Page 194 - 16Neonatal Jaundice_compressed

Appendix C: Economic evaluation of testing strategies for hyperbilirubinaemia

C.8 Discussion

The analysis compared the current testing strategy with an uplift in testing using alternative
strategies. In the base-case analysis, the current strategy of testing only 10% of babies using TSB
was £1.02 million per year. The next cheapest strategy was to use Strategy 3 (TCB to all visually
jaundiced babies followed by TSB if TCB is positive) using a meter that does not require
calibration tips which cost £10.16 million, or £13.25 million using a meter requiring a
calibration tip. Using the TSB more intensively (on 60% of babies who are visibly jaundiced)
would cost £10.22 million per year. The cost difference between TSB and TCB is mainly due to
the increased time to do a blood test compared with a skin test.

An important question is whether any change from current practice can be justified on cost-
effectiveness grounds. In part this depends on the fixed costs, that is, the number of TCB meters
needed to deliver Strategy 3. This determines the incremental costs of increased testing if the
TCB strategy is deemed more cost-effective than TSB, that is, the strategy with the lowest cost
since this is a cost-minimisation analysis. In the base-case analysis, the results estimate that the
maximum incremental cost of more intensive testing is around £9.14 million, which is the
incremental cost of an enhanced testing strategy using TSB alone relative to current practice. If
the strategy using the TCB could be delivered with the purchase of only 1000 additional
bilirubinometers (which would be a highly conservative estimate) then the incremental cost
would be £2.96 million. Figure C.3 suggests that 1.52 cases of kernicterus would have to be
averted for more intensive testing to be considered cost-effective if the incremental testing costs
were £9.14 million. If fewer resources were required (fewer bilirubinometers purchased) then
fewer cases would need to be averted. This assumes a threshold QALY value of £20,000. At a
higher threshold, say £30,000 per QALY, the number of cases of kernicterus averted in order for
more intensive testing to be cost-effective would be fewer.
Figure C.3 shows how this threshold of kernicterus cases that need to be averted for cost-
effectiveness falls as the incremental costs of more intensive testing fall, as is the case with a
smaller number of TCB meters. The evidence base is currently not sufficiently robust to assess
whether more intensive testing would achieve such an incremental gain – there are
approximately five to seven new kernicterus cases per year in England and Wales. However,
given the evidence about the limitations of visual examination, the GDG is opposed to relying
on observations that have been demonstrated to be unreliable in the detection of severe
hyperbilirubinaemia. It does seem plausible that a more intensive testing strategy using tests that
are known to have greater reliability in detection of severe hyperbilirubinaemia would lead to
more appropriate and timely intervention with a concomitant reduction in adverse outcomes.

The costs of the TCB testing strategy vary according to the cost of the meter used. In the absence
of evidence that health outcomes are different between types of meter used, the cheaper
Minolta meter should be preferred. The base-case results (see Figure C.2) suggest that, at current
prices, the Minolta meter would be about £3 million cheaper, assuming that the meters
themselves are similarly priced. Therefore, in the remainder of the discussion it will be assumed
that the analysis is based on the cheaper Minolta TCB meter.
Figures C.1, C.3, C.4 and C.5 all show that the number of meters necessary to deliver the TCB
strategy is important in determining the relative cost-effectiveness of the TCB strategy
(Strategy 3) to the TSB strategy (Strategy 2). In the base-case analysis, TCB is cheaper than TSB
providing the number of TCB meters is less than 9200.
If it is decided that more intensive testing is likely to be cost-effective then a secondary decision
is whether initial testing should be done using TCB or TSB. Factors such as convenience to the
nurse and discomfort to the baby are not irrelevant to the decision but have not been included
explicitly in this analysis because they are difficult to quantify and probably of a relatively small
magnitude. This analysis suggests that the choice between TCB and TSB would depend on the
number of meters that would be required. The NHS staff census as reported on the NHS
Information Centre website (www.ic.nhs.uk/statistics-and-data-collections/workforce/nhs-staff-
numbers, accessed August 2009) reports the ‘head count’ figure for practising midwives as
25 000 with 19 500 full-time equivalents. The base-case analyses suggest that were all midwives
required to have a TCB meter in order to implement a TCB strategy then TSB would be the cost-

163