Table 5.

Incremental cost-effectiveness for clinical practice and patient outcomes

Variable	ΔC	ΔE	ΔC / ΔE	Parametric (95%CI)†	Non-parametric (95%CI)‡
Clinical practice outcomes (NET sample)
PTA1	$169.89 (43, 297)	0.1363 (0.083, 0.190)	$1246	($525, $2055)	($595, $2066)
INFO2	$169.89 (43, 297)	0.0315 (− 0.030, 0.093)	$5393	($1672,− $18,125)	($1909, − $22,865)
SAFED3	$169.89 (43, 297)	0.0349 (0.010, 0.060)	$4868	($1882, $11928)	($2093, $11574)
Clinical outcomes and quality of life (NET-Plus sample)
HADS Anxiety4	$505.06 (96, 915)	− 0.52 (− 1.34, 0.30)	$971^	($132, − $1086)	($215, − $1084)
Rivermead PCS5	$505.06 (96, 915)	− 1.15 (− 2.77, 0.48)	$441^	($54, − $1270)	($90, − $1346)
SF6D HRQoL6	$505.06 (96, 915)	0.030 (− 0.00, 0.06)	$16,948	($2015,-$367,759)	($3796,-$314,637)

^†Parametric CIs derived via Fieller’s Theorem using iprogs.do [25] based on treatment effects reported in Tables 3 and 4 and using standard errors and correlations calculated from non-parametric bootstrap re-estimation

^‡Non-parametric CIs derived via acceptability method using bsceaprogs.do [25] based on the data generated from bootstrap re-estimation of our estimates of incremental cost and incremental effectiveness using bmultiv.do [25]

¹Cost per patient appropriately screened for PTA (primary outcome)

²Cost per patient who received written patient information upon discharge (INFO)

³Cost per safe discharge (SAFED) where SAFED defined as CT scan appropriately provided plus PTA plus INFO

⁴Cost per point improvement on anxiety questions of Hospital Anxiety and Depression Scale (HADS), higher scores indicate higher levels of anxiety

⁵Cost per point improvement on the Rivermead post-concussive symptoms scale, higher scores indicate greater severity

⁶Cost per point improvement in SF6D index scores, higher scores indicate higher HRQoL

^{^}Calculated as ΔC/− ΔE to reflect the fact that higher scores indicate greater severity for HADS and Rivermead