I am going to try and explain some weaknesses and limitations of randomized controlled trials (RCTs) in this post. For even the golden standard of intervention studies, which RCTs are considered to be, does have flaws that are important to know of when reading such articles and when trying to understand evidence.
One must scrutinize the methodological rigor, such as the statistical control and validity in practical settings (Strale F Jr 2024). There are several more limitations, weaknesses and considerations (if not components at least) such as high cost that leads to intervention fidelity where validity and reliability is sacrificed (Cook CE et al 2019), rather short intervention duration, that healthy persons with a given condition participate in clinical trials although they are not representative of the population as a whole, publication bias, not correct use and insufficient knowledge of difference between Bayesian and frequentist statistical approaches, again low external and internal validity and incorrect statistical interference (Kostis JB et al 2020) and inadequate recruitment, underpowering, ceiling and floor effects, high placebo response, high drop-out, poor treatment adherence can scuttle RCTs (Andrade C 2024). Every component is therefore important for the success of a clinical trial (Morley R & Farewell V 2000).
One strength of RCTs is the randomization process which minimizes risk of some forms of bias and balance individual difference which can also give better cause-and-effect relationships (Strale F Jr 2024). It is however impossible to achieve a true randomization, therein lays the first limitation, since it can’t equalize everything (Carter, R. E. & Lubinsky, J. 2016, Deaton A et al 2017, Strale F Jr 2024) and can therefore be faulty which affects the internal validity. There is also always going to be some degree of unmeasured bias such as participants expectations and beliefs or there is contamination of exposure to the other groups intervention etcetera (Cook CE et al 2019). Randomization might induce bias by placebo and nocebo. This happens when participants have preconceived notions about treatments. For example, if participants have strong beliefs about acupuncture and want to be randomized to that group, they get placebo if they do but nocebo for the control intervention if they are randomized to that group. Being randomized to an intervention you do not want is also associated whit an increased risk of dropouts and non-compliance (Andrade C 2024, Fernainy P et al 2024). Neither are we able to control what the patients do outside of the study. It might happen that a control group starts taking the same supplement that is studied (i.e. Omega-3). It is also problematic that not all interventions can be blinded, such as yoga for example. The contamination of post randomization bias is also more likely to occur the longer the intervention is. But in the end, the most acknowledge limitation of RCTs is the narrow study selection (Andrade C 2024).
The narrow study selection is partly due to the difficulty to find and requit large samples of homogenous participants. Participants that because of ethics must volunteer and are therefore self-selected (Carter, R. E. & Lubinsky, J. 2016). Jet another ethical consideration is that it is unethical to perform RCTs if we already see harmful effects in other types of studies (for example observational, on the harms of smoking) (Saldanha IJ et al 2022). Not all potential participants will neither seek to participate, which means that we do not get a representable sample (Carter, R. E. & Lubinsky, J. 2016). Some participants will also be excluded based on limitation in max participants in the study, that meaning that there is not equal opportunity to participate due to RCTs being resource-intensive (Carter, R. E. & Lubinsky, J. 2016, Saldanha IJ et al 2022) and they can also take many years to conduct (Kostis JB et al 2020) which is yet another cost. All this means that generalisability of results is difficult. Another thing that makes generalisability difficult is that RCTs use averages, but there is no average person. Furthermore, result do not become applicable to broader populations due to short durations, tightly control of implementation interventions and comparators, smaller sample size, narrow eligibility criteria for participants (Saldanha IJ et al 2022). Results are also difficult to use since extraneous factors and effect of interventions can not be separated. Finally, some questions that RCTs does not answer is how effective clinical procedures are (Carter, R. E. & Lubinsky, J. 2016)? For whom will it work? (Carter, R. E. & Lubinsky, J. 2016), since some participants always will have effect and some will not in both groups, therefore RCT does not answer why someone improved but only how many/who it works for(Cook CE et al 2019, Deaton A et al 2017)
References
Andrade C. Poorly Recognized and Uncommonly Acknowledged Limitations of Randomized Controlled Trials. Indian J Psychol Med. 2025 Jan;47(1):83-85. doi: 10.1177/02537176241297953. Epub 2024 Nov 20. PMID: 39583300; PMCID: PMC11580118.
Carter, R. E., & Lubinsky, J. (2016). Rehabilitation research : principles and applications (Fifth edition). Elsevier.
Cook CE, Thigpen CA. Five good reasons to be disappointed with randomized trials. J Man Manip Ther. 2019 May;27(2):63-65. doi: 10.1080/10669817.2019.1589697. Epub 2019 Mar 14. PMID: 30935322; PMCID: PMC6484499.
Deaton A, Cartwright N. Understanding and misunderstanding randomized controlled trials. Soc Sci Med. 2018 Aug;210:2-21. doi: 10.1016/j.socscimed.2017.12.005. Epub 2017 Dec 25. PMID: 29331519; PMCID: PMC6019115.
Fernainy P, Cohen AA, Murray E, Losina E, Lamontagne F, Sourial N. Rethinking the pros and cons of randomized controlled trials and observational studies in the era of big data and advanced methods: a panel discussion. BMC Proc. 2024 Jan 18;18(Suppl 2):1. doi: 10.1186/s12919-023-00285-8. Erratum in: BMC Proc. 2024 Aug 16;18(1):16. doi: 10.1186/s12919-024-00299-w. PMID: 38233894; PMCID: PMC10795211.
Morley R, Farewell V. Methodological issues in randomized controlled trials. Semin Neonatol. 2000 May;5(2):141-8. doi: 10.1053/siny.1999.0004. PMID: 10859708.
Saldanha IJ, Skelly AC, Ley KV, et al. Inclusion of Nonrandomized Studies of Interventions in Systematic Reviews of Intervention Effectiveness: An Update [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2022 Sep. 3, Strengths and Limitations of RCTs. Available from: https://www.ncbi.nlm.nih.gov/books/NBK584466/
Strale F Jr. Examining the Efficacy of Control Groups in Achieving Statistical Control: A Critical Look at Randomized Controlled Trials. Cureus. 2024 Sep 30;16(9):e70562. doi: 10.7759/cureus.70562. PMID: 39483947; PMCID: PMC11524751.