Luisa Franzini, PhD

  To investigate the roles of prices, poverty, and health in divergences between Medicare and private spending in Texas.

Retrospective observational design using 2011 Blue Cross Blue Shield of Texas claims data and publicly available Medicare data.

 We measured market-level spending per enrollee among the privately insured. Variation in Medicare and private spending per person are decomposed into prices and quantities, and their associations with poverty are measured. Markets are divided into 4 groups and are compared based on the ratio of Medicare to private spending: “high-private,” “proportional,” “high-Medicare,” and “extremely high-Medicare.”

 Among the privately insured, poverty appears to have large spillover effects; it is strongly associated with lower prices, quantities, and spending. Among Medicare beneficiaries, health status is a key driver of spending variation. The 2 markets with extremely high Medicare-to-private spending ratios (Harlingen and McAllen) are predominantly Hispanic communities with markedly higher rates of poverty and lack of insurance and also extremely low physician supply. The markets with relatively high private spending stand out for having good health-system performance and health outcomes, and higher than average hospital prices.

: Variation in private spending appears to reflect the ability of the local population to pay for healthcare, whereas variation in Medicare is more heavily driven by health status, and presumably, by clinical need. These findings highlight the inadvisability of using Medicare spending as a proxy for systemwide spending, and the need for comprehensive market-level spending data that allow comparisons among populations with different sources of insurance coverage.

In general, our approach begins with the premise that Medicare and private spending patterns diverge, and sets out to explore factors that might account for that divergence.

Markets with relatively high Medicare-to-private spending ratios (Harlingen and McAllen, Texas) have markedly higher rates of poverty and uninsurance and extremely low physician supply.

The markets with relatively high private spending stand out for having good health-system performance and health outcomes, and higher than average hospital prices.

In the privately insured population, there are inhibitory spillover effects of regional poverty on health spending.

Health status explains a larger portion of the Medicare spending variation.

Researchers have studied geographic variation in healthcare spending and practice patterns in the United States for at least 4 decades.

 The centerpiece of this work is the Dartmouth Atlas, a collection of studies based mainly on analyses of claims data from the US Medicare fee-for-service program.

 The Dartmouth researchers’ findings led to the claim that waste accounts for 30% of health spending in the United States.

 Furthermore, Fisher and colleagues examined the relationship between spending, quality, and outcomes, and reported that “residence in higher-spending regions does not cause improved quality, access to care, or survival (and may result in worse survival).”

Medicare claims data lends well to geographic variation analyses, as the program covers individuals nationwide, and claims data are stored in a format that is uniform and relatively easy to work with. However, newer sources of claims data are beginning to be used that include other populations and allow for broader analyses of geographic variation. Most notably, the Institute of Medicine (IOM) sponsored a very ambitious collection of studies that examine geographic variation in healthcare prices, spending, and quality among both Medicare enrollees and the privately insured.

From the existing studies of private claims data, it is clear that Medicare and private spending do not follow the same patterns. Chernew and colleagues, for example, reported that Hospital Referral Region (HRR)-level Medicare spending per beneficiary is negatively correlated with spending per enrollee among the privately insured.

 More recently, Newhouse and Garber, in a summary of the findings from the IOM Committee, noted that “There is almost no correlation between Medicare spending and commercial spending in an area.”

The goals of our study are two-fold: the first goal is to describe the relationship between HRR-level Medicare spending per enrollee and private spending per enrollee. The second goal is to identify the differences between Medicare and private spending in the patterns of variation and, more specifically, the roles played by prices, quantities, poverty, and health status. In general, our approach begins with the premise that Medicare and private spending patterns will diverge, and sets out to explore factors that might account for that divergence.

Our general approach is: 1) to measure spending per enrollee for each market in Texas among Medicare beneficiaries and the privately insured; 2) to decompose Medicare and private spending variation into prices and quantities, and examine their associations with market-level poverty rates; and 3) to compare demographic and health-system characteristics of markets marked by relatively high Medicare spending with those marked by relatively high private spending. Individuals enrolled in employer-sponsored private coverage are rarely in poverty. Based on the Medical Expenditure Panel Survey, among the nonelderly privately insured population, only approximately 5% were in poverty in 2011 (8.8 million out of 179.5 million). Therefore, any associations between market-level poverty rates and private health spending presumably reflect community-level spillover effects. Such a spillover can occur if high concentrations of poverty or lack of insurance reduce what Pauly and Pagan (2012) refer to as “community aggregate demand” for healthcare.

 Because of such spillovers, privately insured individuals in high-poverty markets may end up using less care, or care of a lower quality, even though they themselves have the financial resources to access better care. The economic circumstances of Medicare beneficiaries vary much more widely than those of the privately insured. Therefore, any associations between market-level poverty and Medicare spending likely reflect a mix of direct effects (ie, the effects of poverty among Medicare beneficiaries) and community-level spillover effects, similar to those in the private sector.

This research was approved by a University of Texas Health Sciences Center, Houston, Institutional Review Board. This study used the enrollment and claims files from Blue Cross Blue Shield of Texas (BCBSTX) to measure spending per enrollee in 2011 for each of the 22 HRRs in Texas. HRRs are collections of zip codes that are defined by the Dartmouth Atlas and that represent markets for tertiary hospital care.

 We included all nonelderly enrollees in BCBSTX preferred provider organization (PPO) products, which account for 90% of BCBSTX enrollees. Our final population consisted of 3,484,693 BCBSTX members with valid zip codes, residing in Texas, aged 0 to 64 years, and enrolled in a PPO plan during 2011. For Medicare, we used publicly available 2011 HRR-level spending data produced by CMS

; these are for Medicare fee-for-service beneficiaries 65 years and older, and we included all Medicare spending except for part D drugs.

For each HRR, we calculated 2 spending indexes: one for BCBSTX and one for Medicare. Spending in the BCBSTX data includes all facility and professional claims, and represents the allowed amount, which includes any deductibles, co-pays, and coinsurance paid out of pocket by the patient, as well as payments made by BCBSTX to providers. Spending for Medicare represents only the amount paid by Medicare, and the spending indexes are separated into 3 components: an input prices index, an adjusted price index (adjusted for input prices), and a quantity index. Equal spending per person is divided by the Texas average ($3336 for BCBSTX and $10,165 for Medicare). The input prices index reflects the prices of labor, space, etc, relative to the Texas average. The adjusted price index reflects differences in the prices paid to medical providers above and beyond differences in input prices. The quantity index reflects differences in the number of services used per person. By definition, the overall spending index in each HRR equals the product of that HRR’s input prices index, its adjusted price index, and its quantity index.

The input prices index was calculated for each HRR using Medicare’s hospital wage index for facility claims, and Medicare’s geographic practice cost index for professional claims. To calculate the BCBSTX price index, we used indirect standardization, which involved using all the BCBSTX data to first calculate an average price for each specific type of service within each service category: hospital inpatient, hospital outpatient, and professional services. For this calculation of average prices, facility inpatient services were categorized based on diagnosis-related groups for each admission, while facility outpatient services were categorized based on Current Procedural Terminology (CPT) or Healthcare Common Procedure Coding System (HCPCS) codes and revenue codes. Professional claims were also categorized using the CPT or HCPCS codes. The price index for each service category was computed by comparing the actual allowed amount with the hypothetical allowed amount calculated using the Texas average price for each specific type of service. An overall price index was then calculated for each HRR by summing actual and hypothetical allowed amounts across all service categories and calculating the ratio of actual total spending divided by hypothetical total spending.

 Each HRR’s BCBSTX adjusted price index equals its price index divided by the input prices index. The BCBSTX quantity index is defined as the spending index divided by the price index.

To calculate Medicare price indices, we divided Medicare actual costs by Medicare standardized costs—both of which are reported in the publicly available Medicare data.

 CMS calculates standardized Medicare costs by applying uniform national prices for each type of service, similar to the approach we used to calculate hypothetical BCBSTX spending using Texas average prices. The Medicare quantity index is defined as the Medicare spending index divided by the Medicare price index.

We use health status scores to summarize the health status and expected healthcare utilization of the populations in different markets. The Medicare health status scores are based on mean hierarchical condition categories (HCCs) scores, as reported in the publicly available data. The BCBSTX risk scores are also HCCs, and are calculated from the BCBSTX enrollment and claims files. The Medicare and BCBSTX risk scores reflect age, sex, and diagnoses reported on claims. Health status-adjusted quantity indexes were calculated, for Medicare and BCBSTX, by dividing the quantity index by the health status index.

Variation in spending was allocated to 4 factors—input prices, adjusted prices, health status, and adjusted quantities&mdash;using a 3-step decomposition of variance approach. In the first step, the variation in total spending was allocated among service categories (ie, hospital inpatient, hospital outpatient, and professional services) using a population-weighted variance-covariance matrix. The second step involved decomposing variation within each service category into the 4 factors. In the third step, the variation attributable to each of the 4 factors was summed across service categories.

Poverty clearly has an important and well-documented impact on individual- and community-level health status, and it is also strongly related to input prices in a local market. Our analysis focused on whether poverty also has an association with the 2 “residual” factors: adjusted prices (adjusted for input prices) and adjusted quantities (adjusted for health status). In particular, we wanted to contrast the differing roles of poverty in Medicare versus BCBSTX. To do this, we performed 4 univariate regressions—poverty rates on Medicare prices, on Medicare quantities, on BCBSTX prices, and on BCBSTX quantities&mdash;and measured the 

 and the sign of the estimated coefficients. These univariate regressions allowed us to further decompose the variation in the residual factors into a “poverty effect” (which is in quotation marks because we are not attributing causality) and an “unexplained” factor.

We also performed a descriptive analysis in which we sorted the 22 Texas HRRs into 4 groups based on a ratio of the Medicare spending index to the private spending index. Harlingen and McAllen, both of which have a Medicare-private ratio exceeding 1.2, are placed in their own group (“extremely high-Medicare”). The remaining markets are defined as “high-Medicare” (the Medicare-private ratio exceeds 1.05), “proportional” (the Medicare-private ratio is between 0.95 and 1.05), and “high-private” (the Medicare-private ratio is less than 0.95). We focus our discussion on the markets at the 2 extremes—“extremely high-Medicare” and “high-private.”

We draw on several additional data sources to measure market-level demographics and health system characteristics. These sources include the 2012 Area Health Resources File,

 and the “Prevention and Treatment” dimension score from The Commonwealth Fund Scorecard on Local Health System Performance.

 We created HRR-level measures from county-level data using a population-weighted crosswalk from zip code to county.

Both Medicare and BCBSTX spending vary widely across Texas HRRs (

). The Medicare spending index ranges from 0.84 in San Angelo to 1.27 in McAllen, while the BCBSTX spending index ranges from 0.73 in Harlingen to 1.13 in Beaumont. The correlation between HRR-level Medicare and BCBSTX spending is negative (correlation coefficient = —0.21; 

 = .36), although that relationship is entirely due to 2 markets—Harlingen and McAllen&mdash;that have extremely high Medicare spending and extremely low private spending. If those 2 markets are excluded, the correlation between Medicare and private spending becomes positive (

We find that price differences explain a relatively large share of the variation in BCBSTX spending (33%) compared with Medicare (17%) (

). The IOM found a much larger discrepancy, and reported that prices accounted for 70% of the variation in private spending.

 The IOM study may have picked up much greater variation in prices due to its inclusion of multiple insurers, and markets spanning the country.

In Medicare, price differences are almost entirely explained by differences in input prices. Poverty rates are associated with slight increases in Medicare prices due to explicit adjustments in Medicare payment policy, such as wage index floors and “disproportionate share hospital” payments. (Regressing adjusted Medicare prices on poverty produces a point estimate of 0.00375 (standard error [SE] = 0.00083; 

 = 0.5081). This indicates that a 1 percentage-point increase in poverty is associated with a 0.4% increase in Medicare prices. In contrast, few of the differences in BCBSTX prices can be explained by differences in input prices, and poverty is associated with lower prices. Regressing adjusted BCBSTX prices on poverty produces a point estimate of —0.00375 (SE = 0.00239

 = 0.1100). This indicates that a 1 percentage-point increase in poverty is associated with a 0.4% reduction in private prices. In Medicare, differences in health status play a significant role, explaining 32% of total spending variation. In sharp contrast, differences in health status explain almost none of the spending variation in BCBSTX.

Market-level poverty rates are associated with reductions in the quantities of services received in Medicare and BCBSTX (Figure 2). Regressing adjusted Medicare quantities on poverty produces a point estimate of —0.00542 (SE = 0.00256; 

 = 0.1829). This indicates that a 1 percentage-point increase in poverty is associated with a 0.5% reduction in Medicare quantities. Regressing adjusted BCBSTX quantities on poverty produces a point estimate of —0.00851 (SE = 0.00207; 

 = 0.4577), which indicates that a 1 percentage-point increase in poverty is associated with a 0.9% reduction in BCBSTX quantities. However, the inhibitory role of poverty on quantities is much larger among the privately insured, explaining 30% of overall spending variation. Both Medicare and BCBSTX exhibit large unexplained spending variation, although the unexplained variation in Medicare is essentially all due to quantities whereas in BCBSTX there is unexplained variation in both prices and quantities.

McAllen has garnered an exceptional amount of attention in health policy circles.

 The best known article on McAllen, by Atul Gawande, cited Dartmouth’s Medicare data and attributed the high spending in McAllen to overuse—“patients in McAllen got more of pretty much everything”&mdash;and an ethic of profit maximization in the medical community. Comparing Medicare and private spending data provides a much richer, more nuanced picture of McAllen, and broader patterns in Texas. After Gawande’s article was published, Franzini and colleagues, using BCBSTX data, pointed out that: 1) spending among the privately insured in McAllen was actually a bit lower than in El Paso, McAllen’s allegedly more parsimonious cousin, and 2) McAllen’s demographics are distinctive, and do not simply mirror El Paso’s.

McAllen and Harlingen, which lie side by side along Texas’ border with Mexico, are emblematic of the divergent forces driving Medicare and private spending patterns. Those markets stand out in our data on 3 counts: their Medicare spending is the highest in the state, their private spending is the lowest in the state, and their demographic and health system profiles are exceptional (see 

). The higher Medicare spending is due mainly to higher quantities of services, but Medicare prices are also about 6% higher than the Texas average. The higher Medicare prices in McAllen and Harlingen are due entirely to high inpatient hospital prices (not shown in Table 1). The hospital price boosts in Medicare reflect the fact that Medicare pays higher prices to hospitals that treat low-income patient populations. The exceptionally low private spending is due both to prices and quantities that are lower than the Texas average.

Beyond health spending, profound differences divide McAllen and Harlingen from the rest of Texas. The populations of McAllen and Harlingen are 90% Hispanic, versus less than 40% in the rest of the state. The poverty rate is also twice as high as the rest of the state, and more than half of Medicare beneficiaries are “dual eligibles.” The uninsured, as a share of the population of those aged under 65 years, is almost 40% in McAllen and Harlingen, compared with just over 25% in the rest of the state. McAllen and Harlingen also have an extremely low physician supply and the worst health system performance in the state. And while individuals self-report that they are in considerably worse than average health, mortality rates are low—this is an example of a broader phenomenon known as the “Hispanic paradox.”

Excluding McAllen and Harlingen, the rest of the Texas markets are more uniform in their Medicare and private spending, and in their demographics. But there are still some differences worth noting. Ten markets, including Austin and neighboring Temple, have private spending that is relatively high compared with Medicare spending. Those markets tend to have Medicare spending that is lower than the rest of Texas—due to low quantities&mdash;and private spending that is relatively high&mdash;due to high prices.

The markets with relatively high private spending are socioeconomically similar to the state average, but differ from the rest of Texas on other dimensions. These markets have better than average rates of premature mortality, and adults 50 years and older are more likely to receive screening and recommended preventive care. These markets also tend to score well in The Commonwealth Fund’s “Prevention and Treatment” dimension, which combines measures of access, receipt of recommended care, and risk-adjusted outcomes.

Is Higher Spending Associated With Worse Outcomes?

Several researchers have pointed out that regions with higher Medicare spending tend to have both worse quality of care and outcomes

; however, that finding appears to apply only to Medicare spending, not private spending. In our analysis, areas with the highest private spending (the “high-private” markets) tended to have relatively low rates of hospital readmissions, higher rates of adults receiving recommended services, and better than average health outcomes. This undermines the notion that higher spending somehow leads to poor quality and outcomes, and raises the possibility that the negative relationship between Medicare spending and quality is an artifact of broader socioeconomic patterns.

We found that Medicare and private spending are negatively correlated across Texas HRRs, consistent with the finding of Chernew and colleagues.

 That negative correlation is due to 2 outlier HRRs—McAllen and Harlingen&mdash;that stand out both for their extremely high Medicare spending and their extremely low private spending. We examined those 2 markets in some detail, and found that they are unusual in having extremely high poverty rates and a very limited supply of physicians. Those HRRs could be dismissed as mere flukes, but they may also highlight a broader phenomenon: higher rates of poverty in a market are associated with lower unit prices and lower quantities of care among the privately insured.

Attributing McAllen’s unusually high Medicare spending to a “culture of money” almost certainly oversimplifies the situation. If McAllen and Harlingen were beset by profit-driven medical providers, private spending in those markets would likely be high as well; but in fact, private spending in those markets is remarkably low. Market-level poverty rates appear to play a key role, although further quantitative and qualitative analysis would be needed to unearth the tangled links between McAllen and Harlingen’s unique market features and its Medicare and private spending patterns.

Cooper has argued that health system performance and health outcomes are heavily impacted by what he terms the “affluence-poverty” nexus.

 That nexus may, for reasons that are not fully understood, drive Medicare and private spending in opposite directions.

 Cooper points out that private spending must be financed locally, so the level of spending may primarily reflect the local employed population’s affluence and ability to pay for care. These inhibitory spillover effects of regional poverty on health spending among the privately insured are consistent with our finding that poverty rates are associated with significantly lower private prices, quantities, and spending. Medicare spending, in contrast, is financed mainly by the federal and state governments, so the level of spending depends less on local affluence and more on health status and the incidence of disease. This notion is consistent with our findings that after adjusting for health risk, poverty had a lower impact on spending, prices, and quantities in the Medicare population compared with the privately insured population. Health status, on the other hand, explained a larger portion of the Medicare spending variation.

Wide variation in Medicare spending has been cited in policy debates as indicative of inefficiency in the healthcare system. Our analysis, consistent with the bulk of earlier research, finds significant variation in Medicare spending due to differences in the quantities of services provided. But our analysis also indicates that unexplained spending variation occurs among the privately insured, due to differences in both prices and quantities. This suggests that the policy-making discussion should be broadened to include inefficiencies in pricing by private health plans as well as inefficiencies in practice patterns. Our findings also suggest that the Medicare spending patterns in a market should not be treated as a “biopsy” of the entire healthcare system, and that poverty and other market factors play out differently in Medicare versus other market segments. Finally, one of the key takeaways from the Dartmouth Atlas—that higher spending is not associated with better quality or outcomes&mdash;may not apply to broader measures of healthcare spending.

Our study has the usual limitations of studies using claims data. Even though BCBSTX is the largest insurer in Texas, which is a large and under-studied state with wide diversity in its sociodemographics and composition of healthcare markets, our data represent the population insured by 1 carrier in 1 state. Additional limitations of this study are due to the limitations of BCBSTX and Medicare data, such as the inability to account for Medicaid contribution to healthcare spending on dual eligibles, the inability to control for Hispanic ethnicity in the BCBSTX data, and the inability to control for immigration status effects on enrollment rates in Medicare. Also, the different age ranges (younger than 65 years for BCBSTX and 65 years and older for Medicare) limit the comparability of the populations, but in a sub-analysis using the BCBSTX members aged 50 to 64 years (not shown), the findings were substantially unchanged. Despite these limitations, this study has the opportunity to provide important information on private-sector spending and its relation to Medicare spending.

Articles

Variation in private spending reflects the ability of the local population to pay for healthcare, whereas variation in Medicare is more driven by health status.

Roles of Prices, Poverty, and Health in Medicare and Private Spending in Texas

To explore the feasibility of using a distributed data model for ongoing reporting of local healthcare spending, specifically to investigate the contribution of utilization and pricing to geographic variation and trends in reimbursements for commercially insured beneficiaries younger than 65 years.

Commercial claims were obtained for beneficiaries in 5 states for the years 2008 to 2010 using a distributed data model. Claims were aggregated to the hospital service area (HSA) level and healthcare utilization was quantified using a novel, National Quality Forum—endorsed measure that is independent of price and allows for the calculation of resource use across all services in standardized units. We examined trends in utilization, prices, and reimbursements over time. To examine geographic variation, we mapped resource use by HSA in the 3 states from which we had data from multiple insurers. We calculated the correlation between commercial and Medicare reimbursements and utilization. Medicare claims were obtained from the Dartmouth Atlas.

We found that much of the recent growth in reimbursements for the commercially insured from 2008 to 2010 was due to increases in prices, particularly for outpatient services. As in the Medicare population, resource use by this population varied by HSA. While overall resource use patterns in the commercially insured did not mirror those among Medicare beneficiaries, we observed a strong correlation in inpatient hospital use.

This research demonstrates the feasibility and value of public reporting of standardized area-level utilization and price data using a distributed data model to understand variation and trends in reimbursements.

This study explored the use of a distributed data model and a novel method for calculating utilization to track local trends in spending for commercially insured beneficiaries.

Recent growth in spending for commercially insured beneficiaries was due principally to increases in prices, rather than increases in utilization.

Commercial utilization and spending varied across local areas and was not highly correlated with Medicare utilization and spending.

A distributed model may allow for nationwide reporting of spending and utilization to track the local effects of healthcare reform efforts.

Timely, local data are important to policy-makers, providers, patients, payers, and employers working to slow the growth of healthcare spending, which is a major focus of federal, state, and local healthcare reform initiatives. Community-based multistakeholder coalitions have formed across the country in an effort to influence their local healthcare markets and reduce costs. More than 40 percent of people in the United States live in a community with a multistakeholder coalition aimed at improving health and healthcare, including collaboratives focused on improving the exchange of health information, accelerating engagement by key local opinion leaders and stakeholders, or promoting quality improvement.

 All of these entities, however, lack the local data needed to determine if their efforts are making a difference.        

The factors contributing to rising healthcare spending differ across communities and depend on local context; understanding the drivers of local spending growth is complicated by the variety of inputs. Provider culture and supply, various market segments (outpatient, inpatient, long-term care), payer mix, regulation, and the competitiveness of hospital and physician markets all affect pricing, utilization, and ultimately, the total cost of care.

Research has shown that the relative contribution of these factors varies across markets and that drivers of commercial spending are not necessarily the same as drivers of Medicare spending. Chernew et al found that commercial spending was not correlated with Medicare spending across hospital referral regions.

 Examining the commercial and Medicare populations for El Paso and McAllen, Texas, Franzini et al found there was 86% greater per capita Medicare spending in McAllen than El Paso, but 7 percent less commercial spending.

 A recent Institute of Medicine report found that regional variation in spending in the commercial insurance market is due in large part to differences in price markups by providers. Differences in utilization, however, still explained over 30% of the regional variation in spending.

These diverse results indicate that the factors that contribute to rising healthcare spending are multifactorial and highlight the need for more comprehensive and detailed analysis of commercial data. The 2013 report issued by the Institute of Medicine, Variation in Health Care Spending: Target Decision Making, Not Geography, was unprecedented in that it combined both Medicare and commercial data from a variety of proprietary sources for analysis. The report underscored the importance of available data on both commercial and Medicare utilization and prices at a local level. These data could be useful to engage local stakeholders (such as employers), identify areas of potential waste, point regulators to localities where monopolistic pricing may be occurring, and evaluate the impact of local and national reform initiatives.

There is substantial uncertainty about the best approach to make local data publicly available. Feasible options might include a single federal database,

 combining state all-payer data (currently available in Colorado, Kansas, Maine, Maryland, Massachusetts, Minnesota, New Hampshire, Tennessee, Utah, and Vermont),6 using private data aggregators, or relying on a distributed data model.

 The Health Care Cost Institute has combined data from 4 large commercial insurers and has published broad reports on trends in employer-sponsored insurance, but has not produced any data at a local level.

 There are some community-level efforts under way to track healthcare spending and utilization locally, most notably in California, but they are relatively rare.

 In spite of the Institute of Medicine’s call for making more and better data available on both Medicare and commercial populations, it is not clear how this might be done.

In this paper we explore the feasibility and value of 2 potential approaches for making commercial spending and utilization data available at a local level. First, we work with an all-payer data set for New Hampshire, Maine, and Vermont, aggregated by Onpoint Health Data, a private, nonprofit organization. Second, we use a distributed data model to aggregate data from a single payer in 2 states. Our methods include a standardized approach to measuring utilization, and our partners submitted utilization and spending reports that were stripped of protected health information. These reports allowed us to aggregate data at a hospital service area (HSA) level and adjust for demographic characteristics. The findings demonstrate the feasibility of each of these approaches and the resulting data highlight the potential utility of tracking local healthcare spending.

  To test the distributed data model, we used data from 3 distinct sources. We obtained data on utilization and reimbursements for commercial beneficiaries in Maine, New Hampshire, and Vermont from an all-payer data set (2008-2010), along with data on beneficiaries of Blue Cross Blue Shield of Michigan (2009-2010) and Texas (2008- 2010) plans. The latter 2 states were chosen because of existing research relationships with relevant data providers. Each of the participating analytic teams applied standard software to measure utilization and submitted reports detailing summary information for beneficiaries in each age and gender category for each HSA. HSAs were defined in previous work by identifying the zip codes where the highest proportion of Medicare beneficiaries received their care from a single hospital.

 Data were then aggregated by the study authors and compared with corresponding HSA data from the Medicare program.

Each participating organization applied a standardized approach for defining the commercial population to be included in the analysis. Beneficiaries were required to be enrolled for a minimum of 9 months during the year, unless the member was born during the calendar year or was older than 65 years by the end of the year.

  Participating health plans and data providers used software developed by HealthPartners, a nonprofit healthcare organization and health plan, to measure a utilization amount associated with each insurance claim. The HealthPartners algorithm quantifies utilization using Total Care Relative Resource Values (TCRRVs), which are the basis for the National Quality Forum (NQF)-endorsed Total Resource Use measure.

 TCRRVs, which are expressed in dollars, measure the utilization and intensity of the services delivered to manage a patient’s healthcare needs. We chose to use the HealthPartners measure because it is independent of price and allows for the weighting and calculation of resource use across all medical services in standardized units. We did not want to obscure the effect of regional price differences on overall reimbursement amounts. In addition, we chose to use the measure because it makes resource use amounts equivalent for services offered across multiple settings. 

TCRRVs are unique in that they are relative within and across components of care (inpatient, outpatient, professional, and pharmacy), which allows for the isolation of resource use not only by component, but also on a total, per capita basis. The TCRRVs are made relative within the components of care using the CMS weighting system, including Medicare Severity Diagnosis Related Groups for inpatient care, Ambulatory Payment Classifications for outpatient care, and Relative Value Units for professional office care, while the TCRRVs for pharmaceutical expenditures are made relative by using the median average wholesale price per day for each National Drug Code. TCRRVs are calculated by multiplying the CMS weight by national average paid amount and the relevant service count. While the Health Care Cost Institute also uses CMS weights to measure the intensity of care, its methodology does not involve reporting (1) a resource use amount based on both utilization and intensity, or (2) a measure of total resource use across service lines.

The TCRRV algorithm was applied at the claim level by each data contributor and aggregated at the HSA level for transmission to study authors. (TCRRV weights are updated annually to correspond to updates in the CMS weight files. We used the 2011 TCRRV national weights.) Resource use was capped at $100,000 for each beneficiary in a given year and all components of care were reduced proportionally for both reimbursements and resource use. Very large resource use amounts greatly skew the mean in populations where the majority of beneficiaries have very low utilization and often represent extreme, unavoidable events. Capping reduced average resource use in 2008-2010 by an average of 11% across all HSAs (9.3% and 13.6% for HSAs in the lowest and highest quintile of resource use, respectively).

Both reimbursement and resource use (ie, utilization) amounts for each component of care (inpatient, outpatient, professional, and pharmacy) were transmitted to the study authors, as were denominator data (age in groups, gender, HSA of residence, number of people, and number of person-years). The HealthPartners software performs quality control checks at the claim level by comparing calculated resource use with the reimbursement amount. If the calculated resource use for a particular claim was outside of normal limits, we imputed resource use using the reimbursement amount multiplied by the ratio between resource use and reimbursement amount for all normal claims from that state, year, and component of care. The combination of numerator and denominator data allowed for adjustment and rate calculations. The transmitted data also included the prevalence of prescription drug and mental health carve-outs for each HSA. The Blue Cross Blue Shield of Michigan data did not include HSAspecific information on carve-out levels, but did include the overall state values by year. We used this information and exploratory models from the other states to impute prescription drug reimbursement and utilization values for Michigan HSAs. The imputation assumed a constant carve-out rate across the state each year.

Age, gender, and carve-out adjustment was performed using linear regression across the 5 states. Weighted averages for each HSA were calculated using the adjusted values for each year and component of care. State averages were created by weighting by the number of beneficiaries in each HSA. HSA average relative prices were defined as the ratio of resource use to the reimbursement amount. We normalized relative prices to 1 across all 5 states. Only HSAs with a sum of more than 1000 commercial beneficiaries over the study period were included. Analysis was also limited to those HSAs that were either in 1 of the 5 target states or overlapped the border of 1 of the 5 target states. These 2 exclusion factors reduced the total number of people per calendar year in the analyses by less than 0.1%. 

To examine geographic variation, we mapped resource use for each HSA in Maine, New Hampshire, and Vermont, reflecting procedures and services offered across all 4 components of care. We divided the HSAs into quintiles to display differences in resource use. Maps were not generated for Michigan or Texas, because data were from a single insurer in each state.

To further validate our methodology by attempting to replicate previous findings by examining the association between Medicare and commercial reimbursements and utilization. Age-, sex-, and race-adjusted reimburse- ment and utilization data at the HSA level for the feefor- service Medicare program were downloaded from the Dartmouth Atlas website.

 Utilization data were adjusted for prices across areas and settings.

 Using these data, we examined the association between calculated resource use and reimbursements for the commercially insured and Medicare (aged ≥65 years) populations by calculating a correlation coefficient and plotting the relationship, weighting the regression and the size of the data point by the sum of the number of commercial and Medicare beneficiaries in the HSA.

We identified 388 HSAs and a total of 16,819,237 beneficiary years for inclusion in our analysis (Table 1). The average number of commercial beneficiaries across 2008- 2010 included in the data set for each HSA ranged from 337 to 296,555. The majority of the beneficiaries in each of the 5 states were female. The age distribution was similar across the states, with about half of the population under 40 years and half aged 40 to 65 years. The only exception was Texas, where the population was slightly younger (a greater proportion in the 18-to-39-year age group). The age distributions for our study cohort were similar to those reported by the United States Census Bureau for privately insured individuals in each state.

Trends in Commercial Beneficiary Resource Use,

We examined trends in resource use and relative prices to investigate the extent to which changes in those 2 factors affected changes in reimbursements. The top portion of Figure 1 shows the year-to-year resource use trend per beneficiary by component of care and state. Profes- sional services constituted the highest proportion (42.9%) of resource use on average across the 5 states, followed by outpatient services (21.8%), inpatient services (20.0%), and pharmacy (15.1%). Annual resource use ranged from $4479 per commercial beneficiary in Vermont in 2008 to $5324 per beneficiary in Texas in 2009. The bottom portion of the figure shows calculated changes in resource use, prices, and reimbursements from year to year by component of care. Overall increases in reimbursements were greater in 2008- 2009 than 2009-2010. Reimbursements increased each year in each state, with the exception of Maine in 2010. Total resource use went down in each state between 2009 and 2010, mostly due to declines in inpatient (hospital) services. Increases in prices were larger than the decreases in resource use, however, so reimbursements still increased. The changes in reimbursements in each state can also be compared with changes in the consumer price index (CPI) overall and for medical care. The CPI decreased 0.4% between 2008 and 2009 and increased 1.6% between 2009 and 2010. The medical care CPI increased 3.2% between 2008 and 2009 and 3.4% between 2009 and 2010.15 Overall, growth in prices for commercial beneficiaries in our data set exceeded overall inflation in each state and year. In all but 2 cases, our measure of price increases also exceeded medical inflation.

Variation in Commercial Beneficiary Resource Use,

Resource use, relative prices, and reimbursement amounts varied among commercial beneficiaries across HSAs within our 5 states, and across states (Table 2). Resource use among commercial beneficiaries ranged from $3041 to $6280 across HSAs in the 5 states, which compares to a range of $6322 to $15,049 among Medicare beneficiaries. There was less variation in the commercial population (coefficient of variation = 0.10) than the Medicare population (coefficient of variation = 0.17). Within-state variation in resource use was most marked in Michigan. Relative prices ranged from 0.83 to 1.35 across HSAs in the 5 states, and reimbursements ranged from $1938 to $4626. Relative prices under 1 indicate that, for a given HSA, prices were below the average across all HSAs in the 5 states. The correlation between prices and resource use across HSAs was low (P = .03). 

Variation in resource use is further illustrated in Figure 2, which maps the ratio of resource use in each New England HSA to the weighted average of resource use across the region. Those HSAs with higher resource use relative to the regional average are darker in color. The HSAs with the highest utilization in this area were Rochester, New Hampshire; Ellsworth, Maine; and Augusta, Maine. The HSAs with the lowest utilization in this area were Rumford, Maine; Colebrook, New Hampshire; and St. Johnsbury, Vermont.

Overall resource use in the commercial population was not correlated with resource use in the Medicare population at the HSA level across these 5 states (P = .10, Figure 3), but the correlation across HSAs for inpatient resource use was higher (P = .42). A negative correlation was observed for overall commercial and Medicare reimbursements (P = —.36) and a low correlation for inpatient reimbursements (P = .10).

  This research combined data from 2 private payers and an all-payer data set to produce standardized estimates of utilization, pricing, and reimbursements at the local level across 5 states using a distributed data model. We were able to calculate HSA-level trends and variation for the commercially insured population aged <65 years in the select states using a standardized, NQF-endorsed measure of utilization. Our findings revealed marked variation in both relative prices and utilization, consistent with work by Dunn et al and Baker et al, and confirmed that commercial trends do not necessarily mirror trends in the Medicare population, as previously found by Chernew et al and the Institute of Medicine.

The goal of this research is to help communities understand local healthcare spending patterns and to begin to develop tailored solutions to slow growth in healthcare spending. This line of research differs from previous attempts to examine healthcare costs within the commercially insured population in several significant ways. First, it uses a validated, publicly available measure of utilization to calculate and standardize the magnitude of healthcare resource use across various commercially insured populations and service lines (outpatient, inpatient, professional, and pharmacy), taking into account both the frequency and intensity of care. This measure is independent of price and thus can be used to compare healthcare utilization for different health plans in different parts of the country. Second, we conducted our analysis at the level of the HSA, a relatively small unit that describes the geographic area where patients served by local hospitals reside. Finally, by using a distributed data model, this research demonstrates the feasibility of an approach that could be used either by local multistakeholder initiatives or by a national effort to report on local utilization and spending.

The main limitation of this study stems from it being pilot research: the number of payers included in our analn- ysis for each state varied significantly (61 in Maine, 37 in New Hampshire, 35 in Vermont, and 1 each in Michigan and Texas), along with the insurance products offered by those payers, which reduces comparability across states. The use of a distributed data model presents its own set of potential limitations, since no part of the team had access to claim-level data for all of the states. To address this issue, we created quality control reports to ensure that outputs were accurate and standardized across sites. 

Another limitation was our reliance on risk adjustment using solely age and gender indicators, rather than risk adjustment that included other demographic characteristics (income, education, etc) or information on medical conditions. Diagnosis-related risk adjustment has practical implications and may introduce additional biases

 we sought to demonstrate the viability of the distributed data model before compounding the complexity and cost of data aggregation by implementing a risk adjustment methodology. In future research, we plan to compare results using various risk adjustment methods. Lastly, a potential limitation to this study was our use of HSAs as the unit of analysis. HSAs were defined by Medicare utilization patterns in the 1990s and may not fully capture utilization patterns among commercial beneficiaries.

Key factors contributing to the success of our pilot included the generation of quality control reports for data providers and the data aggregator to ensure data accuracy, as well as the use of mock commercial claims to test aggregation algorithms. While it is possible to apply our methodology using other levels of analysis, we continue to believe that HSAs represent the best option as they, unlike counties, reflect actual patterns of healthcare use, and likely provide more actionable information for community- level interventions than hospital referral regions.

Routine reporting of timely data at a local level can provide actionable information to communities hoping to understand the sources of healthcare spending and growth. This information will be useful for many purposes: to engage employers in reducing healthcare costs, encourage accountability for prices and utilization among providers, encourage community-level shared savings programs (such as the Akron Accountable Care Collaborative

), and facilitate research and evaluation of regional and national reforms. In addition, as consolidation in physician and hospital markets increases, these data can support transparency among health systems and aid regulators in monitoring the effects of changes in payment incentives that might drive consolidation, such as accountable care organizations or bundled payment reforms.

This research provides a unique avenue to develop a nationwide, comprehensive commercial data set in which payers create measures of cost and utilization from their own claims using standard practices, with controls for data quality and consistency. In addition, it allows payers to avoid releasing protected health information (in which patients may be identified) or pricing data, which is commercially valuable and possibly subject to antitrust concerns. The goal is for both large and small payers to contribute to an all-inclusive national data set for public reporting. The barriers to this approach are not trivial: developing it requires the cooperation and effort of multiple competitive payers, as well as entrusting the protection of pricing information to a third party.

The authors are indebted to Sarah Kler of The Dartmouth Institute for Health Policy and Clinical Practice for assistance with mapping, as well as Greg Kotzbauer and Andrew Toler of The Dartmouth Institute, Robyn Rontal of Blue Cross Blue Shield of Michigan, and Marianne Udow-Phillips of the Center for Healthcare Research and Transformation for their insights and advice.

In addition, the authors would like to acknowledge the following entities as the original sources for the New Hampshire and Vermont data sets: the New Hampshire Department of Health and Human Services; the New Hampshire Insurance Department; the Vermont Department of Financial Regulation (formerly the Vermont Department of Banking, Insurance, Securities, and Health Care Administration); and the Green Mountain Care Board. The analyses and conclusions drawn from these data do not necessarily represent those of any agency of the state of New Hampshire or state of Vermont.

The Dartmouth Institute for Health Policy and Clinical Practice (CHC, WLS, DJG, ABM, ESF), the Department of Medicine (ESF), and the Department of Community and Family Medicine (ESF), Geisel School of Medicine at Dartmouth, Lebanon, NH; Blue Cross Blue Shield of Michigan, Detroit, (PGA); Center for Healthcare Research and Transformation, Ann Arbor, MI (NB); Onpoint Health Data, Portland, ME (KF, RS); Division of Management, Policy and Community Health, University of Texas School of Public Health, Houston (LF, RP); and HealthPartners, Inc, Bloomington, MN (GK, SK). Funding Source: This work was funded by grant 69088 from the Robert Wood Johnson Foundation.

The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Concept and design (CHC, DJG, ABM, GK, SK, ESF); acquisition of data (CHC, DJG, ABM, PGA, NB, KF, LF, GK, SK, RS, ESF); analysis and interpretation of data (CHC, WLS, DJG, PGA, KF, LF, GK, SK, RP, RS); drafting of the manuscript (CHC, WLS, ABM, GK, SK, ESF); critical revision of the manuscript for important intellectual content (CHC, WLS, DJG, ABM, NB, LF, GK, SK, ESF); statistical analysis (CHC, WLS, GK, SK, RP); provision of study materials or patients (CHC); obtaining funding (CHC, ESF); administrative, technical, or logistic support (CHC, WLS, DJG, PGA, KF, GK, SK, RP, RS); and supervision (CHC, ESF).

Carrie H. Colla, PhD, 35 Centerra Pkwy, Lebanon, NH 03766. E-mail: carrie.h.colla@dartmouth.edu.

1. Where regional health improvement collaboratives are located. http://www.nrhi.org/about-collaboratives/. Published 2012. Accessed August 7, 2013.

2. Chernew ME, Sabik LM, Chandra A, Gibson TB, Newhouse JP. Geographic correlation between large-firm commercial spending and Medicare spending. Am J Manag Care. 2010;16(2):131-138.

3. Franzini L, Mikhail OI, Skinner JS. McAllen And El Paso revisited: Medicare variations not always reflected in the under-sixty-five population. Health Aff (Millwood). Dec 2010;29(12):2302-2309.

4. IOM Committee on Geographic Variation in Health Care Spending and Promotion of High-Value Care. Report: Variation in Health Care Spending: Target Decision Making, Not Geography. Washington, DC: Institute of Medicine of the National Academies; 2013.

5. Chappel A. Multi-Payer Claims Database (MPCD) for Comparative Effectiveness Research. Presentation to the NCVHS Full Committee, June 16, 2011. http://www.ncvhs.hhs.gov/110616p1.pdf. Accessed August 7, 2013.

6. All-Payers Claims Database Council. Interactive State Report Map. APCD Council website; http://www.apcdcouncil.org/state/map/Maryland. Updated 2013. Accessed July 15, 2013.

7. Quality Alliance Steering Committee. Homepage. QASC website. http://www.healthqualityalliance.org/. Updated 2013. Accessed August 6, 2013.

8. Gottlieb DJ, Zhou W, Song Y, Andrews KG, Skinner JS, Sutherland JM. Prices don’t drive regional Medicare spending variations. Health Aff (Millwood). 2010;29(3):537-543. 

9. Herrera CN, Gaynor M, Newman D, Town RJ, Parente ST. Trends underlying employer-sponsored health insurance growth for Americans younger than age sixty-five. Health Aff (Millwood). 2013;32(10):1715-1722.

10. Rusin G. Total Cost of Vare and Value Based P4P. Presentation of Integrated Healthcare Association, March 19, 2012. http://www.ehcca.com/presentations/pfpsummit7/rusin_pc1.pdf. Accessed October 28, 2013.

11. Appendix on the geography of health care in the United States. In: The Dartmouth Atlas of Health Care.. Lebanon, NH: Dartmouth College; 1999. 289-307. http://www.dartmouthatlas.org/downloads/methods/geogappdx.pdf

12. HealthPartners. Total Cost of Care Population-based PMPM Index. National Quality Forum website. http://www.qualityforum.org/QPS/ 1604. Published January 31, 2012. Accessed June 2013.

13. The Dartmouth Atlas Working Group. The Dartmouth Atlas of Health Care. Hanover, NH: The Dartmouth Institute for Health Policy & Clinical Practice; 201;.http://www.dartmouthatlas.org/. Accessed June 2013.

The authors demonstrate the utility of distributed data models for reporting of local trends and variation in utilization, pricing, and spending for commercially insured beneficiaries.

Tracking Spending Among Commercially Insured Beneficiaries Using a Distributed Data Model

Although large differences by race/ethnicity in breast cancer survival are well established, it is unknown whether disparities in nodal surgery utilization explain the racial/ethnic disparities in survival among women with micrometastasis and macrometastasis in sentinel lymph nodes (SLNs).

Women with breast cancer who underwent sentinel lymph node biopsy (SLNB) and who were found to have nodal metastases were identified from the Surveillance, Epidemiology, and End Results database (1998-2005). Outcomes data were examined for patients who underwent SLNB alone versus SLNB with axillary lymph node dissection (ALND).

Proportions of patients receiving SLNB alone or receiving SLNB with a complete ALND were not statistically different among women of different racial/ethnic backgrounds (

 = .8). Patients of African American descent or Hispanic  origin had reduced overall survival, whereas patients of Hispanic origin had reduced diseasespecific survival after adjusting for selected covariates. Adjusting for nodal surgery did not reduce racial/ethnic disparities in overall survival or disease-specific survival.

 The disparities in survival among African American and Hispanic women with breast cancer are not explained by nodal surgery utilization among women with micrometastasis and macrometastasis in SLNs. 

It is unknown whether nodal surgery utilization explains the racial/ethnic disparities in survival among breast cancer patients with micrometastasis and macrometastasis in sentinel lymph nodes (SLNs). We found that among Medicare beneficiaries, disparities in survival are not explained by nodal surgery utilization among women with micrometastasis and macrometastasis in SLNs.

		Race/ethnicity had no effect on the nodal surgery utilization among Medicare beneficiaries with node-positive breast cancer.

		Adjusting for nodal surgery utilization did not reduce racial/ethnic disparities in breast cancer outcomes for Medicare beneficiaries of Hispanic (disease-specific survival and overall survival) and African American (overall survival) heritage.

Involvement of axillary lymph nodes is the most important prognostic factor in women with invasive breast carcinoma.

 Sentinel lymph node biopsy (SLNB), a rapidly adopted and less morbid alternative to axillary lymph node dissection (ALND) for nodal  staging, has been used since the late 1990s. Based on SLNB, patients are grouped into those with regional nodal metastasis who are likely to benefit from ALND and those without nodal metastases who are unlikely to benefit from ALND.

 Compared with SLNB alone, SLNB with complete ALND was not associated with greatly improved survival for breast cancer (BC) patients with micrometastasis

Recent studies reported that patient race/ethnicity affects receipt and/or quality of sentinel lymph node dissection (SLND).

Variation in compliance with a sentinel lymph node dissection quality measure was supported by a few additional studies. For example, being African American was associated with lower odds of receiving initial SLNB.

 African American or Hispanic women were less likely to undergo initial SLNB with complete ALND,

 and disparities in receipt of SLNB persisted for African American and Hispanic women with BC in a study that looked at temporal differences in use of SLNB from 1998 to 2005.

 Although large differences by race/ethnicity in BC survival are well established,

 it is unknown whether disparities in nodal surgery utilization explain the racial/ethnic disparities in survival among women with micrometastasis and macrometastasis in sentinel lymph nodes.

Because the majority of BC diagnoses occur after age 65 years and because increased life expectancy and advances in treatment mean that more women will be BC survivors,

 we report the impact of nodal surgery utilization on survival among white, black, Hispanic, and Asian women in a large population-based cohort of women with BC diagnosed from 1998 through 2005 who were enrolled in fee-for-service Medicare plans A and B.

We used the Surveillance, Epidemiology, and End Results (SEER) database of the National Cancer Institute to identify BC patients who underwent SLNB alone or SLNB with a complete ALND between 1998 and 2005. As previously described,

 patients were categorized according to their primary surgical procedure (breast-conservation surgery or mastectomy) and according to their axillary lymph node assessment as follows: no nodal evaluation, ALND only, SLNB only (without complete ALND), and SLNB with complete ALND. If a patient underwent SLNB and a modified radical mastectomy, the patient was categorized as having undergone complete ALND. Patients were excluded if they underwent ALND only (without SLNB), did not undergo a lymph node evaluation or if evaluation status was not specified in SEER data, or did not undergo primary surgery. Patients were also excluded if they had stage IV disease. We used the American Joint Committee on Cancer pathologic N category subclassifications to categorize patients according to the degree of nodal metastatic disease as follows: macrometastatic disease (>2.0 mm) and micrometastatic disease (>0.2-2.0 mm). Patients reported only as N1 were classified as having macrometastatic nodal disease as described earlier by Yi and colleagues.

The final sample included 24,961 patients.

We used SEER categories of non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic Asian/Pacific Islander. Patients who underwent surgical resection of the primary breast tumor were categorized as having has mastectomy or breast conservation therapy. Beginning in 1998, SEER started recording information about the type of lymph node assessment. The SEER variable “scope of regional lymph node surgery” was used to capture the patients’ lymph node assessment. Patients were categorized according to their lymph node assessment: no nodal evaluation, ALND only, SLNB without complete ALND (the SLNB-only group), and SLNB with complete ALND (the ALND-group, node-positive women).

 Patients who underwent SLNB and a modified radical mastectomy were considered as having undergone complete ALND. To categorize patients according to their degree of nodal metastatic disease, the American Joint Committee on cancer pathologic N category subclassifications were used as follows: micrometastatic disease (>0.2-2 mm) and macrometastatic disease (>2.0 mm).

Survival was calculated as the number of months between the date of diagnosis and the date of death, the date last known to be alive, or December 31, 2005, whichever occurred first. The day of diagnosis was defined as the 15th of the month, because SEER only reports the month and year of diagnosis. The survival end points for the present study were overall survival (OS) and diseasespecific survival (DSS). Patients who were lost to follow-up or who survived beyond December 31, 2005, were coded as censored observations. Kaplan-Meier survival curves for survival by treatment were parallel per visual inspection as well as by log-rank test supporting the proportionality assumption.There were no significant interactions among the variables.

 testing to compare the differences in categorical variables and proportions between patients who underwent SLNB alone and patients who underwent SLNB with complete ALND. In preliminary analyses, we compared the outcomes of patients  undergoing SLNB alone with those of patients undergoing SLNB with complete ALND. We also performed univariate analyses to determine the influence of patient, tumor, and treatment factors with known or potential prognostic value on OS and DSS as determined by the Kaplan-Meier method. Variables subjected to univariate analysis included age (continuous); socioeconomic  status; comorbidity index

; tumor grade (low/intermediate vs high); tumor size (T1-T3); estrogen receptor and progesterone receptor status (positive, negative, unknown); use of chemotherapy; use of radiotherapy after surgery (yes vs no); year of diagnosis(split, before 2000 and after 2000); and use of ALND. Significant factors from the univariate analysis were included in a multivariate Cox proportional hazard model to identify significant predictors of OS and DSS. In the unadjusted model (model 1) race is the only predictor.The covariates age and tumor size were analyzed as continuous variables in the multivariate models. Covariates added in our regression models included patient and tumor characteristics (age, year of diagnosis, treatment [chemotherapy, surgery, radiation], tumor stage, tumor size, tumor grade and number of positive lymph nodes, diameter of positive lymph nodes, comorbidities, socioeconomic status) (model 2); and receipt of nodal surgery (model 3). Hazard ratios (HRs) and 95% confidence intervals (CIs) were obtained for all regressions. Analyses were performed using SAS release 9.2 (SAS Institute Inc, Cary, North Carolina). All tests were 2-tailed with statistical significance set at 

Of the 24,961 women in the SEER database who underwent SLNB as part of their surgical treatment for BC from 1998 to 2005, 5364 (21%) had nodal metastases and comprised the cohort we analyzed. Of the patients in the cohort, 1028 (19%) underwent SLNB alone and 4336 (81%) underwent SLNB with a complete ALND. Most patients (n = 4098) had macrometastasis, and 1831 patients had micrometastasis.

 = .8) in a cohort of BC patients restricted to those with nodal metastases. A higher proportion of patients underwent SLNB alone if they were diagnosed after 2000 (81.7% vs 18.3%). Patients who underwent SLNB alone rather than SLNB with complete ALND were more likely to have smaller (median tumor size, 15 mm vs 20 mm) or low-grade tumors (69.2% low/intermediate vs 23.6% high grade) (

). Most patients (84.2%) who underwent SLNB alone had breast-conservation surgery. The median number of lymph nodes removed during surgery was 2 in the patients who underwent SLNB alone and 12 in those who underwent SLNB with complete ALND (

 <.0001). The mean number of lymph nodes removed was 3.9 (range, 1-32 nodes) in the SLNB-alone group and 13 (range, 1-52 nodes) in the SLNB plus ALND group.

Of the patients undergoing SLNB alone, 52.4% had micrometastases, compared with only 22.1% of the patients undergoing SLNB with complete ALND. 

shows a multivariate analysis for factors associated with undergoing SLNB alone. Patients were more likely to undergo SLNB alone compared with SLNB with complete ALND if they had smaller or low-grade tumors, had  micrometastases, had positive estrogen receptor status, and were undergoing segmental mastectomy. Of the study population, 551 patients (10.5%) had died of any cause, and 297 patients (5.5%) had died of BC. Overall survival was not greatly different for patients undergoing SLNB alone compared with those undergoing SLNB with complete ALND in the entire cohort and in patients with micrometastases or macrometastases. 

 shows the clinical and pathologic factors affecting OS and DSS.

Patients of African American descent or of Hispanic origin had reduced OS after adjusting for selected covariates. Older women or those with macrometastases, highgrade tumors, larger tumors, negative estrogen receptor status, or more positive lymph nodes found during surgery had reduced OS. Adjusting for nodal surgery did not reduce racial/ethnic disparities in OS. Patients of Hispanic origin who were older or who had high-grade tumors, larger tumors, or negative estrogen receptor or progesterone receptor status; who underwent complete ALND after SLNB; or who had more positive lymph nodes found during surgery had reduced DSS. Disease-specific survival was decreased in patients with macrometastasis (not statistically significant). Adjusting for nodal surgery did not reduce racial/ethnic disparities in DSS.

 recently examined differences in survival for patients with nodal disease undergoing SLNB alone versus SLNB with complete ALND. Similar to their results, we found no significant differences in utilization of SLNB alone or SLNB with a complete ALND between Caucasian and African American older BC patients in a cohort restricted to those with micrometastasis and macrometastasis in sentinel lymph nodes. In our study, this observation also held true for Hispanics and Asians/Pacific Islanders. Our study is unique in that it is one of the first reports of health outcomes in older BC patients with nodal disease for Hispanics, Asians, and Pacific Islanders. Previously mentioned studies did observe racial/ethnic disparities in utilization or quality of nodal surgery.

This difference probably reflects the fact that their purpose was not to address BC survival with receipt of nodal surgery; therefore, they did not restrict their study population to only those with nodal disease. Similar to the findings reported by Bilimoria and colleagues

our study suggests that patients are more likely to receive SLNB alone if they undergo breast-conservation therapy, are older, and have smaller primary tumors.

We observed that adjusting for receipt of nodal surgery did not reduce racial/ethnic disparities. Being of African American descent was associated with reduced OS, while being of Hispanic origin was associated with reduced DSS and OS compared with being Caucasian. This finding is consistent with the findings of Ooi and colleagues,

 who used SEER data to study women diagnosed with invasive breast carcinoma between 2000 and 2006. Ooi and colleagues showed elevated risk of BC-specific mortality among Hispanic whites and blacks that persisted after adjustments for important outcome predictors.

 found a statistically significant hazard of mortality (HR = 1.3; 95% CI, 1.1-1.6) for women aged 18-99 years undergoing a complete ALND compared with those undergoing SLNB only. They attributed this difference to more advanced disease in patients undergoing SLNB with complete ALND. In our study of Medicare beneficiaries 65 years and older, we obtained similar results, with the hazard of BC-specific mortality of a slightly higher magnitude (HR = 1.6; 95% CI, 1.1-2.5).

This study has several limitations, as previously described.

 Medicare beneficiaries in the fee-for-service program represent a self-selected population. Our data set does not include beneficiaries enrolled in health maintenance organizations. Hispanic women are more likely to be in Medicare health maintenance organizations (because of the residence area),

 which might lead to selection bias. Both Hispanic and African American women are less likely than white women to have supplemental coverage, and they are more likely to have Medicaid.

 Comparisons of the accuracy of Medicare’s race codes with self-report (using the Medicare Current Beneficiary Survey data) suggested that the primary error is in mistakenly identifying some Asians, Native Americans, and Hispanics as white.

 The impact of the misclassification has not been examined to date. Because patient comorbidity is identified from diagnoses coded on claim forms, BC survivors’ greater interaction with care providers may explain their higher measured  comorbidity. Income, supplemental insurance, and availability of providers all significantly affect access to care and are not included in our models except as grossly approximated by census tract socioeconomic status quartile and urban-rural status.

In summary, this study examined the effect of nodal surgery, an initial part of longitudinal healthcare for older BC patients, on racial disparities in survival. Overall survival is comparable for those who undergo SLNB alone and those who undergo SLNB with complete ALND among women with BC. African American women have lower OS and women of Hispanic origin have a lower DSS and OS compared with their white counterparts. These disparities in survival are not explained by differences in utilization of ALND among women with micrometastasis and macrometastasis in sentinel lymph nodes. 

From the Divisions of Epidemiology, Human Genetics & Environmental Sciences (ZZN, XLD), Management, Policy and Community Health (LF, ROM), and Health Promotion and Behavioral Sciences (SWV), School of Public Health, University of Texas Health Science Center, Houston, TX.

This work was supported by grant KG090010 (A Transdisciplinary Training Program for Public Health Researchers and Practitioners Wanting to Impact Breast Cancer Disparities) from the Susan G. Komen Breast Cancer Foundation; and in part by grant R01-HS018956 from the Agency for Healthcare Research and Quality.

The authors (ZZN, LF, ROM, SWV, XLD) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Concept and design (ZZM, LF, ROM, XLD); acquisition of data (XLD); analysis and interpretation of data (ZZN, ROM, SWV, XLD); drafting of the manuscript (ZZN, LF, XLD); critical revision of the manuscript for important intellectual content (ZZN, LF, ROM, SWV, XLD); statistical analysis (ZZN, XLD); provision of study materials or patients (XLD); obtaining funding (XLD); administrative, technical, or logistic support (XLD); and supervision (LF, XLD).

Zhannat Z. Nurgalieva, MD, PhD, Division of Epidemiology, Human Genetics, and Environmental Sciences, University of Texas Health Science Center at Houston, 1200 Herman Pressler Dr, RASE631, Houston, TX 77030. E-mail: zhannat.z.nurgalieva@uth.tmc.edu. 1. Rescigno J, Zampell JC, Axelrod D. Patterns of axillary surgical care for breast cancer in the era of sentinel lymph node biopsy. Ann Surg Oncol. 2009;16(3):687-696.

2. Lucci A, McCall LM, Beitsch PD, et al; American College of Surgeons Oncology Group. Surgical complications associated with sentinel lymph node dissection (SLNB) plus axillary lymph node dissection compared with SLNB alone in the American College of Surgeons Oncology Group Trial Z0011. J Clin Oncol. 2007;25(24):3657-3663.

3. Bilimoria KY, Bentrem DJ, Hansen NM, et al. Comparison of sentinel lymph node biopsy alone and completion axillary lymph node dissection for node-positive breast cancer. J Clin Oncol. 2009;27(18): 2946-2953.

4. Yi M, Giordano SH, Meric-Bernstam F, et al. Trends in and outcomes from sentinel lymph node biopsy (SLNB) alone vs. SLNB with axillary lymph node dissection for node-positive breast cancer patients: experience from the SEER database. Ann Surg Oncol. 2010;17(suppl 3):343-351.

5. Olaya W, Wong J, Morgan JW, et al. Factors associated with variance in compliance with a sentinel lymph node dissection quality measure in early-stage breast cancer. Ann Surg Oncol. 2010;17(suppl 3):297-302.

6. Yi M, Meric-Bernstam F, Ross MI, et al. How many sentinel lymph nodes are enough during sentinel lymph node dissection for breast cancer? Cancer. 2008;113(1):30-37.

7. Reeder-Hayes KE, Bainbridge J, Meyer AM, et al. Race and age disparities in receipt of sentinel lymph node biopsy for early-stage breast cancer. Breast Cancer Res Treat. 2011;128(3):863-871.

8. Chen AY, Halpern MT, Schrag NM, Stewart A, Leitch M, Ward E. Disparities and trends in sentinel lymph node biopsy among early-stage breast cancer patients (1998-2005). J Natl Cancer Inst. 2008;100(7):462-474.

9. Ooi SL, Martinez ME, Li CI. Disparities in breast cancer characteristics and outcomes by race/ethnicity. Breast Cancer Res Treat. 2011;127(3):729-738.

The disparities in survival among node-positive breast cancer patients of African American and Hispanic heritage are not explained by nodal surgery utilization.

Luisa Franzini, PhD

Articles

Roles of Prices, Poverty, and Health in Medicare and Private Spending in Texas

Tracking Spending Among Commercially Insured Beneficiaries Using a Distributed Data Model

Utilization of Lymph Node Dissection, Race/Ethnicity, and Breast Cancer Outcomes

Comparing Variation in Medicare and Private Insurance Spending in Texas