Assessing actigraphy performance for daytime sleep detection following stroke: insights from inpatient monitoring in a rehabilitation hospital.

Study Objectives: Stroke can result in or exacerbate various sleep disorders. The presence of behaviors such as daytime sleepiness poststroke can indicate underlying sleep disorders which can significantly impact functional recovery and thus require prompt detection and monitoring for improved care. Actigraphy, a quantitative measurement technology, has been primarily validated for nighttime sleep in healthy adults; however, its validity for daytime sleep monitoring is currently unknown. Therefore this study aims to identify the best-performing actigraphy sensor and algorithm for detecting daytime sleep in poststroke individuals. Methods: Participants wore Actiwatch Spectrum and ActiGraph wGT3X-BT on their less-affected wrist, while trained observers recorded daytime sleep occurrences and activity levels (active, sedentary, and asleep) during non-therapy times. Algorithms, Actiwatch (Autoscore AMRI) and ActiGraph (Cole-Kripke, Sadeh), were compared with on-site observations and assessed using F2 scores, emphasizing sensitivity to detect daytime sleep. Results: Twenty-seven participants from an inpatient stroke rehabilitation unit contributed 173.5 hours of data. The ActiGraph Cole-Kripke algorithm (minute sleep time = 15 minutes, bedtime = 10 minutes, and wake time = 10 minutes) achieved the highest F2 score (0.59). Notably, when participants were in bed, the ActiGraph Cole-Kripke algorithm continued to outperform Sadeh and Actiwatch AMRI, with an F2 score of 0.69. Conclusions: The study demonstrates both Actiwatch and ActiGraph's ability to detect daytime sleep, particularly during bed rest. ActiGraph (Cole-Kripke) algorithm exhibited a more balanced sleep detection profile and higher F2 scores compared to Actiwatch, offering valuable insights for optimizing daytime sleep monitoring with actigraphy in stroke patients.

Implementation of a Cell-Enabled Remote Blood Pressure Monitoring Program During the Postpartum Period at a Safety-Net Hospital.

BACKGROUND: Postpartum hypertension is a risk factor for severe maternal morbidity; however, barriers exist for diagnosis and treatment. Remote blood pressure (BP) monitoring programs are an effective tool for monitoring BP and may mitigate maternal health disparities. We aimed to describe and evaluate engagement in a remote BP monitoring program on BP ascertainment during the first 6-weeks postpartum among a diverse patient population. METHODS AND RESULTS: A postpartum remote BP monitoring program, using cell-enabled technology and delivered in multiple languages, was implemented at a large safety-net hospital. Eligible patients are those with hypertensive disorders before or during pregnancy. We describe characteristics of patients enrolled from January 2021 to May 2022 and examine program engagement by patient characteristics. Linear regression models were used to calculate mean differences and 95% CIs between characteristics and engagement metrics. We describe the prevalence of patients with BP ≥140/or >90 mm Hg. Among 1033 patients, BP measures were taken an average of 15.2 days during the 6-weeks, with the last measurement around 1 month (mean: 30.9 days), and little variability across race or ethnicity. Younger maternal age (≤25 years) was associated with less frequent measures (mean difference, -4.3 days [95% CI: -6.1 to -2.4]), and grandmultiparity (≥4 births) was associated with shorter engagement (mean difference, -3.5 days [95% CI, -6.1 to -1.0]). Prevalence of patients with BP ≥140/or >90 mm Hg was 62.3%, with differences by race or ethnicity (Black: 72.9%; Hispanic: 52.4%; White: 56.0%). CONCLUSIONS: A cell-enabled postpartum remote BP monitoring program was successful in uniformly monitoring BP and capturing hypertension among a diverse, safety-net hospital population.

Disentangling the effects of metabolic cost and accuracy on movement speed.

On any given day, we make countless reaching movements to objects around us. While such ubiquity may suggest uniformity, each movement's speed is unique-why is this? Reach speed is known to be influenced by accuracy; we slow down to sustain high accuracy. However, in other forms of movement like walking or running, metabolic cost is often the primary determinant of movement speed. Here we bridge this gap and ask: how do metabolic cost and accuracy interact to determine speed of reaching movements? First, we systematically measure the effect of increasing mass on the metabolic cost of reaching across a range of movement speeds. Next, in a sequence of three experiments, we examine how added mass affects preferred reaching speed across changing accuracy requirements. We find that, while added mass consistently increases metabolic cost thereby leading to slower metabolically optimal movement speeds, self-selected reach speeds are slower than those predicted by an optimization of metabolic cost alone. We then demonstrate how a single model that considers both accuracy and metabolic costs can explain preferred movement speeds. Together, our findings provide a unifying framework to illuminate the combined effects of metabolic cost and accuracy on movement speed and highlight the integral role metabolic cost plays in determining reach speed.

Machine Learning Models for Parkinson Disease: Systematic Review.

Background: With the increasing availability of data, computing resources, and easier-to-use software libraries, machine learning (ML) is increasingly used in disease detection and prediction, including for Parkinson disease (PD). Despite the large number of studies published every year, very few ML systems have been adopted for real-world use. In particular, a lack of external validity may result in poor performance of these systems in clinical practice. Additional methodological issues in ML design and reporting can also hinder clinical adoption, even for applications that would benefit from such data-driven systems. Objective: To sample the current ML practices in PD applications, we conducted a systematic review of studies published in 2020 and 2021 that used ML models to diagnose PD or track PD progression. Methods: We conducted a systematic literature review in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines in PubMed between January 2020 and April 2021, using the following exact string: "Parkinson's" AND ("ML" OR "prediction" OR "classification" OR "detection" or "artificial intelligence" OR "AI"). The search resulted in 1085 publications. After a search query and review, we found 113 publications that used ML for the classification or regression-based prediction of PD or PD-related symptoms. Results: Only 65.5% (74/113) of studies used a holdout test set to avoid potentially inflated accuracies, and approximately half (25/46, 54%) of the studies without a holdout test set did not state this as a potential concern. Surprisingly, 38.9% (44/113) of studies did not report on how or if models were tuned, and an additional 27.4% (31/113) used ad hoc model tuning, which is generally frowned upon in ML model optimization. Only 15% (17/113) of studies performed direct comparisons of results with other models, severely limiting the interpretation of results. Conclusions: This review highlights the notable limitations of current ML systems and techniques that may contribute to a gap between reported performance in research and the real-life applicability of ML models aiming to detect and predict diseases such as PD.

Wireless wearable sensors can facilitate rapid detection of sleep apnea in hospitalized stroke patients.

STUDY OBJECTIVES: To evaluate wearable devices and machine learning for detecting sleep apnea in patients with stroke at an acute inpatient rehabilitation facility (IRF). METHODS: A total of 76 individuals with stroke wore a standard home sleep apnea test (ApneaLink Air), a multimodal, wireless wearable sensor system (ANNE), and a research-grade actigraphy device (ActiWatch) for at least one night during their first week after IRF admission as part of a larger clinical trial. Logistic regression algorithms were trained to detect sleep apnea using biometric features obtained from the ANNE sensors and ground truth apnea rating from the ApneaLink Air. Multiple algorithms were evaluated using different sensor combinations and different apnea detection criteria based on the Apnea-Hypopnea Index (AHI≥5, AHI≥15). RESULTS: Seventy-one (96%) participants wore the ANNE sensors for multiple nights. In contrast, only forty-eight participants (63%) could be successfully assessed for OSA by ApneaLink; 28 (37%) refused testing. The best-performing model utilized photoplethysmography (PPG) and finger temperature features to detect moderate-severe sleep apnea (AHI≥15), with 88% sensitivity and a positive likelihood ratio (LR+) of 44.00. This model was tested on additional nights of ANNE data achieving 71% sensitivity (10.14 LR+) when considering each night independently and 86% accuracy when averaging multi-night predictions. CONCLUSIONS: This research demonstrates the feasibility of accurately detecting moderate-severe sleep apnea early in the stroke recovery process using wearable sensors and machine learning techniques. These findings can inform future efforts to improve early detection for post-stroke sleep disorders, thereby enhancing patient recovery and long-term outcomes.

Influences of 23 different equations used to calculate gene copies of SARS-CoV-2 during wastewater-based epidemiology.

Following the emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in late 2019, the use of wastewater-based surveillance (WBS) has increased dramatically along with associated infrastructure globally. However, due to the global nature of its application, and various workflow adaptations (e.g., sample collection, water concentration, RNA extraction kits), numerous methods for back-calculation of gene copies per volume (gc/L) of sewage have also emerged. Many studies have considered the comparability of processing methods (e.g., water concentration, RNA extraction); however, for equations used to calculate gene copies in a wastewater sample and subsequent influences on monitoring viral trends in a community and its association with epidemiological data, less is known. Due to limited information on how many formulas exist for the calculation of SARS-CoV-2 gene copies in wastewater, we initially attempted to quantify how many equations existed in the referred literature. We identified 23 unique equations, which were subsequently applied to an existing wastewater dataset. We observed a range of gene copies based on use of different equations, along with variability of AUC curve values, and results from correlation and regression analyses. Though a number of individual laboratories appear to have independently converged on a similar formula for back-calculation of viral load in wastewater, and share similar relationships with epidemiological data, differential influences of various equations were observed for variation in PCR volumes, RNA extraction volumes, or PCR assay parameters. Such observations highlight challenges when performing comparisons among WBS studies when numerous methodologies and back-calculation methods exist. To facilitate reproducibility among studies, the different gc/L equations were packaged as an R Shiny app, which provides end users the ability to investigate variability within their datasets and support comparisons among studies.

Early Prediction of Poststroke Rehabilitation Outcomes Using Wearable Sensors.

OBJECTIVE: Inpatient rehabilitation represents a critical setting for stroke treatment, providing intensive, targeted therapy and task-specific practice to minimize a patient's functional deficits and facilitate their reintegration into the community. However, impairment and recovery vary greatly after stroke, making it difficult to predict a patient's future outcomes or response to treatment. In this study, the authors examined the value of early-stage wearable sensor data to predict 3 functional outcomes (ambulation, independence, and risk of falling) at rehabilitation discharge. METHODS: Fifty-five individuals undergoing inpatient stroke rehabilitation participated in this study. Supervised machine learning classifiers were retrospectively trained to predict discharge outcomes using data collected at hospital admission, including patient information, functional assessment scores, and inertial sensor data from the lower limbs during gait and/or balance tasks. Model performance was compared across different data combinations and was benchmarked against a traditional model trained without sensor data. RESULTS: For patients who were ambulatory at admission, sensor data improved the predictions of ambulation and risk of falling (with weighted F1 scores increasing by 19.6% and 23.4%, respectively) and maintained similar performance for predictions of independence, compared to a benchmark model without sensor data. The best-performing sensor-based models predicted discharge ambulation (community vs household), independence (high vs low), and risk of falling (normal vs high) with accuracies of 84.4%, 68.8%, and 65.9%, respectively. Most misclassifications occurred with admission or discharge scores near the classification boundary. For patients who were nonambulatory at admission, sensor data recorded during simple balance tasks did not offer predictive value over the benchmark models. CONCLUSION: These findings support the continued investigation of wearable sensors as an accessible, easy-to-use tool to predict the functional recovery after stroke. IMPACT: Accurate, early prediction of poststroke rehabilitation outcomes from wearable sensors would improve our ability to deliver personalized, effective care and discharge planning in the inpatient setting and beyond.

Validation of Proprietary and Novel Step-counting Algorithms for Individuals Ambulating With a Lower Limb Prosthesis.

OBJECTIVE: To compare the accuracy and reliability of 10 different accelerometer-based step-counting algorithms for individuals with lower limb loss, accounting for different clinical characteristics and real-world activities. DESIGN: Cross-sectional study. SETTING: General community setting (ie, institutional research laboratory and community free-living). PARTICIPANTS: Forty-eight individuals with a lower limb amputation (N=48) wore an ActiGraph (AG) wGT3x-BT accelerometer proximal to the foot of their prosthetic limb during labeled indoor/outdoor activities and community free-living. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Intraclass correlation coefficient (ICC), absolute and root mean square error (RMSE), and Bland Altman plots were used to compare true (manual) step counts to estimated step counts from the proprietary AG Default algorithm and low frequency extension filter, as well as from 8 novel algorithms based on continuous wavelet transforms, fast Fourier transforms (FFTs), and peak detection. RESULTS: All algorithms had excellent agreement with manual step counts (ICC>0.9). The AG Default and FFT algorithms had the highest overall error (RMSE=17.81 and 19.91 steps, respectively), widest limits of agreement, and highest error during outdoor and ramp ambulation. The AG Default algorithm also had among the highest error during indoor ambulation and stairs, while a FFT algorithm had the highest error during stationary tasks. Peak detection algorithms, especially those using pre-set parameters with a trial-specific component, had among the lowest error across all activities (RMSE=4.07-8.99 steps). CONCLUSIONS: Because of its simplicity and accuracy across activities and clinical characteristics, we recommend the peak detection algorithm with set parameters to count steps using a prosthetic-worn AG among individuals with lower limb loss for clinical and research applications.

AI in Rehabilitation Medicine: Opportunities and Challenges.

Artificial intelligence (AI) tools are increasingly able to learn from larger and more complex data, thus allowing clinicians and scientists to gain new insights from the information they collect about their patients every day. In rehabilitation medicine, AI can be used to find patterns in huge amounts of healthcare data. These patterns can then be leveraged at the individual level, to design personalized care strategies and interventions to optimize each patient's outcomes. However, building effective AI tools requires many careful considerations about how we collect and handle data, how we train the models, and how we interpret results. In this perspective, we discuss some of the current opportunities and challenges for AI in rehabilitation. We first review recent trends in AI for the screening, diagnosis, treatment, and continuous monitoring of disease or injury, with a special focus on the different types of healthcare data used for these applications. We then examine potential barriers to designing and integrating AI into the clinical workflow, and we propose an end-to-end framework to address these barriers and guide the development of effective AI for rehabilitation. Finally, we present ideas for future work to pave the way for AI implementation in real-world rehabilitation practices.

Protocol for a randomized controlled trial to evaluate a year-long (NICU-to-home) evidence-based, high dose physical therapy intervention in infants at risk of neuromotor delay.

INTRODUCTION: Developmental disabilities and neuromotor delay adversely affect long-term neuromuscular function and quality of life. Current evidence suggests that early therapeutic intervention reduces the severity of motor delay by harnessing neuroplastic potential during infancy. To date, most early therapeutic intervention trials are of limited duration and do not begin soon after birth and thus do not take full advantage of early neuroplasticity. The Corbett Ryan-Northwestern-Shirley Ryan AbilityLab-Lurie Children's Infant Early Detection, Intervention and Prevention Project (Project Corbett Ryan) is a multi-site longitudinal randomized controlled trial to evaluate the efficacy of an evidence-based physical therapy intervention initiated in the neonatal intensive care unit (NICU) and continuing to 12 months of age (corrected when applicable). The study integrates five key principles: active learning, environmental enrichment, caregiver engagement, a strengths-based approach, and high dosage (ClinicalTrials.gov identifier NCT05568264). METHODS: We will recruit 192 infants at risk for neuromotor delay who were admitted to the NICU. Infants will be randomized to either a standard-of-care group or an intervention group; infants in both groups will have access to standard-of-care services. The intervention is initiated in the NICU and continues in the infant's home until 12 months of age. Participants will receive twice-weekly physical therapy sessions and caregiver-guided daily activities, assigned by the therapist, targeting collaboratively identified goals. We will use various standardized clinical assessments (General Movement Assessment; Bayley Scales of Infant and Toddler Development, 4th Edition (Bayley-4); Test of Infant Motor Performance; Pediatric Quality of Life Inventory Family Impact Module; Alberta Infant Motor Scale; Neurological, Sensory, Motor, Developmental Assessment; Hammersmith Infant Neurological Examination) as well as novel technology-based tools (wearable sensors, video-based pose estimation) to evaluate neuromotor status and development throughout the course of the study. The primary outcome is the Bayley-4 motor score at 12 months; we will compare scores in infants receiving the intervention vs. standard-of-care therapy.