Key findings

Beneficial impact

Data sharing has a beneficial impact on groups with protected characteristics or who face inequalities. Providing a national data opt-out may have a positive impact: the opt-out may increase trust in data-sharing and empower patients to be able to make choices about their data. Allowing patients to opt out directly via a national system gives them the ability to set a preference once that will be upheld across the health and care system. However, there may be a negative impact on groups with protected characteristics if opt-out rates are disproportionately high (explored in ‘Issues Raised’ below).

Online system

The use of a simple online system through which patients can set a preference has some advantages over previous opt-out systems, e.g. type 1/type 2 opt-outs which were set at GP practices only. For example, not all groups use GP services or trust GPs, whilst in other cases a GP can unduly influence a patient’s decision. The online system empowers many patients by enabling them to set their preference when and where they choose to. It also mitigates patients’ travelling needs, which can often be a larger burden on those who already face disadvantages e.g. those with disabilities, and can be good for mobile communities and for working people. However, the use of an online system also creates a potential barrier for those who are not digitally-skilled or who do not have easy access to the internet.

Digitally assisted system

Offering a digitally assisted channel (via a contact centre) may support people who do not have digital skills, whose online access is limited or who may find it easier to have a conversation about their choice or to ensure they are fully informed.

Non-digital system

Offering a non-digital route where patients can send identity documents to a contact centre helps to mitigate issues faced by people who do not have mobile phones or internet access, or who may need a proxy to set an opt-out preference on their behalf. However, different solutions need to be developed for a small number of people who may not be able to access identity documents in order to set their preference.

Engagement with stakeholders has highlighted the needs and concerns of particular groups, and these are detailed fully in appendix 1. However, there has also been much commonality in the issues that groups with protected characteristics or who face health inequalities have raised. The broader needs of people across several different groups have therefore been considered, as identified from the research above. These are distilled into the list below:

• people who may need support to access the opt-out system, or to understand the scope and purpose of the opt-out sufficiently in order to make an informed decision
• people who are not able to set their preference digitally due to lack of skills or facilities
• people who do not have a mobile phone or email address, or whose details are out-of-date in the health system, and who may therefore have trouble verifying their identity in order to set a preference online
• people with distrust of health services, or who may be more likely to opt out of data sharing

These issues are explored in greater detail below, whilst a fuller list of findings from engagement and research are presented in Appendix 1.

Accessibility has been a concern for all methods of setting the opt-out (digital/digital-assisted/non-digital), and was raised particularly in conversations with Age UK, Change and Sense. This includes both web accessibility for people with physical disabilities or sensory impairments and accessibility of language for those with learning disabilities or limited English. The system needs to be easy to click through and as accessible as possible through a range of devices, e.g. smart phones and tablets rather than just laptops/desktops.

The online system has sought to meet a number of accessibility criteria and is undertaking a Government Digital Service (GDS) assessment. Steps taken have included bringing in accessibility experts, conducting extensive user research and laboratory-testing the system with users across a range of demographics. The web service has also been developed to meet Level AA of the Web Content Accessibility Guidelines (WCAG) 2.0, which includes success criteria such as the use of assistive technologies. Through the public beta phase, the accessibility of the site has continued to be improved to ensure the system is ready to become fully operational. An accessibility audit was carried out by the Digital Accessibility Centre (DAC) user/technical team in November 2018 and as good practice is embedded across the service the website will be retested so that it is eligible for a Digital Accessibility Centre certification.

This approach to accessibility has informed the design of the national data opt-out throughout: work has been carried out to ensure that the question asked of patients can be understood, and the channels available for opting out are easy to use. The national data opt-out must be understood by vulnerable members of the public, who may often have the greatest health and social care needs, and may therefore be disadvantaged if researchers and care planners do not have access to their health information. The Behavioural Insight Team (BIT) online testing, with 12,500 participants, tested comprehension across a representative sample of the population, considering: whether the national data opt-out should be presented as one or two questions; the exact wording of the question, the information given on the landing pages that people would arrive at when setting a preference; and people’s understanding of the exemptions.  User testing has continued during the public beta phase and changes to the landing pages have been implemented as a result of this testing.

Communications materials for the service have been made accessible and have been produced in multiple formats in order to ensure that people are able to make a truly informed decision. The offering for these materials has been based on recommendations from the NHS Accessible Information Standard, the public sector equality duty and from conversations with stakeholders for this EIA assessment and as part of the public engagement strategy. Accessible materials include producing information in a number of different languages, and in a range of formats, e.g. easy-read, large print, Braille, audio and British Sign Language.

Furthermore, the Data Protection Act 2018 (DPA) and EU-wide requirements set out in the General Data Protection Regulation (GDPR) have brought in strengthened requirements for the provision of transparent information to individuals about how their personal data is used: in particular, the need for information to be intelligible, easily accessible and written in clear and plain language.  To meet these requirements and ensure that people can make an informed decision, clear information about what the opt-out is and how it will be applied has been provided. The Accessibility Workshop held in October 2017, and work following on from this session, including the BIT research, has helped to shape the development of communications materials and to improve our understanding of the best ways of communicating the national data opt-out to all audiences.  As the service becomes fully operational, the EIA will feed into a sustainable communications plan that will help to ensure that communications continue to be targeted effectively and key stakeholders are reached.

Information materials have been made available with the expectation they can be adapted and tailored to suit local needs and will be available through local channels, as well as being hosted on, and available through, national websites. The service has also aligned with initiatives such as the Wellcome Trust’s ‘Understanding Patient Data’, which recognises the complex and confusing language used to describe patient data and seeks to introduce vocabulary which is understandable and meaningful to people.

The service also allows an individual to manage or set a choice on behalf of another individual who may not have the ability to manage their choice independently, provided they can prove that they are permitted to act on behalf of the data subject. Where an individual lacks capacity, or cannot otherwise make a decision for themselves, someone with a formal proxy arrangement can register an opt-out on their behalf.  This includes those with a Lasting Power of Attorney (LPA), a Court Appointed Deputy or parental responsibility.

People setting an opt-out on behalf of their children are required to provide documentary evidence to confirm their own identity and to sign a declaration as evidence of their parental responsibility. Following user feedback during the public beta phase, a person with parental responsibility can set a preference on behalf of multiple children at once to make it easier for entire families to opt out. The requirement to provide evidence of parental responsibility was also reviewed and replaced by a signed declaration in order to make the service more user friendly and to recognise the fact that there is no definitive evidence in the UK to prove an adult has parental responsibility.  This brought the service in line with wider guidance in this area. 

For people with a LPA, the service has ensured that disadvantage is minimised by recognising both types of LPA, i.e. “health and welfare” and “property and financial affairs”, recognising that there tend to be higher numbers seeking LPAs for property and financial affairs.  It should be noted that the setting of a national data opt-out preference is not seen as a decision within the scope of these legal powers but rather is offered as a policy in order to minimise disadvantage.  The proxy is required to confirm their own identity and confirm their role to act on behalf of the other person by sending documentation to the contact centre.

Digital exclusion, relating to both physical internet access and digital skills/confidence, has been a particular concern. For example, a recent study has suggested that 21% (11.5m people) of the UK are classified as not having basic digital skills. Certain groups may be disproportionately disadvantaged by the implementation of an online system for the opt-out: particularly the elderly, homeless people and people with disabilities (e.g. learning disabilities & sensory impairments). For example, Age UK figures show that 3.8 million people aged 65+ have never used the internet. Even amongst those who have used the internet, usage is less frequent than in the general population leading to a lack of familiarity and digital skills. Whilst reports suggest that digital technology use by homeless people can be quite high, they face particular issues through their reliance on mobile phones, e.g. reliance on free public Wi-Fi, accessing charging points or keeping technology safe and dry. Digital exclusion also varies significantly amongst ethnic groups, for example ONS figures show that 10% of people of Pakistani/Bangladeshi ethnicities have never used the internet compared to just 2.6% of Chinese ethnicity.

There is wider work on tackling digital exclusion, such as work carried out by the Good Things Foundation to support disadvantaged groups to get online, or the Nominet Trust’s Digital Reach programme which seeks to increase digital skills amongst disadvantaged young people. The service is linked to the work of the Good Things Foundation through the NHS Digital’s Widening Digital Participation team. This work has led, for example, to the development of 20 testbed projects to understand the barriers to digital uptake amongst excluded groups. However, although this work will undoubtedly help people who are digitally excluded, it is inevitable that some people will remain disadvantaged. Offering alternative methods of setting the opt-out, for example a non-digital route where patients can send identity documents to the contact centre, provides some mitigation and has been recognised as a positive step by the groups engaged with. However, it has long been recognised by the programme that there is likely to be no one method that works for everyone. Through the public beta phase, barriers to accessing the service have continued to be tested and changes made as the system becomes fully operational.

During user research, it was raised that requiring a user to enter their NHS number for identity verification could be a barrier to them accessing the national data opt-out service: user behaviour in the private and public beta phases of the programme showed that people may need to go away to find their NHS number. Subsequently the authentication approach has been changed to allow the entry of either postcode or NHS number - this means that users will not necessarily need to know their NHS number to complete the verification process.

In order to verify the identity of a patient, their details need to match those held on the Personal Demographic Service (PDS). The digital solution’s proposed identity verification requires the possession of a mobile phone or an email address registered on the PDS, which is the national electronic database of NHS patient details. As of December 2017 26% of all PDS patient records did not contain a mobile number or email address for contact details, although this figure is gradually decreasing. However, this issue has the potential to disadvantage the elderly, traveller communities, homeless people and service personnel disproportionately, as these groups may be less likely to have mobile phone/email details on the PDS database. For example, a 2015 OFCOM report shows that mobile phone ownership falls off rapidly with age and for those aged 75+ is just 52%. Additionally, young people may not have their own email address or mobile number registered on PDS so they may have barriers in accessing the online solution. Allowing people to verify their identity through a non-digital route does mitigate some groups’ technological barriers.

However, another concern is the accuracy of the contact information on the database where this information exists. Plans are being developed to provide more comprehensive digital services that will allow patients to update their details directly. However, the most feasible route currently for people to update their details on PDS is by visiting their GP surgery, and having up-to-date contact information is beneficial to patients to support their future health journeys and interactions. It is recognised that relying on the use of GP practices as a mechanism for updating PDS could be seen as a burden for patients and GP staff. However, there are wider clinical and administrative benefits to GP Practices, for example, to be able to send text alerts to patients.

Some groups may encounter particular barriers to verifying their identity, e.g. homeless people who are less likely to have up-to-date information on PDS and may not have identity documents to go through the non-digital route. Whilst the mechanism for updating PDS is via the patient’s GP surgery, Groundswell reported that homeless people are often turned away from GP practices although protocol allows them to be registered under the practice address. However, the non-digital route allows people to set a preference without submitting identity documents if they are able to provide their NHS number.

Nonetheless, a minority of groups may be unable to opt out easily using any standard method, e.g. in detained estates such as prisons, where use of phones or online systems are restricted. Through the public beta phase, the programme has developed and implemented a specific approach for the Detailed and Secure Estate where healthcare staff in the Detained and Secure Estate can set an opt-out for their patients via a secure email route.

Certain groups of patients (such as victims of domestic abuse) have very restricted access to their PDS record in order to protect their location and contact details. Because of the way in which the authentication process works for the online and non-digital processes for setting an opt-out, these channels are not available for these groups of patients to set an opt-out. However, patients who have restricted access to their PDS record should not be disadvantaged in being able to set an opt-out. Therefore, specific processes are being put in place to allow such patients to opt out in a way which protects their location and contact details.  These will be fully operational in 2019.

Sharing information carefully offers the potential to improve health and care services and outcomes, e.g. using research to redress health inequalities, especially in terms of how services can be designed or improved so that they meet the needs of vulnerable groups. Enabling the flow of data for research and system improvement can help to explore both the causes of health inequalities and the solutions. Without such data analysis, it is more difficult to reduce health inequalities which in turn would increase the healthcare and financial costs for individuals and society more broadly.

Whilst some analysis can be conducted using anonymised data, which is unaffected by the opt-out, there are some purposes for which identifiable data may still need to be used, e.g. national clinical audits.  If lots of people from a specific group opt out, commissioners may have less information and therefore less understanding of the needs of a specific group; new services are therefore less likely to be commissioned or are less likely to truly meet the needs of patients. This can contribute to a vicious cycle of dissatisfaction with health services. For example, stakeholders raised the lack of services targeted at ethnic minorities, despite some diseases occurring far more frequently within a particular ethnic group. At the accessibility workshop, it was also raised as a risk that data from people within communities who do not access mainstream services, or who may opt out because they do not trust services (for example, Roma and traveller communities), might be lost for research.

High opt-out rates within particular groups can also disadvantage groups with protected characteristics, as research into conditions or treatments that affect these groups disproportionately or uniquely may subsequently be undermined. Many stakeholders raised the need for increased research within their communities. For example, the transgender community, who historically have distrust of health services (e.g. due to perceived barriers being placed in the way of accessing health care) and would benefit from more targeted studies into long-term effects of hormone usage – current research tends to be based upon people who already produce these hormones in their body.

Engagement work as part of the Equality Impact Assessment has helped to identify groups whose opt-out rates may be high, and the online testing work of the Behavioural Insight Team has also provided further insight into willingness to opt-out by demographics, e.g. by age, gender, ethnicity, region and income level. For example, it was identified in the BIT research that opt-out rates might be higher among ethnic minorities; among people with lower incomes; and within London compared to other regions.  The enhancements planned for the national data opt-out publication will also provide some further evidence of any differential opt-out rates and/or will provide researcher with intelligence to enable them to undertake further analysis in this area.  The programme has undertaken to work with researcher and others to monitor the uptake of the national data opt-out publication and to assess the potential demand for any additional analytical products. 

The need to encourage groups whose participation in public life is “disproportionately low” to be able to participate is written into the public sector equality duty. This can be applied to participation in data sharing, which means that messages about the opt-out need to be accompanied by balanced and accessible information about data sharing so that people can make an informed choice. Therefore, information about the national data opt-out has sat alongside communications on data sharing in the health and care system in order to provide an understanding of why data is shared and to help to set information about the opt-out in the context of a wider narrative around the benefits of data.

The service needs to take reasonable steps to meet the needs of all groups, including producing clear messages, so that all members of the public can make an informed choice about whether or not to share their data for reasons other than their individual care.

A public engagement strategy was developed which summarised how the service is meeting the needs of different segments of the population, identified through the Equality Impact Assessment, the Behavioural Insights Team report and the accessibility workshop. The government response to the National Data Guardian Review set out that the voluntary sector would be involved in “taking forward plans to define, communicate, and implement the new approach to opt-outs to understand fully the needs of all and ensure that communications work for all groups” (4.2). The programme has worked with the voluntary sector who are able to act as advocates for specific groups with whom health and social care find it difficult to engage, or where more support is needed (e.g. where the online or telephone system will not be sufficient to enable everyone to set a preference). Accessible information, tailored advice and tools developed with assistance from the local Healthwatch network have been provided to the voluntary sector to enable them to support people in making their own decisions: information has been shared through the local Healthwatch network and also is available online. The accessibility workshop also identified that there are many people who do not access health services who need to know about the opt-out, meaning that communication needs to be wider than across health services.  As the service becomes fully operational, a sustainable communications strategy will be developed to ensure that further communications are targeted effectively.

It is recognised that some people may fit into multiple groups and may therefore face multiple health inequalities. Those who are socially excluded, e.g. people with disabilities, are often the most digitally excluded. Mental ill health is significantly over represented in the homeless population as compared with the general public; with 80% of people experiencing homelessness reporting mental health issues, and 45% having a clinical diagnosis. Age is another such factor. A number of challenges with regard to elderly people being digitally excluded or less likely to possess a mobile phone/email address have been raised in this assessment, whilst it is also recognised that people with limited English would also face particular challenges in setting a preference. However, whether or not a person can speak English is strongly related to age: across all ethnic groups, those aged 65 and over are least likely to speak English and those aged 16 to 24 are most likely.

The government’s Race Disparity Audit, published in October 2017, also illustrates “disparities between ethnic groups in all areas of life affected by public organisations” – this illustrates that ethnic minorities may be over-represented in some of the groups who may find it difficult to set a preference. For example, Black Caribbean adults are the most likely to have been detained under the Mental Health Act, and are over-represented among the homeless and within the prison system. It is therefore important to ensure that those who are disadvantaged by multiple circumstances are considered, rather than solely considering the groups with protected characteristics in isolation. By focusing on the broader needs identified across many groups as well as those raised individually, the service has sought to ensure that the mitigating actions identified work for as many people as possible, including those who are disadvantaged by multiple circumstances.