Hospitalwide Mortality Measurement Project

1
Hospital-wide Mortality Measurement Project

• September 2, 2009

2
What is the charge of the group?

• DHCFP is conducting this project at the request
  of the Massachusetts Health Care Quality and Cost
  Council (HCQCC)
• Purpose is to evaluate and recommend one or more
  valid and reliable measures of whole system,
  hospital mortality that can be used for both
  quality improvement and public reporting on the
  HCQCC website by the middle of 2010

Massachusetts Division of Health Care Finance
and Policy - 2
3
Process for meeting the charge

• Researchers to conduct a head-to-head comparison
  of different vendor methodologies.
• A sample of hospital-specific discrepant cases
  where the mortality predictor is different than
  the outcome will be reviewed by a subset of
  hospitals.
• For example, cases where there was a low
  predictive probability of death and the patient
  died and cases where there was a high probability
  of death and the patient lived.
• Hospitals will review vendor reports and provide
  input on their utility for potential quality
  improvement activities.

4
What progress has been made towards achieving the
goals?

• Developed Issued Request for Information (RFI)
  for vendors to participate in head-to-head
  comparison
• Agreements to participate in the project were
  signed by 5 vendors who responded to the RFI. The
  vendors participating in the project are
• Sir Brian Jarman HSMRs
• UHC/Premier developing new joint measure
• Thomson Reuters
• 3M Health Information Systems
• Developed pilot hospital evaluation plan
  recruited 29 hospitals to participate in pilot
  hospital evaluation
• Vendors received hospital discharge dataset
  produced risk adjusted mortality rates on all
  hospitals as well as reports for the subset of
  pilot hospitals that are participating in the
  pilot hospital evaluation phase
• Researchers have reviewed the different vendor
  conceptual models and have begun to assess the
  vendor datasets reports for completeness and to
  prepare them for the analysis
• Continued ongoing coordination with NQF National
  Priorities Partners and IHI efforts to identify
  measures for use

5
Next Steps

• Conduct head-to-head comparison of different
  models (August October 2009)
• Pilot hospital evaluation (September October
  2009)
• Evaluation results recommendations (December
  2009)
• Results to be shared confidentially with
  hospitals prior to publishing on the web (January
  - June 2010)

6
What are the big challenges over the next year?

• Moving from measure evaluation to selection
• No vendor model may be superior in all respects
• Each may have its particular strengths and
  weaknesses
• If no model is superior, there will need to be a
  thoughtful process for determining which criteria
  will be most important to the measure selection
  process
• Moving from measure selection to measure
  implementation
• Once the measure is selected implementation will
  need to address the following
• Stakeholder engagement
• How to support quality improvement efforts
  through management reports and tools
• Assessment of readiness for public reporting and
  what challenges need to be met to proceed with it
  
• When ready, how to present the information to the
  public