The negative effects of long-term prescription opioid use are well-documented, yet long-term use has increased dramatically in recent years. Clinical guidelines recommend low-severity injury patients not receive opioids early in the course of their care, but supporting evidence is limited. Previous studies have compared those receiving opioids early to those not receiving opioids early, a substantial portion of which never received opioids at all. The present study utilises a sizeable dataset from a large, national workers’ compensation insurer to make the more relevant comparison between those receiving opioids early and those receiving opioids late. Claims data from 123,096 claimants with low-severity back and shoulder injuries from 1999 to 2010 were analysed. Long-term opioid use (>3 months) was compared between those who received opioids early (≤1 month after injury) and late (>1 month). Back-injured claimants with early opioid use had 33% lower (95% CI: 24%–41% lower) odds of long-term opioid use than claimants with late opioid use, after adjusting for key covariates. Shoulder-injured claimants with early opioid use had 29% higher (95% CI: 6%–58% higher) odds of long-term opioid use than those with late opioid use. Stratified analyses indicated that, for both back and shoulder, among claimants with unclear diagnoses within the first month of injury, early opioid use was related to higher risk for long-term opioid use. Among patients with low-severity diagnoses, later opioid use does not appear to decrease risk of long-term use. However, there may still be a benefit to delaying opioid prescription if a diagnosis cannot be ascertained.