Predictive risk modelling in health: options for New Zealand and Australia
Laura E. Panattoni A , Rhema Vaithianathan B D , Toni Ashton A and Geraint H. Lewis C
+ Author Affiliations
- Author Affiliations
A University of Auckland, School of Population Health, Private Bag 92019, Auckland, New Zealand. Email: l.panattoni@auckland.ac.nz; toni.ashton@auckland.ac.nz
B University of Auckland, Economics Department, Private Bag 92019, Auckland, New Zealand.
C The Nuffield Trust, 59 New Cavendish Street, London W1G 7 LP, United Kingdom. Email: geraint.lewis@nuffieldtrust.org.uk
D Corresponding author. Email: r.vaithianathan@auckland.ac.nz
Australian Health Review 35(1) 45-51 https://doi.org/10.1071/AH09845
Submitted: 18 October 2009 Accepted: 7 May 2010 Published: 25 February 2011
Journal Compilation © AHHA 2011
Abstract
Predictive risk models (PRMs) are case-finding tools that enable health care systems to identify patients at risk of expensive and potentially avoidable events such as emergency hospitalisation. Examples include the PARR (Patients-at-Risk-of-Rehospitalisation) tool and Combined Predictive Model used by the National Health Service in England. When such models are coupled with an appropriate preventive intervention designed to avert the adverse event, they represent a useful strategy for improving the cost-effectiveness of preventive health care. This article reviews the current knowledge about PRMs and explores some of the issues surrounding the potential introduction of a PRM to a public health system. We make a particular case for New Zealand, but also consider issues that are relevant to Australia.
What is known about the topic? PRMs are an alternative method to threshold modelling and clinical knowledge for determining a patient’s risk of a future event. PRMs are already in use in New Zealand and Australia to predict the occurrence of a disease. However, Kaiser Permanente in the US, and the UK’s National Health Service are using PRMs to predict health service usage (e.g. risk of future emergency hospitalisation) at the individual level.
What does this paper add? This paper discusses issues including model parameters, data requirements and ethical considerations for using a PRM as a service planning tool in Australia and New Zealand.
What are the implications for practitioners? PRMs could be used as the health service equivalent of disease risk assessments. New Zealand and Australia already have routinely collected data that could be used to predict various adverse, costly and potentially preventable health service events.
References
[1] Reuben DB, Keeler E, Seeman TE, Sewall A, Hirsch SH, Guralnik JM. Identification of risk for high hospital use: cost comparisons of four strategies and performance across subgroups. J Am Geriatr Soc 2003; 51 615–20.| Identification of risk for high hospital use: cost comparisons of four strategies and performance across subgroups.Crossref | GoogleScholarGoogle Scholar | 12752835PubMed |
[2] Hughes JS, Averill RF, Eisenhandler J, Goldfield NI, Muldoon J, Neff JM, Gay JC. Clinical Risk Groups (CRGs): a classification system for risk-adjusted capitation-based payment and health care management. Med Care 2004; 42 81–90.
| Clinical Risk Groups (CRGs): a classification system for risk-adjusted capitation-based payment and health care management.Crossref | GoogleScholarGoogle Scholar | 14713742PubMed |
[3] Improving Chronic Disease Management. Castlefields: United Kingdom Department of Health; 2004. Available at http://www.dh.gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh/@en/documents/digitalasset/dh_4075213.pdf [verified 17 March 2010].
[4] Minutes of the Meeting of the Community and Public Health Advisory Committee for Tuesday 28 May 2002. Counties Manukau: Counties Manukau District Health Board; 2002. Available at http://www.sah.co.nz/About_CMDHB/Meetings/Minutes/CPHAC/2002/MinutesCPHAC28May02.pdf [verified 17 March 2010].
[5] Innovative Care for Chronic Conditions: Building Blocks for Action. Geneva: World Health Organization; 2002. Available at http://www.who.int/entity/diabetesactiononline/about/icccglobalreport.pdf [verified 1 March 2010].
[6] Hutt R, Rosen R, McCauley J. Case-managing long term conditions: What impact does it have in the treatment of older people? London: King’s Fund; 2004. Available at http://www.kingsfund.org.uk/document.rm?id=90 [verified 17 March 2010].
[7] Department of Human Services. Hospital Admission Risk Program (HARP) Background Paper. Victoria: Department of Human Services; 2002. Available at http://www.health.vic.gov.au/harp-cdm/harpbgpap.pdf [verified 10 March 2010].
[8] New Zealand’s 65+ Population 2007. A Statistical Volume. Wellington: Statistics New Zealand; 2007. Available at http://www.stats.govt.nz/reports/analytical-reports/new-zealands-65-plus-population.aspx [verified 10 March 2010].
[9] New Zealand Ministry of Health. Population Ageing and Health Expenditure: New Zealand 2002 – 2051. Wellington: Ministry of Health; 2004. Available at http://www.moh.govt.nz/moh.nsf/0/35F6DE971724D02CCC256F3800792634/$File/populationageingandhealthexpenditure.pdf [verified 28 May 2008].
[10] Older People’s Health Chart Book. Wellington: New Zealand Ministry of Health; 2006.
[11] Curry N, Billings J, Darin B, Curry N, Billings J, Darin B, Dixon J, Williams M, Wennberg D. Predictive Risk Project Literature Review (summary). King’s Fund, Health Dialog, New York University; 2005. Available at http://www.kingsfund.org.uk/document.rm?id=6195 [verified 10 March 2010].
[12] Person-Based Resource Allocation Project. London: Nuffield Trust; 2009. Available at http://www.nuffieldtrust.org.uk/projects/index.aspx?id=338 [verified 10 March 2010].
[13] Parker J, McCombs J, Graddy E. Can pharmacy data improve prediction of hospital outcomes? Comparisons with a diagnosis-based comorbidity measure. Med Care 2003; 41 407–19.
| Can pharmacy data improve prediction of hospital outcomes? Comparisons with a diagnosis-based comorbidity measure.Crossref | GoogleScholarGoogle Scholar | 12618644PubMed |
[14] McCusker J, Bellavance F, Belzile E. Return to the emergency department among elders: patterns and predictors. Acad Emerg Med 2000; 7 249–59.
| Return to the emergency department among elders: patterns and predictors.Crossref | GoogleScholarGoogle Scholar | 1:STN:280:DC%2BD3c7ptlSktw%3D%3D&md5=d299707bf16398110a085644f689162aCAS | 10730832PubMed |
[15] Wahls T, Barnett M, Rosenthal G. Predicting resource utilization in a Veterans Health Administration primary care population: comparison of methods based on diagnoses and medications. Med Care 2004; 42 123–8.
| Predicting resource utilization in a Veterans Health Administration primary care population: comparison of methods based on diagnoses and medications.Crossref | GoogleScholarGoogle Scholar | 14734949PubMed |
[16] Dendukuri N, McCusker J, Belzile E. The identification of seniors at risk screening tool: further evidence of concurrent and predictive validity. J Am Geriatr Soc 2004; 52 290–6.
| The identification of seniors at risk screening tool: further evidence of concurrent and predictive validity.Crossref | GoogleScholarGoogle Scholar | 14728643PubMed |
[17] Selby JV, Karter AJ, Ackerson LM, Ferrara A, Liu J. Developing a prediction rule from automated clinical databases to identify high risk patients in a large population with diabetes. Diabetes Care 2001; 24 1547–55.
| Developing a prediction rule from automated clinical databases to identify high risk patients in a large population with diabetes.Crossref | GoogleScholarGoogle Scholar | 1:STN:280:DC%2BD3MvnvV2msg%3D%3D&md5=851fccbde62927c0e8c278ddf0de88e8CAS | 11522697PubMed |
[18] Barber M, Gregorie D. Looking for the needle in the haystack: techniques for finding. Healthplan 2001; 42 58–62, 64.
| 1:STN:280:DC%2BD2MnhtVSktQ%3D%3D&md5=c667a5b641ccc72987eafd429635d9c1CAS | 16296289PubMed |
[19] Petersen L, Pietz K, Woodard L, Byrne M. Comparison of the predictive validity of diagnosis-based risk adjusters for clinical outcomes. Med Care 2005; 43 61–7.
| 15626935PubMed |
[20] Chronic Disease Management. Castlefields: Castlefields Health Centre; 2004. Available at http://www.natpact.nhs.uk/uploads/Castlefields%20Report.doc [verified 10 March 2010].
[21] Roseman P. Frequent Adult Medical Admissions Final Report. Counties Manukau: Counties Manukau Acute Demand Management Group; 2003. Available at http://www.sah.co.nz/Funded-Services/CCM/docs/reports/FAMAreport.pdf [verified 10 August 2009].
[22] Roland M, Dusheiko M, Gravelle H, Parker S. Follow up of people aged 65 and over with a history of emergency admissions: analysis of routine admissions data. BMJ 2005; 330 289–92.
| Follow up of people aged 65 and over with a history of emergency admissions: analysis of routine admissions data.Crossref | GoogleScholarGoogle Scholar | 15695276PubMed |
[23] Weissert W, Chernew M, Hirth R. Theory and cost benefit analysis to home care policy: titrating versus targeting home care services to frail elderly clients. J Aging Health 2003; 15 99–123.
| Theory and cost benefit analysis to home care policy: titrating versus targeting home care services to frail elderly clients.Crossref | GoogleScholarGoogle Scholar | 12611411PubMed |
[24] Dudley RA, Medlin CA, Hammann LB, Cisternas MG, Brand R, Rennie DJ, Luft HS. The best of both worlds?: Potential of hybrid prospective/concurrent risk adjustment. Med Care 2003; 41 56–69.
| The best of both worlds?: Potential of hybrid prospective/concurrent risk adjustment.Crossref | GoogleScholarGoogle Scholar | 12544544PubMed |
[25] Olfson M, Mechanic D, Boyer CA, Hansell S, Walkup J, Weiden P. Assessing clinical predictions of early rehospitalization in schizophrenia. J Nerv Ment Dis 1999; 187 721–9.
| Assessing clinical predictions of early rehospitalization in schizophrenia.Crossref | GoogleScholarGoogle Scholar | 1:STN:280:DC%2BD3c7islagsA%3D%3D&md5=e8d2b4b0994dec533ff85bbd8033d44cCAS | 10665466PubMed |
[26] Risk Adjustment Methods and their Relevance to “Pay-or-Play”. Washington, DC: Institute of Health Policy Solutions; 2005. Available at http://www.ihps.org/pubs/2005_Apr_IHPS_SB2_ESup_Risk_Adj.pdf [verified 10 March 2010].
[27] Baker S, Priest P, Jackson R. Using thresholds based on risk of cardiovascular disease to target treatment for hypertension: modelling events averted and number treated. BMJ 2000; 320 680–5.
| Using thresholds based on risk of cardiovascular disease to target treatment for hypertension: modelling events averted and number treated.Crossref | GoogleScholarGoogle Scholar | 1:STN:280:DC%2BD3c7nslGkug%3D%3D&md5=e1a9fe371f038e0ca022ec232c757e7dCAS | 10710577PubMed |
[28] Feachem R, Neelam S, White K. Getting more for their dollar: a comparison of the NHS with California’s Kaiser Permanente. BMJ 2002; 324 135–43.
| Getting more for their dollar: a comparison of the NHS with California’s Kaiser Permanente.Crossref | GoogleScholarGoogle Scholar | 11799029PubMed |
[29] Department of Health’s Public Service Agreement Aims and Objectives. United Kingdom Department of Health; 2005.
[30] Billings J, Mijanovich T, Dixon J, Curry N, Wennberg D, Darin R, Steinort K. Case finding algorithms for patients at risk of re-hospitalization PARR1 and PARR2. New York University, King’s Fund, Health Dialog Analytic Solutions; 2006. Available at http://www.kingsfund.org.uk/document.rm?id=6209 [verified 10 March 2010].
[31] Billings J, Zeitel L, Lukomnik J, Carey TS, Blank AE, Newman L. Impact of socioeconomic status on hospital use in New York City. Health Aff (Millwood) 1993; 12 162–73.
| Impact of socioeconomic status on hospital use in New York City.Crossref | GoogleScholarGoogle Scholar | 1:STN:280:DyaK3s3osFGltA%3D%3D&md5=aaa42833997843887f54ee013c1fdb43CAS | 8509018PubMed |
[32] Lewis L. Predictive Risk Stratification Model Project. Wales: Clinical Futures Chronic Disease Management Conference; 2007. Available at http://www.clinicalfuturesgwent.wales.nhs.uk/systems/file_download.aspx?pg=577&ver=2 [verified 10 March 2010].
[33] SPARRA: Scottish patients at risk of readmission and admission. Scotland: NHS National Services Scotland; 2006. Available at http://www.isdscotland.org/isd/files/SPARRA_Report.pdf [verified 10 March 2010].
[34] Mukamel D, Chou C, Zimmer J, Rothenberg B. The effect of accurate patient screening on the cost-effectiveness of case management programs. Gerontologist 1997; 37 777–84.
| 1:STN:280:DyaK1c%2FpsVeqtg%3D%3D&md5=6f46d7e4465b35ed001873481626af97CAS | 9432994PubMed |
[35] Zhao Y, Ellis RP, Ash AS, Calabrese D, Ayanian JZ, Slaughter JP, Weyuker L, Bowen B. Measuring population health risks using inpatient diagnoses and outpatient pharmacy data. Health Serv Res 2001; 36 180–93.
| 1:STN:280:DC%2BD2MvnvVakuw%3D%3D&md5=e0cb3198b9aa8b28ad8c8a03bff89e6fCAS | 16148968PubMed |
[36] Zhao Y, Ash AS, Haughton J, McMillan B. Identifying future high-cost cases through predictive modeling. Dis Manag Health Outcomes 2003; 11 389–97.
| Identifying future high-cost cases through predictive modeling.Crossref | GoogleScholarGoogle Scholar |
[37] Ash AS, Ellis RP, Pope GC, Ayanian JZ, Bates DW, Burstin H, Lezzoni LI, Mackay E, et al Using diagnoses to describe populations and predict costs. Health Care Financ Rev 2000; 21 7–28.
| 1:STN:280:DC%2BD3MvjslSnsg%3D%3D&md5=874443616567cd8833ae120a9346568dCAS | 11481769PubMed |
[38] Ash AS, Zhao Y, Ellis RP, Schlein Kramer M. Finding future high-cost cases: comparing prior cost versus diagnosis-base methods. Health Serv Res 2001; 36 194–206.
| 1:STN:280:DC%2BD2MvnvVarsg%3D%3D&md5=8d8c95ffa659879f953d3c7757ba055fCAS | 16148969PubMed |
[39] Schatz M, Cook EF, Joshua A, Petitti D. Risk factors for asthma hospitalizations in a managed care organization: development of a clinical prediction rule. Am J Manag Care 2003; 9 538–47.
| 12921231PubMed |
[40] Roblin DW, Juhn PI, Preston BJ, Penna RD, Feitelberg SP, Khoury A, Scott JC. A low-cost approach to prospective identification of impending high cost outcomes. Med Care 1999; 37 1155–63.
| A low-cost approach to prospective identification of impending high cost outcomes.Crossref | GoogleScholarGoogle Scholar | 1:STN:280:DC%2BD3c%2FhsFyisg%3D%3D&md5=f0ef610588ac2c56c14b041ec3fafefaCAS | 10549617PubMed |
[41] Meenan RT, Goodman MJ, Fishman PA, Hornbrook MC, O’Keeffe-Rosetti MC, Bachman DJ. Using risk-adjustment models to identify high-cost risks. Med Care 2003; 41 1301–12.
| Using risk-adjustment models to identify high-cost risks.Crossref | GoogleScholarGoogle Scholar | 14583693PubMed |
[42] Dove H, Duncan I, Robb A. A prediction model for targeting low-cost, high risk members of managed care organizations. Am J Manag Care 2003; 9 381–9.
| 12744300PubMed |
[43] Axelrod R, Vogel D. Predictive modelling in health plans. Dis Manag Health Outcomes 2003; 11 779–87.
| Predictive modelling in health plans.Crossref | GoogleScholarGoogle Scholar |
[44] Kass N. An ethics framework for public health. Am J Public Health 2001; 91 1776–82.
| An ethics framework for public health.Crossref | GoogleScholarGoogle Scholar | 1:STN:280:DC%2BD3MngvFOksg%3D%3D&md5=a35b650e6bf7e46f9e7b61460b26097fCAS | 11684600PubMed |
[45] National Health Data Committee. Admitted patient care national minimum data set: nation health data dictionary, version 12. Canberra: Australian Institute of Health and Welfare; 2003. Available at http://www.aihw.gov.au/publications/hwi/apc/apc.pdf [verified 10 March 2010].
