Economic evaluation of the NET intervention versus guideline dissemination for management of mild head injury in hospital emergency departments

Study setting: The NET cluster randomised controlled trial [ACTRN12612001286831]. Study sample: Seventeen EDs were randomised to the control condition and 14 to the intervention. One thousand nine hundred forty-three patients were included in the analysis of clinical practice outcomes (NET sample). A total of 343 patients from 14 control and 10 intervention EDs participated in follow-up interviews and were included in the analysis of patient-reported health outcomes (NET-Plus sample). Outcome measures: Appropriate post-traumatic amnesia (PTA) screening in the ED (primary outcome). Secondary clinical practice outcomes: provision of written information on discharge (INFO) and safe discharge (defined as CT scan appropriately provided plus PTA plus INFO). Secondary patient-reported, post-discharge health outcomes: anxiety (Hospital Anxiety and Depression Scale), post-concussive symptoms (Rivermead), and preference-based health-related quality of life (SF6D). Methods: Trial-based economic evaluations from a health sector perspective, with time horizons set to coincide with the final follow-up for the NET sample (2 months post-intervention) and to 1-month post-discharge for the NET-Plus sample. Results: Intervention and control groups were not significantly different in health service utilisation received in the ED/inpatient ward following the initial mTBI presentation (adjusted mean difference $23.86 per patient; 95%CI − $106, $153; p = 0.719) or over the longer follow-up in the NET-plus sample (adjusted mean difference $341.78 per patient; 95%CI − $58, $742; p = 0.094). Savings from lower health service utilisation are therefore unlikely to offset the significantly higher upfront cost of the intervention (mean difference $138.20 per patient; 95%CI $135, $141; p < 0.000). Estimates of the net effect of the intervention on total cost (intervention cost net of health service utilisation) suggest that the intervention entails significantly higher costs than the control condition (adjusted mean difference $169.89 per patient; 95%CI $43, $297, p = 0.009). This effect is larger in absolute magnitude over the longer follow-up in the NET-plus sample (adjusted mean difference $505.06; 95%CI $96, $915; p = 0.016), mostly due to additional health service utilisation. For the primary outcome, the NET intervention is more costly and more effective than passive dissemination; entailing an additional cost of $1246 per additional patient appropriately screened for PTA ($169.89/0.1363; Fieller’s 95%CI $525, $2055). For NET to be considered cost-effective with 95% confidence, decision-makers would need to be willing to trade one quality-adjusted life year (QALY) for 25 additional patients appropriately screened for PTA. While these results reflect our best estimate of cost-effectiveness given the data, it is possible that a NET intervention that has been scaled and streamlined ready for wider roll-out may be more or less cost-effective than the NET intervention as delivered in the trial.