{"id":7015,"date":"2026-04-24T06:48:02","date_gmt":"2026-04-24T04:48:02","guid":{"rendered":"https:\/\/mediconomics.com\/?post_type=glossary&p=7015"},"modified":"2026-07-13T19:10:01","modified_gmt":"2026-07-13T17:10:01","slug":"crossover-study","status":"publish","type":"glossary","link":"https:\/\/mediconomics.com\/en\/glossar\/crossover-study\/","title":{"rendered":"Crossover Study"},"content":{"rendered":"

Definition and basic principle<\/h2>\n

A crossover study is a study design in which each participating individual receives several treatments or interventions over the course of the trial. Participants switch, after a defined treatment period (period 1), into a further treatment period (period 2, etc.), so that each individual serves as their own control. The sequence is often randomised (e.g. AB vs. BA), and the treatment periods are separated by a washout phase to minimise carryover effects.<\/p>\n

The central advantage: differences between individuals (e.g. disease severity, metabolism, concomitant medication) have a weaker influence on the comparison, because the analysis is primarily carried out within the same individual. This can increase statistical power and reduce the required sample size. Crossover designs are used above all when the condition is stable and the intervention has a reversible, time-limited effect.<\/p>\n

Typical use cases (medicinal products and medical devices)<\/h2>\n

In medicinal product development, crossover studies are particularly common in clinical pharmacology, for example for bioavailability and bioequivalence investigations, food-effect studies, or interaction studies. In such questions, the effect is measurable in the short term (e.g. pharmacokinetic parameters), and the treatment can fully “wash out” after the end of the period.<\/p>\n

For medical devices and digital health applications, crossover designs are used more selectively, for example for short-term-acting, non-permanent interventions (such as certain training or therapy apps). Potential learning effects and the fact that some product effects are not readily reversible are limiting factors here.<\/p>\n

Design elements and statistical particularities<\/h2>\n

A classic 2\u00d72 crossover comprises two treatments (A and B) and two periods. Participants are randomised into sequences (AB or BA). Decisive factors are the determination of the period length, the measurement timepoints, and a sufficiently long washout phase. The washout phase is intended to ensure that residual effects of the preceding treatment do not distort the effect in the following period.<\/p>\n

In the analysis, period, sequence and, where applicable, baseline values are frequently considered in addition to the treatment effect. Mixed models or ANOVA approaches are common, depending on the type of endpoint. A key issue is the testing and handling of carryover effects: if a treatment carries over into the next period, the treatment effect can be biased. Further risks include dropouts (missing data in later periods), insufficient compliance, or changes in concomitant therapies that can accumulate across multiple periods.<\/p>\n

Regulatory and operational requirements (EU\/Germany focus)<\/h2>\n

From a regulatory perspective, the same principles apply to crossover studies as to other interventional trials: scientific justification of the design, risk-adequate quality management, and protection of participants. In the EU, the requirements for clinical trials with medicinal products are anchored in Regulation (EU) No 536\/2014; in day-to-day operations, the principles of Good Clinical Practice are also decisive (e.g. ICH E6, including current revisions).<\/p>\n

For bioequivalence and pharmacokinetic studies, relevant guidelines apply additionally (e.g. EMA guidance on bioequivalence), in particular regarding the design of washout phases, the choice of endpoints, and statistical methodology. From the sponsor’s and CRO’s perspective, randomisation and blinding processes, documentation in the Trial Master File, and traceable data integrity measures should be set up so that inspections can clearly recognise the design rationale and the control of bias risks.<\/p>\n

It is also practically important to align with the clinical study protocol: the visit schedule, laboratory measurements, sample logistics, and the handling of concomitant medication must be planned so that periods remain comparable. Clear handling of protocol deviations (e.g. delayed visits) and consistent data review are essential so that the within-person comparison is not diluted by operational imprecision. A CRO can support this through standardised checklists, risk-based monitoring, and clean documentation pathways within the quality management system.<\/p>\n

Advantages, limitations and common misconceptions<\/h2>\n

A widespread misconception is that crossover designs are “always” more efficient. This is only true if the assumptions are met: stable disease course, reversible effect, sufficiently long washout phase, and low probability of carryover. If any of these prerequisites is not met, a parallel-group design may be methodologically more robust.<\/p>\n

A crossover design is generally unsuitable when the disease is unstable (e.g. acute flares), when the intervention has long-term or irreversible effects (e.g. curative therapies, implants), or when strong learning effects are to be expected. Long trial durations per person can also be problematic, because dropouts and external influences increase. In study planning, it is therefore worthwhile to conduct an early feasibility assessment that systematically evaluates clinical stability, expected duration of effect, recruitability and operational feasibility.<\/p>\n

FAQ<\/h2>\n

Why can crossover studies manage with fewer participants?<\/h3>\n

Because each individual receives several treatments and thus serves as their own control, inter-individual variability is lower. This often increases statistical efficiency, provided that carryover and period effects are well controlled.<\/p>\n

What is a washout phase and how is it determined?<\/h3>\n

The washout phase is a treatment-free period between two periods intended to minimise residual effects. Its duration is typically derived from half-life, duration of effect, and clinical stability assumptions, and is justified in the study protocol.<\/p>\n

What is the main risk with crossover designs?<\/h3>\n

A central risk is carryover effects: if a treatment carries over into the next period, the comparison can be biased. In addition, dropouts in later periods can complicate the analysis.<\/p>\n

Regulatory references (selection):<\/strong><\/p>\n