1. Data minimisation

This is the second of a series of posts exploring patterns that could help solve some of the biggest healthcare data challenges. Read the other posts here:

0. Introduction
1. Data minimisation

Data minimisation: Collect and store only the data required for the thing the user needs to do.

Data minimisation

(credit: Martin Fowler)

The Germans have a word for this. They call it Datensparsamkeit.

It’s a concept enshrined in UK law, in the third principle of the Data Protection Act:

Personal data shall be adequate, relevant and not excessive in relation to the purpose or purposes for which they are processed.

And it’s also a concept in health. By a happy coincidence, it’s the third Caldicott Principle:

Use the minimum necessary patient-identifiable information

It’s the antithesis of the Big Data, 'store all the things' approach.

Putting data minimisation into practice means those designing services doing the hard work to make things run on less data. It means asking 'do we need to ask for this piece of information to meet the user's need?', 'what's the minimum we really need to store?' and 'what can we destroy once we're done with it?'.

Of course, there are objections.

The first is that by limiting the use of data to the purpose it was collected for, you hamper the ability to deliver joined up care. The NHS has no use for data about my allergies right now, but if I was unconscious I'd want the paramedic to have access to it.

The transaction-by-transaction approach to data might be right in other public services but healthcare isn't a transaction, the argument goes.

The second objection, articulated here, is that limiting data collection to the minimum necessary to care for an individual (and preventing any other uses of this data) would impact our ability to advance medical research or do healthcare planning.

Despite these challenges, data minimisation remains a useful mindset for those building services. If our institutions got this bit right at the point at which data is collected and/or stored we'd likely have fewer issues around privacy, consent and security further down the line.

Data minimisation is going to become even more important as it becomes easier to collect data and cheaper to store it. Wearables can collect thousands of data points about us each day and who knows what data Internet-connected devices will be able to collect in the future. What rules will govern this?

For now, next time you add another <input> tag to your digital service just think about it very carefully...

Next up in this healthcare data infrastructure series: permissions