Health Service Research Feature
CSIRO PUBLISHING
www.publish.csiro.au/journals/ahr
Australian Health Review, 2011, 35, 334–340
Effects of the Medicare enhanced primary care program on primary care physician contact in the population of older Western Australians with chronic diseases David A. J. Gibson1,4 BHlthSc(Hons), Research Associate Rachael E. Moorin1,2 PhD, MMRS, GCHlthEcon, Associate Professor, Director of Australian Centre for Economic Research on Health (UWA)
David B. Preen1 BSc(Hons), PhD, Associate Professor, Director of Centre for Health Services Research Jon D. Emery3 MBBCh, MA, FRACGP, MRCGP, DPhil, Winthrop Professor of General Practice C. D’Arcy J Holman1 MBBS, MPH, PhD, GCLaw, Professor of Public Health 1
The University of Western Australia, School of Population Health (M431), 35 Stirling Highway, Crawley, WA 9006, Australia. Email:
[email protected];
[email protected];
[email protected] 2 Curtin University, Health Innovation Research Institute, Centre for Population Health Research, GPO Box U1987, Perth, WA 6845. Australia. 3 University of Western Australia, 328 Stirling Highway, Claremont, WA 6010, Australia. Email:
[email protected] 4 Corresponding author. Email:
[email protected]
Abstract Objective. Assess uptake of Medicare’s enhanced primary care (EPC) services in Western Australia (WA) in 2001 to 2006, evaluating effect of EPC services on the regularity of contact with general practitioners (GPs) in patients aged 65+ years. Method. Whole-population cohort study using linked routinely collected health service data from State and Federal health databases. Analyses include age-standardised rate of EPC services, odds of EPC utilisation relative to other GP services using logistic regression, and total GP service regularity pre- and post-implementation of the EPC program. Results. EPC services provided to WA seniors increased 345% 2001 to 2006, comprising an increasing proportion of the total GP services (1.1 to 3.6%). Uptake of EPC services accelerated abruptly after 2004 due to greater use of ‘care plans’. EPC services were associated with a history of chronic disease, especially type 2 diabetes (OR = 1.74, 95% CI 1.66–1.82). Regularity of total GP services was improved with any EPC service exposure, with greater improvement occurring in the presence of annual EPC service exposure. Conclusions. EPC item uptake responded favourably to item changes from Medicare Australia. Prior exposure to EPC items increased the regularity of GP services, an outcome inversely associated with chronic disease progression. What is known about the topic? The Australian Federal government has invested substantial funds (over $200 million in 2007–08 alone) in the enhanced primary care (EPC) program with the intention of improving patient outcomes, specifically aged patients and those suffering from chronic disease. The EPC program uses high value Medicare items to incentivise GPs to engage in long-term planned care. However, limited research has examined the effect of this program on patient service utilisation and outcomes. What does this paper add? This paper provides a whole-population perspective on the EPC utilisation trends from 2001 to 2006 for Western Australian residents over 65 years of age. The paper then examines the likelihood of using the EPC program on the basis of calendar year, sex, age and chronic disease history. A comparison of the regularity of GP service utilisation for those exposed to the EPC program is also made. What are the implications for practitioners? The adjustments made to the EPC program in 2004, after feedback from GPs, appears to have substantially increased the utilisation of the program. Additionally, patients suffering from several different chronic diseases are more likely to be exposed to the EPC program. Exposure to the EPC program also appears to provide an improvement in regularity of service utilisation which has been associated with improved outcomes in the literature.
AHHA 2011
10.1071/AH09852
0156-5788/11/030334
Effect of EPC on GP service utilisation and regularity
Introduction Current Australian Bureau of Statistics (ABS) projections indicate by the year 2056, 23–25% of Australia’s population will be aged over 65 years, in comparison to 13% in 2007,1 thereby increasing demand on the health system.2,3 In Australia, nearly two-thirds of avoidable hospitalisations for ambulatory care sensitive conditions (ACSCs) are due to chronic illness.4–7 Thus the more effective provision of primary care is central to Australia’s national strategy for chronic disease control both now and into the future. To reduce the burden of avoidable hospitalisations from chronic diseases the enhanced primary care (EPC) program was introduced in November 1999 as a set of claim items registered in Australia’s Medicare Benefits Schedule (MBS).2 The objectives of these government-supported services were to improve the regularity and quality of healthcare provided by general practitioners (GPs) to older Australians and those with chronic diseases.2,8 In particular, the EPC items included annual health assessments for people aged >75 years (>55 years for Indigenous Australians) to assess whether preventive or educational services should be offered to the patient.9 Additionally, EPC multidisciplinary care plans and case conferences became available for patients of any age with chronic or terminal conditions.10 After their introduction, EPC services were promoted to Australian GPs with increasing intensity as a means to improve chronic disease outcomes.11 The cost of EPC items in 2007–08 of 2.064 million items nationally was $203.8 million.12 Despite its high cost, the evaluation of EPC services has been limited to measurements of GP awareness and acceptance,2,13,14 uptake by patients across population subgroups14 and program fidelity.15 No evaluation has occurred to date in terms of population-level uptake, effect on regularity as distinct from frequency of GP visits and the effects on disease progression and hospitalisations in patients with chronic diseases.12 The aim of this study was to assess the use of EPC services in the Western Australian population and to evaluate the effects of EPC services on the regularity of GP visits in patients with chronic diseases. Methods Data sources and study population The study population was defined as individuals aged 65+ years who had been registered continuously in WA by Medicare since its inception in 1984.16 For these individuals de-identified records were extracted and linked at the individual level for: (i) MBS records originating in WA from 2001 to 2006; (ii) WA hospital morbidity data system (HMDS) records from 1999 to 2006; (iii) WA mortality records from 2001 to 2006; and (iv) WA Electoral Roll records from 2001 to 2006. Identification of EPC items and GP consultation records MBS records not relating to EPC or other GP claims were removed from the dataset. EPC or other GP claim records were identified using the MBS item numbers taken from published schedules relevant to each year of study. EPC items were defined as any service listed under Groups A14 and A15 of the relevant MBS schedule.17 GP claims were defined as any item listed in the
Australian Health Review
335
explanatory notes of the relevant MBS as ‘Attendances by General Practitioners’.17 Determination of recent chronic disease history Hospitalisation within the previous 2 years for seven ACSCs – type 2 diabetes, asthma and chronic obstructive pulmonary disease (COPD), ischaemic heart disease (IHD), seizures, dyspepsia, hypertension and heart failure18 – was assessed using the primary and secondary diagnosis fields of HMDS records for each individual. A set of International Classification of Diseases 10th Revision (ICD-10) codes for each disease was devised and independently validated by a clinical consensus panel comprising nine GPs, two medical specialists and two clinical pharmacists. Calculation of primary medical care regularity For each patient in each year, the MBS data were used to calculate the number of days from the start of the year to the first primary care visit (whether EPC or other GP service); between each subsequent pair of primary care visits during the year; and from the last primary care visit to the end of the year. After combining the lead in time to the first visit and the intervals between each subsequent visit, the results provided a distribution of days between visits for each patient in each year. The variance of these distributions were computed and the following equation was used to derive a measure of annual primary medical care regularity: R¼
1 ð1 þ VarðFÞÞ
where R was the regularity and Var(F) was the variance of the days between visits. Thus, R had a range of 0 to 1 with a score of 1 indicating perfect regularity. This method of calculating service regularity has been reported previously.19,20 Definition of EPC exposure status The EPC service exposure of patients was defined using two methods. First, ‘ever exposed’ was defined as exposed for the remainder of the study period after initial EPC claim. Second, ‘yearly exposed’ was defined as exposed for the remainder of the calendar year of at least one EPC claim. Evaluation of patterns of EPC utilisation Annual age-standardised EPC utilisation rates, weighted directly to the population distribution in the first year of the study, were calculated using counts of claims stratified by 5-year age group, sex and MBS item type. Population denominators were obtained from the study data. For each person contributed person-time at risk from 1 January 2001 to 31 December 2006, censored for death and adjusted for out-of-State migration using Electoral Roll information. Multivariate logistic regression modelling was employed to evaluate changes in the odds of EPC service utilisation compared with other GP services during the study period, adjusting for age, sex and recent chronic disease history. Due to the correlated data structure (repeated observations on each individual), a random effects model was employed so the odds of EPC use could be modelled explicitly for each individual, where a marginal effects
336
Australian Health Review
model would have averaged the probability of EPC use across all individuals in each stratum.21,22 Evaluation of the effect of item changes on EPC utilisation Random effects multivariate logistic regression modelling was used to evaluate changes in demographic and prior chronic disease history associated with use of EPC items compared with standard GP consultations in two distinct policy periods (eras), informed by both MBS and EPC item policy changes and the observed changes in annual age-standardised rates. Changes in the odds ratios reported for each demographic and chronic disease history factor were evaluated across the two eras, while simultaneously adjusting for all other covariates in the model. Assessment of the effects of prior EPC exposure on GP regularity To examine changes in GP service regularity associated with prior EPC service exposure, the percentage change in the mean GP service regularity for those exposed compared with those notexposed was calculated for each year under the two exposure definitions described previously and using the following formula: mðRegÞEXP mðRegÞUNEXP 100 mðRegÞUNEXP where m(Reg) was the mean regularity, EXP was the exposed population and UNEXP was the unexposed population. Results The number, proportion and cost of primary medical services are shown in Table 1. The number of EPC claims increased by 345% (n = 43 094) over the study period and comprised a growing percentage of total claims (1.1% in 2001 to 3.6% in 2006). Consequently, the proportion of total primary care costs attributable to EPC services increased from 5.0% in 2001 to 10.2% in 2006. Females consistently used more EPC and GP services, although the sex gap diminished with the proportion of services received by males, which increased from 37.7% in 2001 to 41.5% in 2006. Patients aged 75–79 years used more EPC services than other age groups. During the observation period there was an increase in the proportion of individuals exposed to EPC services with a large increase in this percentage after 2004 (9.1% average) to 2005 and 2006 (14.2 and 18.9% average respectively). A similar trend was observed for patients identified as having a recent history of ACSC hospitalisation. From 2001 to 2004 the percentage of ACSC patients with an EPC item ranged from 12.4 to 15%, whereas in 2005 and 2006 this proportion increased to 22.3 and 28.4% respectively. This trend between ACSC history and EPC exposure was consistent across all seven disease categories investigated. Utilisation of EPC 2001–2006 Fig. 1 shows that the age-standardised rate of EPC claims increased from 1685 per 10 000 person-years in 2001 to 5535 per 10 000 person-years in 2006. The trends were similar in males and females (data not shown), although the claim rates were consistently higher in women than in men. There was an increase of 159%, representing an additional 1365 EPC services, observed after 2004. This was limited to EPC care plans only, with little
D. A. J. Gibson et al.
change in the claim rates for EPC health assessments or case conferences. Table 2 shows the relative odds of using EPC services in each year of the study, adjusted for sex, age and chronic disease history. Consistent with the trend in the age-standardised claim rate, odds of EPC use plateaued between 2002 and 2004 followed by large increases in 2005 and 2006. After adjusting for other variables, only a slightly (3%) but significantly higher odds of EPC use occurred in females compared with males. As would be expected, given the prerequisites for use of EPC items, the adjusted odds of EPC use were significantly associated with a history of chronic disease, with the greatest influence being a history of type 2 diabetes (OR, 1.74; 95% CI, 1.66–1.82). The effect of item changes on the likelihood of EPC utilisation Fig. 2a shows odds ratios for EPC use by age group adjusted for sex and history of chronic disease for the two item eras investigated (2001–04 and 2005–06). In the later item era there was a moderation of the effects of age on utilisation levels, leading to a more even utilisation profile across age groups, albeit that adjusted odds ratios were still elevated by ~240–330% at age 75+ years. Little modification of the effects of most chronic diseases on the adjusted odds of EPC use existed across the two eras (Fig. 2b). The odds of EPC use in patients with a history of type 2 diabetes IHD, asthma/COPD, dyspepsia or hypertension were elevated to equivalent levels in both item eras compared with those without any chronic disease history. Small changes across the item eras did occur in the relative odds of EPC use in patients with prior histories of seizures or heart failure. Prior EPC exposure and GP service regularity Figure 3 shows the percentage increase in annual service regularity in those exposed to EPC services compared to patients with unexposed service regularity. The percentage changes are represented according to the two exposure criteria described as ‘ever exposed’ and ‘yearly exposed’ for males and females. The greatest change in regularity was observed in 2001 (304% increase in females, 346% in males). Higher regularity of GP attendance continued to be associated with having ‘ever’ used EPC services after 2001, increases of 27% in females and 36% in males in 2002 and 65% in females and 73% in males in 2006. The effect on GP service regularity with ‘yearly exposed’ was more consistent and pronounced, the improvement in GP service regularity in 2002 compared with 2001 was less, decreasing to 265% (males) and 203% (females). Exposure to the EPC program in the same year afforded a greater increase in GP service regularity than ‘ever exposed’ except in 2001, where both methods produced the same result due to identical look-back periods. Both figures show that EPC use had similar effects across time in males and females, although the increase in regularity in males was consistently higher in almost all instances. Discussion This study has shown the uptake of EPC services in WA grew rapidly in 2005 and 2006 after modification to Chronic Disease
2002
2578 1677 5529 3291 4340 2011 1183 20 609
12.5 11.5 12.3 14.0 10.8 12.9 12.3 12.4
248 198 647 740 488 214 141 2 676
n at risk
3.9 60.2 64.0
451 234 751 486 563 280 179 2 944
2 003 2 338 6 047 4 148 2 174 873 17 583
n exposed to EPC
8994 13 642 22 636
2.4 59 045 2.9 51 926 13.3 40 658 14.9 25 874 13.9 15 125 10.2 8344 7.5 200 972
%
n 23 669 1 622 069 1 645 738
1385 1516 5182 3555 2092 782 14 512
n exposed to EPC
37.7 62.3 100
6370 10 537 16 907 n at risk
5.0 95.0 100
3.1 58.8 61.9
% 1.1 98.9 100
2001
17 585 1 587 606 1 605 191
Age group (years) 65–69 57 571 70–74 51 616 75–79 38 831 80–84 23 931 85–89 15 006 >90 7 694 Total 194 649 ACSC groups Type 2 diabetes 1986 Asthma and COPD 1721 Ischaemic heart disease and angina 5245 Seizures 5283 Dyspepsia 4517 Hypertension 1664 Heart failure 1149 Any ACSC history 21 565
Number of items EPC GP Total Cost (2008 AU$) (millions) EPC GP Total Sex Male Female Total
n
17.5 14.0 13.6 14.8 13.0 13.9 15.1 14.3
n
2003
n at risk
2886 1621 5617 2204 4149 2344 1159 19 980
447 209 734 318 549 302 158 2 717
1572 1869 6427 4292 2194 991 17 345
n exposed to EPC
8888 13 459 22 347
3.8 61.2 65.0
23 308 1 649 653 1 672 961
3.4 60 848 4.5 51 969 14.9 42 374 16.0 27 680 14.4 15 058 10.5 8929 8.7 206 858
%
39.7 60.3 100
6.0 94.0 100
1.4 98.6 100
%
15.5 12.9 13.1 14.4 13.2 12.9 13.6 13.6
n
2004
n at risk
3206 1624 5962 1637 3990 2548 1123 20 090
594 225 873 255 567 346 147 3 007
1896 2181 7043 4742 2299 1200 19 361
n exposed to EPC
9845 15 078 24 923
4.3 68.6 72.9
26 056 1 722 205 1 748 261
2.6 62 481 3.6 52 278 15.2 43 814 15.5 29 420 14.6 15 029 11.1 9635 8.4 212 657
%
39.8 60.2 100
5.8 94.2 100
1.4 98.6 100
%
18.5 13.9 14.6 15.6 14.2 13.6 13.1 15.0
n
2005
n at risk
3590 1611 6010 1221 3843 2626 1044 19 945
960 356 1 274 268 787 576 235 4 456
3666 4803 9648 6297 3196 1721 29 331
n exposed to EPC
15 888 23 165 39 053
5.8 62.8 68.6
40 916 1 704 962 1 745 878
3.0 51 618 4.2 52 549 16.1 45 527 16.1 30 316 15.3 16 165 12.5 10 269 9.1 206 444
%
39.5 60.5 100
5.9 94.1 100
1.5 98.5 100
%
Table 1. Number and percentage of primary medical services, cost and patient demographics 2001 to 2006 EPC, enhanced primary care; ACSCs, ambulatory care sensitive conditions; COPD, chronic obstructive pulmonary disease
26.7 22.1 21.2 21.9 20.5 21.9 22.5 22.3
n
2006
n at risk
1570 673 2450 502 1493 1098 439 8 225
511 176 702 154 392 287 113 2 335
4131 6990 11 899 8185 4117 2261 37 583
n exposed to EPC
23 994 33 833 57 827
6.8 59.3 66.1
60 679 1 626 247 1 686 926
7.1 38 999 9.1 54 072 21.2 46 060 20.8 32 012 19.8 17 063 16.8 10 747 14.2 198 953
%
40.7 59.3 100
8.4 91.6 100
2.3 97.7 100
%
32.5 26.2 28.7 30.7 26.3 26.1 25.7 28.4
10.6 12.9 25.8 25.6 24.1 21.0 18.9
%
41.5 58.5 100
10.2 89.8 100
3.6 96.4 100
%
Effect of EPC on GP service utilisation and regularity Australian Health Review 337
D. A. J. Gibson et al.
2005–2006
5 4 3 2 65–69 years
1 +
9
90
–8 85
–7
–7 70
4
0
2000
–8
3000
2001–2004
6
80
4000
7 (a)
75
5000
All EPC Health assessments Care plans Case conferences
9
6000
4
Adjusted odds ratio of EPC use
Australian Health Review
Variable Year 2001 2002 2003 2004 2005 2006 Sex Male Female Age 65–70 years 70–74 years 75–79 years 80–84 years 85–89 years 90+ years Chronic disease history No CD history Type 2 diabetes Asthma and COPD Ischaemic heart disease Seizures Dyspepsia Hypertension Heart failure
OR
– 1.17 1.10 1.19 1.93 2.70
– 1.22 1.15 1.24 2.01 2.82
1.00 1.03
– 1.02
– 1.05
1.00 1.51 4.45 4.41 3.96 2.99
– 1.48 4.37 4.32 3.87 2.89
– 1.54 4.53 4.50 4.06 3.08
1.00 1.74 1.29 1.40 1.22 1.27 1.29 1.17
– 1.66 1.20 1.35 1.15 1.22 1.23 1.09
– 1.82 1.39 1.45 1.29 1.32 1.36 1.25
2005–2006
Is
an
d
C ch O a PD di em se ic h as e e ar t Se iz ur es D ys pe ps H ia yp er te ns io H n ea rt fa ilu re
et
es
No chronic disease history
ab
95% confidence interval
1.00 1.20 1.12 1.21 1.97 2.76
a
Table 2. Odds ratios of enhanced primary care (EPC) use by sex, age, year and chronic disease (CD) history adjusted for all other variables OR, odds ratio; COPD, chronic obstructive pulmonary disease
di
Fig. 1. Annual age standardised enhanced primary care (EPC) claim rates in Western Australia 2001–06 for all EPC items and each item group (Health Assessments, Care Plans and Case Conferences).
m
2006
th
2005
Year
2001–2004
As
2004
2003
2
2002
2.0 (b) 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0
pe
2001
Ty
0
Adjusted odds ratio of EPC use
Age group years 1000
ACSCD Fig. 2. Adjusted odds ratio of enhanced primary care (EPC) use across two policy eras (2001–04 and 2005–06) by age group (a) and chronic disease history (b). Model (a) was adjusted for sex, chronic disease history and year of service utilisation. Model (b) was adjusted for sex, age and year of service utilisation.
400
Management (CDM) items. This growth followed a review undertaken by Medicare Australia12 and was largely limited to EPC care plans where GP administrative burden was reduced while item values were also reduced (item 720 total fee value
Percentage change in regularity
Age standardised rate (per 10 000 person years)
338
350
Male – ever exposed
Female – ever exposed
Male – yearly exposed
Female – yearly exposed
300 250 200 150 100 50 0 2001
2002
2003
2004
2005
2006
Year Fig. 3. Annual primary care regularity change between patients with no prior enhanced primary care (EPC) exposure and EPC exposure by the two criteria: ‘ever exposed’ and ‘yearly exposed’ conditions for males and females from 2001 to 2006.
$206.75 in the 2004 MBS whereas the new item 721 in the 2005 MBS was $122.40).17 To improve the uptake of the EPC items and reduce bureaucratic demands, Medicare Australia reviewed the care plan items in consultation with GP groups.12
Effect of EPC on GP service utilisation and regularity
Consequently, the resulting new item numbers and revised fee for EPC care plans in 2005 and 2006 may have provided a strong impetus for supplier-induced utilisation of these services despite a reduction in item value. Our study found the changes made by Medicare Australia to EPC care plan items resulted in relatively more EPC services being delivered to patients aged
