Assessment of Natural Language Processing Methods for Ascertaining the Expanded Disability Status Scale Score From the Electronic Health Records of Patients With Multiple Sclerosis: Algorithm Development and Validation Study

Background The Expanded Disability Status Scale (EDSS) score is a widely used measure to monitor disability progression in people with multiple sclerosis (MS). However, extracting and deriving the EDSS score from unstructured electronic health records can be time-consuming. Objective We aimed to compare rule-based and deep learning natural language processing algorithms for detecting and predicting the total EDSS score and EDSS functional system subscores from the electronic health records of patients with MS. Methods We studied 17,452 electronic health records of 4906 MS patients followed at one of Canada’s largest MS clinics between June 2015 and July 2019. We randomly divided the records into training (80%) and test (20%) data sets, and compared the performance characteristics of 3 natural language processing models. First, we applied a rule-based approach, extracting the EDSS score from sentences containing the keyword “EDSS.” Next, we trained a convolutional neural network (CNN) model to predict the 19 half-step increments of the EDSS score. Finally, we used a combined rule-based–CNN model. For each approach, we determined the accuracy, precision, recall, and F-score compared with the reference standard, which was manually labeled EDSS scores in the clinic database. Results Overall, the combined keyword-CNN model demonstrated the best performance, with accuracy, precision, recall, and an F-score of 0.90, 0.83, 0.83, and 0.83 respectively. Respective figures for the rule-based and CNN models individually were 0.57, 0.91, 0.65, and 0.70, and 0.86, 0.70, 0.70, and 0.70. Because of missing data, the model performance for EDSS subscores was lower than that for the total EDSS score. Performance improved when considering notes with known values of the EDSS subscores. Conclusions A combined keyword-CNN natural language processing model can extract and accurately predict EDSS scores from patient records. This approach can be automated for efficient information extraction in clinical and research settings.

Characteristic Personality Traits of Multiple Sclerosis Patients—An Unicentric Prospective Observational Cohort Study

Background and objectives: Multiple sclerosis (MS) patients tend to present peculiar personality traits that highly impact their quality of life. Our study aimed to determine which personality traits are more common in MS patients compared to a sex- and age-matched control group. Methods and materials: Patients with relapsing–remitting MS along with a sex- and age-matched control group were included. All subjects completed the DECAS Personality Inventory and an additional form including demographic characteristics. Data (including descriptive statistics and univariate and multivariate analysis) were analyzed using SPSS. Results: 122 subjects were included, out of which 61 were in the patient group, mostly females (71.31%) with a mean age of 42.06 ± 10.46 years. Mean duration of disease was 10.18 ± 5.53 years and mean EDSS score was 2.09; 36% of patients were treated with Interferon-beta 1a. Subjects in the patient group presented significantly lower scores for extraversion (p = 0.036), specifically those with higher EDSS score, even after adjusting for possible confounders (age, sex, marital status, early retirement, alcohol, and tobacco consumption). Additionally, regarding orientation in life, MS patients were more often philosophers (p = 0.001), especially young males, whereas the dominant emotional feeling was less common, the actor profile (p = 0.022). Regarding task involvement, MS patients were often passive and compassionate concerning other people. Higher EDSS score also correlated with avoidant (p = 0.006) and melancholic (p = 0.043) personality traits. Subjects with higher education associated more often pragmatic, experimenter, popular, and optimist traits, whereas the elderly had actor, authoritarian, and experimenter profiles. Conclusions: Some MS patients may have reduced levels of extraversion and specific personality traits compared to age- and sex-matched subjects. Determining the exact personality profile might help the neurologist to establish a better therapeutic alliance and to apply specific interventions.

Pregnancy in women with MS: Impact on long-term disability accrual in a nationwide Danish Cohort

Background: Pregnancy is considered to influence the disease course in women with multiple sclerosis (MS). Objective: The aim of this study was to investigate the effect of pregnancy on long-term disability accrual in women with MS. Methods: The Danish Multiple Sclerosis Registry (DMSR) was used to identify women diagnosed with clinically isolated syndrome or relapsing-remitting MS. Cox models with pregnancy as a time-dependent exposure and propensity score (PS) models were used to evaluate time to reach confirmed Expanded Disability Status Scale (EDSS) score of 4 and 6. Results: A total of 425 women became parous and 840 remained nulliparous. When including pregnancy as a time-dependent exposure, a non-significant association with time to reach EDSS 4 (hazard ratio (HR) 0.86, 95% confidence interval (CI) 0.61–1.20) and EDSS 6 (HR 0.70, 95% CI 0.40–1.20) was found. Correspondingly, the PS model showed no association with pregnancy on time to reach EDSS 4 (HR 0.85, 95% CI 0.56–1.28). Conclusion: This study concludes that pregnancy does not affect long-term disability accumulation.

COVID-19 Severity in Multiple Sclerosis

Background and ObjectivesIt is unclear how multiple sclerosis (MS) affects the severity of COVID-19. The aim of this study is to compare COVID-19–related outcomes collected in an Italian cohort of patients with MS with the outcomes expected in the age- and sex-matched Italian population.MethodsHospitalization, intensive care unit (ICU) admission, and death after COVID-19 diagnosis of 1,362 patients with MS were compared with the age- and sex-matched Italian population in a retrospective observational case-cohort study with population-based control. The observed vs the expected events were compared in the whole MS cohort and in different subgroups (higher risk: Expanded Disability Status Scale [EDSS] score > 3 or at least 1 comorbidity, lower risk: EDSS score ≤ 3 and no comorbidities) by the χ2 test, and the risk excess was quantified by risk ratios (RRs).ResultsThe risk of severe events was about twice the risk in the age- and sex-matched Italian population: RR = 2.12 for hospitalization (p < 0.001), RR = 2.19 for ICU admission (p < 0.001), and RR = 2.43 for death (p < 0.001). The excess of risk was confined to the higher-risk group (n = 553). In lower-risk patients (n = 809), the rate of events was close to that of the Italian age- and sex-matched population (RR = 1.12 for hospitalization, RR = 1.52 for ICU admission, and RR = 1.19 for death). In the lower-risk group, an increased hospitalization risk was detected in patients on anti-CD20 (RR = 3.03, p = 0.005), whereas a decrease was detected in patients on interferon (0 observed vs 4 expected events, p = 0.04).DiscussionOverall, the MS cohort had a risk of severe events that is twice the risk than the age- and sex-matched Italian population. This excess of risk is mainly explained by the EDSS score and comorbidities, whereas a residual increase of hospitalization risk was observed in patients on anti-CD20 therapies and a decrease in people on interferon.

Real-World Effectiveness and Safety of Fingolimod in Patients with Multiple Sclerosis in the Czech Republic: Results from Core and Extension Parts of the GOLEMS Study up to 48 Months

Background: Fingolimod, an oral sphingosine 1-phosphate receptor immunomodulator, is approved in Europe for multiple sclerosis (MS) patients with highly active disease despite a full and adequate course of treatment with ≥1 disease-modifying therapy or patients with rapidly evolving severe relapsing–remitting MS. GOLEMS, a 12-month, national, multicenter, non-interventional, single-arm, real-world study showed a favorable benefit–risk profile of fingolimod in patients with MS in the Czech Republic. Here, we evaluated the long-term effectiveness and safety of fingolimod and its impact on disability progression and work capability for up to 48 months in patients with MS.Methods: The endpoints assessed were the incidence and severity of MS relapses in fingolimod-treated patients and the proportion of relapse-free patients up to 48 months of fingolimod treatment, change from baseline in the Expanded Disability Status Scale (EDSS) score, and change from baseline in work capability assessment. Efficacy outcomes were analyzed in the completed and efficacy sets, and safety was evaluated in all the enrolled patients.Results: Of 240 enrolled patients, 237 were included into efficacy set. Patients with a minimum of a 12-month observation period in the core study who continued fingolimod treatment, were eligible to participate in the extension phase. Of 211 patients enrolled in extension study, 155 were evaluated in the completed set. Based on analysis of 48-month period of fingolimod treatment, 95/237 patients (40.1%) in the efficacy set, 54/155 (34.8%) in the completed set were free of relapses. The majority of relapses reported were moderate in intensity. Mean EDSS score remained stable throughout 48-month study period (Baseline, 3.4; Month 48, 3.6). No significant 3 trend was observed in changes in work capability assessment or number of missed days of work. Of 240 enrolled patients, 147 (61.3%) had ≥1 treatment-emergent adverse event (AE) and 20 (8.3%) reported serious AEs. In total, 45 patients (18.8%) permanently discontinued treatment because of AEs related to study drug; two patients reported pregnancy after treatment initiation and subsequently discontinued the treatment; no deaths were reported.Conclusion: GOLEMS study demonstrated the sustained effectiveness and manageable safety profile of fingolimod under real-world conditions over 48 months in patients with MS.Trial registration: Not applicable

Monocyte to High-Density Lipoprotein Ratio: A Novel Predictive Marker of Disease Severity and Prognosis in Patients With Neuromyelitis Optica Spectrum Disorders

Background and Purpose: To investigate the association of monocyte to high-density lipoprotein ratio (MHR) with disease severity and prognosis in patients with neuromyelitis optica spectrum disorders (NMOSD).Methods: This retrospective study included 125 patients with NMOSD. Demographic and clinical parameters, including the MHR, were assessed. The initial Expanded Disability Status Scale (EDSS) score and relapse rate were used to evaluate disease severity and prognosis, respectively. Correlations between MHR and disease severity and relapse rate were analyzed. The predictive value of MHR for prognosis was evaluated using receiver operating characteristic (ROC) curve analysis.Results: Compared with the low MHR group, the initial EDSS score (median 4.5 vs. 5.5%, P = 0.025) and relapse rate (51.61 vs. 30.16%, P = 0.015) were significantly higher in the high MHR group. MHR was positively correlated with the initial EDSS score (r = 0.306, P = 0.001). Multivariate analysis showed that MHR was significantly associated with severity (odds ratio = 7.90, 95% confidence interval [CI] = 1.08–57.82, P = 0.041), and it was a significant predictor of disease prognosis (hazard ratio = 3.12, 95% CI = 1.02–9.53, P = 0.046). The median relapse interval of the high MHR group was 24.40 months. When the MHR was higher than 0.565, the risk of relapse was high [sensitivity, 33.3%; specificity, 91.9%; area under the ROC curve, 0.642 (95% CI = 0.54–0.74, P = 0.007)].Conclusion: MHR is a novel predictive marker of disease severity and prognosis in patients with NMOSD. Early monitoring and reduction of MHR may allow earlier intervention and improved prognosis.

Factors Influencing the Degree of Disability in Patients With Multiple Sclerosis

Objective: To explore the factors influencing the degree of disability in patients with multiple sclerosis (MS), and to provide evidence for its early diagnosis, prognostic evaluation and clinical intervention.Methods: This retrospective observational study included 72 patients with relapsing-remitting multiple sclerosis (RRMS) at the First Hospital of Shanxi Medical University. All patients completed craniocerebral and spinal cord MRI (with or without Gd enhancement) and were evaluated for Expanded Disability Status Score (EDSS) scores before receiving treatment.Results: Among 72 patients with RRMS, 45 (62.5%) had an EDSS score ≤3; A total of 27 patients (37.5%) had an EDSS score &gt;3 points. Univariate analysis showed that age, annual recurrence rate (ARR), drug use, albumin (ALB), triglycerides (TG), and total number of lesions in groups with EDSS score ≤3 were significantly different from those with an EDSS score &gt; 3 points (P &lt; 0.05). Multivariate logistic regression analysis showed that ALB, total number of lesions, and drug use were independent influencing factors of the degree of disability in patients with MS, and the difference was statistically significant (P &lt; 0.05). An ROC curve was constructed using ALB and the total number of lesions. The AUC of ALB was 0.681, P &lt; 0.05, and the best cut-off value was 44.2 g/L. Its sensitivity to predict the degree of disability in patients with multiple sclerosis was 85.2%, while its specificity was 51.1%. The AUC of the total number of lesions was 0.665 (P &lt; 0.05) and the best cut-off value was 5.5. Its sensitivity to predict the degree of disability in patients with multiple sclerosis was 70.4%, while its specificity was 64.4%. The AUC of the combined ALB, total number of lesions, and drug use was 0.795 (P &lt; 0.05), sensitivity was 77.8, and specificity was 73.3%. The optimal diagnostic cut-off value of the regression equation for the EDSS score of patients with multiple sclerosis was 0.420.Conclusion: Serum ALB, total number of lesions, and drug use in patients with multiple sclerosis were independent factors influencing the degree of disability. These findings provide clinical evidence for the prognostic evaluation and early intervention of patients with multiple sclerosis.

Early predictors of disability of paediatric-onset AQP4-IgG-seropositive neuromyelitis optica spectrum disorders

ObjectiveTo describe onset clinical features predicting time to first relapse and time to long-term visual, motor and cognitive disabilities in paediatric-onset aquaporin-4 antibody (AQP4-IgG) neuromyelitis optica spectrum disorders (NMOSDs).MethodsIn this retrospective UK multicentre cohort study, we recorded clinical data of paediatric-onset AQP4-IgG NMOSD. Univariate and exploratory multivariable Cox proportional hazard models were used to identify long-term predictors of permanent visual disability, Expanded Disability Status Scale (EDSS) score of 4 and cognitive impairment.ResultsWe included 49 paediatric-onset AQP4-IgG patients (38.8% white, 34.7% black, 20.4% Asians and 6.1% mixed), mean onset age of 12±4.1 years, and 87.7% were female. Multifocal onset presentation occurred in 26.5% of patients, and optic nerve (47%), area postrema/brainstem (48.9%) and encephalon (28.6%) were the most involved areas. Overall, 52.3% of children had their first relapse within 1 year from disease onset. Children with onset age <12 years were more likely to have an earlier first relapse (p=0.030), despite showing no difference in time to immunosuppression compared with those aged 12–18 years at onset. At the cohort median disease duration of 79 months, 34.3% had developed permanent visual disability, 20.7% EDSS score 4 and 25.8% cognitive impairment. Visual disability was associated with white race (p=0.032) and optic neuritis presentations (p=0.002). Cognitive impairment was predicted by cerebral syndrome presentations (p=0.048), particularly if resistant to steroids (p=0.034).ConclusionsAge at onset, race, onset symptoms and resistance to acute therapy at onset attack predict first relapse and long-term disabilities. The recognition of these predictors may help to power future paediatric clinical trials and to direct early therapeutic decisions in AQP4-IgG NMOSD.

The Relationship between Third Ventricle Diameter, Motor Disability and Cognitive Decline in Patients with Multiple Sclerosis

Background: Multiple Sclerosis (MS) causes brain atrophy at the early stages of the disease which leads to progressive motor and cognitive dysfunction. Brain atrophy can be diagnosed indirectly by measuring the Third Ventricle Diameter (TVD) using Trans Cranial Sonography (TCS). The purpose of the current study was evaluation of TVD in MS patients using TCS to examine its possible correlation with cognitive dysfunction and Expanded Disability Status Scale (EDSS). Methods: Seventy-four patients with a definite diagnosis of MS were enrolled in this study. Transverse diameter of the third ventricle was measured using TCS. All patients were assessed by neurological examination and the level of disability was measured via EDSS. The cognitive performance was assessed by the Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS). Pearson’s correlation was performed to evaluate possible correlations and p-value<0.05 was considered statistically significant. Results: From the total 74 individuals, 58.1% were diagnosed with Relapse-Remitting MS (RRMS) (n=43) and 43.9% with Secondary-Progressive MS (SPMS) (n=31). The mean EDSS score was 1.81±1.38 (Range of 0-10). The mean TVD was 5.61±1.82 cm which had no statistical correlation with the EDSS score, but it was significantly wider in group with EDSS score>3 when compared to the group with EDSS score≤3 (p-value=0.0001). The mean BICAMS score was 0.65±0.57 and there was no statistical correlation between BICAMS score and TVD. Conclusion: Measuring the diameter of the third ventricle using TCS appears to be an appropriate method to diagnose brain atrophy and is associated with cognitive dysfunction in the process of MS. Our data emphasized no association between the diameter of the third ventricle and cognitive dysfunction.

Assessing attacks and treatment response rates among adult patients with NMOSD and MOGAD: Data from a nationwide registry in Argentina

We aimed to examine treatment interventions implemented in patients experiencing neuromyelitis optica spectrum disorders (NMOSD) attacks (frequency, types, and response). Methods Retrospective study. Data on patient demographic, clinical and radiological findings, and administered treatments were collected. Remission status (complete [CR], partial [PR], no remission [NR]), based on changes in the EDSS score was evaluated before treatment, during attack, and at 6 months. CR was analyzed with a generalized estimating equations (GEEs) model. Results A total of 131 patients (120 NMOSD and 11 myelin oligodendrocyte glycoprotein-antibody-associated diseases [MOGAD]), experiencing 262 NMOSD-related attacks and receiving 270 treatments were included. High-dose steroids (81.4%) was the most frequent treatment followed by plasmapheresis (15.5%). CR from attacks was observed in 47% (105/223) of all treated patients. During the first attack, we observed CR:71.2%, PR:16.3% and NR:12.5% after the first course of treatment. For second, third, fourth, and fifth attacks, CR was observed in 31.1%, 10.7%, 27.3%, and 33.3%, respectively. Remission rates were higher for optic neuritis vs. myelitis (p < 0.001). Predictor of CR in multivariate GEE analysis was age in both NMOSD (OR = 2.27, p = 0.002) and MOGAD (OR = 1.53, p = 0.03). Conclusions This study suggests individualization of treatment according to age and attack manifestation. The outcome of attacks was generally poor.