OBJECTIVES : The Treatment Satisfaction Questionnaire for Medication (TSQM) is the most commonly used generic Patient-Reported Outcome (PRO) measuring treatment satisfaction. Its psychometric properties have been assessed in numerous intervention trials using Classical Test Theory (CTT). Given the heterogeneous treatment experience of people taking different medicines in routine health care, the psychometric properties of the TSQM may be challenged in phase 4 and patient preference research. To determine the extent to which the psychometric properties of the TSQM are supported across various conditions in a real world context.

METHODS : Electronic versions of both the TSQM v1.4 and vII were administered to subjects of an online patient community managed by IQVIA (MediGuard) in the US and in the UK. CTT and Rasch measurement theory were used to evaluate the cross-sectional psychometric properties of the TSQM.

RESULTS : The survey was completed by 108 and 354 patients in the UK and in the US, respectively. Most patients who responded had university-level education (67%), were female (68%) and had a mean of 50 (±18.4) years of age. CTT analyses confirmed that the TSQM had excellent internal consistency and construct validity. High floor effects were seen for several items of the TSQM in some disease-level subgroups of patients (e.g. hypothyroidism). Rasch measurement theory (RMT) analyses showed that items were consistently ordered on an interval scale from low to high levels of treatment satisfaction. The Differential Item Functioning and Person-Item Threshold distribution suggests a potential alternate scoring structure that may be more generalizable across therapeutic areas in routine health care.

CONCLUSIONS : The TSQM demonstrates generally high psychometric evidence for its use in a heterogeneous patient population in routine health care. However, the TSQM scoring could be revisited to improve its fit-for-purpose applicability.