The worldwide effort to computerize paper medical records is driven by the need to cut paperwork costs and minimize the chance of medication errors, both of which would benefit patients and clinicians. Sufficiently large collections of patient data could also be a boon to researchers, helping them to investigate a host of epidemiological issues such as quickly identifying rare problems caused by new drugs or tracking the spread of a pandemic (see Nature 458, 278–280; 2009).

As a result, money is pouring into electronic medical records — not least in the United States, where the 2009 stimulus package designated some $19 billion to foster their deployment. The success of such efforts hinges on public acceptance — convincing people that they will see real benefits and that their privacy will be rigorously safeguarded. But this acceptance is seriously under threat in Britain, as the country begins to realize its plans for digitizing patients' records.

Britain is developing a national database of electronic 'summary care records' (SCRs). These will contain limited information, such as details of a patient's prescriptions or allergies, extracted from existing paper records. The UK Department of Health says that there are currently no plans to repackage these data for use by researchers; the goal is simply to provide quicker and more accurate records for health-service workers. Other data will be added to the SCRs later, however, and the research potential will grow accordingly.

Despite government assurances, some patients — and some doctors — feel that personal and potentially sensitive information will not be secure in a centralized, national database. And last week, concerned that patients were not yet fully aware of the system, the British Medical Association called for a halt to the SCR roll-out. It argued that the speed with which the database is being introduced means that patients do not necessarily know about it and that even those who do could not easily exercise their right to opt out of having an SCR.

Those criticisms are well taken. The Department of Health insists that it supports the right of any patient to opt out, and denies that it is rushing implementation of the system. But the roll-out is occurring before an evaluation of the SCR scheme by University College London (UCL) has even been published — and thus before the results of changes made following issues with patient consent raised in a previous UCL evaluation are public. As a result, public confidence in the deployment of the system is at risk of being seriously undermined.

Governments are right to be enthusiastic about such databases, but wrong to rush ahead before they have laid the groundwork for public understanding and trust. This is especially important given that pharmaceutical companies are increasingly showing an interest in the potential of electronic records. Although there is nothing in principle wrong with such use of data, it is easy to see how the public might be suspicious of corporations potentially profiting from highly personal information.

The UK government should suspend the SCR roll-out, at least until UCL's second evaluation is published next month and it is clear that previous concerns have been addressed. It should also ensure that all patients have a clear explanation of the system as well as ready access to the form needed for them to opt out.

For now, the scant data to be held in the UK database will be of limited use to researchers. But those wishing to tap into the potentially rich resources of its successors should work to ensure that the process of medical data collection is not only secure, ethical and beneficial to all concerned, but that it is seen to be so.