Table 2. Comparison of Outcomes for Medicare Beneficiaries and Clinical Trial Participantsa.
| Variable | IDEL for FL | IDEL+R for CLL | ||||
|---|---|---|---|---|---|---|
| Study 101-09 (n = 26) | Medicare (n = 305) | P Value | Study 312-0116 (n = 89) | Medicare (n = 294) | P Value | |
| Duration and dose reductions | ||||||
| Treatment duration, median (IQR), days | 160 (60-250) | 114 (60-240) | .38 | 473 (212-754) | 173 (93-329) | <.001 |
| Dose reduction, % | 38.5 | 16.1 | .02 | 32.6 | 18.0 | .003 |
| Starting dose | ||||||
| 150 mg Twice daily | 100.0 | 83.6 | .02 | 100.0 | 89.8 | <.001 |
| 100 mg Twice daily | 0 | 13.4 | 0 | 9.9 | ||
| 150 mg Once daily | 0 | 2.0 | 0 | 1.0 | ||
| 100 mg Once daily | 0 | 1.0 | 0 | 0.3 | ||
| Reasons for early censoring, by day 180 | ||||||
| Treatment discontinuation, % | 53.8 | 47.2 | .52 | 18.0 | 43.2 | <.001 |
| On-treatment mortality, % | 7.7 | 18.4 | .26 | 4.5 | 9.9 | .11 |
| Treatment discontinuation, HR (95% CI) | ||||||
| All patients, on day 180b | 0.86 (0.48-1.52) | 2.27 (1.58-3.26) | ||||
| Patients with Charlson score ≤3c | 0.76 (0.40-1.43) | 2.32 (1.55-3.47) | ||||
| Overall mortalityd | ||||||
| All patients | 1.39 (0.69-2.78) | 1.40 (0.93-2.11) | ||||
| Patients with Charlson score ≤3 | 1.08 (0.52-2.24) | 1.32 (0.76-2.31) | ||||
| On-treatment Infectionse | ||||||
| Serious infections | ||||||
| Rate per 100 PY | 67.1 | 78.7 | .69 | 81.6 | 80.1 | .90 |
| Overall, % | 30.8 | 39.7 | 64.0 | 48.3 | ||
| Charlson score ≤3, % | 30.4 | 29.9 | 55.9 | 37.3 | ||
| By day 180, % | 19.2 | 32.1 | 38.2 | 34.0 | ||
| Fatal infections | ||||||
| Rate per 100 PY | 18.6 | 27.6 | .54 | 9.8 | 18.4 | .04 |
| Overall, % | 11.5 | 15.1 | 14.6 | 13.3 | ||
| Charlson score ≤3, % | 13.0 | 9.5 | 8.8 | 10.6 | ||
| By day 180, % | 0.0 | 13.1 | 5.6 | 7.5 | ||
Abbreviations: CLL, chronic lymphocytic leukemia; FL, follicular lymphoma; IDEL, idelalisib; IDEL+R, idelalisib with rituximab; HR, hazard ratio; IQR, interquartile range; PY, person-years.
P values calculated using Wilcoxon test for continuous variables and χ2 test for categorical variables.
The HRs were estimated from multivariable Cox proportional hazard models adjusting for age (5-year increments), sex, US location, Charlson score (linear and quadratic terms), and time interaction with indicator of Medicare beneficiaries. Estimates of HR on day 180 were presented here because the HR for treatment discontinuation was not constant over follow-up time. Refer to eFigure 2 in the Supplement.
The HRs were estimated from multivariable Cox proportional hazard models adjusting for age (5-year increments), sex, US location, and Charlson score (1-point increments).
The HRs were estimated from multivariable Cox proportional hazard models adjusting for age (5-year increments) and Charlson score (1-point increments).
See eTable 2 in the Supplement for rates of serious and fatal infections by anatomical location and eTable 3 in the Supplement for counts of serious and fatal infections by infection type.