In total, there were 9013 patients with PsA initiating tofacitinib in the study period from Optum CDM and 11,754 patients from MarketScan (see Table S1 in the electronic supplementary material). Following exclusion per the eligibility criteria, 274 patients from Optum CDM and 395 patients from MarketScan were included in this analysis. Patient baseline demographics and clinical characteristics can be found in Table 1.

In Optum CDM, the mean age was 54.4 years (standard deviation [SD] 12.4), the mean Charlson Comorbidity Index (CCI) score was 0.8 (SD 1.4), and the median follow-up period was 25.4 months (IQR 15.6). In MarketScan, the mean age was 51.4 years (SD 9.7), the mean CCI score was 0.5 (SD 1.0), and the median follow-up period was 15.6 months (IQR 11.0). The patient population was mainly female in both databases (Optum CDM, 70.4%; MarketScan, 68.9%). The insurance type was predominantly commercial in both databases (Optum CDM, 77.4%; MarketScan, 94.9%). The majority of patients were from the Southern region in both databases (Optum CDM, 54.0%; MarketScan, 45.8%). Approximately 96% of patients across both databases had at least one PsA-related comorbidity; the most common comorbidity was respiratory diseases followed by hypertension. In total, 175 (63.9%) patients in Optum CDM and 224 (56.7%) in MarketScan had concomitant psoriasis.

Within the first 90 days post-index, most patients in both database populations were receiving monotherapy vs combination therapy (Optum CDM, 74.1% vs 25.9%; MarketScan, 65.8% vs 34.2%). There were fewer female patients in the monotherapy vs combination therapy cohorts (67.0% vs 80.3%) in Optum CDM. Compared with the monotherapy cohorts, more patients in the combination therapy cohorts had Medicare insurance (monotherapy vs combination therapy: Optum CDM, 19.7% vs 31.0%; MarketScan, 4.6% vs 5.9%) and had a longer follow-up period (monotherapy vs combination therapy median [IQR] Optum CDM, 24.4 [16.6] vs 28.7 [14.8] months; MarketScan, 14.3 [10.6] vs 16.6 [12.2] months).

Fewer patients receiving monotherapy switched to combination therapy (addition of csDMARD) vs switching from combination therapy to monotherapy (removal of csDMARDs) at any time during the analysis in both database populations (see Table S2 in the electronic supplementary material).

In both database populations, the mean MPR was similar for both the monotherapy and combination therapy cohorts and was maintained over time (see Fig. S1a in the electronic supplementary material). At 6 months post-index, the proportion of patients with an MPR ≥ 0.8 was numerically lower in the monotherapy vs combination therapy cohorts in both database populations (Optum CDM, 70.0% vs 73.2%; MarketScan, 66.2% vs 75.6%) (Fig. 1a). This numerically lower pattern was maintained at 24 months post-index (Optum CDM, 62.2% vs 78.7%; MarketScan, 71.4% vs 73.1%). In comparison with the mean MPR, the mean PDC numerically decreased over time in both database populations for monotherapy and combination therapy (see Fig. S1b in the electronic supplementary material). At 6 months post-index, the proportion of patients with a PDC ≥ 0.8 was numerically lower in the monotherapy vs combination therapy cohort in both database populations (Optum CDM, 62.6% vs 67.6%; MarketScan, 58.1% vs 64.4%) (Fig. 1b). At 24 months post-index, this numerically lower pattern was maintained in the Optum CDM population but not in the MarketScan population (Optum CDM, 31.5% vs 44.7%; MarketScan, 31.4% vs 30.8%).

Adherence in patients with PsA initiating tofacitinib evaluated by a MPR ≥ 0.8 and b PDC ≥ 0.8. Optum CDM and MarketScan denote the Optum® Clinformatics® Data Mart and Merative™ MarketScan® databases, respectively. MPR and PDC were capped at 1.0 with a maximum number of 14 days of carryover allowed. N number of evaluable patients, n number of patients with adherence, MPR medication possession ratio, PDC proportion of days covered, PsA psoriatic arthritis

The rates of tofacitinib persistence decreased over time in both the monotherapy and combination therapy cohorts across both databases (6 months post-index, range across both databases 53.1–70.4%; 30 months post-index, range across both databases 20.0–34.0%; Fig. 2a). The overall median time to tofacitinib discontinuation was 10.6 (95% confidence interval [CI] 8.3–13.1) months in the Optum CDM population and 7.7 (95% CI 6.1–9.9) months in the MarketScan population. Lower treatment persistence was observed for the monotherapy vs combination therapy cohorts in both databases (Fig. 2b, c); however, statistical significance (at the 0.05 level of significance) was not found (median [95% CI] Optum CDM, 10.1 [7.4–11.8] vs 16.7 [8.3–26.6] months, log-rank P = 0.0663; MarketScan, 6.9 [5.6–9.4] vs 11.0 [6.1–13.9] months, log-rank P = 0.1152).

a Persistencea in the Optum CDM and MarketScan populations, and time to treatment discontinuation in patients with PsA initiating tofacitinib in the b Optum CDM and c MarketScan populations. Optum CDM and MarketScan denote the Optum® Clinformatics® Data Mart and Merative™ MarketScan® databases, respectively. Shading represents 95% Hall-Wellner confidence bands. aPersistence was estimated in terms of months of continuous medication use. Patients were considered non-persistent if a gap of ≥ 60 days was observed between the prescriptions. N number of evaluable patients, PsA psoriatic arthritis

Within the study period, the all-cause total medical costs for the overall patient populations were $5829 PPPM in Optum CDM and $4805 PPPM in MarketScan (Fig. 3a) and PsA-related total medical costs were $3954 PPPM and $3445 PPPM, respectively (Fig. 3b).

Total medical costs in patients with PsA initiating tofacitinib for a all-cause events stratified by treatment episode, b PsA-related events stratified by treatment episode, c all-cause events stratified by treatment type and episode, and d PsA-related events stratified by treatment type and episode. Optum CDM and MarketScan denote the Optum® Clinformatics® Data Mart and Merative™ MarketScan® databases, respectively. Total cost calculated as the sum of hospital admission, outpatient, and emergency room visit costs, and outpatient pharmacy cost. All costs are presented as PPPM and are inflation-adjusted to 2020 USD. Error bars indicate 95% CI. CI confidence interval, PsA psoriatic arthritis, PPPM per patient per month, USD US dollars

The overall all-cause total medical costs were numerically similar for patients receiving monotherapy and combination therapy in Optum CDM database ($5819 and $5855 PPPM, respectively), but numerically lower for the monotherapy vs combination therapy cohorts ($4539 vs $5270 PPPM) in MarketScan (Fig. 3c).

In comparison, overall PsA-related total medical costs were numerically lower for the monotherapy vs combination therapy cohorts in both Optum CDM ($3869 vs $4173 PPPM) and MarketScan ($3245 vs $3795 PPPM) (Fig. 3d). During off-treatment vs on-treatment periods, numerical cost decreases were observed for all-cause costs (Optum CDM, $5383 vs $6149; MarketScan, $4145 vs $5180) and PsA-related costs (Optum CDM, $3237 vs $4515; MarketScan, $2703 vs $3907), including in patients receiving monotherapy or combination therapy, which was primarily driven by numerical decreases in pharmacy costs off-treatment (see Table S3 in the electronic supplementary material).

All-cause and PsA-related HCRU stratified by treatment episode can be found in Table 2 and stratified by treatment episode and therapy type in Table 3. The number of outpatient visits generally increased numerically during off-treatment vs on-treatment periods for all-cause (Optum CDM, 2.37 vs 2.05 visits PPPM; MarketScan, 2.15 vs 1.99 visits PPPM) and PsA-related (Optum CDM, 0.60 vs 0.46 visits PPPM; MarketScan, 0.47 vs 0.44 visits PPPM) events. Outpatient costs were also numerically increased during off-treatment vs on-treatment periods and can be found in Table S3 in the electronic supplementary material.

In general, all-cause and PsA-related outpatient visits were numerically lower in the monotherapy vs combination therapy cohorts and increased numerically for both therapy types in off-treatment vs on-treatment periods, with exceptions to these observations in the MarketScan PsA-related data (Table 3). The overall number of all-cause medications prescribed was numerically higher during off-treatment vs on-treatment periods in the Optum CDM population (4.56 vs 4.33 prescriptions PPPM), whereas a numerically lower number was observed during off-treatment vs on-treatment periods in the MarketScan population (3.73 vs 4.19 prescriptions PPPM) (Table 2). This increase in medications prescribed in the Optum CDM population was driven by numerical increases in off-treatment vs on-treatment opioid (0.32 vs 0.19 prescriptions PPPM), biologic (0.32 vs no prescriptions PPPM), and non-PsA-related medications (3.36 vs 2.80 prescriptions PPPM) (Table 2).

In comparison, the overall number of PsA-related medications prescribed was numerically lower during off-treatment vs on-treatment periods in both the Optum CDM (1.21 vs 1.53 prescriptions PPPM) and MarketScan populations (1.05 vs 1.48 prescriptions PPPM) (Table 2). The numbers of all-cause and PsA-related emergency room visits were relatively low regardless of therapy type and treatment episode (≤ 0.1 visits PPPM) (Table 3).