Patient characteristics, pain treatment patterns, and incidence of total joint replacement in a US population with osteoarthritis

Table 3 Incidence rates of TJR events per 100 person-years

Cohort	Number of episodes	Episodes with TJR event	Incidence rate ^a (95% confidence interval)
“Any pain medication”, with up to 7 years of follow-up ^b	1 992 670	183 093	3.21 (3.20–3.23)
“Each specific pain medication class”
NSAID ^c	1 947 237	36 369	4.63 (4.58–4.67)
Opioid ^c	2 726 480	89 748	7.45 (7.40–7.49)
Intra-articular corticosteroid ^c	1 320 838	37 980	8.05 (7.97–8.13)

NSAID non-steroidal anti-inflammatory drug
TJR total joint replacement
^aIncidence rate was calculated as episodes with a TJR event divided by person-years at risk times 100
^bFor prevalent users of any prescription pain medication, a follow-up time of up to 7 years after index date was assigned for evaluation of TJR occurrence. Only the first qualifying dispensing (index) for each health plan member was included; cohort re‐entry was not allowed. Person‐years at risk was censored at occurrence of death, disenrollment, query end date, network partner end date, and/or TJR occurrence
^cFor prevalent users of each specific drug class, all qualifying dispensings were identified. TJR was assessed during the treatment duration with a 30-day extension to the end of the treatment episode for the NSAID and opioid subgroups, and using a pre-specified 90-day treatment duration for the intra-articular corticosteroid subgroup with a 30-day extension. Person‐years at risk was censored at evidence of death, disenrollment, query end date, network partner end date, TJR occurrence, discontinuation of treatment, and/or initiation of another class of pain medication (e.g., treatment episodes for NSAID users were censored at initiation of opioid or intra‐articular corticosteroid use)

ISSN: 1471-2474