Test-Retest Reliability of Remote Assessments in Patients With Myotonic Dystrophy Type 1

We previously demonstrated the feasibility of remote assessments in individuals with myotonic dystrophy type 1 (DM1). This study aimed to evaluate test-retest reliability and agreement of remote assessments and the interrater reliability of video-recorded functional assessments in DM1.

Participants were remotely recruited from the National Registry and provided with a toolkit containing a tablet equipped with videoconferencing software and devices for strength and functional assessments. Two remote study visits (RSV1, RSV2) were conducted within 3 months. During each visit, participants completed video-supervised assessments: handgrip, pinch grip (PG), 9-hole peg test (9HPT), video hand opening time (vHOT), timed up and go (TUG), 10-meter walk/run test (10MWRT), sitting and supine forced vital capacity (FVC), sniff nasal inspiratory pressure (SNIP), and tongue and buccal strength tests. Timed tests were video-recorded and scored using standardized protocols. Intraclass correlation coefficients (ICCs) were calculated using 2-way mixed-effects model for test-retest reliability and 2-way random-effects model for interrater reliability. Agreement was evaluated using Bland-Altman plots and measurement sensitivity with minimal detectable differences at 95% confidence (MDD95%). Patient-reported outcomes (PROs) assessing dysphagia (Eating Assessment Tool-10 [EAT-10]) and upper and lower extremity function (Upper Extremity Function Index [UEFI], Lower Extremity Functional Scale [LEFS]) were collected at RSV1 and correlated with quantitative assessments (Spearman coefficient, ρ).

Forty individuals with DM1 (average age 47, 55% female) completed both RSVs. Test-retest reliability (ICC, 95% CI) was excellent for handgrip (0.99, 0.98-0.99), sitting and supine FVC (0.98, 0.96-0.99), 10MWRT (0.96, 0.91-0.98), TUG (0.94, 0.89-0.97), and 9HPT (0.93, 0.86-0.97) with adequate measurement sensitivity (MDD95% <30% for all). ICCs were acceptable for PG (0.96, 0.92-0.98), tongue strength (0.93, 0.86-0.96), right (0.93, 0.85-0.97) and left buccal strength (0.88, 0.75-0.94), and SNIP (0.88, 0.77-0.94) and moderate for vHOT (thumb 0.67, 0.42-0.83; middle finger 0.66, 0.40-0.83), but with inadequate measurement sensitivity (MDD95% >30% for all). Interrater reliability (ICC, 95% CI) was excellent for 9HPT (1), vHOT (thumb 0.99, 0.98-0.99; middle finger 0.98, 0.97-0.99), 10MWRT, and TUG (both 0.99, 0.98-0.99). Bland-Altman plots showed no systematic bias. Correlation (ρ, |95% CI|) was strong between handgrip and UEFI (+0.70, 0.49-0.84), 10MWRT and LEFS (-0.72, 0.86-0.49), and moderate between PG and UEFI (+0.60, 0.30-0.82), TUG and LEFS (-0.53, 0.73-0.23), and tongue strength and EAT-10 (-0.49, 0.71-0.20).

Remote assessments are feasible and safe. Many measurements demonstrate high reliability, show adequate measurement sensitivity, and correlate with PROs. These findings support remote research, facilitating broad participation, and reducing participant burden.