Table S1.
Regression models | Incremental effect on cost | 95% CI max | 95% CI min | AIC |
---|---|---|---|---|
Local costings | ||||
OLS (robust) | −£2,746.58 | −£4,899.02 | −£594.14 | 19.39 |
GLM (Gaussian-Log) | −£3,291.65 | −£4,952.08 | −£2,104.93 | 19.36 |
GLM (Gamma-ID) | −£2,589.00 | −£4,676.21 | −£501.80 | 20.19 |
GLM (Gamma-Log) | −£2,824.93 | −£5,266.25 | −£2,153.15 | 20.20 |
GLM (Inverse Gaussian-ID) | −£2,671.19 | −£4,531.63 | −£810.75 | 28.69 |
GLM (Inverse Gaussian-Log) | −£3,196.00 | −£20,881.79 | −£2,109.59 | 28.69 |
GLM (Poisson-ID) | −£2,609.35 | −£2,667.73 | −£2,550.97 | 1243.50 |
GLM (Poisson-Log) | −£2,970.33 | −£4,794.18 | −£2,083.05 | 1232.81 |
National NHS costing | ||||
OLS (robust) | −£2,715.33 | −£4,849.77 | −£580.88 | 19.37 |
GLM (Gaussian-Log) | −£3,244.57 | −£4,865.83 | −£2,093.79 | 19.35 |
GLM (Gamma-ID) | −£2,557.90 | −£4,635.49 | −£480.31 | 20.21 |
GLM (Gamma-Log) | −£2,790.23 | −£5,155.18 | −£2,134.29 | 20.21 |
GLM (Inverse Gaussian-ID) | −£2,631.80 | −£4,480.63 | −£782.97 | 28.72 |
GLM (Inverse Gaussian-Log) | −£3,051.61 | −£16,820.11 | −£2,118.51 | 28.72 |
GLM (Poisson-ID) | −£2,580.16 | −£2,638.81 | −£2,521.51 | 1214.53 |
GLM (Poisson-Log) | −£2,933.00 | −£4,712.09 | −£2,069.84 | 1204.22 |
Abbreviations: AIC, Akaike information criterion; GLM, generalized linear model; NHS, National Health Service; OLS, ordinary least squares.