| Intervention | Estimation | Drug treatments were made up for each child every 24 hours. No matter how much of the drug was used, a new batch was made up every 24 hours. Time on treatment (from initial loading dose) was recorded by nursing staff for all children. All unused drugs were discarded. The consumables associated with daily drug treatments were estimated by nursing and clinical staff |
| Valuation | Price of clonidine was taken from MIMS (2013).51 There is no entry for clonidine in MIMS 2012. We have assumed that, as the price is very low, it is not unreasonable to assume the same price for 2012
Price of midazolam and morphine were taken from BNF (2012)50
Price of consumables and dextrose were taken from NHS Supply Chain catalogue (2012).52 Consumables include syringe, needle, extension line kit, line filter and line tap |
| Hospital stay | Hospital stays were divided into three categories: per diem, per diem GM ward and per diem HDU
Critical care paediatric bed-days: The PICU cost (£1826) was taken from the NHS Reference Costs 2011–1249 (XB05Z). The HDU cost (£920) was taken from the NHS Reference Costs 2011–1249 (XB07Z). The per diem GM ward cost (£331) was provided by the Finance/Accounts Department of Alder Hey Hospital, Liverpool
Hospital admissions are often made up of stays in different wards. All transfers between wards were recorded on the Patient Transfer form. Of the 108 children in the analysis, 13 did not have a completed Patient Transfer form. Data on LoS in PICU, GM and HDU were then obtained from the completed End of Study form. Only one child did not have this information recorded. For this child, an average of LoS in PICU was estimated using data from the 108 children with completed Patient Transfer forms. LoS in PICU was then subtracted from the total LoS to estimate days in the GM ward
Duration and therefore cost of inpatient stay is a key driver in the economic evaluation, and required careful consideration in the sensitivity analyses, in which various approaches were used to test the robustness of the economic evaluation results to changes in the cost of a hospital inpatient admission. For the most part, LoS was recorded accurately in terms of hours and minutes. However, only discharge dates were recorded (no time). We therefore assumed that all children were discharged from hospital at 23:59
In the base case cost estimates of LoS, if a child had spent > 12 hours in a ward, a full per diem cost was applied. If a child had spent < 12 hours in a ward, a half day cost was applied. Full days incurred the full per diem cost
In the sensitivity analysis, three different approaches to costing LoS were undertaken:
For PICU and HDU: Higher per diem cost [£2002 (PICU), upper quartile unit cost in NHS Reference Costs 2011–12; 49 £1117 (HDU) upper quartile unit cost in NHS Reference Costs 2011–12 49]. For GM: higher per diem cost (£500, assumption) For PICU and HDU: Lower per diem cost [£1554 (PICU), lower quartile unit cost in NHS Reference Costs 2011–12; 49 £785 (HDU) lower quartile unit cost in NHS Reference Costs 2011–12 49]. For GM: lower per diem cost (£225, assumption) Hours and minutes of inpatient stays on all wards were costed exactly, i.e. taking account of proportions of time instead of using half-day or per diem costs
|
| Hospital transfer | All children who were transferred between hospitals during the initial hospital admission were costed using the NHS reference cost of £230 (ASS02). Where no further information was available on LoS, it was assumed that all children had a stay in hospital at least until 14 days post-treatment cessation |
| Additional days in different hospital | Children were sometimes transferred to a different hospital for continuation of treatment. If the extended LoS was known then this estimate was used in the analysis. If the extended LoS was unknown then it was assumed that the child stayed in hospital at least until the time horizon used in the analysis (14 days post-treatment cessation or 14 days postventilation cessation) |
| SAEs | Total length of hospital stay costs already include any additional days in hospital due to a SAE
After careful examination of CRFs, only SAEs pertaining to two children required additional costing over and above the per diem cost. One child went from a GM ward to theatre on two separate occasions for a simple procedure that took 30 minutes. The cost of the SAE for this child was made up of (basic) theatre cost plus surgeon (average) cost per hour. This event was costed in the base case analysis and therefore subsequent sensitivity and scenario analyses. One child suffered a SAE while in the PICU and went to theatre for a cerebral drainage. The cost of the SAE for this child was made up of a (neurosurgery) theatre cost plus (high) surgeon cost per hour. This event was costed only in the sensitivity analysis with the extended time horizon (14 days postventilation cessation)
Cost source: Alder Hey Finance Department (Alder Hey Hospital, personal communication)
Cost source: Managing NHS Hospital Consultants 2012 (www.nao.org.uk/report/managing-nhs-hospital-consultants/)
|
| Death | It was to be assumed that any child who died during the trial within the time horizon of the economic evaluation incurred the cost of a post-mortem as a proxy for the costs associated with dying in hospital. However, none of the children died in the trial during the two time periods of interest
Cost source: Alder Hey Finance Department 2012 (Alder Hey Hospital, personal communication)
|
| Scenario analysis: wider NHS costs (14 days post-treatment cessation) |
|---|
| GP attendance | Cost source: Personal Social Services Research Unit 2012 (Curtis48)
|
| A&E attendance | The cost estimate used in the analysis depended on whether or not the child was admitted to hospital as a result of attendance
Cost source: Personal Social Services Research Unit 2012 (Curtis48)
|
| Hospital admission | The GM per diem cost used in the baseline analysis (£331) was used to estimate the cost of any additional day spent in hospital as part of a re-admission within 14 days post-treatment cessation |
| Additional sensitivity analyses |
|---|
A further three sensitivity analyses were undertaken:
Extended time horizon: from randomisation to 14 days postventilation cessation (one child did not have a record of number of days ventilated, the average number of days ventilated using data from 119 children was therefore estimated and used in the analysis) Wider definition of ‘adequate sedation’: at least 75% of total time spent sedated within a COMFORT range of 17 to 26 Narrower definition of ‘adequate sedation’: at least 85% of total time spent sedated within a COMFORT range of 17 to 26
|
