eMJA     The Medical Journal of Australia

Home | Issues | eMJA shop | Classifieds | Contact | More... | Topics | Search | Login | Buy full access   

Providing costing information to general practitioners -- will this intervention change behaviour and create cost savings?

A systematic review

Justin Beilby and Chris Silagy

Electronically published Monday 28 April 1997. Please submit comments by Monday 26 May 1997.


This article has been accepted for publication in the Medical Journal of Australia after undergoing peer review (you can read the reviews). It has been revised in response to the reviewers' comments, but has not yet been edited by the MJA. Readers are invited to comment on the article. These comments may be considered by the editors and authors in preparing the article for print. Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company
Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>".

Abstract - Introduction - Method - Results - Discussion - Acknowledgements - References - Authors' details


MJA reviewers' comments - Register to be notified of new articles by e-mail - To Papers for review list - ©MJA 1997

 

Abstract

 Objective: To examine how changes in clinical behaviour and reduction in associated costs could be produced by utilising strategies that incorporate information about the costs of pharmaceuticals, pathology, diagnostic imaging and other services generated by general practitioners.
Data sources: For the years 1980 to 1996 we searched (1) computerised database (Medline, CINAHL, Health Plan, Embase) and (2) citations in review articles.
Study selection: The terms used to identify studies were family practice or physicians/family, primary health care, knowledge - attitudes - practice, comprehensive health care combined with costs for fees and charges. Studies were included if they provided costing information to general practitioners with the aim to decrease costs by changing behaviour and utilised a randomised control design, a quasi-randomised controlled trial, crossover designs or controlled time series.
Data extraction: The information extracted included study design, intervention used and objective measures of general practitioner performance/clinical care, which included test ordering, prescribing, hospital and primary care visits and costs.
Data synthesis: The inclusion criteria were met by six studies. The most successful strategies were academic detailing and computerised feedback.
Conclusion: The review has indicated that changes in behaviour among general practitioners using the provision of costing information is possible in all service areas. The sustainability of these charges has not been well studied.
(MJA 1997)

 

Introduction

 In the period from 1986-87 to 1991-92 there has been an 18.3% increase in the number of general practitioners in Australia with a resultant increase in the benefits paid annually for GP consultations from about $1 billion to $1.8 billion.1 There is some evidence that at the same time GP "driven" services have also continued to increase. In the area of pathology and diagnostic imaging ordering the number of tests per capita has risen from 1.7 to 2.3 and 0.4 to 0.5 respectively from 1987-88 to 1992-93.2 This growth has contributed to a 26.5% increase in pathology costs and 28.2% increase in the diagnostic imaging costs over the same period.1 In parallel, GP attendances have increased from 4.4 to 5.3 (per capita). The reasons for these increases are unclear3 but some consideration should be given to the possibility that inappropriate ordering is occurring in parallel to the increased attendances. In the area of prescribing, there is evidence of both increasing costs and substantial variability in behaviour. From 1991-92 to 1992-93, community pharmaceutical costs increased by 23.3%, due mainly to the increased use of expensive medications.1 A recent study has found wide variations in GP prescribing behaviour both overall and for specific medications.4 For example, for people over 65 the prescribing rates for original scripts for non-steroidal anti-inflammatory medication varied from 2.6 to 4.3 per 100 Medicare consultations per full time GP.4 This variability would suggest that inappropriate and inefficient prescribing may be occurring among GPs.

If we require improved efficiency (i.e., the maximisation of health care benefits from the limited health care budget) from general practice, then it is important we examine methods to change the behaviour of general practitioners, particularly if inappropriate behaviour is occurring. In order to facilitate this debate, we systematically searched the literature to identify studies that utilised educational models that incorporated information about costs generated by general practitioners for pathology, diagnostic imaging, pharmaceuticals and other services. Other more general reviews have examined education interventions overall.5,6 There is evidence that primary care physicians know little of the costs of the clinical decisions in the area of prescribing7-11 and the diagnostic tests.12

An important concern around this issue of providing costing information is whether GPs should consider economic decisions within the GP-patient relationship.13 Mooney argues that an efficient health service will require efficient doctor,14 (in this case GPs) and Toon has argued that GPs are already making decisions based on economic considerations.13 As a group they are most aware of available options for their patients and how to choose between them.15 It seems illogical to argue that GPs can be immune from consideration of costs during the consultations.  

Method

  

Criteria for Systematic Review

 For the years 1980 to 1996 we searched (1) computerised databases (Medline, CINAHL, Health Plan, EMBASE) and (2) citations in review articles. The terms used for the computer searches were:- family practice or physicians/family, primary health care, knowledge - attitudes - practice, comprehensive health care combined with cost and/or fees and charges. The terms family practice or physicians/family were used to cover general practitioners in all countries. As this strategy retrieved large number of citations with low precision, it was necessary to manually filter the abstracts. Subsequent analysis of the relevant citations revealed no standard pattern of indexing.  

Study Criteria

 We included any study where costing information had been distributed to general practitioners with the aim to decrease costs by changing physician behaviour. This may have occurred either as a stand alone or as part of a multifaceted strategy. We defined multifaceted as any intervention that included any three of the following: use of opinion leaders, educational material, audit and feedback, guidelines, reminders, actual budgetary management, marketing, individual counselling, patient mediated interventions and providing information on cost effectiveness and clinical effectiveness. The outcomes included any objective measure of health provider performance/clinical care, including test ordering, admissions, consultations, prescribing, hospital costs or patient specific criteria such as quality of life and consumer satisfaction.

We chose to include both randomised control trials which randomised by patient, individual physician and by group and other robust designs, including quasi-randomised controlled trials (such as those using alternative allocation) crossover designs, and controlled time series.  

Objective

 1. The primary hypothesis to be tested is that providing costing information to general practitioners results in a change in their behaviour to produce savings.

2. The secondary hypotheses are: (a) Provision of costing information alone is less effective than when combined with other interventions including peer review, computer generated information, case audit and feedback, individual counselling, use of guidelines, costing, local opinion leaders, academic detailing and other specific interventions.

(b) Provision of costing information as part of a multifaceted interventions is more effective in changing behaviour than providing costing information alone.

Results have been qualitatively synthesised. No attempt has been made to quantitatively combine results of individual studies due to their heterogeneous nature, e.g., the range of providers targeted, multiple interventions utilised and differences in outcome measures chosen.  

Results

 In total six studies met the criteria in the review.16-21 Two studies were aimed at changing behaviour to reduce pharmaceutical costs, three focused on pathology and diagnostic imaging charges, one concentrated on hospital and referral services. The number of participants for each study varied from 2619 to 435.16 The participation rates were over 90% in all studies. The studies are summarised in Table 1.
 

Effects on Pharmaceutical Prescribing

 The two studies that examined pharmaceutical costs used randomised designs and provided cost information using computerised feedback. Positive changes (i.e., reduction in prescribing and/or associated costs) occurred in both studies. Gehlback et al.17 increased generic prescribing by 67% (also by using computerised feedback). In the second study which utilised the most rigorous methodology, Soumerai found that an "academic detailing" model resulted in a mean decrease of US$105 per physician for the nine month study period for three "intervention" drugs and a 13% decrease in prescribing of these three drugs.16  

Effects on Pathology and Diagnostic Imaging Ordering

 Three studies used a mixture of interventions to significantly change pathology and diagnostic imaging ordering behaviour. These intervention included immediate computer screen costing information feedback,18 education for both physician and patient,19 and an educational approach using case studies with costs attached.20 Actual assessment of behaviour change and resultant cost savings is complicated by the different outcomes chosen, but in all three studies there was a significant reduction in pathology and diagnostic imaging costs. Tierney et al., using immediate computerised feedback,18 found that the intervention group ordered 14.3% fewer tests and there was a US$6.68 lower patient charge. Senf and Jogerts found that after educating physician and/or patients about the costs of ordered tests only the physician and patient educated group had significantly lower charges for laboratory costs.19 The third study attached costs for diagnostic tests to a series of case studies and evaluated the tests ordered. The primary care physicians in the intervention group reportedly ordered on average a 31% reduction in the cumulative test costs when compared with the control group.20  

Effects on Referrals, Practical Procedures and Hospitalisation

 In an innovative study21 examining the role of the primary care gatekeeper in which patients were randomised to intervention ("gatekeeper" group) and control ("non-gatekeeper" group), the patients in the intervention group had 6% more visits to the primary care physicians, 9% less visits to specialists, 4% fewer procedures and the ambulatory charges per enrollee were US$21 less than the controls. This latter finding was not significant. The costing information was supplied as part of a risk sharing account (i.e., a specified amount of funds was established for each gatekeeper GP based on the age and sex of the enrollees. Any surplus was split between the GP and the Insurance Plan and if any deficit occurred the GP was required to pay back 10% of the "over spend" from his/her fee for service charges).  

Sustainability

 While the immediate impact of the intervention is important in evaluating its usefulness, the sustainability of any intervention is crucial to retaining any behaviour change and resultant savings over the long term. Of the six studies, four assessed the sustainability of the interventions. The post intervention assessment period varied from 1 month to 12 months. Only Soumerai 16 and Gehlback 14 found significant changes in the intervention groups 9 months or longer. The first used individual feedback by respected peers and the latter used regular computerised feedback on generic prescribing.  

Discussion

 Drawing conclusions from this review is complicated by the range of interventions used. We also note that the most recent study was published in 1990, but believe our search strategy was thorough and robust and has located all relevant articles. While the systematic review by Oxman5 looked at all interventions to change medical professionals' behaviour, our study focused on GPs and the effect of feedback on costs only. As such, we believe the review provides pertinent and specific information on successful strategies that could be utilised to improve the efficiency of Australian GPs. The caveat to this statement is that the local contextual issues in which general practice within this country operates need to be considered. Determining what is feasible and acceptable and likely to have a positive effect on the quality of patient care is critical to the development of new policy initiatives.

Pharmaceutical costs contribute 9.6% to the health care budget with about half being subsidised by the Commonwealth. It is worth noting that although the principal cause of the growth of pharmaceutical costs is the use of more expensive medications, increased prescribing has also contributed. Prescribing by specialists also contributed to these costs. However, with the variability known to exist among GPs4 more appropriate prescribing may reduce both the use of expensive medications and the number of prescriptions ordered. The academic detailing model as illustrated by Soumerai16 has been tried in Australian general practice and found to be successful in changing behaviour 22 and health outcomes. 23 Chapman et al. 24 have shown among UK GPs a six month campaign reduced costs by 200 000 British pounds on targeted drugs, and in Australia, DATIS (Drug and Therapeutic Information Service) documented a 27% reduction in non-steroidal anti-inflammatory drugs (NSAIDS) use in the targeted area.23 What would be an ideal continuation of this model is an intervention based around pharmaceutical costs in a significant number of general practices in a local region or through a Division. This review has also highlighted that computerised feedback25 will decrease pharmaceutical costs. Evidence exists in the United Kingdom that changes in GP prescribing can produce substantial savings26,27 and there is no reason why Australian GP prescribing variability can not be reduced to create pharmaceutical savings possibly using either of these two models.

With pathology and diagnostic imaging services the review identified three strategies. Immediate computerised feedback 18 may be worth exploring but will require better software than is currently available in Australia. Education was only useful when both physician and patient were given information about costs,19 suggesting that providing cost information alone to GPs is unlikely to be successful.

It is likely that consideration of improvements of efficiency in general practice will need to embrace all these areas. What may be worth considering in general practice is a multifaceted model that is first aimed at prescribing practice and then moves on to embrace all services generated by general practitioners. There is evidence that among other medical professionals multifaceted campaigns are successful in changing behaviour and creating savings.28 Modifying prescribing behaviour among Australian GPs to create savings seems to be more acceptable, whereas the costs generated from pathology and diagnostic imaging are less clearly understood perhaps because of the complexity of the Medical Benefits Scheme and the lack of training provided in this area. However, creating more appropriate ordering with the latter two groups is important as they constitute 26% of Medical Benefits payments, i.e., 4.7% of the health care budget.1

There is an important proviso to adopting this type of model. Increases in prescribing costs are related to both the use of more expensive medications and higher prescribing rates, whereas changing behaviour in the area of diagnostic and pathology ordering requires more consideration of appropriateness of ordering. These differences may require the use of different methods within this multifaceted model to change behaviour, e.g., academic detailing for prescribing16 and computerised feedback for pathology and diagnostic imaging. 18

This review has concentrated on interventions based around costing information. Other studies have found that feedback has reduced the number and quality of tests without including costing information. 29,30 These models deserve to be trialed in Australian general practice. In both these studies feedback on the appropriateness of pathology and test ordering was provided in biannual reports. The feedback was provided by a "respected expert (internal medicine specialist)" and was focused on a GP's specific patients. Appropriateness of test ordering was determined after comparison with "accepted guidelines and standards of the Dutch College of General Practitioner".29

It is equally important in any intervention to change behaviour that health outcomes are evaluated and not compromised. Only one of the studies we located for this review measured health outcomes. Tierney et al.18 found no differences in intervention and control groups for number of hospitalisations, emergency room visits and outpatient visits.

This review has looked at one possible method to change behaviour to create cost savings -- an educational model aimed at GPs which incorporates information about costs. All the strategies described are labour intensive and their utilisation may not be cost effective. It may not be possible for all GPs to fully be aware of the costs generated because of their decisions. The use of guidelines particularly for pathology and diagnostic imaging ordering31 that have been developed by respected peers and experts may be a more viable option. Alternatively, in the area of prescribing the use of formularies, 32 or peer led standards 33 or incentive payments 34 combined with information packages may be a more logical use of resources.  

Acknowledgements

 We would like to thank Phillipa Middleton from the Australasian Cochrane Centre who kindly donated her time and expertise to complete the literature searches.  

References
  1. Australian Health 1994. Australian Institute of Health. Canberra: AGPS, 1994.
  2. Doessel D. General practice in perspective: a descriptive analysis of prices and other outcomes under Medicare: a final report on the project "General Practice in Economic Perspective" Report. General Practice Evaluation Program 1995. Canberra: Department of Human Services and Health.
  3. Deeble J. Medical services through Medicare. National Health Strategy. Background Paper No 2, 1991. Treble Press.
  4. Henry D, O'Connell D. Variability in prescribing of pharmaceuticals by general practitioners and factors related to the variability. Final Report of the General Practitioners Prescribing Analysis Project (GPPAP). Canberra: Department of Human Services and Health, January 1996.
  5. Oxman A, Thomson M, Davis D, Haynes B. No magic bullets: a systematic review of 102 trials of interventions to improve professional practice. Can Med Assoc J 1995; 153: 1423-1431.
  6. Davis D, Thomson M, Oxman A, Haynes B. Evidence for the effectiveness of CME. A review of 50 randomized controlled trials. JAMA 1992; 268: 1111-1117.
  7. Walzak D, Swindell S, Bhardway A. Primary care physicians and the costs of drugs. A study of prescribing practices based on recognition and information sources. J Clin Pharmacol 1994; 34: 1159-1163.
  8. Ryan M, Yule B, Bond C, Taylor R. Scottish general practitioner attitudes and knowledge in respect of prescribing costs. BMJ 1990; 300: 1316-1318.
  9. Glickman L, Bruce E, Caro F, Avorn J. Physician knowledge of drug costs for the elderly. J Am Geriatric Soc 1994; 42: 992-996.
  10. Miller L, Blum A. Physician awareness of prescription drug cost: a missing element of drug advertising and promotion. J Fam Pract 1993; 36: 33-36.
  11. Ryan M, Yule B, Bond C, Taylor R. Knowledge of drug costs: a comparison of general practitioners in Scotland and England. Br J Gen Pract 1992; 42: 6-9.
  12. Fowles F. Doctor knowledge of the costs of medical care. Med Educ 1985; 19: 113-117.
  13. Toon P. Justice for gatekeepers. Lancet 1994; 353: 585-587.
  14. Mooney G. Efficiency in health care: just health gains? Aust J Public Health 1995;19: 330-335.
  15. Black D. Paying for health. J Med Ethics 1991; 17: 117-123.
  16. Soumerai S, Avorn J. Economic and policy analysis of university-based drug "detailing". Med Care 1986; 24: 313-331.
  17. Gehlback S, Wilkinson W, Hammond W, et al. Improving drug prescribing in a primary care practice. Med Care 1984; 22: 193-201.
  18. Tierney W, Miller M, McDonald C. The effect of test ordering on informing physicians of the charges for outpatient diagnostic tests. N Engl J Med 1990; 322: 1499-504.
  19. Senf J, Jogerst G. Cost containment: the effect of educating family practice residents and their patients. Fam Pract Res J 1987; 7(1):29-41.
  20. Cummings M, Frisoff K, Long M, Hrynkiewich G. The effects of price information on physician's test ordering behaviour. Med Care 1982; 10: 293-301.
  21. Martin D, Diehr P, Price K, Richardson W. Effect of a gatekeeper plan on health service use and charges: a randomised trial. Am J Public Health 1989; 79: 1628-1632.
  22. Peterson GM, Sugden JF. Educational program to improve the dosage prescribing of allopurinol. Med J Aust 1995; 162: 74-77.
  23. Todd C. Is there a role for academic detailing in today's health care system? In Pharma 1995; October: 7-9.
  24. Chapman SR, et al. Providing independent advice to prescribers: a multinational presentation on academic detailing. Pharmacoepidem Drug Safety 1995; 4 Suppl 1(Aug): 45.
  25. Hershey C, Goldberg H, Cohen D. The effect of compuiterised feedback coupled with a newsletter upon outpatient prescribing charges. A randomised controlled trial. Med Care 1988; 26: 98-93.
  26. Dowell J, Snedden D, Punkar J. Changing to generic formulary: how are fundholding practice reduced prescribing costs. BMJ 1995; 310: 505-508.
  27. Bradlow J, Coulter A. Effect of fundholding and indicative prescribing schemes on general practice prescribing costs. BMJ 1993; 307: 1186-1189.
  28. Manheim L, Feinglass J, Hughes R, et al. Training house officers to be cost conscious. Effects of an eduational intervention on charges and length of stay. Med Care 1990; 28(1): 29-42.
  29. Winkens R, Pop P, Bugter-Maersen A, et al. Randomised controlled trial of routine individual feedback to improve rationality and reduce numbers of tests requests. Lancet 1995; 345: 498-502.
  30. Winkens R, Pop P, Grol R, et al. Effect of feedback on test ordering behaviour of general practitioners. BMJ 1992; 304: 1093-1096.
  31. Grimshaw T, Russell T. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet 1993; 342: 1317-1322.
  32. Hill-Smith T, et al. Sharing resources to create a district-drug formulary: a countrywide controlled trial. Br J Gen Pract 1996; 46: 271-275.
  33. Bateman D, Soutter T, Eccles M, et al. Setting standards of prescribing performance in primary care: use of a consensus group of general practitioners and application of standards to practices in the north of England. Br J Gen Pract 1996; 46: 20-25.
  34. Bateman D, Campbell M, Donaldson L, et al. A prescribing incentive scheme for non-fundholding general practices: an observational study. BMJ 1996; 313: 535-538.

(Received 12 Aug 1996, accepted 20 Mar 1997)
 

Authors' details

Research and Health Promotion Unit, Royal Australian College of General Practitioners, Adelaide, SA 5069.
Justin Beilby, MPH, FRACGP, Director.
Flinders University, Adelaide, SA 5042.
Chris Silagy, FRACGP, FAFPHM, PhD, Professor of General Practice.


MJA reviewers' comments - Register to be notified of new articles by e-mail - To Papers for review list - To top of article - ©MJA 1997

<URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.