统计问题第136问：区组随机化

Question

Researchers evaluated the effectiveness of a newly developed integrated care programme to help patients with chronic low back pain return to work. The programme was a combined patient and workplace directed intervention, delivered in the outpatient setting. A randomised controlled study design was used, with usual care being the control treatment. The study period was one year.

Trial participants were recruited from primary and secondary care if aged between 18 and 65 years and they had had low back pain for more than 12 weeks. In total 134 participants were allocated to the integrated care programme or to control treatment by block randomisation, with a block size of four. The researchers reported that the duration of time until a fully sustained return to work was significantly shorter for patients receiving integrated care than for those undergoing usual care.

Which of the following statements, if any, are true?

·a) Block randomisation ensured similar numbers of patients in the treatment groups.

·b) Participants were randomly allocated four at a time, with all four to the intervention group or all to the control group.

·c) Each participant had an equal probability of being allocated to the intervention or control group.

提示：正确答案不止一个。

Answer

Answers a and c are true, while b is false.

Participants are typically allocated to treatment groups in clinical trials through simple random allocation, often referred to as random allocation or randomisation. There is therefore an equal probability of allocation to any one treatment group. The purpose of random allocation is to achieve similarity of baseline characteristics in the treatment groups. If the treatment groups differ in baseline characteristics, confounding may result. Confounding factors are those that influence treatment and outcome measures and include demographic characteristics, prognostic factors, and other characteristics that may influence someone’s likelihood of participating in or withdrawing from a trial. Therefore, any differences between treatment groups in outcomes may not be due to differences in the treatment received but to differences in baseline characteristics.

A balance in numbers of participants in treatment groups is essential to achieve similarity in baseline characteristics. Although random allocation does not guarantee equal numbers in each treatment group, it is unlikely that the numbers will differ greatly if the total sample size is large. In trials with small samples it is doubtful that there will be an even distribution among groups. Block random allocation, also known simply as block randomisation, may be used in such circumstances to ensure similar numbers of participants in the treatment groups (answer a is true). A total of 134 participants were recruited in the above trial. Although not a multiple of four, the 134 participants were more evenly distributed between the treatment groups than they might have been if simple random allocation had been used.

Block randomisation with blocks of size four was used in the trial to allocate patients to treatment group. This involved selecting groups of four consecutive patients attending primary and secondary care. Within each group of four patients, two would have been allocated to the integrated care programme and two to the control group (answer b is false). However, the order in which treatments were allocated in each block was random. For blocks of size four there are six different ways in which the interventions could have been allocated. If the integrated care programme is denoted by A and the control treatment by B, the six possible permutations of allocation are AABB, ABAB, ABBA, BABA, BAAB, and BBAA. One of these permutations would have been selected at random.

Block randomisation is an example of restricted random allocation—a term to describe a method that controls the procedure of random allocation to achieve greater equivalence between treatment groups in group sizes and baseline characteristics. Despite it being a restricted process of allocation, each patient still has the same probability of being allocated to intervention or control (c is true). Once the allocation sequence for each group of four consecutive patients has been selected, the treatment that each patient receives is predetermined. However, because the allocation sequence is chosen at random from all possible permutations, each patient in a group of four still has an equal probability of being allocated to intervention or control.

Block randomisation ensures that consecutive patients are distributed equally between treatment groups. This would be important if systematic differences existed between patients as they presented and were recruited into the trial. It would also have been important if the skills of those healthcare professionals delivering the treatments changed with time, meaning that the overall outcome, regardless of treatment, might have changed as the trial progressed.

The size of the block in block randomisation can remain fixed or can change during allocation. For example, the block size may be randomly selected from any size that is a multiple of the number of treatment groups. The advantage of varying block sizes is that it preserves the unpredictability of allocation. This is especially important if the trial is not double blinded. Generally, larger block sizes, for example of 20 or more, will help ensure that the allocation process cannot be anticipated.

所以答案是选择 a c

每天学习一点，你会更强大！