Cost of illness of patients with small fiber neuropathy in the Netherlands

Small fiber neuropathy societal costs were estimated at €148 million with daily practice small fiber neuropathy patient data. Health-related quality of life was the key association on health care and societal costs.


Introduction
Small fiber neuropathy (SFN), with a prevalence rate of 53 per 100,000 inhabitants, 25 is a disorder of the thinly myelinated Ad fibers and unmyelinated C fibers and clinically dominated by neuropathic pain and autonomic complaints. 12,29The diagnosis of SFN is based on at least 2 small nerve fiber-related clinical signs of the patient, normal nerve conduction studies (NCS) and an abnormal intraepidermal nerve fiber density (IENFD) in skin biopsy, and/or abnormal quantitative sensory testing (QST). 11,17Diabetes mellitus, autoimmune diseases, and sodium channel gene mutations are the most common conditions observed in patients with SFN, but in 53% of cases, no underlying condition is found. 8In addition to the initial treatment of the underlying condition, neuropathic pain treatment is needed, 15 but generally, these yield disappointing results. 15,29Severe SFN leads to a reduced quality of life (Qol), 2 with commonly associated anxiety and depression greatly interfering with patients' ability to function. 18Higher age and a higher number of comorbidities are prognostic factors for higher health care and productivity costs. 9,32he US annual total health care and patient and family costs of idiopathic painful neuropathy with SFN involvement since 2012 were estimated at $8055 (€7403) per patient. 27Total costs of work productivity loss were estimated to be $13,733 (€12,621) per patient due to a poorer health status, worse sleep outcomes, and loss of productivity. 27A significant association was found between health care and patient and family costs and pain severity, but no statistically significant associations were found in productivity costs. 27However, because of the differences between the United States and the Netherlands in how painful idiopathic SFN is diagnosed, the fact that fewer patients in the United States were in a paid employment and the differences between the healthcare system of the United States and the Netherlands, the study populations and the costs do not lend themselves for straightforward comparison.A cost of illness (COI) study of confirmed SFN has not yet been conducted.Therefore, the costs of SFN and the factors which influence healthcare consumption and productivity costs remain largely unknown.This COI study aims to examine the healthcare, patient and family, and productivity costs of patients with confirmed (by skin biopsy or QST proven) SFN in the Netherlands to estimate the annual SFN costs from a healthcare and a societal perspective.
In addition, the associations of age, pain impact on daily life, anxiety, depression, and health-related Qol on these costs were investigated.

Study design and patients
This COI study was conducted at the diagnostic SFN service of the SFN Center of the Maastricht University Medical Center1 in Maastricht, the Netherlands.The SFN Center is a tertiary referral center for patients with suspected SFN, evaluating approximately 500 patients yearly.The diagnostic SFN service is based on a 1-day stay at the neurological day care unit with time slots reserved for interviewing, examining, diagnostic tests, and analyzing and discussing the findings among a multidisciplinary team.Diagnostic tests include a skin biopsy for identifying abnormalities in IENFD, NCS, and QST.

Patient population and selection
Using the waiting list registration, all patients with suspected SFN and $18 years of age, referred to the SFN Center between April 2017 and February 2020, were invited by e-mail to participate in this study.They were given access to an Internet-based electronic environment to complete the online questionnaires.For those patients not able to complete the online questionnaire, a paper version was provided.All questionnaires were completed before the patients' visit to the SFN Center, so while on the waiting list.Exclusion criteria were declining participation and significant language barrier.Only data of patients with confirmed SFN were analyzed.Between April 2017 and February 2020, 258 patients participated in the study.The flowchart of the waiting list up to and including their visit to the diagnostic SFN service is shown in Figure 1, and 67 patients did not visit the SFN Center yet.In 81.7% of the 191 patients, the diagnosis of SFN was confirmed (n 5 156); in 15.7%, the diagnosis could not be confirmed, and these were omitted from further analysis in this study.

Standard protocol approvals, registrations, and patient consents
The study was approved by the Medical Ethics Committee of Maastricht UMC1 (15-4-004).Informed consent of all patients was obtained before participating in the study, according to the principles of the Declaration of Helsinki. 35

Data collection
Sociodemographic data (eg, age, sex, and education) and data on clinical characteristics (duration of SFN complaints) and patient-related outcome measures (PROMs; eg, pain impact on daily life, anxiety, depression, and health-related QoL) were obtained by the online survey and the electronic patient file.
Pain impact on daily life was measured using the 11-point Pain Impact Numerical Rating Scale (Pain Impact NRS, with 0 meaning having no impact and 10 meaning having the worst imaginable impact). 14nxiety and depression were assessed using the Hospital Anxiety and Depression Scale (HADS) questionnaire, disaggregated for anxiety (HADS-A) and depression (HADS-D).Each subscale consists of 7 questions with answers recorded on a 4-point Likert scale.Scores can range from 0 to 21.Higher scores indicate more symptoms of anxiety and depression.22 The 5-level EuroQol 5D (EQ-5D-5L) was used to measure generic health-related QoL.20 The EQ-5D-5L consists of a Visual Analog Scale (VAS), which ranges from worst (0) to the best imaginable health (100), and 5 additional questions, each representing a health-related QoL dimension.These 5 additional questions covered mobility, self-care, usual activity, pain/ discomfort, and anxiety/depression.20 Each question has 5 response levels, classifying the severity of complaints for that specific dimension and allowing 3125 possible state-of-health combinations. Th were converted into EQ-5D utility scores according to the Dutch tariff.33 Possible EQ-5D utility scores range from 20.446 to 1.00, 33 with 20.446 being the worst imaginable state of health and 1.00 the best.
Healthcare and patient and family costs related to SFN were measured with the iMTA Medical Consumption Questionnaire.
Participants were asked to only report resource use and costs related to neuropathic and autonomic complaints of SFN, within a 3-month recall period. 4The iMTA Medical Consumption Questionnaire includes questions on the utilization of general practitioner (GP) visits, medical specialist visits, other healthcare provider visits (eg, psychologist), emergency room (ER) visits, hospital outpatient visits, and hospitalizations and use of paramedical care, prescription medications, outpatient tests, and procedures.In addition, questions on out-of-pocket costs for medical care and nonmedical resources (help with household or garden work, travel expenses, and help with daily activities, such as cooking) related to SFN are also included.
The iMTA Productivity Cost Questionnaire (iPCQ) was used to measure productivity costs. 5Participants were asked to only report data on paid employment, reduced work schedule, absenteeism, unemployment, and the costs of productivity loss for unpaid employment activities as a result of neuropathic and autonomic complaints of SFN, using a 3-month recall period, in accordance with the Dutch guideline for cost research. 19gure 1.Study flowchart at the time of data analysis on August 21th 2020.SFN, small fiber neuropathy.Subjects were asked to score the pain impact on their working productivity of the last week on the Pain Impact NRS. 14

Statistical analysis
Descriptive statistics were used to present sociodemographic, clinical characteristics, and PROMs data.The Pain Impact NRS scores on daily life were used to categorize subjects into 1 of 3 pain impact groups based on established cut-off points for neuropathic pain (0-3, mild; 4-6, moderate; and 7-10, severe). 36ifferences in demographics, pain impact on daily life, anxiety, depression, and health-related Qol among these groups were tested with an analysis of variance (Kruskal-Wallis test) where appropriate for the continuous variables, whereas a Chi-square test was used for the categorical variables.The continuous variables were tested for normality using a Kolmogorov-Smirnov test.
Healthcare, patient, and family costs were calculated using per unit costs obtained from the Dutch guideline for cost research. 30nit costs were converted to the reference year 2020 by means of index numbers. 6To acquire annual average overall costs per patient, the measured 3-month costs were multiplied by 4.
The productivity costs of paid employment were quantified using the friction cost approach, in which productivity loss is restricted to 85 calendar days (12 weeks). 19The cost of an hour of productivity loss from paid employment was calculated by using the Dutch guideline for cost research based on the average hourly salary costs per paid worker. 19roductivity costs from unpaid employment were valued on the basis of replacement costs for household care.This was equated to a standard hourly rate for cleaning work, as used by the Dutch Central Administration Office (CAK). 19he total societal costs consist of the sum of healthcare, patient and family, and productivity costs.Average total societal costs per patient were multiplied by the prevalence of SFN for the general adult population to estimate the total COI of SFN for the Dutch society.In 2020, the Dutch adult population aged $20 years totaled 15,592,909 residents (Central Bureau of Statistics 2020).By applying the prevalence rate of 53 cases per 100,000 inhabitants, it was calculated that approximately 8264 adults in the Netherlands have SFN.
Usually, cost data are not normally distributed.Therefore, a nonparametric bootstrap resampling procedure with 1000 simulations was performed in SPSS to determine statistical uncertainty of the cost estimates per category.The differences in costs among the pain impact groups were established by calculating confidence intervals (CI) by the bootstrapping procedure.
Table 1 Baseline characteristics of the SFN study population.Multivariate linear regression analyses were performed to estimate the association of age, pain impact on daily life, healthrelated Qol, anxiety and depression, with healthcare, patient and family, productivity, and societal costs.These variables were selected based on their relevance according to the literature. 2,18,32A backward stepwise method was used to test for interaction between the independent variables on all outcomes using an a of 0.05.For statistically significant interactions, results were presented with an interaction term whenever there was at least one continuous variable or stratified per category whenever there was at least one categorical variable.A log transformation was performed on the dependent variables to resemble normality.Where possible, the observed b9s were back transformed into a relative difference (in %) using the formula: (exp[b] 2 1) 3 100%.All analyses were performed using IBM Statistics SPSS version 25.0.

Demographic and clinical characteristics
Demographics, clinical characteristics, and PROMs of the 156 patients with confirmed SFN are shown in Table 1.The median age was 55.2 (IQR 47.2-61.6)years, and the majority (66.7%) was female.

Differences between the 3 pain impact groups
Between all 3 pain impact groups, statistically significant differences were seen in the Pain Impact NRS on daily life and health-related QoL utility scores (P , 0.001).Furthermore, statistically significant higher depression scores were observed in the severe pain group compared with the mild pain impact group (P , 0.001).No statistically significant differences were seen between the 3 pain impact groups regarding age, sex, level of education, duration of SFN complaints, anxiety, and diagnostic tests.

Patient and family costs
The total annual average SFN patient and family costs was €2076 (95% CI €1032-€3759) per patient (Table 3).Personal care was only received by patients in the severe pain impact group, with an annual average of €246 (95% CI €0.0-€705) per patient.Domestic and private paid domestic help only occurred in the moderate and severe pain impact group, with an annual average cost of €532 (95% CI €118-€1219) and €1178 (95% CI €56-€184) per patient, respectively.Half of the patients used informal care, which accounted for .80% of the total patient and family costs (annual average costs per patient €1,739, 95% CI €1181-€2386).The highest travel expenses were found in the severe pain impact group, with an annual average of €253 (95% CI €75-€612) per patient.Over-thecounter medication was used by 44% of the patients, and more than a quarter of the patients bought medical devices, on which an annual average of €550 (95% CI €358-€786) per patient was spent.

Costs of productivity loss
Less than half of the patients were in part-time paid employment (25%) or full-time paid employment (22%), and 21% was disabled.The average SFN productivity costs are presented in Table 4.Among patients in paid employment, the average weekly contract hours were 30.8 hours (95% CI 27.7-32.4),with an average monthly net income of €1387 (95% CI €1087-€1681) based on patients' reported net incomes.Absenteeism in the last quarter occurred in 56% of the patients in paid employment, with an average of 22.9 days (95% CI 16.7-29.1).Costs of productivity loss due to absenteeism per patient in paid employment was €3540 (95% CI €2486-€4676) per quarter.72% of all patients reported a reduction in performing daily household activities due to SFN, with an average quarterly reduction of 517.4 hours (95% CI 388.9-679.4)per patient.Average costs of productivity loss because of limitations in performing daily household tasks due to SFN was €8045 (95% CI €5978-€10,255) per patient.The total average quarterly costs of productivity loss of all patients were €12,167 (95% CI €13,351-€21,926) per patient.

Societal costs
The COI of patients with SFN (€, 2020) in the Netherlands is presented in Figure 2 and will be discussed hereafter.The total average SFN productivity costs accounted for 68% of the total societal costs at the patient level.The total healthcare costs of the adult population with SFN were estimated to be €29.8 million (95% CI: €26.5 million-€33.7 million).Total average societal costs of the adult general population with SFN in the Netherlands were estimated to be €147.7 million (95% CI €120.5 million-€176.3million).

Statistically significant associations with costs
There were no statistically significant associations found between age, pain impact on daily life, health-related Qol, anxiety and depression, and SFN healthcare costs (Table 5).In the SFN patient and family costs, a statistically significant interaction was found between health-related QoL and anxiety (p interaction # 0.001).Therefore, results for patients with mild/moderate anxiety (symptoms # 10) and severe anxiety (symptoms $ 11) are presented separately in Table 5. Health-related QoL was statistically significant inversely associated with patient and family costs in patients with mild or moderate anxiety symptoms (P , 0.01).An increase of 0.1 point on the EQ-5D utility score was associated with a 13.2% decrease (95% CI 220.8 to 4.9) in patient and family costs.No significant association with health-related QoL was observed in patients with severe anxiety symptoms (4.2% increase, 95% CI 210.0 to 20.6, P-value 0.571).For the costs of productivity loss and societal costs, a significant interaction was observed between pain (pain impact NRS # 6 vs pain    impact NRS $ 7) and health-related QoL (per 0.1 point increase, continuous) (P , 0.05), and therefore, these results are presented separately by the pain group in Table 5. Health-related QoL was statistically significant inversely associated with productivity costs in the pain impact group NRS # 6 (P # 0.001).An increase of 0.1 point on the EQ-5D utility score was associated with a 41.4% decrease (95% CI: 257.5 to 219.3) in productivity costs.No statistically significant association with health-related QoL was observed in patients with a pain impact NRS of $ 7 (3.9%decrease, 95% CI 2 17.7% to 12.3%, P-value 0.612).Regarding the societal costs, health-related QoL was statistically significant inversely associated in the pain impact groups NRS # 6 and NRS $ 7 (P , 0.01 vs P , 0.05, respectively).A 0.1 point increase on the EQ-5D utility score was associated with a decrease of 14.6% (223.7% to 24.5%) in societal costs in the pain impact group NRS # 6 and a decrease of 5.8% (210.8% to 20.5%) in societal costs in the pain impact group NRS $ 7.

Discussion
To the best of our knowledge, this is the first study examining the healthcare and societal costs of clinically referred patients with confirmed SFN in the Netherlands.The total healthcare costs to Dutch society for the SFN adult population is estimated to be almost €30 million annually, which is approximately 0.03% of the total healthcare expenditure in the Netherlands in 2020 (€106 billion; Central Bureau of Statistics 2020).Overall, health-related QoL was statistically significant associated with SFN patient and family, productivity, and societal costs.Previous COI of confirmed SFN has not been performed.A cost study of idiopathic painful neuropathy with SFN involvement has been conducted, however, 27 which allows us to compare our results with previous research.Demographic and clinical characteristics of the 2 study populations were similar, but the study population in the previous study 27 was insufficiently defined due to inadequate diagnosis of SFN.Therefore, the results may not be a representative for the SFN population.In addition, data in the previous study was collected over a period of 6 months, which is inconsistent with data collection guidelines for costs studies and may have led to an underestimation or overestimation of costs. 28otal healthcare and patient and family costs (direct costs) of SFN in our study were lower (€5690), and severe pain was associated with statistically significant higher costs.In the previous study, 27 only the direct costs of the mild and moderate pain severity groups were statistically significant higher.
The main contributors to the healthcare costs of SFN in our study were medical specialist care and active medical treatment and were different to main contributors identified in the previous study (prescription drugs and out-of-pocket medical costs). 27In our SFN study, severe pain was associated with statistically significant higher costs of medical specialist care and active medical treatment, whereas in the previous study, no association was found. 27he total productivity costs (indirect costs) of the 2 study populations were similar (approx.€12,000), 27 and the costs of the severe pain group were significantly higher compared with the mild pain group.In the previous study, the indirect costs of the pain severity groups were not statistically significant higher. 27he proportion of patients with SFN in paid employment (47%), who were retired (15%) or disabled (21%) in our study was different compared with the previous study (16%, 49%, and 23%, respectively). 27he main contributor to the high costs of productivity loss of SFN in our study was the cost of lost hours due to being limited in performing daily household tasks, which was significantly higher in the severe pain group compared with the mild pain group.In the cost study of idiopathic painful neuropathy with SFN involvement was the main contributor costs of disability, with no association found. 27ur average SFN healthcare costs are comparable with the UK healthcare costs of painful diabetic peripheral neuropathy 7 and the Dutch healthcare costs of fibromyalgia. 34Furthermore, the use of SFN pain treatments makes up 28% of the SFN healthcare costs, which is comparable with chronic neuropathic pain treatments in other academic pain centers. 31Neuropathic pain is associated with lower health utility scores, and the EQ-5D utility score of our SFN population (0.59) is comparable with diabetic neuropathy (0.61) 13 and fibromyalgia (0.54). 16Comorbidities such as anxiety and depression have a negative effect on Qol in patients with chronic peripheral neuropathic pain, 26 and our study results showed that severe pain was associated with higher depression scores.Health-related Qol is highly correlated with morbidity, mortality, healthcare, and societal costs, 13 and in our study, we used the EQ-5D as a generic instrument to measure health-related Qol.The derived utility scores can also be used in a planned cost-effectiveness study. 10egarding the SFN patient and family costs, over-thecounter-medication (eg, nutraceuticals) was often used in addition to the prescribed medications.Nutraceuticals, such as N-palmitoylethanolamide (PEA) 24 and vitamin D 23 are increasingly used 1 and may play a role in neuropathic pain treatment, but more scientific evidence is needed on their effectiveness.
The leading factor of SFN productivity costs was the limitation in performing daily household tasks due to painful SFN.Important for estimating the COI of SFN is including replacement costs for patients' not performed daily household activities.That is, valuing lost productivity hours from unpaid employment activities, which should not be limited to activities actually taken over by informal care givers. 19Productivity costs in patients with a NRS # 6 were associated with higher pain impact on daily life and lower healthrelated QoL.Our results observed a 41% reduction in productivity costs per 0.1 point increase on the EQ-5D utility score.This study's contribution to the literature is in the detailed insight it provides into the societal COI of patients with confirmed SFN in the Netherlands.This study is based on daily practice data of patients with confirmed SFN, and we were able to investigate a number of associations between costs and patient characteristics.
A limitation of our study was that tertiary care patients were included, who may experience more severe SFN symptoms than patients seeking help in primary or secondary care.However, our sample is a representative for most of the total SFN prevalence figure for 3 reasons: Our study population is comparable with (1) the mean age, the percentage male-female ratio, duration of SFN complaints, and mean average pain of a Swiss study population with SFN, 3 (2) the healthcare cost study from the US, 27 and (3) the inclusion of a considerable number of patients with mild and moderate complaints (n 5 69).Furthermore, because only older prevalence rates were available while rates are likely increasing due to increased global

9 )
Pain impact NRS on daily life, median (

Figure 2 .
Figure 2. Cost of illness overview of SFN in the (2020).

Table 2
Annual SFN-related healthcare costs.

Table 3
Annual SFN-related patient and family costs.

Table 4
Annual SFN-related productivity loss.