Table 4.

Topic groups of the sample identified through structured topic modelling with topic number k = 11 and merged manually

Topic (group)		All topics		Only included topics
	n	%	% cum.	% valid	% valid cum.
Health impacts	1529	23.81	23.8	30.2	30.2
Environmental protection, climate, sustainability	1273	19.82	43.6	25.1	55.3
Risk communication	947	14.74	58.4	18.7	74.0
Prediction models	664	10.34	68.7	13.1	87.1
Air quality management	653	10.17	78.9	12.9	100.0
Substance analysis and filters	998	15.54	94.4
Preclinical analyses	359	5.59	100.0

Note: n = Number of references, with the respective topic as the main topic (based on the highest γ-probability) / % = proportion of the articles in the total sample (n = 6423) / % _valid = proportion of the articles included in the final analysis (n = 5066)