Kuan-Yuan Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Kuan-Yuan Wang has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
- Dae Hyun Kim
Marcus Institute for Aging Research, Hebrew SeniorLife, Boston, Massachusetts
4 shared publications - Cassandra York
Brigham and Women’s Hospital, Boston, MA, USA
2 shared publications - Heidi Zakoul
Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine Brigham and Women's Hospital and Harvard Medical School Boston Massachusetts USA
2 shared publications - John Novoa-Laurentiev
Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA
2 shared publications - Kueiyu Joshua Lin
Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital, Boston, MA, USA
2 shared publications - Li Zhou
Division of General Internal Medicine and Primary Care, Department of Medicine, Brigham and Women's Hospital, Boston, MA, United States
2 shared publications
Latest Publications
- 7 September 2025
- Published by Wiley in Journal of the American Geriatrics Society
- Vol. 73 (11) , 3387-3395
Multi‐Component Prehabilitation Program for Older Adults Undergoing Major Elective Surgery: A Pilot and Feasibility Study
Background: Prehabilitation may help older adults recover after surgery, yet adherence has been variable. We assessed the feasibility of a multi‐component prehabilitation program. Methods: This single‐arm trial was conducted at an academic medical center to test the feasibility of an individualized prehabilitation program before major surgery. The program consisted of twice‐weekly physical therapy, weekly dietician consultations with daily protein supplementation, and weekly group meditation for 3–4 weeks. The primary outcome was adherence, as measured by the proportion of completed sessions. Secondary outcomes included changes in gait speed, chair stands, and grip strength after intervention and Patient‐Reported Outcomes Measurement Information System (PROMIS)‐Physical Function scores at 90 days post‐surgery. Results: Of 95 eligible patients, 30 (31.6%) were enrolled (mean [SD] age, 75.3 [6.6] years; 12 [40.0%] women; 19 cardiac and 11 non‐cardiac procedures) and 28 began the intervention. Adherence was 91.3% (190/208 sessions completed) for physical therapy, 76.9% (80/104 sessions completed) for dietician consultation, and 68.3% (71/104 sessions completed) for meditation, with 67.9% of the participants completing at least 50% of each component. Among 25 participants (89.3%) who completed the intervention, mean walking speed increased (0.2 m/s; 95% CI: 0.0–0.4), and mean time to complete 5 chair stands decreased (3.4 s; 95% CI: −6.9 to 0.0), but mean grip strength did not change significantly (1.6 kg; 95% CI: −2.1 to 5.3). Among 22 out of 23 participants (95.7%) who underwent surgery, the mean PROMIS‐Physical Function score improved by 4.4 points (95% CI: 3.0–5.9) at 90 days post‐surgery compared to baseline. Adverse events included gastrointestinal discomforts (3 events), musculoskeletal pain (3 events), and dizziness (1 event). Conclusions: This study demonstrates the feasibility of multicomponent prehabilitation in older patients scheduled for major elective surgery. Trial Registration: NCT05752474
- 7 September 2025
- Published by Wiley in Journal of the American Geriatrics Society
- Vol. 73 (11) , 3387-3395
Multi‐Component Prehabilitation Program for Older Adults Undergoing Major Elective Surgery: A Pilot and Feasibility Study
Background: Prehabilitation may help older adults recover after surgery, yet adherence has been variable. We assessed the feasibility of a multi‐component prehabilitation program. Methods: This single‐arm trial was conducted at an academic medical center to test the feasibility of an individualized prehabilitation program before major surgery. The program consisted of twice‐weekly physical therapy, weekly dietician consultations with daily protein supplementation, and weekly group meditation for 3–4 weeks. The primary outcome was adherence, as measured by the proportion of completed sessions. Secondary outcomes included changes in gait speed, chair stands, and grip strength after intervention and Patient‐Reported Outcomes Measurement Information System (PROMIS)‐Physical Function scores at 90 days post‐surgery. Results: Of 95 eligible patients, 30 (31.6%) were enrolled (mean [SD] age, 75.3 [6.6] years; 12 [40.0%] women; 19 cardiac and 11 non‐cardiac procedures) and 28 began the intervention. Adherence was 91.3% (190/208 sessions completed) for physical therapy, 76.9% (80/104 sessions completed) for dietician consultation, and 68.3% (71/104 sessions completed) for meditation, with 67.9% of the participants completing at least 50% of each component. Among 25 participants (89.3%) who completed the intervention, mean walking speed increased (0.2 m/s; 95% CI: 0.0–0.4), and mean time to complete 5 chair stands decreased (3.4 s; 95% CI: −6.9 to 0.0), but mean grip strength did not change significantly (1.6 kg; 95% CI: −2.1 to 5.3). Among 22 out of 23 participants (95.7%) who underwent surgery, the mean PROMIS‐Physical Function score improved by 4.4 points (95% CI: 3.0–5.9) at 90 days post‐surgery compared to baseline. Adverse events included gastrointestinal discomforts (3 events), musculoskeletal pain (3 events), and dizziness (1 event). Conclusions: This study demonstrates the feasibility of multicomponent prehabilitation in older patients scheduled for major elective surgery. Trial Registration: NCT05752474
- 2 April 2025
- Published by Wiley in Journal of the American Geriatrics Society
- Vol. 73 (6) , 1836-1846
Home‐Based Rehabilitation After Transcatheter Aortic Valve Replacement ( REHAB ‐ TAVR ): A Pilot Randomized Controlled Trial
Background: The benefit of early cardiac rehabilitation after transcatheter aortic valve replacement (TAVR) is not well established. This pilot study evaluated the feasibility and short‐term effects of a home‐based exercise program, with or without cognitive‐behavioral intervention (CBI). Methods: We randomized 51 patients (mean age, 83.9 years; 19 women) to a home‐based exercise program with CBI (Group A; n = 18) or without CBI (Group B; n = 15), or telephone‐based education control (Group C; n = 18). The exercise program focusing on balance, flexibility, strength, and endurance began within 7 days post‐discharge and was delivered once weekly by a physical therapist for 8 weeks. CBI included discussions on exercise benefits and barriers, goal setting, detailed exercise planning, and a weekly cash adherence incentive. The primary outcome was a disability score (range: 0–22; higher scores indicate greater disability) at 8 weeks. Secondary outcomes included the Short Physical Performance Battery (SPPB) (range: 0–12; higher scores indicate better function), self‐efficacy, and outcome expectation scores. Feasibility outcomes included adherence and drop‐out rates. Results: Fifteen participants (83.3%) in Group A, 10 (58.8%) in Group B, and 10 (52.6%) in Group C completed ≥ 5 of the eight assigned weekly sessions ( p = 0.196). Two participants in each group were lost to follow‐up. At 8 weeks, the home‐based exercise groups (Group A and B combined) demonstrated lower disability scores (mean [SE]: 2.6 [0.3] vs. 4.5 [0.5]; p = 0.042) and higher SPPB scores (9.5 [0.6] vs. 6.5 [0.8]; p = 0.003) compared with the education group (Group C). Group A had lower disability scores than Group B (2.1 [0.4] vs. 3.4 [0.5]; p = 0.047), with no differences in self‐efficacy and outcome expectation scores. Conclusions: An early, home‐based, multi‐domain exercise program appears feasible and may prevent disability and improve physical function in older adults after TAVR. Adding CBI, including a modest cash incentive, showed trends toward improved adherence and reduced disability. Trial Registration: NCT02805309
- 2 April 2025
- Published by Wiley in Journal of the American Geriatrics Society
- Vol. 73 (6) , 1836-1846
Home‐Based Rehabilitation After Transcatheter Aortic Valve Replacement ( REHAB ‐ TAVR ): A Pilot Randomized Controlled Trial
Background: The benefit of early cardiac rehabilitation after transcatheter aortic valve replacement (TAVR) is not well established. This pilot study evaluated the feasibility and short‐term effects of a home‐based exercise program, with or without cognitive‐behavioral intervention (CBI). Methods: We randomized 51 patients (mean age, 83.9 years; 19 women) to a home‐based exercise program with CBI (Group A; n = 18) or without CBI (Group B; n = 15), or telephone‐based education control (Group C; n = 18). The exercise program focusing on balance, flexibility, strength, and endurance began within 7 days post‐discharge and was delivered once weekly by a physical therapist for 8 weeks. CBI included discussions on exercise benefits and barriers, goal setting, detailed exercise planning, and a weekly cash adherence incentive. The primary outcome was a disability score (range: 0–22; higher scores indicate greater disability) at 8 weeks. Secondary outcomes included the Short Physical Performance Battery (SPPB) (range: 0–12; higher scores indicate better function), self‐efficacy, and outcome expectation scores. Feasibility outcomes included adherence and drop‐out rates. Results: Fifteen participants (83.3%) in Group A, 10 (58.8%) in Group B, and 10 (52.6%) in Group C completed ≥ 5 of the eight assigned weekly sessions ( p = 0.196). Two participants in each group were lost to follow‐up. At 8 weeks, the home‐based exercise groups (Group A and B combined) demonstrated lower disability scores (mean [SE]: 2.6 [0.3] vs. 4.5 [0.5]; p = 0.042) and higher SPPB scores (9.5 [0.6] vs. 6.5 [0.8]; p = 0.003) compared with the education group (Group C). Group A had lower disability scores than Group B (2.1 [0.4] vs. 3.4 [0.5]; p = 0.047), with no differences in self‐efficacy and outcome expectation scores. Conclusions: An early, home‐based, multi‐domain exercise program appears feasible and may prevent disability and improve physical function in older adults after TAVR. Adding CBI, including a modest cash incentive, showed trends toward improved adherence and reduced disability. Trial Registration: NCT02805309
- Open Access
- 13 February 2024
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 26, e47739
Identifying Functional Status Impairment in People Living With Dementia Through Natural Language Processing of Clinical Documents: Cross-Sectional Study
Background: Assessment of activities of daily living (ADLs) and instrumental ADLs (iADLs) is key to determining the severity of dementia and care needs among older adults. However, such information is often only documented in free-text clinical notes within the electronic health record and can be challenging to find. Objective: This study aims to develop and validate machine learning models to determine the status of ADL and iADL impairments based on clinical notes. Methods: This cross-sectional study leveraged electronic health record clinical notes from Mass General Brigham’s Research Patient Data Repository linked with Medicare fee-for-service claims data from 2007 to 2017 to identify individuals aged 65 years or older with at least 1 diagnosis of dementia. Notes for encounters both 180 days before and after the first date of dementia diagnosis were randomly sampled. Models were trained and validated using note sentences filtered by expert-curated keywords (filtered cohort) and further evaluated using unfiltered sentences (unfiltered cohort). The model’s performance was compared using area under the receiver operating characteristic curve and area under the precision-recall curve (AUPRC). Results: The study included 10,000 key-term–filtered sentences representing 441 people (n=283, 64.2% women; mean age 82.7, SD 7.9 years) and 1000 unfiltered sentences representing 80 people (n=56, 70% women; mean age 82.8, SD 7.5 years). Area under the receiver operating characteristic curve was high for the best-performing ADL and iADL models on both cohorts (>0.97). For ADL impairment identification, the random forest model achieved the best AUPRC (0.89, 95% CI 0.86-0.91) on the filtered cohort; the support vector machine model achieved the highest AUPRC (0.82, 95% CI 0.75-0.89) for the unfiltered cohort. For iADL impairment, the Bio+Clinical bidirectional encoder representations from transformers (BERT) model had the highest AUPRC (filtered: 0.76, 95% CI 0.68-0.82; unfiltered: 0.58, 95% CI 0.001-1.0). Compared with a keyword-search approach on the unfiltered cohort, machine learning reduced false-positive rates from 4.5% to 0.2% for ADL and 1.8% to 0.1% for iADL. Conclusions: In this study, we demonstrated the ability of machine learning models to accurately identify ADL and iADL impairment based on free-text clinical notes, which could be useful in determining the severity of dementia.
- 13 February 2024
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 26, e47739
Open AccessIdentifying Functional Status Impairment in People Living With Dementia Through Natural Language Processing of Clinical Documents: Cross-Sectional Study
Background: Assessment of activities of daily living (ADLs) and instrumental ADLs (iADLs) is key to determining the severity of dementia and care needs among older adults. However, such information is often only documented in free-text clinical notes within the electronic health record and can be challenging to find. Objective: This study aims to develop and validate machine learning models to determine the status of ADL and iADL impairments based on clinical notes. Methods: This cross-sectional study leveraged electronic health record clinical notes from Mass General Brigham’s Research Patient Data Repository linked with Medicare fee-for-service claims data from 2007 to 2017 to identify individuals aged 65 years or older with at least 1 diagnosis of dementia. Notes for encounters both 180 days before and after the first date of dementia diagnosis were randomly sampled. Models were trained and validated using note sentences filtered by expert-curated keywords (filtered cohort) and further evaluated using unfiltered sentences (unfiltered cohort). The model’s performance was compared using area under the receiver operating characteristic curve and area under the precision-recall curve (AUPRC). Results: The study included 10,000 key-term–filtered sentences representing 441 people (n=283, 64.2% women; mean age 82.7, SD 7.9 years) and 1000 unfiltered sentences representing 80 people (n=56, 70% women; mean age 82.8, SD 7.5 years). Area under the receiver operating characteristic curve was high for the best-performing ADL and iADL models on both cohorts (>0.97). For ADL impairment identification, the random forest model achieved the best AUPRC (0.89, 95% CI 0.86-0.91) on the filtered cohort; the support vector machine model achieved the highest AUPRC (0.82, 95% CI 0.75-0.89) for the unfiltered cohort. For iADL impairment, the Bio+Clinical bidirectional encoder representations from transformers (BERT) model had the highest AUPRC (filtered: 0.76, 95% CI 0.68-0.82; unfiltered: 0.58, 95% CI 0.001-1.0). Compared with a keyword-search approach on the unfiltered cohort, machine learning reduced false-positive rates from 4.5% to 0.2% for ADL and 1.8% to 0.1% for iADL. Conclusions: In this study, we demonstrated the ability of machine learning models to accurately identify ADL and iADL impairment based on free-text clinical notes, which could be useful in determining the severity of dementia.
- Open Access
- 30 March 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Identifying Functional Status Impairment in People Living With Dementia Through Natural Language Processing of Clinical Documents: Cross-Sectional Study (Preprint)
BACKGROUND Assessment of activities of daily living (ADLs) and instrumental ADLs (iADLs) is key to determining the severity of dementia and care needs among older adults. However, such information is often only documented in free-text clinical notes within the electronic health record and can be challenging to find. OBJECTIVE This study aims to develop and validate machine learning models to determine the status of ADL and iADL impairments based on clinical notes. METHODS This cross-sectional study leveraged electronic health record clinical notes from Mass General Brigham’s Research Patient Data Repository linked with Medicare fee-for-service claims data from 2007 to 2017 to identify individuals aged 65 years or older with at least 1 diagnosis of dementia. Notes for encounters both 180 days before and after the first date of dementia diagnosis were randomly sampled. Models were trained and validated using note sentences filtered by expert-curated keywords (filtered cohort) and further evaluated using unfiltered sentences (unfiltered cohort). The model’s performance was compared using area under the receiver operating characteristic curve and area under the precision-recall curve (AUPRC). RESULTS The study included 10,000 key-term–filtered sentences representing 441 people (n=283, 64.2% women; mean age 82.7, SD 7.9 years) and 1000 unfiltered sentences representing 80 people (n=56, 70% women; mean age 82.8, SD 7.5 years). Area under the receiver operating characteristic curve was high for the best-performing ADL and iADL models on both cohorts (>0.97). For ADL impairment identification, the random forest model achieved the best AUPRC (0.89, 95% CI 0.86-0.91) on the filtered cohort; the support vector machine model achieved the highest AUPRC (0.82, 95% CI 0.75-0.89) for the unfiltered cohort. For iADL impairment, the Bio+Clinical bidirectional encoder representations from transformers (BERT) model had the highest AUPRC (filtered: 0.76, 95% CI 0.68-0.82; unfiltered: 0.58, 95% CI 0.001-1.0). Compared with a keyword-search approach on the unfiltered cohort, machine learning reduced false-positive rates from 4.5% to 0.2% for ADL and 1.8% to 0.1% for iADL. CONCLUSIONS In this study, we demonstrated the ability of machine learning models to accurately identify ADL and iADL impairment based on free-text clinical notes, which could be useful in determining the severity of dementia.
- 30 March 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessIdentifying Functional Status Impairment in People Living With Dementia Through Natural Language Processing of Clinical Documents: Cross-Sectional Study (Preprint)
BACKGROUND Assessment of activities of daily living (ADLs) and instrumental ADLs (iADLs) is key to determining the severity of dementia and care needs among older adults. However, such information is often only documented in free-text clinical notes within the electronic health record and can be challenging to find. OBJECTIVE This study aims to develop and validate machine learning models to determine the status of ADL and iADL impairments based on clinical notes. METHODS This cross-sectional study leveraged electronic health record clinical notes from Mass General Brigham’s Research Patient Data Repository linked with Medicare fee-for-service claims data from 2007 to 2017 to identify individuals aged 65 years or older with at least 1 diagnosis of dementia. Notes for encounters both 180 days before and after the first date of dementia diagnosis were randomly sampled. Models were trained and validated using note sentences filtered by expert-curated keywords (filtered cohort) and further evaluated using unfiltered sentences (unfiltered cohort). The model’s performance was compared using area under the receiver operating characteristic curve and area under the precision-recall curve (AUPRC). RESULTS The study included 10,000 key-term–filtered sentences representing 441 people (n=283, 64.2% women; mean age 82.7, SD 7.9 years) and 1000 unfiltered sentences representing 80 people (n=56, 70% women; mean age 82.8, SD 7.5 years). Area under the receiver operating characteristic curve was high for the best-performing ADL and iADL models on both cohorts (>0.97). For ADL impairment identification, the random forest model achieved the best AUPRC (0.89, 95% CI 0.86-0.91) on the filtered cohort; the support vector machine model achieved the highest AUPRC (0.82, 95% CI 0.75-0.89) for the unfiltered cohort. For iADL impairment, the Bio+Clinical bidirectional encoder representations from transformers (BERT) model had the highest AUPRC (filtered: 0.76, 95% CI 0.68-0.82; unfiltered: 0.58, 95% CI 0.001-1.0). Compared with a keyword-search approach on the unfiltered cohort, machine learning reduced false-positive rates from 4.5% to 0.2% for ADL and 1.8% to 0.1% for iADL. CONCLUSIONS In this study, we demonstrated the ability of machine learning models to accurately identify ADL and iADL impairment based on free-text clinical notes, which could be useful in determining the severity of dementia.