The Lancet publishes three studies: from hereditary angioedema to AI in primary care

Within the space of two days, on 8 and 9 October, The Lancet and its sister journal The Lancet Global Health published three pieces of research touching on very different areas of public health and medicine: a rare genetic disease, the use of artificial intelligence in basic diagnosis, and global preparedness for a future epidemic.

The first study, published on 8 October, concerns oral deucrictibant, a drug tested for the on-demand treatment of attacks in hereditary angioedema, a rare genetic disease. It is a phase 3 trial — the stage of clinical testing that precedes possible marketing authorisation — with a multicentre, randomised, double-blind, placebo-controlled, crossover design, in which participants receive both the drug and the placebo at different times. The available abstracts do not report the study’s numerical results in terms of efficacy or safety, which therefore remain to be verified in the full publication.

On the same day, The Lancet published a second paper, described as the first prospective feasibility study on the use of conversational artificial intelligence for diagnosis in outpatient primary care settings — that is, in general practitioners’ surgeries, where most people first come into contact with the healthcare system. Here too it is a preliminary, feasibility stage: the study checks whether the approach is viable in a real-world setting, not yet whether it produces more accurate diagnoses than traditional clinical practice. The abstracts do not report diagnostic accuracy data or comparisons with doctors.

The third paper, published on 9 October in The Lancet Global Health, changes register entirely: it is an epidemiological modelling and cost-benefit analysis that estimates the investment needed to build a global stockpile of mpox vaccines. This is a forecasting exercise based on mathematical models, not a direct field observation: the distinction matters because a cost-benefit model builds scenarios from assumptions, and its reliability depends on the quality of those assumptions, not on data collected during an ongoing epidemic.

Although all three pieces of work appeared within 48 hours in the two Lancet group publications, they sit at different stages of the path separating research from clinical practice or health policy: a phase 3 trial theoretically close to possible regulatory approval; a feasibility study that only verifies the practicability of an approach; a forecasting model that guides investment decisions without stemming from an actual epidemic. On the basis of the available abstracts, none of the three allows definitive conclusions to be drawn on efficacy, accuracy or economic value: the full publications, once available, will make it possible to assess the results with the actual figures in hand.