OUP user menu

A decision analysis model for diagnostic strategies using DNA testing for hereditary haemochromatosis in at risk populations

K. Cooper, J. Bryant, J. Picot, A. Clegg, P.R. Roderick, W.M. Rosenberg, C. Patch
DOI: http://dx.doi.org/10.1093/qjmed/hcn070 631-641 First published online: 3 June 2008

Abstract

Background: New techniques for diagnosing hereditary haemochromatosis (HHC) have become available alongside traditional tests such as liver biopsy and serum iron studies.

Aim: To evaluate DNA tests in people suspected of having haemochromatosis at clinical presentation compared to liver biopsy, and in family members of those diagnosed with haemochromatosis compared to phenotypic iron studies in UK.

Methods: Decision analytic models were constructed to compare the costs and consequences of the diagnostic strategies for a hypothetical cohort of people with suspected haemochromatosis. For each strategy, the number of cases of haemochromatosis identified and treated and the resources used were estimated.

Results: For diagnostic strategies in people suspected clinically of having haemochromatosis, the DNA strategy is cost saving compared to liver biopsy (cost saved per case detected, £123) and continues to be so across all ranges of parameters. For family testing, the DNA strategy is cost saving for the offspring of the proband but not for siblings. If the DNA test cost were to reduce by 40% to £60 or, if in the phenotypic model, those with initially normal iron indices were retested twice instead of once, the DNA strategy would be the cheaper one.

Conclusions: Diagnostic strategies involving DNA testing are likely to be cost saving in clinical cases with iron overload and in the offspring of index cases. This study supports the UK guideline recommendations for the use of DNA testing in UK.

Introduction

Hereditary haemochromatosis (HHC) type 1 (OMIM 235200) results from a genetic disorder of iron metabolism which leads to excessive intestinal absorption of iron and progressive abnormal iron deposition in the liver, heart, pancreas and other vital organs. It can cause morbidity and, if untreated, premature death. HHC is thought to be one of the most common genetic diseases in people of European and Celtic descent with between 0.68% and 1.24% of the UK population homozygous for the C282Y mutation.1,2 Identified genetic mutations account for up to 95% of the disease in Caucasians. Although uncertainty remains as to the incidence and prevalence of the clinical disease due to the variability in its clinical penetrance (i.e. phenotypic expression of the genetic mutation), it is estimated that between 380 000 and 690 000 people in the UK could have the genotype with between 1% and 40% at risk of developing the phenotypic disease.3 HHC is treatable by removing blood (venesection) until blood iron levels fall sufficiently. If treatment is started prior to irreversible end organ damage, it can restore near normal life expectancy.

Although early diagnosis of the condition is essential to limiting the morbidity and mortality from HHC, it has proved difficult due its non-specific symptoms (e.g. weakness, fatigue, joint pain) and associated co-morbidities (e.g. liver disease, diabetes, arthritis, impotence or heart disease). Traditionally biochemical tests, specifically transferrin saturation (TS) and serum ferritin (SF), followed by liver biopsy have provided the basis for diagnosing HHC in people who are suspected of having the clinical features of the condition. The advent of genetic testing for HHC has provided the opportunity to identify the hereditary nature of the primary iron overload.4 This has allowed individuals with signs and symptoms and raised TS and SF levels to be tested, monitored and treated appropriately. Also it has allowed offspring and siblings of those with HHC who are at increased risk of inheriting the susceptibility to the disease to be tested and identified earlier. Those homozygous to the mutation can be monitored and treated appropriately, whilst those not at risk can be excluded from unnecessary follow-up or intervention.

Diagnosis of HHC is complicated by the variable penetrance of the condition. Identification of the genetic mutation for HHC does not necessarily predict those people who will develop the phenotypic disease. Although genetic testing is now used to diagnose HHC in the UK, there is uncertainty about its cost effectiveness compared to other diagnostic techniques.5 Evaluation of the comparative benefits, risks and costs associated with a technology are an important part of assessing a technology prior to recommending the provision of a service. This study was commissioned by the UK National Institute for Health Research Health Technology Assessment Programme to assess the clinical and cost effectiveness of different strategies for diagnosing HHC among those suspected of having the condition and their family members. It evaluated the evidence on the analytic validity, clinical validity, clinical utility, cost effectiveness and the ethical and psychosocial implications of genetic testing. This article summarizes the findings from our economic evaluation of the different diagnostic strategies.3 It addresses the following questions:

  1. What are the costs and benefits of using DNA tests instead of liver biopsy in people suspected of having haemochromatosis at clinical presentation?

  2. What are the cost and benefits of using DNA tests compared to phenotypic iron studies in family members of those diagnosed with haemochromatosis?

Methods

Decision analytic models were constructed in Microsoft Excel according to standard modelling methods.6,7 Costs were derived from primary data from previous studies, and national and local unit costs (Table 1). Only direct National Health Service (NHS) costs were included and hence the model was from the perspective of the UK NHS.6 The time horizon chosen for the model was for the testing and treatment period only and monitoring offspring until the disease manifests. The model did not consider the longer term health outcomes of patients with haemochromatosis. The economic evaluation was a cost effectiveness analysis that focused on estimating the number of cases detected and treated by each diagnostic strategy, and the resources used. The outcome is reported as cost per case detected as this outcome is of most interest to clinicians and medical providers such as the NHS in UK. Differences in costs between the strategies are also reported. The structure and data inputs of all the decision trees were informed by systematic literature reviews,3 clinical guidelines,5 the results of systematic searches and discussion with clinical experts. The decision tree models do not consider long-term cost and consequences of HHC because the data on the long-term cost and consequences were considered of poor quality and this would have necessitated making significant assumptions.

View this table:
Table 1

Key inputs to the decision tree models

Input valueReference/comment
Diagnostic tests (%)
 TS >45%a
        Sensitivity94Olynyk et al.10
        Specificity94
 SF >300 μg/la
        Sensitivity73Moodie et al.11
        Specificity85
 SF >200 μg/la
        Sensitivity73Moodie et al.11
        Specificity70
Prevalence of HHC
    Population with suspected iron overload0.038Moodie et al.11 (from liver clinic)
    Siblings0.25Mendelian
    Offspring0.050.5 × 1 in 10
    Among patients with HHC91.3UK HHC23
    Among the general population0.99UK HHC23
Liver biopsy
    Death0Assumption Gilmore et al.18
    Bleeding requiring transfusion0Assumption Gilmore18
For diagnostic pathways tree (%)
    Proportion with raised SF > 1000 μg/l (DNA positive)39Vantyghem et al.13
    Proportion with raised SF > 300 μg/l and SF < 1000 μg/l (DNA positive)61Vantyghem et al.13
    Proportion with raised SF > 1000 μg/l (DNA negative)24Vantyghem et al.13
    Proportion with raised SF > 300 μg/l and SF < 1000 μg/l (DNA negative)76Vantyghem et al.13
For family testing tree (%)
 Penetrance of HHC
        Men76McCune et al.16
        Women32McCune et al.16
    Proportion of offspring with HHC who initially have iron overload20Assumption
Costs
    DNA test labc£100UKGTN24
    Iron test labd£11.70Shepherd et al.25
    Venesection treatmentc (15 mins)£8.75(Band 5) Curtis and Netten26
    Liver biopsy day casec£388.05Shepherd et al.(2006)25
    Nurse appointment (15 mins)£8.75(Band 5) Curtis and Netten26
    Surgical consultant (30 mins)£39.00Curtis and Netten26
    Number of venesection treatments20Assumption–clinical opinion
    Number of monitoring (blood tests)
        Offspring5Assumption–clinical opinion
        Sibling1
  • aAccuracy of biochemical tests for detecting haemochromatosis.

  • bAccuracy of liver biopsy for detecting haemochromatosis.

  • cCost does not include consultant and nurse appointment costs.

  • dCost for SF or TS saturation test.

Decision tree model for DNA testing instead of liver biopsy to diagnose in individuals suspected of HHC

Prior to the discovery of the common gene mutations, diagnosis of HHC was based on clinical suspicion including persistently raised TS and SF with no other diagnosis followed by liver biopsy to measure hepatic iron. Since the identification of the gene it has been possible to use DNA testing to confirm the diagnosis in those in whom it is suspected. Liver biopsy then becomes a prognostic test in those suspected of having liver damage and potentially can be avoided in those with raised iron levels and no biochemical evidence of liver damage.

Decision models were constructed to compare the costs and consequences of two diagnostic algorithms in people suspected of having HHC on the basis of persistently raised TS > 45% and SF > 300 μg/l;8 liver biopsy for all people in one and genetic testing for all people in the other. The strategies used for the decision tree adopt the current UK recommendations for genetic testing compared to previous strategies.5 The advantage of the genetic testing strategy is that it avoids the unnecessary use of liver biopsy. The decision tree is shown in Figure 1. For the liver biopsy strategy, all people have a liver biopsy and are either confirmed positive or negative for phenotypic haemochromatosis based on the hepatic iron index. Those who are positive will be treated for HHC and those who are negative will not be.

Figure 1.

Decision tree for DNA testing vs. liver biopsy for confirmation of HHC of those suspected (LB liver biopsy, +ve positive, −ve negative).

For the DNA test strategy, all people receive a DNA test which will be either positive (YY, C282Y homozygous) or negative. We omitted compound heterozygotes for C282Y and H63D because the penetrance of HHC in this disease associated genotype is relatively low.9 It is unclear the effect that including compound heterozygotes would have had on the model results due to the uncertainty around this condition. All patients with raised SF (>1000 μg/l) also receive a liver biopsy to check for liver cirrhosis.5,8 All patients with a positive DNA test will be treated with venesection as they are assumed to have phenotypic HHC. Patients with a negative DNA test and SF <1000 μg/l are monitored and receive a repeat SF test. If their SF is stable or decreasing then they are not treated whilst those with increasing SF have a liver biopsy to confirm haemochromatosis. Those with confirmed haemochromatosis (i.e. due to genetic mutations not yet identified) will be treated.

The model can be run for different strategies with different threshold values for TS and SF positive tests. For the purpose of the model a typical patient is defined to represent all patients. For the baseline run, we assume a ‘typical’ patient is a male aged 45-years-old because HHC is more common in men and raised iron levels will typically appear by this age. The effect of age or gender is investigated in the model through the use of sensitivity analyses.

Decision tree model for DNA testing for family members of people diagnosed with HHC

Separate decision models were constructed to compare the costs and consequences of two testing algorithms for siblings and offspring of people diagnosed with HHC. Gene testing is negative for the mutation (YY, C282Y) in 10% of cases and the relatives of index cases are unsuitable for DNA testing and are not considered here. We have not considered compound heterozygote families here as there is lower penetrance and higher uncertainty from iron testing for these people. The algorithms for testing family members are biochemical testing for all vs. DNA testing for all. The strategies used for the decision tree adopt the current UK recommendations for genetic testing compared to previous strategies.5 The decision tree is shown in Figure 2. For the biochemical test strategy relatives have SF and TS tests. If they have raised iron levels according to clinical guidelines,5,8 i.e. TS > 45% and SF > 300 μg/l, they are treated; if not they will be monitored to see if their iron level increases. If the iron levels increase they will be treated for HHC.

Figure 2.

Decision tree for the use of DNA testing in family members (+ve positive, −ve negative).

For the DNA test strategy, relatives have a DNA test. Those who are C282Y homozygous (positive DNA test) have biochemical tests. Those with raised iron levels, i.e. TS > 45% and SF > 300 μg/l,5,8 are treated. Those who do not have raised iron levels are monitored by repeat blood tests to see if their iron level increases. If the iron levels increase they will be treated for HHC and if their iron levels do not increase they will not be treated. Those with negative DNA test result will not be treated or have any further medical investigation for HHC. We assumed that DNA testing is 100% sensitive in families where the index case is homozygous.

For the purpose of the model, a typical patient is defined that is representative of all patients. Siblings will be assumed to be over 45 years of age because most new probands will be over 45 years of age at diagnosis. Those who are monitored will be retested once. Similarly a typical child will be aged 25 years. At this age, most will not have manifestations of iron overload but we assume that those who develop symptoms of iron overload will do so within 20 years and that the proportion of offspring with increased iron levels will rise linearly over this time period. The offspring will be tested every five years until iron overload is detected, i.e. a maximum of five times.

Parameters used in the analysis

The data used in the model has been collected from systematic reviews and systematic searches.3 Data sources were chosen for the model on the basis of appropriateness to the UK setting. Although some non-UK data were used in the model, this only occurred if no UK data were available and the data were considered to be the most appropriate alternative source, through consultation with clinical experts. Table 1 shows the key inputs to the decision tree models. In the literature, there were a range of thresholds used for the tests and a wide range of results in the accuracy of these tests. We used the thresholds suggested in clinical guidelines,5,8 i.e. TS >45% and SF > 300 μg/l. The sensitivity and specificity for TS was taken from Olynyk et al.10 as suggested by the clinical guidelines.8 The sensitivity and specificity of SF was taken from Moodie et al.11 as this UK study reported separately for different thresholds of SF for men and women. Liver biopsy provides a quantitative measurement of hepatic iron concentration on which diagnosis of HHC is based. It is used as the ‘gold standard’ for confirming diagnosis from SF and TS tests and so the model assumed that liver biopsy was 100% accurate in diagnosing HHC.

The decision tree for people suspected of having haemochromatosis required an estimate of the prevalence of true phenotypic HHC in people who are referred with symptoms of HHC. The data for prevalence in this population were scarce with only one relevant study found.11 This study of 427 patients referred for investigation of liver disease from an ethnically mixed population in south London11 was used to estimate the prevalence of HHC in a Caucasian population with suspected iron overload by only including those with northern European or Celtic origin. The prevalence of HHC in relatives is estimated using simple genetic theory. Although liver biopsy was associated with a small risk of death and other complications, for the base case, we assumed that there were no deaths or major events from liver biopsy as in other modelling studies.12 Vantyghem et al.13 described 156 subjects recruited in the Endocrinology and Metabolism Department of Lille University who were referred due to general symptoms of iron overload and abnormal iron levels (SF > 300 μg/l or TS > 45%). Amongst other data, they reported the numbers with high SF (>1000 μg/l) for those who were homozygous for the C282Y mutation which are used here.

Costs

Cost data were obtained from a number of primary and secondary sources (Table 1). The UK Genetic Testing Network (UKGTN) website (www.ukgtn.org) provides information on genetic testing in the UK and includes information on costing for DNA tests from different UK laboratories for different diseases and genes. We assumed patients’ consultations with a nurse or consultant would last 15 and 30 min, respectively, and patients would have one consultation with the nurse and consultant for the DNA test, liver biopsy or iron test. Treatment for individual patients varies according to clinical symptoms. We assumed that an average patient with clinical manifestation of haemochromatosis would require roughly 20 venesections to remove iron from the blood and then maintain iron levels over a 5 year period and they would be seen about seven times by a consultant during this time.14,15 For those with a false positive diagnosis of haemochromatosis, we assumed that this would be corrected after receiving a liver biopsy, as in current clinical practice.

Results

Main analysis

Table 2 shows the results of the diagnostic test decision tree for an average male of 45 years of age. The DNA strategy is cost saving compared to liver biopsy testing for all. The cost saved per case detected was £123 and the cost saved per case tested was £93.

View this table:
Table 2

Base case results for the diagnostic pathways decision tree

Per 100 people with positive testLiver biopsyCases detectedMonitorTotal cost (£)Cost saving/person tested (£)Cost saved/case detected (£)
Liver biopsy100.075.1083 068
DNA strategy41.175.122.973 82392.5123

Table 3 shows the results for male siblings of age 45 years and offspring of age 25 years for the family testing decision tree model. For siblings, the DNA strategy is more expensive than the biochemical strategy even though fewer people are monitored because the cost of monitoring is much cheaper than the cost of the DNA test. For offspring, the DNA test strategy is cheaper than the baseline biochemical testing strategy and there are a similar number of HHC cases detected in both strategies (Table 3). In the biochemical test strategy, there are many more people monitored than in the DNA strategy.

View this table:
Table 3

Base case results for the family testing decision tree

Per 100 people testedCases detectedCases treatedMonitorDNA testTotal cost (£)Cost saved/person tested (£)Cost saved/case detected (£)
Siblings
    Biochemical19.019.786.20.023 628
    DNA tests19.019.111.9100.027 423−37.95−200
Offspring
    Biochemical3.54.498.60.046 753
    DNA tests3.53.64.4100.018 638281.157982

Sensitivity analyses

The parameters in the diagnostic pathways decision tree were varied in a series of sensitivity analyses and the results are shown in Table 4. Where possible, the parameters were varied according to data from other studies identified in the systematic review, otherwise a suitable range was chosen. The sensitivity analyses show that the conclusions from the decision trees are robust across all reasonable parameter ranges. The results from the diagnostic decision tree were most sensitive to the proportion with a positive DNA test for the C282Y mutation, specificity of the TS test and the cost of the liver biopsy. For the siblings, the most sensitive parameters were the cost of the DNA test and the cost of monitoring. If the cost of DNA test were to reduce by 40% to £60, the DNA strategy would be the cheaper one. If those who were monitored were retested twice (instead of only once), the DNA strategy becomes cost saving (£79 per case detected). For offspring, the most sensitive parameters were the number of times the offspring are monitored and the penetrance of the disease. The decision tree models were also run for females, rather than males, with a threshold for SF of 200 μg/l8 and a penetrance of 32%.16 The results from the analyses for females were similar to those for males.

View this table:
Table 4

Sensitivity analyses

VariableBasecaseInputsCost saved per case detected (£)
LowHighLowaHighRange
Diagnostic pathways model
    Raised SF > 1000 DNA positive (%)0.390.220.5619057133
    TS specificity (£)94759922497127
    Costs liver biopsy test (£)388.05310.44465.6662184122
    Prevalence suspect, iron overload0.0380.0160.0616911158
    Costs DNA test (£)100801201509753
Family testing model
Siblings
    Cost DNA test (£)10080120−94−305211
    Cost monitoring (£)71.1556.9285.38−256−144112
    Penetrance (%)766093−253−16390
    TS specificity (%)947599−149−21162
    SF specificity (%)857599−189−21526
Offspring
    Number of monitoring537414411 7867642
    Penetrance (%)76609310 11165233588
    TS specificity (%)947599986774302437
    Cost DNA test (£)10080120855074151135
    SF specificity (%)857599840673651041
  • aResults in the ‘low’ and ‘high’ columns use inputs in corresponding columns.

A probabilistic sensitivity analysis17 was conducted to investigate the uncertainty of the model. The probability distributions were fitted to each of the model parameters using the high and low values from the sensitivity analysis and are shown in Appendix 1. The model used Monte Carlo simulation to randomly sample values for the model inputs and was run for 1000 iterations. In the PSA, the cost saved per case detected for the diagnostic pathways varied between £97 and £187 for the 25th and 75th percentiles, respectively. For the family testing decision tree, the cost saved per case detected varied between £7323 and £9458 for offspring for the 25th and 75th percentiles. For siblings, the extra cost per case detected varied between £145 and £234. More results are shown in Appendix 1.

Discussion

The analyses in this article show that in people suspected of having HHC, DNA testing is cost saving compared to testing using liver biopsy test. Furthermore, results suggest that using a diagnostic strategy which incorporates DNA testing is cost saving in case identification and testing of offspring of haemochromatosis patients. To our knowledge, this is the first attempt to demonstrate the cost effectiveness of these strategies in the UK. Current guidelines hypothesise that ‘the genetic test will probably reduce costs’ for the first analysis based on a cost of £20.5 The analyses in this article have used standard health economic methods to give the best estimate for the cost of these diagnostic strategies and provide rigorous evidence for their use in the current clinical recommendations.5

We have not considered the use of other techniques such as magnetic resonance imaging (MRI). Quantitative MRI can be used as an alternative to liver biopsy to assess the iron liver content but it is not widely available in all centres in Europe and North America and it cannot assess liver fibrosis with any accuracy. There is a significant difference between MRI that is widely available and the quantitative MRI assessment of iron deposition that requires additional equipment which is only maintained at specialist centres.

The analysis did not consider complications from liver biopsy, such as bleeding; this would result in even more favourable results for the DNA testing strategy. Gilmore et al.18 estimated a death rate of 0.13–0.33% and a bleeding rate requiring transfusion of 0.7% for people who have ultrasound guided liver biopsy. In the DNA testing strategy, there were less than half the number of liver biopsies performed than in the liver biopsy testing strategy and so there will be a similar reduction in liver biopsy complications.

The deterministic and probabilistic sensitivity analyses show that the conclusions were robust to a whole range of assumptions about the parameters used. One uncertain parameter is the prevalence of phenotypic HHC in people presenting with symptoms of haemochromatosis. The data used in the model were taken from a study of a hepatology clinic in South London and may not be exactly representative of those with symptoms of HHC. However, the cost saved per case detected varied little when the prevalence was changed and the DNA test strategy remains cost saving for all possible values of prevalence.

The decision tree model does not consider long-term cost and consequences of HHC because the data on the long-term cost and consequences were considered of poor quality and this would have necessitated making significant assumptions. Furthermore, the analyses in this project detected the same number of cases of HHC using both strategies and so a cost-minimalization model is more appropriate.

The decision tree models developed in this article have not included patients who were compound heterozygotes for C282Y and H63D because the penetrance of HHC in this disease associated genotype is relatively low and there were few data for this condition. Furthermore, the clinical treatment of patients with this condition is unclear in the guidelines. More research is needed for compound heterozygotes for C282Y and H63D.

Results suggest that using a diagnostic strategy which incorporates DNA testing is cost saving in case identification and testing of offspring of haemochromatosis patients. The DNA strategy monitored significantly fewer and treated fewer patients who did not have HHC compared to the biochemical strategy. With the biochemical testing strategy, most offspring of probands would have to undergo repeated testing until they reached 45 years of age. In the DNA test strategy, 95% of these offspring will avoid further unnecessary investigations and associated potential long-term uncertainty and anxiety.

The current analysis shows that DNA testing is not a cost saving strategy compared to testing using biochemical tests for screening siblings of a patient with HHC (due to DNA test, costs being higher than the reduced monitoring costs). However, if the cost of the DNA test were to reduce by 40% or if those who were monitored were retested twice (instead of only once) the DNA strategy would be the cheaper one. Also, the model does not estimate the likely inconvenience and anxiety attached to monitoring patients and treating patients without HHC, nor value the reassurance that can be given to those testing negative. As the DNA strategy monitored significantly fewer and treated fewer patients who did not have HHC compared to the biochemical strategy, this would impact on the long-term cost-effectiveness of DNA testing in siblings. As such, the authors suggest that the preferred strategy in practice would be using DNA testing for case identification and for both offspring and siblings of haemochromatosis patients as this can result in reduced anxiety due to definitive reassurance that they do not have the risk of developing HHC which cannot be given using the phenotypic strategy.

A strategy of not screening relatives of patients with HHC would require comparing the long-term costs and consequences of HHC and was not considered for the reasons discussed above. Previous studies have reported the appropriateness of screening first degree relatives of affected patients.19,20 El-Serag et al.19 calculated the cost of screening the family members of a proband that consisted of up to three siblings. Compared to no screening, the screening of family members was dominant for all screening strategies; that is they cost less and yielded greater benefit than no screening. They found that strategies using HFE gene testing were less costly than serum iron studies. In contrast, this article shows that iron studies are cheaper than gene testing. The differences between the results are due to difference in assumptions on the resource costs used and the inclusion of long-term costs in El-Serag et al.19 For example El-Serag et al. uses a lower cost for the DNA test and repeats the iron studies tests if the first result is abnormal.

The model developed here did not include a strategy for testing the spouse first before testing the offspring. There may be mild negative effects on participants with indeterminate results from screening programmes21 and, for this reason, spouses are not currently tested first before offspring, in the UK. However, Adams22 recommended this strategy to avoid unnecessary investigation of offspring of probands and El-Serag et al.19 found that testing the spouse before children was the most cost effective strategy when testing two or more children.

Conclusions

The current UK recommendations for the use of DNA testing for those with clinically suspected HHC rather than routine liver biopsy are confirmed to be cost effective. In families of homozygotes, DNA testing should be used first, together with iron monitoring in those who are gene positive.

Acknowledgements

This project was funded by the NIHR Health Technology Assessment Programme (project number 05/07/04). The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Department of Health. We are very grateful to the advisory panel which provided expert advice and comments on the protocol and/or draft of the systematic review. We would also like to thank staff at the Wessex Institute for Health Research and Development and to the two anonymous referees for their helpful comments and suggestions.

Conflict of interest: None declared.

Appendix 1

Probabilistic Sensitivity Analysis

View this table:
Table A1

Parameters used for the probabilistic sensitivity analysis

Parameter nameMeanDistributionAlphaBeta
Diagnostic pathway
    HHC prevalence0.038Beta11.0278.2
    TS sensitivity0.94Triangle0.750.99
    TS specificity0.94Triangle0.750.99
    SF sensitivity0.73Triangle0.50.99
    SF specificity0.85Triangle0.750.99
    HHC DNA YY0.913Triangle0.91
Proportion raised SF > 1000
    DNA positive0.39Beta11.918.7
    DNA negative0.24Beta26.082.4
Family testing, sibling
    Risk0.25Fixed
    Penetrance0.76Beta20.06.3
    SF > 200 SF sens0.73Triangle0.50.99
    SF > 200 SF spec0.7Triangle0.50.99
Family testing, child
    Risk0.05Fixed
    Initial prop with IO0.2Beta12.148.4
Costs
    DNA lab100Gamma96.041.04
    Nurse8.75Gamma96.040.09
    Consultant39Gamma96.040.40
    Liver biopsy388.05Gamma96.044.04
    Iron lab23.4Gamma96.040.24
    Venesection8.75Gamma96.040.09
    No. monitors child5Gamma24.010.20
    No. treatments20Gamma61.460.32
View this table:
Table A2

Results from the probabilistic sensitivity analysis

Cost per case detected (£)
Diagnostic testingFamily testing, siblingFamily testing, offspring
First quartile97−2347323
Second quartile139−1898311
Third quartile187−1459458
Fourth quartile6912318 196
Figure A1.

PSA results for cases detected and total costs saved for diagnostic testing decision tree.

Figure A2.

PSA results for cases detected and total costs saved for sibling family testing decision tree.

Figure A3.

PSA results for cases detected and total costs saved for offspring family testing decision tree.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
View Abstract