Cross-sectional evaluation of health resource use in patients with functional neurological disorders referred to a tertiary neuroscience centre

Introduction

Functional neurological disorder (FND) represents genuine and involuntary neurological symptoms and signs that have characteristic clinical features and represent a problem of voluntary control and perception despite the normal basic structure of the nervous system.1 Manifestations of FND are varied, such as decreased or increased movement, loss of sensation, difficulties in speech, abnormal gait or posture, cognitive symptoms and seizure-like episodes (functional seizures (FS)).1 FND can have a significant impact on the sufferer’s quality of life.2 Patients often present with comorbid psychiatric conditions, with both depression and anxiety occurring in up to 40% of patients with FND.3 4

The FND of movement and sensation has a prevalence of roughly 50 per 100 000 population and an incidence of 4–12 per 100 000 population per year. FS contributes a further 1.5–4.9 per 100 000 population per year, with a prevalence of 2–33 per 100 000 population.5 Patients with FND make up 9% of neurology admissions6 and 16% of neurology clinic referrals.7 Delayed diagnoses of FND lead to worse outcomes for patients,3 as well as preventable costs, such as missed work, general practitioner (GP) and specialist appointments, and investigations. Diagnostic uncertainty in the midst of ongoing symptoms can lead to intangible costs, such as decreased quality of life (QOL). These costs carry a burden on patients, clinicians, healthcare systems and the economy.

The costs of FND (and other medical conditions) can be separated into direct and indirect costs. Direct costs represent resources used for healthcare (eg, the cost of investigations and time spent on assessments by a doctor), as well as out-of-pocket costs to the patient. Indirect costs represent productivity losses arising from morbidity-related sickness absence (eg, loss of employment and cost of childcare while hospitalised). Direct and indirect costs together constitute the economic burden of FND, which can be estimated by measuring the monetary valuation of healthcare utilisation and lost productivity in patient samples.

The literature concerning the economic cost of FND is sparse, and any conclusions that may be drawn from it are limited by the heterogeneity of the studies that focus on the topic. Studies vary in the costs included in their analysis, with many focusing only on hospital costs.8 However, Stephen et al’s comprehensive study highlights that people with FND accrue similar costs to those with refractory epilepsy and demyelinating disorders. The cost of FND alone was estimated to be $1.2 billion annually in the USA in 2017,9 and these costs appear to depend on the patient’s satisfaction with the explanation of their diagnosis.10 In Denmark, Jennum et al showed a nearly tenfold increase in combined direct and indirect costs in FS patients compared with healthy controls.11

Studies that assessed indirect costs reported these costs as being higher than the direct medical costs resulting from the disorder.8 It has been found that patients with FND are more likely not to be working for health reasons and to be receiving disability-related state financial benefits than patients with other neurological disorders.12 No study has yet assessed whether symptom severity and/or duration impact the economic cost of FND.

In this study, we set out to evaluate the direct and indirect costs associated with FND through a retrospective questionnaire-based assessment of people referred to a tertiary FND specialist assessment clinic.

Methods

Results

Direct health costs

The breakdown of costs by service is given in tables 1 and 2. Despite being used by only 6.36% of patients, the cost associated with intensive care unit admissions had the highest mean cost per patient at £629.15. This was followed by neurology ward admissions (also used by only 6% of patients), which carried a mean cost of £245.38. GP consultations, whether in person (81% of patients) or by phone (46% of patients), were used by most patients and carried a mean cost per patient of £245.

Table 1

View inline|View popup

Inpatient service utilisation and cost in the 6 months prior to new appointments (n=110)

Table 2

View inline|View popup

Outpatient service utilisation and cost in the 6 months prior to new appointments (n=110)

Of note, patients who were depressed (defined as having a PHQ-9 score of >10) incurred greater mean costs to the NHS than those patients who were defined as not depressed (£4380 vs £1503, t(83)=−3.25, p<0.001). The same phenomenon was true for patients defined as anxious (GAD-7 score of >10) versus those defined as not anxious (£4017 vs £1980, t(82)=−2.1, p<0.001).

Many participants incurred substantial out-of-pocket expenses in the form of adaptations made to residences for the purpose of disabled access. The mean out-of-pocket expense of these participants who made modifications was £3499.47 (±£5299.60), while the mean across the full cohort was £570.85 (±£2446.71).

Home-based services are summarised in table 3. A small minority of patients used these services, which perhaps highlights the skewed distribution of the health-resource use of patients with FND.

Table 3

View inline|View popup

Utilisation of home-based services and cost in the 6 months prior to new appointments (n=110)

Total costs to the NHS per patient are displayed in figure 1. These costs include all costs listed in tables 1 and 2, as well as transport to NHS appointments and visits by the district nurse. The mean cost per patient was £3229 (±5,395.93), with a median value of resource use of £1152.27. 12.68% of respondents reported costs of over £5000, predominantly due to inpatient admission.

Request permission

Figure 1

Costs to National Health Service by patient in the 6 months prior to new appointments.

Indirect costs

There was a substantial cost of lost income in the cohort, which was calculated as estimated annual income prior to the onset of their FND, less annual income after onset (table 4), estimated at a total of £758 355 among 115 patients. This represents a mean of £6594.4 (±8503) among all patients. Excluding participants who were unemployed prior to symptom onset, the mean loss of income was £10 821.91 (±8306). The amount of income lost per patient is shown in figure 2.

Table 4

View inline|View popup

Change in employment status due to their FND (n=115)

Request permission

Figure 2

Annual income loss per patient due to their functional neurological disorder.

Only 16.5% of study participants were able to maintain full-time employment, with another 5.2% employed on a part-time basis. Of the 54 and 12 people in full-time and part-time employment before the onset of FND, only 19 (35%) and 6 (50%) people, respectively, remained in employment. With their lack of income from employment, many patients became reliant on government benefits to supplement or replace their income. Of the cohort, 71.8% received welfare benefits over the preceding 6 months, with the mean amount received being £299.50 (±180.76) per week.

Loss of productivity affected not only patients but also their carers, friends and family, as shown in table 5. Patients estimated receiving a mean of almost 20 hours per week (median 13.75 hours per week) of informal care.

Table 5

View inline|View popup

Weekly hours and cost of informal care (n=100)

Intangible costs

The distribution of EQ-5D scores is shown in figure 3. There was no significant relationship between the duration of the disorder and EQ-5D score (p=0.36, Spearman’s r=0.13). After the removal of outliers, defined as >3 SD from the mean, there was a significant relationship between the duration of symptoms and total cost to NHS in the prior 6 months (p=0.04, Spearman’s r=0.226). There was no significant relationship between the severity of symptoms and total cost to NHS in the prior 6 months (p=0.063, Spearman’s r=0.218).

Request permission

Figure 3

Distribution of total EQ-5D scores by patient. EQ-5D, EuroQoL-five dimension.

Discussion

This study highlights the high healthcare costs of patients with FND. These patients were found to have a mean utilisation of health resources valued at £3229 over the 6-month period (£6458 p/a) prior to their initial appointment at the tertiary neuropsychiatry service. Extrapolation of this mean value using an estimated incidence of 4–12 per 100 000 per year7 24 25 gives a total cost of NHS resource use of between £13.5 million and £40.4 million per year. This estimate is nearly eight times that of the national health expenditure per person, almost twice that of estimates of the cost per patient of chronic obstructive airway disease (£3488 p/a) and almost four times that of depression (£1873 p/a) and diabetes (1870 p/a).26 Such comparisons are, however, limited by heterogeneous methods of cost estimation. Jennum et al11 compared the cost of patients with FS to age-matched and location-matched controls and found direct healthcare costs to be 4.8 times greater in the FS group. Both findings demonstrate the high direct healthcare costs of people with FND.

The distribution of these direct costs was positively skewed, resulting in a small number of patients requiring the use of costly interventions, including admissions to hospitals and intensive care units. The most frequently used services were outpatient services, particularly GPs. However, as shown in tables 1 and 2, the most costly resource utilisations in the cohort were, in decreasing order, admission to an intensive care unit, admission to a medical in-patient ward, GP appointments and emergency department visits.

As in other cost of illness analyses, it is difficult to isolate the ‘pure’ cost of FND, that is, the cost that does not stem from any comorbid conditions. Any cost estimates reported in this study represent the yearly direct and indirect costs of patients with FND. Attempting to assess such a pure cost may be an exercise in futility, given the nature of the interaction of FND with its psychiatric comorbidities. Whether FND symptom severity and outcome are positively or negatively affected by a mood or anxiety disorder is unclear.27 28 Our findings suggest that, in any case, symptom severity is not correlated with higher health costs. Our findings of increased costs for patients with FND who also suffer from depression and anxiety may indicate only an added cost of these two disorders, which has been described in the literature,29 30 or they may point to a synergistic relationship. The investigation of this question is beyond the scope of a self-reported, retrospective review but may offer an avenue for future research.

As is the case in previous studies that investigated the indirect costs of FND,8 our findings show that the indirect costs of the disorder dwarf the already considerable direct costs. Total indirect costs per patients were a mean of £15 850. Such indirect costs are borne by both patients and their family or friends, as well as by taxpayers, in supporting patients who are no longer able to gain money from employment. Such high indirect costs are compounded by patients with FND having worse outcomes when in receipt of government welfare.31

Comparing our findings to the literature on the economic costs of FND is challenging given the geographical, clinical and methodological heterogeneity of the studies in this area. Looking specifically at studies investigating adults with FND in countries with a similar public healthcare system, in patients with psychogenic non-epileptic seizures (PNES), Goldstein et al32 in the UK found similar healthcare utilisation costs (£3943), but substantially lower productivity loss (£2953) in a cohort of n=367 in the 6 months prior to treatment. Magee et al33 in Ireland assessed the cost of PNES to taxpayers and reported direct costs of €2714.5 per 6 months, with combined social welfare payments and loss of tax revenue costs calculated at €7783 per 6 months per person. Deleuran et al in Denmark found direct healthcare costs of €2904 over 6 months in patients with FS.34 Finally, Tinazzi et al in Italy reported an average direct hospital cost of €1652 per 6 months for patients referred to a specialist functional movement disorder clinic.35

Limitations

We acknowledge some study limitations. First, the use of the patient-reported CSRI is liable to recall bias. While this limitation should be considered, the results of a 2005 study suggest that retrospective self-report data can be equally reliable as medical database data.39 Also of note, the data gathered on participants’ medication use through self-reporting had a surprisingly low completion rate. Rather than this signifying that fewer patients than expected were taking medication, it could be that participants did not complete this section due to a lack of knowledge of the names of their medications. Therefore, one can argue the data on medication should be treated as a minimum possible value.

A second limitation is the low return rate of the questionnaire, resulting in a relatively small sample size. Furthermore, the large number of questionnaire non-responders (63.99% of the intended cohort) could signify a selection bias in the study, limiting the external validity of these findings.

Third, the patients included in our study are those referred to a tertiary specialist service, and as such, they likely represent a severely affected cohort. Such referral bias would also limit the external validity of the study’s results.

Fourth, our study would ideally have used a comparator group so that the costs associated with FND could have been contextualised. Failing the use of a comparison with healthy controls or another neurological disorder, one possibility might have been to compare the costs of patients before and after their diagnosis of FND, as there has yet to be a study comparing indirect costs before and after patients receive a diagnosis of FND. This is an ongoing project.

Finally, the costs estimated in this study may differ and be lower than those in other countries, such as Australia40 and the USA,41 where unit costs may differ for varying aspects of healthcare.