Page 191 - 16Neonatal Jaundice_compressed

Neonatal jaundice

Figure C.3 The minimum number of kernicterus cases to be averted at different incremental costs of
more intensive testing

C.7 Sensitivity analysis

Sensitivity analysis is used in economic evaluation to assess how sensitive the results of the
model are to the assumptions made about the model parameters, particularly those parameters
where considerable uncertainty exists as to their actual value. One-way sensitivity analysis
involves altering the value of a single parameter, holding all the others constant, to determine
how sensitive the cost-effectiveness conclusion is to the assumptions made about that particular
parameter.
The base-case results above were presented as threshold analyses reflecting uncertainty about
the number of meters that would be needed for the TCB strategy and the number of kernicterus
cases that would need to be averted in order for the additional costs of more intensive costing to
be deemed an efficient use of scarce NHS resources. However, the sensitivity analyses below
explore how changes in other model parameters would affect results.
C.7.1 Varying the cost of meters

In this sensitivity analysis the cost of the meters is varied between £600 and £3,600. It is
assumed that the meter is a Minolta and does not therefore require a new calibration tip for
each test. The analysis additionally considers how this is affected by the number of meters
needed to deliver the TCB testing strategy.
Figure C.4 shows that the cost of meters is likely to be an important determinant of the cost-
effectiveness of TCB relative to TSB. As the cost of meters falls, the number of meters has far less
impact in determining the incremental costs of the TCB strategy.
C.7.2 Varying the mean number of tests per baby tested

In this sensitivity analysis the mean number of tests per baby tested is varied between one and
two tests. The analysis is repeated for different numbers of TCB meters.

Figure C.5 shows that the incremental costs of the TCB test strategy relative to the TSB test
strategy fall as the average number of tests per baby increase. This simply reflects that TSB test
has the higher marginal cost. If just one test per baby were required then the threshold number
of meters for cost neutrality would be approximately 7000. However, if babies were tested
twice on average then that cost neutrality threshold would rise to approximately 14 000 meters.
It should be noted that whilst more testing improves the cost-effectiveness of TCB relative to TSB
the opposite is the case with respect to current practice.

160