More Subjects
Research Methods For The Health Sciences
Research methods for health sciences
[Author Name(s), First M. Last, Omit Titles and Degrees]
[Institutional Affiliation(s)]
Author Note
[Include any grant/funding information and a complete correspondence address.]
Research methods for health sciences
Randomized control trial (RCT) is one of the study designs used to test hypothesis on clinical trials and compare between intervention and control groups. The study design takes the highest rank for quality of research designs. This document aims to explain the use and effectiveness of randomized controlled trials (RCTs). The design has been compared with the other study designs based on methodology, effectiveness, legal, and ethical issues. The internal and external validity has been questioned comparing using available evidence with other research strategies. Lastly, various confounders have been identified and how they affect the internal and external validity of an RCT.
Randomized controlled trials
RCT is a type of experimental study design which involves testing hypothesis by incorporating participants randomly in two different groups; experimental group and control group ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"H6s8q5yr","properties":{"formattedCitation":"(Bhide, Shah, & Acharya, 2018)","plainCitation":"(Bhide, Shah, & Acharya, 2018)","noteIndex":0},"citationItems":[{"id":86,"uris":["http://zotero.org/users/local/YjWHJPzk/items/7QG9VK4G"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/7QG9VK4G"],"itemData":{"id":86,"type":"article-journal","title":"A simplified guide to randomized controlled trials","container-title":"Acta obstetricia et gynecologica Scandinavica","page":"380-387","volume":"97","issue":"4","author":[{"family":"Bhide","given":"Amar"},{"family":"Shah","given":"Prakesh S."},{"family":"Acharya","given":"Ganesh"}],"issued":{"date-parts":[["2018"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Bhide, Shah, & Acharya, 2018). The distribution of participants into two groups is purely random compared to non-randomized trials where participants are chosen and grouped based on personal choice. Because of the selection and distribution process, the non RCTs are subject to selection and outcome, whereas evidence from RCTs is regarded as the highest quality research evidence ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"tseLbhgR","properties":{"formattedCitation":"(Boland, Cherry, & Dickson, 2017)","plainCitation":"(Boland, Cherry, & Dickson, 2017)","noteIndex":0},"citationItems":[{"id":93,"uris":["http://zotero.org/users/local/YjWHJPzk/items/GUTIK5NI"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/GUTIK5NI"],"itemData":{"id":93,"type":"book","title":"Doing a systematic review: A student's guide","publisher":"Sage","ISBN":"1-5264-1658-1","author":[{"family":"Boland","given":"Angela"},{"family":"Cherry","given":"Gemma"},{"family":"Dickson","given":"Rumona"}],"issued":{"date-parts":[["2017"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Boland, Cherry, & Dickson, 2017).
Additionally, randomized trials also provide a strong basis and evidence for systematic reviews and meta-analysis, therefore getting popular in a variety of researches. However, it is important to note that RCTs can be inappropriate and prone to bias if inefficiently done. Another important consideration to make is that the whole process of RCT design, conduction, results, interpretation, and distribution is complicated and requires a lot of effort and time ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"qTU8rKPK","properties":{"formattedCitation":"(Boland et al., 2017)","plainCitation":"(Boland et al., 2017)","noteIndex":0},"citationItems":[{"id":93,"uris":["http://zotero.org/users/local/YjWHJPzk/items/GUTIK5NI"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/GUTIK5NI"],"itemData":{"id":93,"type":"book","title":"Doing a systematic review: A student's guide","publisher":"Sage","ISBN":"1-5264-1658-1","author":[{"family":"Boland","given":"Angela"},{"family":"Cherry","given":"Gemma"},{"family":"Dickson","given":"Rumona"}],"issued":{"date-parts":[["2017"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Boland et al., 2017).
Feasibility/appropriateness
RCTs and other experimental studies are considered more applicable than observational studies. The observational studies are completely based on observation of a phenomenon and presentation of information related to it. The RCTs are most suitable for measurement of effectiveness in pre-clinical and clinical researches. The intervention is followed prospectively after allocation to two or more groups. In the case of clinical trials, the intervention is proposed on the basis of logic, previously done similar researches or laboratory studies. Observational studies show the effectiveness of, but they are highly prone to bias ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"SIcquVQw","properties":{"formattedCitation":"(Bhide et al., 2018)","plainCitation":"(Bhide et al., 2018)","noteIndex":0},"citationItems":[{"id":86,"uris":["http://zotero.org/users/local/YjWHJPzk/items/7QG9VK4G"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/7QG9VK4G"],"itemData":{"id":86,"type":"article-journal","title":"A simplified guide to randomized controlled trials","container-title":"Acta obstetricia et gynecologica Scandinavica","page":"380-387","volume":"97","issue":"4","author":[{"family":"Bhide","given":"Amar"},{"family":"Shah","given":"Prakesh S."},{"family":"Acharya","given":"Ganesh"}],"issued":{"date-parts":[["2018"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Bhide et al., 2018). The questions to be answered in an RCT design are free of participants' opinions, and the research proceeds without any risk of bias.
It must be noted that the RCTs involve putting people on trials of various types, which requires ethical clearance. According to researches, it is ethically not right to put people on risk of a clinical trial if the outcomes are not scientifically contributable. Some researches involving harmful health practices do not get ethical clearance ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"mMpehVjA","properties":{"formattedCitation":"(Boland et al., 2017)","plainCitation":"(Boland et al., 2017)","noteIndex":0},"citationItems":[{"id":93,"uris":["http://zotero.org/users/local/YjWHJPzk/items/GUTIK5NI"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/GUTIK5NI"],"itemData":{"id":93,"type":"book","title":"Doing a systematic review: A student's guide","publisher":"Sage","ISBN":"1-5264-1658-1","author":[{"family":"Boland","given":"Angela"},{"family":"Cherry","given":"Gemma"},{"family":"Dickson","given":"Rumona"}],"issued":{"date-parts":[["2017"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Boland et al., 2017). One such example is oral/intravenous administration of alcohol to pregnant women because of the damages it causes to the mother and developing fetus ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"QBEPQfCQ","properties":{"formattedCitation":"(Bhide et al., 2018)","plainCitation":"(Bhide et al., 2018)","noteIndex":0},"citationItems":[{"id":86,"uris":["http://zotero.org/users/local/YjWHJPzk/items/7QG9VK4G"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/7QG9VK4G"],"itemData":{"id":86,"type":"article-journal","title":"A simplified guide to randomized controlled trials","container-title":"Acta obstetricia et gynecologica Scandinavica","page":"380-387","volume":"97","issue":"4","author":[{"family":"Bhide","given":"Amar"},{"family":"Shah","given":"Prakesh S."},{"family":"Acharya","given":"Ganesh"}],"issued":{"date-parts":[["2018"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Bhide et al., 2018).
In some cases, RCTs are considered inappropriate even after granting of ethical clearance, especially when the research fails to contribute to participant’s treatment. Therefore, it is crucial to carry all steps carefully and responsibly to minimize the risk of harm to participants and produce valuable outcomes ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"447LL7zg","properties":{"formattedCitation":"(Bondemark & Ruf, 2015)","plainCitation":"(Bondemark & Ruf, 2015)","noteIndex":0},"citationItems":[{"id":90,"uris":["http://zotero.org/users/local/YjWHJPzk/items/RF85P3AX"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/RF85P3AX"],"itemData":{"id":90,"type":"article-journal","title":"Randomized controlled trial: the gold standard or an unobtainable fallacy?","container-title":"European Journal of Orthodontics","page":"457–461","volume":"37","issue":"5","source":"Google Scholar","title-short":"Randomized controlled trial","author":[{"family":"Bondemark","given":"Lars"},{"family":"Ruf","given":"Sabine"}],"issued":{"date-parts":[["2015"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Bondemark & Ruf, 2015). Moreover, contrary to all the advantages mentioned above, RCT design is not appropriate to study history or pattern of disease and the diseases which take a long time to develop.
Internal validity/External validity
Internal validity is the amount of trustworthiness a study carries for cause and effect relationship between an outcome and an intervention. Many factors contribute to an increase in internal validity, including randomization, blinding, experimental studies, and study procedures. An RCT design successfully accomplishes most of them compared to other designs like non-randomized controlled trials; which do not follow the random allocation method, observation studies; which do not involve any kind of experiment on the chosen population, case control, or cohort studies. However, the risk of bias still exists in RCTs because of some potential confounding factors, blinding methods, and study procedures ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"g2gWzhRU","properties":{"formattedCitation":"(Cartwright, 2007)","plainCitation":"(Cartwright, 2007)","noteIndex":0},"citationItems":[{"id":88,"uris":["http://zotero.org/users/local/YjWHJPzk/items/XPQTWKU7"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/XPQTWKU7"],"itemData":{"id":88,"type":"article-journal","title":"Are RCTs the gold standard?","container-title":"BioSocieties","page":"11-20","volume":"2","issue":"1","author":[{"family":"Cartwright","given":"Nancy"}],"issued":{"date-parts":[["2007"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Cartwright, 2007).
In contrast to internal validity, external validity refers to the amount of implacability of the research to other study settings. It is measured as generalizability or transferability of the study findings. External validity can be ensured by considering inclusion/exclusion criteria, replication, and field experiment. The selection criteria need to be defined precisely and accurately for the replication of study in a different setting. Field experiment means research should be conducted outside a laboratory in on general population.
However, some situational variables like time and location, sample features, selection bias, and pre-post test effects have a negative impact on external validity of the research. Some other designs can prove to have the highest external validity if confounders are efficiently addressed. Therefore, it is important to recognize the internal and external factors which can affect the outcome of interest in any study design and propose ways to address them.
Confounding/bias
Comparing two or more groups in an observational study, may not yield significantly credible results as the differences observed may not be because of exposure or intervention, instead can be caused by changes in groups themselves. This is called confounding ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"vWdGRUhb","properties":{"formattedCitation":"(Cartwright, 2007)","plainCitation":"(Cartwright, 2007)","noteIndex":0},"citationItems":[{"id":88,"uris":["http://zotero.org/users/local/YjWHJPzk/items/XPQTWKU7"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/XPQTWKU7"],"itemData":{"id":88,"type":"article-journal","title":"Are RCTs the gold standard?","container-title":"BioSocieties","page":"11-20","volume":"2","issue":"1","author":[{"family":"Cartwright","given":"Nancy"}],"issued":{"date-parts":[["2007"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Cartwright, 2007). However, if the sample is sufficiently large, the differences are believed to exist because of intervention and not the sample. The confounding in RCT makes the other study designs like observational studies a better approach. The only way to handle such disparities is random allocation; allocating each participant randomly to either intervention or control group ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"XAh77wsv","properties":{"formattedCitation":"(Howards, 2018)","plainCitation":"(Howards, 2018)","noteIndex":0},"citationItems":[{"id":87,"uris":["http://zotero.org/users/local/YjWHJPzk/items/AXPKYSP8"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/AXPKYSP8"],"itemData":{"id":87,"type":"article-journal","title":"An overview of confounding. Part 1: the concept and how to address it","container-title":"Acta obstetricia et gynecologica Scandinavica","page":"394-399","volume":"97","issue":"4","author":[{"family":"Howards","given":"Penelope P."}],"issued":{"date-parts":[["2018"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Howards, 2018). Randomization can be individual or clustered. Individual randomization refers to assorting all the participants randomly into control and treatment groups, whereas cluster randomization refers to classifying clusters or groups of participants for treatment and control. Cluster randomization does not offer a true or valid comparison between the two study groups, and the research is more susceptible to limitation of confounding ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"A6KMPlHW","properties":{"formattedCitation":"(Miettinen & Cook, 1981)","plainCitation":"(Miettinen & Cook, 1981)","noteIndex":0},"citationItems":[{"id":94,"uris":["http://zotero.org/users/local/YjWHJPzk/items/96VLCLK5"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/96VLCLK5"],"itemData":{"id":94,"type":"article-journal","title":"Confounding: essence and detection","container-title":"American journal of epidemiology","page":"593-603","volume":"114","issue":"4","author":[{"family":"Miettinen","given":"Olli S."},{"family":"Cook","given":"E. Francis"}],"issued":{"date-parts":[["1981"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Miettinen & Cook, 1981). To ensure the credibility of such clustered trials and minimize confounding, it is suggested to keep the sample size higher.
Confounding leads to various types of bias. Bias refers to any type of systematic error that cause incorrect measurement of the effect of exposure/treatment on the results of the study. There can be various factors acting as potential confounders in a study ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"t0e7svyD","properties":{"formattedCitation":"(Brookhart, St\\uc0\\u252{}rmer, Glynn, Rassen, & Schneeweiss, 2010)","plainCitation":"(Brookhart, Stürmer, Glynn, Rassen, & Schneeweiss, 2010)","noteIndex":0},"citationItems":[{"id":96,"uris":["http://zotero.org/users/local/YjWHJPzk/items/Q8JPITAP"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/Q8JPITAP"],"itemData":{"id":96,"type":"article-journal","title":"Confounding control in healthcare database research: challenges and potential approaches","container-title":"Medical care","page":"S114","volume":"48","issue":"6 0","author":[{"family":"Brookhart","given":"M. Alan"},{"family":"Stürmer","given":"Til"},{"family":"Glynn","given":"Robert J."},{"family":"Rassen","given":"Jeremy"},{"family":"Schneeweiss","given":"Sebastian"}],"issued":{"date-parts":[["2010"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Brookhart, Stürmer, Glynn, Rassen, & Schneeweiss, 2010). One mentioned above is the existence of individual differences among groups to be compared. Other than that, some outside variables can be a potential source of impact higher than the intervention itself, causing misleading results.
Blinding help to reduce participant’s or researcher’s bias. Double blinding which is keeping the randomization from both the participants and the observer help to maximize the quality of the results. Single blinding, on the other hand, involves randomization for the subjects, not the researcher ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"t0e7svyD","properties":{"formattedCitation":"(Brookhart, St\\uc0\\u252{}rmer, Glynn, Rassen, & Schneeweiss, 2010)","plainCitation":"(Brookhart, Stürmer, Glynn, Rassen, & Schneeweiss, 2010)","noteIndex":0},"citationItems":[{"id":96,"uris":["http://zotero.org/users/local/YjWHJPzk/items/Q8JPITAP"],"uri":["http://zotero.org/users/local/YjWHJPzk/items/Q8JPITAP"],"itemData":{"id":96,"type":"article-journal","title":"Confounding control in healthcare database research: challenges and potential approaches","container-title":"Medical care","page":"S114","volume":"48","issue":"6 0","author":[{"family":"Brookhart","given":"M. Alan"},{"family":"Stürmer","given":"Til"},{"family":"Glynn","given":"Robert J."},{"family":"Rassen","given":"Jeremy"},{"family":"Schneeweiss","given":"Sebastian"}],"issued":{"date-parts":[["2010"]]}}}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} (Brookhart, Stürmer, Glynn, Rassen, & Schneeweiss, 2010).
Conclusion
To sum up, it can be said that RCTs generate high quality outcomes by accurately and precisely following all the steps. They are the most feasible designs for clinical trials and intervention studies. Internal and external validity of an RCT can be ensured by identifying the possible confounders and addressing them properly. One of the confounding factors is the differences between control and intervention groups. Others include outside variables like the environment. The large sample sizes, precisely defined selection criteria and defined procedures, help to limit the confounding and bias in a RCT. However, the RCT is not a suitable design to measure long term epidemiology of a disease. And even efficiently done, the susceptibility of research to bias cannot be ignored.
References
Boland, A., Cherry, G., & Dickson, R. (Eds.). (2017). Doing a systematic review: A student's guide. Sage.
Bhide, A., Shah, P. S., & Acharya, G. (2018). A simplified guide to randomized controlled
trials. Acta obstetricia et gynecologica Scandinavica, 97(4), 380-387.
ADDIN ZOTERO_BIBL {"uncited":[],"omitted":[],"custom":[]} CSL_BIBLIOGRAPHY Bhide, A., Shah, P. S., & Acharya, G. (2018). A simplified guide to randomized controlled trials. Acta Obstetricia et Gynecologica Scandinavica, 97(4), 380–387.
Boland, A., Cherry, G., & Dickson, R. (2017). Doing a systematic review: A student’s guide. Sage.
Bondemark, L., & Ruf, S. (2015). Randomized controlled trial: The gold standard or an unobtainable fallacy? European Journal of Orthodontics, 37(5), 457–461.
Brookhart, M. A., Stürmer, T., Glynn, R. J., Rassen, J., & Schneeweiss, S. (2010). Confounding control in healthcare database research: Challenges and potential approaches. Medical Care, 48(6 0), S114.
Cartwright, N. (2007). Are RCTs the gold standard? BioSocieties, 2(1), 11–20.
Howards, P. P. (2018). An overview of confounding. Part 1: The concept and how to address it. Acta Obstetricia et Gynecologica Scandinavica, 97(4), 394–399.
Miettinen, O. S., & Cook, E. F. (1981). Confounding: Essence and detection. American Journal of Epidemiology, 114(4), 593–603.
More Subjects
Join our mailing list
© All Rights Reserved 2024