Recommendations for a new consent model for patient data are wide of the target
At Picker, our mission is to encourage high quality person-centred care. At the heart of person-centredness is a respect for people’s individual preferences. This means giving people meaningful choices about how the health service treats them and interacts with them, and is crucial to patients having an active, equal role in care. But unless people are given adequate information when asked to make choices, the act of choice can be tokenistic and detrimental both to patients and the service.
In a new report on data security and data sharing, the National Data Guardian recommends the creation of a single opt-out model for the sharing of patient data. The report aims to ensure patient information is held securely and shared appropriately, to support an effective health service and to build public trust. This is a laudable aim and we support the majority of the recommendations – but a single opt out model for consent would have the perverse outcomes of removing the opportunity for patients to make informed choices and threatening service quality.
Quality in the NHS is seen as comprising of patient safety, clinical effectiveness, and patient experience. Each of these can be measured in multiple ways, but patient experience is unique in that it can only be measured from the user’s perspective.
This means the collection of data on patient experience, vital to understanding healthcare quality, can only be done with people’s consent: there is no way to measure experiences using routine, anonymised data. Patients are invited to give their feedback because of its value for assessing and improving quality; many choose to respond because they recognise this value and want to share their experiences. Others choose not to, as is their right – but by contacting people directly, as is done in national surveys, each individual can be given the opportunity to decide for themselves whether or not to participate in a given survey.
Today, the NHS collects patient experience information in incredible quantities. The GP Patient Survey is sent to 850,000 people every year; the NHS inpatient survey to 190,000; the cancer patient experience survey to more than 110,000. The Friends and Family Test is bigger still, gathering responses from more than a million people every month. The surveys need to be so large to provide robust, reliable data to individual NHS organisations, from GP practices to hospital trusts – and the data they collect provides invaluable insight for regulation, service improvement, and public information alike. Due to the large volume of data collection, the vast majority of NHS providers appoint a survey contractor to send questionnaires and reminder letters to patients. To do this, they have to share patients’ names and addresses.
For national surveys, this sharing is supported by a legal provision most often referred to as ‘section 251’. It allows NHS providers to share patient names and addresses with properly qualified contractors for the purposes of mailing out a survey.
In practice, we see little evidence of concern from patients about these arrangements. From contractors, we understand that the rate of complaints received about data sharing is less than 0.005% – for every 20,000 people contacted, less than one will contact the contractor with concerns about confidentiality. Meanwhile, 40-50% of recipients choose to respond to surveys about their experiences. In other words: a large proportion of people are happy to share their feedback about NHS care when asked, and vanishingly few are upset to receive this request – even when it comes from a third party.
Few studies have looked systematically at the impact of opt-out consent mechanisms, although some (including Angus et al (2003) have shown that the use of more conservative opt-in mechanisms result in significantly reduced overall response rates and increased non response bias. Even opt-out mechanisms are likely to substantially impact representativeness, though. For example, we have seen evidence from one trust who implemented a single opt-out consent model at one site: approximately 10% of their users opted-out of all future data sharing. If opt-outs are non-random, as is overwhelmingly likely, then this would inevitably have a major impact on the representativeness of feedback obtained from the 90%. The 10% may have had very different experiences of care, but they are unable to receive information about surveys to let them decide whether they want to share these experiences.
This example tells us two things. Firstly, clearer national standards for data sharing would be useful to eliminate local variation. Secondly, though, the introduction of a single opt-out consent model would be likely to have a serious effect on patient experience collections. This would undermine our understanding of quality across the system, and have a deleterious effect on the role of the patient voice in service management and improvement.
Fundamental to all of this is the issue of trust. The aim of the National Data Guardian’s report is to find ways “to build trust in how the NHS and social care services look after people’s confidential data and use it appropriately”. Inviting people to opt-out of their information being used in perpetuity creates no further opportunity to build trust. People who opt out will never again see a patient survey or an invitation to participate in research. They would have no chance to consider the costs and benefits of participation in specific studies. A one-off opt-out is a specious solution to the challenge of fair processing, and it is contrary to the report’s call – which we endorse – for the case for data sharing to be made.
Of course, consent is important. People should be able to make meaningful, informed choices about all aspects of their healthcare. But making it easy to opt out of all future contacts takes away information and reduces opportunities to make choices. Instead, we should give people more information and more choice. In the case of patient experience surveys, this could be achieved by giving people more information during their care about how they can expect to be asked for the feedback. NHS providers already display posters informing patients about upcoming surveys, but this approach could be extended to give patients individual information leaflets as part of discharge packs or via national campaigns. This should have the dual benefits of improving awareness and encouraging engagement, and would allow patients to opt out of data sharing for specific collections (as they already can) rather than via a damaging blanket opt-out.
We will be presenting this alternative recommendation as part of our response to the formal consultation on the recommendations, which is open until 7th September and available at https://consultations.dh.gov.uk/information/ndg-review-of-data-security-consent-and-opt-outs/consult_view. We encourage others affected by the proposed changes to do the same. An updated national model for data sharing is therefore an opportunity to improve practice and drive informed choice – but great care is needed to get the model right, and a single opt out is too blunt a tool to benefit patients or providers.