Analysis of the graphic report of results of randomized clinical trials in manual therapy: a methodological review

Tags: Poster
Núñez-Cortés R1, Pérez-Bracchiglione J2, Álvarez G3, Urrútia G4
1Escuela de Kinesiología, Facultad de Medicina, Universidad de Chile, Santiago, 2Interdisciplinary Centre for Health Studies (CIESAL), Universidad de Valparaíso, Valparaíso, 3Iberoamerican Cochrane Centre - Biomedical Research Institute Sant Pau, IIB Sant Pau, Barcelona, Spain; Centre for Osteopathic Medicine - C.O.M.E. Collaboration, Spain National Center, Barcelona, 4Iberoamerican Cochrane Centre - Biomedical Research Institute Sant Pau, IIB Sant Pau, Barcelona, Spain; CIBERESP

Background: the number of randomized controlled trials (RCTs) in manual therapy (MT) has increased exponentially, and graphs are a usual way of summarizing the results. However, the quality of this kind of report has not been assessed yet.

Objectives: to assess the quantity and quality of graphs in RCTs of MT.

Methods: we conducted a methodological review, with an exhaustive search strategy in PubMed and CENTRAL databases of RCTs in the area of ​​MT, published between the year 2000 and 2015. From the total number of articles found, we took a random sample of 100 trials that met the eligibility criteria (RCT design, MT intervention, English language). We identified figures and graphs presenting data related to the primary outcome. We then described the graph type, completeness and visual clarity of each graph in the sample, according to previously defined criteria (Figure 1).

Results: we analyzed a total of 50 RCTs (partial results). Twenty-three out of 50 (46%) reported the primary outcome graphically. The majority of graphs were: repeated measures plots (60.8%); histogram comparative (26.1%) and box and whisker plots (8.7%). None of the graphics were complete according to our criteria. In graph completeness, common areas of underperformance included failing to show number of subjects for each data element (86.9%), absence of title (78.2%) and not defining error bar meaning (60.8%). In visual clarity, most of the graphs (69.5%) had chart junk, distortion or readability issues. The most frequent problem was the readability by the superimposition of data elements (43.4%) and chart junk by excess ink in the form of unnecessary non-data visual elements (21.7%).

Conclusions: despite graphs being able efficiently to convey study findings, less than half of the RCTs in MT use them. In addition, there is a important deficit in the quality of graphic reports. Improvements in graphic reporting are necessary to generate advances in physiotherapy practice.

Patient or healthcare consumer involvement: an incomplete graphic report affects the transmission of the results to the reader, which causes them to lose the ability to create a lasting impression and an adequate interpretation of the results.