Is anyone doing great stuff with friends and family data? Especially the free text comments which are by far the richest part of it. There are over one million pieces of feedback completed by patients each month.
This has been building up for a number of years now at Trusts and GP surgeries. We explored text analysis and thematic coding a couple of years ago, but we’ve dived in again head first. A meeting with the folks at NHS England this week confirmed to us that loads of help is needed and will be welcomed. The time is ripe, patient experiences need to heard and the latest text anaylsis tools are in a great place to handle the data.
Sponsored by a visionary trust, we are doing human-led, machine-trained thematic coding of patient comments and sentiment analysis and this will lead to loads of great stuff.
Here’s our starter list of challenges:
- How does the shape and nature of comments change through time and what are the trends that can’t be seen in fragmented reading of comments?
- What are the variations in themes by ward/hospital/surgery/demographic?
- What can linguistic and discourse analysis of text tell us about the nature of patient/service relationships?
- What’s the relative importance of these macro and micro themes (the nature and personality of your care , your opinion of staff, the facilities, hygiene, speed, efficiency,…) on recommendation likelihood and on actual sentiment?
- What about an easy interface to allow anyone to pull out specific comments by theme and to link to improvement initiatives?
- Can we come up with better designed posters and communications back to staff and patients of the best comments?
- Can we easily and automatically identify serious comments, concerns and complaints into a review pool?
- And similarly gather the positive feedback and named mentions of staff to get them fed back as soon as possible