PN i 3 induced miscarriage p that a woman who has accepted PND has a miscarriage due to the procedure (CVS) 0.015 Step 10 70,238 Fetus diagnosed affected p that a truly affected fetus is diagnosed as affected (1-false negative rate)
Step 3 N o reference, see section 4.4.1
5 SCREENING PROGRAMME COSTS
5.1
Synopsis
This chapter presents the cost information that is required by the model. Only direct health service costs have been considered. These include for antenatal screening: laboratory costs for ascertainment o f parental carriers and at risk couples, and costs for ascertainment o f an affected fetus and termination o f pregnancy; for neonatal screening: laboratory costs for screening and confirmatory ' tests; and for both components: costs associated with education/counselling. The emphasis o f the chapter is on explaining the assumptions that I developed to enable costing. They were programmed into the model by AE Ades. Cost measurement and valuation, carried out by J Kamon and J Brown, are briefly described to complete the chapter. Cost sources included the published literature, local providers, a reference laboratory and manufacturers and are listed in the relevant sections. All data on costs were uprated to 1997 values using the NHS price index. Tables 5.1 and 5.2 at the end o f the chapter summarise the main antenatal and neonatal screening costs used in the analysis.
5.2
Antenatal screening programme costs
5.2
.1 Laboratory costs for ascertainment of parental carriers and at risk couplesThe sequence and type o f laboratory tests required for ascertainment o f parental
carriers and at risk couples has been summarised in section 2.3.5.1 and figure 2.1
(page 27), which depicts the antenatal laboratory algorithm. The parental carrier states
considered in the model, namely S, C, D, E, and trait are ascertained using
different laboratory pathways. In addition, to establish that an individual is not a significant Hb-pathy carrier requires a variable sequence o f laboratory tests. These tests depend on i) whether the individual screened has iron deficiency or trait/homozygous state, conditions that need to be distinguished from significant Hb- pathy traits; and ii) the individual’s ethnic group. As Chinese, Other Asian and
Cypriot have an increased risk o f trait, these groups need additional laboratory
algorithm figure 2.1, page 27), whereas in all other groups such results are interpreted as negative without further tests. Table 5.3 summarises the laboratory tests necessary to identify the different significant Hb-pathy carrier states, iron deficiency, trait/homozygous state and the normal state (with neither o f these conditions) in mothers and their partners.
Table 5.3 Laboratory tests required for ascertainment o f significant
haemoglobinopathy carrier states, iron deficiency, a®**’** trait/homozygous state and normal state in mothers and partners
Individual laboratory tests Significant Hb-pathy carrier states
Confounding conditions Norm al
S
c
D E p lh a l ^ O th a l ^ . . h a i a Iron deficiency^ MCH" ✓ ✓ ✓ /y
y
y
y
y
Characterisation o f Hb variants / ✓ ✓y
y
y
y
y
y
Repeat testing for HbS
y
✓ ✓y
Repeat testing for otherHb variants
✓ ✓
y
H bA2 quantification
y
y
y
y
HbF quantification
y
y
y
y
D N A analysis a d
a ^+thai h om ozygous state.
For w om en the cost o f the MCH test is assumed to be attributed to obstetric care, for partners to the screening
215
programme.
■ The probability o f iron deficiency in partners has been conservatively assumed to be 0.01 See explanation in the text.
If assessment o f an at risk pregnancy depends on a carrier result which cannot reliably be done by phenotyping, DNA analysis is required (section 2.3.5.1). In box 5.1 these cases are specified with the number o f DNA analysis required for a risk assessment.
Table 5.4 summarises the individual test costs depending on laboratory equipment. The two laboratory set-ups considered are the ‘standard’ and ‘H PLC’ set-up, as defined in section 2.3.5.1.
Box 5.1 Assumptions for the need of DNA analysis for risk assessm ent o f carrier couples
• I f either partner is found to have H bD and the other H bS, one D N A analysis is
undertaken to specify the HbD as HbD'"""-’''*’.
• In rare cases o f and Sp"’*’’' trait. To account for these cases, it w as assum ed th at 1%
o f all p"'^' trait results require DNA analysis.
• I f both partners have trait, two DNA analyses are required.
• I f a wom an has trait and her partner has trait/hom ozygous state or iron
deficiency, tw o DNA analyses are required.
• I f a w om an is from a Chinese, O ther Asian or C ypriot ethnic group and has
trait/hom ozygous state or iron deficiency, and her partner has trait,
trait/hom ozygous state or iron deficiency, one D N A analysis is required.
• I f a w om an is from a Chinese, O ther A sian or C ypriot ethnic group and has trait,
trait/hom ozygous state or iron deficiency, and her partner is unavailable, one D N A analysis is required.
Table 5.4 Costs per individual laboratory test according to laboratory
equipment used
Individual laboratory test Standard laboratory set-up HPLC laboratory set-up Method Cost ( £ ’s) M ethod C ost ( £ ’s) M CH Automated 1.55 Automated 1.55 Characterisation o f Hb variants H b-electrophoresis
(cellu lose acetate)
2.56 HPLC 3.11 Repeat testing for HbS Sickle solubility 1.85 S ickle solubility 1.85 Repeat for other Hb variants H b-electrophoresis
(citrate agar)
6.39 HPLC 3.11 H bA ; quantification Elution 7.50 HPLC 0^ HbF quantification Betke 3.39 HPLC
D N A analysis Section 5.2.2 147 Section 5.2.2 147 Source o f cost data; local screening laboratory with an annual throughput o f about 4 ,0 0 0 sam ples. C osts included consum ables, direct labour o f laboratory technicians, general overheads (secretarial support, general repairs, quality control), and sp ecific overheads for the Hb-pathy screening laboratory including the cost o f obtaining the blood sam ples (costing the phlebotom ist’s salary and the associated consum ables).
A ll costs are uprated to 1997 values.
^ HPLC characterises Hb variants and quantifies H bA2 and HbF sim ultaneously.
The above assumptions and costs were programmed into the model for mothers and partners according to ethnic group and carrier status. This allowed calculation o f laboratory costs per women tested by ethnic group, to compare HPLC and standard laboratory equipment (table 5.5). The laboratory costs comprise m other’s carrier test, father’s carrier test and any additional DNA testing required, under baseline assumptions including 10% iron deficiency in the antenatal population. It is evident
that use o f HPLC is cheaper than the standard equipment in every ethnic group. (This is the case at any level o f iron deficiency above 5% - results not shown separately). Hence for inclusion in the model laboratory costs based on HPLC technology have been used. Estimates have been varied by a factor o f 1.5 in a sensitivity analysis.
Table 5.5 Laboratory costs per woman tested, by ethnic group, using HPLC
and standard laboratory equipment, assuming 10% iron deficiency in antenatal population
Ethnic group Laboratory cost, £ ’s per wom an screened HLPC Standard Black Caribbean 5.34 7 .42 B lack African 5.76 7 .94 B lack Other 5.28 7.31 Indian 5.70 10.40 Pakistani 5.52 10.24 Bangladeshi 4.79 6.63 C hinese 37.39 38.4 4 Other A sian 23.58 24.21 Other 3.84 4 .2 7 Cypriot 46.54 4 9 .6 7 Italian 3.84 4 .4 9 North European 3 .50 3 .88 M odel predictions under baseline assumptions (table 4.1, 4.2, pages 91-94).
A ll costs are uprated to 1997 values.
5 .2 .2 Costs for ascertainment of an affected fetus and termination o f
pregnancy
For the purpose o f this study we assumed that fetal material for PND is derived by CVS, the m ost common method used in the UK^^. Cost estimates were taken from the literature^®^ and provided by a hospital trust. A total cost o f £320 was included in the model.
Methods o f fetal diagnosis have been discussed in section 2.3.5.2. We have assumed that fetal diagnosis is performed by DNA analysis o f both parental and fetal samples, following the approach o f the Oxford haemoglobinopathy reference laboratory^ which complies with current guidelines'*^. Nowadays it is always preceded by repeat phenotyping o f both parental blood samples to ensure reliability o f the risk assessment o f the pregnancy
Fetal diagnosis o f each o f four disease groups, namely p thalassaemia m ajor and
other sickle cell disorders require different DNA techniques. Assumptions are
summarised in table 5.6. The total cost estimates per PND according to suspected
disorder are given in summary table 5.1 at the end o f this chapter.
Table 5.6 DNA techniques used for diagnosis of fetal haemoglobinopathies
Suspected disorder Technique used Initial test Repeat test P thalassaem ia major/
HbE P thalassaem ia
ARMS-PCR^ RFLP linkage analysis'^ a" thalassaem ia hydrops fetalis Southern blot Southern blot
S ickle cell P thalassaemia ARMS-PCR^" D d el-P C R Other sickle cell disorders AR M S-PC R D d el-PC R A R M S = A m plification refractory mutation system, PCR = Polym erase chain reaction RFLP = Restriction fragmant length polym orphism, D del = name o f restriction enzym e
In cases o f rare mutations o f 3 thalassaem ia major, a second AR M S-PC R test is undertaken prior to linkage ^ a ly s is . 10% o f mutations have been assumed to be rare (personal communication J Old).
For RFLP linkage analysis, it was assumed that 70% o f cases required analysis o f sam ples from three fam ily m embers, w h ilst the remainder required analysis o f sam ples from six fam ily members (personal com m unication J Old).
T w o initial A R M S-PC R tests are alw ays required for suspected sickle cell P thalassaemia.
PND costs were integrated into the model for each woman accepting PND, according to the genotype o f her fetus.
In the UK the majority o f TOPs following the diagnosis o f a fetus affected by a Hb- pathy are performed in the late first or second trimester^^. The current standard method o f termination for these gestations is surgical r e m o v a l A cost o f £470.50 was included in the model, based on estimates from a hospital trust and a charity.
TOP costs were integrated into the model for each woman accepting termination.
5.3
Neonatal screening programme costs
5.3.1 Laboratory costs for the initial neonatal screening test, repeat test and confirmatory diagnostic test
We have assumed that collection o f neonatal specimens follows the Guthrie card
method with integration o f the screening process into an existing large scale
metabolic neonatal screening programme. The sequence and type o f laboratory tests required for initial neonatal sickle cell screening, repeat testing o f abnormal results
and confirmatory diagnostic tests have been summarised in section 2.4.5.1 and figure
2.2 (page 41), which depicts the neonatal laboratory algorithm. It has been assumed
that the primary laboratory method used for all tests is HPLC, using the equipment to full capacity; confirmatory diagnosis, undertaken in a haematology laboratory, includes phenotyping o f infant and parents and in 1% o f presumptive positive infants (without one or both parents available for testing) DNA analysis. Costs o f confirmatory tests are included in the screening programme costs for false positive sickle cell disease results and non-significant combinations (see table 2.4, page 37) because in these cases no tests would have been performed without a screening programme in place. All other confirmatory tests would have been required regardless o f screening and were hence excluded. To cover uncertainty about laboratory costs, they were varied by a factor o f 1.5 in a sensitivity analysis. Cost estimates o f the
individual tests are listed in summary table 5.2 at the end o f the chapter.
Laboratory costs for neonatal screening tests, repeat tests and confirmatory diagnosis were integrated into the model for all screened newborns according to their genotype.