Category: Statistics

JMIR Hum Factors. 2026 Feb 9;13:e78483. doi: 10.2196/78483.

ABSTRACT

BACKGROUND: Patient autonomy through informed consent is a foundational ethical principle for health care practitioners. Online consent processes risk producing “consent in name only,” using manipulative or confusing user interfaces to extract consent artificially. This presents a significant danger for safe and ethical remote consultations for primary care providers, which often extract significant amounts of sensitive personal data.

OBJECTIVE: This study aims to examine the quality of consent obtained through both currently used and novel consent acquisition interfaces for remote e-consultations between a patient and a primary care provider.

METHODS: A total of 55 adult participants in the United Kingdom completed an interaction with a mock-up e-consultation system’s consent interface for data processing, with 54 completing the full study protocol. The participants were then asked questions regarding what they had provided consent for and the usability of the interface. These responses led to the calculation of an industry-standard System Usability Scale (SUS) score and a novel Quality of Informed Consent Collected Digitally (QuICCDig) score.

RESULTS: Users perceiving interfaces to be more usable (with a greater SUS score) were statistically significantly (n=54; P=.004) correlated with an increase in the quality of consent collected from those users (with a higher QuICCDig score). Nonetheless, both existing and novel user interfaces for collecting e-consultation consent were rated poorly, achieving a maximum SUS letter grade of “F.” In total, 45% (25/55) of all the participants reported not recalling making a privacy-related decision at all during their consultation, and 87% (48/55) did not recall being offered any alternatives to e-consultation.

CONCLUSIONS: The findings demonstrate that current methods for collecting consent in telemedical applications may not be fit for purpose and potentially fail to collect valid informed consent. However, increased usability scores from users do appear to drive improvements in the quality of consent collected. Therefore, decision-makers should place importance on high-quality interface design when building or procuring these systems. We have also provided the QuICCDig score for further use.

PMID:41662699 | DOI:10.2196/78483

JMIR AI. 2026 Jan 30. doi: 10.2196/76928. Online ahead of print.

ABSTRACT

BACKGROUND: Artificial intelligence (AI) models are increasingly being used in medical education. Although models like ChatGPT have previously demonstrated strong performance on USMLE-style questions, newer AI tools with enhanced capabilities are now available, necessitating comparative evaluations of their accuracy and reliability across different medical domains and question formats.

OBJECTIVE: To evaluate and compare the performance of five publicly available AI models: Grok, ChatGPT-4, Copilot, Gemini, and DeepSeek, on the USMLE Step 1 Free 120-question set, checking their accuracy and consistency across question types and medical subjects.

METHODS: This cross-sectional observational study was conducted between February 10 and March 5, 2025. Each of the 119 USMLE-style questions (excluding one audio-based item) was presented to each AI model using a standardized prompt cycle. Models answered each question three times to assess confidence and consistency. Questions were categorized as text-based or image-based, and as case-based or information-based. Statistical analysis was done using Chi-square and Fisher’s exact tests, with Bonferroni adjustment for pairwise comparisons.

RESULTS: Grok got the highest score (91.6%), followed by Copilot (84.9%), Gemini (84.0%), ChatGPT-4 (79.8%), and DeepSeek (72.3%). DeepSeek’s lower grade was due to an inability to process visual media, resulting in 0% accuracy on image-based items. When limited to text-only questions (n = 96), DeepSeek’s accuracy increased to 89.6%, matching Copilot. Grok showed the highest accuracy on image-based (91.3%) and case-based questions (89.7%), with statistically significant differences observed between Grok and DeepSeek on case-based items (p = .011). The models performed best in Biostatistics & Epidemiology (96.7%) and worst in Musculoskeletal, Skin, & Connective Tissue (62.9%). Grok maintained 100% consistency in responses, while Copilot demonstrated the most self-correction (94.1% consistency), improving its accuracy to 89.9% on the third attempt.

CONCLUSIONS: AI models showed varying strengths across domains, with Grok demonstrating the highest accuracy and consistency in this dataset, particularly for image-based and reasoning-heavy questions. Although ChatGPT-4 remains widely used, newer models like Grok and Copilot also performed competitively. Continuous evaluation is essential as AI tools rapidly evolve.

PMID:41662695 | DOI:10.2196/76928

Ann Plast Surg. 2026 Feb 9. doi: 10.1097/SAP.0000000000004663. Online ahead of print.

ABSTRACT

BACKGROUND: Free tissue transfer (FTT) can be used to achieve functional limb salvage in patients undergoing transmetatarsal amputation (TMA). However, the added bulk of the FTT can limit footwear options and hinder ambulation. Dysfunctional gait mechanics may lead to uneven pressure distribution and an increased risk of ulceration. In this proof-of-concept study, wearable sensor data were utilized to compare gait patterns between TMA patients closed with free tissue transfer (FTT) versus those closed with native plantar flaps to assess the biomechanical impact of FTT closure in patients undergoing TMA.

METHODS: Nineteen patients who underwent TMA (n = 14 [73.7% with native plantar flap] and n = 5 [26.3% with FTT]) completed a 120-second gait test with wearable sensors. Gait parameters included cadence (steps per minute), speed (meters per second), stride length (meters), stride duration (seconds), single- and double-limb support (%), elevation at midswing (centimeters), and root mean square (RMS) sway (centimeters), a marker of postural instability. Demographics, comorbidities, and operative details were collected. Groups were compared using univariate and multivariate analyses.

RESULTS: There were no significant differences observed between cohorts with respect to gait speed (0.85 vs 0.85 m/s, P = 0.997), cadence (97.8 vs 98.5 steps/min, P = 0.883), stride duration (P = 0.797), or elevation at midswing (1.64 vs 2.73 cm, P = 0.131). RMS sway was statistically significantly higher in the FTT group (0.35 vs 0.20 cm, P = 0.041), suggesting greater postural instability. On multivariate analysis, flap type was not independently associated with RMS sway (P = 0.839), although higher BMI trended toward significance (P = 0.116). The FTT group had significantly higher BMI (35.5 vs 28.1 kg/m2, P = 0.018) and greater prevalence of peripheral vascular disease (PVD) (80.0% vs 21.4%, P = 0.038).

CONCLUSION: Most spatiotemporal gait parameters were similar between TMA patients reconstructed with FTT and those with a native plantar flap. Although FTT patients had significantly higher RMS sway on univariate analysis, this was not independently associated with flap type after adjusting for confounders. This proof-of-concept analysis suggests that FTT, despite its bulk, may not significantly impair postamputation ambulation and remains a valuable limb salvage option when native plantar flap coverage may not be feasible.

PMID:41662694 | DOI:10.1097/SAP.0000000000004663

Ann Plast Surg. 2026 Feb 9. doi: 10.1097/SAP.0000000000004645. Online ahead of print.

ABSTRACT

BACKGROUND: Social media influences adolescent self-image and confidence, yet its impact on psychosocial outcomes in adolescents with cleft lip and/or palate remains poorly defined. This population may be particularly vulnerable given the importance of craniofacial features to self-perception during adolescence. Negative self-perceptions have been linked to psychosocial distress and interest in revisional surgery, whereas social media may also provide peer support and cleft-related information. This study examined social media use patterns and their association with psychosocial functioning in adolescents with cleft lip and/or palate.

METHODS: A cross-sectional online survey adapted from the CLEFT-Q and Pediatric Quality of Life Inventory (PedsQL) was distributed via REDCap to adolescents aged 10-18 years with cleft lip and/or palate. A parallel survey assessed parental perceptions of their child’s psychosocial functioning. Recruitment occurred at the Lancaster Cleft Palate Clinic between 2023 and 2025. The approximately 10-minute survey assessed demographics, social media use, and psychosocial outcomes across well-being, social functioning, school functioning, and self-perception domains using 5-point Likert scales.

RESULTS: Thirty-one adolescents completed the teen survey (mean [SD] age, 15.2 [2.1] years). Daily social media use was reported as <1 hour by 5/31 (16.1%), 1-3 hours by 14/31 (45.2%), 3-5 hours by 9/31 (29.0%), and ≥5 hours by 3/31 (9.7%). The most commonly used platforms were YouTube 24/31 (77.4%), TikTok 21/31 (67.7%), Snapchat 21/31 (67.7%), and Instagram 16/31 (51.6%). Palate surgery was reported by 29/31 (93.5%), and 17/31 (54.8%) had undergone combined lip, palate, and nasal surgery. Adolescents reporting ≥5 h/d of use demonstrated lower psychosocial functioning scores across domains, including teasing, appearance concerns, and happiness, although differences were not statistically significant. No significant differences were observed between adolescents reporting 1-3 versus 3-5 h/d of use. Surgical history and platform choice were not associated with psychosocial outcomes.

CONCLUSIONS: This pilot study provides preliminary descriptive insight into social media use and psychosocial functioning among adolescents with cleft lip and/or palate. Although no statistically significant associations were identified, consistent directional trends suggested lower psychosocial scores among adolescents reporting heavy social media use. These findings support cautious consideration of social media-based interventions and highlight the need for larger, adequately powered studies.

PMID:41662684 | DOI:10.1097/SAP.0000000000004645

JMIR Cancer. 2026 Feb 9;12:e73058. doi: 10.2196/73058.

ABSTRACT

BACKGROUND: Reliance on telehealth increased dramatically during the COVID-19 pandemic, introducing new opportunities to consider the use of telehealth across the cancer control continuum. However, patient, clinician, and staff perspectives about the types of cancer care appointments that are considered appropriate and the clinical care needs to support expanded remote care services are limited. Understanding older adults’ diverse technology needs and perspectives is especially important given that they comprise a large and growing proportion of patients with cancer.

OBJECTIVE: This study aimed to describe the perceptions and experiences of older patients with cancer and their clinical care team members regarding the expansion of telehealth use across the cancer control continuum and to solicit suggestions about how to support telehealth use for cancer care delivery.

METHODS: Using a convergent mixed methods design, we surveyed and interviewed patients aged ≥60 years, clinicians, and staff at a comprehensive cancer center in the southern United States between December 2020 and November 2021. Interview questions were rooted in the sociotechnical model, which proposes 8 interrelated dimensions representing factors influencing the design, use, and outcomes associated with health information technologies. Patient survey domains included telehealth experience and satisfaction and factors affecting telehealth perceptions and use; clinician survey domains included contexts of telehealth appropriateness, training, and barriers and facilitators to telehealth service provision. Survey data were analyzed using descriptive statistics. Qualitative data were thematically analyzed using a combined deductive and inductive approach.

RESULTS: We received completed surveys from 128 patients (567 invited) and 106 clinicians and staff (146 invited). We completed 14 patient (29 invited) and 20 clinician and staff (22 invited) interviews. Across all participants, most agreed or strongly agreed that multiple cancer care appointment types should be offered via telehealth, including discussing treatment side effects (75/102, 73.5% of patients and 66/94, 70.2% of clinicians and staff), results communication (71/102, 69.6% of patients and 65/94, 69.1% of clinicians and staff), and treatment follow-up (67/102, 65.7% of patients and 52/93, 55.9% of clinicians and staff). In interviews, participants elaborated on factors influencing the appropriateness of telehealth versus in-person appointments, including symptom severity, type of cancer, and purpose of the appointment. Many patient and staff suggestions focused on ways to address digital literacy gaps, while clinicians recommended improving clinic workflows, infrastructure, and training.

CONCLUSIONS: Overall, clinicians, staff, and older patients with cancer all responded positively toward expanding telehealth use across multiple cancer and appointment types across the cancer control continuum. Older adults with cancer are generally interested in telehealth for cancer care, especially if strategies to address digital literacy gaps are incorporated. Clinicians and staff members expressed specialized training and infrastructure needs to optimize telehealth uptake and service delivery.

PMID:41662673 | DOI:10.2196/73058

Int J Group Psychother. 2026 Feb 9:1-38. doi: 10.1080/00207284.2025.2600103. Online ahead of print.

ABSTRACT

Experienced and internalized weight stigma (IWS) are associated with a plethora of health and psychosocial consequences. In this study, we tested the outcomes of a 10-12 week, online, counselor-facilitated support group (We Deserve Space; WDS) for individuals who had experienced weight stigma. Twenty-five large-bodied individuals participated in three iterations of the WDS group. Participants average age was 43.68 years (SD = 13.84, range = 28-78). Revisions to WDS curricula were made iteratively in response to feedback. Paired samples t-tests yielded statistically significant improvements on all variables (IWS, antifat attitudes, depressive symptoms, loneliness, eating-disorder functional impairment, belongingness) from baseline to end-of-group, with effect sizes in the medium to large range. This study provides preliminary evidence that WDS may improve IWS, disordered-eating functional impairment, and psychosocial well-being in large-bodied individuals.

PMID:41662670 | DOI:10.1080/00207284.2025.2600103

Ann Plast Surg. 2026 Feb 9. doi: 10.1097/SAP.0000000000004661. Online ahead of print.

ABSTRACT

BACKGROUND: Nipple-sparing mastectomy (NSM) improves aesthetic outcomes but carries a higher risk of ischemic complications, particularly in patients with large, ptotic breasts. This study evaluates whether reduction mammaplasty prior to NSM with deep inferior epigastric perforator (DIEP) flap reconstruction can help mitigate these risks.

METHODS: All patients who underwent NSM with immediate DIEP flap reconstruction between 2016 and 2024 were identified. Only those with native breast cup size D or larger were included. Patients were divided into 2 cohorts: those who underwent reduction mammaplasty prior to NSM (reduction cohort) and those without prior breast surgery (control cohort).

RESULTS: The reduction group included 21 patients (39 breasts), and the control group included 29 patients (51 breasts). Both groups had an average preoperative breast cup size of DD. Within the reduction group, 41.0% underwent planned staged reductions (average interval, 4.63 months), whereas 59.0% had prior elective reductions (average interval, 13.35 years). Postoperatively, the reduction group experienced significantly lower rates of skin necrosis (5.1% vs 37.3%, P < 0.001) and nipple-areolar complex (NAC) necrosis (2.6% vs 21.6%, P = 0.011). Infection rates were also lower, approaching statistical significance (5.1% vs 19.6%, P = 0.061). On multivariate regression, reduction mammaplasty was a significant protective factor against both skin (β = -2.774, P = 0.002) and NAC necrosis (β = -2.385, P = 0.030), reducing the odds by 93.8% and 90.8%, respectively.

CONCLUSION: Our findings suggest that prior reduction mammaplasty is associated with decreased ischemic complications in large-breasted patients undergoing NSM and DIEP flap reconstruction.

PMID:41662666 | DOI:10.1097/SAP.0000000000004661

JMIR Form Res. 2026 Feb 9;10:e71541. doi: 10.2196/71541.

ABSTRACT

BACKGROUND: Effective diabetes management requires individualized treatment strategies tailored to patients’ clinical characteristics. With recent advances in artificial intelligence, large language models (LLMs) offer new opportunities to enhance clinical decision support, particularly in generating personalized recommendations.

OBJECTIVE: This study aimed to develop and evaluate an LLM-based outpatient treatment support system for diabetes and examine its potential value in routine clinical decision-making.

METHODS: Three compact LLMs (Llama 3.1-8B, Qwen3-8B, and GLM4-9B) were fine-tuned on deidentified outpatient electronic health records using a parameter-efficient low-rank adaptation approach. The optimized models were embedded into a prototype hospital information system via a retrieval-augmented generation framework to generate individualized treatment recommendations, laboratory test suggestions, and medication prompts based on demographic and clinical data.

RESULTS: Among the models evaluated, the fine-tuned GLM4-9B demonstrated the strongest performance, producing clinically reasonable treatment plans and appropriate laboratory test recommendations and medication suggestions. It achieved a mean Bilingual Evaluation Understudy for 4-grams score of 67.93 (SD 2.74) and mean scores of 44.30 (SD 3.91) for Recall-Oriented Understudy for Gisting Evaluation for overlap of unigrams, 27.34 (SD 1.85) for Recall-Oriented Understudy for Gisting Evaluation for overlap of bigrams, and 37.67 (SD 2.88) for Recall-Oriented Understudy for Gisting Evaluation for Longest Common Subsequence.

CONCLUSIONS: The fine-tuned GLM4-9B shows strong potential as a clinical decision support tool for personalized diabetes care. It can provide reference recommendations that may improve clinician efficiency and support decision quality. Future work should focus on enhancing medication guidance, expanding data sources, and improving adaptability in cases involving complex comorbidities.

PMID:41662664 | DOI:10.2196/71541

J Long Term Eff Med Implants. 2026;36(1):1-7. doi: 10.1615/JLongTermEffMedImplants.2025059687.

ABSTRACT

The purpose of this study was to compare outcomes following open reduction and internal fixation (ORIF) of proximal humerus fractures with Zimmer Biomet and DePuy Synthes plating systems. This was a retrospective review of patients who underwent proximal humerus fracture fixation utilizing Zimmer Biomet or DePuy Synthes plating systems between June 2016 and February 2023. Patients without postoperative follow-up were excluded. Complication rates were compared between the two cohorts using chi-square and two-tailed t-tests for categorical and continuous variables, respectively. A total of 86 patients were identified for inclusion and eight were excluded for inadequate follow up. Statistical analysis comprised a total of 78 patients, 41 who underwent ORIF with a Zimmer Biomet implant, and 37 who underwent ORIF with a DePuy Synthes implant. The overall complication rate was 4.88% (n = 2) in the Zimmer Biomet cohort and 13.5% (n = 5) in the DePuy Synthes cohort (P > 0.05). Both Zimmer Biomet patients required revision surgery, compared with two out of five patients in the DePuy-Synthes group. There were no statistically significant differences in rates of avascular necrosis, impingement, malunion, or fixation failure between the two cohorts. Overall complication rate and need for revision surgery were similar between patients who underwent fixation with the Zimmer Biomet or DePuy Synthes plating systems. Our findings support selection of either plating system based on surgeon preference and equipment availability as neither demonstrated clinical superiority.

PMID:41662644 | DOI:10.1615/JLongTermEffMedImplants.2025059687

Am J Clin Oncol. 2026 Feb 5. doi: 10.1097/COC.0000000000001300. Online ahead of print.

ABSTRACT

OBJECTIVES: To evaluate the hepatotoxicity and clinical outcomes following external beam radiation therapy (EBRT) in patients with hepatocellular carcinoma (HCC) previously treated with Yttrium-90 radioembolization (Y-90), a setting in which cumulative liver radiation exposure raises concern for increased toxicity risk.

METHODS: We performed a retrospective review of 94 HCC patients treated with EBRT from 2016 to 2024, including 15 treated with Y-90. Hepatotoxicity was assessed using Albumin-Bilirubin (ALBI) score changes at 3 and 6 months post-EBRT. Secondary outcomes included CTCAE toxicity, local control, and survival. Treatment details, prior locoregional therapies, and dosimetric parameters were collected. Fisher exact tests and Kaplan-Meier analyses were used for statistical comparisons.

RESULTS: Baseline demographics and liver function were similar between groups. Over half the Y-90 cohort (53%) received fractionated proton therapy. At 6 months, grade 1 and 2 ALBI deterioration occurred in 2 (13%) and 4 patients (27%) with prior Y-90, compared with 17 (22%) and 21 patients (27%) without prior Y-90 (P=0.77). Grade ≥2 CTCAE toxicity occurred in 2 patients (13%) with prior Y-90 and in 5 patients (6%) without. Local control was 92% with no local failures in patients treated after Y-90. Median OS was 39 months without prior Y-90 and 34 months with prior Y-90 (P=0.89).

CONCLUSIONS: EBRT following Y-90 was not associated with increased hepatotoxicity or inferior oncologic outcomes compared with EBRT alone. When delivered with individualized planning, including fractionation, EBRT represents a safe and effective salvage modality for patients with residual or recurrent HCC after Y-90.

PMID:41662643 | DOI:10.1097/COC.0000000000001300