Table 3. Estimation results of the two-part models on annual medical costs.
| |
1. Logit (non-zero costs) |
2. ln (costs) |
||
|---|---|---|---|---|
| Coefficient | P>z | Coefficient | P>t | |
| Low adherence (<80%) | −0.4629 | 0.000 | 0.2576 | 0.000 |
| Propensity score |
0.0905 |
0.414 |
0.0868 |
0.026 |
|
Follow-ups (reference Yr2+) | ||||
| Yr1 | 2.9177 | 0.000 | ||
| Local recurrence | 0.5496 | 0.159 | ||
| Axiliary recurrence | 1.5898 | 0.004 | ||
| Distant recurrence | 1.1649 | 0.000 | ||
| Breast cancer death | 2.2796 | 0.000 | ||
| Other-cause death | 1.9951 | 0.000 | ||
| Year of diagnosis |
0.0252 |
0.383 |
0.0097 |
0.491 |
|
Age group (reference <60), years | ||||
| 60–69 | 1.6259 | 0.267 | 1.2613 | 0.016 |
| 70–79 | 3.3071 | 0.118 | 2.1157 | 0.004 |
| 80+ |
3.5656 |
0.125 |
2.6093 |
0.001 |
|
Menopausal (reference pre/peri) | ||||
| Post | 0.3164 | 0.082 | 0.1541 | 0.119 |
| Unknown |
−0.1112 |
0.838 |
−0.3070 |
0.122 |
|
SIMD quintile (reference 1 and 2: deprived) | ||||
| 3–5- affluent | −0.8081 | 0.000 | −0.1402 | 0.007 |
| Unknown |
−0.7696 |
0.233 |
0.0462 |
0.858 |
|
Charlson's index (reference 0) | ||||
| 1–2 | 0.8782 | 0.215 | 0.5497 | 0.022 |
| 3+ |
0.3064 |
0.069 |
0.0631 |
0.381 |
|
Tumour stage (reference 1–2) | ||||
| 3–4 | −1.2306 | 0.367 | −1.0324 | 0.025 |
| Unknown |
−0.4610 |
0.291 |
−0.2107 |
0.186 |
|
Axillary nodes (reference 1) | ||||
| 2–3 | 0.2379 | 0.218 | 0.3075 | 0.000 |
| Unknown |
0.9024 |
0.007 |
0.3007 |
0.031 |
|
Metastases (reference yes) | ||||
| No | −15.5193 | 0.022 | −0.5763 | 0.075 |
| Unknown |
−15.6810 |
0.012 |
−0.1932 |
0.444 |
|
Tumour grade (reference 1–2) | ||||
| 3 | −0.3441 | 0.160 | −0.0187 | 0.840 |
| Unknown |
−0.2229 |
0.470 |
−0.0340 |
0.780 |
|
Oestrogen receptor status (reference negative) | ||||
| Positive | 1.6267 | 0.360 | 1.1324 | 0.073 |
| Unknown | 1.8644 | 0.341 | 1.2796 | 0.066 |
| % Missing months | −1.0209 | 0.000 | −0.0787 | 0.180 |
Coefficients in the first column capture how the probability a patient have non-zero costs within a year post cancer changes with variables; and coefficients in the second column capture, for patients with any costs within a year, how the amount of costs changes with variables.