The published paper featured in this post illustrates many features of a good rehabilitation research paper. (A shame it was not published in a rehabilitation journal!) The blog is categorised as Education and Training, because the paper is useful as an example, and could generate discussion. The paper, available here, concerns the pain and hand dysfunction associated with osteoarthritis of the thumb base.

Comparator
I was really pleased to see the title, and the whole paper refers to ‘the comparator’. The basic design was whatever was going on (e.g. treatments for other conditions, normal life) with all patients having education (self-management and ergonomic principles) and the experimental group having the additional intervention. This was a Randomised Clinical Trial, not Controlled.

This distinction is more than an arcane linguistic distinction. It should alter the perspective of the researchers and the readers, to remove the idea that the comparator group is ‘untreated’ or is necessarily going to do less well. It also emphasises the need to describe, reasonably, sufficient detail about both (or all) groups for the reader to extract what differences there are. It should avoid assumptions about what is or is not being ‘controlled’ for. (see here for more explanation.)

Abstract
The abstract is succinct, written clearly, with masses of specific detail and no unnecessary or wasted words. It uses structured headings, which greatly improve the usefulness of an abstract. The conclusion is related to the objective, answering the implied question posed. The factual specific data in the results allow the conclusion to be drawn. Writing an abstract of this quality will ensure the editor reads the paper; and that the reader who finds the paper will read it too.

Introduction
Three short paragraphs with seven references. It does stray slightly from the point of the paper, introducing background epidemiological data, but it is not lecturing or teaching and is proportionate. Otherwise, the introduction puts forward a cogent argument that the trial is needed and outlines concisely why this intervention was chosen. The references are mostly authoritative overviews and easily accessed.

Methods
There are some areas where I, as editor of Clinical Rehabilitation, would have different preferences. For example, we suggest that all bureaucratic information should go in the first paragraph: registration, ethical agreement, funding, dates, sponsor etc. But the important point about the methods section is that it has a clear structure, with quite short paragraphs each covering a single topic.

It specifically explains how random allocation occurred. Most people simply state that a computer randomised patients. How?! All the interventions are described simply. At no point do the authors start either teaching and explain, or justifying. The measures used are easily found.

Results
The presentation of data is masterful; no wasted space, easily understood and read, precise. There is little numerical data in the text.

One matter I would disagree with is the issue of ‘responders’. Many years ago we showed that ‘improvement’ demonstrated on cognitive tests by using a cut-off value was primarily influenced by regression towards the mean. (here) I have also recently argued that identifying responders could be a ‘statistical mirage’, suggesting that there are two populations, ‘responders’ and ‘non-responders’, when in fact there is one population and everybody gets more or less benefit. (here)

Apart from that small quibble, which is not about the presentation of the results anyway, the presentation is of a standard to be aspired to.

Discussion
The discussion is relatively short and focuses on the trial as a whole and how it fits in. It makes no unjustified claims and is really good at discussing its limitations.

One matter I would feel is missing is any considerations of ‘so what?’. What are the implications of the results? Should more research be carried out? More importantly, what is the immediate research priority? What are the clinical implications? Is it to be recommended as an approach? And, because I am interested in rehabilitation as a process, does this have any implications about the theoretical basis of rehabilitation in general, and to the treatment of hand osteoarthritis in particular?

In conclusion, if you are an author writing up a randomised clinical trial, this paper shows you what you are aiming for. If you are writing up other types of study, the principles that underlie the presentation of a good paper are still illustrated in this paper. For further guidance on some of those principles, read here.