Swiss hospital financing with DRGs: Are there treatments/- combinations that are associated with profitability?

Subelack, Jonas Working Paper Swiss hospital financing with DRGs: Are there treatments/- combinations that are associated with profitability? Working Paper Series in Health Economics, Management and Policy, No. 2025-02 Provided in Cooperation with: University of St.Gallen, School of Medicine, Chair of Health Economics, Policy and Management Suggested Citation: Subelack, Jonas (2025) : Swiss hospital financing with DRGs: Are there treatments/- combinations that are associated with profitability?, Working Paper Series in Health Economics, Management and Policy, No. 2025-02, University of St.Gallen, School of Medicine, Chair of Health Economics, Policy and Management, St.Gallen This Version is available at: https://hdl.handle.net/10419/318261 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ Working Paper Series in Health Economics, Management and Policy 2025 – Nr. 02 2025 Swiss hospital financing with DRGs: Are there treatments/- combinations that are associated with profitability? Jonas Subelack I Working Paper Series in Health Economics, Management and Policy Editor Prof. Dr. Alexander Geissler Professor Chair of Health Economics, Policy, and Management School of Medicine University of St.Gallen Editorial office Jonas Subelack Research assistant Chair of Health Economics, Policy, and Management School of Medicine University of St.Gallen The entire series of publications is available on our website at: https://med.unisg.ch/en/research/healthcare-management/publications/ © 2025. This publication is licensed by the CC license CC-BY-NC-ND 4.0 II Swiss hospital financing with DRGs: Are there treatments/-combinations that are associated with profitability? Keywords: Hospital financing, hospital profitability, hospital reimbursement, DRG, Switzerland JEL Classification: H51, I11, I15, I18, L51 Author: Jonas Subelack Research assistant Chair of Health Economics, Policy, and Management, School of Medicine, University of St.Gallen [email protected] Recommended citation: Subelack, Jonas (2025): Swiss hospital financing with DRGs: Are there treatments/-combinations that are associated with profitability? Working Paper Series in Health Economics, Management and Policy, No. 202502, University of St.Gallen, School of Medicine, Chair of Health Economics, Policy and Management, St.Gallen. 1 Swiss hospital financing with DRGs: Are there treatments/ -combinations that are associated with profitability? Jonas Subelack; University of St.Gallen, School of Medicine Keywords: Hospital financing, hospital profitability, hospital reimbursement, DRG, Switzerland JEL: H51, I11, I15, I18, L51 Declarations Acknowledgements I want to thank Alexander Geissler for inspiring the research question. Furthermore, I want to thank David Ehlig and Justus Vogel for recommending the Apriori algorithm to address the second research question. Also, I want to thank Charlotte Schneider and Daria BukanovaBerend for their valuable support in assigning hospital cases to specific SPLGs. Finally, I want to thank the Zurich Department of Health for providing the SPLG Grouper. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Conflict of interest None. Ethics approval and consent to participate Not applicable. Data availability The first dataset (“Kennzahlen der Schweizer Spitäler“) is publicly available on the website of the Swiss Federal Office of Public Health. The second dataset (“Medizinische Statistik der Krankenhäuser“) is available upon reasonable request from the Swiss Federal Statistical Office. 2 Abstract Objective: This study aims to investigate whether specific treatments or combinations of treatments are significantly associated with the profitability of Swiss acute-care hospitals under the current diagnosis-related group (S-DRG) reimbursement system, while accounting for differences between public and private institutions. Methods: A comprehensive panel dataset of 142 Swiss acute-care hospitals, spanning from 2015 to 2022, was utilized, combining detailed financial and clinical case-level data. Profitability was assessed through hospital-level net financial results excluding deficitcovering payments. All cases were assigned uniquely to a medically homogeneous service group or area, as determined by Swiss hospital capacity planning. Fixed-effects panel regression models analyzed the associations between service areas and profitability, while an Apriori association rule mining algorithm identified service group combinations associated with profitability. Results: From 2015 to 2022, overall hospital profitability margins declined continuously, with public hospitals consistently reporting lower profitability than private hospitals (net profitability margin: 0.75% vs. 1.61%), despite receiving substantial subsidies (CHF 67.1 million vs. CHF 4.1 million). The primary panel regression revealed that three service areas are significantly associated with hospital profitability: Ear, nose and throat (16,778 CHF; p<0.05), gynecology (27,456 CHF; p<0.01), and heart (10,725 CHF; p<0.01). The Apriori algorithm identified that the combination of the following service groups is most strongly linked to profitability: AUG1.2 (orbit, eyelids, tear ducts), BEW10 (plexus surgery), and GEF2 (interventional and endovascular vascular medicine; support: 0.051, confidence: 0.935, lift: 1.615). Conclusion: The analysis of hospital profitability based on the treatments and combinations of treatments performed indicates that the S-DRG reimbursement system is relatively fair. However, across all analyses, the heart service area is primarily associated with profitability, while the serious injury service area is mainly associated with losses. Therefore, minor adjustments to the S-DRG cost weights should be made to reduce this imbalance. 3 Abbreviations CHF Swiss franc CHOP Swiss classification of surgical interventions CMI Case mix index DRG Diagnosis related groups EBITDA Earnings before interest, taxes, depreciation and amortization FTE Full-time equivalent GDP Gross domestic product H+ Association of Swiss hospitals ICD International statistical classification of diseases and related health problems p.a. Per year (per annum) S-DRG Swiss diagnosis related groups SPLB Hospital planning service areas (DE: Spitalplanungs-Leistungsbereiche) SPLG Hospital planning service groups (DE: Spitalplanungs-Leistungsgruppen) VIF Variance inflation factor 4 Table of figures Figure 1: Hospital selection for analyses ........................................................................................... 9 Figure 2: Hospital net profitability margin from 2015 to 2022, stratified by public/ private hospital .................................................................................................................................................. 13 Figure 3: Fixed effects panel regression coefficients ...................................................................... 14 Figure 4: SPLG network diagram ..................................................................................................... 16 Figure 5: Distribution of institutions and cases per legal entity form ......................................... 28 Figure 6: Hospital net profitability margin from 2015 to 2022 ..................................................... 28 Figure 7: Fixed effects panel regression coefficients (incl. transplants) ...................................... 29 Figure 8: SPLG network diagram for private hospitals ................................................................. 29 Figure 9: SPLG network diagram for public hospitals .................................................................. 30 Figure 10: SPLG combinations linked with profitability ............................................................... 30 Figure 11: SPLG combinations linked with profitability for private hospitals .......................... 31 Figure 12: SPLG combinations linked with profitability for public hospitals ............................ 31 Figure 13: SPLG combinations linked with losses .......................................................................... 32 Figure 14: SPLG combinations linked with losses for public hospitals ...................................... 32 Table of tables Table 1: Descriptive Swiss hospital statistics from 2015 to 2022 .................................................. 11 Table 2: Fixed-effects panel regression analysis of absolute hospital profitability (CHF) with cluster-robust standard errors ........................................................................................................... 15 Table 3: Typical SPLG combinations ................................................................................................ 18 Table 4: SPLG combinations linked with profitability ................................................................... 19 Table 5: SPLG combinations linked with losses ............................................................................. 20 Table 6: Fixed-effects panel regression analysis of relative hospital profitability margin with cluster-robust standard errors ........................................................................................................... 33 Table 7: Typical SPLG combinations for private hospitals ........................................................... 34 Table 8: Typical SPLG combinations for public hospitals ............................................................. 35 Table 9: SPLG combinations linked with profitability for private hospitals .............................. 36 Table 10: SPLG combinations linked with profitability for public hospitals ............................. 37 Table 11: SPLG combinations linked with losses ........................................................................... 38 Table 12: SPLG combinations linked with losses for public hospitals ........................................ 39 5 Introduction Swiss healthcare costs rose continuously over the last decades, from 7.3% of the GDP in 1990 to 9.1% in 2000 and to 11.7% in 2022 (1). Hospitals are the largest cost factor in the Swiss healthcare system, accounting for 35.7% of all costs in 2022 (2). Already in 2007, the Swiss government agreed on several initiatives to contain further cost increases through economic incentives, of which most were rolled out in 2012 (3). The key measures included the implementation of Swiss diagnosis-related group reimbursements (S-DRG) for hospitals, hospital capacity planning, dual financing of hospitals from insurances and cantons (federal states), and Swiss-wide free hospital selection for patients (4). The absolute reimbursement for an inpatient case in a Swiss hospital is generally based on a uniform cost weight per S-DRG, multiplied by the base rate of the individual hospital (and insurance company) (5, 6). The cost weight of the S-DRG reflects the average resource consumption of a case, calculated based on historical data from all Swiss hospitals (7). The base weight varies greatly between hospitals, ranging from CHF 8,426 at Diaconis Palliative Care Hospital (cost weight 1.0) to CHF 15,360 at Hochgebirgsklinik Davos in 2023 (8). Today, a growing number of Swiss hospitals are in financial distress, and 90% of Swiss hospitals are operating considerably below the target EBITDA margin of 10%, which is considered sustainable, allowing hospitals to finance long-term investments (9-12). In 2023, the median EBITDA margin of (acute-care) hospitals was 2.5% (6.2% in 2021), and 25% of the Swiss hospitals reported an EBITDA margin of less than 0.1% (12). On the one hand, H+ (the association of Swiss hospitals) argues that the Swiss hospital system is generally underfunded, so the current reimbursement system does not accurately reflect the actual cost increases (9). On the other hand, the question arises as to why some hospitals manage to operate profitably on an ongoing basis, such as Klinik Hirslanden AG, with average net profits of CHF 30,599,045 p.a. from 2015 to 2022 (13). In general, hospitals that operate more efficiently should make a profit within the S-DRG system, but several voices from medical doctors to hospital managers point out that some treatments are more financially lucrative than others (10, 14-17). For example, hospital treatments for children were described as unprofitable (16), while orthopedic treatments were described as quite profitable (14, 15). If this is true and some treatments generally generate higher margins than others, hospitals would achieve unequal 12 Rheumatology 33,276 0.3% 13,276 0.3% 16,246 0.3% Severe injuries 22,629 0.2% 7,461 0.2% 12,636 0.2% Surgical musculoskeletal system 1,478,664 14.2% 854,397 18.8% 371,177 14.2% Thoracic surgery 22,725 0.2% 7,942 0.2% 13,566 0.2% Transplants 4,463 0.0% 685 0.0% 3,648 0.0% Urology 477,075 4.6% 235,846 5.2% 181,976 4.6% Vascular 155,265 1.5% 65,877 1.5% 77,688 1.5% Visceral surgery 181,316 1.7% 88,059 1.9% 77,711 1.7% Notes: All hospitals are the sum of private (R1) and public (R2) hospitals as well as hospitals operated by associations/ foundations (R2) or as sole proprietorships (R3); a) Test for differences performed for year 2022 as latest year with data and most datapoints across years, b) Wilcoxon rank-sum test performed, since neither normal distribution (Shapiro-Wilk-Test) nor variance homogeneity (robust Levene’s test) are given, c) Test for differences performed for year 2020 as latest year with data, d) Chisquared test, e) Excluding deficit coverage departures p.a.; p<0.05), had a significantly higher bed occupation rate (80.9% vs. 70.1%; p<0.05), included significantly more often an emergency department (98.8% vs. 60.5%; p<0.05) and are significantly more often a teaching hospital (94.3% vs. 61.0%; p<0.05). The mean resource consumption per patient, as indicated by the CMI, is quite similar (0.974 vs. 0.956; p = 0.92). Public and private hospitals significantly deviate in terms of hospital type (p<0.05) with most public hospitals (48.8%) being general hospitals center care level 2 (K112), and most private hospitals (30.3%) being special clinics for surgery (K231). Public as well as private hospitals primarily treated patients from the SPLB basic package (all hospitals average: 48.9%), from surgical musculoskeletal system (14.2%), and obstetrics (7.3%). Financially, private hospitals on average reported profits (649,936 CHF) compared to public hospitals, which on average reported significant losses (-2,282,343 CHF; p<0.05). Public hospitals received significantly higher subsidies than private hospitals (67,055,126 CHF vs. 4,103,927 CHF; p<0.05). All hospitals generated the majority of revenues from inpatient treatments (55.2% of all hospitals’ revenues; 74.4% compared to outpatient treatments). Public hospitals relied significantly less on inpatient revenues than private hospitals (p<0.05). In 2015, the median hospital profitability margin was 1.19%, indicating that most hospitals were profitable; however, there was also a great variance, and some hospitals reported substantial losses (see Figure 6). Until 2022, the median hospital profitability margin had consistently decreased to 0.24%, with a dip in 2020, where most hospitals reported losses (median profitability margin: -0.96%). The variance in hospital profitability margins increased from 2015 to 2022. When stratifying hospital profitability margins by public and private hospitals, it is evident that private hospitals continuously reported higher profitability margins (2015: public: 0.75% vs. private: 1.61%; see Figure 2). Both public and private hospitals reported declining profitability margins. Since 2019, public hospitals have reported a negative 13 median hospital profitability margin, whereas private hospitals reported a negative median hospital profitability margin only once in 2020. The variance in the hospital profitability margins is greater for private than for public hospitals. Figure 2: Hospital net profitability margin from 2015 to 2022, stratified by public/ private hospital Note: Margin (%) referring to net profit margin(excluding deficit coverage) divided by hospital revenue; The whiskers in the boxplots represent values within 1.5 times the interquartile range (IQR), extending from the first quartile (Q1) to the third quartile (Q3). Values outside this range (potential outliers) are excluded from the visualization. Table 2 and Figure 3 present the results of the fixed-effects panel regression analysis with cluster-robust standard errors. When focusing on all hospitals and the absolute number of cases treated per SPLB, three significant SPLBs are identified, all of which are positively associated with hospital profitability: Ear, nose and throat (16,778 CHF; p<0.05), gynecology (27,456 CHF; p<0.01), and heart (10,725 CHF; p<0.01). Correspondingly, it can be inferred that an additional patient treated in the SPLB heart is associated with an average increase in annual hospital profit of around 10,725 CHF, ceteris paribus. The largest coefficients, albeit not significant, are nephrology (60,419 CHF) on the positive side and transplants (-182,811 CHF) on the opposing side. For private and public hospitals, the SPLB heart is also significantly positively associated in both subgroups (private: 3,775 CHF, p<0.05; public: 18,379 CHF, p<0.05). Gynecology is only significantly positively associated at public hospitals (43,541 CHF; p<0.01). Ear, nose and throat is not significant in any of the subgroups. Private hospitals also report three significantly negative associated SPLBs with hospital profitability: neurosurgery (-43,505 CHF; p<0.05), severe injuries (-106,727 CHF; p<0.05), and transplants (-542,132 CHF; p<0.01). 14 Figure 3: Fixed effects panel regression coefficients Note: Values indicate coefficient in CHF per additional case, with positive values indicating increasing profitability; excluding SPLB transplants because of extraordinarily high variance. Figure 7 includes all SPLB coefficients, including transplants. When examining all hospitals and analyzing the binary indication of whether a SPLB was treated, three SPLBs were significantly associated with profitability: opening an SPLB unit in surgical musculoskeletal system (-2,626,690 CHF; p<0.05) or transplants (-5,194,664 CHF; p<0.01) resulted in a significant loss. Opening an SPLB unit in visceral surgery (1,798,277 CHF; p<0.05) led to a significant increase in profitability. Surgical musculoskeletal system (-2,097,084 CHF; p<0.05) and visceral surgery (1,891,208 CHF; p<0.01) were similarly significantly associated in private hospitals. Transplants could not be analyzed for any subgroup due to collinearity. Additionally, ear, nose, and throat (-14,700,000 CHF; p < 0.01), endocrinology (5,545,300 CHF; p < 0.05), and gynecology (-11,300,000 CHF; p < 0.05) received statistically significant results for public hospitals. A robustness check with a similar regression, using profitability margin as the dependent variable instead of absolute profit, yielded directionally similar results (see Table 6). Specifically, heart, ear, nose and throat, gynecology, and visceral surgery are also partially significantly associated with the hospital's profitability margin. The Apriori-based network analysis revealed distinct patterns of associations among the SPLGs. The SPLG BP (basic package surgery and internal medicine) emerged as a central and interconnected hub, frequently co-occurring with numerous other SPLGs (see Figure 4). Basic package Dermatology, Hematology, Radio oncology Ear, nose and throat Endocrinology Gastroenterology Gynecology Heart Nephrology Neurology Neurosurgery Obstetrics Ophthalmology Pneumology Rheumatology Severe injuries Surgical musculoskeletal system Thoracic surgery Urology Vascular Visceral surgery -200000 -100000 0 100000 200000 15 Table 2: Fixed-effects panel regression analysis of absolute hospital profitability (CHF) with cluster-robust standard errors Note: a) The SPLBs Dermatology, Hematology, and Radio oncology were merged into the combined category Dermatology/Hematology/Radio oncology. This was necessary because the three SPLBs exhibited very high correlations, leading to problematic multicollinearity (Variance Inflation Factor: VIF > 10). By merging the most strongly correlated SPLBs, the multicollinearity was reduced to acceptable levels (VIF < 10), fulfilling the statistical requirements for the analysis. b) The independent variables are binary, with “1” indicating that a hospital is relevant within the SPLG and “0“ indicating not relevant. A hospital was classified as relevant if it was among the providers that collectively treated at least 97.5% of all cases within each SPLG, excluding providers with only minimal caseloads. If one SPLG was classified as relevant, the SPLB was considered relevant. c) „n.a.“ referring to independent variables that STATA has omitted because of collinearity. Hausman tests were performed to justify the selection of fixed-effects models over random-effects models, indicating significant unobserved heterogeneity at the hospital level. Additionally, modified Wald tests for groupwise heteroscedasticity, Wooldridge tests for autocorrelation, and Pesaran’s tests for cross-sectional dependence were conducted. Due to evidence of heteroscedasticity, autocorrelation, and potential crosssectional dependence, fixed-effects panel regressions with cluster-robust standard errors at the hospital level were applied to ensure robust and unbiased estimates. All hospitals Private hospitals Public hospitals Independent variables: Absolute SPLB case numbers SPLB offered by hospitalb Absolute SPLB case numbers SPLB offered by hospitalb Absolute SPLB case numbers SPLB offered by hospitalb SPLB Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Basic package 788 (1,294) n.a.c 2,104 (1,168) * n.a.c -1,094 (2,825) n.a.c Dermatology/ Hematology/ Radio oncologya -12,980 (14,425) -748,615 (510,610) -2,582 (7,830) -933,174 (538,342) * -12,748 (25,811) 16,300,000 (10,300,000) Ear, nose and throat 16,778 (8,218) ** -1,548,084 (1,026,655) 4,835 (2,954) -234,887 (764,646) 49,512 (25,091) * -14,700,000 (3,211,939) *** Endocrinology -1,997 (18,515) 318,924 (425,556) -1,902 (27,201) -91,198 (651,719) -13,369 (28,427) 5,535,300 (2,586,260) ** Gastroenterology 13,030 (13,103) -985,998 (940,816) 4,352 (12,457) 1,108,777 (902,719) 10,114 (26,507) -12,100,000 (9,493,829) Gynecology 27,456 (9,234) *** -1,363,788 (703,839) * 13,745 (9,981) -767,032 (689,799) 43,541 (14,305) *** -11,300,000 (5,189,834) ** Heart 10,725 (3,990) *** 1,057,310 (918,382) 3,775 (1,616) ** 664,512 (1,110,258) 18,379 (7,411) ** 6,591,464 (5,691,931) Nephrology 60,419 (41,184) -612,451 (998,626) 57,625 (59,098) -142,286 (912,539) 71,539 (58,649) -144,040 (2,118,296) Neurology 15 (13,060) 174,037 (493,022) 9,972 (15,996) -30,954 (489,105) -2,833 (21,268) 8,014,518 (14,500,000) Neurosurgery -10,838 (25,518) 55,838 (638,017) -43,505 (19,935) ** 720,206 (688,105) 82,888 (72,053) -2,135,011 (2,727,181) Obstetrics 4,382 (6,046) -107,049 (2,681,923) 347 (5,705) 1,430,144 (3,019,629) 14,517 (23,306) n.a.c Ophthalmology -23,418 (29,204) 260,918 (646,473) -617 (10,036) -906,898 (825,231) -67,780 (76,814) 808,042 (1,888,360) Pneumology 5,725 (10,899) -632,549 (631,180) 7,864 (6,335) 155,823 (612,505) -13,049 (23,948) -2,320,190 (5,290,946) Rheumatology -6,576 (42,203) 12,993 (522,067) 4,009 (39,187) 248,123 (615,220) -13,642 (112,442) -2,560,479 (8,677,581) Severe injuries -82,661 (44,188) * 59,808 (516,380) -106,727 (51,298) ** -15,608 (630,606) -88,934 (148,476) -4,475,578 (2,905,232) Surgical musculoskeletal system 1,503 (1,769) -2,626,690 (1,112,054) ** 3,944 (2,070) * -2,097,084 (978,876) ** -5,453 (6,074) n.a.c Thoracic surgery -40,366 (59,168) 1,356,002 (863,067) 40,532 (36,045) 1,898,662 (1,099,069) * -118,473 (119,616) 5,551,185 (2,829,515) * Transplants -182,811 (153,026) -5,194,664 (1,175,394) *** -542,132 (92,453) *** n.a.c -125,061 (170,870) n.a.c Urology 10,601 (6,113) * 314,817 (1,331,917) 2,288 (5,601) -24,311 (961,959) 13,596 (14,624) n.a.c Vascular -26,870 (14,092) * -484,036 (926,321) -17,713 (8,955) * -1,329,890 (968,516) -35,692 (30,204) 1,570,279 (2,757,239) Visceral surgery -16,942 (12,206) 1,798,277 (748,464) ** -13,064 (13,223) 1,891,208 (714,574) *** -26,518 (18,771) 7,894,242 (7,576,948) Confounders: Year, bed occupancy rate, share of private patients, share of inpatient revenue ***) p<0.01; **) p<0.05; *) p<0.1 16 Figure 4: SPLG network diagram Note: SPLG network diagram based on the top 100 rules with the highest lift. 17 Despite its central position and numerous connections, the strength of these associations (lift) is low to moderate. In contrast, the peripheral nodes formed several specialized and closely connected clusters, especially around the SPLGs BEW.x (SPLB: Surgical musculoskeletal system), VIS.x (SPLB: Visceral surgery), and URO.x (SPLB: Urology). Table 3 notes the fundamental rules. Here, the rule with the greatest lift (2.907) had the two SPLGs URO1.1 (urology with specialization in 'operative urology') and VIS1.4.1 (complex bariatric surgery) as antecedents for VIS1.4 (bariatric surgery) as consequent. The rule with the greatest lift (2.266) and two antecedents from different SPLGs than the consequent is GEB1 (basic obstetrics care) and URO1.1.4 (isolated adrenalectomy) as antecedents for HNO2 (thyroid and parathyroid surgery) as consequent. When examining the network analysis for the private hospitals (see Figure 8 and Table 7), the SPLG BP also appears as a central and highly connected networked hub with low to moderate lifts. A very dominant cluster with strong confidence and lifts of the BEW.x SPLGs is evident. Further, a more interconnected cluster with GAE.x (SPLB: Gastroenterology) and VIS.x SPLGs is visible. The rule with the greatest lift (2.488) had the two SPLGs KAR1 (cardiology including pacemaker) and NEU1 (neurology) as antecedents for GAE1.1 (specialized gastroenterology) as consequent. When examining the network analysis for public hospitals (see Figure 9, Table 8), it reveals a distinct structure without a clear center. One strong cluster is visible around HAE1.1 (highly aggressive lymphomas and acute leukemias), and several other, more interconnected clusters are visible around BEW7.2.1 (knee prosthesis replacement operations) and URO1.1.7 (implantation of an artificial urinary bladder sphincter). The rule with the greatest lift (3.280) had the two SPLGs AUG1.4 (cataract) and VIS1.2 (liver resection) as antecedents for AUG1.5 (vitreous humor/ retinal problems) as consequent. The rule with the greatest lift (3.114) and two antecedents from different SPLGs than the consequent is NCH1.1 (specialized neurosurgery) and GYNT (gynecological tumors) as antecedents for GEFA (interventions and vascular surgery intraabdominal vessels) as consequent. Further, the Apriori algorithm identified SPLG combinations that are linked with profitability. For all hospitals, the SPLGs that are most strongly linked with profitability (support: 0.051, confidence: 0.935, lift: 1.615) are AUG1.2 (orbit, eyelids, tear ducts), BEW10 (plexus surgery), and GEF2 (interventional and endovascular vascular medicine) (see Table 4). On an aggregated level, diverse combinations of the SPLGs AUG1.2 and BEW10 or AUG1.2 and VIS1.2 or BEW10 and HNO1.3 (middle ear surgery) are associated with profitability (see Figure 10). 18 Table 3: Typical SPLG combinations Antecedents Consequent Support Confidence Lift URO1.1, VIS1.4.1 VIS1.4 0.255 0.993 2.907 GAE1, VIS1.4.1 VIS1.4 0.259 0.986 2.888 PNE1, VIS1.4.1 VIS1.4 0.252 0.986 2.887 GYN1, VIS1.4.1 VIS1.4 0.258 0.983 2.878 DER1, VIS1.4.1 VIS1.4 0.256 0.983 2.877 VIS1.4.1 VIS1.4 0.264 0.980 2.869 VIS1, VIS1.4.1 VIS1.4 0.264 0.980 2.869 URO1, VIS1.4.1 VIS1.4 0.264 0.980 2.869 BEW1, VIS1.4.1 VIS1.4 0.264 0.980 2.869 BP, VIS1.4.1 VIS1.4 0.264 0.980 2.869 BEW5, VIS1.4.1 VIS1.4 0.256 0.980 2.868 BEW2, VIS1.4.1 VIS1.4 0.256 0.980 2.868 BEW6, VIS1.4.1 VIS1.4 0.250 0.979 2.866 GEF1, HNO2 ANG1 0.253 0.979 2.830 GEF1, RAD1 ANG1 0.256 0.976 2.821 GEF1, HNO1.1 ANG1 0.254 0.966 2.792 GAE1.1, GEF1 ANG1 0.261 0.961 2.776 GEB1, GEF1 ANG1 0.273 0.957 2.764 GEF1, GYN2 ANG1 0.270 0.950 2.745 GEF1, HAE3 ANG1 0.270 0.950 2.745 GEB1, URO1.1.4 HNO2 0.254 0.963 2.266 KAR1, URO1.1.4 HNO2 0.250 0.962 2.264 RAD1, VIS1.5 HNO2 0.256 0.960 2.259 GYN2, URO1.1.4 HNO2 0.257 0.957 2.252 KAR1.1, VIS1.5 HNO2 0.257 0.957 2.252 GAE1.1, URO1.1.4 HNO2 0.256 0.957 2.252 KAR1, VIS1.5 HNO2 0.283 0.955 2.247 KAR1.1, VIS1.5 KAR1 0.262 0.974 2.235 KAR1.1, VIS1.4 KAR1 0.258 0.964 2.212 NEU2.1, VIS1.5 KAR1 0.256 0.957 2.196 RAD1, VIS1.5 KAR1 0.255 0.953 2.189 GEF1, NEU1 NEP1 0.256 0.966 2.187 KAR1.1, RAD1 KAR1 0.279 0.952 2.184 HNO2, UNF1 KAR1 0.278 0.951 2.184 HAE1, VIS1.5 KAR1 0.276 0.951 2.183 NEU3, VIS1.5 NEP1 0.282 0.955 2.161 GEF1, NEU3 NEP1 0.256 0.954 2.158 ANG1, NEU1 NEP1 0.274 0.954 2.158 NEU2.1, VIS1.5 NEP1 0.255 0.953 2.158 HNO2, UNF1 NEP1 0.278 0.951 2.153 RHE2, VIS1.5 ONK1 0.259 0.954 2.113 RAD1, VIS1.5 ONK1 0.255 0.953 2.111 BEW1, RAO1 ONK1 0.253 0.953 2.110 RAO1, VIS1 ONK1 0.252 0.953 2.110 HAE3, RAO1 ONK1 0.250 0.953 2.109 BEW7.2.1 BEW7.2 0.297 1.000 2.064 BEW7.1.1, BEW7.2.1 BEW7.2 0.258 1.000 2.064 BEW7.1, BEW7.2.1 BEW7.2 0.295 1.000 2.064 BEW6, BEW7.2.1 BEW7.2 0.296 1.000 2.064 BEW5, BEW7.2.1 BEW7.2 0.295 1.000 2.064 Note: Top 50 rules listed as per defined parameters (see methodology), sorted top down by lift; 16,445 rules identified 19 Table 4: SPLG combinations linked with profitability Antecedents Consequent Support Confidence Lift AUG1.2, BEW10, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, BEW7, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, GEF2, HNO1.2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, BEW3, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, GEF2, VIS1 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, GAE1, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, DER1, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, GEF2, GYN1 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, BEW6, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, GEF2, URO1 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, BEW5, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, BEW2, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW1, BEW10, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, BP, GEF2 Profitability 0.051 0.935 1.615 AUG1.2, BEW10, GEF2, GYN1.3 Profitability 0.051 0.934 1.613 AUG1.2, BEW10, GEF2, HNO1.1 Profitability 0.051 0.934 1.613 AUG1.2, BEW10, GEF2, URO1.1 Profitability 0.051 0.934 1.613 AUG1.2, BEW10, GYN1.4, HNO1.2 Profitability 0.054 0.924 1.595 AUG1.2, BEW10, DER1, GYN1.4 Profitability 0.052 0.922 1.591 AUG1.2, BEW10, BEW8.1, GYN1.4 Profitability 0.051 0.919 1.587 AUG1.2, BEW10, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, BEW7, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, BEW3, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, GYN1, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, BEW6, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, BEW5, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, BEW2, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW1, BEW10, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, BP, GYN1.4 Profitability 0.055 0.912 1.574 AUG1.2, BEW10, GYN1.4, GYN2 Profitability 0.054 0.910 1.571 AUG1.2, BEW10, GYN1.4, URO1 Profitability 0.054 0.910 1.571 AUG1.2, BEW10, GYN1.3, GYN1.4 Profitability 0.053 0.909 1.569 AUG1.2, BEW10, GYN1.4, HNO1 Profitability 0.053 0.909 1.569 AUG1.2, BEW10, GYN1.4, HNO1.1 Profitability 0.052 0.908 1.567 AUG1.2, BEW10, GYN1.4, URO1.1 Profitability 0.052 0.908 1.567 AUG1.2, BEW10, GAE1, GYN1.4 Profitability 0.052 0.908 1.567 AUG1.2, BEW10, BEW8, GYN1.4 Profitability 0.051 0.906 1.564 AUG1.2, BEW10, GEB1, GYN1.4 Profitability 0.051 0.906 1.564 AUG1.2, BEW10, GYN1.4, VIS1 Profitability 0.051 0.906 1.564 AUG1.2, BEW10, BEW8.1, GYN1.3 Profitability 0.051 0.906 1.564 AUG1.2, BEW10, BEW3, GYN1.3 Profitability 0.057 0.901 1.556 AUG1.2, BEW10, GYN1.3, HNO1.2 Profitability 0.056 0.900 1.553 AUG1.2, BEW10, GYN1.3, URO1.1 Profitability 0.055 0.899 1.551 AUG1.2, BEW10, GYN1.3, HNO1.1 Profitability 0.054 0.897 1.548 AUG1.2, BEW10, GYN1.3, PNE1 Profitability 0.051 0.892 1.540 GEF2, HNO1.2, KAR1.1, VIS1.2 Profitability 0.051 0.891 1.537 GEF2, HAE1, HNO1.2, VIS1.2 Profitability 0.051 0.891 1.537 AUG1.2, BEW10, BEW2, GYN1.3 Profitability 0.057 0.889 1.534 AUG1.2, BEW10, DER1, GYN1.3 Profitability 0.055 0.886 1.529 GEF2, HNO1.2, VIS1.2 Profitability 0.052 0.881 1.520 Note: Top 50 rules listed as per defined parameters (see methodology), sorted top down by lift; 187,463 rules identified 20 For private hospitals, the SPLGs that are most strongly linked with profitability (support: 0.052, confidence: 1.0, lift: 1.617) are AUG1.2 and GEF2 (see Table 9). On an aggregated level, diverse combinations with the SPLGs ANG2 (intra-abdominal vascular interventions) and GYN1.2 (malignant neoplasms of the cervix), or ANG2 and GEF2, or GYN1.4 (malignant neoplasms of the ovary) and GEF2 are associated with profitability for this subgroup (see Figure 11). For public hospitals, the SPLGs that are most strongly linked with profitability (support: 0.053, confidence: 1.0, lift: 2.050) were AUG1.3 (specialized anterior segment surgery), HER1.1.3 (surgery and interventions on the thoracic aorta), and URO1.1.6 (plastic reconstruction of the urethra) (see Table 10). On an aggregated level, diverse combinations with the SPLGs HER1.1.3 and AUG1.3 or AUG1.3 and NCH1 (cranial neurosurgery) or AUG1.3 and HAE1.1 are associated with profitability for this subgroup (see Figure 12). Additionally, the Apriori algorithm identified SPLG combinations that are linked with losses. For all hospitals, the SPLGs that are most strongly linked with losses (support: 0.054, confidence: 0.701, lift: 1.667) are UNF1 (trauma surgery polytrauma), UNF2 (severe burns), GEFA, and BEW7.1.1 (hip prosthesis replacement operations) (see Table 11). Table 5: SPLG combinations linked with losses Antecedents Consequent Support Confidence Lift UNF1, UNF2, GEFA, BEW7.1.1 Losses 0.054 0.701 1.667 UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 UNF2, GEFA, GYNT, BEW7.1.1 Losses 0.056 0.700 1.664 UNF2, BEW7.1, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 HAE1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 NEU3, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 NEU2, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 GYN2, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 HAE3, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 GEB1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 RHE1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 HAE2, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 PNE1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 UNF2, VIS1, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 GAE1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 DER1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 GYN1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 BEW6, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 UNF2, URO1, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 BEW5, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 BEW2, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 BEW1, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 BP, UNF2, GEFA, BEW7.1.1 Losses 0.056 0.700 1.664 Note: Top rules listed as per defined parameters (see methodology), sorted top down by lift; 23 rules identified 21 On an aggregated level, combinations of the two SPLGs UNF1 and UNF2, or UNF2 and GEFA, are associated with losses (see Figure 13). For private hospitals, no rule has been identified that links SPLGs with losses at the pre-defined thresholds. For public hospitals, the SPLGs that are most strongly linked with losses (support: 0.053, confidence: 0.929, lift: 1.813) are AUG1.1 (strabology), VIS1.3 (esophageal resection), and BEW7.1.1 (see Table 12). On an aggregated level, diverse combinations of the SPLGs AUG1.1 and VIS1.3, or VIS1.3 and VIS1.4, or VIS1.3 and BEW10, are associated with losses (see Figure 14). Discussion This study aims to investigate whether there is a systematic imbalance in the reimbursement of inpatient hospital cases, specifically between profitable and unprofitable service groups/ areas. In general, the hospital profitability results highlight a worsening situation, as the net profitability of Swiss hospitals declined continuously from 2015 to 2022, with many hospitals reporting losses. While the profitability dip in 2020 can be attributed to the COVID-19 pandemic, the general downward trend is unambiguous (43). Accordingly, the hospital profitability results are in line with the widely discussed deteriorating financial situation of Swiss hospitals (10, 11, 17). Private hospitals, which exhibit distinct structures and are primarily specialty hospitals for surgery, constantly reported higher net profitability margins than public hospitals. Specifically, half of these hospitals still reported a small positive net profit in 2022, while almost 75% of public hospitals reported a loss, despite receiving on average 67,055,126 CHF in subsidies (e.g., for teaching) in addition to the S-DRG reimbursements, which are reflected in the net profitability. The finding that public hospitals are financially in a worse situation than private hospitals is also in line with the public debate (15, 44). Regarding research question one, whether there are significant positive or negative associations between SPLBs and profitability, some SPLBs have been identified with significant associations, although the results differ in part depending on the sub-analysis. The primary causal fixed effects panel regression identified the following SPLBs as significantly positively associated with hospital profitability: heart, gynecology, and ear, nose, and throat. Across the various subgroup analyses, the SPLB heart and visceral surgery were predominantly associated with hospital profitability, and the SPLB gynecology and ear, nose 28 Appendix Figure 5: Distribution of institutions and cases per legal entity form Note: Exemplary for the year 2022; Definitions from Swiss Federal Office of Public Health: Private organization (R1), Association/ foundation (R2), Sole proprietorship (R3), Public organization (R4). Figure 6: Hospital net profitability margin from 2015 to 2022 Note: Margin (%) referring to net profit margin(excluding deficit coverage) divided by hospital revenue; The whiskers in the boxplots represent values within 1.5 times the interquartile range (IQR), extending from the first quartile (Q1) to the third quartile (Q3). Values outside this range (potential outliers) are excluded from the visualization. -10 -5 0 5 10 Margin (%) 2015 2016 2017 2018 2019 2020 2021 2022 excludes outside values 29 Figure 7: Fixed effects panel regression coefficients (incl. transplants) Figure 8: SPLG network diagram for private hospitals Note: SPLG network diagram based on top 100 rules based on highest lift. Basic package Dermatology, Hematology, Radio oncology Ear, nose and throat Endocrinology Gastroenterology Gynecology Heart Nephrology Neurology Neurosurgery Obstetrics Ophthalmology Pneumology Rheumatology Severe injuries Surgical musculoskeletal system Thoracic surgery Transplants Urology Vascular Visceral surgery -600000 -400000 -200000 0 200000 30 Figure 9: SPLG network diagram for public hospitals Note: SPLG network diagram based on top 100 rules based on highest lift. Figure 10: SPLG combinations linked with profitability 31 Figure 11: SPLG combinations linked with profitability for private hospitals Figure 12: SPLG combinations linked with profitability for public hospitals 32 Figure 13: SPLG combinations linked with losses Figure 14: SPLG combinations linked with losses for public hospitals 33 Table 6: Fixed-effects panel regression analysis of relative hospital profitability margin with cluster-robust standard errors All hospitals Private hospitals Public hospitals Absolute SPLB case numbers SPLB offered by hospitalb Absolute SPLB case numbers SPLB offered by hospitalb Absolute SPLB case numbers SPLB offered by hospitalb SPLB Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Coefficient (Standard Error) P-value Basic package 0.001 (0.000) n.a.c 0.001 (0.001) n.a.c 0.000 (0.001) n.a.c Dermatology/ Hematology/ Radio oncologya -0.001 (0.003) -2.493 (1.091) ** -0.001 (0.004) -2.926 (1.333) ** 0.001 (0.006) 7.184 (4.611) Ear, nose and throat 0.004 (0.001) *** -0.186 (0.932) 0.003 (0.002) ** 0.841 (0.941) 0.009 (0.006) -7.499 (4.268) * Endocrinology 0.005 (0.005) 0.482 (0.632) -0.002 (0.008) -0.827 (0.972) 0.004 (0.010) 0.768 (1.444) Gastroenterology 0.004 (0.004) 1.305 (0.732) * -0.004 (0.004) 2.025 (1.061) * 0.005 (0.006) -8.321 (3.911) ** Gynecology 0.004 (0.002) * -0.597 (0.776) 0.006 (0.005) -0.244 (1.161) 0.010 (0.005) ** -2.841 (1.402) * Heart 0.001 (0.001) ** 0.739 (0.658) 0.001 (0.001) 0.239 (0.882) 0.001 (0.001) 1.015 (1.818) Nephrology -0.009 (0.009) -0.152 (1.389) 0.006 (0.018) 0.419 (0.855) -0.019 (0.017) -1.819 (5.096) Neurology 0.000 (0.003) 0.060 (0.737) 0.006 (0.006) -0.204 (1.129) -0.004 (0.004) 31.886 (9.087) *** Neurosurgery 0.002 (0.006) 0.305 (0.343) -0.003 (0.008) 0.563 (0.459) 0.018 (0.015) 0.788 (1.479) Obstetrics 0.001 (0.002) -1.565 (1.747) 0.000 (0.003) -1.212 (2.174) 0.003 (0.006) n.a.c Ophthalmology -0.001 (0.006) 0.391 (0.545) 0.001 (0.007) -0.469 (0.911) -0.009 (0.013) 0.953 (0.884) Pneumology -0.004 (0.003) 0.402 (0.769) 0.000 (0.002) 0.808 (1.157) -0.015 (0.009) -5.929 (1.943) *** Rheumatology 0.000 (0.011) -0.235 (0.754) 0.001 (0.014) 0.656 (1.068) -0.012 (0.025) -16.103 (4.675) *** Severe injuries -0.008 (0.010) 0.360 (0.555) -0.024 (0.016) 0.666 (0.745) 0.010 (0.024) -2.151 (1.384) Surgical musculoskeletal system 0.001 (0.001) -0.610 (0.966) 0.001 (0.001) -0.743 (1.624) 0.003 (0.003) n.a.c Thoracic surgery 0.000 (0.010) 0.562 (0.579) 0.005 (0.016) 1.159 (0.743) -0.005 (0.021) 1.014 (1.052) Transplants -0.030 (0.020) -1.249 (0.999) -0.060 (0.052) n.a.c -0.041 (0.051) n.a.c Urology 0.000 (0.002) 2.316 (1.932) 0.000 (0.004) 2.278 (1.923) -0.002 (0.003) n.a.c Vascular -0.003 (0.003) 0.189 (0.796) -0.008 (0.004) ** -0.760 (0.834) 0.004 (0.010) 3.668 (2.228) Visceral surgery -0.006 (0.005) 2.291 (1.548) 0.002 (0.006) 3.588 (2.105) * -0.008 (0.010) 6.857 (3.229) ** Confounders: Year, bed occupancy rate, share of private patients, share of inpatient revenue ***) p<0.01; **) p<0.05; *) p<0.1 Note: Coefficient provided in percentage points; a) The SPLBs Dermatology, Hematology, and Radio oncology were merged into the combined category Dermatology/Hematology/Radio oncology. This was necessary because the three SPLBs exhibited very high correlations, leading to problematic multicollinearity (VIF > 10). By merging the most strongly correlated SPLBs, the multicollinearity was reduced to acceptable levels (VIF < 10), fulfilling the statistical requirements for the analysis. b) The independent variables are binary, with “1” indicating that a hospital is relevant within the SPLG and “0“ indicating not relevant. A hospital was classified as relevant if it was among the providers that collectively treated at least 97.5% of all cases within each SPLG, excluding providers with only minimal caseloads. If one SPLG was classified as relevant, the SPLB was considered relevant. c) „n.a.“ referring to independent variables that STATA has omitted because of collinearity. 34 Table 7: Typical SPLG combinations for private hospitals Note: Top 50 rules listed as per defined parameters (see methodology), sorted top down by lift; 8,416 rules identified. Antecedents Consequent Support Confidence Lift KAR1, NEU1 GAE1.1 0.280 0.964 2.488 KAR1, NEU3 GAE1.1 0.279 0.959 2.475 KAR1, NEU2.1 GAE1.1 0.253 0.955 2.465 DER1, VIS1.5 GAE1.1 0.259 0.951 2.454 KAR1.1, NEU1 NEU3 0.268 0.989 2.141 KAR1, NEU1 NEU3 0.286 0.985 2.131 NEU1, UNF1 NEU3 0.258 0.983 2.128 HNO1, URO1.1.3 URO1.1 0.253 0.988 2.119 KAR1, NEU2.1 NEU3 0.259 0.978 2.116 NEU3, UNF1 NEU1 0.258 0.994 2.111 NEU1, UNF1.1 NEU3 0.286 0.975 2.110 URO1.1.3, VIS1 URO1.1 0.265 0.983 2.108 GAE1, URO1.1.3 URO1.1 0.264 0.983 2.108 DER1, URO1.1.3 URO1.1 0.261 0.983 2.108 NEU2, UNF1.1 NEU3 0.271 0.973 2.107 HNO1.1, NEU1 NEU3 0.271 0.973 2.107 HAE2, URO1.1.3 URO1.1 0.253 0.983 2.107 URO1.1.1, URO1.1.3 URO1.1 0.252 0.983 2.106 UNF1.1, VIS1 NEU3 0.264 0.973 2.105 NEU2.1, UNF1.1 NEU3 0.259 0.972 2.104 GYN1, UNF1 NEU3 0.253 0.971 2.103 URO1.1.1, VIS1 URO1.1 0.286 0.980 2.100 GAE1.1, NEU2.1 NEU3 0.288 0.970 2.099 DER1, URO1.1.1 URO1.1 0.277 0.979 2.099 GAE1, URO1.1.1 URO1.1 0.277 0.979 2.099 GAE1.1, UNF1.1 NEU1 0.250 0.988 2.098 HAE2, URO1.1.1 URO1.1 0.271 0.978 2.098 PNE1, URO1.1.1 URO1.1 0.264 0.978 2.096 BEW3, URO1.1.1 URO1.1 0.261 0.978 2.096 GEB1, URO1.1.1 URO1.1 0.259 0.978 2.096 NEU1, ONK1 NEU3 0.270 0.968 2.095 END1, URO1.1.1 URO1.1 0.256 0.977 2.095 HAE3, URO1.1.1 URO1.1 0.255 0.977 2.095 GAE1.1, URO1.1.1 URO1.1 0.252 0.977 2.094 KAR1.1, RHE1 NEU3 0.262 0.967 2.093 END1, UNF1 NEU3 0.256 0.966 2.092 KAR1, NEU3 NEU1 0.286 0.985 2.091 GEB1, UNF1.1 NEU3 0.253 0.966 2.091 NEP1, NEU2.1 NEU3 0.252 0.966 2.090 HNO2, NEU3 NEU1 0.282 0.984 2.090 KAR1.1, NEU3 NEU1 0.268 0.984 2.089 URO1.1.3 URO1.1 0.276 0.974 2.087 URO1, URO1.1.3 URO1.1 0.276 0.974 2.087 BP, URO1.1.3 URO1.1 0.276 0.974 2.087 HNO2, NEU1 NEU3 0.282 0.964 2.087 GAE1.1, NEU1 NEU3 0.322 0.964 2.087 BEW1, URO1.1.3 URO1.1 0.274 0.974 2.087 GYN1, UNF1 NEU1 0.256 0.983 2.087 GEB1, UNF1 NEU1 0.250 0.982 2.086 GYN1, URO1.1.3 URO1.1 0.270 0.973 2.086 35 Table 8: Typical SPLG combinations for public hospitals Note: Top 50 rules listed as per defined parameters (see methodology), sorted top down by lift; 151,372 rules identified. Antecedents Consequent Support Confidence Lift AUG1.4, VIS1.2 AUG1.5 0.276 1.000 3.280 AUG1.4, VIS1.1 AUG1.5 0.276 0.986 3.232 ANG3, AUG1.4 AUG1.5 0.264 0.985 3.230 AUG1.2, AUG1.4 AUG1.5 0.256 0.984 3.229 ANG3, KAR1.1.1 KAR1.2 0.264 0.970 3.225 ANG3, KAR1.3 KAR1.2 0.264 0.970 3.225 AUG1.4, THO1.1 AUG1.5 0.272 0.971 3.185 AUG1, AUG1.4 AUG1.5 0.260 0.970 3.181 AUG1.4, THO1.2 AUG1.5 0.276 0.958 3.141 AUG1.4, THO1 AUG1.5 0.276 0.958 3.141 AUG1.4, URO1.1.1 AUG1.5 0.276 0.958 3.141 AUG1.4, BEW8 AUG1.5 0.276 0.958 3.141 AUG1.4, RAO1 AUG1.5 0.272 0.957 3.139 AUG1.4, URO1.1.4 AUG1.5 0.272 0.957 3.139 AUG1.4, RHE2 AUG1.5 0.272 0.957 3.139 AUG1.4, UNF1 AUG1.5 0.272 0.957 3.139 AUG1.4, URO1.1.2 AUG1.5 0.268 0.957 3.137 AUG1.4, URO1.1.8 AUG1.5 0.268 0.957 3.137 AUG1.4, DER1.1 AUG1.5 0.268 0.957 3.137 AUG1.4, HER1 AUG1.5 0.264 0.956 3.135 AUG1.4, KIE1 AUG1.5 0.260 0.955 3.133 AUG1.4, NUK1 AUG1.5 0.252 0.954 3.129 AUG1.4, NCH1.1 AUG1.5 0.252 0.954 3.129 AUG1.4, GEB1.1 AUG1.5 0.252 0.954 3.129 AUG1.4, UNF2 AUG1.5 0.252 0.954 3.129 NCH1.1, GYNT GEFA 0.252 1.000 3.114 HNO1.1.1, GYNT GEFA 0.252 1.000 3.114 THO1.2, GYNT GEFA 0.264 1.000 3.114 NCH1.1, BEW7.1 GEFA 0.252 1.000 3.114 THO1.2, BEW7.1 GEFA 0.264 1.000 3.114 AUG1.5, VIS1.2 AUG1.4 0.276 0.958 3.100 AUG1.5, URO1.1.2 AUG1.4 0.268 0.957 3.096 ANG3, AUG1.5 AUG1.4 0.264 0.956 3.094 AUG1, AUG1.4 ANG3 0.264 0.985 3.067 DER1.1, GYNT GEFA 0.268 0.971 3.022 AUG1.4, NUK1 ANG3 0.256 0.969 3.018 AUG1.4, GEB1.1 ANG3 0.256 0.969 3.018 GEB1.1, GYNT GEFA 0.252 0.969 3.017 AUG1.2, AUG1.4 ANG3 0.252 0.969 3.017 URO1.1.1, VIS1.3 ANG3 0.252 0.969 3.017 VIS1.1, VIS1.2 ANG3 0.305 0.962 2.994 AUG1, VIS1.2 ANG3 0.297 0.961 2.991 AUG1.2, VIS1.1 ANG3 0.297 0.961 2.991 AUG1.2, VIS1.2 ANG3 0.293 0.960 2.989 NUK1, VIS1.2 ANG3 0.285 0.959 2.986 GEB1.1, VIS1.2 ANG3 0.285 0.959 2.986 VIS1.2, VIS1.4 ANG3 0.272 0.957 2.980 DER1.1, BEW7.1 GEFA 0.268 0.957 2.979 AUG1.4, THO1.1 ANG3 0.268 0.957 2.979 AUG1.4, AUG1.5 ANG3 0.264 0.956 2.977 36 Table 9: SPLG combinations linked with profitability for private hospitals Note: Top 50 rules listed as per defined parameters (see methodology), sorted top down by lift; more than 1 million rules identified. Antecedents Consequent Support Confidence Lift AUG1.2, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, HNO1.1 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, URO1.1 Profitability 0.051 1.000 1.617 AUG1.2, BEW7, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, GEF2, VIS1 Profitability 0.052 1.000 1.617 AUG1.2, BEW3, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, GEF2, HNO1.2 Profitability 0.052 1.000 1.617 AUG1.2, DER1, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, GAE1, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, GEF2, GYN1 Profitability 0.052 1.000 1.617 AUG1.2, BEW6, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, GEF2, URO1 Profitability 0.052 1.000 1.617 AUG1.2, BEW5, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, BEW2, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, BEW1, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, BP, GEF2 Profitability 0.052 1.000 1.617 AUG1.2, GEF1, GYN1.3 Profitability 0.052 1.000 1.617 AUG1.2, BEW7, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, GYN1.3, VIS1 Profitability 0.051 1.000 1.617 AUG1.2, BEW3, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, GYN1.3, HNO1.2 Profitability 0.051 1.000 1.617 AUG1.2, DER1, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, GAE1, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, GYN1, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, BEW6, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, GYN1.3, URO1 Profitability 0.051 1.000 1.617 AUG1.2, BEW5, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, BEW2, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, BEW1, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, BP, GEF2, GYN1.3 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, BEW7, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, GEF2, VIS1 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, BEW3, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, GEF2, HNO1.2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, DER1, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, GAE1, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, GEF2, GYN1 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, BEW6, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, GEF2, URO1 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, BEW5, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, BEW2, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW1, BEW10, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, BEW10, BP, GEF2 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, HNO1.1, URO1.1 Profitability 0.051 1.000 1.617 AUG1.2, BEW7, GEF2, HNO1.1 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, HNO1.1, VIS1 Profitability 0.051 1.000 1.617 AUG1.2, BEW3, GEF2, HNO1.1 Profitability 0.051 1.000 1.617 AUG1.2, GEF2, HNO1.1, HNO1.2 Profitability 0.051 1.000 1.617 37 Table 10: SPLG combinations linked with profitability for public hospitals Note: Top 50 rules listed as per defined parameters (see methodology), sorted top down by lift; 641,815 rules identified. Antecedents Consequent Support Confidence Lift AUG1.3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GYN1.2, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GYN1.1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 ANG2, AUG1.3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.5, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GEF2, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GYN1.4, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GYN1.3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HAE1.1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, BEW11, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, AUG1.4, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, AUG1.5, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, KAR1.2, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, BEW9, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.6, VIS1.3 Profitability 0.053 1.000 2.050 ANG3, AUG1.3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.6, VIS1.2 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, KAR1.1.1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, NCH1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, BEW10, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.6, VIS1.4.1 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.6, VIS1.1 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, KAR1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, NUK1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1, AUG1.3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, NCH1.1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.2, AUG1.3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, THO1.1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, HNO1.1.1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, PNE1.1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.2, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, THO1.2, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GEB1.1.1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, PNE1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GEB1.1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, BEW8.1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.6, URO1.1.8 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, NCH3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, HNO1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, RAO1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, DER1.1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, THO1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GEF3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.6, VIS1.4 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.4, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, HER1.1.3, URO1.1.1, URO1.1.6 Profitability 0.053 1.000 2.050 AUG1.3, GEF1, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050 ANG1, AUG1.3, HER1.1.3, URO1.1.6 Profitability 0.053 1.000 2.050