Natalie Parde
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Natalie Parde has a total of 134 co-authors. Here are the top co-authors by number of shared publications:
- Mary A. Khetani
Department of Occupational Therapy, University of Illinois Chicago, Chicago, USA
7 shared publications Rodney Nielsen
4 shared publications- Skye Barbic
University of British Columbia
4 shared publications - Xiaoxu Ding
Department of Occupational Science & Occupational Therapy, Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada
4 shared publications - Brodie M. Sakakibara
University of British Columbia
3 shared publications Gayatri Mehta
University of North Texas
3 shared publications
Latest Publications
- Open Access
- 18 November 2026
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 26 (1)
Usability of an artificial intelligence enabled Participation and Environment Measure prototype for family-centered goal setting support in pediatric rehabilitation
- 18 November 2026
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 26 (1)
Open AccessUsability of an artificial intelligence enabled Participation and Environment Measure prototype for family-centered goal setting support in pediatric rehabilitation
- 13 October 2026
- Published by Springer Nature
- p. 561-581
A Bayesian Correlated Equilibrium for Early Insider-Threat Detection
- 13 October 2026
- Published by Springer Nature
- p. 561-581
A Bayesian Correlated Equilibrium for Early Insider-Threat Detection
- Open Access
- 2 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-3
Designing for Children’s Participation in Everyday Life: Bridging HCI and Occupational Therapy through Roles and Contexts
- 2 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-3
Open AccessDesigning for Children’s Participation in Everyday Life: Bridging HCI and Occupational Therapy through Roles and Contexts
- Open Access
- 4 May 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Multilingual Natural Language Processing for Alzheimer’s Disease Detection: A Scoping Review of Datasets, Modeling Approaches, and Research Opportunities (Preprint)
BACKGROUND Spoken language analysis is a compelling tool for detecting signs of Alzheimer’s disease (AD). However, most language-based AD detection resources are only available in English, leaving multilingual or crosslingual AD detection understudied. OBJECTIVE We review the current state of the field with respect to this topic, compiling recent approaches, constraints, and potential solutions from original work published across engineering, natural language processing, and medical databases from 2004 to the present. METHODS From the 776 search results, we identified 43 articles meeting predefined eligibility criteria and summarized their findings. RESULTS Promising results are reported in almost all studies, but few are integrated into clinical practice. The main limitations of the field are poor standardization, the lack of benchmark data repositories (which in turn hinders direct results comparison), and some disconnect between the study goals and the clinical applications. CONCLUSIONS Active efforts from the research community and industry to close these gaps would support robust integration of research across languages into clinical practice.
- 4 May 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessMultilingual Natural Language Processing for Alzheimer’s Disease Detection: A Scoping Review of Datasets, Modeling Approaches, and Research Opportunities (Preprint)
BACKGROUND Spoken language analysis is a compelling tool for detecting signs of Alzheimer’s disease (AD). However, most language-based AD detection resources are only available in English, leaving multilingual or crosslingual AD detection understudied. OBJECTIVE We review the current state of the field with respect to this topic, compiling recent approaches, constraints, and potential solutions from original work published across engineering, natural language processing, and medical databases from 2004 to the present. METHODS From the 776 search results, we identified 43 articles meeting predefined eligibility criteria and summarized their findings. RESULTS Promising results are reported in almost all studies, but few are integrated into clinical practice. The main limitations of the field are poor standardization, the lack of benchmark data repositories (which in turn hinders direct results comparison), and some disconnect between the study goals and the clinical applications. CONCLUSIONS Active efforts from the research community and industry to close these gaps would support robust integration of research across languages into clinical practice.
- Open Access
Context-Aware Counterfactual Data Augmentation for Gender Bias Mitigation in Language Models
A challenge in mitigating social bias in fine-tuned language models (LMs) is the potential reduction in language modeling capability, which can harm downstream performance. Counterfactual data augmentation (CDA), a widely used method for fine-tuning, highlights this issue by generating synthetic data that may align poorly with real-world distributions or creating overly simplistic counterfactuals that ignore the social context of altered sensitive attributes (e.g., gender) in the pretraining corpus. To address these limitations, we propose a simple yet effective context-augmented CDA method, Context-CDA, which uses large LMs to enhance the diversity and contextual relevance of the debiasing corpus. By minimizing discrepancies between the debiasing corpus and pretraining data through augmented context, this approach ensures better alignment, enhancing language modeling capability. We then employ uncertainty-based filtering to exclude generated counterfactuals considered low-quality by the target smaller LMs (i.e., LMs to be debiased), further improving the fine-tuning corpus quality. Experimental results on gender bias benchmarks demonstrate that Context-CDA effectively mitigates bias without sacrificing language modeling performance while offering insights into social biases by analyzing distribution shifts in next-token generation probabilities.
Open AccessContext-Aware Counterfactual Data Augmentation for Gender Bias Mitigation in Language Models
A challenge in mitigating social bias in fine-tuned language models (LMs) is the potential reduction in language modeling capability, which can harm downstream performance. Counterfactual data augmentation (CDA), a widely used method for fine-tuning, highlights this issue by generating synthetic data that may align poorly with real-world distributions or creating overly simplistic counterfactuals that ignore the social context of altered sensitive attributes (e.g., gender) in the pretraining corpus. To address these limitations, we propose a simple yet effective context-augmented CDA method, Context-CDA, which uses large LMs to enhance the diversity and contextual relevance of the debiasing corpus. By minimizing discrepancies between the debiasing corpus and pretraining data through augmented context, this approach ensures better alignment, enhancing language modeling capability. We then employ uncertainty-based filtering to exclude generated counterfactuals considered low-quality by the target smaller LMs (i.e., LMs to be debiased), further improving the fine-tuning corpus quality. Experimental results on gender bias benchmarks demonstrate that Context-CDA effectively mitigates bias without sacrificing language modeling performance while offering insights into social biases by analyzing distribution shifts in next-token generation probabilities.
- Open Access
Machine learning models to detect opioid misuse in Emergency Department patients at triage
Objective Emergency department (ED) encounters represent valuable opportunities to initiate evidence-based treatments for patients with opioid misuse, but few receive such care. Universal manual screening has been proposed to improve patient identification but is uncommon due to its time and resource-intensive nature. We sought to determine the feasibility of identifying patients with opioid misuse at the time of ED triage using machine learning (ML). Methods We conducted a retrospective cohort study of 1,123 ED encounters (September 2020 – March 2023) at a tertiary hospital. Encounters were enriched for opioid misuse, manually annotated, and chronologically split for training, validation, and testing. Candidate triage-time features included patient demographics, Emergency Severity Index, arrival time of day, chief complaint, comorbidities, and chronic medications. Model performance was evaluated using F1 score, area under the precision–recall curve (AUPRC), accuracy, recall, and AUROC. Post-hoc explainability analyses included SHapley Additive exPlanations (SHAP) and feature importance. Results All models performed comparably to opioid-related diagnosis codes placed at any time during the encounter. Random Forest (F1=0.75 [95%CI 0.70-0.83], AUPRC=0.88 [0.81-0.93], accuracy=0.79 [0.70-0.83]) and Gradient Boosting (F1=0.77 [0.71-0.82], AUPRC=0.89 [0.85-0.93], accuracy=0.81 [0.720.84]) had among the highest F1 score and AUPRC but confidence intervals overlapped with other methods. Explainability analyses highlighted prior drug-use diagnosis codes, triage acuity, and age as top predictors. Conclusion ML classifiers leveraging routinely collected triage data offer a feasible alternative to manual screening in flagging opioid misuse before physician evaluation, potentially enabling early harm-reduction interventions. Prospective multi-site validation, calibration, and bias assessments are warranted.
Open AccessMachine learning models to detect opioid misuse in Emergency Department patients at triage
Objective Emergency department (ED) encounters represent valuable opportunities to initiate evidence-based treatments for patients with opioid misuse, but few receive such care. Universal manual screening has been proposed to improve patient identification but is uncommon due to its time and resource-intensive nature. We sought to determine the feasibility of identifying patients with opioid misuse at the time of ED triage using machine learning (ML). Methods We conducted a retrospective cohort study of 1,123 ED encounters (September 2020 – March 2023) at a tertiary hospital. Encounters were enriched for opioid misuse, manually annotated, and chronologically split for training, validation, and testing. Candidate triage-time features included patient demographics, Emergency Severity Index, arrival time of day, chief complaint, comorbidities, and chronic medications. Model performance was evaluated using F1 score, area under the precision–recall curve (AUPRC), accuracy, recall, and AUROC. Post-hoc explainability analyses included SHapley Additive exPlanations (SHAP) and feature importance. Results All models performed comparably to opioid-related diagnosis codes placed at any time during the encounter. Random Forest (F1=0.75 [95%CI 0.70-0.83], AUPRC=0.88 [0.81-0.93], accuracy=0.79 [0.70-0.83]) and Gradient Boosting (F1=0.77 [0.71-0.82], AUPRC=0.89 [0.85-0.93], accuracy=0.81 [0.720.84]) had among the highest F1 score and AUPRC but confidence intervals overlapped with other methods. Explainability analyses highlighted prior drug-use diagnosis codes, triage acuity, and age as top predictors. Conclusion ML classifiers leveraging routinely collected triage data offer a feasible alternative to manual screening in flagging opioid misuse before physician evaluation, potentially enabling early harm-reduction interventions. Prospective multi-site validation, calibration, and bias assessments are warranted.
- Open Access
- 1 May 2025
- Published by SAGE Publications in Digital Health
- Vol. 11
Unlocking mobile health adoption: A qualitative exploration of user experiences, barriers, and facilitators within integrated youth services in British Columbia, Canada
Background: Early onset of mental health disorders is common, but many cases remain undetected and untreated, highlighting the need for early intervention. In Canada, youth mental health services face challenges, including fragmentation and resource limitations. Integrated youth services (IYS) aim to address these gaps for individuals aged 12–24 years. Mobile health (mHealth) programs, like Foundry Virtual BC, offer potential solutions, yet their integration and sustainability within IYS require further exploration. Objective: This study examined interest-holder perspectives on creating a sustainable, youth-centred mHealth system to improve mental health outcomes. The research focused on three questions: (a) How do users, service providers, and nonclinical staff perceive mHealth's effectiveness and impact? (b) What are the barriers and facilitators to mHealth integration within the Foundry IYS network? and (c) What strategies support the sustainability of mHealth services? Methods: A qualitative study using semi-structured interviews was conducted with 23 interest-holders, including youth users, service providers, and nonclinical staff from the Foundry network. Participants were recruited via social media and snowball sampling. Thematic analysis identified key themes and subthemes. Results: Three themes emerged regarding mHealth perceptions: (a) its own value, (b) its potential to address barriers to in-person care, and (c) its inherent limitations. Barriers and facilitators of mHealth integration were categorized into three domains: (a) design characteristics (e.g., app usability and content quality), (b) individual youth factors (e.g., privacy concern and inner struggle), and (c) external factors (e.g., safe space and support from peers). Sustainability was linked to service quality and external support. Conclusion: This study highlights the complexity of mHealth integration within an IYS network. Interest-holders emphasized addressing user motivations, privacy, and accessibility while advocating for co-design approaches to ensure mHealth meets diverse youth needs. Future research should focus on underrepresented groups to promote equity and improve mental health outcomes through sustainable mHealth solutions.
- 1 May 2025
- Published by SAGE Publications in Digital Health
- Vol. 11
Open AccessUnlocking mobile health adoption: A qualitative exploration of user experiences, barriers, and facilitators within integrated youth services in British Columbia, Canada
Background: Early onset of mental health disorders is common, but many cases remain undetected and untreated, highlighting the need for early intervention. In Canada, youth mental health services face challenges, including fragmentation and resource limitations. Integrated youth services (IYS) aim to address these gaps for individuals aged 12–24 years. Mobile health (mHealth) programs, like Foundry Virtual BC, offer potential solutions, yet their integration and sustainability within IYS require further exploration. Objective: This study examined interest-holder perspectives on creating a sustainable, youth-centred mHealth system to improve mental health outcomes. The research focused on three questions: (a) How do users, service providers, and nonclinical staff perceive mHealth's effectiveness and impact? (b) What are the barriers and facilitators to mHealth integration within the Foundry IYS network? and (c) What strategies support the sustainability of mHealth services? Methods: A qualitative study using semi-structured interviews was conducted with 23 interest-holders, including youth users, service providers, and nonclinical staff from the Foundry network. Participants were recruited via social media and snowball sampling. Thematic analysis identified key themes and subthemes. Results: Three themes emerged regarding mHealth perceptions: (a) its own value, (b) its potential to address barriers to in-person care, and (c) its inherent limitations. Barriers and facilitators of mHealth integration were categorized into three domains: (a) design characteristics (e.g., app usability and content quality), (b) individual youth factors (e.g., privacy concern and inner struggle), and (c) external factors (e.g., safe space and support from peers). Sustainability was linked to service quality and external support. Conclusion: This study highlights the complexity of mHealth integration within an IYS network. Interest-holders emphasized addressing user motivations, privacy, and accessibility while advocating for co-design approaches to ensure mHealth meets diverse youth needs. Future research should focus on underrepresented groups to promote equity and improve mental health outcomes through sustainable mHealth solutions.
- Open Access
- 1 May 2025
- Published by SAGE Publications in Digital Health
- Vol. 11
“You can’t categorize lived experiences”: Understanding youth engagement in mobile health in an integrated youth services setting
Objective: Youth mental health is a global challenge often compounded by fragmented services limiting access. In Canada, Integrated Youth Services (IYS) provide centralized access to youth health services, with youth engagement playing a key role during development and implementation. This study aims to identify key factors influencing youth engagement in mobile health (mHealth) development within an IYS setting. Methods: We conducted 23 semi-structured interviews with youth, clinical service providers, and non-clinical staff involved in mHealth services at Foundry. The Consolidated Framework for Implementation Research guided deductive thematic analysis. Results: We identified key facilitators and barriers across five domains: Recognizing core value of engagement, improving external coordination, addressing internal organizational challenges, mitigating power imbalances, and tailoring implementation strategies. Conclusion: The findings highlight the need for inclusive approaches, regulatory frameworks, and strategies to engage underrepresented youth. Future efforts should prioritize iterative learning systems and safe, supportive spaces for meaningful youth engagement.
- 1 May 2025
- Published by SAGE Publications in Digital Health
- Vol. 11
Open Access“You can’t categorize lived experiences”: Understanding youth engagement in mobile health in an integrated youth services setting
Objective: Youth mental health is a global challenge often compounded by fragmented services limiting access. In Canada, Integrated Youth Services (IYS) provide centralized access to youth health services, with youth engagement playing a key role during development and implementation. This study aims to identify key factors influencing youth engagement in mobile health (mHealth) development within an IYS setting. Methods: We conducted 23 semi-structured interviews with youth, clinical service providers, and non-clinical staff involved in mHealth services at Foundry. The Consolidated Framework for Implementation Research guided deductive thematic analysis. Results: We identified key facilitators and barriers across five domains: Recognizing core value of engagement, improving external coordination, addressing internal organizational challenges, mitigating power imbalances, and tailoring implementation strategies. Conclusion: The findings highlight the need for inclusive approaches, regulatory frameworks, and strategies to engage underrepresented youth. Future efforts should prioritize iterative learning systems and safe, supportive spaces for meaningful youth engagement.
- Open Access
- 25 April 2025
- Published by Wiley in Academic Emergency Medicine
Racial, ethnic, and sex disparities in buprenorphine treatment from emergency departments by discharge diagnosis
Objectives: Racial and sex disparities are noted in the administration and prescribing of buprenorphine from emergency departments (EDs) nationally. It is unknown whether disparities persist when accounting for the specific discharge diagnosis addressed during encounters such as opioid overdose or withdrawal.Methods: We conducted a cross‐sectional analysis of opioid‐related ED encounters from January 2020 through December 2023 using a national database, Epic Cosmos. We analyzed the effect of opioid encounter subtype—overdose or withdrawal—on receipt of buprenorphine using multivariable logistic regression adjusting for demographics and measured confounding variables. Encounter subtypes were defined by diagnosis codes and buprenorphine receipt was defined as administration or prescribing. We evaluated for racial, ethnic, and sex disparities within encounter subtypes for withdrawal and overdose.Results: We examined 1,088,033 opioid‐related encounters. Adjusted odds for buprenorphine receipt were greater for encounters involving withdrawal (odds ratio [OR] 2.22, 95% CI 2.18–2.26), though reduced for overdose (OR 0.52, 95% CI 0.51–0.53) and other opioid complications (OR 0.69, 95% CI 0.64–0.70). Males were more likely to receive buprenorphine (OR 1.18, 95% CI 1.16–1.19) than females. All racial minorities excepting American Indian/Native American patients (OR 1.04, 95% CI 1.00–1.08) were less likely to receive buprenorphine than White patients (Asian OR 0.85, 95% CI 0.79–0.81; Black OR 0.80, 95% CI 0.79–0.81; Native Hawaiian/Pacific Islander OR 0.79, 95% CI 0.71–0.89). Subtype analyses indicated decreased odds for buprenorphine receipt for female patients across all subtypes. An increased odds for buprenorphine receipt among Black patients (OR 1.04, 95% CI 1.01–1.07; ref. White race) was noted in encounters involving opioid withdrawal but disparities persisted for opioid overdose.Conclusions: The administration and prescribing of buprenorphine in the ED is heavily influenced by the presence of opioid withdrawal. Disparities disadvantage female patients and racial minorities. Some racial disparities, particularly among Black patients, are not evident when solely considering encounters involving opioid withdrawal. System‐level interventions are needed to address disparities and improve the equitable uptake of ED‐initiated buprenorphine.
- 25 April 2025
- Published by Wiley in Academic Emergency Medicine
Open AccessRacial, ethnic, and sex disparities in buprenorphine treatment from emergency departments by discharge diagnosis
Objectives: Racial and sex disparities are noted in the administration and prescribing of buprenorphine from emergency departments (EDs) nationally. It is unknown whether disparities persist when accounting for the specific discharge diagnosis addressed during encounters such as opioid overdose or withdrawal.Methods: We conducted a cross‐sectional analysis of opioid‐related ED encounters from January 2020 through December 2023 using a national database, Epic Cosmos. We analyzed the effect of opioid encounter subtype—overdose or withdrawal—on receipt of buprenorphine using multivariable logistic regression adjusting for demographics and measured confounding variables. Encounter subtypes were defined by diagnosis codes and buprenorphine receipt was defined as administration or prescribing. We evaluated for racial, ethnic, and sex disparities within encounter subtypes for withdrawal and overdose.Results: We examined 1,088,033 opioid‐related encounters. Adjusted odds for buprenorphine receipt were greater for encounters involving withdrawal (odds ratio [OR] 2.22, 95% CI 2.18–2.26), though reduced for overdose (OR 0.52, 95% CI 0.51–0.53) and other opioid complications (OR 0.69, 95% CI 0.64–0.70). Males were more likely to receive buprenorphine (OR 1.18, 95% CI 1.16–1.19) than females. All racial minorities excepting American Indian/Native American patients (OR 1.04, 95% CI 1.00–1.08) were less likely to receive buprenorphine than White patients (Asian OR 0.85, 95% CI 0.79–0.81; Black OR 0.80, 95% CI 0.79–0.81; Native Hawaiian/Pacific Islander OR 0.79, 95% CI 0.71–0.89). Subtype analyses indicated decreased odds for buprenorphine receipt for female patients across all subtypes. An increased odds for buprenorphine receipt among Black patients (OR 1.04, 95% CI 1.01–1.07; ref. White race) was noted in encounters involving opioid withdrawal but disparities persisted for opioid overdose.Conclusions: The administration and prescribing of buprenorphine in the ED is heavily influenced by the presence of opioid withdrawal. Disparities disadvantage female patients and racial minorities. Some racial disparities, particularly among Black patients, are not evident when solely considering encounters involving opioid withdrawal. System‐level interventions are needed to address disparities and improve the equitable uptake of ED‐initiated buprenorphine.