1. Performance management using ABF (or managing performance under ABF) Stephen Duckett ABF Workshop

2. 2 Why did you come to this workshop? What are the three things you want to know more about as a result of attending this workshop?

3. 3 How did DRGs arise? Aim: to identify the abnormal for utilisation review Bob Fetter (engineer, married to Audrey Fetter, hospital manager) Reframed: to identify the abnormal, one first needs to identify the normal, then the abnormal is something which is different (statistically) from that What is the normal? Answer: groups of patients which are similar to each other What do you mean by ‘groups of patients’? Answer: groups of patients who have a similar pattern of care

4. The background: Political and economic context of activity based funding

5. 5 Real growth Growth if expenditure a constant % of GDP WelfareHealthEdu’nDefenceInfra-structureAgeing, comm & disability services GovernmentOther Health is the fastest growing segment of government expenditure (2002-3 to 2012-13) $B

6. 6 Hospitals are the fastest growing segment of health expenditure (2002-3 to 2012-13) $B

7. 7 There is significant within-state variation in public hospital costs (2010-11 data) Note: Some small hospitals (total admissions < 4,000 p.a.) not shown NSW VIC QLD WA SA TAS ACT NT Unexplained costs above the lowest level in each state ($ per admission) Hospitals with the lowest unexplained cost in each state = 0 Hospitals Average level of unexplained costs Avoidable costs unexplained costs above the average level

8. 8 ACT TAS QLD WA SA NSW VIC RangeMedian There is huge variation in the cost of treatments, e.g. gall bladder removal … Cost of laparoscopic cholecystectomy (gall bladder removal), unadjusted, 2010-11 $2,000$4,000$6,000$8,000$10,000 Note: H08B, the less complicated DRG category for the procedure

9. 9 ACT TAS QLD SA WA VIC NSW RangeMedian Cost of hip replacement, unadjusted, 2010-11 $0$10,000$20,000$30,000$40,000 Note: I03B, the less complicated DRG category for the procedure … and hip replacements

10. 10 Where are we at with national health reform (ABF emphasis) 2014/15 2015/16 and 2016/17 2017/18 to 2019/20 Commonwealth funds 45% of growth in activity (deemed to occur at National Efficient Price) andr costs (ditto, sort of) Ditto More activity measures Commonwealth share to increase to 50% of growth in activity/costs Commonwealth funds 45% capped at 6.5%

11. 11 COAG Outcomes 1 April 2016 Continued activity based funding from Commonwealth Capped in each state at 6.5% per annum Increase in public hospital expenditure (current $$)

12. 12 COAG outcomes 1 April 2016 10. The Parties, in conjunction with the Australian Commission on Safety and Quality in Health Care (ACSQHC) and the IHPA, will develop a comprehensive and risk-adjusted model to integrate quality and safety into hospital pricing and funding. The model will determine how funding and pricing can be used to improve patient outcomes and reduce the amount that should be paid for specified adverse events, ineffective interventions, or procedures known to be harmful. This could include an adjustment to the amount the Commonwealth contributes to public hospitals for a set of agreed hospital acquired conditions. Any downward adjustment to an individual state would not be deducted from the available pool of funding under the overall cap of 6.5 per cent. 11. The Parties agree to develop the model for implementation by 1 July 2017.

13. 13 Image area Hospital acquired condition list developed for payment purposes: Section 5001(c) of Deficit Reduction Act requires the Secretary to identify conditions that are: (a) high cost or high volume or both, (b) result in the assignment of a case to a DRG that has a higher payment when present as a secondary diagnosis, and (c) could reasonably have been prevented through the application of evidence- based guidelines.

14. 14 HACs which result in DRG re- assignment (2015) Foreign Object Retained After Surgery Air Embolism Blood Incompatibility Stage III and IV Pressure Ulcers Falls and Trauma (Fractures, Dislocations, Intracranial Injuries, Crushing Injuries, Burn, Other Injuries) Manifestations of Poor Glycemic Control (Diabetic Ketoacidosis, Nonketotic Hyperosmolar Coma, Hypoglycemic Coma, Secondary Diabetes with Ketoacidosis, Secondary Diabetes with Hyperosmolarity) Catheter-Associated Urinary Tract Infection (UTI) Vascular Catheter-Associated Infection Surgical Site Infection Following CABG, cardiac implantable device, Bariatric Surgery or Certain Orthopaedic Procedures Deep Vein Thrombosis (DVT)/Pulmonary Embolism (PE) Following Certain Orthopedic Procedures (Total Knee Replacement, Hip Replacement) Iatrogenic Pneumothorax with Venous Catheterization https://www.cms.gov/medicare/medicare- fee-for-service- payment/hospitalacqcond/hospital- acquired_conditions.html

15. 15 Total incidence of CHADx by major class (Source: VAED for FY 2014-15) Major class All Public HospitalsAll Private Hospitals All Victorian Hospitals 01: Post-procedural complications34,10617,80851,914 02: Adverse drug events 14,8586,40221,260 03: Accidental injuries6,0782,1798,257 04: Infections 12,8462,69415,540 05: Cardiovascular complications47,30417,98465,288 06: Respiratory complications23,4998,73732,236 07: Gastrointestinal complications36,81519,11855,933 08: Skin conditions 18,1967,50925,705 09: Genitourinary complications27,5759,75337,328 10: Hospital-acquired psychiatric states16,9595,93422,893 11: Early pregnancy complications2,7107573,467 12: Labour & delivery complications 76,05020,60096,650 13: Perinatal complications40,4584,42444,882 14: Haematological complications12,9943,97016,964 15: Metabolic complications 45,53610,74356,279 16: Nervous system complications4,2451,4295,674 17: Other complications 40,53517,56358,098 Total460,764157,604618,368 ACSQHC ‘Priority complications’Public Hospitals Private Hospitals Pressure injury5,3561,605 Falls with Fracture or ICI362127 Healthcare Assoc Infection16,5975,587 Surgical complications2,5631,099 Respiratory complications2,846554 Venous Thromboembolism1,098429 Renal failure30952 GI bleeding2,099617 Medication complications2,017455 Delirium7,1162,588 Incontinence1,246415 Malnutrition1,564482 Cardiac complications9,8434,194 Iatrogenic pneumothorax requiring intercostal catherer23074 Total count for all major categories 53,246 18,278

16. 16 COAG outcomes 1 April 2016 12. The Parties agree to work together to reduce avoidable readmissions to hospital within 28 days of discharge, with a particular focus on avoidable readmissions within 5 days of discharge, for conditions arising from complications of the management of the original condition that were the reason for the patient’s original hospital stay. 13. The Parties, in conjunction with the ACSQHC and the IHPA, will develop a comprehensive and risk-adjusted strategy and funding model that will adjust the funding to hospitals that exceed a predetermined avoidable readmission rate for agreed conditions and the circumstances in which they occur by 1 July 2017. Hospital responsibility Not hospital responsibility 3 minutes after discharge 3 years after discharge

17. 17 Readmissions good or bad indicator? We found that the validity of readmission rates as a quality indicator is influenced by the clinical process that is assessed, the indicator definition, the extend of case-mix correction, the effect of competing outcomes and the data reliability. Ignoring or poorly handling these aspects may lead to a biased estimation of the overall readmission rate and a biased comparison of readmission rates between hospitals. As a result of variance in handling these methodological threats, studies on the validity of readmission rates as a quality indicator reach conflicting conclusions. We conclude that given the limitations of readmission rates, they need to be used with caution as a measure of in-hospital quality, even more when used as a tool for a pay for performance scheme. Fischer, C. et al (2014) 'Is the readmission rate a valid quality indicator? A review of the evidence', PLoS ONE, 9(11),

18. 18 Progress on Activity based funding relies on alignment of three key factors Technical feasibility Management capacity Political will

19. 19 The questions of the age Should you reward less efficient hospitals by paying them more for treating patients than more efficient hospitals Should you reward poor quality by paying hospitals that have higher rates of adverse events more than hospitals with lower rates Should you reward less efficient hospitals by paying them more for inefficient management of the whole patient pathway? Should hospitals be rewarded more or less if they don’t deliver on their commitments to patients as part of informed consent? YesNo Yes No

20. 20 The questions of the age Should you reward less efficient hospitals by paying them more for treating patients than more efficient hospitals Answer: clearly not. Introduce activity based funding in Victoria in early 1990s, get to laggard states by early 2010s Technical feasibility Management capacity Political will Budget imperative in Victoria, ‘It’s time’ nationally

21. 21 The questions of the age Should you reward poor quality by paying hospitals that have higher rates of adverse events more than hospitals with lower rates Answer: No, as per COAG decision, not from 2017 Technical feasibility Management capacity Political will

22. 22 The questions of the age Should you reward less efficient hospitals by paying them more for inefficient management of the whole patient pathway? Answer: No. But technically more difficult than some of the previous issues. US just starting on this with Accountable Care Organisations. Lot of rhetoric about continuity. When will it be time to put our money where our mouth is? Typical way of phrasing this issue is ‘bundling across care pathway’ Technical feasibility Management capacity Political will

23. 23 Bundling Uncomplicated maternity care IHPA is exploring the feasibility of a bundled price for uncomplicated maternity care services, including antenatal and postnatal services and the admission for birth. Uncomplicated maternity care services are potentially amenable to bundled pricing as they follow a relatively predictable care pathway with clear starting and concluding points to episodes. They are also high volume services, meaning that small improvements in service delivery can result in significant savings to the health system. IHPA has completed a baseline review of the literature which has identified potential variation in the service delivery of different jurisdictions. The Commonwealth Clinical Practice Guidelines – Antenatal Care are nationally agreed guidelines for maternity care. They recommend seven (for subsequent pregnancies) to ten (for a first pregnancy) antenatal visits for a maternity care episode. A review of public data sources has indicated that over 86 per cent of pregnant women in South Australia had seven visits or more and 97 per cent had five or more in 2012. 2 However, approximately 15 per cent of women in the Australian Capital Territory had less than five antenatal visits in 2012.The Commonwealth Clinical Practice Guidelines – Antenatal Care This data suggests that bundled pricing for uncomplicated maternity care could potentially support the implementation of the nationally agreed guidelines. https://www.ihpa.gov.au/publications/pricing-framework-australian-public-hospital-services-2016-17

24. 24 More bundling options Stroke IHPA is exploring the feasibility of bundled pricing for stroke patients across the entire episode of care, including admitted acute, subacute and non- admitted settings. Strokes may be amenable to bundled pricing as they are common, the care episode generally lasts for a definable period of time, and high costs offer potentially significant savings to the health system. Due to differences in the severity of strokes, IHPA is considering bundled price weights which are weighted for complexity and notes that there are a range of issues involved in differentiating between stroke bundles. Joint replacement IHPA is exploring whether joint replacement (particularly for elective hip and knee replacement surgeries) is amenable to bundled pricing for care across settings. IHPA has identified joint replacement surgeries as being potentially amenable to bundled pricing as they are high volume, span multiple settings (non- admitted pre-operative assessment, admitted acute, subacute and follow up) and have a relatively predictable care pathway for most patients.

25. 25 And finally….. The concerns were that it would unduly increase financial risk for jurisdictions as any cost variance would not be reimbursed, that it is not IHPA’s role to use its pricing mechanisms to drive service improvement, that it would introduce more complexity into the National Pricing Model and that it would only cover public hospital services which does not confer significant benefit.

26. 26 The questions of the age Should you reward less efficient hospitals by paying them more for admitting patients ‘unnecessarily’? Answer: No. Technically more difficult than some of the previous issues. Technical feasibility Management capacity Political will Although it is now possible to take both Ambulatory Care and Referral Sensitive Conditions into account in setting activity targets Scoring on incorporating into funding system Scoring on incorporating into target setting

27. 27 Aligning economic and patient value perspectives Should hospitals be rewarded more or less if they don’t deliver on their commitments to patients as part of informed consent? For the 7 difficulty items: Do you have difficulty with the following activities because of your vision? (yes, very great difficulties; yes, great difficulties; yes, some difficulties; no, no difficulties) 1. Reading text in the newspaper 2. Recognizing faces of people you meet 3. Seeing prices of goods when shopping 4. Seeing to walk on uneven ground 5. Seeing to do needlework and handicraft 6. Reading text on television 7. Seeing to carry out a preferred hobby Two global assessment items 8. Do you experience that your present vision gives you difficulties in any way in your daily life? (yes, very great difficulties; yes, great difficulties; yes, some difficulties; no, no difficulties) 9. Are you satisfied or dissatisfied with your present vision? (very dissatisfied; rather dissatisfied; fairly satisfied; very satisfied) Lundström, M., and K. Pesudovs. 2009. Catquest-9SF patient outcomes questionnaire: nine-item short-form Rasch-scaled revision of the Catquest questionnaire. Journal of cataract and refractive surgery 35 (3):504-13. Either use general utility weights or person-specific weights Adjust payment to hospital (?surgeon) if expectations of sight improvement not achieved Can average over class or person-specific adjustments

28. 28 The questions of the age Should hospitals be rewarded more or less if they don’t deliver on their commitments to patients as part of informed consent? Answer: Why not put some consumer focus into the system? (May also help address issues of supplier induced demand). Could only work for a narrow range of procedures, lot of data gaps. A different approach would be to move toward normative pricing where DRG payment weights are adjusted for average expected benefit. There are still data gaps here but probably more feasible. Technical feasibility Management capacity Political will For some procedures

29. 29 What is National Health Reform? (excerpt from 2011 National Health Reform Agreement – pre 2014 Budget) 3. The Commonwealth and the States will work in partnership to implement new arrangements for a nationally unified and locally controlled health system which will a. improve patient access to services and public hospital efficiency through the use of activity based funding (ABF) based on a national efficient price; b. ensure the sustainability of funding for public hospitals by increasing the Commonwealth’s share of public hospital funding through an increased contribution to the costs of growth; c. improve the transparency of public hospital funding through a National Health Funding Pool and a nationally consistent approach to ABF;

30. 30 What is National Health Reform? d. improve standards of clinical care through the Australian Commission on Safety and Quality in Health Care (ACSQHC); e. improve performance reporting through the establishment of the National Health Performance Authority (NHPA); f. improve accountability through the Performance and Accountability Framework; g. improve local accountability and responsiveness to the needs of communities through the establishment of Local Hospital Networks and Medicare Locals; h. improve the provision of GP and primary health care services through the development of an integrated primary health care system and the establishment of Medicare Locals; and i. improve aged care and disability services by clarifying responsibility for client groups.

31. 31 New simplified system http://www.nhfb.gov.au/wp-content/uploads/2013/10/NHFB-Flowchart-cropped-e1382402016590-1024x587.jpg

32. 32 From a hospital/LHN perspective there are two main changes (in terms of national agenda) Increased transparency (of sorts – hospital level) Service category Total NHR payments YTD (ex GST) GST YTD Commonweal th NHR payments YTD (ex GST) State/territory NHR payments YTD (ex GST) ($) Acute admitted - public 364,825,6100140,992,675223,832,935 Acute admitted - private 64,726,808025,014,70739,712,101 Non-admitted 103,426,248039,970,72363,455,525 Emergency department 61,430,160023,740,66537,689,495 Admitted mental health 41,102,923015,884,88025,218,043 Sub-acute 21,371,50908,259,36113,112,149 Total ABF payments (Pool) 656,883,2580253,863,011403,020,247 Non-admitted mental health 22,900,7920 Small rural hospitals 40,523,8810 Teaching, training & research 7,278,9480 Other non-admitted services 13,987,6560 Total Block payments (SMF) 84,691,2770 Total NHR Payments 741,574,5350 NHR payments – Sydney Local Health District, March 2014 (YTD) NB State is key player

33. 33 From a hospital/LHN perspective there are two main changes (in terms of national agenda) Increased transparency (of sorts – hospital level) Changed incentives Changed incentives on Commonwealth Commonwealth now shares in costs of hospital activity growth Stronger efficiency incentives on hospitals/LHNs Because of state adoption of ABF

34. 34 What the Independent Hospital Pricing Authority does Determines National Efficient Price (and Efficient Cost for Block Funded Services) This determines the way Commonwealth funding to states is described (and what each LHN's notional share of that is) and the rate for payment of additional activity State as system managerHospital behaviour

35. 35 What are public hospital services? In patient ED Outpatient (setting independent) In line with the criteria … community mental health, physical chronic disease management and community based allied health programs considered in- scope will have all or most of the following attributes: Be closely linked to the clinical services and clinical governance structures of a public hospital (for example integrated area mental health services, step-up/step-down mental health services and crisis assessment teams); Target patients with severe disease profiles; Demonstrate regular and intensive contact with the target group (an average of eight or more service events per patient per annum); Demonstrate the operation of formal discharge protocols within the program; Demonstrate either regular enrolled patient admission to hospital or regular active interventions which have the primary purpose to prevent hospital admission.

36. 36 Some background Cost/outcomeCost/outputOutcome/output ABF is about technical efficiency

37. 37 Hospitals need to know where to look Cost/outcomeCost/output Cost/unit of service Services/outputOutcome/output

38. 38 Some ‘theory’ about activity based funding Activity-based Funding ABF Funder Population Funding Hospital Funder Area Health Authority etc Who allocates funding to hospitals on different bases Population expenditure Size of (weighted, needs adjusted) population Utilization Rate (conditions per person X admissions per condition) Casemix Services/ admission Cost/service (eg days, tests) =x x x x How successful have any entities been in managing this?

39. 39 Some ‘theory’ about activity based funding Who controls this? Hospital Population expenditure = Size of (weighted, needs adjusted) population x Utilization Rate (conditions per person X admissions per condition ) x Casemix x Cost/service (eg days, tests) x Services/ admission How much of our variation problem relates to this or cost control? Who controls this? How much of our variation problem relates to this or cost control? Different levers for two components Conditions per person: hard to influence Admissions per person: also hard to influence

40. 40 Doctors direct ≈ 80% of hospital expenditures PCSI Summer School 2012 Discharge timing Imaging & test orders Use of theatre Use of ICU Drugs Surgical prostheses Frequency of observations Equipment purchases

41. 41 Appendicectomy without complications DRG G07B Inflamed appendix $3,409 Casemix creates a common language between clinicians and managers (both resource and clinical homogeneity)

42. 42 The fundamental premises of activity based funding The overall goal of activity based funding is usually to address efficiency variation i.e. to have equitable payment so that hospitals (or other services) which do the same thing get paid the same.

43. 43 Clinical Conceptual model to analyse waste Waste in the Health Care System Administrative Transactions - Related Other Waste Cost - Ineffective Detrimental to Health Duplication of Services Inefficient Processes Expensive Inputs Errors Operational Unnecessary paperwork Noah’s Ark Lean Productivity Rework Adverse Events Wages Roles Procurement Excess Diagnostics Excess Procedures Disinvestment Source: Bentley, T., et al. (2008). "Waste in the U.S. Health Care System: A Conceptual Framework." Milbank Quarterly 86(4): 629-659

44. 44 DRGs need to be updated Year 1 Clinical practice reflected in activity and cost data Year 2 Lag in terms of coding, data collection and provision of data to funding authority Year 3 Analysis of collected data to develop new payment relativities Year 4 Year 1 clinical practice reflected in payments

45. 45 DRGs need to be updated New technologies and their impacts https://www.ihpa.gov.au/publications/impact-new-health-technology-framework

46. 46 DRGs need to be updated New technologies and their impacts

47. 47 The challenge of TT + R Direct activities – are distinct and separable activities that occur outside of an episode of care but are directed towards skills and knowledge development (in the case of teaching and training) or the generation of new knowledge (in the case of research). In the teaching and training context, direct activities may include lectures, tutorials and workshops. In the context of research, direct activities relate to the conduct of research. Indirect activities – are those ‘back office’ administrative and coordination activities undertaken by a health service that are essential to facilitate TTR. These activities may include the coordination of student placements, rotations, educational program development or negotiation with higher education providers. Embedded activities – which describe where TTR occurs in conjunction with patient care.

48. ABF challenges for managing (groups of) hospitals

49. 49 What is activity based funding? Funding varies with activity Activity based funding has two components Payment design Payment rules (alongside payment design) Central (state) health authority role shifts from allocating global budgets to allocating (potential) revenue (and monitoring and…) Service management role becomes Determining budget (given likely revenue) Managing costs to budget Managing revenue Watching adherence to the rules

50. 50 Broad choices in hospital funding 50 Determining hospital funding Top down 'Budgets' Bottom-up 'Revenue' (Markets) We are seeing a shift from top down budget allocation to top down activity (aka revenue) allocation

51. 51 Under this (National Health Reform) Agreement, the States will be responsible for: a. system management of public hospitals, including: i. establishment of the legislative basis and governance arrangements of public hospital services, including the establishment of Local Hospital Networks; ii. system-wide public hospital service planning and performance; iii. purchasing of public hospital services and monitoring of delivery of services purchased; iv. planning, funding and delivering capital; v. planning, funding (with the Commonwealth) and delivering teaching, training and research; vi. managing Local Hospital Network performance; and vii. state-wide public hospital industrial relations functions, including negotiation of enterprise bargaining agreements and establishment of remuneration and employment terms and conditions to be adopted by Local Hospital Networks; b. taking a lead role in managing public health; and c. sole management of the relationship with Local Hospital Networks to ensure a single point of accountability in each State for public hospital performance, performance management and planning. So what is the role of state as ‘system manager’?

52. 52 What are the constraints on state variability? The parties agree to contribute funding for hospitals into a single national pool There will be complete transparency and line of sight of respective contributions into the pool and from the pool through State accounts to Local Hospital Networks (LHNs), and of the basis on which the contributions are calculated. The parties agree that funding on the basis of activity from the national pool for public hospital services will flow to State accounts, and from there to LHNs, based on: a. the efficient price set by the Independent Hospital Pricing Authority; b. the volume of services provided in accordance with service agreements agreed by State governments, as the managers of the hospital system, and negotiated with LHNs; and c. State variations in payments to LHNs in that State to reflect differences in the efficiency of public hospital services within that State and other factors States will determine the amount they pay for public hospital services, and will meet the balance of the cost of delivering public hospital services, including any costs over and above the Commonwealth’s contribution. Attachment A to Heads of Agreement

53. 53 ABF is different in each state State is ‘system manager’ State is majority funder so funding, at the margin, will be determined by state State effectively treats Commonwealth funding as revenue target

54. 54 ABF is different in each state Price adjustments Teachingness Remoteness Aboriginality Recognition of multiple products Inpatient, outpatient etc (and in some cases, the activity measures used) Inlier and outlier Private patients One bucket or many? –National Weighted Activity Unit

55. 55 ABF is different in each state: Autonomy choices in top down systems 55 Control of movement between funding lines Control on movement between programs 'Global budget' Should mental health be a (one way) protected funding line? Policy choice: Fully fungible NWAUs or WIES, VACS etc?

56. 56 ‘Adjunct’ incentives: conditionality for additional activity payment (or whatever) Quality Access Prevention Coding quality and timeliness Duckett, S. (2008). Design of price incentives for adjunct policy goals in formula funding for hospitals and health services. BMC Health Services Research, 8(72).

57. 57 Choices about target setting approaches Never change Only with capital enhancement Negotiated annually -Politics -Prior year actuals -Planning ‘catchment area’ growth Take into account activity you want and activity you don’t want? -Readmissions -Ambulatory care sensitive conditions -Referral sensitive conditions

58. 58 The critical ‘system manager’ role: governance The regulatory and compliance framework Addressing gaming Ensuring appropriate financial responses -Short term responses Graded accountability -requirements to file recovery strategies -reductions in autonomy -appointed observers to board -designated management adviser -dismissal of the board and/or management

59. 59 Need to introduce/use graded governance

60. Managing performance at the hospital level: Running the diagnostics - the information tools

61. 61 Payment relativities make a difference http://www.ihpa.gov.au/internet/ihpa/publishing.nsf/Content/nep-determination-2014-15-html~appendix~appendix-f Payment relativity DOES NOT EQUAL profit/surplus

62. 62 Managing performance at the hospital level – key messages Know what the incentives on your network/hospital are Get paid for what you do (i.e. tell people correctly what you do) Manage the average (i.e. benchmarking) Manage variation

63. 63 What information tools do you have? Computerised discharge abstracts -Everyone has this -Data easy to interpret -The problem: cost is not perfectly correlated with length of stay -Especially for surgical cases -Rough rule of thumb: 1 hour in theatre = 1 day of stay ($$) -And ICU Data from clinical support systems -Pathology/radiology -Cost? Volumes √ Cost modeling data -No process of care variation Patient level costing data Round table type data

64. 64 What sort of preliminary (or final) analysis can you do with discharge abstract data? Is your length of stay different? Is the problem in one specialty? -easiest to define this to start with as MDC but computer file often has doctor ID so can get specialty Or with a group of DRGs? Or with a particular clinician? How are you going on hospital acquired diagnoses? Adverse events add 18% to cost of stay What sort of pattern problem is it?

65. 65 State average What is the pattern of the problem?

66. 66 State average What is the pattern of the problem?

67. 67 State average What is the pattern of the problem?

68. Managing cost/resource use variation - 1

69. 69 Pyramid Model of Investigation (adapted to emphasis cost variation investigation) Professional (aka individual physician variation) Process of Care Structure of resources (internal and external) Patient Case Mix Data - really recording

70. 70 Getting started Have you recorded correctly what you do? Is your diagnosis coding reflecting the diagnoses that the patient has and which affected the patient’s treatment while in hospital? NB: Recording more diagnoses doesn’t necessarily change the DRG When a patient’s treatment moves into a rehabilitation program post the acute phase, is that recorded?

71. 71 Coding makes a difference

72. 72 Different proportions of separations in lowest weight DRG

73. 73 Getting ready to look at processes of care Do you have good data analysis support Sophisticated statistical analysis is not required

74. 74 Getting ready to look at processes of care Do you have good data analysis support Sophisticated statistical analysis is not required Visual portrayal, good descriptions ‘inter-ocular traumatic test’

75. 75 Getting ready to look at processes of care Do you have good data analysis support –Sophisticated statistical analysis is not required –Visual portrayal, good descriptions –‘inter-ocular traumatic test’ What type of data do you have access to? –Computerised discharge data  –Cost of intermediate products (lab, DI, patient dependency) –Patient level costing data –More on this later In either case, do the data get fed-back in a sensible form to the right people/groups

76. 76 Getting ready to look at processes of care Is your organisational structure appropriate? -Do managers have budgets? -Do managers have budgets based on activity? -Do managers have budgets which are somewhat related to things they can control? -Do you have the right mechanisms for medical involvement? Do you have processes to review adoption of new technologies? Are your links with other (external) services working?

77. Managing cost/resource use variation - 2

78. 78 A E D C B Costs aren’t driven by scale Cost of gall bladder removal, unadjusted, five high-volume hospitals, 2010-11 Note: H08B, the less complicated DRG category for the procedure

79. 79 A E D C B What is the major management issue? Cost of gall bladder removal, unadjusted, five high-volume hospitals, 2010-11 Note: H08B, the less complicated DRG category for the procedure Groups A-E How can the hospital improve? Management of outliers Management of clinical pathway

80. 80 Prevention and primary care Pre- admission Hospital stay Post- discharge Re-admission Utilisation and management review needs to focus on whole continuum of care Can analyse from cost or stay perspective

81. 81 Length of stay reduction: the good, the bad and the ugly The good Patient preferences Clinical change Benefits of early ambulation Better home care support Greater technical efficiency The bad Potential clinical risk (readmissions) Cost shifting Also unbundling of care Create care discontinuity Avoid (for time being at least) addressing difficult management issues associated with improving technical efficiency

82. 82 If length of stay information is all you’ve got, then need to use what you have, but Remember: length of stay is only a surrogate for the cost of care Inaccurate (and potentially distorting) measurement of the inputs of care Distortion of the outputs of care

83. 83 Prevention and primary care Pre- admission Hospital stay Post- discharge Re-admission Utilisation and management review needs to focus on whole continuum of care Can analyse from cost or stay perspective

84. 84 Utilisation and management review needs to focus on whole continuum of care Prevention and primary care: Is this admission necessary? (i.e. is the utilisation rate too high?) Is hospital in the home an appropriate alternative? Could earlier prevention reduce the likelihood of admission? Ambulatory care sensitive conditions –Conditions for which consistent, high quality primary care (including outpatient care) is thought to be likely to reduce a patient's need for hospitalisation, an increased likelihood of hospitalisation is interpreted as evidence of impaired access to ambulatory care or receipt of poor care.

85. 85 Ambulatory sensitive conditions will reduce admissions by: Preventing the start of the illness or condition (an avoidable condition e.g. vaccine preventable diseases) Controlling an acute episodic illness or condition, or Managing a chronic disease or condition >COPD, heart disease

86. 86 Why care about prevention? Staff often want you to People with ambulatory care sensitive conditions may end up as part of a ‘long tail’ problem Reducing demand for these admissions may help meet other goals for which hospital is rewarded: >ED waits >Elective surgery capacity

87. 87 Utilisation and management review needs to focus on whole continuum of care Pre-admission services Are all the routine tests necessary? Reduce volumes hence (to some extent) costs Can these tests be done prior to admission? Reduce inpatient waits for results When is discharge planning commenced? Discharge planning at start of admission ‘Integrated plan of care’

88. 88 Review every aspect of the stay: Process of care Have we ‘Leaned-out’ the whole patient flow process? Are the theatre allocation and admission process synchronised (i.e. do patients wait for theatre?) Is theatre utilisation optimal? Has the hospital developed a common care path for this condition? Are there variations between practice patterns in the hospital?

89. 89 Is standardisation the future? + vs - vs Every patient is different Standardise and develop care paths Johnson, B Polarity management: identifying and managing unsolvable problems HRD Press 1996 Is there a right answer? vs. How do you manage the polarity?

90. 90 Review every aspect of the stay: Process of care Have we ‘Leaned-out’ the whole patient flow process? Are the theatre allocation and admission process synchronised (i.e. do patients wait for theatre?) Is theatre utilisation optimal? Has the hospital developed a common care path for this condition? Are there variations between practice patterns in the hospital? Could this condition be handled on a day-only basis? What proportion of cases is wholly treated on the appropriate specialty ward/nursing unit? Are all the tests ordered necessary? Are there delays caused by waiting for allied health or rehabilitation services? Are there delays getting access to necessary diagnostic tests or getting results back? Are there other resource co-ordination issues?

91. 91 http://www.ihi.org/knowledge/Knowledge%20Center%20Assets/IHIWhitePapers%20- %20HospitalInpatientWasteIdentificationTool_e016fc50-7128-409f-b6c0- a464e365c533/IHI%20Hospital%20Inpatient%20Waste%20Identification%20Tool%20White%20Paper%2020 11.pdf

92. 92 IHI waste identification tool

93. 93 Waste abounds in hospitals: proportion of beds with identified waste at audit Note: Only one type of waste was recorded for each bed. Resar, R. K., et al. (2011) Hospital inpatient waste identification tool, Institute for Healthcare Improvement

94. 94 Review every aspect of the stay: Process of care Have we ‘Leaned-out’ the whole patient flow process? Are the theatre allocation and admission process synchronised (i.e. do patients wait for theatre?) Is theatre utilisation optimal? Has the hospital developed a common care path for this condition? Are there variations between practice patterns in the hospital? Could this condition be handled on a day-only basis? What proportion of cases is wholly treated on the appropriate specialty ward/nursing unit? Are all the tests ordered necessary? Are there delays caused by waiting for allied health or rehabilitation services? Are there delays getting access to necessary diagnostic tests or getting results back? Are there other resource co-ordination issues? What is the pattern of adverse events in this specialty/DRG? Are there opportunities to reduce the incidence of adverse events?

95. 95 Cases with a diagnosis which arose during course of admission (working definition of ‘adverse event’ sampling frame) represent a large economic burden to the health care system Patients with at least one hospital-onset adverse event: -Stay nearly 10 days longer than other patients -Cost $6826 more per episode (controlling for DRG, age and co-morbidity) (on average $3,000) Extrapolated to entire hospital system: -Adds 18.6% to hospital expenditures -Around $2 bil pa nationally -Even if only 40% preventable $200 mil pa saved in Vic; $800 mil nationally Ehsani JE, Jackson TJ and Duckett SJ. ‘The incidence and cost of adverse events in Victorian hospitals, 2003-04’ Medical Journal of Australia, 184;11; 5 June 2006

96. 96 What is a realistic proportion of cases which should have a hospital acquired diagnosis? 159630 12

97. 97 Review every aspect of the stay: Process of care Have we ‘Leaned-out’ the whole patient flow process? Are the theatre allocation and admission process synchronised (i.e. do patients wait for theatre?) Is theatre utilisation optimal? Has the hospital developed a common care path for this condition? Are there variations between practice patterns in the hospital? Could this condition be handled on a day-only basis? What proportion of cases is wholly treated on the appropriate specialty ward/nursing unit? Are all the tests ordered necessary? Are there delays caused by waiting for allied health or rehabilitation services? Are there delays getting access to necessary diagnostic tests or getting results back? Are there other resource co-ordination issues? What is the pattern of adverse events in this specialty/DRG? Are there opportunities to reduce the incidence of adverse events? Are there systematic reasons for any ‘outliers’? (review all patients who stay > 21 days?)

98. 98 Major CHADx Groupings M CHADx 1 Post-procedural complications M CHADx 2 Adverse drug eventsM CHADx 10 Hospital-acquired psychiatric states M CHADx 3 Accidental injuriesM CHADx 11 Early pregnancy complications M CHADx 4 InfectionsM CHADx 12 Labour & delivery complications M CHADx 5 Cardiovascular complicationsM CHADx 13 Perinatal complications M CHADx 6 Respiratory complicationsM CHADx 14 Haematological complications M CHADx 7 Gastrointestinal complicationsM CHADx 15 Metabolic complications M CHADx 8 Skin conditionsM CHADx 16 Nervous system complications M CHADx 9 Genitourinary complications M CHADx 17 Other complications

99. 99 A more detailed CHADx example … MCHADx 1 Post-procedural complications 1_1 CCs of Infusion /Transfusion 1_2 Gas Embolism 1_3 Failed or Difficult Intubation 1_4 Haemorrhage & haematoma complicating a procedure 1_5 Accidental puncture/lac during proc 1_6 Foreign body or substance left following procedure 1_7 Other comps of Surgical and Medical NEC (Incl Shock T81.1) 1_8 Disruption of wound 1_9 Wound infection (Excl Septicaemia) …1_23 Post-procedural genitourinary

100. 100 Marginal cost of a complication (aka adverse event) by CHADx category

101. 101 Total system cost estimates (probably similar to hospital-level ranking)

102. 102 Pyramid Model of Investigation Professional (aka individual physician variation) Process of Care Structure of resources (internal and external) Patient Case Mix Data - really recording

103. 103 Review every aspect of the stay: Structure of resources (internal): Is there adequate theatre capacity? Are there delays caused by waiting for allied health or rehabilitation services? Are there delays getting access to necessary diagnostic tests? Does scheduling or availability of junior medical staff (interns etc) affect flow? Are care processes followed by junior medical staff (interns etc) appropriate? (e.g. test ordering behaviour) Professional: Are there variations between practice patterns in the hospital?

104. 104 Utilisation and management review needs to focus on whole continuum of care Post discharge services aka structure of resource –external Has discharge planning started early enough? What proportion of patients is waiting transfer to another hospital, or home or residential care placement? Are all patients who have had their acute care finished classified as Nursing Home Type? Would additional home and community care (or hospital in the home services) reduce length of stay or hospital costs? Is residential aged care provision adequate?

105. 105 Utilisation and management review needs to focus on whole continuum of care Readmission Consider whether discharge is occurring too early

106. 106 Why worry about readmissions? National quality indicator Staff often want you to Reducing readmissions may help meet other goals for which hospital is rewarded Re-admissions cost system money*: >16,045 admissions with a PDx of a ‘complication of surgical or medical care’ >$70.6 mil pa public expenditure on these cases * Includes admissions for adverse events from primary care and nursing homes McNair P, Borovnicar D, Jackson TJ. ‘Costs of Victorian admissions for treatment of adverse-event principal diagnoses, 2005/06’ ANZJPH. June 2010

107. 107 External issues Adverse selection -Systematic within-DRG variation >Paediatric care >Socio-economic status/indigeneity Adequate recognition of costs of teaching, training and research Adequacy of outlier policies >LOS only? >Potential to negotiate exceptional cases Where do costs fall for failure in step-down care systems (seniors, transfers) 107

108. 108

109. If you have patient level cost data

110. 110 Patient level costing data Allows more powerful analysis Is constrained by the embedded choices in how you have structured –Chart of accounts –Feeder systems –Overhead allocation –Accuracy of splits across products -Inpatient vs. outpatient vs teaching vs subacute vs … Important to distinguish: –Fixed and variable costs –Average and marginal costs Is essential (today) for managing a large hospital

111. 111 Cost concepts

112. 112 Cost concepts

113. 113 Cost concepts

114. 114 Cost concepts

115. 115 Why is this important? If price < average cost, then increasing volume will help if and only if marginal cost < price –NB: ‘price’ here is the effective price – be it base or marginal If price > average cost, then increasing volume may be a good strategy if marginal cost < price –Depending whether you want volume in this area NB: surplus should not be only criterion –Issue of stability of weights –Hospital role

116. 116 But things aren’t as simple as your costing reports would make you think

117. 117 But things aren’t as simple as your costing reports would make you think

118. 118 A tale of two hospitals

119. 119 Comparative information should be available

120. 120 The costs of DRG I03C (Hip Replacement – Cscc) Direct Ohead Total Ward Medical 1,076159 1,235 Ward Nursing 1,857353 2,210 Non-clinical 418 Pathology 14330 173 Imaging 15630 186 Allied health 28383 366 Pharmacy 18632 218 Critical Care 4812 60 Oper Rooms 2,311550 2,861 Emerg Dept 6115 76 Supplies 239215 454 Spec Proc Suites 13113 144 Prostheses 4,596 On-costs 481 Hotel360 Depreciation237 Total11,5052,57014,075

121. 121 The direct costs of DRG I03C (Hip Replacement – Cscc) Source (so you can do comparisons with your results) : http://www.health.gov.au/internet/main/publishing.nsf/ Content/+Round_9-cost-reports http://www.health.gov.au/internet/main/publishing.nsf/ Content/+Round_9-cost-reports Length of stay driven Clinician preference, purchasing policy driven Theatre, transport and surgeon efficiency, driven

122. 122 Advantage of costing systems Allow you to look at costs, not just stay Allow you to distinguish direct costs By comparing with national or state data, allow you to identify where you are aberrant and prioritise areas for examination

123. 123 The cost estimation continuum ‘Cost’ is a construct, not reality Costing is always ‘cost estimation’

124. 124 Patient costing systems allocate costs to patients and involve choices about precision of allocation of costs and identification of patients to allocate costs to NB: all costs are ‘modelled’ in some way, just some costing systems have greater use of averages/aggregates

125. 125 Evaluating comparability of data Coronary Care Unit Costs Precision of Patient Identification Precision of Cost Identification All inpatientsAll inpatients in clinical unit All inpatients in DRG All inpatients admitted CCU SeparationLEAST PRECISE RVU for separation (eg, National Service Weight) Unweighted LOS Hospital A*Hospitals B,C,D RVU-weighted LOS Hospitals E,F Unweighted shift Hospital G RVU-weighted shift Hospitals H,I** Unweighted actual minutes/costs RVU weighted actual minutes Time unit weighted by actual skill mix Actual time weighted by actual skill mix MOST PRECISE Hosp A: CCU costs allocated across all patients in cardiac-specific DRGs Hosp H: CCU costs allocated only to patients admitted to CCU based on CCU LOS

126. 126 Evaluating comparability of (inpatient) data Coronary Care Unit Costs Precision of Patient Identification Precision of Cost Identification All inpatientsAll inpatients in clinical unit All inpatients in DRG All inpatients admitted CCU SeparationLEAST PRECISE RVU for separation (eg, National Service Weight) Unweighted LOS Hospital A*Hospitals B,C,D RVU-weighted LOS Hospitals E,F Unweighted shift Hospital G RVU-weighted shift Hospitals H,I** Unweighted actual minutes/costs RVU weighted actual minutes Time unit weighted by actual skill mix Actual time weighted by actual skill mix MOST PRECISE Hosp A: CCU costs allocated across all patients in cardiac-specific DRGs Hosp H: CCU costs allocated only to patients admitted to CCU based on CCU LOS The same sort of matrix can be used to evaluate comparability of outpatient data, ED data etc

127. 127 Patient level costing systems Have embedded choices Rely on allocation of costs to a chart of accounts Which may not reflect your organisational hierarchy Report allocation of direct and overhead costs Which may not reflect reality at the current level of marginal activity Are based on feeder systems which assign costs with varying degrees of accuracy Themselves cost money

128. 128 The cost modelling approach ‘Top down’--hospital expenditures apportioned to patient groups Uses formulae to apportion joint costs Uses external relativities (RVUs) to apportion direct care costs: DRG-level or ‘service weights’ Practice pattern changes ‘frozen’ in RVUs (whenever measured) No estimates of inter-patient variability in the DRG

129. 129 Clinical or ‘activity-based’ costing Direct and overhead costs distinguished Overhead (joint) costs absorbed by direct care cost centres by ‘process costing’ (# employees, M2) ‘Feeder systems’ record patient utilisation computerised systems, eg, lab orders, nursing dependency scores by-product of clinical care Total costs estimated by ‘job order’ costing (linked to individual patient by patient/episode identifier) Hospital and department-specific product costs (local RVUs; patient acuity scales) assign costs to individual patients ‘Bottom up’ costing; DRG $ = mean of patients in DRG

130. 130 An example of patient costing using resource counts and local RVUs Patient A: grouped to DRG123 1.5 hr in emergency dept x staff salaries.75 hours in OR x $/hr staff salaries 4 days of stay x meals/day x meal cost 3 high-dependency shifts x $/shift HD ward salaries 9 low-dependency shifts x $/shift ward salaries 10 lab tests x RVU weight x lab ‘base’ or unit cost 1 CT + 2 plain film xrays x RVU weight x imaging base cost Patient costs should reconcile to General Ledger

131. 131 Problems in generalising cost estimates What costs are included? capital? outpatient? medical fees? Are episodes defined in the same way? Rehab? ‘Up’ & ‘down’ transfers? Have ‘outlier’ patients been trimmed from the data?

132. 132 Atypical values affect the mean

133. 133 The problem of outliers The ‘long tail’ of the resource-use histogram ‘Trimming’ used to define ‘typical’ vs ‘atypical’ cases Statistical criteria (IQR, 2SD) used to define and exclude atypical cases from DRGs

134. 134 Patient level costing Have you allocated overheads across products fairly Inpatient, outpatient, ED, mental health Have you got the big cost drivers right Materiality is important

135. 135 Strategies in activity based funding environment: Assume we have activity based funding. Let us assume that the marginal revenue for each hospital is 50% of the average revenue (essentially the Commonwealth share) and that it is uncapped What should Holy Mother’s strategy be in this scenario? What factors should Holy Mother’s management take into account?

136. 136 Group exercise Groups for this exercise are based on people fulfilling like roles with access to similar information systems (e.g. people with costing systems in their hospitals vs. not) 1: Design a suite of mock up reports of information you would like to receive monthly in order to manage to the efficient price 2. Assume there is an untoward variance in one of the metrics (chose any one of the metrics in one of your mock-up reports). How would you approach the next steps? Which of your subordinates would you involve? What committees?