I have read with interest the article by Dr Duterloo, written in response to an editorial by Dr Behrents in the November 2016 issue of the AJO-DO . Dr Duterloo found the editorial to be incomplete, and he wrote to address ethical issues in the context of orthodontic research, the use of controls, and the value of randomized trials compared with retrospective studies.
I also read Dr Behrent’s editorial and found it clear and to the point. The fact that a study should adhere to the ethical guidelines is a given. Dr Behrent’s editorial was a short communication and was not intended to be comprehensive, since it would be impossible to tackle all aspects of clinical studies in a 3-page editorial.
In his editorial, Behrents wrote: “each experimental trial builds direction for the next inquiry.” Dr Duterloo thought that this idea, “although apparently evident and straightforward, raises doubt.” He cited, by way of example, the response to the 2016 Class III early facemask trial by Mandall et al.
The idea of building on the existing evidence when designing and also interpreting new research is universally accepted in the medical field and is fundamental to Cochrane. It is considered unethical to conduct a trial if, for example, the result is already known, because this had negative effects on patients in the past, or to expose patients to a potentially inferior treatment when there is an effective treatment available. Other reasons include unnecessary duplication and waste.
A clinical trial is ethically justified if there is genuine uncertainty (equipoise) in the scientific community about the effectiveness and safety of an intervention. It is important that within the trial context there are provisions under the good clinical practice umbrella that protect the rights and safety of patients. Those rights and safety of the patients are preserved with the signed informed consent, the use of ethics and data-monitoring committees, the management of adverse events and serious adverse events, and patient confidentiality. Those measures are not in contrast with equipoise; Djulbegovic provided an excellent account on this concept.
The contention that later surgery is riskier (Mandall et al ) and therefore that those in an untreated control group are subject to a bad deal is posited on knowledge of the effect of the trial intervention, which is obviously not known at the start of a trial—ie, exposure to additional risks is post hoc thinking. Assessing the ethical justification of a trial after it has been completed and after looking at the results is clearly not correct. Therefore, the patients in the control group were not “ethically harmed” by receiving no treatment as pointed out by Duterloo, because well-conducted studies on the effect of facemasks in young patients were not available at the time. A trial is ethically justified if the answer is unknown. This trial by Mandall et al has informed our view on the effectiveness of the intervention as a well-executed randomized controlled trial (RCT) can.
In addition, not including a “control” group will not allow for the distinction of the effects of the appliance or treatment and growth or time changes, and thus will not allow for a fair and unbiased comparison. A trial with “biased” results is also unethical because it “imposes” a treatment on patients that may be ineffective, and this ineffectiveness perhaps is not evident because of the lack of a control. In clinical trial ethics, there is the notion of what is best for the individual and what is best for the community. Strictly accepting the former will not permit clinical trials on ethical grounds; therefore, there is a trade-off between individual and collective good.
It is important to note that Mandall et al reported no differences in skeletal and occlusal improvement, self-esteem, and oral esthetic impact between treated and untreated patients. If those issues were elements of the patients’ and parents’ chief complaint and were the reasons to seek treatment, then the evaluated facemask protocol for Class III correction would not meet the clinician’s and patient’s treatment expectations. Therefore, patients in the control group were not exposed to the treatment, which did not produce the assumed objectives of such therapy. This is an important finding in the decision-making treatment process and in the era of increased emphasis on patient-important outcomes in medicine and dentistry. This lack of the effect was detected because of this well-designed RCT and the use of controls; those findings should be considered if we really want to meet our patients’ needs.
Orthodontics is an elective treatment, and key components of evidence-based medicine are patient values and preferences. The patient preferences component weights heavy in our treatment choices because we commonly address less weighty problems such as dental and facial esthetics. On the contrary, in medicine, tougher ethical decisions are encountered when, for example, not receiving treatment could mean death. Again, in those circumstances if there is uncertainty about the effectiveness of a therapy, the trial can be ethically justified.
It is common for patients to choose not to receive orthognathic surgery, and it is very likely that patients in the control group classified as orthognathic cases would choose not to have orthognathic therapy. According to Mandall et al, facemask treatment did not produce significant skeletal improvement when compared with the untreated controls in the long term, so the need for orthodontic surgery to improve facial esthetics was not eliminated in all patients. In fact, the results of this study documented that a third of the patients treated with a facemask still needed orthognathic surgery after the completion of growth. Therefore, allocation to the control group would not necessarily compromise them, given the results of the trial. The trial’s findings were in favor of the reduced risk of orthognathic surgery in the rapid palatal expansion and facemask group, however, with a good amount of uncertainty after interpreting the wide 95% confidence intervals (odds ratio = 3.34; 95% CI, 1.21-9.24).
There is no need to conduct an RCT if the answer is already known. For example, no RCT is required to prove that we can align teeth, and there is no need to set up a trial where 1 group is fitted with braces and the other group receives no treatment. Glasziou et al succinctly stated: “A unifying principle is the size of the treatment effect (signal) relative to the expected prognosis (noise) of the condition. A treatment effect is inferred most confidently when the signal to noise ratio is large and its timing is rapid compared with the natural course of the condition.” As such, rapid palatal expansion and facemask treatment is effective in improving occlusal relationships in the short term, and perhaps there is no need to undertake a trial to prove it; however, the effects on long-term occlusal relationships, temporomandibular joints, and self-esteem either are uncertain or have not been assessed. Therefore, the trial of Mandall et al was ethically justified.
Further on, Dr Duterloo brought into the discussion the ethics of the decisions made during peer review and, if I understood correctly, questioned whether the editor and reviewers have the right to reject a paper due to lack of control. Again, those are issues that have been considered by many interested parties and have produced the current standards. There are standards regarding the quality of the evidence, and journals may choose to publish or reject based on the scientific rigor. Ideally, everything should be published, since the All Trials initiative has been advocating this. However, it is not possible to publish everything a journal receives. In addition, there is a need for wider and better understanding among clinicians on how to critically appraise the published literature. Unquestioningly, equating publication with trust in the study findings is likely to result in suboptimal or even harmful treatments.
All studies are useful, and it should be clear that different studies are suitable to answer different or certain questions and can often be complimentary. It is also important that they are placed in the correct hierarchy in terms of their quality and trustworthiness. A recent publication discussed this issue and stated the following in the abstract.
Both randomized and nonrandomized studies [includes retrospective] are integral to orthodontic research and practice because they permit evaluation of relationships between exposures and outcomes, allowing the efficacy, effectiveness, and safety of interventions to be assessed. These designs allow clinical decisions to be informed. Nonrandomized designs include nonrandomized clinical trials, cohort studies, case-control studies, cross-sectional studies, case series, and ecological studies. There is debate surrounding the optimal research design; however, both randomized and nonrandomized designs are important to build a broad, informative evidence base. The designs are therefore complementary, with unique advantages and limitations. The applicability of either approach hinges on the clinical question posed, the feasibility of studying it, and ethical considerations.
The authors closed as follows, quoting Rawlins.
We believe that the following quote fittingly summarizes the synergy between observational and randomized studies: ‘Experiment, observation, and mathematics, individually and collectively, have a crucial role in providing the evidential basis for modern therapeutics. Arguments about the relative importance of each are an unnecessary distraction. Hierarchies of evidence should be replaced by accepting—indeed embracing—a diversity of approaches. This is not a plea to abandon RCTs and replace them with observational studies. Rather, it is a plea to investigators to continue to develop and improve their methods; to decision makers, to avoid adopting entrenched positions about the nature of evidence; and for both, to accept that the interpretation of evidence requires judgment.
Evidence-based medicine and dentistry require the best available evidence and consider all of this. However, they grade the evidence based on experimentally proved criteria. It is possible that a retrospective study will give the same or similar results as an RCT; however, our confidence is much greater if the results come from a better designed study. The bottom line is that a high-quality study is more trusted than a low-quality study, and the reason is that scientific experimentation has shown that findings from high-quality studies are more likely to be correct compared with those of low-quality studies, in terms of both effectiveness and safety.
Lower-level evidence is not necessarily false; in fact, lower-level studies have resulted in important breakthroughs, such as the discovery of penicillin. However, lower-level studies carry a greater risk of “false-positive” results and thus have a higher chance of leading to recommendations that are not in patients’ best interests. In the context of evidence-based dentistry, the position of a study design on the pyramid of evidence does not necessarily indicate the validity of the results but, rather, the priority it is given in decision making for treatment recommendations.