The Step Wedge Design for a Multi-Story Building

Numerous advantages have been claimed for the step wedge trial design. While the design does have some methodological advantages over the ‘standard’ parallel cluster RCT with baseline observations (to be off-set against some disadvantages), the most widely cited rationale for the step wedge is logistical / political / ethical.[1] The design mimics what might naturally happen in non-trial practice where interventions must be phased under logistic constraints and where all eligible clusters might reasonably expect to receive the intervention.

Nevertheless, the step wedge design has been criticised on the grounds that a (non-randomised) adjustment must be made for any temporal trend and, as always, such an adjustment entails certain assumptions. To get around this problem, a hybrid design has been proposed to harvest the advantages of both methods:

This method also yields great precision.[2] However, this design incorporates a control group that does not receive the intervention (at least within the study timeframe), and hence it could be said to vitiate the promised rationale for the step wedge method.

Now, of course, the purpose behind a trial is to obtain information to inform future practice so the control group cluster should (eventually) receive the intervention. Intervention effects though are labile (some may say notoriously so) and have a tendency to fade with time, perhaps because of secular trends.[3] The upshot is that there are often great advantages to prolonged follow up. Enter the ‘two story’ step wedge design:

Such a design accommodates future rollout and also a point (I) for reflection to ensure that further expansion is propitious. It likely has attractive statistical features that will be examined by CLAHRC WM. Moreover, the first story might be a lot lower than the second, as in a certain type of Georgian house. It will not have passed the sagacious News Blog reader that such a study integrates a pilot study into a definitive study, following a pause for reflection (landing in the staircase!) to ensure that continuation is worthwhile – it enables the trial funder to exercise future options.[4]

— Richard Lilford, CLAHRC WM Director

References

  1. Hemming K, Haines TP, ChiltonPJ, Girling AJ, Lilford RJ. The stepped wedge cluster randomised trial: rationale, design, analysis, and reporting. BMJ. 2015; 350: h391.
  2. Hemming K, Lilford R, Girling AJ. Stepped-wedge cluster randomised controlled trials: a generic framework including parallel and multiple-level designs. Stat Med. 2015; 34(2): 181-96.
  3. Chen YF, Hemming K, Stevens AJ, Lilford RJ. Secular trends and evaluation of complex interventions: the rising tide phenomenon. BMJ Qual Saf. 2015. [ePub].
  4. Girling A, Lilford R, Cole A, Young T. Headroom Approach to Device Development: Current and Future Directions. Int J Technol Assess Health Care. 2015; 31(5): 331-8.
Advertisements

2 thoughts on “The Step Wedge Design for a Multi-Story Building”

  1. This might also be a good process for post licensing “place in therapy” studies for new pharmaceuticals. And more radically initial licensing could be “provisional” until place in therapy has been determined in real world conditions

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s