Tag: statistics

Rheumatology (Oxford). 2025 Nov 4:keaf585. doi: 10.1093/rheumatology/keaf585. Online ahead of print.

ABSTRACT

OBJECTIVES: It is unclear whether clinicians agree which manifestations of AAV are associated with necrotizing granulomas or if their presence affects clinical decision-making.

METHODS: We surveyed physicians experienced in caring for individuals with AAV, querying: experience with AAV; beliefs concerning how granulomas affect the diagnosis, treatments, and outcomes of AAV; beliefs concerning the frequency with which granulomas are found in 36 manifestations of AAV; and degree to which granulomas change choice of induction therapy for specific manifestations of AAV. We analyzed responses using descriptive statistics and multivariable linear regression.

RESULTS: We received 142 responses from 35 countries. Responses had a median Likert response ≥5 on a 7-point scale (equal to ‘partially agree’) that granulomatous manifestations respond differently to therapy, increase risk of relapse, and increase organ damage. Four of 36 manifestations were believed to be caused by granulomas in a median of ≥ 75% of cases (on a scale of 0 = never to 100 = always caused by granuloma), 19 in a median of ≤ 25% of cases, and 13 in intermediate medians. The perceived degree to which granulomas caused manifestations was not associated with changes in therapy to induce remission in severe AAV (p-values 0.26-0.93 across scenarios).

CONCLUSIONS: Physicians experienced in vasculitis generally agree on which manifestations of AAV are and are not caused by granulomas and that granulomatous inflammation alters the natural history and treatment of AAV. However, the presence of granulomatous manifestations did not alter treatment choices to induce remission in severe AAV.

PMID:41191924 | DOI:10.1093/rheumatology/keaf585

JMIR Form Res. 2025 Nov 5;9:e69103. doi: 10.2196/69103.

ABSTRACT

BACKGROUND: Cardiopulmonary resuscitation (CPR) is vital for improving patient outcomes in medical emergencies. Both laypersons and health care professionals often seek guidance on performing CPR. In today’s digital age, many turn to easily accessible platforms such as YouTube for practical skills.

OBJECTIVE: This study evaluates the didactic and content quality of CPR videos on YouTube using comprehensive checklists and investigates the association between the assigned quality scores and type of publisher, view count, and video rankings.

METHODS: Videos were included based on defined search terms and exclusion criteria. Two emergency physicians rated each video independently using validated checklists concerning content and didactic quality. Linear regression analysis was performed to assess the relationships between video quality scores and view counts, as well as video rankings.

RESULTS: Of the 250 videos identified, 74 (29.6%) met the inclusion criteria. On the content checklist, videos scored an average of 56.5% (SD 19.2%), and on the didactic checklist, they scored 66.6% (SD 14.3%); none achieved the maximum score. Videos from official medical institutions scored significantly higher in content quality compared to nonofficial sources (P=.04). Video quality scores were not associated with video rankings or view counts.

CONCLUSIONS: The study highlights substantial variability in the didactic and content quality of CPR-related videos on YouTube. For medical educators, this underlines the need to curate and recommend reliable online resources or to develop new high-quality content aligned with established checklists. For the general public, the findings caution against relying on popularity metrics as indicators of accuracy and emphasize the importance of guidance from trusted institutions.

PMID:41191922 | DOI:10.2196/69103

Rheumatology (Oxford). 2025 Nov 4:keaf580. doi: 10.1093/rheumatology/keaf580. Online ahead of print.

ABSTRACT

OBJECTIVES: This study aimed to describe annual trends in sodium-glucose cotransporter-2 inhibitor (SGLT2i) prescriptions among patients with rheumatoid arthritis (RA) and diabetes mellitus (DM), and to assess whether SGLT2i use increases urinary tract infection (UTI) risk, emulating a target trial.

METHODS: An administrative claim database identified RA patients aged ≥18 years with type 2 DM from April 2015 to April 2023. Population 1 included RA patients with DM for assessing diabetes medication status. Population 2 included those with newly initiated first- or second-line antidiabetic medications. The primary outcome was UTI, defined using diagnostic code and antibiotic prescription. For intention-to-treat (ITT) analysis, we used quasi-Poisson regression, whereas for as-treated (AT) analysis, we applied Poisson mixed-effects models.

RESULTS: Among the 26 754 patients in Population 1, SGLT2i prescriptions notably increased, while traditional diabetes medications decreased. Population 2 included 9,772 patients (mean age 69.8 years, 60% women, 13% SGLT2i initiators, 42% on glucocorticoids). During a mean 34-month follow-up, 2,269 UTI events occurred in 1,373 patients. ITT analysis showed no significant difference between SGLT2i and other antidiabetic drug. However, AT analysis demonstrated statistically significant association (adjusted incidence rate ratio 1.64, 95% CI 1.26-2.13). The SGLT2i- daily glucocorticoid dose interaction was not significant in either model.

CONCLUSIONS: Among RA patients with DM, SGLT2i use may inherently increase the risk of UTI compared with other antidiabetics. However, the ITT analysis findings support the safety of SGLT2i selection in routine clinical practice, including in patients receiving glucocorticoids.

PMID:41191914 | DOI:10.1093/rheumatology/keaf580

JMIR Med Inform. 2025 Nov 5;13:e77603. doi: 10.2196/77603.

ABSTRACT

BACKGROUND: In the rapidly evolving landscape of health informatics, adopting a standardized common data model (CDM) is a pivotal strategy for harmonizing data from diverse sources within a cohesive framework. Transitioning regional databases to a CDM is important because it facilitates integration and analysis of vast and varied health datasets. This is particularly relevant in China, where unique demographic and epidemiologic profiles present a rich yet complex data landscape. The significance of this research from the perspective of the Chinese population lies in its potential to bridge gaps among disparate data sources, enabling more comprehensive insights into health trends and outcomes.

OBJECTIVE: This study aimed to understand biomedical professionals’ and trainees’ acceptance of the CDM in medical data management in China and to explore potential advantages and challenges associated with its promotion, implementation, and development in the country.

METHODS: We conducted a questionnaire survey using Sojump and distributed it on WeChat to evaluate the Chinese population’s acceptance of transitioning from local databases to a standardized CDM. The survey assessed participants’ understanding of the CDM and the Observational Medical Outcomes Partnership CDM, as well as their views on the importance of CDM for regional databases in China. Analysis of the survey results revealed the current state, challenges, and trends in CDM application within Chinese health care, providing a foundation for future efforts in data standardization and sharing. The reliability of the questionnaire data was assessed using Cronbach α and Guttman Lambda 6 to determine internal consistency.

RESULTS: Our survey of 418 participants revealed that 41.9% (175/418) were aware of the CDM. Recognition of CDM increased with higher education levels and was notably higher among professionals in contract research organizations and the pharmaceutical industry. Knowledge of CDM was primarily gained through literature and conferences, with formal education less common. Logistic regression analysis indicated that individuals with doctoral degrees, researchers, executives, medical professionals, data engineers, Centers for Disease Control and Prevention staff, and statisticians were more likely to be aware of CDM. Subgroup analyses showed higher awareness among doctoral versus nondoctoral and Beijing-based versus non-Beijing respondents, while perceived necessity was broadly comparable across subgroups. Overall, 94.7% (396/418) of respondents believed CDM integration in China is necessary for standardization and efficiency. Despite 60.7% (254/418) optimism for the Observational Medical Outcomes Partnership as the preferred CDM, challenges such as mapping traditional Chinese medicine or Chinese medical insurance remain.

CONCLUSIONS: A large proportion of respondents expressed a favorable view of implementing the CDM in regional databases in China, with notable endorsement from the doctoral group and professionals in contract research organizations or pharmaceutical sectors; subgroup differences were concentrated in awareness rather than perceived necessity. Participants suggested enhancing CDM-related education and establishing clear data-sharing regulations to support CDM advancement in China.

PMID:41191912 | DOI:10.2196/77603

J Med Internet Res. 2025 Nov 5;27:e77721. doi: 10.2196/77721.

ABSTRACT

The rapid uptake of supervised machine learning (ML) in clinical prediction modelling, particularly for binary outcomes based on tabular data, has sparked debate about its comparative advantage over traditional statistical logistic regression. Although ML has demonstrated superiority in unstructured data domains, its performance gains in structured, tabular clinical datasets remain inconsistent and context dependent. This viewpoint synthesizes recent comparative studies and simulation findings to argue that there is no universal best modelling approach. Model performance depends heavily on dataset characteristics (eg, linearity, sample size, number of candidate predictors, minority class proportion) and data quality (eg, completeness, accuracy). Consequently, we argue that efforts to improve data quality, not model complexity, are more likely to enhance the reliability and real-world utility of clinical prediction models.

PMID:41191908 | DOI:10.2196/77721

JMIR Res Protoc. 2025 Nov 5;14:e73356. doi: 10.2196/73356.

ABSTRACT

BACKGROUND: With current data supporting an increasing population of older people experiencing homelessness (OPEH) requiring unique spatial and placemaking considerations in end-of-life care, understanding the environmental factors that influence their well-being is crucial.

OBJECTIVE: This protocol paper provides a comprehensive overview for evaluating hospice environments tailored to the needs of OPEH.

METHODS: The Aging in the Right Place study aims to address this gap by developing and implementing the Aging in the Right Place-Hospice Environmental Assessment Protocol (AIRP-HEAP) and AIRP-HEAP secondary observation tools. The AIRP-HEAP tool evaluates the built and natural environment within hospice settings. Adaptations were made to ensure alignment with the unique needs of OPEH, such as reconceptualizing spiritual care and expanding the definition of family accommodation. Additionally, the AIRP-HEAP secondary observation tool supplements this by capturing contextual data on the surrounding neighborhood of the hospice site, providing a holistic understanding.

RESULTS: Data were collected at Maggie’s Lodge hospice between November and December 2024 using the AIRP-HEAP and AIRP-HEAP secondary observation tools. The dataset is currently being cleaned, with analysis planned between May and December 2025. The anticipated results will highlight the importance of considering environmental factors in hospice environments and inform recommendations to improve end-of-life care for OPEH.

CONCLUSIONS: Data collected using these audit tools can guide environmental modifications in hospice settings to facilitate aging and end-of-life care in the right place. Thus, this protocol paper aims to promote the adoption of best practices in hospice design to better support this marginalized population.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/73356.

PMID:41191906 | DOI:10.2196/73356

JMIR Res Protoc. 2025 Nov 5;14:e65012. doi: 10.2196/65012.

ABSTRACT

BACKGROUND: Evidence suggests that the communication of type 2 diabetes (T2D) genetic risk alone has limited effectiveness on facilitating behavioral changes among individuals of European descent. Although the use of wearable devices has been associated with changes in behavior, the effects of combining personalized precision medicine with wearable devices on behaviors related to T2D prevention remain unclear. This study aims to assess the novel effects of T2D genetic risk communication and wearable device functions on objectively measured moderate to vigorous physical activity (MVPA) time among East Asian individuals with overweight or obesity.

OBJECTIVE: The objectives of this study are to (1) investigate the effects of communicating T2D genetic risk and (2) examine the effects of combining T2D genetic risk communication with wearable device functions such as step goal setting and activity prompts on objectively measured MVPA time among East Asians with overweight or obesity.

METHODS: In this parallel-group randomized controlled trial, 355 East Asians with overweight or obesity aged between 40 and 60 years are allocated to 1 of 3 groups: 1 control and 2 intervention groups. Blood samples are used for estimation of T2D genetic risk and tested for metabolic risk markers. T2D genetic risk is estimated based on 113 single nucleotide polymorphisms associated with T2D among East Asians using an established method. All 3 groups receive a Fitbit device. Both intervention groups will receive T2D genetic risk estimates along with lifestyle advice, but one of the intervention groups will receive additional Fitbit functions: step goal setting and prompt functions. The intervention materials are delivered weekly using WhatsApp and monthly via email. The primary outcome is MVPA time, which is objectively measured with the built-in accelerometer of the Fitbit Inspire 3 and will be assessed at baseline, immediately after the intervention, 12 months after the intervention, and at 6-month follow-up. Secondary outcomes include other parameters, such as sedentary time, BMI, systolic and diastolic blood pressure, 5 metabolic risk markers, handgrip strength, sleep, activity calories, self-reported physical activity, self-reported fruit and vegetable consumption, smoking status, and psychological variables.

RESULTS: This study was funded in January 2023. Data collection for baseline assessments began in February 2023. Formal data analysis started in April 2025 after the 6-month follow-up assessments were completed.

CONCLUSIONS: To the best of our knowledge, this study will be the first randomized controlled trial to combine T2D genetic risk communication with wearable device functions in any population. Novel findings will be used to inform future lifestyle modification strategies for T2D. We plan to provide a comprehensive report on this study by publishing this analysis plan before the completion of data collection.

TRIAL REGISTRATION: ClinicalTrials.gov NCT05524909; https://www.clinicaltrials.gov/study/NCT05524909.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/65012.

PMID:41191904 | DOI:10.2196/65012

JMIR Hum Factors. 2025 Nov 5;12:e75855. doi: 10.2196/75855.

ABSTRACT

BACKGROUND: Latinas are one of the largest and fastest-growing female ethnic groups in the United States and have high levels of physical inactivity and sedentary behavior (SB), contributing to a disproportionate burden of chronic health conditions. An ecological momentary assessment (EMA) involves the use of smartphone-based data collected in real time to assess health behaviors and outcomes.

OBJECTIVE: We examined the feasibility, validity, and acceptability of an EMA protocol assessing physical activity (PA) and SB in Latina adults.

METHODS: For 7 days, 67 Latinas (average age 39 years, SD = 13.6; n=37, 55.2% earning less than US $50,000/year; n=53, 79.1% foreign-born; and n=49, 73.1% of Mexican or Mexican American origin) completed a signal-contingent EMA protocol with 3 prompts per day and wore an ActiGraph GT3X accelerometer to measure levels of PA and SB. EMA prompts inquired about current behavior, feelings, beliefs, social conditions, and contexts.

RESULTS: Latinas completed 69.7% (892/1279) of EMA prompts. They were more likely to respond to EMA prompts when engaged in more SB (odds ratio [OR] 1.04, 95% CI 1.01-1.06) and less light-intensity PA (OR 0.97, 95% CI 0.94-0.99) in the 30 minutes around the prompt. Accelerometer data validated self-reported occasions of PA and SB via EMA. The majority of participants (>70%) were satisfied with the protocol and expressed interest in participating in future studies.

CONCLUSIONS: EMA is a feasible, valid, and acceptable methodology for capturing movement behaviors among Latinas, which can provide insights into the antecedents and consequences of these behaviors in their daily lives.

PMID:41191874 | DOI:10.2196/75855

JMIR Aging. 2025 Nov 5;8:e79731. doi: 10.2196/79731.

ABSTRACT

BACKGROUND: Given the rapid development of the digital economy and the sustained proliferation of the internet, digital engagement in older adults has garnered mounting attention from the academic community. However, research has yet to systematically examine the impact of digital engagement on sleep in this demographic.

OBJECTIVE: This study aims to examine the association of digital engagement-operationalized as digital access and internet use duration-with the sleep schedules (nocturnal sleep duration, afternoon nap duration, and sleep onset time) of older adults in China, using longitudinal data and robust statistical modeling to explore longitudinal associations and potential mechanisms.

METHODS: Data were derived from 4 waves (2014, 2016, 2018, and 2020) of the China Family Panel Studies, involving 16,784 older adults (≥60 y). We used panel fixed effects models and a random-effects ordered logit model to analyze the effects on continuous outcomes (nocturnal and nap sleep duration), controlling for time-invariant individual characteristics. As sleep onset time is an ordinal variable, a random-effects ordered logit model was used for this outcome. Moderation analyses were conducted by introducing interaction terms (digital engagement×sex and digital engagement×residence) into the models to examine heterogeneity across subgroups (urban or rural, men or women). Mediation analyses were performed using the Sobel test with year-fixed effects and the nonparametric bootstrap method (1000 resamples) to assess the significance of indirect effects via mechanistic pathways (nonfarm employment, protein intake, memory, depressive mood, and instrumental activities of daily living).

RESULTS: The study included a total of 16,784 older adults, with an average age of 69 (SE 6.946) years, including 9100 (54.22%) women and 7684 (45.78%) men. The results showed that both digital access (β=-.15, 95% CI -.25 to -.06; P=.002) and internet use time (β=-.07, 95% CI -.13 to -.01; P=.027) were significantly associated with significantly shorter sleep duration of older adults. Digital access was significantly associated with a significant reduction in the length of afternoon naps among older adults, while internet use did not have this effect; both digital access and internet use were significantly associated with a significant delay in older adults’ sleep onset time. Digital access was associated with older adults’ sleep schedules through its correlations with nonfarm employment, protein intake, memory, depressive mood, and instrumental activities of daily living. Digital access had a greater and more significant impact on men and urban older adults, while internet use had a greater and more significant impact on women and urban older adults.

CONCLUSIONS: The study indicates that digital engagement, such as the use of electronic devices, is associated with a reduction in both daily and nap sleep duration, as well as a delay in sleep onset, among older adults.

PMID:41191871 | DOI:10.2196/79731

JMIR Cardio. 2025 Nov 5;9:e68022. doi: 10.2196/68022.

ABSTRACT

BACKGROUND: Remote care technology has been used to bridge the gap between health care in a clinical setting and in the community, all the more essential post-COVID. Patients with chronic conditions may benefit from interventions that could provide more continuous and frequent monitoring of their disease process and support self-management. A common barrier, however, is the lack of engagement with technological interventions or devices that provide care remotely, which could lead to loss of resources invested and reduced quality of care.

OBJECTIVE: This discrete choice experiment elicits the preferences of patients with heart failure with regard to potential remote care technologies that they would be willing to engage with and, in turn, creates a hierarchy of factors that can affect engagement for use within future technology design.

METHODS: A survey was created using discrete choice design and with input from a patient and public involvement group. It was distributed online via social media to patients with heart failure and to patient support groups. The attributes used for the experiment were based on a previous systematic review looking at factors that affect engagement in remote care and which generated five central themes, each of which was assigned to an attribute directly: communication (increasing interaction between patients and health care staff/carers/other patients), clinical care (improving the quality of care compared to established practice), education (providing tailored information to help with self-care and reduce uncertainty), ease of use (the technical aspects of the intervention are easy to handle without issues), and convenience (the intervention fits well around the patient’s lifestyle and requires minimal effort). Each of the five themes had two levels, positive and negative. The survey presented participants with multiple forced-choice two-alternative scenarios of remote care, which allowed them to trade attributes according to their preference. The results were analyzed using binary logit to obtain preference weights for each attribute.

RESULTS: A total of 93 completed responses were entered into the analysis. The results of the binary logit created coefficients for each attribute, which equated to the relative preference of the associated themes: clinical care, 2.022; education, 1.252; convenience, 1.245; ease of use, 1.155; communication, 1.040. All calculated coefficients were statistically significant (P<.01).

CONCLUSIONS: The results show that, in this cohort of patients with heart failure, the most preferred factor, clinical care, has enough value to be traded for approximately any two other factors. It also shows that the factor of communication is the least preferred attribute. Technology designers can use the associated preference weights to determine the relative increase of value perceived by patients by adding in certain attributes, with the greatest gains achieved by prioritizing clinical care. This would result in increased engagement in a chronic heart failure population that would benefit most from remote care.

PMID:41191865 | DOI:10.2196/68022