Handwork vs machine: a comparison of rheumatoid arthritis patient populations as identified from EHR free-text by diagnosis extraction through machine-learning or traditional criteria-based chart review.

BACKGROUND: Electronic health records (EHRs) offer a wealth of observational data. Machine-learning (ML) methods are efficient at data extraction, capable of processing the information-rich free-text physician notes in EHRs. The clinical diagnosis contained therein represents physician expert opinion and is more consistently recorded than classification criteria components. OBJECTIVES: To investigate the overlap and differences between rheumatoid arthritis patients as identified either from EHR free-text through the extraction of the rheumatologist diagnosis using machine-learning (ML) or through manual chart-review applying the 1987 and 2010 RA classification criteria. METHODS: Since EHR initiation, 17,662 patients have visited the Leiden rheumatology outpatient clinic. For ML, we used a support vector machine (SVM) model to identify those who were diagnosed with RA by their rheumatologist. We trained and validated the model on a random selection of 2000 patients, balancing PPV and sensitivity to define a cutoff, and assessed performance on a separate 1000 patients. We then deployed the model on our entire patient selection (including the 3000). Of those, 1127 patients had both a 1987 and 2010 EULAR/ACR criteria status at 1 year after inclusion into the local prospective arthritis cohort. In these 1127 patients, we compared the patient characteristics of RA cases identified with ML and those fulfilling the classification criteria. RESULTS: The ML model performed very well in the independent test set (sensitivity=0.85, specificity=0.99, PPV=0.86, NPV=0.99). In our selection of patients with both EHR and classification information, 373 were recognized as RA by ML and 357 and 426 fulfilled the 1987 or 2010 criteria, respectively. Eighty percent of the ML-identified cases fulfilled at least one of the criteria sets. Both demographic and clinical parameters did not differ between the ML extracted cases and those identified with EULAR/ACR classification criteria. CONCLUSIONS: With ML methods, we enable fast patient extraction from the huge EHR resource. Our ML algorithm accurately identifies patients diagnosed with RA by their rheumatologist. This resulting group of RA patients had a strong overlap with patients identified using the 1987 or 2010 classification criteria and the baseline (disease) characteristics were comparable. ML-assisted case labeling enables high-throughput creation of inclusive patient selections for research purposes.

Early DAS response after DMARD-start increases probability of achieving sustained DMARD-free remission in rheumatoid arthritis.

BACKGROUND: Sustained DMARD-free remission (SDFR) is increasingly achievable. The pathogenesis underlying SDFR development is unknown and patient characteristics at diagnosis poorly explain whether SDFR will be achieved. To increase the understanding, we studied the course of disease activity scores (DAS) over time in relation to SDFR development. Subsequently, we explored whether DAS course could be helpful identifying RA patients likely to achieve SDFR. METHODS: 772 consecutive RA patients, promptly treated with csDMARDs (mostly methotrexate and treat-to-target treatment adjustments), were studied for SDFR development (absence of synovitis, persisting minimally 12 months after DMARD stop). The course of disease activity scores (DAS) was compared between RA patients with and without SDFR development within 7 years, using linear mixed models, stratified for ACPA. The relation between 4-month DAS and the probability of SDFR development was studied with logistic regression. Cumulative incidence of SDFR within DAS categories (< 1.6, 1.6-2.4, 2.4-3.6, ≥ 3.6) at 4 months was visualized using Kaplan-Meier curves. RESULTS: In ACPA-negative RA patients, those achieving SDFR showed a remarkably stronger DAS decline within the first 4 months, compared to RA patients without SDFR; - 1.73 units (95%CI, 1.28-2.18) versus - 1.07 units (95%CI, 0.90-1.23) (p < 0.001). In APCA-positive RA patients, such an effect was not observed, yet SDFR prevalence in this group was low. In ACPA-negative RA, DAS decline in the first 4 months and absolute DAS levels at 4 months (DAS4 months) were equally predictive for SDFR development. Incidence of SDFR in ACPA-negative RA patients was high (70.2%) when DAS4 months was < 1.6, whilst SDFR was rare (7.1%) when DAS4 months was ≥ 3.6. CONCLUSIONS: In ACPA-negative RA, an early response to treatment, i.e., a strong DAS decline within the first 4 months, is associated with a higher probability of SDFR development. DAS values at 4 months could be useful for later decisions to stop DMARDs.

Do magnetic resonance imaging-detected erosions predict progression to rheumatoid arthritis in patients presenting with clinically suspect arthralgia? A longitudinal study.

Objective: Radiographic joint erosions are a hallmark of rheumatoid arthritis (RA). Magnetic resonance imaging (MRI) is more sensitive than radiographs in detecting erosions. It is unknown whether MRI-detected erosions are predictive for RA development in patients with clinically suspect arthralgia (CSA). Therefore, we investigated the prognostic value of MRI-detected erosions, defined as any MRI erosion, or MRI erosion characteristics that were recently identified as specific for RA in patients with evident arthritis. Method: Patients presenting with CSA (n = 490) underwent contrast-enhanced 1.5 T MRI of the wrist, metacarpophalangeal (MCP) and metatarsophalangeal (MTP) joints. MRIs were scored according to the Rheumatoid Arthritis Magnetic Resonance Imaging Scoring system (RAMRIS). Presence of any MRI erosion (present in < 5% of symptom-free controls) and RA-specific erosion characteristics as identified previously (grade ≥ 2 erosions, erosions in MTP5, erosions in MTP1 if aged < 40 years) were studied with clinically apparent inflammatory arthritis development as outcome. Analyses were corrected for age and MRI-detected subclinical inflammation. Results: Erosions were present in 20%. Presence of any MRI erosion was not associated with arthritis development [multivariable analysis hazard ratio (HR) 0.97 (95% confidence interval 0.59-1.59)]. The different RA-specific erosion characteristics were not predictive [grade ≥ 2 HR 1.05 (0.33-3.34), erosions in MTP5 HR 1.08 (0.47-2.48), and MTP1 if aged < 40 years HR 1.11 (0.26-4.70)]. Erosion scores were higher in anti-citrullinated protein antibody (ACPA)-positive than in ACPA-negative patients (median 2.0 vs 1.0, p = 0.002), and related to more subclinical inflammation. Within both subgroups, MRI erosions were not predictive. Conclusions: MRI-detected erosions in hands and feet were not predictive for inflammatory arthritis development. Therefore, evaluating MRI for erosions in addition to subclinical inflammation does not provide added clinical value in CSA.

Contribution of tenosynovitis of small joints to the symptom morning stiffness in patients presenting with undifferentiated and rheumatoid arthritis.

Objective: Morning stiffness (MS) is characteristic of rheumatoid arthritis (RA). Despite its association with functional disability, the extent to which local inflammatory processes contribute to this symptom is unknown. Magnetic resonance imaging (MRI)-detected tenosynovitis of small joints is recognized as an early feature of RA, which is also associated with functional impairments. It has been proposed that tenosynovitis contributes to MS. Therefore, we assessed the relationship between MS and MRI-detected inflammation, in particular tenosynovitis.Method: In total, 286 consecutive patients newly presenting with undifferentiated arthritis and RA underwent contrast-enhanced 1.5 T MRI of (2-5) metacarpophalangeal, wrist, and (1-5) metatarsophalangeal joints. Scans were scored for tenosynovitis according to Haavardsholm, and for synovitis by Rheumatoid Arthritis Magnetic Resonance Imaging Scoring (RAMRIS). MS was dichotomized as ≥ 60 min or not. Associations between MS and tenosynovitis/synovitis were tested with logistic regression, data were categorized (solitary or simultaneous presence of synovitis/tenosynovitis), and the presence of an additive interaction was assessed.Results: MS was present in 40% of patients. Tenosynovitis was more often present in patients with MS than without MS [80% vs 65%, odds ratio (OR) 2.11, 95% confidence interval (1.21;3.69)]. Synovitis was more often present in patients with MS [58% vs 44%, OR 1.79 (1.11;2.91)]. In categorized analyses, concurrent synovitis and tenosynovitis had the largest association [OR 2.43 (1.30;4.54)], in contrast to solitary synovitis [OR 0.85 (0.21;3.47)]. The additive interaction was non-significant. The variance explained in all analyses was small (range 4-5%).Conclusion: Tenosynovitis, combined with synovitis, at small joints is associated with MS and contributes to the pathophysiology of MS.

Iron absorption in erythropoietin-treated haemodialysis patients: effects of iron availability, inflammation and aluminium.

BACKGROUND: The response to recombinant human erythropoietin (rHuEpo) is determined primarily by the availability of iron. In contrast to i.v., iron, oral iron supplementation is often insufficient for an optimal response. METHOD: We studied iron absorption and the effects of iron status, aluminium status and inflammation in 19 chronic haemodialysis patients on maintenance rHuEpo therapy. Iron mucosal uptake after 24 h, iron retention after 2 weeks and mucosal transfer of iron were determined with a whole-body counter using an oral dose 59Fe. Iron absorption was measured once without, and once after the ingestion of 2 g aluminium hydroxide. RESULTS: On the basis of transferrin saturation, two groups of dialysis patients were distinguished: a group with a functional iron deficiency (n = 9), and an iron-replete group (n = 10). In the iron-deficient dialysis patients group, mucosal uptake, mucosal transfer, and iron retention were 49.9% +/- 29.4, 0.73 +/- 0.29, and 41.6% +/- 32.2, being significantly lower than in a non-uraemic iron deficient population (P < 0.01, P < 0.05, P < 0.01 respectively). In the iron-replete dialysis patients group, mucosal uptake, mucosal transfer, and iron retention were 20.0 +/- 12.3, 0.59 +/- 0.18, and 11.1 +/- 6.7, mucosal uptake and iron retention being lower than in a normal iron-replete population (P < 0.0005 and P < 0.003 respectively). Dialysis patients with high C-reactive protein (CRP) values showed lower iron absorption. Iron absorption data correlated significantly with transferrin saturation and CRP in the iron-deficient group, and with serum ferritin in the iron-replete group. Iron absorption decreased after an aluminium hydroxide challenge in the iron-deficient patients to the lower levels of the iron-replete subjects. Body aluminium stores, estimated by the desferrioxamine test, did not correlate with parameters of iron absorption. CONCLUSION: The absorption of iron in dialysis patients is decreased in haemodialysis patients, which may, at least in part, be due to inflammation. Aluminium ingestion further reduces absorption in functional iron-deficient patients.

The roles of excessive gastro-oesophageal reflux, disordered oesophageal motility and decreased mucosal sensitivity in the pathogenesis of Barrett's oesophagus.

Barrett's oesophagus is often considered an end stage of gastro-oesophageal reflux disease. In its pathogenesis increased oesophageal acid exposure, disturbed oesophageal motility and decreased oesophageal mucosal sensitivity are thought to be of importance. In this review the role of each of these factors will be discussed and an update of the recent literature will be given.