[Figure 2 here]
We identified 78 definitions of LTOT of which ten were used by more than 75% of studies (shown in Table 2). LTOT was most often operationalised as 90 days of cumulative or continuous use within a year (27 studies, 21%). The rationale for this definition is based on clinical judgment and empirical data from the TROUP study, [48]showing one-quarter of people using opioids for more than 90 days in a given year received opioids in the subsequent year for more than 180 days. An additional 16 studies (13%) used 90 days of opioid use or supply as a threshold to define LTOT within various follow-up periods.
Another definition based on empirical data and commonly used across studies is derived from the CONSORT study. [49]The definition combines the length of opioid use episodes (≥ 90 days) with cumulative days of supply (≥ 120 days) or the number of fills (≥ 10) and was based on an increased probability of patients receiving an additional prescription with 10 days of supply surpassing this threshold. We found 23 studies (18%) operationalised this definition according to the original study or with some adaptation.
The use of opioids for more than 180 days within a year (7 studies, 6%), or within another time frame (4 studies, 3%), was based on clinical judgment to distinguish LTOT from acute use of opioids and on the clinical relevance of extending the use beyond 90 days.[50, 51].
Studies that did not evaluate duration of opioid use based on the length of episodes or days of supply often used the number of prescriptions/fills and the distribution of fills within the follow-up period to define LTOT: nine studies (7%) used ≥ 3 fills as a threshold in various follow-up windows and six (5%) used ≥ 6 fills.
Finally, four studies (3%) applied an LTOT definition developed by Svendsen et al [52] that assumes people use opioids at least half of the days of the year (denominated wide definition) and is associated with a higher probability of receiving opioids in the following two years.