J.1. Introduction

This analysis assesses the cost effectiveness of three antibiotics (determined by current prescribing practices and antibiotic resistance patterns of causative organisms in England and Wales) for the treatment of suspected meningococcal disease or suspected bacterial meningitis in children.

In economic evaluation it is necessary to take into account benefits and effects as well as costs. However, the clinical review undertaken for this guideline did not find evidence to support a difference in efficacy between the comparator antibiotics. Where there is no difference in effectiveness between different comparators, a cost-minimisation approach is justified. By selecting the cheapest option more resources are freed up for alternative uses in the NHS without any concomitant loss in health gain in the population of concern.

Therefore, a cost model was developed in Microsoft Excel® to compare the costs of the relevant antibiotics (ceftriaxone, cefotaxime and benzylpenicillin).

This model looks at the cost effectiveness of empiric antibiotics for suspected bacterial meningitis or meningococcal septicaemia. See section J.6 for discussion of the cost of antibiotics for confirmed bacterial meningitis or meningococcal septicaemia.

J.2. Method

The cost analysis is undertaken from the perspective of the NHS and personal social services which is in accordance with NICE guidelines methodology.^* The costing is done using a bottom-up or ’ingredients‘ approach which involves detailing the physical quantity of resources used in providing treatment alongside the unit cost of those resources. From this it is possible to estimate the total cost of treatment. This analysis has restricted itself to pharmaceutical, other consumables and staffing costs. Those costs that are the same across different treatments (such as the occupation of hospital bed) have been omitted as they have no impact on the cost differential between alternatives.

Unit cost data is taken from the most recently available published sources. Other model parameters are estimated using the expert opinion of the GDG.

The model did not address issues of antibiotic resistance which may, of course, have consequences for both health and resource use.

J.3. Model parameters and assumptions

The model’s input values are given in tables J.1 to J.6.

Table J.1

Staff unit costs.

Table J.2

Staff tasks.

Table J.3

Treatment.

Table J.4

Drug costs and dose.

Table J.5

Consumable costs.

Table J.6

Equipment costs.

The purchase of medical equipment also carries an opportunity cost but differs from operating costs, such as labour and consumables, in certain respects. The purchase of equipment often involves an upfront payment (or investment) before use. However, that cost is fixed as it does not vary with the quantity of treatment provided. The equipment can often be used over a number of years before it needs to be replaced.

Capital costs have two facets:

• Opportunity cost: the money spent on the equipment could have been invested in some other venture, yielding positive benefits. This is calculated by applying an interest rate to the sum invested in the equipment.
• Depreciation cost: the equipment has a certain lifespan and depreciates over time. Eventually, the equipment has to be replaced.

In economic evaluation, the usual practice is to annuitise the initial capital outlay over the expected life of the equipment. This gives an ‘equivalent annual cost’ which can then be apportioned to the procedure on a pro rata basis based on the typical equipment use over the course of the year in order to derive a unit cost of using that equipment. Calculating the equivalent annual cost means making an allowance for the differential timing of costs by discounting.

The formula for calculating the equivalent annual cost is:

where:

E = equivalent annual cost

K = purchase price of equipment

S = resale value

r = discount (interest rate)

n = equipment lifespan

A(n, r ) = annuity factor (n years at interest rate r )

Assigning equipment costs to an individual procedure is less straightforward. Firstly, it is necessary to calculate an equivalent annual cost, reflecting the initial purchase cost of the equipment. Table J.6 shows the values that were used to calculate the equipment cost per infusion for an annuity factor of 2.8.

J.4. Results

A comparison of the costs of the different antibiotics is shown in table J.7 and graphically in figure J.1. The calculation of these costs is described here.

Table J.7

Total costs of antibiotic treatment for a 20 kg child.

Figure J.1

Total costs of antibiotic treatment for a 20 kg child.

J.5. Sensitivity analysis

In economic evaluation a technique known as sensitivity analysis is used to assess the importance of uncertainty around baseline parameter values. If the model’s conclusions are not affected by changing assumptions and parameter values then there is greater confidence in the result suggested by the model. On the other hand if the model’s results are particularly sensitive to small changes in some parameter values this may indicate what the key drivers of the results are and where further research is needed to resolve uncertainty.

In this model there is some uncertainty around the timing and frequency of certain tasks. The results may also vary according to the weight of the child as drug dose is a function of weight. Two one-way sensitivity analyses are shown in figures J.2 and J.3 which indicate the effect of changing a single parameter value holding everything else in the model constant.

Figure J.2

Sensitivity analysis: varying child’s weight.

Figure J.3

Sensitivity analysis: varying cannula insertion time.

J.6. Discussion

With the base case assumptions ceftriaxone appears to be the cheapest antibiotic. This is because the saving in staff time associated with a treatment only administered once a day more than offsets the substantially higher cost of the drug itself. Sensitivity analyses generally showed that these results were not sensitive to one-way changes in model parameters, with ceftriaxone remaining the cheapest option under most scenarios. However, an exception was a sensitivity analysis suggesting that the results were sensitive to the weight of the child. Benzylpenicillin was cheaper than cefotaxime in children with a weight greater than 30 kg and cheaper than ceftriaxone in children weighing more than 50 kg.

This analysis strongly suggests that ceftriaxone is the most cost-effective antibiotic for the treatment of suspected meningococcal disease or suspected meningitis in a majority of children as, despite a bi-modal age distribution of disease, peak incidence would occur in children less than 20 kg in weight. However, it should be borne in mind that the cost model did not take into account any complicated ’downstream‘ effects on health or costs arising from antibiotic resistance, patterns of which may vary locally.

Cost of antibiotics for confirmed bacterial meningitis/meningococcal septicaemia

The model is essentially that used for empiric antibiotics for suspected disease. Treatment duration is longer and most costs increase as a linear function of duration. For patients treated with ceftriaxone earlier discharge may be possible, although actual practice varies, as only one dose per day is required. In the event of early discharge antibiotic treatment could be completed either by a home visit from a community nurse or as an out-patient in a ‘day-bed’ area of the hospital. In the absence of any increased risk early discharge is likely to increase the cost effectiveness of ceftriaxone relative to other antibiotic alternatives.

The results shown in figure J.4 indicate why this is likely to be the case.

Figure J.4

Total costs of antibiotic treatment for a 20 kg child with confirmed bacterial meningitis or meningococcal septicaemia.