5 Cost considerations

5 Cost considerations

Cost evidence

5.1 The company identified 7 studies from the clinical evidence search which incorporated an economic analysis. However, it was unable to draw any firm conclusions from these studies. The external assessment centre did not identify any additional economic evidence and agreed with the company that these studies did not provide relevant information.

5.2 The company presented an economic model comparing Spectra Optia with manual exchange. It also included top-up transfusion as a comparator in the model, although this was not specified in the scope decision problem. The population was considered as 12 subgroups based on a mixture of age, clinical indication and degree of iron overload to represent the heterogeneous case‑mix of patients with sickle cell disease and their differing clinical needs and associated costs. The disease complications considered were secondary stroke, painful crises, acute chest syndrome and priapism. The population groups were:

  • Children at high risk of primary stroke, with and without iron overload (mild, moderate and severe).

  • Children having treatment to prevent complications of sickle cell disease, with and without iron overload.

  • Adults having treatment to prevent complications of sickle cell disease, with and without iron overload.

    The structure was a simple costing model which simulated the 'average' cost of chronic sickle cell disease treatment for 1 patient using 1 of 3 modalities: automated exchange with Spectra Optia, manual exchange or top-up transfusion. The time horizon of the economic model was 5 years.

    In the absence of published data relating specifically to the Spectra Optia system, the clinical parameters used in the company's model (table 1) were based on clinical expert opinion, extrapolations from the clinical evidence and data from UK registries and NHS audits. The cost of iron overload was the main driver for the cost of treatment in the company's model.

Table 1 Clinical parameters used in the company's model

Parameter

Spectra Optia

Manual exchange

Top-up transfusion

Patients with iron overload

0%

10%, 30% and 50% after 24 months, 36 months and 48 months

90% after 12 months

Patients able to cease chelation therapy

Mild iron overload: 50% and 100% after 12 months and 24 months

Moderate iron overload: 5%, 15%, 30% and 50% after 12 months, 24 months, 36 months and 48 months

Severe iron overload: 5%, 15% and 30% after 24 months, 36 months and 48 months

N/A, iron overload does not decrease in this arm

N/A, iron overload does not decrease in this arm

Yearly rate of hospital admissions

Children and adults (complications): 0.65

Children (stroke): 0.00

Children and adults (complications): 1.1

Children (stroke): 0.01

Children and adults (complications): 1.1

Children (stroke): 0.07

Procedure time

Adults: 110 minutes

Children: 86 minutes

245 minutes

Adults: 300 minutes

Children: 180 minutes

Procedures per year

8.5

12

13

Packed RBC units per procedure

Adults: 7

Children: 5

4

2

Staffing

1 grade 5

1.5 'highly qualified'

0.5 grade 5

Abbreviations: RBC, red blood cell; N/A, not applicable.

5.3 Parameters relating to adverse events and alloimmunisation were not included as they were considered to be similar for all modalities. The cost of secondary stroke was taken from Cherry et al. (2012) and set as a one‑off payment of £21,807 at 2.5 years (3 months of acute costs and 2.25 years of ongoing care costs); there were insufficient data to support differences in rates of primary stroke between transfusion modalities. The cost per hospital admission included was £1,354, taken from NHS reference costs. Chelation therapy costs were calculated from British national formulary values using average body masses and assumed to be £21,022 per patient per year for adults and £9,954 for children. The cost per packed red blood cell unit was £120 as listed in NHS reference prices. The Spectra Optia exchange set was the only consumable that was not common across all modalities, for which the company used its list price (£167.84).

5.4 The company did 8 univariate deterministic analyses for each of the 12 subgroups. These tested sensitivities to: stroke timing and severity, hospital admissions, cost of medication, staff grades, staff ratios, red blood cell units, procedure length and frequency, and the cost of consumables. Where a parameter change altered the ranking of modalities, threshold analyses were done to inform when the modality orderings changed. The values used for these analyses were informed by values taken from published clinical evidence, clinical advisers, and company and reference sources. The company also did 4 scenario sensitivity analyses: use of depletion exchange protocol with Spectra Optia, resulting in a reduction in the number of packed red blood cell units used in automated exchange by 1; mild iron overload with low chelation costs; severe iron overload with high chelation costs; and an increased rate of patients ceasing chelation therapy for moderate and severe iron overload when having automated exchange.

5.5 The results of the company's base case showed that the Spectra Optia system was always cost saving compared with manual exchange, with savings over 5 years ranging from £360 per adult patient with severe iron overload to £52,516 per adult patient with mild iron overload. The absolute costs of treating sickle cell disease in adult patients over the 5‑year time horizon varied from £48,093 for Spectra Optia in the absence of iron overload to £128,670 for manual exchange in those with iron overload.

5.6 The sensitivity analyses showed that the Spectra Optia system was sensitive to changes in procedural costs (in particular the need for packed red blood cells), and that top-up transfusion was sensitive to changes in chelation costs. Manual exchange had higher procedural costs than top-up transfusion (through staff time and grade, and greater need for red blood cell units) and higher chelation costs than Spectra Optia, and was rarely the lowest cost modality. Stroke and emergency hospital admissions had little impact in the sensitivity analyses except in some extreme threshold scenarios.

Additional work by the external assessment centre

5.7 The external assessment centre revised some input parameters of the model concerning capital, procedural and chelation costs to reflect values which it considered were more plausible. However, it judged that the limited clinical evidence (and the associated uncertainty) should remain a key consideration when interpreting the results.

5.8 The company's model did not include capital costs (£52,052) or maintenance costs (£4,572 per year) for Spectra Optia, although it did report separate analyses on these costs. The external assessment centre considered it wrong to exclude these costs from the model, and so created scenarios to reflect the mixed use functionality of the technology and relatively low demand (due to low prevalence of sickle cell disease) in some areas of the country. The external assessment centre determined that treating 30 sickle cell patients per year would use 50% capacity of each Spectra Optia system, and 15 patients per year would use 30% of capacity. The external assessment centre also included sensitivity analyses in which 100% of the device's capacity was used exclusively for treating sickle cell disease (with patients incurring the full cost) or where 50% of capacity was used elsewhere. In these analyses, the external assessment centre also extended the lifespan of the device from 5 years to 7 years.

5.9 The external assessment centre also noted that the cost of chelation therapy in the company's model did not include diagnostic and monitoring costs. An estimate of the chelation costs used in the model for the monitoring of iron overload was provided in confidence to the committee. The proportion of patients having chelation therapy after top-up transfusion in years 2 to 5 of the model was 90%. These patients entered the model with no iron overload. The external assessment centre revised this value to 75% based on 250 of 332 patients in the Haemoglobinopathy Registry Report having regular chelation therapy (Foster 2014).

5.10 The company had also assumed that adults and children having manual exchange procedures would both need 4 units of packed red blood cells. The external assessment centre reduced this to 3 units in children. Similarly, the procedure time for manual exchange in children was reduced from 254 minutes to 208 minutes. The number of staff needed per patient for manual exchange was also reduced from 1.5 to 1.0, following expert advice.

5.11 Results from the economic modelling showed that:

  • For 30 patients having automated exchange per year at 100% use of the device's capacity, Spectra Optia is cost saving compared with manual exchange in all patients with mild iron overload and in adults without iron overload.

  • For 30 patients having automated exchange per year at 50% use of the device's capacity, Spectra Optia is cost saving compared with manual exchange in all patients with no or only mild iron overload, and in adults with moderate iron overload.

  • For 15 patients having automated exchange per year at 100% use of the device's capacity, Spectra Optia is cost saving compared with manual exchange in all patients with mild iron overload.

  • For 15 patients having automated exchange per year at 50% use of the device's capacity, Spectra Optia is cost saving compared with manual exchange in all patients with mild iron overload, adults with no iron overload, and children with no iron overload having treatment for the secondary prevention of stroke only.

5.12 The committee asked the external assessment centre to make some additional changes to the model. In the company's model, 90% of patients were assumed to have iron overload when beginning treatment with Spectra Optia (and were therefore having chelation therapy) whereas only 80% had iron overload when beginning manual exchange. The company provided no rationale for this difference so the external assessment centre set chelation treatment rates at 90% for both Spectra Optia and manual exchange.

5.13 The external assessment centre decided that it would be most helpful if the cost modelling included a range of scenarios reflecting device‑sharing schemes. The external assessment centre included scenarios for 30 patients using the device at 100% capacity (and so 100% of its capital cost), 50% capacity and 0% capacity (for centres that already have the device but do not use it for red blood cell exchange). From the external assessment centre's subsequent changes to the model, the committee concluded that Spectra Optia is cost saving in the following scenarios:

  • 0% capital cost attributed to sickle cell disease: Spectra Optia is cost saving in all patients except when compared with top-up transfusion in patients with severe iron overload, and transfusion for preventing stroke in children with moderate iron overload.

  • 50% capital cost attributed to sickle cell disease: Spectra Optia is cost saving in all patients with no or mild iron overload, and in some patients with moderate and severe iron overload.

  • 100% capital cost attributed to sickle cell disease: Spectra Optia is cost saving in all patients with mild iron overload, in some patients with no iron overload, and in adults with moderate iron overload. These results are summarised in table 2 and table 3.

Table 2 External assessment centre's revisions to company model (automated compared with manual)

Scenario (capital costs attributed to SCD) and population

No overload

Mild overload

Moderate overload

Severe overload

0%: adults

-£25,011

-£96,512

-£24,874

-£10,867

0%: children (secondary prevention)

-£12,439

-£46,294

-£12,374

-£5,742

0%: children (primary prevention)

-£10,005

-£43,860

-£9,940

-£3,307

50%: adults

-£16,555

-£88,056

-£16,418

-£2,410

50%: children (secondary prevention)

-£3,983

-£37,838

-£3,918

£2,715

50%: children (primary prevention)

-£1,548

-£35,404

-£1,484

£5,149

100%: adults

-£8,099

-£79,600

-£7,961

£6,046

100%: children (secondary prevention)

£4,474

-£29,382

£4,539

£11,171

100%: children (primary prevention)

£6,908

-£26,947

£6,973

£13,605

Abbreviations: SCD, sickle cell disease. Negative values indicate cost savings with Spectra Optia.

Table 3 External assessment centre's revisions to company's model (automated compared with top-up)

Scenario (capital costs attributed to SCD) and population

No overload

Mild overload

Moderate overload

Severe overload

0%: adults

-£77,483

-£81,493

-£9,855

(£4,152)

0%: children (secondary prevention)

-£35,424

-£37,322

-£3,402

(£3,230)

0%: children (primary prevention)

-£31,681

-£33,580

£341

(£6,973)

50%: adults

-£69,027

-£73,037

-£1,399

(£12,609)

50%: children (secondary prevention)

-£26,967

-£28,866

£5,054

(£11,687)

50%: children (primary prevention)

-£23,224

-£25,123

£8,797

(£15,430)

100%: adults

-£60,571

-£64,581

£7,058

(£21,065)

100%: children (secondary prevention)

-£18,511

-£20,409

£13,511

(£20,143)

100%: children (primary prevention)

-£14,768

-£16,667

£17,253

(£23,886)

Abbreviations: SCD, sickle cell disease. Results in brackets are considered clinically unlikely. Negative values indicate cost savings with Spectra Optia.

5.14 The results show that Spectra Optia is cost saving compared with manual exchange and top-up transfusion in most patients with sickle cell disease. The highest cost savings are £96,512 (compared with manual exchange) and £81,493 (compared with top-up transfusion) per patient per year for adults with mild iron overload where an already functioning and purchased device can be used. The lowest cost savings are £1,484 (compared with manual exchange for primary prevention in children with moderate iron overload when the device has been procured at 50% of the capital cost) and £1,399 (compared with top-up transfusion in adults with moderate iron overload) per patient per year.

5.15 Spectra Optia is cost incurring compared with manual exchange for children with severe iron overload if the device is procured at 50% of the capital cost. If the device is purchased at 100% of the capital cost, it is cost incurring (at £2,715 to £13,605 per patient per year) compared with manual exchange for children with no or moderate iron overload and for all patients with severe iron overload. Using Spectra Optia is also cost incurring compared with top-up transfusion in some patients with moderate and severe iron overload. However, this comparator is considered to be a clinically poor treatment option for these patient groups and it is unlikely that it would be the preferred treatment option for people with iron overload.

Committee considerations

5.16 The committee considered that local providers should take into account the availability of existing devices when planning services for people with sickle cell disease; the potential for device‑sharing schemes was discussed and these were represented in the revised cost modelling carried out by the external assessment centre.

5.17 The committee noted that the levels of iron overload in patients included in the model were not clearly defined. The company's cost model stated that these were based on serum ferritin levels but no ranges were provided for the mild, moderate and severe categories used. The expert advisers informed the committee that there were no routinely used and agreed values for this classification.

5.18 The committee noted that top-up transfusion was not included as a comparator in the scope because it is generally used to treat anaemia and emergency crises in patients with sickle cell disease. Top-up transfusion is not suitable as a long‑term regime because it is iron positive, and so is not an appropriate current standard comparator for automated exchange. However, several clinical experts stated that top-up transfusion is sometimes used as a long‑term therapy in hospitals where provision of treatments for sickle cell disease is limited, or in patients who have disease complications that preclude the use of exchange transfusions (such as poor venous access).

5.19 The committee noted that many values in the cost model for Spectra Optia are based on estimates. In particular, it was not possible to define the different categories of iron overload used in the model in terms of serum ferritin levels (or any other kind of measurement of iron).

5.20 Having noted these uncertainties, the committee concluded that Spectra Optia would be cost saving for most patients compared with manual exchange and top-up transfusion.

  • National Institute for Health and Care Excellence (NICE)