Developing ABF in mental health services: time is running out! Joe Scuteri (Managing Director) Health Informatics Conference 2012 Tuesday 31 st July, 2012
The ABF Health Reform From 2014/15 the Commonwealth will contribute 45% of the efficient growth in hospital services using ABF The initiative will deliver a minimum of an additional $16.4 billion for hospital services by 2020, thereby increasing the Commonwealth s share of public hospital funding Hospitals are funded for provision of patient care based on the type and mix of the patients treated A fixed amount is paid to the hospital based on the relative cost of the group to which the patient service event is classified multiplied by the efficient price
ABF Implementation 1 st July 2012 - Tools Classifications for ABF implemented from 1 st July 2012: Acute Admitted: Australian Refined Diagnosis Related Groups (AR-DRG) Version 6.x. Emergency Department: Urgency Related Groups (URG) 1.2 or Urgency and Disposition Groups (UDG) version 1.2; Non Admitted: NHCDC Tier 2 clinic classification National Weighted Activity Units (NWAUs) for all classifications have been developed using 2009/10 data from the National Hospital Cost Data Collection (NHCDC) National Efficient Price (NEP) set at $4,808 for one NWAU
ABF Implementation 1 st July 2013 - Tools Classifications for ABF implemented from 1 st July 2013: Sub-acute: AN-SNAP for admitted sub-acute; and Mental Health: funded using DRGs, URGs, Tier 2 and AN- SNAP depending on type of care National Weighted Activity Units (NWAUs) will be developed using 2010/11 data from the NHCDC There is a lack of appropriate classification systems and associated costing models for mental health services that is the risk to be managed
AR-DRGs for mental health The Australian admitted patient classification system Each group is intended to be clinically meaningful and resource homogenous Uses medical diagnosis, age, and procedure as variables to predict resource consumption Divided into 23 Major Diagnostic Categories (MDCs), 11 pre-mdcs, and 698 DRGs 13 AR-DRGs (two sameday) specific to mental health U40Z to U68Z Poor predictors of resource use, lowest explanation of variance amongst all MDCs in AR-DRGs
AN-SNAP for mental health Australian classification system, developed in 1996, reviewed in 2006, for sub and non-acute patients (overnight and ambulatory) Palliative care Psychogeriatric care Subacute care Rehabilitation medicine Geriatric evaluation and management Non-acute care Maintenance care (e.g. nursing home, convalescent and planned respite care) Each case type is defined based on the characteristics of the patient and goal of care Aside from psychogeriatric care, not developed for mental health (does not cover mental health rehabilitation or maintenance care)
URGs for mental health Australian classification system, originally developed by Jelinek in 1992, for emergency department patients The URG system uses three variables to define 61 classes Disposition (admitted, not-admitted, DOA, DNW) Triage (categories 1 to 5) Principal ED diagnosis (grouped into 27 Major Diagnostic Blocks (MDB)) Only two of the admitted URGs (triage categories 4 and 5) are specific to psychiatric/social problem/other presentation MDB URGs are generally regarded as inadequate to capture the extent of mental health treatment provided in ED
Tier 2 Clinics for mental health Australian classification system, originally developed in 1998, as part of the Developmental Ambulatory Classification Study (DACS) to classify outpatient clinics Tier 2 has 107 classes (based on service characteristics) including: 14 procedural clinics classes 51 medical clinics classes 8 stand-alone diagnostic clinic classes 34 allied health and/or clinical nurse specialist interventions clinic classes Tier 2 includes one placeholder category to accommodate community mental health services Although recently improved, generally accepted that Tier 2 is limited due to poor category differentiation, overlap and ambiguity
Why does mental health need a classification system? Fund: Funding MH services using the general classification systems carries significant risks, as mental health services are poorly differentiated Cost benchmarking: adjustments for casemix differences are needed to enable agencies to compare (and hence improve) their cost, length of stay, and other performance indicators Quality assurance: by analysing differences in service delivery models for the same class, agencies can better focus on improving quality Reviews of consumer outcomes: understanding casemix differences is essential to interpret the variation between agencies in consumer outcomes; Development of clinical pathways/protocols: casemix classes establish a framework for determining what package of services particular consumer groups should receive.
So what are our options? Mental Health Classification and Service Costing (MH-CASC): Based on a 1996 study aimed to develop a casemix classification to define and classify the mix of patients in the specialist MH sector MH Care Clusters: developed for use in the Payment by Results (PbR) system in the UK System for the Classification of In Patient Psychiatry (SCIPP): developed for use in classifying inpatient psychiatric care in Canada Start from scratch: there has been significant investment in establishing the Mental Health National Outcomes and Casemix Collection (MH-NOCC) that could be used as a data source for classification and costing development
What is MH-CASC? One of few studies that attempted to build a classification system specifically for mental health services (the only Australian study) Clinical and service utilisation data were captured on 18,000 mental health episodes of care encompassing acute, rehabilitation, and community components The analysis produced a classification system and associated service costs containing 42 classes 19 for community episodes 23 for inpatient episodes MH-CASC has never been adopted as a classification tool or for funding purposes in Australia
Overview of MH-CASC
Conclusions from the MH-CASC study The developers of MH-CASC claim that the project demonstrated that: a meaningful casemix classification is possible which can be used to describe the complex activities of MH services; clear convergence in the measures used for the classification with those with demonstrated utility in outcome measurement; the value of casemix information tools in highlighting the degree of variation in MH clinical practice; and important new concepts for defining episodes in the community by using a definition based on the concept of period of care.
Issues with MH-CASC classification system Not reflective of current MH practice - many of the classes and the associated costs are no longer relevant Importantly, the classes account for only a relatively low level of variation in observed costs, for inpatients this proportion is not that much different to AR-DRGs The reliance on clinician-rated clinical outcome scores gives rise to the perception that the system is open to gaming The fact that the classification was largely unable to account for the very large provider differences in the resources used by different services to deliver what appeared to be similar services
Payment by Results (PbR) PbR was introduced in the UK to promote: Fairness and equity: support patient choice by allowing the money to follow the patient to different types of provider Efficiency: reward efficiency and quality by allowing providers to retain the difference if they could provide the required standard of care at a lower cost than the national price Effectiveness: reduce waiting times by paying providers for the volume of work done Innovation: refocus discussions between commissioner and provider away from price and towards quality and innovation
Currencies for mental health in PbR Care Clusters The national currency for purchasing mental health services under PbR is the care cluster A standardised care package that includes admitted patients and community care services has been developed for each care cluster known as the Integrated Packages Approach to Care Each of the 21 care clusters defines a group of service users who are relatively similar in their care needs and therefore resource requirements over specific periods of time that range from four weeks to 12 months Care clusters are intended to balance the risks between commissioners and providers - commissioners do not have to pay extra for each contact and intervention and providers know they will be paid for each patient they care for and they also have an incentive to innovate and support the patient in the most cost effective setting.
Mental Health Care Clusters
How does the system work? Individual patients are allocated to a care cluster using the Mental Health Clustering Tool which incorporates items based from the: from the Health of the Nations Outcome Scales (HoNOS); and the Summary of Assessments of Risk and Need (SARN), a computer program using scores input by clinicians to answer 18 questions 2012-13 is the introductory year for a shift from in mental health funding from block grants to PbR currencies, mental health trusts will be paid by need as defined by the care cluster (not paid by activity) different local tariff for each cluster (thus absence of a national tariff) still not clear how the system takes into account the length of treatment (weeks to months) currencies for child and adolescent MH services still need to be developed
System for the classification of In Patient Psychiatry (SCIPP) - Canada Casemix methodology released in 2003 for inpatient care only Developed by interrai, Ontario Hospital Association and Ontario Ministry of Health and Long Term Care (MOHLTC) Based on data for 2,000 patients including assessment and stafftime measures collected 1999 to 2001 There are seven SCIPP categories, each category has a number of SCIPP groups (47 groups total) Categories 1 to 7 are ordered in a clinical hierarchy from most resource intensive to least (based on nursing, non-nursing, and total staff costs)
Overall SCIPP structure SCIPP Categories SCIPP Groups 1 Schizophrenia and other psychotic disorders 18 groups SCIPP Category Hierarchy Provisional diagnosis 2 Cognitive disorders 3 Mood disorders 4 Personality disorders 4 groups 12 groups 2 groups 5 Eating disorders 1 group 6 Substance use disorders 3 groups 7 Other disorders 7 groups 47 groups
SCIPP resource intensity phases There are different resource needs over different periods SCIPP weighting groups correspond to these phases Time Day 1 5 Day 6 730 Day 731 Discharg e Admission phase Post-admission phase Long-term phase Admission phase - highest resource needs Post-admission phase - lower resource needs Long-term phase - different (usually lower) resource needs
Issues with SCIPP Not used for payment purposes in Canada (or anywhere else) Only covers inpatient services Takes account of both clinical and resource factors, while diagnosis is crucial for the initial classification, subdivisions within each category vary by staffing costs Each episode is adjusted for length of stay and location in the treatment pathway so that average daily payments will be lower when length of stay is longer (there may still be gaming) The involvement of interrai means there may be proprietary issues in use outside Canada
Starting from scratch: MH-NOCC? Mainstream hospital and community health services systems generally have incomplete data around mental health services Best data are in the mental health program specific systems used to support Mental Health National Outcomes and Casemix Collection (MH- NOCC) which started in 2003 Focus in on clinical measurement scales (e.g. HoNOS, and Life Skills Profile), other clinical data (e.g. principal and additional diagnoses, and mental health legal status) and consumer selfreport data (e.g. Kessler K10+) Covers psychiatric inpatient episodes, community residential episodes and ambulatory episodes (including sameday admitted)
Starting from scratch: MH-NOCC? Details for each collection episode include: Service unit (e.g. target population, program type, location) Collection occasion (e.g. date, setting, outcome data) Patient details (e.g. identification number, age, sex) Always intended that MH-NOCC be used to develop casemix classifications for mental health, but little progress made Considerable work required, particularly to link to utilisation and costs data for mental health episodes Will not offer a solution for 2013, but should; be considered for longer term to capitalise on substantial investment that has been made in data collection
Conclusions Mental health needs to be part of ABF, but the classification and costing tools and technologies are not there The current proxy arrangements for mental health are inadequate; they are not a solution; and the risks need to be managed IHPA has commissioned a consultancy to define mental health services for activity based funding purposes and identify the associated cost drivers for services that meet the definition Consultants must report by February 2013, but not required to develop a mental health specific classification system We could have drawn on international experiences to move forward but we have run out of time yet again mental health has been left behind