We examined the association between rapid opioid reduction or discontinuation and self-harm, suicide attempt, and suicide death among high-dose long-term opioid therapy (HD-LTOT) patient and examined effect measure modification by individual and neighborhood-level characteristics. Using private insurance data from North Carolina, this retrospective cohort study covered January 2006 to September 2018, with up to four years of follow-up. Participants included patients aged 18-64 years who were prescribed HD-LTOT. Time-varying exposure was ever exposed to rapid opioid reduction or discontinuation vs never exposed. The outcomes were self-harm or suicide attempt, suicide death, and the combined outcome. We estimated cumulative incidence and used Fine-Gray models to estimate sub-distribution hazard ratios (HRs). There were 21,450 HD-LTOT patients. In year 1, rapid opioid reduction or discontinuation was not associated with the combined outcome, HR: 1.09 (95% CI: 0.61-1.96). However, in years 2-4, rapid opioid reduction or discontinuation was associated with higher hazard of the combined outcome, HR: 2.77 (95% CI: 1.45-5.27). This association was stronger among patients with mental health conditions and those residing in underserved neighborhoods. These findings underscore the importance of provider training in adhering to guideline-concordant gradual tapering, offering mental health support, and ensuring patient safety throughout the tapering process.