title
stringlengths
1
1.19k
keywords
stringlengths
0
668
concept
stringlengths
0
909
paragraph
stringlengths
0
61.8k
PMID
stringlengths
10
11
2.4. Outcome Measurement
PMC9964691
2.4.1. Discomfort Scores
NRS
For the assessment of discomfort scores, an 11-point numerical rating scale (NRS) was used. This scale was scored from 0 (without any discomfort) to 10 (extreme discomfort). The discomfort was scored three times, at T1, T4, and T7. Ferreira-Valente et al. (2011) claimed that this NRS can be used to determine specific indications [
PMC9964691
2.4.2. Patients’ Satisfaction
The global perceived effect (GPE) scale was selected to assess the participants’ satisfaction. The GPE is a 10-point scale ranging from −5 (worst symptom) to 0 (no change) to +5 (close to normal improvement). The intraclass correlation coefficient (ICC) of the GPE was between 0.90 and 0.99, which was assessed as excellent [
PMC9964691
2.4.3. Trunk Muscle Fatigue
muscle fatigue
PAD, CONTRACTION
The skin of each participant was cleaned with an alcohol pad before the researcher attached one pair of surface electromyography (sEMG) electrodes (EL 503), with an electrical contact surface area of 1 cmThe electromyography data were recorded at 2000 samples per second using the Wireless Bipolar Cometa Mini Wave Plus 16-channel EMG system (Cometa, Bareggio, Italy), with an online band-pass filter (10–500 Hz) and a 60 Hz notch filter (power line in Thailand). For the experimental task, participants performed the required sitting condition for 60 min (i) in a control condition (sitting without support), (ii) with a foam pillow (sitting with a back pillow made of foam material), and (iii) with a rubber pillow (sitting with a back pillow made of rubber material) in a random order on three consecutive days (The raw EMG signals were full-wave rectified and represented as median frequency (MF) values. EMG normalization was the method by which the magnitude of muscle activation was expressed as a percentage of the muscle’s activity during a calibrated test condition. The current study evaluated the maximum voluntary isometric contraction (MVIC) of the trunk muscle utilizing the methods outlined by Imai et al. (2010) for normalizing data [To determine the MVIC values in the TrA and IO muscles, three muscle tests were performed. A rest period of two minutes was allowed between the tests to avoid muscle fatigue [
PMC9964691
2.5. Statistical Analysis
All analyses were performed using SPSS version 19.0 software (SPSS Inc., Chicago, IL, USA). Mean and standard deviation (SD) were used to assess the participants’ demographics. A Shapiro–Wilk test was performed to check the data distribution. A paired
PMC9964691
3. Results
The participants’ demographic characteristics are shown in
PMC9964691
3.1. Discomfort Score
Three groups of participants who sat for 60 min were evaluated for their back comfort using a discomfort score. As shown in As shown in Participants experienced the most back discomfort while sitting without a back pillow at all time points. Furthermore, when participants used the rubber back pillow, they experienced the least back discomfort compared to the other groups, at a
PMC9964691
3.2. Participants’ Satisfaction
The participants’ satisfaction was measured twice while they sat for 60 min (as shown in When comparing the patients’ satisfaction within groups, the satisfaction of the control group decreased significantly after 60 min of sitting. However, within the rubber group, there was a statistically significant increase in satisfaction when participants sat for 60 min.Compared between groups, participants were more satisfied with back pillows (both types) than without back support at T1. Furthermore, at T1, it was revealed that the satisfaction of participants using rubber back pillows was statistically considerably higher than that of the participants using foam pillows.The effect of satisfaction at T1 was consistent with that observed at T7, in that satisfaction was greater in the back pillow group than in the control group. Participants using rubber pillows reported significantly higher satisfaction levels than those using foam pillows.
PMC9964691
4. Discussion
muscle fatigue, fatigue, pain
CONTRACTION
The current study investigated the effect of using rubber pillows compared to foam pillows on TrA and IO muscle fatigue, discomfort scores, and participant satisfaction. Thirty healthy participants were asked to sit for 60 min on three consecutive days under three sitting conditions. The three conditions were (i) a control group (sitting without support), (ii) a foam pillow group (sitting with a foam pillow), and (iii) a rubber pillow group (sitting with a rubber pillow). Patients’ discomfort scores increased in all groups while sitting for only 30 min, particularly in the control group, where participants sat without a back pillow and had the highest discomfort level compared with the other two pillow groups. Previous studies that showed prolonged sitting can induce discomfort, even in healthy individuals, supported these results [After 30 and 60 min of sitting, the control group had more back discomfort than the foam and rubber pillow groups. The discomfort score was lower in the pillow group, and this finding was concordant with the study of Prommanon et al. (2015), which found decreased pain intensity in participants who received back pillows in addition to physical therapy [Using a back pillow may have psychological effects and cause participants to maintain correct posture and good ergonomics [The foam pillow group had much higher back discomfort than the back rubber pillow group. This may be because the natural rubber pillow had higher resilience, which made participants feel better when using it. Thus, rubber is typically used in mattress and pillow production [Regarding participant satisfaction, the current study discovered that participants who sat with a back support device were more satisfied than those without one at the initiation of sitting (T1) or after 60 min of sitting (T7). Participants were more satisfied with the rubber back pillow than with the foam pillow at both T1 and T7. Kompayak et al. (2016) compared patient satisfaction between the foam pillow and lumbar support, and they reported that participants who used the foam pillow achieved higher satisfaction than lumbar support users [The current study reported that participants without a back pillow (control group) had TrA and IO fatigue at T7 (sitting for 60 min). Sitting for prolonged periods may cause TrA and IO muscle fatigue due to the continuous contraction of the TrA and IO muscles in seated postures [However, the foam and rubber groups showed no difference in MDF values over 60 min of sitting. The TrA and IO muscles play a crucial stabilizing role in the lumbopelvic region and reduce stress on spinal structures [
PMC9964691
5. Conclusions
trunk muscle fatigue, muscle fatigue
The current study was the first to compare the effects of foam and rubber pillows on TrA and IO muscle fatigue, discomfort scores, and patient satisfaction. The authors recommend that individuals who sit for prolonged periods during the day use back support to delay deep trunk muscle fatigue and reduce discomfort scores. Furthermore, rubber pillows may provide enhanced comfort due to their softness and flexibility compared to foam pillows.
PMC9964691
Author Contributions
Conceptualization, R.P.; Methodology, T.C. and P.S.; Formal analysis, T.C. and P.S.; Investigation, P.S. and A.L.; Data curation, A.L.; Writing—original draft, T.C. and A.L.; Writing—review & editing, R.P. and T.C.; Project administration, R.P. All authors have read and agreed to the published version of the manuscript.
PMC9964691
Institutional Review Board Statement
Pain
This randomized crossover study was conducted at the Research Center in the Back, Neck, Other Joint Pain, and Human Performance (BNOJPH) laboratory at Khon Kaen University. The Khon Kaen University Ethics Committee (HE 632261, Khon Kaen, Thailand, 17 December 2020) approved the current study.
PMC9964691
Informed Consent Statement
Informed consent was obtained from all subjects involved in the study.
PMC9964691
Data Availability Statement
The data will be available for anyone who wishes to access them for any purpose and contract should be made via the corresponding author (thiwch@kku.ac.th).
PMC9964691
Conflicts of Interest
The authors declare no conflict of interest.
PMC9964691
References
Trunk muscle, Muscle fatigue, fatigue
Overview of the study.The assessment times for each outcome.General characteristics of 30 healthy participants.MaleFemaleHigh schoolBachelor’s degreeMaster’s degreeYesNoYesNoYesNoAbbreviations: SD, standard deviation; BMI, Body Mass Index.Comparison of discomfort scores in the back region within the group and between the experimental groups.Note: Discomfort score reported as median (interquartile range); The comparison of patients’ satisfaction within and between groups.Note: Patient’s satisfaction reported as median (interquartile range); Trunk muscle (TrA and IO) fatigue while sitting for 60 min.Note: Muscle fatigue was reported as median (interquartile range);
PMC9964691
Purpose
myocardial and cerebrovascular infarction, hypotension, hypotensive, Hypotension, acute kidney injury
ADVERSE EVENTS, EVENTS, HYPOTENSIVE
Intraoperative hypotension is linked to increased incidence of perioperative adverse events such as myocardial and cerebrovascular infarction and acute kidney injury. Hypotension prediction index (HPI) is a novel machine learning guided algorithm which can predict hypotensive events using high fidelity analysis of pulse-wave contour. Goal of this trial is to determine whether use of HPI can reduce the number and duration of hypotensive events in patients undergoing major thoracic procedures.
PMC10061960
Methods
esophageal or lung resection, hypotensive
ADVERSE EVENTS, EVENTS, HYPOTENSIVE
Thirty four patients undergoing esophageal or lung resection were randomized into 2 groups -“machine learning algorithm” (AcumenIQ) and “conventional pulse contour analysis” (Flotrac). Analyzed variables were occurrence, severity and duration of hypotensive events (defined as a period of at least one minute of MAP below 65 mmHg), hemodynamic parameters at 9 different timepoints interesting from a hemodynamics viewpoint and laboratory (serum lactate levels, arterial blood gas) and clinical outcomes (duration of mechanical ventilation, ICU and hospital stay, occurrence of adverse events and in-hospital and 28-day mortality).
PMC10061960
Results
hypotensive, hypotension, AUT
EVENTS, HYPOTENSIVE
Patients in the AcumenIQ group had significantly lower area below the hypotensive threshold (AUT, 2 vs 16.7 mmHg x minutes) and time-weighted AUT (TWA, 0.01 vs 0.08 mmHg). Also, there were less patients with hypotensive events and cumulative duration of hypotension in the AcumenIQ group. No significant difference between groups was found in terms of laboratory and clinical outcomes.
PMC10061960
Conclusions
hypotensive
EVENTS, HYPOTENSIVE
Hemodynamic optimization guided by machine learning algorithm leads to a significant decrease in number and duration of hypotensive events compared to traditional goal directed therapy using pulse-contour analysis hemodynamic monitoring in patients undergoing major thoracic procedures. Further, larger studies are needed to determine true clinical utility of HPI guided hemodynamic monitoring.
PMC10061960
Trial registration
Date of first registration: 14/11/2022Registration number: 04729481-3a96-4763-a9d5-23fc45fb722d
PMC10061960
Keywords
PMC10061960
Introduction
cerebral and myocardial infarction, myocardial infarction, IOH includes systolic blood pressure, hypovolemia, cerebrovascular insult, IOH, hypotension, postoperative organ dysfunction, Hypotension, acute kidney injury
MYOCARDIAL INFARCTION, POSTOPERATIVE COMPLICATIONS, ADVERSE EVENTS, HEMODYNAMIC INSTABILITY, COMPLICATIONS, EDWARDS
In Europe, approximately 20 million major surgical procedures are performed annually. Up to 5% of these patients will die and 10–15% will develop adverse events that could have been prevented in more than 33% of cases [Intraoperative hypotension (IOH) may increase mortality and morbidity rates in the postoperative period, and is associated with adverse events such as acute kidney injury, cerebral and myocardial infarction [A recent randomized controlled trial reported that preventing intraoperative hypotension reduces the risk of postoperative organ dysfunction by about a quarter [Absolute threshold in defining IOH includes systolic blood pressure (SBP) less than 90 mmHg or mean arterial pressure (MAP) less than 65 mmHg, while a 20% reduction in baseline SBP or a 30% reduction in MAP represents the relative threshold. Most patients in noncardiac surgery experience at least one episode during which MAP decreases to < 65 mmHg, and known causes are anesthetic drugs, uncorrected hypovolemia, preexisting comorbidities, and surgical manipulation [The “Hypotension Prediction Index” (HPI) is a novel hemodynamic monitoring tool that predicts episodes of intraoperative hypotension before they occur. It has recently been shown that applying Acumen (Edwards Lifesciences, Irvine, Ca, USA) HPI software algorithm on data obtained by invasive blood pressure monitoring sensor during noncardiac surgery enables earlier assessment of causes of impending hypotension and timely response to possible hemodynamic instability [Given the significant association between hypotension and postoperative complications and adverse outcomes such as increased incidence of myocardial infarction and cerebrovascular insult, as well as increased hospital stay [The aim of this study was to examine whether use of machine-learning algorithm guided intraoperative patient hemodynamic optimization reduces duration and severity of hypotension and its complications during and after thoracic surgery compared to conventional pulse contour analysis goal-directed hemodynamic optimization.
PMC10061960
Patients, materials and methods
PMC10061960
Participants and group allocation
heart failure, heart defects, coin toss, valvular anomalies
HEART FAILURE, PERSISTENT ATRIAL FIBRILLATION, HEART
By design this study is a prospective, randomized, single blinded study.Participants are patients over 18 years of age which were scheduled for elective major thoracic procedure (lung resection, pleurectomy or resection of the esophagus) with planned thoracotomy and intraoperative period of one lung ventilation with planned postoperative admission to the ICU. Exclusion criteria were persistent atrial fibrillation, structural heart defects (shunting or moderate to severe valvular anomalies), preoperative serum hemoglobin levels < 120 g/L and severe heart failure classified as New York Heart Association (NYHA) grade IV.In the anesthesia preparation room, after informed consent was obtained, patients were randomized using coin toss into two groups:“machine learning algorithm”
PMC10061960
Study protocol
PMC10061960
Hemodynamic measurements
tumor
STERILE, TUMOR, INFILTRATION, DECUBITUS
In the operating room and before induction of anesthesia, arterial and central venous lines were placed after sterile skin preparation and local anesthetic infiltration, zeroed to account for hydrostatic pressure differences between the pressure transducer and the heart, and monitoring was initiated.Following hemodynamic parameters were measured: cardiac index (CI, l/min/mHemodynamic measurements were recorded at following time points: at baseline (T0), during induction of anesthesia - 3 min after administration of muscle relaxant (T1), 1 min after intubation (T2), 1 min after placing the patient in lateral decubitus position (T3), 1 min after skin incision (T4), 1 min after thoracotomy and initiation of one-lung ventilation (T5), 1 min after removal of tumor (T6), 3 min after skin closure (T7) and 1 min after placing the patient back in the supine position (T8).
PMC10061960
Anesthesia protocol and hemodynamics related interventions
Induction of anesthesia was performed using propofol (1–1.5 mg/kg), sufentanil (0.1–0.5 mcg/kg) and rocuronium bromide (0.6 mg/kg). After intubation, placement of Robert Shaw tube was verified capnographically and the correct position of endobronchial lumen was verified using a flexible fiber bronchoscope. Anesthesia was maintained using oxygen/air/sevoflurane mixture at minimal alveolar concentration (MAC) 0.8–1 and FiOIn patients with thoracic epidural catheter in place, continuous infusion of 1 mg/ml levobupivacaine and 1 mcg/ml sufentanil was administered at rate of 1–2 ml/h, according to attending anesthesiologist’s assessment of patient’s nociceptive response, and in those without an epidural catheter, 5—10 mcg sufentanil i.v. boluses were administered as needed.Patients were ventilated (Draeger Perseus, Draeger Medical AG, Lübeck, Germany) using a tidal volume of 8 ml/kgIn the In the In supine patients SVV ≥ 13% defined volume responsiveness [
PMC10061960
Endpoints
hypotension, AUT
HYPOTENSIVE EPISODE
Primary endpoint is time weighted average of area spent under 65 mmHg for MAP per patient (AUT - depth of hypotension below a MAP of 65 mmHg × time spent below a MAP of 65 mm Hg divided by procedure duration in minutes) of AUT (TWA-AUT).Secondary endpoints are AUT, number of hypotensive episodes (defined as MAP < 65 mmHg for at least 1 min) during the procedure and cumulative duration of hypotension during the procedure.Volume of i.v. fluids (ml/kg/h) and norepinephrine (mcg/kg/min) administered in the operating theater, urine output (ml/kg/h), postoperative arterial blood pH, arterio-venous CO
PMC10061960
Discussion
blood loss, hypotension, AUT, hypotensive, Hypotension
BLOOD LOSS, EVENTS, HYPOTENSIVE EPISODE, HYPOTENSIVE
In the studied cohort, use of machine learning algorithm guided hemodynamic optimization (HPI) significantly reduced the number and duration of intraoperative hypotensive episodes, as well as AUT and the primary outcome—TWA of AUT.These results are in agreement with results obtained by Wijnberge et al. [A recent study by Enevoldsen et al. [Our results show that while incidence and duration of hypotension were significantly reduced in patients that used HPI algorithm, there were no significant differences in clinical outcomes, however these results should be interpreted with caution due to sample size and inadequate statistical power. Results of a large retrospective study by Gregory et al. [There was one patient in the While hemodynamic stability was much better preserved in the Hypotension prediction index has proven to be reliable enough in prediction of intraoperative hypotensive events even when used with non-invasive volume clamp method, with similar sensitivity and specificity levels as when used with arterial lines [Since most elective thoracic procedures (lung and esophagus resection) are generally not associated with major blood loss and fluid shift, most hemodynamic changes are due to surgical manipulation, the patient’s position and changes in cardiac output during OLV, which rises after pleural opening [One probable reason why both incidence and duration of hypotensive events were significantly lower in the Further extension of machine learning algorithms in order to preserve intraoperative hemodynamic stability is application of closed loop systems that automatically administer fluids and vasopressors according to measured hemodynamic variables, which have reduced duration of intraoperative hypotension (< 65 mmHg) by 21.1%, as well as a reduction of intraoperative fluid balance (+ 1600 vs + 2050 ml) and lower serum lactate levels (1.2 vs 2.7 mmol/L) in patients undergoing moderate to high-risk surgery as demonstrated by Joosten et al. [It must be noted that most patients in both groups were graded as ASA 3, and although significant comorbidities are present (most of which are associated with smoking), careful multidisciplinary approach and preoperative assessment of cardiac and lung reserve are essential in decreasing the duration of mechanical ventilation, ICU and hospital stay, and improving patient outcomes in general [
PMC10061960
Study limitations
hypotension
SECONDARY, MYOCARDIAL INFARCTION
Since surgical manipulation (such as heart, lung or major vascular structure compression leading to preload drop) in thoracic procedures will result in rapid hemodynamic changes that cannot be predicted by an algorithm, there were concerns that use of HPI to guide hemodynamic optimization in such procedures is futile. However, obtained results show that even with that shortcoming, there were significant differences between groups in hypotension related outcomes.One other question that is usually raised in studies which rely on arterial waveform pulse contour analysis is their utility in patients with an open thorax and/or reduced left ventricular ejection fraction, as studied by Vetrugno et al. in patients that were monitored using second generation algorithm [In terms of the number of participants included in this study, sample size was determined for primary outcome measure, for which the null hypothesis was rejected. However, in terms of secondary clinical outcomes, the study is underpowered, and a much larger sample is needed to draw relevant conclusions. For example, while there was no significant difference in lactate levels between groups (0.9 vs 1.1 mmol/L, Secondly, we decided not to routinely check for serum troponin levels to determine whether the patient developed myocardial infarction, since elevated serum troponin is common after lung resection (up to 49% patients) [Also, during statistical analysis, some concerns were raised about the validity of the coin-toss method of randomization. However, since there were no significant differences in relevant baseline characteristics between groups, choice of randomization method did not result in sampling bias, as can be seen in Table
PMC10061960
Acknowledgements
N/A
PMC10061960
Authors’ contributions
JoP - data acquisition
AŠ - study concept and design, data acquisition, statistical analysis, manuscript preparation, ISJ - data acquisition, manuscript preparation, JM - data acquisition, HA - data acquisition, IB - data acquisition, JoP - data acquisition, JaP - final manuscript draft supervision. The author(s) read and approved the final manuscript.
PMC10061960
Funding
None received
PMC10061960
Availability of data and materials
The datasets generated and/or analyzed during the current study are not publicly available due to authors’ fear that they will be used without consent but are available from the corresponding author on reasonable request after the manuscript has been published.
PMC10061960
Declarations
PMC10061960
Ethics approval and consent to participate
Methods employed adhere to the relevant ethical guidelines and study protocol was approved by the University Hospital Dubrava ethics committee under ID 2022/1807–03 and ClinicalTrials.gov under ID NCT05615168.Informed consent was obtained in the anesthesia preparation room by all participants.
PMC10061960
Consent for publication
Not applicable.
PMC10061960
Competing interests
The authors declare no competing interests.
PMC10061960
References
PMC10061960
Key Points
PMC9982698
Question
Can natural language processing (NLP) be used to measure clinical trial outcomes?
PMC9982698
Findings
In this diagnostic study evaluating the performance, feasibility, and power implications of using deep-learning NLP to measure the outcome of documented goals-of-care discussions in a 2512-patient pragmatic trial, NLP-screened human abstraction measured the outcome with 92.6% sensitivity, substantial savings in abstractor-hours, and minimal loss of power, compared with manual abstraction.
PMC9982698
Meaning
The findings suggest that NLP may facilitate measurement of previously inaccessible outcomes in clinical trials and that incorporation of misclassification-adjusted power calculations into the design of studies using NLP may be beneficial.
PMC9982698
Importance
Many clinical trial outcomes are documented in free-text electronic health records (EHRs), making manual data collection costly and infeasible at scale. Natural language processing (NLP) is a promising approach for measuring such outcomes efficiently, but ignoring NLP-related misclassification may lead to underpowered studies.
PMC9982698
Objective
To evaluate the performance, feasibility, and power implications of using NLP to measure the primary outcome of EHR-documented goals-of-care discussions in a pragmatic randomized clinical trial of a communication intervention.
PMC9982698
Design, Setting, and Participants
This diagnostic study compared the performance, feasibility, and power implications of measuring EHR-documented goals-of-care discussions using 3 approaches: (1) deep-learning NLP, (2) NLP-screened human abstraction (manual verification of NLP-positive records), and (3) conventional manual abstraction. The study included hospitalized patients aged 55 years or older with serious illness enrolled between April 23, 2020, and March 26, 2021, in a pragmatic randomized clinical trial of a communication intervention in a multihospital US academic health system.
PMC9982698
Main Outcomes and Measures
Main outcomes were natural language processing performance characteristics, human abstractor-hours, and misclassification-adjusted statistical power of methods of measuring clinician-documented goals-of-care discussions. Performance of NLP was evaluated with receiver operating characteristic (ROC) curves and precision-recall (PR) analyses and examined the effects of misclassification on power using mathematical substitution and Monte Carlo simulation.
PMC9982698
Results
A total of 2512 trial participants (mean [SD] age, 71.7 [10.8] years; 1456 [58%] female) amassed 44 324 clinical notes during 30-day follow-up. In a validation sample of 159 participants, deep-learning NLP trained on a separate training data set identified patients with documented goals-of-care discussions with moderate accuracy (maximal F
PMC9982698
Conclusions and Relevance
In this diagnostic study, deep-learning NLP and NLP-screened human abstraction had favorable characteristics for measuring an EHR outcome at scale. Adjusted power calculations accurately quantified power loss from NLP-related misclassification, suggesting that incorporation of this approach into the design of studies using NLP would be beneficial.
PMC9982698
Introduction
Natural language processing (NLP) of free-text electronic health records (EHRs) presents rich opportunities for measuring outcomes that would otherwise require costly, laborious medical record abstraction.Researchers in the fields of palliative care and serious illness communication have shown interest in using NLPIn this study, we used deep-learning NLP to measure the primary outcome of EHR-documented goals-of-care discussions in a large pragmatic trial of a communication-priming intervention for hospitalized patients.
PMC9982698
Methods
chronic life-limiting illness, ADRD, Alzheimer disease, dementias
ALZHEIMER DISEASE
This diagnostic study was conducted to inform selection of an outcome measurement strategy for a pragmatic randomized clinical trial of a communication-priming intervention for hospitalized patients (Project to Improve Communication About Serious Illness—Hospital Study: Pragmatic Trial 1 [PICSI-H Trial 1]).The trial enrolled patients hospitalized at any of 3 study hospitals who either had advanced age (≥80 years) or were aged 55 years or older and had a chronic life-limiting illness as defined by diagnosis codes used by the Dartmouth Atlas Project to study end-of-life care (eTable 1 in During planning, the trial was specified to use NLP to measure its primary outcome, with human abstraction as a backup strategy. However, because NLP approaches were developed concurrently with enrollment, the expected degree of NLP-related misclassification was not known at the time the trial began. The trial was initially specified to target a sample size of 2000 participants, which would result in 80% power to detect a difference in proportions of at least 6.2% (assuming a control arm proportion of 0.54 and 2-sided α of 0.05). However, this sample size determination did not consider the potential effect of NLP-related misclassification. Development, training, and testing of NLP models continued throughout enrollment, using data sources gathered from outside the trial.Between April 23, 2020, and March 26, 2021, the trial enrolled 2512 patients. The prespecified enrollment target of 2000 was exceeded to increase the number of participants with Alzheimer disease and related dementias (ADRD), a prespecified subgroup. Following conclusion of enrollment and prior to unblinding and primary analyses, we froze our NLP program, evaluated NLP performance within a validation sample collected from the trial, and reevaluated the statistical power, human abstraction burden, and pragmatic implications of 3 strategies for measuring the primary outcome: (1) conventional manual abstraction, (2) NLP alone, and (3) NLP-screened human abstraction, in which only EHR passages scored by NLP above a predefined threshold would be reviewed by human abstractors for documented goals-of-care discussions.
PMC9982698
NLP Development and Model Training
We collected a training data set of 4642 EHR notes from 150 participants in a previous pilot trial of a similar patient- and clinician-facing communication-priming intervention at the same study hospitals (Project to Improve Communication About Serious Illness—Pilot Study [PICSI-P]) (eTable 1 in We used the training data set to train Bio+ClinicalBERT, a publicly available and freely distributed deep-learning NLP model, to predict the presence of documented goals-of-care discussions in EHR text (eMethods in
PMC9982698
Training, Prediction, and Validation of the Natural Language Processing (NLP) Model
EHR indicates electronic health record; GOC, goals of care; PICSI-H Trial 1, Project to Improve Communication About Serious Illness—Hospital Study: Pragmatic Trial (Trial 1)
PMC9982698
Trial Data Set and Validation Sample
Following conclusion of the PICSI-H Trial 1 outcome assessment period, we used automated database queries to collect all EHR notes authored by attending and trainee physicians, subintern medical students, nurse practitioners, and physician assistants between the date of randomization and 30 days thereafter. This yielded a PICSI-H Trial 1 data set of 44 324 notes from 2512 patients (
PMC9982698
Statistical Analysis
PMC9982698
Evaluating NLP Performance
ADRD
Following model training and manual abstraction of the validation sample, we used the trained BERT NLP model to predict the probability of documented goals-of-care discussions in all 2.64 million EHR passages (44 324 notes) of the PICSI-H Trial 1 data set. To characterize the expected performance of BERT NLP in the trial data set, we resampled the validation sample to reflect the prevalence of enrolled patients with ADRD (11%) and compared NLP-predicted probabilities for each note and patient against manual abstraction using ROC curves and precision-recall analyses. For all analyses, we report note- and patient-level performance, defining note- and patient-level probability as the maximum predicted probability for all constituent passages and defining the gold-standard label for each note and patient as the union of labels for all constituent passages. Statistical analyses were conducted using Stata/MP, version 17.0 (StataCorp LLC)
PMC9982698
Results
Between April 23, 2020, and March 26, 2021, PICSI-H Trial 1 enrolled 2512 patients, whose baseline characteristics are presented in eTable 1 in
PMC9982698
Abstraction and Composition of Training and Validation Data Sets
In manual abstraction of 4642 EHR notes in the training data set, 340 notes (7%; belonging to 34 of 150 patients [23%]) contained documented goals-of-care discussions (eTable 2 in In manual abstraction of 2840 EHR notes in the validation sample, 268 notes (9%; belonging to 54 of 159 patients [34%]) contained documented goals-of-care discussions (eTable 2 in
PMC9982698
BERT NLP Performance in the Validation Sample
In comparing BERT NLP predictions with manual abstraction in the validation sample, BERT NLP demonstrated note- and patient-level areas under the ROC curve of 0.962 and 0.924, respectively (
PMC9982698
Performance of BERT Natural Language Processing in Classifying 30-Day Documented Goals-of-Care Discussions at Note and Patient Levels in a 159-Patient, 2480-Note Internal Validation Sample
Representative values of sensitivity, specificity, positive and negative predictive values, and F1 scores from these curves are presented in the Table. B, Dotted lines indicate nondiscriminating classifiers. AUC indicates area under the receiver operating characteristic curve; AUPRC, area under the precision-recall curve; and PPV, positive predictive value.
PMC9982698
Performance Metrics for BERT NLP in Classifying 30-Day Documented Goals-of-Care Discussions at Note and Patient Levels in a 159-Patient, 2480-Note Internal Validation Sample
Abbreviations: AUC, area under the receiver operating characteristic curve; AUPRC, area under the precision-recall curve; NLP, natural language processing; NPV, negative predictive value; PPV, positive predictive value.Performance metrics are shown at observed discrimination thresholds with sensitivities closest to prespecified values of 70%, 80%, and 90%.
PMC9982698
Power Estimates With and Without Misclassification
In conventional power analysis (which assumes no misclassification), we calculated that the trial (N = 2512; 1:1 allocation; assumed
PMC9982698
Detectable Risk Difference Over Classifier Performance
Assumptions: n1 = 1256; n2 = 1256; p1 = 0.335; power = 0.8; and 2-sided α = .05. An interactive 3-D plot is available at
PMC9982698
Comparison of Outcome Measurement Strategies
fatigue
We evaluated 3 strategies for measuring the primary outcome: manual human abstraction, BERT NLP alone, and BERT NLP–screened human abstraction. The first approach, manual human abstraction, is the de facto gold standard and would power the study to detect a risk difference of 5.4%. Our experience suggested that individual abstractors could perform this abstraction task for up to 3 hours per day before experiencing excessive fatigue. To estimate the number of hours required for complete manual abstraction of the trial data set, we scaled the known abstractor-hours required to collect data for the validation sample to the entire trial data set, yielding an estimate of approximately 3000 abstractor-hours (ie, 67 work weeks for a team of 3 abstractors devoting 3 hours per day to this task; costing $195 000 at a rate of $65 per hour). Based on estimates from the validation sample, constraining abstractors to records from randomization to the first EHR-documented goals-of-care discussion (or 30 days if none was present) would reduce this estimate to approximately 2000 abstractor-hours (ie, 45 work weeks for a team of 3 abstractors, or $130 000 at a rate of $65 per hour).In the second approach, BERT NLP alone, the primary outcome could be measured with any combination of sensitivity and specificity represented on the ROC curve in In the third approach, BERT NLP–screened human abstraction, only EHR passages that were scored by NLP above a predefined threshold would be reviewed by human abstractors for documented goals-of-care discussions.
PMC9982698
Discussion
In this diagnostic study, we examined the novel use of deep-learning–based NLP to measure a complex outcome from unstructured EHR text in a large pragmatic clinical trial. We also demonstrated and validated the use of statistical methods to quantitatively assess the effects of NLP-related misclassification on study power at a given sample size.Natural language processing–screened human abstraction represents an efficient and useful approach for measuring EHR outcomes in large pragmatic studies and is increasingly used to measure palliative care outcomes similar to the one we examined.Researchers considering NLP or NLP-aided approaches to measuring outcomes should consider the effects of the outcome measurement strategy on the statistical power, costs, and validity of the trial. In this study, we demonstrated 2 simple methods for researchers to perform misclassification-adjusted power calculations, and we encourage the uptake of this approach in the design of future trials.Perhaps the most conspicuous limitation of using NLP to measure clinical research outcomes is the investment required to implement NLP. Although this investment may be minimal for outcomes that are easily detected using rule-based NLP, identifying more complex constructs such as the one investigated here may require substantial software development and acquisition of training and validation data. Our research group has spent more than 500 developer-hours implementing this NLP model, in addition to 491 abstractor-hours spent collecting training and validation data. We anticipate that the cost of implementing high-performing pretrained NLP models will decrease as the field matures. Additionally, many NLP development costs are fixed with respect to the number of study participants, and many aspects of NLP development are transferrable to other studies and research questions. Although our BERT NLP model is not yet portable due to the privacy risks associated with training on identifiable protected health information,
PMC9982698
Limitations
Our study has several important limitations. First, our model was trained and validated on data from a single health system, and its performance may not generalize to other systems. Future external validation and efforts to improve explainability of deep learning models
PMC9982698
Conclusions
In this diagnostic study evaluating the use of deep-learning NLP to measure EHR-documented goals-of-care discussions, we measured the primary outcome of a large pragmatic clinical trial using NLP-screened human abstraction with acceptable sensitivity and substantial savings in abstractor-hours. Our experience demonstrated that NLP may facilitate clinical research studies that would otherwise be infeasible due to the costs of manual medical record abstraction. Misclassification-adjusted power calculations quantified power loss from NLP-related misclassification, suggesting that incorporation of this approach into the design of future studies that use NLP to measure outcomes would be beneficial.
PMC9982698
2. Materials and Methods
PMC10456951
2.2. Functional Validation of the CPAP-AirFlife™ Flow Generator
MSA, HIT
This section provides a comprehensive overview of the tests conducted to validate the flow generator, a crucial component responsible for controlling flow and oxygen concentrations in the device. The testing process involved meticulously examining the flow generator valve performance, oxygen concentration, flow rates, and PEEP pressure parameters within a closed circuit with limited ventilation and a PEEP valve. A well-designed test rig was assembled to ensure accurate and reliable results, as depicted in During the testing procedure, various measurement instruments were utilized. The SIARGO MF5712 flowmeter (Siargo, Santa Clara, CA, USA) served as the primary tool for measuring flow rates and was cross-referenced with a trusted calibration standard, the TSI 4000 series Mass Flowmeter (TSI Incorporated, Shoreview, MN, USA). To assess oxygen concentration accurately, the digital oxygen measuring instrument CY-12C (CJCMALL, Wuxi, China) was employed and compared with the calibration standard MiniOX I—Model 473030 (MSA, Pittsburgh, PA, USA). The HIT-1890 differential pressure sensor (HIT, Nangang, Harbin, Heilongjiang, China) was used to measure and compare pressure values with the calibration standard TSI 4000 series Mass Flowmeter (TSI Incorporated, Shoreview, MN, USA).To ensure the integrity and effectiveness of the test setup, two HEPA-type antimicrobial filters, a vent hose, and a 10 cmHBefore initiating the characterization process, it is crucial to ensure that the flow generator valves, namely, valves A and B, are fully closed, as depicted in
PMC10456951
2.3. Clinical Trial Design
To ascertain the clinical efficacy of the device, a rigorous randomized non-inferiority and randomized crossover clinical trial was conducted at Servicios Especiales de Salud (S.E.S) Hospital Universitario de Caldas, situated in Manizales, Caldas, Colombia. The testing procedures took place within a specially adapted hospital environment to cater to COVID-19 patients. For comparative analysis of ventilation technologies, the S.E.S Hospital Universitario de Caldas provided the General Electric ventilator model R860 alongside the CPAP-AirFlife™ system.
PMC10456951
2.3.1. Outcomes
The main focus of this study was to assess the primary outcome variable, SpO
PMC10456951
2.3.2. Eligibility Criteria for Participants
PATHOLOGY
Inclusion criteria: healthy adult volunteers aged 18 years or older.Exclusion criteria: individuals with any form of respiratory pathology or other underlying comorbidities.
PMC10456951
2.3.3. Interventions
The study included a sample of 19 healthy adult volunteers, who were divided into two groups. Group 1 initiated with conventional CPAP (
PMC10456951
2.3.4. Sample Size
The objective of this experimental validation is not to establish the superiority of CPAP-AirFlife™ over a conventional CPAP delivery device (such as the General Electric model R860 ventilator) but rather to demonstrate that it is not inferior or of lower quality than the latter. Based on the null hypothesis Ho, which assumes that CPAP-AirFlife™ is inferior to conventional CPAP
PMC10456951
2.3.5. Randomization
The randomization process for determining the starting order of each CPAP technology (conventional vs. CPAP-AirFlife™) was conducted using a simple randomization method. This method ensured that each patient had an equal probability of beginning with either option. To maintain the concealment of randomization, sequentially numbered envelopes that were non-transparent were used. An external epidemiologist generated the sequence of randomization. The healthy adult volunteers who participated in the trial were enrolled in the study and kept masked until the assignment of the starting order for each CPAP technology was determined.
PMC10456951
2.4. Statistical Methods
For this study, continuous data such as age were reported as means and standard deviations following a normal distribution based on the Shapiro–Wilk test. These data were compared using Student’s Here,
PMC10456951
3. Results and Discussion
PMC10456951
4. Conclusions
The study findings demonstrate that the low-cost CPAP-AirFlife™ CPAP device incorporating aerosol control headgear is not inferior to conventional equipment across various efficacy measures. These measures include hemoglobin oxygen saturation levels, exhaled PCOOne significant advantage of the CPAP-AirFlife™ device is its accessibility, particularly in resource-limited areas and lower-level hospitals where the cost and availability of treatments present significant challenges. The device’s affordability and local availability can significantly improve access to effective respiratory treatments, particularly in rural regions. This has profound implications for enhancing the quality of life and overall health of individuals with respiratory conditions requiring specialized care.Moreover, the portability and ease of use of the CPAP-AirFlife™ device make it a suitable option for patients needing CPAP therapy during transportation to specialized medical facilities. Its independence from external sources of medical air or electricity, along with its ergonomic and user-friendly design, allow for straightforward operation without requiring specialized personnel. Consequently, the CPAP-AirFlife™ device emerges as an efficient and convenient alternative for providing respiratory support in emergencies and during patient transfers to other medical centers.In summary, this study underscores the CPAP-AirFlife™ device as a viable and effective solution for addressing respiratory problems, particularly in resource-constrained areas. Its affordability, accessibility, and portability make it a valuable tool in improving respiratory care and overcoming the challenges faced by patients in underserved regions. By bringing effective respiratory treatments within reach, the CPAP-AirFlife™ device has the potential to improve the well-being of individuals in need significantly.
PMC10456951
Author Contributions
Conceptualization, H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S. and O.J.-R.; methodology, H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S., O.J.-R. and O.D.A.-O.; software, H.A.T., L.P.-H., J.A.H.-C. and J.F.E.-S.; validation, H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S., O.J.-R. and O.D.A.-O.; formal analysis, H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S., O.J.-R. and O.D.A.-O.; investigation H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S., O.J.-R., O.D.A.-O., M.H.-H. and J.L.-G.; resources, H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S., O.J.-R., O.D.A.-O., M.H.-H. and J.L.-G.; writing—original draft preparation, H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S., O.J.-R., O.D.A.-O., M.H.-H. and J.L.-G.; writing—review and editing, H.A.T., L.P.-H., J.A.H.-C., J.F.E.-S., O.J.-R., O.D.A.-O., M.H.-H. and J.L.-G.; visualization, H.A.T., L.P.-H. and J.L.-G.; supervision, H.A.T., L.P.-H., O.J.-R. and O.D.A.-O.; project administration, H.A.T. and L.P.-H.; funding acquisition, H.A.T. and L.P.-H. All authors have read and agreed to the published version of the manuscript.
PMC10456951
Institutional Review Board Statement
The study was conducted in accordance with the Declaration of Helsinki, and approved by the bioethics Committee of UNIVERSIDAD AUTONOMA DE MANIZALES for the project 754-118 approved on 11 August 2020.
PMC10456951
Informed Consent Statement
Informed consent was obtained from all subjects involved in the study.
PMC10456951
Data Availability Statement
The data collected in this research are available when be required.
PMC10456951
Conflicts of Interest
The authors declare no conflict of interest.
PMC10456951
References
(Configuration for validation of the CPAP-AirFlife™ flow generator.CPAP-AirFlife™ System flow generator valves.Comparison between conventional (Variation of OClinical trial protocol for healthy patients.Variations in the flow and pressure parameters.Variations in Oxygen (OClinical and sociodemographic characteristics.* Fisher’s exact test ** Student’s SpOQualitative results for comfort.* Fisher’s exact test.
PMC10456951
Introduction
chronic stroke, stroke, death, Stroke
ADVERSE EFFECTS, STROKE, STROKE
Stroke is a leading cause of death and long-term disabilityRobot-assisted therapy is an effective intervention for promoting upper limb function in stroke survivors since robot-assisted therapy provides patients with intense, repetitive practice which is considered a key element for motor trainingMirror therapy is an easy-to-use and cost-effective intervention in neurorehabilitation and it has been shown to improve motor function of the upper limb in stroke survivorsThe goals of stroke rehabilitation are not only to improve motor function but also to help stroke patients regain independence and daily life participation. Limitation of activity participation may cause adverse effects on life satisfaction and affect quality of lifeThe purpose of this study was to examine whether mirror therapy would augment therapeutic effects of robot-assisted therapy on motor function, daily function, and self-efficacy in chronic stroke survivors. Mirror therapy was applied prior to robot-assisted therapy as a priming technique and sham mirror therapy with robot-assisted therapy was used as a control condition in this study. We hypothesized that sequential combination of mirror therapy and robot-assisted therapy would lead to greater improvements in the objective and subjective health-related outcomes than sham mirror therapy with robot-assisted therapy.
PMC10558527
Methods
PMC10558527
Study design and participants
post-stroke, spasticity, cognitive impairment, chronic stroke, stroke, peripheral polyneuropathy, dementia, neurological diseases, aphasia
STROKE, NEUROLOGICAL DISEASE
This study was a single-blinded, randomized controlled trial to investigate whether using mirror therapy as a priming strategy would augment therapeutic effects of robot-assisted therapy on motor function, daily function, and self-efficacy in chronic stroke survivors (ClinicalTrials.gov Identifier: NCT03917511, registered on 17/04/2019). The institutional review boards of Chang Gung Memorial Hospital approved the trials (IRB No. 201801025B0C603), and all participants provided written informed consent before participating. All methods were performed in accordance with relevant guidelines and regulations. The sample size of this study was estimated based on the systematic review and meta-analysis of robot-assisted therapy on upper limb recovery after stroke, which showed medium to large effect sizes measured by Fugl-Meyer AssessmentParticipants were recruited from medical centers in Taiwan, who attended for post-stroke rehabilitation between December 2018 and April 2021. The inclusion criteria included: (1) unilateral stroke ≥ 3 months prior to study enrollment; (2) Fugl-Meyer Assessment for Upper Extremity (FMA-UE) score < 60; (3) without excessive spasticity in any of the UE joint (modified Ashworth scale ≤ 3); (4) Mini Mental State Exam (MMSE) score > 24, indicating no serious cognitive impairment; and (5) between the ages of 20 and 75 years. The exclusion criteria included: (1) histories of other neurological diseases such as dementia and peripheral polyneuropathy; (2) difficulties in following and understanding instructions such as global aphasia; (3) enroll in other rehabilitation or drug studies simultaneously; (4) receiving Botulinum toxin injections within 3 months. The research design and flow process are shown in Fig. Flow diagram illustrating the flow of participants through each stage of the study.
PMC10558527
Intervention protocol
upper extremity impairment
Participants were stratified into four strata based on the lesion side (left and right) and the initial upper extremity impairment levels (Fugl-Meyer Assessment for Upper Extremity score < 35 and ≥ 36)
PMC10558527
Mirror therapy protocol
hand movements
A mirror was placed in the participants’ midsagittal plane to create a visual illusion of a paretic limb by using the mirror reflection of the non-paretic arm. The paretic arm was placed behind a mirror without being seen by the participant. During the 20-min mirror therapy, a robotic hand was attached to the paretic hand and provided continuous passive motion including 10-min grasp and release motion and 10-min pinch and release motion. Participant were instructed to look at the reflection of the non-paretic arm in the mirror, imagined it as the paretic arm and perform bilateral hand movements as symmetrically as possible. For the sham mirror priming group, participants underwent the same protocol except the mirror was covered with black fabric.
PMC10558527
Robot-assisted therapy protocol
hand movements
RECRUITMENT
The 40-min robot-assisted therapy consisted of 10-min active-assisted training and 30-min interactive training using the Hand of Hope (HOH) robotic hand system (Rehab-Robotics Co. Ltd, Hongkong, China). HOH is an exoskeleton type of robot with 2 surface electromyography (EMG) sensors that detect the level of motor unit recruitment. In active-assisted training, the robot provided participants with assistive movements that guided the fingers to complete grasp and release motion or pinch and release motion, once the EMG signal exceeds the predetermined threshold. In interactive training, the therapist selected 3 interactive games from the robot system and chose the level of difficulty based on participants' upper limb functional status. Participants were instructed to coordinate arm and hand movements to complete game missions. A detailed description of the robot was presented in a previous paper
PMC10558527
Outcome measurements
stroke
STROKE
We used clinical assessments to examine three domains of therapeutic effects of sequential combination of mirror therapy and robot-assisted therapy: (1) motor function, (2) independence in daily function, and (3) self-efficacy. Clinical assessments included Fugl-Meyer Assessment for Upper Extremity (FMA-UE), Wolf Motor Function Test (WMFT), Nottingham Extended Activities of Daily Living Scale (NEADL), the stroke self-efficacy questionnaires (SSEQ) and Daily Living Self-Efficacy Scale (DLSES). Participants were assessed within 1 week before the intervention (baseline assessment), and after the 18-session intervention (post-assessment). All participants were assessed by a certified occupational therapist who was unaware of the group to which the participant had been allocated.
PMC10558527
Domain of motor function
upper extremity motor ability, stroke
STROKE
Fugl-Meyer Assessment for Upper Extremity (FMA-UE): The FMA-UE includes 33 items assessing movements, reflexes, and coordination of upper limbs. Each item is measured on a 3-point ordinal scale and the total score ranges from 0 to 66Wolf Motor Function Test (WMFT): The WMFT assesses upper extremity motor ability by measuring the performance time (WMFT-Time) and functional ability rating scale (WMFT-FAS) in required task. Participants were timed and rated by using a 6-point ordinal scale. The WMFT is valid and reliable in assessing motor function in stroke patients
PMC10558527
Domain of independence in daily function
Nottingham Extended Activities of Daily Living Scale (NEADL): The NEADL is a measure of independence in 4 areas of daily life, including mobility, kitchen, domestic, and leisure activities. It includes 22 items, and each item is measured on a 4-point scale. The total score ranges from 0 to 66 and a higher score indicates better daily functional ability. The psychometric properties of the NEADL have been well established
PMC10558527
Domain of self-efficacy
stroke
STROKE
The stroke self-efficacy questionnaires (SSEQ): The SSEQ measures an individual's confidence in relation to functional performance and self-management after stroke. It includes 13 items, and each item is rated on a 10-point scale from 0 (not at all confident) to 10 (very confident). The reliability and validity of the SSEQ are well establishedDaily Living Self-Efficacy Scale (DLSES): The DLSES measures self-efficacy of daily functioning, including psychosocial functioning and activities of daily living. The scale consists of 12 items, and each item is measured on a 100-point scale with 10-unit intervals (0 = cannot do at all, 100 = highly certain can do). A higher score indicates higher level of self-efficacy. The DLSES is a psychometrically sound measure of self-efficacy in stroke survivors
PMC10558527
Statistical analysis
The
PMC10558527
Results
PMC10558527
Daily function
Results of NEADL showed no statistically significant interaction effect between group and intervention (
PMC10558527
Self-efficacy
Results of SSEQ showed no statistically significant interaction effect between group and intervention (Results of DLSES showed no statistically significant interaction effect between group and intervention (
PMC10558527
Discussion
post-stroke, chronic stroke, stroke, self-efficacy post-stroke, left hemisphere damage, hand movements, hemispheric lesions, right hemisphere damage, MR
STROKE
In this study, we used a randomized controlled trial to examine whether mirror therapy would augment therapeutic effects of robot-assisted therapy on motor function, daily function, and self-efficacy in chronic stroke survivors. Based on the timeline of stroke recovery framework proposed by SRRR group, chronic stroke is defined as 6 months after stroke onset and 3 to 6 months post-stroke refers to as late sub-acute phaseIn motor function domain, results of FMA-UE and WMFT-FAS demonstrated the positive effects of robot-assisted therapy on upper limb motor recovery and functional performance, and the effect size of intervention was large based on benchmarks suggested by CohenIn contrast to improvements in FMA-UE and WMFT-FAS scores, participants did not improve their performance time for completing the tasks after the interventions in both groups based on the results of WMFT-Time scores. One explanation is that movement speed was not the primary focus of our intervention in this study. Our robot-assisted therapy consisted of 10-min active-assisted training and 30-min interactive training. The primary focus of active-assisted training was to assist patients in precisely recruiting the desired muscle group and enhance muscle activation during functional movements. Additionally, interactive training emphasized on coordination of arm and hand movements to improve endpoint accuracy during functional tasks. Participants were not asked to move in a fast pace during training and therefore the speed of task performance did not significantly improve. While temporal efficiency is identified as an aspect of movement qualityOur results in motor function domain seemed to conflict with the findings from Rong et al.'s study, which showed mirror therapy could augment therapeutic effects of robot-assisted therapy in motor recovery measured by FMA-UEIn daily function domain, our findings showed that participants successfully transferred gains in upper limb motor function to daily functional ability after robot-assisted interventions, measured by NEADL. Regarding our robot-assisted intervention protocols, active-assisted training mainly focused on neuromuscular control of hand movement while interactive training required coordinated motions for arm and hand. Integration of distal and proximal upper limb training has been advocated as a key for enhancing functional gainsLimited research has been conducted to identify stroke survivors' self-efficacy after receiving robot-assisted therapy using standardized scales. We used two standardized scales to measure self-efficacy since they capture different aspects of self-efficacy in stroke population. Whereas SSEQ was developed to measure one's confidence in relation to functional performance following stroke, DLSES captures self-efficacy in a broader sense including psychosocial functioning and activities of daily living. Results of both measures suggest that applying mirror therapy prior to robot-assisted therapy could be advantageous for enhancing self-efficacy post-stroke. Applying mirror therapy prior to robot-assisted therapy led to significant improvements in stroke survivors' self-efficacy based on the results of SSEQ and DLSES scores. According to Bandura's self-efficacy theory, self-efficacy can be developed by four main sources of influence, including mastery experience , vicarious experience, verbal persuasion, and emotional arousalThis study has some limitations. First, there were more participants with left hemisphere damage in MR group and more participants with right hemisphere damage in SMR group. Although there were no statistically significant differences in participants' demographic and clinical characteristics between the MR and SMR groups, side of hemispheric lesions could affect stroke rehabilitation in upper limb training
PMC10558527