Search

Modifications to Fried’s frailty phenotype (FFP) are common. We evaluated a self-reported modified frailty phenotype (Mod-FP) used among people with HIV (PWH). Among 522 PWH engaged in two longitudinal studies, we assessed validity of the four-item Mod-FP compared with the five-item FFP. We compared the phenotypes via receiver operator characteristic curves, agreement in classifying frailty, and criterion validity via association with having experienced falls. Mod-FP classified 8% of PWH as frail, whereas FFP classified 9%. The area under the receiver operator characteristic curve for Mod-FP classifying frailty was 0.93 (95% CI = 0.91–0.96). We observed kappa ranging from 0.64 (unweighted) to 0.75 (weighted) for categorizing frailty status. Both definitions found frailty associated with a greater odds of experiencing a fall; FFP estimated a slightly greater magnitude (i.e., OR) for the association than Mod-FP. The Mod-FP has good performance in measuring frailty among PWH and is reasonable to use when the gold standards of observed assessments (i.e., weakness and slowness) are not feasible.

ImportanceDiscussions about goals of care are important for high-quality palliative care yet are often lacking for hospitalized older patients with serious illness.ObjectiveTo evaluate a communication-priming intervention to promote goals-of-care discussions between clinicians and hospitalized older patients with serious illness.Design, Setting, and ParticipantsA pragmatic, randomized clinical trial of a clinician-facing communication-priming intervention vs usual care was conducted at 3 US hospitals within 1 health care system, including a university, county, and community hospital. Eligible hospitalized patients were aged 55 years or older with any of the chronic illnesses used by the Dartmouth Atlas project to study end-of-life care or were aged 80 years or older. Patients with documented goals-of-care discussions or a palliative care consultation between hospital admission and eligibility screening were excluded. Randomization occurred between April 2020 and March 2021 and was stratified by study site and history of dementia.InterventionPhysicians and advance practice clinicians who were treating the patients randomized to the intervention received a 1-page, patient-specific intervention (Jumpstart Guide) to prompt and guide goals-of-care discussions.Main Outcomes and MeasuresThe primary outcome was the proportion of patients with electronic health record–documented goals-of-care discussions within 30 days. There was also an evaluation of whether the effect of the intervention varied by age, sex, history of dementia, minoritized race or ethnicity, or study site.ResultsOf 3918 patients screened, 2512 were enrolled (mean age, 71.7 [SD, 10.8] years and 42% were women) and randomized (1255 to the intervention group and 1257 to the usual care group). The patients were American Indian or Alaska Native (1.8%), Asian (12%), Black (13%), Hispanic (6%), Native Hawaiian or Pacific Islander (0.5%), non-Hispanic (93%), and White (70%). The proportion of patients with electronic health record–documented goals-of-care discussions within 30 days was 34.5% (433 of 1255 patients) in the intervention group vs 30.4% (382 of 1257 patients) in the usual care group (hospital- and dementia-adjusted difference, 4.1% [95% CI, 0.4% to 7.8%]). The analyses of the treatment effect modifiers suggested that the intervention had a larger effect size among patients with minoritized race or ethnicity. Among 803 patients with minoritized race or ethnicity, the hospital- and dementia-adjusted proportion with goals-of-care discussions was 10.2% (95% CI, 4.0% to 16.5%) higher in the intervention group than in the usual care group. Among 1641 non-Hispanic White patients, the adjusted proportion with goals-of-care discussions was 1.6% (95% CI, −3.0% to 6.2%) higher in the intervention group than in the usual care group. There was no evidence of differential treatment effects of the intervention on the primary outcome by age, sex, history of dementia, or study site.Conclusions and RelevanceAmong hospitalized older adults with serious illness, a pragmatic clinician-facing communication-priming intervention significantly improved documentation of goals-of-care discussions in the electronic health record, with a greater effect size in racially or ethnically minoritized patients.Trial RegistrationClinicalTrials.gov Identifier: NCT04281784

CONTEXT: Documented goals-of-care discussions are an important quality metric for patients with serious illness. Natural language processing (NLP) is a promising approach for identifying goals-of-care discussions in the electronic health record (EHR). OBJECTIVES: To compare three NLP modeling approaches for identifying EHR documentation of goals-of-care discussions and generate hypotheses about differences in performance. METHODS: We conducted a mixed-methods study to evaluate performance and misclassification for three NLP featurization approaches modeled with regularized logistic regression: bag-of-words (BOW), rule-based, and a hybrid approach. From a prospective cohort of 150 patients hospitalized with serious illness over 2018–2020, we collected 4,391 inpatient EHR notes; 99 (2.3%) contained documented goals-of-care discussions. We used leave-one-out cross-validation to estimate performance by comparing pooled NLP predictions to human abstractors with receiver-operating-characteristic (ROC) and precision-recall (PR) analyses. We qualitatively examined a purposive sample of 70 NLP-misclassified notes using content analysis to identify linguistic features that allowed us to generate hypotheses underpinning misclassification. RESULTS: All three modeling approaches discriminated between notes with and without goals-of-care discussions (AUC(ROC): BOW, 0.907; rule-based, 0.948; hybrid, 0.965). Precision and recall were only moderate (precision at 70% recall: BOW, 16.2%; rule-based, 50.4%; hybrid, 49.3%; AUC(PR): BOW, 0.505; rule-based, 0.579; hybrid, 0.599). Qualitative analysis revealed patterns underlying performance differences between BOW and rule-based approaches. CONCLUSION: NLP holds promise for identifying EHR-documented goals-of-care discussions. However, the rarity of goals-of-care content in EHR data limits performance. Our findings highlight opportunities to optimize NLP modeling approaches, and support further exploration of different NLP approaches to identify goals-of-care discussions.

Background and Objective Care partners of people living with dementia require support to knowledgeably navigate decision making about how and when to use monitoring technologies for care purposes. We conducted a pilot study of a novel self-administered intervention, “Let’s Talk Tech,” for people living with mild dementia and their care partners. This paper presents preliminary efficacy findings of this intervention designed to educate and facilitate dyadic communication about a range of technologies used in dementia care and to document the preferences of the person living with dementia. It is the first-of-its-kind decision-making and planning tool with a specific focus on technology use. Research Design and Methods We used a 1-group pretest–post-test design and paired t tests to assess change over 2 time periods in measures of technology comprehension, care partner knowledge of the participant living with mild Alzheimer’s disease’s (AD) preferences, care partner preparedness to make decisions about technology use, and mutual understanding. Thematic analysis was conducted on postintervention interview transcripts to elucidate mechanisms and experiences with Let’s Talk Tech. Results Twenty-nine mild AD dementia care dyads who live together completed the study. There was statistically significant improvement with medium and large effect sizes on outcome measures of care partners’ understanding of each technology, care partners’ perceptions of the person living with dementia’s understanding of each technology, knowledge of the person living with dementia’s preferences, decision-making preparedness, and care partners’ feelings of mutual understanding. Participants reported that it helped them have important and meaningful conversations about using technology. Discussion and Implications Let’s Talk Tech demonstrated promising preliminary efficacy on targeted measures that can lead to informed, shared decision making about technologies used in dementia care. Future studies should assess efficacy with larger samples and more diverse sample populations in terms of race, ethnicity, and dementia type.

Making EHR Data More Available for Research and Public Health (MedMorph) is a Centers for Disease Control and Prevention-led initiative developing and demonstrating a reference architecture (RA) and implementation, including Health Level Seven International Fast Healthcare Interoperability Resources (HL7 FHIR) implementation guides (IGs), describing how to leverage FHIR for aligned research and public health access to clinical data for automated data exchange. MedMorph engaged a technical expert panel of more than 100 members to model representative use cases, develop IGs (architectural and content), align with existing efforts in the FHIR community, and demonstrate the RA in research and public health uses. The RA IG documents common workflows needed to automatically send research data to Research Patient Data Repositories for multiple use cases. Sharing a common RA and canonical data model will improve data sharing for research and public health needs and generate evidence. MedMorph delivers a robust, reusable method to utilize data from electronic health records addressing multiple research and public health needs.

Duncan Short,1 Xueqi Wang,2 Shivali Suri,3 Thomas K Hsu,2 Bryn Jones,1 Rob J Fredericksen,4 Heidi M Crane,4 Alexandra Musten,5 Jean Bacon,5 Yongwei Wang,2 Kevin A Gough,3 Moti Ramgopal,6 Jeff Berry,7 William B Lober4 1ViiV Healthcare, Brentford, UK; 2ViiV Healthcare, Durham, NC, USA; 3St Michael’s Hospital, Toronto, ON, Canada; 4Center for AIDS Research, University of Washington, Seattle, WA, USA; 5Ontario HIV Treatment Network, Toronto, ON, Canada; 6Midway Specialty Care Center, Fort Pierce, FL, USA; 7TPAN, Chicago, IL, USACorrespondence: Duncan Short, ViiV Healthcare, 980 Great West Road, Brentford, Middlesex, TW8 9GS, UK, Tel +44 7827 282971, Email duncan.x.short@viivhealthcare.comPurpose: Use of patient-reported outcomes assessments (PROs) can improve patient–provider communication and focus provider attention on current health issues. This analysis examines the association between suboptimal antiretroviral therapy (ART) adherence and factors obtained through PROs among people with HIV (PWH) at 2 North American outpatient clinics.Patients and Methods: Immediately before a clinic visit, PWH completed self-administered PROs. Unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were estimated from logistic regression models to identify sociodemographic and health-related factors (satisfaction with ART, difficulty meeting housing costs, depression, intimate partner violence, risk of malnutrition, smoking status, alcohol use, and substance use) associated with suboptimal adherence (defined as self-reporting < 95% or < 80% adherence). Multiple imputation was performed to account for missing data in the multivariate analyses.Results: Of 1632 PWH, 1239 (76%) responded to the adherence assessment; of these, 268 (22%) and 106 (9%) reported < 95% and < 80% adherence, respectively. Of 1580 PWH who responded, 354 (22%) were dissatisfied with their HIV medication. Of responding PWH, 19% reported moderate-to-severe depression, 23% indicated they were at risk of malnutrition, 34% were current smokers, and 62% reported substance use in the past 3 months. Dissatisfaction with ART was significantly associated with < 95% and < 80% adherence in the unadjusted analysis (unadjusted OR [95% CI], 3.38 [2.51– 4.56] and 4.26 [2.82– 6.42], respectively) and adjusted analysis (adjusted OR [95% CI], 2.76 [1.91– 4.00] and 3.28 [1.95– 5.52], respectively); significance remained after multiple imputation. In adjusted analyses, no risk of malnutrition was significantly associated with reduced odds of < 95% adherence after multiple imputation (adjusted OR [95% CI], 0.714 [0.511– 0.997]); no other factors were associated with < 95% or < 80% adherence.Conclusion: These results suggest that implementation of PROs evaluating treatment satisfaction may provide value to adherence management in routine HIV care.Keywords: patient satisfaction, antiretroviral therapy, highly active, treatment adherence, implementation science, quality of life

Goals-of-care discussions are important for patient-centered care among hospitalized patients with serious illness. However, there are little data on the occurrence, predictors, and timing of these discussions.To examine the occurrence, predictors, and timing of electronic health record (EHR)-documented goals-of-care discussions for hospitalized patients.This retrospective cohort study used natural language processing (NLP) to examine EHR-documented goals-of-care discussions for adults with chronic life-limiting illness or age ≥80 hospitalized 2015-2019. The primary outcome was NLP-identified documentation of a goals-of-care discussion during the index hospitalization. We used multivariable logistic regression to evaluate associations with baseline characteristics.Of 16,262 consecutive, eligible patients without missing data, 5,918 (36.4%) had a documented goals-of-care discussion during hospitalization; approximately 57% of these discussions occurred within 24 hours of admission. In multivariable analysis, documented goals-of-care discussions were more common for women (OR=1.26, 95%CI 1.18-1.36), older patients (OR=1.04 per year, 95%CI 1.03-1.04), and patients with more comorbidities (OR=1.11 per Deyo-Charlson point, 95%CI 1.10-1.13), cancer (OR=1.88, 95%CI 1.72-2.06), dementia (OR=2.60, 95%CI 2.29-2.94), higher acute illness severity (OR=1.12 per National Early Warning Score point, 95%CI 1.11-1.14), or prior advance care planning documents (OR=1.18, 95%CI 1.08-1.30). Documentation of these discussions was less common for racially or ethnically minoritized patients (OR=0.823, 95%CI 0.75-0.90).Among hospitalized patients with serious illness, documented goals-of-care discussions identified by NLP were more common among patients with older age and increased burden of acute or chronic illness, and less common among racially or ethnically minoritized patients. This suggests important disparities in goals-of-care discussions.

Substance use in the U.S. varies by geographic region. Opioid prescribing practices and marijuana, heroin, and methamphetamine availability are evolving differently across regions. We examined self-reported substance use among people living with HIV (PLWH) in care at seven sites from 2017-2019 to understand current regional substance use patterns. We calculated the percentage and standardized percentage of PLWH reporting current drug use and at-risk and binge alcohol use by U.S. Census Bureau geographic region and examined associations in adjusted logistic regression analyses. Among 7,686 PLWH, marijuana use was the most prevalent drug (30%), followed by methamphetamine/crystal (8%), cocaine/crack (7%), and illicit opioids (3%). One-third reported binge alcohol use (32%). Differences in percent of current use by region were seen for marijuana (24-41%) and methamphetamine/crystal (2-15%), with more use in the West and Northeast, and binge alcohol use (26-40%). In adjusted analyses, PLWH in the Midwest were significantly less likely to use methamphetamine/crystal (aOR: 0.13;0.06-0.25) or illicit opioids (aOR:0.16;0.05-0.53), and PLWH in the Northeast were more likely to use cocaine/crack (aOR:1.59;1.16-2.17), compared to PLWH in the West. Understanding differences in substance use patterns in the current era, as policies continue to evolve, will enable more targeted interventions in clinical settings among PLWH.

BACKGROUND In an effort to increase vaccination coverage in low-resource settings, digital tools have been introduced to better track immunization records, improve data management practices, and provide improved access to vaccination coverage data for decision-making. Despite the potential of these electronic systems to improve the provision of health services, few digital health interventions have been institutionalized at scale in low- and middle-income countries. OBJECTIVE In this paper, we aimed to describe how health care workers in Kenya had integrated an electronic immunization registry into their immunization clinic workflows and to use these findings to inform the development of a refined program theory on the registry’s usability. METHODS Informed by realist methodology, we developed a program theory to explain usability of the electronic immunization registry. We designed a qualitative study based on our theory to describe the barriers and facilitators influencing data entry and use. Qualitative data were collected through semistructured interviews with users and workflow observations of immunization clinic sessions. Our findings were summarized by context-mechanism-outcome relationships formed after analyzing our key themes across interviews and workflow observations. Using these relationships, we were able to identify common rules for future implementers. RESULTS Across the 12 facilities included in our study, 19 health care workers were interviewed, and 58 workflow sessions were observed. The common rules developed from our qualitative findings are as follows: rule 1—ensure that the users complete training to build familiarity with the system, understand the value of the system and data, and know where to find support; rule 2—confirm that the system captures all data needed for users to provide routine health care services and is easy to navigate; rule 3—identify work-arounds for poor network, system performance, and too few staff or resources; and rule 4—make users aware of expected changes to their workflow, and how these changes might differ over time and by facility size or number of patients. Upon study completion, we revised the program theory to reflect the importance of the goals and workflows of electronic immunization registries aligning with reality. CONCLUSIONS We created a deeper understanding of the underlying mechanisms for usability of the registry. We found that the electronic immunization registry had high acceptability among users; however, there were numerous barriers to using the system, even under ideal conditions, causing a misalignment between the system and the reality of the users’ workflows and their environment. Human-centered design and human-factors methods can assist during pilot stages to better align systems with users’ needs and again after scale-up to ensure that interventions are suitable for all user settings.

We assessed acceptability/usability of tablet-based patient-reported outcome (PRO) assessments among patients in HIV care, and relationships with health outcomes using a modified version of the 6-item Acceptability E-Scale (AES) within a self-administered PRO assessment. Using multivariable linear regression, we measured associations between patient characteristics and continuous combined AES score. Among 786 patients (median age=48; 91% male; 49% white; 17% Spanish-speaking) overall mean score was 26/30 points (SD: 4.4). Mean scores per dimension (max 5, 1=lowest acceptability, 5=highest): ease of use 4.7, understandability 4.7, time burden 4.3, overall satisfaction 4.3, helpfulness describing symptoms/behaviors 4.2, and enjoyability 3.8. Higher overall score was associated with race/ethnicity (+1.3 points/African-American patients (95%CI:0.3-2.3); +1.6 points/Latino patients (95%CI:0.9-2.3) compared to white patients). Patients completing PROs in Spanish scored +2.4 points on average (95%CI:1.6-3.3). Higher acceptability was associated with better quality of life (0.3 points (95%CI:0.2-0.5)) and adherence (0.4 points (95%CI:0.2-0.6)). Lower acceptability was associated with: higher depression symptoms (−0.9 points (95%CI:-1.4 to −0.4)); recent illicit opioid use (−2.0 points (95%CI:-3.9 to −0.2)); multiple recent sex partners (−0.8 points (95%CI:−1.5 to −0.1)). While patients endorsing depression symptoms, recent opioid use, condomless sex, or multiple sex partners found PROs to be less acceptable, overall, patients found self-administered, tablet-based PRO assessments to be highly acceptable and easy to use.

Introduction: Many patients use mobile devices to track health conditions by recording patient-generated health data. However, patients and clinicians may disagree how to use these data. Objective: To systematically review the literature to identify how patient-generated health data and patient-reported outcomes collected outside of clinical settings can affect patient–clinician relationships within surgery and primary care. Methods: Six research databases were queried for publications documenting the effect of patient-generated health data or patient-reported outcomes on patient–clinician relationships. We conducted thematic synthesis of the results of the included publications. Results: Thirteen of the 3204 identified publications were included for synthesis. Three main themes were identified: patient-generated health data supported patient–clinician communication and health awareness, patients desired for their clinicians to be involved with their patient-generated health data, which clinicians had difficulty accommodating, and patient-generated health data platform features may support or hinder patient–clinician collaboration. Conclusion: Patient-generated health data and patient-reported outcomes may improve patient health awareness and communication with clinicians but may negatively affect patient–clinician relationships.

BackgroundPeople with HIV (PWH) are at increased risk of cardiovascular comorbidities and substance use is a potential predisposing factor. We evaluated associations of tobacco smoking and alcohol use with venous thromboembolism (VTE) in PWH.MethodsWe assessed incident, centrally adjudicated VTE among 12957 PWH within the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort between January 2009 and December 2018. Using separate Cox proportional hazards models, we evaluated associations of time-updated alcohol and cigarette use with VTE, adjusting for demographic and clinical characteristics. Smoking was evaluated as pack-years and never, former, or current use with current cigarettes per day. Alcohol use was parameterized using categorical and continuous alcohol use score, frequency of use, and binge frequency.ResultsDuring a median of 3.6years of follow-up, 213 PWH developed a VTE. One-third of PWH reported binge drinking and 40% reported currently smoking. In adjusted analyses, risk of VTE was increased among both current (HR: 1.44, 95% CI: 1.02-2.03) and former (HR: 1.44, 95% CI: 0.99-2.07) smokers compared to PWH who never smoked. Additionally, total pack-years among ever-smokers (HR: 1.10 per 5 pack-years; 95% CI: 1.03-1.18) was associated with incident VTE in a dose-dependent manner. Frequency of binge drinking was associated with incident VTE (HR: 1.30 per 7days/month, 95% CI: 1.11-1.52); however, alcohol use frequency was not. Severity of alcohol use was not significantly associated with VTE.ConclusionsCurrent smoking and pack-year smoking history were dose-dependently associated with incident VTE among PWH in CNICS. Binge drinking was also associated with VTE. Interventions for smoking and binge drinking may decrease VTE risk among PWH.

ImportanceMany clinical trial outcomes are documented in free-text electronic health records (EHRs), making manual data collection costly and infeasible at scale. Natural language processing (NLP) is a promising approach for measuring such outcomes efficiently, but ignoring NLP-related misclassification may lead to underpowered studies.ObjectiveTo evaluate the performance, feasibility, and power implications of using NLP to measure the primary outcome of EHR-documented goals-of-care discussions in a pragmatic randomized clinical trial of a communication intervention.Design, Setting, and ParticipantsThis diagnostic study compared the performance, feasibility, and power implications of measuring EHR-documented goals-of-care discussions using 3 approaches: (1) deep-learning NLP, (2) NLP-screened human abstraction (manual verification of NLP-positive records), and (3) conventional manual abstraction. The study included hospitalized patients aged 55 years or older with serious illness enrolled between April 23, 2020, and March 26, 2021, in a pragmatic randomized clinical trial of a communication intervention in a multihospital US academic health system.Main Outcomes and MeasuresMain outcomes were natural language processing performance characteristics, human abstractor-hours, and misclassification-adjusted statistical power of methods of measuring clinician-documented goals-of-care discussions. Performance of NLP was evaluated with receiver operating characteristic (ROC) curves and precision-recall (PR) analyses and examined the effects of misclassification on power using mathematical substitution and Monte Carlo simulation.ResultsA total of 2512 trial participants (mean [SD] age, 71.7 [10.8] years; 1456 [58%] female) amassed 44 324 clinical notes during 30-day follow-up. In a validation sample of 159 participants, deep-learning NLP trained on a separate training data set identified patients with documented goals-of-care discussions with moderate accuracy (maximal F1 score, 0.82; area under the ROC curve, 0.924; area under the PR curve, 0.879). Manual abstraction of the outcome from the trial data set would require an estimated 2000 abstractor-hours and would power the trial to detect a risk difference of 5.4% (assuming 33.5% control-arm prevalence, 80% power, and 2-sided α = .05). Measuring the outcome by NLP alone would power the trial to detect a risk difference of 7.6%. Measuring the outcome by NLP-screened human abstraction would require 34.3 abstractor-hours to achieve estimated sensitivity of 92.6% and would power the trial to detect a risk difference of 5.7%. Monte Carlo simulations corroborated misclassification-adjusted power calculations.Conclusions and RelevanceIn this diagnostic study, deep-learning NLP and NLP-screened human abstraction had favorable characteristics for measuring an EHR outcome at scale. Adjusted power calculations accurately quantified power loss from NLP-related misclassification, suggesting that incorporation of this approach into the design of studies using NLP would be beneficial.

Background Digital health interventions (DHI) have the potential to improve the management and utilization of health information to optimize health care worker performance and provision of care. Despite the proliferation of DHI projects in low-and middle-income countries, few have been evaluated in an effort to understand their impact on health systems and health-related outcomes. Although more evidence is needed on their impact and effectiveness, the use of DHIs among immunization programs has become more widespread and shows promise for improving vaccination uptake and adherence to immunization schedules. Methods Our aim was to assess the impact of an electronic immunization registry (EIR) using an interrupted time-series analysis to analyze the effect on proportion of on-time vaccinations following introduction of an EIR in Tanzania. We hypothesized that the introduction of the EIR would lead to statistically significant changes in vaccination timeliness at 3, 6, and > 6 months post-introduction. Results For our primary analysis, we observed a decrease in the proportion of on-time vaccinations following EIR introduction. In contrast, our sensitivity analysis estimated improvements in timeliness among those children with complete vaccination records. However, we must emphasize caution interpreting these findings as they are likely affected by implementation challenges. Conclusions This study highlights the complexities of using digitized individual-level routine health information system data for evaluation and research purposes. EIRs have the potential to improve vaccination timeliness, but analyses using EIR data can be complicated by data quality issues and inconsistent data entry leading to difficulties interpreting findings.

Background: An optimal adherence to antiretroviral therapy (ART) is fundamental for suppression of HIV viral load and favourable treatment outcomes. Patient-reported outcomes (PROs) are effective tools for improving patient–provider communication and focusing providers’ awareness on current health problems. The objectives of this analysis were (1) to determine the feasibility of implementing an electronic screening tool to measure PROs in a Canadian HIV clinic to obtain information on ART adherence and related factors and (2) to determine the factors related to sub-optimal adherence. Methods: This implementation research with a convenience sample of 600 people living with HIV (PLWH) was conducted in a busy, academic, urban HIV clinic in Toronto, Canada. PLWH were approached to participate in PRO assessments just prior to their in-clinic appointments, including health-related domains such as mental health, housing, nutrition, financial stress and medication adherence, and responses were summarized on a single sheet available for providers to review. Feasibility of implementing PROs was assessed by quantifying response rate, completion rate, time taken and participation rate. Medication adherence was elicited by self-report of the percentage of prescribed HIV medications taken in the last month. Unadjusted and adjusted odds ratios were estimated from logistic regression models to identify factors associated with adherence of