Using government data for good

The mishandling of NHS’ care.data has created much interest about the risks and potential abuse of  sharing personal data. The focus tends to be on two nightmare scenarios. The first is that someone—think of an unscrupulous, profiteering pharmaceutical or insurance company, or a blackmail-minded technician—gets holds of the data and uses it to unfairly squeeze money out of or deny services to unsuspecting, innocent people. The second is that this personal data is leaked, causing embarrassment.

Both are indeed possible—and it’s helpful that these issues are brought to light and debated—but are over-blown by campaigners worried by the government’s plan to share personal medical data. These nightmare scenarios are not so clear cut upon examination. Most of us are probably happy that people caught speeding pay higher car insurance premiums than people who don’t speed. But we would be less comfortable if health insurance companies charged premiums for people who have genetic disorders over which they have no control. Comparing the costs and benefits of sharing personal medical data is full of ethical conundrums like this. There are safeguards in place, but they’re not fool proof—and that’s why the issue is complicated. (If you’re not sure about your position on whether the NHS should be sharing personal medical data, I suggest you read Ben Goldacre’s article).

I will not be opting out of sharing my personal medical data, primarily because I see making it available as an act of altruism. I know analysing large data sets will generate knowledge that will improve and save lives. Yes, there are risks of profiteering and leaks, but the good it can do will always trump such concerns.

If you want to get a sense of the potential benefits of sharing government-held personal data look at the Justice Data Lab, a Ministry of Justice initiative that NPC inspired. By using the lab, organisations can find out how many of the offenders they worked with (the “intervention” group) have been subsequently convicted of a crime. They can also learn how many offenders in a control group (ie, similar offenders they did not work with) were convicted. The difference between the two conviction rates—that of the intervention group and that of the control group—is the difference that the organisation has made. These results, which are made public on the Ministry of Justice’s website each month, help inform what works, and what does not, in reducing re-offending. This will reduce crime. As the data on individuals stays within government walls, and only aggregate data is reported out, there is only a miniscule risk of anyone being able to identify which individuals have been re-convicted.

Since April 2013, the Justice Data Lab has produced 55 reports of quasi-experimental studies. In the world of evaluation, these are good research designs. If the Ministry of Justice had independently commissioned the 55, it would probably pay ten times what it costs to run the Justice Data Lab. The Justice Data Lab is like a golden research goose, generating high-quality research studies every month and creating value for money for the taxpayer.

The Ministry of Justice is currently deciding whether to extend the life of the Justice Data Lab beyond its one year pilot. NPC has recently written to Jeremy Wright, one of the Ministers, both to congratulate the Ministry of its pioneering approach, and to suggest how the service can be improved. Our letter is here.

The large sets of personal data held by governments are incredibly useful but are woefully under-used. If the data is more accessible, and if safeguards are put in place, we will all be better off. Read about the risks and make sure you understand them. But also make sure you understand the benefits. If you do, the choice should be clear.