identifying, evaluating and implementing cost-effective skill mix

6
Journal of Nursing Management, 1999, 7, 265–270 Identifying, evaluating and implementing cost-effective skill mix G. RICHARDSON msc Research Fellow in Health Economics, Centre for Health Economics, University of York, York, UK Correspondence richardson g. (1999) Journal of Nursing Management 7, 265–270 G. Richardson Identifying, evaluating and implementing cost-effective skill mix Research Fellow in Health Economics Background The British National Health Service (NHS) employs a large number of Centre for Health Economics individuals, at great monetary cost, to provide direct care to patients. Changes in the University of York combinations of staC, including nurses, nurse practitioners and midwives, delivering this York YO1 5DD care have been shown to be eCective in many settings. UK Findings The (opportunity) cost implications of such changes in the skill mix are rarely evaluated adequately. The impact of releasing professionals’ time has not been estimated and therefore determining whether changes are cost-eCective is diBcult; these diBculties have often been increased by poor study design. Conclusions Economic evaluation has been under-utilized in studies of skill mix. If economic evaluation demonstrates that skill mix changes reduce cost and improve or maintain patient outcomes, this is strong evidence that these changes should be implemented. Incentives may be required to attract the necessary personnel. This in itself may influence the cost of changing the skill mix and therefore the situation should be monitored as both costs and eCectiveness can alter over time. Accepted for publication: 13 May 1999 Introduction eBciency should probably be implemented. This paper highlights some of the practical problems involved in The NHS employed almost 550 000 whole-time equivalent performing economic evaluations of skill mix alterations direct care staC in England alone in 1996 (Government and of implementing cost-eCective changes in the employ- Statistical Service 1997) costing an estimated £11 billion ment and deployment of nursing staC. in 1998 (based on figures from DoH 1998). Medical staC in primary and secondary care account for approximately 78 600 of this total, while the number of nurses in these Nursing and economic evaluation environments is over 340 000 Despite the large numbers of individuals providing It has been demonstrated that nurses and other health direct care and the associated burden on NHS resources, professionals can substitute for doctors without detrimen- and the fact that there are a large number of studies tal eCect on patient outcomes in a variety of settings such demonstrating the eCectiveness (or consequences) of as radiology (Thompson 1974) paediatrics (Honigfeld changes in skill mix, economic evaluations of changes in et al. 1990), obstetrics (McKee 1992), intensive care the skill mix remain relatively few in number. Economic (Dubaybo et al. 1991), neurology (Gunderson et al. 1984), evaluation is concerned with the costs and consequences orthopaedics (Harris & McCollister-Evarts 1990), mid- wifery (MacDorman & Singh 1998) and anaesthesia of an intervention; changes that demonstrate increased 265 © 1999 Blackwell Science Ltd

Upload: richardson

Post on 06-Jul-2016

222 views

Category:

Documents


4 download

TRANSCRIPT

Page 1: Identifying, evaluating and implementing cost-effective skill mix

Journal of Nursing Management, 1999, 7, 265–270

Identifying, evaluating and implementing cost-effective skill mix

G. RICHARDSON msc

Research Fellow in Health Economics, Centre for Health Economics, University of York, York, UK

Correspondence richardson g. (1999) Journal of Nursing Management 7, 265–270G. Richardson Identifying, evaluating and implementing cost-effective skill mixResearch Fellow in HealthEconomics Background The British National Health Service (NHS) employs a large number ofCentre for Health Economics individuals, at great monetary cost, to provide direct care to patients. Changes in theUniversity of York combinations of staC, including nurses, nurse practitioners and midwives, delivering thisYork YO1 5DD

care have been shown to be eCective in many settings.UK

Findings The (opportunity) cost implications of such changes in the skill mix are rarelyevaluated adequately. The impact of releasing professionals’ time has not been estimatedand therefore determining whether changes are cost-eCective is diBcult; these diBcultieshave often been increased by poor study design.Conclusions Economic evaluation has been under-utilized in studies of skill mix. Ifeconomic evaluation demonstrates that skill mix changes reduce cost and improve ormaintain patient outcomes, this is strong evidence that these changes should beimplemented. Incentives may be required to attract the necessary personnel. This in itselfmay influence the cost of changing the skill mix and therefore the situation should bemonitored as both costs and eCectiveness can alter over time.

Accepted for publication: 13 May 1999

Introduction eBciency should probably be implemented. This paperhighlights some of the practical problems involved in

The NHS employed almost 550 000 whole-time equivalentperforming economic evaluations of skill mix alterations

direct care staC in England alone in 1996 (Governmentand of implementing cost-eCective changes in the employ-

Statistical Service 1997) costing an estimated £11 billion ment and deployment of nursing staC.in 1998 (based on figures from DoH 1998). Medical staCin primary and secondary care account for approximately78 600 of this total, while the number of nurses in these

Nursing and economic evaluationenvironments is over 340 000Despite the large numbers of individuals providing It has been demonstrated that nurses and other health

direct care and the associated burden on NHS resources, professionals can substitute for doctors without detrimen-and the fact that there are a large number of studies tal eCect on patient outcomes in a variety of settings suchdemonstrating the eCectiveness (or consequences) of as radiology (Thompson 1974) paediatrics (Honigfeldchanges in skill mix, economic evaluations of changes in et al. 1990), obstetrics (McKee 1992), intensive carethe skill mix remain relatively few in number. Economic (Dubaybo et al. 1991), neurology (Gunderson et al. 1984),evaluation is concerned with the costs and consequences orthopaedics (Harris & McCollister-Evarts 1990), mid-

wifery (MacDorman & Singh 1998) and anaesthesiaof an intervention; changes that demonstrate increased

265© 1999 Blackwell Science Ltd

Page 2: Identifying, evaluating and implementing cost-effective skill mix

G. Richardson

(SchaCner et al. 1995). In addition, other studies have cost per death avoided or setting out the increased costagainst the improved outcome (lower mortality). It isdemonstrated that other professionals can perform specific

tasks (that have been the domain of doctors) within then a matter for decision makers to conclude whethersuch substitution is the best use of scarce resources; thatspecialties including endoscopy (Lieberman & Ghormley

1992), coronary arteriography (Demots et al. 1987) and is whether it is worth spending £x to gain an extra ylife years.sigmoidoscopy (Rosevelt & Frankl 1984). A variety of

professionals were analysed in these studies including For changes in the skill mix to be described as cost-eCective it is necessary to consider both the costs andphysician assistants, midwives, nurses, nurse clinicians

and nurse practitioners. consequences of such changes.The other forms of evaluation do not demonstrateThese studies were solely addressing the issue of the

eCectiveness of substituting health professionals. equivalence of outcomes nor do they only consider asingle eCect (such as the change in mortality or life yearsEconomic evaluation is concerned with the costs and

consequences (eCectiveness) of a given activity. It is gained). Cost benefit and cost utility analyses compareinterventions where more than one eCect occurs in bothrequired because resources are scarce and choices must

be made between competing alternatives (Drummond of the comparisons or where programmes with diCerentoutcomes of interest are being compared. Consider foret al. 1987). There are four types of economic evaluation

namely cost minimization, cost-eCectiveness, cost-benefit example that we wish to compare two programs, onewhich looks at the eCectiveness of district nurse visits inand cost-utility. The simplest of these is cost minimization

where the consequences of competing alternatives are reducing complications post-discharge, the other at apractice nurse carrying out injections to immunize againstidentical in all relevant respects. For example, if it were

demonstrated that coronary arteriography could be per- influenza. Clearly the outcomes of interest are diCerent,and cannot be summarized in one measure; therefore aformed by nurses with the same rates of myocardial

infarctions, ischemic attacks and other complications straightforward cost-eCectiveness comparison is inappro-priate. The solution is to use either cost-benefit or cost-(such as emboli and infected punctures) as when per-

formed by doctors, then the most eBcient use of scarce utility analysis. Cost-benefit analysis attempts to place amonetary value on the benefits of a programme, whereasresources would be the least cost alternative. It should be

noted however, that the relevant comparison is the cost cost-utility analysis attempts to value the worth of aspecific health state often using a measure comprising aper task not the cost of employing the individual for a

given period; for instance, even if one individual cost half quality of life component. There are instances where thesubstitution of a nurse for a doctor may not lead toas much to train and employ as the other it would not

be cost-eCective to use them for this task if they required improvements in health outcome, but may increase wel-fare in general, for example by increasing the uptake of10 times longer to carry out a task.

The most commonly used evaluation is cost- benefits or the level of household appliances provided.As both the costs and consequences of personnel caneCectiveness where the costs of an intervention (for

example a nurse providing the service rather than a change over time and between locations, it should berecognized that economic evaluation is a dynamic ratherdoctor) are compared with the consequences which are

measured by a single eCect such as mortality, life years than a static measure. Such changes need to be monitored.Several factors which could change the costs and/or thegained or the change in the score of some preordained

measure (such as the Barthel Index). As an example of eCectiveness of substitution, together with practical prob-lems in the evaluation and implementation of skill mixthis consider again the case where the alternatives are a

nurse performing coronary arteriography or a doctor changes are considered below.performing the same task, where the outcome we areconcerned with is mortality due to complications. If the

The relative cost of employing healthnurse performs this procedure at a lower cost per task

professionalsand also lowers the rate of mortality due to complicationsthen the case for the substitution is described as dominant The true measure of the cost of any activity is the

opportunity cost which is defined as value of the activityas both the costs and consequences favour the nurse. Ifthe cost per procedure is higher for the nurse to perform which needs to be given up in order to achieve something.

While in some cases the monetary cost may be a reason-the task (for example due to more time being required)but the level of mortality due to complications is lower able approximation of the opportunity cost, this is not

always the case. Consider, for example, the case where athen the additional cost of using the nurse is comparedwith the additional eCect, resulting in a ratio showing the nurse performs the task traditionally carried out by a

266 © 1999 Blackwell Science Ltd, Journal of Nursing Management, 7, 265–270

Page 3: Identifying, evaluating and implementing cost-effective skill mix

Cost-effective skill mix

doctor. If the doctor has nothing else to do with his/her inputs. If dated technology requires a more skilled andexperienced user it is unlikely that substitution could ortime, then his/her opportunity cost is zero. In such

instances, it is not cost-eCective to substitute a nurse for should take place. If newer technology makes the tasksimpler the potential for substitution is increased.a doctor as the nurse (with presumably a positive oppor-

tunity cost) is more costly. In practice the consequence ofthe above scenario is that though the same task is being

Evaluating substitutioncompleted, the employer pays both the doctor and thenurse to perform it; therefore it cannot be cost-eCective Some tasks have been shown to be eCective and cost-

eCective when performed by a nurse rather than a doctor.(assuming the outcome is the same).However, in reality doctors are unlikely to have nothing However, the cost-eCectiveness studies are in general

weak and suCer many methodological problems includingto do with their time, though their productivity may belower once they are released from certain tasks. In most small sample sizes, single site studies, lack of randomiz-

ation, inadequate measures of outcome and limitedinstances where nurses substitute for doctors, the nursesare likely to be significantly less costly to employ. In the follow-up period. For a detailed description of the studies

and their weaknesses see Richardson and Maynard (1995).primary care environment, the doctor substitute is com-monly the practice nurse. If, for instance, a general These flaws are not diBcult to address in theory, though

in practice they may prove problematic. Other, morepractitioner (GP) costs approximately three times moreto employ than a nurse practitioner (and in this instance complex issues in the design of such studies are dis-

cussed below.assuming that wage rates reflect opportunity cost), thensubstitution will be cost-eCective as long as outcomes areequivalent (or better) and the time taken to perform the

Evaluating released timetask by the nurse is not more than three times longer.Whatever the setting, the costs of employment can change In most cases, studies do not address the issue of how

released time is put to use, for example what the doctorover time, thus aCecting the cost-eCectiveness of substi-tution. However, as doctors and nurses are paid on actually does with the extra time he/she has if a nurse

performs tasks previously performed by doctors. It isdiCerent bases, and are compensated in various ways forout-of-hours or overtime activities, it is also necessary to implicitly assumed in most evaluations that a doctor will

perform tasks which are equal in value to those previouslyconsider the timing of substitution.To illustrate the point that cost diCerentials change performed. However, there are likely to be instances

where the doctor performs other tasks such as adminis-markedly over time, ScheAer et al. (1996) compared thecosts of employing physicians, physician assistants and tration which (at least in terms of remuneration rates)

are less highly valued. In these instances the cost ofnurse practitioners over the period 1975–92. In 1975, thephysician cost 3.5 times as much as the physician assistant changing the skill mix is the cost of employing the nurse

to perform the task plus the diCerence between the actualor the nurse practitioner; by 1992 the cost diCerential haddecreased to the point where the physician cost twice as value of the doctor’s time and the actual value of the task

the doctor now performs (i.e. the opportunity cost of themuch to employ, i.e. the diCerence in earnings had almosthalved over the period. Clearly such changes could impact doctor in the new task). Thus the higher the value of the

task the doctor now performs, the more likely thatupon the cost-eCectiveness of a strategy; considering againthe example of nurses performing coronary arteriography, substitution will be cost-eCective.

This problem is diBcult to solve given current trialif the outcomes were the same but the use of the nursewas slightly cheaper per procedure. If the diCerentials practice where a fixed sample of patients are evaluated.

Where substitution occurs, if the doctor performs otherbetween the costs of the doctor and that of the nurse arereduced then it is possible that the cost per procedure tasks for other patients, then throughput will be increased

(presumably benefiting other patients not in the trial) orwould be more expensive when carried out by a nurse.waiting times shortened. These benefits to other patientsneed to be incorporated into the analysis. One method of

The relative effectiveness of healthevaluating the eCects of substitution on doctors’ time

professionalswould be to perform an audit alongside the clinical trial.If the trial were performed in a primary care setting, thePatient outcomes vary between health professionals and

also over time. In addition, the technology that individuals hypothesis could be: that the practice nurse(s) treat(s) thesimple cases as eCectively as the GP(s). This releases thehave at their disposal will aCect their levels of productivity

and may also aCect the potential for substituting labour GP(s) to deal with the more complex (or more serious)

267© 1999 Blackwell Science Ltd, Journal of Nursing Management, 7, 265–270

Page 4: Identifying, evaluating and implementing cost-effective skill mix

G. Richardson

cases. Thus the aggregate outcomes of the practice costs are certain to increase. Where these costs are higher,and patient outcomes are not improved, decision makersshould increase.

This would be adequate so long as the doctor continues need to ask themselves whether they are prepared to payfor an intervention which has no impact on overall patientto treat only patients from the practice. It should be

noted, however, that such an approach would measure health outcome, but merely improves the process of care.If such interventions are implemented, it will be at theonly the actual substitution rather than the potential for

substitution. For a discussion of the potential for del- expense of other interventions which may have a positiveimpact on patient outcomes.egation within the primary care environment see Jenkins-

Clarke et al. (1997.)An additional complication for the economic evaluation

would occur if the doctor carried out activities which Implementing cost-effective substitutionmay prevent the onset of disease. In such instances where

If it is demonstrated that skill mix changes are cost-the eCect of released time may only be realized years

eCective, this provides strong evidence that such changeslater, a long-term follow-up of the ward or unit would

should be implemented in order to improve eBciency.be required. Alternatively an intermediate measure of

However, a number of factors will influence the degreeoutcome may be used as long as there is strong evidence

to which these changes in the skill mix occur in practice.of a relationship between the intermediate outcome meas-

In terms of substituting between doctors and other healthure and the final outcome measure. For instance in

professionals, these can be separated broadly into factorsevaluating the eCectiveness of nurse or GP counselling on

aCecting the demand and those aCecting the supply ofheart disease the quit rate amongst smokers may be a

health professionals.useful intermediate outcome.

Evaluation of non-health outcomes Demand side factors

The demand for other professionals to perform the tasksCost-eCectiveness analysis is the most common form ofeconomic evaluation. In many cases, the measure of traditionally carried out by doctors is constrained in the

UK by the doctors themselves. The Medical WorkforceeCectiveness is a form of health outcome such as anActivities of Daily Living (ADL) Index. However, in many Standing Advisory Committee (MWSAC) uses a medical

manpower approach to advise government on the numbercases it is likely that substituting a nurse for a doctormay result in an increase in costs due to the employment of doctors required by the nation. Though this committee

acknowledges the importance of skill mix issues (MWSACof the nurse as well as the resource consequences ofemploying the nurse. For instance, consider an inter- 1997), such factors are not incorporated into the calcu-

lation of doctor numbers. The potential for replacingvention where a specialist nurse providing a service topatients with a chronic condition may increase the amount some doctors with other health professionals (or in theory,

vice versa) is therefore not being met and the number ofof appliances or benefits received. These resource conse-quences need to be recorded during the trial and ultimately individuals performing traditionally doctor tasks is likely

to be suboptimal. For instance, if it is shown that practiceincluded in the cost of the intervention. The resourceconsequences may not necessarily be associated with an nurses can perform a particular task more cost-eCectively

than a GP, then it is feasible that the number of GPsimprovement in clinical outcome, but may increase qualityof life in general (e.g. by increasing the level of benefit ought to decrease. However, if this substitution were to

suggest a number of doctors which is less than the ‘need’received and therefore the individual’s well-being and/orsocial functioning). In these instances cost-utility analysis for GPs (as determined by the medical manpower model),

the policy currently would not be implemented.is the suggested form of evaluation.Richardson et al. (1998) state that in recent years there This situation is in contrast with that in the US where

the Council on Graduate Medical Education (COGME)has been a world-wide policy shift towards substitutingother health professionals, and in particular nurses, for and National Advisory Council on Nurse Education and

Practice (NACNEP) have developed the Integrateddoctors in certain tasks. However, the authors point outthat this so-called substitution may not be real; that Requirements Model (COGME & NACNEP 1995) to

project the numbers of primary care practitioners, nurseincreased roles for non-physician personnel may result inservice enhancement rather than substitution. In these practitioners, physicians assistants and certified nurse

midwives, that will be required under a variety of assump-situations, patient health outcomes are unlikely to increase(although patients may prefer the enhanced service), but tions including the degree to which other health

268 © 1999 Blackwell Science Ltd, Journal of Nursing Management, 7, 265–270

Page 5: Identifying, evaluating and implementing cost-effective skill mix

Cost-effective skill mix

professionals are able to substitute for physicians. This Thus the labour market cannot adjust instantaneously toshortages in a particular field. OCering higher wagesmodel could be used as the basis for the UK’s approach.would not solve the problem in the short term; costscould increase and eCectiveness would be reduced if the

Supply side factorspersonnel were insuBciently trained. Thus there wouldbe a lag between the identification and implementationThe supply of nurses and other health professionals will

also aCect whether cost-eCective substitution takes place. of a cost-eCective change in skill mix.More recently, the government has revealed plans toOne obvious factor which could aCect the number of

nurses and other health professionals is the wage rate, increase the number of doctors by 7000 over the nextthree years and the number of nurses by 15 000 over theboth within the NHS and in organizations or occupations

which are suitable for such staC. Other labour markets same period (DoH 1998). Apart from the question ofwhere these individuals are going to come from (MaynardaCect the market for nurses and other health professionals.

For nurses, two particular markets could have a detrimen- 1998), how were the figures arrived at? Certainly, theseestimates were not achieved through any research intotal eCect on the supply of NHS nurses. Firstly, many opt

to further their careers abroad. Secondly, the NHS is the most eBcient (cost-eCective) skill mix. Translating theresults of evaluations of skill mix into the numbers oflosing many trained nursing staC to the private sector

and this trend towards private nursing is increasing (Cole doctors, nurses and other health professionals required isa complex issue. However, in addition to carrying out1996). This may have an impact on the level of cost-

eCective substitution if the opportunities for such substi- large trials with careful monitoring of the costs andoutcomes over an adequate follow-up period, it may betution are lower in the private sector.

An alternative to increasing the salaries of nurses in appropriate to attempt to model these requirements as inthe USA, rather than radically alter the skill mix in anthe short term is to improve their career structure. This

could increase the supply of trainee nurses and would arbitrary manner.also have the knock-on eCect of reducing the flow oftrained nurses to the private sector. The longer termimplications are likely to be an increase in the average

Conclusionwage rate for nurses, a reduction in their cost-eCectivenessand possibly an increase in the level of substitution of Economic evaluation has been underused in skill mix

studies, despite the fact that what little evidence there isother health care staC for nurses. This hypothesis issupported by evidence from the USA (Aiken et al. 1995) suggests that substitution between health professions in a

variety of settings can be cost-eCective. More evaluationswhere it has been suggested that increasing the educationand career prospects of registered nurses is likely to are required in order to move towards an eBcient skill

mix.increase the substitution of other workers for nurses.Attracting non-practising nurses to return to the pro- If changes in skill mix are identified which are lower

in (properly measured) cost and produce equal or betterfession is another possible means of increasing the nursingsupply. As pregnancy is a common reason for leaving outcomes, there is compelling evidence that these changes

should be implemented. Maintaining current skill mixnursing (Bentham & Haynes 1990), the provision offacilities and conditions to allow nurses to work hours to patterns would be an ineBcient use of scarce resources.

However, it is recognized that there are barriers tosuit those with a young family could be considered(Bentham & Haynes 1990). A small study carried out in implementation of such changes. Perhaps the greatest

barrier to substitution lies in the omission of skill mixthe UK in 1990 suggests that creche facilities may be apotential aid to recruitment especially if used in job factors from the medical manpower estimation of doctor

requirements. While doctors may fear that their numbersadvertisements (Hurst et al. (1990).In the USA several policies have been implemented would be threatened, in the interests of attaining the most

eBcient use of scarce labour resources, skill mix must bewith the aim of attracting and retaining nurses. Thesepolicies include professionally orientated recognition and included in these estimates.

Where skill mix changes are demonstrated to be cost-compensation programmes, educational leave and tuitionreimbursement (Havens & Mills 1992). There is some eCective, incentives may be required to attract the neces-

sary personnel. However, these incentives will have anevidence to suggest that these strategies have provedsuccessful in maintaining nurse numbers (Keleher 1993). impact on the cost of nursing and other direct health care

staC and therefore on their cost-eCectiveness, which willEven where individuals can be attracted to fill positions,training individuals to perform a specific task takes time. need to be monitored.

269© 1999 Blackwell Science Ltd, Journal of Nursing Management, 7, 265–270

Page 6: Identifying, evaluating and implementing cost-effective skill mix

G. Richardson

compensation for RNs: 1990 and 1995. Nursing Economics, 10Acknowledgements(1), 15–20.

The author would like to thank those who commented Honigfeld L., PerloC J. & Barzansky B. (1990) Replacing the workof pediatric residents: strategies and issues. Pediatrics, 85 (6),on earlier drafts of this paper as well as those who969–976.contributed to the paper through various informal dis-

Hurst K., Jones R., Pulleyblank P. & Hickling D. (1990) Retainingcussions, notably Ken Wright, Sue Jenkins-Clarke, LindastaC through creche provision. Nursing Times, June 13–19, 54–55.

Davies, Stephen Palmer and Karen Spilsbury. Jenkins-Clarke S., Carr-Hill R., Dixon P. & Pringle M. (1997) Skillmix in primary care. A study of the interface between the GeneralPractitioner and other members of the Primary Health Care

References Team. Centre for Health Economics, University of York, York.Keleher K.C. (1993) Sabbatical leaves for nurse midwives in clinical

Aiken L.H., Gwyther M.E. & Friese C.R. (1995) The registered practice. Journal of Nurse Midwifery, 38 (3), 165–7.nurse workforce: infrastructure for health care reform. Statistical Lieberman D.A. & Ghormley J.M. (1992) Physicians Assistants inBulletin of the Metropolitan Insurance Company, 76 (1), 2–9. Gastroenterology: Should they perform endoscopy? American

Bentham G. & Haynes R. (1990) Attitudes to a return to nursing: Journal of Gastroenterology, 87 (8), 940–943.a survey in the Norwich Health District. International Journal of MacDorman M.F. & Singh G.K. (1998) Midwifery care, social andNursing Studies, 27 (3), 287–96. medical risk factors, and birth outcomes in the USA. Journal of

Cole A. (1996) Number crunching. Health Services Journal, Epidemiology and Community Health, 52, 310–17.October, 10. Maynard A. (1998) Debunking the magic £21bn. Health Services

Council on Graduate Medical Education and National Advisory Journal. 20th August.Council on Nurse Education and Practice (1995) Report on McKee M., Priest P., Ginzler M. & Black N. (1992) Can out-of-Primary Care Workforce Projections. US Dept of Health and hours work by junior doctors in obstetrics be reduced? BritishHuman Services, Health Resources and Services Administration, Journal of Obstetrics and Gynaecology, 99, 197–202.Bureau of Health Professions, Rockville, MD, USA. Medical Workforce Standing Advisory Committee. (1997) Planning

Demots H., Coombs B., Murphy E. & Palac R. (1987) Coronary the Medical Workforce, Third Report. Department of Health,arteriography performed by a physician assistant. The American London.Journal of Cardiology, 60, 784–787. Richardson G. & Maynard A. (1995) Fewer Doctors? More Nurses?

Department of Health (1998) Submission of evidence to pay review a Review of the Knowledge Base of Doctor-Nurse Substitution.bodies. Press Release 98/423. Centre for Health Economics Discussion Paper 135.

Department of Health (1998) Biggest Programme of renewal and Richardson G., Maynard A., Cullum N. & Kindig D. (1998) Skillmodernisation since NHS was founded. Press Release 98/294. mix changes: substitution or service development? Health Policy,

Drummond M., Stoddart G. & Torrance G. (1987) Methods of 45, 119–132.economic evaluation of health care programmes. Oxford Medical Rosevelt J. & Frankl H. (1984) Colorectal Cancer Screening byPublications, Oxford. Nurse Practitioner Using 60-cm Flexible fiberoptic Sigmoidoscope.

Dubaybo B.A., Samson M.K. & Carlson R.W. (1991) The Role of Digestive Diseases and Sciences, 29 (2), 161–163.Physician Assistants in Critical Care Units. Chest, 99 (1), 89–91. SchaCner J.W., Ludwig-Beymer P. & Wiggins J. (1995) Utilization

Government Statistical Service (1998) Health and Personal Social of advanced practice nurses in healthcare systems and multispeci-Services for England 1997. The Stationery OBce, London. alty group practice. Journal of Nursing Administration, 25 (12),

Gunderson C.H. & Kampen D. (1984) Utilization of nurse clinicians 37–43.and physician assistants by active members and fellows of the ScheAer R.M., Waitzman N.J. & Hillman J.M. (1996) The pro-American Academy of Neurology. Neurology, 38, 156–60. ductivity of physician assistants and nurse practitioners and health

Harris C.M. & McCollister-Evarts C. (1990) The Relationship of force work policy in the era of managed health care. Journal ofPhysician Assistants to an Orthopaedic Residency Program. Allied Health, 25 (3), 207–17.Clinical Orthopaedics and Related Research, 252, 252–261. Thompson T. (1974) The evaluation of Physician’s Assistants in

radiology. Radiology, 111, 603–606.Havens D.S. & Mills M.E. (1992) Professional recognition and

270 © 1999 Blackwell Science Ltd, Journal of Nursing Management, 7, 265–270