Last week, the Chinese government officially launched a mobile application which allows locals to check if they have been in close contact to someone with COVID-2019 (previously known as 2019-nCOV), the novel coronavirus originated from Wuhan, China which infected nearly 47,000 people and killed more than 1300 patients.
Compromising public trust and risks of singling out a particular community
Users are required to surrender their identities (i.e., name, ID, phone number) and scan a QR code on their smartphones to make an inquiry on the new “close contact detector”. If they were found to be in close proximity with COVID-2019 patients, the app will advise them to stay home for self-quarantine and alert the local health authorities. Users can also use the same app to check on the status of three other individuals.
According to Xinhua, the state-run press agency, the detector was co-developed by the General Office of the State Council, the National Health Commission and China Electronics Technology Group Corporations which employed “accurate, reliable and authoritative data” provided by various channels comprising the Ministry of Transport, China Railway and Civil Aviation Administration of China.
At the same time, the Chinese government had also changed its way of recording new COVID-2019 cases in Hubei province, where the concentration of outbreak is. It will now include individuals with demonstrated symptoms, such as cough, fever and shortness of breath, but have not been officially tested. This had not only caused a temporary spike in reported cases but also a concern of false positive.
While all these may help prevent the virus from spreading to a certain extent, marking individuals as potentially contagious via a new mobile app may still cause unnecessary stigma and panic, in view of the social reactions to this disease. Some believe this is taking place at the expense of public trust as individuals may choose to hide their symptoms or risk of singling out a particular community or neighborhood; another example of technology used without attention paid to public concerns.
Poorly regulated and cannot be relied upon
The close contact detector is definitely not the only mobile app under scrutiny recently. Researchers from the University of Birmingham’s Institute of Applied Health Research and the Centre of Evidence-Based Dermatology at the University of Nottingham conducted nine studies to assess the reliability of six different mobile applications targeted at skin cancer detection. They found most of these algorithm-based smartphone apps, including two which are regulated as Class 1 medical device with CE marking, either failed to uncover malignancy or yield false positives.
In one of the studies, researchers found the Class 1 medical device approved app missed 12% patients with cancerous moles and inaccurately identified 21% of harmless moles as cancerous. All results were published in The British Medical Journal (BMJ). Although all applications come with disclaimers indicating they exist as guides and cannot replace actual human healthcare advice, researchers remain disappointed as there’s not enough evidence to evaluate the efficacy and accuracy of mobile health applications even though they are becoming ubiquitous. They believe it is important for healthcare professionals to be aware of the existence and limitations of most of these technologies.
Ways to do better
Ben Goldacre, DataLab Director and Jessica Morley, DataLab Policy Lead at the Nuffield Department of Primary Care, University of Oxford and Luciano Floridi, Professor of Philosophy and Ethics of Information, Oxford Internet Institute, University of Oxford said it’s important to have a collective understanding of what are good evidence to assess health apps and how they should be monitored and updated regularly after they have been released.
Naturally, regulators play a huge part in this but users should also be aware that classifying a healthcare mobile app into a low-risk category does not mean it is risk-free. At times, some of the overlooked shortcomings in these apps may have serious implications on patients and the general public. Misdiagnosing or over-diagnosing will both put needless pressure on clinicians, patients and the health system as resources are inappropriately allocated. In the long run, the digital health industry may lose its credibility.
As such, Goldacre, Morley and Floridi recommended official guides on how to help the public as well as the industry to critically appraise on different digital health products and know each of their limits. It is also crucial to establish a cultural shift; one which encourages healthy competitions, error corrections and reproducibility to uphold the quality of digital health solutions in the market.