Research

Do family physicians know the costs of medical care?

Survey in British Columbia

G. Michael Allan, MD, CCFP   Grant D. Innes, MD, CCFP, FRCPC

This article has been peer reviewed. Cet article a fait l’objet d’une évaluation externe. Can Fam Physician 2004;50:263-270.

Dr Allan is an Assistant Professor in the Department of Family Medicine at the University of Alberta in Edmonton. Dr Innes is an Associate Professor of Medicine at the University of British Columbia and Chair of the Department of Emergency Medicine at St Paul’s Hospital in Vancouver, BC.

The gap between health care costs and budgetary limitations continues to widen, and many suggest that the current system is unsustainable.1 Diagnostic services and therapies account for a large portion of health care expenditures; the cost of therapies is growing fast.2 Physicians’ orders largely direct use of these health care resources, and physicians are under growing pressure to practise cost effectively.3,4 Strategies to limit use of resources by modifying physicians’ ordering behaviour often focus on family doctors, who are the primary contact for patients and comprise 50% of the physicians in Canada.5

Physicians’ ability to make cost-effective decisions is likely limited by their lack of knowledge of health care costs. Many studies in the United States and Europe show that physicians know little about these costs,6-19 but no such studies have been done in Canada. There is evidence that physicians modify their ordering behaviour and reduce costs when they have information about costs,20,21 and some Canadian researchers have begun to investigate strategies for promoting cost awareness.22 Before implementing such strategies, however, it is important to assess Canadian physicians’ knowledge of health care costs and try to determine whether they are likely to benefit from additional education.

Our primary objectives were to assess British Columbia doctors’ awareness of costs and to determine the costs of high-volume investigations and therapies. Other objectives were to assess family doctors’ attitudes to cost awareness, its importance, and education and access to information about costs. Our hypothesis was that most physicians would be unable to estimate the costs of commonly used investigations and therapies within 25% of their true value.

METHODS

Survey development and distribution

We selected a random sample of 600 general practitioners from the British Columbia Medical Association’s (BCMA) membership database. We used the Dillman total design method23 to develop and distribute a questionnaire. Part 1 of the questionnaire used a 10-point rating scale to assess physicians’ level of agreement with five statements about health care costs. Parts 2 and 3 asked physicians to estimate the costs of 23 investigations and 23 therapies frequently ordered by BC family physicians.24 Five practising family physicians with urban, rural, and academic experience had selected the 46 items by consensus.

In part 2 of the survey (investigations), physicians were asked to assume that each test was ordered independently. In part 3 (therapeutics), they were asked to estimate medication cost without dispensing fee and to assume a 10-day course for antibiotics and a 1-month course for other therapies.

The questionnaire was pilot-tested on five family physicians, modified to improve clarity, and distributed to family practice residents at the University of British Columbia.25 It was then modified slightly again and sent to the 600 family physicians. To maximize response, there were three mailings.

Cost determination

The “true” cost of radiologic studies was defined as the amount in Canadian dollars paid by the BC Medical Services Plan (MSP). These costs, negotiated by the BC Medical Association (BCMA) and MSP, include technical and professional components and are specified in the BCMA Guide to Fees.26 Because computed tomography scans are done only in hospitals, there is no agreed-upon technical fee, so the “true” cost of a CT scan of the head was determined by averaging the cost estimates of three Vancouver hospital finance departments. Estimates included capital costs; costs of clerical, technical, and maintenance staff; radiologists’ fees; and cost of materials.

True costs for laboratory tests were derived from the BCMA Guide to Fees,26 with the exception of Helicobacter pylori serology, which has no MSP billing code. The cost of this test was obtained from the BC Centre for Disease Control (CDC). Because the cost of a positive urine culture varies depending on plating and sensitivity testing, true cost was based on the cost of a negative urine culture.

Actual wholesale drug costs were obtained from MEDIS/NWD27 and Alpharma,28 pharmaceutical distributors in British Columbia. The only markup on the wholesale price is the dispensing fee; the “true” dispensing fee was defined as the average dispensing fee for the province.29 The “true” cost of six sessions of physiotherapy or massage therapy was the established MSP fee for “opted-in” therapists.

Statistical analysis

Mean scores and 95% confidence intervals (CI) were calculated for the five statements assessing physicians’ attitudes and knowledge (Table 1). Each physician’s estimate of cost was compared with the true cost, and the absolute error (true cost minus estimated cost) was calculated. As with previous studies, the proportion of estimates accurate within 25%6-11 and 50%8,12 of true cost was reported. For each item, the median error and the proportion of physicians who estimated above and below true cost were determined.

Two-sample t tests were used to determine whether rural or urban practice, sex, presence or absence of College of Family Physicians of Canada certification (CCFP), or involvement in teaching were associated with awareness of costs. One-way analysis of variance was used to determine whether time in practice was associated with awareness of costs. An interclass correlation analysis was done to determine the relationship between cost awareness and physicians’ responses to questions about costs. To assess the precision of important means, medians, and proportions, 95% CIs were calculated. Partially completed surveys were included, and missing data were reported as such. Ethics approval was obtained through the University of British Columbia.

RESULTS

A total of 283 surveys were returned, for a response rate of 47.2%. Twenty-one surveys were returned blank; in 12 cases, physicians no longer practised family medicine, and in two cases, physicians had relocated without forwarding addresses. Two surveys were illegible, and one arrived late, leaving 259 for analysis. Of these, 61 had missing data, including demographics (n = 11), attitudes to cost awareness (n = 6), and cost estimates of investigations (n = 20) and therapies (n = 45) (some had data missing in more than one category).

Table 2 shows characteristics of respondents. Table 1 shows that physicians thought costs influenced their prescribing behaviour, better knowledge of costs would change their prescribing behaviour, and they had received inadequate education on costs. Table 3 summarizes the true costs of the items surveyed.

Figures 1 and 2 show the proportion of high estimates and the percentage of estimates accurate within 25% and 50% of actual costs for investigations and therapies, respectively. Table 4 shows the median estimation error (with interquartile range) for both investigations and therapies.

In general, physicians underestimated costs of expensive therapies and overestimated costs of inexpensive therapies. There was slightly more variability and larger median errors in their estimates of the cost of therapies.

Awareness of costs did not correlate with sex, practice location, CCFP certification, faculty appointment, or years in practice. Level of agreement with statement4 (“I have adequate access to investigation/therapeutic cost information”) correlated with accuracy of cost estimate for an investigation (intraclass correlation coefficient [ICC] 0.22, P < .01) or a therapy (ICC 0.14, P < .05). Responses to other questions did not correlate with cost awareness.

Many physicians commented on their lack of awareness of costs. Three reported that they guessed at most of their responses; 15 used phrases such as “no idea,” “no clue,” “don’t know,” or “aargh.” In 17 cases, physicians responded to items with only a question mark, indicating they could not even estimate a cost. Three noted that physicians should have a better understanding of costs, two were frustrated by colleagues’ excessive test ordering, and three thought that costs should not influence ordering of investigations, which should be guided only by necessity and a test’s effectiveness. One respondent suggested that test costs should be printed with test results to raise physicians’ awareness, and 10 requested a copy of our “true” costs.

DISCUSSION

This study suggests that BC family doctors are inadequately educated about, and have limited knowledge of, the costs of medical care. Previous studies, many also using 20% to 25% as acceptable estimates,6-11,13 found that most physicians could not estimate the costs of investigations,8,10,16,18 therapies,6-9,11-14,17,18 or both.15 This study found marked variation in physicians’ estimates and a degree of error similar to that found in past studies.19,25 It also suggests that BC family physicians’ awareness of costs is only slightly better than that of family practice residents.25

Physicians seemed particularly unaware of the large differences in drug costs and tended to overestimate the cost of inexpensive drugs and underestimate the cost of expensive drugs.6,7,12,14 In this study, the cost of all 13 drugs under $40 were overestimated and the cost of six out of seven drugs over $40 were underestimated. The most expensive drug on our list, lovastatin, is 180 times more costly than the least expensive, furosemide, yet the average cost estimate was only six times higher. Physicians in this study also tended to under­-estimate the cost of laboratory tests; however, this could relate partly to the fact that we incorporated phlebotomy costs in the true cost of the test.

Predictors of cost awareness

In this study, cost awareness correlated only with a positive response to statement 4. Although that could mean that physicians with greater access to cost information use those resources and have a slightly better understanding of cost, the correlation is weak, and we are uncertain of its importance. Cost awareness did not vary with sex, practice location, CCFP certification, faculty appointment, or time in practice, or with responses to the other four statements shown in Table 1. One previous study18 found that first-year medical students estimated costs less accurately than residents and faculty, but most other studies6-10,13-16,19 found no difference in accuracy among medical students, residents, faculty physicians, non–faculty physicians, specialists, and physicians from diverse practice settings. Also, previous investigators found no significant difference in cost awareness among physicians, nurses,8 and ward clerks.16

Previous studies indicate that physicians have limited access to cost information,6,7,30 that they would like more information,6,8,14,31 and that improved knowledge changes their ordering behaviour and reduces costs.20,21 Our study suggests that BC physicians have not received adequate education about the costs of medical care, that they have limited awareness of these costs, and that they think that more knowledge would improve their ordering behaviour.

These findings have implications for program directors at medical school, residency, and post­doctoral levels. In addition to cost-awareness programs, other strategies could include educational programs,32 clinical guidelines,33 computer-based ordering,34 feedback and audits,35 physician incentives,36 and fundholding or formulary restrictions.37-40 The greatest cost saving seems to result from forced restrictions,37,38 but such strategies might have a negative effect on health outcomes, and this could lead to higher long-term costs to the system.39

Future research into cost awareness and cost-control strategies should target higher-cost investigations and therapies. Research should focus on the effect of such programs on ordering behaviour, on direct and downstream costs, and on meaningful patient outcomes.

Limitations

The main limitation of this study is its relatively low response rate, which might be related to the length of the questionnaire and to physicians’ frustration with their lack of knowledge about costs (as several respondents indicated). In a similar study involving residents,25 the most common reason for not completing the survey was frustration with lack of knowledge of costs. We recognize that our 47% response rate could have introduced sampling bias; however, the similarity of our findings to those of previous investigators’ suggests the results are valid.

Another limitation relates to the interpretation of error estimates for high- and low-cost items (eg, lovastatin estimates were considered “accurate” if they were within $28 of the true monthly cost of $113.32, but furosemide estimates had to be within $0.16 of the true monthly cost of $0.63). Clearly, errors of the latter magnitude are not important, either for a third-party payer or a patient; therefore educational programs addressing very low-cost items would be of questionable value.

Conclusion

Physicians in British Columbia have not received adequate education about medical care costs, have limited access to and awareness of these costs, and think increased knowledge of costs would improve their ordering behaviour.

Acknowledgment

We thank Dr Carol Herbert for her assistance in attaining funding for the project, Penny Miller for her assistance in determining the true cost of the medications and dispensing fee, and Jonathan Berkowitz for his statistical skills and assistance with analyzing the data. We thank the Vancouver Foundation for a $5000 grant for this study.

Contributors

Drs Allan and Innes conceived and designed the study, analyzed and interpreted the data, and prepared the paper for publication.

Competing interests

None declared

Correspondence to: Dr G.M. Allan, Department of Family Medicine, 12-103 Clinical Sciences Bldg, University of Alberta, Edmonton, AB T6G 2G3; telephone (780) 472-5054; fax (780) 472-5192; e-mail M&B_Allan@telus.net

References

1. Chwialkowski L. Health costs on steep rise: study. National Post 2001 May 9;Sect. A:1,11.
2. Canadian Institute for Health Information. Canada to spend more on health care in 1998 says Canadian Institute for Health Information. Ottawa, Ont: Canadian Institute for Health Information; 1998.
3. Sulmasy DP. Physicians, cost control, and ethics. Ann Intern Med 1992;116(11):920-6.
4. Grumback K, Bodenheimer T. Reins or fences: a physician’s view of cost containment. Health Aff (Milwood) 1990;9(4):120-6.
5. Teasdale C. New concerns over FP numbers. Fam Pract 1998;10(21):1,6.
6. Glickman L, Bruce EA, Caro FG, Avorn J. Physicians’ knowledge of drug costs for the elderly. J Am Geriatr Soc 1994;42(9):992-6.
7. Ryan M, Yule B, Bond C, Taylor R. Scottish general practitioner attitudes and knowledge in respect to prescribing costs. BMJ 1990;300:1316-8.
8. Robertson WO. Costs of diagnostic tests: estimates by health professional. Med Care 1980;18:556-9.
9. 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.
10. Long MJ, Cummings KM, Frisof KB. The role of perceived price in physician demand for diagnostic tests. Med Care 1983;21:243-50.
11. Silcock J, Ryan M, Bond CM, Taylor RJ. The cost of medicines in the United Kingdom. A survey of general practitioners’ opinions and knowledge. Pharmacoeconomics 1997;11(1):56-63.
12. Schlunzen L, Simonsen MS, Spangsberg NL, Carlsson P. Cost consciousness among anaesthetic staff. Acta Anaesthesiol Scand 1999;43(2):202-5.
13. Conti G, Dell’Utri D, Pelaia P, Rosa G, Cogliati AA, Gasparetto A. Do we know the costs of what we prescribe? A study on awareness of the cost of drugs and devices among ICU staff. Intensive Care Med 1998;24(11):1194-8.
14. Walzak D, Swindells S, Bhardwaj A. Primary care physicians and the cost of drugs: a study of prescribing practices based on recognition and information sources. J Clin Pharmacol 1994;34(12):1159-63.
15. Fowkes F. Doctor knowledge of the costs of medical care. Med Educ 1985;19:113-7.
16. Shulkin DJ. Cost estimates of diagnostic procedures. N Engl J Med 1988;319:1291.
17. Miller L, Blum A. Physician awareness of prescription drug costs: a missing element of drug advertising and promotion. J Fam Pract 1993;36:33-6.
18. Thomas DR, Davis KM. Physician awareness of cost under prospective reimbursement systems. Med Care 1987;25:181-4.
19. Kuiken T, Prather H, Bloom S. Physician awareness of rehabilitation costs. Am J Phys Med Rehabil 1996;75(6):416-21.
20. Beilby JJ, Silagy CA. Trials of providing costing information to GPs. Med J Aust 1997;167(2):89-92.
21. Hart J, Salman H, Bergman M, Neuman V, Rudniki C, Gilenberg D, et al. Do drug costs affect physicians’ prescription decisions? J Intern Med 1997;241(5):415-20.
22. Hux JE, Melady MP, DeBoer D. Confidential prescriber feedback and education to improve antibiotic use in primary care: a controlled trial. CMAJ 1999;161(4):388-92.
23. Dillman DA. Mail and telephone surveys: total design method. New York, NY: John Wiley & Sons; 1978.
24. British Columbia Ministry of Health. Pharmacare. Top 100 drugs for seniors (Pharmacare Plan A) in the first 6 months of 1997. Victoria, BC: Pharmacare; 1997.
25. Allan GM, Innes G. Family practice residents’ awareness of medical care costs in British Columbia. Fam Med 2002;34(2):106-11.
26. British Columbia Medical Association. Guide to fees. Vancouver, BC: British Columbia Medical Association; 1997.
27. MEDIS/NWD. Quoted wholesale price list. Vancouver, BC: MEDIS/NWD; 1998.
28. Alpharma. Quoted wholesale price list. Vancouver, BC: Alpharma; 1998.
29. British Columbia Pharmacy Association. Annual provincial survey. Vancouver, BC: British Columbia Pharmacy Association; 1998.
30. Anderson GM, Lexchin J. Strategies for improving prescribing practice. CMAJ 1996;154(7):1013-7.
31. Barclay LP, Hatton RC, Doering PL, Shands JW Jr. Physicians’ perceptions and knowledge of drug costs: results of a survey. Formulary 1995;30(5):268-70,272,277-9.
32. Davis DA, Thomson MA, Oxman AD, Haynes RB. Evidence for the effectiveness of CME: a review of 50 randomized controlled trials. JAMA 1992;268:1111-7.
33. Grimshaw JM, Russell IT. Effects of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet 1993;342:1317-22.
34. Sullivan F, Mitchell E. Has general practitioner computing made a difference to patient care? A systematic review of published reports. BMJ 1995;311:848-52.
35. Winkens RA, Pop P, Grol RP, Kester AD, Knottnerus JA. Effect of feedback on test ordering behaviour of general practitioners. BMJ 1992;304:1093-6.
36. Egdahl RH, Taft CH. Financial incentives to physicians. N Engl J Med 1986;315:59-61.
37. Woodward RS, Medoff G, Smith MD, Gray JI. Antibiotic cost saving from formulary restrictions and monitoring in a medical school–affiliated hospital. Am J Med 1987;83:817-23.
38. Clapham CE, Hepler CD, Reinders TP, Lehman ME, Peske L. Economic consequences of two drug-use control systems in a teaching hospital. Am J Hosp Pharm 1988;45:2329-40.
39. Soumerai SB, Ross-Degnan D, Avorn J, McLaughlin TJ, Choodnovskiy I. Effects of Medicaid drug-payment limits on admission to hospitals and nursing homes. N Engl J Med 1991;325:1072-7.
40. Maxwell M, Heaney D, Howie JGR, Noble S. General practice fundholding: observations of prescribing patterns and costs using the defined daily dose method. BMJ 1993;307:1190-5.