Would you recommend a trip to A&E?

Dan Wardle is our survey guru – who better to ask about the ‘recommend’ question now being put about in the NHS as a key measure of user satisfaction. We have in the past been fans of an approach that uses this question and now we need to know how it stands up under this latest glare of publicity. Does Dan recommend the ‘recommend’ question? Over to you, Dan

If you go down to your A&E department today, you could be asked a puzzling question:

“how likely is it that you would recommend this service to friends and family?”

Happy patient (image from bricklink.com)

From this month, NHS trusts are implementing their new friends and Family test, which in essence is using the NetPromoter® recommend question intended to highlight areas of excellence and Trusts, departments and wards that are struggling.

Not #Tripdvisor for the #NHS

The plans were reported in the Guardian at the end of 2012 (New mothers will rate midwives and doctors on childbirth care) and while I am all for transparency and scrutiny, I really struggle to see how this particular patient feedback programme is going to help the NHS.

The story was a source of amusement for the comedians on Mock the Week

“OUCH! I’ve cut off my arm. Hang on a sec while I log onto Trip Advisor to choose a hospital.”

“I’m terribly sorry for your loss. Before George passed away, did he say anything about the food? Do you think he thought it was excellent, very good, or just ok?”

In an emergency, choice doesn’t really come into play – it is the luck of the draw where you are, and external factors such as having a consultant cardiologist in the crowd or your accident on Friday night coinciding with the booze rush-hour. Besides, patients don’t think that there is a choice, so why ask the question? In fact, can patients answer this very question? (more issues are raised in the Guardian: The friends and Family test is unfit for purpose)

Our own experience measuring customer satisfaction and/or willingness to recommend in business, when customers have generally received a helpful service for free, are mostly scores 10 out of 10 and a few 9’s. The fact there were problems in the beginning, that “we got there in the end”, or any event that isnit intended part of the customer experience are not reflected in the score.

The problem with Benchmarking

While poor scores can be used as a “lever” to encourage (or force) improvements, a score alone doesn’t help explain the why (root cause analysis, anyone?). This is why services like the very good Patient Opinion (which is a Trip Advisor for the NHS, but used by the NHS to fix issues) are popular and effective. In the absence of actionable data, perhaps departments will decide it is easier to game the system to boost their scores? And the danger with benchmarking is people become obsessed with Are we on the bottom? No? Well, that’s all right then.

Ultimately, if the average score across the UK hovers around 65 (out of 100), who decides this is a good thing or needs improving?

The guidance for NHS trusts acknowledges that the friends and family test does not replace existing feedback being collected on a local level, and leaves the decision on how to implement the patient survey with each trust. If the results are intended to be collated centrally this makes no sense. Surveys are not neuroscience but it’s easy to accidentally (or deliberately) artificially improve on customer feedback ratings, and this particular test really feels like a case of comparing apples to pears (and all other fruit).

Surveys can be powerful tools and help drive real improvement and (in business speak) “improve the bottom line”. But a saying from my old company won’t go away

“do it right, or don’t do it at all”

You can go home when you’ve filled this form in

When the nurse hands you your feedback form and tells you to fill it in before you can be discharged, tell the truth and just hope it’s not their shift when you come back to have your stitches taken out.

Leave a Reply

Your email address will not be published. Required fields are marked *