Algorithms -- In today's world are they a good thing? (2024)

So what is an algorithm? The Cambridge dictionary defines it as “a set of mathematical instructions or rules that, especially if given to a computer, will help to calculate and answer to a problem". Today we live in an era where technology and algorithms impact our daily lives in ways we are both aware of and not.

I love the music recommendations Spotify makes for me, does it mix things up and give me a genre I’d never considered before…No! But as long as I’m conscious of that I can make sure I search around to diversify what I’m hearing.

Do I appreciate the ads on social media trying to make me read articles it thinks I’ll love or buy this season's on-trend dress… less so.

Algorithms can also have a huge social impact, they may be responsible for taking away someone's job or help support someone to become mobile and spend much needed time with friends and family.

In healthcare software as a medical device (SaMD) defined by the International Medical Device Regulators Forum (IMDRF) as "software intended to be used for one or more medical purposes that perform these purposes without being part of a hardware medical device”. Is a super hot area with many companies small and large investing in it. We’re in the middle of a perfect algorithm storm:

1. There are endless challenges facing patients, healthcare professionals, healthcare systems, governments, payers that need to be solved
2. Fit-for-purpose, diverse data is becoming more readily available
3. Advanced analytics is becoming easier to do, both in terms of the number of talented people skilled to do it, as well as the infrastructure and systems to be able to run such analytics
4. We are becoming better able to interpret and explain sometimes extremely complex techniques to build confidence and understanding in the results, without people needing a degree in data science
5. Results can be displayed in visually easy to action ways, meaning it can be quick and simple to use such algorithms

So while the environment is hot for making huge advancements in healthcare through the use of algorithms. The most popular question I hear is can we trust them?

Healthcare in this sense is thankfully highly regulated and logical. The more decision making power and/or critical the decision being made by an algorithm the more evidence you need to provide and the more confidence you need to show in that decision.

While the sensitivity (the ability to correctly identify patients who are positive for what you’re looking for e.g. pregnant women classed as pregnant) and specificity (the ability to correctly identify patients who are negative for what you’re looking for e.g. non-pregnant women classed as not pregnant) of the algorithms can be pre-specified, tested and will always be published to be transparent with users.

One aspect that is critical in the trust question is the data the algorithms need. With poor quality or missing data, algorithms may not be able to provide an answer at all or the reliability may be called into question. That’s why it’s critical to have a broad range of diverse data to be able to train and validate algorithms on to ensure they are representative of the populations they are going to be used in and we can all gain confidence in the robustness of the algorithms being generated from these data. More to come on data centric artificial intelligence soon.

To me, we need to think of algorithms and healthcare holistically to make sure we’re making thoughtful decisions about the impact they create and what it takes to have confidence in the results they provide. While in general life I’m often nervous about how algorithms are influencing our life, in healthcare I am not. Instead, I am really excited about the opportunities we have to co-create solutions (including algorithms) with patients and across healthcare systems that make a positive and meaningful impact for society.