Factors associated with delayed diagnosis of migraine: A hospital-based cross-sectional study.

CONTEXT: Migraine is common debilitating disorders, affecting 10 to 20% of the world's population. However, proper diagnosis is delayed due to many factors. AIMS: To determine various factors associated with delayed diagnosis of migraine. SETTINGS AND
DESIGN: Hospital-based cross-sectional study.
MATERIALS AND METHODS: Patients attending Neurology OPD of AIIMS Bhopal and satisfying diagnostic criteria of the International Headache Society (ICHD-3β) for migraine were selected for study. MIDAS, MINI, and ROME-III were used. First diagnosis was considered as "Appropriate" if patients were previously diagnosed as "migraine" or getting treatment for it; otherwise labeled as "Inappropriate." STATISTICAL ANALYSIS: Associations were tested by Chi-square, t-test, or Mann-Whitney test. Logistic regression analysis was used for identifying independent factors associated with inappropriate diagnosis.
RESULTS: Hundred patients (female = 77) of migraine were included. Mean age (SD) was 32.42 (10.74). Diagnosis was "inappropriate" in 65 patients. Number of inappropriate diagnosis/appropriate diagnosis was 10/40 (25%) by neurologists; 35/39 (89.3%) by physicians; 18/18 (100%) by ophthalmologists. Factors associated with "Inappropriate Diagnosis" were "Neurologist vs Other Doctors" {10 (25%) vs 55 (91.7%), P < 0.001}; throbbing vs other types of headache {51 (60.7%) vs 14 (87%), P = 0.047}; and temporal vs other sites {9 (42.9%) vs 56 (70.9%), P = 0.017}. Patients with "Inappropriate Diagnosis" had to expend more money {7000 (4,500; 12,500) vs 4000 (1000, 6000), P < 0.01; median (interquartile range) all in INR}. Other clinical parameters including vertigo, cervical pain, anxiety, depression, and functional gastrointestinal symptoms were not associated with delayed diagnosis.
CONCLUSION: Delayed diagnosis and misdiagnosis is very frequent in migraine, leading to financial burden to patients. Management of common disorders like migraine should be addressed in undergraduate medical teaching curriculum.

Introduction

Migraine is one of the most common chronic debilitating disorders, affecting approximately 10 to 20% of the world's population.[12345] There is wide range of presentations of migraine due to different characters of headache, various types of auras, associated symptoms such as nausea, vomiting, loss of appetite, and other gastrointestinal symptoms; irritability due to photophobia and phonophobia; dizziness and vertigo, as well as associated neuropsychiatric comorbidities.[67891011] Due to wide range of presentation and lack of awareness to these associated symptoms, number of patients with migraine are often subjected to unnecessary investigations, frequent reference to different specialists, and sometimes even receive nonspecific and only symptomatic treatment thus facing delayed or inappropriate diagnosis.[12131415]

There are limited Indian studies evaluating functional gastrointestinal (GI) and psychiatric comorbidities among migraineurs.[16] Thus present study was planned to assess the association of functional GI symptoms and neuropsychiatric comorbidities with migraine as well as to determine the factors associated with delayed diagnosis or misdiagnosis of migraine.

Materials and Methods

It was a cross-sectional hospital-based observational study conducted at tertiary care center. Adult patients (age >18 years) attending Neurology OPD from April 2016 to July 2016, satisfying operational diagnostic criteria of the International Headache Society (International Classification of Headache Disorders- 3 beta, ICHD-3β) were enrolled in study after a written informed consent.[17] The study was approved by institutional human ethics committee. Patients with (i) serious systemic disease; (ii) history of abdominal surgery (except appendectomy); (iii) active biliary, pancreatic, gastrointestinal disorders, and (iv) significant head and neck trauma in previous 3 months were excluded from the study.

Two investigators (one trained neurologist NKR, with more than 10 years of experience in neurology; and other was senior resident doctor (DSP), trained to evaluate headache disorder as per ICHD-3β independently evaluated patients with headache attending neurology OPD. Patients with migraine as per our diagnostic criteria were included in study only if both investigators independently agreed for the diagnosis. In case of disagreement between the two investigators, those patients were excluded. As per our inclusion and exclusion criteria, consecutive 100 patients were included in the study. Further, we categorized these patients into “Appropriate Diagnosis” if they were either diagnosed as “migraine” and/or they received common prophylactic treatment for migraine (namely propranolol, valproate/divalproex, tricyclic antidepressant, or topiramate) by previous treating doctors; otherwise they were considered as patients with “Inappropriate Diagnosis.”

Detailed clinical data of every patient were collected. Total expenditure for the treatment of the migraine was recorded as per their memory and tried to verify with available records. However, patient's word regarding total expenditure was accepted as the final word. Following questionnaires were used to assess severity, disability, and associated comorbidities in migraineurs:

For headache-related disability: MIDAS (Migraine Disability Assessment)[18]

For functional GI symptoms: Rome III questionnaire[19]

For Nueropsychiatric Comorbidities: The Mini-International Neuropsychiatric Interview (MINI) questionnaire.[20]

The Migraine Disability Assessment (MIDAS) Questionnaire is validated and having good internal consistency, test–retest reliability, developed to assess headache-related disability in a patient with migraine. It contains five questions, calculates number of working days missed, or activity was reduced in the past 3 months due to migraine. These features make it very suitable for use in clinical practice.[18]

Rome III is very comprehensive and authoritative criteria for the subject with functional gastrointestinal disorders (FGID). It is easily comprehensible, well organized, and easy to administer. It is developed in such a manner that most of the clinicians, researchers, basic scientists, and patients with FGIDs can be benefited.[19]

The Mini-International Neuropsychiatric Interview (M.I.N.I.) is a validated, short structured diagnostic interview, which was developed for DSM-IV and ICD-10 psychiatric disorders. It is very good, and well-structured tool for clinical trials and epidemiological studies as well as in various clinical settings with an administration time of approximately 15 minutes only.[20]

Statistical analysis

All statistical analysis was done by IBM SPSS 21. Data were summarized as nominal variables by expressing their frequency and percentage and numerical variables as mean and standard deviation or median and range. Association of “Inappropriate Diagnosis” with sociodemographic and clinical factors was tested by Chi-square test for nominal variables and t-test or Mann-Whitney test for numerical variables. Then we have used logistic regression analysis for identifying independent factors associated with “Inappropriate Diagnosis.” Inappropriate diagnosis was used as dependent variable and independent variables were, duration of illness, localization of pain, type of first consultation physician, and presence of any neuropsychiatric condition, character of pain, autonomic dysfunction, and MIDAS score. Selection of independent variables was based on mixed approach, i.e., all variables where univariate P value is less than 0.25 and clinically important variables were selected. P Value less than 0.05 was considered statistically significant.

Results

The clinical characteristic of patients with migraine is shown in Tables 1 and 2. Migraine was about three and half time more frequent in women than men. Only one third of patients with migraine came to us with “Appropriate Diagnosis.” Forty patients were earlier evaluated by neurologist, out of which “Appropriate Diagnosis,” was made in 30 patients (75%). Nonneurologists treated rest 60 patients and could make “Appropriate Diagnosis” in only five patients (8.3%). Patients with “Inappropriate Diagnosis” had longer duration of illness and expended more money for treatment [Table 3]. On univariate analysis, duration of illness, character of headache (throbbing vs not-throbbing), localization of headache (temporal vs others), and speciality of first consulted doctor (neurologist vs others) were significantly associated with appropriate diagnosis. Further, Hosmer–Lemeshow Goodness-of-Fit test indicated fitness of data (P = 0.763). Omnibus test of model coefficient was significant (P < 0.001) and Nagelkerke R-value was 0.73. This indicates logistic regression model could be used and can explain 73% of variation in dependent variable. After logistic regression, statistically significant independent predictors evident in this study were, if first treating doctor was a neurologist (P < 0.001), and if the localization of headache was temporal (P = 0.01).

Table 1

Characteristics of all patients (n=100) included in the study

S. No.	Characteristics
1	Age; years mean (SD)	32.4 (10.7)
2	Gender Female: Male	77:23
3	Education
3.1	Illiterate	09
3.2	Literate	91
4	Duration of illness, median (interquartile range); (in months)	48 (12, 93)
5	Appropriate diagnosis of migraine	35
6	Inappropriate diagnosis	65
6.1	No diagnosis	27
6.2	Nonspecific headache	10
6.3	Tension-type headache (TTH)	07
6.4	Refractory error	19
6.5	Sinus-related headache	02
7	First attending doctor
7.1	Neurologist	40
7.2	Primary physician	39
7.3	Ophthalmologist	18
7.4	AYUSH	03
9	Expenses before appropriate diagnosis (n=77); INR; median (IQR)	1000.00 (0.00; 55,000.00)
10	Expenses after appropriate diagnosis (n=98); INR; median (IQR)	1000.00 (500.00, 4,000.00)
11	Total expenses; INR; median (IQR)	5750.00 (3000.00, 11,000.00)

INR: Indian rupees; AYUSH: Ayurveda-Yoga-Unani-Siddha-Homeopathy; IQR: interquartile range

Table 2

Clinical features and comorbidities in migraine (n=100)

S. No.	Clinical features	Number
1	Localization of headache
1.1	Frontal	66
1.2	Temporal	21
1.2	Occipital	09
1.4	Orbital	02
2	Lateralization
2.1	Unilateral	55
2.2	Bilateral	45
3	Character
3.1	Throbbing	84
3.2	Heaviness	07
3.3	Stabbing	09
4	Cervical pain	27
5	Aura (Any kind)	20
6	Nausea	83
7	Vomiting	58
8	Photophobia	87
9	Phonophobia	88
10	Worsening with outdoor activities	99
11	Vertigo/dizziness	26
12	Autonomic feature	32
13	Medication overuse headache (MOH)	07
13.1	Duration of MOH months; median (interquartile range)	24 (6,60)
14	Family H/o migraine	32
15	Headache Impact Test (HIT); median (interquartile range)	62.5 (50, 65)
16	MIDAS score; median (interquartile range)	15 (8;21)
17	Hypertension	09
18	Diabetes mellitus	07
19	Any psychiatric	41
19.1	Anxiety	35
19.2	Depression	32
20	Functional gastrointestinal	25
20.1	Dyspepsia	12
20.2	Esophageal	16
20.3	Stomach	10
20.4	Rectum & anal canal	02

Table 3

Comparison of various parameters between “Appropriate” and “Inappropriate” diagnosis

Variables		Appropriate Dx	Inappropriate Dx	Univariate P	Multivariate P
Age; mean (SD)		32.4 (10.94)	32.43 (10.72)	0.99
Gender	Female (77)	26 (33.8%)	51 (66.2%)	0.64
	Male (23)	9 (39.1%)	14 (60.9%)
Residence	Urban (80)	29 (36.3%)	51 (63.9%)	0.60
	Rural (20)	6 (30%)	14 (70%)
Education	Illiterate (9)	2 (22.2%)	7 (77.8%)	0.40
	Literate (91)	33 (36.3)	58 (63.7%)
Duration of illness (months); median, (IQR)		24 (12,72)	60 (24,96)	0.02
Time to appropriate Dx; (months); median (IQR)		00 (00, 12)	36 (12, 72)	<0.01
Expenses before appropriate Dx; (INR) median (IQR)		00 (00, 00)	5,000 (3,000, 10,000)	<0.01
Expenses after appropriate Dx; (INR); median (IQR)		3,000 (1,000, 5000)	1,000 (1,000, 2,000)	<0.01
Total expenses; (INR); median (IQR)		4,000 (1000;6000)	7,000 (4,500; 12,500)	<0.01
Freq of headache per month; median (IQR)		5 (4,8)	6 (4,10)	0.38
First doctor	Neurologist (n=40)	30 (75%)	10 (25%)	<0.01	<0.001
	Others (n=60)	5 (8.3)	55 (91.7)
Character	Throbbing (84)	33 (39.3%)	51 (60.7%)	0.047	0.051
	Others (16)	2 (12.5%)	14 (87.5%)
Localization	Temporal (21)	12 (57.1%)	9 (42.9%)	0.017	0.010
	Others (79)	23 (29.1%)	56 (70.9%)
Aura	Present (20)	8 (40%)	12 (60%)	0.60
	Absent (80)	27 (33.7%)	53 (66.3%)
Lateralization	Unilateral (55)	17 (30.9%)	38 (69.1%)	0.34
	Bilateral (45)	18 (40%)	27 (60%)
Cervical pain	Present (27)	9 (33.3%)	18 (66.7%)	0.83
	Absent (73)	26 (35.6%)	47 (64.4%)
Vertigo	Present (26)	8 (30.8%)	18 (69.2%)	0.6
	Absent (74)	27 (36.5)	47 (43.5)
Autonomic features	Present (32)	8 (25%)	24 (75%)	0.15	0.684
	Absent (68)	27 (39.7%)	41 (60.3%)
Family h/o migraine	Present (32)	12 (37.5%)	20 (62.5%)	0.72
	Absent (68)	23 (33.8%)	45 (66.2%)
Any neuropsychiatric comorbidity	Present (41)	11 (26.8%)	30 (73.2%)	0.15	0.450
	Absent (59)	24 (40.7%)	35 (59.3%)
Depression	Present (32)	9 (28.1%)	23 (71.9%)	0.32
	Absent (68)	26 (38.2%)	42 (61.8%)
Anxiety	Present (35)	10 (28.6%)	25 (71.4%)	0.32
	Absent (65)	25 (38.5%)	40 (61.5%)
Any function GI disorder	Present (25)	8 (32%)	17 (38%)	0.72
	Absent (75)	27 (36%)	48 (64%)
MIDAS grade	No or mild Disabilities (38)	17 (44.7%)	21 (55.3%)	0.11	0.840
	Moderate to Severe Disabilities (62)	18 (29%)	44 (71%)
MIDAS score; median (IQR)		12 (6;20)	15 (10,24)	0.11

GI: gastrointestinal; INR: Indian rupee; SD: standard deviation; IQR: interquartile range

Discussion

In present hospital-based study, majority of patients with migraine (60%) first consulted nonneurologists for their headache and appropriate diagnosis could be made only in few patients (8.3%) by them. Interestingly, neurologist also could not make appropriate diagnosis in 25% of patients with migraine, when consulted as a first doctor. Patients with “inappropriate diagnosis” had prolonged duration of illness and had more expenditure for their disease.

Migraine, a common disorder is reported to be associated with various co-morbidities including functional GI and psychiatric comorbidities such as depression and anxiety.[678910112122] It has been suggested that along with clinical skills, associated co-morbidities may be responsible for diagnostic delay in migraine.[1223] Neuropsychiatric comorbidities, mainly depression and anxiety are very common in migraineurs and are reported to have bidirectional relationship.[2425] It is also important to highlight that these mood disorders are not only restricted to migraine or other headache disorders but also in many other chronic conditions.[2627] Hence, it needs to be explored that whether the association of migraine and mood disorder has a pathological basis or it is secondary to migraine like other chronic illnesses. Anxiety and depression were very frequent in our study and about one third of our patient had anxiety and depression. However, the mood disorder was not associated with delayed diagnosis of migraine. One recent Indian study reported relatively less frequent mood disorders in migraine.[28] The different frequency of these co-morbidities in migraineurs in various studies may be due to different study design including diagnostic criteria; selection of patients, study setup, and scales used to assess these comorbidities. It needs to be highlighted again that due to significant association between migraine and mood disorders, these comorbidities should be explored during management of migraine.

GI symptoms, an another common comorbidity with migraine, were evaluated in a recent study suggested a complex relationship between them.[29] Both mood disorder and migraine were associated with IBS but mood disorder coexisting with migraine did not increased risk of IBS. In a recent review of migraine and GI disorders, authors concluded that data were too heterogeneous to get a definite meaningful relationship.[10] In our study, about one-fourth of our patient had functional GI symptoms. Upper GI symptoms related to esophagus and stomach was much more common than lower GI symptoms. Whether these GI symptoms were associated “functional GI disorder” or “nausea and/or vomiting” of migraine needs to be explored.

Despite a common disorder, migraine is often not diagnosed properly across the globe. In a population-based survey for migraine diagnosis, about 50% of the migraineurs were not aware about their diagnosis. Interestingly, primary physicians also reported incorrect diagnosis of migraine in a large proportion of migraineurs.[233031] Along with physician's skill, other possible causes for incorrect diagnosis were attributed to multiple complaints, less time during consultation, evaluation of patients during asymptomatic period, and poor health-care seeking behavior. Despite the fact that unilateral location is one of the four characteristics of headache as per ICHD-3β, it was present only in 55% of our patients. So other characteristics of headache need to be explored during clinical evaluation. One-fourth of patients had brief vertigo/dizziness, cervical pain, and autonomic features, suggesting that these features may be seen in a patient with migraine.

Various studies tried to simplify the diagnosis criteria of migraine to save time without compromising sensitivity and specificity.[32333435] Most of these studies tried to select few sensitive and specific clinical features for diagnosis of migraine. Prolonged duration of severe headache (4-72 hours), which limits physical or intellectual activity, photophobia and phonophobia, and nausea and/or vomiting were reported by most of these studies as useful questions to diagnose migraine. In our study, 99% of our patients had worsening of headache during outdoor activities and most of them had nausea and/or vomiting and photophobia or phonophobia. This observation re enforces the statement that diagnosis of migraine should be considered in a person with recurrent headache for at least more than 3 months; of moderate-to-severe intensity; duration (4-72 hours); restricting physical/intellectual activity associated with nausea/vomiting, and photophobia/phonophobia.[32333435] Inappropriate diagnosis has its financial burden on patients. In our study, patients with delayed diagnosis paid more money from their pocket than those with early diagnosis. Of note, that only 77% of our patient could recollect money spent for the treatment. Considering limitations of memory-based data, we tried to objectivize the total expenditure by verifying records, receipt, and other information available with the patients.

Presence of the comorbidities is reported as a possible cause of delayed diagnosis.[12] However, in present study, these comorbidities were not associated with delayed diagnosis or inappropriate diagnosis of migraine. It is worth mentioning that correct diagnosis was not mentioned in about half of patients by first treating physicians and probably these associated comorbidities were also not explored by them. Another important reason for this observation may be due to our study design, in which we included patients on the basis of strict screening as per ICHD-3β by two experts rather than including all suspected migraineurs. So-called Sinus Headache, a commonly reported misdiagnosis of migraine, was reported in only two patients, while diagnosis of “refractory error” was relatively high in our study, probably due to not committing any specific diagnosis by primary physicians or effect of locally prevalent belief/understanding regarding cause of headache.[36] Tension-type headache is another common causes of misdiagnosis of migraine and was relatively less common in this study.[13]

Despite the fact that only few patients in this study had atypical features of migraine-like medication overuse headache or complicated auras; the “Appropriate Diagnosis” was missed by more than 90% of the nonneurologist. Although, correct diagnosis of migraine is often missed in about half of patients even in developed countries,[2331] but scenario is alarming in our study, specially due to limited number of neurologists and large number of patients with migraine in our country. It needs to be highlighted that in present study, trained neurologists also could not make appropriate diagnosis in one-fourth patients. Common factors responsible for misdiagnosis are complex presentation of migraine, strict diagnostic criteria, or less time given to patient during clinical evaluation. However, in present study, all patients were included as per strict ICHD-3β criteria, and only few patients had atypical features, hence less time given to patients by a neurologist appears to be an important factor for misdiagnosis. Due to larger patient load, it would be impossible for a neurologist/headache expert to evaluate in detail every patient with migraine or similar common disorders in our country. Thus, primary physicians will have to be trained to diagnose migraine for managing simple migraine within community. Uncomplicated migraine may be managed in community without brain imaging as it is usually not recommended for all patients with chronic headache. However, imaging may be required in selected patients with focal neurological deficit, papilledema, or atypical headache or in patient with some additional risk factors.[3738]

Our study was done in single center with limited number of patients, it paves a way to conduct larger multicentric study, evaluating knowledge and efficiency of primary physicians for managing migraine in community, and for needful modification of undergraduate teaching and training curriculum.

Financial support and sponsorship

ICMR STS fellowship to Medical Student

Conflicts of interest

There are no conflicts of interest.