Citing and endpoints

To interpret composite endpoints, reporting of contributing items are needed.

In an article in this issue of the Journal, Jakobsson [1] address the citing of scientific data, specifically composite endpoints in six different randomized clinical trials, and how composite endpoints should be reported [1].

The correctness of citing is often in the eyes of the beholder. What is cited and the correctness have been sparsely investigated. Sometimes, many of us have, to our disappointment, found that our own pertinent data are not mentioned in more or less similar articles. It has been suggested that citing is regional i.e. articles from the author´s region are more often cited than papers from other regions. This was investigated by Grange, the technical editor of the BJU International at that time [2]. He found some evidence of national differences in citing patterns in the BJU International and the Journal of Urology but concluded that there were many different possible reasons for this.

The present paper investigates the citing pattern of the composite endpoints in six different urological randomized controlled trials in the 531 articles citing them. In 58% of the citing articles, the composite endpoint was not mentioned. In these studies, citing was used to, for instance, underline general statements. Of the citing articles, 42% cited the composite endpoint and it was correctly cited in 97%, which is reassuring.

Citing analysis in this sense is to some extent new and as the authors of the present paper point out, there are yet no validated instrument for this.

The authors also investigated how the composite endpoints were reported, i.e. whether the contributing items comprising the composite endpoint were reported or not. Only 42% of the citing articles reported data on the contributing items. Without data on the contributing items, there is no way to know what is driving the composite endpoint. This is illustrated below:

Table

Metric, Ohio	Metric, California
Founded 1850	Founded 1925
Inhabitants 2450	Inhabitants 6365
Altitude 4000 feet	Altitude 10 feet
Sum 8300	Sum 8300

Although the composite endpoint (sum) is equal, the towns are obviously different in all ingoing aspects of the sum. Composite endpoints are often quite easy to analyze statistically but may be quite meaningless if the ingoing item are not given. I can only agree with the authors that the items that are used to construct a composite endpoint should be reported.

Thus, authors and citing authors – please define your composite endpoints and report the ingoing items, referees – look for and ask for this, and editors – make this a requirement for publishing.

Jan Adolfsson
The CLINTEC Department
Karolinska Institutet, Stockholm, Sweden
[email protected]

Disclosure statement

No potential conflict of interest was reported by the author(s).