Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx

Contents lists available at ScienceDirect

Best Practice & Research Clinical

Rheumatology
journal homepage: www.elsevierhealth.com/berh

Time to diagnosis of fibromyalgia and factors

associated with delayed diagnosis in primary
care
Omer Gendelman a, 1, Howard Amital a, b, *, 1, Yael Bar-On c,
Dana Ben-Ami Shor d, Daniela Amital e, Shmuel Tiosano a, b,
Varda Shalev b, f, Gabriel Chodick b, f, Dahlia Weitzman b, f
a
Department of Medicine ‘B’, Sheba Medical Center, Tel-Hashomer, Israel
b
Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
c
Department of Medicine E, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel
d
Department of Gastroenterology, Sheba Medical Center, Tel-Hahsomer, Israel
e
Beer-Yaakov/Ness-Ziona Mental Health Center, Beer-Yaacov, Israel
f
MK and Maccabi Research & Innovation Institute, Maccabi Healthcare Services, Tel Aviv, Israel

a b s t r a c t
Keywords:
Diagnosis Fibromyalgia is a complex disorder, and its diagnosis often com-
Fibromyalgia prises a challenge for physicians.
Anxiety We aimed to assess the time from the appearance of early symp-
Widespread pain toms to the definite diagnosis of fibromyalgia and to delineate
Rheumatic patient and physician factors associated with this delay.
The database of a large healthcare services provider was searched
for all fibromyalgia patients (age 21 years or more) diagnosed
during 2008e2011. Patients with a diagnosis confirmed by a
physician with a relevant specialization were compared with age-
and sex-matched nonfibromyalgia controls to retrospectively
identify an initial complaint pattern characteristic of fibromyalgia.
The time from initial complaints to fibromyalgia diagnosis, as well
as the fraction of time while the patient was treated continuously
by the same primary physician as at time of diagnosis, was
assessed among all eligible fibromyalgia patients.
An initial complaint pattern was identified in 67.9% of 2,055 pa-
tients with confirmed fibromyalgia vs. 27.6% of 9,172 controls.
Among 2,369 (69.0%) of all 3,434 eligible patients with

* Corresponding author. Department of Medicine B, Sheba Medical Center, Tel Hashomer, 52621, Israel.
E-mail address: [email protected] (H. Amital).
1
The first two authors share equal contribution.

https://doi.org/10.1016/j.berh.2019.01.019
1521-6942/© 2019 Published by Elsevier Ltd.

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
2 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx

fibromyalgia, mean (SD) total time to diagnosis was 6.42 (3.57)

years and mean (SD) time while treated by the same primary
physician as at diagnosis was 3.75 (3.26) years. Comorbidity,
younger patient, and older physician age were associated with
longer time to diagnosis.
By analysis of “real-world” data, fibromyalgia remains a complex
and elusive diagnosis, taking years to be properly diagnosed. There
is a need to increase the awareness of this syndrome among
physicians, especially of older age.
© 2019 Published by Elsevier Ltd.

Introduction

Fibromyalgia (FM) is an idiopathic chronic pain disorder characterized by widespread nonarticular

musculoskeletal pain, allodynia, and hyperesthesia [1]. It is a well-known medical condition among the
central sensitivity syndromes, with overlapping aspects with many other medical conditions [2]. Fe-
male preponderance in FM has been well established, with female prevalence being up to 8 times
higher than that in males [1,3,4].
Alongside the chronic pain sensation, patients tend to complain of various sleeping disorders, fa-
tigue, memory problems, and concentration difficulties [5,6]. Depression and other psychiatric
comorbidities are often reported in patients with FM; the constellation of numerous somatic, func-
tional, and emotional symptoms often causes a decline in patients’ overall social and vocational
functioning, thereby leading to the aggravation of their depressive symptoms, turning their treatment
more difficult and challenging [7,8]. Consequently, the economic and financial burden of FM is sub-
stantial [9], including more frequent visits to general practitioners and specialists, excessive usage of
laboratory and imaging tests, and medications for pain relief as well as for coexistent illnesses [10,11].
Despite the refining of the diagnostic tools that currently include parameters beyond pain, such as
unrefreshing sleep, fatigue, and cognitive dysfunction, the diagnosis of FM remains challenging and
elusive [12e14]. Often, a long duration precedes proper diagnosis of patients with FM [15]. The absence
of specific clinical characteristics and the impact of patients' and physicians’ attitudes toward the FM
construct all contribute to inadequate and deferred diagnosis [16].
This research aimed to investigate the time interval from the appearance of early symptoms and
complaints to the definite diagnosis of FM by a physician. In addition, we studied the patient and
physician parameters that play a role in this delay.

Patients and methods

Study design and population

This retrospective cohort study utilized the computerized database of Maccabi Healthcare Services
(MHS), the second largest healthcare services provider in Israel. MHS comprehensive database is
derived from a central electronic medical record, containing longitudinal data since 1998 on a stable
population of more than 2 million people.
The study was carried out in two stages (Fig. 1). First, we characterized an initial complaint pattern
among adult patients (age
21 years) eventually diagnosed with FM by a rheumatologist, neurologist,
psychiatrist, or during hospitalization (“confirmed FM”, n ¼ 2055). Second, the initial complaint
pattern was used to assess the time to FM diagnosis among the “confirmed FM” group as well as among
patients diagnosed with FM by a primary care physician (in at least 2 visits on different dates) but not
confirmed elsewhere (“Primary care diagnosed,” n ¼ 1379). International Classification of Diseases code

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx 3

Stage I - DefiniƟon of an iniƟal complaint paƩern

“Confirmed”
FM paƟents
TIME N=2,055

Age & sex

RetrospecƟve search for an iniƟal
1998 Index date matched non-
complaint paƩern frequent among FM controls
(2008-2011)
FM paƟents and infrequent among N=9,172
non-FM controls
Rheumatoid
ArthriƟs
paƟents
N=856

Stage II - Assessment of Ɵme to FM diagnosis

TIME “Confirmed” +
“primary-care
diagnosed” FM
paƟents for
Index date whom the
First 6-month (2008-2011) iniƟal
interval with the complaint
iniƟal complaint paƩern was
paƩern defined in FracƟon of the TOTAL TIME found
B
stage I (A) while the paƟent was N=2,369
associated with the same
primary-care physician to
whom he/she was
associated with at index

A TOTAL TIME from iniƟal complaints to FM

Fig. 1. Illustration of study design.

version 9 (ICD-9) of 729.1 or an equivalent internal MHS code was used to identify FM. For all FM
patients, the first date of diagnosis was defined as index date.
Two comparison groups served as negative controls (Fig. 1): a) Rheumatoid arthritis (RA) patients
(n ¼ 856) diagnosed by a rheumatologist, at hospital, or at 2 or more primary care visits and b)
randomly selected non-FM enrollees (n ¼ 9172). The latter were frequency matched by sex and age to
the “confirmed FM” patient group with a 1:5 ratio. Age matching was done by assigning each control an
average diagnosis date of all FM patients born in the same year. Initially, we included patients with
index dates between 2003 and 2011 and
5 years continuous enrollment with MHS. However, initial
analysis showed longer time to diagnosis than that anticipated; therefore, patients with index dates
<2008 and < 10 years of continuous enrollment were excluded from the study, resulting in a ratio of
~4.4 controls per FM case. Age and sex distribution of FM cases and controls remained similar (Table 1).

Data collection

Data included age, sex, socio-economic status (SES), and diagnoses of FM and of comorbidities. A list
of diagnoses potentially comprising “initial FM-related complaints” was constructed (Supplementary
data, Table 1) based on the American College of Rheumatology criteria for diagnosing FM [13] and
after reviewing the medical files of 50 FM patients. For all diagnoses, date, source (primary or sec-
ondary physicians, or at hospital discharge), and specialties of diagnosing physicians were abstracted.
SES was defined according to the poverty index of the member's residence area as defined during the
2008 national census. The poverty index is based on several parameters including household income,
education, crowding, material conditions, and car ownership and ranges from 1 to 20, based on cluster
analysis, with 1 as the lowest SES and 20 the highest [17].

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
4 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx

Table 1
Characteristics of FM patients, RA patients, and general controls aged
21 years (Index dates 2008e2011,
10 years at MHS)
(N ¼ 13,462)a.

Confirmed FM FM diagnosed only at primary care Rheumatoid arthritis General control

N ¼ 2055 N ¼ 1379 N ¼ 856 N ¼ 9172
% (n) % (n) % (n) % (n)

Age (years), Mean (SD) 50.5 (11.4) 50.3 (11.6) 57.8 (14.2) 51.0 (12.8)
Female gender 91.1% (1873) 87.7% (1209) 71.4% (611) 90.7% (8321)
SES score (1e10)b, Mean (SD) 6.0 (2.1) 6.3 (2.2) 6.2 (2.2) 6.5 (2.2)
Comorbidity
Anxiety 20.8% (428) 20.3% (280) 10.6% (91) 7.6% (696)
Both anxiety and depression 3% (62) 4.1% (57) 1.4% (12) 0.6% (59)
Depression 18.8% (386) 19.9% (275) 8.1% (69) 6.2% (567)
Anxiety and/or Depression 32.4% (666) 33.6% (464) 16.0% (137) 11.8% (1086)
Pain disorder 20.6% (424) 22% (303) 9.1% (78) 6% (550)
Sleep disorder 11.4% (234) 11.6% (160) 9.2% (79) 4.7% (428)
OA 10.1% (208) 8.2% (113) 12.4% (106) 3.1% (280)
CFS 6.5% (133) 8.2% (113) 1.9% (16) 3% (276)
Migraine 5.9% (122) 5% (69) 2.6% (22) 3% (272)
IBS 4.4% (90) 5% (69) 1.6% (14) 1.4% (129)
TMJ disorder 3.3% (67) 3.2% (44) 1.8% (15) 0.7% (66)
PTSD 3.0% (62) 2.8% (39) 1.2% (10) 0.6% (54)
Tinnitus 1.5% (30) 0.9% (12) 0.4% (3) 0.5% (42)
SLE 1.3% (27) 0.3% (4) 0.5% (4) 0.1% (13)
Other neurological dis. 0.8% (16) 0.7% (10) 0.5% (4) 0.2% (18)
Sjogren's syndrome 0.7% (15) 0.4% (6) 1.1% (9) 0.1% (9)
Tendinitis 0.3% (6) 0.1% (1) 0.1% (1) 0.1% (10)
RLS 0.0% (1) 0.1% (2) 0.0% (0) 0.1% (6)

IBS-Irritable Bowel Syndrome; RLS-Restless Leg Syndrome; TMJ- Temporomandibular Joint Disorder; OA-Osteoarthritis; SLE-
Systemic Lupus Erythematosus; PTSD-Post Traumatic Stress Disorder.
a
All p-values<0.001 except for tendinitis (p ¼ 0.197) and RLS (p ¼ 0.585).
b
SES was missing for 1.8% of patients.

Administrative information included period of enrollment with MHS and dates of association with
each primary care physician. In MHS, patients may choose their primary care physician, and a new
choice can be made every quarter of a year. Data pertaining to the primary care physicians included age,
gender, country of medical studies, and medical specialization.

Definition of variables

Identification of a pattern of FM-related initial complaints

Several patterns of initial complaints were retrospectively searched for within the period between
establishment of the database (1998 and in some cases as early as 1993) and the index date. Complaints
were categorized into the following distinct categories: pain, mood, neurological, sleep disturbances,
gastroenterological, and fatigue and malaise (Supplementary data, Table 2). The minimal time interval
for counting complaints proved to be 6 months, whereas a minimal set of 3 or 4 complaints proved to
be substantial enough to produce differences between groups. In due course, we created four most
relevant complaint models, each comprising a minimum set of complaints reported within a time
interval of 6 months (see Table 2). Combinations of these patterns were assessed as well. The pro-
portion of FM patients in whom a specific pattern was determined and regarded as “sensitivity” and the
proportion of general-control enrollees for whom the pattern was not found was regarded as “speci-
ficity.” The pattern with the lowest proportion of the sum of false-positive and false-negative patients
was chosen for further analysis of lagged time until diagnosis.

Assessing time to diagnosis

Time to diagnosis was assessed among all FM patients for whom we found the complaint pattern
that was eventually selected, i.e., complaint pattern B (see Table 2). These patients were termed

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx 5

Table 2
Complaint patterns found prior to FM diagnosis among “confirmed FM” patients, RA patients, and general controls (N ¼ 12,083)a.

Initial complaint pattern within any period of 6 months: Confirmed FM Rheumatoid arthritis General controls
N ¼ 2055 N ¼ 856 N ¼ 9172
% (n) % (n) % (n)

Pattern A
Two complaints from different categories of complaints 76.0% (1562) 49.9% (427) 35.8% (3286)
Pattern B
Any 4 different complaints from the detailed list of complaintsa 67.9% (1395) 49.1% (420) 27.6% (2529)
Pattern C
One pain complaint and two different nonpain complaints 35.0% (719) 16.6% (142) 8.2% (752)
Pattern D
One pain complaint and one nonpain complaint 70.5% (1448) 46.6% (399) 30.8% (2825)
a
All p-values<0.001.

“Complainers.” Time to diagnosis was assessed from 2 perspectives: a) “Total time to diagnosis”
(denoted “A” in Fig. 1, stage II) was defined as time from the beginning of the period of 6 months with
complaint pattern B to the diagnosis of FM, and b) “Time to diagnosis while associated with primary
care physician at diagnosis” (denoted “B” in Fig. 1, stage II) defined as the fraction of time while patients
were associated with the same primary care physician to whom they were associated with at the time
of diagnosis. This measure reflected the primary care physician's perspective.

Statistical analysis

The chi-square test and KruskaleWallis test were used for comparison of categorical and contin-
uous variables between study groups, respectively. Univariate and multivariate assessments of patient
characteristics associated with “total time to diagnosis” were made using generalized linear models
with gamma distribution and log link function. Assessment of “time to diagnosis by primary care
physicians” was made using generalized linear mixed models, taking into account the correlation
between patients associated with the same physician. Selection of variables for the final models was
based on minimal Akaike's information criterion and clinical relevance. The 2 FM patient groups were
unified in these models as the FM group variable did not contribute to model fit. Analyses were
conducted in IBM SPSS Statistics for Windows, Version 22.0 (Armonk, NY: IBM Corp.).
The study has been approved by research ethics board of MHS (approval number: 2/2013).

Results

Identification of a pattern of FM-related initial complaints

Characteristics of all study groups are presented in Table 1. FM patients were on average 7 years
younger than RA patients. A clear female preponderance was evident, more so among “confirmed FM”
and “primary care diagnosed FM” patients as compared with RA patients (91.1% and 87.7% vs. 71.4%,
respectively, p < 0.001). Most comorbid conditions were much more prevalent among FM patients than
both RA patients and general controls. Diagnoses of anxiety and/or depression were found in one third of
FM patients, twice as high as among RA patients, and thrice as that found in the general control group.
Rheumatic autoimmune conditions were also more prevalent among FM patients, with “confirmed FM”
patients generally presenting higher rates than “primary care diagnosed” FM patients (Table 1).
Table 2 presents the different complaint patterns that we identified as candidates for being initial
FM-related complaint patterns and their sensitivity (i.e., proportion of FM patients positive for each
pattern) and specificity (i.e., the proportion of general controls without each pattern). Pattern B,
namely, any constellation of at least four complaints during 6 months, was chosen as the initial
complaint pattern for further analysis, as it had the best combination of sensitivity (67.9%) and spec-
ificity (72.4%).

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
6 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx

Time to diagnosis

Time from initial complaints to FM diagnosis was assessed among 2,369 FM patients with complaint
pattern B who consisted 67.9% (N ¼ 1,440) of the “confirmed FM” group and 70.6% (N ¼ 974) of the
“primary care diagnosed FM” group. Approximately half (53.3%) of “complainers” were first diagnosed
with FM by a primary care physician, 42.8% by a rheumatologist, 1.6% by a physician with another
relevant specialization, and 2.2% during hospitalization. “Complainers” were on average ~3.5 years
older than “noncomplainers” (mean (SD): 51.6 (11.6) vs. 47.8 (10.8) years, respectively, p < 0.001), and
with ~0.3 point lower SES on a 1e10 scale (p < 0.001).
The mean (SD) total time to diagnosis among all FM “complainers” was 6.42 (3.57) years, with a
slightly short period among “primary care diagnosed” vs. “confirmed” FM patients (Table 3). Because
patients may switch their primary care physician, we also assessed the fraction of time to diagnosis
while the patients were under the continuous care of the same primary care physician as was at
diagnosis. This primary care physician was either the diagnosing physician or had referred patients to
secondary or tertiary care for diagnosis. This measure of time interval until diagnosis was shorter
among “primary care diagnosed” patients (3.54 years) than among “confirmed FM” patients (3.90
years) with p ¼ 0.003.

Associations between patient characteristics and time to diagnosis

In the unified FM patient group of “complainers” (N ¼ 2,369), total time to diagnosis decreased with
increasing patient age, with a mean of 4.2 years among those aged
75 vs. 7.3 years among patients
under 35 years (Table 4). Comorbidity with various conditions, of which some may actually be
symptoms of FM, was associated with greater time to diagnosis. Migraines were associated with the
largest increase in time to diagnosis, with a ratio of 1.33 in those with vs. without migraines, and sleep
and pain disorders showed the lowest effect (ratios of 1.14 and 1.12 in those with vs. without these
conditions, respectively (Table 4)).

Physician characteristics e time to diagnosis while associated with the same physician as at diagnosis

Using MHS database, we were able to identify 642 primary care physicians taking care of 2,112 FM
“complainers” at the time of FM diagnosis. To allow for enough time for diagnosis, only 549 physicians
born prior to 1970 were included in further analyses, of whom 56% were females, 45% completed
medical studies in Eastern Europe or former USSR, 31% in Israel, and others in Western Europe (11%) or
elsewhere (13%); 30% were general practitioners, 30% were family physicians, 24% held an internal
medicine specialization, and others had a different specialty or added a specialty during the study years
(16%). During the period between initial complaints and FM diagnosis, patients were associated, on
average, with 2.48 (SD ¼ 1.93) different primary care physicians, with 21.5% associated with
4

Table 3
Proportion of FM patients with initial complaint pattern B (“complainers”) and time from initial complaints to diagnosis among
“confirmed” and “primary care diagnosed” FM patients (N ¼ 3,434).

“Confirmed” FM “Primary-care diagnosed” FM Total p-value

N ¼ 2055 N ¼ 1379 N ¼ 3434

“Complainers”, % (n) 67.9% (1395) 70.6% (974) 69.0% (2369) 0.088

Time from initial complaints to diagnosis (years)
Total time
Mean (SD) 6.49 (3.55) 6.32 (3.60) 6.42 (3.57) 0.237
Median (IQR) 6.54 (3.39e9.41) 6.20 (3.26e9.27) 6.39 (3.32e9.37)
Time while patient was 3.90 (3.28) 3.54 (3.21) 3.75 (3.26) 0.003
associated with the same 2.99 (1.13e6.04) 2.46 (0.97e5.37) 2.80 (1.05e5.76)
primary physician as at time
of diagnosis

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx 7

Table 4
Association between patient characteristics and total time to diagnosis among all FM study patients with initial complaint
pattern B (N ¼ 2,369).

N Crude time till diagnosis (years) Mutually adjusted ratio of mean p-value
Mean (SD) years until diagnosis (95% CI)

Total 2369 6.42 (3.57)

Age at the beginning of initial complaints period (years)
<34 471 7.27 (3.48) 1.0 (ref.)
35-44 708 6.25 (3.64) 0.85 (0.79; 0.91) <0.001
45-54 746 6.35 (3.57) 0.85 (0.79; 0.92) <0.001
55-64 332 6.02 (3.40) 0.80 (0.73; 0.88) <0.001
65-74 88 5.93 (3.38) 0.76 (0.66; 0.88) <0.001
75þ 24 4.18 (2.86) 0.61 (0.47; 0.80) <0.001
Comorbidity
Anxiety and/or depression Yes 990 7.30 (3.52) 1.24 (1.17; 1.30) <0.001
No 1379 5.79 (3.47)
CFS Yes 206 7.62 (3.58) 1.17 (1.06; 1.28) 0.001
No 2163 6.31 (3.55)
Migraines Yes 167 8.45 (3.41) 1.33 (1.21; 1.47) <0.001
No 2202 6.27 (3.53)
OA Yes 262 7.27 (3.28) 1.21 (1.11; 1.32) <0.001
No 2107 6.32 (3.59)
Pain disorder Yes 582 7.12 (3.45) 1.12 (1.05; 1.18) <0.001
No 1787 6.19 (3.58)
Sleep disorder Yes 334 7.34 (3.33) 1.14 (1.06; 1.23) 0.001
No 2035 6.27 (3.58)

different physicians. Of all above-mentioned physician characteristics and patients' age, sex, and SES,
the only variable found to be associated with “time to diagnosis while associated with physician at
diagnosis” was physicians’ age, with a sharp decrease from ~4 years for those born prior to 1955 to 2.1
years for those born between 1965 and 1969 (Fig. 2).

Discussion

Over the decades, FM has been associated with both patient and physician beliefs and concepts,
leading its diagnosis to be overestimated, delayed or avoided [21,23e26]. Our study was designated to

6
same primary physician as at diagnosis (years)
Time to diagnosis while associated with the

Physician's
birth year

# Physicians 12 72 118 147 114 86

# Paents 39 238 458 524 365 287

Fig. 2. Time to diagnosis while associated with the same primary physician as at diagnosis by physicians' birth year: Generalized
linear mixed model (N ¼ 549 physicians of 1911 patients)1. Bars represent 95% confidence intervals.

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
8 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx

determine the current interval from initial FM-related complaints to the diagnosis of FM and to identify
patient and physician attributes that deter proper diagnosis.
The mean age at diagnosis of FM patients was 50.6 years, which is in accordance with previous
studies showing that the occurrence of FM is higher at the middle age (30e50 years) [27,28], while
other publications show a higher prevalence of FM over the age of 50 years old [19,29,30], these dis-
crepancies possibly reflect differences among populations and countries. The well-described female
predominance in chronic pain disorders and specifically in FM is clearly corroborated in our study,
where more than 90% of the patients with FM are women [18,31,32].
Our study results corroborate previous studies, indicating that despite their relatively younger
mean age, FM patients are more likely to present with more comorbid conditions than the general
population or than the patients with other conditions such as RA or OA [33e35]. In accordance with
previous studies, the prevalence of psychiatric comorbidity was significantly higher in the FM group
than that in RA patients and controls, with over 30% of the FM patients suffering from anxiety and/or
depression.
Similar to other clinical conditions establishing, the diagnosis of FM is not only crucial for effective
management but provides a relief for many patients. Understanding that a diagnosis exists and the fact
that other diagnoses are ruled out is often reassuring and may limit further medical expenses
[10,36,37]. White et al. [38] suggested that following the diagnosis of FM, patients showed significant
improvement in their satisfaction with health and reported fewer symptoms over a long period of time.
Interestingly, there is a paucity of data establishing the time to diagnosis of FM. In one international
study [25], FM patients had 7.3 symptoms on average out of 14 investigated. The mean time from
presenting these symptoms to a care provider to diagnosis of FM was 2.3 years. A similar survey by
Clark et al. [39], comprising 900 patients with FM and 1824 physicians across Latin America and
Europe, reported a longer average duration until diagnosis: 42.3 months among Latin Americans and
31.3 months among Europeans.
To the best of our knowledge, no research has investigated the time to FM diagnosis using a rela-
tively objective source, such as our study, nor delineated factors associated with it. The novelty and
challenge of this research were the attempt to find a pattern defining the time of initial complaints as
the starting point for calculating the time to diagnosis of FM.
The mean time to diagnosis in our study was 6.42 (SD 3.57) years, longer than shown in the previous
studies [25,39]. Our results indicate that FM patients tend to switch their health caregiver quite often,
perhaps also because of the frustration resulting from the delayed diagnosis and the process they go
through. Therefore, to assess the physician's perspective, we measured the extent of time patients were
continuously associated with by the same primary care physician to whom they were associated when
diagnosed. From the physician's perspective, time to diagnosis was significantly shorter, 3.75 years on
average.
We found patient's age to be crucial for time to diagnosis; patients younger than 34 years were
found to be diagnosed almost twofold as quickly as patients older than over 75 years. One possible
explanation could be that older individuals tend to present with more comorbidities, turning the
physician to be reluctant to use the diagnosis of FM. Another explanation is lack of knowledge
regarding the natural history of the disease, as reflected in many studies, or poor communication in
physicians' interactions with the elderly or that younger patients are probably much more likely to
make a self-diagnosis of FM by independent research of their condition online [2,3,18,40].
Another relevant variable in the complex route to diagnosis of FM is the health care providers.
Previous studies have shown several physician-related factors contributing to diagnostic inaccuracy
and delays, including lack of knowledge of the ACR diagnostic criteria, difficulty discussing the clinical
manifestations with patients, difficulty in distinguishing the symptoms of FM from other conditions,
and lack of confidence giving the diagnosis of FM [21,40]. Our study focused on some basic physician
characteristics; we found no association between physicians' gender and region of medical education
with the time to FM diagnosis. However, physician's age was closely related with time to diag-
nosisdthe older the physician, the longer it took to diagnose FMdwhereas physicians born before
1955 diagnosed almost twice slower than those born between before 1969. The promising diagnostic
outcomes of younger physicians probably reflect the contribution of inclusion of FM as part of the
syllabi of medical training in many schools and specialization programs around the globe.

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx 9

The main strength of our study originates from the large and comprehensive database we used.
These databases are “real-life” clinical records, encompassing more ten to 14 years of outpatient clinical
encounters. The main limitation of our study lies in the method we developed to identify when FM
patients first turned to health care givers with initial FM symptoms. Although the identified pattern of
complaints was much more prevalent in FM patients than in non-FM controls, it could only be found
among 70% of all FM patients. Thus, our assessment of time to diagnosis can only apply to these pa-
tients. However, it calls for alarm even if only 70% of FM patients are diagnosed with a mean time of
over 6 years. We were, however, able to detect a large difference in frequency of FM-related diagnoses
between FM patients and non-FM controls. We may speculate that some non-FM patients for whom
the complaint pattern defining initial FM-related complaints may actually be undiagnosed FM patients.
Further follow-up can further provide more insight into this patient group.
Another constitutional limitation in the study is its retrospective design being based on diagnosed
FM patients, thus limiting us to the physicians who diagnose the syndrome. This generates a certain
biased selection of the physician population, excluding physicians who do not diagnose FM at all.
In conclusion, FM remains a very complex and elusive diagnosis. This research demonstrates that
specific patient- and physician-dependent factors that are related with the time to diagnosis, including
patient's age and comorbidity, as well as diagnoses of migraines, anxiety and depression, and the
physician's age. The novelty of this study is the use of a model of initial complaints that can serve as an
approximate baseline to predict and possibly expedite diagnosis of FM.

Practice points

 In many patients, a long duration precedes proper diagnosis with FM.

 The purpose of this study was to investigate the “average” time to diagnosis in a data
analysis of “real-life” population of an Israeli HMO.

Research agenda

 To pursue ways to facilitate proper diagnosis of patients with pain syndromes in general and
of fibromyalgia in particular.
 To provide pragmatic tools that will improve the methodology of primary care providers to
diagnose properly fibromyalgia.

Funding statement

The study was supported by a research grant by Pfizer, Israel, AbbVie, Israel, and Janssen, Israel.

Conflicts of interest

None of the authors has any conflict of interests regarding this publication.

Disclosures and acknowledgments

The study was supported by an educational grant by Pfizer Israel and also AbbVie Israel and Janssen
Israel. However, this should not constitute a financial conflict of interest in relation to this manuscript.
Personal conflicts of interest: Omer Gendelman e none declared; Howard Amital e none declared; Yael
Bar-On e none declared; Dana Ben-Ami Shor e none declared; Daniela Amital e none declared;
Shmuel Tiosano e none declared; Varda Shalev- none declared; Gabriel Chodick e none declared;
Dahlia Weitzman e none declared.

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
10 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx

Provenance and peer-review

Commissioned, externally peer-reviewed.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.berh.2019.01.019.

References

[1] Clauw DJ. Fibromyalgia and related conditions. Mayo Clin Proc 2015;90:680e92.
[2] Yunus MB. Editorial review: an update on central sensitivity syndromes and the issues of nosology and psychobiology.
Curr Rheumatol Rev 2015;11:70e85.
[3] Walitt B, Nahin RL, Katz RS, Bergman MJ, Wolfe F. The prevalence and characteristics of fibromyalgia in the 2012 national
health interview survey. PLoS One 2015;10, e0138024.
[4] Wolfe F, Ross K, Anderson J, Russell IJ, Hebert L. The prevalence and characteristics of fibromyalgia in the general pop-
ulation. Arthritis Rheum 1995;38:19e28.
[5] Perez dH-T, Huertas-Hoyas E, Maximo-Bocanegra N, Palacios-Cena D, Fernandez-De-Las-Penas C. Cognitive performance
in women with fibromyalgia: a case-control study. Aust Occup Ther J 2016;63:329e37.
[6] Tran ST, Guite JW, Pantaleao A, Pfeiffer M, Myer GD, Sil S, et al. Preliminary outcomes of a cross-site cognitive-behavioral
and neuromuscular integrative training intervention for juvenile fibromyalgia. Arthritis Care Res (Hoboken) 2017;69:
413e20.
[7] Arnold BS, Alpers GW, Suss H, Friedel E, Kosmutzky G, Geier A, et al. Affective pain modulation in fibromyalgia, soma-
toform pain disorder, back pain, and healthy controls. Eur J Pain 2008;12:329e38.
[8] Arnold LM, Palmer RH, Gendreau RM, Chen W. Relationships among pain, depressed mood, and global status in fibro-
myalgia patients: post hoc analyses of a randomized, placebo-controlled trial of milnacipran. Psychosomatics 2012;53:
371e9.
[9] Spaeth M. Epidemiology, costs, and the economic burden of fibromyalgia. Arthritis Res Ther 2009;11:117.
[10] Berger A, Dukes E, Martin S, Edelsberg J, Oster G. Characteristics and healthcare costs of patients with fibromyalgia
syndrome. Int J Clin Pract 2007;61:1498e508.
[11] Berger A, Sadosky A, Dukes E, Martin S, Edelsberg J, Oster G. Characteristics and patterns of healthcare utilization of
patients with fibromyalgia in general practitioner settings in Germany. Curr Med Res Opin 2008;24:2489e99.
[12] Wolfe F, Clauw DJ, Fitzcharles MA, Goldenberg DL, Katz RS, Mease P, et al. The American College of Rheumatology pre-
liminary diagnostic criteria for fibromyalgia and measurement of symptom severity. Arthritis Care Res (Hoboken) 2010;62:
600e10.
[13] Wolfe F, Clauw DJ, Fitzcharles MA, Goldenberg DL, Hauser W, Katz RS, et al. Fibromyalgia criteria and severity scales for
clinical and epidemiological studies: a modification of the ACR Preliminary Diagnostic Criteria for Fibromyalgia.
J Rheumatol 2011;38:1113e22.
[14] Wolfe F, Clauw DJ, Fitzcharles MA, Goldenberg DL, Hauser W, Katz RL, et al. 2016 Revisions to the 2010/2011 fibromyalgia
diagnostic criteria. Semin Arthritis Rheum 2016;46:319e29.
[15] Hayes SM, Myhal GC, Thornton JF, Camerlain M, Jamison C, Cytryn KN, et al. Fibromyalgia and the therapeutic relationship:
where uncertainty meets attitude. Pain Res Manag 2010;15:385e91.
[16] Arnold LM, Clauw DJ, McCarberg BH. Improving the recognition and diagnosis of fibromyalgia. Mayo Clin Proc 2011;86:
457e64.
[17] Central Bureau of Statistics. Characterization and classification of geographical units by the socio-economic level of the
population. Jerusalem: Central Bureau of Statistics; 2013. Publication no. 1530.
[18] Buskila D, Neumann L, Odes LR, Schleifer E, Depsames R, Abu-Shakra M. The prevalence of musculoskeletal pain and fi-
bromyalgia in patients hospitalized on internal medicine wards. Semin Arthritis Rheum 2001;30:411e7.
[19] Wolfe F, Brahler E, Hinz A, Hauser W. Fibromyalgia prevalence, somatic symptom reporting, and the dimensionality of
polysymptomatic distress: results from a survey of the general population. Arthritis Care Res (Hoboken) 2013;65:777e85.
[21] Perrot S, Choy E, Petersel D, Ginovker A, Kramer E. Survey of physician experiences and perceptions about the diagnosis
and treatment of fibromyalgia. BMC Health Serv Res 2012;12:356.
[23] Arshad A, Kong KO. Awareness and perceptions of fibromyalgia syndrome: a survey of southeast asian rheumatologists.
J Clin Rheumatol 2007;13:59e62.
[24] Bjorkegren K, Wallander MA, Johansson S, Svardsudd K. General symptom reporting in female fibromyalgia patients and
referents: a population-based case-referent study. BMC Public Health 2009;9:402.
[25] Choy E, Perrot S, Leon T, Kaplan J, Petersel D, Ginovker A, et al. A patient survey of the impact of fibromyalgia and the
journey to diagnosis. BMC Health Serv Res 2010;10:102.
[26] Fitzcharles MA, Esdaile JM. The overdiagnosis of fibromyalgia syndrome. Am J Med 1997;103:44e50.
[27] Mas AJ, Carmona L, Valverde M, Ribas B. Prevalence and impact of fibromyalgia on function and quality of life in in-
dividuals from the general population: results from a nationwide study in Spain. Clin Exp Rheumatol 2008;26:519e26.
[28] Senna ER, De Barros AL, Silva EO, Costa IF, Pereira LV, Ciconelli RM, et al. Prevalence of rheumatic diseases in Brazil: a study
using the COPCORD approach. J Rheumatol 2004;31:594e7.
[29] Branco JC, Bannwarth B, Failde I, Abello Carbonell J, Blotman F, Spaeth M, et al. Prevalence of fibromyalgia: a survey in five
European countries. Semin Arthritis Rheum 2010;39:448e53.
[30] Makela M, Heliovaara M. Prevalence of primary fibromyalgia in the Finnish population. BMJ 1991;303:216e9.

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019
 O. Gendelman et al. / Best Practice & Research Clinical Rheumatology xxx (xxxx) xxx 11

[31] Haviland MG, Banta JE, Przekop P. Fibromyalgia: prevalence, course, and co-morbidities in hospitalized patients in the
United States, 1999e2007. Clin Exp Rheumatol 2011;29:S79e87.
[32] Quintero OL, Amador-Patarroyo MJ, Montoya-Ortiz G, Rojas-Villarraga A, Anaya JM. Autoimmune disease and gender:
plausible mechanisms for the female predominance of autoimmunity. J Autoimmun 2012;38:J109e19.
[33] Brooks L, Hadi J, Amber KT, Weiner M, La Riche CL, Ference T. Assessing the prevalence of autoimmune, endocrine, gy-
necologic, and psychiatric comorbidities in an ethnically diverse cohort of female fibromyalgia patients: does the time
from hysterectomy provide a clue? J Pain Res 2015;8:561e9.
[34] Rehm SE, Koroschetz J, Gockel U, Brosz M, Freynhagen R, Tolle TR, et al. A cross-sectional survey of 3035 patients with
fibromyalgia: subgroups of patients with typical comorbidities and sensory symptom profiles. Rheumatology (Oxford)
2010;49:1146e52.
[35] Wolfe F, Hawley DJ. Evidence of disordered symptom appraisal in fibromyalgia: increased rates of reported comorbidity
and comorbidity severity. Clin Exp Rheumatol 1999;17:297e303.
[36] Goldenberg DL. Fibromyalgia syndrome: an emerging but controversial condition. JAMA 2014;311:1570.
[37] Rahman A, Underwood M, Carnes D. Fibromyalgia. BMJ 2014;348:g1224.
[38] White KP, Nielson WR, Harth M, Ostbye T, Speechley M. Does the label "fibromyalgia" alter health status, function, and
health service utilization? A prospective, within-group comparison in a community cohort of adults with chronic
widespread pain. Arthritis Rheum 2002;47:260e5.
[39] Clark P, Paiva ES, Ginovker A, Salomon PA. A patient and physician survey of fibromyalgia across Latin America and Europe.
BMC Muscoskelet Disord 2013;14:188.
[40] Buskila D, Neumann L, Sibirski D, Shvartzman P. Awareness of diagnostic and clinical features of fibromyalgia among
family physicians. Fam Pract 1997;14:238e41.

Please cite this article as: Gendelman O et al., Time to diagnosis of fibromyalgia and factors associated
with delayed diagnosis in primary care, Best Practice & Research Clinical Rheumatology, https://doi.org/
10.1016/j.berh.2019.01.019