Professor Alan Maynard on data use, incentives, PROMs, private sector HES and a 1974 Burlington.

The reformed NHS has some “nice” data gaps which need to be plugged quickly, if performance management and transparency are to be improved.

Without improvements in data collection and analysis, NHS managers will not be able to manage efficiently and the NHS Commissioning Board will not be able to commission efficiently and equitably.

In the hospital system, Hospital Episode Statistics (HES) are being used extensively by the research community. Its members now recognise that these data are of good quality and offer significant publication and employment opportunities to academics!

Complexity, data and incentives
However this work raises some nice issues about dissemination. The quantitative analyses by academics at places such as the LSE, Imperial, Manchester and York are complex. Highly pertinent results from this work for NHS managers and clinician leaders may take time to affect minds and policy.

This is a result of perverse incentives. Managers and clinicians may not have the skills and the openness of mind to access and use such material

To market, to market?
Academics are rewarded for publications in peer-reviewed journals, not for trotting round the NHS and selling the results of their labours! Both parties need to do better, perhaps?

The nice related issue about HES is how to increase its use in the NHS. For instance hospitals are continually investing in ‘initiatives’ to reduce admissions. By using “before and after” data analysis with controls (commonly called “differences in differences analysis” in the trade!), the effects of such investments could be evaluated. Why is the NHS so slow in exploiting these opportunities?

With good data available, it is frustrating for all that investments such as these are either not evaluated or subject to review based on opinion!

Those investing in such policies are potentially wasting NHS resources. They should be required to demonstrate value for money; both to vindicate their behaviour, and to inform colleagues about what works and what does not work. To do this, they must invest in analytical capacity.

The HES gives detailed information about patients flowing through NHS hospitals. It is time the Department of Health required the collection of similar data for private hospitals.

With increasing numbers of private providers, it is extraordinarily inefficient that integrated NHS and private HES data is not yet compulsory.

Hopefully, the Department of Health will require the collection of HES data by the private sector soon. Why is government so slow in remedying this gross omission?

Equally hopefully, the NHS Commissioning Board will focus on data gaps in primary care - since it has to commission the work of GPs and their teams.

’Improvements in NHS activity data have been significant in the last decade. Comparative outcome data remains poor.‘

GP commissioning consortia may begin to use more fully and aggressively the HES data on hospital referrals and the prescribing data.

Each GP gets information about their relative prescribing activity regularly. Can this be better managed in relation to the evidence base? The HES gives data on referrals - and again, there may be unused potential in using this data to performance-manage GPs.

However the Commissioning Board will also need more information about what GPs do e.g. the numbers they treat and their disease categories. Who knows but the Board may also press for IT to replace those lovely brown envelopes in which records are kept; which were introduced following the Lloyd George 1911 National Insurance Act!

Any such investments should be targeted at producing a national system of data collection, and avoid perpetuating islands of excellence in a sea of mediocrity

Liberating the nurses
Lying beneath these data issues are major policy choices around skill mix. In 1974 the results of the “Burlington experiment” in Ontario were published in the New England Journal of Medicine and Annals of Internal Medicine. This randomised controlled trial showed that nurse practitioners were equally as good at providing patient care as physicians.

In times of “austerity”, nurse substitution in primary care may be attractive financially and offer improved access to care of good quality. Without good data and analytical capacity such changes in skill mix will not be evaluated well. That would be criminal!

Improvements in NHS activity data have been significant in the last decade. Comparative outcome data remains poor. Patient-reported outcomes data (PROMs) is expanding now, and will offer some much-needed insights into relative outcome performance at the clinician and hospital level.

These and HES and GP data improvements are costly and will need to be exploited fully by managers and academics.

As ever, the price of knowledge is high, but the cost of ignorance may be even greater! With budgets highly constrained for the next four years, managers, clinical and non-clinical, cannot afford to ignore the potential of investing in better data collection and its analysis.