Kerly Guevara Maldonado

Kerly Guevara Maldonado

Auto-generated profile, based on publicly available information

Kerly Guevara Maldonado

Knowledge and Evaluation Research Unit, Mayo Clinic, Rochester, Minnesota

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 September 2025 in Patient Preference and Adherence
Verified
Journal Article
Open access
PDF file
MZMisk Al Zahidy
Misk Al Zahidy
KMKerly Guevara Maldonado
Kerly Guevara Maldonado
SSSuvyaktha Simha
Suvyaktha Simha
MBMariana Borras-Osorio
Mariana Borras-Osorio
MBMegan Branda
Megan Branda
VMVictor M Montori
Victor M Montori
Background: Patients with diabetes manage complex treatment regimens that include the use of digital medicine tools. Existing instruments do not explicitly capture treatment burden, i.e., workload and its effect on patient’s quality of life, from using digital medicine tools. Objective: To engage patients and clinical experts in adapting the Treatment Burden Questionnaire (TBQ) to capture digital treatment burden. The adapted instrument underwent cognitive testing and refinements to ensure it captures the burden of using digital medicine tools in diabetes self-management. Methods: This two-phase study was conducted with adults with diabetes at the Division of Endocrinology at Mayo Clinic (Rochester, MN). First, we mapped themes from prior concept elicitation interviews to existing TBQ items to identify content gaps related to digital burden. Based on these gaps, the study team and expert panel generated new items and adapted existing ones to better reflect the workload and burdens from using digital medicine tools. The resulting instrument underwent three rounds of cognitive testing with adult patients living with diabetes, using a think-aloud protocol to assess clarity, relevance, and comprehensiveness. Results of cognitive testing informed iterative refinements across three rounds of interviews, leading to improved clarity, reduced redundancy, and improved relevance of items. Results: The final TBQ+D retained the original 15-item TBQ structure, added 8 new items, and modified 8 extant ones to capture burden of digital care (e.g, syncing issues, discomfort from sensors, and device malfunctions). Cognitive testing demonstrated strong content relevance and patient comprehension. Conclusion: The TBQ+D can measure digital treatment burden in patients with diabetes. Limitations include a relatively homogeneous sample drawn from a single center. Next steps include field testing for validation across diverse populations and settings.

24 March 2026 in JMIR Research Protocols
Verified
Journal Article
Open access
MZMisk Al Zahidy
Misk Al Zahidy
KMKerly Guevara Maldonado
Kerly Guevara Maldonado
LALuis Vilatuna Andrango
Luis Vilatuna Andrango
APAna Cristina Proano
Ana Cristina Proano
ACAna Gabriela Claros
Ana Gabriela Claros
JBJuan P Brito
Juan P Brito
Background: The promise of artificial intelligence (AI) in medicine depends on its ability to learn from data that reflect what matters to patients and clinicians in the care process. Most existing models are trained on electronic health records (EHRs), which primarily capture biological measures but rarely the interactions and relationships between patients and clinicians. These relationships, central to how care is understood, negotiated, and delivered, unfold across multiple modalities, including voice, text, and video, yet remain largely absent from current datasets. As a result, AI systems trained solely on EHRs risk perpetuating a narrow biomedical view of medicine and overlooking the lived exchanges that define clinical encounters. Objective: This study aims to design, implement, and evaluate the feasibility of a longitudinal, multimodal system for capturing patient-clinician encounters, linking 360° video or audio recordings with postvisit surveys and EHR data, to create a foundational dataset for downstream AI research. Methods: This single-site study was conducted in an academic outpatient specialty clinic (Division of Endocrinology, Mayo Clinic, Rochester, Minnesota, United States). Adult patients attending in-person visits with participating clinicians were invited to enroll. Encounters were recorded using a 360° 2D monocular video camera and dual-channel audio. After each visit, patients completed a brief survey assessing relational empathy, satisfaction, visit pace, and treatment burden. Demographic and clinical data were extracted from the EHR. Feasibility was assessed using 5 prespecified end points: clinician consent, patient consent, recording success, survey completion, and data linkage across modalities. Results: Recruitment began in January 2025. By August 2025, 35 of 36 (97%) eligible clinicians and 212 of 281 (75%) approached eligible patients had consented. Of the consented encounters, 162 (76%) resulted in a complete 360° video recording, and the postvisit surveys were completed for 204 of 212 (96%) consented encounters, reflecting 1 survey per encounter. Data collection is ongoing as of December 2025, and further analyses will be reported in subsequent publications. Conclusions: This protocol describes a longitudinal multimodal encounter capture system that links 360° audio or video with postvisit surveys and EHR data. The study specifies operational definitions, workflows, feasibility end points, and governance procedures to support implementation and replication in other clinical settings.


13 May 2025 in medRxiv
Verified
Preprint
Open access
PDF file
MZMisk Al Zahidy
Misk Al Zahidy
KGKerly Guevara
Kerly Guevara
SSSue Simha
Sue Simha
MBMariana Borras-Osorio
Mariana Borras-Osorio
MBMegan E. Branda
Megan E. Branda
VMVictor M. Montori
Victor M. Montori
Background Patients with diabetes often manage complex treatment regimens that increasingly include the use of digital medicine tools. While several instruments measure treatment burden, none capture the specific burden introduced by digital medicine tools used in self-care. Development of patient reported measures to capture digital treatment burden are needed to assess the lived experiences and challenges patients face when using digital medicine tools. Objective To engage patients and clinical experts in adapting an existing measure of treatment burden—the Treatment Burden Questionnaire (TBQ). The adapted instrument underwent cognitive testing and refinements to ensure it could capture the burden of using digital medicine tools in the self-management and care of diabetes. Methods This two-phase study was conducted with adults with diabetes and other chronic conditions at the Division of Endocrinology at Mayo Clinic in Rochester, Minnesota. First, we mapped themes from prior concept elicitation interviews to existing TBQ items to identify content gaps related to digital burden. Based on these gaps, the study team and expert panel followed an item development guide to generate new items and adapt existing ones to better reflect the workload and burdens imposed by digital medicine tools. The resulting adapted instrument underwent three rounds of cognitive testing with adult patients living with diabetes, using a think-aloud protocol to assess clarity, relevance, and comprehensiveness. Results of cognitive testing informed iterative refinements across three rounds of interviews, leading to improved clarity, reduced redundancy, and improved relevance of items. Results The final TBQ+D retained the structure of the original 15-item TBQ and included 8 new items and modified 8 extant ones to capture burden of digital care (e.g., syncing issues, discomfort from sensors, and device malfunctions). Cognitive testing demonstrated strong content relevance and patient comprehension. Conclusion The TBQ+D appears able to measure digital treatment burden in patients with diabetes. Planned next steps include field testing the instrument to test validity hypotheses, and if successful, extending this evaluation to diverse populations and clinical and research settings.

3 September 2026 in Journal of Medical Internet Research
Verified
Journal Article
Open access
MJMaria Lizarazo Jimenez
Maria Lizarazo Jimenez
ACAna Gabriela Claros
Ana Gabriela Claros
KGKieran Green
Kieran Green
DTDavid Toro-Tobon
David Toro-Tobon
FLFelipe Larios
Felipe Larios
OPOscar J Ponce-Ponte
Oscar J Ponce-Ponte
Background: Large language models (LLMs) are increasingly demonstrating the potential to reach human-level performance in generating clinical summaries from patient-clinician conversations. LLMs are usually evaluated against clinical summaries that focus mainly on patients’ biology and not on their biography (eg, preferences, values, wishes, and concerns). To achieve patient-centered care, artificial intelligence clinical summarization must incorporate patient-centered domains, implemented through patient-centered summaries (PCSs). Objective: This study aimed to develop a framework to generate PCS that capture patients’ values, preferences, and wishes while ensuring clinical utility for clinicians, and assess if current open-source LLMs can achieve human-level performance in generating PCS. Methods: We developed a 4-step mixed methods process to define and evaluate PCS. First, 2 patient and public involvement and engagement groups were convened in the United Kingdom (10 patients and 8 clinicians), who participated in semistructured interviews exploring what personal and contextual information should be included in clinical summaries and how it should be structured for clinical use. Second, findings were translated into an annotation guideline, which was used by 8 clinician annotators to generate gold standard PCS from 88 transcribed patient-clinician consultations about the management of atrial fibrillation. Third, 16 consultations were used to iteratively develop and refine a prompt aligned with the annotation guideline. Finally, 5 LLMs (Llama-3.2-3B [Meta AI], Llama-3.1-8B [Meta AI], Mistral-8B [Mistral AI], Gemma-3-4B [Google DeepMind], and Qwen3-8B [Alibaba]) generated summaries from 72 consultations using zero-shot and few-shot prompting, which were evaluated against gold standard PCS using ROUGE-L (Recall-Oriented Understudy for Gisting Evaluation–Longest Common Subsequence) and BERTScore (Bidirectional Encoder Representations from Transformers Score) and assessed for correctness, completeness, conciseness, patient-centeredness, and fluency. Results: Patients emphasized that summaries should include (1) lifestyle routines and daily functioning as indicators of independence or disruption; (2) the presence and role of social support systems, especially during crises; (3) recent life events or stressors, such as trauma, loss, or caregiving demands; and (4) care preferences, values, and communication styles that provide meaning or reflect autonomy. Clinicians sought summaries that included a concise functional baseline, psychosocial context, and emotional cues, preferably in a structured, clinically digestible format. In the 72 consultations (mean age 70, SD 11 y; 32/72, 44.4% female), the best zero-shot performance was observed with Mistral-8B (ROUGE-L 0.189) and Llama-3.1-8B (BERTScore 0.673). The best few-shot prompting was found with 3 examples using Llama-3.1-8B (ROUGE-L 0.206 and BERTScore 0.683). Conclusions: The open-source LLMs we evaluated did not achieve human-level performance in generating PCSs. Without task-specific fine-tuning, current open-source LLMs cannot reach human-level performance in this task. Our framework serves as an innovative guideline for developing gold standard PCS for artificial intelligence clinical tasks.

9 February 2026 in medRxiv
Verified
Preprint
Open access
PDF file
VMVictor Montori
Victor Montori
FLFelipe Larios
Felipe Larios
SBSatya Sai Sri Bandi
Satya Sai Sri Bandi
APAna Cristina Proano
Ana Cristina Proano
KGKerly Guevara
Kerly Guevara
KLKasia J. Lipska
Kasia J. Lipska
Background The self-management of type 2 diabetes (T2D) typically requires enacting various lifestyle changes, which can challenge people living with T2D. Clinical encounters between people with T2D and their clinicians, however, are often focused on metabolic management, leaving less time available for other self-management topics. The QBSAFE cards help patients articulate aspects of their experience with diabetes and prioritize issues for discussion. Methods This report details secondary outcomes of a randomized controlled trial; primary outcomes are reported elsewhere. All data was collected at Fair Haven Community Health Care, a federally qualified primary care clinic. 11 clinicians were randomly assigned to provide either usual care or usual care with QBSAFE cards to 155 of their patients with type 2 diabetes and hemoglobin A1c >8%. All patient encounters were video recorded for analysis. Patients and clinicians were not blinded to arm allocation but were kept unaware of the specific aims of the trial. Encounter video reviewers were blinded to arm allocation, but not to specific aims of the trial. The outcomes of interest for this report were the extent to which the QBSAFE cards were used as intended, their effect on the topics of discussion, and whether they enabled clinicians to notice and respond to each patient's situation; comparisons between arms were conducted by a linear mixed model with fixed effect of arm and cluster effect of clinician, analyzed in both intent-to-treat and per-protocol populations. Findings 12 patients were excluded post-randomization (A1c <8%). Of 143 eligible patients, 137 encounters (65 in the usual care arm, 72 in QBSAFE) yielded evaluable videos. QBSAFE was used as intended in 61 (85%) QBSAFE arm encounters. Conversations about burden of treatment related to non-pharmacological interventions (17 vs 33, p= 0.04) and taking medications (11 vs 33, p= 0.0008) and about the patient's challenging environment (2 vs 10, p= 0.04) were more prevalent in the QBSAFE group. There was no difference in the rate of conversations about metabolic management or of new care plans as a result of conversations between groups. Interpretation While there was a difference in the types of conversations observed between the two study arms, this difference was small and only apparent in a few domains. Future work could aim to modify the QBSAFE cards to more effectively stimulate patient-centered discussions and to further prepare clinicians to respond to a variety of issues raised during the clinical visit. Funding 3 This work was supported by funding from the