FEATURE25 November 2015

The missing link

Features UK

Linking answers from surveys with administrative data sets – like your NHS records or Student Loans information – is becoming increasingly common practice, and promises to hugely enrich the data we collect.


At the same time smaller research budgets, particularly in the public sector, mean there is more pressure to extract as much value as possible from existing data sets.

NatCen Social Research delivers a number of national surveys, mostly for government and academic clients, like the Health Survey for England, the Family Resources Survey and the English Housing Survey. All these studies ask participants for consent to data linkage of some sort.

For example, by linking data on fruit and vegetable consumption from Health Survey for England to health and mortality records, we have been able to show that eating your five-a-day is correlated with positive long-term health outcomes.

Getting consent

Consent rates vary substantially across studies, depending on how data is collected and which datasets are involved. For example, consent rates are highest when data is collected face-to-face, and rates tend to be lower for more sensitive information like criminal records.

For example, 80% of adults taking part in the Health Survey for England consent to linking their answers to hospital episodes stats and mortality data. While for a similar study on finance, only 65% of people agreed to linking to records held by the Department for Work and Pensions. Both studies are carried out using face-to-face interviews.

What’s more, the number of data linkage consents requested is multiplying. We are currently running the fieldwork for a major national study which asks for linkage to ten different data sets, including criminal records.

With data linkage becoming an increasingly important part of our work, we’ve conducted various pieces of qualitative and development work to understand the process better. As you might expect, certain members of the public, while willing to take part in an anonymous questionnaire, stop short of consenting to data linkage. Concerns about data security are regularly cited as the key reason for not granting permission.

Understanding the consent process is a complicated business. Here’s how it goes: we take a few pieces of information from your survey data which identify you, like your name and date of birth; alongside a unique identifier, these are passed to the body which holds the data; they match your records, destroy the identifying data and pass the linked data set back. Still with me? In the face of complicated explanations people often shut down.

Our qualitative work plotted people’s attitudes to consent on two scales: comprehension of the process, and willingness to consent. Fortunately it indicated that higher comprehension meant people were more likely to agree, which means efforts to explain the process clearly are worth it. Worryingly, from an ethical point-of-view, there are a group of people who do not fully understand what they are signing up to, but are still willing to consent.

Another issue we explored was the move to obtaining consent to data linkage online. In the past many government surveys have been carried out by an interviewer face-to-face or over the phone, so the rapport with a participant has played an important role in securing consent. Without the human touch it becomes more difficult. In a recent national, longitudinal study carried out online, only 47% of participants agreed to linking their NHS records.

Ethical implications

The move online also has ethical implications. There are serious questions around whether we should be asking people to sign away access to their criminal records at the click of a button. Of course, we always send a letter and email to confirm what people have agreed to, letting them know they can change their mind at any time, but is this really enough? Indeed, participants in our qualitative interviews had mixed feelings about consent at the click of a button.

This is why we have created an animation – to help simplify the issue for participants. We already create easy-to-understand leaflets with clear diagrams and a low-level of reading comprehension, as well as supporting online content. But we know participants do not always read them in full. We felt that a short animation was a great format to explain this complex process in a simple, engaging way, that didn’t get bogged down in the detail. Plus this can be used face-to-face on interviewer tablets, or embedded in online questionnaires.

As an industry, it’s our ethical duty to reassure the public on this issue in terms they understand, and it’s also in the best interests of us and our clients. Better informed people feel more comfortable agreeing to data linkage.

Kirby Swales is survey centre director and Matt Jonas is campaigns manager at NatCen Social Research