Guideline-concordant breast cancer care by patient race and ethnicity accounting for individual-, facility- and area-level characteristics: a SEER-Medicare study.

PURPOSE: To examine racial-ethnic variation in adherence to established quality metrics (NCCN guidelines and ASCO quality metrics) for breast cancer, accounting for individual-, facility-, and area-level factors. METHODS: Data from women diagnosed with invasive breast cancer at 66+ years of age from 2000 to 2017 were examined using SEER-Medicare. Associations between race and ethnicity and guideline-concordant diagnostics, locoregional treatment, systemic therapy, documented stage, and oncologist encounters were estimated using multilevel logistic regression models to account for clustering within facilities or counties. RESULTS: Black and American Indian/Alaska Native (AIAN) women had consistently lower odds of guideline-recommended care than non-Hispanic White (NHW) women (Diagnostic workup: ORBlack 0.83 (0.79-0.88), ORAIAN 0.66 (0.54-0.81); known stage: ORBlack 0.87 (0.80-0.94), ORAIAN 0.63 (0.47-0.85); seeing an oncologist: ORBlack 0.75 (0.71-0.79), ORAIAN 0.60 (0.47-0.72); locoregional treatment: ORBlack 0.80 (0.76-0.84), ORAIAN 0.84 (0.68-1.02); systemic therapies: ORBlack 0.90 (0.83-0.98), ORAIAN 0.66 (0.48-0.91)). Commission on Cancer accreditation and facility volume were significantly associated with higher odds of guideline-concordant diagnostics, stage, oncologist visits, and systemic therapy. Black residential segregation was associated with significantly lower odds of guideline-concordant locoregional treatment and systemic therapy. Rurality and area SES were associated with significantly lower odds of guideline-concordant diagnostics and oncologist visits. CONCLUSIONS: This is the first study to examine guideline-concordance across the continuum of breast cancer care from diagnosis to treatment initiation. Disparities were present from the diagnostic phase and persisted throughout the clinical course. Facility and area characteristics may facilitate or pose barriers to guideline-adherent treatment and warrant future investigation as mediators of racial-ethnic disparities in breast cancer care.

A Nationwide Emergency Department Data Analysis to Predict Beers List Medications Use Among Older Adults.

BACKGROUND: The use of potentially inappropriate medications (PIMs) is considered an important quality indicator for older adults seen in the ambulatory care setting. STUDY OBJECTIVES: To evaluate the pattern of potentially inappropriate medication (PIMs) use as specified in the Beers Criteria, for older adults during emergency department (ED) visits in the United States. METHODS: Using data from the National Hospital Ambulatory Care Survey (NHAMCS) we identified older adults (age 65 or older) discharged home from an ED visit in 2019. We defined PIMs as those with an 'avoid' recommendation under the American Geriatrics Society (AGS) 2019 Beers Criteria in older adults. Logistic regression models were used to assess demographic, clinical, and hospital factors associated with the use of any PIMs upon ED discharge. RESULTS: Overall, 5.9% of visits by older adults discharged from the ED included administration or prescriptions for PIMs. Among those who received any PIMs, 25.5% received benzodiazepines, 42.5 % received anticholinergics, 1.4% received nonbenzodiazepine hypnotics, and 0.5% received barbiturates. A multivariable model showed statistically significant associations for age 65 to 74 (OR 1.91, 95% CI 1.39-2.62 vs. age >=75), dementia (OR 0.45, 95% CI 0.21-0.95), lower immediacy (OR 2.45, 95% CI 1.56-3.84 vs. higher immediacy), and Northeastern rural region (OR 0.34, 95% CI 0.21-0.55 vs. Midwestern rural). CONCLUSION: We found that younger age and lower immediacy were associated with increased prescriptions of PIMs for older adults seen, while dementia and Northeastern rural region was associated with reduced use of PIMs seen and discharged from EDs in United States.

Evaluating the performance of machine learning methods for risk estimation of delirium in patients hospitalized from the emergency department.

INTRODUCTION: Delirium is a cerebral dysfunction seen commonly in the acute care setting. It is associated with increased mortality and morbidity and is frequently missed in the emergency department (ED) and inpatient care by clinical gestalt alone. Identifying those at risk of delirium may help prioritize screening and interventions in the hospital setting. OBJECTIVE: Our objective was to leverage electronic health records to identify a clinically valuable risk estimation model for prevalent delirium in patients being transferred from the ED to inpatient units. METHODS: This was a retrospective cohort study to develop and validate a risk model to detect delirium using patient data available from prior visits and ED encounter. Electronic health records were extracted for patients hospitalized from the ED between January 1, 2014, and December 31, 2020. Eligible patients were aged 65 or older, admitted to an inpatient unit from the emergency department, and had at least one DOSS assessment or CAM-ICU recorded within 72 h of hospitalization. Six machine learning models were developed to estimate the risk of delirium using clinical variables including demographic features, physiological measurements, medications administered, lab results, and diagnoses. RESULTS: A total of 28,531 patients met the inclusion criteria with 8057 (28.4%) having a positive delirium screening within the outcome observation period. Machine learning models were compared using the area under the receiver operating curve (AUC). The gradient boosted machine achieved the best performance with an AUC of 0.839 (95% CI, 0.837-0.841). At a 90% sensitivity threshold, this model achieved a specificity of 53.5% (95% CI 53.0%-54.0%) a positive predictive value of 43.5% (95% CI 43.2%-43.9%), and a negative predictive value of 93.1% (95% CI 93.1%-93.2%). A random forest model and L1-penalized logistic regression also demonstrated notable performance with AUCs of 0.837 (95% CI, 0.835-0.838) and 0.831 (95% CI, 0.830-0.833) respectively. CONCLUSION: This study demonstrated the use of machine learning algorithms to identify a combination of variables that enables an estimation of risk of positive delirium screens early in hospitalization to develop prevention or management protocols.

Telehealth Utilization Is Associated with Lower Risk of Discontinuation of Buprenorphine: a Retrospective Cohort Study of US Veterans.

OBJECTIVE: Treatment for opioid use disorder (OUD) may include a combination of pharmacotherapies (such as buprenorphine) with counseling services if clinically indicated. Medication management or engagement with in-person counseling services may be hindered by logistical and financial barriers. Telehealth may provide an alternative mechanism for continued engagement. This study aimed to evaluate the association between telehealth encounters and time to discontinuation of buprenorphine treatment when compared to traditional in-person visits and to evaluate potential effect modification by rural-urban designation and in-person and telehealth combination treatment. METHODS: A retrospective cohort study of Veterans diagnosed with OUD and treated with buprenorphine across all facilities within the Veterans Health Administration (VHA) between 2008 and 2017. Exposures were telehealth and in-person encounters for substance use disorder (SUD) and mental health, treated as time-varying covariates. The primary outcome was treatment discontinuation, evaluated as 14 days of absence of medication from initiation through 1 year. RESULTS: Compared to in-person encounters, treatment discontinuation was lower for telehealth for SUD (aHR: 0.69; 95%CI: 0.60, 0.78) and mental health (aHR: 0.69; 95%CI: 0.62, 0.76). There was no evidence of effect modification by rural-urban designation. Risk of treatment discontinuation appeared to be lower among those with telehealth only compared to in-person only for both SUD (aHR: 0.48, 95%CI: 0.37, 0.62) and for mental health (aHR: 0.46; 95%CI: 0.33, 0.65). CONCLUSIONS: As telehealth demonstrated improved treatment retention compared to in-person visits, it may be a suitable option for engagement for patients in OUD management. Efforts to expand services may improve treatment retention and health outcomes for VHA and other health care systems.

Gynecologic oncologist impact on adjuvant chemotherapy care for stage II-IV ovarian cancer patients.

OBJECTIVE: We aim to evaluate the impact gynecologic oncologists have on ovarian cancer adjuvant chemotherapy care from their role as surgeons recommending adjuvant chemotherapy care and their role as adjuvant chemotherapy providers while considering rural-urban differences. METHODS: Multivariable adjusted logistic regressions and Cox proportional hazards models were developed using a population-based, retrospective cohort of stage II-IV and unknown stage ovarian cancer patients diagnosed in Iowa, Kansas, and Missouri in 2010-2012 whose medical records were abstracted in 2017-2018. RESULTS: Gynecologic oncologist surgeons (versus other type of surgeon) were associated with increased odds of adjuvant chemotherapy initiation (adjusted odds ratio (OR) 2.18; 95% confidence interval (CI) 1.10-4.33) and having a gynecologic oncologist adjuvant chemotherapy provider (OR 10.0; 95% CI 4.58-21.8). Independent of type of surgeon, rural patients were less likely to have a gynecologic oncologist chemotherapy provider (OR 0.52; 95% CI 0.30-0.91). Gynecologic oncologist adjuvant chemotherapy providers (versus other providers) were associated with decreased surgery-to-chemotherapy time (rural: 6 days; urban: 8 days) and increased distance to chemotherapy (rural: 22 miles; urban: 11 miles). Rural women (versus urban) traveled 38 miles farther when their chemotherapy provider was a gynecologic oncologist and 27 miles farther when it was not. CONCLUSION: Gynecologic oncologist surgeons may impact adjuvant chemotherapy initiation. Gynecologic oncologists serving as adjuvant chemotherapy providers were associated with some care benefits, such as reduced time from surgery-to-chemotherapy, and some care barriers, such as travel distance. The barriers and benefits of having a gynecologic oncologist involved in adjuvant chemotherapy care, including rural-urban differences, warrant further research in other populations.

Self-controlled assessment of thromboembolic event (TEE) risk following intravenous immune globulin (IGIV) in the U.S. (2006-2012).

Since 2013, the U.S. Food and Drug administration (FDA) has required that intravenous immune globulin (IGIV) products carry a boxed warning concerning the risk of thromboembolic events (TEEs). This study assessed the incidence of TEEs attributable to IGIV in a large population-based cohort. A self-controlled risk interval design was used to quantify the transient increase in TEE risk during the risk interval (days 0-2 and 0-13 following IGIV for arterial and venous TEEs, respectively) relative to a later control interval (days 14-27 following IGIV). Potential IGIV-exposed TEE cases from 2006 to 2012 were identified from the FDA-sponsored Sentinel Distributed Database and confirmed through medical record review. Inpatient IGIV exposures were not included in the venous TEE analysis due to concerns about time-varying confounding. 19,069 new users of IGIV who received 93,555 treatment episodes were included. Charts were retrieved for 62% and 70% of potential venous and arterial cases, respectively. There was a transient increase in the risk of arterial TEEs during days 0-2 following IGIV treatment (RR = 4.69; 95% CI 1.87, 11.90; absolute increase in risk = 8.86 events per 10,000 patients, 95% CI 3.25, 14.6), but no significant increase in venous TEE risk during days 0-13 following outpatient IGIV treatments (RR = 1.07, 95% CI 0.34, 3.48). Our results suggest there is a small increase in the absolute risk of arterial TEEs following IGIV. However, lower-than-expected chart retrieval rates and the possibility of time-varying confounding mean that our results should be interpreted cautiously. Continued pharmacovigilance efforts are warranted.

Rural disparities in surgical care from gynecologic oncologists among Midwestern ovarian cancer patients.

OBJECTIVE: Up to one-third of women with ovarian cancer in the United States do not receive surgical care from a gynecologic oncologist specialist despite guideline recommendations. We aim to investigate the impact of rurality on receiving surgical care from a specialist, referral to a specialist, and specialist surgery after referral, and the consequences of specialist care. METHODS: We utilized a retrospective cohort created through an extension of standard cancer surveillance in three Midwestern states. Multivariable adjusted logistic regression was utilized to assess gynecologic oncologist treatment of women 18-89 years old, who were diagnosed with primary, histologically confirmed, malignant ovarian cancer in 2010-2012 in Kansas, Missouri and Iowa by rurality. RESULTS: Rural women were significantly less likely to receive surgical care from a gynecologic oncologist specialist (adjusted odds ratio (OR) 0.37, 95% confidence interval (CI) 0.24-0.58) and referral to a specialist (OR 0.37, 95% CI 0.23-0.59) compared to urban women. There was no significant difference in specialist surgery after a referral (OR 0.56, 95% CI 0.26-1.20). Rural women treated surgically by a gynecologic oncologist versus non-specialist were more likely to receive cytoreduction and more complete tumor removal to ≤1 cm. CONCLUSION: There is a large rural-urban difference in receipt of ovarian cancer surgery from a gynecologic oncologist specialist (versus a non-specialist). Disparities in referral rates contribute to the rural-urban difference. Further research will help define the causes of referral disparities, as well as promising strategies to address them.

Risk Factors and Survival Analysis of Spinal Cord Stimulator Explantation.

OBJECTIVE: The treatment failure rate for spinal cord stimulators (SCS) remains unacceptably high, with reports of removal in up to 30% of patients. The purpose of this study is to perform survival and multivariate regression analyses of patients who have undergone SCS explantation in order to identify patient characteristics that may predict treatment failure. MATERIALS AND METHODS: We identified 253 patients who underwent SCS placement using current procedural terminology codes in a private health insurance data base spanning 2003-2016. Patient demographics, opioid use, surgical indications, as well as comorbidities were noted. At least 6 months of continuous claims data before and after implantation were required for inclusion. Patients who underwent explantation were defined as those who underwent removal without replacement within 90 days and had at least 90 days of continuous insurance eligibility following removal. Those who underwent removal for infectious reasons were identified with corresponding diagnosis codes. RESULTS: Of the 252 patients who met the inclusion criteria, 17 (6.7%) underwent SCS explantation. Median follow-up time was 2.0 years. Of those who had their system explanted, six patients (2.8%) had their systems removed for infection and 11 (4.3%) for noninfectious reasons. Bivariate analysis revealed that younger age and tobacco use were associated with an increased likelihood of explantation. The Cox proportional hazards analysis demonstrated that younger age, tobacco use, and the presence of "other" mental health disorders were predictive of explantation. CONCLUSIONS: In a cohort of SCS patients from multiple institutions, this study demonstrates that explantation for noninfectious reasons is more likely in younger patients, tobacco users, and those with certain psychiatric conditions. With an estimated 10% of patients opting to have their devices removed within 5 years of implantation, refining the ability of clinicians to predict who will see benefit from SCS treatment remains necessary.

Association of Mood and Anxiety Disorders and Opioid Prescription Patterns Among Postpartum Women.

BACKGROUND AND OBJECTIVES: Postpartum women represent a large population with opioid exposure who also have an increased risk of experiencing mood and anxiety disorders. However, the effect that mood and anxiety disorders have on opioid use postpartum has received little attention in the literature. Therefore, the objective of this study was to examine the association of mood and anxiety disorders with filling opioid prescriptions within the first 3 months postpartum. METHODS: A retrospective cohort study (n = 25 279) was completed using claims data for a sample of privately insured women who gave birth in the state of Iowa. The interactive effects of mood and anxiety disorders and delivery mode on filling at least one and two or more opioid prescriptions were examined in logistic regression models. RESULTS: The presence of mood and anxiety disorders among women who delivered vaginally increased their odds of filling at least one opioid fill by nearly 50% (odds ratio [OR]: 1.48, 95% confidence interval [CI]: 1.35-1.63) and by 20% (OR: 1.20, 95% CI: 1.00-1.43) among women with cesarean delivery. DISCUSSION AND CONCLUSION: Postpartum women with mood and anxiety disorders were more likely to fill opioid prescriptions postpartum compared to women without these conditions. SCIENTIFIC SIGNIFICANCE: This study extends prior research by examining the intersection of risk of mood and anxiety disorders and opioid use postpartum. Findings from this study support the need for future research to identify the drivers of increased opioid use among postpartum women with mood and anxiety disorders. (Am J Addict 2020;29:463-470).

Recognition, prevention, and treatment of delirium in emergency department: An evidence-based narrative review.

BACKGROUND: Delirium is an acute disorder of attention and cognition that is common, serious, costly, under-recognized, and potentially fatal. Delirium is particularly problematic in the emergency department (ED) care of medically complex older adults, who are being seen in greater numbers. OBJECTIVE: This evidence-based narrative review focuses on the key components of delirium screening, prevention, and treatment. DISCUSSION: The recognition of delirium requires a systematic approach rather than a clinical gestalt alone. Several delirium assessment tools with high sensitivity and specificity, such as delirium triage screen and brief Confusion Assessment Method, can be used in the ED. The prevention of delirium requires environmental modification and unique geriatric care strategies tailored to the ED. The key approaches to treatment include the removal of the precipitating etiology, re-orientation, hydration, and early mobilization. Treatment of delirium requires a multifaceted and comprehensive care plan, as there is limited evidence for significant benefit with pharmacological agents. CONCLUSION: Older ED patients are at high risk for current or subsequent development of delirium, and a focused screening, prevention, and intervention for those who are at risk for delirium and its associated complications are the important next steps.

Increased opioid use following rotator cuff repair associated with prior opioid use and surgeon prescription patterns.

BACKGROUND: Prescription opioids are standard of care for postoperative pain management after musculoskeletal surgery, but there is no guideline or consensus on best practices. Variability in the intensity of opioids prescribed for postoperative recovery has been documented, but it is unclear whether this variability is clinically motivated or associated with provider practice patterns, or how this variation is associated with patient outcomes. This study described variation in the intensity of opioids prescribed for patients undergoing rotator cuff repair (RCR) and examined associations with provider prescribing patterns and patients' long-term opioid use outcomes. METHODS: Medicare data from 2010 to 2012 were used to identify 16,043 RCRs for patients with new shoulder complaints in 2011. Two measures of perioperative opioid use were created: (1) any opioid fill occurring 3 days before to 7 days after RCR and (2) total morphine milligram equivalents (MMEs) of all opioid fills during that period. Patient outcomes for persistent opioid use after RCR included (1) any opioid fill from 90 to 180 days after RCR and (2) the lack of any 30-day gap in opioid availability during that period. Generalized linear regression models were used to estimate associations between provider characteristics and opioid use for RCR, and between opioid use and outcomes. All models adjusted for patient clinical and demographic characteristics. Separate analyses were done for patients with and without opioid use in the 180 days before RCR. RESULTS: In this sample, 54% of patients undergoing RCR were opioid naive at the time of RCR. Relative to prior users, a greater proportion of opioid naive users had any opioid fill (85.7% vs. 75.4%), but prior users received more MMEs than naive users (565 vs. 451 MMEs). Providers' opioid prescribing for other patients was associated with the intensity of perioperative opioids received for RCR. Total MMEs received for RCR were associated with higher odds of persistent opioid use 90-180 days after RCR. CONCLUSIONS: The intensity of opioids received by patients for postoperative pain appears to be partially determined by the prescribing habits of their providers. Greater intensity of opioids received is, in turn, associated with greater odds of patterns of chronic opioid use after surgery. More comprehensive, patient-centered guidance on opioid prescribing is needed to help surgeons provide optimal postoperative pain management plans, balancing needs for short-term symptom relief and risks for long-term outcomes.

Predictors of Reduced Opioid Use With Spinal Cord Stimulation in Patients With Chronic Opioid Use.

OBJECTIVES: Spinal cord stimulation (SCS) has gained traction as an alternative to chronic opioid therapy in light of the opioid crisis. Prior reports vary widely in their estimates of its effect on opioid consumption. We therefore aimed to address the following questions: 1) Does chronic opioid use change after SCS? 2) Which patient characteristics predict reduced opioid consumption after SCS? MATERIALS AND METHODS: Claims from a private health insurance company were used to identify patients with SCS implantation from 2003 to 2014. We required 12 months of continuous data before and after surgery (i.e., a minimum total observation period of two years), and at least two opioid prescription fills in the six months before surgery. Daily morphine equivalent dose (MED) was calculated from prescription medication claims. Diagnosis codes identified common comorbidities. RESULTS: Hundred forty-five patients met inclusion criteria. MED of 65 was the most statistically meaningful preoperative dose threshold. Approximately half of patients decreased opioid use >20% after SCS implantation. Logistic regression analysis revealed age (p = 0.0362), gender (p = 0.0076), and preoperative daily MED < 65 (p = 0.0322) as predictors of meaningful reduction, which was defined as a 20% reduction in MED. CONCLUSIONS: With only half of chronic opioid users demonstrating meaningful opioid reduction after SCS implantation, we demonstrate that current SCS technology does not reliably help a larger number of patients reduce opioid usage. Women, older age, and preoperative MED < 65 are predictive of meaningful opioid reduction but only one of these is modifiable. As not all patients saw benefit from their therapies, there is still much room for improvement in the treatment of refractory chronic pain that is associated with failed back surgery syndrome and chronic regional pain syndrome.

Impact of preexisting mental illness on breast cancer endocrine therapy adherence.

PURPOSE: Patients with estrogen receptor positive (ER+) breast cancer are often non-adherent to endocrine therapies, despite clear survival benefits. We utilized a nationally representative cancer cohort to examine the role of specific mental illnesses on endocrine therapy adherence. METHODS: Using the SEER-Medicare database, we included 21,894 women aged 68+ at their first surgically treated stage I-IV ER+ breast cancer during 2007-2013. All had continuous fee-for-service Medicare Parts A and B for 36+ months before, 18+ months after diagnosis, and continuous Part D for 4+ months before, 18+ after diagnosis. Mental illness was defined as occurring in the 36 months prior to cancer onset. We analyzed endocrine therapy adherence, initiation, and discontinuation using longitudinal linear and Cox regression models. RESULTS: Unipolar depression (11.0%), anxiety (9.5%), non-schizophrenia psychosis (4.6%), and dementias (4.6%) were the most prevalent diagnoses. Endocrine therapies were initiated by 80.0% of women. Among those with at least one year of use, 28.0% were non-adherent (< 0.80 adherence, mean = 0.84) and 25.7% discontinued. Patients with dementia or bipolar depression/psychotic/schizophrenia disorders had lower adjusted initiation probabilities by year one of follow-up, versus those without these diagnoses [0.74 95% CI (0.73-0.74) and 0.73 (0.72-0.73), respectively, reference 0.76 (0.76-0.77)]. Patients with substance use or anxiety disorders less frequently continued endocrine therapy for at least one year, after adjustment, [0.85 95% CI (0.85-0.86) and 0.88 (0.87-0.88), respectively, reference 0.90 (0.89-0.90)]. Patients with substance use disorders had 2.3% lower adherence rates (p < 0.001). CONCLUSIONS: Nearly one-quarter of female Medicare beneficiaries have diagnosed mental illness preceding invasive breast cancer. Those with certain mental illnesses have modestly reduced rates of initiation, adherence, and discontinuation and this may help define patients at higher risk of treatment abandonment. Overall, endocrine therapy adherence remains suboptimal, unnecessarily worsening recurrence and mortality risk.

Impact of Iowa's Prescription Monitoring Program on Opioid Pain Reliever Prescribing Patterns: An Interrupted Time Series Study 2003-2014.

OBJECTIVE: To evaluate the impact of Iowa's prescription monitoring program (PMP), implemented in 2009, on opioid pain reliever (OPR) prescribing patterns. METHODS: We conducted interrupted time series analyses using 2003-2014 health insurance claims from a private health insurer in Iowa. OPR prescriptions for all beneficiaries were included. Another data set included only OPR prescription for new opioid users required to have six months of insurance coverage. We evaluate four OPR prescribing patterns: 1) average daily dosage in morphine milligrams equivalents (MME), 2) MME per prescription, 3) average days' supply per prescription, and 4) prescription rate per 1,000 insured person-years. We examined confounding and effect measure modification of the relationship between PMP and prescribing patterns by age and sex. RESULTS: During the 12 years of follow-up, 1,512,388 insured Iowans contributed 6,169,634.92 person-years of follow-up. Of these, 505,274 patients filled 2,401,818 OPR prescriptions and 360,688 new OPR users filled as many first OPR prescriptions. The increasing trend of OPR prescription rates from 2003 to 2009 declined post-PMP. Similarly, there was a large decline in MME per day and MME per prescription. The OPR days' supply kept increasing post-PMP implementation, albeit at a slightly slower rate than pre-PMP implementation. There was no confounding by age and sex; however, we observed heterogeneity by age and sex; patients aged ≥50 years and females received higher doses and more prescriptions pre-PMP and experienced the greatest declines post-PMP. CONCLUSIONS: Our study suggests that Iowa PMP implementation may have resulted in declines in OPR prescribing, and this impact varies by patient age and sex.

Intravenous immune globulin and thromboembolic adverse events in patients with hematologic malignancy.

In patients with hypogammaglobulinemia secondary to chronic lymphocytic leukemia (CLL) or multiple myeloma (MM), intravenous immune globulin (IVIg) may be administered to reduce the risk of infection. Since 2013, IVIg products have carried a boxed safety warning about the risk of thromboembolic events (TEEs), with TEEs reported in 0.5% to 15% of patients treated with IVIg. In this retrospective cohort study of older patients with CLL or MM identified from the Surveillance, Epidemiology, and End Results-Medicare Linked Database, we assessed rates of clinically serious TEEs in 2724 new users of IVIg and a propensity-matched comparison group of 8035 nonusers. For the primary end point, arterial TEE, we observed a transient increased risk of TEE during the day of an IVIg infusion and the day afterward (hazard ration = 3.40; 95% confidence interval [CI]: 1.25, 9.25); this risk declined over the remainder of the 30-day treatment cycle. When considered in terms of absolute risk averaged over a 1-year treatment period, the increase in risk attributable to IVIg was estimated to be 0.7% (95% CI: -0.2%, 2.0%) compared with a baseline risk of 1.8% for the arterial TEE end point. A statistically nonsignificant risk increase of 0.3% (95% CI: -0.4%, 1.5%) compared with a baseline risk of 1.1% was observed for the venous TEE end point. Further research is needed to establish the generalizability of these results to patients receiving higher doses of IVIg for other indications.

Bipolar Disorder in Nursing Homes: Impact on Antipsychotic Use, Diagnosis Patterns, and New Diagnoses in People with Dementia.

OBJECTIVES: Nursing home quality measures include the proportion of residents who receive antipsychotics. Residents with bipolar disorder are included even though antipsychotics are FDA-approved for this indication. We evaluated how including residents with bipolar disorder impacted the antipsychotic use quality measure for long-stay residents. We evaluated the agreement of minimum data set (MDS) bipolar disorder diagnoses with Medicare data, whether dementia was diagnosed before bipolar disorder, and how less-specific bipolar disorder diagnoses impacted findings. DESIGN: Cross-sectional study. SETTING: Nursing homes in Iowa. PARTICIPANTS: 21,955 long-stay nursing home residents in the first quarter of 2014. MEASUREMENTS: We identified antipsychotic use and bipolar disorder using MDS data. We compared MDS bipolar disorder diagnoses with Chronic Conditions Warehouse (CCW) "ever" bipolar disorder indicators, and prior year claims. We compared CCW condition onset dates to identify bipolar disorder diagnosed after dementia. RESULTS: The mean (SD) proportion receiving antipsychotics was 19.6% (11.1%) with bipolar disorder and 18.3% (10.8%) without. The positive predictive value (PPV) of MDS bipolar disorder diagnoses was 80.2% versus CCW lifetime indicators, and 74.6% versus claims. PPV decreased by 27.1% when "bipolar disorder, unspecified" and "other bipolar disorders" diagnoses were excluded. Nearly three-quarters of residents with bipolar disorder had dementia. Over half of those with dementia had dementia first per CCW records. This proportion was lower among those with more specific bipolar disorder diagnoses or MDS bipolar disorder indicators. CONCLUSIONS: Bipolar disorder in nursing home residents is often first diagnosed after dementia using nonspecific diagnoses. This practice deserves further evaluation.

Evaluation of the US Food and Drug Administration sentinel analysis tools in confirming previously observed drug-outcome associations: The case of clindamycin and Clostridium difficile infection.

PURPOSE: The Food and Drug Administration's Sentinel System developed parameterized, reusable analytic programs for evaluation of medical product safety. Research on outpatient antibiotic exposures, and Clostridium difficile infection (CDI) with non-user reference groups led us to expect a higher rate of CDI among outpatient clindamycin users vs penicillin users. We evaluated the ability of the Cohort Identification and Descriptive Analysis and Propensity Score Matching tools to identify a higher rate of CDI among clindamycin users. METHODS: We matched new users of outpatient dispensings of oral clindamycin or penicillin from 13 Data Partners 1:1 on propensity score and followed them for up to 60 days for development of CDI. We used Cox proportional hazards regression stratified by Data Partner and matched pair to compare CDI incidence. RESULTS: Propensity score models at 3 Data Partners had convergence warnings and a limited range of predicted values. We excluded these Data Partners despite adequate covariate balance after matching. From the 10 Data Partners where these models converged without warnings, we identified 807 919 new clindamycin users and 8 815 441 new penicillin users eligible for the analysis. The stratified analysis of 807 769 matched pairs included 840 events among clindamycin users and 290 among penicillin users (hazard ratio 2.90, 95% confidence interval 2.53, 3.31). CONCLUSIONS: This evaluation produced an expected result and identified several potential enhancements to the Propensity Score Matching tool. This study has important limitations. CDI risk may have been related to factors other than the inherent properties of the drugs, such as duration of use or subsequent exposures.

Chart validation of inpatient ICD-9-CM administrative diagnosis codes for acute myocardial infarction (AMI) among intravenous immune globulin (IGIV) users in the Sentinel Distributed Database.

BACKGROUND: The Sentinel Distributed Database (SDD) is a large database of patient-level administrative health care records, primarily derived from insurance claims and electronic health records, and is sponsored by the US Food and Drug Administration for medical product safety evaluations. Acute myocardial infarction (AMI) is a common study endpoint for drug safety studies that rely on health records from the SDD and other administrative databases. PURPOSE: In this chart validation study, we report on the positive predictive value (PPV) of inpatient International Classification of Diseases, Ninth Revision, Clinical Modification AMI administrative diagnosis codes (410.x1 and 410.x0) in the SDD. METHODS: As part of an assessment of thromboembolic adverse event risk following treatment with intravenous immune globulin, charts were obtained for 103 potential post-intravenous immune globulin AMI cases. Charts were abstracted by trained nurses and physician-adjudicated based on prespecified diagnostic criteria. RESULTS: Acute myocardial infarction status could be determined for 89 potential cases. The PPVs for the inpatient AMI diagnoses recorded in the SDD were 75% overall (95% CI, 65-84%), 93% (95% CI, 78-99%) for principal-position diagnoses, 88% (95% CI, 72-97%) for secondary diagnoses, and 38% (95% CI, 20-59%) for position-unspecified diagnoses (eg, diagnoses originating from separate physician claims associated with an inpatient stay). Of the confirmed AMI cases, demand ischemia was the suspected etiology more often for those coded in secondary or unspecified positions (72% and 40%, respectively) than for principal-position AMI diagnoses (21%). CONCLUSIONS: The PPVs for principal and secondary AMI diagnoses were high and similar to estimates from prior chart validation studies. Position-unspecified diagnosis codes were less likely to represent true AMI cases.

Prospective surveillance pilot of rivaroxaban safety within the US Food and Drug Administration Sentinel System.

PURPOSE: The US Food and Drug Administration's Sentinel system developed tools for sequential surveillance. METHODS: In patients with non-valvular atrial fibrillation, we sequentially compared outcomes for new users of rivaroxaban versus warfarin, employing propensity score matching and Cox regression. A total of 36 173 rivaroxaban and 79 520 warfarin initiators were variable-ratio matched within 2 monitoring periods. RESULTS: Statistically significant signals were observed for ischemic stroke (IS) (first period) and intracranial hemorrhage (ICH) (second period) favoring rivaroxaban, and gastrointestinal bleeding (GIB) (second period) favoring warfarin. In follow-up analyses using primary position diagnoses from inpatient encounters for increased definition specificity, the hazard ratios (HR) for rivaroxaban vs warfarin new users were 0.61 (0.47, 0.79) for IS, 1.47 (1.29, 1.67) for GIB, and 0.71 (0.50, 1.01) for ICH. For GIB, the HR varied by age: <66 HR = 0.88 (0.60, 1.30) and 66+ HR = 1.49 (1.30, 1.71). CONCLUSIONS: This study demonstrates the capability of Sentinel to conduct prospective safety monitoring and raises no new concerns about rivaroxaban safety.

Sentinel Modular Program for Propensity Score-Matched Cohort Analyses: Application to Glyburide, Glipizide, and Serious Hypoglycemia.

Sentinel is a program sponsored by the US Food and Drug Administration to monitor the safety of medical products. We conducted a cohort assessment to evaluate the ability of the Sentinel Propensity Score Matching Tool to reproduce in an expedited fashion the known association between glyburide (vs. glipizide) and serious hypoglycemia. Thirteen data partners who contribute to the Sentinel Distributed Database participated in this analysis. A pretested and customizable analytic program was run at each individual site. De-identified summary results from each data partner were returned and aggregated at the Sentinel Operations Center. We identified a total of 198,550 and 379,507 new users of glyburide and glipizide, respectively. The incidence of emergency department visits and hospital admissions for serious hypoglycemia was 19 per 1000 person-years (95% confidence interval = 17.9, 19.7) for glyburide users and 22 (21.6, 22.7) for glipizide users. In cohorts matched by propensity score based on predefined variables, the hazard ratio (HR) for glyburide was 1.36 (1.24, 1.49) versus glipizide. In cohorts matched on a high-dimensional propensity score based on empirically selected variables, for which the program ran to completion in five data partners, the HR was 1.49 (1.31, 1.70). In cohorts matched on propensity scores based on both predefined and empirically selected variables via the high-dimensional propensity score algorithm (the same five data partners), the HR was 1.51 (1.32, 1.71). These findings are consistent with the literature, and demonstrate the ability of the Sentinel Propensity Score Matching Tool to reproduce this known association in an expedited fashion.See video abstract at, http://links.lww.com/EDE/B275.