OBJECTIVES: Establish an MCID for the TDIS using data from two 52-week, phase 3 trials of valbenazine in US adults with tardive dyskinesia (TD): placebo-controlled KINECT3 (NCT02274558) and open-label KINECT4 (NCT02405091)

METHODS: Both anchor- and distribution-based approaches were used to estimate an MCID for TDIS, a reliable and valid 11-item questionnaire that evaluates how TD symptoms affect daily functioning. TDIS total score ranges from 0–44; higher scores represent greater impairment. Single-item Clinical Global Impression of Change (CGI-TD) and Patient Global Impression of Change (PGIC), with 7 responses, varying from very much improved (1) to no change (4) to very much worsened (7), were used as anchors. TDIS score changes from baseline to Week 6 in KINECT3 and to Week 8 in KINECT4 were computed from patients who received 40 mg or 80 mg valbenazine or placebo and had CGI-TD and PGIC minimal responses (score=3, representing minimally improved). The standardized response mean (SRM) and the standard error of measurement (SEM) were used in the distribution-based approaches.

RESULTS: In KINECT3 (n=201), 41% (n=83) of patients had a PGIC score of 3 and 43% (87) had a CGI-TD score of 3 at Week 6, while 32% (47) had a PGIC score of 3 and 38% (56) had a CGI-TD score of 3 at Week 8 in KINECT4 (n=148). Change from baseline in TDIS total score in both trials ranged from 3–4 points in patients with CGI-TD and PGIC scores equal to 3. The SEM was −1.66 and −1.68 and the SRM was −0.58 and −0.74 for KINECT3 and KINECT4, respectively. Combining data from anchor- and distribution-based results yielded an MCID estimate of 3-4 points.

CONCLUSIONS: An MCID of 3-4 points based on anchor- and distribution-based analyses from 2 separate studies is useful for interpreting how much change in TDIS scores is considered clinically meaningful.