[통계, 펌글] Randomized controlled trial from wikipedia

A randomized controlled trial (RCT) (or randomized control trial or randomized comparative trial) is a specific type of scientific experiment, and the gold standard for a clinical trial. RCT are often used to test the efficacy and/or effectiveness of various types of medical intervention within a patient population. RCT may also provide an opportunity to gather useful information about adverse effects, such as drug reactions.

The key distinguishing feature of the usual RCT is that study subjects, after assessment of eligibility and recruitment, but before the intervention to be studied begins, are randomly allocated to receive one or other of the alternative treatments under study. Random allocation in real trials is complex, but conceptually, the process is like tossing a coin. After randomization, the two (or more) groups of subjects are followed in exactly the same way, and the only differences between the care they receive, for example, in terms of procedures, tests, outpatient visits, follow-up calls etc. should be those intrinsic to the treatments being compared. The most important advantage of proper randomization is that it minimizes allocation bias, balancing both known and unknown prognostic factors, in the assignment of treatments."^[2]

The terms "RCT" and randomized trial are often used synonymously, but some authors distinguish between "RCTs" which compare treatment groups with control groups not receiving treatment (as in a placebo-controlled study), and "randomized trials" which can compare multiple treatment groups with each other.^[3] RCTs are sometimes known as randomized control trials.^[4] RCTs are also calledrandomized clinical trials or randomized controlled clinical trials when they concern clinical research;^[5][6][7] however, RCTs are also employed in other research areas, including many of the social sciences, where their relevance and the advantages claimed for them have been contested in the literature.

History

Randomized experiments first appeared in psychology, where they were introduced by Charles Sanders Peirce,^[8] and in education.^[9][10][11] Later, randomized experiments appeared in agriculture, due to Jerzy Neyman^[12] and Ronald A. Fisher. Fisher's experimental research and his writings popularized randomized experiments.^[13]

The first published RCT appeared in the 1948 paper entitled "Streptomycin treatment of pulmonary tuberculosis", which described a Medical Research Council investigation.^[14][15][16] One of the authors of that paper was Austin Bradford Hill, who is credited as having conceived the modern RCT.^[17]

By the late 20th century, RCTs were recognized as the standard method for "rational therapeutics" in medicine.^[18] As of 2004, more than 150,000 RCTs were in the Cochrane Library.^[17] To improve the reporting of RCTs in the medical literature, an international group of scientists and editors published Consolidated Standards of Reporting Trials (CONSORT) Statements in 1996, 2001, and 2010 which have become widely accepted.^[1][2] Randomization is the processs of assigning trial subjects to treatment or control groups using an element of chance to determine the assignments in order to reduce the bias

Classification of RCT

By study design[edit]

One way to classify RCTs is by study design. From most to least common in the medical literature, the major categories of RCT study designs are:^[27]

• Parallel-group – each participant is randomly assigned to a group, and all the participants in the group receive (or do not receive) an intervention.
• Crossover – over time, each participant receives (or does not receive) an intervention in a random sequence.^[28][29]
• Cluster – pre-existing groups of participants (e.g., villages, schools) are randomly selected to receive (or not receive) an intervention.
• Factorial – each participant is randomly assigned to a group that receives a particular combination of interventions or non-interventions (e.g., group 1 receives vitamin X and vitamin Y, group 2 receives vitamin X and placebo Y, group 3 receives placebo X and vitamin Y, and group 4 receives placebo X and placebo Y).

An analysis of the 616 RCTs indexed in PubMed during December 2006 found that 78% were parallel-group trials, 16% were crossover, 2% were split-body, 2% were cluster, and 2% were factorial

Randomization[edit]

The advantages of proper randomization in RCTs include:^[32]

• "It eliminates bias in treatment assignment," specifically selection bias and confounding.
• "It facilitates blinding (masking) of the identity of treatments from investigators, participants, and assessors."
• "It permits the use of probability theory to express the likelihood that any difference in outcome between treatment groups merely indicates chance."

There are two processes involved in randomizing patients to different interventions. First is choosing a randomization procedure to generate an unpredictable sequence of allocations; this may be a simple random assignment of patients to any of the groups at equal probabilities, may be "restricted," or may be "adaptive." A second and more practical issue is allocation concealment, which refers to the stringent precautions taken to ensure that the group assignment of patients are not revealed prior to definitively allocating them to their respective groups. Non-random "systematic" methods of group assignment, such as alternating subjects between one group and the other, can cause "limitless contamination possibilities" and can cause a breach of allocation concealment.^[32]

Randomization procedures[edit]

The treatment allocation is the desired proportion of patients in each treatment arm.

An ideal randomization procedure would achieve the following goals:^[33]

• Maximize statistical power, especially in subgroup analyses. Generally, equal group sizes maximize statistical power, however, unequal groups sizes maybe more powerful for some analyses (e.g., multiple comparisons of placebo versus several doses using Dunnett’s procedure^[34] ), and are sometimes desired for non-analytic reasons (e.g., patients maybe more motivated to enroll if there is a higher chance of getting the test treatment, or regulatory agencies may require a minimum number of patients exposed to treatment).^[35]
• Minimize selection bias. This may occur if investigators can consciously or unconsciously preferentially enroll patients between treatment arms. A good randomization procedure will be unpredictable so that investigators cannot guess the next subject's group assignment based on prior treatment assignments. The risk of selection bias is highest when previous treatment assignments are known (as in unblinded studies) or can be guessed (perhaps if a drug has distinctive side effects).
• Minimize allocation bias (or confounding). This may occur when covariates that affect the outcome are not equally distributed between treatment groups, and the treatment effect is confounded with the effect of the covariates (i.e., an "accidental bias"^[32][36]). If the randomization procedure causes an imbalance in covariates related to the outcome across groups, estimates of effect may be biased if not adjusted for the covariates (which may be unmeasured and therefore impossible to adjust for).

However, no single randomization procedure meets those goals in every circumstance, so researchers must select a procedure for a given study based on its advantages and disadvantages.

Simple randomization[edit]

This is a commonly used and intuitive procedure, similar to "repeated fair coin-tossing."^[32] Also known as "complete" or "unrestricted" randomization, it is robust against both selection and accidental biases. However, its main drawback is the possibility of imbalanced group sizes in small RCTs. It is therefore recommended only for RCTs with over 200 subjects.^[37]

Restricted randomization[edit]

To balance group sizes in smaller RCTs, some form of "restricted" randomization is recommended.^[37] The major types of restricted randomization used in RCTs are:

• Permuted-block randomization or blocked randomization: a "block size" and "allocation ratio" (number of subjects in one group versus the other group) are specified, and subjects are allocated randomly within each block.^[32] For example, a block size of 6 and an allocation ratio of 2:1 would lead to random assignment of 4 subjects to one group and 2 to the other. This type of randomization can be combined with "stratified randomization", for example by center in a multicenter trial, to "ensure good balance of participant characteristics in each group."^[2] A special case of permuted-block randomization is random allocation, in which the entire sample is treated as one block.^[32] The major disadvantage of permuted-block randomization is that even if the block sizes are large and randomly varied, the procedure can lead to selection bias.^[33] Another disadvantage is that "proper" analysis of data from permuted-block-randomized RCTs requires stratification by blocks.^[37]
• Adaptive biased-coin randomization methods (of which urn randomization is the most widely known type): In these relatively uncommon methods, the probability of being assigned to a group decreases if the group is overrepresented and increases if the group is underrepresented.^[32] The methods are thought to be less affected by selection bias than permuted-block randomization.^[37]

Adaptive[edit]

At least two types of "adaptive" randomization procedures have been used in RCTs, but much less frequently than simple or restricted randomization:

• Covariate-adaptive randomization, of which one type is minimization: The probability of being assigned to a group varies in order to minimize "covariate imbalance."^[37] Minimization is reported to have "supporters and detractors";^[32] because only the first subject's group assignment is truly chosen at random, the method does not necessarily eliminate bias on unknown factors.^[2]
• Response-adaptive randomization, also known as outcome-adaptive randomization: The probability of being assigned to a group increases if the responses of the prior patients in the group were favorable.^[37] Although arguments have been made that this approach is more ethical than other types of randomization when the probability that a treatment is effective or ineffective increases during the course of an RCT, ethicists have not yet studied the approach in detail.^[38]

Allocation concealment[edit]

"Allocation concealment" (defined as "the procedure for protecting the randomisation process so that the treatment to be allocated is not known before the patient is entered into the study") is important in RCTs.^[39] In practice, in taking care of individual patients, clinical investigators in RCTs often find it difficult to maintain impartiality. Stories abound of investigators holding up sealed envelopes to lights or ransacking offices to determine group assignments in order to dictate the assignment of their next patient.^[32] Such practices introduce selection bias and confounders (both of which should be minimized by randomization), thereby possibly distorting the results of the study.^[32] Adequate allocation concealment should defeat patients and investigators from discovering treatment allocation once a study is underway and after the study has concluded. Treatment related side-effects or adverse events may be specific enough to reveal allocation to investigators or patients thereby introducing bias or influencing any subjective parameters collected by investigators or requested from subjects.

Some standard methods of ensuring allocation concealment include sequentially numbered, opaque, sealed envelopes (SNOSE); sequentially numbered containers; pharmacy controlled randomization; and central randomization.^[32] It is recommended that allocation concealment methods be included in an RCT's protocol, and that the allocation concealment methods should be reported in detail in a publication of an RCT's results; however, 2005 study determined that most RCTs have unclear allocation concealment in their protocols, in their publications, or both.^[40] On the other hand, a 2008 study of 146 meta-analyses concluded that the results of RCTs with inadequate or unclear allocation concealment tended to be biased toward beneficial effects only if the RCTs' outcomes were subjective as opposed to objective.

The types of statistical methods used in RCTs depend on the characteristics of the data and include:

• For dichotomous (binary) outcome data, logistic regression (e.g., to predict sustained virological response after receipt of peginterferon alfa-2a for hepatitis C^[48]) and other methods can be used.
• For continuous outcome data, analysis of covariance (e.g., for changes in blood lipid levels after receipt of atorvastatin after acute coronary syndrome^[49]) tests the effects of predictor variables.
• For time-to-event outcome data that may be censored, survival analysis (e.g., Kaplan–Meier estimators and Cox proportional hazards models for time to coronary heart disease after receipt ofhormone replacement therapy in menopause^[50]) is appropriate.

Regardless of the statistical methods used, important considerations in the analysis of RCT data include:

• Whether a RCT should be stopped early due to interim results. For example, RCTs may be stopped early if an intervention produces "larger than expected benefit or harm," or if "investigators find evidence of no important difference between experimental and control interventions."^[2]
• The extent to which the groups can be analyzed exactly as they existed upon randomization (i.e., whether a so-called "intention-to-treat analysis" is used). A "pure" intention-to-treat analysis is "possible only when complete outcome data are available" for all randomized subjects;^[51] when some outcome data are missing, options include analyzing only cases with known outcomes and using imputed data.^[2] Nevertheless, the more that analyses can include all participants in the groups to which they were randomized, the less bias that an RCT will be subject to.^[2]
• Whether subgroup analysis should be performed. These are "often discouraged" because multiple comparisons may produce false positive findings that cannot be confirmed by other studies