Developing and validating measures of take-home methadone with administrative data

Take-home methadone (THM) flexibility has increased since 2020, representing innovation in opioid use disorder treatment. There are no established approaches to measuring THM using insurance claims data. We proposed and validated candidate measures of THM.

Using 2020 Medicaid data from 4 states, we constructed treatment episodes for enrollees aged 18-64. Episodes started after July 1, 2020 and lasted at least 60 days. We labelled individuals as receiving THM if they received ≥6 consecutive days of THM in their 2nd month of treatment, as defined by presence of claims with a modifier code indicating THM (the "gold-standard" indicator). We defined 4 candidate indicators of THM based on intervals between in-clinic methadone administrations. We assessed performance of each candidate indicator against the gold-standard. We assessed the extent to which between-program variation explained total variation in measured THM.

The study sample included 4836 episodes for 4801 individuals. THM was present in 14 % of episodes. Sensitivity of candidate indicators ranged from 65 to 100 %, with the most sensitive being an indicator that was true if any two adjacent in-clinic service dates had a gap of ≥7 days. Specificity ranged from 80 to 96 %, with the most specific measure being one requiring 2 consecutive intervals of ≥7 days that were of the same length. Between-program variation explained 38.6-48.3 % of variation in THM receipt.

Two indicators of THM using Medicaid data presented excellent performance when evaluated against a gold-standard indicator. Our approach can be used to assess uptake and outcomes of THM.