Quality registries in ICU Nicolette de Keizer PhD Dept Medical - - PowerPoint PPT Presentation

Quality registries in ICU

Nicolette de Keizer PhD Dept Medical Informatics Academic Medical Center Amsterdam, The Netherlands

Knowledge

• No standard treatment
• Very expensive
• Shortage of beds

Effective and Efficient?

Evaluating Quality of Care

• Randomized Controlled Trials

– Not ethical

• Benchmarking

– Compare to national average or best ICU – Monitor performance in time – Quality registry

Quality registry

• Continuous data collection
• Predefined data set
• Multiple centers
• Feedback and benchmark

Examples of quality registries

Project Impact e.g. TRACER

State of the art registries

• Many national quality registries
• Some international initiatives

– Project-based

• Overlap in data collection
• Differences in data items and data

definitions

• All have the same goal

Goal of ICU quality registries

• To get insight into quality of intensive care

– Quality indicators

• To foster improvements in the organization

and practice of intensive care

– Benchmarking – Case mix correction – Statistical Process Control – Collaborative network

Goal of ICU quality registries

• To get insight into quality of intensive care

– Quality indicators

• To foster improvements in the organization

and practice of intensive care

– Benchmarking – Case mix correction – Statistical Process Control – Collaborative network

Quality management

Structure Process Outcome

Quality Indicator

“An measurable variable (or characteristic) that is used to determine the level of quality achieved“

Selection of Quality Indicators (1)

• Association with outcome or a process/ structure

indicator that is related to outcome

• Relevant for clinical practice

– frequency

• Results in quality improvement activities
• Easily measurable
• Fast available
• Applicable in several institutions

Example DMV

• Mean duration of mechanical ventilation (DMV)
• f patients admitted at the ICU
• Interpretation:
• Mechanical ventilation?
• Calculation:

1. Mean difference in fractional days between start and end time of (non-) invasive mechanical ventilation during an ICU admission 2. Same as 1, but also including non-ventilated patients (DMV=0) 3. Same as 1, but using calendar days instead of fractional days

Example indicator mean duration

• f mechanical ventilation

fractional days*

• incl. DMV=0*

calendar days* p-value ** DMV (days) 1.95 [0.59] 0.08 [0.13] 3.86 [0.70] <0.001

State of the art indicators

• Many quality indicators
• Ambigously defined
• Indicators should be easily measurable, fast

available, applicable in several institutions

• -> Requires formalisation and automatic

derivation of indicators based on routinely collected data

Goal of ICU quality registries

• To get insight into quality of intensive care

– Quality indicators

• To foster improvements in the organization and

practice of intensive care

– Benchmarking – Case mix correction – Statistical Process Control

Measuring quality of care by indicators

• External benchmark

– Compare to peers

• Internal benchmark

– Compare own performance in time

Measuring quality of care by indicators

• External benchmark

Measuring quality of care by indicators

• External benchmark
• Standardized mortality ratio=
• bserved mortality / expected mortality

Case-mix corrected

Case mix adjustment

• Many different prognostic models:

– APACHEII,III, IV, SAPSII,III, MPM0/24, LODS, SOFA, Euroscore etc. – Logistic regression models – Data of first 24 hours of ICU admission

• Measure performance (discrimination,

calibration, accuracy)

• Recalibrate if needed

Statistical Process Control

• Internal benchmark

– Compare own performance in time – Variation in proces is noise or special cause – Detection rules e.g. 1 point cross UCL/LCL, 8 consequetive points below or under mean

MIE 2011, 30 Aug, Oslo, Norway Reza.Shahpori@AlbertaHealthServices.Ca

22

$2.338.000 $1.344.000 $1.190.000 $574.000 $434.000 $378.000

$0 $500.000 $1.000.000 $1.500.000 $2.000.000 $2.500.000 $0 $50.000 $100.000 $150.000 $200.000 $250.000 $300.000 apr./05 jun./05 aug./05

• kt./05

dec./05 feb./06 apr./06 jun./06 aug./06

• kt./06

dec./06 feb./07 apr./07 jun./07 aug./07

• kt./07

dec./07 feb./08 apr./08 jun./08 aug./08

• kt./08

dec./08 feb./09 apr./09 jun./09 aug./09

• kt./09

dec./09 feb./10 apr./10 jun./10 aug./10 Total Case Cost (Cnd$) / Fiscal Year VAP Case Cost (Cnd$)

Months of the Year

Cost of VAP in Calgary Health Region

Each case of VAP = $14,000 extra expense for Health Care (Safer Healthcare Now)

VAP Cost VAP Total Cost / Year

State of the art SPC

• Many types of charts

– EWMA, CUSUM, RSPRT, P-charts etc – Risk adjusted or not

• Limitly applied in critical care
• Unknown what is the best chart to detect a

decrease or increase in quality

Points of interest / discussion

• Share datasets / indicator sets

– Formalisation – Terminology

• How to move from quality assessment to

quality improvement

– Internal or external benchmarking – Statistical process control – Collaborative networks