The above study is to compare an intervention group utilising an accelerated exercise programme post injury against the standard protocol of rest, ice, compression and elevation (RICE). The following is to critically appraise the above quantitative piece of research with some guidance from the Critically Appraisal Skills Programme tool. The abstract gives clear objectives to the aims of the researchers and the chosen setting to allocate participants is pertinent to the research question.The samples for the trial are chosen at random a method considered as gold standard due to their design in attempt to remove bias, a technique considered to provide internal validity (Sibbald and Roland, 1998). The primary and secondary outcomes to be measured in the abstract are clearly focused, but to a non-statistician, the results section appear to be a barrage of figures and do not initially provoke further reading.
The introduction informs the reader of the estimated cases of ankle injury in the U. K. nd its implications but gives no indication to the costs incurred to the NHS. Unfortunately it reports the estimated costs in the Netherlands and fails to explain the costs to the Royal Victoria Hospital in Belfast, or the sports injury clinic at the University of Ulster, where the research had been carried out. The authors made it clear of the prevalence of the injury that occurs in the UK but a more profound impact to engage the reader could be achieved if the costs to the UK rather than the Netherlands had been discussed.As the article was printed in the British Medical Journal in 2010, utilising a electronic data base search to ascertain these costs, the reader could have further insight to the impact of ankle trauma and its relevant literature regarding resources spent on this outpatient injury (Burns and Grove, 2001).
Having said that, to understand the above article it is imperative to read the authors former article preceding this study that informs the reader that in the UK’s accident and emergency departments, 302,000 admissions are due to ankle injuries alone (Bleakley et al, 2007).From this preceding article the researchers continued to follow their developed protocol. Sample size was determined with parameters of statistical significance set at 0. 05 and a power calculation of 80% in addition to standard deviations for the continuous primary outcome. Confidence intervals to estimate the mean differences between the samples was more than adequate set at 98.
75%. The methodology to recruit the samples utilised a computer generated randomisation sequence to obtain the participants with a stratified method to balance the randomisation.Stratification is utilised to ensure that equal numbers of participants share characteristics that may otherwise influence results in each comparison group (Salmond, 2008). Of the inclusion criteria the sample sizes were almost identical with 51 participants allocated to the control group and 50 in the standard group.
One of the participants did not receive intervention in the control group contributing to a total balance of the samples. The authors do not say why one of the participants allocated to this control group did not receive treatment.Although it could appear to be selective rather than truly randomised, stratification strives to provide unbiased samples in order to produce valid results (Schulz and Grimes, 2002). Within this piece of research, randomisation without the stratification of athletes and non athletes could have a high impact on the findings as it is likely that the athletic participants would be more compliant in the exercise group and possibly skew the results. Initially, the exclusion criteria seems quite vast but justified as it would be a burden on resources to include the need of interpreters or have participants that could prove difficult to access.Within random control trials it is also necessary that the treatment groups share baseline characteristics to avoid selection bias.
Interesting to note that the baseline comparisons shown in table 1 expresses grade (1) injuries 18:11 in the standard group and a ratio of 33:39 of grade (2) injuries between the groups. Although exclusion criteria was a grade (3) injury involving rupture, bone or other parts of the anatomy, the differentiation of a grade 1 or 2 injury is not fully made clear and a minor injury may not warrant any intervention at all affecting the overall outcomes.Another interesting baseline comparison can be seen in the standard group is that 41 of right dominant limb participants incurred equally 41 inversion injuries compared to the exercise group of 32 right dominant limb injuries accounting for 37 inversion injuries at baseline. Suggesting that only 5 injuries occur on the weaker side. It could be argued the dominant limb is more susceptible to injury but how this was measured is not known. In addition it would have been interesting to see which re-injuries of the two in each study group at baseline had occurred on the dominant side.
The research design is robust as it initiates block randomisation for each strata and attempts to gather evidence in a highly controlled way to produce data from each of the variables under investigation. The variables of the groups are assessed as primary and secondary outcomes to ascertain if there are any significant differences. Stratified sampling is a design that takes groups into account rather than individual sampling and analysis of the data gathered from these groups may be more valid and reliable (Lee, 2006).As there are several outcomes in this RCT the sample size has to be considered and this pragmatic approach was conducted when the researchers were at their planning stage of this investigation, as more participants are required when several outcomes are to be measured (Roland and Torgerson, 1998). Within the paper, a researcher (MAT) is said to be blinded to the intervention group to record outcomes at weeks 1 to 4 but in the previous PRICE paper it postulates to conduct a strict double- blinded trial.
When participant and researcher are unaware to treatment modalities the reporting of participants and the assessment of the researcher when collecting data, is more controlled and limits bias (Chan, 2003). There is no clarification to whether the groups were blinded to treatments in the above article and in addition there is no mention of the attention given to the standard group. Although an attempt was made to be consistent at the initial stage of post injury, giving general advice and treatment to both groups was carried out by the same research physiotherapist.Non-blinded to the two treatment modalities of both groups provokes doubt, as potentially the therapist will have their own preferences, thus inducing bias from the outset. Confusing, that after the first week both groups received exactly the same advice and treatment including an exercise regime. Why the trial continued to measure seems ironic when intervention only lasted for one week and treatment thereafter was standardised between the two groups.
Accepted that the research question was to ascertain the affects of accelerated exercise on ankle joint function, the fact that the trial continued after there was no longer a control group suggests there is nothing to compare. Hence the results of the findings with no differences between the two groups thereafter. The outcomes to be measured were primary and secondary measurements and taken at the adjusted baseline and at weeks 1 to 4 followed by a 16 week follow up. Primary outcome measures were to assess ankle joint mobility using the lower extremity functional scale.This lower extremity functional scale has been found to have excellent test-retest reliability suitable for research purposes and clinical decision-making (Binkley et al, 1999).
However the scale is a self completed questionnaire based on 20 functional leg activities and is subjective. The secondary outcomes are concerned with pain at rest, pain on activity, again very subjective with the aid of a pain analogue scale which is a unamenable tool to rely on as pain is dependant each individual’s pain thresholds. Other secondary outcome measures were swelling and physical activity levels.Questionnaires as in this piece of quantitative research uses specific questions that can be answered with a score to generate inferential statistics that provides the assessor with data for analysis (Punch, 2004). The repetition of the measuring outcomes running consistently through to the follow-up supports reliability of the research producing test-retest reliability (Bowers, 1996).
Although it could be argued that anything tested over time repeatedly will eventually lead to some significance. As stated measurement outcomes were consistently run at consecutive weeks. Fig 1 however presents confusing follow up recordings.Fig 1 presents a flow chart of participants through the trial at follow up stages.
The control group at week 1 identifies 44 participants (n=44) but at week 16 (n=46) giving an increase of two. Whether this is due to hidden circumstances that two of the participants could not be accounted for is disturbing, as this is the start of the trial and indicates a recording error of the researchers or it just could simply be a misprint in the published article. At weeks 1-4, the number of participants tally in the control group as there are 39 remaining from the sample of 50 after excluding the 11 lost during the trial.The results revealed that a total of 15 participants dropped out during the trial.
There is no explanation as to why these participants dropped out once they had signed and consented to take part. Although the target sample was not reached from the population it was still big enough for a constrained linear mixed model analysis. The researchers adopted a rigorous and appropriate approach to analyse data by utilising the ANCOVA software. The analysis of covariance allows to compare a single variable in 2 or more groups taking into account variability of other covariates (Adams, 2007).
The results identify at week one a confidence level of 98. 75% and a p value (p=0. 0077) proving that the exercise group at week 1 significantly had increased joint mobility. Again at week 2 the exercise group significantly had increased mobility compared to the standard group with a p value 0.
008%. Confidence level not stated. Increased activity levels were also significantly higher in the exercise group but there were no significant differences at any other time for pain at rest or during activity or swelling.At the sixteen week follow-up although no differences were found by using the mean value, missing data was also included.
In RCT’s where there is an intention to treat, including missing data by taking into account the last observation in their analyses conforms to the Consolidated Standards of Reporting Trials guidelines (Hollis, 1999). Accepted that the research trial is an excellent attempt with a well-planned design with methodological rigour to control randomisation. However, as aforementioned there are discrepancies and it fails to be a double blinded trial that it initially purports to deliver.Chris M Bleakley, the research associate deserves credit where due as this is his 4th publication all active in injury and rehabilitation in the interest of patients. Ethically, the research has been approved by the research ethics committee Northern Ireland and follows the guidelines of the A&E Department. As registered therapists, the researchers have to conduct in accordance to the physiotherapists regulatory body, the Health Professional’s Council.
The research design is reproducible with room for improvement possibly by categorising and analysing grade 1 and 2 ankle injuries separately as they differ in severity.The inflammatory response to injury can subside with rest, ice, compression and elevation and some clinicians in fact consider impairment of function to be the fifth cardinal sign of acute inflammation (Marieb, 1998). RICE is the treatment modality advocated through podiatric practice although the significant results may prove useful in the field of physiotherapy depending on severity of the injury and the therapists’ personal preference. This reason alone may have contributed to the 15 participants lost during the trial as more importantly eleven of these were lost from the intervention group and only four from the standard group.Possibly this loss may have been contributed to less severe injuries being misgraded with more randomly allocated to the intervention group.
If the reason for the drop out was ascertained and found to be due to a rapid recovery the results for no longer taking part in the trial could be used to further support that accelerated exercise intervention does work. Acknowledged by the authors themselves that physiotherapy within A&E is not standard practise, the results although internally significant, cannot be generated to a wider population and therefore do not provide external validity.
