Combined bias suppression in single-arm therapy studies

For therapy evaluation studies, control groups are sometimes not feasible. In single-arm studies, various bias factors apart from the test therapy can affect clinical outcomes. The objective of this analysis was to improve the methods to minimize bias in single-arm studies.

We present a procedure for combined suppression of several bias factors, using two methods: sample restriction to patients unaffected by bias, and score adjustment. The procedure was used for a secondary analysis of disease score (doctors' global rating, 0-10) in a cohort of patients receiving anthroposophic therapies for chronic diseases. Four bias factors were suppressed stepwise: attrition bias (by replacing missing values with the baseline value carried forward), bias from natural recovery (by sample restriction to patients with disease duration of >/=12 months), regression to the mean due to symptom-driven self-selection (by replacing baseline scores with scores three months before enrolment) and bias from adjunctive therapies (by sample restriction to patients not using adjunctive therapies).

In the cohort analysed, these four bias factors could together explain a maximum of 37% of the 0- to 6-month improvement of disease score.

Combined bias suppression, using sample restriction and score adjustment, is a transparent procedure to minimize bias in single-arm therapy studies. Further applicability of the procedure should be tested in future studies.