Multivariable model for the receipt of first-line immunochemotherapy (rather than chemotherapy alone) for PCNSL
| Variable . | Relative risk . | 95% CI . | P . |
|---|---|---|---|
| Age, per 10 y | 0.98 | (0.97-1.00) | .08 |
| Sex, female | Ref. | .010 | |
| Male | 0.95 | (0.91-0.99) | |
| Race/ethnicity, WNH | Ref. | .0001 | |
| Hispanic (any race) | 0.87 | (0.79-0.96) | |
| Black | 0.87 | (0.78-0.98) | |
| Asian/other | 0.96 | (0.88-1.04) | |
| Unknown | 0.66 | (0.46-0.93) | |
| HIV status, negative | Ref. | .0076 | |
| Positive | 0.79 | (0.67-0.94) | |
| Charlson-Deyo score, 0 | Ref. | .21 | |
| 1 | 1.01 | (0.96-1.07) | |
| 2 | 1.08 | (1.01-1.15) | |
| ≥3 | 1.02 | (0.94-1.11) | |
| Primary location, brain | Ref. | .10 | |
| Spinal cord | 1.01 | (0.90-1.15) | |
| Eye | 1.16 | (0.85-1.58) | |
| Meninges | 1.17 | (0.91-1.49) | |
| Other/unspecified | 1.05 | (0.99-1.12) | |
| Prior malignancy, none | Ref. | .19 | |
| Yes | 1.03 | (0.98-1.09) | |
| Chemotherapy, single agent | Ref. | .019 | |
| Multiple agents | 1.07 | (0.99-1.14) | |
| Unspecified | 0.78 | (0.59-1.03) | |
| Insurance, private | Ref. | .036 | |
| Medicaid | 0.94 | (0.86-1.04) | |
| Medicare | 0.94 | (0.87-1.01) | |
| Other government | 1.12 | (1.00-1.26) | |
| Not insured | 1.06 | (0.92-1.24) | |
| Unknown | 1.09 | (0.92-1.29) | |
| Hospital, academic | Ref. | .43 | |
| Community | 0.97 | (0.89-1.05) | |
| 0-5 PCNSL cases per y | Ref. | . | |
| >5 PCNSL cases per y | 1.17 | (1.07-1.27) | .0003 |
| Year of diagnosis, 2013 | Ref. | < .0001 | |
| 2014 | 1.31 | (1.18-1.46) | |
| 2015 | 1.49 | (1.34-1.65) | |
| 2016 | 1.63 | (1.45-1.82) | |
| 2017 | 1.73 | (1.56-1.92) | |
| 2018 | 1.69 | (1.52-1.88) | |
| Random intercept, specific hospital∗ | < .0001 |
| Variable . | Relative risk . | 95% CI . | P . |
|---|---|---|---|
| Age, per 10 y | 0.98 | (0.97-1.00) | .08 |
| Sex, female | Ref. | .010 | |
| Male | 0.95 | (0.91-0.99) | |
| Race/ethnicity, WNH | Ref. | .0001 | |
| Hispanic (any race) | 0.87 | (0.79-0.96) | |
| Black | 0.87 | (0.78-0.98) | |
| Asian/other | 0.96 | (0.88-1.04) | |
| Unknown | 0.66 | (0.46-0.93) | |
| HIV status, negative | Ref. | .0076 | |
| Positive | 0.79 | (0.67-0.94) | |
| Charlson-Deyo score, 0 | Ref. | .21 | |
| 1 | 1.01 | (0.96-1.07) | |
| 2 | 1.08 | (1.01-1.15) | |
| ≥3 | 1.02 | (0.94-1.11) | |
| Primary location, brain | Ref. | .10 | |
| Spinal cord | 1.01 | (0.90-1.15) | |
| Eye | 1.16 | (0.85-1.58) | |
| Meninges | 1.17 | (0.91-1.49) | |
| Other/unspecified | 1.05 | (0.99-1.12) | |
| Prior malignancy, none | Ref. | .19 | |
| Yes | 1.03 | (0.98-1.09) | |
| Chemotherapy, single agent | Ref. | .019 | |
| Multiple agents | 1.07 | (0.99-1.14) | |
| Unspecified | 0.78 | (0.59-1.03) | |
| Insurance, private | Ref. | .036 | |
| Medicaid | 0.94 | (0.86-1.04) | |
| Medicare | 0.94 | (0.87-1.01) | |
| Other government | 1.12 | (1.00-1.26) | |
| Not insured | 1.06 | (0.92-1.24) | |
| Unknown | 1.09 | (0.92-1.29) | |
| Hospital, academic | Ref. | .43 | |
| Community | 0.97 | (0.89-1.05) | |
| 0-5 PCNSL cases per y | Ref. | . | |
| >5 PCNSL cases per y | 1.17 | (1.07-1.27) | .0003 |
| Year of diagnosis, 2013 | Ref. | < .0001 | |
| 2014 | 1.31 | (1.18-1.46) | |
| 2015 | 1.49 | (1.34-1.65) | |
| 2016 | 1.63 | (1.45-1.82) | |
| 2017 | 1.73 | (1.56-1.92) | |
| 2018 | 1.69 | (1.52-1.88) | |
| Random intercept, specific hospital∗ | < .0001 |
Ref, references.
The main model presented is a general linear model with a Poisson link and robust standard error, which provides direct estimates of relative risk. However, because this type of model does not produce a likelihood measurement, the P value for the random intercept was obtained from an analogous logistic model via a likelihood ratio test of the mixed-effects model vs the nonhierarchical model.