Edinburgh Research Explorer

Screening for functional neurological disorders by questionnaire

Citation for published version:

Shipston-sharman, O, Hoeritzauer, I, Edwards, M, Reuber, M, Carson, A & Stone, J 2019, 'Screening for
functional neurological disorders by questionnaire', Journal of Psychosomatic Research, vol. 119, pp. 65-73.
https://doi.org/10.1016/j.jpsychores.2019.02.005

Digital Object Identifier (DOI):

10.1016/j.jpsychores.2019.02.005

Link:
Link to publication record in Edinburgh Research Explorer

Document Version:
Peer reviewed version

Published In:
Journal of Psychosomatic Research

General rights
Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s)
and / or other copyright owners and it is a condition of accessing these publications that users recognise and
abide by the legal requirements associated with these rights.

Take down policy

The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer
content complies with UK legislation. If you believe that the public display of this file breaches copyright please
contact [email protected] providing details, and we will remove access to the work immediately and
investigate your claim.

Download date: 11. Sept. 2024

1 Screening for Functional Neurological Disorders by
2 Questionnaire
3
4 Running head
5 The Edinburgh Neurosymptoms Questionnaire
6
7 Authors
8 Oliver Shipston-Sharman1, Ingrid Hoeritzauer1, Mark Edwards2, Markus Reuber3, Alan
9 Carson1,4, Jon Stone1.
10
11 Author Affiliations
12 1. Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh, United Kingdom
13 2. Neuroscience Research Centre, Institute of Molecular and Clinical Sciences, St George’s
14 University of London, London, United Kingdom
15 3. Academic Neurology Unit, University of Sheffield, Royal Hallamshire Hospital, Glossop
16 Road, Sheffield, S10 2JF, United Kingdom
17 4. Scottish Neurobehavioural Rehabilitation Unit, Royal Edinburgh Hospital, Edinburgh,
18 United Kingdom
19
20 Corresponding Author
21 Jon Stone; [email protected]; +44 (0)131 537 1167; The University of Edinburgh, Centre
22 for Clinical Brain Sciences, Chancellor's Building, 49 Little France Crescent, Edinburgh, EH16
23 4SB.
24
25 Abstract: 232 words
26 Article: 4257 words
27
28 Target: Full length paper in The Journal of Psychosomatic Research
29
30
31 Abstract
32 Objective: Diagnostic screening for functional neurological disorders (FNDs) continues to
33 pose a challenge. Simple symptom counts fail clearly to discriminate patients with FND but
34 there is increasing recognition of ‘positive’ features which are useful diagnostically during
35 face-to-face assessments. A self-completed screening questionnaire evaluating specific
36 features of FNDs would be useful for screening purposes in clinical and research settings.
37 Methods: The Edinburgh Neurosymptoms Questionnaire (ENS) is a 30-item survey of
38 presence and nature of: blackouts, weakness, hemisensory syndrome, memory problems,
39 tremor, pain, fatigue, globus, multiple medical problems, and operations constructed via
40 literature review and expert consensus. We conducted a pilot of the ENS on new general
41 neurology clinic attendees at a large regional neuroscience centre. Patients were grouped
42 according to consultant neurologist impression as having symptoms that were ‘Not at
43 all’, ’Somewhat’, ’Largely’ or ’Completely’ due to a functional disorder.
44 Results: Blackouts, weakness and memory questions provided reasonable diagnostic utility
45 (AUROC = 0.94, 0.71, 0.74 respectively) in single symptom analysis. All other symptoms
46 lacked discriminating features. A multivariate linear model with all symptoms predicted
47 functional classification with moderate diagnostic utility (AUROC = 0.83), specificity of 0.97,
48 sensitivity of 0.47. Pain and blackout scores provided the most accurate predictor of
49 functional classification.
50 Conclusion: The ENS questionnaire provides some utility in differentiating patients
51 presenting with functional blackouts but failed to provide diagnostic value in other types of
52 FND highlighting the limitations of this self-report tool.
53
54 Key Words: Functional Neurological Disorders, Symptom Count, Screening Questionnaire,
55 Neurological Symptoms, Neuropsychiatry.
56
57 Highlights:
58 • A novel screening questionnaire for functional neurological disorders (FNDs).

59 • Gross symptom count provided no diagnostic utility in FNDs (AUC = 0.60).

60 • Questions regarding positive features of FND provide modest utility (AUC = 0.83).
61 Introduction
62 Functional Neurological Disorders (FNDs) have historically been considered a common but
63 challenging diagnosis [1], with a considerable impact on patient quality of life [2]. They are
64 characterised by a deficit in neurological functioning rather than a pathophysiological lesion
65 which may affect any faculty including movement, sensation or cognition. Patients with
66 symptoms without a pathophysiological cause comprise 30% of general neurology
67 outpatients [3] and between 16-34% of primary care attendees [4–6]. They are commonly
68 undiagnosed [7–10], over-investigated [7,11,12], and report poor clinical outcomes
69 [2,13,14].
70
71 Although challenging for a variety of reasons [7], there is a growing body of
72 literature describing the reliable diagnosis of FNDs if undertaken by clinicians appropriately
73 trained in neurological assessment [15]. It is a diagnosis based upon positive signs of
74 neurological deficit, inconsistent with pathophysiologically explained neurological disease.
75 Examples include: Hoover’s Sign, in which a deficit in voluntary hip extension is reversed
76 with contralateral hip flexion; the tremor entrainment test, in which the frequency of
77 tremor may be entrained to that of an externally cued rhythmic movement of the
78 contralateral arm. Recent work [16–18] has described the diagnostic value of these and
79 other signs, which in a pilot sample provided specificities and sensitivities of 100% and 95%
80 respectively, for a variety of functional disorders [19]. However, the dependency of
81 diagnosis being based on a clinical assessment by an experienced clinician trained in
82 neurological examination, limits reliability of diagnosis in primary care [20], and is financially
83 prohibitive to the conduct of large cohorts studies. Ideally a brief questionnaire is needed
84 with acceptable specificity and sensitivity for community epidemiology and to improve pre-
85 test probability in primary care, but no such scale currently exists.
86
87 There have been several self-report questionnaire approaches to assessing somatic
88 symptoms [21], the Patient Health Questionnaire-15 (PHQ-15) [22] being perhaps the most
89 widely used, including in the validation of DSM-5 cross-cutting assessments [23,24]. These
90 scores, although not initially intended for diagnostic use, have been applied [25,26] to the
91 prediction of somatoform disorder, but seldom tested against gold standard clinical
92 assessments. In FNDs specifically however, these tools fail to discriminate
93 pathophysiological or “organic” from functional neurological disorders and perform little
94 better than chance when tested against clinical examination by a neurologist [27]. The
95 performance of such symptom counts was not enhanced by the addition of items measuring
96 various features of psychopathology.
97
98 Alternate approaches assessing specific clinical features by questionnaire have been
99 more promising. Self-reported features of transient loss of consciousness using an extensive
100 86-item tool could predict, with accuracy, a diagnosis of syncope, psychogenic non-epileptic
101 seizures and epilepsy with sensitivities and specificities ranging from 80-95% and 74-93%
102 between diagnoses [28]. Erba et al [29] similarly showed diagnostic utility in a range of self-
103 reported seizure features in patients with epilepsy vs psychogenic non-epileptic seizures,
104 including: triggering headache; premonitory racing heart or numbness/tingling; post-ictal
105 physical pain and a history of head injury with loss of consciousness > 5mins, physical abuse
106 or fatigue. There have so far been no attempts to construct a short, self-report
107 questionnaire for the prediction of a functional neurological disorders in general. Such a
108 questionnaire could be used to increase pre-test probabilities of a functional disorder
109 diagnosis and assist in epidemiological research. We would not expect that a questionnaire
110 would, or should, replace clinical diagnosis.
111
112 We therefore piloted a 30-item questionnaire that synthesised recognised diagnostic
113 features of the neurological history in people with FND with the aim of developing a
114 screening tool.
115

116 Methods
117 Patients

118 We recruited from consecutive, newly referred general neurology patients who attended a
119 clinic appointment at the Department of Clinical Neurosciences, Western General Hospital,
120 Edinburgh in a 4-week period between September and October 2017. Prospective
121 participants were sent an information letter in the post with their appointment describing
122 the aims and nature of the study. All patients were approached and consented in the
123 waiting room. Patients were excluded if: they were under 16, they did not attend their
124 appointment, they had cognitive impairment or insufficient English language skills to
125 provide informed consent or completion of the survey. Ethical approval for the study was
126 granted by South East Scotland Research Ethics Committee.

127 Survey Design

128 Expert consensus between authors JS, ME, MR, IH and AC with extensive clinical experience
129 of the patients and a literature review [30] was used to construct a 30-item questionnaire of
130 possible discriminating questions (Appendix A) which could be completed in under 10
131 minutes. Symptom features identified from the literature with evidence of positive
132 diagnostic utility were:
133 - Blackouts: Lying still or shaking; Episodes in a medical setting [31]; More than two
134 seizures lasting more than 10 minutes [32–34]; Ability to hear but not respond during a
135 blackout [18]; Pre-ictal dissociative symptoms [35]; Postictal crying/upset [32].
136 - Weakness: Dropping things frequently; Variable severity; Worsening of weakness with
137 attention [36]; Prodromal anxiety [37,38]; Associated depersonalisation [39];
138 - Memory Problems: Forgetting important details of everyday life[40]; Blank spells
139 occurring during the day [40]; Oneself more bothered than others;
140 - Tremor: Sudden onset [41]; Precipitating traumatic event [37]; Variable severity [41];
141 Distractibility [42].
142 - Pain: Variable location and severity [43].
143 - Fatigue: Worsened by activity [43].
144 Patients only had to complete sub-questions regarding a symptom if they had reported
145 experiencing the symptom as a “stem” question.
146
147 We also included questions about the presence of certain symptoms and features of
148 clinical history that in themselves may be predictive of a functional disorder. These included
149 hemisensory syndrome (‘Do you have numbness or altered sensation that makes you feel
150 like your body is cut in half?’) [44], globus [45], stutter [46,47], multiple medical problems
151 [48], and particular operations such as hysterectomy, appendicectomy, laparoscopy or
152 tonsillectomy [49,50]. These items did not have differentiating sub-questions. Demographic
153 data including sex and age were also collected.
154

155 Diagnosis and Rating of explanation with respect to functional disorder

156 We asked neurologists to provide 1) their provisional diagnosis and 2) their assessment of
157 the extent to which the patients’ symptoms were related to a functional disorder.
158 Functional neurological disorders remain a taxonomic challenge and often exist in a
159 spectrum, concomitant with pathophysiological disease. For this reason, patients were
160 scored according to a 4-point Likert scale: ‘Not at All’, ‘Somewhat’, ‘Largely’ and
161 ‘Completely’ by clinicians in response to the question: “To what extent do you think the
162 patient’s clinical symptoms are explained by a functional disorder?”. Wording of this
163 question encompassed the entire clinical presentation not just the presenting symptom.
164 Definitions of functional disorders were supplied to clinicians as a guide to diagnostic
165 categorisation (Appendix B). A graded classification like this allows for a broader evaluation
166 of patients which may have symptoms without a pathophysiological cause but not a primary
167 functional diagnosis. Note this question was an evolution of previous categorisations from
168 our research group as 'not explained by disease' [3]. We were keen to move away from
169 defining disorders by the absence of disease since they have their own positive diagnostic
170 features, now recognised in DSM-5 criteria for Functional Neurological Symptom Disorder.
171 Clinical assessment of these features by an experienced neurologist is the current diagnostic
172 gold standard for FND [51], with misdiagnosis rates quantified as 0.4% in a large cohort of
173 new neurology outpatients [3].

174 Questionnaire Analysis

175 For the purposes of analysis, patients were grouped into having symptoms classed as ‘Not at
176 all/Somewhat’ and ‘Largely/Completely’ due to a functional disorder. Univariate analysis
177 was undertaken on individual questions by cross-tabulation and significance testing using
178 Fisher’s Exact test. Symptom and gross ENS score were assessed using two-tailed Student’s
179 t-tests. Multivariate analysis was undertaken via logistic regression. We first analysed the
180 diagnostic utility of sub-questions in predicting classification of ‘Largely’ or ‘Completely’
181 functional for reporters of a particular symptom. Linear models for each symptom were
182 used to return a score for likelihood of functional classification. Scores from these
183 symptoms were then combined in an aggregate model with symptoms and features that did
184 not have sub-questions and demographic data to provide an overall score. This method
185 introduces a significant positive bias into the second round of modelling, as symptoms with
186 sub-questions have already been weighted towards predicting a functional outcome.
187 Alternative options such as hierarchical logistic regression and stratifying patients by
188 reported symptoms were prohibited by sample size and the number of potential symptom
189 combinations. We justify this method as exploratory and speculative in the context of a pilot
190 that aims to obtain a broad picture of the potential utility of a general screening tool.
191 Questions which provided perfect or quasi-separation were excluded from multivariate
192 analysis and their contribution assessed during univariate analysis only. All analysis was
193 conducted in MATLAB© Release 2015b using custom written scripts.

194 Results
195 Data were gathered on 165 patients, 56 (34%) participants had data missing and were
196 excluded leaving 109 (Age = 44.6 ± 17.1 years; Female:Male Ratio = 1.53:1) responses
197 available for analysis. 104/109 (95%) of those surveyed responded having at least one of the
198 symptoms included in the questionnaire.
199
200 73/109 (67%) patients were classed as having symptoms ‘Not at All/Somewhat (N/S)’
201 and 36/109 (33%) as ‘Largely/Completely (L/C)’ due to a functional disorder. The most
202 common diagnoses made in those classified as ‘Not at All/Somewhat’ were: Epilepsy 16/109
203 (15%), Migraine 11/109 (10%), peripheral neuropathy or radiculopathy 9/109 (8%),
204 headache syndromes 6/109 (6%), first seizure 6/109 (6%) and demyelinating disease 5/109
205 (5%). In those classified as ‘Largely/Completely’: dissociative seizures 9/109 (8%), functional
206 weakness 3/109 (3%), functional sensory changes 3/109 (3%), anxiety related symptoms
207 3/109 (3%), functional memory symptoms 1/19 (1%) and FND not otherwise specified 2/109
208 (2%) were the most common diagnoses. Female:Male ratio differed significantly between
209 groups (N/S = 1.09:1; L/C = 3.5:1; Fisher’s Exact p = 0.01) whilst age did not (N/S = 46 ± 17.5;
210 L/C = 41.6 ± 16.2; two-tailed Student’s t-test p = 0.20).
211
212 The 56 participants excluded from analysis due to incomplete questionnaires or
213 consultant diagnosis were marginally older than those included (47.15 ± 17.1 vs 44.6 ± 16.83
214 years; Student’s t-test p = 0.36) and had a greater F:M ratio (2.31:1 vs 1.53:1; Chi-square p =
215 0.72). 15/56 were excluded for lack of diagnostic outcome data, of those remaining 28/41
216 (68%) were classed as having symptoms ‘Not at all/Somewhat’ due to a functional disorder
217 and 13/41 (32%), similar proportions to those included in analysis (Chi-square p = 0.88).
218

219 Univariate Analysis: Few questions provide diagnostic utility and gross scores fail to
220 discriminate patients.

221 Answers to all symptom questions and sub-questions are displayed in Table 1. Some
222 symptoms were reported significantly more frequently by those classed as
223 ‘Largely/Completely’ functional, including: hemisensory disturbance (N/S = 8/73 (11%); L/C
224 = 11/36 (31%); p = 0.02), tremor (N/S = 19/73 (11%); L/C = 17/36 (31%); p = 0.02), pain (N/S
225 = 24/73 (33%); L/C = 22/36 (61%); p = 0.007), fatigue (N/S = 40/73 (55%); L/C = 28/36 (78%);
226 p = 0.02).
227
228 5/20 symptom features were reported significantly more often by patients classed as
229 ‘Largely/Completely’ related to a functional disorder including: having had a blackout in a
230 medical setting (N/S = 1/21 (5%); L/C = 5/9 (56%); p = 0.005); being able to hear others but
231 not respond during a blackout (N/S = 5/21 (24%); L/C = 8/9 (89%); p = 0.002); crying or being
232 upset after a blackout (N/S = 5/21 (24%); L/C = 6/9 (67%); p = 0.04); having blank spells
233 occurring throughout the day if also experiencing memory problems (N/S = 12/39 (31%); L/C
234 = 15/22 (68%); p = 0.007) and experiencing pain that is variable in severity and location (N/S
235 = 10/24 (42%); L/C = 16/22 (73%); p = 0.04).
236
237 Gross symptom count was significantly different between ‘N/S’ and ‘L/C’ patients
238 (N/S = 3.15 ± 2.07; L/C = 4.33 ± 2.27; 2-Tailed Student’s t-test p = 0.008) (Figure 1A) but
239 without diagnostic utility (Receiver-operator characteristic area under the curve (AUC) =
240 0.595). Raw Edinburgh Neurosymptom Score (ENS) scores, which include the addition of
241 sub-questions designed to provide a positively discriminating score, yields greater gross
242 scores for ‘L/C’ patients, again significantly so (N/S = 7.95 ± 5.48; L/C = 11.69 ± 7.27; 2-Tailed
243 Student’s t-test p = 0.003) (Figure 1B) but again without diagnostic utility (AUC = 0.602).
244

245 Multivariate sub-question analysis: Blackouts may be amenable to questionnaire

246 diagnosis, but other symptom groups lack discriminating questions.

247 Logistic regression analysis of individual “common” symptoms is described in Figure 2. Only
248 three sub-questions obtained significance during multivariate analysis. Q1d: “Have you ever
249 been able to hear people but not respond to them during your blackout?” (p = 0.047; OR =
250 20.72 (0.88-487.97)), Q4c: “Do you have blank spells which occur during the day?” (p = 0.02;
251 OR = 4.066 (1.23-13.45)), and Q6a: “Is your pain worse in different parts of your body on
252 different days?” (p = 0.04; OR = 3.73 (1.04-13.37)). Diagnostic utility (AUC) of sub-questions
253 for each symptom were: blackouts = 0.94, weakness = 0.71, memory problems = 0.74,
254 tremor = 0.63, pain = 0.66 and fatigue = 0.6.
255

256 Aggregate symptom score modestly predicts functional classification.

257 Scores from symptom sub-question modelling were input into an aggregate model with
258 other symptoms, features of clinical history, sex and age. Variable coefficients for the
259 resulting model are shown in Figure 3. Only adjusted pain score (p = 0.047) and adjusted
260 blackout score (p = 0.02) achieved significance in the model, with odds ratios 26.80 (2.00-
261 359.59) and 40.15 (1.73-930.21) respectively.
262
263 Resulting aggregate scores were capable of predicting functional disorder likelihood
264 with modest utility (Figure 4) (AUC = 0.83) and “optimal” operating point, as determined by
265 minimising false positive rate, resulting in specificity and sensitivity of 0.99 and 0.47
266 respectively. Positive and negative predictive values were 0.94 and 0.79. The model
267 accounted for little of the variability in the outcome (Adjusted R2 = 0.23) but performed
268 better than the constant model (Chi-square test vs Constant model p < 0.001).
269
270 Symptom ‘networks’ may aid in differentiating patients with a functional disorder.

271 As sample size precluded interactional analysis between reported symptoms, and as FND
272 encompasses a heterogenous collection of symptoms, we sought to characterise the
273 possible predictive utility of symptom pairs, thereby providing a coarse assessment of
274 symptom interaction. That is, if a patient reports more than one symptom what symptom is
275 that likely to be? Are there basic ‘syndromes’ that particularly delineate patients with a
276 functional disorder? Of the 110 possible bidirectional symptom pairings, patients classed as
277 ‘Largely/Completely’ functional were more likely to report one symptom after reporting
278 another when compared to those classed as ‘Not at All/Somewhat’ in 76/110 pairings. This
279 reflects the greater average symptom counts in this group. Figure 5 exhibits how fatigue
280 plays a central role in these interactions, being reported by more than 80% of those also
281 reporting: stutter, memory problems, pain, weakness, blackouts, globus, altered sensation,
282 tremor and multiple medical problems. Only one symptom pair (P(Memory problems |
283 Multiple medical problems)) reaches this threshold in those with symptoms not explained
284 by a functional disorder and none do so when paired with fatigue.

285 Discussion
286 This is the first reported pilot of a general screening questionnaire to improve the pre-test
287 probability of a diagnosis of functional neurological disorders. We found that the total
288 number of symptoms, in the subset we investigate here, failed to distinguish cases from
289 controls, replicating a previous study [27]. The addition of items in our novel questionnaire
290 about features reportedly specific to functional disorders also commonly failed to
291 distinguish patient groups in our sample. There were however some exceptions, where
292 patients classified as having functional symptoms more commonly reported features of:
293 Blackouts (having had a blackout in a medical setting, being able to hear people but not
294 respond during a blackout, being upset following an episode); Memory problems (having
295 associated blank spells during the day); Pain (reporting variability in bodily location and
296 severity.
297
298 Symptoms scores weighted according to these features in an aggregate model show
299 good specificity (0.99) but poor sensitivity (0.47) when compared to consultant neurologist
300 impression as measured on a 4-point Likert scale. Resulting positive and negative predictive
301 values (0.94 and 0.79 respectively) were however, promising, and had greater utility as a
302 pre-screening diagnostic tool for FND than measures based on symptom counts such as
303 PHQ-15 [25,27]. Although more effective for excluding those deemed to have symptoms of
304 an “organic” cause, our linear score failed to reliably identify patients with FND from a
305 general neurology outpatient population. Our speculative assessment of symptom
306 interactions suggests that non-linear methods that take account of multivariate higher order
307 interactions may prove a more valuable approach.
308

309 Eliciting self-reported positive features of functional disorders is challenging.

310 Although many discriminating features of history have been described in the literature and
311 anecdotally, our data show that these are difficult to translate into specific and sensitive
312 questions for patients to answer in an unguided way. A comparative analysis of self-
313 reported vs clinical record extracted seizure features [52] recently highlighted that using a
314 self-report questionnaire [29] is associated with a greater quantity of reported features, and
315 greater detail regarding premonitory or triggering features as compared to clinician enquiry
316 which was more effective at eliciting historical predictors. In keeping with our data, these
317 questions also showed generally good specificities and poor sensitivities. The corollary being
318 that although our understanding of the semiology and history of functional symptoms has
319 improved, the ability to extract that from patients with a functional disorder in a meaningful
320 way is still the remit of an experienced diagnostic interview and physical examination. This is
321 reflected in diagnostic criteria for functional neurological symptoms disorder in DSM-5
322 which mandates the importance of physical signs typical of the disorders, and not the
323 subjective experience of the patient.
324
325 Capturing the recognised linguistic features of FND descriptions is a core problem in
326 constructing a viable self-reported screening questionnaire. There is now a significant body
327 of work highlighting these discriminating features: Poor formulation effort [53], inconsistent
328 metaphorical conceptualisation [54], and vague seizure experience descriptions in
329 psychogenic non-epileptic seizures; preserved working memory, the ability to process
330 compound questions and good recollection of personal information in functional memory
331 disorders [55]; post-exertional malaise in fatigue [56]. However, those studies were all done
332 on the basis of interactive conversation analysis. Self-report tools implicitly rely on a
333 particular symptom being amenable to self-recognition. Transposing clinical observations
334 into questions capable of eliciting introspection and ‘accurate’ response is a clear limitation
335 to such an enquiry. It may be that questionnaire items need to be refined or that
336 questionnaires are, themselves, too crude a tool.
337
338 Perhaps a surprising finding in this population is that questions regarding functional
339 symptoms such as globus and stutter show poor diagnostic utility in both univariate and
340 multivariate analysis. Although globus and adult onset stutter are generally considered to
341 relate to a functional disorder [57] they were reported with similar frequency in both
342 functional and non-functional groups, albeit in small numbers. There were also interesting
343 responses in those with symptoms unexplained by a functional disorder to questions that
344 are commonly associated with functional disorders. For example, 8 out of 73 patients
345 reported that they had numbness or altered sensation that made them feel ‘like your body
346 is cut in half’ [44] and 5 out of 21 patients reported tearfulness after blackouts [18].
347 Questions about movement disorders also indicated the difficulty of using questionnaires to
348 elicit a history. All 19 patients who reported an abnormal movement such as tremor in the
349 structural group said it came on suddenly. But what a neurologist understands as sudden,
350 for example not there at 10.58am and present at 11.00am – may not be the same as how a
351 patient understands that word – ‘I didn’t have it last year and suddenly this year I do’. It was
352 also surprising how many movement disorder patients said that their movements could go
353 away for hours or days (16/19).
354

355 The importance of diagnostic tools and more effective diagnostic procedures in FNDs

356 A standardised and easily administrable tool for the screening of functional disorders has
357 the potential to enhance clinicians’ pre-test probability for making a diagnosis of functional
358 disorder and, as a consequence of earlier intervention, reduce iatrogenic harm. A shorter
359 duration of symptoms prior to diagnosis often predicts a favourable prognosis in FNDs
360 [2,14]. Early identification of patients with likely functional symptoms could also assist in
361 quantifying their prevalence and demographics at an epidemiological scale. So far this has
362 been unattainable with the present non-specific tools and the expense of definitive clinical
363 diagnosis.
364

365 Limitations

366 This was a pilot study of a new approach to FND diagnosis, with a relatively small sample
367 size. Our reported predictive values are dependent on prevalence calculated on a relatively
368 small population which, for certain symptoms, failed to meet the generally accepted rule of
369 5-10 participants per predictor variable [58]. The large variances observed during linear
370 modelling may be a reflection of this, or a reflection of the variable nature of functional
371 disorders. There is a risk that some patients were classified in to the wrong diagnostic group
372 by the neurologists seeing them, although a similar study found a very low rate of
373 misdiagnosis at 18 months follow up [3]. We also don’t know whether, even if the
374 neurologist rated the main diagnosis as “organic”, the symptom the patient gave their
375 responses about would have received the same rating. We are also cautious to highlight the
376 limitations of the present two-stage modelling. Ideally, sub-question coefficients should be
377 computed on a separate population from the overall aggregate score to prevent a
378 significant bias in favour of symptoms with sub-questions in the final model.
379
380 Our final model is biased to a degree by case deletion of those with incomplete
381 questionnaires. 109 individuals were included in the final analysis, with 56 (34%) of the 165
382 participants excluded. Given this significant proportion, we sought to establish whether
383 their inclusion in analysis might mitigate some of the bias case deletion introduces. As we
384 first model symptom sub-questions on a subset of those reporting that symptom, we were
385 able to include every participant who had at least answered a single symptom’s sub-
386 questions completely in the first stage of modelling. Using symptom scores derived from
387 this more inclusive criterion, we then reran the aggregate model with the 109 respondents
388 who had complete questionnaires. Resulting sub-question coefficients were similar with
389 Q1d: “Have you ever been able to hear people but not respond to them during your
390 blackout?” and Q4c: “Do you have blank spells which occur during the day?” remaining
391 significant with p values in the new model 0.039 and 0.006 respectively. And Q6a: “Is your
392 pain worse in different parts of your body on different days?” becoming less significant (p
393 = 0.05). In the final aggregate model, blackout scores become insignificant (OR = 7.97 (0.57-
394 111.68)) but pain scores remain predictive (OR = 21.87 (1.34-358.05). Aggregate scores
395 however retain similar discriminate utility (AUC = 0.80) and sensitivity of 0.64 and specificity
396 of 0.84 at the ‘optimal’ operating point.
397
398 We also found that many of our questions, or question wordings, although
399 constructed to elicit positive answers in those experiencing functional symptoms, failed to
400 do so on many occasions. Only blackouts, memory problems and pain domains had sub-
401 questions answered significantly more often by patients deemed ‘Largely/Completely’
402 functional. The heterogeneity of both FND and neurological pathology in general may be the
403 limiting factor to such a broad goal. Future approaches to this issue would require more
404 rigorous testing of questionnaire comprehension, investigating educational background and
405 reading level of the participants. It is clear that if the present tool is to be developed, and
406 sensitivities greater than 0.47 are to be achieved, question wording and inclusion needs to
407 be adjusted considerably.
408
409 Readers may also wonder why we didn’t study the performance of the relevant
410 subsections of the questionnaire for diagnostic categories (for example functional gait
411 disorder, non-epileptic seizures). This was firstly because the numbers involved would have
412 been too small and secondly because patients with functional neurological disorders often
413 have mixed symptoms which are not always picked up on diagnostically by neurologists.
414

415 Conclusions
416 Despite limitations, this pilot version of an ENS questionnaire was, in its complete form,
417 surprisingly capable of reliably excluding patients diagnosed by neurologists as not having a
418 functional disorder. Although showing utility in capturing functional/dissociative blackouts,
419 we failed to distinguish many other symptoms, highlighting the linguistic and interpretive
420 difficulties in eliciting functional vs structural symptom experience. The use of specific
421 positive features of functional disorder in an aggregate model rather than linear summation
422 of symptom counts has shown promising utility. Future work could aim to investigate more
423 systematically how those who experience functional symptoms, outside the domain of
424 blackouts, report their disorder and therefore how to improve question wording.
425

426 Acknowledgements

427 Thanks to Dr Declan Ahern, Dr Richard Davenport, Dr Louise Davidson, Dr Christopher Derry,
428 Dr Susan Duncan, Dr Robin Grant, Dr Mireia Moragas-Garrido, Dr Colin Mumford, Dr Belinda
429 Weller and Dr Peter Foley for providing diagnoses and functional classification.

430 Funding

431 IH is supported by an Association of British Neurologists/Patrick Berthoud Charitable Trust

432 Clinical Research Training Fellowship. JS is supported by an NHS Scotland NRS Career
433 Fellowship.

434 Competing Interests

435 All authors have completed the Unified Competing interest form at
436 http://www.icmje.org/coi_disclosure.pdf and declare the following non-financial interests
437 that may be relevant to the present work. ME reports royalties from Oxford Specialist
438 Handbook of Parkinson's Disease and Other Movement Disorders from Oxford University
439 Press and honoraria for educational events from Merz Pharma, Boehringer Ingelheim and
440 UCB. JS reports independent expert testimony work for personal injury and medical
441 negligence claims, royalties from UpToDate for articles on functional neurological disorder
442 and runs a free non-profit self-help website, www.neurosymptoms.org. Dr. Carson gives
443 independent testimony in court on a range of neuropsychiatric topics, is director of Carson
444 (Edinburgh) Ltd, a personal services company for medical reports and is paid editor of
445 Journal of Neurology, Neurosurgery, and Psychiatry.
446 Table 1

Symptoms explained by a func�onal disorder:

Not at All/Somewhat Largely/Completely p-value
N 73/109 (67%) 36/109 (33%)
Sex F:M = 1.09:1 F:M = 3.5:1 0.01*
Age ( Mean ± SD) 46 ± 17.5 41.6 ± 16.2 0.20
Symptom Count ( Mean ± SD) 3.15 ± 2.07 4.33 ± 2.27 0.008**
Gross ENS Score ( Mean ± SD) 7.95 ± 5.48 11.69 ± 7.27 0.003**
Q1: During the last 6 months have you been bothered by blackouts? 21/73 (29%) 9/36 (25%) 0.83
Lie Still: 5/21 (24%) Lie Still: 3/9 (33%)
Q1a: During you blackouts do you get told you lie still or shake? Shake: 13/21 (62%) Shake: 4/9 (44%) 0.67
Unsure: 3/21 (14%) Unsure: 2/9 (22%)

Q1b: Have you ever had a blackout in a medical setting e.g. visiting the 1/21 (5%) 5/9 (56%) 0.005**
hospital/GP/another doctor?

Q1c: Have you had more than two seizures during which you shook without
stopping for more than 10 minutes? (This does not include the time taken for you 2/21 (10%) 2/9 (22%) 0.56
to come round after the seizure had finished)

Q1d: Have you ever been able to hear people but could not respond to them 5/21 (24%) 8/9 (89%) 0.002**
during your blackout?

Q1e: Do you ever have moments before your blackouts of losing track of what is
going on, of “blanking out” or “spacing out” or in some way feeling that you are 13/21 (62%) 9/9 (100%) 0.07
not part of what is going on?

Q1f: Are you told that after an attack you cry or are upset? 5/21 (24%) 6/9 (67%) 0.04*

Q2: During the last six months have you been bothered by weakness in one or more 30/73 (41%) 20/36 (56%) 0.22
limb e.g. arm(s) or leg(s)?

Q2a: Do you drop things frequently? 13/30 (43%) 13/20 (65%) 0.16
Q2b: Does your limb weakness get worse or better at different times of the 14/30 (47%) 10/20 (50%) >0.99
day?
Q2c: Does concentrating on trying to move make the limb weakness worse? 6/30 (20%) 9/20 (45%) 0.11

Q2d: At the start of your limb weakness did you feel your heart pounding or 9/30 (30%) 10/20 (50%) 0.24
did you feel frightened, anxious or very uneasy?

Q2e: Does your weak limb feel like it does not fully belong to you? 13/30 (43%) 11/20 (55%) 0.57

Q3: Do you have numbness or altered sensation that makes you feel like your body 8/73 (11%) 11/36 (31%) 0.02*
is cut in half?

Q4: During the last six months have you been bothered by memory problems? 39/73 (53%) 22/36 (61%) 0.54
Family: 3/39 (8%) Family: 4/22 (18%)
Q4a: Who is most bothered by your memory problems, you or your Me: 32/39 (82%) Me: 16/22 (73%) 0.47
partner/family/friends?
Unsure: 4/39 (10%) Unsure: 2/22 (9%)
Q4b: Are you bothered by forgetting important details such as the name of a 17/39 (44%) 14/22 (64%) 0.18
family member or your PIN number?
Q4c: Do you have blank spells which occur during the day? 12/39 (31%) 15/22 (68%) 0.007**

Q5: During the last six months have you been bothered by tremor or an abnormal 19/73 (26%) 17/36 (47%) 0.03*
movement in one or more limb e.g. arm (s) or leg(s)?

Q5a: Did your tremor or abnormal movement come on suddenly? 19/19 (100%) 15/17 (88%) 0.22
Q5b: Did your tremor or abnormal movement come on after an injury or 2/19 (11%) 3/17 (18%) 0.65
accident?
Q5c: Can your tremor or abnormal movement go away completely for hours 16/19 (84%) 16/17 (94%) 0.61
to days only to return again?

Q5d: Does your tremor or abnormal movement ever stop when you are 3/19 (16%) 5/17 (29%) 0.43
distracted or concentrating on something else?

Q6: During the last three months have you had pain almost every day in more than 24/73 (33%) 22/36 (61%) 0.007**
one part of your body?
Q6a: Is your pain worst in different parts of your body on different days? 10/24 (42%) 16/22 (73%) 0.04*

Q7: Have you been lacking energy every day or almost every day for the last six 40/73 (55%) 28/36 (78%) 0.02*
months?
Q7a: Does activity make your fatigue worse? 25/40 (63%) 23/28 (82%) 0.11

Q8: In the last five years have you had to see doctors in the hospital for different
problems more than four times? (E.g. problems with your heart, your joints, your 27/73 (37%) 16/36 (44%) 0.53
brain and gut)

Q9: Do you get a feeling that there is a lump in your throat or something stuck when 18/73 (25%) 8/36 (22%) >0.99
you are trying to eat or drink?

Q10: Do you have a stutter which started after you were more than 16 years old? 4/73 (5%) 3/36 (8%) 0.68

Q11: Have you needed any operations? 40/73 (55%) 16/36 (44%) 0.42

447
448 Figure 1

A 10 p = 0.008
B p = 0.003
C1
ENS Gross Sum
** **
9 30 Symptom Count

Neurosymptoms Score
8 0.8
25
7
Symptom Count

6 20

Sensitivity
0.6
5
15
4 0.4
3 10
2
5 0.2
1 Symptom Count AUC = 0.595
0 0 Raw ENS Score AUC = 0.602
0
N/S L/C N/S L/C 0 0.2 0.4 0.6 0.8 1

Functional Group Functional Group 1- Specificity

449
450 Figure 1: Comparison of gross scores. A - Boxplot of symptom counts separated by
451 functional classification. Symptom counts are significantly greater in patients with functional
452 disorder. B - Boxplot of gross scores for full 30-point ENS questionnaire. The addition of
453 discriminating sub-questions yields greater scores for ‘Largely/Completely’ explained by
454 functional disorder. C - ROC curve of symptom count and gross sum. Symptom count and
455 raw ENS scores fail to provide diagnostic utility (N/S = Not at All/Somewhat; L/C =
456 Largely/Completely explained by a functional disorder).

457

458

459
460
461
462
463
464
465
466
467 Figure 2

Blackouts N = 30 Weakness N = 50 Memory Problems N = 61

14.59 1.2
Q1f Q2e Q4c
4.07
20.72
Q1d Q2d
1.74
2.41
Q1c
0.16
2.77
Q4b
21.63 Q2c 0.24
Q1b 0.97 Q4a: Not Sure
1.13 Q2b
Q1a: Not Sure 0.28
0.26 Q2a
2.08 Q4a: Me
Q1a: Shake 0.53
(Intercept)
0.03 (Intercept) 0.24 (Intercept)

-5 0 5 -4 -2 0 2 4 -4 -2 0 2 4
log(Odds Ratio) log(Odds Ratio) log(Odds Ratio)

Tremor N = 36 Pain N = 46 Fatigue N = 68

1.99
Q5d
3.73 2.76
2.87 Q6a Q7a
Q5c
2.15
Q5b 0.43 0.33
(Intercept) (Intercept)
0.27
(Intercept)

-4 -2 0 2 4 -4 -2 0 2 4 -4 -2 0 2 4
log(Odds Ratio) log(Odds Ratio) log(Odds Ratio)

Legend:
OR = Odds Ratio > 1 = Odds Ratio < 1

468
469 Figure 2: Results of multivariate sub-question analysis. Sub-questions were input as
470 predictor variables and the resulting coefficients, confidence intervals and odds ratios are
471 displayed above. Only Q1d, Q4c and Q6a achieve significance in their respective models.
472 Most sub-questions provide, as expected, a positive predictive value for functional
473 classification, but only 3 did so with odds ratios significantly greater than 1.
474
475
476
477
478
479 Figure 3
1 Legend:
Age 17-32
OR
0.75
Age 33-48 = Odds Ratio > 1
0.88 = Odds Ratio < 1
Age 49-64
0.73
Age 65-80
3.3
Female Sex
1.42
Multiple Operations
0.39
Stutter
0.44
Globus
0.43
Multiple Medical Problems
1.33
Fatigue*
26.8
Pain*
4.79
Adjusted Scores*

Tremor*
9.93
Memory Problems*
2.71
Altered Sensation*
0.47
Weakness*
40.15
Blackouts*
-4 -2 0 2 4 6
log(Odds Ratio)
480
481 Figure 3: Aggregate score coefficients. Forest plot showing linear coefficients and
482 confidence intervals for each variable in the aggregate model. “Common” symptoms have
483 been replaced by the linear predictor scores from sub-question modelling. Odds ratios are
484 displayed for each coefficient above the bar. Adjusted scores for pain and blackouts achieve
485 significance and drastically increase the odds of correct classification.
486
487 Figure 4
A 1
B 1

Linear Predictor Score

0.8 0.8
Sensitivity

0.6 0.6

0.4 0.4

AUC = 0.83
0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 Not at All/Somewhat Largely/Completely

488 1 - Specificity Outcome

489 Figure 4: Diagnostic utility of the ENS questionnaire. A - ROC curve of aggregate linear
490 model scores predicting consultant classification of patients with symptoms ‘Not at
491 All/Somewhat’ or ‘Largely/Completely’ functional. The optimal operating point is displayed
492 as a red circle on the curve. Predictor scores were capable of achieving an AUC of 0.83. B -
493 Scatter plot of aggregate model scores separated by functional classification. The
494 corresponding optimal score identified in ROC analysis is displayed as a grey dotted line. The
495 model is capable of excluding patients with a pathophysiological disorder effectively, but
496 many with a functional disorder are missed with the ‘optimal’ threshold.
497
498
499
500
501
502
503
504
505
506
507 Figure 5

Largely/Completely Functional Not at All/Somewhat Functional

Memory Problems Memory Problems
Altered Sensation Altered Sensation

Tremor 61%
Tremor 53%
11%
31%

47% Weakness 26% Weakness

56% 41%

Pain Pain
33%
61%

25% 29%

Blackouts Blackouts
78%
55%

Fatigue Fatigue
44% 55%

Multiple Operations Multiple Operations

44% 37%
8%
Multiple Medical Multiple Medical 5%
Problems 22% Problems 25%
Stutter
Globus Stutter
Globus
Key: Symptom
0.5-0.8
Conditional
Probability % of group
reporting
>0.8 symptom

508
509
510 Figure 5: Symptom interactions. Paired conditional probabilities of symptoms occurring if
511 another symptom is reported. Red lines indicate a symptom pair in which there is a more
512 than 80% likelihood of a co-occurrence. Grey lines indicate co-occurrence > 0.5 and are
513 weighted linearly between 0.5-0.8. Patients with functional disorders reported symptom
514 networks that are far more connected than patients with a pathophysiological disorder.
515 Fatigue is more commonly reported as a comorbidity in patients with a functional rather
516 than structural disorder. (Red: Functional class = ‘Largely/Completely’; Blue: Functional class
517 = ‘Not at All/Somewhat’).
518
519
520
521
522
523
524
525

526 References
527 [1] T.R.J. Nicholson, J. Stone, R.A.A. Kanaan, Conversion disorder: A problematic
528 diagnosis, J. Neurol. Neurosurg. Psychiatry. 82 (2011) 1267–1273.
529 doi:10.1136/jnnp.2008.171306.
530 [2] J. Gelauff, J. Stone, M. Edwards, A. Carson, The prognosis of functional (psychogenic)
531 motor symptoms: A systematic review, J. Neurol. Neurosurg. Psychiatry. 85 (2014)
532 220–226. doi:10.1136/jnnp-2013-305321.
533 [3] J. Stone, A. Carson, R. Duncan, R. Coleman, R. Roberts, C. Warlow, C. Hibberd, G.
534 Murray, R. Cull, A. Pelosi, J. Cavanagh, K. Matthews, R. Goldbeck, R. Smyth, J. Walker,
535 A.D. MacMahon, M. Sharpe, Symptoms “unexplained by organic disease” in 1144
536 new neurology out-patients: How often does the diagnosis change at follow-up?,
537 Brain. 132 (2009) 2878–2888. doi:10.1093/brain/awp220.
538 [4] N. Steinbrecher, S. Koerber, D. Frieser, W. Hiller, The Prevalence of Medically
539 Unexplained Symptoms in Primary Care, Psychosomatics. 52 (2011) 263–271.
540 doi:10.1016/j.psym.2011.01.007.
541 [5] M.W.M. De Waal, I.A. Arnold, J.A.H. Eekhof, A.M. Van Hemert, Somatoform disorders
542 in general practice: Prevalence, functional impairment and comorbidity with anxiety
543 and depressive disorders, Br. J. Psychiatry. 184 (2004) 470–476.
544 doi:10.1192/bjp.184.6.470.
545 [6] H. Haller, H. Cramer, R. Lauche, G. Dobos, Somatoform disorders and medically
546 unexplained symptoms in primary care., Dtsch. Arztebl. Int. 112 (2015) 279–87.
547 doi:10.3238/arztebl.2015.0279.
548 [7] A.M. Murray, A. Toussaint, A. Althaus, B. Löwe, The challenge of diagnosing non-
549 specific, functional, and somatoform disorders: A systematic review of barriers to
550 diagnosis in primary care, J. Psychosom. Res. 80 (2016) 1–10.
551 doi:10.1016/j.jpsychores.2015.11.002.
552 [8] J.E. Dimsdale, F. Creed, J. Escobar, M. Sharpe, L. Wulsin, A. Barsky, S. Lee, M.R. Irwin,
553 J. Levenson, Somatic symptom disorder: An important change in DSM, J. Psychosom.
554 Res. 75 (2013) 223–228. doi:10.1016/j.jpsychores.2013.06.033.
555 [9] J.C. Hamilton, M. Eger, S. Razzak, M.D. Feldman, N. Hallmark, S. Cheek, Somatoform,
556 Factitious, and Related Diagnoses in the National Hospital Discharge Survey:
557 Addressing the Proposed DSM-5 Revision, Psychosomatics. 54 (2013) 142–148.
558 doi:10.1016/j.psym.2012.08.013.
559 [10] K. Leaver, E. Yi, B. Evans, K. Bullock, J. Dearlove, Documentation Bias in Functional
560 Neurological Symptom Disorder: Comparing the Prevalence and Documentation of
561 Functional Neurological Symptom Disorder and Parkinson’s Disease (P4.050),
562 Neurology. 86 (2016).
563 http://n.neurology.org/content/86/16_Supplement/P4.050.abstract.
564 [11] J. Shaw, F. Creed, The cost of somatization, J. Psychosom. Res. 35 (1991) 307–312.
565 doi:10.1016/0022-3999(91)90085-3.
566 [12] A. Ring, C.F. Dowrick, G.M. Humphris, J. Davies, P. Salmon, The somatising effect of
567 clinical consultation: What patients and doctors say and do not say when patients
568 present medically unexplained physical symptoms, Soc. Sci. Med. 61 (2005) 1505–
569 1515. doi:10.1016/J.SOCSCIMED.2005.03.014.
570 [13] J. Stone, M. Sharpe, P.M. Rothwell, C.P. Warlow, The 12 year prognosis of unilateral
571 functional weakness and sensory disturbance, J. Neurol. Neurosurg. Psychiatry. 74
572 (2003) 591–596. doi:10.1136/jnnp.74.5.591.
573 [14] M. Sharpe, J. Stone, C. Hibberd, C. Warlow, R. Duncan, R. Coleman, R. Roberts, R. Cull,
574 A. Pelosi, J. Cavanagh, K. Matthews, R. Goldbeck, R. Smyth, A. Walker, J. Walker, A.
575 MacMahon, G. Murray, A. Carson, Neurology out-patients with symptoms
576 unexplained by disease: Illness beliefs and financial benefits predict 1-year outcome,
577 Psychol. Med. 40 (2010) 689–698. doi:10.1017/S0033291709990717.
578 [15] A.J. Carson, S. Best, K. Postma, J. Stone, C. Warlow, M. Sharpe, The outcome of
579 neurology outpatients with medically unexplained symptoms: A prospective cohort
580 study, J. Neurol. Neurosurg. Psychiatry. 74 (2003) 897–900.
581 doi:10.1136/jnnp.74.7.897.
582 [16] C. Daum, M. Hubschmid, S. Aybek, The value of “positive” clinical signs for weakness,
583 sensory and gait disorders in conversion disorder: A systematic and narrative review,
584 J. Neurol. Neurosurg. Psychiatry. 85 (2014) 180–190. doi:10.1136/jnnp-2012-304607.
585 [17] P. Schwingenschuh, T.A. Saifee, P. Katschnig-Winter, A. Macerollo, M. Koegl-Wallner,
586 V. Culea, C. Ghadery, E. Hofer, T. Pendl, S. Seiler, U. Werner, S. Franthal, N.M.
587 Maurits, M.A. Tijssen, R. Schmidt, J.C. Rothwell, K.P. Bhatia, M.J. Edwards, Validation
588 of “laboratory-supported” criteria for functional (psychogenic) tremor, Mov. Disord.
589 31 (2016) 555–562. doi:10.1002/mds.26525.
590 [18] A. Avbersek, S. Sisodiya, Does the primary literature provide support for clinical signs
591 used to distinguish psychogenic nonepileptic seizures from epileptic seizures?, J.
592 Neurol. Neurosurg. Psychiatry. 81 (2010) 719–725. doi:10.1136/jnnp.2009.197996.
593 [19] C. Daum, F. Gheorghita, M. Spatola, V. Stojanova, F. Medlin, F. Vingerhoets, A.
594 Berney, M. Gholam-Rezaee, G.E. Maccaferri, M. Hubschmid, S. Aybek, Interobserver
595 agreement and validity of bedside “positive signs” for functional weakness, sensory
596 and gait disorders in conversion disorder: A pilot study, J. Neurol. Neurosurg.
597 Psychiatry. 86 (2015) 425–430. doi:10.1136/jnnp-2013-307381.
598 [20] A.J. Carson, B. Ringbauer, J. Stone, L. McKenzie, C. Warlow, M. Sharpe, Do medically
599 unexplained symptoms matter? A prospective cohort study of 300 new referrals to
600 neurology outpatient clinics., J. Neurol. Neurosurg. Psychiatry. 68 (2000) 207–210.
601 doi:10.1136/jnnp.68.2.207.
602 [21] W.L. Zijlema, R.P. Stolk, B. Löwe, W. Rief, P.D. White, J.G.M. Rosmalen, How to assess
603 common somatic symptoms in large-scale studies: A systematic review of
604 questionnaires, J. Psychosom. Res. 74 (2013) 459–468.
605 doi:10.1016/j.jpsychores.2013.03.093.
606 [22] K. Kroenke, R.L. Spitzer, J.B.W. Williams, The PHQ-15: Validity of a new measure for
607 evaluating the severity of somatic symptoms, Psychosom. Med. 64 (2002) 258–266.
608 doi:10.1097/00006842-200203000-00008.
609 [23] D.A. Regier, W.E. Narrow, D.E. Clarke, H.C. Kraemer, S.J. Kuramoto, E.A. Kuhl, D.J.
610 Kupfer, DSM-5 field trials in the United States and Canada, part II: Test-retest
611 reliability of selected categorical diagnoses, Am. J. Psychiatry. 170 (2013) 59–70.
612 doi:10.1176/appi.ajp.2012.12070999.
613 [24] W.E. Narrow, D.E. Clarke, S.J. Kuramoto, H.C. Kraemer, D.J. Kupfer, L. Greiner, D.A.
614 Regier, DSM-5 Field Trials in the United States and Canada, Part III: Development and
615 Reliability Testing of a Cross-Cutting Symptom Assessment for DSM-5, Am. J.
616 Psychiatry. 170 (2013) 71–82. doi:10.1176/appi.ajp.2012.12071000.
617 [25] H. Van Ravesteijn, K. Wittkampf, P. Lucassen, E. Van De Lisdonk, H. Van Den Hoogen,
618 H. Van Weert, J. Huijser, A. Schene, C. Van Weel, A. Speckens, Detecting somatoform
619 disorders in primary care with the PHQ-15, Ann. Fam. Med. 7 (2009) 232–238.
620 doi:10.1370/afm.985.
621 [26] S. Körber, D. Frieser, N. Steinbrecher, W. Hiller, Classification characteristics of the
622 Patient Health Questionnaire-15 for screening somatoform disorders in a primary
623 care setting, J. Psychosom. Res. 71 (2011) 142–147.
624 doi:10.1016/j.jpsychores.2011.01.006.
625 [27] A.J. Carson, J. Stone, C.H. Hansen, R. Duncan, J. Cavanagh, K. Matthews, G. Murray,
626 M. Sharpe, Somatic symptom count scores do not identify patients with symptoms
627 unexplained by disease: A prospective cohort study of neurology outpatients, J.
628 Neurol. Neurosurg. Psychiatry. 86 (2014) 295–301. doi:10.1136/jnnp-2014-308234.
629 [28] M. Reuber, M. Chen, J. Jamnadas-Khoda, M. Broadhurst, M. Wall, R.A. Grünewald, S.J.
630 Howell, M. Koepp, S. Parry, S. Sisodiya, M. Walker, D. Hesdorffer, Value of patient-
631 reported symptoms in the diagnosis of transient loss of consciousness., Neurology. 87
632 (2016) 625–33. doi:10.1212/WNL.0000000000002948.
633 [29] G. Erba, E. Bianchi, G. Giussani, J. Langfitt, A. Juersivich, E. Beghi, Patients’ and
634 caregivers’ contributions for differentiating epileptic from psychogenic nonepileptic
635 seizures. Value and limitations of self-reporting questionnaires: A pilot study, Seizure.
636 (2017). doi:10.1016/j.seizure.2017.11.001.
637 [30] M. Hallett, J. Stone, A. Carson, Symptoms (including signs and investigations), in:
638 Funct. Neurol. Disord. 1st Ed., 1st ed., 2016: pp. 169–520.
639 https://www.sciencedirect.com/handbook/handbook-of-clinical-neurology/vol/139
640 (accessed November 29, 2018).
641 [31] A. McGonigal, M. Oto, A.J.C. Russell, J. Greene, R. Duncan, Outpatient video EEG
642 recording in the diagnosis of non-epileptic seizures: A randomised controlled trial of
643 simple suggestion techniques, J. Neurol. Neurosurg. Psychiatry. 72 (2002) 549–551.
644 doi:10.1136/jnnp.72.4.549.
645 [32] R. Alessi, S. Vincentiis, P. Rzezak, K.D. Valente, Semiology of psychogenic nonepileptic
646 seizures: Age-related differences, Epilepsy Behav. 27 (2013) 292–295.
647 doi:10.1016/j.yebeh.2013.02.003.
648 [33] L. Plug, M. Reuber, Making the diagnosis in patients with blackouts: It’s all in the
649 history, Pract. Neurol. 9 (2009) 4–15. doi:10.1136/jnnp.2008.161984.
650 [34] M. Reuber, R. Pukrop, A.J. Mitchell, J. Bauer, C.E. Elger, Clinical significance of
651 recurrent psychogenic nonepileptic seizure status, J. Neurol. 250 (2003) 1355–1362.
652 doi:10.1007/s00415-003-0224-z.
653 [35] J. Stone, Dissociation: What is it and why is it important?, Pract. Neurol. 6 (2006) 308–
654 313. doi:10.1136/jnnp.2006.101287.
655 [36] I. Pareés, P. Kassavetis, T.A. Saifee, A. Sadnicka, M. Davare, K.P. Bhatia, J.C. Rothwell,
656 S. Bestmann, M.J. Edwards, Failure of explicit movement control in patients with
657 functional motor symptoms, Mov. Disord. 28 (2013) 517–523.
658 doi:10.1002/mds.25287.
659 [37] I. Pareés, M. Kojovic, C. Pires, I. Rubio-Agusti, T.A. Saifee, A. Sadnicka, P. Kassavetis, A.
660 MacErollo, K.P. Bhatia, A. Carson, J. Stone, M.J. Edwards, Physical precipitating factors
661 in functional movement disorders, J. Neurol. Sci. 338 (2014) 174–177.
662 doi:10.1016/j.jns.2013.12.046.
663 [38] J. Stone, A. Carson, H. Aditya, R. Prescott, M. Zaubi, C. Warlow, M. Sharpe, The role of
664 physical injury in motor and sensory conversion symptoms: A systematic and
665 narrative review, J. Psychosom. Res. 66 (2009) 383–390.
666 doi:10.1016/j.jpsychores.2008.07.010.
667 [39] J. Stone, C. Warlow, M. Sharpe, Functional weakness: Clues to mechanism from the
668 nature of onset, J. Neurol. Neurosurg. Psychiatry. 83 (2012) 67–69. doi:10.1136/jnnp-
669 2011-300125.
670 [40] K. Schmidtke, B. Metternich, Validation of two inventories for the diagnosis and
671 monitoring of functional memory disorder, J. Psychosom. Res. 67 (2009) 245–251.
672 doi:10.1016/j.jpsychores.2009.04.005.
673 [41] C. Kenney, A. Diamond, N. Mejia, A. Davidson, C. Hunter, J. Jankovic, Distinguishing
674 psychogenic and essential tremor, J. Neurol. Sci. 263 (2007) 94–99.
675 doi:10.1016/j.jns.2007.06.008.
676 [42] L.S. Roper, T.A. Saifee, I. Parees, H. Rickards, M.J. Edwards, How to use the
677 entrainment test in the diagnosis of functional tremor, Pract. Neurol. 13 (2013) 396–
678 398. doi:10.1136/practneurol-2013-000549.
679 [43] R. Baker, E.J. Shaw, Diagnosis and management of chronic fatigue syndrome or
680 myalgic encephalomyelitis (or encephalopathy): summary of NICE guidance, BMJ. 335
681 (2007) 446–448. doi:10.1136/bmj.39302.509005.AE.
682 [44] C. Toth, Hemisensory syndrome is associated with a low diagnostic yield and a nearly
683 uniform benign prognosis, J. Neurol. Neurosurg. Psychiatry. 74 (2003) 1113–1116.
684 doi:10.1136/jnnp.74.8.1113.
685 [45] R. Finkenbine, V.J. Miele, Globus hystericus: A brief review, Gen. Hosp. Psychiatry. 26
686 (2004) 78–82. doi:10.1016/S0163-8343(03)00089-6.
687 [46] J. Baumgartner, J.R. Duffy, Psychogenic stuttering in adults with and without
688 neurologic disease., J. Med. Speech. Lang. Pathol. 5 (1997) 75–95.
689 https://www.researchgate.net/profile/Joseph_Duffy/publication/279897923_Psycho
690 genic_stuttering_in_adults_with_and_without_neurologic_disease/links/565493ca08
691 aeafc2aabbe37d.pdf%0Ahttp://myaccess.library.utoronto.ca/login?url=http://search.
692 ebscohost.com/log.
693 [47] J.R. Duffy, Functional speech disorders, Handb. Clin. Neurol. 139 (2016) 379–388.
694 doi:10.1016/B978-0-12-801772-2.00033-3.
695 [48] K. McGorm, C. Burton, D. Weller, G. Murray, M. Sharpe, Patients repeatedly referred
696 to secondary care with symptoms unexplained by organic disease: Prevalence,
697 characteristics and referral pattern, Fam. Pract. 27 (2010) 479–486.
698 doi:10.1093/fampra/cmq053.
699 [49] P. Fink, Surgery and medical treatment in persistent somatizing patients, J.
700 Psychosom. Res. 36 (1992) 439–447. doi:10.1016/0022-3999(92)90004-L.
701 [50] G.F. Longstreth, J.F. Yao, Irritable bowel syndrome and surgery: A multivariable
702 analysis, Gastroenterology. 126 (2004) 1665–1673. doi:10.1053/j.gastro.2004.02.020.
703 [51] A. Carson, A. Lehn, Epidemiology, Handb. Clin. Neurol. 139 (2016) 47–60.
704 doi:10.1016/B978-0-12-801772-2.00005-9.
705 [52] E. Bianchi, G. Erba, E. Beghi, G. Giussani, Self-reporting versus clinical scrutiny: the
706 value of adding questionnaires to the routine evaluation of seizure disorders. An
707 exploratory study on the differential diagnosis between epilepsy and psychogenic
708 nonepileptic seizures, Epilepsy Behav. (2018). doi:10.1016/j.yebeh.2018.11.040.
709 [53] M. Schwabe, M. Reuber, M. Schöndienst, E. Gülich, Listening to people with seizures:
710 How can linguistic analysis help in the differential diagnosis of seizure disorders?,
711 Commun. Med. 5 (2008) 59–72. doi:10.1558/cam.v5i1.59.
712 [54] L. Plug, B. Sharrack, M. Reuber, Seizure metaphors differ in patients’ accounts of
713 epileptic and psychogenic nonepileptic seizures, Epilepsia. 50 (2009) 994–1000.
714 doi:10.1111/j.1528-1167.2008.01798.x.
715 [55] D. Jones, P. Drew, C. Elsey, D. Blackburn, S. Wakefield, K. Harkness, M. Reuber,
716 Conversational assessment in memory clinic encounters: Interactional profiling for
717 differentiating dementia from functional memory disorders, Aging Ment. Heal. 20
718 (2016) 500–509. doi:10.1080/13607863.2015.1021753.
719 [56] A. Keech, C.X. Sandler, U. Vollmer-Conna, E. Cvejic, A.R. Lloyd, B.K. Barry, Capturing
720 the post-exertional exacerbation of fatigue following physical and cognitive challenge
721 in patients with chronic fatigue syndrome, J. Psychosom. Res. 79 (2015) 537–549.
722 doi:10.1016/j.jpsychores.2015.08.008.
723 [57] A. Baumann, P.O. Katz, Functional disorders of swallowing, in: 2016: pp. 483–488.
724 doi:10.1016/B978-0-12-801772-2.00039-4.
725 [58] L.L. Kupper, K.B. Hafner, How appropriate are popular sample size formulas?, Am.
726 Stat. 43 (1989) 101–105. doi:10.1080/00031305.1989.10475628.
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745 Appendix A: Edinburgh Neurosymptoms Questionnaire

Neurological Symptom Questions

Many people experience neurological symptoms. These can be disabling and distressing. This
survey asks about common neurological symptoms you may be experiencing. 5. During the last six months have you been bothered by tremor or an
Yes No
abnormal movement in one or more limb e.g. arm (s) or leg(s)?
1. During the last six months have you been bothered by blackouts? Yes No ¨1 ¨0
If NO go to question 2. If YES please answer questions below. ¨1 ¨0 If NO go to question 6. If YES please answer questions below.
a. During your blackout have you been told that you lie still or shake? Lie Still Shake ¨1 ¨0
a. Did your tremor or abnormal movement come on suddenly?
¨1 ¨2
b. Have you ever had a blackout in a medical setting e.g. visiting the
¨1 ¨0
b. Did your tremor or abnormal movement come on after an injury or ¨1 ¨0
hospital/GP/another doctor? accident?
c. Have you had more than two seizures during which you shook
without stopping for more than 10minutes? (This does not include ¨1 ¨0
c. Can your tremor or abnormal movement go away completely for ¨1 ¨0
the time taken for you to come round after the seizure had finished) hours to days only to return again?
d. Have you ever been able to hear people but could not respond to ¨1 ¨0 d. Does your tremor or abnormal movement ever stop when you are
them during your blackout? ¨1 ¨0
distracted or concentrating on something else?
e. Do you ever have moments before your blackouts of losing track of
what is going on, of “blanking out” or “spacing out” or in some way ¨1 ¨0 During the last three months have you had pain almost every day in more
6. Yes No
feeling that you are not part of what is going on? than one part of your body?
f. Are you told that after an attack you cry or are upset? ¨1 ¨0 ¨1 ¨0
If NO go to question 7. If YES please answer questions below.
2. During the last six months have you been bothered by weakness in one or
Yes No a. Is your pain worst in different parts of your body on different days? ¨1 ¨0
more limb e.g. arm(s) or leg(s)?
¨1 ¨0
If NO go to question 3. If YES please answer questions below.
7. Have you been lacking energy every day or almost every day for the last
a. Do you drop things frequently? ¨1 ¨0 Yes No
six months?
¨1 ¨0
b. Does your limb weakness get worse or better at different times of ¨1 ¨0
If NO go to question 8. If YES please answer questions below.
the day?
a. Does activity make your fatigue worse? ¨1 ¨0
c. Does concentrating on trying to move make the limb weakness ¨1 ¨0
worse? 8. In the last five years have you had to see doctors in the hospital for
Yes No
d. At the start of your limb weakness did you feel your heart pounding different problems more than four times?
¨1 ¨0 ¨1 ¨0
or did you feel frightened, anxious or very uneasy? (E.g. problems with your heart, your joints, your brain and gut)
e. Does your weak limb feel like it does not fully belong to you? ¨1 ¨0
9. Do you get a feeling that there is a lump in your throat or something stuck Yes No
3. Do you have numbness or altered sensation that makes you feel like your Yes No when you are trying to eat or drink? ¨1 ¨0
body is cut in half? ¨1 ¨0
10. Do you have a stutter which started after you were more than 16years Yes No
4. During the last six months have you been bothered by memory problems? Yes No old? ¨1 ¨0
If NO go to question 5. If YES please answer questions below. ¨1 ¨0
Yes No
Me Partner/ 11. Have you needed any operations? If Yes, please circle all that apply:
a. Who is most bothered by your memory problems, you or your ¨1 ¨0
partner/family/friends? Family
¨2 ¨1
Appendix Tonsils Laparoscopy to investigate pain Hysterectomy
b. Are you bothered by forgetting important details such as the name ¨1 ¨0
of a family member or your PIN number? Operation for adhesions
c. Do you have blank spells which occur during the day? ¨1 ¨0
Other Operations:
19/10/2017 v2.2 5
1. 2. 3. _______________________

746 Thank you for filling out the questionnaire

747 19/10/2017 v2.2 6

748
749
750
751
752
753
754
755
756
757
758
759
760 Appendix B: Consultant diagnostic guidance
761 “What we mean by a functional disorder

762 The following is meant as a guide for this study and we are aware that any divisions like

763 this are imperfect. Many patients have a mixture of symptoms, syndromes or disease and

764 the final coding is your decision based on these guidelines

765

766 ‘Functional disorder’ for the purpose of this study: Tension Headache; Aetiologically

767 controversial symptom ‘syndromes’ (e.g. Chronic fatigue syndrome, Fibromyalgia, Irritable

768 Bowel Syndrome); Physiologically explained processes which are thought to be linked to

769 emotional symptoms (e.g. Hyperventilation); Chronic pain or dizziness which is

770 unexplained by a clear structural cause.

771

772 ‘Organic disease’ for the purpose of this study: Migraine; Any neurological disorder with

773 a known pathological basis; Neurological disorders with defined and characteristic

774 features but without a clear pathological basis (e.g. Gilles de la Tourette syndrome,

775 Idiopathic focal dystonia); Physiological explained processes NOT linked to emotional

776 symptoms (e.g. micturition syncope); Psychotic disorder.”

777
778
779