Methods
Article
Factors associated with bulk billing: experience from a general practitioners’ survey in New South Wales Asaduzzaman Khan, Rafat Hussain, David Plummer and Victor Minichiello School of Health, University of New England, New South Wales
Abstract Objective: To assess whether some demographic and practice characteristics of general practitioners (GPs) are associated with the use of bulk billing.
B
ulk billing for consultations by general practitioners (GPs) is one of the central features of Australia’s universal health care system. Under bulkbilling arrangements, people can see their doctors when they need without any direct cost since their doctors accept the Medicare rebate as full payment for their consultations. But for privately billed services, patients are required to pay the difference between the schedule fee and the Medicare rebate, known as co-payment or out-of-pocket expense. Recent evidence confirms that bulk billing has declined throughout Australia. Over the past three years the percentage of consultations bulk billed by GPs has steadily declined from 77.8% in the September quarter 2000 to 66.7% in the September quarter 2003.1 With the decline in bulk billing there has been a concomitant increase in co-payments for GP consultations. Recent Medicare figures indicate that the average co-payments by patients for all GP consultations has increased from $2.26 in the September quarter 2000 to $4.34 in the September quarter 2003, rising by 92% in three years.2 The decline in bulk billing and the rise in copayments for all GP consultations over the past three years is illustrated in Figure 1. There has been a growing emphasis by the Commonwealth Government on increasing patient contributions towards medical care. However, contributing even a modest amount towards healthcare costs on a regular basis can be unaffordable for some.3 In addition, paying fees upfront for consultations that are not bulk billed is considered a potential bar rier to access to Medicare services. 4
Therefore, medical services provided by the GPs who do not bulk bill are becoming less affordable, particularly for economically disadvantaged patients. Universal access to health care is being compromised as a result. Recent research also shows that higher out-of-pocket expenses are associated with lower use of general practice services.5 If people face out-of-pocket expenses that they cannot afford, they risk not accessing basic primary and/or preventive health care. Thus, there are concerns that the recent gains in primary health care in Australia may not be sustainable in the face of a decline in bulk billing and subsequent increases in out-ofpocket expenses.6 It is, therefore, important to gain a better understanding of bulk-billing practices of GPs. While some studies have explored linkage between bulk billing and factors such as access to health care, out-of-pocket expenses, and frequency of medical consultations,4,7-9 few studies in Australia to date have examined whether particular demographic and practice characteristics of GPs are associated with the use of bulk billing in general practice. The aim of this study was to assess factors associated with bulk billing by GPs in New South Wales (NSW), Australia.
Methods: A cross-sectional postal survey was conducted in late 2002 with a 15% stratified random sample, based on sex and area of practice, of currently practising GPs in New South Wales. Multinomial logistic regression was used to look at GPs’ characteristics associated with their selfreported use of bulk billing.
Results: Of the 494 GPs who participated in the study, 44% bulk billed for all patient consultations, 34% for selective patients, while 22% did not bulk bill for any patient. Multivariate analysis revealed that GPs practising in metropolitan areas were six times more likely to bulk bill for all patients compared with GPs in rural areas (OR 6.7, 95% CI 3.8-11.9). Overseas-trained GPs were twice as likely to bulk bill for all patients compared with locally trained GPs (OR 2.3, 95% CI 1.2-4.3). The likelihood of bulk billing for all patients also increased with an increase in GPs’ caseload.
Conclusions: This paper discusses some of the policy and programmatic implications of the changing pattern of bulk billing. Special efforts are needed to provide increased practice support for GPs in rural and remote areas in order to ensure affordable and accessible GP services. (Aust N Z J Public Health 2004; 28: 135-9)
Methods The data for this study were extracted from a cross-sectional postal survey on management of sexually transmitted infections (STIs) by GPs practising in metropolitan and rural areas of NSW in 2002. The study
Correspondence to: Mr Asaduzzaman Khan, School of Health, University of New England, Armidale, NSW 2351. Fax: (02) 6773 3666; e-mail:
[email protected]
Submitted: October 2003 Revision requested: December 2003 Accepted: February 2004
2004 VOL. 28 NO. 2
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
135
Khan et al.
Article
sample was drawn from a database of medical practitioners maintained by the Australian Medical Publishing Company (AMPCo), a commercial subsidiary of the Australian Medical Association. The rationale for using a commercial organisation such as AMPCo was that the records maintained by AMPCo were updated almost on a daily basis from a range of sources, and thus were less likely to lead to missing responses. At the time of sample selection (April 2002) there were 6,800 GPs in NSW listed on the AMPCo database. A 15% stratified random sample, based on sex and area of practice (metropolitan versus rural), of GPs in NSW was selected for the study that yielded a total sample of 1,020 GPs. An additional 2% stratified random sample of GPs was used for fieldtesting the draft versions of questionnaires through a pilot study. Ethical approval for the study was granted by the Human Research Ethics Committee of the University of New England. A nine-page self-administered questionnaire along with a covering letter, a self-addressed reply paid envelope and information sheet were mailed to 1,020 GPs across NSW in October 2002. A total of 120 GPs were excluded from the survey because they were either no longer in general practice (n=74), not contactable on available mailing address (n=23), had retired (n=12), were on extended leave (n=8), or had moved to another State or overseas (n=3). Therefore, the final sample size for the study comprised 900 GPs practising in NSW. After the initial mail-out, two reminder letters were sent at an interval of three weeks each. Along with the second reminder letter, a brief one-page questionnaire, titled non-response questionnaire, was sent to all non-responding GPs. The purpose of the non-response questionnaire was to explore possible reasons for non-participation by GPs in the main survey. In both the questionnaires, STI and non-response, a common question was asked about usual practice of bulk billing by GPs. A total of 409 GPs returned the completed STI questionnaire and 116 returned the completed non-response questionnaire.
Statistical analysis The outcome measure of interest for the present paper was use of bulk billing by GPs. The GPs were asked about their usual practice of bulk billing with three possible answers: ‘No’, ‘Yes’, and ‘Yes, but for certain patients only’. Eight explanatory variables, available in both STI and non-response questionnaires, were initially considered in the present analysis: GPs’ age, sex, place of graduate training in medicine, number of years practising medicine, area of practice (metropolitan versus rural), type of practice, number of patients seen per week (caseload), and hours worked per week. Of the 525 GPs who returned the completed questionnaires, either STI or non-response, 31 had missing responses to the question on bulk billing and, as such, were excluded from the analysis sample. Therefore, the present paper is based on responses obtained from 494 GPs currently practising in NSW. Bivariate analysis was carried out by using χ2 tests to assess crude associations between the outcome measure and the explanatory variables. Only variables found to be associated with bulk 136
Figure 1: GPs’ bulk billing (%) and patients’ co-payments ($) for all GP consultations, September 2000 to September 2003.
(%)
80
5
78
4 .5
76
4
74
3 .5
72
3
70
2 .5
68
2
66
($)
1 .5
64
GPs' Bulk billing(% )
1
Patients' co-payme nts($)
62
0 .5
60
0 S e p- 0 0
D e c- 0 0
M a r-0 1
Ju n -0 1
S e p- 0 1
D e c- 0 1
M a r-0 2
Ju n -0 2
S e p- 0 2
D e c- 0 2
M a r-0 3
Ju n -0 3
S e p- 0 3
Source: Commonwealth Department of Health and Ageing Source: Commonwealth Department of Health and Aging 2003a, 2003b 1,2
billing at p≤0.10 at the bivariate level were considered for multivariate analysis. Before conducting multivariate analysis, the data were examined for possible collinearity between the explanatory variables. As the outcome variable had three response categories (bulk billing by GPs: ‘No’, ‘Selective’ or ‘All’ patients), multinomial logistic regression10 was used to identify factors that were independently associated with bulk billing by GPs after adjusting for other factors. This procedure of modelling the outcome variable with three categories involved the estimation of two equations, each with parameter estimates for every explanatory variable. These equations estimated the effect of the explanatory variables on the following: • The likelihood of bulk billing by GPs for selective patients versus the likelihood of not bulk billing. • The likelihood of bulk billing by GPs for all patients versus the likelihood of not bulk billing. The multinomial logistic regression provided estimates of odds ratios and their 95% confidence intervals (CIs) for factors associated with bulk billing by GPs after adjusting for all other variables. Statistical analysis was performed using SPSS V11.5.11
Results Study participants Half (50%) of the GPs in the analysis sample were in age group 40-55 years, while a quarter (26%) were aged over 55 years. Approximately 39% of the GPs were female. Just over a quarter (27%) of the GPs had their graduate training in medicine from overseas. Two-thirds (65%) of the GPs had practised medicine for 16 years or more. The majority (78%) of the GPs were employed in group practice and two-thirds (66%) were in metropolitan areas. Median number of patients seen per week was 120 and median number of hours worked per week was 42. Comparison of characteristics of the study sample with the GP population in NSW, provided by the General Practice Branch of the Commonwealth Department of Health and Ageing
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
2004 VOL. 28 NO. 2
Methods
Factors associated with GPs’ bulk billing
(A. Calcine, personal communication, 23 January 2004), revealed no signif icant differences for GP sex and place of graduation (see Table 1). However, study participants were significantly more likely to be younger and less likely to work in metropolitan areas.
Table 1: Comparison of study sample of GPs with NSW GPs.
Bulk billing in general practice
Age (years) 25-39 40-55 56+
23.7 50.1 26.2
20.6 46.5 32.9
0.0068
Sex Male Female
60.7 39.3
64.7 35.3
0.071
Area of practice Metropolitan Rural
65.6 34.4
77.5 22.5