Plots are an elegant and effective way to represent data. At their best they encourage the reader and promote comprehension. A graphical representation can give a far more intuitive feel to the pattern of results in the study than a list of numerical data, or the result of a statistical calculation.

The temptation to exaggerate differences or relationships between variables by using broken axes, overlaid axes, or inconsistent scaling between plots should be avoided.

A plot should be self-explanatory and not complicated. It should make good use of the available space. The axes should be scaled appropriately and labelled with an appropriate dimension.

Plots are recognised statistical methods of presenting data and usually require specialised statistical software to create them. The statistical analysis and methods to generate the plots are as important as the methodology of the study itself. The software, including dates and version numbers, as well as statistical tests should be appropriately referenced.

Following some of the guidance provided in this article will enhance a manuscript.

Cite this article: Bone Joint J 2015;97-B:1593–1603.

We examined the rates of infection and colonisation by methicillin-resistant Staphylococcus aureus (MRSA) between January 2003 and May 2004 in order to assess the impact of the introduction of an MRSA policy in October 2003, which required all admissions to be screened. Emergency admissions were treated prophylactically and elective beds ring-fenced. A total of 5594 admissions were cross-referenced with 22 810 microbiology results. The morbidity, mortality and cost of managing MRSA-carrying patients, with a proximal fracture of the femur were compared, in relation to age, gender, American Society of Anaesthesiologists grade and residential status, with a group of matched controls who were MRSA-negative.

In 2004, we screened 1795 of 1796 elective admissions and MRSA was found in 23 (1.3%). We also screened 1122 of 1447 trauma admissions and 43 (3.8%) were carrying MRSA. All ten ward transfers were screened and four (40%) were carriers (all p < 0.001). The incidence of MRSA in trauma patients increased by 2.6% per week of inpatient stay (r = 0.97, p < 0.001). MRSA developed in 2.9% of trauma and 0.2% of elective patients during that admission (p < 0.001). The implementation of the MRSA policy reduced the incidence of MRSA infection by 56% in trauma patients (1.57% in 2003 (17 of 1084) to 0.69% in 2004 (10 of 1447), p = 0.035). Infection with MRSA in elective patients was reduced by 70% (0.56% in 2003 (7 of 1257) to 0.17% in 2004 (3 of 1806), p = 0.06). The cost of preventing one MRSA infection was £3200.

Although colonisation by MRSA did not affect the mortality rate, infection by MRSA more than doubled it. Patients with proximal fractures of the femur infected with MRSA remained in hospital for 50 extra days, had 19 more days of vancomycin treatment and 26 more days of vacuum-assisted closure therapy than the matched controls. These additional costs equated to £13 972 per patient.

From this experience we have been able to describe the epidemiology of MRSA, assess the impact of infection-control measures on MRSA infection rates and determine the morbidity, mortality and economic cost of MRSA carriage on trauma and elective orthopaedic wards.