Full multilevel models# of the association between yearly hospital surgical lung biopsy (SLB) volume and 30-day post-operative mortality in patients with interstitial lung disease undergoing SLB in the overall cohort and stratified by nonelective and elective patients
| Variable | Overall cohort | Nonelective patients¶ | Elective patients¶ | |||
| OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |
| Subjects | 3039 | 858 | 2181 | |||
| Yearly hospital SLB volume+ | 0.84 (0.73–0.97) | 0.02 | 0.84 (0.69–1.02) | 0.08 | 0.94 (0.74–1.18) | 0.57 |
| Nonelective | 8.76 (6.19–12.39) | <0.0001 | ||||
| Open thoracotomy | 2.70 (1.95–3.75) | <0.0001 | 3.29 (2.18–4.95) | <0.0001 | 1.35 (0.70–2.61) | 0.37 |
| Long-term oxygen therapy | 1.93 (1.19–3.13) | 0.008 | 1.05 (0.55–2.03) | 0.88 | 5.48 (2.70–11.12) | <0.0001 |
| Male sex | 1.64 (1.20–2.22) | 0.002 | 1.51 (1.04–2.17) | 0.03 | 2.05 (1.08–3.91) | 0.03 |
| Age group years | 1.36 (1.20–1.55) | <0.0001 | 1.38 (1.19–1.61) | <0.0001 | 1.34 (1.00–1.79) | 0.05 |
| Charlson Comorbidity Index | 1.17 (1.09–1.26) | <0.0001 | 1.12 (1.03–1.22) | 0.007 | 1.37 (1.19–1.56) | <0.0001 |
| Procedure year | 1.08 (1.04–1.13) | 0.0005 | 1.09 (1.03–1.15) | 0.002 | 1.03 (0.94–1.13) | 0.53 |
| Income quintile | 0.95 (0.86–1.05) | 0.33 | 0.96 (0.85–1.09) | 0.52 | 0.90 (0.73–1.12) | 0.35 |
| ICC§ | 0.02 | 0.05 | 0.00 | |||
ICC: intraclass correlation coefficient. There were 18, 2 and 16 patients excluded from the overall cohort, nonelective patients and elective patients, respectively, due to missing data. OR >1 represents higher odds of 30-day post-operative mortality, OR <1 represents lower odds of 30-day post-operative mortality. Age and income were analysed as ordinal variables, with increasing age group being associated with higher 30-day post-operative mortality in the overall cohort and nonelective patients. Age groups used were 18–29, 30–39, 40–49, 50–59, 60–69 and ≥70 years. Charlson Comorbidity Index and procedure year were analysed as continuous variables, with higher Charlson Comorbidity Index being associated with higher 30-day post-operative mortality in all three groups, and later procedure year being associated with higher 30-day post-operative mortality in the overall cohort and nonelective patients. All other variables were analysed as categorical (yes versus no). #: the full multilevel model includes hospital (yearly hospital SLB volume) and patient-level covariates; ¶: an interaction term for yearly hospital SLB volume and nonelective versus elective patients was nonsignificant; +: unit for OR is per 50 SLBs performed; §: ICC represents the variability in 30-day post-operative mortality that is accounted for by hospital differences (e.g. in the overall cohort, 2% of the variability in 30-day post-operative mortality is accounted for by between-hospital differences, leaving 98% of the variability attributable to patient characteristics).