Using clinical indicators

1
Using clinical indicators to drive improvement
Heather Dawe heather.dawe_at_ic.nhs.uk
www.ic.nhs.uk
2
The NHS Information Centre

• What do we do?
• Control the data standards required to
  systematically manage the provision of NHS and
  social care
• Collect and publish information across the wide
  arena of health and social care
• Assist the NHS and policy makers helping them
  to use their data to inform decision making

3
Summary

• What is quality and how can it be measured
• Clinical indicators and their uses in assessing
  quality of care
• The challenges of advocating the effective use of
  clinical indicators in an information hungry age
• The research potential of this relatively young
  area

4
Measuring Quality of Care

• And what is good, Phædrus,And what is not
  good...Need we ask anyone to tell us these
  things?
• Zen and the Art of Motorcycle Maintenance An
  Enquiry Into Values, Robert M. Pirsig

5
NHS Next Stage Review Lord Darzis definition
of quality

• Patient safety
• Safe clean
• Minimising avoidable harm (e.g. drug errors, HAI)
• Patient experience
• Compassion, dignity respect
• Patient satisfaction
• Effectiveness of care
• Survival rates
• Complication rates
• Measures of clinical improvement
• Patient Reported Outcome Measures

6
Effectiveness of care survival rates

• How do you assess outcomes?
• Monitoring patient survival rates over time
• Comparing survival rates for given treatment
  across NHS organisations
• Using data derived from operational NHS systems
• To enable both requires standardisation
• Survival rates are adjusted for patient casemix
• The Standard Mortality Ratio (SMR) is a clinical
  indicator used internationally to provide a means
  of comparing and monitoring
• SMR (observed/expected) x 100

7
Using SMRs...
...enquiry initiated by high SMRs
8
Where do we need to be?

• Using clinical indicators effectively to
• Assist NHS managers and clinical governance
• Help the regulator
• Provide patients with useful information to
  assist them as they choose which hospital to
  attend for treatment
• (At least) three different audiences with at
  times subtle but fundamental differences in their
  needs

9
Anything slowing this down?

• Data quality
• Politics and the media
• Public and clinical trust in statistics

10
Solutions? Data quality...

• Data sourced from operational NHS information
  systems and utilised for secondary uses are
  sometimes of dubious quality
• Treating the issue at source is usually the best
  way to reach an effective resolution
• Also a good idea not to ask the very busy NHS to
  fill in forms they can see no use for sell them
  the benefits!
• NHS IC has a data quality programme focused on
  working with the NHS to improve this area

11
Solutions? Politics and the media...

• It can be all too easy to ignore the risks of
  confusing and alienating the clinical and patient
  communities
• League tables and targets are never far away. Is
  indicator just another word for target? It
  should not be
• Taking time to advocate and help to encourage the
  effective use of clinical indicators is key
• An indicator is just an indicator after all!
• Indicators of potential good practice as well as
  potential bad practice

12
Solutions? Lies, damn lies ...

• The methods used to derive SMRs are relatively
  complex
• There are a number of ways to derive risk
  adjusted clinical indicators. Which way is
  correct?
• Intellectual Property. Who owns the methodology?
• Complexity and a lack of clarity leads to a loss
  of trust
• Complete transparency of methodology. Should it
  be owned by the NHS?
• Continually open to peer challenge
• Acknowledgement that this is a young area and
  methods will change over time

13
Statistical research potential...

• The methods used to derive SMRs are relatively
  complex but still fairly rudimentary
• Risk modelling in healthcare could learn from
  other sectors
• Alternatives to traditional logistic regression
  and other GLMs
• Bayesian decision frameworks can incorporate
  prior knowledge incorporating specialist
  knowledge could make the baseline models more
  robust and defensible
• Lots of potential. If youre interested or can
  help then please let me know!

14
Social research...
... What do patients want to know ...
15
... And how do they want to learn?

• Can I park my car for free?
• Whats the food like?
• Am I going to die?
• Digital (I like to twitter...)
• Word-of-mouth
• Trust their GP to best advise
• These decisions implement an important government
  policy NHS patient choice

16
Policy aside the bottom line
17
Is it a middle line?

• Bristol
• Shipman
• And now Mid-Staffordshire NHS Foundation Trust
• Clinical indicators are not the complete
  solution. They can however form part of it. We
  must use them effectively for each of their
  different audiences and help to improve the
  quality of care.

18
(No Transcript)