In 2015, Google-owned DeepMind announced a new partnership with the Royal Free NHS Trust (RFT) that many predicted would herald an exciting new era of healthcare. They would apply machine learning algorithms to NHS data to create the healthcare app 'Streams' - an alert, diagnosis, and detection system for acute kidney injury. A spokesman for RFT explained: ‘The RFT approached DeepMind with the aim of developing an app that improves the detection of AKI (Acute Kidney Injury) by immediately reviewing blood test results for signs of deterioration and sending an alert and the results to the most appropriate clinician via a dedicated handheld device.’
Since then, Streams has been rolled out across a number of NHS trusts. DeepMind, meanwhile, has continued to apply its machine learning technology in other clinical trials, including early detection of diabetic retinopathy and the treatment of head and neck cancers. However, it has not all been plain sailing. An investigation by the Information Commissioner’s Office (ICO) into the data-sharing agreement, without which the work would not have been possible, found that the Royal Free had failed to comply with the data protection act when it handed over details of 1.6 million patients to DeepMind. Elizabeth Denham, the information commissioner said of the findings that, ‘Our investigation found a number of shortcomings in the way patient records were shared for this trial… Patients would not have reasonably expected their information to have been used in this way, and the Trust could and should have been far more transparent with patients as to what was happening. We’ve asked the Trust to commit to making changes that will address those shortcomings, and their co-operation is welcome.’
While the ICO should be praised for its light touch and considered approach, the issue highlights what is likely to be an ongoing problem for healthcare providers looking to use the wealth of data they hold about their patients to its fullest potential. This was also evidenced by last year's decision by UK government ministers to scrap the care data plan to link GP records, following a public outcry about whether patients had been given the chance to opt out. This is despite an independent report into the growth of AI commissioned by the UK government, 'Growing the Artificial Intelligence Industry in the UK', recommending the secure sharing of anonymized data from patients’ health records with private firms if the technology is going to be successfully applied in the sector.
There is an understandable sensitivity around the collection and sharing of medical data, yet for organizations to truly benefit from data analytics, they need to take an open approach. This means doing everything possible to encourage patients to share their data, as well as sharing data with different hospitals and technology companies. It also means, the report states, using data from supermarkets, transport organizations and town planning to better understand the root causes of ill health in society and better encourage them to live well.
Healthcare organizations have been placing greater importance on big data analytics for a number of years now. In a recent survey from SAP and Oxford Economics, almost 70% of healthcare leaders say that health IT tools are 'essential' for organizational growth and improving the consumer experience. Three-quarters of respondents said they are planning to accelerate their investments in big data analytics tools over the next two years to cement their place as top performers in their field. Clinical and data analytics also topped the HealthCare Executive Group's list of the top 10 challenges the healthcare industry will face in 2018.
The benefits they can gain from their data are many. For example, next-generation sequencing and large-scale clinical genomics are driving a major expansion into precision medicine, while the availability of real-world healthcare data from medical and hospital practices is providing new insights into how patients actually fare on a given therapy. Parkland Health and Hospital System in Dallas, Texas have also developed a validated algorithm based on Electronic Health Records (EHR) that predicts readmission risk in patients with heart failure. Those found to be at high risk of readmission are educated on how best to minimize the changes by a multidisciplinary team, given follow-up support via telephone within two days of discharge to ensure medication adherence, have an outpatient follow-up appointment within seven days, and a non-urgent primary-care appointment. The initiative was found to have slashed readmissions by 26%
The amount of data in healthcare is set to explode as we realize the potential of wearables and IoT to monitor people's health. This data has tremendous implications for AI and could provide a real boost to preventative medicines. But if we companies are not able to use the data, this is clearly going to be impossible.
The main problems arise when data is not collected transparently. As with all information sharing agreements between the NHS and external organizations, patients are able to opt out of any data-sharing system by contacting the trust’s data protection officer. However, there are two problems with this. Firstly, there are concerns around how easy it is, and many will be unaware that their data is being shared in the first place. Secondly, not enough is being done to educate people around the anonymity that is afforded and the benefits of sharing your personal information. People just see that they are sharing their information and decide it's not worth the risk. More needs to be done to let them know it is, but this also requires both the NHS and technology companies to be completely transparent in the data they are using and how they are using it. Trust is a two way street. We need to trust organizations to use our most private information responsibly, but we also need to appreciate that sharing it benefits everyone and pay attention to what they are trying to do, not just ignore it.