A1c Testing Research Articles

Quality of Chronic Kidney Disease Management in Canadian Primary Care

Although patients with chronic kidney disease (CKD) are routinely managed in primary care settings, no nationally representative study has assessed the quality of care received by these patients in Canada. To evaluate the current state of CKD management in Canadian primary care practices to identify care gaps to guide development and implementation of national quality improvement initiatives. This cross-sectional study leveraged Canadian Primary Care Sentinel Surveillance Network data from January 1, 2010, to December 31, 2015, to develop a cohort of 46 162 patients with CKD managed in primary care practices. Data analysis was performed from August 8, 2018, to July 31, 2019. The study examined the proportion of patients with CKD who met a set of 12 quality indicators in 6 domains: (1) detection and recognition of CKD, (2) testing and monitoring of kidney function, (3) use of recommended medications, (4) monitoring after initiation of angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin II receptor blockers (ARBs), (5) management of blood pressure, and (6) monitoring for glycemic control in those with diabetes and CKD. The study also analyzed associations of divergence from these quality indicators. The cohort comprised 46 162 patients (mean [SD] age, 69.2 [14.0] years; 25 855 [56.0%] female) with stage 3 to 5 CKD. Only 4 of 12 quality indicators were met by 75% or more of the study cohort. These indicators were receipt of an outpatient serum creatinine test within 18 months after confirmation of CKD, receipt of blood pressure measurement at any time during follow-up, achieving a target blood pressure of 140/90 mm Hg or lower, and receiving a hemoglobin A1c test for monitoring diabetes during follow-up. Indicators in the domains of detection and recognition of CKD, testing and monitoring of kidney function (specifically, urine albumin to creatinine ratio testing), use of recommended medications, and appropriate monitoring after initiation of treatment with ACEIs or ARBs were not met. Only 6529 patients (18.4%) with CKD received a urine albumin test within 6 months of CKD diagnosis, and 3954 (39.4%) had a second measurement within 6 months of an abnormal baseline urine albumin level. Older age (≥85 years) and CKD stage 5 were significantly associated with not satisfying the criteria for the quality indicators across all domains. Across age categories, younger patients (aged 18-49 years) and older patients (≥75 years) were less likely to be tested for albuminuria (314 of 1689 patients aged 18-49 years [18.5%], 1983 of 11 919 patients aged 75-84 years [61.6%], and 614 of 5237 patients aged ≥85 years [11.7%] received the urine albumin to creatinine ratio test within 6 months of initial estimated glomerular filtration rate <60 mL/min per 1.73 m2; P < .001). Patients aged 18 to 49 years were less commonly prescribed recommended medications (222 of 2881 [7.7%]), whereas patients aged 75 to 84 years were prescribed ACEIs or ARBs most frequently (2328 of 5262 [44.2%]; P < .001). The findings suggest that management of CKD across primary care practices in Canada varies according to quality indicator. This study revealed potential priority areas for quality improvement initiatives in Canadian primary care practices.

Text messaging to engage friends/family in diabetes self-management support: acceptability and potential to address disparities.

Explore acceptability of engaging family/friends in patients' type 2 diabetes (T2D) self-management using text messaging. Participants (N = 123) recruited from primary care clinics for a larger trial evaluating mobile phone support for T2D completed self-report measures and a hemoglobin A1c test and then had the option to invite an adult support person to receive text messages. We examined characteristics and reasons of participants who did/did not invite a support person, responses to the invitation, and feedback from patients and support persons. Participants were 55.9 ± 10.1 years old, 55% female, 53% minority, and 54% disadvantaged (low income, less than high school degree/GED, uninsured, and/or homeless). Participants who invited a support person (48%) were slightly younger, more likely to be partnered, and reported more depressive symptoms and more emergency department visits in the year prior to study enrollment as compared to participants who did not (all p <.05). Participants' reasons for inviting a support person included needing help and seeing benefits of engaging others, while reasons for not inviting a support person included concerns about being a "burden" and support person's ability or desire to text. Support persons reported the texts increased awareness, created dialogue, and improved their own health behaviors. Patients inviting a support person had higher need and thus may stand to benefit most. Most support persons were open to engagement via text messages. Across race and socioeconomic status, text messaging may engage support persons to increase health-related support-particularly for patients with higher levels of need. Clinicaltrials.gov NCT02409329.

A quality improvement project to increase compliance with diabetes measures in an academic outpatient setting

BackgroundAmerican Diabetes Association (ADA) sets annual guidelines on preventative measures that aim to delay the onset of severe diabetes mellitus complications. Compared to private internal medicine clinics, resident clinics provide suboptimal diabetic preventative care as evidenced by decreased compliance with ADA guidelines. The purpose of our study is to improve diabetic care in resident clinics through quality improvement (QI) projects, with A1C value as primary outcome and other ADA guidelines as secondary outcomes.MethodsOur resident clinic at Beaumont Hospital, Royal Oak consists of 76 residents divided in 8 teams. In November 2016, baseline data on ADA guideline measures was obtained on 538 patients with diabetes mellitus. A root cause analysis was conducted. 5 teams developed a QI intervention plan to improve their diabetes care and 3 teams served as comparisons without intervention plans. In November 2017, post-intervention data was collected.ResultsBaseline characteristics demonstrate mean age of intervention groups at 60.9 years and of comparison groups at 58.9 years. The change in A1C value from baseline to post-intervention was + 0.09 vs. + 0.322 in the intervention and comparison groups respectively (p = 0.174). As a group, the changes in secondary outcome measures were as follows: eye examinations (+ 5% in intervention vs. -7% in comparison, p < 0.01), foot examinations (+ 13% vs. + 5%, p = 0.09), lipid panel testing (+ 7% vs. -5%, p < 0.01), micro-albumin/creatinine ratio testing (+ 4% vs. + 1%, p = 0.03), and A1C testing (+8% vs. + 5%, p = 0.24).ConclusionsWhile the QI project did not improve A1C value, it did have significant improvement in several secondary outcomes within intervention groups. One resident team implemented an intervention involving protected half-day blocks to identify overdue examinations and consequently had the largest improvements, thus serving as a potential intervention to further study. Given our study results, we believe that QI interventions improve preventative care for patients with diabetes in resident clinics.

Description of a pharmacist-led diabetes prevention service within an employer-based wellness program

ObjectivesTo describe a pharmacist-led diabetes prevention service piloted within an employer-based wellness program. Practice descriptionA pharmacist-led ambulatory care clinic within a school of pharmacy that provides wellness services to university employees. Practice innovationImplementation of a diabetes prevention service using opportunistic A1C screening within a biometric screening program. Patients with a prediabetes-level A1C from July 2016 to March 2019 were invited to participate in the National Diabetes Prevention Program (NDPP). EvaluationComparison of baseline characteristics of participants with normal and elevated A1C. Evaluation of participation in the NDPP and changes in clinical values at the subsequent biometric screening appointment for individuals with a prediabetes-level AlC. ResultsA1C testing of 740 individuals identified 69 participants (9.3%) with a prediabetes-level A1C and 7 (1.0%) with a diabetes-level A1C. Compared with those with a normal A1C (< 5.7%), participants with an elevated A1C were more likely to be older, nonwhite, obese, and physically inactive, to have a sibling with diabetes, higher random blood sugar (RBS), lower high-density lipoprotein (HDL), and more likely to have hypertension. Twelve patients participated in the NDPP, although most attended only 1 session. Attenders had a significantly lower baseline weight and body mass index (BMI). There were no significant differences in the changes in A1C, BMI, weight, RBS, or HDL between attenders and nonattenders approximately 1 year later. ConclusionThis pilot demonstrated that opportunistic A1C testing could be incorporated into an ambulatory care clinic within a pharmacist-led employer-based wellness program. Uptake and retention of the NDPP were poor. Barriers to NDPP participation need to be investigated and addressed to improve service impact.

Prediabetes Knowledge, Attitudes, and Practices at an Academic Family Medicine Practice.

Lifestyle change programs are an effective but underutilized approach to prevent or delay type 2 diabetes in people with prediabetes. Understanding clinician prediabetes knowledge, attitudes, and practices can inform implementation efforts to increase lifestyle change program referrals. We surveyed clinicians at an academic family medicine clinic about their prediabetes knowledge, attitudes, and practices. From the same clinic, we reviewed electronic health records to assess prediabetes screening, diagnosis, and treatment coverage in the cohort of adults seen from 2015 to 2017. Thirty-one clinicians (69.6%) completed the survey. Clinicians believed prediabetes was an important health issue (n = 29; 93.7%) and that prediabetes screening (n = 20, 64.5%) and diagnosis (n = 31, 100%) were important for prediabetes management. About half of the respondents (n = 14; 45.2%) reported familiarity with the National Diabetes Prevention Program (DPP). Electronic chart review included 15,520 adult patients. Most of the 5360 nondiabetic patients meeting US Preventive Services Task Force diabetes screening guidelines (n = 4068; 75.9%) received a hemoglobin A1c test. Of the 1437 patients with an A1c result diagnostic of prediabetes, 729 (50.7%) had the diagnosis in their chart. Prediabetes patients receiving point-of-care A1c testing instead of laboratory testing had 4.7 increased odds (95% CI, 3.5 to 6.4) of metformin prescription. No patients were referred to a DPP. Clinicians' positive attitudes toward prediabetes screening, moderate knowledge of prediabetes management, and low awareness of DPPs were reflected by high diabetes screening coverage, limited prediabetes diagnosis, and no DPP referrals. We will tailor our implementation strategy to overcome these prediabetes care barriers.

Prevalence and related factors on diabetes among HIV/AIDS receiving antiretroviral therapy in Dehong Dai and Jingpo Autonomous Prefecture

Objective: To study the prevalence and correlates of diabetes among HIV/AIDS who were on antiretroviral therapy (ART) in Dehong Dai and Jingpo autonomous prefectures (Dehong), Yunnan province. Methods: The database of HIV/AIDS receiving ART in Dehong was downloaded by using the basic information system of AIDS prevention and control in China. In this cross-sectional study, HIV/AIDS patients who were currently on ART and aged 18 years or above, were consecutively recruited, between July 2017 and June 2018, in Dehong. All the subjects underwent hemoglobin A1c (HbA1c) testing. Patient with diabetes was defined as meeting any of these indicators (HbA1c ≥6.5%, baseline FPG ≥7.0 mmol/L, FPG ≥7.0 mmol/L in the most recent visit). Both univariate and multivariate logistic regression analysis were carried on to evaluate the correlates of diabetes among the HIV/AIDS patients. Results: In total of 4 376 HIV/AIDS patients were included for analysis, with the average age as (43.7±10.1) years, proportion of males as 53.8% (2 356/4 376) and the HCV positive rate as 24.1% (1 055/4 376). The mean years was (8.9±3.8) years after the HIV diagnosis was made, and the mean duration on treatment was (6.8±2.9) years. The prevalence of diabetes was 11.4% (500/4 376). Through multivariate logistic regression analysis, data showed that the risk factors of diabetes of HIV/AIDS on ART were: aged 40 years or above, being male, HCV positive, baseline body mass index ≥24.0 kg/m(2), elevated TG ≥1.70 mmol/L in the most recent visit and baseline antiretroviral regimens under Efavirenz (EFV). Conclusions: Prevalence rate of diabetes appeared higher in HIV/AIDS patients who were on ART in Dehong. Prevention and control measures should be targeted on HIV/AIDS patients who were with risk factors of diabetes as being elderly, male, HCV positive, overweight and higher TG. Further esearch is needed to evaluate the association between the use of EFV and diabetes.

Sleep Quality and Blood Lipid Composition Among Patients with Diabetes.

BackgroundRecent literature has mentioned that people with sleep disorder, experience insulin sensitivity reduction and accordingly higher levels of blood glucose.ObjectivesThis study aimed to investigate the relationship between sleep quality and blood lipid composition in patients with diabetes referring to Minoodar health center in Qazvin, Iran in 2017.MethodsSleep duration and quality were assessed in 347 patients with diabetes using the Pittsburgh sleep quality index (PSQI). The glycosylated hemoglobin A1c (HbA1c) test was used to measure the glycemic control and total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), and high-density lipoprotein cholesterol (HDL-C) were used to determine blood lipid composition of the patients. Multiple regression analyses were applied to examine the associations between sleep measures and HbA1c and lipid parameters using SPSS version 20.ResultsThe patients in the poor sleep quality group had higher levels of fasting blood sugar (FBS) (146.07 ± 57.06 versus 132.8 ± 53.3 mg/dL, P = 0.02), body mass index (BMI) (29.1 ± 3.9 versus 27.6 ± 4.2 kg/m2, P = 0.005) and total cholesterol (209.9 ± 53.4 versus 193.4 ± 45.8, P = 0.02). Furthermore, the patients with short sleep duration had higher total cholesterol level compared with long sleep and medium sleep duration group (202.3 ± 50.2 versus 196.6 ± 47.7 and 195.7 ± 47.4, respectively, P = 0.05). Among different PSQI measures, subjective sleep quality was associated with lower TC and TG in unadjusted models (β = -0.0.1, P = 0.05). Furthermore, greater sleep disturbance was positively linked with higher levels of TC and TG (β = 0.1, P = 0.01 and β = 0.02, P = 0.05).ConclusionsIn an Iranian population with diabetes living in Qazvin city, sleep disorder is common and as study findings revealed sleep quality was recognized as an influencing factor on some of the lipid profiles, including TC and TG. Thus sleep assessment of patients with type 2 diabetes to find the early recognition of their sleep disorder should be considered an important part of the patients’ treatment.

343-OR: Trends in A1C and Lipid Testing, Antilipid Therapy, and Target Attainment following First-Time Glucose-Lowering Drug Treatment in Type 2 Diabetes, Denmark, 2000-2017

Objective: To examine 18-year trends in early A1C and lipid control among patients initiating first-time glucose-lowering drug therapy (GLD) for type 2 diabetes. Materials and Methods: Population-based sequential cross-sectional study using healthcare databases covering Northern Denmark (∼1.8 million persons). For all 94,175 GLD initiators during 2000-2017, we examined likelihood of A1C and lipid testing, antilipid therapy, and 1-year glycemic and LDL cholesterol target achievement. Results: The proportion of patients with at least one A1C test increased from 53% in 2000 to 95% in 2017 (Figure, Panel A). Lipid testing increased from 82% to 98% and lipid lowering therapy initiation quintupled from 12% to 61%. Mean pre-treatment A1C declined substantially, from 9.2% (77 mmol/mol) to 7.3% (56 mmol/mol) in 2011, then rose to 7.9% (63 mmol/mol) in 2017 (Panel B). Post-treatment A1C decreased less; pre- and post-treatment LDL cholesterol declined similarly. The proportion of patients achieving glycemic and cholesterol targets increased substantially over time (Panel C). Conclusion: Monitoring and treatment of A1C and cholesterol in early type 2 diabetes have improved considerably over the past 18 years, while quality improvement was somewhat flat after the 2011 change in diagnostic criteria. Disclosure J.S. Knudsen: None. D.R. Witte: None. A. Hulman: None. P.F. Rønn: Research Support; Self; Amgen AB. Stock/Shareholder; Spouse/Partner; Novo Nordisk A/S. T. Lauritzen: Advisory Panel; Self; AstraZeneca. Board Member; Self; Novo Nordisk A/S. H.T. Sôrensen: None. R.W. Thomsen: None. Funding Aarhus University

888-P: High Engagement over 12 Months with a Text Messaging Intervention for Type 2 Diabetes

The extent to which users engage with an intervention can reveal why and when it is effective. However, studies reporting on engagement with text messaging interventions are often short-term and frequency/dose varies widely. Further, limited research has examined engagement in diabetes text-based interventions with racially and socioeconomically diverse patients. We examined engagement with a 12-month text messaging intervention for T2D as part of a larger RCT. We recruited adults with T2D from community health centers and primary care clinics. At baseline, patients completed a survey and an A1c test. Patients randomized to intervention received daily one-way and interactive texts for the first 6 months, and then had the option to receive fewer texts (i.e., low-dose) for the remaining 6 months. We examined associations between patient characteristics and low-dose choice, and the association between choice and subsequent engagement, as measured by odds of response to interactive texts. The sample (N=250) had a mean age of 55.9±9.8 years, was 55% female, and 52% non-white; 41% had ≤ a high school degree, 41% had incomes &amp;lt;$25K, and mean A1c was 8.6±1.8%. On average, patients responded to 81% (IQR: 75-96%) of texts over 12 months. At 6 months, 44% chose to continue receiving daily texts, whereas 56% chose low-dose for the remaining 6 months. No patient characteristics were associated with choice, and choice was not associated with later engagement (b=-0.20, p=0.52). Odds of responding was lower over time in both groups (p&amp;lt;.001) but did not differ across groups. Average response rate was 73% (61-97%) in months 8-12. Among diverse patients with T2D, engagement with text messages remained high over 12 months. Almost half the sample elected to continue receiving daily texts for the entire 12 months. Engagement was similar in both groups in the final 6 months. Our findings support using texts to engage diverse patients in mobile health interventions. Providing an option for text frequency may help sustain engagement. Disclosure L.A. Nelson: None. R.A. Greevy: None. A.J. Spieker: None. K. Wallston: None. L.S. Mayberry: None. Funding National Institute of Diabetes and Digestive and Kidney Diseases (R01DK100694); Center for Diabetes Translational Research (P30DK92986)

Validity of Predictive Equations for Metabolic Basal Rate in Brazilian Patients with Type 2 Diabetes (P12-039-19)

ObjectivesTo evaluate in Brazilian patients with type 2 diabetes which of the seven predictive equations selected to estimate basal metabolic rate (BMR) is the best alternative considering calorimetry indirect (CI) as a reference method. MethodsA cross-sectional study was conducted with 62 patients (31 men and 31 women) with type 2 diabetes. Clinical and laboratorial variables were evaluated as well as body composition by electrical bioimpedance. The BMR was measured by IC (Analisador MedGraphics Cardiorespiratory Diagnostic Systems, model CCM Express®) and estimated by prediction equations. Dietary intake was evaluated by a food frequency questionnaire (FFQ). Data were analyzed using Bland–Altman plots, paired t-tests, and Pearson’s correlation coefficients. ResultsOur patients were aged 63.1 ± 5.2 years, had diabetes duration of 11 (1–36) years and A1C test 7.6 ± 1.2%. The body composition contained a fat free mass of 35.2 ± 11.8 kg and a fat mass of 29.1 ± 8.8 kg. The energy intake by FFQ was 1826.9 ± 628.1 Kcal/day and the BMR by IC was 1644.6 ± 310.6 kcal/day. There was a wide variation in the accuracy of BMR values predicted by equations when compared to IC BMR measurement. FAO/WHO/UNO and Oxford equations produced the smallest differences to IC. For women, the FAO/WHO/UNO equation provided the best BMR prediction in comparison to measured BMR (–1.8% of the difference). For men, the equation of the Oxford was the one closest to the BMR values as measured by IC (–1.3% of the difference). ConclusionsIn Brazilian patients with type 2 diabetes, the predictive equations by FAO/WHO/UNO (for women) and Oxford (for men) were the best to estimate BMR. Funding SourcesFundação do Incentivo à Pesquisa (FIPE) of Hospital de Clinicas de Porto Alegre (#protocol 15.0625) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior.

2434-PUB: Comparison of Diabetes Prevalence and Severity between Chinese and Non-Chinese Populations in New York City

Diabetes mellitus (DM) is the 7th leading cause of death in the U.S. in 2015. According to Center of Disease Control (CDC), prevalence of DM has increased in recent years affecting 9.4% of U.S. population and 12.5% of New York City (NYC) population in 2015. Increased DM prevalence is a worldwide trend associated with increasing age of the population. Aim: To determine prevalence and severity of diabetes/prediabetics in NYC Chinese population in comparison with those of non-Chinese patients. Methods: We obtained and analyze HMO A1C data of 276,854 Chinese (age 18-98 with median of 51 years, 55% female, 45% male, representing the largest population studies) and non-Chinese sampling population in NYC from January 2015 to October 2018 using BioReference Lab. Patients (57,612 Chinese, 49,548 non-Chinese) classified as prediabetics with A1C between 5.7-6.4% upon initial A1C test had follow-up test after management according to CDC Prediabetic management protocol. Results: Of 197,487 Chinese who had A1C test indicated prevalence of 18.5% diabetes, 45.6% of prediabetic and 35.6% nondiabetics. The male group had a higher incidence of diabetes than females (21% vs. 16.5%). Among 1,341,054 non-Chinese clients, there were 51.6% nondiabetics, 31.4% prediabetics and 18.0% diabetes. The severity of diabetes with respect to A1C data in 36,596 Chinese were 6.5-7.4, 57%; 7.5-9.0, 27%; &amp;gt;9.0, 16% whereas in 241,180 non-Chinese they were 44%, 29% and 28%, respectively. Among Chinese prediabetics, 30% reversed to nondiabetics and 5.7% progressed to diabetic categories, whereas among non-Chinese prediabetics, 25.8% reverse to nondiabetics and 11.1% progressed to diabetics during the study time. Conclusion: Prevalence of diabetes in NYC population is higher than that found in 2015. More severe diabetes with higher A1C levels was found in non-Chinese than Chinese patients. The NYC Chinese had higher incidence of prediabetics but had better management outcome than non-Chinese community. Disclosure G. Liu: None. J.D. George: None. J.W. Sharp: None. A. Hu: None. T. Chang: None.

1283-P: PTS Diagnostics A1CNow+: A Valuable Tool in a Value-Based World

Background: Value-based medicine is a new healthcare model to shift physician focus from the quantity of services and revenue generated to quality and value of healthcare provided. The basis for this model is the idea of the right care for the right patient in the right environment at the right time, also known as Patient Centered Medical Home (PCMH). Point-of-care A1C testing can maximize patient and provider engagement by obtaining the results in real-time which allows for effective communication, healthy discussions, and improved healthcare. This study evaluated the accuracy of A1CNow+ as a tool to assist with PCMH for diabetes management. Method: Blood was collected from PTS employees. A1C values ranged from 4.4 - 12.4% A1C for a total of 90 data points. Capillary blood was tested on the PTS Diagnostics A1CNow+ system while venous blood was tested at PTS Diagnostics on the Tosoh G8 and Roche Cobas® Integra® 400 Plus as clinical laboratory (CL) comparators. For reference, venous blood was tested at LabCorp on the Roche Cobas c513. Deming regression analysis was used to evaluate accuracy and paired differences to measure bias. Clinical risk was assessed using A1C clinical cut points of &amp;lt; 5.7, 5.7 - 6.4, and ≥ 6.5% A1C. Fisher’s exact test was used to assess differences amongst methods. Results: Average paired percent bias was -0.83% for the A1CNow+ and +0.17% for the CL analyzers. Slopes were 0.92 and 0.93 (p=0.35) and intercepts were 0.42 and 0.44 (p=0.098) for the A1CNow+ and CL, respectively. Clinical risk agreement was 89% for A1CNow+ and 97% for CL analyzers (p=0.08). Conclusion: The A1CNow+ was shown to be as accurate as the clinical lab in measuring A1C. Risk analysis showed no statistical difference between the CL and A1CNow+ in classifying A1C clinical cut points. The A1CNow+ is a valuable tool to help meet the goals of the Quadruple Aim which seeks to improve the health of the population, improve the patient experience of care, improve the provider experience of giving care, and reduce the cost of healthcare in a value-based world. Disclosure C.J. Szablowski: Employee; Self; PTS Diagnostics. K. Davis: None. J.H. Anderson: Board Member; Self; Generex Biotechnology Corp., VoCare. Consultant; Self; Know Labs. Employee; Self; PTS Diagnostics. Other Relationship; Self; Lilly Diabetes.

Association of Habitual Coffee Consumption and HECTD4 Polymorphism with Risk of Prediabetes and Type 2 Diabetes (P18-071-19)

ObjectivesEpidemiological studies suggested the evidence that coffee consumption decreased risk of type 2 diabetes. Recently, Japanese genome-wide association studies (GWAS) of coffee consumption has identified rs2074356 (G > A) at 12q24.12–13 in HECTD4. This study aims to examine the associations of habitual coffee consumption with prediabetes and type 2 diabetes, and whether this association is modified by rs2074356 variant in Korean adults. MethodsA total of 4010 participants (1890 men and 2120 women) who had genetic information from Korea Association REsource (KARE) study were included. Habitual coffee consumption was assessed through a food frequency questionnaire and categorized into five categories (non-consumers, <1 cup/day and ≥1 cups/day of black coffee, and <1 cup/day and ≥1 cups/day of sugared coffee). Incident prediabetes or type 2 diabetes were defined according to the self-report of physician-diagnosis, oral glucose tolerance test (OGTT) or hemoglobin A1c (HbA1c) test. Multivariable logistic regression models were used to calculate odds ratio (OR)s and 95% confidence interval (CI)s. ResultsA total of 2916 participants (72.7%) have developed prediabetes during a follow-up of 15 years. We found that black coffee consumption lowered risk of prediabetes and type 2 diabetes combined among men and women combined (OR = 0.63; 95% CI = 0.44–0.91 for ≥1 cups/day black coffee vs. non-consumers). When we separated men and women, compared with non-consumers, ORs (95% CIs) for ≥1 cups/day of black coffee were 0.50 (0.27–0.93) among men and 0.72 (0.45–1.14) among women and ORs (95% CIs) for ≥1 cups/day of sugared coffee were 1.41 (0.91–2.18) among men and 1.12 (0.80–1.58) among women. We observed a suggestive difference by rs2074356 (GG vs. AG + AA). Compared with non-consumers, participants with AG + AA genotypes consumed ≥ 1 cups/day of black coffee had a 60% lower risk of prediabetes and type 2 diabetes combined (95% CI 0.20–0.78), but we found a weaker association among those with GG genotype (OR = 0.81; 95% CI = 0.51–1.28). ConclusionsWe observed an inverse association between black coffee consumption and prediabetes and type 2 diabetes combined in Korean population. This association was more pronounced among carriers of minor allele of HECTD4 rs2074356 (AG/AA). Funding SourcesNone.

1552-P: Relationship between Diabetes Status and Chronic Kidney Disease in New York City Chinese and Non-Chinese Populations

Diabetes mellitus (DM) and prediabetes incidence have increased in the U.S. in recent years. In our center we have observed higher prevalence of diabetes/prediabetes in NYC Chinese-Americans than the general population. Diabetes patients are often associated with cardiovascular disease and chronic kidney disease (CKD). Aim: To determine the relationship between diabetes status and CKD status in NYC Chinese population compared to the general population so that proper medical management can be achieved. Method: Using A1C test to determine the diabetes status, with A1C value of &amp;lt;5.7 classified as nondiabetics, 5.7-6.4 as prediabetics and &amp;gt;6.5 as diabetics. Estimated glomerular filtration rate (eGFR) were determined and values &amp;lt;60 mL/min/1.73 m2 were regarded as having CKD. The data of 274,784 NYC Chinese and 1,271,642 non-Chinese were reviewed and compared in 18-64 years and ≥65 years groups. Results: In both populations the A1C levels tended to increase with increasing age, whereas their mean eGFR tended to decrease with age. The CKD prevalence was 8.9% in Chinese and 9.5% in non-Chinese. In 18-64 years groups CKD was 2.4% in Chinese and 2.7% in non-Chinese, whereas in ≥65 years group they were 29% in Chinese and 25.6% in non-Chinese. When A1C and eGFR were reviewed together, CKD prevalence was 3.1% of nondiabetics, 10.3% of prediabetics, 22.6% of A1C 6.5-6.9 DM, 24.1% of A1C 7.0-8.0 DM, 29.6% of A1C 8.1-8.9 DM and 27.6% of A1C &amp;gt;9.0 in Chinese population. In non-Chinese population, the CKD prevalence was 5.3%, 13.2%, 21.1%, 23.5%, 23.8%, and 18.6%, respectively for the corresponding nondiabetics, prediabetics and various DM subgroups. Conclusion: Significant prevalence of CKD were found in prediabetics of NYC Chinese and general populations. There is a high prevalence of CKD associated with diabetes mellitus in both populations. Prevalence of CKD in NYC Chinese is higher than the general population, particularly in the ≥65 years group and those having severer DM with A1C &amp;gt;8.0. Disclosure T. Chang: None. G. Liu: None. J.D. George: None. J.W. Sharp: None. A. Hu: None.

819-P: Personalized Texts Improve Adherence and A1c over 6 Months among High-Risk Adults with Type 2 Diabetes

Medication nonadherence and suboptimal glycemia are common among adults with T2D, particularly minorities and people with low SES. Text messaging is used ubiquitously without disparities. REACH (Rapid Education/Encouragement And Communications for Health) is a theory-based text messaging intervention that iteratively personalizes content to participants’ diabetes medicines and barriers to adherence. All participants have completed 6-month follow-up of a 15-month pragmatic RCT evaluating REACH. Adults (N=512) with T2D were recruited from community health centers and medical center primary care clinics. All participants received educational newsletters, completed A1c tests (venipuncture or mail-in kit) and surveys every 3 months; those assigned to REACH received daily texts. Surveys included a barriers assessment and the Summary of Diabetes Self-Care Activities medications subscale. Regression models examined change in barriers, adherence, and A1c (adjusted for test type) relative to controls. Participants were predominantly non-white (53%) and low SES (56% annual incomes &amp;lt;$35K; 49% underinsured; 42% &amp;lt;high school degree; 11% homeless/unstably housed), with an average age 56.0±9.5 years; 54% female; 49% on insulin; and baseline A1c 8.6±1.8%. At 6 months, retention was 90% and average response rate to daily interactive texts was mean 84% [IQR 79-90]. The 6-month overall REACH effect on A1c was -0.3% (p&amp;lt;.05), but reductions were greater among participants with higher baseline A1c (interaction term p&amp;lt;.05) - demonstrating A1c reductions ≥1.0% among participants with baseline A1c≥12.5% and ≥0.5% among participants with baseline A1c≥8.0% (both p&amp;lt;.05). Relative to controls, REACH also reduced adherence barriers (p&amp;lt;.05) and improved adherence (.53 days per week, p&amp;lt;.001). REACH engaged a diverse and high-risk group with daily texts, reduced barriers to adherence, improved adherence, and substantively improved A1c among patients with elevated glycemia at baseline. Disclosure L.S. Mayberry: None. L.A. Nelson: None. R.A. Greevy: None. K. Wallston: None. S. Kripalani: None. C.Y. Osborn: Employee; Self; One Drop. E.M. Bergner: None. L. LeStourgeon: None. S.E. Williamson: None. A.J. Spieker: None. C. Gentry: None. A.W. Brown: None. B.P. Gregory: None. S. Acuff: None. T. Elasy: None. Funding National Institutes of Health (R01DK100694)

2287-PUB: Association of Depression Treatment and Glycemic Control in Adults with Diabetes

Patients with diabetes and comorbid depression have an increased risk of complications. In a repeated panel cross-sectional study, we investigated the effect of depression care on glycemic control for adults with diabetes. Local health plan provided secondary data (N=29,227) that contained depression information and overall health risk calculated by Johns Hopkins ACG software using diagnosis code, visits, and medication claims. This data was matched with A1c, demographics, and neighborhood characteristics. Study outcome is A1c change. Predictor of interest is depression treatment: (1) no depression in either year, (2) new onset is not treated in prior year but treated in sequential year, (3) past is treated in one year and none in sequential year, and (4) persistent is received treatment in both years. Covariates are age, sex, health risk, insurance types, and neighborhood characteristics. The analysis is a multiple linear regression model using a difference-in-differences framework. Adjusting for covariates, new onset of depression treatment was associated with better diabetes control, A1c decreased by 0.072 (p=0.025), compared to no depression. Compared to no depression, past depression treatment was associated with lower diabetes control, A1c increased by 0.071 (p=0.007). Persistent depression treatment had A1c change not different from no depression group (p=0.731). Among patients with diabetes, past depression treatment appeared to have poorer glycemic control compared with those not depressed. New onset of depression treatment appeared to improve glycemic control. New diagnosis of depression increased patient contact with provider. A decline in diabetes control among patients received depression treatment in the past, may be related to a change in contact with provider. Future studies could investigate whether patient visit with providers was affected by a change in depression and examine if A1c testing rate correlates with health service utilization. Disclosure H.N. Fu: None. A.M. Raiter: None. C. Carlin: None. K.A. Peterson: None. Funding National Institute of Diabetes and Digestive and Kidney Diseases (R18DK110732)

1313-P: Comparison of Diabetes Health-Care Quality between Endo ECHO and the Academic Medical Center

Background: Patients receiving diabetes specialty care experience higher quality care, fewer complications and improved survival; however, access to specialists is limited in medically underserved areas. The Endo ECHO program connects specialists at academic medical centers with PCPs through virtual clinics. We report the comparability of quality measures between Endo ECHO patients and those receiving diabetes specialty care. Methods: Patients with complex diabetes from 10 Endo ECHO clinics in NM completed surveys assessing healthcare quality one year after program enrollment. Patients at the University of New Mexico Diabetes Comprehensive Care Center (DCCC) served as a comparison group and completed the same survey. Results: Five hundred thirty-three ECHO patients and 139 DCCC patients completed surveys. ECHO patients were much less likely to identify English as their primary language (69% ECHO vs. 91% DCCC, p&amp;lt;0.0001) and reported lower levels of educational attainment (40% ECHO with at least some college vs. 60% DCCC, p&amp;lt;0.0001). DCCC patients were more likely to report a positive overall health status (42% ECHO vs. 68% DCCC, p&amp;lt;0.0001), yet there was no difference in overall concern about health across groups. ECHO patients reported more frequent visits to their usual source of diabetes care (72% ECHO with 4 or more annual visits vs. 58% DCCC, p&amp;lt;0.0001). There were no differences in the frequency of A1C testing or feet checks by a provider; however, more DCCC patients reported regular eye exams (69% ECHO vs. 87% DCCC, p&amp;lt;0.0001) and that their provider had assessed smoking status (70% ECHO vs. 82% DCCC, p&amp;lt;0.0001). Conclusions: Endo ECHO patients faced greater medical literacy barriers yet reported similar outcomes for several, but not all, diabetes care quality measures as compared to those reported by patients at the DCCC. Endo ECHO may be a suitable alternative to specialty care for patients in medically underserved areas; ongoing evaluation will determine its impact on clinical outcomes. Disclosure M.M. Paul: None. A. Davila Saad: None. J. Billings: None. S. Blecker: Advisory Panel; Self; Medtronic. M.F. Bouchonville: None. B. Hager: None. S. Arora: None. C. Berry: None. Funding The Leona M. and Harry B. Helmsley Charitable Trust

Prevalence of hematologic conditions in samples received for hemoglobin A1c testing

Fetal hemoglobin (HbF) is a strong modifier of sickle cell disease (SCD) severity and is associated with 3 common genetic loci. Quantifying the genetic effects of the 3 loci would specifically address the benefits of HbF increases in patients. Here, we have applied statistical methods using the most representative variants: rs1427407 and rs6545816 in BCL11A, rs66650371 (3-bp deletion) and rs9376090 in HMIP-2A, rs9494142 and rs9494145 in HMIP-2B, and rs7482144 (Xmn1-HBG2 in the β-globin locus) to create g(HbF), a genetic quantitative variable for HbF in SCD. Only patients aged ≥5 years with complete genotype and HbF data were studied. Five hundred eighty-one patients with hemoglobin SS (HbSS) or HbSβ0 thalassemia formed the “discovery” cohort. Multiple linear regression modeling rationalized the 7 variants down to 4 markers (rs6545816, rs1427407, rs66650371, and rs7482144) each independently contributing HbF-boosting alleles, together accounting for 21.8% of HbF variability (r2) in the HbSS or HbSβ0 patients. The model was replicated with consistent r2 in 2 different cohorts: 27.5% in HbSC patients (N = 186) and 23% in 994 Tanzanian HbSS patients. g(HbF), our 4-variant model, provides a robust approach to account for the genetic component of HbF in SCD and is of potential utility in sickle genetic and clinical studies.

Findings From a National Diabetes Survey: Highlighting Progress and Opportunities for Diabetes Prevention and Care.

The National Diabetes Education Program (NDEP) implemented periodic deployment of the NDEP National Diabetes Survey (NNDS) in 2006 to collect data on diabetes-related knowledge, attitudes, and behaviors among U.S. adults with diabetes and on risk-reduction attitudes and behaviors among people at risk for type 2 diabetes. Because of similarities in the nationally representative samples of the past three surveys, trends between 2011 and 2016 can be assessed. We present key findings of the 2016 NNDS, focusing on areas of interest to diabetes educators. The 2016 NNDS identified both progress and gaps. Eighty-nine percent of people with diabetes are aware of the A1C test, and disparities by race and ethnicity are narrowing. Awareness of the link between cardiovascular disease and diabetes has plateaued at 75% of respondents. Most people with diabetes do not feel comfortable managing hyper- or hypoglycemia (65 and 60%, respectively), and 60-69% report some level of diabetes distress. Among people without diabetes, awareness of personal risk of type 2 diabetes has increased from 30% in 2011 to 45% in 2016. A diagnosis of prediabetes significantly increases awareness of personal risk of diabetes (65% of people with prediabetes vs. 45% of those at risk), likelihood of receiving counseling about reducing risk (45 vs. 11%), and likelihood of taking action to reduce risk (71 vs. 52%). However, 33% of people without diabetes report not planning to take action to lower their risk of type 2 diabetes. The 2016 NNDS results suggest that outreach and educational efforts have increased knowledge about diabetes and diabetes risk, but also point to the need for more education and support for people with diabetes and those at risk.

PDB85 IMPACT OF TYPE OF INSURANCE COVERAGE ON QUALITY OF DIABETES CARE

To evaluate differences in receipt of diabetes care services: A1C test, foot examination, dilated eye examination, flu vaccination and serum cholesterol test between privately health-insured and publicly health-insured diabetic subjects. Using the household component Medical Expenditure Panel Data Survey (MEPS) 2015 consolidated data, a secondary data analysis was carried out. Logistic regression was used to separately model the effect of insurance type on the receipt of each of the diabetes-care quality indicators: dilated eye examination, serum cholesterol test, foot examination, flu vaccination and A1C test. For subjects aged 18-64 yrs., private insurance-only versus public insurance-only users were compared using 348 eligible subjects, while for subjects above 64 yrs., Medicare-only versus Medicare plus private insurance users were compared using sample size of 240. Effects of age, education level, gender, income and race were controlled. For subjects 18-64 yrs., the receipt of eye examination (OR 0.85, 95% CI: 0.51-1.41); foot examination (OR 0.716, 95% CI: 0.42-1.22), A1C test (OR 0.73, 95% CI: 0.40-1.34); flu vaccination (OR 0.901, 95% CI: 0.55-1.48); and serum cholesterol test (OR 1.35, 95% CI: 0.64-2.87); were not significantly associated with type of insurance coverage. Comparing Medicare versus Medicare plus private coverage for subjects above 64 yrs., the receipt of eye examination (OR 1.76, 95% CI: 0.88-3.50); feet examination (OR 1.13, 95% CI: 0.57-2.26), A1C test (OR 0.89, 95% CI: 0.38-2.08); and flu vaccination (OR 1.26, 95% CI: 0.63-2.50) were not significantly associated with type of coverage while odds of blood cholesterol test is higher (OR 1.35, 95% CI: 0.64-2.87) with Medicare plus private coverage users. Receipt of most key preventive and monitoring services among people with diabetes is similar between private and public insurance users, however elderly diabetic Medicare users with any additional private coverage are a little more likely to receive blood cholesterol test than Medicare-only users.