Abstract
Objective. To develop feasible indices as clues to comorbid fibromyalgia (FM) in routine care of patients with various rheumatic diseases based only on self-report multidimensional Health Assessment Questionnaire (MDHAQ) scores, which are informative in all rheumatic diagnoses studied.
Methods. All patients with all diagnoses complete an MDHAQ at each visit; the 2011 FM criteria questionnaire was added to the standard MDHAQ between February 2013 and August 2016. The proportion of patients who met 2011 FM criteria or had a clinical diagnosis of FM was calculated. Individual candidate MDHAQ measures were compared to 2011 FM criteria using receiver-operating characteristic (ROC) curves; cutpoints to recognize FM were selected from the area under the curve (AUC) for optimal tradeoff between sensitivity and specificity. Cumulative indices of 3 or 4 MDHAQ measures were analyzed as fibromyalgia assessment screening tools (FAST).
Results. In 148 patients, the highest AUC in ROC analyses versus 2011 FM criteria were seen for MDHAQ symptom checklist, self-report painful joint count, pain visual analog scale (VAS), and fatigue VAS. The optimal cutpoints were ≥ 16/60 for symptom checklist, ≥ 16/48 for self-report painful joint count, and ≥ 6/10 for both pain and fatigue VAS. Cumulative FAST indices of 2/3 or 3/4 MDHAQ measures correctly classified 89.4–91.7% of patients who met 2011 FM criteria.
Conclusion. FAST3 and FAST4 cumulative indices from only MDHAQ scores correctly identify most patients who meet 2011 FM criteria. FAST indices can assist clinicians in routine care as clues to FM with a general rheumatology rather than FM-specific questionnaire.
Fibromyalgia (FM) is a common disorder in the general population1, with a considerably higher prevalence in people with rheumatic diseases2,3,4,5. As with most rheumatic diseases, no gold standard diagnostic marker is available for FM. In an effort to standardize the identification of FM, the American College of Rheumatology (ACR) developed FM classification criteria in 19906, based on the clinical information of 558 consecutive patients, 293 with FM according to experts in FM versus 265 controls. The combination of widespread pain and > 11 over 18 tender points identified on physical examination yields a sensitivity of 88% and specificity of 81%. These criteria were modified in 2010, with no tender point examination, but rather a patient questionnaire for widespread body pain and symptom severity as well as physician ratings, and designated as “preliminary diagnostic criteria” rather than classification criteria7. A further modification was introduced in 2011, based entirely on the patient self-report widespread pain index and symptom severity scale questionnaire, without any requirement for physician assessment, and was designed for “epidemiologic and clinical studies” but not intended for diagnosis of FM8.
FM criteria are not used in most routine clinical care, other than by subspecialists. It is not feasible to ask patients with different diagnoses to complete different patient self-report questionnaires in busy clinical settings. As a consequence, clinicians may both fail to identify or incorrectly identify patients as having FM9.
A multidimensional Health Assessment Questionnaire (MDHAQ) has been found informative in most rheumatic diseases studied10, based on a RAPID3 (Routine Assessment of Patient Index Data 3 — an index within the MDHAQ). Additional MDHAQ scales have been reported to provide clues to FM, including a high pain score relative to physical function score11 and a high number of positive responses on a symptom checklist12.
This study examines an additional feature of the MDHAQ to provide clues to the presence of comorbid FM, without the need for administration of a separate, condition-specific FM questionnaire, recognizing that a clinical FM diagnosis requires the judgment of a clinician to synthesize all available information about a patient.
MATERIALS AND METHODS
Patients
All patients with all diagnoses at Liverpool Hospital, Australia, complete an MDHAQ at all visits before seeing the rheumatologist. The treating clinician assigns 1 or more diagnoses at the first and subsequent patient visits. Most of the patients included in this study (79%) had their first clinic visit and diagnosis(es) assigned prior to the study onset. All clinical diagnoses and all MDHAQ data are included in a database. All data were exported for analyses to STATA 12.0 for Mac (StataCorp LP).
Patients seen between February 2013 and August 2016 by KAG at a weekly clinic also completed a 2011 FM criteria questionnaire at the same time as the MDHAQ questionnaire to compare with candidate FM Assessment Screening Tool (FAST) indices. Patients included in the study were older than 18 years, had a primary diagnosis of any rheumatic condition except FM, and had complete MDHAQ and 2011 FM criteria cross-sectional data at a random visit during the study period. The primary diagnosis was assigned by the treating physician, who did not examine the 2011 FM criteria results when each diagnosis was assigned. Approval by the district Human Research and Ethics Committee was obtained for this study (LNR/13/LPOOL/370 Local project number 13/229LNR). All patients consent to the use of their anonymized data for research purposes, including publication.
Design of MDHAQ
The MDHAQ (Figure 1) is designed for use in routine care with a primary purpose to improve the quality of clinical care and patient outcomes13. It includes 10 queries concerning activities of daily living to evaluate physical function (FN) and 3 (0–10) visual numerical scales (VNS) for pain, patient’s global assessment (PtGA), and fatigue. RAPID3 is a composite index that includes the 3 rheumatoid arthritis (RA) core dataset measures, FN, pain, and PtGA, each scored 0–10 for a total of 0–30. MDHAQ/RAPID3 has been found informative in osteoarthritis (OA), systemic lupus erythematosus (SLE), ankylosing spondylitis (AS), psoriatic arthritis (PsA), vasculitis, gout, and FM, in addition to RA10,12,14–20.
The form for the multidimensional Health Assessment Questionnaire. Copyright: Health Report Services. Used with permission.
The MDHAQ also includes 3 queries to rate how difficult it was to “get a good night’s sleep” (sleep quality), to “deal with feelings of anxiety or being nervous” (anxiety), and to “deal with feelings of depression or feeling blue” (depression) in the traditional, patient-friendly HAQ format21. A self-report RA Disease Activity Index (RADAI) painful joint count is recorded22 on the MDHAQ. The RADAI queries patients to score pain in 16 specific joint groups, 8 each on the right and left sides: fingers, wrists, elbows, shoulders, hips, knees, ankles, and toes. Scoring options are 0 (none), 1 (mild), 2 (moderate), or 3 (severe) pain; total scores range from 0 to 48. RADAI self-report joint counts have been shown to be useful in patients with different rheumatic diseases23. In addition, the MDHAQ includes a 60-symptom checklist12 and recent medical history information24. Demographic data include sex, date of birth, ethnicity, and years of formal education25.
2011 FM criteria
The 2011 FM criteria8 were developed as a patient self-report questionnaire for epidemiologic and clinical studies, and consist of 2 scales: the Widespread Pain Index (WPI) and the symptom severity scale (SSS). The WPI component queries patients to indicate whether they have had pain or tenderness over the previous week in specific joints and areas, including the shoulder girdle, hip, jaw, upper back, lower back, upper arm, upper leg, chest, neck, abdomen, lower arm, and lower leg. Patients grade the right and left side of the body separately from 0 or 1 with a maximum total score of 19. The SSS queries patients to indicate the severity of 6 symptoms over the previous week. Three symptoms [fatigue, trouble thinking or remembering, and waking up tired (unrefreshed)] are scored 0–3: 0, no problem; 1, slight or mild problems, generally mild or intermittent; 2, moderate, considerable problems, often present and/or at a moderate level; and 3, severe, continuous, life-disturbing problems. The other 3 symptoms [pain or cramps in the lower abdomen, depression, or headache during the previous 6 months] are scored 0–1 (0 No, 1 Yes). When summed, the 6 items result in a total score between 0 and 12. A patient meets 2011 FM criteria if the following conditions are met: WPI ≥ 7 and SSS ≥ 5 or WPI between 3–6 and SSS ≥ 9, symptoms have been present at a similar level for at least 3 months, and the patient does not have a disorder that would otherwise sufficiently explain the pain. These criteria are used in our study as the reference standard against which the proposed MDHAQ/FAST indices are compared.
FAST indices based on MDHAQ scores
FAST composite cumulative indices were developed from MDHAQ candidate measures, including FN, pain visual analog scale (VAS), PtGA VAS, RAPID3, sleep quality, anxiety, depression, fatigue VAS, RADAI painful joint count, and a 60-symptom checklist. Receiver–operating characteristic (ROC) curves were constructed to recognize those that provided the highest area under the curve (AUC) compared to the 2011 FM criteria as the reference standard (Figure 2). ROC curves were computed for RAPID3 and PtGA, but these variables were not regarded as candidates for FAST indices because they appeared clinically more likely to be associated with somatic symptoms that could be observed by a clinician, such as dyspnea or joint swelling, and were not included in the 2011 FM criteria. The cutpoint for each score to be included in a FAST cumulative index was selected on the basis of the best tradeoff values between sensitivity and specificity on the ROC curve. An optimal cutpoint was identified for each measure on the basis of the ROC curve for that measure.
Receiver-operating characteristic (ROC) curve analysis of multidimensional Health Assessment Questionnaire (MDHAQ) variables versus 2011 FM criteria. FM: fibromyalgia; PN: pain; PATGL: patient’s global assessment; RAPID3: Routine Assessment of Patient Index Data; RADAI: Rheumatoid Arthritis Disease Activity Index; JC: joint count; FN: function; FT: fatigue.
Statistical analysis
The proportions of patients who met 2011 FM criteria or a clinical diagnosis of FM [assigned by the treating rheumatologist (KAG)] were calculated. Means and SD of demographic measures and clinical characteristics in patients who did and did not meet FM by criteria were compared using t tests, and percentages using chi-square tests.
FAST3 and FAST4 composite cumulative indices were constructed from the optimal cutpoints of the candidate MDHAQ measures with the highest AUC under the ROC curves as described above. Agreement of FAST indices with the 2011 FM criteria as the reference standard was analyzed using the κ statistic26 and ROC curves (of the indices) for AUC.
RESULTS
Patient characteristics
The study included 148 patients: 55 (37%) with RA, 21 (14%) with OA, 14 (10%) with PsA, and the remaining 58 (39%) with other rheumatologic diagnoses (Table 1). Among these patients, 24% with RA, 38% with OA, and 21% with PsA met 2011 criteria, while 22%, 38% and 29%, respectively, had a clinical diagnosis of comorbid FM. There were 13%, 29%, and 14%, respectively, who both met 2011 FM criteria and had a clinical diagnosis of FM (Table 1). Among all patients, 29 (20%) met the 2011 FM criteria, 31 (21%) were assigned a clinical diagnosis of FM, 18 (12%) had both a clinical diagnosis and met criteria, while 106 (72%) met neither a clinical diagnosis nor FM criteria (Table 1). Agreement between clinical and 2011 FM criteria was moderate (83.8%, κ 0.50, p < 0.001).
Proportion of patients with different diseases meeting 2011 FM criteria and/or a clinical diagnosis of FM.
The majority of patients (80.4%) was female and white (68%) or Asian (12%). No statistically significant differences were seen in ethnicity or education level between those who did or did not meet 2011 FM criteria in this study (Table 2).
Characteristics of patients according to 2011 FM criteria.
Comparison of MDHAQ scores with 2011 FM criteria
Patients who met 2011 FM criteria had significantly poorer scores on all MDHAQ scales, including for physical function, pain VAS, patient global VAS, self-report joint counts, symptom checklist, and fatigue VAS, compared to patients who did not meet criteria (p < 0.001; Table 2). ROC analyses indicated that the highest AUC values were seen for symptom checklist (0.926), self-report joint count (RADAI; 0.917), RAPID3 (0.881), pain VAS (0.887), PtGA (0.831), and fatigue VAS (0.811). Lower AUC versus the 2011 FM criteria (range 0.716–0.799) were seen for remaining MDHAQ scores including physical function, anxiety, depression, and sleep quality (Figure 2). These variables were not analyzed further despite being significantly different from scores in people with no FM, to develop an optimal, feasible index for clinical settings. PtGA and RAPID3 were not regarded as candidates for FAST indices.
The optimal cutpoints were ≥ 16 for symptom checklist and self-report painful joint count and ≥ 6 for pain and fatigue VAS. FAST3 cumulative indices include a cumulative score of the symptom checklist, painful joint count, and either a pain VAS called FAST3-P, or fatigue VAS called FAST3-F. A FAST4 cumulative index includes a cumulative score of the symptom checklist, painful joint count, and both pain and fatigue VAS (Figure 1).
FAST3-P of ≥ 2 correctly classified 90.9% of patients versus 2011 FM criteria, with a sensitivity of 81.5% and a specificity of 93.3% (Table 3). FAST3-F ≥ 2 correctly classified 89.4% of patients versus 2011 FM criteria, with a sensitivity of 77.8% and a specificity of 92.4% (Table 3). FAST4 ≥ 3 correctly classified 91.7% of patients versus 2011 FM criteria, with a sensitivity of 70.4% and a specificity of 97.1% (Table 3).
Sensitivity and specificity of MDHAQ measures and FAST indices versus 2011 FM criteria.
Agreement with a clinical diagnosis of FM was 81.1% for FAST3-P, 85.6% for FAST3-F, and 58.8% for FAST4 (Table 4). The κ values for FAST3–P, FAST3–F, and FAST4 versus 2011 criteria were 0.73, 0.68, and 0.72, respectively, and versus the clinical diagnosis were 0.44, 0.57, and 0.51, respectively (Table 4).
Classification of FM according to FAST3-P, FAST3-F, and FAST4 compared to 2011 FM criteria and clinical diagnosis (n = 132).
DISCUSSION
We found that FAST indices based on MDHAQ scales provide clues to the presence of FM in patients with other rheumatic diseases compared to the 2011 FM criteria as the reference standard, against which the proposed FAST criteria were examined in this study. The 2011 FM criteria are designated in the report title as “for clinical and epidemiological studies,”8 but also may provide clues to the diagnosis of FM in the clinic. The total 2011 FM criteria score derived from this questionnaire as a continuous variable also provides information regarding the degree of “fibromyalgianess” that patients experience27.
The 2011 report indicated that 21.1% of patients with RA, 16.8% with OA, and 36.7% with SLE met the revised FM criteria8. The symptom severity burden is similar in patients with other rheumatic diagnoses, and it has been suggested that a distinction between “primary” and “secondary” FM is not needed28. This recommendation supports use of these criteria to classify patients who have other rheumatic diagnoses and FM in the current report.
Agreement between the 2011 FM criteria and the clinical diagnosis of FM in this study was moderate (83.8%, κ 0.50, p < 0.001), a level somewhat higher than reported in another study (79.2%, κ 0.41)9. It is noteworthy that about 25% of previously diagnosed patients with FM did not satisfy the ACR 1990 classification criteria at the time of the 2010 study7, and the modified ACR 2010 criteria were satisfied by 60% with a prior diagnosis of FM in the 2011 study8. These minor differences reflect differences between the clinicians, patients, and clinical environments in these studies, but general agreement in the range of 70–85% has been found.
One advantage of using the MDHAQ to identify FM is that it has been found informative in all rheumatic diseases studied10. It is not feasible in most busy clinical settings to add administration of another questionnaire, such as an additional FM-specific questionnaire. In addition, the FAST3 and FAST4 scores are easily calculated directly from the MDHAQ during the consultation.
Previous reports have identified clues to diagnosis and characterization of FM exclusively by self-report on the Fibromyalgia Impact Questionnaire29 and MDHAQ11,12. FAST3 and FAST4 indices provide relatively high sensitivity and specificity for alerting the clinician to a possible diagnosis of FM compared to the 2011 FM criteria. An additional report from Rush University supports the use of MDHAQ indices to provide clues to the diagnosis of FM30.
This study has a number of limitations. First, it is a cross-sectional study and patient symptoms may fluctuate over time; further longitudinal data may help to recognize the possible stability or sensitivity to change of the FAST indices, as well as whether a particular FAST3 or FAST4 may perform substantially better than others. Second, dichotomous criteria for the presence or absence of FM may obscure FM as a “spectrum” disorder of “fibromyalgianess”27, although cutpoints may serve as convenient “anchors” for many continuous variables, including laboratory tests such as erythrocyte sedimentation rate, serum glucose, etc. Third, this study does not use the 2016 modification of the FM criteria31, which was reported after commencement of this study, although no patients in the study had regional pain syndromes; therefore, it appears unlikely the modifications would have meaningfully changed the results. Fourth, FAST3 and FAST4 scores misclassify some patients according to the 2011 criteria; and further investigation in larger populations of patients with rheumatic disease is required.
It is not suggested that FAST indices are in any way a substitute for a careful evaluation to determine whether a patient has FM, or that there may not exist other reasons for certain patients to meet cutpoint values for FAST3 and FAST4. A definitive diagnosis of FM requires a careful history and examination of the patient, and interpretation of laboratory tests and other ancillary data. Nevertheless, it may be of considerable value to assess the presence of FM in patients with other rheumatic diseases using quantitative criteria from the 2011 FM questionnaire or the FAST indices to assist clinicians in recognizing this frequent diagnosis.
We have developed simple FAST cumulative, composite indices that provide clues to the presence of FM in patients with primary diagnoses of other rheumatic diseases. These indices are based entirely on the patient self-report MDHAQ questionnaire, which is used in routine care in the setting of the research and other rheumatology settings, and does not require a disease-specific FM questionnaire. FAST3 provides greater sensitivity but lower specificity than FAST4 for clues to FM; it may be desirable to use a FAST3-P index as the first screening tool for greater sensitivity, and check positive patients according to FAST4 for greater specificity. The FAST indices can be incorporated easily into routine care and may assist clinicians to identify patients with FM in the context of other comorbid rheumatic diseases. They remain a tool to support clinical judgment and are not a substitute for appropriate and thorough clinical history, physical examination, and comprehensive patient assessment for a diagnosis of FM.
Footnotes
Dr. Pincus holds a copyright and trademark on MDHAQ and RAPID3 (Routine Assessment of Patient Index Data 3), for which he receives royalties and license fees, all of which are used to support further development of quantitative questionnaire measurements for patients and doctors in clinical rheumatology care.
- Accepted for publication August 6, 2019.
