Initial development of tools to identify child abuse and neglect in pediatric primary care.

BACKGROUND: Child abuse and neglect (CAN) is prevalent, associated with long-term adversities, and often undetected. Primary care settings offer a unique opportunity to identify CAN and facilitate referrals, when warranted. Electronic health records (EHR) contain extensive information to support healthcare decisions, yet time constraints preclude most providers from thorough EHR reviews that could indicate CAN. Strategies that summarize EHR data to identify CAN and convey this to providers has potential to mitigate CAN-related sequelae. This study used expert review/consensus and Natural Language Processing (NLP) to develop and test a lexicon to characterize children who have experienced or are at risk for CAN and compared machine learning methods to the lexicon + NLP approach to determine the algorithm's performance for identifying CAN. METHODS: Study investigators identified 90 CAN terms and invited an interdisciplinary group of child abuse experts for review and validation. We then used NLP to develop pipelines to finalize the CAN lexicon. Data for pipeline development and refinement were drawn from a randomly selected sample of EHR from patients seen at pediatric primary care clinics within a U.S. academic health center. To explore a machine learning approach for CAN identification, we used Support Vector Machine algorithms. RESULTS: The investigator-generated list of 90 CAN terms were reviewed and validated by 25 invited experts, resulting in a final pool of 133 terms. NLP utilized a randomly selected sample of 14,393 clinical notes from 153 patients to test the lexicon, and .03% of notes were identified as CAN positive. CAN identification varied by clinical note type, with few differences found by provider type (physicians versus nurses, social workers, etc.). An evaluation of the final NLP pipelines indicated 93.8% positive CAN rate for the training set and 71.4% for the test set, with decreased precision attributed primarily to false positives. For the machine learning approach, SVM pipeline performance was 92% for CAN + and 100% for non-CAN, indicating higher sensitivity than specificity. CONCLUSIONS: The NLP algorithm's development and refinement suggest that innovative tools can identify youth at risk for CAN. The next key step is to refine the NLP algorithm to eventually funnel this information to care providers to guide clinical decision making.

Enhancing research data infrastructure to address the opioid epidemic: the Opioid Overdose Network (O2-Net).

Opioid Overdose Network is an effort to generalize and adapt an existing research data network, the Accrual to Clinical Trials (ACT) Network, to support design of trials for survivors of opioid overdoses presenting to emergency departments (ED). Four institutions (Medical University of South Carolina [MUSC], Dartmouth Medical School [DMS], University of Kentucky [UK], and University of California San Diego [UCSD]) worked to adapt the ACT network. The approach that was taken to enhance the ACT network focused on 4 activities: cloning and extending the ACT infrastructure, developing an e-phenotype and corresponding registry, developing portable natural language processing tools to enhance data capture, and developing automated documentation templates to enhance extended data capture. Overall, initial results suggest that tailoring of existing multipurpose federated research networks to specific tasks is feasible; however, substantial efforts are required for coordination of the subnetwork and development of new tools for extension of available data. The initial output of the project was a new approach to decision support for the prescription of naloxone for home use in the ED, which is under further study within the network.

Automatically identifying opioid use disorder in non-cancer patients on chronic opioid therapy.

Background: Using the International Classification of Diseases (ICD) codes alone to record opioid use disorder (OUD) may not completely document OUD in the electronic health record (EHR). We developed and evaluated natural language processing (NLP) approaches to identify OUD from the clinal note. We explored the concordance between ICD-coded and NLP-identified OUD.Methods: We studied EHRs from 13,654 (female: 8223; male: 5431) adult non-cancer patients who received chronic opioid therapy (COT) and had at least one clinical note between 2013 and 2018. Of eligible patients, we randomly selected 10,218 (75%) patients as the training set and the remaining 3436 patients (25%) as the test dataset for NLP approaches.Results: We generated 539 terms representing OUD mentions in clinical notes (e.g., "opioid use disorder," "opioid abuse," "opioid dependence," "opioid overdose") and 73 terms representing OUD medication treatments. By domain expert manual review for the test dataset, our NLP approach yielded high performance: 98.5% for precision, 100% for recall, and 99.2% for F-measure. The concordance of these NLP and ICD identified OUD was modest (Kappa = 0.63).Conclusions: Our NLP approach can accurately identify OUD patients from clinical notes. The combined use of ICD diagnostic code and NLP approach can improve OUD identification.

Correction to: Automatically identifying social isolation from clinical narratives for patients with prostate Cancer.

Following publication of the original article [1], the authors reported an error in one of the authors' names.

Automatically identifying social isolation from clinical narratives for patients with prostate Cancer.

BACKGROUND: Social isolation is an important social determinant that impacts health outcomes and mortality among patients. The National Academy of Medicine recently recommended that social isolation be documented in electronic health records (EHR). However, social isolation usually is not recorded or obtained as coded data but rather collected from patient self-report or documented in clinical narratives. This study explores the feasibility and effectiveness of natural language processing (NLP) strategy for identifying patients who are socially isolated from clinical narratives. METHOD: We used data from the Medical University of South Carolina (MUSC) Research Data Warehouse. Patients 18 years-of-age or older who were diagnosed with prostate cancer between January 1, 2014 and May 31, 2017 were eligible for this study. NLP pipelines identifying social isolation were developed via extraction of notes on progress, history and physical, consult, emergency department provider, telephone encounter, discharge summary, plan of care, and radiation oncology. Of 4195 eligible prostate cancer patients, we randomly sampled 3138 patients (75%) as a training dataset. The remaining 1057 patients (25%) were used as a test dataset to evaluate NLP algorithm performance. Standard performance measures for the NLP algorithm, including precision, recall, and F-measure, were assessed by expert manual review using the test dataset. RESULTS: A total of 55,516 clinical notes from 3138 patients were included to develop the lexicon and NLP pipelines for social isolation. Of those, 35 unique patients (1.2%) had social isolation mention(s) in 217 notes. Among 24 terms relevant to social isolation, the most prevalent were "lack of social support," "lonely," "social isolation," "no friends," and "loneliness". Among 1057 patients in the test dataset, 17 patients (1.6%) were identified as having social isolation mention(s) in 40 clinical notes. Manual review identified four false positive mentions of social isolation and one false negatives in 154 notes from randomly selected 52 controls. The NLP pipeline demonstrated 90% precision, 97% recall, and 93% F-measure. The major reasons for a false positive included the ambiguities of the experiencer of social isolation, negation, and alternate meaning of words. CONCLUSIONS: Our NLP algorithms demonstrate a highly accurate approach to identify social isolation.

Natural language processing and machine learning algorithm to identify brain MRI reports with acute ischemic stroke.

BACKGROUND AND PURPOSE: This project assessed performance of natural language processing (NLP) and machine learning (ML) algorithms for classification of brain MRI radiology reports into acute ischemic stroke (AIS) and non-AIS phenotypes. MATERIALS AND METHODS: All brain MRI reports from a single academic institution over a two year period were randomly divided into 2 groups for ML: training (70%) and testing (30%). Using "quanteda" NLP package, all text data were parsed into tokens to create the data frequency matrix. Ten-fold cross-validation was applied for bias correction of the training set. Labeling for AIS was performed manually, identifying clinical notes. We applied binary logistic regression, naïve Bayesian classification, single decision tree, and support vector machine for the binary classifiers, and we assessed performance of the algorithms by F1-measure. We also assessed how n-grams or term frequency-inverse document frequency weighting affected the performance of the algorithms. RESULTS: Of all 3,204 brain MRI documents, 432 (14.3%) were labeled as AIS. AIS documents were longer in character length than those of non-AIS (median [interquartile range]; 551 [377-681] vs. 309 [164-396]). Of all ML algorithms, single decision tree had the highest F1-measure (93.2) and accuracy (98.0%). Adding bigrams to the ML model improved F1-mesaure of naïve Bayesian classification, but not in others, and term frequency-inverse document frequency weighting to data frequency matrix did not show any additional performance improvements. CONCLUSIONS: Supervised ML based NLP algorithms are useful for automatic classification of brain MRI reports for identification of AIS patients. Single decision tree was the best classifier to identify brain MRI reports with AIS.

Identifying Falls Risk Screenings Not Documented with Administrative Codes Using Natural Language Processing.

Quality reporting that relies on coded administrative data alone may not completely and accurately depict providers' performance. To assess this concern with a test case, we developed and evaluated a natural language processing (NLP) approach to identify falls risk screenings documented in clinical notes of patients without coded falls risk screening data. Extracting information from 1,558 clinical notes (mainly progress notes) from 144 eligible patients, we generated a lexicon of 38 keywords relevant to falls risk screening, 26 terms for pre-negation, and 35 terms for post-negation. The NLP algorithm identified 62 (out of the 144) patients who falls risk screening documented only in clinical notes and not coded. Manual review confirmed 59 patients as true positives and 77 patients as true negatives. Our NLP approach scored 0.92 for precision, 0.95 for recall, and 0.93 for F-measure. These results support the concept of utilizing NLP to enhance healthcare quality reporting.

The Impact of Censoring Drug Switching in Medication Adherence Measures of Chronic Single Ingredient Oral Drugs.

We explored how drug switching impacts adherence measures for common chronic oral medications. Switching between ingredients with the same indication was detected within a 30-day grace period. The proportion of days covered (PDC) and adherent status (cutoff 0.8) for each ingredient was calculated and compared between different censoring approaches: censoring drug switching (PDCswitch), censoring the end of dispensing (PDCend), and fixed 365-day period (PDC365). Overall, 854,380 (15.9%) patients in the Optum ClinFormatics (Optum) and 150,785 (22.0%) patients in the MarketScan Multi-state Medicaid (MDCD) had at least one switch within one year. Compared with PDC365 in Optum, PDCswitch means were higher: 0.85 vs. 0.41 for antihypertensive, 0.82 vs. 0.46 for antihyperglycemics, and 0.84 vs. 0.33 for antihyerlipidemia. Further, the percentages of adherent patients were higher: 95.8% vs. 17.9% for antihypertensive, 85.5% vs. 18.9% for antihyperglycemics, and 72.1% vs. 5.3% for antihyerlipidemia. Significant and modest changes were observed between PDCswitch and PDCend.

Depression among people with type 2 diabetes mellitus, US National Health and Nutrition Examination Survey (NHANES), 2005-2012.

BACKGROUND: Depression in people with diabetes can result in increased risk for diabetes-related complications. The prevalence of depression has been estimated to be 17.6 % in people with type 2 diabetes mellitus (T2DM), based on studies published between 1980 and 2005. There is a lack of more recent estimates of depression prevalence among the US general T2DM population. METHODS: The present study used the US National Health and Nutrition Examination Survey (NHANES) 2005-2012 data to provide an updated, population-based estimate for the prevalence of depression in people with T2DM. NHANES is a cross-sectional survey of a nationally representative sample of the civilian, non-institutionalized US population. Starting from 2005, the Patient Health Questionnaire (PHQ-9) was included to measure signs and symptoms of depression. We defined PHQ-9 total scores ≥ 10 as clinically relevant depression (CRD), and ≥ 15 as clinically significant depression (CSD). Self-reported current antidepressant use was also combined to estimate overall burden of depression. Predictors of CRD and CSD were investigated using survey logistic regression models. RESULTS: A total of 2182 participants with T2DM were identified. The overall prevalence of CRD and CSD among people with T2DM is 10.6 % (95 % confidence interval (CI) 8.9-12.2 %), and 4.2 % (95 % CI 3.4-5.1 %), respectively. The combined burden of depressive symptoms and antidepressants may be as high as 25.4 % (95 % CI 23.0-27.9 %). Significant predictors of CRD include age (younger than 65), sex (women), income (lower than 130 % of poverty level), education (below college), smoking (current or former smoker), body mass index (≥30 kg/m(2)), sleep problems, hospitalization in the past year, and total cholesterol (≥200 mg/dl). Significant predictors of CSD also include physical activity (below guideline) and cardiovascular diseases. CONCLUSIONS: The prevalence of CRD and CSD among people with T2DM in the US may be lower than in earlier studies, however, the burden of depression remains high. Further research with longitudinal follow-up for depression in people with T2DM is needed to understand real world effectiveness of depression management.

Changes in Body Weight Among People With Type 2 Diabetes Mellitus in the United States, NHANES 2005-2012.

PURPOSE: To understand weight loss strategies, weight changes, goals, and behaviors in people with type 2 diabetes mellitus (T2DM) and whether these differ by ethnicity. METHODS: T2DM was identified by self-reported diagnosis using the NHANES 2005-2012 data, which also included measured and self-reported current body weight and height, self-reported weight the prior year, and self-reported aspired weight. Nineteen weight loss strategies were evaluated for association with ≥5% weight loss or weight gain versus <5% weight change. RESULTS: Among people with T2DM, 88.0% were overweight/obese (body mass index [BMI] ≥25 kg/m(2)) in the prior year and 86.1% the current year. About 60% of the overweight/obese took weight loss actions, mostly using diet-related methods with average weight lost <5%. Two most "effective" methods reported (smoking, taking laxatives/vomiting) are also potentially most harmful. Similar BMI distributions but different goals and behaviors about weight and weight loss were observed across ethnicity. Only physical activity meeting the recommended level and changing eating habits were consistently associated with favorable and statistically significant weight change. CONCLUSIONS: Weight management in T2DM is an ongoing challenge, regardless of ethnicity/race. Among overweight/obese T2DM subjects, recommended level of physical activity and changing eating habits were associated with statistically significant favorable weight change.

Nonadherence to Oral Antihyperglycemic Agents: Subsequent Hospitalization and Mortality among Patients with Type 2 Diabetes in Clinical Practice.

Using real-world clinical data from the Indiana Network for Patient Care, we analyzed the associations between non-adherence to oral antihyperglycemic agents (OHA) and subsequent diabetes-related hospitalization and all-cause mortality for patients with type 2 diabetes. OHA adherence was measured by the annual proportion of days covered (PDC) for 2008 and 2009. Among 24,067 eligible patients, 35,507 annual PDCs were formed. Over 90% (n=21,798) of the patients had a PDC less than 80%. In generalized linear mixed model analyses, OHA non-adherence is significantly associated with diabetes related hospitalizations (OR: 1.2; 95% CI [1.1,1.3]; p<0.0001). Older patients, white patients, or patients who had ischemic heart disease, stroke, or renal disease had higher odds of hospitalization. Similarly, OHA non-adherence increased subsequent mortality (OR: 1.3; 95% CI [1.02, 1.61]; p<0.0001). Patient age, male gender, income and presence of ischemic heart diseases, stroke, and renal disease were also significantly associated with subsequent all-cause death.

Effect of Opioids on Testosterone Levels: Cross-Sectional Study using NHANES.

OBJECTIVE: Opioids can suppress gonadal hormone production, which may result in low testosterone levels. To date, there have been no large-scale population-based studies examining the extent to which opioid use may contribute to changes in testosterone levels. DESIGN: Cross-sectional study. SETTING: 2011-2012 National Health and Nutrition Examination Survey. SUBJECTS: Participants 17 years and older who had data on prescription medication usage and serum testosterone levels available. Participants were divided in two groups, opioid exposed and unexposed. METHODS: Testosterone levels of participants who responded that they had been exposed (n = 320) to prescription opioids in the past 30 days were compared with those who were unexposed (n = 4909). The number of participants with low testosterone levels was calculated and unadjusted and adjusted analyses were performed. RESULTS: Participants on opioids had higher odds of having low testosterone levels than those unexposed, odd ratio (OR) = 1.40, 95% confidence interval (CI) (1.07-1.84). After controlling for opioid exposure, as the age and the number of comorbidities increased, the odds of having low testosterone levels significantly increased in all categories. Compared with participants between 17 and 45 years of age, participants >70 years had OR = 1.70, 95% CI (1.16-2.50). Compared with participants with no comorbidities, participants with >2 comorbidities had OR = 1.69 95% CI (1.24-2.30). CONCLUSION: When assessing the impact of opioids on testosterone, the effects of age and medical conditions should be considered.

Feasibility and utility of applications of the common data model to multiple, disparate observational health databases.

OBJECTIVES: To evaluate the utility of applying the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) across multiple observational databases within an organization and to apply standardized analytics tools for conducting observational research. MATERIALS AND METHODS: Six deidentified patient-level datasets were transformed to the OMOP CDM. We evaluated the extent of information loss that occurred through the standardization process. We developed a standardized analytic tool to replicate the cohort construction process from a published epidemiology protocol and applied the analysis to all 6 databases to assess time-to-execution and comparability of results. RESULTS: Transformation to the CDM resulted in minimal information loss across all 6 databases. Patients and observations excluded were due to identified data quality issues in the source system, 96% to 99% of condition records and 90% to 99% of drug records were successfully mapped into the CDM using the standard vocabulary. The full cohort replication and descriptive baseline summary was executed for 2 cohorts in 6 databases in less than 1 hour. DISCUSSION: The standardization process improved data quality, increased efficiency, and facilitated cross-database comparisons to support a more systematic approach to observational research. Comparisons across data sources showed consistency in the impact of inclusion criteria, using the protocol and identified differences in patient characteristics and coding practices across databases. CONCLUSION: Standardizing data structure (through a CDM), content (through a standard vocabulary with source code mappings), and analytics can enable an institution to apply a network-based approach to observational research across multiple, disparate observational health databases.

Chronic kidney disease in US adults with type 2 diabetes: an updated national estimate of prevalence based on Kidney Disease: Improving Global Outcomes (KDIGO) staging.

BACKGROUND: Kidney Disease Improving Global Outcomes (KDIGO) 2013 updated the classification and risk stratification of chronic kidney disease (CKD) to include both the level of renal function and urinary albumin excretion (UAE). The update subclassifies the previous category of moderate renal impairment. There is currently limited information on the prevalence of CKD based on this new classification in United States (US) adults with type 2 diabetes mellitus (T2DM). The objective of this study was to provide such estimates, for T2DM both overall and in those ≥ 65 years of age. We used the continuous National Health and Nutrition Examination Survey (NHANES) 1999-2012 to identify participants with T2DM. Estimated glomerular filtration rate (eGFR) and UAE were calculated using laboratory results and data collected from the surveys, and categorized based on the KDIGO classification. Projections for the US T2DM population were based on NHANES sampling weights. RESULTS: A total of 2915 adults diagnosed with T2DM were identified from NHANES, with 1466 being age ≥ 65 years. Prevalence of CKD based on either eGFR or UAE was 43.5% in the T2DM population overall, and 61.0% in those age ≥ 65 years. The prevalence of mildly decreased renal function or worse (eGFR < 60/ml/min/1.73 m2) was 22.0% overall and 43.1% in those age ≥ 65 years. Prevalence of more severe renal impairment (eGFR < 45 ml/min/1.73 m2) was 9.0% overall and 18.6% in those age ≥ 65 years. The prevalence of elevated UAE (> 30 mg/g) was 32.2% overall and 39.1% in those age ≥ 65 years. The most common comorbidities were hypertension, retinopathy, coronary heart disease, myocardial infarction, and congestive heart failure. CONCLUSIONS: This study confirms the high prevalence of CKD in T2DM, impacting 43.5% of this population. Additionally, this study is among the first to report US prevalence estimates of CKD based on the new KDIGO CKD staging system.

Exposure to oral fluoroquinolones and the risk of retinal detachment: retrospective analyses of two large healthcare databases.

BACKGROUND: A recent Canadian case-control study reported a 4.5-fold increased risk of retinal detachment (RD) during oral fluoroquinolone use. Of the fluoroquinolone-exposed cases, 83 % were exposed to ciprofloxacin. We sought to replicate this finding, and assess whether it applied to all fluoroquinolones. METHODS: In two large US healthcare databases, we performed three case-control analyses: one replicating the recent study; one addressing additional potential confounders; and one that increased sample size by dropping the Canadian study's requirement for a prior ophthalmologist visit. We also performed a self-controlled case-series (SCCS) analysis in which each subject served as his or her own comparator. RESULTS: In the replication case-control analyses, the adjusted odds ratios (ORs) for any exposure to fluoroquinolones or ciprofloxacin were approximately 1.2 in both databases, and were statistically significant, and the ORs for current exposure were modestly above 1 in one database, modestly below 1 in the other, and not statistically significant. In the other case-control analyses, the ORs were close to 1. In a post hoc age-stratified case-control analysis, we observed an association of RD with fluoroquinolone exposure among older subjects in one of the two databases. All estimates from the SCCS analyses were below 1.2 and none was statistically significant. CONCLUSION: The present study does not confirm the recent Canadian study's finding of a strong relationship between RD and current exposure to fluoroquinolones. Instead, it found a modest association between RD and current or any exposure to fluoroquinolones in the case-control analyses, and no association in the SCCS analyses.

A Comparison of Data Driven-based Measures of Adherence to Oral Hypoglycemic Agents in Medicaid Patients.

We evaluated and compared different methods for measuring adherence to Oral Antihyperglycemic Agents (OHA), based on the correlation between these measures and glycated hemoglobin A1C (HbA1c) levels in Medicaid patients with Type 2 diabetes. An observational sample of 831 Medicaid patients with Type 2 diabetes who had HbA1c test results recorded between January 1, 2001 and December 31, 2005 was identified in the Indiana Network of Patient Care (INPC). OHA adherence was measured by medication possession ratio (MPR), proportion of days covered (PDC), and the number of gaps (GAP) for 3, 6, and 12-month intervals prior to the HbA1c test date. All three OHA adherence measurements showed consistent and significant correlation with HbA1c level. The 6-month PDC showed the strongest association with HbA1c levels in both unadjusted (-1.07, P<0.0001) and adjusted (-1.12, P<0.0001) models.

Data for drugs available through low-cost prescription drug programs are available through pharmacy benefit manager and claims data.

BACKGROUND: Observational data are increasingly being used for pharmacoepidemiological, health services and clinical effectiveness research. Since pharmacies first introduced low-cost prescription programs (LCPP), researchers have worried that data about the medications provided through these programs might not be available in observational data derived from administrative sources, such as payer claims or pharmacy benefit management (PBM) company transactions. METHOD: We used data from the Indiana Network for Patient Care to estimate the proportion of patients with type 2 diabetes to whom an oral hypoglycemic agent was dispensed. Based on these estimates, we compared the proportions of patients who received medications from chains that do and do not offer an LCPP, the proportion trend over time based on claims data from a single payer, and to proportions estimated from the Medical Expenditure Panel Survey (MEPS). RESULTS: We found that the proportion of patients with type 2 diabetes who received oral hypoglycemic medications did not vary based on whether the chain that dispensed the drug offered an LCPP or over time. Additionally, the rates were comparable to those estimated from MEPS. CONCLUSION: Researchers can be reassured that data for medications available through LCPPs continue to be available through administrative data sources.

Evaluation of a clinical decision support algorithm for patient-specific childhood immunization.

OBJECTIVE: To evaluate the effectiveness of a clinical decision support system (CDSS) implementing standard childhood immunization guidelines, using real-world patient data from the Regenstrief Medical Record System (RMRS). METHODS: Study subjects were age 6-years or younger in 2008 and had visited the pediatric clinic on the campus of Wishard Memorial Hospital. Immunization records were retrieved from the RMRS for 135 randomly selected pediatric patients. We compared vaccine recommendations from the CDSS for both eligible and recommended timelines, based on the child's date of birth and vaccine history, to recommendations from registered nurses who routinely selected vaccines for administration in a busy inner city hospital, using the same date of birth and vaccine history. Aggregated and stratified agreement and Kappa statistics were reported. The reasons for disagreement between suggestions from the CDSS and nurses were also identified. RESULTS: For the 135 children, a total of 1215 vaccination suggestions were generated by nurses and were compared to the recommendations of the CDSS. The overall agreement rates were 81.3% and 90.6% for the eligible and recommended timelines, respectively. The overall Kappa values were 0.63 for the eligible timeline and 0.80 for the recommended timeline. Common reasons for disagreement between the CDSS and nurses were: (1) missed vaccination opportunities by nurses, (2) nurses sometimes suggested a vaccination before the minimal age and minimal waiting interval, (3) nurses usually did not validate patient immunization history, and (4) nurses sometimes gave an extra vaccine dose. CONCLUSION: Our childhood immunization CDSS can assist providers in delivering accurate childhood vaccinations.

Race and medication adherence and glycemic control: findings from an operational health information exchange.

The Central Indiana Beacon Community leads efforts for improving adherence to oral hypoglycemic agents (OHA) to achieve improvements in glycemic control for patients with type 2 diabetes. In this study, we explored how OHA adherence affected hemoglobin A1C (HbA1c) level in different racial groups. OHA adherence was measured by 6-month proportion of days covered (PDC). Of 3,976 eligible subjects, 12,874 pairs of 6-month PDC and HbA1c levels were formed between 2002 and 2008. The average HbA1c levels were 7.4% for African-Americans and 6.5% for Whites. The average 6-month PDCs were 40% for African-Americans and 50% for Whites. In mixed effect generalized linear regression analyses, OHA adherence was inversely correlated with HbA1c level for both African-Americans (-0.80, p<0.0001) and Whites (-0.53, p<0.0001). The coefficient was -0.26 (p<0.0001) for the interaction of 6-month PDC and African-Americans. Significant risk factors for OHA non-adherence were race, young age, non-commercial insurance, newly-treated status, and polypharmacy.