Table 3. Model Results of Primary Analysis of Adherence vs Models Using Alternate Definitions of Expected Follow-up and Sensitivity Analyses.
| Analysis | Evaluation, No. (%) | Total, No. | |||
|---|---|---|---|---|---|
| Early | Expected | Late | No evaluation | ||
| Primarya | 2296 (8.1) | 17 863 (63.1) | 3696 (13.1) | 4439 (15.7) | 28 294 |
| Alternate definitions of adherenceb | |||||
| Stringent | 2418 (8.3) | 14 486 (49.7) | 6951 (23.9) | 5301 (18.2) | 29 137 |
| Liberal | 1986 (7.6) | 20 579 (78.8) | 1290 (4.9) | 2259 (8.7) | 26 114 |
| Sensitivity | |||||
| Exclude veterans with CAN score >95c | 2254 (8.0) | 17 737 (63.2) | 3662 (13.0) | 4415 (15.7) | 28 068 |
| Exclude veterans age <65 y | 1403 (8.9) | 10 297 (65.4) | 1863 (11.8) | 2173 (13.8) | 15 736 |
| Restricted to study period 2015-2018 | 1273 (8.8) | 9306 (64.5) | 1500 (10.4) | 2350 (16.3) | 14 429 |
Abbreviations: CAN, Care Assessment Needs; Lung-RADS, Lung CT Screening Reporting & Data System.
In our primary analysis, expected follow-up is defined as occurring within 3 months of the recommended Lung-RADS evaluation for Lung-RADS 1-3 and within 2 months for Lung-RADS 4. All models defined an early evaluation as occurring more than 2 months before the recommended Lung-RADS.
In the stringent model, expected follow-up is defined as occurring within 1 month of the Lung-RADS recommended interval; in the liberal model, expected follow-up is defined as occurring within 12 months of the recommended evaluation for Lung-RADS 1 and 2, 6 months for Lung-RADS 3, and 3 months for Lung-RADS 4.
CAN score of 95 or greater indicates estimated 20% risk of death within 1 year.