05/05/2026
The most thorough review of AI in primary care to date has just been published in The Lancet Primary Care, led by Dr Liliana Laranjo and Dr Ana Luisa Neves. Both have long been part of WONCA's work on digital health.
Three findings deserve attention.
First, AI scribes. Studies show mixed results for time saved on documentation, an increase in note length, and improved satisfaction at work. The gap between perception and measurement is one of the most important findings.
Second, evidence base. Most AI being deployed in primary care is pre-generative. Trial evidence for the newer tools, including LLMs, is thin. The performance gap between simulated conditions and real practice is wider than the marketing implies.
Third, equity. AI is trained on data from the past, embedding biases. Algorithms developed on data over-representing European-ancestry populations perform worse for under-represented populations. This is not a future risk. It is a current property of tools already in use.
Read the review here:
https://www.thelancet.com/journals/lanprc/article/PIIS3050-5143(25)00078-0/fulltext
