Predicting the 6-month risk of severe hypoglycemia among adults with diabetes: Development and external validation of a prediction model.

AIMS: To develop and externally validate a prediction model for the 6-month risk of a severe hypoglycemic event among individuals with pharmacologically treated diabetes. METHODS: The development cohort consisted of 31,674 Kaiser Permanente Colorado members with pharmacologically treated diabetes (2007-2015). The validation cohorts consisted of 38,764 Kaiser Permanente Northwest members and 12,035 HealthPartners members. Variables were chosen that would be available in electronic health records. We developed 16-variable and 6-variable models, using a Cox counting model process that allows for the inclusion of multiple 6-month observation periods per person. RESULTS: Across the three cohorts, there were 850,992 6-month observation periods, and 10,448 periods with at least one severe hypoglycemic event. The six-variable model contained age, diabetes type, HgbA1c, eGFR, history of a hypoglycemic event in the prior year, and insulin use. Both prediction models performed well, with good calibration and c-statistics of 0.84 and 0.81 for the 16-variable and 6-variable models, respectively. In the external validation cohorts, the c-statistics were 0.80-0.84. CONCLUSIONS: We developed and validated two prediction models for predicting the 6-month risk of hypoglycemia. The 16-variable model had slightly better performance than the 6-variable model, but in some practice settings, use of the simpler model may be preferred.

Automating assessment of lifestyle counseling in electronic health records.

BACKGROUND: Numerous population-based surveys indicate that overweight and obese patients can benefit from lifestyle counseling during routine clinical care. PURPOSE: To determine if natural language processing (NLP) could be applied to information in the electronic health record (EHR) to automatically assess delivery of weight management-related counseling in clinical healthcare encounters. METHODS: The MediClass system with NLP capabilities was used to identify weight-management counseling in EHRs. Knowledge for the NLP application was derived from the 5As framework for behavior counseling: Ask (evaluate weight and related disease), Advise at-risk patients to lose weight, Assess patients' readiness to change behavior, Assist through discussion of weight-loss methods and programs, and Arrange follow-up efforts including referral. Using samples of EHR data between January 1, 2007, and March 31, 2011, from two health systems, the accuracy of the MediClass processor for identifying these counseling elements was evaluated in postpartum visits of 600 women with gestational diabetes mellitus (GDM) compared to manual chart review as the gold standard. Data were analyzed in 2013. RESULTS: Mean sensitivity and specificity for each of the 5As compared to the gold standard was at or above 85%, with the exception of sensitivity for Assist, which was 40% and 60% for each of the two health systems. The automated method identified many valid Assist cases not identified in the gold standard. CONCLUSIONS: The MediClass processor has performance capability sufficiently similar to human abstractors to permit automated assessment of counseling for weight loss in postpartum encounter records.

Geographic variation in cardiovascular procedure use among Medicare fee-for-service vs Medicare Advantage beneficiaries.

IMPORTANCE: Little is known about how different financial incentives between Medicare Advantage and Medicare fee-for-service (FFS) reimbursement structures influence use of cardiovascular procedures. OBJECTIVE: To compare regional cardiovascular procedure rates between Medicare Advantage and Medicare FFS beneficiaries. DESIGN, SETTING, AND PARTICIPANTS: Cross-sectional study of Medicare beneficiaries older than 65 years between 2003-2007 comparing rates of coronary angiography, percutaneous coronary intervention (PCI), and coronary artery bypass graft (CABG) surgery across 32 hospital referral regions in 12 states. MAIN OUTCOMES AND MEASURES: Rates of coronary angiography, PCI, and CABG surgery. RESULTS: We evaluated a total of 878,339 Medicare Advantage patients and 5,013,650 Medicare FFS patients. Compared with Medicare FFS patients, Medicare Advantage patients had lower age-, sex-, race-, and income-adjusted procedure rates per 1000 person-years for angiography (16.5 [95% CI, 14.8-18.2] vs 25.9 [95% CI, 24.0-27.9]; P < .001) and PCI (6.8 [95% CI, 6.0-7.6] vs 9.8 [95% CI, 9.0-10.6]; P < .001) but similar rates for CABG surgery (3.1 [95% CI, 2.8-3.5] vs 3.4 [95% CI, 3.1-3.7]; P = .33). There were no significant differences between Medicare Advantage and Medicare FFS patients in the rates per 1000 person-years of urgent angiography (3.9 [95% CI, 3.6-4.2] vs 4.3 [95% CI, 4.0-4.6]; P = .24) or PCI (2.4 [95% CI, 2.2-2.7] vs 2.7 [95% CI, 2.5-2.9]; P = .16). Procedure rates varied widely across hospital referral regions among Medicare Advantage and Medicare FFS patients. For angiography, the rates per 1000 person-years ranged from 9.8 to 40.6 for Medicare Advantage beneficiaries and from 15.7 to 44.3 for Medicare FFS beneficiaries. For PCI, the rates ranged from 3.5 to 16.8 for Medicare Advantage and from 4.7 to 16.1 for Medicare FFS. The rates for CABG surgery ranged from 1.5 to 6.1 for Medicare Advantage and from 2.5 to 6.0 for Medicare FFS. Across regions, we found no statistically significant correlation between Medicare Advantage and Medicare FFS beneficiary utilization for angiography (Spearman r = 0.19, P = .29) and modest correlations for PCI (Spearman r = 0.33, P = .06) and CABG surgery (Spearman r = 0.35, P = .05). Among Medicare Advantage beneficiaries, adjustment for additional cardiac risk factors had little influence on procedure rates. CONCLUSIONS AND RELEVANCE: Although Medicare beneficiaries enrolled in capitated Medicare Advantage programs had lower angiography and PCI procedure rates than those enrolled in Medicare FFS, the degree of geographic variation in procedure rates was substantial among Medicare Advantage beneficiaries and was similar in magnitude to that observed among Medicare FFS beneficiaries.

A population-based cohort study of undervaccination in 8 managed care organizations across the United States.

OBJECTIVES To examine patterns and trends of undervaccination in children aged 2 to 24 months and to compare health care utilization rates between undervaccinated and age-appropriately vaccinated children. DESIGN Retrospective matched cohort study. SETTING Eight managed care organizations of the Vaccine Safety Datalink. PARTICIPANTS Children born between 2004 and 2008. MAIN EXPOSURE Immunization records were used to calculate the average number of days undervaccinated. Two matched cohorts were created: 1 with children who were undervaccinated for any reason and 1 with children who were undervaccinated because of parental choice. For both cohorts, undervaccinated children were matched to age-appropriately vaccinated children by birth date, managed care organization, and sex. MAIN OUTCOME MEASURES Rates of undervaccination, specific patterns of undervaccination, and health care utilization rates. RESULTS Of 323 247 children born between 2004 and 2008, 48.7% were undervaccinated for at least 1 day before age 24 months. The prevalence of undervaccination and specific patterns of undervaccination increased over time (P < .001). In a matched cohort analysis, undervaccinated children had lower outpatient visit rates compared with children who were age-appropriately vaccinated (incidence rate ratio [IRR], 0.89; 95% CI, 0.89- 0.90). In contrast, undervaccinated children had increased inpatient admission rates compared with age-appropriately vaccinated children (IRR, 1.21; 95% CI, 1.18-1.23). In a second matched cohort analysis, children who were undervaccinated because of parental choice had lower rates of outpatient visits (IRR, 0.94; 95% CI, 0.93-0.95) and emergency department encounters (IRR, 0.91; 95% CI, 0.88-0.94) than age-appropriately vaccinated children. CONCLUSIONS Undervaccination appears to be an increasing trend. Undervaccinated children appear to have different health care utilization patterns compared with age-appropriately vaccinated children.

Interventional radiologic treatment of hepatocellular carcinoma-a cost analysis from the payer perspective.

PURPOSE: To determine whether there is a cost advantage for one of the three commonly performed interventional radiology (IR) procedures (chemoembolization, selective internal radiation therapy [SIRT], radiofrequency ablation [RFA]) in the treatment of hepatocellular carcinoma (HCC). MATERIALS AND METHODS: A cost analysis from the payer perspective was performed. Primary data were collected from a university hospital, and sensitivity testing was done by comparing coding information obtained at two other tertiary care medical facilities. Medicare allowable reimbursements were used to estimate costs. Decision analytic models using decision tree analysis and Monte Carlo simulations were used to compare alternatives. Simulations were performed comparing all three procedures, followed by a two-way comparison of chemoembolization and SIRT. RESULTS: Simple decision tree analyses showed that RFA was less expensive compared with chemoembolization and SIRT. Monte Carlo simulations showed average reimbursements for each of the three procedures that was largely dependent on the number of repeat procedures required ($9,362 vs $30,107 vs $35,629 for RFA, chemoembolization, and SIRT; P < .001). When comparing only chemoembolization and SIRT, chemoembolization was the lower cost strategy in most scenarios, but SIRT was lower in cost in more than one-third of the simulations. CONCLUSIONS: RFA was the least costly of the three IR strategies in nearly all scenarios studied in these models. Although chemoembolization was less expensive than SIRT in most instances, Monte Carlo simulation showed a preference for SIRT in more than one-third of all scenarios. Sensitivity analyses showed that the most important variables assessed were the need for repeat procedures.

Safety of trivalent inactivated influenza vaccine in children aged 24 to 59 months in the vaccine safety datalink.

OBJECTIVES: To evaluate the safety of trivalent inactivated influenza vaccine (TIV) in children aged 24 to 59 months and to evaluate the risk of medically attended events (MAEs) in a subcohort of children who had multiple annual doses of TIV over their lifetimes. DESIGN: Self-controlled screening study. SETTING: Seven US managed care organizations from October 1, 2002, to March 31, 2006. PARTICIPANTS: Children aged 24 to 59 months who received at least 1 TIV dose (66 283 children and 91 692 TIV doses). EXPOSURE: Vaccination with TIV. MAIN OUTCOME MEASURES: Medically attended events in inpatient and emergency department settings in one of the following risk windows: 0 to 2, 1 to 14, or 1 to 42 days after vaccination. All MAEs that met the screening criteria of incidence rate ratios (IRRs) exceeding 1.0 and P ≤ .05 or IRRs exceeding 2.0 and P < .20 underwent medical record review. A secondary analysis examined the risk of MAEs in children who had multiple annual lifetime TIV doses. RESULTS: Nine diagnoses met the screening criteria. After medical record review, gastrointestinal tract symptoms (IRR, 1.18; 95% confidence interval [CI], 1.10-1.25), gastrointestinal tract disorders (7.70; 1.11-53.52), and fever (1.71; 1.64-1.80) remained significantly associated with vaccination. None of the events seemed to be serious, and none had complications. In the secondary analysis, there was an apparent dose response for vaccine and allergic reactions in the 1- to 3-day risk window. CONCLUSIONS: There was no evidence of serious MAEs following vaccination with TIV among children aged 24 to 59 months. Further studies are warranted to evaluate the risk of MAEs in children with multiple lifetime TIV doses.