Estimating the impact of physician risky-prescribing on the network structure underlying physician shared-patient relationships

Social network analysis and shared-patient physician networks have become effective ways of studying physician collaborations. Assortative mixing or "homophily" is the network phenomenon whereby the propensity for similar individuals to form ties is greater than for dissimilar individuals. Motivated by the public health concern of risky-prescribing among older patients in the United States, we develop network models and tests involving novel network measures to study whether there is evidence of geographic homophily in prescribing and deprescribing in the specific shared-patient network of physicians linked to the US state of Ohio in 2014. Evidence of homophily in risky-prescribing would imply that prescribing behaviors help shape physician networks and could inform interventions to reduce risky-prescribing (e.g., should interventions target groups of physicians or select physicians at random). Furthermore, if such effects varied depending on the structural features of a physician's position in the network (e.g., by whether or not they are involved in cliques - groups of actors that are fully connected to each other - such as closed triangles in the case of three actors), this would further strengthen the case for targeting of select physicians for interventions. Using accompanying Medicare Part D data, we converted patient longitudinal prescription receipts into novel measures of the intensity of each physician's risky-prescribing. Exponential random graph models were used to simultaneously estimate the importance of homophily in prescribing and deprescribing in the network beyond the characteristics of physician specialty (or other metadata) and network-derived features. In addition, novel network measures were introduced to allow homophily to be characterized in relation to specific triadic (three-actor) structural configurations in the network with associated non-parametric randomization tests to evaluate their statistical significance in the network against the null hypothesis of no such phenomena. We found physician homophily in prescribing and deprescribing in both the state-wide and multiple HRR sub-networks, and that the level of homophily varied across HRRs. We also found that physicians exhibited within-triad homophily in risky-prescribing, with the prevalence of homophilic triads significantly higher than expected by chance absent homophily. These results may explain why communities of prescribers emerge and evolve, helping to justify group-level prescriber interventions. The methodology could be applied to arbitrary shared-patient networks and even more generally to other kinds of network data that underlies other kinds of social phenomena.

Racial differences in low value care among older adult Medicare patients in US health systems: retrospective cohort study

OBJECTIVE

To characterize racial differences in receipt of low value care (services that provide little to no benefit yet have potential for harm) among older Medicare beneficiaries overall and within health systems in the United States.

DESIGN

Retrospective cohort study SETTING: 100% Medicare fee-for-service administrative data (2016-18).

PARTICIPANTS

Black and White Medicare patients aged 65 or older as of 2016 and attributed to 595 health systems in the United States.

MAIN OUTCOME MEASURES

Receipt of 40 low value services among Black and White patients, with and without adjustment for patient age, sex, and previous healthcare use. Additional models included health system fixed effects to assess racial differences within health systems and separately, racial composition of the health system's population to assess the relative contributions of individual patient race and health system racial composition to low value care receipt.

RESULTS

The cohort included 9 833 304 patients (6.8% Black; 57.9% female). Of 40 low value services examined, Black patients had higher adjusted receipt of nine services and lower receipt of 20 services than White patients. Specifically, Black patients were more likely to receive low value acute diagnostic tests, including imaging for uncomplicated headache (6.9% v 3.2%) and head computed tomography scans for dizziness (3.1% v 1.9%). White patients had higher rates of low value screening tests and treatments, including preoperative laboratory tests (10.3% v 6.5%), prostate specific antigen tests (31.0% v 25.7%), and antibiotics for upper respiratory infections (36.6% v 32.7%; all P<0.001). Secondary analyses showed that these differences persisted within given health systems and were not explained by Black and White patients receiving care from different systems.

CONCLUSIONS

Black patients were more likely to receive low value acute diagnostic tests and White patients were more likely to receive low value screening tests and treatments. Differences were generally small and were largely due to differential care within health systems. These patterns suggest potential individual, interpersonal, and structural factors that researchers, policy makers, and health system leaders might investigate and address to improve care quality and equity.

Sources of Low-Value Care Received by Medicare Beneficiaries and Associated Spending Within US Health Systems

This cross-sectional study examines referrals for low-value health care services and associated spending by ordering clinician among Medicare beneficiaries.

Who's Accountable? Low-Value Care Received By Medicare Beneficiaries Outside Of Their Attributed Health Systems

Policy makers and payers increasingly hold health systems accountable for spending and quality for their attributed beneficiaries. Low-value care-medical services that offer little or no benefit and have the potential for harm in specific clinical scenarios-received outside of these systems could threaten success on both fronts. Using national Medicare data for fee-for-service beneficiaries ages sixty-five and older and attributed to 595 US health systems, we describe where and from whom they received forty low-value services during 2017-18 and identify factors associated with out-of-system receipt. Forty-three percent of low-value services received by attributed beneficiaries originated from out-of-system clinicians: 38 percent from specialists, 4 percent from primary care physicians, and 1 percent from advanced practice clinicians. Recipients of low-value care were more likely to obtain that care out of system if age 75 or older (versus ages 65-74), male (versus female), non-Hispanic White (versus other races or ethnicities), rural dwelling (versus metropolitan dwelling), more medically complex, or experiencing lower continuity of care. However, out-of-system service receipt was not associated with recipients' health systems' accountable care organization status. Health systems might improve quality and reduce spending for their attributed beneficiaries by addressing out-of-system receipt of low-value care-for example, by improving continuity.

Provision of Digital Health Technologies for Opioid Use Disorder Treatment by US Health Care Organizations

Importance

Digital health technologies may expand organizational capacity to treat opioid use disorder (OUD). However, it remains unclear whether these technologies serve as substitutes for or complements to traditional substance use disorder (SUD) treatment resources in health care organizations.

Objective

To characterize the use of patient-facing digital health technologies for OUD by US organizations with accountable care organization (ACO) contracts.

Design, Setting, and Participants

This cross-sectional study analyzed responses to the 2022 National Survey of Accountable Care Organizations (NSACO), collected between October 1, 2021, and June 30, 2022, from US organizations with Medicare and Medicaid ACO contracts. Data analysis was performed between December 15, 2022, and January 6, 2023.

Exposures

Treatment resources for SUD (eg, an addiction medicine specialist, sufficient staff to treat SUD, medications for OUD, a specialty SUD treatment facility, a registry to identify patients with OUD, or a registry to track mental health for patients with OUD) and organizational characteristics (eg, organization type, Medicaid ACO contract).

Main Outcomes and Measures

The main outcomes included survey-reported use of 3 categories of digital health technologies for OUD: remote mental health therapy and tracking, virtual peer recovery support programs, and digital recovery support for adjuvant cognitive behavior therapy (CBT). Statistical analysis was conducted using descriptive statistics and multivariable logistic regression models.

Results

Overall, 276 of 505 organizations responded to the NSACO (54.7% response rate), with a total of 304 respondents. Of these, 161 (53.1%) were from a hospital or health system, 74 (24.2%) were from a physician- or medical group-led organization, and 23 (7.8%) were from a safety-net organization. One-third of respondents (101 [33.5%]) reported that their organization used at least 1 of the 3 digital health technology categories, including remote mental health therapy and tracking (80 [26.5%]), virtual peer recovery support programs (46 [15.1%]), and digital recovery support for adjuvant CBT (27 [9.0%]). In an adjusted analysis, organizations with an addiction medicine specialist (average marginal effect [SE], 32.3 [4.7] percentage points; P < .001) or a registry to track mental health (average marginal effect [SE], 27.2 [3.8] percentage points; P < .001) were more likely to use at least 1 category of technology compared with otherwise similar organizations lacking these capabilities.

Conclusions and Relevance

In this cross-sectional study of 276 organizations with ACO contracts, organizations used patient-facing digital health technologies for OUD as complements to available SUD treatment capabilities rather than as substitutes for unavailable resources. Future studies should examine implementation facilitators to realize the potential of emerging technologies to support organizations facing health care practitioner shortages and other barriers to OUD treatment delivery.

Racial Inequality in Receipt of Medications for Opioid Use Disorder

BACKGROUND

Since 2010, Black persons in the United States have had a greater increase in opioid overdose-related mortality than other groups, but national-level evidence characterizing racial and ethnic disparities in the use of medications for opioid use disorder (OUD) is limited.

METHODS

We used Medicare claims data from the 2016-2019 period for a random 40% sample of fee-for-service beneficiaries who were Black, Hispanic, or White; were eligible for Medicare owing to disability; and had an index event related to OUD (nonfatal overdose treated in an emergency department or inpatient setting, hospitalization with injection drug use-related infection, or inpatient or residential rehabilitation or detoxification care). We measured the receipt of medications to treat OUD (buprenorphine, naltrexone, and naloxone), the receipt of high-risk medications (opioid analgesics and benzodiazepines), and health care utilization, all in the 180 days after the index event. We estimated differences in outcomes according to race and ethnic group with adjustment for beneficiary age, sex, index event, count of chronic coexisting conditions, and state of residence.

RESULTS

We identified 25,904 OUD-related index events among 23,370 beneficiaries, with 3937 events (15.2%) occurring among Black patients, 2105 (8.1%) among Hispanic patients, and 19,862 (76.7%) among White patients. In the 180 days after the index event, patients received buprenorphine after 12.7% of events among Black patients, after 18.7% of those among Hispanic patients, and after 23.3% of those among White patients; patients received naloxone after 14.4%, 20.7%, and 22.9%, respectively; and patients received benzodiazepines after 23.4%, 29.6%, and 37.1%, respectively. Racial differences in the receipt of medications to treat OUD did not change appreciably from 2016 to 2019 (buprenorphine receipt: after 9.1% of index events among Black patients vs. 21.6% of those among White patients in 2016, and after 14.1% vs. 25.5% in 2019). In all study groups, patients had multiple ambulatory visits in the 180 days after the index event (mean number of visits, 6.6 after events among Black patients, 6.7 after events among Hispanic patients, and 7.6 after events among White patients).

CONCLUSIONS

Racial and ethnic differences in the receipt of medications to treat OUD after an index event related to this disorder among patients with disability were substantial and did not change over time. The high incidence of ambulatory visits in all groups showed that disparities persisted despite frequent health care contact. (Funded by the National Institute on Drug Abuse and the National Institute on Aging.).

Receipt of Medications for Chronic Disease During the First 2 Years of the COVID-19 Pandemic Among Enrollees in Fee-for-Service Medicare

Importance

During the first 2 years of the COVID-19 pandemic, inpatient and ambulatory care declined dramatically. Little is known about prescription drug receipt during this period, particularly for populations with chronic illness and with high risk of adverse COVID-19 outcomes and decreased access to care.

Objective

To investigate whether receipt of medications was maintained during the first 2 years of the COVID-19 pandemic among older people with chronic diseases, particularly Asian, Black, and Hispanic populations and people with dementia, who faced pandemic-related care disruptions.

Design, Setting, and Participants

This cohort study used a 100% sample of US Medicare fee-for-service administrative data from 2019 to 2021 for community-dwelling beneficiaries aged 65 years or older. Population-based prescription fill rates were compared for 2020 and 2021 vs 2019. Data were analyzed from July 2022 to March 2023.

Exposure

The COVID-19 pandemic.

Main Outcomes and Measures

Age- and sex-adjusted monthly prescription fill rates were calculated for 5 groups of medications commonly prescribed for chronic disease : angiotensin-converting enzyme inhibitors and angiotensin receptor blockers, 3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) reductase inhibitors (statins), oral diabetes medications, asthma and chronic obstructive pulmonary disease medications, and antidepressants. Measurements were stratified by race and ethnicity group and dementia diagnosis. Secondary analyses measured changes in the proportion of prescriptions dispensed as a 90-day or greater supply.

Results

Overall, the mean monthly cohort included 18 113 000 beneficiaries (mean [SD] age, 74.5 [7.4] years; 10 520 000 females [58.1%]; 587 000 Asian [3.2%], 1 069 000 Black [5.9%], 905 000 Hispanic [5.0%], and 14 929 000 White [82.4%]); 1 970 000 individuals (10.9%) were diagnosed with dementia. Across 5 drug classifications, mean fill rates increased by 2.07% (95% CI, 2.01% to 2.12%) in 2020 and decreased by 2.61% (95% CI, -2.67% to -2.56%) in 2021 compared with 2019. Fill rates decreased by less than the mean overall decrease for Black enrollees (-1.42%; 95% CI, -1.64% to -1.20%) and Asian enrollees (-1.05%; 95% CI, -1.36% to -0.77%) and people diagnosed with dementia (-0.38%; 95% CI, -0.54% to -0.23%). The proportion of fills dispensed as 90-day or greater supplies increased during the pandemic for all groups, with an increase per 100 fills of 3.98 fills (95% CI, 3.94 to 4.03 fills) overall.

Conclusions and Relevance

This study found that, in contrast to in-person health services, receipt of medications for chronic conditions was relatively stable in the first 2 years of the COVID-19 pandemic overall, across racial and ethnic groups, and for community-dwelling patients with dementia. This finding of stability may hold lessons for other outpatient services during the next pandemic.

Low-Value Care at the Actionable Level of Individual Health Systems

IMPORTANCE

Low-value health care remains prevalent in the US despite decades of work to measure and reduce such care. Efforts have been only modestly effective in part because the measurement of low-value care has largely been restricted to the national or regional level, limiting actionability.

OBJECTIVES

To measure and report low-value care use across and within individual health systems and identify system characteristics associated with higher use using Medicare administrative data.

DESIGN, SETTING, AND PARTICIPANTS

This retrospective cohort study of health system-attributed Medicare beneficiaries was conducted among 556 health systems in the Agency for Healthcare Research and Quality Compendium of US Health Systems and included system-attributed beneficiaries who were older than 65 years, continuously enrolled in Medicare Parts A and B for at least 12 months in 2016 or 2017, and eligible for specific low-value services. Statistical analysis was conducted from January 26 to July 15, 2021.

MAIN OUTCOMES AND MEASURES

Use of 41 individual low-value services and a composite measure of the 28 most common services among system-attributed beneficiaries, standardized to distance from the mean value. Measures were based on the Milliman MedInsight Health Waste Calculator and published claims-based definitions.

RESULTS

Across 556 health systems serving a total of 11 637 763 beneficiaries, the mean (SD) use of each of the 41 low-value services ranged from 0% (0.01%) to 28% (4%) of eligible beneficiaries. The most common low-value services were preoperative laboratory testing (mean [SD] rate, 28% [4%] of eligible beneficiaries), prostate-specific antigen testing in men older than 70 years (mean [SD] rate, 27% [8%]), and use of antipsychotic medications in patients with dementia (mean [SD] rate, 24% [8%]). In multivariable analysis, the health system characteristics associated with higher use of low-value care were smaller proportion of primary care physicians (adjusted composite score, 0.15 [95% CI, 0.04-0.26] for systems with less than the median percentage of primary care physicians vs -0.16 [95% CI, -0.27 to -0.05] for those with more than the median percentage of primary care physicians; P < .001), no major teaching hospital (adjusted composite, 0.10 [95% CI, -0.01 to 0.20] without a teaching hospital vs -0.18 [95% CI, -0.34 to -0.02] with a teaching hospital; P = .01), larger proportion of non-White patients (adjusted composite, 0.15 [95% CI, -0.02 to 0.32] for systems with >20% of non-White beneficiaries vs -0.06 [95% CI, -0.16 to 0.03] for systems with ≤20% of non-White beneficiaries; P = .04), headquartered in the South or West (adjusted composite, 0.28 [95% CI, 0.14-0.43] for the South and 0.22 [95% CI, 0.02-0.42] for the West compared with -0.09 [95% CI, -0.26 to 0.08] for the Northeast and -0.44 [95% CI, -0.60 to -0.28] for the Midwest; P < .001), and serving areas with more health care spending (adjusted composite, 0.23 [95% CI, 0.11-0.35] for areas above the median level of spending vs -0.24 [95% CI, -0.36 to -0.12] for areas below the median level of spending; P < .001).

CONCLUSIONS AND RELEVANCE

The findings of this large cohort study suggest that system-level measurement and reporting of specific low-value services is feasible, enables cross-system comparisons, and reveals a broad range of low-value care use.

Trends in Testosterone Prescriptions for Older Men Enrolled in Commercial Insurance and Medicare Advantage

This cohort study of health claims data examines trends in testosterone prescriptions for older men in the US following 2014 changes to the US Food and Drug Administration’s stance on testosterone’s effects and safety risks.

Racial Inequality in Prescription Opioid Receipt - Role of Individual Health Systems

BACKGROUND

Historically, the receipt of prescription opioids has differed among racial groups in the United States. Research has not sufficiently explored the contribution of individual health systems to these differences by examining within-system prescription opioid receipt according to race.

METHODS

We used 2016 and 2017 Medicare claims data from a random 40% national sample of fee-for-service, Black and White beneficiaries 18 to 64 years of age who were attributed to health systems. We identified 310 racially diverse systems (defined as systems with ≥200 person-years each for Black and White patients). To test representativeness, we compared patient characteristics and opioid receipt among the patients in these 310 systems with those in the national sample. Within the 310 systems, regression models were used to explore the difference between Black and White patients in the following annual opioid measures: any prescription filled, short-term receipt of opioids, long-term receipt of opioids (one or more filled opioid prescriptions in all four calendar quarters of a year), and the opioid dose in morphine milligram equivalents (MME); models controlled for patient characteristics, state, and system.

RESULTS

The national sample included 2,197,153 person-years, and the sample served by 310 racially diverse systems included 896,807 person-years (representing 47.4% of all patients and 56.1% of Black patients in the national sample). The national sample and 310-systems sample differed meaningfully only in the percent of person-years contributed by Black patients (21.3% vs. 25.9%). In the 310-systems sample, the crude annual prevalence of any opioid receipt differed slightly between Black and White patients (50.2% vs. 52.2%), whereas the mean annual dose was 36% lower among Black patients than among White patients (5190 MME vs. 8082 MME). Within systems, the adjusted race differences in measures paralleled the population trends: the annual prevalence of opioid receipt differed little, but the mean annual dose was higher among White patients than among Black patients in 91% of the systems, and at least 15% higher in 75% of the systems.

CONCLUSIONS

Within individual health systems, Black and White patients received markedly different opioid doses. These system-specific findings could facilitate exploration of the causes and consequences of these differences. (Funded by the National Institute on Aging and the Agency for Healthcare Research and Quality.).

A proposed taxonomy for population-level prescription use patterns

Recent and increasing discussion of prescription price transparency highlights the importance of defining, measuring and communicating prescription drug value. To help advance these goals, the authors propose a taxonomy of population-level prescription drug use patterns. The taxonomy assigns prescription use to one of five categories according to likely population-level health impact. The categories include effective, potentially discretionary, potentially harmful, wasteful, and lifestyle. The authors hope the proposed taxonomy will inform discussion of prescription drug value by providing estimates of population impact, especially the balance of anticipated benefit and harm.

Association of Receiving Multiple, Concurrent Fracture-Associated Drugs With Hip Fracture Risk

IMPORTANCE

Many prescription drugs increase fracture risk, which raises concern for patients receiving 2 or more such drugs concurrently. Logic suggests that risk will increase with each additional drug, but the risk of taking multiple fracture-associated drugs (FADs) is unknown.

OBJECTIVE

To estimate hip fracture risk associated with concurrent exposure to multiple FADs.

DESIGN, SETTING, AND PARTICIPANTS

This cohort study used a 20% random sample of Medicare fee-for-service administrative data for age-eligible Medicare beneficiaries from 2004 to 2014. Sex-stratified Cox regression models estimated hip fracture risk associated with current receipt of 1, 2, or 3 or more of 21 FADs and, separately, risk associated with each FAD and 2-way FAD combination vs no FADs. Models included sociodemographic characteristics, comorbidities, and use of non-FAD medications. Analyses began in November 2018 and were completed April 2019.

EXPOSURE

Receipt of prescription FADs.

MAIN OUTCOMES AND MEASURES

Hip fracture hospitalization.

RESULTS

A total of 11.3 million person-years were observed, reflecting 2 646 255 individuals (mean [SD] age, 77.2 [7.3] years, 1 615 613 [61.1%] women, 2 136 585 [80.7%] white, and 219 579 [8.3%] black). Overall, 2 827 284 person-years (25.1%) involved receipt of 1 FAD; 1 322 296 (11.7%), 2 FADs; and 954 506 (8.5%), 3 or more FADs. In fully adjusted, sex-stratified models, an increase in hip fracture risk among women was associated with the receipt of 1, 2, or 3 or more FADs (1 FAD: hazard ratio [HR], 2.04; 95% CI, 1.99-2.11; P < .001; 2 FADs: HR, 2.86; 95% CI, 2.77-2.95; P < .001; ≥3 FADs: HR, 4.50; 95% CI, 4.36-4.65; P < .001). Relative risks for men were slightly higher (1 FAD: HR, 2.23; 95% CI, 2.11-2.36; P < .001; 2 FADs: HR, 3.40; 95% CI, 3.20-3.61; P < .001; ≥3 FADs: HR, 5.18; 95% CI, 4.87-5.52; P < .001). Among women, 2 individual FADs were associated with HRs greater than 3.00; 80 pairs of FADs exceeded this threshold. Common, risky pairs among women included sedative hypnotics plus opioids (HR, 4.90; 95% CI, 3.98-6.02; P < .001), serotonin reuptake inhibitors plus benzodiazepines (HR, 4.50; 95% CI, 3.76-5.38; P < .001), and proton pump inhibitors plus opioids (HR, 4.00; 95% CI, 3.56-4.49; P < .001). Receipt of 1, 2, or 3 or more non-FADs was associated with a small, significant reduction in fracture risk compared with receipt of no non-FADs among women (1 non-FAD: HR, 0.93; 95% CI, 0.90-0.96; P < .001; 2 non-FADs: HR, 0.84; 95% CI, 0.81-0.87; P < .001; ≥3 non-FADs: HR, 0.74; 95% CI, 0.72-0.77; P < .001).

CONCLUSIONS AND RELEVANCE

Among older adults, FADs are commonly used and commonly combined. In this cohort study, the addition of a second and third FAD was associated with a steep increase in fracture risk. Many risky pairs of FADs included potentially avoidable drugs (eg, sedatives and opioids). If confirmed, these findings suggest that fracture risk could be reduced through tighter adherence to long-established prescribing guidelines and recommendations.

Validating Publicly Available Crosswalks for Translating ICD-9 to ICD-10 Diagnosis Codes for Cardiovascular Outcomes Research

Background On October 1, 2015, the Center for Medicare and Medicaid Services transitioned from the International Classification of Diseases, Ninth Revision ( ICD-9) to the ICD, Tenth Revision ( ICD-10) compendium of codes for diagnosis and billing in health care, but translation between the two is often inexact. Here we describe a validated crosswalk to translate ICD-9 codes into ICD-10 codes, with a focus on complications after carotid revascularization and endovascular aortic aneurysm repair. Methods and Results We devised an 8-step process to derive and validate ICD-10 codes from existing ICD-9 codes. We used publicly available sources, including the General Equivalence Mapping database, to translate ICD-9 codes used in prior work to ICD-10 codes. We defined ICD-10 codes as validated if they were concordant with the initial ICD-9 codes after manual comparison by two physicians. Our primary validation measure was the percent of valid ICD-10 codes out of the total ICD-10 codes obtained during translation. We began with 126 ICD-9 diagnosis codes used for complication identification after carotid revascularization procedures, and 97 ICD-9 codes for complications after endovascular aortic aneurysm procedures. Translation generated 143 ICD-10 codes for carotid revascularization, a 14% increase from the initial 126 codes. Manual comparison demonstrated 98% concordance, with 99% agreement between the reviewers. Similarly, we identified 108 ICD-10 codes for endovascular aortic aneurysm repair, an 11% increase from the initial 97 ICD-9 codes. We again noted excellent concordance and agreement (98% and 100%, respectively). Manual review identified 4 ICD-10 codes incorrectly translated from ICD-9 codes for carotid revascularization, and 3 codes incorrectly translated for endovascular aortic aneurysm repair. Conclusions Algorithms to crosswalk lists of ICD-9 codes to ICD-10 codes can leverage electronic resources to minimize the burden of code translation. However, manual review for code validation may be necessary, with collaboration across institutions for researchers to share their efforts.

Prevalence and Cost of Care Cascades After Low-Value Preoperative Electrocardiogram for Cataract Surgery in Fee-for-Service Medicare Beneficiaries

IMPORTANCE

Low-value care is prevalent in the United States, yet little is known about the downstream health care use triggered by low-value services. Measurement of such care cascades is essential to understanding the full consequences of low-value care.

OBJECTIVE

To describe cascades (tests, treatments, visits, hospitalizations, and new diagnoses) after a common low-value service, preoperative electrocardiogram (EKG) for patients undergoing cataract surgery.

DESIGN, SETTING, AND PARTICIPANTS

Observational cohort study using fee-for-service Medicare claims data from beneficiaries aged 66 years or older without known heart disease who were continuously enrolled between April 1, 2013, and September 30, 2015, and underwent cataract surgery between July 1, 2014 and June 30, 2015. Data were analyzed from March 12, 2018, to April 9, 2019.

EXPOSURES

Receipt of a preoperative EKG. The comparison group included patients who underwent cataract surgery but did not receive a preoperative EKG.

MAIN OUTCOMES AND MEASURES

Cascade event rates and associated spending in the 90 days after preoperative EKG, or in a matched timeframe for the comparison group. Secondary outcomes were patient, physician, and area-level characteristics associated with experiencing a potential cascade.

RESULTS

Among 110 183 cataract surgery recipients, 12 408 (11.3%) received a preoperative EKG (65.6% of them were female); of those, 1978 (15.9%) had at least 1 potential cascade event. The comparison group included 97 775 participants (63.1% female). Those who received a preoperative EKG experienced between 5.11 (95% CI, 3.96-6.25) and 10.92 (95% CI, 9.76-12.08) additional events per 100 beneficiaries relative to the comparison group. This included between 2.18 (95% CI, 1.34-3.02) and 7.98 (95% CI, 7.12-8.84) tests, 0.33 (95% CI, 0.19-0.46) treatments, 1.40 (95% CI, 1.18-1.62) new patient cardiology visits, and 1.21 (95% CI, 0.62-1.79) new cardiac diagnoses. Spending for the additional services was up to $565 per Medicare beneficiary (95% CI, $342-$775), or an estimated $35 025 923 annually across all Medicare beneficiaries in addition to the $3 275 712 paid for the preoperative EKGs. Among preoperative EKG recipients, those who were older (adjusted odds ratio [aOR] for patients aged 75 to 84 years vs 66 to 74 years old, 1.42; 95% CI, 1.28-1.57), had more chronic conditions (aOR for each additional Elixhauser condition, 1.18; 95% CI, 1.14-1.22), lived in more cardiologist-dense areas (aOR, 1.05; 95% CI, 1.02-1.09), or had their preoperative EKG performed by a cardiac specialist rather than a primary care physician (aOR, 1.26; 95% CI, 1.10-1.43) were more likely to experience a potential cascade.

CONCLUSIONS AND RELEVANCE

Care cascades after preoperative EKG for cataract surgery are infrequent but costly. Policy and practice interventions to reduce low-value services and the cascades that follow could yield substantial savings.

ICD-10 Coding Will Challenge Researchers: Caution and Collaboration may Reduce Measurement Error and Improve Comparability Over Time

BACKGROUND

The October 1, 2015 US health care diagnosis and procedure codes update, from the 9th to 10th version of the International Classification of Diseases (ICD), abruptly changed the structure, number, and diversity of codes in health care administrative data. Translation from ICD-9 to ICD-10 risks introducing artificial changes in claims-based measures of health and health services.

OBJECTIVE

Using published ICD-9 and ICD-10 definitions and translation software, we explored discontinuity in common diagnoses to quantify measurement changes introduced by the upgrade.

DESIGN

Using 100% Medicare inpatient data, 2012-2015, we calculated the quarterly frequency of condition-specific diagnoses on hospital discharge records. Years 2012-2014 provided baseline frequencies and historic, annual fourth-quarter changes. We compared these to fourth quarter of 2015, the first months after ICD-10 adoption, using Centers for Medicare and Medicaid Services Chronic Conditions Data Warehouse (CCW) ICD-9 and ICD-10 definitions and other commonly used definitions sets.

RESULTS

Discontinuities of recorded CCW-defined conditions in fourth quarter of 2015 varied widely. For example, compared with diagnosis appearance in 2014 fourth quarter, in 2015 we saw a sudden 3.2% increase in chronic lung disease and a 1.8% decrease in depression; frequency of acute myocardial infarction was stable. Using published software to translate Charlson-Deyo and Elixhauser conditions yielded discontinuities ranging from -8.9% to +10.9%.

CONCLUSIONS

ICD-9 to ICD-10 translations do not always align, producing discontinuity over time. This may compromise ICD-based measurements and risk-adjustment. To address the challenge, we propose a public resource for researchers to share discovered discontinuities introduced by ICD-10 adoption and the solutions they develop.

Trends in Testosterone Prescribing for Age-related Hypogonadism in Men With and Without Heart Disease

This cohort study examines the trend in prescription testosterone use among male Medicare beneficiaries with and without coronary artery disease aged 50 years or older.

Medicaid Expansion and Prescription Trends: Opioids, Addiction Therapies, and Other Drugs

BACKGROUND

Opioid overdose deaths in the United States have climbed since 1999. In 2014, the Affordable Care Act prompted some states to expand Medicaid programs, providing low-cost prescription access to millions of Americans. Some have questioned whether Medicaid expansion might worsen the opioid crisis.

OBJECTIVE

To test the association between the expansion of state Medicaid programs and Medicaid-paid prescriptions of opioid pain relievers and opioid addiction therapies.

RESEARCH DESIGN

We analyzed the 2010-2016 Medicaid State Drug Utilization Data using a difference-in-differences regression approach, comparing prescriptions per enrollee between states that expanded Medicaid in 2014 and states that did not. We compared opioid pain relievers and opioid addiction therapies to 5 other commonly prescribed drug types important to the Medicaid expansion population (antidepressants, antihypertensives, diabetes medications, cholesterol treatments, and contraceptives) and to overall prescription volume. A secondary analysis compared opioid pain relievers and opioid addiction therapies, between states with high and low overdose death rates.

RESULTS

We found overall prescription use per enrollee was higher after 2014. Relative growth in opioid pain reliever prescriptions was modest compared with growth in medications for depression, hypertension, diabetes, and high cholesterol. Growth in prescriptions used to treat opioid use disorder greatly outpaced other drugs, suggesting important gains in access to addiction treatments; growth was higher in states with higher pre-2014 overdose death rates.

CONCLUSIONS

Our results suggest Medicaid expansion benefited a population with unique needs, and that Medicaid expansion could be a valuable tool in addressing the opioid overdose epidemic.

Overuse and insurance plan type in a privately insured population

OBJECTIVES

A substantial portion of healthcare spending is wasted on services that do not directly improve patient health and that cause harm in some cases. Features of health insurance coverage, including enrollment in high-deductible health plans (HDHPs) or health maintenance organizations (HMOs), may provide financial and nonfinancial mechanisms to potentially reduce overuse of low-value healthcare services.

STUDY DESIGN

Using 2009 to 2013 administrative data from 3 large commercial insurers, we examined patient characteristics and health insurance plan types associated with overuse of 6 healthcare services identified by the Choosing Wisely campaign.

METHODS

We explored associations between overuse and patient characteristics using multivariate logistic regression models, including patient age, gender, enrollment in an HMO, enrollment in an HDHP, an indicator of primary care fragmentation, and number of outpatient visits as explanatory variables.

RESULTS

Measurement of services highlighted as potential overuse by the Choosing Wisely recommendations revealed low to moderate prevalence, depending on the service. HMO coverage and enrollment in HDHPs were significantly associated with differences in prevalence of all 6 services, albeit differently in terms of the direction of the effects. Primary care fragmentation was significantly associated with higher rates of overuse.

CONCLUSIONS

Neither HDHPs nor HMO plans, with their closed networks and referral requirements, consistently reduced overuse, although HMO plans were never associated with higher rates of overuse. As policy makers seek levers for reducing low-value healthcare utilization, health insurance plan features may prove a valuable target, although the effect may be complicated by other factors.

Direct oral anticoagulant and antiplatelet combination therapy: Hemorrhagic events in coronary artery stent recipients

Direct oral anticoagulant (DOAC) use is growing as monotherapy and combined with platelet inhibitors. The safety of such combination therapy, especially in comparison to regimens including warfarin, in real world populations remains uncertain. We investigated hemorrhage associated with DOAC and antiplatelet combination therapy in a cohort of elderly coronary artery stent recipients. We employed Medicare data 2010-2013 for a 40% random sample of beneficiaries enrolled in inpatient, outpatient and prescription benefits. We used Cox proportional hazards models to examine the association of the combination anticoagulant (DOAC or warfarin) plus antiplatelets with major hemorrhage events (upper gastrointestinal or intracranial) in the 12 months following stent placement. We identified 70,900 stent recipients. 14.4% had atrial fibrillation (AF) diagnosis preoperatively. Among the 24.5 million observation days, exposure distribution was: 73.8% antiplatelets only, 4.7% antiplatelets plus warfarin, 0.6% antiplatelets plus DOAC, 2.2% warfarin only, 0.3% DOAC only and 18.4% no observed antiplatelets or anticoagulant. Overall, 8,029 patients (11.3%) experienced major hemorrhage. Among AF patients, compared to antiplatelets only, DOAC plus antiplatelets was associated with increased hemorrhage risk (HR, 1.94; 95%CI, 1.48-2.54); warfarin plus antiplatelets conferred comparable bleed risk (HR, 1.69; 95%CI, 1.47-1.94). In the non-AF group, compared to antiplatelets alone, combination DOAC plus antiplatelets (HR, 3.09; 95%CI, 2.15-4.46), and warfarin plus antiplatelets (HR, 2.21; 95%CI, 1.97-2.48) conferred greater bleed risk. Among elderly coronary artery stent recipients with AF, the two drug combinations, DOAC plus antiplatelets and warfarin plus antiplatelets, were associated with similarly increased risk of major hemorrhage compared to antiplatelets alone.

Measuring overuse with electronic health records data

OBJECTIVES

To measure overuse of low-value care using electronic health record (EHR) data and manual chart review and to evaluate whether certain low-value services are better captured using EHR data.

STUDY DESIGN

We implemented algorithms to extract performance on 13 Choosing Wisely-identified healthcare services using EHR data at a large physician practice group between 2011 and 2013.

METHODS

We calculated rates of overuse using automated EHR extracts. We manually reviewed the charts for 200 cases of overuse for each measure to determine if they had clinical risk factors that could explain use of the low-value service and then calculated adjusted rates of overuse. We explored trends in overuse for each low-value service in the 3-year duration using logistic regression.

RESULTS

Unadjusted rates of overuse ranged from 0.2% to 92%. Automated EHR extracts and manual chart review identified explanatory risk factors for most measures, although the magnitude varied: for some measures (eg, bone densitometry exam for women younger than 65 years), manual chart review did not identify many additional risks (3.0%). In contrast, in patients who had sinus computed tomography or an antibiotic prescription for uncomplicated acute rhinosinusitis, manual chart review identified more explanatory risk factors (22.5%) than the automated EHR extract (9.5%). Adjusted rates of overuse ranged from 0.2% to 61.9%. Eight services demonstrated a statistically significant decrease in overuse over 3 years, while 1 increased significantly.

CONCLUSIONS

The use of EHR data, both extracted and manually abstracted, provides an opportunity to more accurately and reliably identify overuse of low-value healthcare services.

Atrial Fibrillation and Stent Selection (Bare Metal vs Drug Eluting) (from Medicare Claims)

In treating coronary artery disease, many factors influence the choice of bare metal stent (BMS) or drug-eluting stent (DES), including bleed risk and suitability for prolonged dual antiplatelet therapy. Atrial fibrillation (AF) further complicates this choice, due to common use of anticoagulation. We examined stent selection in the United States by AF status and across academic medical centers (AMCs) to explore how cardiologists are managing this complex choice. Using a 100% Medicare denominator file and associated claims (2008 to 2012), we identified patients over age 65 receiving inpatient coronary artery stents. We measured BMS and DES use in patients with AF and in patients without AF and assessed variation in stent choice across AMCs, adjusting for differences in age, gender, and race. We identified 898,788 stent episodes among elderly Medicare beneficiaries. BMS, as a percentage of total inpatient stent episodes, decreased from 2008 to 2012, in patients with AF (42% to 34%) and in patients without AF (32% to 23%). Across AMCs, adjusted stent choice varied substantially, but more so for patients with AF (2008 to 2012 median BMS 44% to 39%, annual interquartile ratios range 1.8 to 2.3) than patients without AF (2008 to 2012 median BMS 33% to 25%, annual interquartile ratios range 2.0 to 1.8). In conclusion, among stent recipients, patients with AF are more likely to receive BMS than patients without AF but treatment varies across systems, suggesting a lack of consensus. Studies of stent choice and outcomes among patients with AF are needed to guide care decisions and optimize outcomes.

For Selected Services, Blacks And Hispanics More Likely To Receive Low-Value Care Than Whites

US minority populations receive fewer effective health services than whites. Using Medicare administrative data for 2006-11, we found no consistent, corresponding protection against the receipt of ineffective health services. Compared with whites, blacks and Hispanics were often more likely to receive the low-value services studied.

The growing problem of co-treatment with opioids and benzodiazepines

A powerful example of potentially dangerous low value care

Payer Type and Low-Value Care: Comparing Choosing Wisely Services across Commercial and Medicare Populations

OBJECTIVE

To compare low-value health service use among commercially insured and Medicare populations and explore the influence of payer type on the provision of low-value care.

DATA SOURCES

2009-2011 national Medicare and commercial insurance administrative data.

DESIGN

We created claims-based algorithms to measure seven Choosing Wisely-identified low-value services and examined the correlation between commercial and Medicare overuse overall and at the regional level. Regression models explored associations between overuse and regional characteristics.

METHODS

We created measures of early imaging for back pain, vitamin D screening, cervical cancer screening over age 65, prescription opioid use for migraines, cardiac testing in asymptomatic patients, short-interval repeat bone densitometry (DXA), preoperative cardiac testing for low-risk surgery, and a composite of these.

PRINCIPAL FINDINGS

Prevalence of four services was similar across the insurance-defined groups. Regional correlation between Medicare and commercial overuse was high (correlation coefficient = 0.540-0.905) for all measures. In both groups, similar region-level factors were associated with low-value care provision, especially total Medicare spending and ratio of specialists to primary care physicians.

CONCLUSIONS

Low-value care appears driven by factors unrelated to payer type or anticipated reimbursement. These findings suggest the influence of local practice patterns on care without meaningful discrimination by payer type.

Use of Guideline-Directed Medications for Heart Failure Before Cardioverter-Defibrillator Implantation

BACKGROUND

Guideline-directed medical therapy (GDMT) for heart failure with reduced ejection fraction (HFrEF) is recommended before primary prevention implantable cardioverter-defibrillator (ICD) placement. Adherence to this recommendation and associated outcomes are unknown.

OBJECTIVES

This study examined the use of GDMT (≥1 prescription filled for both a renin-angiotensin inhibitor [RAI] and a heart failure-approved beta-blocker [HFBB]) within 90 days before primary prevention ICD placement in patients with HFrEF.

METHODS

Data from the National Cardiovascular Data Registry ICD Registry were merged with a 40% random sample of Medicare administrative data. Prescription fills for recipients of primary prevention ICD between 2007 and 2011 were examined, analyzing GDMT overall and for each U.S. hospital referral region. We identified characteristics associated with GDMT and the association with 1-year mortality.

RESULTS

Among 19,733 patients with HFrEF and primary prevention ICD, 61.1% filled any GDMT before implantation. Across hospital referral regions, GDMT was applied in 51% to 71%. The strongest predictors of any GDMT included absence of chronic renal disease or nonsustained ventricular tachycardia, low-income prescription benefits subsidy, and less recent left ventricular ejection fraction evaluation. Patients receiving GDMT versus those without had a lower 1-year mortality rate after ICD implantation (11.1% vs. 16.2%), even after adjustment for comorbidities, left ventricular ejection fraction, and functional heart failure class.

CONCLUSIONS

Rates of GDMT for HFrEF before primary prevention ICD implantation were low, and failure to achieve GDMT was associated with significantly decreased 1-year survival.

State Legal Restrictions and Prescription-Opioid Use among Disabled Adults

BACKGROUND

In response to rising rates of opioid abuse and overdose, U.S. states enacted laws to restrict the prescribing and dispensing of controlled substances. The effect of these laws on opioid use is unclear.

METHODS

We tested associations between prescription-opioid receipt and state controlled-substances laws. Using Medicare administrative data for fee-for-service disabled beneficiaries 21 to 64 years of age who were alive throughout the calendar year (8.7 million person-years from 2006 through 2012) and an original data set of laws (e.g., prescription-drug monitoring programs), we examined the annual prevalence of beneficiaries with four or more opioid prescribers, prescriptions yielding a daily morphine-equivalent dose (MED) of more than 120 mg, and treatment for nonfatal prescription-opioid overdose. We estimated how opioid outcomes varied according to eight types of laws.

RESULTS

From 2006 through 2012, states added 81 controlled-substance laws. Opioid receipt and potentially hazardous prescription patterns were common. In 2012 alone, 47% of beneficiaries filled opioid prescriptions (25% in one to three calendar quarters and 22% in every calendar quarter); 8% had four or more opioid prescribers; 5% had prescriptions yielding a daily MED of more than 120 mg in any calendar quarter; and 0.3% were treated for a nonfatal prescription-opioid overdose. We observed no significant associations between opioid outcomes and specific types of laws or the number of types enacted. For example, the percentage of beneficiaries with a prescription yielding a daily MED of more than 120 mg did not decline after adoption of a prescription-drug monitoring program (0.27 percentage points; 95% confidence interval, -0.05 to 0.59).

CONCLUSIONS

Adoption of controlled-substance laws was not associated with reductions in potentially hazardous use of opioids or overdose among disabled Medicare beneficiaries, a population particularly at risk. (Funded by the National Institute on Aging and others.).

Physician perceptions of Choosing Wisely and drivers of overuse

OBJECTIVES

Little is known regarding physicians' views on health service overuse or their awareness of the American Board of Internal Medicine Foundation's Choosing Wisely campaign. Through the Survey on Overuse and Knowledge of Choosing Wisely, we assessed physician views on hypothesized drivers of overuse and Choosing Wisely.

STUDY DESIGN

We designed the survey to investigate physicians' knowledge of, awareness of, and feelings toward Choosing Wisely, along with their concerns about malpractice, perception of patient demand, discomfort with uncertainty, and cost-consciousness. Where possible, we used pre-validated survey instruments.

METHODS

We distributed the survey to clinicians practicing at Atrius Health, the largest ambulatory care provider in Massachusetts. We analyzed 584 responses (72% response rate) and calculated 3 previously validated scales.

RESULTS

Primary care physicians reported significantly greater awareness of Choosing Wisely (47.2%) than medical specialists (37.4%) and surgical specialists (27%). A majority (62%) of all respondents reported they found uncertainty involved in providing care disconcerting. Approximately one-third felt it unfair to ask physicians to be both cost-conscious and concerned with welfare, thought too much emphasis was placed on costs, and thought doctors were too busy to worry about costs. Surgical specialists were more concerned about malpractice, whereas primary care physicians reported feeling significantly more pressure from patients for tests and procedures.

CONCLUSIONS

Knowledge of Choosing Wisely is limited, but primary care physicians are more aware of the campaign than specialists. Although hypothesized drivers of overuse are prevalent, most physicians support cost-consciousness in medicine and embrace their responsibility in reducing costs.

Diabetes Pharmacotherapies and Bladder Cancer: A Medicare Epidemiologic Study

OBJECTIVE

Patients with type II diabetes have an increased risk of bladder cancer and are commonly treated with thiazolidinediones and angiotensin receptor blockers (ARBs), which have been linked to cancer risk. We explored the relationship between use of one or both of these medication types and incident bladder cancer among diabetic patients (diabetics) enrolled in Medicare.

RESEARCH DESIGN AND METHODS

We constructed both a prevalent and incident retrospective cohort of pharmacologically treated prevalent diabetics enrolled in a Medicare fee-for-service plan using inpatient, outpatient (2003-2011) and prescription (2006-2011) administrative data. The association of incident bladder cancer with exposure to pioglitazone, rosiglitazone and ARBs was studied using muitivariable Cox's hazard models with time-dependent covariates in each of the two cohorts.

RESULTS

We identified 1,161,443 prevalent and 320,090 incident pharmacologically treated diabetics, among whom 4433 and 1159, respectively, developed incident bladder cancers. In the prevalent cohort mean age was 75.1 years, mean follow-up time was 38.0 months, 20.2% filled a prescription for pioglitazone during follow-up, 10.4% received rosiglitazone, 31.6% received an ARB and 8.0% received combined therapy with pioglitazone + ARB. We found a positive association between bladder cancer and duration of pioglitazone use in the prevalent cohort (P for trend = 0.008), with ≥24 months of pioglitazone exposure corresponding to a 16% (95% confidence interval 0-35%) increase in the incidence of bladder cancer compared to no use. There was a positive association between bladder cancer and rosiglitazone use for <24 months in the prevalent cohort, but no association with ARB use. There were no significant associations in the incident cohort.

CONCLUSIONS

We found that the incidence of bladder cancer increased with duration of pioglitazone use in a prevalent cohort of diabetics aged 65+ years residing in the USA, but not an incident cohort.

Prescription Use among Children with Autism Spectrum Disorders in Northern New England: Intensity and Small Area Variation

OBJECTIVE

To measure prescription use intensity and regional variation of psychotropic and 2 important nonpsychotropic drug groups among children with autism spectrum disorders (ASDs) compared with children in the general population.

STUDY DESIGN

Cross-sectional study of ambulatory prescription fills from Maine, Vermont, and New Hampshire all-payer administrative data, 2007-2010.

RESULTS

Overall there were 13,100 children diagnosed with ASD (34,584 person years [PYs]) and 936,721 (1.7 million PYs) without ASD diagnosis. The overall prescription fill rate was 16.6 per PY in children with ASD and 4.1 per PY in the general population. Psychotropic use among children with ASDs was 9-fold the general population rate (7.80 vs 0.85 fills per PY); these children comprised 2.0% of the pediatric population but received 15.6% of psychotropics. Nonpsychotropic drug use was also higher in the population with ASD, particularly the youngest: among those under age 3 years, antibiotic use was 2-fold and antacid use nearly 5-fold the general population rate (3.2 vs 1.4 and 1.0 vs 0.2 per PY, respectively). Among children with ASDs, prescription use varied substantially across hospital service areas, as much as 3-fold for antacids and alpha agonists, more than 4-fold for benzodiazepines (5th to 95th percentile).

CONCLUSIONS

The overall psychotropic and nonpsychotropic prescription intensity among children with ASDs is characterized by broad regional variation, suggesting diverse provider responses to pharmacotherapeutic uncertainty. This variation highlights a need for more research, practice-based learning, and shared decision making with caregivers surrounding therapy for children with ASDs.

Physician response to patient request for unnecessary care

OBJECTIVES

Evaluating unnecessary US medical practices, and the strategies that reduce them, are increasingly recognized as crucial to healthcare financing sustainability. Provider factors are known to affect unnecessary medical practices, yet little is known about how physician responses to patient requests for unnecessary care affect these practices. Among primary care physicians (PCPs), we investigated 2 types of unnecessary medical practices triggered by patient requests: a) unnecessary specialty referrals and b) prescriptions for brand-name drugs when generic alternatives are available.

STUDY DESIGN AND METHODS

We used data from a survey of a nationally representative sample of 840 US PCPs in 2009. Response rates for family practice (n = 274), internal medicine (n = 257), and pediatrics (n = 309) were 67.5%, 60.8%, and 72.7%, respectively.

RESULTS

In response to patient requests, 51.9% of PCPs reported making unnecessary specialty referrals and 38.7% prescribed brand-name drugs. Family physicians (odds ratio [OR], 2.77; 95% CI, 1.77-4.34) and internal medicine physicians (OR, 4.51; 95% CI, 2.87-7.06) were more likely than pediatricians to prescribe brand-name drugs. PCP specialty was similarly associated with unnecessary referrals. Other predictors of acquiescence to patient requests included interactions with drug/device representatives, more years of clinical experience, seeing fewer safety net patients, and solo/2-person practice organizations. Area-level Medicare spending was not associated with the 2 unnecessary practices.

CONCLUSIONS

Many PCPs reported acquiescing to patient requests for unnecessary care. Provider and organizational factors predicted this behavior. Policies aimed at reducing such practice could improve care quality and lower cost. Patient and physician incentives that can potentially reduce unnecessary medical practices warrant exploration.

Abstract A27: Association between diabetes pharmacotherapies and bladder cancer: A Medicare epidemiologic study

Background: Patients with type two diabetes mellitus (T2DM) have an increased risk of bladder cancer; they are also commonly treated with pioglitizone and angiotensin receptor blockers (ARBs), medications associated with an increased cancer risk in some studies. This study aimed to further explore the relationship between these medications and bladder cancer in an older population with higher baseline risk of this cancer and common exposure to these drugs. We examine the risk associated with exposure to the individual drug/drug class, and risk associated with exposure to both to assess for a potential cancer promoting interaction.

Methods: Using a 40% Medicare random sample denominator and corresponding inpatient, outpatient (2004-2012) and prescription (2006-2012) claims, we studied fee-for-service beneficiaries age 67 and older. We identified pharmacotherapeutically treated diabetics (2006 to 2011) defined as patients filling one or more prescription for an oral hypoglycemic drug. Individual-level, time varying exposure to all oral and injectable hyopglycemics as well as ARBs and ACE-Is was calculated as count of drug-specific daily doses; daily doses were derived empirically as the median daily dose (MDDs) for this population. Cox proportional hazards models, adjusted for patient characteristics, comorbidities and diabetes complications, were used to examine the risk of incident bladder cancer associated with cumulative observed exposure to pioglitazone and ARBs. To address possible indication bias for the exposures of interest, we also estimated risk associated ACE-I exposure and non-pioglitazone TZD exposure, substitutes for the exposures of interest not suspected of increasing cancer risk. Interaction terms estimated the effect of exposure to ARBs with pioglitazone and ARBs with non-pioglitazone TZD. Secondary models repeated main models stratified on gender due to the higher baseline risk of bladder cancer in men.

Results: Overall, we identified 1,161,443 diabetics, 62.5% female, 10.7% black, 9.8% Hispanic. Mean age at the beginning of follow-up was 75.1 years (SD 7.5); 18.9% had evidence of current or past tobacco use. Mean follow time was 42.2 months (SD 26.4). Overall, 20.2% of the cohort used some pioglitazone (mean exposure 642 MDDs); 31.7% used some ARB (Mean exposure 745 MDDs); 10.4% used some non-pioglitazone TZD (mean exposure 503 MDDs); 59.1% used some ACE-I (mean exposure 855 MDDs). 8.0% used both pioglitazone and ARBs; 13.2% used both pioglitazone and ACE-I; 4.1% used non-pioglitazone TZD and ARB; and 6.8% used non-pioglitazone TZD and ACE-I. In total, 5,874 bladder cancers were identified, 69% amongst men, an incidence of 0.28/1000 PY in men and 0.07/1000 PY in women. Compared to no observed exposure, the risk of bladder cancer associated with each additional year of exposure at median daily dose was: 0.97 (95% CI 0.89, 1.06) for pioglitazone, 1.04 (95% CI 0.97, 1.12) for ARBs; 0.99 (95% CI 0.95, 1.03) for ACE-Is and 0.98 (95% CI 0.86, 1.12) for non- pioglitizone TZDs. The risk of bladder cancer associated with each additional year of exposure to both pioglitizone and ARB was 0.97 (95% CI 0.91, 1.04); for pioglitazone and ACE-I, 1.01 (95% CI 0.98, 1.03); for non-pioglitazone TZD and ARB, 0.99 (95% CI 0.91, 1.08); and for non-pioglitazone TZD and ACE-I, 0.99 (95% CI 0.93, 1.04). Results did not differ when stratified by sex.

Conclusions: We found no increased risk of bladder cancer associated with pioglitizone use, ARB use or exposure to both over an average observation time of 3.5 years. While the findings of this observational study should be confirmed in a prospective trial of longer duration, overall the results should reassure diabetics and their clinicians using these products to minimize the substantial health risks associated with diabetes.

Citation Format: Jeremy Smith, Rebecca Zaha, Todd A. Mackenzie, Margaret R. Karagas, Nancy E. Morden. Association between diabetes pharmacotherapies and bladder cancer: A Medicare epidemiologic study. [abstract]. In: Proceedings of the Thirteenth Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2014 Sep 27-Oct 1; New Orleans, LA. Philadelphia (PA): AACR; Can Prev Res 2015;8(10 Suppl): Abstract nr A27.

Statins and subarachnoid hemorrhage in Medicare patients with unruptured cerebral aneurysms

BACKGROUND

Statins have been shown to decrease aneurysm progression and rupture in two experimental settings: animals with cerebral aneurysm and humans with abdominal aortic aneurysms.

AIMS

To investigate statin use and outcomes in humans with unruptured cerebral aneurysms through Medicare administrative data.

METHODS

We used a 40% random sample Medicare denominator file and corresponding inpatient, outpatient (2003-2011), and prescription (2006-2011) claims to conduct a retrospective cohort study of patients diagnosed with unruptured cerebral aneurysms, between 2003 and 2011. We used propensity score-adjusted models to investigate the association between statin use and risk of subarachnoid hemorrhage. Secondary analyses repeated the main models stratified on tobacco use status and separately assessed other composite outcomes.

RESULTS

We identified 28 931 patients with unruptured cerebral aneurysms (average age 72·0 years, 72·6% female); mean follow-up was 30·0 months; 41·3% used statins. Overall, 593 patients developed subarachnoid hemorrhage, and 703 underwent treatment before subarachnoid hemorrhage. Current or recent statin use was not associated with a difference in subarachnoid hemorrhage risk (odds ratio, 1·03; 95% conflict of interest 0·86-1·23); models stratified on tobacco use status were nearly identical. No association was observed between statin use and the composite outcome of subarachnoid hemorrhage or aneurysm treatment (odds ratio, 0·94; 95% conflict of interest, 0·84-1·06). The risk of subarachnoid hemorrhage or out-of-hospital death was lower among statin users (odds ratio, 0·69; 95% conflict of interest, 0·64-0·74).

CONCLUSIONS

Statin use by patients with unruptured cerebral aneurysms was not associated with subarachnoid hemorrhage risk. Given the prior animal experimental studies demonstrating a protective effect, further prospective studies are needed to investigate the potential relationship.

Role of pharmacy services in accountable care organizations

BACKGROUND

The accountable care organization (ACO) model being adopted across the United States aims to improve patient care and reduce costs. Little is known about whether commercial ACO contracts include accountability for prescription drug spending or how ACOs are engaging outpatient pharmacies and managing prescription drug use.

OBJECTIVE

To explore how ACOs are addressing drug spending and pharmacy services-a potentially important determinant of quality and total spending.

METHODS

We used data from 2 waves of the National Survey of Accountable Care Organizations (N = 270), a survey completed by ACOs that were established prior to July 2013. ACO executives were asked about ACO engagement of pharmacy services, pharmacy-related health information technology capabilities, and ACO accountability for prescription drug spending. 

RESULTS

Among ACOs with commercial contracts, 77% reported being held responsible for prescription spending by their largest contract. Considering all ACOs (Medicare, Medicaid, and/or commercial contracts), 45% reported at least 1 contract included prescription drug spending responsibility. Nearly half of ACOs reported a formal relationship with a pharmacy; 26% included a pharmacy within the ACO; and 19% had contracted pharmacy services. On average, compared with those that do not, ACOs that engage pharmacies have a broader range of services and provider types, commercial and public contracts, and greater experience with payment reform. 

CONCLUSIONS

Management of pharmacy services and prescription spending will likely influence commercial ACO contract success. Given the broad potential impact of prescription use on overall spending and quality, payers might encourage integration of pharmacy services in ACOs through prescribing quality and prescription spending performance measures.

Oral bisphosphonates and upper gastrointestinal toxicity: a study of cancer and early signals of esophageal injury

SUMMARY

We evaluated the association between bisphosphonate use and (1) upper gastrointestinal cancer, (2) upper endoscopy, (3) incident Barrett's esophagus, and (4) prescription antacid initiation among Medicare beneficiaries. We found no bisphosphonate-cancer association and negative bisphosphonate-Barrett's association.

INTRODUCTION

Bisphosphonates can irritate the esophagus; a cancer association has been suggested. Widespread bisphosphonates use compels continued investigation of upper gastrointestinal toxicity.

METHODS

Using a 40% Medicare random sample denominator, inpatient, outpatient (2003-2011), and prescription (2006-2011) claims, we studied patients age 68 and older with osteoporosis and/or oral bisphosphonate use. Inverse propensity weighting estimated marginal structural models for the effect of bisphosphonate intensity (pills per month) and cumulative bisphosphonate pills received on upper gastrointestinal cancer risk. Secondary analyses of sub-cohorts without past bisphosphonates or upper endoscopy assessed bisphosphonate initiation and risk of (1) upper endoscopy, (2) incident Barrett's esophagus, and (3) prescription antacid initiation.

RESULTS

The cohort included 1.64 million beneficiaries: 87.9% women, mean age, 76.8 (standard deviation (SD) 9.3); mean follow-up, 39.6 months; 38.1% received oral bisphosphonates. Cumulative bisphosphonate receipt, among users, ranged from 4 to 252 pills (5th to 95th percentile). We identified 2,308 upper gastrointestinal cancers (0.43/1000 person years). We found no association between cumulative bisphosphonate pills and cancer, odds ratio (OR) for each additional pill 1.00 (95% confidence interval (CI) 1.00, 1.00). In sub-cohorts, compared to none, lowest cumulative bisphosphonate use (one to nine pills) was associated with higher risk of endoscopy (OR 1.11, 95% CI 1.08-1.14) and antacid initiation (OR 1.13, 95% CI 1.10-1.16); higher intensity conferred no increased risk. Higher intensity and higher cumulative bisphosphonate category were associated with lower Barrett's risk.

CONCLUSIONS

We found no bisphosphonate-cancer association and negative bisphosphonate-Barrett's association. Bisphosphonate initiation appears to identify patients susceptible to early irritating effects; clinicians might offer alternatives and delay endoscopy or antacids.

Choosing wisely: prevalence and correlates of low-value health care services in the United States

BACKGROUND

Specialty societies in the United States identified low-value tests and procedures that contribute to waste and poor health care quality via implementation of the American Board of Internal Medicine Foundation's Choosing Wisely initiative.

OBJECTIVE

To develop claims-based algorithms, to use them to estimate the prevalence of select Choosing Wisely services and to examine the demographic, health and health care system correlates of low-value care at a regional level.

DESIGN

Using Medicare data from 2006 to 2011, we created claims-based algorithms to measure the prevalence of 11 Choosing Wisely-identified low-value services and examined geographic variation across hospital referral regions (HRRs). We created a composite low-value care score for each HRR and used linear regression to identify regional characteristics associated with more intense use of low-value services.

PATIENTS

Fee-for-service Medicare beneficiaries over age 65.

MAIN MEASURES

Prevalence of selected Choosing Wisely low-value services.

KEY RESULTS

The national average annual prevalence of the selected Choosing Wisely low-value services ranged from 1.2% (upper urinary tract imaging in men with benign prostatic hyperplasia) to 46.5% (preoperative cardiac testing for low-risk, non-cardiac procedures). Prevalence across HRRs varied significantly. Regional characteristics associated with higher use of low-value services included greater overall per capita spending, a higher specialist to primary care ratio and higher proportion of minority beneficiaries.

CONCLUSIONS

Identifying and measuring low-value health services is a prerequisite for improving quality and eliminating waste. Our findings suggest that the delivery of wasteful and potentially harmful services may be a fruitful area for further research and policy intervention for HRRs with higher per-capita spending. These findings should inform action by physicians, health systems, policymakers, payers and consumer educators to improve the value of health care by targeting services and areas with greater use of potentially inappropriate care.

Small geographic area variations in prescription drug use

BACKGROUND

Despite the frequency of pediatric prescribing little is known about practice differences across small geographic regions and payer type (Medicaid and commercial).

OBJECTIVE

The goal of this research was to quantify variation in prescription drug use among northern New England children.

METHODS

Northern New England, all-payer administrative data (2007-2010) permitted study of prescriptions for 949 821 children ages 0 to 17 years (1.75 million person-years [PYs]; 54% Medicaid, 46% commercial). Age- and gender adjusted overall and drug group-specific prescription use was quantified according to payer type (Medicaid or commercial) and within payer type across 69 hospital service areas (HSAs). We measured prescription fills per PY (rate) and annual, mean percentage of the population with any drug group-specific fills (prevalence).

RESULTS

Overall mean annual prescriptions per PY were 3.4 (commercial) and 5.5 (Medicaid). Generally, these payer type differences were smaller than HSA-level variation within payer type. HSA-level rates of attention-deficit/hyperactivity disorder drug use (5th-95th percentile) varied twofold in Medicaid and more than twofold in commercially insured children; HSA-level antidepressant use varied more than twofold within each payer type. Antacid use varied threefold across HSAs and was highest in infants where commercial use paradoxically exceeded Medicaid. Prevalence of drug use varied as much as rates across HSAs.

CONCLUSIONS

Prescription use was higher among Medicaid-insured than commercially insured children. Regional variation generally exceeded payer type differences, especially for drugs used in situations of diagnostic and therapeutic uncertainty. Efforts should advance best pediatric prescribing discussions and shared decision-making.

Overuse of short-interval bone densitometry: assessing rates of low-value care

We evaluated the prevalence and geographic variation of short-interval (repeated in under 2 years) dual-energy X-ray absorptiometry tests (DXAs) among Medicare beneficiaries. Short-interval DXA use varied across regions (coefficient of variation = 0.64), and unlike other DXAs, rates decreased with payment cuts.

INTRODUCTION

The American College of Rheumatology, through the Choosing Wisely initiative, identified measuring bone density more often than every 2 years as care "physicians and patients should question." We measured the prevalence and described the geographic variation of short-interval (repeated in under 2 years) DXAs among Medicare beneficiaries and estimated the cost of this testing and its responsiveness to payment change.

METHODS

Using 100 % Medicare claims data, 2006-2011, we identified DXAs and short-interval DXAs for female Medicare beneficiaries over age 66. We determined the population rate of DXAs and short-interval DXAs, as well as Medicare spending on short-interval DXAs, nationally and by hospital referral region (HRR).

RESULTS

DXA use was stable 2008-2011 (12.4 to 11.5 DXAs per 100 women). DXA use varied across HRRs: in 2011, overall DXA use ranged from 6.3 to 23.0 per 100 women (coefficient of variation = 0.18), and short-interval DXAs ranged from 0.3 to 8.0 per 100 women (coefficient of variation = 0.64). Short-interval DXA use fluctuated substantially with payment changes; other DXAs did not. Short-interval DXAs, which represented 10.1 % of all DXAs, cost Medicare approximately US$16 million in 2011.

CONCLUSIONS

One out of ten DXAs was administered in a time frame shorter than recommended and at a substantial cost to Medicare. DXA use varied across regions. Short-interval DXA use was responsive to reimbursement changes, suggesting carefully designed policy and payment reform may reduce this care identified by rheumatologists as low value.

Using instrumental variables to estimate a Cox's proportional hazards regression subject to additive confounding

The estimation of treatment effects is one of the primary goals of statistics in medicine. Estimation based on observational studies is subject to confounding. Statistical methods for controlling bias due to confounding include regression adjustment, propensity scores and inverse probability weighted estimators. These methods require that all confounders are recorded in the data. The method of instrumental variables (IVs) can eliminate bias in observational studies even in the absence of information on confounders. We propose a method for integrating IVs within the framework of Cox's proportional hazards model and demonstrate the conditions under which it recovers the causal effect of treatment. The methodology is based on the approximate orthogonality of an instrument with unobserved confounders among those at risk. We derive an estimator as the solution to an estimating equation that resembles the score equation of the partial likelihood in much the same way as the traditional IV estimator resembles the normal equations. To justify this IV estimator for a Cox model we perform simulations to evaluate its operating characteristics. Finally, we apply the estimator to an observational study of the effect of coronary catheterization on survival.

Choosing wisely--the politics and economics of labeling low-value services

More than 40 medical specialties have identified "Choosing Wisely" lists of five overused or low-value services. But these services vary widely in potential impact on care and spending, and specialty societies often name other specialties' services as low value.

Quality of osteoporosis care of older Medicare recipients with fragility fractures: 2006 to 2010

OBJECTIVES

To assess uptake of postfracture care guidelines in community-dwelling Medicare recipients with fractures.

DESIGN

Retrospective observational cohort study.

SETTING

Claims-based study using U.S. Medicare administrative inpatient, outpatient (2003-2010), and prescription (2006-2010) data.

PARTICIPANTS

Individuals aged 68 and older who survived at least 12 months after a fracture of the hip, radius, or humerus.

MEASUREMENTS

Poisson regression modeled factors, including participant characteristics, comorbidities and hospital referral region (HRR), associated with bone density testing or osteoporosis pharmacotherapy in the 6 months after fracture. Models were repeated for participants with no osteoporosis care observed before fracture (attention naïve).

RESULTS

In 61,832 individuals with fractures, mean age was 80.6, 87.0% were female, 88.5% were white, 2.6% were black, and 62.1% were attention naïve at the time of fracture; 21.8% received testing, pharmacotherapy, or both in the 6 months after fracture. In adjusted models, factors associated with significantly lower likelihood of receiving this care were black race, male sex, and an upper extremity fracture (vs hip). In models restricted to attention-naïve participants, the same factors were associated with lower relative risk of receiving care. Adjusted HRR-level care rates ranged from 14.7% to 22.9% (10th to 90th percentile). The proportion receiving care increased from 2006 to 2009.

CONCLUSION

Postfracture osteoporosis care was uncommon, particularly in black and male participants. Care increased over time, but for most, a fracture was insufficient to trigger effective secondary prevention, especially for participants who were not receiving prefracture osteoporosis attention. Clinicians and policy-makers must consider effective remedies to this persistent care gap.