Quality

Diagnostic journey

Goal: to understand a clinical process so we know where the best opportunities are for improving clinical care

Define your goal/aim/measurement.
1. Problem statement (aka 'Elevator speech' speech), which can be used for a project title:
  - "I would have written a shorter letter, but I did not have the time" - Blaise Pascal (often misattributed to Mark Twain)
  - Wreden N. How to Make Your Case in 30 Seconds or Less. Harvard Business Review 2002
  - Example: Our aim is to increase/decrease X
2. AIM statement:
  - Project Charter form example and template
  - Example: 'Our aim is to increase parsable
    
    What happens if the documentation is not parsable?
    
    Can the EHR provide decision support?
    
    Can your EHR digitally communicate your care to another provider?
    
    If you are in a fee-for-service system, can you bill for service that is not documented?
    
    documentation of fulfilling patient choice of
    
    Should patients be able to choose whether to comply?
    
    Surprisingly, almost all Centers for Medicare and Services PQRS measures do not account for patient choice
    
    What happens when compliance might harm a patient? PMID: 19861674
    
    What happens when noncompliance might harm other patients by increasing costs to the cohort? DOI: 10.1007/s11606-013-2645-4
    
    X according to someone's evidence-based guidelines
    
    Whose authority guides you?
    Is there an evidence-based practice guideline available that:
    
    Executed a systematic review of the literature?
    
    Objectively rated the strength of clinical recommendations?
    
    Used formal group decision making such as Delphi method or Nominal Group?
    
    Was without conflict of interest?
    
    by Y% over T time.'
  - Doran's SMART criteria guide development of an aim statement:
    - S=Specific
    - M= Measurable
    - A = Achievable
      For target rates, you can either
      
      Use an arbitrary practical goal of 10% improvement - as done by HealthPeople.gov.
      
      Use an arbitrary 'stretch' goal of 50% improvement; however, stretch goals have been criticized.
      
      Find benchmarks with the list below
      
      Search PubMed for observed improvement from implementations of projects similar to yours
      - Benchmark sources (link to this list): existing quality measures and benchmarks are available. These may help specify your measure.
        
        Dartmouth Atlas
        
        Data.CMS.gov
        
        eCQI Resource Center ( coordinated by the Centers for Medicare & Medicaid Services (CMS) and the Office of the National Coordinator for Health Information Technology (ONC); U.S. Department of Health and Human Services)
        
        HealthyPeople.gov
        
        HEDIS Measures
        
        MIPS (Merit-Based Incentive Payment System)
        
        MN Community Measurement
        
        Hospital compare (see also data.medicare.gov and list at UpToDate
        
        Inpatient measures (AHRQ)
      - Benchmark calculator
    - R = Relevant
    - T = Time (time frame for completion)
    Problem: whether to count patients who refuse or have other barriers can be debated see PMID 18635432
    Problem: the A (target amount of improvement) and T (time-frame) may not be estimable until the diagnostic journey is complete.
Measure your baseline conformity:
- Process control charts template and generator (AKA statistical process control charts or SPC)
- Show details of interpreting an SPC chart
  
  Hide details of interpreting an SPC chart
  1. Are all points within the 3-sigma control limits? If so, the process is in control.
  2. Is there a signal that an effective change has occurred (if a prior QI effort has been done)?
    - Shift 6 or more consecutive points above or below the median
    - Trend - 5 or more consecutively increasing or decreasing points
    - Runs - Are there too many or too few for just common cause variation?
    - Astronomical Point-A dramatically different value
    - From: Run Chart Tool. Institute for Health Care Innovation (IHI)
    - More detail from Wikipedia: Control Charts: Western Electric rules and Nelson rules
Analyze your system for opportunities to improve:
- Non-quantitative analysis
  
  Cause and effect (AKA Fishbone or Ishikawa) diagram
  
  Stakeholders analysis form
  
  Online template QI Tool (free)
  
  Online template LucidChart (registration required)
  
  PowerPoint download
  
  Flow charts and Process maps
  
  Form for clinical site observation
  
  Templates
  
  Online template QI Tool (free)
  
  Online template LucidChart (registration required)
  
  PowerPoint download
  
  Surveys
  
  Morrel-Samuels. Getting the Truth into Workplace Surveys. HBR 2002 PMID 11894677
  
  Morrel-Samuels. Web Surveys Hidden Hazards. HBR 2003 PMID 12858707
- Quantitative analysis
  - Tally sheets (data collection at chart reviews)
    - http://redcap.kumc.edu
  - Pareto charts
    - Templates
      - Online template QI Tool (free)
      - Online template LucidChart (registration required)
      - PowerPoint download

Treatment (remedial) journey

Improve the clinical process.
1. Searching for evidence

Goal: to use existing publications to find the best solutions to the opportunities identified above.

Clinical science
- PubMed searching
- Evidence tables examples/template
Implementation science

Interventions
- Summary of redesign methods (See "Process Re-design. Lecture A" - 25 minute audiovisual presentation)
- Systematic reviews of interventions in implementation science
  - Cochrane Effective Practice and Organisation of Care Group (EPOC) (click 'Full list')
  - SUPPORT Summaries
- Online repositories of studies
  - Knowledge Translation+ (KT+))
  - RX for Change
- PubMed collections
  - Directed at Patients
    - Patient education
    - Motivational interviewing
  - Directed at health care providers
    - Audit and feedback
    - Financial incentives
    - Checklists
    - Flow sheets (flowsheets)
    - Panel management
    - Pharmacists
    - Positive deviance (more detail at WikiDoc)
    - Problem lists
    - Process Control
    - Teamwork
    - Order sets / bundles / checklist (Definition: "Order sets are groups of medical orders that work to standardize diagnosis and treatment following pre-established clinical guidelines or protocols." PMID 22980699)
    - Other EHR-based interventions
  - Other interventions
- Innovations Exchange
Process outcomes (if you need a surrogate outcome that can be measured quicker than waiting on clinical follow-up)
- Intention to change (Selected helpful articles in a PubMed Collection)
2. Designing inventions
- A checklist for making checklists
- Usability testing with heatmaps
- Order set template using Excel
- Measurable documentation (use the Allscripts file and make sure interventions and documentation can be measured by Allscripts)
- Patient education and decision aids
  - Lists of existing decision aids
    - Informed Medical Decisions Foundation
    - Ottawa Hospital Research Institute
  - Templates for creating new decision aids, designed based on empiric research (see also, PubMed collection for Patient education above)
  - Icon array (pictograph) maker by the University of Michigan
  - Improving drug information for patients This interesting, 30 minute video on YouTube is by a pair of researchers presenting at the Institute of Medicine their story of spending 10 years on this project, culminating with their work being incorporated into Federal law. Along the way, their talk describes 1) how to improve the communication of probabilities of risk and benefit to patients, 2) examples of important misinformation in publications
3. Financial projections
- Financial projection templates (KUSMW)
- Financial projection templates (Society for Hospital Medicine) See their link labeled 'ROI Calculations'.
Control the clinical process.
1. Analysis did your interventions make a difference?

Simple before and after comparisons are inadequate.(NEJM 2007. PMID: 17687138)
- Internal validity: for uncontrolled before-after comparisons, use (PMID 21098772):
  - Interpret SPC charts with Western Electric rules and Nelson rules
  - Time series analyses or segmented regression analyses (PMID: 24943919). This is built into the QI Tools at OpenCPU
  - Linear regression analysis to account for confounding variables such as secular trends. This is built into the QI Tools at OpenCPU
  - Regression Discontinuity Design (JAMA. 2020 PMID: 32614409).
  - Additional quasi-experimental designs PMID 16221933
- Benchmarking to other external measures such as CMS data, published reports, etc. Patient mix can affect measurment (PMID: 20823437).
  - Stratified analyses: Romano. Medical Care. 2000. PMID 11021670
  - Risk adjustments: All Patient Refined DRGs (APR-DRG) are a good method among no perfect method

Quality Improvement

Diagnostic journey

What happens if the documentation is not parsable?

Should patients be able to choose whether to comply?

Whose authority guides you?

Treatment (remedial) journey

1. Searching for evidence

Clinical science

Implementation science

Interventions

Process outcomes (if you need a surrogate outcome that can be measured quicker than waiting on clinical follow-up)

2. Designing inventions

3. Financial projections

1. Analysis did your interventions make a difference?

Publishing QI projects