LAUSR

Abstract Objective This study measures the relative preference for attributes of device-aided therapies (DATs) for advanced Parkinson’s Disease (PD) from the perspective of Japanese neurologists. Methods Attributes and levels were elicited based on literature and interviews with certified neurologists experienced with DATs. An online survey including a discrete choice experiment (DCE) was developed, pilot tested, and distributed through an online panel to neurologists treating advanced PD patients. Participants were asked to choose treatments among several choice sets of two hypothetical DATs described only by the attributes, or no DAT (continuing oral treatment). A conditional logit model using the Bayesian framework was developed to estimate the marginal utilities of attributes’ levels, and the relative utility of treatments available to Japanese advanced PD patients or being developed in Japan was assessed. Results The DCE survey completed by 308 neurologists showed that the attributes with the greatest influence on DAT selection were surgery requirement (relative importance of 28%), average increase in the duration of daily “on” time without dyskinesia which affects daily activities (15%), average change in cognitive function related to treatment introduction (15%), device management frequency (14%), average number of pills of oral PD medication after treatment introduction (13%), average influence of treatment on symptoms of depression (12%), and type of device (large/small) (3%). All attributes significantly influenced respondents’ choices, except for external device type. Experience with DATs did not influence the directions of preferences. Out of treatment profiles representing DATs, continuous subcutaneous infusion of levodopa-carbidopa had a higher preference score than levodopa-carbidopa intestinal gel infusion and deep brain stimulation. Conclusions Our findings suggest that Japanese neurologists would prefer a DAT without surgery requirement. Other factors related to efficacy, safety, and administration mode have a significant, but a smaller influence on prescription choices.