The ability of rheumatologists blinded to prior workup to diagnose rheumatoid arthritis only by clinical assessment: a cross-sectional study

The ability of rheumatologists blinded to prior workup to diagnose rheumatoid arthritis only by... Abstract Objectives We aimed to study the ability of board-certified rheumatologists, blinded to all prior diagnostic test results, to establish the presence/absence of an inflammatory rheumatic disease (IRD) or RA among polyarthralgia or arthritis patients, solely relying on clinical assessment. Methods We performed a prospective, examiner-blinded, cross-sectional study documenting the diagnostic work in four sequential steps (medical history, physical examination, musculoskeletal ultrasonography and laboratory tests) of board-certified rheumatologists in a convenience cohort of 100 patients referred for inpatient diagnostic workup to a tertiary care rheumatology centre. Results The ability to correctly identify patients with or without an IRD (diagnostic accuracy) increased from 27% after the clinical assessment to 53% after the ultrasonography and to 70% after taking laboratory test results into account. The corresponding values for correctly identifying patients with or without RA were 19, 42 and 60%, respectively. Therefore the diagnostic accuracy of solely clinical assessment for determining the diagnosis of IRD or RA compared with the diagnosis established by a consecutive thorough in-patient workup was only 27 and 19% in our cohort, respectively. Pretreatment with corticosteroids (in the prior 7 days) vs none did not alter these results substantially (20 vs 29% for IRD, 15% vs 20% for RA). Conclusion Experienced rheumatologists, if deprived of information on prior external imaging and laboratory workup by blinding, were not able to correctly classify the majority of patients presenting with polyarthralgia or arthritis symptoms for inpatient workup, relying only on a brief symptom-focused medical history and physical examination. arthritis, diagnosis, medical history taking, physical examination, rheumatoid, ultrasonography Rheumatology key messages Blinded to prior workup, rheumatologists could not correctly classify most patients when relying only on medical history and examination. The addition of ultrasonography and laboratory tests increased rheumatologists’ diagnostic accuracy for RA from 27 to 70%. Pretreatment with corticosteroids vs none did not significantly worsen rheumatologists’ diagnostic performance. Introduction Most clinicians would agree that taking a thorough medical history and performing a physical examination are the key components of establishing the correct diagnosis for many diseases [1–5]. In the advent of ever more precise laboratory markers and better imaging techniques, rheumatologists sometimes wonder to what extent these diagnostic tools should augment or even have already partially replaced their clinical judgement in daily clinical practice. With the introduction of scanners that enable the depicting of soft tissues with high resolution, musculoskeletal ultrasonography has become a powerful tool to instantly image inflammatory and structural changes of joints and tendons, and often these results are available even prior to receiving the results of ordered laboratory tests. Hence some rheumatologists, especially in Germany, with >80% of respondents in an international survey published in 2012, perform ultrasonography examinations of joints as an integral part of the diagnostic workup in clinical practice [6]. This widespread adoption of imaging and laboratory workup early in the management of patients with joint symptoms has led to a blurred line between clinical and imaging examinations in daily clinical practice, where these examinations are performed simultaneously by some rheumatologists. Therefore, nowadays, the exact diagnostic yield of solely clinically assessing patients without musculoskeletal ultrasonography to definitively diagnose inflammatory rheumatic diseases (IRDs), and especially RA, has been debated among experts [4, 7]. While some emphasize that visualizing the amount of synovitis in joints and tendon sheets of clinically symptomatic and asymptomatic joints can greatly enhance the ability to diagnose RA [7, 8], others point out that in most of the examined patients the diagnosis could have been reached without performing musculoskeletal ultrasonography [4]. As specialists, rheumatologists are rarely the first physicians patients turn to with new symptoms indicating a possible IRD. Therefore most patients seen in daily clinical practice provide the rheumatologist with the results of externally obtained tests and imaging. Thus the diagnostic accuracy of a solely clinical assessment (only obtaining a medical history and performing a physical examination) to diagnose a patient with RA cannot be studied reliably by just observing rheumatologists in daily clinical practice, because of bias introduced to clinical judgement by preeminent external laboratory test and imaging results. We aimed to prospectively study the contribution of solely clinical assessment (taking a medical history and performing a physical examination focused on joint symptoms) compared with a diagnostic workup including ultrasonography of selected joints and results of laboratory tests for establishing the diagnosis of an IRD and the diagnosis and classification of RA in patients with thus far unexplained joint symptoms. Methods Study design We performed a prospective, examiner-blinded study documenting the diagnostic work of board-certified rheumatologists with long-standing experience in musculoskeletal ultrasonography in a convenience cohort of 100 patients referred for diagnostic workup to a tertiary care rheumatology centre. Patients were recruited into the study at the time of their hospital admission if they had not yet received a rheumatologist-established diagnosis of an IRD, had not been treated with conventional or biologic DMARDs and reported a history of or present arthralgia of at least five peripheral joints and/or reported a history of or present swelling of at least one joint with an overall symptom duration of no longer than 2 years. Prior treatment with NSAIDs and corticosteroids was permitted and the extent of such pretreatment was categorically recorded. All patients provided written consent after receiving oral and written information about the study by one of the four authors. The study was reviewed and approved by the institutional review board/ethics committee of the University Medical Center of Regensburg (07/125). The study was conducted in compliance with the Declaration of Helsinki, International Conference on Harmonization Good Clinical Practice guidelines and local country regulations. Participants were instructed not to provide any unsolicited information pertaining to their prior external diagnostic workup, other pre-existing rheumatic diseases or prior drug treatment to the second rheumatologist conducting the diagnostic evaluation. The second rheumatologist (one of the authors: B.E., G.P. or W.H., all board-certified rheumatologists), with a mean of 16 years of clinical experience and a mean of 10 years of experience in musculoskeletal ultrasonography, was always different from the rheumatologist that had obtained the informed consent and was therefore blinded to all prior externally conducted diagnostic workups and performed and documented the following four sequential steps: (1) obtaining a medical history focused on the joint-related symptoms, (2) performing a physical examination focused on joint-related signs, (3) conducting musculoskeletal ultrasonography of selected peripheral joints and (4) arriving at a final judgement, including results of laboratory tests. After each of the four steps the blinded rheumatologist reported the time required for the step and the degree of certainty for establishing the diagnosis of an IRD and of RA using four pre-defined categories (definite, probable, possible, absent) and the presence of probable (>10% possibility) differential diagnoses. Furthermore, for the last three of the four evaluation steps, fulfilment of the ACR/EULAR RA classification criteria from 2010 was recorded [9]. After the evaluation by the blinded examiner, patients received their scheduled thorough workup as part of the standard in-patient clinical care in our institution. We defined the rheumatic diagnosis reported on the discharge summary report as the final diagnosis for each patient. After having completed and documented their four-step research evaluation, the three blinded rheumatologists (B.E., G.P. and W.H.) contributed later to the routine clinical care of some of the patients, but to prevent bias, all discharge diagnoses were adjudicated by the fourth author (M.F.) as part of his responsibilities as head of the department during routine clinical care. Medical history The medical history obtained by the blinded rheumatologist focused on symptoms pertaining to rheumatic inflammatory diseases, especially typical symptoms of RA. The examiners were instructed to not inquire about externally performed diagnostic tests, suspected or given diagnoses of rheumatic diseases and specific information about prior medications to treat rheumatic symptoms of the patient. Despite these protocol-imposed limitations, the pre-recorded categorical information about current or prior use of NSAIDs and corticosteroids was available to the blinded examiner. The examiner recorded the time spent and information obtained regarding involved joints, symmetrical joint symptoms, morning stiffness of joints, presence of inflammatory back pain, psoriasis, RP, preceding infections, classical symptoms of gouty arthritis and family history for RA in pre-specified categories on a single-page study documentation sheet. Physical examination The physical examination performed by the blinded examiner was focused on signs of peripheral arthritis. Tender (68 joints) and swollen (66 joints) joint counts were documented on provided graphical representations. Additionally, the absence or presence of clinical signs of tenosynovitis for each hand and each foot was recorded. If deemed clinically necessary by the examiner, Menell’s test for sacroiliitis could be performed and documented for each sacroiliac joint. Musculoskeletal ultrasonography All of the ultrasonography examinations were performed on a Logiq E9 device (GE Healthcare, Buckinghamshire, UK) with an ML6-15 linear probe (6–15 MHz frequency) for larger joints and a linear hockey stick probe (8–18 MHz frequency) for joints of the fingers and toes. The technical parameters of the examination included a greyscale frequency of 15 MHz, pulse repetition frequency of 600 Hz and a colour Doppler frequency of 7.5 MHz, with highest gain level without background noise and a low wall filter. Examiners were asked to evaluate at least the seven joints included in the published US7 score [10] of the clinically most affected hand and foot: the wrist joint, the MCP joints and PIP joints of digits 2 and 3 and the MTP joints of digits 2 and 5. If deemed necessary in view of the particular symptoms of the patient, the blinded examiner could evaluate additional joints. For each joint evaluated, a semi-quantitative score (range 0–3) was used to rate the extent of synovitis detected by greyscale ultrasonography and hyperperfusion by colour Doppler ultrasonography. Additionally, the presence or absence of erosions and/or tenosynovitis of tendons adjacent to the joint were also documented on the single-page documentation sheet [10]. Laboratory tests All patients were admitted for inpatient routine clinical care to our rheumatology department and therefore all study participants received a preset core panel of laboratory tests including, among other parameters, ESR [abnormal values stratified by age (younger/older than 50 years): >15/>20 for men and >20/>30 mm/h for women], CRP (abnormal values >5.0 mg/l), RF (abnormal values >14 U/l), antibodies against anti-CCP (abnormal values >17 U/l), antibodies against anti-mutated citrullinated vimentin (abnormal values >20 U/l), ANA (abnormal titre >1:40) and ANCA. Statistical analyses Due to the exploratory nature of our study, we performed no formal sample size calculation prior to and no screening log during enrolment of the convenience sample of 100 patients. The two co-primary endpoints of our study were to determine the ability of board-certified rheumatologists to definitively diagnose or exclude (accuracy) an IRD and RA solely relying on clinical assessment without utilizing the results of musculoskeletal ultrasonography and laboratory tests. We defined accuracy in the setting of our study as the sum of the proportion of patients correctly identified as having a definite diagnosis of the disease combined with the proportion of patients correctly identified as not having the diagnosis (‘absent’ diagnosis). Secondary analyses were performed by dichotomizing the observed data [definite or probable (diagnosis) vs possible or absent (no diagnosis)] and comprehensive diagnostic test statistics (sensitivity, specificity, positive and negative likelihood ratios and predictive values and κ statistic) were reported for each of the sequential diagnostic steps of the blinded rheumatologists. Additionally, demographic factors, inclusion criteria, pretreatment and information obtained and time spent during the four steps of the blinded diagnostic workup were compared among patients receiving the final discharge diagnosis of RA, IRD other than RA (IRD not RA) and non-IRDs (non-IRD). Statistical comparisons were performed utilizing chi-square or Fischer’s exact test for categorical variables and non-parametric Mann–Whitney U test for continuous variables. Results In the period from 8 September 2014 to 15 October 2015, 100 patients on 70 separate days (1–3 patients of the on average 12 patients admitted per weekday to our institution) were recruited into the study. Demographic data, symptoms pertaining to the specified inclusion criteria, pretreatment with NSAIDs and/or corticosteroids and the final diagnosis reported in the discharge summary of the routine clinical inpatient workup are presented in Table 1. Table 1 Demographics, pretreatment and final rheumatic diagnosis obtained during the inpatient workup of the 100 patients studied Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  CPPD: calcium pyrophosphate deposition disease. Table 1 Demographics, pretreatment and final rheumatic diagnosis obtained during the inpatient workup of the 100 patients studied Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  CPPD: calcium pyrophosphate deposition disease. Of the 75 patients (75%) with a final discharge diagnosis comprising an IRD, 43 were diagnosed with RA. Of these 43 patients, 11 (25.6%), despite receiving the diagnosis, did not fulfil the ACR/EULAR classification criteria for RA from 2010. Reasons given for establishing the diagnosis despite the lack of fulfilment of classification criteria in these 11 discharge summaries were typical findings for RA on MRI of a hand (n = 6), polyarthritis of small joints not involving >10 joints in anti-CCP- and RF-negative patients (n = 4) and lack of synovitis due to prednisolone pretreatment (n = 1). The most common diagnoses among patients with an IRD other than RA were crystal arthropathy, undifferentiated and PsA and among patients with non-IRDs the most common diagnoses were OA, FM and arthralgia of unknown cause, despite the thorough inpatient workup (Table 1). Three of the authors performed the blinded stepwise evaluation of the patients and spent a median time of 5 min (range 3–8) taking the medical history, 5 min (range 2–10) on performing the physical examination and 10 min (range 5–16) min for the musculoskeletal ultrasonography. The main findings in each diagnostic step are presented stratified by the type of final diagnosis in Table 2. Findings that differed highly significantly (P < 0.01) between patients with RA and IRD other than RA were a history of PIP joints ever involved, the ratio of joints examined by greyscale sonography with synovitis and laboratory evidence of RF, anti-CCP and anti-MCV antibodies. Findings that differed highly significantly (P < 0.01) between patients with an IRD and non-IRD were the age and symptom duration at presentation, the self-reported swelling of at least one joint, finding of a swollen joint(s), tenosynovitis and a positive Menell’s test on the physical examination, the ratio of diseased among all or all of the US score evaluating 7 joints (clinically dominant hand: wrist joint, MCP and PIP joints of digits 2 and 3; clinically dominant foot: MTP joints of digits 2 and 5) examined by ultrasonography or the detection of any joints with synovitis (by greyscale sonography) or hyperperfusion (by colour Doppler sonography) and laboratory evidence of elevated ESR or CRP. Table 2 Inclusion criteria and selected documented findings stratified by the type of final discharge diagnosis obtained during routine clinical inpatient evaluation Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Values are given as n (%) unless stated otherwise. Statistical comparison by chi-square or Fisher’s exact test (categorical variables) or Mann–Whitney U test (continuous variables): for patients with RA vs patients with IRD not RA; for patients with RA vs patients with non-IRD; for patients with IRD (RA and other than RA) vs patients with non-IRD. IRD not RA: inflammatory rheumatic diseases other than RA; non-IRD: non-inflammatory rheumatic diseases; GSUS: greyscale ultrasonography; CDUS: colour Doppler ultrasonography; US7: ultrasound score evaluating seven joints (clinically dominant hand: wrist joint, MCP and PIP joints of digits 2 and 3; clinically dominant foot: MTP joints of digits 2 and 5); MCV; mutated citrullinated vimentin. Table 2 Inclusion criteria and selected documented findings stratified by the type of final discharge diagnosis obtained during routine clinical inpatient evaluation Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Values are given as n (%) unless stated otherwise. Statistical comparison by chi-square or Fisher’s exact test (categorical variables) or Mann–Whitney U test (continuous variables): for patients with RA vs patients with IRD not RA; for patients with RA vs patients with non-IRD; for patients with IRD (RA and other than RA) vs patients with non-IRD. IRD not RA: inflammatory rheumatic diseases other than RA; non-IRD: non-inflammatory rheumatic diseases; GSUS: greyscale ultrasonography; CDUS: colour Doppler ultrasonography; US7: ultrasound score evaluating seven joints (clinically dominant hand: wrist joint, MCP and PIP joints of digits 2 and 3; clinically dominant foot: MTP joints of digits 2 and 5); MCV; mutated citrullinated vimentin. Although not part of the blinded evaluation and the study protocol, the results of radiographs of the hands and feet and MRI of joints performed for routine clinical care were retrieved from charts. Hand radiographs were performed for the majority of patients (n = 90) and depicted erosions proving RA in 1 patient and were suggestive for RA in an additional 17 patients (13 of those were later diagnosed with RA). Radiographs of both feet were performed in 91 patients and depicted erosions proving RA in 1 patient and were suggestive for RA in an additional 7 patients (6 of those were later diagnosed with RA). Twelve patients with RA, 9 patients with IRD other than RA and 14 patients with non-IRD received MRI of selected joints, of whom 6, 0 and 0 showed typical signs and 4, 1 and 0 had suggestive signs of RA, respectively. The observed diagnostic certainties of the rheumatologists after each step of the blinded workup to diagnose RA or an IRD are reported in Table 3 and to delineate RA patients only among the subset of IRD patients are reported in Table 4. The ability to correctly arrive at or rule out the diagnosis of an IRD (accuracy) increased from 14% after the medical history, 27% after the physical examination and 53% after the US to 70% after taking laboratory test results into account. Of note, only 1 (4%) of the 25 patients with a final diagnosis of a non-IRD in the discharge summary was falsely classified as having an IRD in the clinical assessment of the blinded rheumatologist. Table 3 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or IRDs as established consecutively by in-patient workup in routine clinical care among 100 patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients); sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly); De: definite; Pr: probable; Po: possible; Ab: absent. Table 3 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or IRDs as established consecutively by in-patient workup in routine clinical care among 100 patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients); sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly); De: definite; Pr: probable; Po: possible; Ab: absent. Table 4 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA among the subset of 75 patients with inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  The ability of rheumatologists, blinded to all prior external diagnostic results to detect/rule out RA among the subset of 75 patients with IRDs as established consecutively by in-patient workup in routine clinical care among patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients; sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly]; De: definite; Pr: probable; Po: possible; Ab: absent. Table 4 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA among the subset of 75 patients with inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  The ability of rheumatologists, blinded to all prior external diagnostic results to detect/rule out RA among the subset of 75 patients with IRDs as established consecutively by in-patient workup in routine clinical care among patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients; sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly]; De: definite; Pr: probable; Po: possible; Ab: absent. The corresponding values for each step for correctly identifying patients with or without RA were 8, 19, 42 and 60%, respectively. Therefore the diagnostic accuracy of solely clinical assessment for determining the definite diagnosis of RA in our cohort was only 19%. Pretreatment with corticosteroids (in the prior 7 days) vs none did not alter these results substantially (15 vs 20%). Taking the results of ultrasonography and laboratory tests into account, the diagnostic accuracy for RA increased to 42 and 60%, respectively. In the secondary analyses utilizing dichotomized results, the diagnostic test statistics also displayed a marked improvement of the precision of the diagnoses after taking the ultrasonography results into account. The positive likelihood ratio to detect RA increased from step 2 (clinical assessment) to step 3 (clinical assessment and ultrasonography) from 3.1 to 4.8, but the next step—adding the laboratory results—only lead to a marginal further increase to 5.2. After each step (physical examination, ultrasonography and taking laboratory test results into account) the blinded rheumatologists documented if the ACR/EULAR classification criteria for RA had been met. Of the 43 patients receiving a final diagnosis of RA, 10 (23%), 11 (26%) and 26 (61%) fulfilled the classification criteria, respectively. Of note, some patients with the final diagnosis of an IRD other than RA [4 of 32 (13%)] and with non-IRD [6 of 25 (24%)] formally fulfilled these criteria. Discussion Although most clinicians agree that taking a thorough medical history and performing a physical examination are the key components of establishing the correct diagnosis for many diseases [1, 2], the widespread adoption of imaging and laboratory workup in the management of patients presenting with joint symptoms has led to a blurred line between a purely clinical assessment and the combined evaluation of clinical findings in the context of imaging and laboratory test results in daily routine practice. The exact contribution of just clinically assessing patients without relying on results of imaging and laboratory tests to diagnose IRD, and especially to diagnose RA, has been subject to debate among experts [1–5, 7]. To our knowledge, this is the first examiner-blinded study evaluating patients referred to a tertiary rheumatology centre for a thorough inpatient workup of joint symptoms to elucidate the contribution of each workup step separately for establishment of the diagnosis of IRD in general and RA in particular. We found that in the context of patients referred to a tertiary referral centre for the evaluation of thus far unexplained joint symptoms lasting not longer than 2 years, performance of a clinical assessment by an experienced rheumatologist relying only on a joint-focused medical history and physical exam established or ruled out the correct diagnosis in only a minority of patients. Contrary to our expectation, these results did not improve substantially in the subgroup of patients who had not received any pretreatment with corticosteroids. As expected from other research, the ratio of correct classification in regard to IRD and RA improved markedly when taking the results of ultrasonography into account, and even further when laboratory test results were included [11]. Despite a thorough literature review, we were unable to find other studies explicitly blinding examiners to external laboratory and imaging results prior to the clinical evaluation of arthralgia patients with which to compare our results. Although not formally blinded to prior diagnostic results, a retrospective Dutch study utilizing data from the Leiden University rheumatic outpatient clinic from 2012 to 2013 evaluated 1934 patients with new-onset arthralgia, of whom 145 patients were classified only by clinical evaluation to have clinical significant arthralgia and 1791 unexplained arthralgia without utilizing imaging or laboratory results [5]. In the following 12-month period 4.0 vs 0.2%, respectively, of these patients progressed to RA, classified by the 1987 ACR criteria, demonstrating the ability of rheumatologists do define high-risk patients among arthralgia patients only by clinical evaluation. Most studies evaluating the utility of ultrasonography for diagnosing IRD or RA were conducted among outpatients presenting to an early arthritis clinic (review of published studies 1982–2015 [7]). While one British study failed to demonstrate a significant improvement in the prediction of which arthralgia patients progressed to persistent inflammatory arthritis when ultrasonography was added to the clinical and laboratory workup [4], all the other 14 studies found meaningful improvement of the diagnostic accuracy utilizing ultrasonography [3, 8, 12–23]. For example, in a recent Swedish study [8], the diagnostic certainty for the diagnosis of an inflammatory arthritis increased from 33 to 72% when adding ultrasonography to the clinical and laboratory evaluation of 103 patients, demonstrating a finding similar to our study. The strength of our study was the examiner-blinded design and that only experienced, board-certified rheumatologists performed all steps of the evaluation. Additionally, the way all evaluations (symptom-focused history taking, physical examination focused on joint symptoms, ultrasonography informed by findings of the physical examination) were performed similarly to routine clinical care, thus increasing the applicability of our results. The generalizability of our findings was limited due to the recruitment of only patients sent for an inpatient workup to a single tertiary referral centre. Therefore the studied cohort might represent a more diagnostically challenging cohort and may have had a much higher prevalence of RA and IRD than one would find in the classical setting of an early arthritis outpatient clinic. Nevertheless, the median symptom duration of 4 months among the 43 patients diagnosed with RA confirmed that the studied cohort contained a majority of patients with very early RA. The final diagnosis used in this cross-sectional study was established by retrieving the discharge diagnosis of the included patients. In comparison to other studies, no longitudinal follow-up of the patients was performed to further strengthen the validity of these discharge diagnoses. Therefore some of the patients in our study incorrectly diagnosed by the blinded rheumatologist as having RA could develop more clinical evidence later and would then receive an RA diagnosis during follow-up. Taking the results of a recent publication of MRI findings of the wrists in healthy adults into account (not yet available during the evaluation of the patients in this study) [24], an overinterpretation of mild signs of synovitis and limited numbers of erosions as proof of manifest RA and therefore misclassification of a minority of the included patients is possible. Conclusion We conclude that experienced rheumatologists, in the setting of our cross-sectional study, only judging from a brief joint-symptom focused medical history and clinical examination and being deprived of information on external imaging and laboratory workup by blinding, were unable to correctly classify the majority of patients presenting with arthralgia or arthritis symptoms for inpatient workup as suffering from inflammatory arthritis or RA. Acknowledgements We wish to thank all the patients participating in this study for their support. All the authors (B.E., G.P., M.F. and W.H.) contributed to the study design, delineated suitable patients for the study and obtained informed consent from patients. Three authors (B.E., G.P. and W.H.) performed the blinded evaluations of patients. B.E. entered the data. All authors (B.E., G.P., M.F. and W.H.) analysed the data, participated in writing the manuscript and approved the final version of the manuscript. All authors meet the criteria for authorship as stipulated in the guidelines of the Recommendations for the Conduct, Reporting, Editing and Publication of Scholarly Work in Medical Journals (International Committee of Medical Journal Editors Recommendations). The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article. Disclosure statement: The authors have declared no conflicts of interest. References 1 Castrejon I, McCollum L, Tanriover MD, Pincus T. Importance of patient history and physical examination in rheumatoid arthritis compared to other chronic diseases: results of a physician survey. Arthritis Care Res  2012; 64: 1250– 5. 2 Jauhar S. The demise of the physical exam. N Engl J Med  2006; 354: 548– 51. Google Scholar CrossRef Search ADS PubMed  3 Nakagomi D, Ikeda K, Okubo A et al.   Ultrasound can improve the accuracy of the 2010 American College of Rheumatology/European League against rheumatism classification criteria for rheumatoid arthritis to predict the requirement for methotrexate treatment. Arthritis Rheum  2013; 65: 890– 8. Google Scholar CrossRef Search ADS PubMed  4 Pratt AG, Lorenzi AR, Wilson G, Platt PN, Isaacs JD. Predicting persistent inflammatory arthritis amongst early arthritis clinic patients in the UK: is musculoskeletal ultrasound required?. Arthritis Res Ther  2013; 15: R118. Google Scholar CrossRef Search ADS PubMed  5 van Steenbergen HW, van der Helm-van Mil AH. Clinical expertise and its accuracy in differentiating arthralgia patients at risk for rheumatoid arthritis from other patients presenting with joint symptoms. Rheumatology  2016; 55: 1140– 1. Google Scholar CrossRef Search ADS PubMed  6 Mandl P, Naredo E, Conaghan PG et al.   Practice of ultrasound-guided arthrocentesis and joint injection, including training and implementation, in Europe: results of a survey of experts and scientific societies. Rheumatology  2012; 51: 184– 90. Google Scholar CrossRef Search ADS PubMed  7 Lage-Hansen PR, Lindegaard H, Chrysidis S, Terslev L. The role of ultrasound in diagnosing rheumatoid arthritis, what do we know? An updated review. Rheumatol Int  2017; 37: 179– 87. Google Scholar CrossRef Search ADS PubMed  8 Rezaei H, Torp-Pedersen S, af Klint E et al.   Diagnostic utility of musculoskeletal ultrasound in patients with suspected arthritis—a probabilistic approach. Arthritis Res Ther  2014; 16: 448. Google Scholar CrossRef Search ADS PubMed  9 Aletaha D, Neogi T, Silman AJ et al.   2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis  2010; 69: 1580– 8. Google Scholar CrossRef Search ADS PubMed  10 Backhaus TM, Ohrndorf S, Kellner H et al.   The US7 score is sensitive to change in a large cohort of patients with rheumatoid arthritis over 12 months of therapy. Ann Rheum Dis  2013; 72: 1163– 9. Google Scholar CrossRef Search ADS PubMed  11 Kennish L, Labitigan M, Budoff S et al.   Utility of the new rheumatoid arthritis 2010 ACR/EULAR classification criteria in routine clinical care. BMJ Open  2012; 2: e001117. Google Scholar CrossRef Search ADS PubMed  12 Broll M, Albrecht K, Tarner I, Muller-Ladner U, Strunk J. Sensitivity and specificity of ultrasonography and low-field magnetic resonance imaging for diagnosing arthritis. Clin Exp Rheumatol  2012; 30: 543– 7. Google Scholar PubMed  13 Filer A, de Pablo P, Allen G et al.   Utility of ultrasound joint counts in the prediction of rheumatoid arthritis in patients with very early synovitis. Ann Rheum Dis  2011; 70: 500– 7. Google Scholar CrossRef Search ADS PubMed  14 Freeston JE, Wakefield RJ, Conaghan PG et al.   A diagnostic algorithm for persistence of very early inflammatory arthritis: the utility of power Doppler ultrasound when added to conventional assessment tools. Ann Rheum Dis  2010; 69: 417– 9. Google Scholar CrossRef Search ADS PubMed  15 Kawashiri SY, Suzuki T, Okada A et al.   Musculoskeletal ultrasonography assists the diagnostic performance of the 2010 classification criteria for rheumatoid arthritis. Mod Rheumatol  2013; 23: 36– 43. Google Scholar CrossRef Search ADS PubMed  16 Minowa K, Ogasawara M, Murayama G et al.   Predictive grade of ultrasound synovitis for diagnosing rheumatoid arthritis in clinical practice and the possible difference between patients with and without seropositivity. Mod Rheumatol  2016; 26: 188– 93. Google Scholar PubMed  17 Navalho M, Resende C, Rodrigues AM et al.   Bilateral evaluation of the hand and wrist in untreated early inflammatory arthritis: a comparative study of ultrasonography and magnetic resonance imaging. J Rheumatol  2013; 40: 1282– 92. Google Scholar CrossRef Search ADS PubMed  18 Ozgul A, Yasar E, Arslan N et al.   The comparison of ultrasonographic and scintigraphic findings of early arthritis in revealing rheumatoid arthritis according to criteria of American College of Rheumatology. Rheumatol Int  2009; 29: 765– 8. Google Scholar CrossRef Search ADS PubMed  19 Rakieh C, Nam JL, Hunt L et al.   Predicting the development of clinical arthritis in anti-CCP positive individuals with non-specific musculoskeletal symptoms: a prospective observational cohort study. Ann Rheum Dis  2015; 74: 1659– 66. Google Scholar CrossRef Search ADS PubMed  20 Salaffi F, Ciapetti A, Gasparini S et al.   A clinical prediction rule combining routine assessment and power Doppler ultrasonography for predicting progression to rheumatoid arthritis from early-onset undifferentiated arthritis. Clin Exp Rheumatol  2010; 28: 686– 94. Google Scholar PubMed  21 Tamas MM, Rednic N, Felea I, Rednic S. Ultrasound assessment for the rapid classification of early arthritis patients. J Investig Med  2013; 61: 1184– 91. Google Scholar CrossRef Search ADS PubMed  22 van de Stadt LA, Bos WH, Meursinge Reynders M et al.   The value of ultrasonography in predicting arthritis in auto-antibody positive arthralgia patients: a prospective cohort study. Arthritis Res Ther  2010; 12: R98. Google Scholar CrossRef Search ADS PubMed  23 Wakefield RJ, Green MJ, Marzo-Ortega H et al.   Should oligoarthritis be reclassified? Ultrasound reveals a high prevalence of subclinical disease. Ann Rheum Dis  2004; 63: 382– 5. Google Scholar CrossRef Search ADS PubMed  24 Mangnus L, van Steenbergen HW, Reijnierse M, van der Helm-van Mil AH. Magnetic resonance imaging-detected features of inflammation and erosions in symptom-free persons from the general population. Arthritis Rheumatol  2016; 68: 2593– 602. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Rheumatology Oxford University Press

The ability of rheumatologists blinded to prior workup to diagnose rheumatoid arthritis only by clinical assessment: a cross-sectional study

Loading next page...
 
/lp/ou_press/the-ability-of-rheumatologists-blinded-to-prior-workup-to-diagnose-0xwN0nfes9
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: journals.permissions@oup.com
ISSN
1462-0324
eISSN
1462-0332
D.O.I.
10.1093/rheumatology/key127
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives We aimed to study the ability of board-certified rheumatologists, blinded to all prior diagnostic test results, to establish the presence/absence of an inflammatory rheumatic disease (IRD) or RA among polyarthralgia or arthritis patients, solely relying on clinical assessment. Methods We performed a prospective, examiner-blinded, cross-sectional study documenting the diagnostic work in four sequential steps (medical history, physical examination, musculoskeletal ultrasonography and laboratory tests) of board-certified rheumatologists in a convenience cohort of 100 patients referred for inpatient diagnostic workup to a tertiary care rheumatology centre. Results The ability to correctly identify patients with or without an IRD (diagnostic accuracy) increased from 27% after the clinical assessment to 53% after the ultrasonography and to 70% after taking laboratory test results into account. The corresponding values for correctly identifying patients with or without RA were 19, 42 and 60%, respectively. Therefore the diagnostic accuracy of solely clinical assessment for determining the diagnosis of IRD or RA compared with the diagnosis established by a consecutive thorough in-patient workup was only 27 and 19% in our cohort, respectively. Pretreatment with corticosteroids (in the prior 7 days) vs none did not alter these results substantially (20 vs 29% for IRD, 15% vs 20% for RA). Conclusion Experienced rheumatologists, if deprived of information on prior external imaging and laboratory workup by blinding, were not able to correctly classify the majority of patients presenting with polyarthralgia or arthritis symptoms for inpatient workup, relying only on a brief symptom-focused medical history and physical examination. arthritis, diagnosis, medical history taking, physical examination, rheumatoid, ultrasonography Rheumatology key messages Blinded to prior workup, rheumatologists could not correctly classify most patients when relying only on medical history and examination. The addition of ultrasonography and laboratory tests increased rheumatologists’ diagnostic accuracy for RA from 27 to 70%. Pretreatment with corticosteroids vs none did not significantly worsen rheumatologists’ diagnostic performance. Introduction Most clinicians would agree that taking a thorough medical history and performing a physical examination are the key components of establishing the correct diagnosis for many diseases [1–5]. In the advent of ever more precise laboratory markers and better imaging techniques, rheumatologists sometimes wonder to what extent these diagnostic tools should augment or even have already partially replaced their clinical judgement in daily clinical practice. With the introduction of scanners that enable the depicting of soft tissues with high resolution, musculoskeletal ultrasonography has become a powerful tool to instantly image inflammatory and structural changes of joints and tendons, and often these results are available even prior to receiving the results of ordered laboratory tests. Hence some rheumatologists, especially in Germany, with >80% of respondents in an international survey published in 2012, perform ultrasonography examinations of joints as an integral part of the diagnostic workup in clinical practice [6]. This widespread adoption of imaging and laboratory workup early in the management of patients with joint symptoms has led to a blurred line between clinical and imaging examinations in daily clinical practice, where these examinations are performed simultaneously by some rheumatologists. Therefore, nowadays, the exact diagnostic yield of solely clinically assessing patients without musculoskeletal ultrasonography to definitively diagnose inflammatory rheumatic diseases (IRDs), and especially RA, has been debated among experts [4, 7]. While some emphasize that visualizing the amount of synovitis in joints and tendon sheets of clinically symptomatic and asymptomatic joints can greatly enhance the ability to diagnose RA [7, 8], others point out that in most of the examined patients the diagnosis could have been reached without performing musculoskeletal ultrasonography [4]. As specialists, rheumatologists are rarely the first physicians patients turn to with new symptoms indicating a possible IRD. Therefore most patients seen in daily clinical practice provide the rheumatologist with the results of externally obtained tests and imaging. Thus the diagnostic accuracy of a solely clinical assessment (only obtaining a medical history and performing a physical examination) to diagnose a patient with RA cannot be studied reliably by just observing rheumatologists in daily clinical practice, because of bias introduced to clinical judgement by preeminent external laboratory test and imaging results. We aimed to prospectively study the contribution of solely clinical assessment (taking a medical history and performing a physical examination focused on joint symptoms) compared with a diagnostic workup including ultrasonography of selected joints and results of laboratory tests for establishing the diagnosis of an IRD and the diagnosis and classification of RA in patients with thus far unexplained joint symptoms. Methods Study design We performed a prospective, examiner-blinded study documenting the diagnostic work of board-certified rheumatologists with long-standing experience in musculoskeletal ultrasonography in a convenience cohort of 100 patients referred for diagnostic workup to a tertiary care rheumatology centre. Patients were recruited into the study at the time of their hospital admission if they had not yet received a rheumatologist-established diagnosis of an IRD, had not been treated with conventional or biologic DMARDs and reported a history of or present arthralgia of at least five peripheral joints and/or reported a history of or present swelling of at least one joint with an overall symptom duration of no longer than 2 years. Prior treatment with NSAIDs and corticosteroids was permitted and the extent of such pretreatment was categorically recorded. All patients provided written consent after receiving oral and written information about the study by one of the four authors. The study was reviewed and approved by the institutional review board/ethics committee of the University Medical Center of Regensburg (07/125). The study was conducted in compliance with the Declaration of Helsinki, International Conference on Harmonization Good Clinical Practice guidelines and local country regulations. Participants were instructed not to provide any unsolicited information pertaining to their prior external diagnostic workup, other pre-existing rheumatic diseases or prior drug treatment to the second rheumatologist conducting the diagnostic evaluation. The second rheumatologist (one of the authors: B.E., G.P. or W.H., all board-certified rheumatologists), with a mean of 16 years of clinical experience and a mean of 10 years of experience in musculoskeletal ultrasonography, was always different from the rheumatologist that had obtained the informed consent and was therefore blinded to all prior externally conducted diagnostic workups and performed and documented the following four sequential steps: (1) obtaining a medical history focused on the joint-related symptoms, (2) performing a physical examination focused on joint-related signs, (3) conducting musculoskeletal ultrasonography of selected peripheral joints and (4) arriving at a final judgement, including results of laboratory tests. After each of the four steps the blinded rheumatologist reported the time required for the step and the degree of certainty for establishing the diagnosis of an IRD and of RA using four pre-defined categories (definite, probable, possible, absent) and the presence of probable (>10% possibility) differential diagnoses. Furthermore, for the last three of the four evaluation steps, fulfilment of the ACR/EULAR RA classification criteria from 2010 was recorded [9]. After the evaluation by the blinded examiner, patients received their scheduled thorough workup as part of the standard in-patient clinical care in our institution. We defined the rheumatic diagnosis reported on the discharge summary report as the final diagnosis for each patient. After having completed and documented their four-step research evaluation, the three blinded rheumatologists (B.E., G.P. and W.H.) contributed later to the routine clinical care of some of the patients, but to prevent bias, all discharge diagnoses were adjudicated by the fourth author (M.F.) as part of his responsibilities as head of the department during routine clinical care. Medical history The medical history obtained by the blinded rheumatologist focused on symptoms pertaining to rheumatic inflammatory diseases, especially typical symptoms of RA. The examiners were instructed to not inquire about externally performed diagnostic tests, suspected or given diagnoses of rheumatic diseases and specific information about prior medications to treat rheumatic symptoms of the patient. Despite these protocol-imposed limitations, the pre-recorded categorical information about current or prior use of NSAIDs and corticosteroids was available to the blinded examiner. The examiner recorded the time spent and information obtained regarding involved joints, symmetrical joint symptoms, morning stiffness of joints, presence of inflammatory back pain, psoriasis, RP, preceding infections, classical symptoms of gouty arthritis and family history for RA in pre-specified categories on a single-page study documentation sheet. Physical examination The physical examination performed by the blinded examiner was focused on signs of peripheral arthritis. Tender (68 joints) and swollen (66 joints) joint counts were documented on provided graphical representations. Additionally, the absence or presence of clinical signs of tenosynovitis for each hand and each foot was recorded. If deemed clinically necessary by the examiner, Menell’s test for sacroiliitis could be performed and documented for each sacroiliac joint. Musculoskeletal ultrasonography All of the ultrasonography examinations were performed on a Logiq E9 device (GE Healthcare, Buckinghamshire, UK) with an ML6-15 linear probe (6–15 MHz frequency) for larger joints and a linear hockey stick probe (8–18 MHz frequency) for joints of the fingers and toes. The technical parameters of the examination included a greyscale frequency of 15 MHz, pulse repetition frequency of 600 Hz and a colour Doppler frequency of 7.5 MHz, with highest gain level without background noise and a low wall filter. Examiners were asked to evaluate at least the seven joints included in the published US7 score [10] of the clinically most affected hand and foot: the wrist joint, the MCP joints and PIP joints of digits 2 and 3 and the MTP joints of digits 2 and 5. If deemed necessary in view of the particular symptoms of the patient, the blinded examiner could evaluate additional joints. For each joint evaluated, a semi-quantitative score (range 0–3) was used to rate the extent of synovitis detected by greyscale ultrasonography and hyperperfusion by colour Doppler ultrasonography. Additionally, the presence or absence of erosions and/or tenosynovitis of tendons adjacent to the joint were also documented on the single-page documentation sheet [10]. Laboratory tests All patients were admitted for inpatient routine clinical care to our rheumatology department and therefore all study participants received a preset core panel of laboratory tests including, among other parameters, ESR [abnormal values stratified by age (younger/older than 50 years): >15/>20 for men and >20/>30 mm/h for women], CRP (abnormal values >5.0 mg/l), RF (abnormal values >14 U/l), antibodies against anti-CCP (abnormal values >17 U/l), antibodies against anti-mutated citrullinated vimentin (abnormal values >20 U/l), ANA (abnormal titre >1:40) and ANCA. Statistical analyses Due to the exploratory nature of our study, we performed no formal sample size calculation prior to and no screening log during enrolment of the convenience sample of 100 patients. The two co-primary endpoints of our study were to determine the ability of board-certified rheumatologists to definitively diagnose or exclude (accuracy) an IRD and RA solely relying on clinical assessment without utilizing the results of musculoskeletal ultrasonography and laboratory tests. We defined accuracy in the setting of our study as the sum of the proportion of patients correctly identified as having a definite diagnosis of the disease combined with the proportion of patients correctly identified as not having the diagnosis (‘absent’ diagnosis). Secondary analyses were performed by dichotomizing the observed data [definite or probable (diagnosis) vs possible or absent (no diagnosis)] and comprehensive diagnostic test statistics (sensitivity, specificity, positive and negative likelihood ratios and predictive values and κ statistic) were reported for each of the sequential diagnostic steps of the blinded rheumatologists. Additionally, demographic factors, inclusion criteria, pretreatment and information obtained and time spent during the four steps of the blinded diagnostic workup were compared among patients receiving the final discharge diagnosis of RA, IRD other than RA (IRD not RA) and non-IRDs (non-IRD). Statistical comparisons were performed utilizing chi-square or Fischer’s exact test for categorical variables and non-parametric Mann–Whitney U test for continuous variables. Results In the period from 8 September 2014 to 15 October 2015, 100 patients on 70 separate days (1–3 patients of the on average 12 patients admitted per weekday to our institution) were recruited into the study. Demographic data, symptoms pertaining to the specified inclusion criteria, pretreatment with NSAIDs and/or corticosteroids and the final diagnosis reported in the discharge summary of the routine clinical inpatient workup are presented in Table 1. Table 1 Demographics, pretreatment and final rheumatic diagnosis obtained during the inpatient workup of the 100 patients studied Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  CPPD: calcium pyrophosphate deposition disease. Table 1 Demographics, pretreatment and final rheumatic diagnosis obtained during the inpatient workup of the 100 patients studied Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  Demographics  Values  Sex, female, n  48  Age, median (range), years  55.5 (18–84)  BMI, median (range), kg/m2  28.1 (18.0–43.5)  Symptom duration, median (range), months  4.0 (0.5–24.0)  Referral by, n        Primary care physician  80      Rheumatologist  8      Other specialist  12  Pretreatment, n        Any NSAID in previous 7 days  64      Any corticosteroid in previous 7 days  20      Any corticosteroid in previous 3 month  53  Final rheumatic diagnosis, n        RA  43          ACR/EULAR classification criteria for RA fulfilled  32      IRD not RA  32          Gouty arthritis  8          Undifferentiated arthritis  6          PsA  4          CPPD  3          SpA  3          CTDs  3          PMR  2          ReA  1          Whipple’s disease  1          Lyme arthritis  1      Non-IRD, n  25          OA  10          FM  3          Post-traumatic stress disorder  1          Arthralgias of unknown cause despite thorough workup  11  CPPD: calcium pyrophosphate deposition disease. Of the 75 patients (75%) with a final discharge diagnosis comprising an IRD, 43 were diagnosed with RA. Of these 43 patients, 11 (25.6%), despite receiving the diagnosis, did not fulfil the ACR/EULAR classification criteria for RA from 2010. Reasons given for establishing the diagnosis despite the lack of fulfilment of classification criteria in these 11 discharge summaries were typical findings for RA on MRI of a hand (n = 6), polyarthritis of small joints not involving >10 joints in anti-CCP- and RF-negative patients (n = 4) and lack of synovitis due to prednisolone pretreatment (n = 1). The most common diagnoses among patients with an IRD other than RA were crystal arthropathy, undifferentiated and PsA and among patients with non-IRDs the most common diagnoses were OA, FM and arthralgia of unknown cause, despite the thorough inpatient workup (Table 1). Three of the authors performed the blinded stepwise evaluation of the patients and spent a median time of 5 min (range 3–8) taking the medical history, 5 min (range 2–10) on performing the physical examination and 10 min (range 5–16) min for the musculoskeletal ultrasonography. The main findings in each diagnostic step are presented stratified by the type of final diagnosis in Table 2. Findings that differed highly significantly (P < 0.01) between patients with RA and IRD other than RA were a history of PIP joints ever involved, the ratio of joints examined by greyscale sonography with synovitis and laboratory evidence of RF, anti-CCP and anti-MCV antibodies. Findings that differed highly significantly (P < 0.01) between patients with an IRD and non-IRD were the age and symptom duration at presentation, the self-reported swelling of at least one joint, finding of a swollen joint(s), tenosynovitis and a positive Menell’s test on the physical examination, the ratio of diseased among all or all of the US score evaluating 7 joints (clinically dominant hand: wrist joint, MCP and PIP joints of digits 2 and 3; clinically dominant foot: MTP joints of digits 2 and 5) examined by ultrasonography or the detection of any joints with synovitis (by greyscale sonography) or hyperperfusion (by colour Doppler sonography) and laboratory evidence of elevated ESR or CRP. Table 2 Inclusion criteria and selected documented findings stratified by the type of final discharge diagnosis obtained during routine clinical inpatient evaluation Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Values are given as n (%) unless stated otherwise. Statistical comparison by chi-square or Fisher’s exact test (categorical variables) or Mann–Whitney U test (continuous variables): for patients with RA vs patients with IRD not RA; for patients with RA vs patients with non-IRD; for patients with IRD (RA and other than RA) vs patients with non-IRD. IRD not RA: inflammatory rheumatic diseases other than RA; non-IRD: non-inflammatory rheumatic diseases; GSUS: greyscale ultrasonography; CDUS: colour Doppler ultrasonography; US7: ultrasound score evaluating seven joints (clinically dominant hand: wrist joint, MCP and PIP joints of digits 2 and 3; clinically dominant foot: MTP joints of digits 2 and 5); MCV; mutated citrullinated vimentin. Table 2 Inclusion criteria and selected documented findings stratified by the type of final discharge diagnosis obtained during routine clinical inpatient evaluation Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Criteria  Final diagnosis obtained in routine clinical care   Statistical comparisons   RA n = 43  IRD, not RA n = 32  Non-IRD n = 25  RA vs IRD, no RA  RA vs non-IRD  IRD vs non-IRD  Demographics/inclusion criteria/pretreatment      Sex, female  22 (51.2)  9 (28.1)  17 (68.0)  0.045  0.176  0.021      Age, median (range), years  60 (32–78)  50.5 (21–84)  49 (18–79)  0.041  0.000  0.005      Symptom duration, median (range), months  4.0 (1.0–12.0)  3.0 (0.7–10.0)  8.0 (0.5–24.0)  0.100  0.008  0.000      Patient-reported arthralgias of at least five joints  35 (81.4)  18 (56.3)  20 (80.0)  0.018  1.000  0.363      Patient-reported swelling of at least one joint  40 (93.0)  29 (90.6)  14 (56.0)  1.000  0.000  0.000      Any NSAID in previous 7 days  34 (79.1)  18 (56.3)  12 (48.0)  0.034  0.008  0.054      Any corticosteroid in previous 7 days  10 (23.3)  9 (28.1)  1 (4.0)  0.632  0.045  0.021      Any corticosteroid in previous 3 months  23 (53.5)  19 (59.4)  11 (44.0)  0.611  0.451  0.298  Medical history      Wrist joints ever involved  33 (76.7)  16 (50.0)  15 (60.0)  0.016  0.144  0.630      MCP joints ever involved  28 (65.1)  15 (46.9)  15 (60.0)  0.114  0.673  0.815      PIP finger joints ever involved  31 (72.1)  12 (37.5)  17 (68.0)  0.003  0.721  0.346      MTP joints ever involved  19 (44.2)  11 (34.4)  5 (20.0)  0.391  0.044  0.069      Symmetric joint symptoms  29 (67.4)  15 (46.9)  16 (64.0)  0.143  0.386  0.523      Maximal duration of morning stiffness, median (range), h  1.0 (0.0–24.0)  0.9 (0.0–24.0)  0.5 (0.0–14.0)  0.780  0.111  0.208      Family history of RA  4 (9.3)  2 (6.3)  4 (16.0)  0.834  0.709  0.512  Physical examination      Gaenslen’s sign positive, hands  16 (37.2)  8 (25.0)  13 (52.0)  0.262  0.234  0.073      Gaenslen’s sign positive, feet  10 (23.3)  5 (15.6)  7 (28.0)  0.414  0.663  0.403      Tender joint count (68 joints), median (range)  6 (0–36)  3 (0–16)  6 (0–26)  0.010  0.964  0.431      Swollen joint count (66 joints), median (range)  1 (0–29)  0.5 (0–12)  0 (0–6)  0.045  0.000  0.001      Clinical signs of tenosynovitis  13 (30.2)  7 (21.9)  0 (0.0)  0.418  0.001  0.004      Menell’s test [n positive (% of examined)/n examined]  0 (0%)/11  0 (0%)/15  7 (58%)/12  1.000  0.000  0.000  Musculoskeletal ultrasonography      Number of joints evaluated, median (range)  10 (7–14)  11 (7–14)  9 (7–12)  0.394  0.204  0.089      Ratio of joints with synovitis in GSUS, median (range)  0.45 (0.0–1.0)  0.29 (0.0–0.6)  0.11 (0.0–0.5)  0.003  0.000  0.000      Ratio of joints with hyperperfusion in CDUS, median (range)  0.11 (0.0–0.8)  0.12 (0.0–0.4)  0.00 (0.0–0.2)  0.922  0.000  0.000      Detection of any synovitis by GSUS  42 (97.7)  28 (87.5)  16 (64.0)  0.081  0.000  0.001      Detection of any hyperperfusion by CDUS  32 (74.4)  21 (65.6)  4 (16.0)  0.408  0.000  0.000      Detection of any erosions  6 (14.0)  6 (18.8)  0 (0.0)  0.575  0.078  0.034      Detection of any tenosynovitis  21 (48.8)  9 (28.1)  3 (12.0)  0.070  0.002  0.010      US7 joints with synovitis by GSUS, median (range)  3 (0–7)  1 (0–5)  1 (0–2)  0.000  0.000  0.000      US7 joints with hyperperfusion by CDUS, median (range)  1 (0–6)  0 (0–3)  0 (0–1)  0.038  0.000  0.001  Laboratory test results      Elevated ESR  21 (48.8)  19 (59.4)  1 (4.0)  0.366  0.000  0.000      Elevated CRP  29 (67.4)  21 (65.6)  7 (28.0)  0.869  0.002  0.001      Positive RF  24 (45.8)  2 (6.2)  3 (12.0)  0.000  0.001  0.031      Positive anti-CCP antibody test  23 (53.5)  0 (0.0)  2 (8.0)  0.000  0.001  0.023      Positive anti-MCV antibody test  21 (48.8)  3 (9.4)  3 (12.0)  0.000  0.002  0.051  Values are given as n (%) unless stated otherwise. Statistical comparison by chi-square or Fisher’s exact test (categorical variables) or Mann–Whitney U test (continuous variables): for patients with RA vs patients with IRD not RA; for patients with RA vs patients with non-IRD; for patients with IRD (RA and other than RA) vs patients with non-IRD. IRD not RA: inflammatory rheumatic diseases other than RA; non-IRD: non-inflammatory rheumatic diseases; GSUS: greyscale ultrasonography; CDUS: colour Doppler ultrasonography; US7: ultrasound score evaluating seven joints (clinically dominant hand: wrist joint, MCP and PIP joints of digits 2 and 3; clinically dominant foot: MTP joints of digits 2 and 5); MCV; mutated citrullinated vimentin. Although not part of the blinded evaluation and the study protocol, the results of radiographs of the hands and feet and MRI of joints performed for routine clinical care were retrieved from charts. Hand radiographs were performed for the majority of patients (n = 90) and depicted erosions proving RA in 1 patient and were suggestive for RA in an additional 17 patients (13 of those were later diagnosed with RA). Radiographs of both feet were performed in 91 patients and depicted erosions proving RA in 1 patient and were suggestive for RA in an additional 7 patients (6 of those were later diagnosed with RA). Twelve patients with RA, 9 patients with IRD other than RA and 14 patients with non-IRD received MRI of selected joints, of whom 6, 0 and 0 showed typical signs and 4, 1 and 0 had suggestive signs of RA, respectively. The observed diagnostic certainties of the rheumatologists after each step of the blinded workup to diagnose RA or an IRD are reported in Table 3 and to delineate RA patients only among the subset of IRD patients are reported in Table 4. The ability to correctly arrive at or rule out the diagnosis of an IRD (accuracy) increased from 14% after the medical history, 27% after the physical examination and 53% after the US to 70% after taking laboratory test results into account. Of note, only 1 (4%) of the 25 patients with a final diagnosis of a non-IRD in the discharge summary was falsely classified as having an IRD in the clinical assessment of the blinded rheumatologist. Table 3 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or IRDs as established consecutively by in-patient workup in routine clinical care among 100 patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients); sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly); De: definite; Pr: probable; Po: possible; Ab: absent. Table 3 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  Gold-standard  Observed data   Dichotomized analyses, definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists   Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA vs. no RA (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  8.0 (3.5, 15.2)  55.8 (39.9, 70.9)  82.5 (70.1, 91.3)  3.2 (1.7, 5.9)  0.5 (0.4, 0.8)  70.6 (56.3, 81.7)  71.2 (63.4, 77.9)  39.3 (20.7, 57.9)  No RA (57)  0  10  39  8  RA (43)  2: MH and PE  5  18  17  3  19.0 (11.8, 28.1)  53.5 (37.7, 68.8)  82.5 (70.1, 91.3)  3.1 (1.6, 5.7)  0.6 (0.4, 0.8)  69.7 (55.1, 81.2)  70.2 (62.5, 76.8)  37.0 (18.1, 55.9)  No RA (57)  0  10  33  14  RA (43)  3: MH and PE and US  12  17  8  6  42.0 (32.2, 52.3)  67.4 (51.5, 80.9)  86.0 (74.2, 93.7)  4.8 (2.5, 9.4)  0.4 (0.2, 0.6)  78.4 (64.9, 87.7)  77.8 (69.2, 84.5)  54.3 (37.5, 71.2)  No RA (57)  3  5  19  30  RA (43)  4: MH and PE and US and LT  25  10  6  2  60.0 (49.7, 69.7)  81.4 (66.6, 91.6)  84.2 (72.1, 92.5)  5.2 (2.8, 9.6)  0.2 (0.1, 0.4)  79.6 (67.7, 87.1)  85.7 (76.1, 91.9)  65.4 (50.4, 80.4)  No RA (57)  0  9  13  35    IRD vs. no IRD (n)  Diagnostic steps  Diagnosis of IRD (n)                  De  Pr  Po  Ab  IRD (75)  1: MH  14  41  19  1  14.0 (7.9, 22.4)  73.3 (61.9, 82.9)  60.0 (38.7, 78.9)  1.8 (1.1, 3.0)  0.4 (0.3, 0.7)  84.6 (77.0, 90.1)  42.8 (31.4, 55.1)  29.4 (8.3, 50.5)  No IRD (25)  0  10  15  0  IRD (75)  2: MH and PE  25  29  17  4  27.0 (18.6, 36.8)  72.0 (60.4, 81.8)  68.0 (46.5, 85.1)  2.3 (1.3, 4.1)  0.4 (0.3, 0.7)  87.1 (78.9, 92.4)  44.7 (34.0, 56.0)  34.1 (13.9, 54.3)  No IRD (25)  1  7  15  2  IRD (75)  3: MH and PE and US  47  15  13  0  53.0 (42.8, 63.1)  82.7 (72.2, 90.4)  88.0 (68.8¸ 97.5)  6.9 (2.4, 20.0)  0.2 (0.1, 0.3)  95.4 (87.7, 98.4)  62.9 (50.3, 73.9)  63.4 (45.5, 79.3)  No IRD (25)  1  2  16  6  IRD (75)  4: MH and PE and US and LT  60  7  6  2  70.0 (60.0, 78.8)  89.3 (80.1, 95.3)  84.0 (63.9, 95.5)  5.6 (2.3¸13.8)  0.1 (0.1, 0.3)  94.4 (87.2, 97.6)  72.4 (57.2, 83.8)  69.6 (53.5, 85.7)  No IRD (25)  0  4  11  10  The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA or IRDs as established consecutively by in-patient workup in routine clinical care among 100 patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients); sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly); De: definite; Pr: probable; Po: possible; Ab: absent. Table 4 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA among the subset of 75 patients with inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  The ability of rheumatologists, blinded to all prior external diagnostic results to detect/rule out RA among the subset of 75 patients with IRDs as established consecutively by in-patient workup in routine clinical care among patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients; sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly]; De: definite; Pr: probable; Po: possible; Ab: absent. Table 4 The ability of rheumatologists, blinded to all prior external diagnostic results, to detect/rule out RA among the subset of 75 patients with inflammatory rheumatic diseases Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  Gold-standard  Observed data   Dichotomized analyses definite or probable (diagnosis) vs possible or absent (no diagnosis)   Final diagnosis at discharge  Evaluation by blinded rheumatologists  Accuracy % (95% CI)  Sensitivity % (95% CI)  Specificity % (95% CI)  +LR (95% CI)  −LR (95% CI)  PPV % (95% CI)  NPV % (95% CI)  κ % (95% CI)  RA, only among IRD patients (n)  Diagnostic steps  Diagnosis of RA (n)                  De  Pr  Po  Ab  RA (43)  1: MH  0  24  16  3  5.3 (1.5, 13.1)  55.8 (39.9, 70.9)  78.1 (60.0, 90.7)  2.6 (1.3, 5.2)  0.6 (0.4, 0.8)  77.4 (62.8, 87.4)  56.8 (47.3, 65.9)  32.4 (11.4, 53.4)  IRD not RA (32)  0  7  21  4  RA (43)  2: MH and PE  5  18  17  3  14.7 (7.6, 24.7)  53.5 (37.7, 68.8)  81.3 (63.6, 92.8)  2.9 (1.3, 6.2)  0.6 (0.4, 0.8)  79.3 (63.9, 89.3)  56.5 (47.5, 65.1)  32.9 (12.1, 53.8)  IRD not RA (32)  0  6  20  6  RA (43)  3: MH and PE and US  12  17  8  6  34.7 (24.0, 46.5)  67.4 (51.5, 80.9)  75.0 (56.6, 88.5)  2.7 (1.4, 5.1)  0.4 (0.3, 0.7)  78.4 (65.8, 87.3)  63.2 (51.6, 73.4)  41.5 (20.9, 62.0)  IRD not RA (32)  3  5  10  14  RA (43)  4: MH and PE and US and LT  25  10  6  2  58.7 (46.7, 69.9)  81.4 (66.6, 91.6)  81.3 (63.6, 92.8)  4.3 (2.1, 9.1)  0.2 (0.1, 0.4)  85.4 (73.7, 92.4)  76.5 (63.0, 86.1)  62.2 (44.3, 80.0)  IRD not RA (32)  0  6  7  19  The ability of rheumatologists, blinded to all prior external diagnostic results to detect/rule out RA among the subset of 75 patients with IRDs as established consecutively by in-patient workup in routine clinical care among patients admitted for arthralgia and/or arthritis symptoms. The observed data are provided stratified by the four diagnostic steps performed [medical history (MH), physical examination (PE), musculoskeletal ultrasonography (US) and laboratory tests (LT)]. Accuracy: proportion of patients with either correct definite diagnosis (n, formatted bold) or correct absent diagnosis (n, formatted bold italic) among all patients; sensitivity: true positive rate; specificity: true negative rate; +LR: positive likelihood ratio; −LR: negative likelihood ratio; PPV: positive predictive value; NPV: negative predictive value; κ: kappa statistic [proportion (0–100%) above chance that blinded rheumatologists classified patients correctly]; De: definite; Pr: probable; Po: possible; Ab: absent. The corresponding values for each step for correctly identifying patients with or without RA were 8, 19, 42 and 60%, respectively. Therefore the diagnostic accuracy of solely clinical assessment for determining the definite diagnosis of RA in our cohort was only 19%. Pretreatment with corticosteroids (in the prior 7 days) vs none did not alter these results substantially (15 vs 20%). Taking the results of ultrasonography and laboratory tests into account, the diagnostic accuracy for RA increased to 42 and 60%, respectively. In the secondary analyses utilizing dichotomized results, the diagnostic test statistics also displayed a marked improvement of the precision of the diagnoses after taking the ultrasonography results into account. The positive likelihood ratio to detect RA increased from step 2 (clinical assessment) to step 3 (clinical assessment and ultrasonography) from 3.1 to 4.8, but the next step—adding the laboratory results—only lead to a marginal further increase to 5.2. After each step (physical examination, ultrasonography and taking laboratory test results into account) the blinded rheumatologists documented if the ACR/EULAR classification criteria for RA had been met. Of the 43 patients receiving a final diagnosis of RA, 10 (23%), 11 (26%) and 26 (61%) fulfilled the classification criteria, respectively. Of note, some patients with the final diagnosis of an IRD other than RA [4 of 32 (13%)] and with non-IRD [6 of 25 (24%)] formally fulfilled these criteria. Discussion Although most clinicians agree that taking a thorough medical history and performing a physical examination are the key components of establishing the correct diagnosis for many diseases [1, 2], the widespread adoption of imaging and laboratory workup in the management of patients presenting with joint symptoms has led to a blurred line between a purely clinical assessment and the combined evaluation of clinical findings in the context of imaging and laboratory test results in daily routine practice. The exact contribution of just clinically assessing patients without relying on results of imaging and laboratory tests to diagnose IRD, and especially to diagnose RA, has been subject to debate among experts [1–5, 7]. To our knowledge, this is the first examiner-blinded study evaluating patients referred to a tertiary rheumatology centre for a thorough inpatient workup of joint symptoms to elucidate the contribution of each workup step separately for establishment of the diagnosis of IRD in general and RA in particular. We found that in the context of patients referred to a tertiary referral centre for the evaluation of thus far unexplained joint symptoms lasting not longer than 2 years, performance of a clinical assessment by an experienced rheumatologist relying only on a joint-focused medical history and physical exam established or ruled out the correct diagnosis in only a minority of patients. Contrary to our expectation, these results did not improve substantially in the subgroup of patients who had not received any pretreatment with corticosteroids. As expected from other research, the ratio of correct classification in regard to IRD and RA improved markedly when taking the results of ultrasonography into account, and even further when laboratory test results were included [11]. Despite a thorough literature review, we were unable to find other studies explicitly blinding examiners to external laboratory and imaging results prior to the clinical evaluation of arthralgia patients with which to compare our results. Although not formally blinded to prior diagnostic results, a retrospective Dutch study utilizing data from the Leiden University rheumatic outpatient clinic from 2012 to 2013 evaluated 1934 patients with new-onset arthralgia, of whom 145 patients were classified only by clinical evaluation to have clinical significant arthralgia and 1791 unexplained arthralgia without utilizing imaging or laboratory results [5]. In the following 12-month period 4.0 vs 0.2%, respectively, of these patients progressed to RA, classified by the 1987 ACR criteria, demonstrating the ability of rheumatologists do define high-risk patients among arthralgia patients only by clinical evaluation. Most studies evaluating the utility of ultrasonography for diagnosing IRD or RA were conducted among outpatients presenting to an early arthritis clinic (review of published studies 1982–2015 [7]). While one British study failed to demonstrate a significant improvement in the prediction of which arthralgia patients progressed to persistent inflammatory arthritis when ultrasonography was added to the clinical and laboratory workup [4], all the other 14 studies found meaningful improvement of the diagnostic accuracy utilizing ultrasonography [3, 8, 12–23]. For example, in a recent Swedish study [8], the diagnostic certainty for the diagnosis of an inflammatory arthritis increased from 33 to 72% when adding ultrasonography to the clinical and laboratory evaluation of 103 patients, demonstrating a finding similar to our study. The strength of our study was the examiner-blinded design and that only experienced, board-certified rheumatologists performed all steps of the evaluation. Additionally, the way all evaluations (symptom-focused history taking, physical examination focused on joint symptoms, ultrasonography informed by findings of the physical examination) were performed similarly to routine clinical care, thus increasing the applicability of our results. The generalizability of our findings was limited due to the recruitment of only patients sent for an inpatient workup to a single tertiary referral centre. Therefore the studied cohort might represent a more diagnostically challenging cohort and may have had a much higher prevalence of RA and IRD than one would find in the classical setting of an early arthritis outpatient clinic. Nevertheless, the median symptom duration of 4 months among the 43 patients diagnosed with RA confirmed that the studied cohort contained a majority of patients with very early RA. The final diagnosis used in this cross-sectional study was established by retrieving the discharge diagnosis of the included patients. In comparison to other studies, no longitudinal follow-up of the patients was performed to further strengthen the validity of these discharge diagnoses. Therefore some of the patients in our study incorrectly diagnosed by the blinded rheumatologist as having RA could develop more clinical evidence later and would then receive an RA diagnosis during follow-up. Taking the results of a recent publication of MRI findings of the wrists in healthy adults into account (not yet available during the evaluation of the patients in this study) [24], an overinterpretation of mild signs of synovitis and limited numbers of erosions as proof of manifest RA and therefore misclassification of a minority of the included patients is possible. Conclusion We conclude that experienced rheumatologists, in the setting of our cross-sectional study, only judging from a brief joint-symptom focused medical history and clinical examination and being deprived of information on external imaging and laboratory workup by blinding, were unable to correctly classify the majority of patients presenting with arthralgia or arthritis symptoms for inpatient workup as suffering from inflammatory arthritis or RA. Acknowledgements We wish to thank all the patients participating in this study for their support. All the authors (B.E., G.P., M.F. and W.H.) contributed to the study design, delineated suitable patients for the study and obtained informed consent from patients. Three authors (B.E., G.P. and W.H.) performed the blinded evaluations of patients. B.E. entered the data. All authors (B.E., G.P., M.F. and W.H.) analysed the data, participated in writing the manuscript and approved the final version of the manuscript. All authors meet the criteria for authorship as stipulated in the guidelines of the Recommendations for the Conduct, Reporting, Editing and Publication of Scholarly Work in Medical Journals (International Committee of Medical Journal Editors Recommendations). The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article. Disclosure statement: The authors have declared no conflicts of interest. References 1 Castrejon I, McCollum L, Tanriover MD, Pincus T. Importance of patient history and physical examination in rheumatoid arthritis compared to other chronic diseases: results of a physician survey. Arthritis Care Res  2012; 64: 1250– 5. 2 Jauhar S. The demise of the physical exam. N Engl J Med  2006; 354: 548– 51. Google Scholar CrossRef Search ADS PubMed  3 Nakagomi D, Ikeda K, Okubo A et al.   Ultrasound can improve the accuracy of the 2010 American College of Rheumatology/European League against rheumatism classification criteria for rheumatoid arthritis to predict the requirement for methotrexate treatment. Arthritis Rheum  2013; 65: 890– 8. Google Scholar CrossRef Search ADS PubMed  4 Pratt AG, Lorenzi AR, Wilson G, Platt PN, Isaacs JD. Predicting persistent inflammatory arthritis amongst early arthritis clinic patients in the UK: is musculoskeletal ultrasound required?. Arthritis Res Ther  2013; 15: R118. Google Scholar CrossRef Search ADS PubMed  5 van Steenbergen HW, van der Helm-van Mil AH. Clinical expertise and its accuracy in differentiating arthralgia patients at risk for rheumatoid arthritis from other patients presenting with joint symptoms. Rheumatology  2016; 55: 1140– 1. Google Scholar CrossRef Search ADS PubMed  6 Mandl P, Naredo E, Conaghan PG et al.   Practice of ultrasound-guided arthrocentesis and joint injection, including training and implementation, in Europe: results of a survey of experts and scientific societies. Rheumatology  2012; 51: 184– 90. Google Scholar CrossRef Search ADS PubMed  7 Lage-Hansen PR, Lindegaard H, Chrysidis S, Terslev L. The role of ultrasound in diagnosing rheumatoid arthritis, what do we know? An updated review. Rheumatol Int  2017; 37: 179– 87. Google Scholar CrossRef Search ADS PubMed  8 Rezaei H, Torp-Pedersen S, af Klint E et al.   Diagnostic utility of musculoskeletal ultrasound in patients with suspected arthritis—a probabilistic approach. Arthritis Res Ther  2014; 16: 448. Google Scholar CrossRef Search ADS PubMed  9 Aletaha D, Neogi T, Silman AJ et al.   2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis  2010; 69: 1580– 8. Google Scholar CrossRef Search ADS PubMed  10 Backhaus TM, Ohrndorf S, Kellner H et al.   The US7 score is sensitive to change in a large cohort of patients with rheumatoid arthritis over 12 months of therapy. Ann Rheum Dis  2013; 72: 1163– 9. Google Scholar CrossRef Search ADS PubMed  11 Kennish L, Labitigan M, Budoff S et al.   Utility of the new rheumatoid arthritis 2010 ACR/EULAR classification criteria in routine clinical care. BMJ Open  2012; 2: e001117. Google Scholar CrossRef Search ADS PubMed  12 Broll M, Albrecht K, Tarner I, Muller-Ladner U, Strunk J. Sensitivity and specificity of ultrasonography and low-field magnetic resonance imaging for diagnosing arthritis. Clin Exp Rheumatol  2012; 30: 543– 7. Google Scholar PubMed  13 Filer A, de Pablo P, Allen G et al.   Utility of ultrasound joint counts in the prediction of rheumatoid arthritis in patients with very early synovitis. Ann Rheum Dis  2011; 70: 500– 7. Google Scholar CrossRef Search ADS PubMed  14 Freeston JE, Wakefield RJ, Conaghan PG et al.   A diagnostic algorithm for persistence of very early inflammatory arthritis: the utility of power Doppler ultrasound when added to conventional assessment tools. Ann Rheum Dis  2010; 69: 417– 9. Google Scholar CrossRef Search ADS PubMed  15 Kawashiri SY, Suzuki T, Okada A et al.   Musculoskeletal ultrasonography assists the diagnostic performance of the 2010 classification criteria for rheumatoid arthritis. Mod Rheumatol  2013; 23: 36– 43. Google Scholar CrossRef Search ADS PubMed  16 Minowa K, Ogasawara M, Murayama G et al.   Predictive grade of ultrasound synovitis for diagnosing rheumatoid arthritis in clinical practice and the possible difference between patients with and without seropositivity. Mod Rheumatol  2016; 26: 188– 93. Google Scholar PubMed  17 Navalho M, Resende C, Rodrigues AM et al.   Bilateral evaluation of the hand and wrist in untreated early inflammatory arthritis: a comparative study of ultrasonography and magnetic resonance imaging. J Rheumatol  2013; 40: 1282– 92. Google Scholar CrossRef Search ADS PubMed  18 Ozgul A, Yasar E, Arslan N et al.   The comparison of ultrasonographic and scintigraphic findings of early arthritis in revealing rheumatoid arthritis according to criteria of American College of Rheumatology. Rheumatol Int  2009; 29: 765– 8. Google Scholar CrossRef Search ADS PubMed  19 Rakieh C, Nam JL, Hunt L et al.   Predicting the development of clinical arthritis in anti-CCP positive individuals with non-specific musculoskeletal symptoms: a prospective observational cohort study. Ann Rheum Dis  2015; 74: 1659– 66. Google Scholar CrossRef Search ADS PubMed  20 Salaffi F, Ciapetti A, Gasparini S et al.   A clinical prediction rule combining routine assessment and power Doppler ultrasonography for predicting progression to rheumatoid arthritis from early-onset undifferentiated arthritis. Clin Exp Rheumatol  2010; 28: 686– 94. Google Scholar PubMed  21 Tamas MM, Rednic N, Felea I, Rednic S. Ultrasound assessment for the rapid classification of early arthritis patients. J Investig Med  2013; 61: 1184– 91. Google Scholar CrossRef Search ADS PubMed  22 van de Stadt LA, Bos WH, Meursinge Reynders M et al.   The value of ultrasonography in predicting arthritis in auto-antibody positive arthralgia patients: a prospective cohort study. Arthritis Res Ther  2010; 12: R98. Google Scholar CrossRef Search ADS PubMed  23 Wakefield RJ, Green MJ, Marzo-Ortega H et al.   Should oligoarthritis be reclassified? Ultrasound reveals a high prevalence of subclinical disease. Ann Rheum Dis  2004; 63: 382– 5. Google Scholar CrossRef Search ADS PubMed  24 Mangnus L, van Steenbergen HW, Reijnierse M, van der Helm-van Mil AH. Magnetic resonance imaging-detected features of inflammation and erosions in symptom-free persons from the general population. Arthritis Rheumatol  2016; 68: 2593– 602. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

RheumatologyOxford University Press

Published: May 29, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off