Taking much better care of health data | @guardianletters

The proposal to website link hospital and common practice data via the intended care.information technique has created intense discussion (Care.data is in chaos, 1 March). This has so far been confined to issues above inappropriate industrial exploitation of the data and leakage of confidential data.

Although these are essential elements, we also have concerns relating to what occurs when information are not linked accurately. There is escalating global evidence that the inevitable mistakes happening in the course of information linkage can distort sorts of analyses that care.information aims to support. Information such as NHS number and date of birth, which are utilized to link information, are never ever best and typically it is specific kinds of men and women, for instance ethnic minorities, who fail to get linked and hence fall outside the program. Failing to hyperlink information for the identical person or wrongly linking different people can make critically misleading benefits, even when only a modest minority are wrongly linked. More transparency about the nature and extent of linkage processes and linkage error would support health care researchers assess prospective distortions and aid services providers to enhance information good quality.

As for confidentiality, this is best preserved by proper monitoring of patient records, with rigid safety controls on accessibility. Proposals to “scramble” patient identifiers prior to data leave the GP record programs are not the remedy, and would in fact make issues worse by rising the numbers of wrongly matched data. In our view, the current debate needs to contain a total discussion of all these linkage quality difficulties.
Professor Harvey Goldstein University College London &amp University of Bristol, Professor Ruth Gilbert University University London, Dr Katie Harron University University London, Dr Gareth Hagger-Johnson University College London, Dr Mario Cortina University College London , Dr Nirupa Dattani City University

• There is a basic step which can be taken to address the care.data controversy. The Health and Social Care Info Centre must give a public undertaking that it will supply only analysis of GP patient information to outdoors bodies. Patient data would by no means be disclosed to outdoors bodies. Ideally this need to be formalised by statute and utilized to all patient data gathered by NHS companies. Any examination requested could be subject to ethics committee overview and, when total, identified on the HSCIC site. These methods, swiftly taken and appropriately publicised, may well assistance public acceptance of the creation of a very worthwhile national healthcare asset.
Malcolm Rogan
Nomansland, Wiltshire

• Your piece on use of patient record data (MPs’ anger at missing information on who has witnessed patient data, 26 February) quotes Dr Stephanie Bown as saying that GPs “be concerned that patients’ considerations about care.data could avoid them from speaking openly to their medical doctor”. This kind of worries already do so, every time an individual is looking for daily life insurance. The insistence of the insurance coverage sector on an applicant’s agreeing to their GP’s report frequently prospects to reluctance on the element of sufferers to share symptoms they believe might affect their premiums. Related damaging reluctance to engage in healthcare could precede holidays when travel insurance coverage cover depends on absence of pending hospital appointments.

The public advantage that will follow profitable implementation of the care records program is most likely to be huge and the challenge to confidentiality nugatory. By contrast, there is no identifiable basic public advantage in the long-established collusion between the insurance coverage business and the NHS. It should be banned.
Professor Robert Boyd
Adlington, Cheshire

• Bettering health companies, as far as the government is concerned, will also suggest cutting costs by delaying treatment method, minimizing eligibility for pricey drugs or operations, increasing charges, and furthering privatisation. The care.data computer method would enable it to keep track of whether medical doctors and nurses carry out central instructions. Whether or not it would actually be much use for evaluating therapy – given that so many elements influence outcomes, such as way of life and whether or not individuals actually consider all the medication they are prescribed – is another query when it comes to justifying the cost of but an additional large personal computer method. The Workplace for National Statistics is trusted and has wonderful encounter in analysing the census and other huge data sets. It would be much preferred to any private firm. Otherwise the income might be far better spent by the Health-related Research Council.
Dr Richard Turner
Harrogate

Taking much better care of health data | @guardianletters