Classification of hospital admissions into emergency and elective care: a machine learning approach

Abstract

Rising admissions from emergency departments (EDs) to hospitals are a primary concern for many healthcare systems. The issue of how to differentiate urgent admissions from non-urgent or even elective admissions is crucial. We aim to develop a model for classifying inpatient admissions based on a patient’s primary diagnosis as either emergency care or elective care and predicting urgency as a numerical value. We use supervised machine learning techniques and train the model with physician-expert judgments. Our model is accurate (96%) and has a high area under the ROC curve (>.99). We provide the first comprehensive classification and urgency categorization for inpatient emergency and elective care. This model assigns urgency values to every relevant diagnosis in the ICD catalog, and these values are easily applicable to existing hospital data. Our findings may provide a basis for policy makers to create incentives for hospitals to reduce the number of inappropriate ED admissions.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Notes

  1. 1.

    For example, the NHS introduced the “marginal rate rule” in 2010 in response to growth in the volume of patients being admitted to the hospital as emergencies. The rule sets a baseline value for income from each provider’s emergency admissions. For emergency admissions above this baseline, the provider receives only 30% of the normal price [6, 7] regardless of urgency.

  2. 2.

    If a hospital treats more cases than negotiated with the SHI, it may receive reductions in the DRG payments for additional patients [23]. Therefore, it may be beneficial to code elective admissions as emergencies since every acute care hospital in Germany is required by law to admit all emergency cases unless it has reached capacity [24].

  3. 3.

    Urgency interpreted from greatest to least harm as a consequence of deferring treatment [29].

  4. 4.

    ICD and OPS codes were harmonized to the same version over the years using official mapping tables by DIMDI (German Institute of Medical Documentation and Information).

  5. 5.

    The operating definition of interaction is that variables m and k interact if a split on one variable, say m, in a tree makes a split on k either systematically less possible or more possible.

  6. 6.

    We have training data from all ICD chapters except for III (diseases of the blood and blood-forming organs and certain disorders involving the immune system), V (mental and behavioral disorders) and XX (external causes of morbidity and mortality), which together make up less than 5% of cases. We exclude chapter XV (pregnancy, childbirth and puerperium).

  7. 7.

    The 50 diagnoses in Tables 2 and 3 represent approximately 1/4 of Germany’s market volume. Top 25 diagnoses in number of cases per decile stratified urgency.

  8. 8.

    Additionally, superficial partial thickness (2nd degree a) has a lower urgency then deep partial thickness (2nd degree b).

  9. 9.

    Excluding maternity and birth.

  10. 10.

    For illustration purposes, we used logarithmic values for the hospital utilization in the lower part because of the steep age gradient.

  11. 11.

    The large increase of inpatient cases with a length of stay of <48 h in Germany is in line with this observation.

References

  1. 1.

    Morse A (2013) Emergency admissions to hospital: managing the demand. The Stationery Office, National Audit Office, London

    Google Scholar 

  2. 2.

    Schuur JD, Venkatesh AK (2012) The growing role of emergency departments in hospital admissions. N Engl J Med 367(5):391–393. https://doi.org/10.1056/NEJMp1204431

    Article  Google Scholar 

  3. 3.

    Meacock R, Doran T, Sutton M (2015) What are the costs and benefits of providing comprehensive seven-day services for emergency hospital admissions? Health Econ 24(8):907–912. https://doi.org/10.1002/hec.3207

    Article  Google Scholar 

  4. 4.

    Roland M, Abel G (2012) Reducing emergency admissions: are we on the right track? BMJ 345:e6017. https://doi.org/10.1136/bmj.e6017

    Article  Google Scholar 

  5. 5.

    Wallace E, Smith SM, Fahey T, Roland M (2016) Reducing emergency admissions through community based interventions. BMJ 352:h6817. https://doi.org/10.1136/bmj.h6817

    Article  Google Scholar 

  6. 6.

    Macnabb J, Rowan L, Heneghan C, Onakpoya I (2014) How England's emergency departments are being penalised. BMJ 348:g1604. https://doi.org/10.1136/bmj.g1604

    Article  Google Scholar 

  7. 7.

    Monitor and NHS England (2013) Monitor and NHS England’s review of the marginal rate rules. Monitor and NHS England, London

    Google Scholar 

  8. 8.

    Adan IJBF, Vissers JMH (2002) Patient mix optimisation in hospital admission planning: a case study. Int J Oper Prod Manag 22(4):445–461. https://doi.org/10.1108/01443570210420430

    Article  Google Scholar 

  9. 9.

    Mackway-Jones K, Marsden J, Windle J (2013) Emergency triages. Wiley, Chichester. https://doi.org/10.1002/9781118299029

    Google Scholar 

  10. 10.

    Propper C, Sutton M, Whitnall C, Windmeijer F (2010) Incentives and targets in hospital care: evidence from a natural experiment. J Public Econ 94(3-4):318–335. https://doi.org/10.1016/j.jpubeco.2010.01.002

    Article  Google Scholar 

  11. 11.

    Wilper AP, Woolhandler S, Lasser KE, McCormick D, Cutrona SL, Bor DH, Himmelstein DU (2008) Waits to see an emergency department physician: U.S. trends and predictors, 1997-2004. Health Aff 27(2):w84–ww95. https://doi.org/10.1377/hlthaff.27.2.w84

    Article  Google Scholar 

  12. 12.

    Addison PD, Getgood A, Paterson-Brown S (2001) Separating elective and emergency surgical care (the emergency team). Scott Med J 46(2):48–50. https://doi.org/10.1177/003693300104600207

    Article  Google Scholar 

  13. 13.

    Blunt I, Bardsley M, Grove A, Clarke A (2015) Classifying emergency 30-days readmissions in England using routine hospital data 2004-2010: what is the scope for reduction? Emerg Med J 32(1):44–50. https://doi.org/10.1136/emermed-2013-202531

    Article  Google Scholar 

  14. 14.

    McCormack LA, Jones SG, Coulter SL (2016) Demographic factors influencing nonurgent emergency department utilization among a Medicaid population. Health Care Manag Sci 20(3):395–402. https://doi.org/10.1007/s10729-016-9360-8

    Article  Google Scholar 

  15. 15.

    Moskop JC, Sklar DP, Geiderman JM, Schears RM, Bookman KJ (2009) Emergency department crowding, part 2--barriers to reform and strategies to overcome them. Ann Emerg Med 53(5):612–617. https://doi.org/10.1016/j.annemergmed.2008.09.024

    Article  Google Scholar 

  16. 16.

    Pines JM, Hilton JA, Weber EJ, Alkemade AJ, Al Shabanah H, Anderson PD, Bernhard M, Bertini A, Gries A, Ferrandiz S, Kumar VA, Harjola VP, Hogan B, Madsen B, Mason S, Ohlén G, Rainer T, Rathlev N, Revue E, Richardson D, Sattarian M, Schull MJ (2011) International perspectives on emergency department crowding. Acad Emerg Med 18(12):1358–1370. https://doi.org/10.1111/j.1553-2712.2011.01235.x

    Article  Google Scholar 

  17. 17.

    Baker LC, Baker LS (1994) Excess cost of emergency department visits for nonurgent care. Health Aff 13(5):162–171. https://doi.org/10.1377/hlthaff.13.5.162

    Article  Google Scholar 

  18. 18.

    Cunningham PJ (2006) What accounts for differences in the use of hospital emergency departments across U.S. communities? Health Aff 25(5):w324–ww336. https://doi.org/10.1377/hlthaff.25.w324

    Article  Google Scholar 

  19. 19.

    Ballard DW, Price M, Fung V, Brand R, Reed ME, Fireman B, Newhouse JP, Selby JV, Hsu J (2010) Validation of an algorithm for categorizing the severity of hospital emergency department visits. Med Care 48(1):58–63. https://doi.org/10.1097/MLR.0b013e3181bd49ad

    Article  Google Scholar 

  20. 20.

    Weinick RM, Burns RM, Mehrotra A (2010) Many emergency department visits could be managed at urgent care centers and retail clinics. Health Aff 29(9):1630–1636. https://doi.org/10.1377/hlthaff.2009.0748

    Article  Google Scholar 

  21. 21.

    Young GP, Wagner MB, Kellermann AL, Ellis J, Bouley D (1996) Ambulatory visits to hospital emergency departments. Patterns and reasons for use. 24 hours in the ED study group. JAMA 276(6):460–465. https://doi.org/10.1001/jama.276.6.460

    Article  Google Scholar 

  22. 22.

    Johar M, Jones GS, Savage E (2013) Emergency admissions and elective surgery waiting times. Health Econ 22(6):749–756. https://doi.org/10.1002/hec.2849

    Article  Google Scholar 

  23. 23.

    Busse R, Geissler A, Quentin W (2011) Diagnosis-related groups in Europe: moving towards transparency, efficiency and quality in hospitals. McGraw-Hill Education, Maidenhead

    Google Scholar 

  24. 24.

    Schwierz C, Augurzky B, Focke A, Wasem J (2012) Demand, selection and patient outcomes in German acute care hospitals. Health Econ 21(3):209–221. https://doi.org/10.1002/hec.1706

    Article  Google Scholar 

  25. 25.

    Schreyögg J, Bäuml M, Krämer J, Dette T, Busse R, Geissler A (2014) Forschungsauftrag zur Mengenentwicklung nach.§ 17 b Abs. 9 KHG. Endbericht Juli 2014s, Hamburg

  26. 26.

    Alessandrini EA, Alpern ER, Chamberlain JM, Shea JA, Gorelick MH (2010) A new diagnosis grouping system for child emergency department visits. Acad Emerg Med 17(2):204–213. https://doi.org/10.1111/j.1553-2712.2009.00635.x

    Article  Google Scholar 

  27. 27.

    Gale SC, Shafi S, Dombrovskiy VY, Arumugam D, Crystal JS (2014) The public health burden of emergency general surgery in the United States: a 10-years analysis of the nationwide inpatient sample--2001 to 2010. J Trauma Acute Care Surg 77(2):202–208. https://doi.org/10.1097/TA.0000000000000362

    Article  Google Scholar 

  28. 28.

    Unger B, Afilalo M, Boivin JF, Bullard M, Grafstein E, Schull M, Lang E, Colacone A, Soucy N, Xue X, Segal E (2010) Development of the Canadian emergency department diagnosis shortlist. CJEM 12(04):311–319. https://doi.org/10.1017/S1481803500012392

    Article  Google Scholar 

  29. 29.

    Green SM, Roback MG, Miner JR, Burton JH, Krauss B (2007) Fasting and emergency department procedural sedation and analgesia: a consensus-based clinical practice advisory. Ann Emerg Med 49(4):454–461. https://doi.org/10.1016/j.annemergmed.2006.08.017

    Article  Google Scholar 

  30. 30.

    Herwartz H, Strumann C (2012) On the effect of prospective payment on local hospital competition in Germany. Health Care Manag Sci 15(1):48–62. https://doi.org/10.1007/s10729-011-9180-9

    Article  Google Scholar 

  31. 31.

    Huang XM (1995) A planning model for requirement of emergency beds. IMA J Math Appl Med Biol 12(3-4):345–353. https://doi.org/10.1093/imammb/12.3-4.345

    Article  Google Scholar 

  32. 32.

    Hughes D, McGuire A (2003) Stochastic demand, production responses and hospital costs. J Health Econ 22(6):999–1010. https://doi.org/10.1016/S0167-6296(03)00048-1

    Article  Google Scholar 

  33. 33.

    Hof S, Fügener A, Schoenfelder J, Brunner JO (2015) Case mix planning in hospitals: a review and future agenda. Health Care Manag Sci 20(2):1–14. https://doi.org/10.1007/s10729-015-9342-2

    Article  Google Scholar 

  34. 34.

    Clemens J, Gottlieb JD (2014) Do physicians' financial incentives affect medical treatment and patient health? Am Econ Rev 104(4):1320–1349. https://doi.org/10.1257/aer.104.4.1320

    Article  Google Scholar 

  35. 35.

    Dolton P, Pathania V (2016) Can increased primary care access reduce demand for emergency care? Evidence from England's 7-days GP opening. J Health Econ 49:193–208. https://doi.org/10.1016/j.jhealeco.2016.05.002

    Article  Google Scholar 

  36. 36.

    Obermeyer Z, Emanuel EJ (2016) Predicting the future - big data, machine learning, and clinical medicine. N Engl J Med 375(13):1216–1219. https://doi.org/10.1056/NEJMp1606181

    Article  Google Scholar 

  37. 37.

    Breiman L (2001) Random forests. Mach Learn 45(1):5–32. https://doi.org/10.1023/A:1010933404324

    Article  Google Scholar 

  38. 38.

    Dietterich TG (2000) An experimental comparison of three methods for constructing ensembles of decision trees: bagging, boosting, and randomization. Mach Learn 40(2):139–157. https://doi.org/10.1023/A:1007607513941

    Article  Google Scholar 

  39. 39.

    Liaw A, Wiener M (2002) Classification and regression by random forest. R News 2:18–22

    Google Scholar 

  40. 40.

    Malley JD, Kruppa J, Dasgupta A, Malley KG, Ziegler A (2012) Probability machines: consistent probability estimation using nonparametric learning machines. Methods Inf Med 51(1):74–81. https://doi.org/10.3414/ME00-01-0052

    Article  Google Scholar 

  41. 41.

    Strobl C, Malley J, Tutz G (2009) An introduction to recursive partitioning: rationale, application, and characteristics of classification and regression trees, bagging, and random forests. Psychol Methods 14(4):323–348. https://doi.org/10.1037/a0016973

    Article  Google Scholar 

  42. 42.

    Fernandez-Delgado M, Cernadas E, Barro S, Amorim D (2014) Do we need hundreds of classifiers to solve real world classification problems? J Mach Learn Res 15:3133–3181

    Google Scholar 

  43. 43.

    Díaz-Uriarte R, Alvarez de Andrés SA (2006) Gene selection and classification of microarray data using random forest. BMC Bioinformatics 7(1):3. https://doi.org/10.1186/1471-2105-7-3

    Article  Google Scholar 

  44. 44.

    Billings J, Parikh N, Mijanovich T (2000) Emergency department use in New York City: a survey of Bronx patients. Issue Brief 435:1–5

    Google Scholar 

  45. 45.

    Smulowitz PB, Lipton R, Wharam JF, Adelman L, Weiner SG, Burke L, Baugh CW, Schuur JD, Liu SW, McGrath ME, Liu B, Sayah A, Burke MC, Pope JH, Landon BE (2011) Emergency department utilization after the implementation of Massachusetts health reform. Ann Emerg Med 58(3):225–234. https://doi.org/10.1016/j.annemergmed.2011.02.020

    Article  Google Scholar 

  46. 46.

    Taubman SL, Allen HL, Wright BJ, Baicker K, Finkelstein AN (2014) Medicaid increases emergency-department use: evidence from Oregon's health insurance experiment. Science 343(6168):263–268. https://doi.org/10.1126/science.1246183

    Article  Google Scholar 

  47. 47.

    Wharam JF, Landon BE, Galbraith AA, Kleinman KP, Soumerai SB, Ross-Degnan D (2007) Emergency department use and subsequent hospitalizations among members of a high-deductible health plan. JAMA 297(10):1093–1102. https://doi.org/10.1001/jama.297.10.1093

    Article  Google Scholar 

  48. 48.

    Alessandrini EA, Alpern ER, Chamberlain JM, Shea JA, Holubkov R, Gorelick MH, Pediatric Emergency Care Applied Research Network (2012) Developing a diagnosis-based severity classification system for use in emergency medical services for children. Acad Emerg Med 19(1):70–78. https://doi.org/10.1111/j.1553-2712.2011.01250.x

    Article  Google Scholar 

  49. 49.

    Australian Institute of Health and Welfare (2012) Australian hospital statistics, 2010–11s. Australian Institute of Health and Welfare, Canberra

    Google Scholar 

  50. 50.

    Troyanskaya O, Cantor M, Sherlock G, Brown P, Hastie T, Tibshirani R, Botstein D, Altman RB (2001) Missing value estimation methods for DNA microarrays. Bioinformatics 17(6):520–525. https://doi.org/10.1093/bioinformatics/17.6.520

    Article  Google Scholar 

  51. 51.

    Hanley JA, McNeil BJ (1982) The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143(1):29–36. https://doi.org/10.1148/radiology.143.1.7063747

    Article  Google Scholar 

  52. 52.

    Clopper CJ, Pearson ES (1934) The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika 26(4):404–413. https://doi.org/10.1093/biomet/26.4.404

    Article  Google Scholar 

  53. 53.

    Bäuml M, Dette T (2014) Hospital responses to price changes: evidence from the German DRG system. Mimeo

  54. 54.

    Riessen R, Gries A, Seekamp A, Dodt C, Kumle B, Busch H-J (2015) Positionspapier für eine reform der medizinischen Notfallversorgung in Deutschen Notaufnahmen. Notfall Rettungsmed 18(3):174–185. https://doi.org/10.1007/s10049-015-0013-0

    Article  Google Scholar 

  55. 55.

    Wilson M, Cutler D (2014) Emergency department profits are likely to continue as the affordable care act expands coverage. Health Aff 33(5):792–799. https://doi.org/10.1377/hlthaff.2013.0754

    Article  Google Scholar 

  56. 56.

    Pitts SR, Pines JM, Handrigan MT, Kellermann AL (2012) National trends in emergency department occupancy, 2001 to 2008: effect of inpatient admissions versus emergency department practice intensity. Ann Emerg Med 60(6):679–686.e3. https://doi.org/10.1016/j.annemergmed.2012.05.014

    Article  Google Scholar 

  57. 57.

    OECD (2017) Hospital beds (indicator). https://doi.org/10.1787/0191328e-en

  58. 58.

    Corbella X, Ortiga B, Juan A, Ortega N, Gomez-Vaquero C, Capdevila C, Bardes I, Alonso G, Alonso G, Ferre C, Soler M, Mañez R, Jaurrieta E, Pujol R, Salazar A (2012) Alternatives to conventional hospitalization for improving lack of access to inpatient beds. A 12-years cross-sectional analysis. JHA 2(2):9–21. https://doi.org/10.5430/jha.v2n2p9

    Article  Google Scholar 

  59. 59.

    Athey S (2017) Beyond prediction: Using big data for policy problems. Science (New York, NY) 355(6324):483–485. https://doi.org/10.1126/science.aal4321

    Article  Google Scholar 

  60. 60.

    Kleinberg J, Ludwig J, Mullainathan S, Obermeyer Z (2015) Prediction policy problems. Am Econ Rev 105(5):491–495. https://doi.org/10.1257/aer.p20151023

    Article  Google Scholar 

Download references

Author information

Affiliations

Authors

Corresponding author

Correspondence to Jonas Schreyögg.

Ethics declarations

Conflict of interest

The authors have no conflicts of interest to declare.

Appendix

Appendix

Table 5 Learning data set
Table 6 Summary of predictor variables
Table 7 Mapping ICD-10-GM to ICD-10
Table 8 Shock classification table
Table 9 Burn classification table
Table 10 Diabetes classification table
Fig. 6
figure6

Changes in emergency and elective care from 2005 to 2013, classification result (Class.) and administrative type of admission (Admin)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Krämer, J., Schreyögg, J. & Busse, R. Classification of hospital admissions into emergency and elective care: a machine learning approach. Health Care Manag Sci 22, 85–105 (2019). https://doi.org/10.1007/s10729-017-9423-5

Download citation

Keywords

  • Emergency care
  • Elective care
  • Hospital
  • Machine learning
  • Classification
  • Random forest