K.2.3.1. High frequency chest wall oscillation (HFCWO) vests

According to the committee, HillRom is the most widely used vest in the UK. In light of this, HillRom was approached to provide accurate costing information on the vest. According to HillRom, the vest has an upfront capital cost of £6,995 (excluding VAT), this cost also includes a garment and an additional larger garment for a child as they grow. HillRom advised that the vest should last for at least 10 years before it needs to be replaced. Table 3 below presents the parameters used to calculate the equivalent annual cost.

In addition to the initial capital outlay, the committee advised that the vest would be serviced annually. Conversely, HillRom stated that the vest does not need annual servicing, but out of warranty repairs would incur a cost of £395 (excluding VAT). HillRom added that the number of repairs a device will require in a lifetime is impossible to say as they vary from never to once a year. However, 1 repair every 3 years would be reasonable assumption. Furthermore, additional garments would cost £295 (excluding VAT), but HillRom noted that most people would have enough garment provision in their purchase package to last 10 years.

Assuming a vest is purchased by a hospital for use across patients at the cystic fibrosis centre, a unit cost can be calculated based on the typical use of the vest over a period of time. As a result, the unit cost would depend on the usage of the vest, for example the more the vest is used the lower the cost per use.

However, current practice in the UK would be for the person with cystic fibrosis to purchase the vest themselves for home use. The reasons for this are to minimise the risk of cross-infection, reduce staff time and reduce the burden of clinic visits to use the vest by enabling the vest to be used as-and-when required.

K.2.3.2. Manual chest physiotherapy

Manual chest physiotherapy can include a variety of techniques such as chest shaking / vibrations or chest percussion. However, these techniques would not differ hugely in the time required to perform. According to NHS Reference Costs 2015/16, 1 manual chest physiotherapy session would cost £45 (WF01A, Non-Consultant-Led, Non-Admitted, Follow-up, 650). However, the number of sessions over a time frame would be individualised to the person with cystic fibrosis according to their severity and other treatment schedules.

K.2.3.3. Active cycle of airway breathing techniques (ACBT)

ACBT would require 1 initial appointment with a physiotherapist to show the person with cystic fibrosis how to perform the technique at a cost of £57 (NHS Reference Costs 2015/16: WF01B, Non-Consultant-Led, Non-Admitted, First, 650). Following this, ACBT could be replicated at home. Assuming the monitoring schedules across the techniques are similar, ACBT would cost less than the alternative techniques that require ongoing resources from staff time and equipment.

K.2.3.4. Oscillating and positive expiratory pressure (PEP) devices

The committee advised that oscillating devices and PEP devices are normally included in hospital equipment budgets and provided by cystic fibrosis centre. However, some people with cystic fibrosis may choose to replace their device themselves which, with regular use, would be every 2 to 5 years.

Table 4 below presents the cost of the most widely used oscillating and PEP devices available to the NHS, but other manufacturers, such as Astra, are available. In addition to the device, people may require accessories and replacement parts, for example the PARI O-PEP may require a nose clip and pressure hose to optimise the technique. The cost of PARI O-PEP accessories, are reproduced in Table 5 to provide an estimate of the total cost.

Assuming each new device requires a visit to a physiotherapist, to issue the device and teach them how to replicate the technique at home, the equivalent annual cost across a 5 year lifespan, ranges from £20.86 to £27.39, for oscillating devices and PARI O-PEP, respectively (Table 6). Further follow-up visits will be needed during the lifespan of the devices. However, it is assumed this is equivalent across the airway clearance techniques as there is no opportunity cost created by switching from one technique to another.

K.2.3.5. Non-invasive ventilation (NIV)

The cost of NIV depends upon the specification of the device and the consumables used. However, the committee noted that the most common device provided to people with cystic fibrosis in the UK is the NIPPY ventilator. The typical cost of a Nippy ventilator is £4,000 according to the NHS Supply Chain (Table 7). However, as previously stated, other manufacturers are available.

For illustrative purposes a cost £4,000 has been used to calculate the equivalent annual cost (£465). In addition to that initial capital outlay, the equipment also requires consumables such as a mask, or mouthpiece, and filters. According to members of the committee these would be replaced annually at a cost of approximately £100, leading to a total annual cost of £565 (Table 8).

The committee advised, in most cases, the cystic fibrosis centre would purchase NIV equipment and lend it to the person with cystic fibrosis for however long it was needed. They do not provide each person with their own personal device to keep over the equipment’s lifespan. Ideally, when NIV is required by someone with cystic fibrosis, it is initiated over a few days of coaching with a physiotherapist in an inpatient setting at a cost of £300/day (NHS Reference Costs 2015/16, DZ30Z, elective inpatient, chest physiotherapy attendance). The NIV equipment could then be used at home without assistance from a healthcare professional, but would be reviewed regularly at future attendances to the cystic fibrosis centre.

K.3.1.1. Moodie 2014

The trial by Wainwright 2011, included in the clinical evidence review, found no between-group difference for the 2 primary outcomes. Consequently, the original proposal for cost-effectiveness analysis (which addressed whether the incremental benefit of BAL-directed therapy was worth its incremental cost, measured against standard therapy) was no longer warranted.

Instead Moodie 2014 assessed the difference in costs between the 2 groups by using patient level data to ascertain whether BAL-directed therapy, rather than standard therapy, was still justified on the grounds of costs and whether BAL-directed therapy reduced treatment costs, by decreasing hospital days. Costs were valued in Australian dollars using a 2010 cost year. The costs included in the analysis were hospital admissions, BAL procedures, pharmaceutical costs, professional attendance, pathology tests and other procedures.

They found that the additional cost of BAL (A$11,880) was not offset by reductions in other health care expenditure. The mean total costs/child during the 5-year study period were A$92,860 in BAL-directed group and A$90,958 in standard group (mean difference [MD] A$1,902, 95% confidence interval [CI] −27,782 to 31,586, P = 0.90).

Moreover, there was no significant difference in mean hospital admission costs between the 2 groups. The mean hospital costs/child during the 5-year study period were A$ 57,302 in the BAL-directed group and A$66,590 in the standard group (MD A$-9,288; 95% CI −35,252 to 16,676, P =0.48).

It is important to note that both of these estimates are subject to wide Cis, reducing their credibility. However, considering that BAL-directed treatment offered no clinical advantage over standard therapy, BAL-directed therapy should not be recommended as a cost-effective monitoring strategy.

K.3.1.2. Etherington 2008

Etherington 2008 examined the cost and clinical impact of reducing the number of routine susceptibility tests conducted on isolates of P aeruginosa obtained from chronic infections in adults with cystic fibrosis.

This study was undertaken at a hospital in Leeds in 119 participants chronically infected with P aeruginosa. Their initial policy was to collect sputum samples at each clinic visit, every 8 weeks, and at the beginning and end of every course of intravenous (IV) antibiotics (routine therapy every 3 to 4 months). This is in accordance with the UK’s Cystic Fibrosis Trust recommendations that respiratory samples should be obtained every 4 to 8 weeks.

The application of a new protocol whereby isolates were only taken at the commencement of antibiotic therapy, there was evidence of clinical failure of therapy or routinely if not tested in the previous 3 months reduced the number of susceptibility tests by 56%.

No significant differences in median change of FEV₁%, forced vital capacity (FVC), C-reactive protein, white cell count, weight or duration of IV antibiotics were observed following the new protocol. However, the projected savings (cost year 2008) were €3,500 in consumables and 170 hours (costed at €6,500) of laboratory staff time per annum, a total annual saving of €10,000 (£6,500). However, the study does not report the number of times samples were taken when there was evidence of clinical failure of therapy. For this reason, the savings reported may be overestimated because it is unclear which participants failed treatment and followed an alternative treatment strategy that could incur additional assessments. Moreover, details regarding resource and cost use were not provided beyond the results stated above.

Overall, this study showed that the number of routine susceptibility tests conducted on P aeruginosa isolates can be reduced without adversely affecting clinical outcomes of IV antibiotic therapy. However, the relevance of this analysis is questionable given that the study compared the number of susceptibility tests of P aeruginosa, rather than monitoring techniques for pulmonary disease.

K.3.3.1. Non-invasive microbiological investigation

Non-invasive microbiological investigation techniques include an induced sputum sample, cough swab, throat swab and nasopharyngeal aspiration. Each of these investigations would be performed by a specialist nurse at the cystic fibrosis centre, with results available 2 to 3 days later.

According to NHS Reference Costs 2015/16 the average cost of directly accessed pathology services relating to microbiology is £8 (currency code DAPS07), whilst the cost of a nurses time is £33 for a 15 minute consultation (PSSRU 2016, Band 7, £130 per hour of patient contact) leading to a cost of £41 for a non-invasive microbiological investigation, excluding any subsequent visits for treatment.

K.3.3.2. Invasive microbiological investigation

BAL is an invasive microbiological investigation performed as an inpatient procedure. The results are not instantaneous and could take up to 5 days to process and report. The committee advised that BAL is performed in current UK clinical practice if the person with cystic fibrosis could not produce adequate quantities of mucus through coughing, as the procedure is invasive and costly. The cost of BAL is presented in Table 10.

K.3.3.3. Lung physiological function tests

Lung physiological function tests include spirometry, lung clearance index (LCI) and cardio-pulmonary exercise testing (CPEX); all of which, provide instantaneous results.

The committee advised that spirometry is regularly performed at the annual review, whereas LCI and CPEX are not. The committee added that CPEX is only performed in current UK clinical practice in people with established lung disease.

The cost of a spirometry is relatively inexpensive and could be performed by a specialist nurse at the cystic fibrosis centre. A procedure using spirometry is not provided in NHS Reference Costs 2015/16, but the chronic obstructive pulmonary disease (COPD) costing model for spirometry and pulmonary rehabilitation estimated a cost of £5.53 (inflated to 2015/16 prices) based on the work-up reproduced in Table 11.

However, committee communications with their cystic fibrosis centres found a much higher cost than that estimated by the Department of Health for COPD. The committee advised that the tariff for spirometry in people with cystic fibrosis (using a semi-portable machine as opposed to a hand-held device) is £42, but local Clinical Commissioning Groups (CCGs) could negotiate this price with their providers.

LCI and CPEX on the other hand, are more costly and would be performed by a pulmonary function technician who has expertise in performing and interpreting the tests (Table 12).

K.3.3.4. Imaging techniques

According to the committee, imaging techniques, including chest x-rays and CT scans, are used to monitor people with cystic fibrosis without clinical signs or symptoms, or those experiencing an acute exacerbation, but not in people with established lung disease, where management is unlikely to be influenced from the scans. The unit costs of those imaging techniques are presented in Table 13.

It is also important to note that generally, all children less than 4 years of age would require a general anaesthetic such as propofol (BNF November 2016: 0.5% emulsion for injection 20ml ampoules; 5 ampoules/£14.71) from an anaesthetist (NHS Reference Costs 2015/16: WF01B, 190, Non-Admitted Face to Face Attendance, First, Non-consultant led, £90) to produce good images of their lungs. The committee also noted that administering an anaesthetic can lead to longer than ideal waiting times which reduces the number of CT scans performed for a timely assessment.

K.4.3.1. Oral glucose tolerance test (OGTT)

The costs of an OGTT test comprises of the laboratory test costs for each blood sample, the costs of the glucose solution and the costs of staff time in administering the OGTT. Although practice will not be the same everywhere, it was assumed that as part of the test it would be necessary to provide some explanation of the test, obtain patient consent, prepare the glucose solution, take blood samples and inform the patient of the result. The blood tests are often taken by a healthcare assistant but a diabetic specialist nurse will often be responsible for explaining the test and providing them with the results.

After consultation with the committee, it was assumed that a 2-sample OGTT will take 30 minutes of a healthcare assistant’s time and 5 minutes of a nurse’s time.

The committee advised that someone with cystic fibrosis would be seen in a dedicated clinic room for the whole of their attendance at the clinic for an OGTT (2 hours). Moreover, when OGTTs are taken in children a band 6 nurse would fit them with a cannula, which could take up to 30 minutes. Following this, committee members communicated a cost of approximately £50 per OGTT from their hospital for adults with cystic fibrosis and a cost of approximately £70 for children, if a cannula is fitted. The breakdown of those costs are provided in Table 14.

The committee noted that a single-point OGTT would not be sufficient to monitor for the onset of CFRD as diet and lifestyle choice are variable within individuals. To obtain a more complete picture of blood glucose levels, the committee agreed that the OGTT should be repeated after a few days leading to a total cost of £98 for adults and £143 for children who require a cannula.

K.4.3.2. Continuous glucose monitoring (CGM)

The cost of CGM used to inform NICE NG17 (August 2015, Type 1 diabetes in adults: diagnosis and management) was based on the average of 3 of the main technologies available in the UK: Dexcom G4, Abbott Freestyle, and Medtronic RT Guardian. The items included in the estimation of the annual cost were the receiver, sensors, transmitters, and calibration (self-blood tests). NICE DG21 (February 2016, Integrated sensor-augmented pump therapy systems for managing blood glucose levels in type 1 diabetes) also provided a cost analysis of those 3 CGM technologies.

One noteworthy discrepancy between those analyses was the cost of a Dexcom G4 receiver (NICE NG17, £1,750; DG21, £750) and the exclusion of calibration from DG21. Details of the analyses by NICE NG17 and NICE DG21 are reproduced in Table 15 and Table 16, respectively. Despite this, the annual cost to provide CGM is approximately £3,500.

It is important to note that those costs in Table 15 and Table 16 are based on continuous use once diabetes has been diagnosed. When screening for CFRD, the committee advised that a person with cystic fibrosis would use a monitoring system for up to 1 week to clarify a diagnosis of CFRD. For this reason, the monitoring system would be shared across people at the clinic screened for CFRD; subsequently lowering the cost/person.

If the annual cost to provide a CGM system is approximately £3,500, the cost/person could be as low as £67 if one person with cystic fibrosis utilised one system a week. However, the committee added that administration to track the equipment (to facilitate returns and monitor lending history) would be required and this, combined with cleaning, could lead to a delay between uses of up to a week. As a result, the committee expected the equipment to be shared by up to 25 people with cystic fibrosis each year leading to a cost from £135/year to provide the CGM system.

Additional consultations to provide the CGM system and discuss the results would also be incurred (PSSRU 2016, Nurse advanced band 7, per hour of patient contact, £130) leading to a total cost of up to £200 if up to an additional 30 minutes of staff time is required.

It is important to note that for CGM to be viable, a centre would need access to several systems as to allow more than one person with cystic fibrosis to be monitored at any period. As a result, the implementation cost could be substantial if centres do not currently have access to CGM systems to monitor for the onset of CFRD.

K.5.3.1. Tests currently performed at annual review

As described in the clinical evidence review, a new definition of cystic fibrosis-related liver disease has come into practice using recommendations based on the tests performed at the annual review. These tests include an ultrasound scan, clinical examination (hepatomegaly, splenomegaly) and liver function blood tests (Table 17).

As can be seen from Table 17, ultrasound scans are more costly than liver function blood tests. Moreover, if ultrasound scans produce unclear images, the additional information they provide, may be reduced. This was highlighted by Mueller-Abt 2008, who reported that the highest kappa values (a statistical measure of agreement between observers) were obtained for nodularity, attenuation and spleen size (0.76–0.94). However, kappa values for hepatic homogeneity/coarseness were relatively low, indicating high variance in interpretation.

K.5.3.2. Transient elastography (FibroScan®)

Purchasing the FibroScan® is a capital cost, requiring an up-front payment. The National Horizon Screening Centre estimated the cost of FibroScan® to be £49,950 in 2008 prices (excluding VAT).

The unit cost of FibroScan® according to the Resource Impact Report for NICE NG50 (Cirrhosis in over 16s: assessment and management) is £164; comprising an ultrasound scan more than 20 minutes (£56; Health Resource Group, HRG code RA24Z) and a follow-up appointment as a hepatology outpatient (£108, HRG code WF01B).

For completeness, the equivalent annual cost (Table 18) has also been calculated, given that many centres may need an injection of resources to implement FibroScan® in their centre.

In addition to the initial capital outlay there are several other parameters to consider when estimating the unit cost (i.e. total cost/scan), these are presented in Table 19.

The cost/scan will vary depending on the usage of the machine, for example, the more the machine is used the lower the cost/person. The CF Trust Standard of Care 2011 advise a minimum of 75 adults or children to be managed by a specialist CF centre. However, a small number of ‘tertiary’ cystic fibrosis centres may manage 200 to 250 people with cystic fibrosis.

According to the NHCS paper, and members of the committee, FibroScan® can be performed by trained medical or paramedical staff within 15 minutes and would not require any additional disposables. Based on this, one scan could be performed by a specialist nurse within one surgery consultation at a cost of £33. Therefore, the total cost/scan in centre managing 150 people with cystic fibrosis would be £98. On the other hand, if a hepatologist (£108 HRG code WF01B) is believed to be more appropriate, a cost of £173 would closely reflect the cost reported by NG50.

It is important to note that FibroScan® could be used to diagnose liver fibrosis outside of people with cystic fibrosis, subsequently lowering the average cost/scan. However, more FibroScan® machines may be required, for supply to equal demand.

K.5.3.3. CT and MRI

Although CT and MRI scans are not invasive and are significantly cheaper than a liver biopsy, the committee noted that they are not routinely performed to detect liver disease. This was reflected in the clinical evidence review that found no evidence for MRI or CT scanning used as the reference standard.

In addition to the costs reported in Table 20, the committee advised that young children would require a general anaesthetic, such as propofol (BNF November 2016: 0.5% emulsion for injection 20ml ampoules; 5 ampoules/£14.71), from an anaesthetist (NHS Reference Costs 2015/16, WF01B, 190, Non-Admitted Face to Face Attendance, First, Non-consultant led, £90) to produce good images and this could lead to longer than ideal waiting times.

The committee stated that the images obtained from CT and MRI are not always clear-cut and may be subject to variance in interpretation, which questions their cost-effectiveness relative to liver function blood tests that are cheaper and objective.

K.5.3.4. Liver biopsy

Liver biopsies are generally performed under local anaesthesia and require a short hospital stay. The committee also advised that this invasive technique is associated with serious adverse events due to bleeding and other complications. For those reasons, liver biopsies should only be performed when the benefits outweigh the risks in terms of changing the disease outcome.

Unlike clinical examinations and imaging, the results from a biopsy are not instantaneous and could take up to 2 weeks. Moreover, the results could be inconclusive resulting in the need for a repeat biopsy. Some clinicians would take a “dual pass” to reduce this uncertainty, this was demonstrated by Lewindon 2011 who reported that a dual pass biopsy improved the detection of fibrosis in their trial. Their first pass detected liver fibrosis in 26 people with cystic fibrosis and the second detected liver fibrosis in another 5.

For these reasons, liver biopsies can have a negative impact on quality of life due to their invasiveness, potential scarring and delayed results which can cause anxiety and distress. Moreover, if complications or subsequent procedures are incurred, those costs would be greater. The cost of a single biopsy is reported in Table 21.

If biopsies do not provide additional diagnostic information to inform the patient’s management strategy, because the aetiology (secondary biliary cirrhosis) is generally, already known, it is clear that biopsies would be dominated (more expensive and less effective) by the current annual review as they would not produce any additional benefit to outweigh the expected cost and expected QALY losses associated with the procedure itself and potential adverse effects of the procedure.

However, the cost-effectiveness of a biopsy is less certain if there was reason to suggest a non-cystic fibrosis cause of liver disease as their management strategy may differ to that for secondary biliary cirrhosis.

K.6.3.1. Acetylcysteine

The recommended dosages reported in the BNF for acetylcysteine according to indication and age are presented below:

• Treatment of DIOS:

  • Child 1 month–2 years 0.4–3 g as a single dose;
  • Child 2–7 years 2–3 g as a single dose;
  • Child 7–18 years 4–6 g as a single dose.
• Prevention of DIOS:

  • Child 1 month–2 years 100–200 mg tds (3 times daily);
  • Child 2–12 years 200 mg tds;
  • Child 12–18 years 200–400 mg tds.

Table 22 presents the cost of acetylcysteine for the treatment of DIOS based the cheapest available manufacturer for tablet and capsule preparations. However, other forms are available from special-order manufacturers such as granules and oral solution.

K.6.3.2. Osmotic laxative containing polyethylene glycol

The recommended dosages reported in the BNF for lactulose to treat chronic constipation are presented below, according to age:

• Movicol-Paediatric®:

  • Child 2–5 years: 1 sachet daily, adjust dose to produce regular soft stools; maximum 4 sachets per day
  • Child 6–11 years: 2 sachets daily, adjust dose to produce regular soft stools; maximum 4 sachets per day.
• Macrogol:

  • Child 12–17 years: 1–3 sachets daily in divided doses usually for up to 2 weeks; maintenance 1–2 sachets daily
  • Adult: 1–3 sachets daily in divided doses usually for up to 2 weeks; maintenance 1–2 sachets daily.

Based on those dosages, Table 23 presents the cost of Movicol-Paediatric® oral powder and Macrogol oral powder compound for children and adult, respectively.

K.6.3.3. Sodium meglumine diatrizoate (Gastrogafin®)

The cost of sodium meglumine diatrizoate (Gastrografin®) is reported in Table 24 for an indication of DIOS in children with cystic fibrosis.

K.6.3.4. Lactulose

The recommended dosages reported in the BNF for lactulose to treat constipation according to age are presented below:

• Child 1–11 months: 2.5 mL bd, adjusted according to response;
• Child 1–4 years: 2.5–10 mL bd, adjusted according to response;
• Child 5–17 years: 5–20 mL bd, adjusted according to response;
• Adult: Initially 15 mL bd, adjusted according to response.

Table 25 summaries those dosages in to 2 categories, to illustrate the plausible costs.

K.6.3.5. Phosphates enema

Enemas are relatively more invasive than oral preparations requiring application to the rectum. The committee noted that enemas are often administered under radiological supervision to ensure the required site is reached, at a cost of £89 (NHS Reference Costs 2015/16, WF01A, Non-Admitted Non-Face to Face Attendance, Non-consultant led, Follow-up, Colorectal Surgery, 104).

Table 26 presents the acquisition cost of phosphates enema according to the dosages reported in the BNF:

• Child 3–6 years: 45–65 ml once daily;
• Child 7–11 years: 65–100 ml once daily;
• Child 12–17 years: 100–128 ml once daily;
• Adult: 128ml daily.

K.6.3.6. Stimulant laxatives

Stimulant laxatives include bisacodyl, sodium picosulfate, and members of the anthraquinone group, senna and dantron, but the committee regarded sodium picosulfate and senna to be the most common laxatives prescribed to people with cystic fibrosis to treat DIOS. For this reason only senna and sodium picosulfate are presented in Table 27 based on the following dosages reported in the BNF for constipation:

• Senna:

  • Child 2–3 years: 3.75–15 mg once daily, adjusted according to response;
  • Child 4–5 years: 3.75–30 mg once daily, adjusted according to response;
  • Child 6–17 years: 7.5–30 mg once daily, adjusted according to response;
  • Adult: 7.5–15 mg daily (max. per dose 30 mg daily), dose usually taken at bedtime; initial dose should be low then gradually increased, higher doses may be prescribed under medical supervision.
• Sodium picosulfate:

  • Child 1 month–3 years: 2.5–10 mg once daily, adjusted according to response;
  • Child 4–17 years: 2.5–20 mg once daily, adjusted according to response;
  • Adult: 5–10 mg once daily, dose to be taken at bedtime.

K.6.3.7. Surgery (distal ileal resection)

The committee advised that surgery is currently limited to very specific cases and only when first and second line treatments have failed. The cost of an elective inpatient procedure according to NHS Reference Costs 2015/16 is presented in Table 28 for the possible complexities.

K.7.3.1. Anti-acid drugs

The committee considered the addition of acid suppression such as H2 receptor antagonists (ranitidine or cimetidine) or proton pump inhibitors (omeprazole) in those with persistent symptoms of malabsorption. Table 29 presents the cost of anti-acid drugs over the course of 1 week and a typical monthly cost of continued use based on the following dosages reported by the BNF:

• Omeprazole: 20 mg once daily increased to 40 mg once daily if necessary, child over 1 year can receive up to 40 mg once daily;
• Rantidine: 150 mg bd (morning and night) or 300 mg at night, child 3–12 years can receive up to 150mg bd;
• Cimetidine: 400 mg bd (morning and night) or 800 mg at night; when necessary the dose may be increased to a maximum of 400 mg qds, child 1–12 years can receive up to 400 mg qds.

The preparation of anti-acid drugs received can impact the cost of the therapy, for example, oral solutions and effervescent (rapid dissolving) tablets, are more costly than standard tablets and capsules. As can be seen from Table 29, anti-acids provided over a short time-period can be relatively inexpensive if the lowest cost preparation is chosen.

However, the clinical review does not compare the efficacy according to the preparation; hence, the preparation received must be informed by clinical judgment. Moreover, the dose will depend on response, age and weight.

K.7.3.2. High-dose versus low dose PERT

Table 30 presents the cost of PERT over the course of 1 week and a typical monthly cost of continued use. Cost data were taken from the BNF November 2016, when dose ranges were reported the mid-point was taken for costing purposes. The clinical evidence suggested that there was a dose response, but in clinical practice the optimal dose may depend on the size of the person and how well the drug mixes with food. As a result, it is important to consider the concentration of enzymes, also reported in Table 30.

Taking PERT can be burdensome if it is taken before every meal or snack. As a result, choosing a PERT regimen that will have less of an impact lifestyle may increase adherence and effectiveness. However, this is not necessarily the cheapest option or the most effective.

As can be seen in Table 30 Creon® Micro granules and Creon® 10,000 capsules are the cheapest preparation of low-dose PERT at a cost of £14.36 per month and £17.69 per month, respectively. The cost of high-dose PERT is greater than Creon® Mirco granules and Creon® 10,000 capsules, ranging from £38.65 to £56.99 per month, but without knowing the benefits of high-dose PERT for certain we cannot know if high-dose is cost-effective compared to low-dose.

K.8.3.1. Tube feeding

Tube feeding can be used as an adjunct to oral feeding, or if there is clinical concern about the safety of swallowing they can replace oral feeding. Long term interventions to optimise nutritional status include gastrostomy or jejunostomy tube feeding, whereas nasogastric tube feeding would be used on a shorter term basis. The former are surgical procedures associated with a high cost, whereas the latter can be performed by a nurse as an outpatient procedure. However, there are specific clinical implications for long term nasogastric tube placement that mean they are not the preferred route of enteral feeding beyond short term use.

The costs associated with long-term nutritional supplementation via gastrostomy or nasogastric tube feeding, are outside the scope of NHS Reference Costs and should remain within primary medical services (Department of Health, Reference Costs Guidance 2015–16). For this reason, currency codes related to endoscopic insertions from NHS Reference Costs 2015/16 are presented in Table 31 as a proxy. With regards to nasogastric tube feeding, costs were reported solely for babies under special care (XA03Z); these were considered irrelevant to this review and are not reported.

The randomised study by Corry 2008 was identified as a relevant source of costing data on tube feeding through ad-hoc searches. This study was included in the Cochrane review on tube feeding for adults with swallowing disturbances. However, it is important to note that Corry 2008 was based on patients with head and neck cancer who required enteral feeding. These costs may not be generalisable to people with cystic fibrosis. They stated that the insertion costs are significantly different as nasogastric tubes are inserted by nursing staff as an outpatient attendance, including the cost of chest X-ray. Whereas, percutaneous endoscopic gastrostomy tubes are inserted by surgeons in theatre. Table 32 below reports the costs by Corry 2008 alongside inflated sterling prices.

In addition to the procedure, the committee advised that some people with cystic fibrosis would undergo an intense monitoring schedule during the first few days or weeks with a specialist. Thereafter, they would be monitored on a similar frequency to those receiving oral supplementation or appetite stimulants, with gastrostomy or jejunostomy incurring 1 additional visit with their surgeon each year at a cost of £132 (NHS Reference Costs 2015/16, 301, Consultant-Led, Non-Admitted Face to Face attendance, Follow-Up, Gastroenterology, WF01A).

In addition to the monetary cost of tube feeding, the committee advised that some qualitative reviews show tube feeding can negatively impact a person’s quality of life by affecting social interactions at meal times. Moreover, if the procedure and use of tube feeding is associated with adverse effects, they can incur further treatment costs and decrements in quality of life.

Nasogastric tubes frequently fall out and require the cost of a healthcare professional to reapply the tube if the family or carer were unable to do so. The committee added that there are a number of clinical concerns to their long term usage. Equally, gastrostomy and jejunostomy tubes need routine and, on occasion, emergency replacement which may need professional, rather than family, intervention. It was noted that tube feeding, when used appropriately, positively impacts on clinical wellbeing and health, improving quality of life and so justifying the high costs tube feeding can entail in those cases.

Compared to usual care, 1 cohort study (Bradley 2012) included in the clinical evidence review found a clinically significant difference in the indices of nutrition and growth for weight and BMI. However, there was no clinically significant difference in height, or FEV₁%, between the group receiving gastrostomy and those who received usual care at 6 month and 1 year follow-up. Based on this, gastrostomy could be considered cost-effective, when the aim is to improve weight and BMI, as improvements may reduce later downstream costs from hospital attendances related to malnutrition. However, gastrostomy is unlikely to be considered cost-effective when the aim is to improve lung function as the same benefits can be obtained from usual care which is a lot less costly and not subject to the decrements in quality of life previously outlined.

K.8.3.2. Oral supplementation

Table 33 reports the cost of high-energy feed supplements for fat, carbohydrate and protein and other nutritional supplements that may be considered as first line supplements for people with cystic fibrosis.

The cost of supplements undoubtedly varies by brand, but each brand can provide a different make up of nutrients. Consequently, the cheapest brand may not be suitable for all deficiencies; hence, supplements should be individualised to the person with cystic fibrosis.

The cost of supplements will also depend on the frequency those supplements are administered. If they were used to substitute rather than complement diet at home, the cost could be substantial. However, the person’s diet should be reviewed and modified prior to consideration of supplements and the healthcare professional should determine the appropriate quantity and frequency.

Compared to usual care, the clinical evidence review found no clinically significant difference for any indices of nutrition and growth. Moreover, 1 of 2 RCTs (Hanning 1993, Poustie 2006) showed a clinically significant decrease in FEV₁% in the group of participants receiving oral calorie supplementation compared to the participants in the control group receiving usual care at 3 month follow-up, but no significant difference was found by that study at 1-year follow-up. Based on this evidence the benefits of oral supplementation are unlikely to justify the costs in most cases.

Compared to nutritional advice, the clinical evidence review found 1 study (Kalnins 2005) that showed no clinically significant difference for any indices of nutrition and growth. However, the same study showed a clinically significant decrease in FEV₁% in the group of participants receiving oral calorie supplementation for 3 months compared to the participants receiving nutritional advice at 3 month follow-up and at 6 month follow-up.

K.8.3.3. Appetite stimulants

The acquisition cost of appetite stimulants are reported in Table 34 over a course of 1 day and 1 month of continued use. It is evident from Table 34 that cyproheptadine hydrochloride is half the cost of megace; therefore, if there is no evidence to suggest cyproheptadine hydrochloride is any less effective, cyproheptadine hydrochloride should be offered instead of megace if stimulants can improve quality of life and reduce the downstream costs associated with a reduced appetite.

Compared to placebo, 3 RCTs (Homnick 2004, Eubanks 2002, Marchand 2000) showed a clinically significant benefit in the indices of nutrition and growth. However, none of the RCTs found a clinically significant difference in FEV₁%, the number of exacerbations or adverse events (constipation). Therefore, if improvements in the indices of nutrition and growth are considered to be of greater importance than improvements in lung function or the number of exacerbations, appetite stimulants could be considered cost-effective.

K.8.3.4. Psychological and behavioural interventions

According to NHS Reference Costs 2015/16 the cost of a psychotherapy attendance is £158 (WF01A, 713, Consultant-Led, Non-Admitted Face to Face Attendance, Follow-Up). The cost of psychotherapy will ultimately depend on the number of sessions the person requires. The committee advised that sessions would be performed intensively for the first few months on a weekly basis, with further follow up sessions as required. Due to the difficulties of accessing these service promptly on the NHS, teaching psychological strategies to families to support the person with nutritional difficulties after the course of intensive healthcare professional input would be advantageous. However, the committee noted that not all families could engage with this.

Compared to usual care, the clinical evidence review found 1 RCT (Stark 1996) that showed no clinically significant difference for any indices of nutrition and growth or lung function. Based on these findings, the cost of psychological or behavioural interventions cannot be justified if usual care can provide the same benefits at a much cheaper cost. However, this is 1 small RCT with 9 participants that may not be representative of the population with cystic fibrosis in the UK today.

K.9.1.1. Dornase alfa

Menzin 1996 presented a cost-benefit analysis for daily dornase alfa versus placebo, based on clinical evidence in the US from Fuchs 1994. They suggested that dornase alfa therapy may reduce the cost of respiratory tract infection-related care. However, the cost of dornase alfa was not included in their analysis so we cannot know if the cost of dornase alfa treatment is completely offset from those savings.

Christopher 1999, also informed by Fuchs 1994, undertook a cost-effectiveness analysis for daily dornase alfa versus placebo. They found that the cost/life year gained in the subgroup of participants with initial FEV₁ <70% was a lot less (£16,000) than the whole population (£52,500). However, they noted that the short term evidence used to inform their analysis questioned the credibility of their findings.

McIntyre 1999 presented a cost-benefit analysis for daily dornase alfa versus placebo based on assumptions for disease progression, survival and cost savings from reduced respiratory tract infection-related care inferred from several clinical trials. In the best (worst) case scenario dornase alfa might maintain a person with cystic fibrosis in a mild state of disease for an additional 4 years and extend their life by up to 7 (2) years at a cost of £6,084 (£45,234) per life year gained. Suri 2002 undertook a clinical crossover trial to compare the total health service costs of daily dornase alfa with alternate day dornase alfa and hypertonic saline. They found that daily dornase alfa was more effective than hypertonic saline, but significantly increased health care costs. Moreover, administering dornase alfa on alternate days, rather than daily, was as effective, with a potential for cost savings.

Grieve 2003 performed a cost-effectiveness analysis based on the 12 week trial by Suri 2002. Benefits were based on the change in effectiveness measured by FEV₁% which was similar between daily dornase alfa (14% change) and alternate day dornase alfa (12% change) and absent from hypertonic saline (0% change). Using this data they estimated the cost/ 1% gain in FEV₁% and presented cost-effectiveness acceptability curves (CEACs) and net benefit statistics for each comparison. They showed that unless decision makers are willing to pay over £200/ 1% gain in FEV₁%, the probability of daily dornase alfa proving more cost-effective than alternate day dornase alfa was less than 50%. Moreover, the mean net benefit for the daily regimen compared to alternate days at that threshold was negative. As a result, they concluded that it may be more cost-effective if dornase alfa was prescribed on an alternate day basis rather than a daily basis.

The findings from those 5 studies suggest that daily dornase alfa is more costly and more effective than no dornase alfa and hypertonic saline. Moreover, if the effectiveness of daily and alternate day dornase alfa is similar, and alternate day dornase alfa is less expensive than daily dornase alfa, alternate day dornase alfa would be considered cost-effective compared to daily dornase alfa. However, those 5 studies do not report cost per QALY ICERs which makes their interpretation relative to NICE’s threshold difficult.

It is important to note that all 5 economic evaluations were subject to short trial durations. Therefore, if the effectiveness of dornase alfa is expected to decrease over time, the cost savings relative to no dornase alfa or hypertonic saline reported by the studies may be overestimated when longer time horizons are considered.

K.9.1.2. Mannitol dry powder for inhalation (Bronchitol®)

The sixth and final study included in this review was the cost-effectiveness evidence for a NICE Health Technology Appraisal (HTA, NICE TA266) submitted by Pharmaxis, who intended to make the following 2 comparisons:

1. mannitol versus BSC;
2. mannitol + rhDNase (dornase alfa) versus BSC + rhDNase (dornase alfa).

The cost-utility analysis developed for the HTA was a patient-level simulation Markov model. Here, the progression of each individual patient is modelled, rather than the progression of a whole patient cohort at once. As a result, individual patient characteristics are taken into account when determining the transition probabilities, hence the path through the tree. A schematic presentation of the relationship between treatment (black), time (black), clinical endpoints (grey) and economic endpoints (blue) is show in Figure 2. A green arrow indicates a positive relationship and a red arrow indicates a negative relationship.

Figure 2Relationships in the model, reproduced from the submission

Source: Manufacturer’s submission for TA266, February 2011, Figure 15

The effect of mannitol was assumed to be the same in rhDNase users and non-users. The only difference in the model between the 2 populations relates to the cost of rhDNase. The manufacturer concluded that mannitol treatment for people with cystic fibrosis in addition to BSC and regardless of rhDNase use was effective. There was no clear statement regarding the cost-effectiveness, but the ICERs (cost per QALY gained) for each comparison were similar and above NICE’s threshold for cost-effectiveness:

1. mannitol versus BSC: £41,074;
2. mannitol + rhDNase versus BSC + rhDNase: £47,095.

The Technology Assessment Group (TAG, Riemsma 2011) questioned the assumption that mannitol use was completely independent of rhDNase use. This means that any benefit of mannitol did not depend on whether a patient used, or did not use, rhDNase.

As a result, the TAG suggested during clarification that the economic results should be reported separately for the 2 comparisons according to rhDNase suitability, where the rate ratios (RRs) for exacerbations are population specific (according to rhDNase suitability). The manufacturer subsequently provided the following results to the TAG:

1. mannitol + rhDNase versus rhDNase + BSC for people with cystic fibrosis using rhDNase (i.e. mannitol as add-on therapy): £74,140;
2. mannitol versus BSC for people with cystic fibrosis who are ineligible, intolerant or inadequately responsive to rhDNase (i.e. mannitol as second-line therapy): £19,828.

Based on the manufacturer’s initial submission, the findings of the TAG, and manufacturer’s response to clarification, the TAG presented an alternative base case with the following amendments:

• the cost-effectiveness of mannitol is analysed separately for the 2 populations/comparisons according to rhDNase suitability where RRs for exacerbations are population specific;
• treatment independent and improvement specific values for costs and utilities are used to value health states.

These amendments led to the following estimates of cost-effectiveness for the 2 populations:

1. mannitol + rhDNase versus rhDNase + BSC for people with cystic fibrosis using rhDNase (i.e. mannitol as add-on therapy): £80,098;
2. mannitol versus BSC for people with cystic fibrosis who are ineligible, intolerant or inadequately responsive to rhDNase (i.e. mannitol as second-line therapy): £29,883.

In the manufacturer’s base case analysis, mannitol is almost equally cost-effective for both comparisons. However, under the TAG’s scenario, mannitol would only be considered cost-effective in people with cystic fibrosis who are ineligible, intolerant or inadequately responsive to rhDNase (i.e. mannitol as second-line therapy).

Box 1 below presents the Appraisal Committee’s discussion of cost-effectiveness based on the manufacturer’s submission and TAG’s report.

Following the discussion of the evidence, the Appraisal committee explored if there was a group of adults with cystic fibrosis in which treatment with mannitol would provide a cost-effective use of NHS resources. They believed that if mannitol treatment was offered only to people with cystic fibrosis with a rapid decline in lung function, the ICER would most likely be lower than in the whole population because of this group’s lower quality of life and lung function, and a greater potential to improve. The committee concluded that the ICER for mannitol in people with cystic fibrosis for whom hypertonic saline is not considered appropriate, who cannot use rhDNase because of ineligibility, intolerance or inadequate response to rhDNase, and whose lung function is rapidly declining would be under £30,000 per QALY gained; hence, they concluded that mannitol should be recommended as an acceptable use of NHS resources for those people with cystic fibrosis.

K.9.3.1. Mannitol dry powder for inhalation (Bronchitol®)

Table 37 reproduces the manufacturer’s expected unit costs of Bronchitol®, whilst Table 38 provides the latest acquisition cost from the BNF November 2016.

K.9.3.2. Dornase alfa

Table 39 below presents the acquisition cost (NHS Electronic Drug Tariff November 2016) of dornase alfa, alongside the cost/day and cost/year. The dose to calculate those costs was taken from the BNF for an indication to manage people with cystic fibrosis with FVC >40% of predicted to improve pulmonary function. Dornase alfa would also incur the cost of a nebuliser (Table 36) if the person with cystic fibrosis did not already acquire one.

K.9.3.3. Nebulised hypertonic sodium chloride

Table 40 below present the acquisition cost of nebulised sodium chloride for each available concentration reported in the NHS Electronic Drug Tariff November 2016. The cost/day and cost/year are also presented based on BNF recommendations for an indication to mobilise lower respiratory tract secretions in mucous consolidation (e.g. cystic fibrosis). Nebulised sodium chloride would also incur the cost of a nebuliser if the person with cystic fibrosis did not already acquire one (Table 36).

K.9.3.4. Acetylcysteine

The BNF does not report an indication or dose for acetylcysteine relevant to cystic fibrosis and the committee noted that acetylcysteine is generally no longer used as a mucolytic agent in cystic fibrosis. Where acetylcysteine has been used, the committee advised:

• Adults, 20% (200mg/mL) injection solution diluted (50–100mg/mL) and nebulised bd to qds;
• Children, 3 to 5mL of 20% (200mg/mL) injection solution diluted to 100mg/mL and nebulised bd to qds.

Table 41 below present the unit cost of acetylcysteine (NHS Electronic Drug Tariff November 2016) and the cost/day and cost/year based on 1 ampoule per dose, on the assumption that ampoules cannot be carried over. Acetylcysteine would also incur the cost of a nebuliser (Table 36) if the person with cystic fibrosis did not already acquire one.

K.10.3.1. Combination antibiotics

The committee advised the following dosages of co-amoxiclav and co-trimoxazole to prevent pulmonary bacterial colonisation with S aureus in people with cystic fibrosis:

• As co-amoxiclav

  • Child 1 month – 1 year: 0.5mL/kg tds of 125/31mg/5mL suspension;
  • Child 1 – 6 years: 5mL tds of 250/62mg/5mL suspension;
  • Child 6 – 12 years: 10mL tds of 250/62mg/5mL suspension or 625mg tablet 3 times a day;
  • Child >12 years and adult: 625mg tds.
• As Augmentin Duo® preparation (400/57mg/5mL)

  • Child 2 month – 2 years: 0.3mL/kg bd;
  • Child 2–6 years (13 – 21kg): 5mL bd;
  • Child 7 – 12 years (22 – 40kg): 10mL bd;
  • Child >12 years (>40kg): 10mL tds.
• Co-trimoxazole

  • Child 6 weeks – 5 months: 120mg bd;
  • Child 6 months – 5 years: 240mg bd;
  • Child 6 – 12 years: 480mg bd;
  • Child > 12 years and adult: 960mg bd.

Table 42 below presents the cost of 2 combination antibiotics: co-amoxiclav and cotrimoxazole.

For adults, the cost of trimoxazole would be £0.46/day (960mg tablet ×2) for a tablet preparation, or £3.98 for a sugar free oral solution (40ml). Co-amoxiclav would be cheaper at a cost £0.28/day for tablets (625mg tablet ×3) and £2.48/day for oral solution (10ml tds).

K.10.3.2. Beta-lactam antibiotics

The recommended dosages reported in the BNF for beta-lactam antibiotics according to indication and age are:

• Tetracyline for susceptible infections:

  • Child 12–17 years: 250 mg qds, increased if necessary to 500 mg tds to qds, increased dose used in severe infections.
  • Adult: 250 mg qds, increased if necessary to 500 mg tds to qds, increased dose used in severe infections.
• Cefradine for prevention of S aureus lung infection in cystic fibrosis

  • Child 7–17 years: 2 g bd.
• Ceflexin for susceptible infections:

  • Child 1–4 years: 12.5 mg/kg bd, alternatively 125 mg tds;
  • Child 5–11 years: 12.5 mg/kg bd, alternatively 250 mg tds;
  • Child 12–17 years: 500 mg bd to tds;
  • Adult: 250 mg every 6 hours, alternatively 500 mg every 8–12 hours; increased to 1–1.5 g every 6–8 hours, increased dose to be used for severe infections.

Table 43 below presents the cost of 3 beta-lactam antibiotics specified in the protocol: tetracycline, cefradrine and cephalexin. Based on the dosages above, for adults, the cost of tetracycline would be £0.28/day (250mg tablet ×4), cefradine £1.08/day (500mg capsule ×8) and cefalexin £0.14/day if the cheapest preparation is chosen (500mg capsules ×2) or £0.46/day if the more expensive oral solution is chosen (20ml).

K.10.3.3. Flucloxacillin

The BNF advised the following indications and dosages of flucloxacillin for people with cystic fibrosis:

• Prevention of S aureus lung infection in cystic fibrosis—primary prevention

  • Child 1 month–3 years: 125 mg bd.
• Prevention of S aureus lung infection in cystic fibrosis—secondary prevention

  • Child and adult: 50 mg/kg bd (max. per dose 1 g bd).

From Table 44 it is evident that capsules are cheaper than oral solutions, and non-sugar free solutions are cheaper than sugar free. For adults receiving 1g bd, the cost would range from £10.59 (40ml, 250mg/5ml oral solution sugar free) to £0.31 (500mg capsule ×4).

K.10.3.4. Azithromycin

Azithromycin can be used to prevent pulmonary bacterial colonisation with S aureus, and also, as an immunomodulator.

The committee advised the following for children over 6 years and adults for this indication:

• <40kg: 250mg 3 times a week;
• >40kg: 500mg 3 times a week (250mg daily may also be used if necessary).

It is evident that from Table 45 that tablets are cheaper than capsules. For adults the cost would be £1.32/week for a tablet preparation and £15.13 for a capsule preparation.

K.11.1.1. Wolter 1998

Wolter 1998 conducted a cost–consequence analysis in Australia based on an RCT that included 17 adult participants. The perspective of the study was not clearly stated, but authors appear to include costs incurred by the hospital and costs incurred by people with cystic fibrosis and their families. Costs were valued in Australian dollars (A$) at 1992 prices. Hospital-based treatment costs were calculated using inpatient costs from the Prince Charles Hospital and from projected diagnostic-related group reimbursement figures. Home-care IV therapy costs were calculated based on hospital acquisition costs and consumption of resource. Staff costs spent on education and home visits were calculated from hourly wages. Travel costs were determined according to a standard allowance per km. Other patient and family costs were determined by interview. Mean total costs included the costs of home physiotherapy, home visits, training, equipment, drugs and bed occupancy.

The authors concluded that home-care IV therapy was considerably less expensive for families than hospitalisation per day of hospitalisation (A$15.08 versus A$23.77). Moreover, the estimated cost saving for managing exacerbations at home compared with hospital was estimated to be A$2,552. These estimates should be approached with some caution owing to the small sample size within the study. It is also unclear whether or not these findings would hold in a UK setting today.

K.11.1.2. Thornton 2005

Thornton 2005 performed a cost-effectiveness analysis based on data collected retrospectively from 116 adults with cystic fibrosis at the Manchester cystic fibrosis unit over 1 year. Participants were categorised as belonging to the ‘home’ or ‘hospital’ group if they received >60% of treatment at home or in hospital, respectively. The study was conducted from a NHS non-societal perspective and costs were valued in UK pounds sterling at 2002 prices.

Unit costs were calculated from the NHS Trust, the cystic fibrosis unit’s budget, the BNF and the hospital-supplied catalogue. Resource use and costs were estimated for IV antibiotics, disposable equipment, home kits, sputum microbiology and sensitivity and blood drug level assays. The time spent with each person with cystic fibrosis was estimated using a time sheet completed by each staff member attending to the patient. Staff costs were obtained from the CF Unit budget. Clinical records were used to determine the number of days people with cystic fibrosis spent in hospital relating to IV antibiotic treatment. Fixed costs for the ward and outpatient clinic were calculated from the cystic fibrosis unit’s budget; these were used to estimate a fixed cost / hour related to an inpatient stay or clinic visit. A standard time per home visit was determined by interviewing staff. Travel time from the clinic to each patient’s home was estimated using data from the Automobile Association. The cost of travel for each home visit was calculated using a standard mileage allowance obtained from the hospital payroll department.

Treatment was defined as effective if lung function was maintained at the baseline ‘best’ FEV₁% level i.e. percentage decline in FEV₁ was ≤0%. An additional analysis with a less stringent definition of effectiveness of percentage decline in FEV₁ of ≤2% was also performed.

The authors reported that hospital-based care was more effective in terms of FEV₁ but also more expensive compared with home-care IV therapy. Hospital-based care may be cost-effective with a 95% probability at a willingness to pay of £262,500 for 1 extra patient with a decline in FEV₁ of ≤2%. However, using a stricter definition of lung function (decline in FEV₁ of ≤0%) the probability that hospital-based care is cost-effective at a willingness to pay of £10 million per patient is <0.05.

Wages, salary on-costs and staff overheads

Wages, salary on-costs and staff overheads will be incurred by each model of care under consideration. However, if supplementary models of care require a substantial increase in administration due to scheduling and communication issues across the models, staff overheads may increase. Providing an additional administrator would incur a mean annual basic pay of £27,134 (PSSRU 2016: non-medical occupational groupings, administration and estates, full-time equivalent).

Non-staff overheads and capital overheads

Home-care and telemedicine do not require an on-site attendance. For this reason, non-staff overheads and capital overheads may be reduced under these models of care.

Travel

Home-care (excluding self-administered treatment) and Outreach Care incur an opportunity cost in terms of travel, as it is necessary for the MDT from the Specialist Centre to travel off-site. Shared Care will also involve travel when visits include the MDT from the Specialist Centre. There is also the opportunity cost of staff time in making a visit, as the number of people they can see in a given time frame will be reduced.

The mean travel cost/visit can be estimated by multiplying the distance by a mileage allowance. From July 2014, NHS reimbursement has been based on a single rate for the first 3,500 miles travelled of 56p/mile, and a reduced rate thereafter of 20p/mile, irrespective of the type of car or fuel use. This approach can be used to explore how the geographical spread of a region would influence the costs of a model as rural areas would typically involve longer travel distances.

The committee advised that nurses and physiotherapists travel within a 40 mile radius on an almost daily basis to see individual people with cystic fibrosis. This was regarded as a stretch for those healthcare professionals as it would take up most of their day. In terms of time, the committee reported an average duration of 2 hours for a home-visit, with a range of 1 to 3 hours depending on geographics.

K.11.3.1. Multidisciplinary teams of various compositions

MDTs are of various compositions including a core team and, on occasion, an extended team. The cost of providing healthcare professionals within the core and extended MDT are presented in Table 47.

The composition of the MDT will depend on the model of care and the number of people with cystic fibrosis managed by that model. For these reasons, the costing tool allows the user to define the composition of the MDT, number of whole time equivalents (WTEs) and number of people with cystic fibrosis they manage. Figure 4 below provides an example of the user form which allows the user to determine how the MDT composition is to be specified. The number of WTEs utilised in Figure 4 is based on the CF Trust Standards of 2011 for a clinic managing 250 people with cystic fibrosis.

Figure 4MDT configuration taken from the costing tool

Note: specialist pharmacist WTE increased from 1 to 2 to reflect committee consensus on current clinical practice in England, a specialist centre is led by a principal physiotherapist (band 8) and all remaining physiotherapists in the MDT are band 7

K.11.3.2. Models of care

Estimating the annual cost/person for each model requires the following inputs for each model of care:

• clinic size;
• number of clinic visits/ person/ year;
• core members of the MDT;
• WTEs for the size of the clinic.

In addition to their work at the Specialist Centre, core members of the MDT would be responsible for providing Outreach Care and telemedicine. Home-care (excluding self-administered treatment), on the other hand, would only be provided by dietitians, physiotherapists and nurses. Therefore, it is evident telemedicine and home-care alone cannot provide sufficient care for all people with cystic fibrosis (the models of care are not mutually exclusive). Therefore, supplementary models of care should not be compared to the recognised models of care as all people with cystic fibrosis will continue to need the support of the full MDT. Moreover, the frequency of healthcare professional contact via the supplementary models will vary considerably from person to person which is difficult to represent in the costing tool. For these reasons, the only valid comparison (annual cost) is between the Specialist Centre, Shared Care and Outreach Care.

Based on the committee’s feedback, it was agreed the composition of the MDT would be the same for the Specialist Centre and Shared Care, and reflect the core members in their recommendations for that review question.

In the base case, a clinic size of 150 was utilised for Shared Care and 250 for the Specialist Centre, according to committee opinion that Shared Care would usually manage a smaller number of people with cystic fibrosis than the Specialist Centre.

However, a scenario using a clinic size of 250 for both Shared Care and the Specialist Centre was also undertaken, as the committee stated that a smaller clinic size may be driven by restricted resources rather than ideal practice.

WTE recommendations reported in the CF Trust Standards of Care 2011 are presented in Table 48 alongside the estimated annual cost to provide full-time MDTs at Shared Care and the Specialist Centre. This does not include the cost of providing joint-clinics.

The committee noted that the increasing complexity and cost of treatments, coupled with increasing longevity, polypharmacy, chronic use of potentially toxic drugs and the increasing medicines optimisation agenda all supported the need for additional pharmacist resource. For these reasons, the committee regarded the ‘economy of scale’ for pharmacist time seen in the current CF Trust Standards of Care 2011 to no longer be appropriate. Instead, the committee suggested that pharmacist resource should increase in line with other healthcare professionals, such as physiotherapists and dietitians, who have a mandate to see all people with cystic fibrosis regularly. For this reason, the committee agreed it was necessary to increase the WTE of pharmacists from 1 to 2, for clinic sizes ≥150.

The committee advised that “sick” people with cystic fibrosis and babies in their first year of life would need to be seen more frequently than those who are “well”. However, for simplicity, the committee agreed 6 clinics/year would be reasonable to inform the model.

Under the Specialist Centre model, all 6 of those clinics would be conducted at the Specialist Centre. Based on the adult WTEs reported in Table 48, this would cost £2,311,670/year (£9,247/person) for a clinic size of 250.

To calculate the cost of Shared Care (n=150), it is assumed one half of people with cystic fibrosis (n=75) attend the Specialist Centre whilst the other half (n=75) are managed under Shared Care. The Specialist Centre would incur the cost for all 150 people with cystic fibrosis as they supervise Shared Care. Based on the paediatric WTEs reported in Table 48, the Specialist Centre MDT would cost £1,384,445 /year, whilst the Shared Care MDT managing 75 people with cystic fibrosis would cost an additional £747,951 /year.

As previously stated, every person with cystic fibrosis is seen approximately 6 times a year by their MDT. Under a Shared Care model, 4 of those reviews would be undertaken by the local clinic (the Shared Care MDT) and 2 would be undertaken by the Shared Care MDT and Specialist Centre MDT together as a joint clinic.

Each visit to Shared Caremade by the Specialist Centre MDT (assuming 1 of each MDT speciality takes part) costs approximately £470/day in travel costs (120 miles/healthcare professional/visit @ 56p/mile).

Assuming the Specialist Centre MDT can review 12 people/day at Shared Care, a centre managing 75 people with cystic fibrosis would require 13 days (150/12) from the Specialist Centre MDT, to conduct 2 reviews for each person, incurring travel costs of £6,115 (£470 ×13).

The committee also advised that Shared Care would incur additional administration due to the communication with the Specialist Centre. As a result, the committee agreed the cost of an administrator (0.5 WTE) (£13,567/year) should be added.

Including the cost of the Specialist Centre MDT (£1,398,431), Shared Care MDT (£747,951), additional administration (£13,567) and reviews undertaken by the Specialist Centre MDT at the Shared Care Centre (£6,115), the cost to provide Shared Care is £2,166,064/year, or £14,440/person based on a clinic size of 150.

For Outreach Care, the committee advised that the core MDT from the Specialist Centre (assuming 1 of each MDT speciality takes part) would travel 120–200 miles (including return travel) 6 times/year to perform a cystic fibrosis clinic in a local hospital. If the MDT can perform up to 12 reviews each day, it would take approximately 0.5 days to undertake 6 reviews for a person with cystic fibrosis each year. Therefore, the MDT would incur travel costs of £314/year (£627*0.5) if they travelled 160 miles/day to the local clinic for each person with cystic fibrosis.

It is important to note that Outreach Care requires economies of scale to make it viable as the resources and costs to provide Outreach Care for a single person with cystic fibrosis would be substantial. For this reason, the clinic should be organised to ensure several people with cystic fibrosis attend the clinic on the day the MDT are scheduled to visit. Ideally, the committee advised a minimum of 12 people/day.

Similarly to Shared Care, the committee agreed Outreach Care would increase administration costs. For each person with cystic fibrosis this would cost an additional £60/year (£121/day/administrator). Outreach care would also incur additional capital costs and non-staff overheads as additional clinic space is needed. Based on the non-staff overheads and capital overheads reported in the PSSRU for each healthcare professional attending the clinic, this would cost an additional £440 /person /year.

From July 2014, NHS reimbursement has been based on a single rate for the first 3,500 miles travelled of 56p/mile, and a reduced rate thereafter of 20p/mile. Given that the MDT from the Specialist Centre MDT will not travel a distance greater than 3,500 miles under a Shared Care or Outreach Care model, the reimbursement rate is set at 56p/mile. Conversely, healthcare professionals performing home-care will travel more frequently. Assuming a healthcare professional providing home-care travels 40 miles each working day (225 days), the first 3,500 miles would be reimbursed at 56p a day whilst the remaining 5,500 miles would be reimbursed at 20p a day leading to a total annual travel cost of £3,060.

It is assumed each healthcare professional incurs a travel cost, but in reality they may car share or use alternative transport, this will, for example, reduce the cost from £470 for a MDT travelling 120 miles with seven cars to £67 with 1 car, or even less under a car sharing scheme. For a clinic managing 250 people with cystic fibrosis, this is a small cost/person (£1.90 versus £0.30) so is unlikely to change our decision.

Based on the assumptions outlined above, Table 49 shows that the cheapest model of care is the Specialist Centre (£9,247) followed by Outreach Care (£10,126) and Shared Care (n=150, £14,440; n=250, £13,220). However, it is important to note that if demand increases for Shared Care or Outreach Care, the WTEs utilised in the model may underestimate their numbers, as the current WTEs may not consider the time staff spend outside of the Specialist Centre.

Recommending the Specialist Centre model alone would require the person with cystic fibrosis, and their families or carers, to travel greater distances than they would under a Shared Care or Outreach Care model. However, the NICE reference case states that economic analyses should only include costs borne by the NHS. Therefore, costs incurred by people with cystic fibrosis, and their families or carers, using services that are not reimbursed by the NHS should not be included. This position is based on the argument that including the costs of lost productivity discriminates between people based on their capacity for work and income.

However, if the committee felt that the journeys were sufficiently burdensome to negatively impact quality of life, or reduce hospital attendance resulting in later downstream costs, this should be acknowledged. To quantify this reduction in a person’s quality of life we can estimate the QALY gain necessary to determine the additional (incremental) benefit that would be needed for each of the models to be considered as the most cost-effective option.

The 2014 NICE Guidelines Manual advises that an intervention will generally be considered cost-effective if the ICER is £20,000 per QALY or less. The cost per QALY (incremental cost-effectives ratio, ICER) is given by:

Incremental cost ÷ incremental QALY gain = incremental cost per QALY

Or, rearranging:

Incremental cost ÷ £20,000 = incremental QALY gain

To estimate the QALY gain necessary for an intervention to be considered the first step is to calculate the incremental cost of the models being compared. If we compare Shared Care to the Specialist Centre, the incremental cost is £3,973 (£13,220 - £9,247) and if we compare Outreach Care to the Specialist Centre, the incremental cost is £879 (£10,126 - £9,247).

Table 50 suggests what additional benefit each model is required to provide in order to be considered cost-effective relative to the comparator. Hence, despite higher costs, a model could be considered cost-effective if those QALY gains can be achieved. This is not to say that a model is cost-effective, but rather it gives the level of clinical effectiveness relative to the comparator that would be necessary given the current differential in cost and NICE’s threshold of £20,000 per QALY.

To interpret those QALY gains in Table 50 with regards to patient travel and potential time off work, we can take the usual activities domain of the EQ-5D as a proxy. The EQ-5D is NICE’s preferred measure of health-related quality of life in adults comprising the following 5 dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Each dimension has 3 levels: no problems, some problems, extreme problems. The resulting value sets (disutilities) for usual activities estimated from a representative sample of the UK population are summarised in Table 51.

If the Specialist Centre negatively impacts usual activities at either level 2 or level 3, Outreach Care could be considered cost-effective compared to the Specialist Centre, as the disutilities in Table 51 could offset the QALY gains in Table 50. However, neither level 2 or level 3 would be sufficient to justify Shared Care on the basis of usual activities alone, as the disutilities in Table 51 do not offset a QALY gain of 0.20 (Table 50).

K.12.5.1. Strategy costs

K.12.5.2. Cost to treat infection

When the committee advised on the appropriate management strategies for each of the infections they considered the recommendations they made on antibiotics for the treatment of acute pulmonary infection or exacerbation for consistency. Drug costs are taken from the November 2016 BNF Drug Tariff, unless unreported and otherwise stated. People with cystic fibrosis who become infected with a pathogen are assumed to visit their specialist nurse (band 7) for a 20 minute consultation at the clinic at a cost of £43 (£130 per hour of patient contact PSSRU 2016) and receive microbiology tests at a cost of £17 (£8, NHS Reference Costs 2015/16, DAPS07, microbiology; £9 /10 minutes, Band 7 microbiologist, £54 /hour, PSSRU 2016).

In addition to the acquisition cost of drugs, the use of nebulisers for their delivery is associated with fixed costs related to equipment purchase and ongoing costs associated with maintenance. For example a Pari BOY mobile S nebuliser has an upfront cost of £285 (Table 36), but many people with cystic fibrosis would already possess a nebuliser for other treatments. Clinical experts advised Tappenden 2013 and Tappenden 2014 (see Section K.14.1) that the cost to cover replacement aerosol heads and filters would be approximately £200 /year. In the base case, it is assumed people with cystic fibrosis already possess a nebuliser, but a sensitivity analysis has been explored that includes this cost (Section K.12.7).

K.12.9.1. Base case

It is important to note that the total costs incorporate the cost of the strategy to prevent cross-infections with transmissible pathogens (Section K.12.5.1) plus the expected cost to treat the infection (Section K.12.5.2).

As the strategies are applied to a clinic that manages many people with cystic fibrosis the results are reported based on a clinic size of 250. Regardless of the number of people with cystic fibrosis, the incremental costs and QALYs will produce the same ICER, as the relative difference is the same.

Cohort segregation

Segregating a cohort would be considered cost-effective to prevent the transmission of chronic P aeruginosa, as the new strategy is less expensive and more effective than the old strategy, subsequently dominating the old strategy. The cost-effectiveness plane illustrated in Figure 7 shows that the ICERs for each study lie in the south-east quadrant of the cost-effectiveness plane.

However, cohort segregation to prevent intermittent P aeruginosa would not be considered cost-effective as the ICER is substantially greater than NICE’s threshold or dominated (less effective and more effective than no cohort segregation). This is illustrated in Figure 8 with points in the north quadrants above NICE’s advisory threshold of £20,000 per additional QALY.

The cost-effectiveness with regards to intermittent BCC is uncertain with ICERs above and below NICE’s threshold in Figure 9. Based on the probability of infection reported by Whitford 1995, cohort segregation is more effective and less expensive than no cohort segregation; however, according to France 2008, cohort segregation is less effective and more expensive which is expected given that the probability of transmission in France 2008 is greater following the new strategy (16.3% versus 3 to 5%).

Table 70Cohort segregation results (new strategy versus old strategy)

View in own window

Study	Pathogen	Total £ S1	Total £ S2	Total QALYs S1	Total QALYs S2	Inc. costs	Inc. QALYs	ICER
France 2008	Intermittent BCC	£1,357,488	£895,680	210.69	212.05	£461,808	−1.37	Dominated
Whitford 1995	Intermittent BCC	£914,521	£1,050,954	212.26	211.50	−£136,433	0.76	Dominant
Federiksen 1999	Intermittent PA	£1,116,939	£827,033	210.00	208.79	£289,906	1.20	£241,185
Lee 2004 (prevalence)	Intermittent PA	£1,137,291	£814,992	208.66	209.58	£322,299	−0.92	Dominated
Lee 2004 (incidence 1990 versus 2000)	Intermittent PA	£1,19,659	£818,045	209.16	209.38	£311,615	−0.22	Dominated
Lee 2004 (incidence 1990 versus average 1991–2000)	Intermittent PA	£1,144,923	£818,045	208.16	209.38	£326,879	−1.22	Dominated
Federiksen 1999	Chronic PA	£1,639,441	£1,870,334	207.60	202.84	−£230,893	4.76	Dominant
Hoiby 1989	Chronic PA	£4,136,118	£5,922,151	185.77	167.23	−£1,786,033	18.54	Dominant
Lee 2004 (prevalence)	Chronic PA	£2,079,079	£2,117,784	203.75	200.66	−£38,705	3.09	Dominant

BCC, B cepacia complex; ICER, idomncremental cost-effectiveness ratio; PA, P aeruginosa; QALYs, quality-adjusted life years; S1, new strategy; S2, old strategy

Figure 7CE plane for chronic PA cohort segregation

Figure 8CE plane for intermittent PA cohort segregation

Figure 9CE plane for intermittent BCC segregation

Protective equipment

Equipment such as masks, gloves and aprons are relatively inexpensive. According to Chen 2001 and Savant 2014 the probability of infection decreased following their new strategies that included protective equipment. It is important to note that Chen 2001 also cohorted their patients according to B cepacia infection; hence the cost of this strategy has also been included. Savant 2014 on the other hand, did not cohort their patients according to pathogen status, but explicitly stated that a “no-waiting” room policy was applied. Given that Chen’s strategy was a lot less costly than Savant’s with little differences in their effectiveness, Chen’s strategy would be preferable.

Both of these studies show that protective equipment dominates no protective equipment (more effective and less expensive). This is illustrated by the points in Figure 10 that lie in the south-east quadrant of the cost-effectiveness plane. However, it is important to note that it is unclear as to whether the addition of protective equipment is driving the reduction in transmissible pathogens.

Table 71Protective equipment results (new strategy versus old strategy)

View in own window

Study	Pathogen	Total £ S1	Total £ S2	Total QALYs S1	Total QALYs S2	Inc costs	Inc QALYs	ICER
Chen 2001a	Intermittent BCC	£878,272	£1,045,827	212.40	211.52	−£167,554	0.88	Dominant
Savant 2014b	Intermittent PA	£61,116	£78,016	208.49	207.38	−£16,900	1.11	Dominant

BCC, B cepacia complex; ICER, incremental cost-effectiveness ratio; PA, P aeruginosa; QALYs, quality-adjusted life years; S1, new strategy; S2, old strategy

(a)

Includes cohort segregation

(b)

Does not include cohort segregation

Figure 10CE plane for protective equipment

Inpatient beds

It is clear that single rooms are more effective than beds in shared rooms as the ICERs lie in the east quadrants of the cost-effectiveness plane in Figure 11. However, cost-effectiveness depends largely on the cost used to inform the additional cost of single beds, relative to beds in shared rooms.

As stated in Section K.12.5.1 the additional cost of a single inpatient room compared to a bed in a shared room is uncertain. When the model is informed by the additional cost reported by NHS Estates 2005, single rooms dominate beds in shared rooms as they are more effective and less expensive with points in the south-east quadrant of the cost-effectiveness plane in Figure 11.

However, when the costs are based on the charges by South Devon NHS (50% greater) the ICERs are substantially above NICE’s advisory cost-effectiveness threshold of £20,000 per additional QALY.

When a 25% uplift is considered, the ICER based on France 2008 is dominant (single beds are less expensive and more effective) whereas the ICER based on Chen 2001 is above NICE’s upper threshold for cost-effectiveness. This is unsurprising given the greater reductions in transmission reported by France 2008 than Chen 2001 following the introduction of single rooms (16.3% to 2.9% versus 15% to 7%).

The total costs and QALYs for these comparisons are provided in Table 72. Estimating the cross-over point for a threshold of £20,000 per QALY, the additional cost of single rooms to be considered cost-effective would need to be less than £100/day for France 2008 and less than an additional £60/day for Chen 2001.

Figure 11CE plane for single rooms versus beds in shared rooms

Table 72Single inpatient room versus ward bed results

View in own window

Study	Pathogen	Total £ S1	Total £ S2	Total QALYs S1	Total QALYs S2	Inc costs	Inc QALYs	ICER
NHS Estates 2005 costs (single rooms cost an additional £10/night relative to beds in shared rooms)
Chen 2001	Intermittent BCC	£1,113,792	£1,316,823	211.72	210.83	−£203,031	0.89	Dominant
France 2008	Intermittent BCC	£985,541	£1,357,488	212.18	210.69	−£371,947	1.49	Dominant
50% uplift (beds in shared rooms £275/night versus single rooms £413/night)
Chen 2001	Intermittent BCC	£1,678,452	£1,316,823	211.72	210.83	£361,629	0.89	£406,154
France 2008	Intermittent BCC	£1,550,201	£1,357,488	212.18	210.69	£192,713	1.49	£129,218
25% uplift (beds in shared rooms £275/night versus single rooms £344/night)
Chen 2001	Intermittent BCC	£1,373,627	£1,316,823	211.72	210.83	£56,804	0.89	£63,798
France 2008	Intermittent BCC	£1,245,376	£1,357,488	212.18	210.69	−£112,112	1.49	Dominant

BCC, B cepacia complex; ICER, incremental cost-effectiveness ratio; QALYs, quality-adjusted life years; S1, new strategy; S2, old strategy

K.12.9.2. Sensitivity analysis

Sensitivity analysis regarding the cost of inpatient beds has already been reported in the previous sub-section for ease of reference with the base case. For Lee 2004, sensitivity analysis was only conducted on prevalence data since prevalence data was utilised in the clinical evidence review and the same ICERs were found for incidence data (dominated). Due to the large number of comparisons only ICERs are reported in Table 74.

Overall, the results were robust to the analyses undertaken.

Using a cheaper treatment to treat the infection (scenario 1) or reducing the cost to treat an exacerbation (scenario 4) favours the less effective strategy and increases the incremental cost between the old and new strategy. For Lee 2004 this increases the ICER from dominant to positive (£19,021 and £21,525, respectively). However, the ICERs are still around NICE’s advisory lower threshold for cost-effectiveness; hence, cohort segregation could still be recommended as a cost-effective strategy to prevent cross-infections.

Conversely, including the cost of a nebuliser (scenario 2) increases the cost of the less effective strategy, favouring the more effective strategy by decreasing the incremental costs. Given that the new strategy is more effective, the incremental cost decreases, producing a lower ICER, strengthening our decision that the new strategy is cost-effective. Overall, the results are not sensitive to the additional cost of a nebuliser.

Increasing the disutility from an intermittent infection (scenario 3) favours the more effective strategy by increasing the incremental QALYs. However, the ICER produced for Federiksen 1999 remains substantially above NICE’s upper threshold. All remaining comparisons were robust to a change in the disutility.

Shared Care is more expensive to provide than a Specialist Centre, consequently the cost of the new strategy increases when the former is assumed (if the same number of days are segregated in both models). However, given that the prevalence of P aeruginosa is lower in children than adults, less days at the clinic will be segregated under a Shared Care model. Overall, the service delivery model does not change the ICER from the base case.

Scenario 6 has not been undertaken in the model because the QALY gain from Jones 2005 have been used to estimate the acceptable additional cost given NICE’s threshold of £20,000 per QALY. However, it is clear that increasing the cost to treat the infection will favour the more effective strategy. This may reduce the incremental cost between en suite bathroom facilitates and shared bathroom facilitates increasing the likelihood that en suite bathroom facilities are cost-effective (given than en suite bathroom facilities are more effective).

Scenario 8 has identified the scope in the probability of infection (associated with the new strategy), required for an ICER to be on the boarder of NICE’s threshold for cost-effectiveness (£20,000 to 30,000 per QALY). The results of this analysis are illustrated in Figure 12, Figure 13, Figure 14 and Figure 15. Large changes in the probability (associated with the new strategy) from the base case, provide greater confidence in the base case ICER, as the ICER is unlikely to change when alternative, plausible probabilities are applied. Smaller changes on the other hand, show that the probability of infection is a key driver of cost-effectiveness.

From Figure 12 it is clear that substantial changes from the study are needed for intermittent P aeruginosa (Federiksen 1999, Lee 2004) to be considered cost-effective. It is important to note that absolute changes can appear small when considered on a 0 to 100% scale, but the relative changes from the base case probability are somewhat larger; for example, for Whitford 1995 a change in probability from 2.1% to 6.7% is over 3 times more likely, but only 4.5% points greater.

Those studies that require less than a 50% relative change from the base case which could potentially question the generalisability of the cost-effectiveness estimates in clinical practice include Lee 2004 (cohorts by pathogen, chronic infection) and Savant 2014 (protective equipment, intermittent B cepacia complex). However, the relative change required to be considered certain or uncertain, is subjective.

Figure 12Threshold analysis, cohorts by pathogen, intermittent infection

Note: Excel produces negative percentages to obtain the requested ICERs, given that negative percentages are not plausible with regards to the incidence of infection, 0% is produced as the lower limit

Figure 13Threshold analysis, cohorts by pathogen, chronic infection

Figure 14Threshold analysis, protective equipment

Figure 15Threshold analysis, inpatient beds

K.13.3.1. Omalizumab

The dose of omalizumab received is based on the level of immunoglobulin E antibodies (IgE level) and bodyweight and is given as subcutaneous injections every 2 to 4 weeks, but the maximum recommended dose is 600 mg every 2 weeks (8× 150 mg prefilled syringe, £2,049 per month). Table 76 presents the recommended dosages taken from the Summary of Product Characteristics (SPC) for a 4 week administration schedule, and Table 77 presents the cost of syringes available. In addition to the acquisition cost, a day case visit would be required to administer the injection at a cost of £570 (NHS Reference Costs 2015/16, DZ19K, Other Respiratory Disorders with Single Intervention, with CC Score 0–4). The committee stated that a day case tariff would be appropriate as the patient may require access to a bed for the duration of the administration and would require observations before and after administration; hence, an outpatient clinic attendance would not be appropriate.

K.13.3.2. Intravenous corticosteroids

The dose of IV methylprednisolone as an immunomodulatory agent for people with cystic fibrosis has been informed by the committee. Unlike oral corticosteroids and inhaled corticosteroids, IV methylprednisolone is administered monthly over 3 days.

For illustrative purposes, the cost of IV methylprednisolone calculated for 20kg and 60kg bodyweights are presented in Table 78. In addition to the acquisition cost, a day case visit would be required, each day to administer the injection, for the same reasons previously outlined for omalizumab.

K.13.3.3. Inhaled corticosteroids

In addition to the acquisition cost of inhaled corticosteroids, the use of inhalers or nebulisers for the delivery of immunomodulatory agents is associated with fixed costs related to equipment purchase and ongoing costs associated with maintenance and replacement parts. Tappenden 2013 and Tappenden 2014 (see Section K.14.1) included costs associated with nebuliser use when they constructed their economic model, based on expert opinion that it would cost approximately £100 per year to cover replacement heads and filters. The committee also advised that inhalers would be included in the majority of prescriptions, but if inhalers were not included, the cost would be negligible. For example, inhalers such as the Spacer cost of £7.73 (NHS Supply Chain 2015).

The committee advised that the dosages reported in the BNF for an indication of asthma would also apply when they are prescribed as immunomodulatory agents. The acquisition cost of budesonide, beclomethasone and fluticasone are presented in Table 79 according to age and method of inhalation.

Fluticasone, specifically by inhalation of dry powder, was the only inhaled corticosteroid included in the model as no clinical effectiveness data on the remaining inhaled corticosteroids in the protocol was identified.

Fluticasone by inhalation of dry powder was included in the NMA, based on the effectiveness reported by Boeck 2007. They administered 500ug bd which mirrors the dosages in the BNF (Table 79^d).

K.13.3.4. Macrolides

According to the committee, azithromycin can be used as an immunomodulatory agent and as an antibiotic to prevent pulmonary bacterial colonisation with S aureus, using similar dosages:

• Bodyweight 25–40kg, 250mg 3 times/week;
• Bodyweight >40kg, 500mg 3 times/week.

Those dosages were also administered by the trials included in the NMA (Clement 2006 and Saiman 2010) and subsequently, the dosages in the model. Based on those dosages the cost/month is presented in Table 80.

K.13.3.5. Nonsteroidal anti-inflammatory drugs (NSAIDs)

The following dosages were provided by the committee to represent ibuprofen prescribed as an immunomodulatory agent for people with cystic fibrosis:

• Child 1 to 2 years, 50mg bd to qds (2 to 4 times daily);
• Child 3 to 7 years, 100mg bd to qds;
• Child 8 to 12 years, 200mg bd to qds;
• Adults, 400mg tds.

Similarly, the trials included in the NMA (Konstan 1995 and Lands 2007) administered 20 to 30 mg/ kg/ day to a maximum of 1600mg /day. All available preparations of oral ibuprofen are presented in Table 81 for dosages of 800mg/day and 1,200mg/day.

K.13.3.6. Oral corticosteroids

Two trials included in the clinical evidence review (prednisone, Auberch 1985; prednisolone, Greally 1984) administered up to 60mg of oral corticosteroids daily. However, the committee stated that in clinical practice today, the maximum dose is 40mg, due to concerns on drug toxicity.

The committee agreed that there was no evidence to suggest prednisolone and prednisone differ in their effects and given that prednisolone in its non-proprietary form, is a lot cheaper than prednisone, prednisolone is used to inform the model, to reflect the best price available to the NHS (NICE 2013 Guides to the methods of technology appraisal). Available preparations of prednisolone are presented in Table 82 for dosages of 30mg/day and 40mg/day.

K.13.5.1. Drug acquisition costs

People with cystic fibrosis enter the model aged 5; hence it is important to consider their weight as they get older because this will influence the dose and hence cost of treatment they receive. Three of the 4 immunomodulatory agents included in the model rely primarily on body weight to determine the appropriate dose:

• azithromycin ≤40kg, 250mg; >40kg, 500mg;
• prednisolone 1–2mg/kg;
• ibuprofen 20–30mg/kg.

The CF Registry 2014 found in relation to UK growth data, that the median weight percentiles among children and adults (reported up to 19 years of age) with cystic fibrosis reduces, as age increases. Therefore general population estimates may overestimate the weight of people with cystic fibrosis, and more so as they age. In the absence of data, it is assumed the data at 19 years can be used to approximate adults over the age of 19. Even though their BMI may not differ, their weight (and height) would still be lower than average.

The median percentiles over the age of 15 years in the CF Registry 2014 were around the 30^th percentile. The 30^th percentile implies 30% of the general population at the same age are the same weight or lighter than someone with cystic fibrosis, 70% of the general population weigh more than someone with cystic fibrosis. In the model people with cystic fibrosis over the age of 15 are assumed to weigh 60kg based on the average 30^th percentile reported in the GIRLSUK Growth chart 2–20 years and BOYSUK Growth Chart 2–20 years at 20 years of age. The weight associated with children aged 14 to 15 was also adjusted using equivalent methods.

For ages 5 to 13, the percentiles reported in the CF Registry 2014 were not considered large enough to warrant adjustments from the general population as the median percentile did not drop below 40. Table 85 below summarises the weight of children in the general population based on World Health Organisation (WHO) Child Growth Standards (50^th percentile) and the median percentiles among children and adults (reported up to 19 years of age) with cystic fibrosis used to inform the body weights in the model.

The model also uses a lower dose of fluticasone in children under 16 years of age (200ug bd) to reflect prescribing practices. Table 86 below presents the acquisition cost of immunomodulatory agents for adults in the model, assuming 100% adherence to treatment.

More than 1 preparation of ibuprofen and prednisolone is available, and those preparations can vary in their cost. As outlined in the NICE 2013 Guide to the methods of technology appraisal, the reduced price should be used in the base-case analysis to best reflect the price available to the NHS. Therefore, the lowest cost preparation (tablets) has been used to inform the model for prednisolone and ibuprofen. For completeness, oral solution preparations of ibuprofen and prednisolone are explored in sensitivity analysis.

K.13.5.2. Treatment related adverse events

K.13.5.3. Exacerbations

NHS Reference Costs do not report costs specific to cystic fibrosis related exacerbations. As a result, Tappenden 2013 and Tappenden 2014 (see Section K.14.1) took the cost of asthma complications as a proxy to inform their economic model. They assumed that the Reference Costs (2010/11) for asthma with major complications without intubation (DZ15D, long stay, £1,500) reflected the cost of major exacerbations due to cystic fibrosis, whilst the cost of asthma complications without intubation (DZ15E, short stay, £403) reflected the cost of minor exacerbations. Table 89 below presents the latest currency codes related to asthma complications from NHS Reference Costs 2015/16.