Thread by @RammyaMathew

1. Measurement is a fundamental part of quality improvement in healthcare.

2. Here’s a #tweetorial on measurement for improvement, mostly addressing common pitfalls and how to avoid them

🌿Inspired by this excellent paper from @MaryDixonWoods and team

qualitysafety.bmj.com/content/30/2/1…

#QI #quality #meded @TheQCommunity @MedTweetorials

3. Measurement IS an essential part of any improvement project because:

The only way to know if a change in an improvement is by measuring the impact of it

4. There are 3 basic types of measures:
🪂Outcome measures–demonstrate the impact of your intervention
🪂Process measures–identify if your change interventions are happening as planned
🪂Balancing measures-determine if your intervention is having any unintended consequences

5. Here's an example to demonstrate the different types of measures -->

6. As part of your PLAN (Remember Plan-Do-Study-Act) – you should decide on your measures and have a clear measurement implementation plan

7. 🛑Common pitfalls to avoid include:

8. Where possible, try and choose measures which are routinely collected, at least for your outcome measures

If choosing ex novo outcome measures, ask yourself why.

9.🟠Weigh up the need to capture these outcomes vs the extra work required

🟠You likely have a fixed amount of resource (people/time/money) to do your project.

🟠So > resource required to do measurement, equates to < resource available to plan and implement your change ideas

10. If you intend to do random sampling to capture your data – outline how this is going to be done so that your methods are consistent and there is less room for measurement bias.

11. 🟠You also need to determine the frequency of measurement.

🟠Having fewer measures, which you collect & review frequently will allow you to be more responsive to the data

12. Next, remember that for the purpose of identifying whether a change is an improvement, you need to have sufficient baseline data (usually a minimum of 6 data points prior to implementing your change)

13. Sometimes baseline data can be collected retrospectively, but not always – another thing to be aware of, if choosing ex novo measures

14. 🪂To ensure data collection is as accurate as possible, you can use templates/proformas/protocols etc – but the difficult part is ensuring that these get used

🪂IT solutions can be helpful.

15. Using the bowel cancer screening example, you can use programmes that enter a code in the patients digital GP record when SMS reminders are sent.

This can then be searched for electronically to identify the no. of patients who received a SMS and then took up screening

16. But no method is fool proof – so try & validate your data collection method & determine what degree of error you are willing to accept

17. Lastly, it’s easier than you might imagine to invest in measurement, but then fail to review or act on what the data is telling you.

Build in time as a team to do this, using the data to decide whether to adopt, adapt or abandon your change ideas.

@threadreaderapp unroll