NPs,both the amount of time they spend to performa given service and the cost per unit of time forNPs must be compared with the figures for phy-sicians. However, available cost data arelimited, and what data exist often come fromstudies of small samples that are not com-parable. An additional difficulty in comparingcosts arises because most physicians are self-employed and compensated for their services ona fee-for-service basis, while virtually all NPsare salaried.

The basic costs of employing an NP includesalary, fringe benefits, and physician super-vision. The average salary of NPs was estimatedin 1978 to be about $13,800 (105). Using 1975data, CBO (18) determined that the medianhourly wage for physician extenders was about$6 as compared with $24 for physicians. Record(70) found salary (or income) differentials bet-ween physicians and physician extenders to beclose to $36,000 per provider per year. Severalstudies stress the importance of such differen-tials in the physician’s decision to employ an NP(33,82). In their systems analysis of the use ofphysician extenders, Schneider and Foley (82)

Table 2.—First Comparison of Practices With and Withouta PA When Supervision Is Considered

Assumptions

With PA/4 minutes supervisionWithout PA 242 patients 264 patients 286 patients 308 patients220 patients per week per week per week per week

Measures per week (10%-gain) (20%-gain) (30%-gain) (40%-gain)—Average physician-patient contact time per

day, Including supervision (hours) 5.93 4 34’ 5.24b 5.55 6.15Average time last patient leaves office

(minutes past 4 p m ) . 30 25b 27 37b 66a

Average total wait per patient (minutes) 17 16 18 23b 39a

Average percent of patients who wait fora d m i s s i o n t o e x a m i n a t i o n r o o m 39 52b 54b 62b 81a

Maximum number of patients in waiting roomo n a v e r a g e d a y 3 3 5 10a

Average annual net Income before taxes $34,735 $31,970 $35,102 $38,250 $41 ,615b


Table 3.—Second Comparison of Practices With and Withouta PA When Supevision is Considered

Assumptions

With PA

4 minutes 2 minutes 1 minutesupervision supervision supervision No supervision

Without PA 264 patients 286 patients 308 patients 330 patients220 patients per week per week per week per week

Measures per week (20%-gain) (30%-gain) (40%-gain) (50%-gain)

Average physician-patient contact time perd a y , i n c l u d i n g s u p e r v i s i o n ( h o u r s ) 5.93 5.24’ 4.75b 4.84b 4.75b

Average time last patient leaves office(minutes past 4 p.m.) . . . . . . . . . . . . . . . . . . . . 30 27 27 26a 31

Average total wait per patient (minutes) . . . . . . 17 18 17 20 18Average percent of patients who wait for

a d m i s s i o n t o e x a m i n a t i o n r o o m . . 39 53a 53a 72zb 7 l b

Maximum number of patients in waiting roomon average day . . . . . . . . . . . . . . . . . . . 5

$343735

6a 7b

Average annual net income before taxes. . . . . . $35:102 $37,567 $42,291’ $46,855b

SOURCE J C Hershey and D H Kropp, ‘A Re-Appratsal of the Productlvlty Potential and Economic Benefits of Physiclan’s Assistant s,” Med. Care 17602, 1979

found physician extenders’ use to be unaffectedwhen their salaries remained below 47 percentof physician salaries. Once the physician ex-tender’s salary reached 62 percent of the phy-sician’s salary, however, physician extendersrapidly lost their cost effectiveness in the model.

The amount of time reported for physiciansupervision and consultation varies consider-ably among practices. Supervision time varieswith the type of practitioner involved. Legal re-quirements also determine the time devoted tosupervision. Forty-three States require directsupervision of PAs and Medex; only 11 Stateshave similar requirements for NPs (18). Withina given practice, variation in consulting time is afunction of the reason for the consultation,whether the physician sees the patient or onlyconfers with the NP, and the practice experienceof the NP. The actual time per consultationranges from less than a minute to approximately8 minutes. Consultation on initial visits requiresabout half as much time as consultation on con-tinuing visits (105). Record, et al. (72) foundthat in the 12 percent of cases where the PAsunder study requested physician consultation,the physicians were likely to spend as much timewith the patient as if they had been the initialprovider. In such cases, the extra cost to the

practice is represented by the cost of the PA andnot that of the physician. Because of the cost im-plications, direct referral to the physician wouldbe optimal but usually is not possible in suchcases.

In its review of available studies, CBO de-termined that supervision and consultation withphysician extenders require between 10 and 20percent of physician time (18). With CBO esti-mates of hourly earnings, this adds between $3and $5 to the physician extender’s hourly salarycost. CBO further determined that direct com-pensation accounted for 54 to 72 percent of thetotal cost of a physician extender, while super-vision costs made up the remainder (18). Table 4compares the time and costs of physicians andphysician extenders, according to patient visits.If both salary and supervisory costs are con-sidered, the hourly costs of NPs range from one-third to one-half of physician costs ($9 to $12/hour as compared to $24/hour).

It should be recognized that the salary level ofNPs is in part a function of the demand for theirservices. As Berki (7) points out, with somecaveats, demand for physician primary careservices can be considered a direct demand ex-pressed by the patient. The demand for NPs is

14 ● Background Paper #2: Case Studies of Medical Technologies

Table 4.—Patient Care Time and Cost of Physician Extendersand Physicians, 1975 a

Physician extender

PA -

NP and Medex PhysicianMinutes per patient visit. . . . . . . . . . . . 19.4 13.2 1 1 . 7Cost per hour . . . . . . . . . . . . . . . . . . . . $9.43-$12.22 $8.36-$10.73 $23.90Direct compensation. . . . . . . . . . . . . . $6.63b $5.98b $23.90’Supervision. . . . . . . . . . . . . . . . . . . . . . $2.80-$ 5.60d $2.38-$4.75d N.A.Cost per patient visit . . . . . . . . . . . . . . $3.04-$3.94 $1.84-$2.36 $4.66

one that the physician must “initiate,” “ex-press, ” and “legitimate,” making this demand aderived one (7). Goldfarb (30) notes that themarket for physician extenders is not suffi-ciently competitive to raise the relatively lowwages of physician extenders to the wage levelthat would prevail if more physicians expresseddemand for their services. She finds that the fac-tors that depress wages are stronger than thosethat raise them, resulting in a prevailing wagelevel that bears little relationship to pro-ductivity and, on average, leaves them under-paid. Obviously, an increase in demand for NPsresulting from such systemwide changes as en-actment of national health insurance or expan-sion of HMOS, or a decrease in demand re-sulting from changes in physician distributionthrough the National Health Service Corps orthe potential oversupply of physicians would re-quire a reassessment of salary costs.

The costs associated with NP practice go be-yond direct compensation and supervision. Theneed for additional staff support, space, andequipment may accompany the decision to hirean NP. Moreover, the style of NP practice hascost implications. A number of studies havefound that physician extenders perform morediagnostic tests than physicians and have dif-ferent patterns of medication use (18,26,43,71).One recent study (60) found that NPs performed53 percent more diagnostic tests per patient thanthe physicians for whom they worked and 46percent more than the physicians who did notwork with any physician extenders at all. Insome cases, these increased lab tests result inidentification of additional pathology (43). Al-though the benefit to the patient in such in-stances may counter the additional cost, patientdiscomfort and time plus the cost of false posi-tives or negatives must be considered. Available


evidence also suggests that the use of protocolscan diminish the tendency toward excess use ofdiagnostic tests (33,34).

In addition to spending more time per patient,NPs may log more visits per patient in a giventime period. Since the salary and supervisioncosts of NPs are significantly lower than thesalaries of physicians, the increased return visitsare not necessarily a financial problem from thepractice’s perspective, although they do con-sume the patient’s time. However, as Spector, etal. (92) discovered, a disproportionate increasein visits can raise the overall cost per patientbeyond the point where the use of NPs is cost ef-fective for the practice. Others stress that interms of overall medical expenditures such prac-tice patterns may be cost effective if they reducehospitalizations (78).

When all of these cost factors have been con-sidered, NPs have been found to perform com-parable medical care tasks at a lower total costthan physicians. Lewis and Resnik (51) foundthis to be true for inpatient and ambulatoryservices for all patients. In the Kaiser-Per-manente Medical Care Program, which insti-tuted a Health Evaluation Service (HES) op-erated by NPs, entry costs (health appraisal,followup, and referral) for the HES group were$43.09 as compared to $61.41 for patients usingphysicians as the point of entry. Costs of overallmedical resources used over 12 months by co-horts of patients with comparable health statuswere $98.63 for the HES group and $131.18 forthe physician group (22). In another Kaiser-Permanente facility, overall combined. costs ofNPs were 20 percent lower than physician costs(33). Studies on private physician practices,while not specifically addressing the cost issue,indicate similar experiences (66,86).

Average Expenses per Patient Visit

Even if one allows for supervision costs, NPscan provide selected services at less cost thanphysicians. This lesser cost does not necessarilytranslate into lower average expenses per pa-tient visit, the latter of which are a function oftotal practice expenses and patient volume (18).A number of Studies have documented increases

in patient visits which occurred in practicesusing physician extenders (38,65). Annual pa-tient visits in the University of Southern Califor-nia survey (103) were 50 to 60 percent higher inpractices with physician extenders. System Sci-ences, Inc. (102) found that practices in-corporating physician extenders provided 12more patient visits per $1,000 of cost than prac-tices without them. CBO reports that the Sys-tem Sciences study showed practice expenses forphysicians with physician extenders to be 74percent higher than for solo physicians (18). Atthe same time, patient volume in the physi-cian/physician extender practices was 146 per-cent higher, resulting in an average expense perpatient visit 29 percent lower than that for solophysicians.

While experience has shown that NPs andother physician extenders can lower average ex-penses per patient visit by as much as one-third,the manner in which the physician or institutionuses them and the way in which time freedthrough task delegation is used will determinewhether the potential saving is realized. If NPsare used to provide services complementary tothose of the physician rather than services sub-stituting for the physician’s, the potential re-duction in average per-visit expenses may bediminished or lost. In such cases, however, thecomplementary services often imply quality en-hancement, a different (and implicitly better)visit for the same cost.

With the addition of NPs to a practice, physi-cians may choose to maintain, increase, or de-crease their level of effort. Komaroff, et al.,caution (43):

Over and above any efficiencies introducedby this or any other system, the time a physicianaverages with patients on a given day is power-fully influenced by two additional factors: thevolume of patients to be seen and the “style” andinterests of the individual physician. It istherefore unwise to expect too much from, or toattribute too much to, organizational changes ofthis kind in the absence of strong pressures tooptimize efficiency.

In order to achieve the saving that comes fromsuch efficiency, the practice employing an NP

must either expand its volume of patient visitsor maintain its volume and reduce its physicianinput. The latter option is obviously more feasi-ble in an institutional setting where physiciansare hired as salaried employees than in physi-cian office practices where physicians are self-employed (and also) would be less likely to hirean NP).

If physicians continue to see the same numberof patients, NPs may reduce average per-visitexpenses by increasing patient volume suf-ficiently to more than cover costs accom-panying their introduction into the practice.Reduction in physician effort in terms of patientvisits, however, may occur for several reasons:1) physicians may be seeing patients with morecomplex problems that demand more time pervisit; 2) they may devote more time to hospi-talized patients; and 3) the presence of NPs mayallow physicians to take more leisure time. Fur-thermore, some physician time is required forsupervision and consultation with the NP. Forwhatever reason, if physicians reduce their pa-tient load, the average expense per patient visitincreases, as illustrated in table 5. However, itshould be recognized that if the number of pa-tients seen by the physician per day decreasesbecause of more time spent per patient or the de-livery of more complex services, the “patientvisit” produced becomes a different product thatmay justify a higher cost.

Contrary to what had been hoped for, the re-ductions in average expense per patient visit

achieved by solo practices have not been real-ized in many group practices. Using System Sci-ences, Inc., data, CBO reports average per-visitexpenses in group practices employing physi-cian extenders to be only 1 percent lower thanthose in group practices without them, althoughsolo practices with physician extenders have ex-perienced as much as a 30-percent reduction inaverage practice expenses (18). CBO speculatesthat the differential may be due less to practiceorganization than to the type of physician at-tracted to each practice arrangement, with phy-sicians in group practice more highly valuingleisure time and using physician extenders to re-duce their workload.

Training Costs

Training costs indirectly affect employmentcosts and the costs to society of NPs. These costsare important because much of them are pub-licly subsidized and they may have some in-fluence on salary expectations. NPs obviouslybenefit from public subsidies for their training,just as do other health professionals. They enterthe job market with a lower personal investmentin training than would have existed withoutsubsidization. This, in turn, means that they areseeking a return on an investment that does notreflect the full costs of their training, thus ben-efiting their employers through lower salarycosts (30).

Total training expenditures for physician ex-tenders are substantially lower than those for

Table 5.--Change in Average Expense per Patient Visit With Reduced Physician Effort

Number ofHours per day Hourly cost Total cost patients Cost per patient visit

Physician . . . . . . .— .—

8 x $24 = $192 - 10 = $19.20NP . . . . . . . . . . . . . 8 x $12 = $96 - 6 = $16.00

$288 - 16 = $18.00(average expense per

patient visit)

Physician . . . . . . . 8 x $24 = $192 - 7 = $27.40NP . . . . . . . . . . . . . 8 x $12 = $96 - 6 = $16.00

$288 - 13 = $22.10(average expense per

patient visit)


physicians, as shown in table 6. Total expend-itures for the segment of training that qualifiesone to practice as a physician extender are lowerfor NPs than for PAs (18). In the case of NPs,the figures in table 6 do not account for the costsof education required to obtain an RN license asa prerequisite to NP training. Depending on thetype of initial nursing education program (seeappendix) and the duration of the NP trainingprogram, the costs of NP training could equal orexceed the costs of PA training. Estimates of theaverage annual costs of nursing education in1979 are $5,901 for baccalaureate programs (4years), $4,974 for diploma programs (3 years),and $4,912 for associate degree programs (2years) (47). Developing comparable full trainingcost figures for NPs, PAs, and physicians iscomplicated by the different education re-

Table 6.-Training Expenditures” for Physicians andPhysician Extenders, Academic Year 1978-79

Cost per student

Annual costb Total costc

PhysicianMean . . . . . . . . . . . . . . . . . . . $14,200 $60,700Median . . . . . . . . . . . . . . . . . N A d N A d

Range . . . . . . . . . . . . . . . . . . $7,600 -$ 20,800$30,200-$83,000

Physician extenderePA and Medex

Mean $6,800 $11,900Median $7.400 $9,900Range $4400-$9.900 $7,100-$17,200

NPMean $12,900 $10,300Median $10,100 $8,000Range $5,300-$31.000 $3,000-$32,000

quirements for entry into NP, PA, or medicaltraining programs. Since a baccalaureate degreeis a prerequisite for medical school, comparablefigures for physician training must include thecosts of baccalaureate education if the costs ofobtaining the RN license are included for NPs.Program requirements for PAs do not nec-essarily include postsecondary education, al-though most PAs have had at least 3 years ofcollege. Whether any of these education costsshould be accounted for (as a reflection of reali-ty rather than formal entrance requirements) isunclear.

Looking solely at the expenditures for the NPor PA phase of training (see table 6), higheraverage annual expenditures for NPs are in partexplained by the fact that PA training occurs inmedical schools, using their faculty and re-sources. Thus, there is an indirect subsidy of PAtraining in medical schools that is met more di-rectly in NP training programs. Moreover,while the average annual cost of NP training ishigher than for PAs, the shorter length of NPtraining results in lower average total costs forNPs.

The total Federal contribution to basic healthprofessions education and training is highest formedical education. This reflects both the costsof that training and the fact that virtually allmedical schools receive Federal funds. Theamount of Federal support per student and itsrelation to average annual training costs, asshown in table 7, is significantly higher forphysician extender training.

Table 7.—Federal Support for Physician andPhysician Extender Training, Fiscal Year 1978

Percent ofaverage

Total annualFederal Amount per trainingsupport student cost s

Physician. ... ... .$190 million $3,000 20%Physicianextender. . . . . . . . . $22 million $4,000- 50-700/0

$5,000

18 ● Backround Paper #2: Case Studies of Medical Technologies

The public cost of producing each additionalphysician and physician extender who wouldnot have been trained in the absence of Federalsupport has been estimated by CBO (18). In-cluding both construction and operating sup-port, the Federal cost of each additional physi-cian trained between 1969 and 1978 was esti-mated to range from $40,000 to $60,000. TheFederal cost for each physician extender duringthe same time period ranged from $10,000 to$20,000. 2 A significant factor in the differentialbetween medical education and physician ex-tender training, CBO noted, is the Federal strat-egy of providing seed money but not operating

subsidies to physician extender programs.

NP programs are less dependent on Federalsubsidies than PA programs. In fiscal year 1978,only 40 percent of NP programs received Fed-eral support, as opposed to 90 percent of PAprograms. Because they are less dependent ofFederal subsidies, NP programs are potentiallyless vulnerable to changes in Federal trainingsupport, although it can be assumed that manyNP programs do receive public subsidies at theState level that also might diminish if Federalpolicy no longer encouraged the training anduse of new health professionals.

Annual training expenditures for physicianextenders are significantly lower than those forphysicians. As Scheffler (80) argues, however,even if the cost of training physician extenderswere the same as that for physicians, physicianextenders would still be a good investment be-cause of their shorter training period. Schefflerestimated that three PAs can be trained for thecost of training one physician and together canproduce 1.8 times more visits than one physi-cian. Although analyses like Scheffler’s begin tograpple with the issue of differences in trainingcosts between physician extenders and physi-

cians, their findings do not define the costsassociated with the specific services that eithergroup can provide. Detailed cost data on train-ing for specific sets of primary care services arenonexistent. There is no agreement on a uni-form set of services that any type of physicianextender can provide. Some consensus is a pre-requisite to refining currently available data.Without it, only gross comparisons of totaltraining costs are possible.

Medical Care Prices

Lower costs associated with NPs do not nec-essarily translate into lower prices for their serv-ices. Moreover, productivity gains for physi-cians who employ NPs may not lead to a re-duction in physician fees. Therefore, consumerscannot necessarily expect to benefit from a re-duction in average charges in practices withNPs. Because the market for medical care serv-ices does not conform to the competitive model,a reduction in expenses need not be followed bya decline in prices (7,18,74,90). As seen in table8, a System Sciences, Inc., survey (102) did re-veal lower average per-visit charges of about 21percent in practices with physician extenders.Similar experience has been documented else-where (66). However, the note accompanyingthe table suggests that the location of many ofthe physician/physician extender practices in

Table 8.—Average Charge per Patient Visit inPractices With and Without a Physician

Extender, 1977

Average chargeper patient visitb

Type of practlcea and provider (in dollars)

Phystcian with physician extender ., .. .$13.00NP ., . . . . . 8.13PA or Medex 12.02Physician ., ., . . . . . . . . . . . 15.06

PhysicIan without physician extender 16.48


lower income areas with fewer available healthresources could account for the charge differ-ential.

Few physicians surveyed by the General Ac-counting Office reported any reduction in feesfor physician extender services (27). Moreover,CBO cautions that use of NPs to provide com-plementary services rather than to increasevolume could lead to even higher average per-visit charges (18). Existing data lead to the con-clusion that where prices for patient services dodecline after introduction of a physician ex-tenders, the change is insufficient to lead tomore than a modest reduction in averagecharges per patient visit. The tendency for phy-sician extenders to order more diagnostic testscan further increase practice revenues if the testsare performed in the physician’s office (84). Inaddition, if the physician extender assumessome of the nonreimbursable physician services(e.g., telephone consultation, prescription re-fills), that frees physician time to provide reim-bursable services. Therefore, in the majority ofcases (no price change or modest price change),the physician extender’s income generation po-tential is enhanced by the fact that additionalreimbursable services are being provided andthat prices may be excessive in relation to thecosts of physician/physician extenders prac-tices. Whether the introduction of physician ex-tenders might cause fees to increase less rapidly,thus leading to a relative decline in prices, hasnot been examined.

There is little reason to expect physicians tocharge a lower price for physician extender serv-ices or to reduce their fees if the use of physicianextenders leads to a reduction in physician timerequired per patient visit. There is no incentivefor physicians to do so. Bicknell, et al., write(9):

On grounds that they bear ultimate responsi-bility, most U.S. physicians employing assist-ants expect to continue receiving their cus-tomary fees from all patients, including those ex-amined and treated by the assistant. The resultbetrays the promise of primary care assistants.Instead of bringing a reasonable dimension toprimary care costs, inadvertently the assistantsmay maximize the worst in the fee-for-servicesystem . . . .

Berki (7) argues that current fee levels do not re-flect physician time inputs, citing a study bySchonfeld, et al. (83) that found office visit feesto be significantly higher than could be justifiedby the value of physician time devoted to suchvisits. Berki also describes the possible dilem-mas of fee differentiation. While physicianscould achieve financial benefits from the em-ployment of NPs, even if they received a lowerfee for NP services, the incentive to incorporatethis innovation into their practices diminisheswith any reductions in NP revenue-generatingpotential. Such fee differentiation also mightengender resistance among patients who viewthe lower price of NP services as a signal thatcare provided by NPs is of lower quality thanthat rendered by physicians.

With current physician pricing behavior, NPsare not only a cost-effective addition to theirpractices, but often a profitable investment forthe physician. NPs’ income-generation potentialgives the physicians more flexibility in maxi-mizing their combined income and leisure ob-jectives. Table 9 illustrates the physician’s abili-ty to work less time after employing a physicianextender while suffering no loss in income.Rather than working less time and maintainingtheir incomes, physicians could choose to in-crease their incomes by maintaining their pre-vious level of effort and by generating evenmore revenue from an increased volume of pa-tients. This alternative appears to be much moreprevalent (103). Most studies of revenue gen-eration and profitability found higher expenses

Table 9.—Physician Time Input and IncomeWith and Without Employment of a New Type

of Health Manpower (NTM)

One NTMNo NTM employed—-——

Physician practice hours/week . . . 60 60Weeks/year . . . . . . . . . . . . . . . . . . . 50 46.7Office vists/week . . . . . . . . . . . . . . 120 150Visits/year . . . . . . . . . . . . . . . . 6,000 7,000Fee/visit. . . . . . . . . . . . . . . . . . . . . . $10 $10Fee/volume/year . . . . . . . . . . . . . . . $60,000 $70,000Market cost/NTM ... . . . . . . . . $10,000Income/year. . . . . . . . . . . . . . . . . . . $60,000 $60,000


for practices incorporating physician extendersthan for physician-only practices, but alsofound physician/physician extenders’ revenuesto be sufficiently higher to show greater profits(21,4O,65,66,8O,81,86,1O8).

The amount of profit realized through the em-ployment of NPs varies considerably amongpractices. In a separate study of 26 NPs inpediatric practices reported in 1972 byYankauer, et al. (108), average annual grossrevenues generated by the NPs exceeded theirexpenses by an average of $2,500. In a study ofpediatric NPs in 1969, Schiff, et al. (81) foundthat the NP generates net revenues of about$6,000 after 1 year of practice. Schwartz (86) ex-amined the revenue-generating experiences ofthree different types of practices employing NPsin California. He found that the average annualnet revenue of an NP in a rural solo privatepractice in 1974 was $18,653, resulting in an in-crease by more than one-third in the employing

physician’s income. Kane, et al. (40) found thatthe higher profits earned by physician/Medexpractices were in part due to the physician’sbeing relieved by the Medex to spend time onmore highly remunerative activities, such asspecialized procedures and inpatient care. Evenwith limited third-party reimbursement for phy-sician extender services, only one major studycited this as a financial problem for physiciansemploying NPs (93).

From the physician’s perspective, the NP is acost-effective addition to the practice. The factthat the financial benefits gained from the use ofNPs are not passed on to consumers is well doc-umented. As noted earlier, the potential in-crease in income afforded the physician by theNP is a major incentive for NP employment.Given that income increases are usually the re-sult of expansion in the volume of patient visits,consumers may benefit from improved access toservices. While this has been the experience ofpractices in areas with few available health re-sources, it is unwise to assume that increasedvolume, regardless of its character, generallyleads to improved access.

Finally, the price effects of physician extenderemployment in organized systems with cap-

itation should be treated separately. The pro-ductivity gains from physician extender em-ployment are likely to result in cost savings insuch systems. Given that they provide specifiedservices in return for monthly cavitation pay-ments, such organizations may choose to ex-pand their scope of benefits or increase physi-cian salaries or leisure rather than to pass thesavings to the consumer in the form of reducedpremiums. Which action such organizationschoose depends on how they assess their com-petitive advantage over similar practices in thearea and standard insurance carriers (7).

Medical Care Expenditures

Practices employing physician extenders gen-erally see more patients than those without phy-sician extenders, increasing volume by as muchas 50 to 60 percent (18). Prices charged for serv-ices in physician/physician extender practicesdo not differ significantly from physician-onlypractices even though average per-visit practiceexpenses tend to be lower. Given current em-ployment and pricing patterns, NPs and PAs doincrease medical expenditures beyond the ex-penditures that would have occurred withoutthem. Because of their small numbers (physi-cians outnumber physician extenders by 18 to1), physician extenders’ current impact on totalexpenditures is marginal. If physician extenderswere slated to play a substantially larger role,with no other changes, in the current healthservices delivery system, their impact on overallmedical expenditures would grow.

CBO estimated that in 1977 medical expend-itures for practices employing physician ex-tenders were 19 to 24 percent higher than thosefor practices without physician extenders (seetable 10). If 50 percent of physician practicesemployed physician extenders, extrapolation ofcurrent trends for both pricing and patientvolume increases would result in at least a 10-percent increase in total expenditures related tophysician practices, Given the expanding supplyof physicians, it is uncertain whether the in-crease in volume of patient visits required tosupport this higher number of physician/physi-cian extender practices is possible.

Total expenditures associated with physicianextenders must include training expenditures,much of which are subsidized with public funds.The costs of physician extender training are sub-stantially lower than those for physician train-ing. To the extent that physician extenders sub-stitute for physicians, and that substitution isreflected in the respective numbers of physicianextenders and physicians trained, the lowerphysician extender training costs could reducetotal training expenditures. For now, the ex-tension of public support for physician extenderprograms means an increase in public expend-itures for health professions training. Since thepotential savings from this public investmentare not passed on to the consumer, medical careexpenditures also are inflated by the intro-duction of physician extenders.

The increase in medical care expendituresassociated with physician extenders may be out-weighed by the benefits their presence bringsthrough increased access. In a recently reportedsurvey, 57 percent of physicians who employNPs cited the extension of services to more peo-ple as the NP’s major contribution to medicalservices delivery (55). Improved access occursnot only as the result of the general increase involume of patient visits in practices with NPs,but also because NPs tend to serve more low-income and nonmetropolitan patients who tra-ditionally have had diminished access to physi-cians (61,98). NPs in underserved areas can freeoverextended physicians to focus on complexmedical problems and consultant services (104).Furthermore, with the NP to provide followup,such physicians can discharge hospital patients

sooner (91). Improved access to primary careservices may reduce expenditures for costly spe-cialized services and hospitalization. To the ex-tent that NPs contribute to the expansion of pri-mary care services, the increase in expendituresaccompanying their use could be offset by suchsavings.

Cost Effectiveness:Actual or Potential?

That individual NPs can be cost effective isdocumented in numerous studies. Generalizingthat experience to the total NP or physician ex-tender population or basing future projectionson individual experiences is more difficult.Although the number of physician extenders isexpected to more than double in the next decade(see table 11), physicians will still outnumberphysician extenders by 10 to 1.

Given the nature of the U.S. health care sys-tem and realistic expectations for future struc-tural change, it is uncertain how many NPs andPAs will find employment commensurate withtheir training and whether society will benefitfrom their cost-effectiveness potential. In theirlongitudinal study of NPs, Sultz, et al. (99)found that only 50 percent of employed NPsfunctioned purely as NPs (see table 12). Otherswere either performing mixed functions or pro-viding only traditional nursing services. It mustbe noted that this study was based on 1974 data,thus reflecting the early experience of NPs.Many master’s students at that time were facul-ty in schools of nursing preparing to teach in NPprograms. This, in part, may explain the seem-ingly low percentage of NPs functioning fulltime in the NP role. Data on more recent ex-perience are essential in order to more accurate-ly estimate the education costs of one fully func-tional NP, the increased availability of NP serv-ices resulting from additions to the supply in thefield, and the prospects for future employment.Given their small numbers and current employ-ment experience, the total impact of physicianextenders, even if beneficial, will be modest.

The structure of the U.S. health care system isoften cited as the major factor inhibiting theachievement by NPs of their full potential. Ac-cording to Bicknell, et al. (9), the “hospital-based, specialist-intensive, resource-rich” sys-

tem is incompatible with primary care practiceand resistant to innovations that augment pri-mary care capacity. Patterns of financing thatcover the costs of an inefficient delivery process,medical education that discourages delegationof patient responsibility, and the prominent role

of physicians in defining the boundaries of prac-tice for other health professionals inhibit thegrowth and efficient utilization of a professionthat may invade territory traditionally confinedto physicians.

Within the existing structure, incentives toemploy NPs vary according to practice arrange-ment, physician payment mechanisms, andbudget constraints. In general, the financial in-centive for physicians in private practice to hirean NP or PA is diminished, because physiciansearn high incomes and are not constrained bycompetitive market forces to produce services inthe most cost-effective manner. On the otherhand, employment of physician extenders canoffer attractive benefits to physicians. Physicianextenders allow physicians to expand their prac-tices to improve patient access and increase in-come. They provide the physician an opportuni-ty for more leisure time or a more leisurely workpace. Finally, many physicians who emplo y

physician extenders stress their contribution toupgrading the quality and comprehensiveness ofcare provided by their practices (90).

Multiple considerations enter into the de-cision to hire a physician extender. The phy-sician must be convinced that sufficient demandexists for the planned expansion in patientvolume. Smith (90) cautions, “If the present hin-drances of access to medical care are, as somehave suggested, merely geographical, we mighthave the ironical situation of there being nophysician to hire the practititioners where theyare needed and no need for them where there is aphysician to hire them. ” Legal and reimburse-ment policies must not constrain physician ex-tenders’ potential to meet the physician’s expec-

table 12. —Employed Graduates of NP Programs by Role and Type of NP Program

Type of program

Certificate Master’s Total

Role Number Percent Number Percent Number Percent. NP role only . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

— . . — . — — . — —257 55.2°10 67 35.8%. 324 49.70/0

Traditional nursing role only, . . . . . . . . . . . . . . . . . . 51 11.0 71 38.0 122 18.7NP role and traditional nursing role, . . . . . . . . . . . 157 33.8 49 26.2 206 31.6

Totala. . . . . . . . . . . . . . . . ., . . . . . . . . . . . . . . . . 465 100.0 ”/0 187 ‘- - 100.0% 652 100.0%


tations. Physicians with minimal experience intask delegation may not feel comfortable usingphysician extenders in the manner required tobenefit fully from their employment. Some ini-tial investment for salary and perhaps addi-tional overhead accompanies the introductionof the physician extender into the practice.Sultz, et al. (98) reported that 92 percent ofphysicians employing NPs thought the benefitsof their employment outweighed the costs. Forthe physician contemplating whether to hire, theincentives also must seem to outweigh the risks.

The rapid expansion in medical school enroll-ments and projected increases in physician sup-ply add a new dimension that may overshadowother factors influencing physician extenders’employment opportunities. Physician supply isgrowing and there are thousands more in theeducational pipeline who must be absorbed intothe system. The effects of this increase in physi-cians are now only being contemplated, butthey could be profound. Among them could bea restriction in employment opportunities forNPs and PAs. Some argue that NPs will farebetter because of their nursing background andscope of practice (18). Regardless, there arealready signs that physicians, particularlyspecialists, are redefining the scope of theirpractices in response to diminished numbers ofpatients requiring their specialized skills. Aiken,et al. (3) recentl y concluded, “Despite the cur-rent shortage of generalist-physician services,continuing specialist participation in primarycare will lead to sufficient generalist medicalservices by the mid-1980’ s.” Physicians may berecapturing primary care responsibilities thatnot so long ago they considered delegating.Moreover, it has been suggested that practicingphysicians who perceive this oversupply mayhire young physicians to perform the tasks thatphysician extenders can handle (64).

Organized settings that operate on fixedbudgets (prepaid group practices and some clin-ics) have a much greater incentive to employNPs and PAs. It is to their financial advantageto produce services with the most efficient com-bination of inputs, substituting lower pricedphysician extenders for higher priced physicianswhenever possible. Such organizations will dis-

play more efficient staffing patterns unless theyare able to pass through the costs of more in-efficient personnel configurations to third-partypayers who reimburse on a cost basis. The valuesuch organizations place on physician extendersis reflected in the fact that half of all PAs and 80percent of all NPs are employed in organizedsettings (18). Future changes in practice ar-rangements and the preferences of new physi-cians toward entering organized practice set-tings or private practice, therefore, will have animpact on future opportunities for physician ex-tenders.

Prepaid group practices (PPGPs) haveperhaps the strongest incentive to employ physi-cian extenders. Operating revenues derive fromfixed cavitation payments for plan enrollees.The PPGPs seek to minimize expenses in rela-tion to revenues by using more cost-effectivemeans of achieving comparable outcomes. Thisincludes the substitution of NPs and PAs forphysicians wherever possible. While they haveyet to follow their own maximum substitutionmodels, HMOS, the prototype PPGP, are lead-ers in employing physician extenders and con-ducting research on their actual and potentialutilization (33,69, 71,94). CBO reported that in1977 HMOS provided care to their memberswith an FTE of 0.44 physician extenders for eachphysician, as opposed to 0.07 physician ex-tenders for each office-based physician in theUnited States (18).

HMOS have the advantage of greater flexi-bility in modifying personnel arrangements togain the benefits from substitution. Physiciansin private practice, beyond fulfilling their objec-tives for leisure time, are unlikely to reduce theirtime inputs to achieve a more efficient opera-tion. Where physicians are salaried employees,however, the efficiency objectives of the em-ploying organization may lead to a reduction intheir time, numbers, or income. In line with theHMOS’ incentive to minimize total salaries,Greenfield, et al. (33) report on the effects of anexperimental physician extender protocol sys-tem introduced at the Southern CaliforniaKaiser-Permanente Facility and the adaptationof the organization to a more cost-effectivemodel. Before the study, 10 physicians and 3

physician extenders saw 2,700 patients permonth, 70 percent of whom had acute illness.Two years after the study, 6.5 physicians and 6physician extenders saw 2,900 patients permonth, 70 percent of whom had acute illnesses.

The use of NPs and PAs results in pro-ductivity gains and cost reductions. Yet, theirfuture participation in medical care delivery isuncertain. Available evidence indicates thatphysician extenders’ incorporation into orga-nized settings, particularly HMOS, has contrib-uted to more cost-effective service delivery.However, the experience in private physicianpractices is less promising. Demand for physi-cian extenders in that setting has been limited.Although the public benefits from increasedavailability of services, the cost effectiveness ofphysician extenders in such settings has not re-duced prices. Moreover, since most physicianextenders are employed in organized settings,employment opportunities are limited by thefact that the majority of physicians are in theprivate practice, fee-for-service sector.

Government subsicly of NP and PA traininghas not been accompanied by policies to ensurethe promise of these health professionals oncethey are in practice. Moreover, policies thatmay inhibit the use of physician extenders, suchas those supporting expansion in physician sup-ply, have been enacted simultaneously withpolicies encouraging their development.Changes could be made-short of a nationalhealth insurance scheme with incentives torestructure health services delivery—that wouldfacilitate the efficient use of physician extenders.Modifying current reimbursement policies tocompensate for physician extender services in amanner that reflects their lower costs would af-fect prices, although it might also reduce op-portunities for physician extender placement inphysician private practices. Opportunities forNPs to practice more independently through re-moval of legal and reimbursement constraintscould provide consumers with a lower cost al-ternative for receiving the primary care servicesthat both physician extenders and physicians

can provide. The expansion of PPGPs and othermedical care organizations that operate on fixedbudgets would provide more employment op-portunities for physician extenders in settingsthat use them more efficiently, As Berki argues(7):

Where medical care is provided by hierar-chical organizations on a cavitation basis, bothservice expansion and price reduction are likely.Thus, not only is effective use of new types ofhealth manpower enhanced by the structuralcharacteristics of hierarchical organizations, butalso the gains flowing from their employmentare less likely to accrue to the providers, and themore likely it is that consumers will benefit byincreased availability of services at lower prices.

In evaluating the role of physician extenders,it is insufficient to assess their cost effectivenesswithout also looking at who gains from the sav-ings. Are the financial benefits of lower trainingand employment costs to be shared with thepublic or reaped only by providers? Under thecurrent fee-for-service system, are the modestsalaries of physician extenders exploitative,given their income-generation capacity? Theorganization and financing of health services inthe United States encourage inefficiency in thedelivery of medical care. Reforms that wouldoptimize the efficient use of physician extendershave implications for the cost effectiveness ofother components of the system. Physician ex-tenders can be integrated into the existingsystem, as they have for the past decade, withperpetuation of existing inefficiencies. From apublic policy perspective, it must be askedwhether improved access, resulting from em-ployment of physician extenders, but unaccom-panied by 1ower prices, is worth the further in-flation in medical care expenditures. This is thechoice as the system currently functions. Giventhe overriding concern for containing healthcare costs, an assessment of NPs and PAsshould consider not only their performancewithin the constraints of the current system, butalso their potential role in an integrated strategyof reform to meet public policy objectives.

Case Study #16. The Costs and Effectiveness of Nurse Practitioners ● 25

APPENDIX: BACKGROUND INFORMATION ONPHYSICIAN EXTENDERS

Training

Although physician extender training existed on avery limited scale as early as the 1930’s, the majorthrust in training both NPs and PAs in the UnitedStates began in 1965 (5). Duke University establishedthe first primary care PA training program. PAs gen-erally receive 2 years of academic and clinical train-ing in a medical school setting. Although postsec-ondary education or previous experience in a healthprofession is not specifically required, most PAs havehad 3 or more years of college-level education orseveral years’ experience in a related health field suchas medical technology, physical therapy, or voca-tional nursing. While the initial intention of PA pro-grams was to provide training in primary care, abouthalf of all PAs now specialize in other fields such asophthalmology, urology, orthopedics, internal medi-cine, and other medical and surgical subspecialties(46).

Medex often are considered in the same category asPAs because they practice under the same legal au-thority and receive similar preparation. Medex train-ing was conceived as a way to allow former militarycorpsmen to apply their skills to rural medical prac-tice. The first training program was established in1969 at the University of Washington. Medex receive15 months of training, including preceptorship,often with rural physicians. Unlike PAs, Medex aretrained almost exclusively in primary care.

The concept of expanded function nursing grewout of a program to train nurses to become pediatricNPs at the University of Colorado in 1965 (49). Alsoreferred to as nurse clinicians, nurse associates,clinical nurse specialists, or nurse generalists, these

Specialty

Pediatric. . .Midwi fery .M a t e r n i t yFamily. . . . .Adult . . . . . .Psychiatric.

Total .

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

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

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

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

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

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

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

individuals all belong to the general category knownas NPs, The NP receives additional training beyondthat required for an RN license3 in a particular spe-cialty such as pediatrics, family practice, maternity,adult practice, or psychiatry. There are two types oftraining programs: those that offer an NP certificateand those that award a master’s degree. Table A-1shows the distribution of different NP training pro-grams by both specialty and type of degree offered.

Until standardization of the length of NP cer-tificate programs to 1 year in the Nurse Training Actof 1976, NP training varied between master’s and cer-tificate programs as well as among specialties. Mas-ter’s degree programs require a previous baccalaure-ate RN license and, on average, require more than ayear of training; certificate programs are now 1 yearin length. Table A-2 shows the average length of dif-ferent NP training programs.

Legislation Supporting PhysicianExtender Training

Federal support for physician extender trainingwas limited before 1970. Some early NP training pro-grams received assistance through special projectgrants provided under the Nurse Training Act of1964 (Public Law 88-581) and, later, title 11 of theHealth Manpower Act of 1968 (public Law 90-490)(103). The National Center for Health Services Re-search funded the first Medex training program at the

Table A-1 .—NP Programs by Specialty and Type of Program

T y p e o f p r o g r a m “-


Number Percent Number Percent Number Percent

42 48.8 ”/0 ‘8 17.8°\o 50. 38.2-0/05 5.8 6 13.3 11 8.4.7 8.1 7 15.5 14 10.7.

17 19.8 12 26.7 29 22.115 17.5 8 17.8 23 17.6. .

— — 4 8.9 4 3.0. .86 1OO.O O/O 45 100.OYO 131. . 1OO.OO/O


University of Washington. By the late 1960’s, PAtraining programs were receiving funding from avariety of Federal sources, including the Office ofEconomic Opportunity, the Model Cities Program,the Veterans’ Administration, the Public HealthService, the Department of Defense, and the Depart-ment of Labor (52). However, most physician ex-tender training programs during this period de-pended on institutional or private resources.

In the early 1970’s, the Federal Government be-came more interested in the potential of physician ex-tenders to address health manpower problems. In-creasing concern over rising costs and the continuedshortage of physicians in primary care was reflectedin two major pieces of legislation aimed specificallyat increasing the number of NPs and PAs. The Com-prehensive Health Manpower Act of 1971 (PublicLaw 92-157) provided the first large Federal pro-vision for NP and PA training programs (35). TheNurse Training Act of 1971 (Public Law 92-150) pro-vided broadened authority for special project grantsand contracts including support for training pro-grams for NPs (99). Passage of the Nurse TrainingAct of 1975 further reinforced the Federal commit-ment by establishing a new, separate section for sup-port of NP training. Further, in 1977, the Health Pro-fessions Educational Assistance Act of 1976 (PublicLaw 94-484) was amended by the Health Services Ex-tension Act (Public Law 95-83) to provide additionalgrants and contracts for physician extender trainingprograms (64). Although the Nurse Training Act of


Table A-3.—Practice Setting Location of NPs by Type of NP Programa

Type of program


Practice setting location Number Percent Number Percent Number Percent

Inner city . . . . . . . . . . . . . . . . . . 126 31 .40/0 41 42.4°10 167 33.6°\oOther urban . . . . . . . . . . . . . . . . 65 16.3 24 24.7 89 17.9Suburban . . . . . . . . . . . . . . . . . . 35 8.8 11 11.3 46 9.3Rural . . . . . . . . . . . . . . . . . . . . . . 77 19.3 7 7.2 84 16.9Combination. . . . . . . . . . . . . . . 16 4.0 3 3.1 19 3.8OtherC. . . . . . . . . . . . . . . . . . . . . 81 20.2 11 11.3 92 18.5

Totald . . . . . . . . . . . . . . . . . . . 400 1 OO.OO/O 97 1 OO.OO/O 497 1 OO.OO/O

The tendency of NPs to locate more often in inner-city areas may be partially explained by the fact thatNP training programs are largely based in urban cen-ters (see table A-4). Moreover, most nurses, whocomprise the applicant pool for NP programs, areemployed in metropolitan areas initially and thus aremore likely to remain there (18).

In the past, it has been difficult for many under-served areas to attract and maintain physicians.However, practice in these areas may have specialadvantages for physician extenders. One study of 85NPs in rural areas (97) showed that 40 percent choserural practice because it “offered a creative approachto health care delivery. ” Another 25 percent were inrural settings because of “the opportunity for roleautonomy. ” Both of these responses can be inter-preted as reflections of the inadequacy of physiciansupply in relation to consumer needs. What may bedefined by physicians as problems of rural practicemay instead present themselves as opportunities for

physician extenders which may not be available inareas with an adequate or excess supply of physicians(8,15,37,73,97),

Several training programs have been developedspecifically to prepare NPs to practice in underservedareas. One of the largest such programs in the coun-try is the FamiIy Nurse Practitioner Program at theUniversity of California. Students in this programare recruited from rural areas and trained with a self-selected physician preceptor in a rural practice. Ad-ministrators of the program feel that this system hashelped to promote a positive relationship betweenphysicians and NPs and to encourage their furtheruse in rural areas (88). Other institutions, includingthe University of North Carolina and the Universityof Minnesota, operate similar training programswhich place NPs in Area Health Education Centersand other rural health clinics in their States (28). InKentucky, the Frontier Nursing Service trains nurse-midwives and family NPs to provide services to ruralareas of southeastern Kentucky and is the only avail-able source of primary care to most residents in thisarea (62). These programs have succeeded in placingand maintaining a high percentage of their graduatesin rural areas.

The practice location of NPs varies by both theirspecialty and type of training program from whichthey graduated. The specialties most likely to berepresented in inner-city locations are pediatrics,midwifery, and maternity, while family NPs are themost likely to be represented in rural areas (see tableA-5). Moreover, a larger percentage of practitionersfrom certificate programs (19 percent) than graduatesof master’s degree programs (7 percent) is in ruralareas.

Although physician extenders have made theirpresence known in underserved communities, and, as

28 . Backround Paver #2: Case Studies of Medical Technologies

Table A-S.—Practice Setting Location of NPs by Specialty and Type of NP Programa

Type of Pediatric Midwifery Maternity Family Adult Psychiatric Totalprogram/practice - -— -- ---—– —–– -——-– ——–-.–– —. ---setting location Number Percent Number Percent Number Percent Number Percent Number Percent Number Percent Number Percent

— — —

CertificateInner city. ., .,Other urban. . . .S u b u r b a n .R u r a l . . .C o m b i n a t l o n b.O t h e rc

Total. . . .

Master’sInner city. .Other urban, .,S u b u r b a n .Rural . . ., . .C o m b i n a t i o n b.Other c . . . . . .

Total. . . . . . . .

TotaldInner city, . . .,Other urban. , . . .Suburban. . . . . .Rural . . . .C o m b i n a t i o n b.Otherc . . .

Total. ., . . .,

702820261115

170

3111

—17

732921271116

177

—

41.1 0/0

16.511.815.36.58.8

100.0

42.80/014.314,314.3

—

14.3100.0

41 .2’7016.411.915.36.29.0

“100,0

3212

—

19

1263

—

22

25

1584223

34

810

14—

225

31————4

1111

14—

229

2814

529

521

102

1612

5317

44

44261032

628

146—.

———————

——21

——

3

——

21

——

3

———————

——

66.7 %33.3——

100.0

——

66.70/033.3——

100.0—

1266535771681

400

412411

73

1197

1678946841992

497

31 .40/o16.38.8

19.34.0

20.2100.0

42.4%24,711.37.23.1

11.3100.0

33.60/’17.99.3

16.93.8

18.5100.0.

a group, they seem most easily attracted to theseareas, it should not be assumed that this trend willcontinue indefinitely. As noted earlier, under currentlegal arrangements, physician backup must be avail-able to physician exterders. It is uncertain how manyphysician extenders can be absorbed in underservedareas given this constraint. Also, until now, manyrural practices have been supported by Federal andState funds. If these resources evaporate in thefuture, such rural practices would have to becomefinancially self-sufficient in order to retain existingpersonnel or to employ new personnel. A study ofphysician extender practices in rural communities inCalifornia has shown that freestanding rural prac-tices can become financially viable with continuedFederal and State support (31). But without such sup-port or third-party reirnbursement, economic viabili-ty is much less likely (63).

Physician and Consumer Acceptance

From the beginning, physician and consumer ac-ceptance of NPs has been a major issue. The wide-spread use of NPs depends on the willingness of phy-sicians to hire them and patient receptiveness to the

kinds of services they provide, particularly in privatephysician practices.

Current Federal reimbursement policies, which donot include payment for services provided by physi-cian extenders, serve as a disincentive for physiciansto employ physician extenders. Among the other rea-sons most frequently cited by physicians for not hir-ing NPs are legal restrictions, limitations on space orfacilities, and resistance from other health providers(100). Concerns about liability and malpracticecoverage also may discourage physicians from hiringNPs (45,56,107). However, the use of NPs also pro-vides advantages for physicians to expand the num-ber of patients in their practices, increase their in-come, broaden the scope of their services, or acquiremore flexibility with their time. Physician extendersmay be useful particularly to rural physicians, who,because of their fewer numbers, must see consid-erably more patients and work longer hours thantheir urban counterparts (46).

The Institute of Medicine (IOM) reported on aseries of studies in the last decade which attempted todocument both physician approval of the concept ofNPs and physician willingness to hire them (77).While the findings varied significantly among the

Case Study #16 The Costs and Effectiveness of Nurse Practitioners ● 29

studies, the general pattern showed that physicianapproval of the concept of NPs often was not accom-panied by a willingness to hire them. While one sur-vey cited by IOM found that as many as 70 percentof physicians would hire an NP (67), most othersfound considerably lower interest. For example,Lawrence, et al. (45) found that only 24 percent ofthose surveyed would hire an NP, although 86 per-cent expressed approval of the concept (45).

Physician attitudes toward NPs are perhaps morestrongly influenced by their medical education ex-perience than any other single factor. Medical educa-tion does not encourage delegation of patient-care-related tasks to nonphysicians and emphasizes ulti-mate physician responsibility in all aspects of care.Physicians may feel that by using NPs they will sac-rifice quality of care and important elements of thephysician-patient relationship. As a result, physi-cians may be particularly reluctant to delegate tasksto a large degree or to accept the NP as a professionalcolleague (19,96).

In its review of studies on task delegation, IOMconcluded that while physicians who employed NPswere satisfied with their performance, they were“more disposed to delegate duties that are in therealm of nursing practice and not in the realm ofmedical practice” (105). While confident of the NP’sability to take medical histories, provide counseling,and perform other routine tasks relating to primarycare, many physicians express hesitation in allowingNPs to perform physical examinations or other moretechnical procedures. In the past, pediatricians havebeen the most receptive to NPs functioning in an ex-panded capacity, but the recent increase in the num-ber of pediatricians and pediatric NPs, combinedwith the declining birth rate, may alter this pattern inthe future (1).

Because exposure to team practice is a recent andstill limited innovation in medical education, olderphysicians (particularly those in long-standing soloprivate practices) are somewhat less likely thanyounger ones to be receptive to NPs, Physicians ingroup practices generally are more willing to employNPs than solo practice physicians. Several explana-tions are offered for this pattern, including the sug-gestion that members of group practices are more fa-miliar with team practice and more willing to dele-gate tasks to gain time for other professional ac-tivities or for leisure (45,54).

Because of the reluctance of physicians in privatepractice to employ NPs, most NPs practice in orga-nized settings. As seen in table A-6, only 14.2 percentof NPs surveyed by Sultz, et al. (100) were employedin private physician practices.

Consumer acceptance of NPs is essential to the in-tegration of NPs into medical services delivery. The

patient’s confidence in the NP’s professional com-petence, the quality of communication between pa-tient and NP, and improved access to services due toNP participation in the practice are key factors inachieving consumer acceptance.

From the beginning, consumers have shown lessresistance than physicians to NPs. Studies on con-sumer acceptance reviewed by IOM show that pa-tients seem to perceive little or no difference betweenphysicians and NPs once an ongoing relationship isestablished (76). In fact, several studies have foundpatients preferring NPs for many services previouslyprovided by physicians (41,48,53,57). Other studieshave reported that patients under the care of NPsbroke fewer appointments and complied more strict-ly with prescribed treatments than those of physi-cians (14,51,81,85). NPs tend to spend more time perpatient visit and place greater emphasis on counsel-ing and education. These aspects of NP practiceenhance their attractiveness to patients.

As with physicians, consumer acceptance of NPsincreases with exposure to them. It is here that thephysician plays a critical role. The patient often usesthe physician’s attitude toward the NPs as a signal toaccept or reject NP services (68). Moreover, reportsof experimental projects introducing NPs noted thenecessity of assuring patients assigned to NPs thatsuch assignment would not restrict their access to thephysician. As a result, patient confidence in theavailability of the physician was generally sufficientfor them to receive the majority of their care from theNP (17).

Obviously, acceptance by physicians and consum-ers is just as great an issue for PAs as for NPs.Because of the differences in their training and skills,one group may be preferred by physicians in certainsituations. Whereas the PA usually functions moredirectly with the physician or as a physician sub-stitute, the NP may bring a broader spectrum of serv-ices to a practice. This is an area which has yet to beexplored in depth.

Legal Restrictions

Since licensure is a State responsibility, each Statehas legal jurisdiction to regulate the practice of healthprofessionals. As new professions emerge, they seekrecognition by the State. In such instances, the Statemust act with little precedent or experience to guideits actions. The lack of uniformity in State regulation

of NPs and PAs reflects their relatively recent in-troduction and the continued uncertainty about theirappropriate role in health services delivery. More-over, as pointed out by CBO (18), changes in medicalpractice legislation generally occur in consultation

30 . Backround Paper #2: Case Studies of Medical Technologies

Table A-6.—NPs by the One Employment Setting in Which They Spent Most of Their Timeas NPs and Type of NP Programa

Type of program


Employment setting Number Percent Number Percent Number Percent

2217

5248

591575

986

80

4

3243

142

402

——

47

In-hospital practice. . . . . . . . . . . . . ‘- — --- .- . - - —

Patient unit . . . . . . . . . . . . . . . . .Emergency room. . . . . . . . . . . . .

Ambulatory clinical practice . . . . .Private practice. . . . . . . . . . . . . .Prepaid group practice. . . . . . . .Hospital-based clinic . . . . . . . . .Community-based clinic

or center. . . . . . . . . . . . . . . . . .Other ambulatory practice. . . . .

Nonhospital institutional setting .School for mentally or

physically handicapped. . . . .Grades 1-12, public school

system. . . . . . . . . . . . . . . . . . .College health programs . . . . . .Other non hospital institutional

setting . . . . . . . . . . . . . . . . . . .Nonhospital community setting . .

Health department orhome health agency. . . . . . . .

Social services or agency . . . . .Other nonhospital community

setting . . . . . . . . . . . . . . . . . . .School of nursing . . . . . . . . . . . . . .Extended care facility. . . . . . . . . . .Other b . . . . . . . . . . . . . . . . . . . . . . .

Overall . . . . . . . . . . . . . . . . . . . . . 403 100.070 97 100.070 500 1OO.OO/O

5.5%4.21.3

61.514.63.7

18.6

23.11.5

19.9

1.0

7.910.7

0.310.4

9.90.5

1312

16912

527

2321

1

8

61

13.4”/012.4

1.071.212.45.2

27.8

3529

6317

7120

102

7.0 ”/05.81.2

63.414.24.0

20.4

23.72.11.0

1.0

1168

81

4

3244

23.21.6

16.2

0.8

6.48.8

8.21

500.2

10.0

6.21.0

453

9.20.6

1.01.7

1231

1.02.13.11.0

1278

0.20.41.41.6

with the medical profession. The requirements forphysician supervision, restrictions on functions suchas drug prescribing, and prohibition of independentpractice all reflect the influence of organizedmedicine on this process.

In 1978, Miller and Byrne, Inc. (58) published asurvey of State legislation governing the practice ofNPs and PAs. Each State regulates NPs through theState’s nurse practice act. In some States, NPs areable to practice without significant changes in ex-isting statutes. Although they perform functionsbeyond that of traditional nursing, their expandedrole is considered an extension of nursing allowed bythe statute, rather than a significant departure requir-ing new regulations. In States that prohibited nursesfrom engaging in diagnosis and prescription of treat-ment, NPs could not practice without new statutoryauthority. The response of these States has beeneither to replace previous statutes with new defini-tions of nursing roles or to amend existing law to ac-commodate expanded role nursing. The response ofsome States has been to provide more open-ended

authority in recognition of evolving nursing roles.For example, the California Nurse Practice Act statesthat nursing is “a dynamic field . . . which is con-tinually evolving to include more sophisticated pa-tient care, ” and also recognizes “the existence ofoverlapping functions between physicians and reg-istered nurses” and permits “sharing of functions”5(58). While States have generally recognized the par-ticipation of nurses in activities previously restrictedto physicians, they have maintained prohibitionsagainst nurses functioning independently in the med-ical sphere.

The PA, unlike the NP, represented a new type ofhealth professional previously not covered by Statelaw (58). For this reason, new statutory authoritywas required in every State where PAs were allowedto practice. Initially, States responded by expandingphysician delegator authority under the MedicalPractice Acts to allow PAs to work under physiciansupervision. The majority of States, however, have

Case Studv #16: The Costs and Effectiveness of Nurse Practitioners ● 31

enacted regulatory statutes giving the State board ofmedical examiners authority over training and em-ployment of PAs. As with NPs, the laws regardingPAs tend to vary from State to State. For example,legislation in New Mexico, Ohio, and South Carolinacontains extensive lists of specific medical taskswhich PAs may or may not perform. In other States,such as Oregon, where the laws are much less ex-plicit, the PA may perform whatever tasks are per-mitted by the supervising physician. There is alsowide variation regarding the particular type of super-vision required for PAs. Some States permit tele-phone consultation as a sufficient means of supervi-sion while others also require the physician to es-tablish written protocols and review all patientrecords on a regular basis. Very few States requiredirect “over-the-shoulder” supervision, but in remoteareas where the physician and the PA may be in dif-ferent locations, the physician is often required toregularly visit the facility where the PA works. Whenthe Miller and Byrne study (58) was published, fourStates had no guidelines to regulate PAs.

The legal restrictions placed on NP and PA prac-tice are a significant barrier to their integration intomedical delivery. Such restrictions can make it im-possible for physician extenders to practice at a levelcommensurate with their training. Moreover, thelack of uniformity in State laws limits the mobility ofphysician extenders. While all training programsmust meet minimum standards for accreditation, thecontent of such programs tends to reflect the law ofthe States in which they are located. While an NP orPA trained in one State may be capable of perform-ing authorized functions in other States, their specifictraining experience may not fulfill licensing re-quirements outside their State of training. Althoughthis does not prohibit relocation, it does create a ma-jor disincentive. Moreover, since most physician ex-tender training programs are concentrated on the eastand west coasts, it may be difficult to achieve a dis-tribution of graduates from areas where they weretrained to other areas of the country where their serv-ices are most needed.

Reimbursement

Reimbursement remains a major obstacle to the ex-panded use of physician extenders. Insurers are gen-erally reluctant to extend coverage to new serviceproviders because of either legal restrictions or desireto control costs and to avoid encouragement of otherhealth workers who provide similar services fromseeking compensation (18,58). While some private

third-party payers support the concept of reimburse-ment for physician extender services, virtually noneprovide payment.

Approximately half the States provide some typeof reimbursement for physician extender servicesunder their medicaid programs. In all cases, paymentis made to the supervising physician or institution.Federal reimbursement under the medicare programhas allowed institutions to include physician extendercompensation in their calculation of reasonable costfor reimbursement purposes. Federal payments forprimary care services provided by physician ex-tenders, however, have been restricted by provisionsenacted before these new professions were estab-lished. In most cases, services traditionally, per-formed by physicians are not reimbursable underFederal programs when provided by physician ex-tenders (64). Under medicare part B, reimbursementfor medical services rendered by physician extendersis restricted to those “furnished as an incident to aphysician’s professional services, of kinds which arecommonly furnished in physicians’ offices and arecommonly either rendered without charge or in-cluded in physicians’ bills” (213).6

In 1977, the Rural Health Clinic Services Act(Public Law 95-210) waived such restrictions in themedicare and medicaid programs for physician ex-tenders practicing in certified rural health clinicslocated in designated underserved areas. The Actprovides payment for physician extender serviceseven when not directly supervised by a physician atall times. This allows such clinics staffed only byphysician extenders, with physician backup, to pro-vide reimbursable primary care services to medicareand medicaid beneficiaries. Payment is on a reason-able cost basis and is restricted to those physician ex-tender services authorized under State legislation.Because of variation in legal and reimbursement pol-icies among States, the impact of Public Law 95-210on each State will differ (105).

Changes in reimbursement policy, while signifi-cant, have come slowly and on a very limited basis.The result has been to tie physician extender practiceto the supervising physician or institution. Becausemost physician extenders are currently employed inorganized settings, reimbursement restrictions havenot prevented the growth of their professions. With-out expanded employment opportunities in physicianprivate practices, however, physician extenders willnot be able to assume an important role in medical


care delivery. Current reimbursement restrictionsmake such an expansion unlikely. Modifications ofcurrent reimbursement policy could address thisproblem. A number of questions including the scopeof services to be reimbursed, method of payment tophysician extenders in physician offices, level of pay-ment for services that both physician extenders andphysicians can provide, payment levels in under-served v. adequately served areas, and the recipientof payment for physician extender services must firstbe resolved (64).

Practice With Physicians v.Independent Practice

The functions of NPs are directly influenced bytheir employment setting. NP training encourages in-dependence, responsibility, and autonomy as impor-tant aspects of professional development. As a result,NPs generally prefer employment settings which of-fer increased self-sufficiency and greater decision-making responsibility in patient care. Organized set-tings tend to offer less opportunity to meet these ex-pectations, but because such settings are more recep-tive to hiring NPs, the major portion of NPs areemployed in them (1o00. NPs might prefer to work inphysician private practices or in independent prac-tices, but the opportunities to do so are limited. Themajor portion of physicians are in private solo orgroup practices, and yet a 1976 survey by Sultz, et al.(99) found only 14 percent of NPs employed in physi-cian private practices. While the actual number ofNPs in independent practice is not known, the Amer-ican Nursing Association estimates that there areabout 300 private independent NP practices (18).

For many NPs, employment in physician privatepractices presents several perceived advantages overless flexible institutional settings. In particular, grouppractices which emphasize a “team” approach seemto allow NPs increased participation in many aspectsof practice (11,12), However, physicians may be dis-couraged from using NPs in their private practicesbecause of legal restrictions or ambiguities and lackof third-party reimbursement for NP services, Also,the increase in the number of medical school grad-uates may further reduce physician interest in NPs orPAs.

The experience of NPs in physician private prac-tices varies considerably, since the physician deter-mines the exact nature of their responsibilities. Insome situations, NPs may actively participate indiagnosis and treatment of illness, provide counsel-ing and patient education, and perform a wide rangeof other duties. In other settings, however, the physi-

cian may be unwilling to delegate a broad spectrumof tasks, and the functions of the NP may be limitedto more traditional nursing tasks (11,12).

As an alternative to their interest in employment inphysician private practices, a small number of NPshave attempted to establish independent practices.There are many difficulties in undertaking such prac-tices. Initially, an independent practice requires asubstantial financial investment. Startup costs can be$15,000 or more, and the NP must be prepared tooperate at a deficit until an adequate clientele can bedeveloped (2). Moreover, since in the limited caseswhere third-party reimbursement is available for NPservices, the payment is made to the supervisingphysician or institution, rather than directly to theNP, virtually all independent practices must rely onout-of-pocket payments from patients. As a result,very few independent practices have attained finan-cial self-sufficiency. Those that have been successfulusually have either been located in metropolitanareas or have provided a particular service not wide-ly available otherwise in the community, such ashome visits (2,110). However, more often NPs in pri-vate independent practice report that they must sup-plement their incomes through speaking engage-ments, teaching, and other nursing employment (2,4,6,23,95).

In addition to serious financial problems, the ap-propriateness of independent practice by NPs is ques-tioned by physicians, nurses, consumers, and policy-makers. It has been noted that nurses may perceiveNPs as taking on a more dominant role, similar tothat of physicians, and therefore may be reluctant tosupport independent practice (18). In one study, 73percent of nurses and 2 percent of physicians felt thatthe proper role for an NP is to practice with a physi-cian (18). Since NPs’ independent practices would de-pend on physician referrals to establish a clientele,the concerns expressed by physicians suggest thatsuch referrals generally will not be forthcoming.Consumer reluctance to use independent NPs comesfrom inherent resistance to any new provider, fromunfamiliarity with a nonphysician providing medicalservices, and because, when given the choice, manywould continue to prefer a physician. Some inde-pendent NPs’ practices have waited several monthsfor their first patients (87).

Independent practices will not develop under exist-ing characteristics of NP practice. As a result of cur-rent legal specification of the kinds of services NPscan perform without the supervision of a physician,the majority of NPs in independent practice providetraditional nursing care rather than primary medicalcare (50).

the costs and effectiveness of nurse practitioners (part 3

Documents