Fixing the ugly asthma blog post

A couple of weeks ago I posted a blog about asthma phenotypes which was lacking in graphics. To try to make up for this, today's blog is all about the missing graphics.

That post discussed a paper from the blue journal from the USA, which surveyed hundreds of asthmatics and found that they clustered in five quite distinctive clinical groups. Details are in the previous blog, or indeed the original paper.

Although a wide range of variables were used in the analysis to assign people to clusters (initially 34, whittled down to 11 ultimately). However, using three variables only (basline FEV1, maximum FEV1 after maximal bronchodilator, and age of onset of asthma) as follows...

...patients could be assigned to clusters with 80% accuracy. In this diagram, the blobs represent assignment of cluster based on the three-variable approach as opposed to the 11 variable approach. (ie for cluster 4, 72% of patients assigned to cluster 4 based on 11 variables were also assigned to cluster 4 based on the three variables only).

Just FYI, the colours correspond to clinical state as follows:

Blue = mild atopic asthma
Green = mild to moderate atopic asthma
Yellow = late onset non-atopic asthma
Orange = severe atopic asthma
Red = severe asthma with fixed airflow obstruction.

I think that the possibility of differentiating between clinical phenotypes of asthma based on three simple clinical parameters is very attractive, and could have significant implications for both research and clinical practice.

Andrew