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.

Comparison of two propensity score-based methods for balancing covariates: the overlap weighting and fine stratification methods in real-world claims data.

BACKGROUND: Two propensity score (PS) based balancing covariate methods, the overlap weighting method (OW) and the fine stratification method (FS), produce superb covariate balance. OW has been compared with various weighting methods while FS has been compared with the traditional stratification method and various matching methods. However, no study has yet compared OW and FS. In addition, OW has not yet been evaluated in large claims data with low prevalence exposure and with low frequency outcomes, a context in which optimal use of balancing methods is critical. In the study, we aimed to compare OW and FS using real-world data and simulations with low prevalence exposure and with low frequency outcomes. METHODS: We used the Texas State Medicaid claims data on adult beneficiaries with diabetes in 2012 as an empirical example (N = 42,628). Based on its real-world research question, we estimated an average treatment effect of health center vs. non-health center attendance in the total population. We also performed simulations to evaluate their relative performance. To preserve associations between covariates, we used the plasmode approach to simulate outcomes and/or exposures with N = 4,000. We simulated both homogeneous and heterogeneous treatment effects with various outcome risks (1-30% or observed: 27.75%) and/or exposure prevalence (2.5-30% or observed:10.55%). We used a weighted generalized linear model to estimate the exposure effect and the cluster-robust standard error (SE) method to estimate its SE. RESULTS: In the empirical example, we found that OW had smaller standardized mean differences in all covariates (range: OW: 0.0-0.02 vs. FS: 0.22-3.26) and Mahalanobis balance distance (MB) (< 0.001 vs. > 0.049) than FS. In simulations, OW also achieved smaller MB (homogeneity: <0.04 vs. > 0.04; heterogeneity: 0.0-0.11 vs. 0.07-0.29), relative bias (homogeneity: 4.04-56.20 vs. 20-61.63; heterogeneity: 7.85-57.6 vs. 15.0-60.4), square root of mean squared error (homogeneity: 0.332-1.308 vs. 0.385-1.365; heterogeneity: 0.263-0.526 vs 0.313-0.620), and coverage probability (homogeneity: 0.0-80.4% vs. 0.0-69.8%; heterogeneity: 0.0-97.6% vs. 0.0-92.8%), than FS, in most cases. CONCLUSIONS: These findings suggest that OW can yield nearly perfect covariate balance and therefore enhance the accuracy of average treatment effect estimation in the total population.

Moderation of associations between weight discrimination and diabetes status by psychosocial factors.

Weight discrimination has adverse effects on health that include increasing the risk factors for developing type 2 diabetes. Preliminary evidence suggests a positive association between weight discrimination and diagnosed diabetes; however, it is unknown whether psychosocial resources may buffer this association. In logistic regressions stratified by gender, we examined links between weight discrimination and diabetes among a nationally representative sample of U.S. adults (the National Social Life, Health, and Aging Project; N = 2,794 adults age 50 and older in 2015-16). We also tested the extent to which trait-resilience and social support from a spouse/partner, family, and friends buffered any observed association. We adjusted for known predictors of diabetes (age, race/ethnicity, Body Mass Index) and conducted sensitivity analyses restricted to men and women with obesity. Net of covariates, in the overall sample, weight discrimination was associated with significantly greater odds of having ever had diabetes among women (OR = 2.00, 95% CI [1.15, 3.47]), but not men. Among women with obesity, weight discrimination was only significantly associated with greater odds of diabetes for those with low resilience (OR = 1.84, 95% CI [1.01, 3.35]). Among men overall, weight discrimination was associated with lower odds of diabetes for those with high family support (OR = 0.03, 95% CI [0.003, 0.25]) as well as those with high friend support (OR = 0.34, 95% CI [0.13, 0.91]); similar effects were observed in men with obesity. These novel findings evince a role for psychosocial resources in buffering associations between weight discrimination and diabetes.

Costs and Health Care Utilization Analysis of Medical Group Visits for Adults With Type 2 Diabetes in Community Health Centers.

OBJECTIVE: We evaluated the economic impact of group visits (GVs) in adults with uncontrolled diabetes in community health centers (CHCs) in the United States. RESEARCH DESIGN AND METHODS: In this prospective controlled trial, we implemented 6 monthly GV sessions in 5 CHCs and compared intervention patients (n=49) to control patients (n=72) receiving usual care within the same CHCs. We conducted patient chart reviews to obtain health care utilization data for the prior 6 months at baseline, 6 months (during the GV implementation), and 12 months (after the implementation). We also collected monthly logs of CHC expenses and staff time spent on activities related to GVs. Per-patient total costs included CHCs' expenses and costs associated with staff time and patients' health care use. For group comparison, we used the Wilcoxon rank-sum test and the bootstrapping method that was to bootstrap generalized estimating equation models. RESULTS: The GV group had fewer 6-month hospitalizations (mean: GV: 0.06 vs. control: 0.24, rate: 6.1% vs. 19.4%) ( P ≤ 0.04) and similar emergency department visits at 12 months than the control group. Implementing GV incurred $1770 per-patient. The intervention cost $1597 more than the control at 6 months ($3021 vs. $1424) but saved $1855 at 12 months ($857 vs. $2712) ( P =0.002). CONCLUSIONS: The diabetes GV care model reduced hospitalizations and had cost savings at 12 months, while it improved patients' diabetes-related quality of life and glucose control. Future studies should assess its lifetime cost-effectiveness through a randomized controlled trial.

Electronic Health Records (EHRs) Can Identify Patients at High Risk of Fracture but Require Substantial Race Adjustments to Currently Available Fracture Risk Calculators.

BACKGROUND: Osteoporotic fracture prediction calculators are poorly utilized in primary care, leading to underdiagnosis and undertreatment of those at risk for fracture. The use of these calculators could be improved if predictions were automated using the electronic health record (EHR). However, this approach is not well validated in multi-ethnic populations, and it is not clear if the adjustments for race or ethnicity made by calculators are appropriate. OBJECTIVE: To investigate EHR-generated fracture predictions in a multi-ethnic population. DESIGN: Retrospective cohort study using data from the EHR. SETTING: An urban, academic medical center in Philadelphia, PA. PARTICIPANTS: 12,758 White, 7,844 Black, and 3,587 Hispanic patients seeking routine care from 2010 to 2018 with mean 3.8 years follow-up. INTERVENTIONS: None. MEASUREMENTS: FRAX and QFracture, two of the most used fracture prediction tools, were studied. Risk for major osteoporotic fracture (MOF) and hip fracture were calculated using data from the EHR at baseline and compared to the number of fractures that occurred during follow-up. RESULTS: MOF rates varied from 3.2 per 1000 patient-years in Black men to 7.6 in White women. FRAX and QFracture had similar discrimination for MOF prediction (area under the curve, AUC, 0.69 vs. 0.70, p=0.08) and for hip fracture prediction (AUC 0.77 vs 0.79, p=0.21) and were similar by race or ethnicity. FRAX had superior calibration than QFracture (calibration-in-the-large for FRAX 0.97 versus QFracture 2.02). The adjustment factors used in MOF prediction were generally accurate in Black women, but underestimated risk in Black men, Hispanic women, and Hispanic men. LIMITATIONS: Single center design. CONCLUSIONS: Fracture predictions using only EHR inputs can discriminate between high and low risk patients, even in Black and Hispanic patients, and could help primary care physicians identify patients who need screening or treatment. However, further refinements to the calculators may better adjust for race-ethnicity.

Development and Validation of the Life Expectancy Estimator for Older Adults with Diabetes (LEAD): the Diabetes and Aging Study.

BACKGROUND: Estimated life expectancy for older patients with diabetes informs decisions about treatment goals, cancer screening, long-term and advanced care, and inclusion in clinical trials. Easily implementable, evidence-based, diabetes-specific approaches for identifying patients with limited life expectancy are needed. OBJECTIVE: Develop and validate an electronic health record (EHR)-based tool to identify older adults with diabetes who have limited life expectancy. DESIGN: Predictive modeling based on survival analysis using Cox-Gompertz models in a retrospective cohort. PARTICIPANTS: Adults with diabetes aged ≥ 65 years from Kaiser Permanente Northern California: a 2015 cohort (N = 121,396) with follow-up through 12/31/2019, randomly split into training (N = 97,085) and test (N = 24,311) sets. Validation was conducted in the test set and two temporally distinct cohorts: a 2010 cohort (n = 89,563; 10-year follow-up through 2019) and a 2019 cohort (n = 152,357; 2-year follow-up through 2020). MAIN MEASURES: Demographics, diagnoses, utilization and procedures, medications, behaviors and vital signs; mortality. KEY RESULTS: In the training set (mean age 75 years; 49% women; 48% racial and ethnic minorities), 23% died during 5 years follow-up. A mortality prediction model was developed using 94 candidate variables, distilled into a life expectancy model with 11 input variables, and transformed into a risk-scoring tool, the Life Expectancy Estimator for Older Adults with Diabetes (LEAD). LEAD discriminated well in the test set (C-statistic = 0.78), 2010 cohort (C-statistic = 0.74), and 2019 cohort (C-statistic = 0.81); comparisons of observed and predicted survival curves indicated good calibration. CONCLUSIONS: LEAD estimates life expectancy in older adults with diabetes based on only 11 patient characteristics widely available in most EHRs and claims data. LEAD is simple and has potential application for shared decision-making, clinical trial inclusion, and resource allocation.

First-Line Therapy for Type 2 Diabetes With Sodium-Glucose Cotransporter-2 Inhibitors and Glucagon-Like Peptide-1 Receptor Agonists : A Cost-Effectiveness Study.

BACKGROUND: Guidelines recommend sodium-glucose cotransporter-2 (SGLT2) inhibitors and glucagon-like peptide-1 (GLP1) receptor agonists as second-line therapy for patients with type 2 diabetes. Expanding their use as first-line therapy has been proposed but the clinical benefits may not outweigh their costs. OBJECTIVE: To evaluate the lifetime cost-effectiveness of a strategy of first-line SGLT2 inhibitors or GLP1 receptor agonists. DESIGN: Individual-level Monte Carlo-based Markov model. DATA SOURCES: Randomized trials, Centers for Disease Control and Prevention databases, RED BOOK, and the National Health and Nutrition Examination Survey. TARGET POPULATION: Drug-naive U.S. patients with type 2 diabetes. TIME HORIZON: Lifetime. PERSPECTIVE: Health care sector. INTERVENTION: First-line SGLT2 inhibitors or GLP1 receptor agonists. OUTCOME MEASURES: Life expectancy, lifetime costs, incremental cost-effectiveness ratios (ICERs). RESULTS OF BASE-CASE ANALYSIS: First-line SGLT2 inhibitors and GLP1 receptor agonists had lower lifetime rates of congestive heart failure, ischemic heart disease, myocardial infarction, and stroke compared with metformin. First-line SGLT2 inhibitors cost $43 000 more and added 1.8 quality-adjusted months versus first-line metformin ($478 000 per quality-adjusted life-year [QALY]). First-line injectable GLP1 receptor agonists cost more and reduced QALYs compared with metformin. RESULTS OF SENSITIVITY ANALYSIS: By removing injection disutility, first-line GLP1 receptor agonists were no longer dominated (ICER, $327 000 per QALY). Oral GLP1 receptor agonists were not cost-effective (ICER, $823 000 per QALY). To be cost-effective at under $150 000 per QALY, costs for SGLT2 inhibitors would need to be under $5 per day and under $6 per day for oral GLP1 receptor agonists. LIMITATION: U.S. population and costs not generalizable internationally. CONCLUSION: As first-line agents, SGLT2 inhibitors and GLP1 receptor agonists would improve type 2 diabetes outcomes, but their costs would need to fall by at least 70% to be cost-effective. PRIMARY FUNDING SOURCE: American Diabetes Association.

Quality of Care for US Adults With Medicaid Insurance and Type 2 Diabetes in Federally Qualified Health Centers Compared With Other Primary Care Settings.

OBJECTIVE: The objective of this study was to evaluate indicators of diabetes quality of care for US nonelderly, adult Medicaid enrollees with type 2 diabetes and compare federally qualified health centers (FQHCs) versus non-FQHCs. RESEARCH DESIGN AND METHODS: We analyzed diabetes process measures and acute health services utilization with 2012 US fee-for-service and managed care Medicaid claims in all 50 states and DC. We compared FQHC (N=121,977) to non-FQHC patients (N=700,401) using propensity scores to balance covariates and generalized estimating equation models. RESULTS: Overall, laboratory-based process measures occurred more frequently (range, 65.7%-76.6%) than measures requiring specialty referrals (retinal examinations, 33.3%; diabetes education, 3.4%). Compared with non-FQHC patients, FQHC patients had about 3 percentage point lower rates of each process measure, except for higher rates of diabetes education [relative risk=1.09, 95% confidence interval (CI): 1.03-1.16]. FQHC patients had fewer overall [incident rate ratio (IRR)=0.87, 95% CI: 0.86-0.88] and diabetes-related hospitalizations (IRR=0.79, 95% CI: 0.77-0.81), but more overall (IRR=1.06, 95% CI: 1.05-1.07) and diabetes-related emergency department visits (IRR=1.10, 95% CI: 1.08-1.13). CONCLUSIONS: This national analysis identified opportunities to improve diabetes management among Medicaid enrollees with type 2 diabetes, especially for retinal examinations or diabetes education. Overall, we found slightly lower rates of most diabetes care process measures for FQHC patients versus non-FQHC patients. Despite having higher rates of emergency department visits, FQHC patients were significantly less likely to be hospitalized than non-FQHC patients. These findings emphasize the need to identify innovative, effective approaches to improve diabetes care for Medicaid enrollees, especially in FQHC settings.

Longer-term Benefits and Risks of Sodium-Glucose Cotransporter-2 Inhibitors in Type 2 Diabetes: a Systematic Review and Meta-analysis.

BACKGROUND: Sodium-glucose cotransporter-2 inhibitors (SGLT2Is) are a recent class of medication approved for the treatment of type 2 diabetes (T2D). Previous meta-analyses have quantified the benefits and harms of SGLT2Is; however, these analyses have been limited to specific outcomes and comparisons and included trials of short duration. We comprehensively reviewed the longer-term benefits and harms of SGLT2Is compared to placebo or other anti-hyperglycemic medications. METHODS: We searched PubMed, Scopus, and clinicaltrials.gov from inception to July 2019 for randomized controlled trials of minimum 52 weeks' duration that enrolled adults with T2D, compared an SGLT2I to either placebo or other anti-hyperglycemic medications, and reported at least one outcome of interest including cardiovascular risk factors, microvascular and macrovascular complications, mortality, and adverse events. We conducted random effects meta-analyses to provide summary estimates using weighted mean differences (MD) and pooled relative risks (RR). The study was registered a priori with PROSPERO (CRD42018090506). RESULTS: Fifty articles describing 39 trials (vs. placebo, n = 28; vs. other anti-hyperglycemic medication, n = 12; vs. both, n = 1) and 112,128 patients were included in our analyses. Compared to placebo, SGLT2Is reduced cardiovascular risk factors (e.g., hemoglobin A1c, MD - 0.55%, 95% CI - 0.62, - 0.49), macrovascular outcomes (e.g., hospitalization for heart failure, RR 0.70, 95% CI 0.62, 0.78), and mortality (RR 0.87, 95% CI 0.80, 0.94). Compared to other anti-hyperglycemic medications, SGLT2Is reduced cardiovascular risk factors, but insufficient data existed for other outcomes. About a fourfold increased risk of genital yeast infections for both genders was observed for comparisons vs. placebo and other anti-hyperglycemic medications. DISCUSSION: We found that SGLT2Is led to durable reductions in cardiovascular risk factors compared to both placebo and other anti-hyperglycemic medications. Reductions in macrovascular complications and mortality were only observed in comparisons with placebo, although trials comparing SGLT2Is vs. other anti-hyperglycemic medications were not designed to assess longer-term outcomes.

The Longer-Term Benefits and Harms of Glucagon-Like Peptide-1 Receptor Agonists: a Systematic Review and Meta-Analysis.

BACKGROUND: Previous meta-analyses of the benefits and harms of glucagon-like peptide-1 receptor agonists (GLP1RAs) have been limited to specific outcomes and comparisons and often included short-term results. We aimed to estimate the longer-term effects of GLP1RAs on cardiovascular risk factors, microvascular and macrovascular complications, mortality, and adverse events in patients with type 2 diabetes, compared to placebo and other anti-hyperglycemic medications. METHODS: We searched PubMed, Scopus, and clinicaltrials.gov (inception-July 2019) for randomized controlled trials ≥ 52 weeks' duration that compared a GLP1RA to placebo or other anti-hyperglycemic medication and included at least one outcome of interest. Outcomes included cardiovascular risk factors, microvascular and macrovascular complications, all-cause mortality, and treatment-related adverse events. We performed random effects meta-analyses to give summary estimates using weighted mean differences (MD) and pooled relative risks (RR). Risk of bias was assessed using the Cochrane Collaboration risk of bias in randomized trials tool. Quality of evidence was summarized using the Grading of Recommendations, Assessment, Development, and Evaluation approach. The study was registered a priori with PROSPERO (CRD42018090506). RESULTS: Forty-five trials with a mean duration of 1.7 years comprising 71,517 patients were included. Compared to placebo, GLP1RAs reduced cardiovascular risk factors, microvascular complications (including renal events, RR 0.85, 0.80-0.90), macrovascular complications (including stroke, RR 0.86, 0.78-0.95), and mortality (RR 0.89, 0.84-0.94). Compared to other anti-hyperglycemic medications, GLP1RAs only reduced cardiovascular risk factors. Increased gastrointestinal events causing treatment discontinuation were observed in both comparisons. DISCUSSION: GLP1RAs reduced cardiovascular risk factors and increased gastrointestinal events compared to placebo and other anti-hyperglycemic medications. GLP1RAs also reduced MACE, stroke, renal events, and mortality in comparisons with placebo; however, analyses were inconclusive for comparisons with other anti-hyperglycemic medications. Given the high costs of GLP1RAs, the lack of long-term evidence comparing GLP1RAs to other anti-hyperglycemic medications has significant policy and clinical practice implications.

Estimating the Impact of Medicaid Expansion and Federal Funding Cuts on FQHC Staffing and Patient Capacity.

Policy Points In the preexpansion period, federally qualified health centers (FQHCs) in Medicaid expansion states were significantly different from those in nonexpansion states. This gap widened as revenues in expansion states continued to grow at a faster rate after the expansion. If Medicaid expansion had occurred nationwide, FQHCs' revenue and capacity could have increased substantially. Over time, Medicaid could play a bigger role as it becomes a more stable funding source to allow for capital investments. Section 330 grants appear to have a larger impact on access to care. Given the varying levels of reliance on Medicaid, investing through federal grants might be more effective and equitable. CONTEXT: The Health Resources and Services Administration's Health Center Program (HCP) plays a critical role as the national ambulatory safety net, delivering services to patients in medically underserved areas, regardless of their ability to pay. As the program has grown, health policy initiatives may have altered access to care for the underserved population. Understanding how federally qualified health centers (FQHCs) have been affected by past policies is important for anticipating the effects of future policies. METHODS: By analyzing a national data set from the Uniform Data System, we examined, using two sets of random effects regressions, the potential impact of alternative policy actions affecting FQHCs. Our primary equation models the number of full-time equivalent staff, of patients served, and of visits provided in the subsequent year as a function of Medicaid revenues, Section 330 grants, and other revenues. Our secondary equation is a difference-in-differences analysis that models Medicaid revenues as a function of the states' status of Medicaid expansion. FINDINGS: The expansion of Medicaid in nonexpansion states could have increased Medicaid revenues by 138%, staffing by 25%, and patients' visits by 24% in 2017. Compared to the impact of a "repeal" of Medicaid expansion, the percentage of reductions in staffing would be similar to those predicted by a 50% cut in Medicaid revenues or in Section 330 grants. On a dollar-for-dollar basis, the effects of one dollar of Section 330 grants were more than double that of one dollar of Medicaid revenue. CONCLUSIONS: Both Medicaid eligibility and Section 330 funding support are important to the HCP, and Section 330 grants are particularly closely related to staffing and the provision of services. States' decisions not to participate in or to repeal Medicaid expansion, to reduce Medicaid payment rates, and federal funding cuts all could have a negative impact on FQHCs, resulting in thousands of low-income patients losing access to primary care.

Health Centers and Value-Based Payment: A Framework for Health Center Payment Reform and Early Experiences in Medicaid Value-Based Payment in Seven States.

Policy Points As essential access points to primary care for almost 29 million US patients, of whom 47% are Medicaid enrollees, health centers are positioned to implement the population health management necessary in value-based payment (VBP) contracts. Primary care payment reform requires multiple payment methodologies used together to provide flexibility to care providers, encourage investments in infrastructure and new services, and offer incentives for achieving better health outcomes. State policy and significant financial incentives from Medicaid agencies and Medicaid managed care plans will likely be required to increase health center participation in VBP, which is consistent with broader state efforts to expand investment in primary care. CONTEXT: Efforts are ongoing to advance value-based payment (VBP), and health centers serve as essential access points to comprehensive primary care services for almost 29 million people in the United States. Therefore, it is important to assess the levels of health center participation in VBP, types of VBP contracts, characteristics of health centers participating in VBP, and variations in state policy environments that influence VBP participation. METHODS: This mixed methods study combined qualitative research on state policy environments and health center participation in VBP with quantitative analysis of Uniform Data System and health center financial data in seven vanguard states: Oregon, Washington, California, Colorado, New York, Hawaii, and Kentucky. VBP contracts were classified into three layers: base payments being transformed from visit-based to population-based (Layer 1), infrastructure and care coordination payments (Layer 2), and performance incentive payments (Layer 3). FINDINGS: Health centers in all seven states participated in Layer 2 and Layer 3 VBP, with VBP participation growing from 35% to 58% of all health centers in these states from 2013 to 2017. Among participating health centers, the average percentage of Medicaid revenue received as Layer 2 and Layer 3 VBP rose from 6.4% in 2013 to 9.1% in 2017. Oregon and Washington health centers participating in Layer 1 payment reforms received most of their Medicaid revenue in VBP. In 2017, VBP participation was associated with larger health center size in four states (P <.05), and higher average number of days cash on hand (P <.05) in three states. CONCLUSIONS: A multilayer payment model is useful for implementing and monitoring VBP adoption among health centers. State policy, financial incentives from Medicaid agencies and Medicaid managed plans, and health center-Medicaid collaboration under strong primary care association and health center leadership will likely be required to increase health center participation in VBP.

Building and experimenting with an agent-based model to study the population-level impact of CommunityRx, a clinic-based community resource referral intervention.

CommunityRx (CRx), an information technology intervention, provides patients with a personalized list of healthful community resources (HealtheRx). In repeated clinical studies, nearly half of those who received clinical "doses" of the HealtheRx shared their information with others ("social doses"). Clinical trial design cannot fully capture the impact of information diffusion, which can act as a force multiplier for the intervention. Furthermore, experimentation is needed to understand how intervention delivery can optimize social spread under varying circumstances. To study information diffusion from CRx under varying conditions, we built an agent-based model (ABM). This study describes the model building process and illustrates how an ABM provides insight about information diffusion through in silico experimentation. To build the ABM, we constructed a synthetic population ("agents") using publicly-available data sources. Using clinical trial data, we developed empirically-informed processes simulating agent activities, resource knowledge evolution and information sharing. Using RepastHPC and chiSIM software, we replicated the intervention in silico, simulated information diffusion processes, and generated emergent information diffusion networks. The CRx ABM was calibrated using empirical data to replicate the CRx intervention in silico. We used the ABM to quantify information spread via social versus clinical dosing then conducted information diffusion experiments, comparing the social dosing effect of the intervention when delivered by physicians, nurses or clinical clerks. The synthetic population (N = 802,191) exhibited diverse behavioral characteristics, including activity and knowledge evolution patterns. In silico delivery of the intervention was replicated with high fidelity. Large-scale information diffusion networks emerged among agents exchanging resource information. Varying the propensity for information exchange resulted in networks with different topological characteristics. Community resource information spread via social dosing was nearly 4 fold that from clinical dosing alone and did not vary by delivery mode. This study, using CRx as an example, demonstrates the process of building and experimenting with an ABM to study information diffusion from, and the population-level impact of, a clinical information-based intervention. While the focus of the CRx ABM is to recreate the CRx intervention in silico, the general process of model building, and computational experimentation presented is generalizable to other large-scale ABMs of information diffusion.

Cost-Effectiveness of 5-Aminosalicylate Therapy in Combination With Biologics or Tofacitinib in the Treatment of Ulcerative Colitis.

INTRODUCTION: Medications are major cost drivers in the treatment of patients with inflammatory bowel disease. Recent analyses suggest that there is no added efficacy in continuing nor harm in stopping 5-aminosalicylate (ASA) therapy in patients with inflammatory bowel disease escalated to biological therapies or tofacitinib. We assessed the cost-effectiveness of discontinuing 5-ASA therapy in patients with ulcerative colitis on biological therapies or tofacitinib, compared with continuing 5-ASA therapy. METHODS: We performed a cost-effectiveness analysis of 5-ASA with biologic therapy and tofacitinib compared with the same treatment without 5-ASA. Our primary outcome was to determine whether biologic/tofacitinib monotherapy was cost-effective compared with biologic/tofacitinib and 5-ASA combination therapy using the incremental cost-effectiveness ratio at a willingness to pay of $50,000/quality-adjusted life year. Owing to the uncertainty surrounding outcome probabilities, probabilistic sensitivity analyses with 10,000 simulations were also performed. We conducted a sensitivity analysis comparing biologic/tofacitinib and 5-ASA therapy compared with biologic/tofacitinib monotherapy, whereby vedolizumab was the first biologic used, followed by infliximab and finally tofacitinib. RESULTS: Our model shows that biologic/tofacitinib monotherapy dominates (cheaper and more effective) combination therapy of biologics/tofacitinib with 5-ASA. Probabilistic sensitivity analyses simulations resulted in biologic/tofacitinib monotherapy dominating 100% of the scenarios, with mean cost savings of $24,483.01 over 2 years. When vedolizumab was the first-line therapy in the sensitivity analysis, biologic/tofacitinib monotherapy continued to dominate the combination of 5-ASA and biologic/tofacitinib therapy. DISCUSSION: This analysis in patients with ulcerative colitis who require treatment with biologics or tofacitinib demonstrates that continuing 5-ASA therapy is not a cost-effective strategy. Discontinuation of 5-ASA therapy in these patients is safe and less expensive and should be recommended.

Pembrolizumab and lenvatinib versus carboplatin and paclitaxel as first-line therapy for advanced or recurrent endometrial cancer: A Markov analysis.

OBJECTIVE: To determine the cost effectiveness of pembrolizumab/lenvatinib (P/L) versus standard-of-care carboplatin/paclitaxel (C/T) as first-line systemic therapy for patients with advanced/recurrent endometrial cancer. METHODS: We designed a Markov model to simulate treatment outcomes for advanced/recurrent endometrial cancer patients whose tumors are either microsatellite stable (MSS) or have high microsatellite instability (MSI-high). We adopted a healthcare sector perspective for the analysis. Model inputs for costs, health utility, and clinical estimates were obtained from the literature including data from GOG0209 and KEYNOTE-146. Primary outcomes included costs of care, quality-adjusted life years (QALYs), and the incremental cost-effectiveness ratio (ICER). The time-horizon was three years and the discount rate was 3% annually. RESULTS: In a MSS cohort, compared to C/T, first-line treatment with P/L increased treatment costs by $212,670 and decreased QALYs by 0.28 per patient. In a MSI-high cohort, compared to C/T, P/L increased costs by $313,487 and increased QALYs by 0.11 per patient, representing an ICER of $2,849,882 per QALY. Sensitivity analyses found that the price of the new drugs was the most important determinant of the ICER and that the price of the new drugs would need to decrease by 85% to $2817 per cycle to reach a $150,000/QALY threshold. CONCLUSION: In the MSS model, we found that first-line therapy for advanced or recurrent endometrial cancer with P/L increased costs and worsened outcomes compared to C/T. In the MSI-high model, P/L improved survival and QALYs compared to C/T but was not cost-effective at the current cost of the drugs.

Cost and healthcare utilization analysis of culturally sensitive, shared medical appointment model for Latino children with type 1 diabetes.

OBJECTIVE: This study evaluated costs and healthcare utilization associated with a culturally-sensitive, medical and education program for pediatric Latino patients with type 1 diabetes. RESEARCH DESIGN AND METHODS: Program participants included Latino children ages 1-20 years old diagnosed with type 1 diabetes (n = 57). Control subjects with type 1 diabetes were matched by age, sex, and zip code to intervention participants from the Colorado All Payer Claims Database. Data included emergency department (ED) visits, hospitalizations, demographic information, and health insurance claims data 180 days prior to program start/index date through 1 year after program start/index date. We tracked program staff time and estimated costs for healthcare utilization using data from the scientific literature. Generalized Estimating Equation (GEE) models with logit link were used to estimate group differences in probabilities of ED visits and hospitalizations over 6-month periods pre/post-study, accounting for correlation of within-subject data across time points. Sensitivity analyses modeled longer-term cost differences under different assumptions. RESULTS: The intervention group had fewer hospitalizations, 2% versus 12% of controls (p = 0.047,OR = 0.13;95%CI: 0.02-0.97) for 6 months following start date. The intervention group had fewer ED visits, 19% versus 32% in controls (n.s.; p = 0.079,OR = 0.52;95%CI:0.25-1.08) and significantly fewer hospitalizations, 4% versus 15% of controls (p = 0.039,OR = 0.21;95%CI: 0.05-0.93) 6-12 months post-start date. One-year per-patient program costs of $633 and healthcare cost savings of $2710 yielded total per-patient savings of $2077, or a 5-year cost savings of $14,106. CONCLUSION: This unique type 1 diabetes management program altered health service utilization of program participants, reducing major healthcare cost drivers, ED visits, and hospitalizations.

The Development and Validation of a Machine Learning Model to Predict Bacteremia and Fungemia in Hospitalized Patients Using Electronic Health Record Data.

OBJECTIVES: Bacteremia and fungemia can cause life-threatening illness with high mortality rates, which increase with delays in antimicrobial therapy. The objective of this study is to develop machine learning models to predict blood culture results at the time of the blood culture order using routine data in the electronic health record. DESIGN: Retrospective analysis of a large, multicenter inpatient data. SETTING: Two academic tertiary medical centers between the years 2007 and 2018. SUBJECTS: All hospitalized patients who received a blood culture during hospitalization. INTERVENTIONS: The dataset was partitioned temporally into development and validation cohorts: the logistic regression and gradient boosting machine models were trained on the earliest 80% of hospital admissions and validated on the most recent 20%. MEASUREMENTS AND MAIN RESULTS: There were 252,569 blood culture days-defined as nonoverlapping 24-hour periods in which one or more blood cultures were ordered. In the validation cohort, there were 50,514 blood culture days, with 3,762 cases of bacteremia (7.5%) and 370 cases of fungemia (0.7%). The gradient boosting machine model for bacteremia had significantly higher area under the receiver operating characteristic curve (0.78 [95% CI 0.77-0.78]) than the logistic regression model (0.73 [0.72-0.74]) (p < 0.001). The model identified a high-risk group with over 30 times the occurrence rate of bacteremia in the low-risk group (27.4% vs 0.9%; p < 0.001). Using the low-risk cut-off, the model identifies bacteremia with 98.7% sensitivity. The gradient boosting machine model for fungemia had high discrimination (area under the receiver operating characteristic curve 0.88 [95% CI 0.86-0.90]). The high-risk fungemia group had 252 fungemic cultures compared with one fungemic culture in the low-risk group (5.0% vs 0.02%; p < 0.001). Further, the high-risk group had a mortality rate 60 times higher than the low-risk group (28.2% vs 0.4%; p < 0.001). CONCLUSIONS: Our novel models identified patients at low and high-risk for bacteremia and fungemia using routinely collected electronic health record data. Further research is needed to evaluate the cost-effectiveness and impact of model implementation in clinical practice.

Neighborhood Disadvantage and Hospital Quality Ratings in the Medicare Hospital Compare Program.

BACKGROUND: The Centers for Medicare and Medicaid Services provide nationwide hospital ratings that may influence reimbursement. These ratings do not account for the social risk of communities and may inadvertently penalize hospitals that service disadvantaged neighborhoods. OBJECTIVE: This study examines the relationship between neighborhood social risk factors (SRFs) and hospital ratings in Medicare's Hospital Compare Program. RESEARCH DESIGN: 2017 Medicare Hospital Compare ratings were linked with block group data from the 2015 American Community Survey to assess hospital ratings as a function of neighborhood SRFs. SUBJECTS: A total of 3608 Medicare-certified hospitals in 50 US states. MEASURES: Hospital summary scores and 7 quality group scores (100 percentile scale), including effectiveness of care, efficiency of care, hospital readmission, mortality, patient experience, safety of care, and timeliness of care. RESULTS: Lower hospital summary scores were associated with caring for neighborhoods with higher social risk, including a reduction in hospital score for every 10% of residents who reported dual-eligibility for Medicare/Medicaid [-3.3%; 95% confidence interval (CI), -4.7 to -2.0], no high-school diploma (-0.8%; 95% CI, -1.5 to -0.1), unemployment (-1.2%; 95% CI, -1.9 to -0.4), black race (-1.2%; 95% CI, -1.7 to -0.8), and high travel times to work (-2.5%; 95% CI, -3.3 to -1.6). Associations between neighborhood SRFs and hospital ratings were largest in the timeliness of care, patient experience, and hospital readmission groups; and smallest in the safety, efficiency, and effectiveness of care groups. CONCLUSIONS: Hospitals serving communities with higher social risk may have lower ratings because of neighborhood factors. Failing to account for neighborhood social risk in hospital rating systems may reinforce hidden disincentives to care for medically underserved areas in the United States.

Impact of a Low-Intensity Resource Referral Intervention on Patients' Knowledge, Beliefs, and Use of Community Resources: Results from the CommunityRx Trial.

BACKGROUND: Connecting patients to community-based resources is now a cornerstone of modern healthcare that supports self-management of health. The mechanisms that link resource information to behavior change, however, remain poorly understood. OBJECTIVE: To evaluate the impact of CommunityRx, an automated, low-intensity resource referral intervention, on patients' knowledge, beliefs, and use of community resources. DESIGN: Real-world controlled clinical trial at an urban academic medical center in 2015-2016; participants were assigned by alternating week to receive the CommunityRx intervention or usual care. Surveys were administered at baseline, 1 week, 1 month, and 3 months. PARTICIPANTS: Publicly insured adults, ages 45-74 years. INTERVENTION: CommunityRx generated an automated, personalized list of resources, known as HealtheRx, near each participant's home using condition-specific, evidence-based algorithms. Algorithms used patient demographic and health characteristics documented in the electronic health record to identify relevant resources from a comprehensive, regularly updated database of health-related resources in the study area. MAIN MEASURES: Using intent-to-treat analysis, we examined the impact of HealtheRx referrals on (1) knowledge of the most commonly referred resource types, including healthy eating classes, individual counseling, mortgage assistance, smoking cessation, stress management, and weight loss classes or groups, and (2) beliefs about having resources in the community to manage health. KEY RESULTS: In a real-world controlled trial of 374 adults, intervention recipients improved knowledge (AOR = 2.15; 95% CI, 1.29-3.58) and beliefs (AOR = 1.68; 95% CI, 1.07-2.64) about common resources in the community to manage health, specifically gaining knowledge about smoking cessation (AOR = 2.76; 95% CI, 1.07-7.12) and weight loss resources (AOR = 2.26; 95% CI 1.05-4.84). Positive changes in both knowledge and beliefs about community resources were associated with higher resource use (P = 0.02). CONCLUSIONS: In a middle-age and older population with high morbidity, a low-intensity health IT intervention to deliver resource referrals promoted behavior change by increasing knowledge and positive beliefs about community resources for self-management of health. NIH TRIAL REGISTRY: NCT02435511.

Provider and Staff Morale, Job Satisfaction, and Burnout over a 4-Year Medical Home Intervention.

BACKGROUND: The patient-centered medical home (PCMH) is a widely adopted primary care model. However, it is unclear whether changes in provider and staff perceptions of clinic PCMH capability are associated with changes in provider and staff morale, job satisfaction, and burnout in safety net clinics. OBJECTIVE: To determine how provider and staff PCMH ratings changed under a multi-year PCMH transformation initiative and assess whether changes in provider and staff PCMH ratings were associated with changes in morale, job satisfaction, and burnout. DESIGN: Comparison of baseline (2010) and post-intervention (2013-2014) surveys. SETTING: Sixty clinics in five states. PARTICIPANTS: Five hundred thirty-six (78.2%) providers and staff at baseline and 589 (78.3%) post-intervention. INTERVENTION: Collaborative learning sessions and on-site coaching to implement PCMH over 4 years. MEASUREMENTS: Provider and staff PCMH ratings on 0 (worst) to 100 (best) scales; percent of providers and staff reporting good or better morale, job satisfaction, and freedom from burnout. RESULTS: Almost half of safety net clinics improved PCMH capabilities from the perspective of providers (28 out of 59, 47%) and staff (25 out of 59, 42%). Over the same period, clinics saw a decrease in the percentage of providers reporting high job satisfaction (- 12.3% points, p = .009) and freedom from burnout (- 10.4% points, p = .006). Worsened satisfaction was concentrated among clinics that had decreased PCMH rating, with those clinics seeing far fewer providers report high job satisfaction (- 38.1% points, p < 0.001). LIMITATIONS: Control clinics were not used. Individual-level longitudinal survey administration was not feasible. CONCLUSION: If clinics pursue PCMH transformation and providers do not perceive improvement, they may risk significantly worsened job satisfaction. Clinics should be aware of this potential risk of PCMH transformation and ensure that providers are aware of PCMH improvements.