This site is intended for health professionals only

At the heart of general practice since 1960

Should medical researchers get access to patients' care records?

With proper safeguards, access to patient records can open up a new era of medical research in which GPs take the lead rather than relying on academics and hospital doctors, argues Professor Frank Sullivan. But Dr Mark McCartney argues that using confidential personal details risks undermining the doctor-patient relationship

With proper safeguards, access to patient records can open up a new era of medical research in which GPs take the lead rather than relying on academics and hospital doctors, argues Professor Frank Sullivan. But Dr Mark McCartney argues that using confidential personal details risks undermining the doctor-patient relationship


As GPs we often need better evidence to inform our decisions. Until recently we have had to rely on laboratories and hospitals to provide that evidence, but that need no longer be the case.

Some of the evidence we need is already in the practice - locked away in our computer systems. As electronic patient records become ubiquitous, they also become a unique and valuable resource for research that could improve the care we deliver. This is particularly important because the information held in GP records is representative of the whole population, so research findings based on our data is often relevant across the UK and in similar populations elsewhere. The observations we make also can also, potentially, help to improve understanding of disease and determine the safety and effectiveness of new treatments used by our hospital colleagues.

I say potentially because there has been uncertainty, and a lack of consistency, about the processes that should be used to ensure that researcher access to electronic patient records is conducted in a way we and our patients would consider appropriate. The safeguards most GPs and members of the public believe are reasonable include the need to check that the research is asking an important, relevant question, that data will be held securely and that patients are approached only by someone who has clinical responsibility for their care.

The academics and GPs I work with use standard operating procedures to ensure that work undertaken in the Scottish primary care research network adheres to these standards. A number of studies our research group has been involved in recently have required access to GP prescribing data. These have included a study aiming to identify adolescents with diabetes who could have their insulin stopped when a genetic defect is corrected with a small daily dose of a sulphonylurea, and a study showing the increased mortality in 50% of all women with breast cancer treated with tamoxifen who stopped treatment within two-and-a-half years.

The Wellcome Trust has recently developed a consensus statement for best practice in consultation with many GPs, academics and patients. The statement has been endorsed by the RCGP1 and has also been sent for consideration to the BMA, primary care computer suppliers, the Academy of Medical Sciences, Connecting for Health's research capability programme and a range of other interested parties.

During the past 30 years we have seen GPs take control of our own postgraduate education and we are playing an increasingly important role in undergraduate education. Should we still rely on basic scientists and hospital colleagues to tell us how to investigate and treat our patients based on evidence produced in labs and hospitals? Or can we make research something that is expected, valued and rewarded in everyday practice?

As long as adequate safeguards are provided for data extraction and processing, I believe GPs should facilitate access to records. It works for extraction of QOF data through the Quality Management and Analysis System (QMAS). There is no reason to suppose we can't make it work for other purposes such as research that will help us make the best decisions for our patients.

Professor Frank Sullivan is director of the Scottish school of primary care and a GP in Dundee


As a nation we have become obsessed with numbers and statistics - not a day passes without some meaningless survey being published to promote one or other side of an argument. Some of these are based on biased questions and may not even be statistically valid.

The opportunity to perform large-scale scientific research on the vast array of data in patients' confidential medical records is already proving too tempting for many, including the pharmaceutical industry, which is keen to make a quick buck, and the Government itself, which may use the idea to find a way to pay for the immense expensive white elephant that is the National Programme for IT.

I have serious concerns about the use of confidential medical records in this way. These records have been created in the delivery of patient care, not for research purposes. Often patients will talk about matters of an intimate and personal nature which they do not wish to have stored on a record. The quality of the information on the records may not even be adequate for reliable research. How can we trust research when there will be no way of drilling down to the exact data on which it is based?

The central Summary Care Record and Detailed Care Record plans are still the subject of great concern to many patients and GPs. The use of so-called sealed envelopes to protect highly sensitive information has not been proven to be effective or workable. It is also not clear how research using central records would overcome the problems presented by information held in sealed envelopes, if they can ever be made to work.

Connecting for Health has suggested that anonymisation or pseudonymisation of the record prior to secondary use for research could overcome the need to get informed consent from patients.

I have concerns about this because the anonymisation process has not been validated and it may be possible to identify patients from details such as postcodes or combinations of supposedly anonymous data.

The potential breach of confidentiality by the use of centralised medical records for research purposes is high and the routine use of these records will undermine the GP-patient relationship. After research, where will the use of confidential patient information end?

The Government wants to access medical details for other reasons too, including crime, counter-terrorism and the investigation of tax or benefit fraud.

This Government has a very poor record on data protection and we should resist further attempts to ride roughshod over patients' concerns. The health select committee was critical of the National Programme for IT in this respect and suggested that the issues of consent and pseudonymisation needed to be addressed urgently2. There is no evidence of this happening yet. There is no strong independent organisation to protect the interests of patients.

I support the use of robust scientific data for the delivery of evidence-based medicine and I can see why many researchers are keen on using electronic patient records to achieve this end. But it would not be right for the Government to decide to sell off the data from our medical records to raise money and supply information that would be financially advantageous to the big pharmaceutical companies.

Our medical records were created for the purpose of aiding medical care and should not be treated primarily as a research tool. Any further use should require individual patient consent and the guarantee of a robust and effective anonymisation process, supported by a strong independent patient advocacy organisation.

Dr Mark McCartney is a GP in Pensilva, Cornwall, and a vocal critic of the National Programme for IT

Should medical researchers get access to patients' care records? yes

Some of the evidence we need is already there, locked away in our computer systems


We should resist further attempts by the Government to ride roughshod over our patients

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say