The prevalence of opioid use nationwide, coupled with the large share of new Social Security Disability Insurance (SSDI) awardees who have conditions associated with opioid use, suggests that opioid use may be common and increasing among SSDI applicants. But statistics on opioid use among these applicants, who often report medications to SSA in an open-ended text field, have been unavailable because of the format of the data. We filled this knowledge gap by using a machine-learning method to identify opioids in the medication text fields in SSDI administrative data. We found that from 2007 through 2017, over 30% of SSDI applicants reported using one or more opioids. Consistent with nationwide prescribing trends, the rate of opioid use rose until 2012 and then declined through 2017. This rate also varied by demographic and other beneficiary characteristics.