Putting patients in control of data from electronic health recordsBMJ 2018; 360 doi: https://doi.org/10.1136/bmj.j5554 (Published 02 January 2018) Cite this as: BMJ 2018;360:j5554
All rapid responses
Oops! The Bmj has slipped badly in publishing this ignorant article that is full of basic errors, perhaps not surprisingly considering that the authors fail to include any GP’s or any epidemiologists, or any European author despite making much of the North European example of shared health records.
First of all, ‘primary care has used computerised record keeping since 2000’! Really ! We stopped using paper records in 1990, as did most GP practices. Ironically, considering the subject of the article, systems such as VAMP were offered very cheaply to practices by firms hoping to sell the data to research and big
Secondly they fail to mention that all GP prescriptions in Denmark have been on a Central Database for well over 10 years as one can see from the article in the current New England Journal which is discussed by Richard Lehman in his current BMJ blog. This data, combined with their population register and good disease and mortality data has produced a wealth of useful conclusions on drug safety, for example on oral contraceptives, HRT, SSRI’s in pregnancy etc, etc.
Since 2009, Danes have had access via www.sundhed.dk (some of it is in English) to their medication records , and hospital discharge letters. They can also use it to book appointments.
This, as in the swedish system is an ‘opt out’ system because anything else would be hugely costly and produce a poor result.
The article asserts that opt in systems are better but produces no evidence that this has benefits. It also says that QOF has improved diagnostic coding, although the evidence (and my experience) is that for non QOF diseases, this has not improved.
We need to follow the Danish and Swedish example, and not waste time and effort in local schemes that depend on drug company funding and ignore the basic problem that around 10% of people move house every year. If we avoid the pitfalls of allowing access to personal information, and allowing big pharma to access the data, it could be a huge advance.
I expect their will be Scandinavian readers who will offer their own responses, if they think this effort deserves a response. It might be better if the BMJ Commissions someone who is better informed to write an article describing their systems. We might even have a sort of Danish Swedish rivalry as in ‘The Bridge’ now showing on BBC. ( Who have learned to borrow from them! )
Competing interests: No competing interests
Thankyou John New et al for raising further awareness on this important issue of sharing electronic health records. Health systems should share records with healthcare providers and with patients and nominated family and carers. Your own experience of patients opting-in to data sharing for research purposes required time and staff to explain this. As they gain a better understanding of the needs, in general patients were reassured. Improving outcomes requires more activated patients and staff working in organisations that value and invest in wider health literacy in the neighbourhood and community. Over 89% of families now have access to the internet and hence the opportunity to become active partners in their care. We call this "Responsible Sharing". Our own experience where 83% of our type 1 diabetes patients (and generally 62% of all patients http://www.htmc.co.uk/pages/pv.asp?p=htmc0328) now have full access to their GP electronic health record and can share parts of it with whom they like whenever the need arises points to what is possible today. Recently we reached a significant milestone where over 10,000 patients completed questionnaires for signing up for this in one GP practice http://www.htmc.co.uk/pages/pv.asp?p=htmc0711 You can learn more about what such patients do and what "responsible sharing" means here http://www.htmc.co.uk/pages/pv.asp?p=htmc0701 Given the current #NHSCrisis, how much longer will it take before we realise we need to work together with patients and staff to help families and communities do more for themselves and support the research agenda too. Patients should have access to electronic health records and understanding in order to provide explicit consent for data sharing with other organisations where possible and organisations need to be supported to make this a reality.
Competing interests: I have lectured and written numerous articles, been on TV and the popular press describing our experience of giving patients access to their records and understanding as a full-time General Practitioner and what this can mean for the delivery of care
We found that patients found it simple to say which pieces of data they would not want to be shared when they were given two complete copies of their records and correspondence. Patients were asked to mark errors in blue pen and data that they would not want shared in red pen. The results can be seen in the paper "Sharing electronic health records: the patient view" By John Powell, Richard Fitton, Caroline Fitton in "Informatics in Primary care b2006 14.55-7 PHCSG British Computer Society. We produced algorithms and a patient portal that allowed patients to direct permissions to access to data in chapters/folders of their records which had ben created from sorting the Read/ICD classiification codes into their main chapter headings (Infectious diseases, Growths, Endocrine, Blood disorders, mental health, Neurlogical, Cardiovascular, Respiratory, Genitourinary (includes breast and gynaecology), Pregnancy, skin, musculoskeletal, Neonatal, Accident and injury.
Generally patients selected sexual (includind STDS), pregnancy, mental health including drug and aldcohl problems, employment issues, genetics and stigmatising social data. I believe patients in the future wil see the data at source and state which data they dont want shared.
Here is the paper:
Sharing electronic health records: the patient view.
Dr John Powell
Dr Richard Fitton
The introduction of electronic health records to the UK NHS has raised public and professional concerns with issues of data accuracy, security and confidentiality.[ref] Such issues are not unique to electronic records, but have been highlighted by the intention to create a nationally held spine of information which will be contributed to by existing primary care records.[ref] There are fears that inaccurate information could be uploaded to the central database and previous work has shown that existing primary care records do contain inaccuracies.[ref:pyper] There are also fears that there could be illegitimate access to personal information, either due to professional misconduct by a health worker or caused by a malicious hacker attack.[ref] In the UK there is debate over whether the consent to the sharing of confidential information on the national database should be implied, but with the choice to “opt-out”, or whether explicit consent to “opt in” should be obtained from every individual.[ref] A 2002 report commissioned by the NHS Information Authority showed that more than a third of patients would like to be consulted on every occasion that their details are shared.[ref] There has been discussion over whether the enormous practical difficulties that this would cause should influence the ethical arguments around patient confidentiality in the new era of national electronic records.[ref] One initiative designed to safeguard against the unwanted sharing of sensitive information without explicit consent is the proposed provision of electronic “sealed envelopes”.[ref] Individuals can opt for certain information which they do not want to be routinely shared to be placed in these envelopes. There is speculation that the information that people would not want to share would include records of mental health and sexual health issues,[ref] but there is a dearth of research evidence to guide policymakers in this area. We have therefore analysed an audit of electronic health records held in one general practice to establish how many patients would want to restrict access to some or all of their records, and what the nature of this information would be.
Fifty consecutive attenders to one general practitioner [RF] who agreed to take part were given a paper printout of their full primary care electronic health record. Patients were excluded if they had a severe acute medical or psychological problem, or were unable to complete the task, as judged by the practitioner. Each patient was given an information sheet about the project and two coloured pens. They were asked to highlight information which they considered to be incorrect with one colour, and information that they would not want to be shared on the national electronic database of records with the other colour. For the latter task they were told “At the moment your record is only stored at this Medical Centre. In the near future parts of your record will be copied to the National Care Record. This will contain medical details of all National Health Service patients within a central database. Please circle or underline the parts which you would not want to be shared with the National Care Record.”
There was a 62% response rate (31/50). The response was 54% for males (14/26) and 71% for females (17/24). The mean age of respondents was 35. Five of the 31 patients (16%) identified information that they would not want to be shared on the national record system. The items identified by them as not wishing to be shared are summarised below.
Items patients did not want shared on the national record.
Five patients marked data items that they did not wish to be shared to the National Care Record. One hundred and ten items were marked. The majority of items identified were related to mental health, sexual health and genito-urinary problems, The items concerned;
emergency contraception, dietician appointment for advice to lose weight, contraception, cervical smear appointments and findings, stress incontinence, vaginal discharge, life insurance, anxiety, asthma, overdose attempt, termination of pregnancy, medication history, lack of libido, erectile dysfunction, impotence, referral to sex therapy, overdose attempt, depression, anxiety, personal problems, being signed off work.
Items that patients marked as incorrect
Ten respondents (32%) each identified one piece of incorrect information in their records. Further investigation showed that three of these ‘inaccuracies’ were disagreements by patients over diagnostic labels given to them, but the data had been recorded as intended by the practitioner. (Deleted diagnoses here) One inaccuracy was of a patient believing they had a drug sensitivity when the notes identified them as having no sensitivities. The errors identified by the other six respondents were audited by cross-checking with paper records and correspondence, and all were found to be due to simple errors in manual data transcription or data entry. Three related to incorrect dates. The other three concerned incorrect clinical information.
The primary aim of this study was to identify the extent to which primary care patients will allow their local electronic record data to be shared on a national database. Eighty-four percent of the patients in this study were happy to have their whole record shared. Of the five patients who felt that there was at least some information in their primary care record that they would not want to be shared, the items they identified related almost entirely to matters of pregnancy, contraception, sexual health and mental health. This fits with the assumption that matters related to sensitive or embarrassing issues are most likely to be censored by patients.[ref] However previous work has tended to ask hypothetical questions concerning data sharing rather than examine a real situation.[ref]
A secondary finding of this work was to support the results of other studies showing that existing records contain errors and these will be duplicated if they are to be used to populate the new national spine. Giving patients access to their records is a useful way of identifying errors that have occurred through problems with manual data transcription or data entry.
It is important to note that this was a preliminary study on a small number of patients from one general practice. The sample was not random but consecutive, and acutely unwell or distressed patients were excluded. The results should be interpreted accordingly and not treated as representative of the situation across UK primary care. They are intended to highlight some issues with have to date been under-researched and which require further work to guide policymakers making key decisions over the practicalities and ethics in the sharing of personal data in national electronic health records, including the decision to have an “opt out” or “opt in” system. A larger study of a representative sample of patients across several practices is needed to further investigate issues of data sharing and consent.
Competing interests: No competing interests