Few Disparities in Baseline Laboratory Testing After the Diuretic or Digoxin Initiation by Medicare Fee-For-Service Beneficiaries
Background—Despite the persistence of significant disparities, few evaluations examine disparities in laboratory testing by race/ethnicity, age, sex, Medicaid eligibility, and number of chronic conditions for Medicare fee-for-service beneficiaries’ newly prescribed medications. In Medicare beneficiaries initiating diuretics or digoxin, this study examined disparities in guideline-appropriate baseline laboratory testing and abnormal laboratory values.
Methods and Results—To evaluate guideline-concordant testing for serum creatinine and serum potassium within 180 days before or 14 days after the index prescription fill date, we constructed retrospective cohorts from 10 states of 99 711 beneficiaries who had heart failure or hypertension initiating diuretic in 2011 and 8683 beneficiaries who had heart failure or atrial fibrillation initiating digoxin. Beneficiaries initiating diuretics were less likely to have testing if they were non-Hispanic Black (relative risk [RR], 0.99; 95% confidence interval [CI], 0.98–0.99) than non-Hispanic White. Beneficiaries initiating diuretics and beneficiaries initiating digoxin were more likely to have testing if they had multiple chronic conditions relative to 0 to 1 conditions. Beneficiaries initiating diuretics with laboratory values were more likely to have an abnormal serum creatinine value at baseline if they were non-Hispanic Black (RR, 2.57; 95% CI, 1.91–3.44), other race (RR, 2.11; 95% CI, 1.08–4.10), or male (RR, 2.75; 95% CI, 2.14–3.52) or an abnormal serum potassium value if they were aged ≥76 years (RR, 1.29; 95% CI, 1.09–1.51) or male (RR, 1.17; 95% CI, 1.03–1.33).
Conclusions—Testing rates were consistently high, so there were negligible disparities in guideline-concordant testing of creatinine and potassium after the initiation of digoxin or diuretics by Medicare beneficiaries.
WHAT IS KNOWN
Disparities in a broad range of access and quality measures have not improved in recent years; for Medicare beneficiaries who initiate diuretics or digoxin, laboratory monitoring is important for patient safety to identify contraindications before initiation and serious adverse drug reactions after initiation via periodic testing
WHAT THE STUDY ADDS
Beneficiaries initiating diuretics were more likely to have serum creatinine or serum potassium testing if they had ≥2 chronic conditions but less likely to have testing if they were non-Hispanic Black or other race. Nearly 14% of beneficiaries who received baseline testing had a corresponding laboratory value, but few had abnormal serum creatinine or serum potassium values.
Beneficiaries were more likely to have an abnormal serum creatinine value at baseline if they were non-Hispanic Black, other race, or male; beneficiaries were more likely to have an abnormal serum potassium value if they were aged ≥76 or male but less likely to have an abnormal value if they were non-Hispanic Black.
Beneficiaries initiating digoxin were more likely to have testing for serum creatinine or serum potassium if they were dually enrolled in Medicaid or had ≥2 chronic conditions; just over 12% of beneficiaries who received baseline testing had a corresponding laboratory value, but few had abnormal serum creatinine or serum potassium values.
These analyses demonstrate the potentially important performance assessments that are possible when laboratory results are linked to Medicare claims data.
According to the 2014 National Healthcare Quality and Disparities report, racial/ethnic disparities in a broad range of access and quality measures did not improve substantially compared with previous years.1 Blacks fared worse than Whites in almost half of all access to care measures, whereas Hispanics had worse access to care than Whites for two thirds of access measures. Minority beneficiaries also reported having less access to drug information, needed drugs, customer service, or needed care than White beneficiaries in 2008.2 Reducing disparities in access to drug information and essential medications is important because minorities and other vulnerable beneficiaries had a higher prevalence of ≥6 chronic conditions in 2010.3 Further, black and Hispanic beneficiaries in Medicare Advantage plans were less likely to have adequately controlled hypertension, diabetes mellitus, or dyslipidemia than white enrollees,4 potentially because of lower rates of medication refill adherence.5
For beneficiaries who do access essential medications, laboratory monitoring is important for patient safety to identify contraindications before initiation and serious adverse drug reactions after initiation via periodic testing. Timely laboratory monitoring could have economic implications as well because medications requiring outpatient laboratory monitoring account for nearly half of all unintended drug overdoses that result in emergency room visits,6 including significant emergency room utilization for digoxin toxicity.7 Assessment of laboratory monitoring on initiation of digoxin and diuretics is important to examine because these medications are used to treat Medicare fee-for-service (FFS) beneficiaries’ most prevalent (ie, hypertension and dyslipidemia) and costly (ie, atrial fibrillation and heart failure) conditions.3
The purpose of the study was to examine rates of guideline-appropriate laboratory testing, prevalence of abnormal values at baseline, and disparities in testing and abnormal values for FFS beneficiaries initiating diuretics and beneficiaries initiating digoxin in 2011. We report the prevalence of baseline abnormal values and disparities in abnormal values because these outcomes have never been reported for FFS beneficiaries because of the lack of laboratory results data. We define groups at risk for serious adverse drug reactions (eg, hyperkalemia and renal failure) to include beneficiaries who are racial and ethnic minorities, dually eligible for Medicare and Medicaid, older age, disabled, male, or have multiple chronic conditions.
Data and Participants
The cohort of Medicare FFS beneficiaries in this study was obtained from a convenience sample in 10 eastern United States (NY, NJ, MD, DE, VA, NC, SC, GA, FL, and AL) for a novel data linkage project.8 Medicare Part B claims for laboratory services were linked to outpatient laboratory values processed by a national laboratory vendor in 2011 for beneficiaries residing in these 10 states. Beneficiaries were retained if they were ≥65 years on January 1, 2011, were enrolled in Parts A, B, and D the entire year (2011), and were alive on December 31, 2011.
Using the Medicare Part D events file, we then identified 2 cohorts of beneficiaries: (1) those who initiated a diuretic and had heart failure or hypertension (Figure S1 in the Data Supplement) and (2) those who initiated digoxin and had atrial fibrillation or heart failure (Figure S2 in the Data Supplement). For each medication, a beneficiary was determined to have initiated if they had a drug claim from July 1 to November 30, 2011 and no claims for these medications between January 1 and June 30, 2011. Diuretics were identified in the Part D data by looking for any medications having relevant generic drug names (Table S1 in the Data Supplement). The conditions of interest—heart failure, hypertension, and atrial fibrillation—were determined using the comorbid condition indicators in the Medicare Enrollment Master Beneficiary Summary File.
For analysis of test results, we used data from the subset of these FFS beneficiaries who had serum creatinine and serum potassium values in the laboratory vendor records. There were 13 468 of the 99 711 beneficiaries initiating diuretics who had laboratory results from the large national laboratory vendor. Only 1074 of the 8683 beneficiaries (39 Hispanic, 938 non-Hispanic White, 81 non-Hispanic Black, and 61 others) initiating digoxin had laboratory results, which was an insufficient sample size to examine disparities.
Definition of Baseline Testing and Abnormal Laboratory Values
Consistent with the previous analyses,9,10 we defined guideline-concordant testing to have occurred if there was an outpatient or carrier claim for either a serum creatinine test or serum potassium test within 180 days before the index fill or 14 days after the index fill. We defined indicators of guideline-concordant testing for each test (serum potassium and serum creatinine) and timeframe (180 day before and 14 days after the drug initiation) to examine whether guideline-concordant testing varied by clinical marker or timeframe. Medicare data do not contain claims for inpatient laboratory testing, so beneficiaries who were hospitalized during the study period were credited with appropriate testing, with the exception of patients whose initial fill occurred within 3 days of being discharged from a hospitalization. We assumed that these patients started the medication during hospitalization and were credited with 1 test for both potassium and creatinine in the early follow-up period.
Abnormal baseline values that would suggest the need for intensive monitoring were defined for serum creatinine (>2.5 mg/dL) and serum potassium (>5.0 mmol/L), based on the thresholds defined previously.11
For each cohort, we summarize demographic characteristics and baseline comorbidities using medians and interquartile ranges for continuous variables and frequencies for categorical variables. We defined race/ethnicity (non-Hispanic White, Hispanic, non-Hispanic Black, and other), Medicaid dual eligibility, age groups (65–70, 71–75, and ≥76), eligibility (age versus disability), and sex from the Beneficiary Summary File. Multiple chronic conditions (0–1, 2–3, 4–5, and ≥6) were constructed from the Chronic Condition Warehouse indicators for acquired hypothyroidism, atrial fibrillation, anemia, asthma, benign prostatic hyperplasia, cancer (as a combination of indicators for breast, colorectal, prostate, lung, and endometrial cancer), chronic kidney disease, chronic obstructive pulmonary disease, dementia/Alzheimer’s disease/related conditions, depression, diabetes mellitus, heart failure, hyperlipidemia, hypertension, ischemic heart disease, osteoporosis, rheumatoid arthritis/osteoarthritis, and stroke. The race/ethnicity variables were revised in 2008 to increase sensitivity without loss of specificity, which increased the identification of Hispanic, Asian, and Pacific Islander beneficiaries.12
We present observed rates of baseline laboratory testing for serum potassium and serum creatinine, and the combination of both tests by time period, overall and by disparity group. To understand which Medicare FFS beneficiaries initiating diuretics or digoxin were more likely or less likely to receive guideline-concordant monitoring of both serum potassium and serum creatinine, we used log-binomial regression to estimate the association between the binary testing outcome and age (reference=aged 65–70 years), sex (reference=female), race/ethnicity (reference=non-Hispanic Whites), Medicaid eligibility, eligibility (reference=age eligible), and multiple chronic conditions (reference=0–1 comorbidities). We reported 95% confidence intervals (CIs) and used α=0.05 to establish the statistical significance of tests. All tests were 2-sided.
Among the subset of beneficiaries who had a Part B laboratory claim with a corresponding laboratory value, we examined the proportion with abnormal values at baseline. In the diuretic cohort, we used log-binomial regression to examine disparities in abnormal laboratory values at baseline. The institutional review board of the Duke University Health System approved the study. SAS Enterprise Guide 7.1 (SAS Institute Inc, Cary, NC) was used for all analyses.
Disparities in Guideline-Concordant Testing Related to Diuretic Initiation
There were 99 711 Medicare FFS beneficiaries who initiated diuretics between July 1 and November 30, 2011 (Table 1). The median age was 77 years, 55% were aged 76 or older years, 75.8% were non-Hispanic White, and 34% were male. The vast majority (86.9%) was Medicare eligible because of age, and only 28.8% were dually enrolled in Medicaid. A plurality (47.4%) had ≥6 chronic conditions, 31.3% had 4 to 5 conditions, and 19.2% had 2 to 3 conditions.
The observed rates of baseline laboratory testing in 2011 were 89.6% for serum creatinine testing and 89.3% for serum potassium (Table 2), and most testing occurred in the 6 months before initiation. Unadjusted rates of guideline-concordant testing for both serum creatinine and serum potassium were similar for beneficiaries aged 71 to 75 years (87.8%), beneficiaries aged ≥76 years (89.6%), and beneficiaries aged 65 to 70 years (86.2%) and were similar for female and male beneficiaries (88.5% versus 88.3%). When considering testing rates by race/ethnicity, non-Hispanic Blacks had the lowest unadjusted testing rates (87.6%). Disability-eligible beneficiaries had modestly higher testing rates (89.7% versus 88.2%) than age-eligible beneficiaries, and dually eligible beneficiaries had slightly lower testing rates (87.9% versus 89.5%) than beneficiaries only enrolled in Medicare. Final, unadjusted baseline testing rates were much higher for beneficiaries with ≥6 conditions (93.6% versus 70.0% than beneficiaries with 0 to 1 conditions.
In log-binomial regression of guideline-concordant testing (Table 3), Medicare FFS beneficiaries initiating diuretics were more likely to have serum creatinine or serum potassium testing if they had ≥2 chronic conditions (relative risk [RR]: 1.14, 95% CI: 1.11–1.17 for 2–3 conditions; RR: 1.24, 95% CI: 1.20–1.27 for 4–5 conditions; and RR: 1.32, 95% CI: 1.28–1.36 for ≥6 conditions) when compared with 0 to 1 conditions. Beneficiaries were less likely to have testing if they were non-Hispanic Black (RR, 0.99; 95% CI, 0.98–0.99) or Other race (RR, 0.98; 95% CI, 0.97–0.99) than non-Hispanic Whites.
Abnormal Serum Creatinine and Serum Potassium Values Among Tested Beneficiaries
Among the 99 711 beneficiaries initiating diuretics, there were 13 468 (13.5%) who received a baseline creatinine and potassium testing and had a corresponding laboratory value. Of these beneficiaries, 261 (2.0%) had an abnormal serum creatinine value at baseline (<2.5 mg/dL) and 927 (7.0%) had an abnormal serum potassium (<5 mmol/L) (Figure 1A and 1B). In log-binomial regression (Table S3 in the Data Supplement), beneficiaries were more likely to have an abnormal serum creatinine value at baseline if they were non-Hispanic Black (RR, 2.57; 95% CI, 1.91–3.44), Other race (RR, 2.11; 95% CI, 1.08–4.10), or male (RR, 2.75; 95% CI, 2.14–3.52). Beneficiaries were more likely to have an abnormal serum potassium value if they were aged ≥76 years (RR, 1.29; 95% CI, 1.09–1.51) or male (RR, 1.17; 95% CI, 1.03–1.33). Beneficiaries were less likely to have an abnormal potassium value if they were non-Hispanic Black (RR, 0.64; 95% CI, 0.51–0.81).
Disparities in Guideline-Concordant Testing Related to Digoxin Initiation
There were 8683 Medicare FFS beneficiaries who initiated digoxin between July 1 and November 30, 2011 (Table 1). The median age was 80 years, 69% were aged ≥76 years, 84.1% were non-Hispanic White, and 39% were male. The vast majority (85.8%) was Medicare eligible because of age, and only 28.6% were dually enrolled in Medicaid. A majority (76.3%) had ≥6 chronic conditions, 17.5% had 4 to 5 conditions, only 5.8% had 2 to 3 conditions, and <1% had 0 to 1 conditions.
The observed rates of appropriate laboratory testing were high (95.6% for serum creatinine testing and 95.5% for serum potassium, Table S2 in the Data Supplement), and most testing occurred in the 6 months before initiation. Unadjusted testing rates were similar for beneficiaries aged 71 to 75 years (95.6%), beneficiaries aged ≥76 years (95%), and beneficiaries aged 65 to 70 years (93.4%) and were similar for female and male beneficiaries (95.3% versus 94.3%). Non-Hispanic Whites had the lowest unadjusted testing rates (94.6%). Disability-eligible beneficiaries had similar testing rates (95.7% versus 94.8%) as age-eligible beneficiaries, whereas dually eligible beneficiaries had higher testing rates (97.1% versus 94%) than beneficiaries only enrolled in Medicare. Final, unadjusted baseline testing rates were much higher for beneficiaries with ≥6 conditions than beneficiaries with 0 to 1 condition (97.1% versus 67.6%).
In log-binomial regression of testing (Table 3), FFS beneficiaries initiating digoxin were more likely to have testing for serum creatinine or serum potassium if they were dually enrolled in Medicaid (RR, 1.02; 95% CI, 1.01–1.02) or had ≥2 chronic conditions (RR, 1.35; 95% CI, 1.08–1.69 for 4–5 conditions and RR, 1.44; 95% CI, 1.15–1.80 for ≥6 conditions) when compared with 0 to 1 conditions.
Abnormal Serum Creatinine and Serum Potassium Values Among Tested Beneficiaries
Among the 8683 beneficiaries initiating digoxin, there were 1074 (12.4%) beneficiaries who received baseline testing for serum creatinine and serum potassium and had a corresponding laboratory value. Of these 1074 beneficiaries, 27 (2.5%) had an abnormal serum creatinine value at baseline (<2.5 mg/dL) and 89 (8.4%) had an abnormal serum potassium (<5 mmol/L) at baseline (Figure 2A and 2B).
Given the widespread use of digoxin to manage atrial fibrillation and ongoing debate about the value of digoxin to manage heart failure,13 it was important to examine disparities in serum creatinine and serum potassium testing before initiation in 2011. Overall testing rates and disparities in testing among FFS beneficiaries have not been previously published, and this analysis shows that guideline-concordant rates of baseline testing were consistently high across all subgroups. This result suggests that there are negligible disparities in testing of serum creatinine and serum potassium after the initiation of digoxin or diuretics by Medicare beneficiaries. Ninety percent of FFS beneficiaries in these 10 states received serum creatinine and serum potassium on initiating diuretics in 2011, which was much higher than the 67% testing rate from a study of Medicare Advantage enrollees initiating diuretics in 1999 to 20019 but similar to the 91% testing rate in Medicare Advantage plans in 2011.14 Ninety-six percent of beneficiaries in these 10 states received both tests on initiating digoxin in 2011, which is higher than the 75% testing rate from a study of Medicare Advantage enrollees initiating digoxin in 1999 to 2001.9
Previous evaluation of testing rates in Medicare Advantage plans did not consider disparities by race/ethnicity or other factors that we considered here. The differences by race for beneficiaries initiating diuretics were statistically significant because of the large sample size but not meaningfully different. We also compared testing rates by age groups, sex, age versus disability eligibility for Medicare, dual eligibility, and number of chronic conditions. Multiple chronic conditions has received increased attention in recent years because multimorbid beneficiaries have worse functional status15 than patients with fewer conditions and incur a disproportionate share of inpatient admission rates,5,6 number of prescriptions,16 number of specialist visits,17 and health expenditures.18 Multiple chronic conditions were positively associated with guideline-concordant testing, which is consistent with the previous work that found that quality of care19 and care experiences2 improve with more chronic conditions. This suggests that providers may be ensuring that the beneficiaries at highest risk for adverse events are receiving testing or simply that beneficiaries with more chronic conditions have more visits, creating more opportunities for testing.
Although there were negligible disparities in testing, there were notable disparities in abnormal laboratory values at baseline even though abnormal values for serum creatinine (2.0% for diuretic initiators and 2.5% for digoxin initiators) and serum potassium were rare (7.0% for diuretic initiators and 8.4% for digoxin initiators). Abnormal serum creatinine values were much more common for non-Hispanic Black, Other race, or male beneficiaries; abnormal serum potassium values were more common among the oldest old (aged ≥76 years) or male beneficiaries. We were unable to assess whether physicians responded to these abnormal values because we lacked 2012 data to determine whether physicians changed dosage, switched medication, or discontinued therapy altogether. More effort is needed to reduce testing disparities among Medicare FFS beneficiaries initiating diuretics. In future work, longitudinal data will allow us to conduct these assessments and also address important questions about how often longitudinal laboratory monitoring is conducted after the initiation of diuretics or digoxin. We were also unable to assess disparities in abnormal values because of the limited sample size, so future work with a national sample is needed to explore this issue.
Advances in electronic health record (EHR) functionality that links pharmacy and laboratory ordering systems may reduce disparities observed here because a previous trial of EHR-embedded physician reminders was effective at increasing testing rates.20 Disparities in testing rates may decrease if EHRs include physician reminders to ask patients whether they have followed through with ordered laboratory testing because adverse drug reactions are most likely to arise shortly after the initiation of medications.21 This is critical for ensuring patient safety because a majority of adverse drug events may be prevented by appropriately selecting the medication to be initiated.22 The likelihood that physicians will act on abnormal outpatient test results increases if they use an EHR that includes progress notes and test results.23 Recent work has shown that disparities in annual testing of low-density lipoprotein cholesterol and glycohemoglobin levels have declined in the past decade, such that there are no significant disparities in 2011 among Medicare Advantage enrollees with prevalent dyslipidemia or diabetes mellitus.4 Despite the lack of disparities in testing, black and Hispanic beneficiaries were less likely to have adequate control of blood pressure, cholesterol, or glycohemoglobin.4 Future research with more current laboratory results data should examine if the increased implementation of EHRs following the 2009 Health Information Technology for Economic and Clinical Health (HITECH) Act reduced these remaining disparities. Future research should also examine whether there are disparities in testing and abnormal values for Medicare beneficiaries initiating other chronic medications (eg, metformin, methotrexate, and statins).
These results also have implications of Medicare policy and performance measurement, if the Centers for Medicare and Medicaid Services can obtain laboratory data on most/all reimbursed tests. With such data, Centers for Medicare and Medicaid Services would eventually be able to generate laboratory-based performance measures on Accountable Care Organizations. Public reporting of disparities in laboratory testing can provide accountability and actionable information to providers to actively reduce disparities where identified.24 Interventions to reduce disparities may need to be multifactorial and involve multidisciplinary teams, and the health system must value health equity as a central tenet of its organizational culture of quality improvement.25–27
There are several limitations that must be acknowledged, including that lack of generalizability to Medicare FFS beneficiaries outside of these 10 states or to beneficiaries requiring testing after the initiation of other essential medications. The statistical power to examine disparities in abnormal values was limited because the laboratory results data were available from 1 national laboratory vendor, albeit the vendor with the largest market share in these 10 states.8 Future work with laboratory data from multiple vendors would enable disparities evaluation, which would increase generalizability of results in these 10 states, increase statistical power of comparisons, and enable assessment of disparities in abnormal values among digoxin initiators. Serum creatinine and serum potassium values were only available in this sample if a physician orders the test, a Medicare FFS beneficiary follows through and had the test done, and the test was processed by a clinical facility affiliated with this single laboratory vendor.28
The ability to link laboratory results to Medicare claims represents a unique opportunity to examine disparities in care in a new domain of care. Awareness of the importance of laboratory monitoring for patient safety increased in 2006 when the National Committee for Quality Assurance created a Healthcare Effectiveness Data and Information Set measure, Annual Monitoring for Persistent Medications, for patients on diuretics, digoxin, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and other medications. Rates of guideline-concordant testing for serum creatinine and serum potassium of Medicare FFS beneficiaries initiating digoxin or diuretics were high in 2011, and there were negligible disparities in testing as a result. This analysis begins to fill a significant void in understanding disparities in the medication management of Medicare FFS beneficiaries. With more complete laboratory data, Medicare will be able to assess the quality of care provided to its beneficiaries initiating the essential medications.
We thank Mary Barton and Sarah Scholle at National Committee for Quality Assurance for their helpful comments.
Sources of Funding
The research in this article was supported by the Centers for Medicare and Medicaid Services under Contract Number HHSM-500-2014-00442G with the National Committee for Quality Assurance. This work was also supported by the Office of Research and Development, Health Services Research and Development Service, Department of Veterans Affairs, and Dr Maciejewski was also supported by a Research Career Scientist award from the Department of Veterans Affairs (RCS 10-391).
Dr Maciejewski reported receiving research support from the Department of Veterans Affairs and the Agency for Healthcare Research and Quality and owning Amgen stock because of his spouse’s employment. Dr Curtis reported receiving research support from the Agency for Healthcare Research and Quality, Boston Scientific, Bristol-Myers Squibb, GE Healthcare, GlaxoSmithKline, Johnson & Johnson, Medtronic, Merck, and Novartis. Dr Hammill reported receiving research support from Agency for Healthcare Research and Quality. The other authors report no conflicts.
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veteran Affairs, Duke University, the National Committee for Quality Assurance, the Centers for Medicare and Medicaid Services, or the US Department of Health and Human Services.
The Data Supplement is available at http://circoutcomes.ahajournals.org/lookup/suppl/doi:10.1161/CIRCOUTCOMES.116.003052/-/DC1.
- Received December 16, 2015.
- Accepted September 9, 2016.
- © 2016 American Heart Association, Inc.
- 1.↵2015 National Healthcare Quality and Disparities Report and 5th Anniversary Update on the National Quality Strategy. Rockville, MD: Agency for Healthcare Research and Quality; 2016.
- 3.↵Centers for Medicare and Medicaid Services. Chronic Conditions Among Medicare Beneficiaries. Chartbook, 2012 Ed. Baltimore, MD: 2012.
- See I,
- Shehab N,
- Kegler SR,
- Laskar SR,
- Budnitz DS
- Simon SR,
- Andrade SE,
- Ellis JL,
- Nelson WW,
- Gurwitz JH,
- Lafata JE,
- Davis RL,
- Feldstein A,
- Raebel MA
- Raebel MA,
- Lyons EE,
- Andrade SE,
- Chan KA,
- Chester EA,
- Davis RL,
- Ellis JL,
- Feldstein A,
- Gunter MJ,
- Lafata JE,
- Long CL,
- Magid DJ,
- Selby JV,
- Simon SR,
- Platt R
- Allen LA,
- Shetterly SM,
- Peterson PN,
- Gurwitz JH,
- Smith DH,
- Brand DW,
- Fairclough DL,
- Rumsfeld JS,
- Masoudi FA,
- Magid DJ
- Hernandez AF,
- DeVore AD
- 14.↵National Committee for Healthcare Quality. The State of Healthcare Quality. 2012. http://www.ncqa.org/report-cards/health-plans/state-of-health-care-quality. Accessed March 17, 2016.
- Bayliss EA,
- Ellis JL,
- Steiner JF
- Anderson G
- Starfield B,
- Lemke KW,
- Herbert R,
- Pavlovich WD,
- Anderson G
- Fisher TL,
- Burnet DL,
- Huang ES,
- Chin MH,
- Cagney KA
- Chin MH,
- Walters AE,
- Cook SC,
- Huang ES
- Schneeweiss S,
- Rassen JA,
- Glynn RJ,
- Myers J,
- Daniel GW,
- Singer J,
- Solomon DH,
- Kim S,
- Rothman KJ,
- Liu J,
- Avorn J