Table 3.
Effects of hospital volume (as a categorical variable) on outcomes in patients with severe acute pancreatitis (n = 22,551)
| Outcomes | Hospital volume | Crude OR or percent change (95% CI) |
Adjusted OR or percent change (95% CI) |
|
|---|---|---|---|---|
| Model 1* | Model 2† | |||
|
Hospital LOS |
|
|
|
|
| |
Quartile 1 |
Ref. |
Ref. |
Ref. |
| |
Quartile 2 |
−2.52% (-12.78%, 8.94%) |
−15.64% (-24.13%, -6.21%) |
−12.95% (-21.22%, -3.82%) |
| |
Quartile 3 |
0.00% (-11.15%,12.56%) |
−16.78% (-25.57%, -7.02%) |
−14.95% (-23.46%, -5.50%) |
| |
Quartile 4 |
0.00% (-4.50%, 23.62%) |
−16.36% (-25.05%, -6.05%) |
−14.80% (-23.84%, -4.70%) |
| |
P for trend |
0.909 |
0.005 |
0.007 |
| |
R-squared |
-- |
23.81% |
30.36% |
|
Hospital charges |
|
|
|
|
| |
Quartile 1 |
Ref. |
Ref. |
Ref. |
| |
Quartile 2 |
8.65% (-4.50%, 23.62%) |
−9.99% (-18.65%, -0.41%) |
−5.15% (-13.38%, 3.86%) |
| |
Quartile 3 |
7.39% (-7.63%, 24.85%) |
−10.03% (-19.18%, 0.14%) |
−5.60% (-14.28%, 3.96%) |
| |
Quartile 4 |
9.25% (-6.79%, 28.05%) |
−11.42% (-20.90%, -0.80%) |
−6.79% (-15.92%, 3.33%) |
| |
P for trend |
0.398 |
0.053 |
0.205 |
| |
R-squared |
-- |
53.89% |
64.16% |
|
Hospital mortality |
|
|
|
|
| |
Quartile 1 |
Ref. |
Ref. |
Ref. |
| |
Quartile 2 |
0.98 (0.73, 1.31) |
0.58 (0.42, 0.82) |
0.75 (0.49, 1.15) |
| |
Quartile 3 |
1.09 (0.83, 1.45) |
0.62 (0.45, 0.88) |
0.83 (0.54, 1.26) |
| |
Quartile 4 |
1.02 (0.77, 1.35) |
0.58 (0.40, 0.83) |
0.75 (0.49, 1.16) |
| |
P for trend |
0.720 |
0.012 |
0.281 |
| c statistic | -- | 85.4% | 93.0% | |
OR: odds ratio; CI: confidence interval; LOS: length of stay.
* Covariates in model 1 included age (as a continuous variable), sex, year of admission, Charlson-Comorbidity Index (categorized as 0,1,2 and ≥3), urbanization, hospital level, the ownership of hospital, the region of hospital, causes of acute pancreatitis (categorized into biliary, alcohol-related, both or others), intensive care unit admission, organ failure, gastrointestinal bleeding and local complication.
† Model 2 enrolled all covariates of model 1 and additional treatment covariates, including cholecystectomy, total parenteral nutrition, vasopressors, hemodialysis and mechanical ventilation.