P Smith SHEA seminar Nov English Health Reforms: What is in them for Sweden? Peter C. Smith Professor of Health Policy Imperial College Business School
P Smith SHEA seminar Nov Some recent initiatives in England Competition – Mixed market of public and private provision – Competition in the market and competition for the market – Cooperation and Competition Panel Patient choice – Choice of non-emergency specialist (any willing provider) – To be extended to general practitioners and community care Capitated budgets – Indicative budgets for general practitioners – ‘Hard’ budgets for general practitioner consortiums – Experiments with patient-held budgets Clinical guidelines – NICE building quality into its guidelines, and extending coverage to established treatments – Move towards ‘value based pricing’
P Smith SHEA seminar Nov Performance measurement: a requirement for all reforms Helping practitioners identify ‘what works’ Identifying good and bad practitioners Stimulating performance improvement Promoting choice for patients and purchasers Regulating the health system Designing health system reform Promoting accountability
P Smith SHEA seminar Nov The universal role of performance information … to enable actors throughout the system to make better decisions The thesis: – that measuring and reporting performance offers one of the most powerful instruments for performance improvement – it has hitherto been largely unexploited.
P Smith SHEA seminar Nov Securing attention
P Smith SHEA seminar Nov SCOTTISH CLINICALOUTCOMES INDICATORS “It is stressed that no direct inferences about the quality of care should be drawn from the indicators. They are intended rather to highlight issues that may require further attention” Source: Kendrick, S., Cline, D. And Finlayson, A., “Clinical outcomes indicators in Scotland: lessons and prospects”, in Davies, H. T. O. et al (1999), Managing quality: strategic issues in health care management, Aldershot: Ashgate.
P Smith SHEA seminar Nov Why did Scotland’s clinical indicators scheme have low impact? poor credibility, relevance and timeliness of the data, a lack of awareness and expertise on the part of clinical staff a lack of incentives and effective external scrutiny. Mannion, R. and M. Goddard (2001). "Impact of published clinical outcomes data: case study in NHS hospital trusts." British Medical Journal 323,
P Smith SHEA seminar Nov NHS Star Ratings Prepared for every NHS organization starting in 2001 Every organization ranked on a scale of zero to three stars Objective was to inform the public of the performance of their local health care organizations Complex composite measure of performance reflecting centrally determined objectives (especially waiting times) Organizations with higher scores were given increased freedoms Jobs of chief executives at risk in organizations with poorer scores.
P Smith SHEA seminar Nov York Hospital Star Rating
P Smith SHEA seminar Nov Inpatient waiting list by length of wait, England,
P Smith SHEA seminar Nov Propper et al (2008): England vs Scotland Waiting more than 6 months Waiting more than 9 months
P Smith SHEA seminar Nov Some unintended consequences Ignoring ‘untargeted’ outputs Misrepresentation Distorted behaviour (‘gaming’)
P Smith SHEA seminar Nov Post-operative mortality rate by star rating 2001/02
P Smith SHEA seminar Nov Hospital readmission rate by star rating 2001/02
P Smith SHEA seminar Nov Ambulance response times: no evidence of discontinuity Commission for Health Improvement (2004), “Analysis of ambulance CAD data to consider robustness of reporting of response times”, London: CHI
P Smith SHEA seminar Nov Ambulance response times: marginal discontinuity at 8 minutes Commission for Health Improvement (2004), “Analysis of ambulance CAD data to consider robustness of reporting of response times”, London: CHI
P Smith SHEA seminar Nov Ambulance response times: pronounced discontinuity at 8 minutes Commission for Health Improvement (2004), “Analysis of ambulance CAD data to consider robustness of reporting of response times”, London: CHI
P Smith SHEA seminar Nov Better measurement and risk adjustment
P Smith SHEA seminar Nov Measuring patient-reported outcomes From April 2009 all hospitals will report on patient-reported outcome measures (PROMs) for patients undergoing – hip replacements – knee replacements – groin hernia repair – varicose vein ligation.
P Smith SHEA seminar Nov EQ-5D: A Generic Quality of Life Measure Kind, P. et al. BMJ 1998;316:
P Smith SHEA seminar Nov Quality measures: some desiderata? sustained commitment to the performance measures, at the very highest level; choice of quality measures that are aligned with system objectives; central specification, updating and audit of measures; consistent reporting across all relevant organizations (yardstick competition); analytic capacity to address issues of risk adjustment and attribution; integration of measures into local governance and external accountability arrangements; alignment with proportionate scrutiny and incentives; engagement with relevant stakeholders, including user groups and professional organizations.
P Smith SHEA seminar Nov Paying for performance
P Smith SHEA seminar Nov Quality and Outcomes Framework (QOF) Developed in negotiation between government and primary care physicians Implemented in April 2004 Major emphasis on clinical quality About 20% of income determined by quality incentives About 150 performance measures Major reliance on self-reporting (with external audit).
P Smith SHEA seminar Nov Quality and Outcomes Framework 2004/05: Indicators and points at risk Area of practicePIsPoints Clinical76550 Organizational56184 Additional services1036 Patient experience4100 Holistic care (balanced clinical care)-100 Quality payments (balanced quality)-30 Access bonus-50 Maximum
P Smith SHEA seminar Nov Quality and Outcomes Framework: Clinical indicators DomainPIsPoints CHD including LVD etc15121 Stroke or transient ischaemic attack1031 Cancer212 Hypothyroidism28 Diabetes1899 Hypertension5105 Mental health541 Asthma772 COPD845 Epilepsy416 Clinical maximum76550
P Smith SHEA seminar Nov Hypertension: indicators, scale and points at risk Records MinMaxPoints BP 1. The practice can produce a register of patients with established hypertension 9 Diagnosis and initial management BP 2.The percentage of patients with hypertension whose notes record smoking status at least once BP 3.The % of patients with hypertension who smoke, whose notes contain a record that smoking cessation advice has been offered at least once Ongoing Management BP 4.The % of patients with hypertension in which there is a record of the blood pressure in the past 9 months BP 5. The % of patients with hypertension in whom the last blood pressure (in last 9 months) is 150/90 or less
P Smith SHEA seminar Nov Threshold indicator BP5 Points earned Achievement % (55-20)/(70-20) x 56 = 39.2
P Smith SHEA seminar Nov Coronary heart disease CHD6Blood Pressure < 150/90 in last 15 months CHD8Cholesterol < 5 mmo/l in last 15 months QOF Copyright © 2007 QRESEARCH (Version 12) and The Information Centre for health and social care.
P Smith SHEA seminar Nov Stroke STROKE6BP < 150/90 in last 15 months STROKE8Cholesterol < 5 mmo/l in the last 15 months QOF Copyright © 2007 QRESEARCH (Version 12) and The Information Centre for health and social care.
P Smith SHEA seminar Nov Hypertension HBP4Blood pressure recorded in last 9 months HBP5Blood pressure < 150/90 in the last 9 months QOF Copyright © 2007 QRESEARCH (Version 12) and The Information Centre for health and social care.
P Smith SHEA seminar Nov Trends in six QOF indicators Copyright © 2007 QRESEARCH (Version 12) and The Information Centre for health and social care. CHD Coronary heart disease STROKEStroke HBP Hypertension QOF
P Smith SHEA seminar Nov Gravelle H, Sutton M, Ma A. (2010). Doctor behaviour under a pay for performance contract: treating, cheating and case finding? Economic Journal, 120, F129–F156. Examines ‘exception reporting’: patients excluded from the calculation of performance because the GP judges them to be unsuitable for inclusion Exception reporting an important ‘safety mechanism’, especially when there is no risk adjustment for complex cases Sample of 903 practices in Scotland Exception reporting of 41 indicators with continuous performance measure
P Smith SHEA seminar Nov Exception Reporting ASTHMA02, Scotland, 2005/06 The percentage of patients aged eight and over diagnosed as having asthma from 1 April 2003 where the diagnosis has been confirmed by spirometry or peak flow measurement Exceptions Mean 6.3% Median 4.1% Maximum 75.9%
P Smith SHEA seminar Nov Conclusions from Gravelle, Sutton and Ma (2010) Little evidence of gaming of prevalence reporting Some evidence of gaming of exception reporting – Correlation with practice characteristics (former fundholders and those experiencing more competition) – Higher increase in exception reporting when the practice was below upper scoring threshold in the previous year
P Smith SHEA seminar Nov P4P design issues At whom to aim incentives – Organizations, teams, individuals Scope of scheme – Comprehensive (QOF) or piecemeal Power and size of incentives – Use of thresholds in QOF Difficulty of targets – Too easy in QOF? How certain should outcome be? – Competition for limited funds or guaranteed reward schedule? Risk adjustment for disadvantaged populations – Exception reporting Avoidance of gaming and other adverse outcomes – Information systems – Audit
P Smith SHEA seminar Nov Clinical engagement
P Smith SHEA seminar Nov Royal College of Physicians “We call on employers of doctors and on medical colleges... to fashion a healthcare system in which every doctor is supported to reflect on and improve their performance.”
P Smith SHEA seminar Nov Challenges for clinical engagement in performance measurement Data and analytic requirements for patient safety and protection are quite distinct from those for continuous quality improvement Clinicians need to ‘own’ and trust data if they are to act on it There can be a tension between public reporting and anonymous feedback and peer review Performance measurement should be a central element of modern professional practice How can defensive professional responses be avoided? How are performance data to be integrated into re- accreditation of practitioners?
P Smith SHEA seminar Nov The future
P Smith SHEA seminar Nov Health System Performance monitoring Accountability inputoutput A Model of Health System Leadership and Governance Setting priorities targets, standards, elections, markets
P Smith SHEA seminar Nov The National Role Development of a clear conceptual framework and a clear vision of the purpose of performance measurement; Mandating data collection mechanisms; Information assurance and governance; Development of analytic devices and capacity to help understand the data; Development of appropriate data presentational methods; Design of incentives to act on performance measures; Proper evaluation of performance measurement instruments.
P Smith SHEA seminar Nov Without performance measurement and comparison: No means of identifying good and bad delivery practice (‘what works’) No means of identifying good and bad practitioners or organizations No protection for patients or payers No evidence with which to design health system reforms No case for investing in health care No accountability.
P Smith SHEA seminar Nov Performance Measurement for Health System Improvement Edited by – Peter C. Smith – Elias Mossialos – Irene Papanicolas – Sheila Leatherman Cambridge University Press February 2010