The relationship between outpatient department utilisation and non-hospital ambulatory care in Austria

European Journal of Public Health, Feb 2017

Background: Coordinated health service utilisation in the ambulatory care sector is of major interest from a health policy perspective. This ecological study investigates the interplay between medical care utilisation in hospital outpatient departments and in freestanding physician practices by drawing on the example of the Austrian healthcare system, which is standing out due to three features: ambulatory care is provided by both free-standing public (contract) and private (non-contract) practitioners; medical specialists operate in free-standing physician practices and in hospital outpatient departments; essentially, no gatekeeping is in place. As the ongoing health care reform aims to strengthen the primary care sector, we investigate whether in the current system care in general practitioner and specialist physician practices is in a substitutive, complementary or independent relation with medical care in outpatient departments. Methods: Hypotheses were tested using ordinary least square regression analysis based on administrative data of all Austrian districts with a hospital department in 2010, including a proxy for actual utilisation rather than physician headcount. Results: Controlling for socio-demographic and geographic characteristics and inpatient activity, we find that a higher level of care provision by contract GPs is associated with lower use of hospital outpatient departments on the district level. In contrast, a higher level of care by non-contract specialists is related to a higher utilization in outpatient departments. Conclusion: While care by non-contract specialists seems to be in a complementary and potentially demand-inducing relation with outpatient departments, primary care by contract GPs appears to be capable of replacing care in outpatient departments.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://academic.oup.com/eurpub/article-pdf/27/1/20/10301786/ckw153.pdf

The relationship between outpatient department utilisation and non-hospital ambulatory care in Austria

Background: Coordinated health service utilisation in the ambulatory care sector is of major interest from a health policy perspective. This ecological study investigates the interplay between medical care utilisation in hospital outpatient departments and in freestanding physician practices by drawing on the example of the Austrian healthcare system, which is standing out due to three features: ambulatory care is provided by both free-standing public (contract) and private (non-contract) practitioners; medical specialists operate in free-standing physician practices and in hospital outpatient departments; essentially, no gatekeeping is in place. As the ongoing health care reform aims to strengthen the primary care sector, we investigate whether in the current system care in general practitioner and specialist physician practices is in a substitutive, complementary or independent relation with medical care in outpatient departments. Methods: Hypotheses were tested using ordinary least square regression analysis based on administrative data of all Austrian districts with a hospital department in 2010, including a proxy for actual utilisation rather than physician headcount. Results: Controlling for socio-demographic and geographic characteristics and inpatient activity, we find that a higher level of care provision by contract GPs is associated with lower use of hospital outpatient departments on the district level. In contrast, a higher level of care by non-contract specialists is related to a higher utilization in outpatient departments. Conclusion: While care by non-contract specialists seems to be in a complementary and potentially demand-inducing relation with outpatient departments, primary care by contract GPs appears to be capable of replacing care in outpatient departments. Introduction Coordinated health service utilisation in the ambulatory care sector is key to both improved health care delivery and efficiency enhancement. This is especially important for healthcare systems providing specialist care not only in hospitals, but also in the non-hospital-based sector, as it is the case for example in Germany, France and Austria. If, at the same time, access to hospital outpatient departments is mostly free and without co-payments, the interplay of care utilisation in such a set-up is a question of major political interest. In two-tiered healthcare systems, where public health care is complemented with privately funded care, this issue gains even more importance. Related international literature mostly either focusses on the relationship between inpatient and outpatient care1–7 or public versus private health care utilisation.8–11 In comparison, the specific link between the utilisation of outpatient departments and ambulatory care services provided in free-standing physician practices both in the public and private sector of a healthcare system, seems less researched in the European context (details on search in Supplementary Appendix S1). This study aims to fill this gap for Austria. To this end, we refer to the Austrian healthcare system. Its organization of ambulatory healthcare stands out due to three features: First, ambulatory care (i.e. care provided outside the hospital) is provided by self-employed, free-standing physicians, of which some have a contract with Social Health Insurance (SHI) (contract physicians) and some do not (non-contract physicians). Fees charged by the latter have to be paid up front by patients and can be submitted to SHI for partial reimbursement. Second, medical specialists are located in free-standing physician practices and in hospital outpatient departments, thus providing for large parts of ambulatory secondary care a duplicated setting with different financing regimes. Third, almost no gatekeeping is in place in the ambulatory care sector. Access to hospital outpatient departments is presuppositionless and free of charge. This makes Austria a prime example for free access to ambulatory health care, providing insights into interactions elsewhere mitigated by co-payments and regulations. By international comparison, the Austrian healthcare system is exceedingly hospital-centred.12 The ongoing healthcare reform (2013–2016) thus aims at shifting medical care away from the hospital, especially to (reorganised) physician practices. In practice, however, the value of strengthening ambulatory care is highly dependent on whether these physicians can effectively substitute for the presumably more costly13 care provided in hospital outpatient departments.1 Gächter et al., using a panel dataset of Austrian districts between 2002 and 2008, explain the relationship between non-contract and contract physicians based on a Hotelling-like spatial competition model (according to which physicians compete over locations to be close to their potential patients): a higher density of non-contract specialists is related to a higher density of non-contract general practitioners (GPs).14 For non-contract specialists, a negative association with contract GPs is identified, which also applies to non-contract and contract specialists. According to the authors, this hints at competitive forces between the public and private outpatient sector. However, the relationship between outpatient departments and physicians in free-standing practices remains uncharted in their analysis. This study thus addresses the following research question: Is ambulatory care utilisation (provided by free-standing contract and non-contract GPs and specialists) and utilisation in hospital outpatient departments in a substitutive relation (i.e. utilisation of one provider reduces utilisation at the level of a different provider), a complementary relation (i.e. utilisation of one provider increases utilisation at the level of another provider) or independent from one another? By drawing on a unique administrative, district level dataset comprising utilisation information (rather than provider headcounts) from 2010 covering all national health insurance funds we are first to fully address the interdependencies of utilisation in the Austrian ambulatory care sectors, i.e. a question that is also crucial for similar healthcare systems. Setting and hypotheses The Austrian healthcare system is based on a mandatory social insurance model, covering 99.9% of the population. Affiliation with one of the 19 statutory health insurance funds is determined by the insuree’s place of residence or occupation.15 Insurees are given free choice of physicians in the ambulatory care sector. Care is provided by free-standing ‘public’ and ‘private’ physicians, who are self-employed and mostly operate in single practices. The former are called contract physicians, as they are in a contractual relationship with at least one health insurance fund, and the latter non-contract physicians. Since 2007, the number of contract physicians has slightly decreased (2007: 7770; 2013: 7657) with the contrary applying to non-contract physicians (2007: 9018; 2013: 10 883).16 Specialist care may be received both in these physician practices as well as in hospital outpatient departments. To regulate ambulatory care provided by contract physicians, the Chambers of Physicians in each state and the health insurance funds for the Main Association of the Austrian Social Security Institutions annually negotiate general contracts. In these contracts, the number of contract physicians and their spatial distribution are set. In contrast, non-contract physicians can chose their location freely. Indeed, non-contract GPs compared to contract GPs are more likely concentrated in richer districts, according to an analysis of Vienna.17 A full contract with a health insurance normally comprises 20 weekly office hours. As the vast majority of contract physicians works in single practices, this constitutes a problem with accessibility not only restricted regarding out-of-hours care, whereas outpatient departments are accessible without appointment during the day, and in a more limited form also out-of-hours. Contract physicians are financed based on a mixed system under which physicians receive a base payment for each quarter a patient actually visits, and fee-for-service. While the ratio is typically 70:30 in favour of the base payment for GPs, it is roughly 30:70 for specialists. De facto, the remuneration varies between social insurance institutions, states and medical disciplines.15 While insurees are free to consult any physician in the ambulatory care sector, full cost coverage by SHI is only provided for treatments by contract physicians (for around 80% of the insurees; three public health insurance funds, however, stipulate a 20% co-payment per consultation). Non-contract physicians charge their patients directly. If the respective service is included in the patient’s insurance fund’s benefit catalogue, the patient may file the settled invoice with their health insurance fund, which reimburses 80% of the tariff paid for the same service provided by a contract physician. The intention is to curb the costs in the growing sector of non-contract physicians by effectively imposing user charges. Costs for non-contract physicians are potentially covered by complementary private health insurance. While patient satisfaction with the public healthcare system is very high,18 around one in three Austrians has some form of private health insurance.15 Mostly, these insurance plans focus on free choice of hospital physicians and upgraded amenities in hospitals. In fact, private health insurance contributions constitute an important source of income for many hospital physicians, especially since remuneration in the public hospital sector is considered low and provides little incentives to highly-qualified physicians. Commonly, physicians operate both in the inpatient and non-hospital based outpatient sector.14 In 2010, 20% of all contract physicians and 50% of all non-contract physicians moonlighted as hospital-based physicians, especially specialists: the share of non-contract specialists working in public and private hospitals varies between 44% (dermatologists) and 68% (surgeons).14 Basically no gatekeeping is in place in the ambulatory care sector (exceptions: referrals to practice-based physicians are necessary for laboratory diagnostics and radiology).19 This also applies to hospital outpatient departments, which may be visited without referral and at no cost (a fee introduced in 2001 was withdrawn in 2003). Every public acute care hospital holds an outpatient department. By law, these outpatient departments offer emergency care and services that may not be sufficiently provided in other outpatient facilities.20 Services in outpatient departments are funded from all levels of government and SHI through a global budget to each outpatient department, albeit far from cost-covering. SHI contributes a lump sum to hospital financing, but has to pay fees to contract physicians. This creates a strong incentive for SHI to ration services outside the hospital and thus shift them to outpatient departments, where patients do not cause additional expenditures. Against this institutional background, we hypothesise as follows: Care provided by contract GPs is expected to substitute for care in hospital outpatient departments, whereas non-contract specialists are expected to induce demand in hospital outpatient departments. Method Dataset and statistical analysis This study is based on SHI administrative data complemented with data from Statistics Austria (details in Supplementary Appendix S2). The initial dataset includes information on all 121 Austrian districts in 2010 (table 2). One district (1090 Vienna) was excluded due to its outliner position, i.e. accommodating Austria’s biggest hospital. Hypotheses were tested using ordinary least square regression analysis, limited to districts with an outpatient department. This effectively reduced the dataset to 89 observations (table 3). Table 2 Descriptive statistics of dependent and independent variables (district level)  N Mean VC Min Max Hospital department outpatient frequencya 90 2.294 0.768 0.033 7.672 Initial consultations with contract physicians: GPsa 120 3.156 0.280 0.245 5.873 Initial consultations with contract physicians: specialistsa 117 1.866 0.414 0.001 4.272 AVE with non-contract physicians: GPsa 120 0.009 0.920 0.000 0.058 AVE with non-contract physicians: specialistsa 117 0.048 0.674 0.003 0.195 Hospitals inpatient staysa 90 0.362 0.690 0.009 1.449 Share of highly educated population 120 0.136 0.604 0.063 0.431 Share of unemployed population 120 0.029 0.394 0.011 0.074 Share of retired population 120 0.221 0.123 0.161 0.285 District area (100 km2) 120 6.990 0.900 0.011 32.683   N Mean VC Min Max Hospital department outpatient frequencya 90 2.294 0.768 0.033 7.672 Initial consultations with contract physicians: GPsa 120 3.156 0.280 0.245 5.873 Initial consultations with contract physicians: specialistsa 117 1.866 0.414 0.001 4.272 AVE with non-contract physicians: GPsa 120 0.009 0.920 0.000 0.058 AVE with non-contract physicians: specialistsa 117 0.048 0.674 0.003 0.195 Hospitals inpatient staysa 90 0.362 0.690 0.009 1.449 Share of highly educated population 120 0.136 0.604 0.063 0.431 Share of unemployed population 120 0.029 0.394 0.011 0.074 Share of retired population 120 0.221 0.123 0.161 0.285 District area (100 km2) 120 6.990 0.900 0.011 32.683  AVE, outpatient care units; N, number of districts; VC, coefficient of variation. a: Per commuter-adjusted population. Table 2 Descriptive statistics of dependent and independent variables (district level)  N Mean VC Min Max Hospital department outpatient frequencya 90 2.294 0.768 0.033 7.672 Initial consultations with contract physicians: GPsa 120 3.156 0.280 0.245 5.873 Initial consultations with contract physicians: specialistsa 117 1.866 0.414 0.001 4.272 AVE with non-contract physicians: GPsa 120 0.009 0.920 0.000 0.058 AVE with non-contract physicians: specialistsa 117 0.048 0.674 0.003 0.195 Hospitals inpatient staysa 90 0.362 0.690 0.009 1.449 Share of highly educated population 120 0.136 0.604 0.063 0.431 Share of unemployed population 120 0.029 0.394 0.011 0.074 Share of retired population 120 0.221 0.123 0.161 0.285 District area (100 km2) 120 6.990 0.900 0.011 32.683   N Mean VC Min Max Hospital department outpatient frequencya 90 2.294 0.768 0.033 7.672 Initial consultations with contract physicians: GPsa 120 3.156 0.280 0.245 5.873 Initial consultations with contract physicians: specialistsa 117 1.866 0.414 0.001 4.272 AVE with non-contract physicians: GPsa 120 0.009 0.920 0.000 0.058 AVE with non-contract physicians: specialistsa 117 0.048 0.674 0.003 0.195 Hospitals inpatient staysa 90 0.362 0.690 0.009 1.449 Share of highly educated population 120 0.136 0.604 0.063 0.431 Share of unemployed population 120 0.029 0.394 0.011 0.074 Share of retired population 120 0.221 0.123 0.161 0.285 District area (100 km2) 120 6.990 0.900 0.011 32.683  AVE, outpatient care units; N, number of districts; VC, coefficient of variation. a: Per commuter-adjusted population. Table 3 Influence of ambulatory care utilization, socio-demographic, geographic and inpatient activity variables on hospital department utilization (district level)  Dependent variable: Hospital department outpatient frequency Independent variables B SE t value  Tolerance VIF Initial consultations with contract GPa) −0.412 0.180 −2.293 ** 0.331 3.020 Initial consultations with contract specialistsa) −0.209 0.163 −1.284  0.449 2.228 AVE with non-contract GPa) −7.075 9.762 −0.725  0.801 1.248 AVE with non-contract specialistsa) 6.600 3.364 1.962 ** 0.691 1.446 Hospital Inpatient staysa) 6.143 0.433 14.180 *** 0.549 1.822 Share of highly educated population −6.397 1.686 −3.793 *** 0.402 2.485 Share of unemployed population 10.497 8.614 1.219  0.661 1.512 Share of retired population −0.463 3.759 −0.123  0.627 1.594 District area (100 km2) −0.031 0.017 −1.852 * 0.491 2.039 Constant 2.366 1.072 2.207 **     Dependent variable: Hospital department outpatient frequency Independent variables B SE t value  Tolerance VIF Initial consultations with contract GPa) −0.412 0.180 −2.293 ** 0.331 3.020 Initial consultations with contract specialistsa) −0.209 0.163 −1.284  0.449 2.228 AVE with non-contract GPa) −7.075 9.762 −0.725  0.801 1.248 AVE with non-contract specialistsa) 6.600 3.364 1.962 ** 0.691 1.446 Hospital Inpatient staysa) 6.143 0.433 14.180 *** 0.549 1.822 Share of highly educated population −6.397 1.686 −3.793 *** 0.402 2.485 Share of unemployed population 10.497 8.614 1.219  0.661 1.512 Share of retired population −0.463 3.759 −0.123  0.627 1.594 District area (100 km2) −0.031 0.017 −1.852 * 0.491 2.039 Constant 2.366 1.072 2.207 **    AVE, outpatient care units; B, unstandardized coefficient; SE, standard error; VIF, variance inflation factor. a: Per communter-adjusted population. Level of significance: ***= 1%, **= 5%, *= 10%; Adj. R2 = 0.818, n = 89 (excluding 1090 Vienna). Table 3 Influence of ambulatory care utilization, socio-demographic, geographic and inpatient activity variables on hospital department utilization (district level)  Dependent variable: Hospital department outpatient frequency Independent variables B SE t value  Tolerance VIF Initial consultations with contract GPa) −0.412 0.180 −2.293 ** 0.331 3.020 Initial consultations with contract specialistsa) −0.209 0.163 −1.284  0.449 2.228 AVE with non-contract GPa) −7.075 9.762 −0.725  0.801 1.248 AVE with non-contract specialistsa) 6.600 3.364 1.962 ** 0.691 1.446 Hospital Inpatient staysa) 6.143 0.433 14.180 *** 0.549 1.822 Share of highly educated population −6.397 1.686 −3.793 *** 0.402 2.485 Share of unemployed population 10.497 8.614 1.219  0.661 1.512 Share of retired population −0.463 3.759 −0.123  0.627 1.594 District area (100 km2) −0.031 0.017 −1.852 * 0.491 2.039 Constant 2.366 1.072 2.207 **     Dependent variable: Hospital department outpatient frequency Independent variables B SE t value  Tolerance VIF Initial consultations with contract GPa) −0.412 0.180 −2.293 ** 0.331 3.020 Initial consultations with contract specialistsa) −0.209 0.163 −1.284  0.449 2.228 AVE with non-contract GPa) −7.075 9.762 −0.725  0.801 1.248 AVE with non-contract specialistsa) 6.600 3.364 1.962 ** 0.691 1.446 Hospital Inpatient staysa) 6.143 0.433 14.180 *** 0.549 1.822 Share of highly educated population −6.397 1.686 −3.793 *** 0.402 2.485 Share of unemployed population 10.497 8.614 1.219  0.661 1.512 Share of retired population −0.463 3.759 −0.123  0.627 1.594 District area (100 km2) −0.031 0.017 −1.852 * 0.491 2.039 Constant 2.366 1.072 2.207 **    AVE, outpatient care units; B, unstandardized coefficient; SE, standard error; VIF, variance inflation factor. a: Per communter-adjusted population. Level of significance: ***= 1%, **= 5%, *= 10%; Adj. R2 = 0.818, n = 89 (excluding 1090 Vienna). Variables The dependent variable, demand for care in outpatient departments, was approximated by the number of contacts by outpatients in ‘fund hospitals’, which are representatives of the acute care sector.20 Physician utilisation was retrieved from a database administered by the Main Association of the Austrian Social Security Institutions. Therein, two variables are used as proxies for demand for physician services: initial consultations and outpatient care units (‘Ambulante Versorgungseinheiten’, AVE). In our analysis, demand for care provided in contract physician practices by district is based on patient consultations. Consultations are defined as face-to-face contacts between physician and patient. Both contacts in hospital outpatient departments and consultations are measures for demand in terms of pure volume and thus considered best comparable. However, as data on consultations is not available for non-contract physicians, demand for these providers is approximated by the second-best option, i.e. AVE. AVE of non-contract physicians are calculated from the ratio between patients’ first consultations in a region and the Austrian mean per ‘full contract physician’ (see Supplementary Appendix S2). A full contract physician is defined as a physician having an all-year contract with all health insurance funds. Consequently, the concept of AVE may be paraphrased with ‘full contract equivalent’. Compared to physician headcount, AVE and initial consultations yield significant benefits. The activity of physicians in providing care to their regional population is not picked up by headcounts. This seems especially problematic for non-contract physicians, with often only marginal opening hours. Their potential to provide care is thus better captured based on actual utilization of their services instead of their mere presence as mirrored in headcounts. Summing up, as explanatory variables, care provided by outpatient physicians is incorporated as follows: First, in terms of consultations for both contract GPs and contract specialists, with the latter including paediatricians, surgeons, internists, gynaecologists, neurologists, psychologists, dermatologists, ophthalmologists, otologists, urologists, pneumologists, orthopaedists and emergency surgeons. Second, for both non-contract GPs and non-contract specialists in terms of AVE. To control for socioeconomic and geographic characteristics, which in line with existing literature (details in Supplementary Appendix S1) are also expected to explain variations in outpatient demand,5,6,21,22 we consider the districts’ educational level (measured by the share of the population with tertiary education), economic situation (measured by the share of the unemployed population), demographics (measured by the share of retired people) and the district size (in 100 km2). The variable inpatient stays was included as a proxy for hospital activity. While higher education is closely correlated with an individual’s income and thus affordability of non-contract physicians, low socioeconomic status is correlated with poorer health, and retirement with higher healthcare utilization. District size serves as a proxy for travel costs. Finally, to adjust for the size of the regional population, utilization variables were divided by the commuter-adjusted population (table 2). Commuter-adjusted population was preferred over standard population as according to SHI expert opinion people typically utilise medical resources close to their workplace rather than their home when working in a larger city. Descriptions and details of the dataset are provided in tables 1 and 2. Table 1 Pearson correlation coefficients of dependent and independent variables (district level)  (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) Hospital department outpatient frequencya 1          (2) Initial consultations with contract physicians: GPb −0.465** 1         (3) Initial consultations with contract physicians: specialistsc 0.447** −0.325** 1        (4) AVE with non-contract physicians: GPd 0.097 0.002 −0.126 1       (5) AVE with non-contract physicians: specialistse 0.268* −0.210* 0.225* 0.286** 1      (6) Hospitals inpatient staysf 0.894** −0.451** 0.527** 0.093 0.242* 1     (7) Share of highly educatedg 0.198 −0.725** 0.252** −0.118 0.257** 0.283** 1    (8) Share of unemployedh 0.174 −0.344** 0.412** −0.136 −0.083 0.152 0.241** 1   (9) Share of retiredi 0.220* 0.268** 0.191* −0.056 −0.176 0.293** −0.227* −0.154 1  (10) District areaj (100 km2) −0.309** 0.43** −0.387** 0.230* 0.036 −0.309** −0.532** −0.201* −0.075 1   (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) Hospital department outpatient frequencya 1          (2) Initial consultations with contract physicians: GPb −0.465** 1         (3) Initial consultations with contract physicians: specialistsc 0.447** −0.325** 1        (4) AVE with non-contract physicians: GPd 0.097 0.002 −0.126 1       (5) AVE with non-contract physicians: specialistse 0.268* −0.210* 0.225* 0.286** 1      (6) Hospitals inpatient staysf 0.894** −0.451** 0.527** 0.093 0.242* 1     (7) Share of highly educatedg 0.198 −0.725** 0.252** −0.118 0.257** 0.283** 1    (8) Share of unemployedh 0.174 −0.344** 0.412** −0.136 −0.083 0.152 0.241** 1   (9) Share of retiredi 0.220* 0.268** 0.191* −0.056 −0.176 0.293** −0.227* −0.154 1  (10) District areaj (100 km2) −0.309** 0.43** −0.387** 0.230* 0.036 −0.309** −0.532** −0.201* −0.075 1  Notes: Level of significance: **= 1%, * = 5%. a: Frequency of patients in hospital outpatient departments, per capita (commuter-adjusted). b: Number of initial consultations in (free-standing) practices of general practitioners with one (or more) health insurance contract(s), per capita (commuter-adjusted). c: Number of initial consultations in (free-standing) practices of specialists with one (or more) health insurance contract(s), per capita (commuter-adjusted). d: Number of AVE (outpatient care units, ‘Ambulante Versorgungseinheiten’) of general practitioners without a health insurance contract, per capita (commuter-adjusted). e: Number of AVE of specialists without a health insurance contract, per capita (commuter-adjusted). f: Number of inpatient hospital stays (admissions + discharges + deaths)/2), per capita (commuter-adjusted). g: Number of graduates of universities (of applied sciences) and related institutions, as a share of the total population. h: Number of unemployed (economically active population), as a share of the total population. i: Number of retirees (economically inactive population), as a share of the total population. j: District area, in 100 km2. Table 1 Pearson correlation coefficients of dependent and independent variables (district level)  (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) Hospital department outpatient frequencya 1          (2) Initial consultations with contract physicians: GPb −0.465** 1         (3) Initial consultations with contract physicians: specialistsc 0.447** −0.325** 1        (4) AVE with non-contract physicians: GPd 0.097 0.002 −0.126 1       (5) AVE with non-contract physicians: specialistse 0.268* −0.210* 0.225* 0.286** 1      (6) Hospitals inpatient staysf 0.894** −0.451** 0.527** 0.093 0.242* 1     (7) Share of highly educatedg 0.198 −0.725** 0.252** −0.118 0.257** 0.283** 1    (8) Share of unemployedh 0.174 −0.344** 0.412** −0.136 −0.083 0.152 0.241** 1   (9) Share of retiredi 0.220* 0.268** 0.191* −0.056 −0.176 0.293** −0.227* −0.154 1  (10) District areaj (100 km2) −0.309** 0.43** −0.387** 0.230* 0.036 −0.309** −0.532** −0.201* −0.075 1   (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) Hospital department outpatient frequencya 1          (2) Initial consultations with contract physicians: GPb −0.465** 1         (3) Initial consultations with contract physicians: specialistsc 0.447** −0.325** 1        (4) AVE with non-contract physicians: GPd 0.097 0.002 −0.126 1       (5) AVE with non-contract physicians: specialistse 0.268* −0.210* 0.225* 0.286** 1      (6) Hospitals inpatient staysf 0.894** −0.451** 0.527** 0.093 0.242* 1     (7) Share of highly educatedg 0.198 −0.725** 0.252** −0.118 0.257** 0.283** 1    (8) Share of unemployedh 0.174 −0.344** 0.412** −0.136 −0.083 0.152 0.241** 1   (9) Share of retiredi 0.220* 0.268** 0.191* −0.056 −0.176 0.293** −0.227* −0.154 1  (10) District areaj (100 km2) −0.309** 0.43** −0.387** 0.230* 0.036 −0.309** −0.532** −0.201* −0.075 1  Notes: Level of significance: **= 1%, * = 5%. a: Frequency of patients in hospital outpatient departments, per capita (commuter-adjusted). b: Number of initial consultations in (free-standing) practices of general practitioners with one (or more) health insurance contract(s), per capita (commuter-adjusted). c: Number of initial consultations in (free-standing) practices of specialists with one (or more) health insurance contract(s), per capita (commuter-adjusted). d: Number of AVE (outpatient care units, ‘Ambulante Versorgungseinheiten’) of general practitioners without a health insurance contract, per capita (commuter-adjusted). e: Number of AVE of specialists without a health insurance contract, per capita (commuter-adjusted). f: Number of inpatient hospital stays (admissions + discharges + deaths)/2), per capita (commuter-adjusted). g: Number of graduates of universities (of applied sciences) and related institutions, as a share of the total population. h: Number of unemployed (economically active population), as a share of the total population. i: Number of retirees (economically inactive population), as a share of the total population. j: District area, in 100 km2. Results As shown in table 3, we find the following statistically significant results (α = 5%): districts with a higher level of care provision by contract GPs show lower use of outpatient departments. In contrast, a higher level of care provision by non-contract specialists on the district level is related to higher utilization of care in outpatient departments. The number of inpatient stays is highly positively associated with outpatient department contacts. In contrast, a higher share of highly educated people is negatively associated with hospital outpatient department utilization. The same applies (α = 10%) to the district size: the larger the district, the lower the utilisation of outpatient departments. Discussion This ecological study showed that in Austria, care by contract GPs on the district level is in a substitutive relation with utilisation in outpatient departments, while care provided by non-contract specialists is complementary to care in outpatient departments. The first result is in line with earlier findings: 60% of all patients showing up with no referral in an Austrian outpatient department could have alternatively been treated by GPs in the ambulatory care sector.22 The substitutive relationship between GPs and hospital outpatient departments is also supported by (correlation) analyses from Denmark4 and Norway6, but contrasts a recent regression analysis,5 finding a positive relationship between GP utilisation and outpatient secondary care use for the elderly Norwegian population. Another analysis for Norway shows a negative impact of GPs on private specialist utilization but no association between GP and hospital contacts.11 However, differences in primary and secondary care make international comparisons of such results difficult.7 This article is first to identify the interrelation between the utilisation of these sub-sectors of outpatient and ambulatory care provision in Austria. Nevertheless, some limitations apply. First, districts without outpatient departments, which are mostly rural, were excluded from the analysis. However, given their small population size, the exclusion effect on the results is expected to be marginal. Second, hospital stays are a major determinant of utilisation in hospital outpatient departments, as hospitals with high inpatient activity are also expected to provide more care in outpatient facilities. In a sensitivity analysis excluding the inpatient variable, the adjusted R2 decreased from 0.819 (table 3) to 0.364 (Supplementary Appendix S3). The observed associations in the main analysis regarding contract GPs and non-contract specialists, however, remain the same (α = 5%). Additionally, a substitutive association between contract specialists and hospital outpatient departments emerges (α = 10%), which seems plausible since both provide partly similar services. Third, as the use of the commuter-adjusted population to calculate densities is not common, a sensitivity analysis was carried out with the standard population (Supplementary Appendix S4). This analysis confirms the complementary relationship between non-contract specialists and hospital outpatient departments whereas the association with the contract GP is not statistically significant. Fourth, alas we were provided with data for one year only, making this analysis cross-sectional. Further analyses of several years might elevate statistical power and enable a causal analysis. Finally, utilization in outpatient departments should be but might not actually be counted the same way in all hospitals. Health administration in Austria is currently working on this issue, however, without a definitive timeline. Overall, the findings provide vital insights into the relationship between the different ambulatory care pillars. On the one hand, the utilisation of contract GPs seems to replace and thus reduce the utilisation in outpatient departments. This effect is likely to be observed in many countries with strong primary healthcare sectors and as the Austrian results show, even holds for a country with unlimited access. On the other hand, utilisation of non-contract specialists seems to increase the utilisation in outpatient departments. This makes sense given that many non-contract specialists additionally work in hospitals; these physicians might avoid the more complex diagnostics in their practices by sending the patients to ‘their’ hospital instead. This suggests that the complementary relation between these non-contract specialists might be demand-inducing. In terms of control variables, a higher share of academics correlates with lower outpatient department utilisation. This points at different utilisation behaviours emanating from socially more advantaged and disadvantaged areas and is in line with a recent study identifying different access points to outpatient care by socioeconomic position:21,23 For instance, Austrians born in Turkey were found to be more likely to report an outpatient department visit without a prior GP visit.21 Finally, our observed positive relationship between district size and outpatient department utilisation hints at a potential travel cost effect as supported by international evidence:5,6 in larger districts, outpatient department use causes higher travel costs and might thus be considered less attractive. Interestingly, despite their importance in the context of the Austrian system, these findings are the first to provide an actual evidence-base for policy measures. In a setting where access to ambulatory care is unrestricted, interdependencies are highly relevant and a potential cause for misallocation of resources. The ongoing healthcare reform specifically aims to establish a primary healthcare system living up to international standards. This study supports the notion of improving the role of GPs in the ambulatory care sector, while the growth of the non-contract specialist sector calls for action. Conclusion While care by non-contract specialists was found to be in a complementary and potentially demand-inducing relation with outpatient departments in Austria, care by contract GPs replaced care in outpatient departments. If combined with measures to increase the attractiveness of the ambulatory care sector compared to outpatient departments, strengthening this sector may thus actually results in increased take-up of ambulatory care, as intended by the currently ongoing Austrian healthcare reform. These findings are also an indication that channelling patient flows might prove cost-effective. Moreover, further thoughts are indicated concerning physicians simultaneously working in hospitals and their own private practice. Policy-makers internationally should also be aware of the interrelations between public and private provision in ambulatory care. For example, Germany tends to increase competition between outpatient care in hospitals and non-hospital based settings, the gains of which could be mitigated by demand induction. This is also the case for Beveridge-type countries like England and Sweden, where the private sector is increasingly used to provide alternate (ambulatory) care settings with the aim of fostering competition, but at the same time facilitating the mechanisms that are the subject matter of this study. Supplementary data Supplementary data are available at EURPUB online. Acknowledgements We would like to thank the Main Association of Austrian Social Security Institutions and the Ministry of Health for providing data. We thank the participants at the 10th World Congress in Health Economics Conference (iHEA) in Dublin (July 13–16, 2014) and at the 1st Austrian Health Economics Association Conference (ATHEA) in Vienna (November 28-29, 2014) as well as three anonymous reviewers for constructive comments and feedback. Funding This work was supported by the Main Association of Austrian Social Security Institutions. Conflict of interest: The Institute for Advanced Studies has a framework contract with the Main Association of Austrian Security Institutions under which data is provided and projects are conducted. Key points In the European context, relatively little is known about the interdependencies in the utilisation of care provided in hospital outpatient departments and free-standing physician practices, especially in two-tiered healthcare systems. This question is particularly vital for care settings with outpatient specialist care provided both in hospitals and the non-hospital based ambulatory care sector and by public and private providers. This ecological study is first to find that in Austria, ambulatory care provided by public (contract) physicians is associated with lower use of care in hospital outpatient departments, while a higher level of care by private (non-contract) specialists is associated with higher utilization in outpatient departments. This points a potentially demand-inducing relationship between non-contract specialists and outpatient hospital departments, while primary care by contract general physicians appears to be able to replace care in outpatient departments and attractiveness of the primary care sector should thus be further increased in Austria. For countries with similar outpatient and ambulatory care systems the results imply that the gains from fostering competition between public and private providers might be mitigated by demand induction. References 1 Fortney JC Steffick DE Burgess JF , et al. . Are primary care services a substitute or complement for specialty and inpatient services? Health Serv Res   2005 ; 40(5p1) : 1422 – 42 . Google Scholar Crossref Search ADS   2 Davis K Russell LB . The substitution of hospital outpatient care for inpatient care . Rev Econ Stat   1972 : 109 – 20 . 3 Adhikari SR . An assessment of a substitute or complement for inpatient and outpatient care of visceral leishmaniasis in Nepal . J Vector Borne Dis   2012 ; 49 : 242 – 48 . Google Scholar PubMed   4 Groenewegen PP . Substitution of primary care and specialist care: a regional analysis in Denmark . Soc Sci Med   1991 ; 33 : 471 – 76 . Google Scholar Crossref Search ADS PubMed   5 Deraas TS Bertsen GR Hasvold T Ringberg U Forde OH . Is a high level of general practitioner consultations associated with low outpatients specialist clinic use? A cross-sectional study . BMJ Open   2013 ; 3 : e002041 . Google Scholar Crossref Search ADS PubMed   6 Lindstrom K Engström S Bengtsson C , et al. . Determinants of hospitalisation rates: does primary health care play a role? Scand J Prim Health Care   2003 ; 21 : 15 – 20 . Google Scholar Crossref Search ADS PubMed   7 Tjerbo T . Does competition among general practitioners increase or decrease the consumption of specialist health care? Health Econ Policy Law   2010 ; 5 : 53 – 70 . Google Scholar Crossref Search ADS PubMed   8 Propper C . The demand for private health care in the UK . J Health Econ   2000 ; 19 : 855 – 76 . Google Scholar Crossref Search ADS PubMed   9 Fabbri D Monfardini C . Public vs. private health care services demand in Italy . Giornale degli economisti e annali di economia   2003 ; 93 – 123 . 10 Atella V Deb P . Are primary care physicians, public and private sector specialists substitutes or complements? Evidence from a simultaneous equations model for count data . J Health Econ   2008 ; 27 : 770 – 85 . Google Scholar Crossref Search ADS PubMed   11 Iversen T Kopperud GS . The impact of accessibility on the use of specialist health care in Norway . Health Care Manag Sci   2003 ; 6 : 249 – 61 . Google Scholar Crossref Search ADS PubMed   12 WHO . European health for all database. Available at: http://data.euro.who.int/hfadb/ (8 November 2015, date last accessed). 13 Rechnungshof. Finanzierung und Kosten von Leistungen in Spitalsambulanzen und Ordinationen [Funding and costs of services in hospital outpatient departments and physician practices]. Report 2011/3. 14 Gächter M Schwazer P Theurl E , et al. . Physician density in a two-tiered health care system . Health Policy   2012 ; 106 : 257 – 68 . Google Scholar Crossref Search ADS PubMed   15 Hofmarcher MM Quentin W . Austria: health system review . HiT   2013 ; 15 : 1 – 292 . Google Scholar PubMed   16 Statistik Austria . Jahrbuch der Gesundheitsstatistik 2014 [Yearbook of health statistics 2014]. Statistik Austria, 2014 . 17 Piribauer F Thaler K Harris MF . Covert checks by standardised patients of general practitioners' delivery of new periodic health examinations: clustered cross-sectional study from a consumer organisation . BMJ Open   2012 ; 2 : e000744 . Google Scholar Crossref Search ADS PubMed   18 Eurobarometer . Patient safety and quality of healthcare  . Brussels : TNS Opinion & Social , 2010 . 19 Stigler FL Starfield B Sprenger M , et al. . Assessing primary care in Austria: room for improvement . Fam Pract   2013 ; 30 : 185 – 9 . Google Scholar Crossref Search ADS PubMed   20 Czypionka T Kraus M Mayer S , et al. . Efficiency, ownership, and financing of hospitals: The case of Austria . Health Care Manag Sc   2014 ; 17 : 331 – 47 . Google Scholar Crossref Search ADS   21 Hoffmann K Stein KV Maier M , et al. . Access points to the different levels of health care and demographic predictors in a country without a gatekeeping system. Results of a cross-sectional study from Austria . Eur J Public Health   2013 ; 23 : 933 – 9 . Google Scholar Crossref Search ADS PubMed   22 Haidinger G Eckert-Graf L Wirgler PE , et al. . Selbstzuweiser im Spital – wie viele könnten im primär-medizinischen Bereich behandelt werden? [Patients with self-referral to the hospital – how many could have been treated by the primary sector?] Z Allgemeinmed   2013 ; 89 : 41 – 6 . 23 Hoffmann K Stein KV Dorner TE . Differences in access points to the ambulatory health care system across Austrian federal states . Wiener Med Wochenschr   2014 ; 164 : 152 – 9 . Google Scholar Crossref Search ADS   © The Author 2016. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.


This is a preview of a remote PDF: https://academic.oup.com/eurpub/article-pdf/27/1/20/10301786/ckw153.pdf

Czypionka, Thomas, Röhrling, Gerald, Mayer, Susanne. The relationship between outpatient department utilisation and non-hospital ambulatory care in Austria, European Journal of Public Health, 2017, 20-25, DOI: 10.1093/eurpub/ckw153