Oxford and Kings International Health Conference 2016 Presentation 2 - Mental Health Mortality (Abstract 0282) Final

IHC2016,AbstractRef0282 Shervais,Baker-Glenn,DickersonandSpencerJune2016 Slide1Shervais,Baker-Glenn,DickersonandSpencerJune2016

Presentation 2

The Limitations Inherent in using “benchmark” outcomes to estimate NHS health service safety

IHC2016,AbstractRef0282 Shervais,Baker-Glenn,DickersonandSpencerJune2016 Slide2

Background – People involved in the study

Dr Jennifer Spencer Healthcare Fellow University of Cambridge, Cambridge UK

Dual CCT RCPsych CAMH and ID psychiatry, MRCPsych, MB, BAO, BCh, BMedSci, BA,

Dr Elena Baker-GlennDual CCT in training RCPsych General & Old Age psychiatry, MRCPsych, BSC MBBCHIR MMEDSCI

Cambridgeshire and Peterborough NHS Foundation Trust

Dr Terry Dickerson Assistant Director EDC, University of Cambridge, Cambridge UK

PhD, MiMechE, CEng, BSc

Professor Stephen ShervaisAssociate Professor of Management Information Systems

Accounting and Information Systems, College of Business and Public Administration, Eastern Washington UPhD, MS, MA, BA


Background – What motivated the study

• Clinicians have been complaining that it is difficult to ensure people with mental health disorders obtain the care they need when they are acutely physically ill.

• Benchmark goals have shifted numerous times over the years, thus long term monitoring of appropriate outcome measures has been difficult for NHS trusts to accomplish.

• We wished to see if any appropriate long term outcome measures demonstrated an association with governmental policies regarding the NHS over time.


Theory/Framework Design Research Methodology

Blessing etal,2009


Mental illness

Encompasses a range of diagnoses including dementia, substance misuse, depression, anxiety, mania, psychosis, eating disorders, and personality disorders.• Patients with mental illness have lower life expectancy than the rest of the

population• Differences are more marked in younger adults

5


Benchmark outcomes

• Benchmarking is intended to help managers implement best practice at best cost• In the UK bemnchmarks are currently used as tools to monitor impact of

governance, management, clinical outcomes and logistics

6


Importance of the Measure• Relevance to stakeholders• Health importance• Applicability to measuring the equitable

distribution of health care (for health delivery measures)• or of health (for population health

measures)• Potential for improvement• Susceptibility to being influenced by the

health care systemScientific Soundness: Clinical Logic

• Explicitness of evidence• Strength of evidence

• Scientific Soundness: Measure Properties• Reliability• Validity• Allowance for patient/consumer

factors as required• Comprehensible

• Feasibility• Explicit specification of numerator and

denominator• Data availability

Desirable attributes of a Quality Measure

https://www.qualitymeasures.ahrq.gov/tutorial/attributes.aspx


General Health Care Delivery MeasuresClinical quality Measures• Process• Access• Outcome• Structure• Patient Experience

Related Health Care Delivery Measures• User-Enrollee Health State• Management• Use of Services• Clinical Efficacy Measures• EfficiencyPopulation Health Measure Domains

Domains of Measurement


NHS Outcomes Framework


NHS Outcomes Framework GoalsDomain 1: Preventing People from Dying Prematurely

• Maximising the contribution that the NHS can make to preventing disease• Finding the ‘missing millions’ and diagnosing earlier and more

accurately• Treating people in an appropriate and timely way• Addressing unwarranted variation in mortality and

survival rates• Reducing deaths in babies and young children


NHS outcome framework for mental illness 2015/16

• Percentage of adults receiving secondary mental health services living independently • Proportion of all people in prison who have a mental illness • Percentage of adults in contact with secondary mental health services in

paid employment• Excess mortality rate in adults with serious mental illness, aged

under 75, per 100,000 population• Age–standardised mortality rate from suicide and injury of

undetermined intent per 100,000 population (in development)

11


Methodology

1. Benchmarks used by mental health NHS trusts were identified from the National Quality Measures website

2. A highly respected “Big Data” database was selected (The World Health Organization Mortality Database)

3. Mental Health mortality and self harm rates as well as general population mortality rate were collected from the UK between the years 1990-2014.

4. Data was graphically depicted to look for trends5. Further data was then collected from the WHO site from countries with similar economic and

health care systems6. Statistical analyses were conducted to determine the statistical differences between similar

parameters. Tests were kept to a minimum to avoid Type II error.


Results – Initial GraphUK Mental Health Mortality rate, Self Harm Mortality rate and NHS Healthcare Policy changes

0

5000

10000

15000

20000

2500019

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

0720

0820

0920

1020

1120

1220

1320

14

UK total deaths attributable to mental ill healthUK female deaths attributable to mental ill healthUK male deaths attributable to mental ill healthUK total deaths caused by Intentional self-harm

UK male deaths caused by Intentional self-harm

UK female deaths caused by Intentional self-harm

Sir LiamDonaldsonpublishes"Anorganisation withaMemory"

Deloitte,MonitorandParliamentimplementtheFoundationTrustprogramme

Careinthecommunityimplemented&atypicalantipsychoticsonthemarket.

13


Further Data Collection and Preparation

• WHO mortality data was obtained for 36 countries in Europe including the UK from 1991– 2014 using ICD-10 diagnoses for:• Total deaths from all causes• Deaths due to mental and behavioural disorders• Deaths due to self harm

• Data was converted to the rate per 100,000 population and then graphed


Total Population Mortality Rates in Europe

0

500

1000

1500

2000

2500

30001990

1991

1992

1993

1994

1994

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Death

sper

100,0

00

TotalPopulationMortalityRatesinEurope(UKMortalityRateinpurple)AustriaBelarusBelgiumBulgariaCroatiaCyprusCzechRepublicDenmarkEstoniaFinlandFranceGeorgiaGermanyGreeceHungaryIcelandIrelandItalyLatviaLithuaniaLuxembourgMaltaNetherlandsNorwayPolandPortugalRomaniaRussianFederationSerbiaSlovakiaSloveniaSpainSwedenSwitzerlandUkraineUnitedKingdom


European Mortality Rates for People with Mental Health Conditions

0

20

40

60

80

100

120

140

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Morta

lityRa

tefor

peop

lewith

aMen

talHe

alth

Cond

ition(

per1

00,00

0totalpo

pulat

ion)

MortalityRateforpeoplewithaMentalHealthCondition(UKinpurple)

AustriaBelarusBelgiumBulgariaCroatiaCyprusCzechRepublicDenmarkEstoniaFinlandFranceGeorgiaGermanyGreeceHungaryIcelandIrelandItalyLatviaLithuaniaLuxembourgMaltaNetherlandsNorwayPolandPortugalRomaniaRussianFederationSerbiaSlovakiaSloveniaSpainSwedenSwitzerlandUkraineUnitedKingdom


European Mortality Rates secondary to Self Harm

0

10

20

30

40

50

60

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Mortal

ityRa

teduet

oSelf

Harm

(pe

r100,00

0tota

lpopula

tion)

EuropeanMortalityRatesduetoSelfHarm(UKisinpurple)AustriaBelarusBelgiumBulgariaCroatiaCyprusCzechRepublicDenmarkEstoniaFinlandFranceGeorgiaGermanyGreeceHungaryIcelandIrelandItalyLatviaLithuaniaLuxembourgMaltaNetherlandsNorwayPolandPortugalRomaniaRussianFederationSerbiaSlovakiaSloveniaSpainSwedenSwitzerlandUkraineUnitedKingdom


Statistical Analyses

We performed a series of one-tail t-tests on the years prior to and following each policy change. Sample size was adjusted based on the number of years available.

Policy Year Years Before Years After n t-test result1993 1990-1992 1994-1996 3 0.00372000 1994-1999 2001-2006 6 0.000032008 2003-2007 2009-2013 5 0.0146

We also averaged the mortality rates for the years prior to and after each policy change and performed a paired t-test (n=3) on the result. The result was 0.216, which is not statistically significant. We are extending the range of sample measurements and repeating the test.


Conclusion

• There appears to be an association between government mandated changes in the way NHS services operate, and an increase in the mortality rate of people with mental health disorders • The way benchmark outcomes are currently used may not be

providing decision makers with enough information to create and design safe services.


References

• Agency for Healthcare Research and Quality US Department of Health and Human ServicesNational Quality Measures Clearinghouse(https://www.qualitymeasures.ahrq.gov/tutorial/index.aspx , page last viewed 18 June 2016 and https://www.qualitymeasures.ahrq.gov/tutorial/selecting.aspx page last reviewed 22 June 2016)

• Krousel-Wood. Practical Considerations in the Measurement of Outcomes in Healthcare. 187-194. October 1999.

• NHS Outcomes Framework Measurement (https://www.england.nhs.uk/resources/resources-for-ccgs/out-frwrk/dom-1/Porter)

• What Is Value in Health Care? New England Journal of Medicine. 2477-2481. December 2010• World Health Organisation Mortality Database, accessed November 2015– June 2016

(http://apps.who.int/healthinfo/statistics/mortality/whodpms/ )

Oxford and Kings International Health Conference 2016 Presentation 2 - Mental Health Mortality (Abstract 0282) Final

Documents