NHS Health Apps Library full of data-spaffing apps, claims studies

Researchers from Imperial College London have published three studies in the journal BMC Medicine, provoking serious concerns about the mobile health apps approved by the NHS and provided through its Health Apps Library.

The studies show serious issues in the software tools, which can provide diabetic users with inappropriate insulin doses, asthmatics with shoddy peak flow calculators, and widespread hackneyed security controls.

NHS Choices claims all of the apps that appear in its library have been reviewed and found to be clinically safe, but are not formally "accredited" or "endorsed".

Twenty-eight per cent of apps included in the studies lacked any sort of privacy policy, while one which promised anonymity actually spaffed off personally identifying information without encryption.

A commentary on the studies, titled 'Trust but Verify' – five approaches to ensure safe medical apps, and also published (PDF) in BMC Medicine, has summarised the studies by stating that health apps are "consistently poor" in quality.

For example, a dermatology app claiming to have been downloaded over 35,000 times purported to identify pre-cancerous moles. However, on testing it was found to have just 10 per cent sensitivity to classify biopsy-proven melanomas correctly.

A highly downloaded rheumatology app featuring various calculators was withdrawn for giving users 15 to 20 per cent inaccurately higher scores on a disease activity score for one formula, and a 10 to 15 per cent lower score than was accurate for another.

Although the NHS reviews the apps on its library to ensure compliance with the Data Protection Act, and "to ensure they are clinically safe," on examination 70 of 79 tested apps did transmit data over the internet, with 38 of those not providing any information about what data would be sent.

Kit Huckvale, lead researcher for the project, told the Press Association that it is "known that apps available through general marketplaces had poor and variable privacy practices, for example, failing to disclose personal data collected and sent to a third party".

"However, it was assumed that accredited apps – those that had been badged as trustworthy by organisational programmes such as the UK’s NHS health apps library – would be free of such issues," he added.

He continues that the ICL study "suggests that the privacy of users of accredited apps may have been unnecessarily put at risk, and challenges claims of trustworthiness offered by the current national accreditation scheme being run through the NHS."

The results of the study should provide an opportunity to address these concerns, stated Huckvale, and to "minimise the risk of a future privacy breach."

The researchers have already provided their findings and data to the NHS Choices, which is in charge of the Health Apps Library.

A spokesperson for NHS Choices said: "It's important that all of the apps listed on the NHS Health Apps Library meet the criteria of being clinically safe, relevant to people living in England and compliant with the Data Protection Act."

They added: "We were made aware of some issues with some of the featured apps and took action to either remove them or contact the developers to insist they were updated. A new, more thorough NHS endorsement model for apps has begun piloting this month."

The abstract to the commentary on the studies states that "App store owners could ensure transparency of algorithms (whiteboxing), data sharing, and data quality. While a proper balance must be struck between innovation and caution, patient safety must be paramount."

Serious concerns were raised in July this year over the credibility of the NHS Health Apps Library, with privacy campaign group MedConfidential identifying at least 60 apps that call into question the body's approval process.®