Cost of illness associated with Niemann-Pick disease type C in the UK

Abstract

Objective: Niemann-Pick disease type C (NP-C) is a rare and devastating genetic disorder characterised by a range of progressive neurological symptoms, which imposes a burden on patients, family members, the healthcare system and society overall. The objective of this study was to assess direct and indirect costs associated with NP-C in the UK.

Methods: This was a non-interventional, retrospective, cross-sectional cohort study based on responses from patients and/or their carers/guardians recruited from a UK NP-C database. Resource use and direct medical, direct non-medical and indirect costs were evaluated using data collected via postal survey in October 2007, which included a Medical Resource Use questionnaire. Total annual costs per patient were estimated.

Results: In total, 18 Medical Resource Use questionnaires (29% response rate) were received and analysed. The mean total annual cost (SD) of NP-C per patient was £39,168 (£50,315); 46% were direct medical costs, to which home visits and residential care contributed 68% and 15%, respectively. Direct non-medical costs accounted for 24% of the average annual cost per patient, mainly due to specialist education, and indirect costs 30%. If only direct medical costs were considered, the mean annual cost (SD) per patient was reduced to £18,012 (£46,536).

Conclusions: The direct annual per-patient cost of NP-C illness in 2007 appears moderate when compared with other rare and severely disabling diseases. However, cost estimates may be conservative, since findings are limited by a small sample size, low survey response rate and potential recall bias. As demonstrated by this study, a substantial proportion of the cost is shifted from the healthcare system to the patient, family and non-medical providers. These findings highlight the need for treatments that can slow or stop disease progression in NP-C.

Key words::

• cost
• direct
• indirect
• Niemann-Pick disease type C
• resource use

Introduction

Niemann-Pick disease type C (NP-C) is a very rare, devastating genetic disorder. The disease is characterised by a range of progressive and disabling neurological symptoms, including difficulty with muscle co-ordination, impaired gait and balance, and learning and cognitive deficits^1–4. At the late stages of disease, before premature death, patients are frequently bedridden, tube-fed, have little muscle control, and are intellectually impaired^1,2,5.

NP-C is caused by pathological mutations in either of two genes (NPC1 and NPC2), which result in severe impairment of intracellular lipid trafficking^6,7. This leads to the accumulation of cholesterol and other lipids inside the cells in various tissues, including the liver, spleen, lung, bone marrow and in particular, the central nervous system (CNS)^8–10. NP-C is a multi-systemic disease with a highly variable clinical symptomatology⁹.

Patients are often categorised according to the age at onset of neurological symptoms^4,5; the most common presentations are the late-infantile (<6 years) and juvenile forms (6–12 years), but patients can also be diagnosed well into adulthood^11,12. Age of symptom onset has been found to be a major risk factor for progression to severe disease: in general, if symptoms arise early in life, the rate of deterioration is usually faster and premature death occurs earlier^2,13. In infants with very early-onset disease, disease progression tends to be particularly rapid, with death often occurring before the age of 5 years. Most patients with childhood-onset disease die before they reach 20 years of age, although there are reports of some of these patients surviving into their thirties³.

NP-C is pan-ethnic and generally arises sporadically across populations, regardless of race⁴. According to recent reviews of epidemiological data, NP-C is prevalent in at least 1 in 150,000 live births^14–16. However, the number of patients in the UK currently known to have NP-C is far lower than might be expected from this rate, i.e. 67 diagnosed patients as of February 2008¹⁷. This is partly due to the early mortality associated with the disease, and partly due to an unknown number of patients remaining undiagnosed. The signs/symptoms and biochemical findings of NP-C are often not disease-specific, which can delay suspicion and, therefore, diagnosis of NP-C. Further, while NP-C can now be diagnosed reliably using specific genetic and biochemical tests^4,8,18, these tests were, at the time of this research, only available at three specialist centres in the UK. Consequently, it can be expected that many patients already displayed significant disease progression by the time a diagnosis was established.

To date, the cost of illness of NP-C has not been researched. This is probably explained by the rarity of the disease and the fact that there were, until recently, no approved treatments that could effectively have an impact on disease progression and therefore potentially reduce the future cost of the disease. The aim of the present study was to assess the societal and healthcare costs associated with NP-C in a cohort of UK-based patients.

Methods

Design and patients

This was a non-interventional, retrospective, cross-sectional cohort study of NP-C patients aimed at assessing the direct medical costs, the direct non-medical costs, and the indirect costs in terms of inability to work. Analyses were performed from the societal perspective, i.e. considering all costs, including those paid by other public bodies and the patient or caregiver and indirect costs due to lost productivity, as well as the National Health Service (NHS) perspective, i.e. considering direct costs paid by the NHS only.

In October 2007, data were collected for patients who had a confirmed clinical diagnosis of NP-C and were listed on a national NP-C database held at the Royal Manchester Children's Hospital, Manchester, UK. This database, which contains some demographic, but only minimal clinical data, included as of October 2007 a total of 67 diagnosed patients¹⁷. However, postal questionnaires were sent to only 63 patients, excluding those patients for whom the nurse considered participation in the study to be too disruptive. A total of 56% (n=35) of these patients were children aged 14 or under, while 44% (n=28) were adult/adolescent; the majority (60%, n=38) were female, and most (97%, n=61) were cared for at home. Patients and/or carers (as well as parents/guardians for patients aged ≤16 years) were required to provide written, informed consent using consent forms included with questionnaires.

Medical Resource Use questionnaire

The Medical Resource Use questionnaire, which was sent to patients and their carers as part of the survey, was designed to assess healthcare use associated with NP-C. It was developed in consultation with external experts (Chris Hendriksz, Consultant Paediatrician at Birmingham Children's Hospital; expert in health services research), and was initially tested in a pilot study with parents/guardians of NP-C patients. The questionnaire was mainly intended to be completed by the parents/guardians of patients with NP-C, but could also be completed by patients able to consent for themselves.

Information on medical healthcare resource use related to NP-C was collected, including: investigations and tests, medications, home visits (community and social services), consultations (nurses and doctors), hospitalisations, other services provided by health professionals (e.g., physiotherapists), residential care, and out-of-pocket expenses (e.g., non-prescription medications) (Table 1). In addition, NHS-funded as well as out-of-pocket non-medical resources (e.g., transport, aids/appliances), the impact of NP-C on patient and carer employment or education, and basic patient demographic characteristics (date of birth, gender and location of residence) were recorded.

In order to avoid recall bias, different recall periods were applied to the different resource categories (Table 1). For instance, a 6-month recall period was used for infrequent but potentially expensive resources such as hospital visits or aids/appliances (adaptations to houses and/or cars). On the other hand, a 1-day recall period was applied to frequently-used resources such as prescription medications, which are generally taken daily, while a 1-month recall period was used for occasional resource use items, e.g. non-prescription medications.

Unit costs

Wherever possible, unit costs for resources that were paid by the NHS or otherwise publicly funded were taken from respective official sources including: NHS Reference Costs¹⁹, the Personal Social Services Research Unit (PSSRU)^20,21, the British National Formulary (BNF)²² and the Office of National Statistics (ONS)²³. If no published unit cost was available from any of these sources, costs were estimated based on local prices at Manchester Children's Hospital, expert opinion, or professional association websites. For out-of-pocket expenses, unit costs were collected in the Medical Resource Use questionnaire, and the hourly cost for hours of work missed was based on the median salary for the UK population according to the Annual Survey of Hours and Earnings (ASHE) 2007²³. All unit costs refer to 2007 and are listed in the Supplementary material.

Data analysis

All questionnaires received were entered into a Microsoft Access 2003 database by two independent researchers; any discrepancies were reviewed and resolved with a third researcher. Data analyses were conducted in Microsoft Excel 2003. For the resource use analysis, mean (SD) resource use per patient for the respective period of recall was calculated for users (i.e., patients reporting use of a resource), as well as for the entire cohort. In the cost analysis, mean (SD) costs for each resource were also calculated for users and for the entire cohort; mean costs for the respective period of recall were converted to monthly costs. Total mean monthly and annual costs per patient were analysed by each type of resource and each resource category.

Missing cost values, where respondents indicated resource use but gave no further detail, were imputed based on the mean per-patient cost for the respective resource calculated for the entire patient cohort (derived from complete responses).

The use of miglustat (Zavesca, Actelion Pharmaceuticals, Ltd., Allschwil, Switzerland) was excluded from the main medication cost analysis. Miglustat is a small iminosugar molecule that reversibly inhibits the enzyme catalysing the first committed step of glycosphingolipid synthesis. Miglustat was developed as a potential therapy for NP-C based on preclinical studies which demonstrated its ability to cross the blood–brain barrier²⁴, and reduce excess storage of brain gangliosides and prolong survival in animal models of NP-C²⁵. A randomised, controlled clinical trial and subsequent open-label extension indicated that miglustat slowed disease progression in both adults/juveniles and paediatric patients with NP-C^26,27, with therapeutic effects maintained in adults/juveniles during up to 66 months of treatment^28,29. On the basis of these data, miglustat was approved in Europe for the treatment of progressive neurological manifestations in paediatric and adult patients with NP-C in January 2009. However, at the time of the study, miglustat was an investigational drug for NP-C and patients who were taking miglustat at the time of this study were doing so as part of a clinical trial. A sensitivity analysis was conducted, which compared patients taking miglustat with those not taking miglustat, in order to assess whether results were different.

Results

Sample population

From a total of 63 questionnaires sent out to patients and carers, 18 were returned (29% response rate), providing data for a relatively equal number of male/female patients (9/8). Gender information was not available for one patient, who did not complete the demographic section of the questionnaire. The response rate for adult/adolescent patients was 43%, and that of children aged 14 or under was 17%. Questionnaires were completed by patients themselves in 28% of cases, and otherwise by their carers.

The overall mean ± SD (range) age of respondents was 22.9 ± 16.6 (2–53) years. One-third of patients (n=6) were children aged 14 or under, with a mean ± SD (range) age of 4.5 ± 2.3 (2–9) years. A total of 61% of patients (n=11) were adolescents/adults with a mean ± SD (range) age of 32.9 ± 11.2 (16–53) years. Most patients (n=16; 89%) resided with their families. Four out of 18 patients (22%) were taking miglustat for the treatment of NP-C. Patients prescribed miglustat tended to be slightly older than those not receiving miglustat, with a mean age of 31 years compared with 20 years for the non-miglustat group. One patient taking miglustat (25%) was female compared with 50% of the non-miglustat patients. Due to very small patient samples, the two subgroups should not be compared and no conclusions can be drawn in this regard.

Resource use analysis

Direct medical resource use

Direct medical resource use encompassed all NHS-funded healthcare services. Fourteen patients (78%) had consulted healthcare professionals in the previous 6 months, mainly hospital consultants, GP surgeries, and hospital-based specialist NP-C nurses (Figure 1a), with a mean (SD) number of visits per patient in the previous 6 months across the whole cohort of 0.94 (0.80), 0.67 (1.28), and 0.28 (0.46), respectively. In addition, 12 patients (67%) had accessed ‘other medical support services’ in the previous 6 months, especially speech therapists, occupational therapists and educational psychologists (Figure 1b). Other therapists consulted, in order of frequency, included swallowing therapists, ‘other’ NHS-funded services (‘life-time nursing’ service, ‘home portage – learning’, and optician/dentist), social workers, dieticians, physiotherapists, hearing specialists, and geneticists. Notably, eight patients (44%) had consulted more than one of these specialists. In terms of the mean (SD) number of visits per patient in the previous 6 months across the whole cohort, the top three resources under the category, ‘other medical support services’ were: speech therapists, 0.83 (1.95); physiotherapists, 0.61 (1.96); and occupational therapists, 0.5 (0.99).

Apart from outpatient visits, 11 patients (61%) had received different investigations and tests in the previous 6 months, the most frequent being blood tests and eye examinations (Figure 1c). Altogether, eight patients (44%) had taken prescribed medications on the day before completing the questionnaire (mostly antidepressants, sleep medication, anxiolytics and analgesics), and six (33%) used over-the-counter medications within the prior month, four of whom had also taken prescribed medications on the previous day.

Some patients had required hospitalisation, residential care, or home visits during the previous 6 months. Overall, there were two in-patient admissions, four out-patient admissions (i.e., a scheduled hospital visit with no overnight stay) and three admissions to an A&E department for emergency treatment, and a total of three admissions were to paediatric wards. Five patients (28%) were in residential care, which lasted between 4 and 182 days (one patient was in permanent residential care). Three patients (17%) had home visits from a care worker or occupational therapists, which ranged from 4 hours per day on 4 days per month to 24 hours per day on a daily basis (in one patient).

Direct non-medical resource use

Direct non-medical resource use included educational support, travelling to/from healthcare services, and use of aids/appliances. Nine patients (50%) were in education, of whom five (56%), all in primary/secondary education, had required specialist educational support on an individual or group basis in the previous week, with mean (SD) numbers of days of support required per child of 4.17 (2.04) and 2.50 (2.74), respectively. The mean number of visits to educational support professionals (psychologists, school nurses, and community paediatricians), which was measured over 6 months to capture sufficient data, ranged between 0.17 and 31 times per child per 6 months.

Overall, respondents reported relatively low use of transport to attend healthcare consultations or to visit patients. Based on reported use of transportation in the previous month from two respondents, the average distance travelled per month was estimated at 13 miles per person across the entire cohort. The use of aids/appliances for NP-C was also infrequent, with four respondents (22%) reporting having installed home adaptations or using walking aids, continence products, or protective bedding during the past 6 months.

Indirect resources

Indirect costs occur because of patients and/or carers not being able to work, and indirect cost data were therefore collected for adult respondents only. Seven out of 11 adult patients (64%) reported being unemployed due to their disease; they had not been employed for a mean (SD) of 8 (5.73) years. Only two (18%) adult patients were employed, one of whom had to reduce work time from 37 down to 12 hours per week, while the other worked 40 hours per week.

The employment status of carers was also adversely affected by patients' disease. From a total of 13 carers responding to the questionnaire, over half (seven carers; 54%) were not currently in paid work due to their caring responsibilities, and reported having been in this position for an average of 6.3 years (range 2–14 years). An additional three carers (23%) had reduced their working hours by an average of 17 hours per week (range 4–25 hours), and two other employed carers (15%) had taken an average of 2 days' leave in the last 3 months.

Cost analysis

Total costs

The total mean annual cost (SD) of NP-C per patient across the entire cohort was estimated at £39,168 (£50,315), which included both direct and indirect costs (Table 2). The distribution of the patient population by total annual cost is shown in Figure 2. Direct medical costs represented 46% (£18,088 per year) of total costs, while direct non-medical costs and indirect costs contributed 23.4% (£9,442 per year) and 29.7% (£11,639 per year), respectively (Figure 3a). If the limited NHS perspective was applied (i.e., excluding indirect costs and resources not paid by the NHS), the mean annual cost (SD) per patient was £18,012 (£46,536). Mean annual costs (SD) paid out-of-pocket by patients and families amounted to a total of £350.76 (£650.4), including £67.2 (£105.6) for non-prescription medications, £10.8 (£46.8) for other health services and £273.6 (£559.2) for travel.

Direct medical costs

The proportional breakdown of direct medical costs (Figure 3b) showed that the majority (72.5%) of overall cost was due to home visits, followed by residential care, tests and investigations, and consultations. Although home visits from care workers or occupational therapists were infrequent, they were associated with high mean costs. However, the annual cost of care worker visits across the entire cohort (mean [SD] £13,104 [£45,384]) was highly influenced by one patient who received 24-hour daily care. Residential care was also infrequent but costly (mean annual cost [SD] of £444 [£1,884] per patient <14 years old staying in children's hospices, and £2,388 [£8,544] for patients ≥14 years old staying in adult hospices).

The most frequently prescribed investigations were blood tests, but these were associated with low overall costs across the cohort (mean annual cost [SD] £14.0 [£29.88] per patient). On the other hand, monitoring techniques such as videofluoroscopy assessments of swallowing and eye examinations for abnormal saccades were rarely used, but expensive, resulting in mean annual costs (SD) per patient across the entire cohort of £70.68 (£205.68) and £517.2 (£1,073.76), respectively. MRI scans for atrophic changes in the brain in one patient cost £265, giving a mean annual cost (SD) per patient across the cohort of £176.52 (£749.04).

The total mean annual cost of prescribed medications was £207.72 (SD £726.72) per patient. The most costly prescribed medications, in terms of mean annual cost per user, were hyoscine patches (£2,352), Movicol (£336) and penicillin (£192) (each used by one patient), and imipramine (£72), which was used by two patients. NHS prescription charges paid by the patient/carer were not recorded.

Direct non-medical costs

The main contributor to the total mean annual direct non-medical cost of NP-C was special education, accounting for 97% (Table 2). Only the incremental cost of specialist education has been included in the analysis presented (mean annual cost [SD] £9,168 [£14,976] per patient per year across the cohort), as the cost of attendance at general schools is borne by society for all children.

Travel expenses were generally low (3% of total direct non-medical costs) and were paid out-of-pocket by patients/carers; the mean annual cost (SD) across the cohort was £273.6 (£559.2). Under the category aids/appliances, modifications to the home contributed the greatest single total cost per patient across the whole cohort (mean [SD] £361 [£1,532] for items funded by the NHS and £261 [£1,108] for those paid by patients/carers). However, as aids/appliances were one-time only expenses, they were excluded from the total annual cost per patient. If aids/appliances were included in the cost analysis, the total mean (SD) annual cost of NP-C per patient would increase from £39,168 (£50,315) to £40,513 (£52,983).

Indirect costs

The indirect costs of NP-C were high and stemmed from productivity losses due to reduced working hours, absenteeism and unemployment, both for patients and carers. The mean (SD) indirect cost per patient across the entire cohort was £10,008 (£11,736) per year, while for carers it was £1,632 (£4,356) (through either reduced hours or leave taken).

Sensitivity analysis

The results of the sensitivity analysis, which assessed mean costs in the subgroups of patients taking and not taking miglustat, demonstrated that costs were generally similar between the two subgroups (Table 3); the overall mean annual costs in patients not taking miglustat was £39,804, compared with £36,948 in those taking miglustat.

Discussion

Findings from this study, the first assessment of resource use and associated costs in NP-C, indicate that the mean annual cost per patient was rather moderate (£39,168), considering the severity of the disease. Less than half of this cost (£18,012) was borne by the healthcare system, while the remainder affected other publicly funded systems as well as patients and their carers. Due to the fact that, until recently, there were no approved treatment options for NP-C, healthcare costs were expected to be moderate, a phenomenon that has been observed for other rare diseases such as cystic fibrosis, for which there is currently no approved targeted treatment³⁰.

For example, a study of the cost of cystic fibrosis, which was carried out in patients attending a cystic fibrosis centre in the UK in 1992, estimated the average annual direct cost per patient at £8,241, of which the cost of medication (predominantly intravenous antibiotics) made up £4,718³⁰. A later study in the US reported annual healthcare costs of cystic fibrosis of $13,300 per patient, of which $2,394 was related to treatment with the targeted therapy dornase alfa (Pulmozyme, Genentech, San Francisco, CA, USA)³¹. In his review of the economic evidence on cystic fibrosis, Krauth noted that ‘according to the literature, direct costs [of cystic fibrosis] amount to between US$6,200 and 16,300 (1996 values) per patient per year… most studies likely underestimated the actual costs (e.g., by disregarding provision of certain healthcare services)'³². In this NP-C study, an attempt has been made to measure all costs related to NP-C, including those not borne by the NHS. Contrary to cystic fibrosis, drug costs of NP-C seem to be rather limited, reflecting the fact that patients received only symptomatic treatment. Another rare, chronic disease, for which therapy is currently limited to symptomatic treatment of affected organs, is scleroderma. A 1997 COI study in the US found the total annual cost of scleroderma to be $14,959 per patient, of which direct medical costs comprised only 31%³³. It should be noted that comparisons between these COI studies with the one reported here are limited by differences in the healthcare systems as well as in the years during which studies were conducted.

Despite improved knowledge on the underlying pathophysiology of NP-C, only palliative therapies were generally available at the time of this study. These were prescribed for the alleviation of feeding and movement symptoms, seizures, sleep disorders, gastrointestinal symptoms, and, in rare cases, lung involvement^5,34. Due to the multi-organ involvement of NP-C, patients generally require multidisciplinary healthcare as well as various forms of non-medical support. However, in this study, the number of healthcare visits by NP-C patients was found to be relatively low, especially in view of the severity and impact of the disease. A possible explanation is the fact that patients and carers, who know that only symptomatic therapies are currently available, might undertake relatively few healthcare visits. Findings from the study seem to indicate that there is a shift in the overall cost of care from the NHS to other public bodies as well as carers.

While home visits and residential care were relatively infrequent in this sample of NP-C patients, these resources were the main drivers of the overall cost. Indeed, a single patient incurred very high home visit costs due to 24-hour daily care, which had a large influence on the overall mean direct medical cost (£13,104 per patient per year); excluding this patient resulted in a lower estimate (£7,500 per patient per year). Residential care also contributed high costs, although again, a single patient living in permanent residential care skewed the cost results in this respect. However, the estimated mean direct cost for the whole sample is likely to be a more reliable estimate than that calculated after exclusion of the single patient requiring more intensive care, since in clinical practice there are known to be severe patients requiring intensive home or residential care. These can be expected to be underrepresented in this study, assuming that carers for severely afflicted patients are less likely to respond to a survey. In addition, in the current study, patients were excluded if the nurse considered participation in the study to be too disruptive for the patient.

Together, non-healthcare costs, i.e. direct non-medical costs, non-medical out-of-pocket expenses, and indirect costs contributed over half of the total mean monthly costs associated with NP-C. Specialist educational support was the main source of direct non-medical costs, primarily due to its intensive nature. Indirect costs due to lost employment or reduced working hours, both among patients with NP-C and their carers, were also found to be a major component of total costs. It should be noted that indirect costs could only be estimated in this study, since data collected by the postal questionnaire were limited to the length of unemployment, the number of reduced working hours per week, and the number of days of work missed during the past 3 months. In the absence of information on completed education and/or previous type of employment, which were not collected due to the sensitivity of these data, indirect costs were estimated based on UK median wages. In addition, productivity losses associated with carer unemployment were not calculated because it was not known whether respondents had needed to give up employment due to their caring responsibilities. Also, lost school time for those patients still in education was not considered as a cost. Thus, the indirect cost of NP-C is likely to have been significantly underestimated.

Since the questionnaire used in this study was designed to assess only resources associated with NP-C, estimated costs can be considered a reliable, although conservative, reflection of the overall NP-C-related cost to the NHS and society. Nevertheless, the research is subject to a few methodological limitations, with the low response rate being the main one. In the absence of registries on NP-C containing detailed healthcare resource use information, respective estimates could only be obtained by means of a survey. While this approach has the advantage that non-healthcare costs can also be obtained, it is subject to patient/caregiver willingness to participate in the study. Since NP-C is a very severe disease, which imposes a high burden on caregivers, the response rate to this survey was expected to be low. Additional reasons for the low response rate may include the lack of rigid pre-planned follow-up of non-responders to the questionnaire. Furthermore, patients with low resource use may have been reticent about returning a questionnaire due to concerns that it appeared to be incomplete or might not meet the perceived expectations of the researchers.

In general, the group of 18 survey respondents appears to be comparable to the 63 NP-C patients initially sent the questionnaire. According to information from the primary author (JI), who knows all patients personally, the cohort of 18 respondents included patients with a wide range of levels of disease severity representative of the total population of 63 patients. Nevertheless, the small size of the NP-C patient cohort is the main limiting factor. The sample may not be completely representative due to the low survey response rate, especially for children aged 14 and under (17 versus 43% for the adolescents and adults). Compared with the UK NP-C population, the percentage of children was only 35 versus 58%¹⁷, indicating that findings may not be generalisable to the overall UK NP-C population. In addition, results are relevant to the UK setting only and are not generalisable to other countries due to difference in healthcare systems. In general, the study sample can be expected to reflect less severely-affected patients; carers for more severe patients may have been less likely to respond to a survey because of time constraints or fear of distress generated by the questionnaire. This, together with the fact that more severely-affected patients are more likely to require residential or intensive home care, implies that that the cost estimates reported here are probably conservative.

It should also be taken into consideration that no information is available on the number of undiagnosed cases of NP-C in the UK. This, taken together with the lack of generalisability of the 18 patients, implies total costs for the UK cannot be reliably estimated. In addition, the cross-sectional design of this study does not allow for calculation of life-time costs; since the sample is likely not representative of patients at different disease stages, only the current annual costs of NP-C have been estimated. This is an underestimate of the life-time costs, which would include future indirect costs due to reduced earning potential.

The use of different recall periods and the fact that recall bias cannot be excluded, especially for the infrequent but costly resource items like home adaptations or specialist consultations, has likely contributed further to the conservative nature of the cost estimates. Since many drugs are not routinely taken, the cost of medications may have been severely underestimated. In addition, exacerbations of symptoms requiring drug treatment are unlikely to be appropriately captured, since these are infrequent, and carers and patients are unlikely to complete a questionnaire during these periods.

The subgroup analysis comparing non-miglustat-treated patients with those receiving miglustat showed that overall costs (excluding miglustat drug costs) were similar, and that pooling of all patient responses was appropriate. Since this study was not designed to assess differences in costs between the two subgroups, no conclusions can be drawn in this regard. The overall small sample size did not allow for any further subgroup analyses, e.g. by levels of disease severity or age at symptoms onset. Longitudinal, long-term studies are required to assess how healthcare costs might vary with different clinical phenotypes and natural history of disease.

Despite these limitations, the findings reported here highlight the fact that NP-C results in significant costs, not only for the healthcare system, but also for patients, their carers, and society overall. While this cross-sectional, retrospective survey can only provide a snapshot of the costs of NP-C across a broad patient population, it should be borne in mind that NP-C is a progressive, chronic disease, and that without effective treatment, costs can be expected to increase over time. Since patients experience more severe functional disability as the disease progresses^5,35, this will have an impact on costs, with patients requiring more and more medical and non-medical support. There is therefore a need for effective, disease-specific therapies that can delay the progression of neurological symptoms in NP-C, thereby potentially reducing the long-term cost of the disease.

Conclusion

The direct cost of NP-C illness in 2007 appeared moderate when compared with other rare and severely disabling diseases; a substantial proportion of the cost appears to be shifted from the healthcare system to the patient, family and non-medical providers. These findings highlight the need for treatments that can slow or stop disease progression in NP-C.

Appendix

The table dealing with ‘Supplementary material: Unit cost listings by category’ for this article is available as electronic supplementary data (doi: 10.3111/13696990903245863).

Figure 1. Percentage of survey respondents (n=18) (a) consulting each type of healthcare professional, (b) accessing each type of other healthcare service, and (c) undergoing each type of test or investigation (in previous 6 months). Total number of respondents reporting any use of each resource category: n=14 for consultations, n=12 for ‘other funded services’, n=11 for tests and investigations.

Figure 2. Patient distribution by total annual cost of NP-C.

Figure 3. Breakdown of (a) mean total annual cost of NP-C per patient and (b) mean annual direct medical costs for NP-C patients. OoP, out-of-pocket; NHS, National Health Service; PSS, Personal Social Services.

Table 1. Recall period and unit cost sources for resources captured in the NP-C Medical Resource Use questionnaire.

Payer	Resource use	Recall period	Source of unit cost data
Direct medical
NHS	Hospital admissionsHospital outpatient visits*Home visitsInvestigations/testsPrescription drugsHospice and respite care	6 months6 months1 month6 months1 day6 months	NHS reference costs; PSSRU 2007PSSRU 2007PSSRU 2007NHS Reference Costs 2006-07BNFPSSRU 2007
Patient (OoP)	Private specialist consultationsNon-prescription drugs	6 months1 month	Collected in questionnaireCollected in questionnaire
Direct non-medical
DWP, DH, DfES	Travel expenditureSocial servicesAids/appliances^†Specialist educational support	1 month6 months6 months6 months	PSSRU, internet and other sourcesPSSRU 2007PSSRU 2006<www.sheffield.gov.uk/?pgid=64269& fs=b> accessed 16.01.2008;<www.dfes.gov.uk/trends/index.cfm? fuseaction=home.showChart&cid=2&iid=6& chid=27> accessed 01.07.2008; PSSRU 2007; NHS Reference Costs 2006–07
Patient (OoP)	Private home helpTravel expenditureAids/appliances^†	6 months1 month6 months	Collected in questionnaireCollected in questionnairePSSRU 2006
Indirect
Society	Productivity losses (absenteeism) due to inability to work	3 months	ASHE 2007, ONS

*GP/nurse/AHP (allied health professional) consultations; ^†adaptations to house/car. Abbreviations: ASHE, Annual Survey of Hours and Earnings; BNF, British National Formulary; DWP, Department of Work and Pensions; DH, Department of Health; DfES, Department for Education and Skills; GP, General Practitioner (primary healthcare physician); NHS, National Health Service; OoP, out-of-pocket (expenses); ONS, Office of National Statistics; PSSRU, Personal Social Services Research Unit.

Table 2. Total mean direct and indirect costs per NP-C patient per month (2007).

Cost category	Range of mean (SD) costs for individual resource items in respective category (£)	Mean (SD) monthly cost* per patient per resource category (£)
Direct medical costs
Tests and investigations	0.02 (0.05)–43.1 (89.5)	74.9 (155.3)
Medication (prescribed)	0.1 (0.4)–10.9 (46.2)	18.3 (60.4)
Medication (non-prescribed; OoP)	—	5.6 (8.8)
Home visits	4.6 (19.3)–1,087.8 (3,762.8)	1,092.3 (3,781.7)
Healthcare consultations	0.08 (0.4)–34.4 (29.2)	47.1 (35.5)
Hospitalisations	3.1 (7.1)–9.4 (27.2)	17.9 (26.8)
Other healthcare services^†	0.2 (0.6)–2.7 (8.2)	14.2 (22.6)
Other healthcare services (OoP)	—	0.9 (3.9)
Residential care	37 (157.1)–298.6 (866.2)	236.1 (718)
Total		1,507.3 (3,884.2)
Direct non-medical costs
Special education	0.3 (1.3)–422.2 (700.6)	764 (1,247.2)
Travel (OoP)	3.9 (12.4)–10.6 (33)	22.8 (46.6)
Total		786.8 (1,241.3)
Indirect costs
Productivity losses (patient)	63.5 (269.4)–770.1 (993.4)	833.6 (977.7)
Productivity losses (carer)	6.8 (267.7)–129.8 (347.2)	136.3 (362.6)
Total		969.9 (1,079.8)
Overall total		3,264 (4,192.9)

*Across entire sample population; ^†including social worker support, occupational therapy, dietician consultation, swallowing therapy, geneticist consultation, speech therapy, physiotherapy, educational psychology and hearing specialist consultation; OoP, out-of-pocket.

Table 3. Sensitivity analysis: total costs in miglustat-treated versus non-miglustat-treated patients*.

Costs	Mean (SD) total monthly cost per patient per resource category (£)
	Patients not taking miglustat (n=14)	Patients taking miglustat (n=4)
Direct medical costs	1,599.8 (4,374.7)	1,183.4 (1,543.8)
Direct non-medical costs	752.8 (1,139.5)	905.8 (1,755.1)
Indirect costs	964.2 (1,105.9)	990.2 (1,143.4)
Overall total	3,316.8 (4,679.4)	3,079.4 (2,163.0)

*Cost of miglustat was excluded from main analysis, as patients taking miglustat received it as part of a clinical trial.

Acknowledgments

Declaration of interest: This study was funded by Actelion Pharmaceuticals Ltd, Allschwil, Switzerland. J.I. has received research support from Actelion for participation in clinical trial programmes, and speaker fees for participation in sponsored events. C.G. and E.H. are employees of Actelion Pharmaceuticals Ltd. K.G. and K.L. are employees of HERON Evidence Development Ltd, which received consultancy fees from Actelion Pharmaceuticals Ltd. Alpha-Plus Medical Communications Ltd provided writing assistance, paid for by Actelion Pharmaceuticals Ltd.

Notes