Food Insecurity and Hypoglycemia among Older Patients with Type 2 Diabetes Treated with Insulin or Sulfonylureas: The Diabetes & Aging Study.

BACKGROUND: Severe hypoglycemia is a serious adverse drug event associated with hypoglycemia-prone medications; older patients with diabetes are particularly at high risk. Economic food insecurity (food insecurity due to financial limitations) is a known risk factor for hypoglycemia; however, less is known about physical food insecurity (due to difficulty cooking or shopping for food), which may increase with age, and its association with hypoglycemia. OBJECTIVE: Study associations between food insecurity and severe hypoglycemia. DESIGN: Survey based cross-sectional study. PARTICIPANTS: Survey responses were collected in 2019 from 1,164 older (≥ 65 years) patients with type 2 diabetes treated with insulin or sulfonylureas. MAIN MEASURES: Risk ratios (RR) for economic and physical food insecurity associated with self-reported severe hypoglycemia (low blood glucose requiring assistance) adjusted for age, financial strain, HbA1c, Charlson comorbidity score and frailty. Self-reported reasons for hypoglycemia endorsed by respondents. KEY RESULTS: Food insecurity was reported by 12.3% of the respondents; of whom 38.4% reported economic food insecurity only, 21.1% physical food insecurity only and 40.5% both. Economic food insecurity and physical food insecurity were strongly associated with severe hypoglycemia (RR = 4.3; p = 0.02 and RR = 4.4; p = 0.002, respectively). Missed meals ("skipped meals, not eating enough or waiting too long to eat") was the dominant reason (77.5%) given for hypoglycemia. CONCLUSIONS: Hypoglycemia prevention efforts among older patients with diabetes using hypoglycemia-prone medications should address food insecurity. Standard food insecurity questions, which are used to identify economic food insecurity, will fail to identify patients who have physical food insecurity only.

Willingness to take less medication for type 2 diabetes among older patients: The Diabetes & Aging Study.

BACKGROUND: To examine the willingness of older patients to take less diabetes medication (de-intensify) and to identify characteristics associated with willingness to de-intensify treatment. METHODS: Survey conducted in 2019 in an age-stratified, random sample of older (65-100 years) adults with diabetes on glucose-lowering medications in the Kaiser Permanente Northern California Diabetes Registry. We classified survey responses to the question: "I would be willing to take less medication for my diabetes" as willing, neutral, or unwilling to de-intensify. Willingness to de-intensify treatment was examined by several clinical characteristics, including American Diabetes Association (ADA) health status categories used for individualizing glycemic targets. Analyses were weighted to account for over-sampling of older individuals. RESULTS: A total of 1337 older adults on glucose-lowering medication(s) were included (age 74.2 ± 6.0 years, 44% female, 54.4% non-Hispanic white). The proportions of participants willing, neutral, or unwilling to take less medication were 51.2%, 27.3%, and 21.5%, respectively. Proportions of willing to take less medication varied by age (65-74 years: 54.2% vs. 85+ years: 38.5%) and duration of diabetes (0-4 years: 61.0% vs. 15+ years: 44.2%), both p < 0.001. Patients on 1-2 medications were more willing to take less medication(s) compared with patients on 10+ medications (62.1% vs. 46.6%, p = 0.03). Similar proportions of willingness to take less medications were seen across ADA health status, and HbA1c. Willingness to take less medication(s) was similar across survey responses to questions about patient-clinician relationships. CONCLUSIONS: Clinical guidelines suggest considering treatment de-intensification in older patients with longer duration of diabetes, yet patients with these characteristics are less likely to be willing to take less medication(s).

Care Partner Engagement in Secure Messaging Between Patients With Diabetes and Their Clinicians: Cohort Study.

BACKGROUND: Patient engagement with secure messaging (SM) via digital patient portals has been associated with improved diabetes outcomes, including increased patient satisfaction and better glycemic control. Yet, disparities in SM uptake exist among older patients and racial and ethnic underserved groups. Care partners (family members or friends) may provide a means for mitigating these disparities; however, it remains unclear whether and to what extent care partners might enhance SM use. OBJECTIVE: We aim to examine whether SM use differs among older patients with diabetes based on the involvement of care partner proxies. METHODS: This is a substudy of the ECLIPPSE (Employing Computational Linguistics to Improve Patient-Provider Secure Emails) project, a cohort study taking place in a large, fully integrated health care delivery system with an established digital patient portal serving over 4 million patients. Participants included patients with type 2 diabetes aged ≥50 years, newly registered on the patient portal, who sent ≥1 English-language message to their clinician between July 1, 2006, and December 31, 2015. Proxy SM was identified by having a registered proxy. To identify nonregistered proxies, a computational linguistics algorithm was applied to detect words and phrases more likely to appear in proxy messages compared to patient-authored messages. The primary outcome was the annual volume of secure messages (sent or received); secondary outcomes were the length of time to the first SM sent by patient or proxy and the number of annual SM exchanges (unique message topics generating ≥1 reply). RESULTS: The mean age of the cohort (N=7659) at this study's start was 61 (SD 7.16) years; 75% (n=5573) were married, 15% (n=1089) identified as Black, 10% (n=747) Chinese, 12% (n=905) Filipino, 13% (n=999) Latino, and 30% (n=2225) White. Further, 49% (n=3782) of patients used a proxy to some extent. Compared to nonproxy users, proxy users were older (P<.001), had lower educational attainment (P<.001), and had more comorbidities (P<.001). Adjusting for patient sociodemographic and clinical characteristics, proxy users had greater annual SM volume (20.7, 95% CI 20.2-21.2 vs 10.9, 95% CI 10.7-11.2; P<.001), shorter time to SM initiation (hazard ratio vs nonusers: 1.30, 95% CI 1.24-1.37; P<.001), and more annual SM exchanges (6.0, 95% CI 5.8-6.1 vs 2.9, 95% CI 2.9-3.0, P<.001). Differences in SM engagement by proxy status were similar across patient levels of education, and racial and ethnic groups. CONCLUSIONS: Among a cohort of older patients with diabetes, proxy SM involvement was independently associated with earlier initiation and increased intensity of messaging, although it did not appear to mitigate existing disparities in SM. These findings suggest care partners can enhance patient-clinician telecommunication in diabetes care. Future studies should examine the effect of care partners' SM involvement on diabetes-related quality of care and clinical outcomes.

Severe hypoglycemia and falls in older adults with diabetes: The Diabetes & Aging Study.

Objective: To estimate rates of severe hypoglycemia and falls among older adults with diabetes and evaluate their association. Research Design and Methods: Survey in an age-stratified, random sample adults with diabetes age 65-100 years; respondents were asked about severe hypoglycemia (requiring assistance) and falls in the past 12 months. Prevalence ratios (adjusted for age, sex, race/ethnicity) estimated the increased risk of falls associated with severe hypoglycemia. Results: Among 2,158 survey respondents, 79 (3.7%) reported severe hypoglycemia, of whom 68 (86.1%) had no ED visit or hospitalization for hypoglycemia. Falls were reported by 847 (39.2%), of whom 745 (88.0%) had no fall documented in outpatient or inpatient records. Severe hypoglycemia was associated with a 70% greater prevalence of falls (adjusted prevalence ratio = 1.7 (95% CI, 1.3-2.2)). Conclusion: While clinical documentation of events likely reflects severity or care-seeking behavior, severe hypoglycemia and falls are common, under-reported life-threatening events.

Glycemic control and diabetes complications across health status categories in older adults treated with insulin or insulin secretagogues: The Diabetes & Aging Study.

BACKGROUND: For older adults with type 2 diabetes (T2D) treated with insulin or sulfonylureas, Endocrine Society guideline recommends HbA1c between 7% to <7.5% for those in good health, 7.5% to <8% for those in intermediate health, and 8% to <8.5% for those in poor health. Our aim was to examine associations between attained HbA1c below, within (reference), or above recommended target range and risk of complication or mortality. METHODS: Retrospective cohort study of adults ≥65 years old with T2D treated with insulin or sulfonylureas from an integrated healthcare delivery system. Cox proportional hazards models of complications during 2019 were adjusted for sociodemographic and clinical variables. Primary outcome was a combined outcome of any microvascular or macrovascular event, severe hypoglycemia, or mortality during 12-month follow-up. RESULTS: Among 63,429 patients (mean age: 74.2 years, 46.8% women), 8773 (13.8%) experienced a complication. Complication risk was significantly elevated for patients in good health (n = 16,895) whose HbA1c was above (HR 1.97, 95% CI 1.62-2.41) or below (HR 1.29, 95% CI 1.02-1.63) compared to within recommended range. Among those in intermediate health (n = 30,129), complication risk was increased for those whose HbA1c was above (HR 1.45, 95% CI 1.30-1.60) but not those below the recommended range (HR 0.99, 95% CI 0.89-1.09). Among those in poor health (n = 16,405), complication risk was not significantly different for those whose HbA1c was below (HR 0.98, 95% CI 0.89-1.09) or above (HR 0.96, 95% CI 0.88-1.06) recommended range. CONCLUSIONS: For older adults with T2D in good health, HbA1c below or above the recommended range was associated with significantly elevated complication risk. However, for those in poor health, achieving specific HbA1c levels may not be helpful in reducing the risk of complications.

Data-driven classification of health status of older adults with diabetes: The diabetes and aging study.

BACKGROUND: We set out to identify empirically-derived health status classes of older adults with diabetes based on clusters of comorbid conditions which are associated with future complications. METHODS: We conducted a cohort study among 105,786 older (≥65 years of age) adults with type 2 diabetes enrolled in an integrated healthcare delivery system. We used latent class analysis of 19 baseline comorbidities to derive health status classes and then compared incident complication rates (events per 100 person-years) by health status class during 5 years of follow-up. Complications included infections, hyperglycemic events, hypoglycemic events, microvascular events, cardiovascular events, and all-cause mortality. RESULTS: Three health status classes were identified: Class 1 (58% of the cohort) had the lowest prevalence of most baseline comorbidities, Class 2 (22%) had the highest prevalence of obesity, arthritis, and depression, and Class 3 (20%) had the highest prevalence of cardiovascular conditions. The risk for incident complications was highest for Class 3, intermediate for Class 2 and lowest for Class 1. For example, the age, sex and race-adjusted rates for cardiovascular events (per 100 person-years) for Class 3, Class 2 and Class 1 were 6.5, 2.3, and 1.6, respectively; 2.1, 1.2, 0.7 for hypoglycemia; and 8.0, 3.8, and 2.3 for mortality. CONCLUSIONS: Three health status classes of older adults with diabetes were identified based on prevalent comorbidities and were associated with marked differences in risk of complications. These health status classes can inform population health management and guide the individualization of diabetes care.

Association of the 2020 US Presidential Election With Hospitalizations for Acute Cardiovascular Conditions.

Importance: Prior studies found a higher risk of acute cardiovascular disease (CVD) around population-wide psychosocial or environmental stressors. Less is known about acute CVD risk in relation to political events. Objective: To examine acute CVD hospitalizations following the 2020 presidential election. Design, Setting, and Participants: This retrospective cohort study examined acute CVD hospitalizations following the 2020 presidential election. Participants were adult members aged 18 years or older at Kaiser Permanente Southern California and Kaiser Permanente Northern California, 2 large, integrated health care delivery systems. Statistical analysis was performed from March to July 2021. Exposure: 2020 US presidential election. Main Outcomes and Measures: Hospitalizations for acute CVD around the 2020 presidential election were examined. CVD was defined as hospitalizations for acute myocardial infarction (AMI), heart failure (HF), or stroke. Rate ratios (RR) and 95% CIs were calculated comparing rates of CVD hospitalization in the 5 days following the 2020 election with the same 5-day period 2 weeks prior. Results: Among 6 396 830 adults (3 970 077 [62.1%] aged 18 to 54 years; 3 422 479 [53.5%] female; 1 083 128 [16.9%] Asian/Pacific Islander, 2 101 367 [32.9%] Hispanic, and 2 641 897 [41.3%] White), rates of hospitalization for CVD following the election (666 hospitalizations; rate = 760.5 per 100 000 person-years [PY]) were 1.17 times higher (95% CI, 1.05-1.31) compared with the same 5-day period 2 weeks prior (569 hospitalizations; rate = 648.0 per 100 000 PY). Rates of AMI were significantly higher following the election (RR, 1.42; 95% CI, 1.13-1.79). No significant difference was found for stroke (RR, 1.02; 95% CI, 0.86-1.21) or HF (RR, 1.18; 95% CI, 0.98-1.42). Conclusions and Relevance: Higher rates of acute CVD hospitalization were observed following the 2020 presidential election. Awareness of the heightened risk of CVD and strategies to mitigate risk during notable political events are needed.

Association of Low-Density Lipoprotein Testing After an Atherosclerotic Cardiovascular Event with Subsequent Statin Adherence and Intensification.

PURPOSE: This study aimed to evaluate associations between outpatient low-density lipoprotein cholesterol (LDL-C) testing and subsequent statin adherence and intensification in patients after an atherosclerotic cardiovascular (ASCVD) event. METHODS: This was a longitudinal study of adult members of Kaiser Permanente Northern California hospitalized with an ASCVD event (myocardial infarction or stroke) during January 01, 2016, to December 31, 2017, with follow-up through December 31, 2019. Outcomes were statin adherence (estimated using continuous medication gap [CMG]) and intensification (defined by an increased dose or switch to a higher-intensity statin) based on pharmacy dispensing. The exposure of interest was first outpatient LDL-C test after an ASCVD event. Baseline for follow-up was LDL-C test date or a date assigned using incidence density sampling. Multivariate logistic regression models were specified to estimate the odds ratios for statin adherence or intensification among those with vs without an LDL-C test, with adjustment for age, sex, race/ethnicity, smoking, hypertension, diabetes, body mass index, and estimated glomerular filtration rate. RESULTS: There were 19,604 adults hospitalized with ASCVD, including 7054 adults not on high-intensity statins. The mean age was 69.5 years and 33.0% were female. Prevalence of good adherence (continuous medication gap ≤20%) was significantly higher (80.2% vs 75.9%; odds ratio 1.38; 95% confidence interval, 1.28-1.49; P <.001) among participants who had an LDL-C test compared with participants who did not. LDL-C testing was associated with significantly higher rates of treatment intensification (16.1% vs 10.7%; odds ratio 1.51; 95% confidence interval,1.29-1.76; P <0.001). CONCLUSIONS: Low-density lipoprotein cholesterol testing is recommended for patients with a history of ASCVD and may be a high-value and low-cost intervention to improve adherence and statin management.

GRASP55 restricts early-stage autophagy and regulates spatial organization of the early secretory network.

There is great interest in understanding the cellular mechanisms controlling autophagy, a tightly regulated catabolic and stress-response pathway. Prior work has uncovered links between autophagy and the Golgi reassembly stacking protein of 55 kDa (GRASP55), but their precise interrelationship remains unclear. Intriguingly, both autophagy and GRASP55 have been functionally and spatially linked to the endoplasmic reticulum (ER)---Golgi interface, broaching this compartment as a site where GRASP55 and autophagy may intersect. Here, we uncover that loss of GRASP55 enhances LC3 puncta formation, indicating that GRASP55 restricts autophagosome formation. Additionally, using proximity-dependent biotinylation, we identify a GRASP55 proximal interactome highly associated with the ER-Golgi interface. Both nutrient starvation and loss of GRASP55 are associated with coalescence of early secretory pathway markers. In light of these findings, we propose that GRASP55 regulates spatial organization of the ER-Golgi interface, which suppresses early autophagosome formation.

Descriptive examination of secure messaging in a longitudinal cohort of diabetes patients in the ECLIPPSE study.

The substantial expansion of secure messaging (SM) via the patient portal in the last decade suggests that it is becoming a standard of care, but few have examined SM use longitudinally. We examined SM patterns among a diverse cohort of patients with diabetes (N = 19 921) and the providers they exchanged messages with within a large, integrated health system over 10 years (2006-2015), linking patient demographics to SM use. We found a 10-fold increase in messaging volume. There were dramatic increases overall and for patient subgroups, with a majority of patients (including patients with lower income or with self-reported limited health literacy) messaging by 2015. Although more physicians than nurses and other providers messaged throughout the study, the distribution of health professions using SM changed over time. Given this rapid increase in SM, deeper understanding of optimizing the value of patient and provider engagement, while managing workflow and training challenges, is crucial.

Employing computational linguistics techniques to identify limited patient health literacy: Findings from the ECLIPPSE study.

OBJECTIVE: To develop novel, scalable, and valid literacy profiles for identifying limited health literacy patients by harnessing natural language processing. DATA SOURCE: With respect to the linguistic content, we analyzed 283 216 secure messages sent by 6941 diabetes patients to physicians within an integrated system's electronic portal. Sociodemographic, clinical, and utilization data were obtained via questionnaire and electronic health records. STUDY DESIGN: Retrospective study used natural language processing and machine learning to generate five unique "Literacy Profiles" by employing various sets of linguistic indices: Flesch-Kincaid (LP_FK); basic indices of writing complexity, including lexical diversity (LP_LD) and writing quality (LP_WQ); and advanced indices related to syntactic complexity, lexical sophistication, and diversity, modeled from self-reported (LP_SR), and expert-rated (LP_Exp) health literacy. We first determined the performance of each literacy profile relative to self-reported and expert-rated health literacy to discriminate between high and low health literacy and then assessed Literacy Profiles' relationships with known correlates of health literacy, such as patient sociodemographics and a range of health-related outcomes, including ratings of physician communication, medication adherence, diabetes control, comorbidities, and utilization. PRINCIPAL FINDINGS: LP_SR and LP_Exp performed best in discriminating between high and low self-reported (C-statistics: 0.86 and 0.58, respectively) and expert-rated health literacy (C-statistics: 0.71 and 0.87, respectively) and were significantly associated with educational attainment, race/ethnicity, Consumer Assessment of Provider and Systems (CAHPS) scores, adherence, glycemia, comorbidities, and emergency department visits. CONCLUSIONS: Since health literacy is a potentially remediable explanatory factor in health care disparities, the development of automated health literacy indicators represents a significant accomplishment with broad clinical and population health applications. Health systems could apply literacy profiles to efficiently determine whether quality of care and outcomes vary by patient health literacy; identify at-risk populations for targeting tailored health communications and self-management support interventions; and inform clinicians to promote improvements in individual-level care.

Investigating the reaction and substrate preference of indole-3-acetaldehyde dehydrogenase from the plant pathogen Pseudomonas syringae PtoDC3000.

Aldehyde dehydrogenases (ALDHs) catalyze the conversion of various aliphatic and aromatic aldehydes into corresponding carboxylic acids. Traditionally considered as housekeeping enzymes, new biochemical roles are being identified for members of ALDH family. Recent work showed that AldA from the plant pathogen Pseudomonas syringae strain PtoDC3000 (PtoDC3000) functions as an indole-3-acetaldehyde dehydrogenase for the synthesis of indole-3-acetic acid (IAA). IAA produced by AldA allows the pathogen to suppress salicylic acid-mediated defenses in the model plant Arabidopsis thaliana. Here we present a biochemical and structural analysis of the AldA indole-3-acetaldehyde dehydrogenase from PtoDC3000. Site-directed mutants targeting the catalytic residues Cys302 and Glu267 resulted in a loss of enzymatic activity. The X-ray crystal structure of the catalytically inactive AldA C302A mutant in complex with IAA and NAD+ showed the cofactor adopting a conformation that differs from the previously reported structure of AldA. These structures suggest that NAD+ undergoes a conformational change during the AldA reaction mechanism similar to that reported for human ALDH. Site-directed mutagenesis of the IAA binding site indicates that changes in the active site surface reduces AldA activity; however, substitution of Phe169 with a tryptophan altered the substrate selectivity of the mutant to prefer octanal. The present study highlights the inherent biochemical versatility of members of the ALDH enzyme superfamily in P. syringae.

The LC3-conjugation machinery specifies the loading of RNA-binding proteins into extracellular vesicles.

Traditionally viewed as an autodigestive pathway, autophagy also facilitates cellular secretion; however, the mechanisms underlying these processes remain unclear. Here, we demonstrate that components of the autophagy machinery specify secretion within extracellular vesicles (EVs). Using a proximity-dependent biotinylation proteomics strategy, we identify 200 putative targets of LC3-dependent secretion. This secretome consists of a highly interconnected network enriched in RNA-binding proteins (RBPs) and EV cargoes. Proteomic and RNA profiling of EVs identifies diverse RBPs and small non-coding RNAs requiring the LC3-conjugation machinery for packaging and secretion. Focusing on two RBPs, heterogeneous nuclear ribonucleoprotein K (HNRNPK) and scaffold-attachment factor B (SAFB), we demonstrate that these proteins interact with LC3 and are secreted within EVs enriched with lipidated LC3. Furthermore, their secretion requires the LC3-conjugation machinery, neutral sphingomyelinase 2 (nSMase2) and LC3-dependent recruitment of factor associated with nSMase2 activity (FAN). Hence, the LC3-conjugation pathway controls EV cargo loading and secretion.

Risk of Cardiovascular Events in Patients With Type 2 Diabetes and Metabolic Dyslipidemia Without Prevalent Atherosclerotic Cardiovascular Disease.

BACKGROUND: The relationship between achieved low-density lipoprotein cholesterol (LDL-C) levels and risk of incident atherosclerotic cardiovascular disease events among patients with diabetes and metabolic dyslipidemia has not been well described. METHODS: We conducted an observational cohort study of statin-treated adults (ages 21-90 years) with type 2 diabetes without established atherosclerotic cardiovascular disease (as of January 1, 2006) who had metabolic dyslipidemia (elevated triglycerides ≥150 mg/dL and low high-density lipoprotein cholesterol, <50 mg/dL [women] and <40 mg/dL [men]). All subjects were members of Kaiser Permanente Northern California, an integrated health care delivery system. Adjusted multivariable Cox models were specified to estimate hazard ratios (HRs) for incident atherosclerotic cardiovascular disease events by achieved LDL-C levels (<50, 50-<70, 70-<100, and ≥100 mg/dL). Incident atherosclerotic cardiovascular disease events were defined as a composite of nonfatal myocardial infarction, ischemic stroke, or coronary heart disease death through December 31, 2013. RESULTS: A total of 19,095 individuals met the selection criteria. Mean age was 63.4 years, 53.5% were women, and the mean follow-up was 5.9 years. Unadjusted rates of atherosclerotic cardiovascular disease events were not significantly different across specified LDL-C categories. In models adjusted for demographics and clinical characteristics, the risk was significantly lower with decreasing achieved LDL-C levels (P <0.0001 for trend). Relative to achieved LDL-C ≥100 mg/dL, LDL-C <50 mg/dL had an hazard ratio of 0.66 (95% confidence interval [CI] 0.52-0.82). CONCLUSION: In a large, contemporary cohort of statin-treated patients with type 2 diabetes and metabolic dyslipidemia without established atherosclerotic cardiovascular disease, lower achieved LDL-C levels were associated with a monotonically lower risk of incident atherosclerotic cardiovascular disease events. The benefits of achieving very-low LDL-C (<50 mg/dL) in this population requires further evaluation in prospective interventional studies.

Secure Messaging with Physicians by Proxies for Patients with Diabetes: Findings from the ECLIPPSE Study.

BACKGROUND: Little is known about patients who have caregiver proxies communicate with healthcare providers via portal secure messaging (SM). Since proxy portal use is often informal (e.g., sharing patient accounts), novel methods are needed to estimate the prevalence of proxy-authored SMs. OBJECTIVE: (1) Develop an algorithm to identify proxy-authored SMs, (2) apply this algorithm to estimate predicted proxy SM (PPSM) prevalence among patients with diabetes, and (3) explore patient characteristics associated with having PPSMs. DESIGN: Retrospective cohort study. PARTICIPANTS: We examined 9856 patients from Diabetes Study of Northern California (DISTANCE) who sent ≥ 1 English-language SM to their primary care physician between July 1, 2006, and Dec. 31, 2015. MAIN MEASURES: Using computational linguistics, we developed ProxyID, an algorithm that identifies phrases frequently found in registered proxy SMs. ProxyID was validated against blinded expert categorization of proxy status among an SM sample, then applied to identify PPSM prevalence across patients. We examined patients' sociodemographic and clinical characteristics according to PPSM penetrance, "none" (0%), "low" (≥ 0-50%), and "high" (≥ 50-100%). KEY RESULTS: Only 2.3% of patients had ≥ 1 registered proxy-authored SM. ProxyID demonstrated moderate agreement with expert classification (Κ = 0.58); 45.7% of patients had PPSMs (40.2% low and 5.5% high). Patients with high percent PPSMs were older than those with low percent and no PPSMs (66.5 vs 57.4 vs 56.2 years, p < 0.001) had higher rates of limited English proficiency (16.1% vs 3.2% vs 3.5%, p < 0.05), lower self-reported health literacy (3.83 vs 4.43 vs 4.44, p < 0.001), and more comorbidities (Charlson index 3.78 vs 2.35 vs 2.18, p < 0.001). CONCLUSIONS: Among patients with diabetes, informal proxy SM use is more common than registered use and prevalent among socially and medically vulnerable patients. Future research should explore whether proxy portal use improves patient and/or caregiver outcomes and consider policies that integrate caregivers in portal communication.

Using natural language processing and machine learning to classify health literacy from secure messages: The ECLIPPSE study.

Limited health literacy is a barrier to optimal healthcare delivery and outcomes. Current measures requiring patients to self-report limitations are time-consuming and may be considered intrusive by some. This makes widespread classification of patient health literacy challenging. The objective of this study was to develop and validate "literacy profiles" as automated indicators of patients' health literacy to facilitate a non-intrusive, economic and more comprehensive characterization of health literacy among a health care delivery system's membership. To this end, three literacy profiles were generated based on natural language processing (combining computational linguistics and machine learning) using a sample of 283,216 secure messages sent from 6,941 patients to their primary care physicians. All patients were participants in Kaiser Permanente Northern California's DISTANCE Study. Performance of the three literacy profiles were compared against a gold standard of patient self-reported health literacy. Associations were analyzed between each literacy profile and patient demographics, health outcomes and healthcare utilization. T-tests were used for numeric data such as A1C, Charlson comorbidity index and healthcare utilization rates, and chi-square tests for categorical data such as sex, race, poor adherence and severe hypoglycemia. Literacy profiles varied in their test characteristics, with C-statistics ranging from 0.61-0.74. Relations between literacy profiles and health outcomes revealed patterns consistent with previous health literacy research: patients identified via literacy profiles indicative of limited health literacy: (a) were older and more likely of minority status; (b) had poorer medication adherence and glycemic control; and (c) exhibited higher rates of hypoglycemia, comorbidities and healthcare utilization. This represents the first successful attempt to employ natural language processing to estimate health literacy. Literacy profiles can offer an automated and economical way to identify patients with limited health literacy and greater vulnerability to poor health outcomes.