Retrieved from https://studentshare.org/statistics/1483815-rct-research-design-methodology
https://studentshare.org/statistics/1483815-rct-research-design-methodology.
A critical review of the two main features of a randomised control trial with reference to an original article by Ebenbichler, G.R., Erdogmus, C.B.,Resch, K.L., Funovics, M.A., Kainberger, F., Barisani, G., Aringer, M., Nicolakis, P., Wiesinger, G.F., Baghestanian, M., Preisinger, E. and Fialka-Moser, V. (1999) ‘Ultrasound therapy for calcific tendenitis of the shoulder’. The aim of this essay is to discuss the significant features of a randomised control trial (RCT) methodological design through the evaluation of a study undertaken by Ebenbichler et al. (1999) in ‘Ultrasound Therapy for Calcific Tendinitis of the Shoulder’.
Their rationale was to rigorously evaluate the efficacy of using ultrasound therapy to treat calcific tendinitis of the shoulder since it has not been explored as yet. I will discuss the randomisation (random sampling and random allocation) and blinding process used throughout the article as they are essential features of a successful RCT. This will be followed by an exploration of the way the researchers have used these features in their study and if the methodological design was applicable to their study.
The essay will also examine the fundamental principles of each feature and the significance of their contribution to the results of the study. Finally, the limitations of the RCT used in the study are discussed. RCTs help in determining whether an intervention and its outcome have a cause and effect relationship or not. RCTs are used by many disciplines due to their efficacy in measuring the true impact of an intervention within the clinical setting whilst minimising bias. Bias refers to subjective prejudice in operations.
It can be both conscious and subconscious but impacts results greatly. Randomisation, which is a key feature in a RCT, offers a number of considerable benefits to the researchers. In a RCT, all participants have equal chances of being associated to any particular group. Some of the main benefits of randomisation include elimination of bias, reduced chances of participants influence on treatment selection, production of more comparable groups and ensured validity of statistical tests (Rosenberger & Lachin 2002).
In a research procedure, the first step is randomisation of the selection of participants (random sampling). The second step is the random allocation of the participants to one of the groups in the study. The principle of random sampling is that everyone or every sampling unit in the study population should have a predetermined (non-zero) chance, usually an equal chance of being selected out of the population being researched. This is probably the most important element of RCTs that contributes to eliminating bias and improving validity of the study, both internally and externally.
In this particular study by Ebenbichler et al. (1999), the sampling technique was described as being a multicentre sample of volunteers from the participation of outpatient clinics. The researchers referred the ones who showed interest in treatment to the outpatient clinic of the university hospital of Vienna. To be entitled to participate in the study, they needed to satisfy the inclusion criteria of idiopathic calcific tendinitis type one or type two. Unfortunately the authors did not provide sufficient information on the selection process from the target population.
Based on the information provided in the paper, I believe their chosen method of sampling was convenient sampling in which researchers gather data from one or multiple places, settings or sources. Researchers often choose convenience sampling because it is a cost-effective method of choosing an institution or area. Despite the advantages for the authors, this type of sampling is also a weak method as it can introduce bias and limits the transferability of outcomes to the target population. During the sampling process in Ebenbichler et al. (1999), confounding factors may have occurred.
In this study, there was a difference in the baseline values. Although the difference was minimal in terms of the characteristics of patients that started and those that completed the study. To increase external validity, I believe the sampling method should have been stratified because it could have ensured a better representation of the target population. This method would have enabled future users of the study to know the effect of ultrasound treatment in the reduction of pain in patients with calcific tendinitis of the shoulder.
While talking about the sampling technique, we should also discuss random allocation of subjects, which is one of the main features of randomisation. Researchers use random allocation to ensure that there is a complete similarity between the control groups and the intervention with the exception of the preventative or healing measure being tested. The choice of comparison treatments can also include either a standard way of treatment or a placebo in which there are no active ingredients. There are numerous types of random allocation methods used to allocate subjects having different purposes.
The study by Ebenbichler et al. (1999) stated a list of random numbers was generated by a spreadsheet programme that assigned participants to the trial group or the control group in the randomisation stage. This excluded all chances of bias by selecting participants according to shoulders. Due to the fact that patients involved in the study could be experiencing calcific tendinitis on either shoulder, the random allocation was based on the shoulder rather than the patient. The significance of this is that a patient could receive a sham treatment on one shoulder and a genuine ultrasound treatment on the other.
I think one obvious inadequacy in the study again is that they failed to specify the type of random allocation they adopted and how they had narrowed the number of participants down to seventy. The researchers state that they enrolled seventy patients of whom sixty-three participated in the study. There were thirty-two shoulders that participated in the active ultrasound treatment and twenty-nine in the sham ultrasound treatment out of the sixty-one that completed the study. I can presume that they used the simple random allocation method which is less reliable as compared to the stratified block randomisation that looks appropriate in such circumstances where different variables may influence the outcome of either the control or treatment group.
This method follows the fundamental principle of randomisation in which every individual has the same chance of being allocated into groups. Stratified random sampling is more preferable as it requires smaller sample sizes that translate into reduced costs for the study. Stratified sampling technique always gives a higher precision statistic as compared to simple sampling due to the fact that within the subgroups or subsections, there is low variability as compared with when one uses an entire population often associated with simple random sampling.
Blinding is another fundamental feature in RCTs. Blinding is very important to do because without it the chances of bias related to treatment increases (Chow & Liu 2004). Bias refers to unfairness or prejudice in operations, whereas blinding refers to the elimination of the chances of bias from the study by hiding the clinical information on treatment from the subjects as well as from the observers. In epidemiological studies, this subjectivity is often eliminated through blinding during experimental population sampling, allocation and clinical trial.
For research to be considered both internally and externally valid, as well as transferable to the target population, blinding should be rigorous throughout. Blinding also helps in hiding the identity of the treatment from all individuals involved, such as participants and researchers. In a single blinded study, subjects are not informed which group they have been assigned to, but the researchers are aware of whether they have been assigned to the control or treatment group. On the other hand, if we talk about double blinded studies, we come to know that in such studies, both subjects and observers are not aware of the nature of the assigned task, which makes such studies more accurate and reliable as compared to single blinded studies.
Double blind studies have other issues. They are expensive, take time to set up, and it is possible that the population that participates in the study does not actually represent a population as a whole. This type of study does not reveal causation either. However, the main thing that makes double-blind studies significant is that it removes subconscious bias from the researchers. Ebenbichler et al. (1999) clearly state that the study was double-blind. In order to ensure that this was a double-blind study, a therapist who was not involved in the study distributed sealed assignment envelops to the groups.
None of the participants (patients, evaluator or ultrasound technicians) were aware of the group the individual was assigned to. Hence this allowed this therapist to be the only one who knew which shoulder would receive real treatment and which one would receive sham treatment. Since both treatments felt the same, none of the subjects were able to determine, during testing, if they were receiving a real or sham treatment. Despite the advantages of double blinding studies, they have other issues.
They are expensive, they take time to set up, and it is possible that the population that participates in the study does not really represent a population as a whole. Ebenbicher et al. (1999) stated the use of double-blind comparison of sham and effective treatment in their study. This was evident as the treatment allocation was determined by a therapist who was not actively involved in the study. Later, a radiograph was performed on the subjects at each follow-up visit by two radiologists who were not aware of what treatment each participant received.
Further bias was minimised as the radiographs obtained were made under standardised conditions, and the same machine was used with the same settings and exposure for all participants in an attempt to ensure reliability and validity of the results. The therapist who switched the subjects’ treatment from an active one to a sham one also carried out the random allocation process. I think that although they were not blinded, they did not have an opportunity to effect results. These data were then passed onto researchers who were blinded to allocation and may have had an opportunity to implement bias during the processing stage.
I am not of the opinion that the results of the RCT by Ebenbichler et al. (1999) were the only method of determining the effectiveness of the particular intervention. Additionally, other research methods might have obvious differences that suggest diverse hypotheses and therefore, without the application of the detailed criteria set by RCT, the degree of certainty of an outcome can be limited. For example, the researchers could have used a cohort study to get more comparable results on the use of ultrasound therapy to treat calcific tendinitis of the shoulder.
Cohort studies would have also provided the researchers with the facility to investigate the causes and prognosis. The discrepancy that may be found in different aspects of the study that should be reported and those that are reported should be known (Begg et al. 1996, p. 637). For example, this study by Ebenbichler et al. (1999) does not clearly report or state if the study was actually an RCT. They did not mention the type of randomisation they used either. In conclusion, Ebenbichler et al. (1999) conducted an RCT to explore the essential role of ultrasound therapy in the treatment of calcific tendinitis of the shoulder.
The sample chosen does not fully represent those who suffered from the condition and from the treatment regime they chose to adopt. The randomisation process is not representative of all variables. For example, those using other treatment such as surgery and other interventions might not have been sampled. Considering a few constraints experienced by Ebenbichler et al. (1999) in terms of the limitation on the number of subjects or researchers and time constraints, future healthcare interventions should be evaluated through a comprehensive randomised controlled trial.
However, it is also important to note that the blinding feature of randomisation was used successfully as it was maintained throughout the study. I would say that the outcome of the study is valid and reliable. Moreover, it can also be used for future studies in the field of treatment of calcific tendinitis to the shoulders using ultrasound therapy. Word count : 1951 References Begg, C, Cho, M, Eastwood, S, Horton, R, Moher, D, Olkin, I et al. 1996, ‘Improving the quality of reporting of randomized controlled trials.
The consort statement’, JAMA: the Journal of the American Medical Association, vol. 276, no. 8, pp. 637–639. Chow, S & Liu, J 2004, Design and analysis of clinical trials: concepts and methodologies, 2nd ed, John Wiley & Sons, New Jersey. Ebenbichler, G, Erdogmus, C, Resch, K, Funovics, M, Kainberger, F, Barisani, G et al. 1999, ‘Ultrasound therapy for calcific tendinitis of the shoulder’, The New England Journal of Medicine, vol. 340, no. 20, pp. 1533–1538. Rosenberger, W & Lachin, J 2002, Randomization in clinical trials: theory and practice, John Wiley & Sons, New York.
Read More