Health Apps and Evidence: We Can't Be 'appy all the Time

The number of healthcare apps has exploded (there are now more than 300,000 on the market). From tracking fertility cycles to enabling remote consultations, we have entered what our wonderful OHT pals have cleverly called an 'Appy' time.

Given these apps are used in healthcare, it seems natural to assume that these pieces of technology are making people better, and definitely making sure they 'do no harm'. But is this as simple as we may think? 

Would Hippocrates, the father of modern medicine, who coined this philosophical code of conduct be heading to the app store to track his (fictitious) high blood pressure?

In this short and sweet blog post we want to look at why we raise one eyebrow at a health app utopia and what it might take for an app to ‘do no harm’, bringing our beloved friends evidence and ethics to the main stage. 


Are we being cynical, are all apps just ‘appy’ after all?

In traditional evidence-based medicine, the way to be sure that a health intervention won’t cause harm is to test its effects, and then test them again, and again, and again (ideally in a randomised controlled trial). Each time writing the results up in public, and being honest about what was found. There are many flaws in this process and not everyone plays by the same rules, but overall it works. In app world this process of rigorous testing and evaluation is, by and large, not happening. 


Most app studies are small non-randomised pilot studies that have been conducted by the developer themselves, introducing many opportunities for bias. Where Randomised Controlled Trials have been conducted, these are small, unreplicated, unblinded, single-site, and have not progressed beyond phase two trials (testing with a very small number of people, to see if it can have an effect in a lab setting). Such issues seriously call into question the reliability of the evidence for the clinical use of health apps. 


This problem is made worse, by the fact that nobody is helping appless app browsers identify which apps genuinely might help, and which might hurt them. In traditional evidence based medicine clinicians critically appraise the validity and applicability of available evidence and combine this with clinical experience and expertise to make clinical judgements. In this sense, the clinician acts as the gatekeeper between the patient and the evidence behind their clinical care. In the realm of apps this is no longer the case. Hundreds of thousands of apps are freely available in the public domain for use by patients, without being subject to any quality-control or approval process, and there are almost no resources to help end users evaluate the quality and suitability of these products. A minority, such as the Sleepio app, may be evidence-based, safe and effective but these are few and far between. In short, there is currently no way for individuals to tell which apps have been clinically proven to be safe, and which may potentially cause harm. 


Still not convinced that this situation could cause a problem?? Let us tell you a cautionary tale…..


A cautionary tale 

In 2013 Dermatologists published a paper in the Journal of American Dermatology highlighting their assessment of a health app to monitor moles so as to prevent skin cancer. 


They showed the app 93 biopsy-proven melanoma photos from online Dermatology sources (granted they weren’t iphone photos as the app asks for, but you imagine a bunch of dermatologists know when the resolution of a skin cancer photo isn’t good enough right?!). The app had a sensitivity (how good it is to rule in a diagnosis) to report a melanoma as high risk of just 10.8%. Say what?! Yet 35,000 people had downloaded this app. 


This left everyone feeling rather concerned. 


But then we thought, hey - that was 2013 things must have improved since then??? 


So imagine how happy we were to see a systematic review from 2020 looking at exactly this! It set out to see how 6 different dermatology smartphone apps performed when used by the public to check out their moles. Unfortunately, the review found that recommendations between different apps contradicted each other when it came to histopathological results and recommendations. 


This creates a scenario where, people may be turning to apps to determine whether they should seek medical attention and either being falsely alarmed - which is damaging psychologically for the individual and damaging for the health service in terms of wasted resources - or falsely reassured potentially leading to later diagnoses. 


So from this we can conclude that the future is coming, but it's not here now. 


What we can do to make sure they are ‘appy’ after all

This leaves us in a conundrum. We want everyone to be appy all the time, and yet it seems we can’t achieve this yet, so what needs to be put in place: 

  1. A robust regulatory framework that clearly establishes what ‘good’ evidence looks like for different digital health interventions that pose different levels of risks to users. These standards shouldn’t vary across different individual app store providers.

  2. Greater transparency around what is and is not currently assessed by app store providers (and how assessments are conducted) so that users do not fall into a trap of thinking apps are subject to stricter evaluations than they are in reality. In other words, we need a ‘buyer beware’ market. 

  3. A clear process for alerting app store providers and/ or regulators to apps that are found to be ineffective or potentially harmful so that they can be removed safely & swiftly 

  4. Mechanisms for effectively integrating apps into care pathways so that users and/or patients have a seamless experience of care, with no risk of things falling through the gaps 

  5. An agreed upon ethical framework to help App developers think through the difference between what could be built and what should be 


Luckily most of these things are being developed and so we are well on our way to an appier future. But, it will take time to get things ‘right’ and there will be a need for flexibility and iteration as technology changes and develops. So, in the meantime we urge everyone to be optimistically skeptical: assume that apps can be a force for good, and that they have been designed with good intentions, but question whether they can really achieve what they claim to. Adopt a madeye-moody approved stance of ‘constant vigilance:’ don’t take anything at face value, remain cautious, curious and critical about what apps you download and use, and we will find an appy medium. 

Jess Morley & Amelia Martin

Amelia is working as Clinical Experience lead at a health tech start up called Nye. She is a Junior Doctor and previously worked at NHS England looking at the role of technology to enable alternate models of Outpatient appointments. 

Jess is policy lead at University of Oxford’s DataLab, and a student researcher in the Oxford Internet Institute’s Digital Ethics Lab. They are both passionate about ethics and global health equity and together curate the Oxford One HealthTech Hub.

@jessRmorley






Support Us

We are a volunteer-led organisation and reliant on the support of partners. For more details on partnering email us or contact your local hub.

Contact Us

Ask us anything!

 

info@onehealthtech.com

Connect with Us

71-75 Shelton Street,

London

WC2H 9JQ