Trusting the numbers

Ed Humpherson argues that health and care data needs to improve if it is to fully serve its purpose in driving quality.

Blog post

Published: 30/01/2018

Statistics on health and care matter. They are an essential building block of public confidence in the UK's health and care system and an important tool to drive improved outcomes.

Given the subject’s importance it is right that it occupies a significant amount of my time. A key part of our role as the Office for Statistics Regulation is to highlight how to improve the public value of health and care statistics. Recently there have been two prominent instances where I have had to intervene to support this aim.

Last week I wrote to statisticians in NHS England about changes to A&E statistics. I wanted to see them explain the potential impact of new guidelines on the way A&E data are recorded. The published statistics did not provide information about the impact, or lack of it, on a very prominent set of statistics. I explained that “Changes in the way that statistics are collected, compiled and presented should be clearly explained, reassuring all users that the production and quality of the published data is trustworthy”.

Similarly, in December, my team were following a public debate between Ralf Little and the Secretary of State for Health about mental health workforce statistics. One aspect of the debate revolved around changes in mental health staffing numbers over time. It is clear that the answers differ depending on the definitions used. We felt that it would be better if the statisticians could provide further insight. We asked the statisticians to publish more information to inform future debates on this important issue. They responded by publishing a supplementary analysis alongside their NHS Workforce statistics, which should help people better navigate the complexities involved.

Public debate and holding government to account

Statistics are important – correctly reported, they help stimulate public debate and help to hold government to account. I believe that statistics, well explained, can reflect events really effectively. We are willing to act as an impartial, expert voice to speak out in the public interest. However, statistics about health and care need to improve dramatically if they are to serve the public properly.

In 2015, we identified issues affecting health and care statistics in England, worked with producers and published an agreed direction of travel in early 2016. Since there was no single individual or organisation with clear leadership responsibility for health and care statistics, this has led to problems with the coherence and accessibility of these statistics – they are hard to find and can be confusing when you do find them. I can imagine a curious member of the public giving up in despair if they tried to navigate the data to compare waiting time performance in England, Wales, Scotland and Northern Ireland.

A clearer picture

To galvanise the leadership that the system requires, we organised and led three strategic Round Table meetings and two conferences. We involved several think tanks in these meetings as we fully realise the valuable role they have in using health and care statistics to develop evidence-based analysis. Since undertaking this work we have seen slow, but steady, progress to provide a clearer picture. NHS Digital and ONS have collaborated to publish Cancer Survival in England, NHS Digital’s supplementary analysis of mental health and learning disabilities workforce (noted above) shows responsiveness to public debate and NHS England and NHS Digital are collaborating to improve the timeliness of the annual Hospital Accident and Emergency Activity publication.

However, many users are unaware of all the positive work that statisticians are doing to improve health and social care statistics. This affects confidence in the system and we can see that users are still frustrated with the pace of change.

I am pleased that the Office for National Statistics has committed to lead the continuing improvement of health and care statistics - and not just in England, but also by working collaboratively with statisticians in Scotland, Wales and Northern Ireland.

To make progress, the key is to recognise that serving the public isn’t just to ‘collect and count’, but about statistics providing insight and answering key questions that the public have.

It all comes down to a simple question: who are the statistics for? We believe the answer is simple: It is a fundamental responsibility of the health and care system to describe itself coherently to the citizens who depend on it.