1. Cluster Randomised Trials

2. Background In most RCTs people are randomised as individuals to treatment. Whilst this method is appropriate for many interventions (e.g. drug trials), in some types of intervention individuals cannot be randomised. An alternative approach is randomise groups of individuals or ‘clusters’.

3. History Cluster trials originated from educational research. Intact classes or schools were randomised to an intervention or no intervention. Sadly educational researchers have all but abandoned RCTs in favour of qualitative research.

4. Rationale For Cluster Randomisation Some interventions have to be delivered at a group level. Guidelines for clinicians Interventions to reduce infectious diseases Practical considerations Potential for treatment contamination

5. Clusters A cluster can take many forms: GP practice or patients belonging to an individual practitioner; Hospital ward; A period of time (week; day; month); Geographical area (village; town; postal district).

6. Cluster allocation Because the unit of allocation is the cluster and the sample size of clusters tends to be small care needs to be taken with cluster allocation. With typically ‘only’ 10 or so clusters simple randomisation is likely to lead to chance imbalance.

7. Cluster allocation Need to use some form of stratification. Pairing is often used – match clusters on an important co-variate and randomly allocate a member of each pair to the intervention. Stratification using blocking or the use of minimisation is an alternative.

8. Problems with Cluster Randomisation Possible Selection Bias; Inadequate uptake of intervention by either group reduces study power; Sample size needs to be increased (typically between 50% to 100%), which will often increase the cost and complexity of a trial.

9. Selection Bias - A Reminder This is where individuals who are using a treatment have some difference, unrelated to the treatment, that affects outcome. For example, women using HRT take more exercise, are slimmer, have higher social class compared with those who do not - may explain cardiovascular benefit.

10. Randomisation Randomisation, or similar procedure, will balance known and unknown co- variates or confounders across the groups and therefore selection bias should not occur. Thus, in an HRT trial women in treatment and placebo groups will have the same weight, exercise levels etc.

11. Selection Bias in Randomised Trials This should not occur in an individually randomised trial unless the randomisation has been subverted. However, in cluster trials it is possible for selection bias to occur after successful cluster randomisation. This defeats the objective of randomisation.

12. Selection Bias in Cluster Trials Given enough clusters bias should not occur in cluster trials as randomisation will deal with this. HOWEVER, the clusters are balanced at the individual level ONLY if all eligible people, or a random sample, within the cluster are included in the trial.

13. Recruitment into cluster trials A key issue is individual participant recruitment into cluster trials. There are a number of ways where biased participant recruitment can occur, which can lead to baseline imbalances in important prognostic factors.

14. Participant flow in cluster trial: sources of bias

15. Identification Problems For example, in a cluster trial of backpain equal number of patients with same severity of back pain will be present in both clusters. The problem lies in how to identify such patients to include them in the interventions. Unless one is very careful different numbers and types of patient can be selected.

16. UK BEAM Trial The UKBEAM pilot study used a cluster design. Eligible patients were identified by GPs for trial inclusion. GP practices were randomised to usual care or extra training. The ‘primary care team’ were trained to deliver ‘active’ management of backpain.

17. UK BEAM Selection bias The pilot showed that practices allocated to ‘active management’ were more adept at identifying patients with low back pain and including them in the trial. Patients had different characteristics in one arm than the other.

18. UK BEAM participant recruitment P = 0.025 P = 0.001 P = 0.03

19. UKBEAM pilot study.

20. UK BEAM Because of the selection bias in the cluster design that element of the trial was abandoned and the trial reverted to completely individual allocation.

21. Cluster Trials: Rule 1 All eligible patients or a random sample ideally MUST be identified BEFORE clusters are randomised. Alternatively systems must be put into place to PREVENT selective recruitment.

22. Trial Consent Problems Even when it is possible to identify all eligible members of a cluster some may not consent to take part in the trial. If there is differential consent, in particular, this can lead to selection bias again.

23. Hip Protector Trial Kannus. N Eng J Med 2000;343:1506. At this point trial is balanced for all co-variates

24. First Rule Kannus trial DID identify all eligible patients at baseline, thus, fulfilling first rule of cluster randomisation.

25. Hip Protector Trial Selection Bias

26. Fracture risk: Important Co- variates Most important risk factors for hip fracture are (in order of importance): Being Female; Age; Body Weight

27. Important Co-variates Balanced at baseline?

28. Results of Trial. Hip fractures were reduced by 60% (95% CI 0.2 to 0.8) HOWEVER, arm fractures were also reduced by 30% (0.3 to 1.5). Suggesting that some or all of the hip fracture effect could have been due to selection bias.

29. Hip Protector Trial

30. Cluster Trials: Rule 2 As in individually randomised trials imperative to use intention to treat analysis.

31. Inadequate uptake of intervention Because a robust cluster trial consent to randomisation is not given only consent to treatment this results in a proportion of eligible participants declining the intervention BUT have to stay in the trial for intention to treat analysis and this reduces study power. This also leads to DILUTION BIAS.

32. Accident prevention In a cluster trial of accident prevention among young children 25% of parents in the experimental arm did not receive the intervention. Clearly this will reduce the power of that trial AND dilute any likely ‘treatment’ effect. Kendrick et al. BMJ 1999;318:980.

33. Cluster Trials: Rule 3 Increase sample size to compensate for less than 100% uptake of intervention. Or alternatively and in conjunction identify and consent before randomisation and then only use those participants who have expressed a willingness to take part in the trial.

34. Review of Cluster Trials Because of the ‘BEAM’ problem we decided to undertake a methodological review of cluster trials. We identified all cluster trials published in the BMJ, Lancet, NEJM since 1997. Puffer et al. BMJ 2003;327:785.

35. Results We identified 36 relevant trials. ONLY 13 had identified participants prior to randomisation. Of the 23 not identifying participants a priori 7 showed evidence of differential recruitment or consent. Other biases included differential of inclusion criteria or attrition. In total 14 (39%) showed evidence of bias.

36. Underestimate of problem Only in 5 papers did authors alert reader to possible problem. Subsequently one of the trials that ‘looked’ OK was published elsewhere where recruitment bias was admitted to have occurred. Cluster trials are DIFFICULT to undertake robustly. Is there an ALTERNATIVE?

37. Cluster Sample Size Usual sample size estimaes assume independence of observations. When people are members of the same cluster (e.g., classroom, GP surgery) they are more related than we would expect to be at random. This is the intra-cluster correlation co- efficient.

38. ICC The ICC needs to incorporated into the sample size calculations. The formula is as follows: Design effect = 1 + (m – 1) X ICC. Design effect is the size the sample needs to be inflated by. M is the number of people in the cluster.

39. Sample size example. Let’s assume for an individually randomised trial we need 128 people to detect 0.5 of an effect size with 80% power (2p = 0.05). Now assume we have 24 groups with 7 members. The ICC is 0.05, which is quite high. 1+ (7 – 1) x 0.05 = 1.3, we need to increase the sample size by 30%. Therefore, we will need 166 participants.

40. What happens if cluster gets bigger? If our cluster size is twice as big (14), things begin to get really interesting. 1+(14-1)x0.05 = 1.65. What about 30? (1+(30-1)x 0.05 = 2.45 (I.e, 314 participants).

41. What makes the ICC large? If the treatment is applied to health care provider (e.g., guidelines will increase ICCs for patients). If cluster relates to outcome variable (e.g., smoking cessation and schools) If members of cluster are expected to influence each other (e.g., households).

42. AuthorsSourceYears Clustering allowed for in sample size Clustering allowed for in analysis Donner et al. (1990) 16 non-therapeutic intervention trials 1979 – 1989 <20%<50% Simpson et al. (1995) 21 trials from American Journal of Public Health and Preventive Medicine 1990 – 1993 19%57% Isaakidis and Ioannidis (2003) 51 trials in Sub-Saharan Africa 1973 – 2001 (half post 1995) 20%37% Puffer et al. (2003) 36 trials in British Medical Journal, Lancet, and New England Journal of Medicine 1997 – 2002 56%92% Eldridge et al. (Clinical Trials 2004) 152 trials in primary health care 1997 - 2000 20%59% Reviews of Cluster Trials

43. Sample Size Problems Cluster Trials Demand Larger Sample Sizes

44. Summary of sample size The KEY thing is the size of the cluster. It is nearly always best to get lots of small clusters than a few large ones (e.g, a trial with small hospital wards, GP practices, classrooms will, ceteris paribus, be better than large clusters). BUT if the ICC is tiny may not affect the sample too much.

45. Analysis Many cluster randomised health care trials have been INCOMPETENTLY analysed. Most analyses use t-tests, chi-squared tests, which assumes independence of observations, which are violated in a cluster trial. This leads to spurious p values and narrow CIs. Various methods exist, e.g., multilevel models, comparing means of clusters, which will produce correct estimates.

46. Cluster Trials: Should I do one? If possible avoid like the plague. BUT although they are difficult to do, properly, they WILL give more robust answers than other methods, (e.g., observational data), when done properly. Is it possible to avoid doing them and do an individually randomised trial?

47. Contamination An important justification for their use is SUPPOSED ‘contamination’ between participants allocated to the intervention with people allocated to the control.

48. Spurious Contamination? Trial proposal to cluster randomise practices for a breast feeding study – new mothers might talk to each other! Trial for reducing cardiac risk factors patients again might talk to each other. Trial for removing allergens from homes of asthmatic children.

49. Contamination Contamination occurs when some of the control patients receive the novel intervention. It is a problem because it reduces the effect size, which increases the risk of a Type II error (concluding there is no effect when there actually is).

50. Patient level contamination In a trial of counselling adults to reduce their risk of cardiovascular disease general practices were randomised to avoid contamination of control participants by intervention patients. Steptoe. BMJ 1999;319:943.

51. Accepting Contamination We should accept some contamination and deal with it through individual randomisation and by boosting the sample size rather than going for cluster randomisation Torgerson BMJ 2001;322:355.

52. Counselling Trial Steptoe et al, wanted to detect a 9% reduction in smoking prevalence with a health promotion intervention. They needed 2000 participants (rather than 1282) because of clustering. If they had randomised 2000 individuals this would have been able to detect a 7% reduction allowing for a 20% CONTAMINATION. Steptoe. BMJ 1999;319:943.

53. Comparison of Sample Sizes NB: Assuming an ICC of 0.02.

54. Misplaced contamination The ONLY study, I’m aware of to date, to directly compare an individually randomised study with a cluster design, showed no evidence of contamination. In an RCT of nurse led cardiovascular risk factor screening some ‘intervention’ clusters had participants allocated to no treatment. NO contamination was observed.

55. Cluster Trials Can cluster trials give different results? All things being equal this shouldn’t happen (except for a more imprecise estimate). BUT because of the greater potential for selection bias cluster trials MAY give the ‘wrong’ answer.

56. An example. There are 14 RCTs of hip protectors for the prevention of hip fracture. Nine RCTs are individually randomised trials, whilst 5 are cluster trials (e.g., hospital ward, nursing home). Cluster trials, without exception show a benefit of hip protectors.

57. Hip Protector Trials Individual RCTSCluster RCTs 1.19 (0.8 to 1.7)0.34 0.94 (0.5 to 1.7)0.53* 0.93 (0.5 to 1.7)0.44 1.17 (0.4 to 3.0)0.34 0.39 (0.1 to 1.4)0.11 0.20 (0.0 to 1.6)All Cluster trials, bar *, significant, No individual trial was significant 1.49 (0.3 to 7.1) 3.03 (0.6 to 14.8)

58. Hip Protector Trials: Cluster vs Individually Randomised.

59. Age differences between ‘good’ cluster and poor cluster trials. Data from Puffer et al.

60. Guidelines Royal College of Physician guidelines for fracture prevention grade hip protectors as grade A evidence based on flawed, cluster trial, evidence.

61. Cluster Trials- What Should We Do? Identify ALL eligible people if possible BEFORE randomisation ALWAYS use Intention To Treat analysis INCREASE sample size not only for cluster effects but also because of treatment refusal

62. Cluster designs

65. Summary Cluster Trials are currently very trendy Whilst in principle they are a robust. design in practice fraught with difficulty. If possible avoid and opt for individual randomisation If cluster trial is necessary follow rules to avoid bias.