The Cost Containment Dividend: What would you have done with an extra $392—or $727– last year? by Nancy Turnbull

As the need to control health care costs gets more attention, various groups have proposed specific cost containment goals. For example, the Connector Board (of which I am a member) voted to push insurers to limit rate increases for the unsubsidized Commonwealth Choice plans to no more than five percent for next year. A number of parties, including Health Care for All and Senate President Murray, have proposed holding rate hearings for health insurance premium increases at or above 7%. The Quality and Cost Council (QCC) has established a goal of reducing the annual rise in health care costs to no more than the unadjusted growth in Gross Domestic Product (GDP) by 2012.

These goals are ambitious and welcome, but theoretical. What would achieving such goals mean for people with private health insurance? I wanted to find out. So, I looked at the financial reports health insurers file with the Division of Insurance for the three largest private health insurers in Massachusetts—Blue Cross Blue Shield (including HMO Blue), Harvard Pilgrim Health Care, and Tufts Health Plan. I looked at reports for 2003-2006 and analyzed how much medical costs and health plan administrative expenses had increased for people with private insurance (i.e., I excluded the Medicare line of business). Since the Division of Insurance doesn’t require health plans to provide information for people covered by self-insured plans, the financial information that’s publicly available for the health insurers includes only people covered by fully-insured plans—about 2 million people for the three health insurers whose data I reviewed. I analyzed medical and administrative costs on a per member basis, to adjust for any enrollment changes at the health plans.

I assessed two different scenarios: First, I figured out how much lower expenses would have been if they had increased at the same rate as the growth in nominal GDP (i.e., growth in GDP not adjusted for inflation), the QCC’s goal. Then I looked at what would have happened if costs had grown at the same pace as the increase in the Consumer Price Index (CPI) for the area that includes Boston, since this is one widely-used measure of the inflation rate consumers are experiencing for other goods and services.

The growth in the two indices was very different over this period: according to the U.S. Bureau of Economic Analysis, the unadjusted growth in GDP was 20.9% from 2003-2006, while data from the Bureau of Labor Statistics show that the CPI for the Boston area increased a total of 9.4%. So I’ll discuss the results of both scenarios.

Here’s what I found: Aggregated across the three insurers, medical expense per member per year increased from $2,585 in 2003 to $3,429 in 2006. This is an increase of 33% over this three-year period, or about 10% per year. (The rate of increase was almost identical at each of the three plans.) If medical expenses had increased only by the rate of growth in GDP—20.9% instead of 33%– per member per year medical costs in 2006 would have been only $3,125, or about $305 less. Total medical spending for the approximately two million people who had health insurance through the three insurers would have been $629 million lower than it actually was in 2006. If I assume self-insured employers had the same rate of increase in medical expenses as those with insured plans, and add in the 1.8 million people covered by self-insured plans at these same three insurers, the total savings would have been nearly double—or almost $1.2 billion in 2006.

If medical expenses had increased only by the rate of growth in CPI—9.4% instead of 33%– per member per year medical costs in 2006 would have been only $2,828, or about $602 less. Total medical spending would have been reduced by $1.2 billion for people with insured plans, and $2.3 billion if I add in people covered by self-insured plans.

Surprising to me, the rate of increase in health insurer administrative costs per member was even greater than the increase in medical expenses over this period. Average administrative expenses per member per year increased from $330 in 2003 to $486 in 2006. That’s a total increase of 47% over the three-year period, or about 13.5% per year. (The rate of increase was quite different from insurer to insurer, indicating that some plans have done a much better job controlling their administrative expenses than others, or that plans might make different decisions about how to allocate administrative expenses across lines of business. Again, their DOI reports reflect only their fully insured lines of business) If the rate of growth in administrative expenses had been constrained to the same growth rate as GDP, per member administrative costs in 2006 would have been only $399, or $87 lower per member. This would have saved the two million insured people a total of $179 million in 2006. And if people covered by self-insured plans are added in, the total savings would have nearly doubled—to about $336 million.

Using the lower growth rate in the CPI produced more savings: average health plan administrative costs per member would have been only $361, a savings of $125 per year. This produces total savings for insured people of $257 million, and more than $482 million if I include those covered by self-insured plans.

So, constraining medical and administrative cost increases to the same rate as GDP growth would have resulted in total annual savings in 2006 of about $392 per person. That’s enough to pay for three months of a Young Adult Plan or two month of a Bronze Plan with drug coverage. Controlling cost growth to the increase in the CPI results in total annual savings of $727. That’s nine months of automobile insurance premiums for the average driver in Massachusetts, a significant portion of a monthly mortgage or rent payment, or a good contribution to savings for a kid’s future college education.

My little analysis shows what we all know: the biggest savings in health insurance premiums will come from controlling medical expenses. This will be quite challenging to accomplish in the short-term—unless we have the political will and resolve to control prices. But my calculations also suggest that there are huge savings to be reaped if health insurers can control the growth in their own administrative expenses. This is an area in which it should be possible to make much faster progress. The Boston area CPI increased 9.4% from 2003-2006. Why did average health plan administrative expenses per member increase at five times this rate? As health insurers and the rest of us rightly pressure providers to control medical expenses and demand more transparency about prices and quality, we need to be sure to take an equally transparent—and hard–look at health plan administrative costs.

Nancy Turnbull
Harvard School of Public Health

A final post-script:

‘Twas just days before New Year
No word from the state:
Next year’s penalty schedule?
Why such a long wait?

Please follow our community rules when engaging in comment discussion on this site.
  • Janet Weathers

    I would be interested to know how much of the health care costs are attributable to medication. This analysis is helpful, and I am grateful to the attention to both medical costs and administrative costs. However, to just meet a monetary limit in expenditures, it is too easy for health insurers to just cut reimbursements to providers. A real solution is much more complicated and may need to include such radical measures as negotiating with drug companies for much greater reductions in the costs of drugs, instituting a 20x rule (in which the CEO earns no more than 20 times the lowest paid employee), and as a society making decisions about what will be covered or not covered.

  • Nancy Turnbull

    Allen, thanks for sharing the BCBS numbers. As I said in my original post, there was wide variation in administrative expense trends among the plans in my analysis. The BCBS trend was the lowest, and is about the same as the increase in nominal GDP. But it’s still more than twice the increase in the CPI. And we could probably have a healthy debate about whether nominal GDP growth is an appropriate target for health care cost growth.

    Thanks for pointing out that BCBS’s administrative expenses include contributions to the Foundation and the E-Health Collaborative, both of which are great initiatives. (My opinion on the Foundation is hardly surprising since I worked there and was on its board.) But excluding the company’s contributions to the Foundation in 2003 and 2006 makes no difference to the trend because the contributions are so small compared to overall administrative expenses (e.g., $12 million in 2006 vs. nearly $600 million in administrative expenses for insured private business). I can’t find any explicit line item for the E-Health Collaborative in BCBS’s DOI filings, so I can’t make any adjustment for that expense to judge whether it would have a material impact on the trend. I suspect not, particularly if the company’s $50 million contribution to the Collaborative has been expensed over multiple years.

    I’m happy to hear that BCBS is going to start posting financial information on its website. I hope the informaiton will be easy to find and understand for anyone who wants to review it. I’ll look forward to seeing what you share.

    Charlie, I appreciate the clarification about the impact on HPHC of the change in the classification of cost containment expenses. I did not mean to imply that I was “blowing off” this change—just looking for more public information about it. When I look at the HPHC filing for 2006, I see about $50 million in “containment expenses.” There’s no comparable line item in 2003. Assuming this is what you’re talking about, I removed this $50 million from 2006 administrative expenses and redid my analysis. (This may be overestimating the impact of the alloation change since I don’t know if all of the $50 million was previously in medical expenses, and some of this expense may be more appropriately allocated to HPHC’s Medicare line of business). Removing the $50 million does result in a significant reduction in the growth in HPHC’s administrative expenses per member from 2003-2006. But the adjusted trend is still nearly three times the growth rate in GDP and about six times the increase in the CPI.

    What you, and Allen, write about the relative importance of administrative expenses compared to medical expenses is certainly true. As you note, I made this point in my original post, and I have also made it in subsequent postings here and on the Health Care for All blog. That being said, there are lots of premium dollars going to administrative expenses too (according to Marylou’s post on the Health Care for All blog, total health plan administrative expenses were $1.4 billion in 2006 for the ten commercial plans just for their insured business—and this digure would be much largeer if it included the administrative expenses paid to health insurers by self-insured accounts, MassHealth and Commonwealth Care). And these expenses are rising at a far faster clip than the cost of most other goods and services. In this new era of health reform, when the state is requiring many people to buy the product of the health plans, HPHC and other plans can expect more demands for information and accountability for administrative expense trends. You’ve long been one of the most passionate and articulate voices for pulling back the curtain on costs in the health care system. This scrutiny is badly needed, and I look forward to someone beginning a public discussion of the financial trends and results for hospitals and other providers. As I’ve heard you say, this type of transparency may not be the solution to rising costs, but it can’t hurt.

  • Allen Maltz

    Happy New Year to everyone.

    Nancy, I read with great interest the information you shared on medical expenses and administration costs of Massachusetts-based health plans. As the CFO of Blue Cross and Blue Shield of Massachusetts, I am concerned about the affordability of health care as well. While I can’t speak for the other plans, with respect to Blue Cross the actual results are quite different and it would appear substantially better than the state average. Here’s what Blue Cross reported:

    BCBSMA reported a 19.5% growth in per member per month administrative expenses between 2003 and 2006 for all products. That number compares favorably to the 47% figure you quote as the average for Massachusetts-based health plans and the 20.9% figure you cite from the U.S. Bureau of Economic Analysis as the unadjusted growth in GDP. Please keep in mind that the BCBSMA administrative cost total for this time period includes our corporate contributions to the Massachusetts eHealth Collaborative (a $50 million pilot program designed to facilitate the implementation of a state-wide system of electronic medical records) and $51 million to the Blue Cross Blue Shield of Massachusetts Charitable Foundation. Actually, we’re very proud of the charitable contributions included in our “administration” costs.

    Still, your analysis correctly identifies the control of medical expenses as the key to making health care more affordable. For Blue Cross customers, almost 90 cents of the insurance premium dollar goes directly to pay health care expenses. That’s why it was also appropriate that your analysis considered self-insured customers. The fact is, whether an insurer takes the risk or a customer pays their own claims, medical care expenses drive the rate of increase. Medical expense increases are made up of two important categories of roughly equal weight:

     The first category is unit price increases (that is, what does an office visit cost this year compared to the same office visit last year). It’s reasonable to compare the increase in unit costs to CPI.

     The second component is the rate of increase in medical care utilization and severity. Controlling that is more challenging and involves everyone. Increases in utilization and severity include things such as the increase in the demand for prescription drugs (there’s a reason all those ads are on TV) and the use of an MRI in place of an x-ray. Those things in many cases represent improvements in medicine, but they also increase cost. It’s not clear whether capping the use of MRIs is a good or a bad thing.

    As a company, we feel strongly that the most effective way to moderate health care costs is to make care better by eliminating care that is unnecessary or even possibly harmful. We believe that changes in medicine that improve health care quality and outcomes are cost effective. On the other hand, we work with doctors and hospitals to reduce unnecessary overuse, under use and mis-use of medical care. This is where administrative costs and health care costs blur. Administrative expenses include the cost of implementing and running health plan care management programs. Our disease management programs reduce BCBSMA member health care costs by an average of $100 million dollars per year. This happens because our members with chronic conditions receive more coordinated treatment for their conditions and, therefore, suffer fewer catastrophic medical events and hospitalizations. Our actuaries factor these care management programs into BCBSMA premiums so they help offset increasing medical expenses. The impact these programs have on administrative cost is a fraction of the amount of money they save on medical expense increases.

    BCBSMA is constantly researching ways to control and minimize administrative expenses. As a company, it is one of the many efforts we are undertaking to help slow the increasing cost of health insurance in Massachusetts. We are happy to participate in conversations like this because we believe that in order to determine the best ways to spend health care dollars all stakeholders need to understand where those dollars are currently being spent. Thanks for facilitating this conversation with your analysis.

    Nancy, I also wanted to let you know that we are in the process of posting information about our revenues, expenses, and membership on

  • charlie

    Nancy – a thoughtful – and understandable – set of comments on a confusing and arcane set of insurance issues. And you draw the right conclusion – “…the biggest savings in health insurance premiums will come from controlling medical expenses.” Absolutely correct.
    That said, I don’t think you can totally blow off the rules change implemented by the MA DOI in 2003. Speaking only for Harvard Pilgrim, I can tell you that that particular change moved a lot of spending out of our medical expense line and into our administrative expense line. When we made the change, it increased administrative expenses by 30(!) percent and reduced medical expenses by 3 percent. That’s because medical spending represents 88-90 percent of total spending, and administrative spending is only 11-12 percent. Move a big number out of medical spending and it barely gets noticed that it’s gone, but move it into administrative spending and BAM! – it makes a big difference.
    I don’t know, however, if this change had as material an impact on the other carriers as it had on us. If it did, it would not change the medical expense trend very much, but would represent a big piece of the increase in administrative spending.
    Finally, administrative spending in 2006 is still only 11-12 percent of total spending. Take it down by 10 percent – a relatively big number – and you save 1 percent of premium. By contrast, take the 88-90 percent that’s spent on medical services down by 10 percent – also a big number – and you save 8-9 percent of premium.
    As Nancy said in the beginning, “the biggest savings in health insurance premiums will come from controlling medical expenses.” Right.

  • Nancy Turnbull

    Paul and Beth, I forgot to answer your questions:

    Paul, there was an enormous difference in the increase in health plan administrative expense per member (several orders of magnitude), and there was a negative correlation with size of plan (i.e., bigger plan had smaller increase).

    Beth, I suspect your physician’s office has three people at the front desk at least in part because the office has to deal with such a large number of insurance plans, each with different administrative requirements and complexities. This is one of the enormous prices we pay for having such a fragmented financing system in the US. But there may also be a lack of economy of scale in staffing that results from the doctor being in solo practice, as opposed to a group practice arrangement. But I’ll leave it to people who are experts in practice management to comment more knowledgeably than I can.

  • Nancy Turnbull

    To John and John: Thanks for your questions. Most of the plans report information only on their insured business, and they are supposed to allocate and report only the administrative expenses associated with these members. (The detailed administrative expense schedule in the DoI filings requires a specific offset for “reimbursements by uninsured plans.” Of course, in a competitive market, different customers and lines of business have differential ability to bargain about prices, including administrative loads, and plans make different pricing decisions depending on how much they want certain customers. So whether the expense allocation decisions that are made are “fair” or “equitable” is hard to judge, at least from the public filings. I assume that large self-insured accounts have much more clout and get a much better deal than smaller employers, and certainly than individuals. But it’s hard to support that with any data from the public filings. I have seen plan filings that suggest that certain health plans generally experience overall losses on self-insured accounts, which suggests that self-insured groups are being subsidized by other lines of business.

    Thanks to Jim and Marylou for their very thoughtful comments. A few comments in response:

    I agree completely that health plan administrative expenses are only one part of the challenge of controlling health insurance premiums. But they are an important component and one which often does not receive sufficient discussion. So I welcome the opportunity to talk about this issue.

    The 2007 health plan financial reports will be filed at the end of February, so we’ll be able to see the impact of the changes Jim notes have occurred at Tufts on its medical expenses and administrative costs, particularly on a per member basis. I’ll be happy to share my analysis of this issue with Commonhealth readers when the data are available.

    There are many challenges in comparing expense trends and ratios at the plans, including changes in reporting requirements and differences in expense allocation methods. From my review of the filings, I don’t believe that the 2005 requirement that plans report “cost containment expenses” had any material impact on my analysis, since the vast majority of the expenses that are being reported as cost containment expenses by the plans are salary, occupancy and outsourced services (such as EDP), which presumably were classified before as administrative expenses. But I would love to have more publicly available information with which to understand this issue. If the change resulted in a significant reclassification of medical management services from medical expenses to administrative expense, this could be important. The issue of classification of health plan expenses is one that my colleague Nancy Kane and I have examined in some detail, particularly as it relates to the reporting of medical expense ratios. (For the stout of heart, you could check out N. Turnbull, and N. Kane, “The Impact of Accounting and Actuarial Practice Differences on Medical Loss Ratios: An Exploratory Study of Five HMOs,” Inquiry 36, no. 3 (Fall 1999): 343-352.) We need to address the differences in health plan allocation methods if we’re going to move to using health plan “medical expense ratios” in any important regulatory way.

    I commend MAHP for making it easy to obtain information on the financial performance of its member health plans. Blue Cross Blue Shield, which is not a member of MAHP, should make comparable information publicly available in an easier to digest format. I also agree completely that hospitals should be required to report publicly. The Division of Health Care Finance and Policy has compiled and reported hospital information at various points in the past, although its report could be more user-friendly. The credibility and usefulness of any public reports of this type would be enhanced greatly if the health plans and hospitals consulted with outside folks, particularly from consumer groups such as Health Care for All, in designing these report.

    Although no one asked, I have a few suggestions for improving the MAHP report:

    #1: make the report easier to find on the MAHP website–not everyone will know to look in “Cost Control Initiatives”–if you’re looking for the report, you can find it at

    #2: Provide information on self-insured membership. I have written before on this blog about my belief that health insurers should be reqired to report on self-insured business so that we can understand the entire financial picture at the plans, including getting a clearer understanding of how expenses are being allocated to each line of business. The Division of Insurance permitted the plans to stop reporting this information several years ago; I believe this was a mistake that should be corrected by the current Administration. Until it does, MAHP’s member plans could provide this information voluntarily.

    #3: Report net income for each plan, not just underwriting gain/loss. All of the plans have significant investment income and this is an important part of the the financial performance of the plans. Without showing net income, and overall profit margin, we can’t get a complete picture of how the plans are doing. (This is comparable to the “surplus revenue” figure that Marylou wants hospitals to report.)

    #4:Report on health plan reserves/net worth,total capital and surplus. Again, this is a critical figure to understand the finanical position of health plans (and it’s comparable to the “endowment levels” that Marylou rightly recommends hospitals should report publicly).

    I look forward to continuing this conversation. Nancy

  • Dr. Marylou Buyse

    We’re pleased that Nancy and so many others are focused on controlling health care costs. Understanding why health care costs are rising is the first step to controlling them. It is also an essential step.

    While the analysis focused on health plan administrative expenses for the three largest health plans, it doesn’t focus on where the money is: the bulk of the health care dollar that goes to pay for medical care. Medical care is expensive and health plans provide protection against the unexpected expenses of serious illness, accident or injury. Without this protection, a stay in the hospital or an operation can mean financial ruin to many Massachusetts residents and families.

    Massachusetts health plans are national leaders when it comes to quality. That is an indisputable fact, and they expend their resources known as “administrative expenses” towards improving health care and keeping costs as low as possible. It is also important to note that the state’s Division of Insurance changed the reporting rules for administrative costs between 2004 and 2005, requiring all plans to re-classify some costs that had been thought of as “medical and medical cost management” in 2003 and 2004 and these changes moved them into the administrative expense line beginning in 2005.

    Focusing on health plan administrative expenses provides a very limited view of what’s going on in health care. The focus needs to be on everyone in the health care system and what each is doing to keep health care affordable and quality high.

    In Massachusetts, for the six major commercial carriers, nearly 90 cents (87.3) of every premium dollar goes to pay for medical services, such as doctor visits, prescription drugs, and hospital costs. Roughly 10 cents (11.1) is allocated toward administrative costs, such as care management programs for individuals with chronic conditions and other services to support consumers and providers, along with costs associated with complying with requirements mandated by state and federal agencies. The remaining 1.6 cents generally is directed into health plan reserves, money set aside to ensure that medical claims are paid for catastrophic medical expenses or if a natural or man-made disaster or some other unforeseen event were to occur.

    We’re happy to answer for consumers and employers the question, “Where does the money go?” Last month, MAHP along with several of our member health plans outlined a comprehensive cost control agenda. An important component of this was publicly disclosing health plan revenue and expense data in a consumer-friendly format. This data is available on the MAHP website ( and includes the following information:

    1. The amount paid for medical expenses, such as hospital and medical benefits, prescription drugs, and bonuses paid to providers;
    2. Total administrative costs; and
    3. Health plan surpluses and reserve levels

    However, health plans are just one component of the health care industry. We are the one part of the health care system that is transparent, that does publicly report its expenses and that works tirelessly day after day to find ways to lower costs and improve quality. With the bulk of the premium dollar going to pay for medical care, we think it is important that there be similar disclosure from others in the health care industry.

    For example, hospitals should make public in an easy to understand format, available data disclosing total inpatient and outpatient service revenue, total patient expenses, total capital expenses, total administrative expenses, and surplus revenue and endowment levels, similar to what we have released.

    Our cost control agenda also includes a proposal that the state’s Health Care Quality and Cost Council convene annual hearings to examine the factors driving health care costs. Health plans would be required to outline the factors contributing to any changes in premiums, including their projected medical expenses due to provider reimbursement rates, patient utilization, administrative costs, capital investments, and efforts to reduce the rate of growth. To provide a complete picture, the Council should also require other entities in the health care industry – health centers, hospitals, physician practices, and pharmacies – to participate in these hearings and explain the factors contributing to their rising costs, whether it is greater use of technology, increased consumer demand, or if they’re seeking higher reimbursement rates.

    Enhanced public reporting of health care revenues and expenses and public hearings on cost drivers will shed light on the why costs are rising and what can be done about it.

  • Jim Roosevelt

    Nancy is right that the overwhelming majority of cost, nearly 90 cents on the dollar, is due to rising medical costs. While Tufts Health Plan takes pride in our position as a market leader in medical management efforts, which requires significant investment of administrative dollars, we recognize that these programs alone will not achieve the savings necessary to have a substantial impact on the medical cost problem. We need the help of the entire health care community, including hospitals, providers, pharmaceutical manufactures and government, to truly reign in costs in way that will affect premiums.

    With that being said, administrative costs should not be ignored, which is why Tufts Health Plan, along with Massachusetts Association of Health Plans, is advocating for transparency in health care costs – administrative costs included. While I can’t speak for other plans, it is well known to those who work closely with health plans, that our organization has worked aggressively in the past 18 months to reduce administrative costs. While not easy, this effort speaks to our commitment to keeping health care premiums as affordable as possible.

    Our successful turnaround plan has been predicated on reducing medical and administrative costs. Therefore, we not only achieved double digit decreases, but we also reconfigured the way we do business. Our premiums are competitive and the marketplace is responding by buying our products. Moreover, we continue to receive national attention for clinical excellence. We are staffed appropriately, and continue to review our processes in order to find opportunities for efficiencies. We understand that close attention to administrative costs must remain a top priority.

    I appreciate the dissection of premium costs and welcome the dialogue on ways that we can all contribute to reining in unsustainable costs. The exciting thing, to me, is that others are eager for the same discussion.

  • John Greenbaum

    Nancy, a thoughtful analysis of some of the outward drivers in the increase in the cost of healthcare. I also find it interesting that a wide disaparity exists in administrative cost increases. I wonder how it relates to each of the insurers contributions to reserves. In my experience self-insured plans have a much lower rate of administrative inflation than insured plans as the unbundling of expenses is decidely more revealing.

  • John Andrewes

    Hi Nancy,

    You bring up good points and I enjoyed your article.
    When you adjust your calculations for Medicare and self insured for comparison, do we lose the true administrative cost since the overhead, such as building cost, salaries, etc. for the associates and IT infrastructure supporting those business lines are in fact included in the overall administrative cost?

    Do the DOI filings accurately distinguish the admin cost so true admin analysis can be made?

  • Brian Rosman

    Nancy asks:

    ‘Twas just days before New Year
    No word from the state:
    Next year’s penalty schedule?
    Why such a long wait?

    Reading Nancy’s lament, Deval’s team gave a whistle,
    They sent out a release, and posted an epistle
    But I heard him exclaim, ‘ere he drove out of sight,
    “Happy New Year to all, health reform is all right!”

  • Beth Vance

    Thank you, Turnbull.

    Now, can you please tell me why my private, single practice MD needs three people at the front desk?

  • Paul Levy

    Terrific summary, Nancy, and a great indication of why controlling costs isn’t something that can be done by administrative fiat. Also, yet another imperative for hospitals and other providers to seek quality and safety improvements to help reduce cost increases.

    Just curious. On this point — “The rate of increase was quite different from insurer to insurer” — Would you like to tell us what the range of growth rates was, and whether they are correlated with the size of the companies or any other such factors?