Purpose: This study assessed the impact of adherence definitions and apriori assumptions on common persistence and adherence measures from utilization data on the infusible anti-inflammatory biologic infliximab.
Method: Patients were identified through electronic health records (EHR) with ICD-9 diagnoses of rheumatoid arthritis (714.xx), psoriatic arthritis (696.xx), ulcerative colitis (556.xx), or Crohn's disease (555.xx). Incident infliximab-treated patients between January 1, 2007 and June 30, 2011, ≥18 years of age, with ≥1 maintenance dose were included. Infliximab dose specifics were extracted from medical chart review and EHR. Kaplan-Meier (KM) analyses estimated median time to treatment discontinuation. Medication possession ratio (MPR) was calculated as the sum of prescribed infusion frequency intervals divided by days from first infusion to last infusion (+ day supply). Proportion of days covered (PDC) was calculated as sum of prescribed infusion frequency intervals divided by a fixed time interval of 12 months. Adherence was defined as MPR ≥0.80 and persistence as PDC ≥0.80. Sensitivity analyses were performed for various definitions of “treatment gap” and “discontinuation” for MPR and with or without treatment continuation for PDC.
Result: 122 patients met study inclusion criteria. Mean age of patients was 45 years and 61% were female. A total of 100 patients had a calculable MPR. Mean MPR, when allowing for different treatment discontinuation gaps of 0 days (unweighted MPR), ≥60 and ≥90 days (weighted MPR), were 0.93, 0.94 and 0.94, respectively, however fewer than 9% of patients had either 60 or 90 day gaps in treatment. Mean PDC was 0.71 and 0.92, depending on whether the denominator included 12 months of follow up or 12 months of treatment, respectively (Table). KM estimated a median treatment duration of 19 months (N=88) or 29 months (N=60) if the allowance for gaps in treatment were ≥60 or ≥90 days respectively.
Conclusion: Patients receiving infliximab were adherent and persistent with therapy overall. PDC and KM results varied depending on the underlying definitions used. Transparency of underlying definitions in reporting adherence or persistence results is recommended.
Table
Proportion of days covered (PDC) and persistence | |||||
| PDC through 12 months follow up |
| PDC through 12 months treatment | ||
| No censoring |
| Censor at 90day gap |
| No censoring |
| N=88 |
| N=88 |
| N=60 |
|
|
|
|
|
|
Mean PDC | 0.713 |
| 0.710 |
| 0.915 |
% Persistent (PDC≥0.80) | 59% |
| 59% |
| 88% |