Gastroenterological Endoscopy. Группа авторов
Чтение книги онлайн.
Читать онлайн книгу Gastroenterological Endoscopy - Группа авторов страница 30
Название: Gastroenterological Endoscopy
Автор: Группа авторов
Издательство: Ingram
Жанр: Медицина
isbn: 9783131470133
isbn:
Statisticians play a key role in the design and analysis of studies. A common mistake is to involve statisticians only at the end of the study when analysis is needed. A much more effective strategy is to involve statisticians at the planning stages. In this way, variables can be carefully defined and chosen in a way that will optimize data analysis. Statisticians can also significantly improve the overall study design. For example, simple changes in study design can substantially alter the sample size needed.6
Finally, partnering with editorial assistants may be highly valuable for some investigators. The skill of writing manuscripts is very different than the skill of conducting research. Many large academic centers have medical editors who can facilitate how we communicate our scientific discovery with the broader community.
2.3 Designing Clinical Trials
The field of endoscopy has matured substantially over the past 20 to 30 years. From one where simply describing our observations and experience was enough to be published, to now where competitive journals typically only publish well-designed controlled clinical trials and cohort studies. In order to be published, such high-quality clinical studies should be carefully designed to achieve our primary goal of seeking scientific truth. Designing clinical trials follows a general pattern from generating ideas, to study design, to grant writing, and finally completion of the study. Each of these is discussed further.
2.3.1 Generating Ideas
Generating ideas should be the easiest of all research activities. All those involved in patient care know that many decisions we make, for both diagnosis and treatment, have only a limited amount of scientific evidence. Thus, in almost every patient encounter, we can identify opportunities for research.
2.3.2 Refining Ideas
Many studies can take months or even years to complete, so the long list of possible research topics needs to be refined based on several key factors:
• Is the topic of high interest to the investigator?
• Are the resources to study the question available to the investigator, including adequate numbers of patients, access to large databases, collaborators with sufficient expertise, and funding sources?
Research questions should be further defined based on a very detailed review of the current literature. Ultimately, research is aimed at extending the envelope of knowledge beyond what is currently known. Many resources are available (PubMed, Google Scholar, Medline, etc.) to identify current knowledge and its gaps including review of published research and consultation with other experts. Almost all published studies end with a statement such as “further research is necessary to confirm/clarify….” These statements offer excellent clues on how to further refine a specific research question. Moreover, some outcomes need to be confirmed or even excluded because they seem clinically not rational.
2.3.3 Clinical Trial Design
Clinical trial design balances precision and accuracy of a particular research question with available resources. Theoretical and practical issues must be taken into consideration. For treatment trials, the definitive randomized control trial is often not feasible, cost-effective, or even ethical.
Studies of new diagnostic technologies, which are particularly common in the field of endoscopy, often begin with a pilot study assessing the general safety and efficacy of a new device such as a new endoscopic imaging technology. This should initially be compared to historical controls. If promising, further studies should then be performed comparing the new method to the current standard in a controlled cohort or randomized study. Many diagnostic trials follow a crossover design where each patient undergoes both procedures, either back-to-back or in a sequential crossover design. Such methods may reduce by 10- to 20-fold the number of patients needed compared with a simple randomized design since each patient serves as his/her own control, thus minimizing variability.6 Nonetheless, particularly back-to-back studies are prone to bias if not well conducted, for example, when the same investigator performs both procedures in the same subject.
Treatment studies also began typically as an initial safety study. Ideally, these should be compared to historical controls and, if the data are promising, lead to prospective randomized controlled trials. The classic randomized controlled trial is well-suited in this area and can be done efficiently, particularly when there are significant improvements with a new technology compared to existing technology. Such studies have led to major landmark publications and have set new standards of care for endoscopy.7
Trials of causation and association, such as the link between Helicobacter pylori and gastric ulcers and gastric lymphoma,8 often cannot be addressed through prospective clinical trials. In these cases, large cohort or case–control studies may be better suited. Cohort studies are valuable for common conditions such as the association between nonsteroidal anti-inflammatory drug use and peptic ulcer disease. However, for more rare conditions, case–control study design is more efficient. This includes studies such as the association between gastroesophageal reflux disease and esophageal adenocarcinoma.9
2.3.4 Grant Writing
The skill of grant writing is similar to the skill of publishing clinical trials. A well-written grant must convince the funding agency that:
• the question is important.
• your methods are well-suited.
• the question and approach is novel.
• your team is the best one to answer the question.
Most successful research teams build on a long arc of successful investigation in the specific area. Because of their established record, they can achieve each of the elements mentioned above and continue to push the frontier of knowledge forward. Having said this, it also means that the initial steps on the research path are often not easy!
2.3.5 Conducting Clinical Trials
Once the study is designed and funded, the day-to-day work of completing the study is often assigned to study coordinators, with oversight from the principal investigator. If the idea was well formed and refined, the trial carefully designed, and the team of collaborators well chosen, clinical trial conduct usually proceeds without difficulties. However, even in these circumstances, regular meetings to review enrollment and identify any problems with study data collection are very important. In some studies, where the outcomes are highly uncertain, a planned interim analysis should be considered to allow for appropriate adjustment.
2.3.6 Presentation and National Meetings
Once the study is complete, or during a planned interim analysis, the communication of study results is often done in a multifactorial way including presentation at national meetings. This is often the first opportunity to communicate important results to colleagues and to receive feedback. By their nature, presentations at national meetings are very short relative to the full publication. Most large meetings include options СКАЧАТЬ