Sonja I. Van Nes

Modifying the Medical Research Council grading system through Rasch analyses

The Medical Research Council grading system has served through decades for the evaluation of muscle strength and has been recognized as a cardinal feature of daily neurological, rehabilitation and general medicine examination of patients, despite being respectfully criticized due to the unequal width of its response options. No study has systematically examined, through modern psychometric approach, whether physicians are able to properly use the Medical Research Council grades. The objectives of this study were: (i) to investigate physicians’ ability to discriminate among the Medical Research Council categories in patients with different neuromuscular disorders and with various degrees of weakness through thresholds examination using Rasch analysis as a modern psychometric method; (ii) to examine possible factors influencing physicians’ ability to apply the Medical Research Council categories through differential item function analyses; and (iii) to examine whether the widely used Medical Research Council 12 muscles sum score in patients with Guillain–Barré syndrome and chronic inflammatory demyelinating polyradiculoneuropathy would meet Rasch models expectations. A total of 1065 patients were included from nine cohorts with the following diseases: Guillain–Barré syndrome (n = 480); myotonic dystrophy type-1 (n = 169); chronic inflammatory demyelinating polyradiculoneuropathy (n = 139); limb-girdle muscular dystrophy (n = 105); multifocal motor neuropathy (n = 102); Pompes disease (n = 62) and monoclonal gammopathy of undetermined related polyneuropathy (n = 8). Medical Research Council data of 72 muscles were collected. Rasch analyses were performed on Medical Research Council data for each cohort separately and after pooling data at the muscle level to increase category frequencies, and on the Medical Research Council sum score in patients with Guillain–Barré syndrome and chronic inflammatory demyelinating polyradiculoneuropathy. Disordered thresholds were demonstrated in 74–79% of the muscles examined, indicating physicians’ inability to discriminate between most Medical Research Council categories. Factors such as physicians’ experience or illness type did not influence these findings. Thresholds were restored after rescoring the Medical Research Council grades from six to four options (0, paralysis; 1, severe weakness; 2, slight weakness; 3, normal strength). The Medical Research Council sum score acceptably fulfilled Rasch model expectations after rescoring the response options and creating subsets to resolve local dependency and item bias on diagnosis. In conclusion, a modified, Rasch-built four response category Medical Research Council grading system is proposed, resolving clinicians’ inability to differentiate among its original response categories and improving clinical applicability. A modified Medical Research Council sum score at the interval level is presented and is recommended for future studies in Guillain–Barré syndrome and chronic inflammatory demyelinating polyradiculoneuropathy.

Revised normative values for grip strength with the Jamar dynamometer

The Jamar dynamometer has been widely used in various chronic illnesses and has demonstrated its strength as a potential prognostic indicator. Various stratified normative values have been published using different methodologies, leading to conflicting results. No study used statistical techniques considering the non‐Gaussian distribution of the obtained grip strength (GS) values. Jamar GS was assessed in 720 healthy participants, subdivided into seven age decade groups consisting of at least 50 men and 50 women each. Normative values (median and fifth values) were calculated using quantile regressions with restricted cubic spline functions on age. Possible confounding personal factors (hand dominance, length, weight, hobby, and job categorization) were examined. Clinically applicable revised normative values for the Jamar dynamometer, stratified for age and gender, are presented. Hand dominance had no influence. Other personal factors only minimally influenced final values. This study provides revised normative GS values for the Jamar dynamometer.

Changing outcome in inflammatory neuropathies Rasch–comparative responsiveness

Objectives: We performed responsiveness comparison between the patient-reported Inflammatory Rasch-built Overall Disability Scale (I-RODS) and the widely used clinician-reported Inflammatory Neuropathy Cause and Treatment–Overall Neuropathy Limitation Scale (INCAT-ONLS) in patients with Guillain-Barré syndrome (GBS), chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), and immunoglobulin M–monoclonal gammopathy of undetermined significance related polyneuropathy (IgM-MGUSP). Methods: One hundred thirty-seven patients (GBS: 55, CIDP: 59, IgM-MGUSP: 23) with a new diagnosis or clinical relapse assessed both scales. Patients with GBS/CIDP were examined at 0, 1, 3, 6, and 12 months; patients with IgM-MGUSP at 0, 3, and 12. We subjected all data to Rasch analyses, and calculated for each patient the magnitude of change on both scales using the minimal clinically important difference (MCID) related to the individual standard errors (SEs). A responder was defined as having an MCID-SE ≥1.96. Individual scores on both measures were correlated with the EuroQoL thermometer (heuristic responsiveness). Results: The I-RODS showed a significantly higher proportion of meaningful improvement compared with the INCAT-ONLS findings in GBS/CIDP. For IgM-MGUSP, the lack of responsiveness during the 1-year study did not allow a clear separation. Heuristic responsiveness was consistently higher with the I-RODS. Conclusion: The I-RODS more often captures clinically meaningful changes over time, with a greater magnitude of change, compared with the INCAT-ONLS disability scale in patients with GBS and CIDP. The I-RODS offers promise for being a more sensitive measure and its use is therefore suggested in future trials involving patients with GBS and CIDP.

Outcome measures in immune-mediated neuropathies: the need to standardize their use and to understand the clinimetric essentials

Abstract  Peripheral neurological disorders like neuropathies may cause impairments (such as weakness and sensory deficits), which may lead to problems in daily life and social functioning with a possible decrement in quality of life expectations. Choosing the proper outcome measure to evaluate the therapeutic efficacy of an intervention at one of these levels of outcome should therefore be considered as fundamental to the design of randomized trials in peripheral neurological disorders. However, these choices are dependent not only on the proposed research purposes but also, and perhaps more importantly, on the fulfillment of the scientific needs of these measures. With an increasing demand for accuracy, a thorough and comprehensive evaluation of an outcome measure is needed to determine its simplicity, communicability, validity, reliability, and responsiveness before being clinically applicable, techniques that are being captured by the science of clinimetrics. Most neurologists are still unfamiliar with these rigorous methodological essentials or overlook some of them in their trial preparations because these are considered time consuming and mind numbing. This review will highlight, against the background of the international classification framework and clinimetric needs for outcome measures, the selected scales applied in published randomized controlled trials in patients with Guillain‐Barré syndrome, chronic inflammatory demyelinating polyradiculoneuropathy, multifocal motor neuropathy, and gammopathy‐related neuropathies. The need for comparison responsiveness studies between equally valid and reliable measures and to standardize their use is emphasized in these conditions. Finally, specific recommendations are given to move from classic to modern clinimetric approach when constructing, evaluating, and selecting outcome measures using new methods like Rasch analysis, accentuating the need of shifting toward a more modern era.

Improving fatigue assessment in immune‐mediated neuropathies: the modified Rasch‐built fatigue severity scale

Abstract Fatigue is a major disabling complaint in patients with immune‐mediated neuropathies (IN). The 9‐item fatigue severity scale (FSS) has been used to assess fatigue in these conditions, despite having limitations due to its classic ordinal construct. The aim was to improve fatigue assessment in IN through evaluation of the FSS using a modern clinimetric approach [Rasch unidimensional measurement model (RUMM2020)]. Included were 192 stable patients with Guillain‐Barré syndrome (GBS), chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) or polyneuropathy associated with monoclonal gammopathy of undetermined significance (MGUSP). The obtained FSS data were exposed to RUMM2020 model to investigate whether this scale would meet its expectations. Also, reliability and validity studies were performed. The original FSS did not meet the Rasch model expectations, primarily based on two misfitting items, one of these also showing bias towards the factor ‘walking independent.’ After removing these two items and collapsing the original 7‐point Likert options to 4‐point response categories for the remaining items, we succeeded in constructing a 7‐item Rasch‐built scale that fulfilled all requirements of unidimensionality, linearity, and rating scale model. Good reliability and validity were also obtained for the modified FSS scale. In conclusion, a 7‐item linearly weighted Rasch‐built modified FSS is presented for more proper assessment of fatigue in future studies in patients with immune‐mediated neuropathies.

Comparing the NIS vs. MRC and INCAT sensory scale through Rasch analyses

We performed a comparison between Neuropathy Impairment Scale‐sensory (NISs) vs. the modified Inflammatory Neuropathy Cause and Treatment sensory scale (mISS), and NIS‐motor vs. the Medical Research Council sum score in patients with Guillain‐Barré syndrome (GBS), chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), and IgM monoclonal gammopathy of undetermined significance‐related polyneuropathy (MGUSP). The ordinal data were subjected to Rasch analyses, creating Rasch‐transformed (RT)‐intervals for all measures. Comparison between measures was based on validity/reliability with an emphasis on responsiveness (using the patients level of change related to the individually obtained varying SE for minimum clinically important difference). Eighty stable patients (GBS: 30, CIDP: 30, and MGUSP: 20) were assessed twice (entry: two observers; 2–4 weeks later: one observer), and 137 newly diagnosed or relapsing patients (GBS: 55, CIDP: 59, and IgM‐MGUSP: 23) were serially examined with 12 months follow‐up. Data modifications were needed to improve model fit for all measures. The sensory and motor scales demonstrated approximately equal and acceptable validity and reliability scores. Responsiveness scores were poor but slightly higher in RT‐mISS compared to RT‐NISs. Responsiveness was equal for the RT‐motor scales, but higher in GBS compared to CIDP; responsiveness was poor in patients with MGUSP, suggesting a longer duration of follow‐up in the latter group of patients.

Rasch-built Overall Disability Scale for Multifocal motor neuropathy (MMN-RODS(©) ).

Clinical trials in multifocal motor neuropathy (MMN) have often used ordinal‐based measures that may not accurately capture changes. We aimed to construct a disability interval outcome measure specifically for MMN using the Rasch model and to examine its clinimetric properties. A total of 146 preliminary activity and participation items were assessed twice (reliability studies) in 96 clinically stable MMN patients. These patients also assessed the ordinal‐based overall disability sum score (construct, sample‐dependent validity). The final Rasch‐built overall disability scale for MMN (MMN‐RODS©) was serially applied in 26 patients with newly diagnosed or relapsing MMN, treated with intravenous immunoglobulin (IVIg) (1‐year follow‐up; responsiveness study). The magnitude of change for each patient was calculated using the minimum clinically important difference technique related to the individually obtained standard errors. A total of 121 items not fulfilling Rasch requirements were removed. The final 25‐item MMN‐RODS© fulfilled all Rasch models expectations and showed acceptable reliability and validity including good discriminatory capacity. Most serially examined patients improved, but its magnitude was low, reflecting poor responsiveness. The constructed MMN‐RODS© is a disease‐specific, interval measure to detect activity limitations in patients with MMN and overcomes the shortcomings of ordinal scales. However, future clinimetric studies are needed to improve the MMN‐RODS©s responsiveness by longer observations and/or more rigorous treatment regimens.