PDCA: Check
Check Overview
Did your implemented solution achieve the Aim Statement?
Check is the proof which proclaims success was attained. It also prepares for the ongoing measurement and adjustment of the change to assure it is sustained.
Data Don'ts (12:30)
Facilitated by Wayne Fischer, manager, the Institute for Healthcare Excellence.
Guidelines on Check
Check is the evaluation of the implemented solution, either the success of the pilot or the entire implementation. Perform the same measurements used to determine the baseline measures so you can determine if you reached the goal in your Aim Statement.
- The ideal is to collect a series of data points on a run chart such as eight weeks of average data for the baseline and the same measurement series to determine sustained improvement
- Hold off drawing a conclusion of the effectiveness or the solution until the new way is settled and producing at peak level
- If a baseline measure was not taken, then conduct an implementation post survey. In short, is it better after the implementation?
- Continue taking measurements (of the Aim Statement) to prove the gain from the solution is sustained over time
Tools to Download on Check
Here are tools to download on Check:
- Before and After Measures (doc) - Rule of Thumb for baseline and post measures
- Signs of Good Measure (doc) - Consideration in selecting a measure
- Metric Evaluation (doc) - Filtering questions in evaluating a metric
- X-Bar Chart (xls) - X-Bar Chart instructions, data entry, graph
Cautions on Check
- Be sure the metric to track your solution encompasses the Aim Statement
- Make sure the metric is specific, sensitive, targeted, focused
- Take the post measure after the installed solution has had time to normalize
- Avoid taking only aggregate measurements that can hide trends and variation that better show the change over time
- To graphically analyze the data, use the proper chart based on the type of data being measured
- Don't draw an inference on less than 10 data points. In fact, 30 to 50 data points is recommended
- Beware of sampling bias. Random samples are best
- Be aware of influential factors outside the study that could affect the data. Bad data collection leads to wasted time and questionable conclusions
- Don't display data without footnoting their specifics such as source, collection period, sample size and technique
Action Items on Check
- If this measures the pilot, re-scale the measure or cover the roll out out in its entirety
Remember that rolled up global (summary of all clinics) measures are not as sensitive nor diagnostic as local ones (per clinic). - Adjust the outcome and internal process metrics if the solution goes varies from the original Aim Statement
- Ready the measurement for long term reporting. Use statistical control charts where possible
Special or Common Cause (6:02)
Review the difference between a special or a common cause. Treating a special cause like a common cause adds unwanted variation to a system.
- Define the frequency, distribution and thresholds for each metric for the long haul. Consult Metric Evaluation (doc)
To continue instruction, go to PDCA: Act.