Fig. 1.

Study Design. ^a The index date was the date of dementia onset for cases, while for matched controls it was a study visit date at similar ACT follow-up time as the case’s dementia onset (within +/− 1 year). The anchor date was defined as three years before the index date. Participants in the antihypertensive analysis were also required to have at least one ICD-9 claim for a hypertension diagnosis in the 24 months prior to the anchor date. Computerized pharmacy dispensing data were used to identify people with prevalent use of antihypertensives or a statin as of the anchor date. Group-based trajectory modeling was employed to identify distinct patterns of medication adherence for antihypertensive and statin PDC data over the 3 years between anchor and index date