Consumer Driven Health Plans – Working as Intended

By Greg Scandlen

Consumer Driven Health Plans have been around for about five years now. They began in June 2002, when the Internal Revenue Service released its first guidance on Health Reimbursement Arrangements (HRAs), but really got a boost in December 2003 when Congress enacted and the President signed legislation enabling Health Savings Accounts (HSAs) as part of the Medicare Modernization Act. Both programs reduce the amount of services covered by an insurance plan, but supplement the insurance with an account of money that receives the same beneficial tax treatment as the insurance portion of the coverage. Money not spent in one year may be rolled over and used for future expenses. 

The two approaches – HRAs and HSAs – are the lynchpins of consumer driven health, but they are not the only elements. Other approaches include “defined contribution” plans, where employers make a fixed sum of money available to employees who then use that money to buy their own coverage and “flexible spending accounts” that enable workers to set aside some money to pay directly for the care they need.  

Outside of these tax-favored programs, many employers have raised deductibles and co-payments, which require employees to pay directly for services though without any tax advantage. Other employers have simply stopped providing coverage at all. In both cases, consumers are responsible for making their own decisions on purchasing a health insurance policy or paying directly for the care they consume. 

These other developments certainly increase consumer cost awareness and demand for information that helps stretch their health care dollars. But the core of consumerism in health care is associated with HRAs and HSAs, known together as “account based plans.” 

When these programs were being considered, proponents said they would have several beneficial effects:[1]

1. They would increase the awareness of consumers to the cost of the care they used.
2. That awareness and the fact that unspent money could be saved for future use would prompt consumers to be more cautious in the use of services.
3. That new caution would result in a lower use of unnecessary care.
4. Less unnecessary care would lower the rate of growth in health care costs.
5. Being at financial risk would prompt consumers to pay more attention to and become more invested in the treatment programs prescribed by their physicians.
6. Being in a position to choose how to spend their money would prompt consumers to demand better information to make better choices.
7. Word of these favorable trends would spread, causing more employers and more consumers to embrace these programs.
8. And, finally, consumers controlling more of the money and demanding better information would prompt new demands for health services delivery that is more accountable, more efficient, more convenient, and of better quality that what we have had in the past.

Five years later, all of these predictions are being realized:

1. Patient behavior is changing and people are being more cautious about needless use of services.
2. Consumers are more compliant with treatment regimens, especially those with chronic conditions who are high utilizers of services.
3. The rate of increase in health care costs is down substantially for people and groups in these plans.
4. The demand for information, transparent prices, and patient support services is high.
5. The adoption rate in the benefits market is sizzling.
6. The transformation of service delivery is beginning, though still very formative. Early indicators include the growth of retail clinics, concierge medicine practices, and medical tourism.

Not all of the evidence agrees with these conclusions. But the counter evidence tends to be based on three surveys that are of questionable value. Let’s look at those first.

NEGATIVE REPORTS 

KFF/HRET Annual Survey of Employer Health Benefits

This is an annual telephone survey of some 2,000 employers. It has been conducted every year for some 20 years, and is very comprehensive. In fact, its comprehensiveness may be a weakness. It asks some 400 detailed questions of respondents, testing the limits of telephone survey research.  The respondents must be extremely knowledgeable about their benefit programs and articulate enough to answer questions verbally. Sample sizes are also a problem for some of the market segments it surveys. While it interviews 321 employers with more than 5,000 employees (out of a universe of 8,114 such firms), it interviews only 113 companies with 3-9 employees, out of a universe of 2 million firms of that size. 

The enrollment information collected by KFF/HRET is questionable to start with. Its 2005 survey found 2.4 million workers in CDHPs (0.8 million in HSAs and 1.6 million in HRAs), growing to only 2.7 million in 2006 (1.4 million in HSAs and 1.3 million in HRAs), and 3.8 million in 2007 (with 1.9 million in each type). This is counting workers only, not family members and not those with non-group coverage, so total enrollment in early 2007 would equal some 11 million, which is consistent with what most industry watchers estimate. But still, the survey pretends a level of precision that is unwarranted. It is questionable, for instance, that HRA enrollment dipped in 2006 and then rose again in 2007.  

Some of the more detailed information contained within the report can be spun in a negative or positive light, depending on the point of view of the reader. The 2006 survey found that 19 percent of workers with a choice of plan chose the CDHP. Is that high or low? It can be characterized either way. One critic emphasized that 39 percent of the people in a CDHP were offered no other choice, but neglected to mention that much higher percentages of people in HMOs and PPOs were offered no other choice of plan. [2] 

KFF Survey of Enrollees in Consumer-Directed Health Plans

The Kaiser Family Foundation also published a survey in November 2006 that takes a closer look at the experience of people in Consumer Driven Plans. It found generally that people in CDHPs are more cost conscious and more likely to use information tools than others, but are also less happy with their plan and more likely to skip needed care. However there are some major problems with this survey. 

For one thing, it interviewed only 272 individuals with CDHP, and 715 in a “control group.” But 21 percent of the control group said they had deductibles that were high enough to qualify as a CDHP and 14 percent said they had a “personal savings account that they can use for health related expenses.” There is no explanation why these people were included in the control group rather than the CDHP group. Eighty-two percent of those with a CDHP said they were in a “health savings account” (rather than an HRA), even though population-wide the enrollment in the two programs is about equal. All of  the control group but only 78 percent of the CDHP had employer-sponsored coverage, which alone could account for the difference in satisfaction. (We are happier when someone else pays the bills.) Respondents were not asked if they had a choice of plan, which could be important when they are later asked whether they would switch plans if they could. Also, the people in the CDHP group were far more likely to have been in their plan for one year or less (48 percent vs. 20 percent), so the natural sorting process had not worked itself out yet. 

There are a host of other problems with the survey, but the problems of the small sample size could have been overcome if the researchers had made sure they were testing people in similar situations. They did not, so the results are questionable. [3] 

EBRI/Commonwealth Survey

Even more problematic is “The 2nd Annual EBRI/Commonwealth Fund Consumerism in Health Care Survey, 2006: Early Experience With High-Deductible and Consumer-Driven Health Plans” that was conducted jointly by the Employee Benefits Research Institute and the Commonwealth Fund in December 2006. This is an online survey of 3,000 people drawn from an online database of Internet users who have agreed to participate in online surveys. In their initial sampling they found only 21 people who were enrolled in a CDHP. So they went back and “oversampled” to get 150 or so respondents. This is a survey whose sole purpose was to look at the experience people were having with CDHPs, so it is hard to imagine why they would be satisfied with such a small number of people with relevant experience. Interestingly, in their “chartpack” of results, they never mention the actual numbers of respondents, just the percentages. Reliable research will always tell the reader what the “n” (number of respondents) equals. 

Not surprisingly, the results are terribly skewed. It estimates the number of people enrolled nationally at just 1.3 million in 2006. Even the KFF/HRET survey cited above reported 2.7 million workers (6.75 million people) enrolled in 2006. And contrary to most other reports, the survey report finds less satisfaction, more “missed care,” less impact on the uninsured, and less access to information than “comprehensive care” provides. But given how unreliable the simple count of enrollees is, there is no reason to think the rest of this survey is any better. [4] 

These three surveys are the basis for virtually all of the negative reports and articles that have been written about consumer driven health care in the past year. Reporters, and even academics rarely look any further, and they end up quoting each other to beef up their references, even though it is all based on just three very flawed public opinion surveys. 

Completely ignored is a wealth of other, more reliable information that is not hard to find. In fact, there are surveys that are more comprehensive and vendor reports that are more accurate in terms of enrollment, utilization changes and cost effects than these limited public opinion surveys could ever be.  

POSITIVE REPORTS 

Far more convincing is the empirical evidence coming from employers and health plans about the actual changes in consumer behavior. Public opinion surveys may be easily skewed to achieve the results the surveyors are hoping for. They rely on a very small number of respondents who may or may not be representative of the wider population, and may or may not be well informed about their own health care situation and the terminology used to describe alternative plan designs. 

The reports from health plans and employers know with precision whether an individual is enrolled in an HRA or an HSA. They can actually count the number of services used before and after enrollment. They can compare the experience of CDHP enrollees with a large population of people in more traditional forms of coverage, and adjust the effects for differences in demographics and health status. 

Enrollment Trends

Let’s begin with the available information on enrollment. 

America’s Health Insurance Plans (AHIP) is the trade association of the insurance industry. Not all of the insurance plans and HMOs belong to it, but the vast majority do. Every year since March 2005 it has surveyed its members about HSA enrollment. It does not ask about HRAs or other programs that might be included in a definition of “consumer driven health care.” These companies do not have to guess. They know precisely how many of their members are enrolled in HSA programs versus HRA, FSA or any other incarnation.  

AHIP’s nose count found 1 million enrolled in HSA programs as of March 2005, increasing to 3.2 million in January 2006 and 4.5 million in January 2007. These are three times as many people in HSAs alone as EBRI counted for both HSAs and HRAs.  

AHIP also found that people enrolled in HSA programs have widespread access to preventive services, disease management programs, and information and patient support tools. The vast majority have access to account information online (93 percent of all HSA enrollees), health education information (99 percent), physician-specific information (97 percent percent), hospital-specific quality information (86 percent), and health care cost information (88 percent). The companies offer coverage of disease management for diabetes (91 percent), coronary artery disease (90 percent), congestive heart failure (89 percent), and asthma (87 percent). Other programs aren’t as widely offered, but are growing.[5] 

More recent surveys find CDHPs have continued to grow rapidly in 2007. United Benefits Advisors (UBA) surveyed 10,000 employers (five times the number surveyed by KFF/HRET) and found that 56 percent more companies offered CDHPs in 2007 than in 2006, and 76 percent more people were enrolled. It also reported that this growth is concentrated in the 25–100 employee group market, the very market segment KFF/HRET was most likely to miss. [6]  

Recent press reports suggest that growth in HSAs is likely to double between January 2007 and January 2008 to 10 million. One report cited interviews of a number of benefit consultants who attribute the growth to a greater number of employees offering the programs and a greater proportion of employees enrolling in them. [7] 

Cost Trends

The growth in enrollment is fueled largely by favorable cost trends. The UBA survey cited above found that the cost of CDHPs went up just 2.7 percent in 2006, compared with 7.2 percent for all health plans. This finding is supported by many other reports: 

• Deloitte reports that cost trend for CDHPs in 2006 was 2.6 percent, as opposed to 7.4 percent for HMOs, 7.5 percent for PPOs, 7.3 percent for POS, and 6.6 percent for traditional indemnity coverage. [8]
• Cigna reports an overall trend of 10.3 percent in 2005, but only 4.8 percent for its HRA products and minus 1.2 percent for its HSAs. [9]
• An updated report from Cigna (October 2007) found that medical trend for its CDHP enrollees was less than half the trend for its PPO and HMO enrollees, even though out-of-pocket costs were similar for the two groups. [10]
• Minneapolis-based HealthPartners reported in October 2007 that medical costs for its CDHP enrollees were 4.4 percent lower than for people in traditional coverage, even after adjusting for health status. [11]
• In the non-group market, eHealthInsurance reported that premium costs for HSAs dropped 17 percent for individuals and 4.6 percent for families from 2004 to 2005. [12]
• Aetna reported on four years of experience with HRAs and found a 1 percent annual increase for full-replacement employers and 6.7 percent for employers that offered them as an option. [13]

Clearly something important is happening here. The same phenomenon is being reported by many different and independent sources. The cause is not a mystery. It comes from very favorable utilization changes.  

Utilization Trends

Enrollment is going up and costs are stabilizing because Consumer Driven Health Plans are doing exactly what they promised to do – change patient behavior.

UnitedHealth Group has recently reported that people in CDHPs are –

• Far more likely to see a doctor for diabetes (73 percent vs. 54 percent) and 16 percent more likely to receive HbA1c tests if they have diabetes.
• 22 percent more likely to have lipid tests if they have coronary artery disease.
• 6 percent more likely to use ACE inhibitors, 41 percent more likely to get creatinine tests and 26 percent more likely to receive potassium tests if they have congestive heart failure.
• 16 percent more likely to get cervical and prostate screening.
• 10 percent more likely to get cholesterol screening. [14]

The Blue Cross Blue Shield Association reported in 2006 that people with HSAs are more likely to –

• Use nurse hotlines (10 percent vs. 6 percent)
• Participate in wellness programs (20 percent vs. 8 percent)
• Use provider information tools (39 percent vs. 10 percent)
• Use Rx cost & comparison tools (42 percent vs. 19 percent)
• Use Web site based coverage information (53 percent vs. 32 percent)  [15]

A more recent report from the Blue Cross Blue Shield Association confirms these findings. Jennifer Vachon, Executive Director of the Marketing and Consulting Services Group at the Association, summed up the new survey by saying, “These findings show us that CDHPs are beginning to deliver on their promise. Our survey shows that CDHPs empower consumers and help them become more engaged in their health care decisions.”

Some of the information provided includes the following:

HSA enrollees are much more likely to research health information, including:

• Doctor quality: 20 percent of HSA enrollees; 14 percent of non-CDHP enrollees
• Doctor costs: 14 percent HSAs; 4 percent non-CDHPs
• Hospital quality: 12 percent HSAs; 7 percent non-CDHPs
• Hospital costs: 10 percent HSAs; 3 percent non-CDHPs
• Insurance information: 25 percent HSAs; 17 percent non-CDHPs

HSA enrollees are much more likely to plan and save for future health care expenses:

• Track health care expenses: 63 percent of HSAs; 43 percent of non-CDHPs
• Estimate future health care expenses: 38 percent of HSAs; 19 percent of non-CDHPs
• Save for future health care expenses: 47 percent of HSAs; 18 percent of non-CDHPs

HSA enrollees are much more likely to participate in wellness programs:

• Smoking Cessation: 20 percent of HSAs; 6 percent of non CDHPs
• Stress Management: 22 percent of HSAs; 8 percent of non-CDHPs
• Nutrition Programs: 27 percent of HSAs; 12 percent of non-CDHPs
• Exercise Programs: 29 percent of HSAs; 12 percent of non-CDHPs

HSA enrollees are no more likely to forego care due to cost:

• Did Not Go To Doctor: 18 percent of HSAs; 18 percent of non-CDHPs
• Delayed Treatment: 17 percent of HSAs; 17 percent of non-CDHPs
• Delayed Prescription: 15 percent of HSAs; 15 percent of non-CDHPs

The survey was conducted by Knowledge Networks of 3,000 people enrolled in Blue and non-Blue CDHPs and non-CDHPs. [16]

Cigna studied the experience of 38,211 “Choice Fund” (including both HSAs and HRAs) enrollees and compared it with the experience of 231,680 people enrolled in their PPO and HMO products. It found the Choice Fund enrollees had 11 percent lower costs for pharmaceuticals, 24 percent lower for inpatient care, and 10.7 percent lower for outpatient care. It found these savings were not the result of healthier enrollment. It also found that Choice Fund enrollees were 12 percent more likely to use preventive care and that, “Choice Fund members are more compliant with medications that manage ongoing conditions, and more discerning in their use of medications with over-the-counter alternatives.”  [17] 

These findings were confirmed by Cigna in October 2007 in a follow-up report that said, “First-year member preventive visits increased and second-year member visits remained significantly higher than those among traditional plan members (and) use of maintenance medications that support chronic conditions increased while costs decreased.”  [18] 

McKinsey & Company reports that people in CDHPs are:

• More likely to comply with treatments than people in traditional plans  (36 percent vs. 27 percent for diabetes, and 51 percent vs. 31 percent for HBP)
• 25 percent more likely to engage in healthy behaviors and 30 percent more likely to get an annual physical. [19]

A study in the Journal of the American Medical Association (March 14, 2007) found that people in CDHPs have 10 percent fewer emergency room visits overall and 25 percent fewer repeat visits, almost entirely for non-severe conditions. “Our study showed that for most members, the high-deductible plan seemed to work as intended,” said Frank Wharam, MD, MPH, research fellow in the Department of Ambulatory Care and Prevention at the Harvard Medical School and the study’s lead author. “Patients went to the emergency room less frequently for non-emergency conditions.”  [20] 

Now, it must be said that most of these reports are also limited. They measure only what is measured and do not report on other indicators. They don’t always control for differences in population characteristics and they often don’t present their methodology. These are not randomized control trials, but neither are they opinion surveys.

In fact, they are not aimed at public policy at all. This is market research. It is how business, not government, solves problems. Public policy research tries to identify issues on which there is broad agreement. It is looking for 50 percent-plus support of an idea that Congress or state legislators will be comfortable voting on. Once enacted into law a new idea or program will rarely be repealed or substantially changed. The citizenry is stuck with it, for better or worse. So, it had better have widespread support beforehand. 

Markets work very differently. A new product or idea does not need widespread support. It simply needs enough buyers to be potentially profitable. The first iteration is a test. People buy it because they like new things and they think it will improve their lives in some manner. Business looks at how well the new product is doing, what are the strengths and weaknesses, what works and doesn’t work, and revises it to eliminate the problems and build on the strengths.

The second iteration – Version 2.0 – is almost always better and cheaper than the first. Version 2.0 will appeal to a broader population and it, too, will be measured and improved. And so it goes until it reaches market maturity, growth levels off and people wonder how they ever did without it in the old days.

 WHERE IS ACADEMIA? 

With the exception of the JAMA article above, none of this information is peer-reviewed, scholarly research. Unfortunately, such research is almost always behind the times and doomed to be looking in a rear-view mirror. Typically, it takes two or three years for academic research to be conducted and published. In a fluid market, three years is ancient history. 

One example is the special edition on consumer driven plans published by the journal Health Services Research in June 2004. [21] This was a mere six months after HSAs became law so the information on Consumer Driven Plans was confined to a few companies that had pioneered the Health Reimbursement Arrangement concept. A few of the articles looked at the program Humana had first offered in 2001, but Humana had already discovered the problems with that design and had revised its product by the time the papers were published. 

Another example would be a literature review by Melinda Beeuwkes Buntin and colleagues at the RAND Corporation published in Health Affairs in October 2006. Buntin does not attempt to offer original research, nor does she evaluate the quality of the available literature. Instead, she provides a reasonably balanced, though uncritical, synopsis of what has been published, including many, though not all, of the studies reported on above.  

Unfortunately, this results in speculative work from 1996 being given equal weight to current reports from the field. Of course, in 1996 the original “Medical Savings Account” law (the predecessor of current approaches), had not even been written, let alone enacted. Projections from the time would have been entirely hypothetical and were aimed at influencing Congressional deliberations. 

Even more disturbing is the continued reliance on the RAND Health Insurance Experiment (HIE) from 1971-1982. The HIE was a powerful research project that actually informed much of the interest in Consumer Driven Health Care today. But its limitations are profound. The federal HMO Act hadn’t been passed when the project was designed, PPOs didn’t exist, communications tools were primitive, and it predated much of the treatment protocols and medications that are standard today. It is disturbing that academics still reach back 30 years to predict what effects these programs will have on today’s market, especially when there is an enormous real-life experiment involving many millions of Americans taking place before their eyes. [22] 

Academic versus Market Analysis

In her conclusion, Buntin calls (naturally enough for a researcher) for more research, especially research that “adopts rigorous analytic techniques and methods that will produce reliable and generalizable conclusions.”  

Once again, this call underscores the differences between academic, public policy, and market research. [23] “Generalizable” to what? The entire population? The 50 percent-plus-1 of the population needed to pass new laws? Or some market segment, such as single mothers with high school educations living in the Midwest?  

Marketeers know that different population segments behave differently and want different products. No one product is right for everybody. The trick is to differentiate products so that every market segment is happy. Not everyone likes cell phones, so land lines and CB radios continue to be available on the market. But cell phones are still an extremely successful innovation.  

Similarly, not everyone will like Consumer Driven Health Care. It takes some effort to make choices and many people are passive. They will prefer HMO coverage. How big is the population that prefers each approach? It is impossible to say, at least for another 10 years or so, once the newer approach begins to reach maturity. By then there will be even newer products available and some small number of “early adopters” will try out the newer products to see how they work. 

People in business are entirely comfortable with a fluid environment like this. Academics and bureaucrats are not. They want programs that are universal and unchanging. Social Security today is very much like Social Security was 70 years ago, and if some people aren’t pleased with it, too bad. Most people like it fine and minorities of voters don’t matter.  

Meanwhile, we are in the midst of a transformation in American health care. Not everything about consumer directed health care will succeed, but the overwhelming preponderance of the evidence says it is working exactly as it was intended to work. Policy-makers and academics who ignore or deny this development are missing out on the most significant change in health care of their lifetimes.

ADDENDUM

The data below was not included in the study as it was originally published in November 2007 by Consumers for Health Care Choices.  

The Mercer consulting firm has issued a report on a new survey it conducted on employer benefits. It finds that the increases in benefits costs has declined substantially since 2002, and leveled off at around 6 percent annually – still double inflation. Smaller employers (under 200 workers) are terminating coverage. The percentage providing benefits has dropped from 69 percent in 2001 to 61 percent in 2007. The remaining employers have slowed trend through the use of health management and employee consumerism.

CDHPs (both HSAs and HRAs) have grown substantially from 1 percent of covered workers in 2005 to 3 percent in 2006 to 5 percent in 2007, and are likely to grow further in 2008. Employers with fewer than 500 workers increased their offering of CDHPs from 5 percent in 2006 to 7 percent in 2007, to a projected 11 percent in 2008. For large employers (500–20,000 workers), it has gone from 11 percent in 2006, to 14 percent in 2007 and 18 percent in 2008. For jumbos (over 20,000 workers) it was 37 percent in 2006, 41 percent in 2006, and 43 percent in 2008.

The survey says, “Evidence that the (CDHP) plans are cost-effective is accumulating. CDHPs delivered substantially lower cost per employee than either PPOs or HMOs in 2007. CDHP cost averaged $5,970 per employee, compared to $7,120 for HMOs and $7,352 for PPOs. Of the two types of CDHPs, HSA-based plans were less expensive than HRA-based plans ($5,679 compared to $6,224).”

Trend between 2006 and 2007 was also favorable to CDHPs, suggesting it is not just a one-time saving, but an ongoing advantage. The trend for CDHPs was 3.5 percent, compared to 7.6 percent for HMOs, 6.1 percent for PPOs, and 11.2 percent for POS plans.

Mercer says this cost difference is not due only to higher deductibles, but because the “account feature encourages more careful health spending.” Comparing CDHPs to high-deductible PPOs still shows a savings of $700 per employee per year.

When given a choice of plan, a greater percentage of employees are choosing the CDHP, up from 21 percent in 2005 to 29 percent in 2007.

The survey also asked about employer support of reform proposals. Only 23 percent support a “pay or play” mandate on employers and only the same number (23 percent) support an individual mandate. Over twice as many oppose these ideas.

SOURCE:
http://www.mercer.com/referencecontent.jhtml?idContent=1287790

Endnotes

Greg Scandlen is the president and founder of Consumers for Health Care Choices, a non-partisan, non-profit membership organization aimed at empowering consumers in the health care system, and a senior fellow for the Georgia Public Policy Foundation, an independent think tank that proposes practical, market-oriented approaches to public policy to improve the lives of Georgians. Nothing written here is to be construed as necessarily reflecting the views of the Georgia Public Policy Foundation or as an attempt to aid or hinder the passage of any bill before the U.S. Congress or the Georgia Legislature.

© Georgia Public Policy Foundation (March 17, 2008). Permission to reprint in whole or in part is hereby granted, provided the author and his affiliations are cited.

By Greg Scandlen

Consumer Driven Health Plans have been around for about five years now. They began in June 2002, when the Internal Revenue Service released its first guidance on Health Reimbursement Arrangements (HRAs), but really got a boost in December 2003 when Congress enacted and the President signed legislation enabling Health Savings Accounts (HSAs) as part of the Medicare Modernization Act. Both programs reduce the amount of services covered by an insurance plan, but supplement the insurance with an account of money that receives the same beneficial tax treatment as the insurance portion of the coverage. Money not spent in one year may be rolled over and used for future expenses. 

The two approaches – HRAs and HSAs – are the lynchpins of consumer driven health, but they are not the only elements. Other approaches include “defined contribution” plans, where employers make a fixed sum of money available to employees who then use that money to buy their own coverage and “flexible spending accounts” that enable workers to set aside some money to pay directly for the care they need.  

Outside of these tax-favored programs, many employers have raised deductibles and co-payments, which require employees to pay directly for services though without any tax advantage. Other employers have simply stopped providing coverage at all. In both cases, consumers are responsible for making their own decisions on purchasing a health insurance policy or paying directly for the care they consume. 

These other developments certainly increase consumer cost awareness and demand for information that helps stretch their health care dollars. But the core of consumerism in health care is associated with HRAs and HSAs, known together as “account based plans.” 

When these programs were being considered, proponents said they would have several beneficial effects:[1]

1. They would increase the awareness of consumers to the cost of the care they used.
2. That awareness and the fact that unspent money could be saved for future use would prompt consumers to be more cautious in the use of services.
3. That new caution would result in a lower use of unnecessary care.
4. Less unnecessary care would lower the rate of growth in health care costs.
5. Being at financial risk would prompt consumers to pay more attention to and become more invested in the treatment programs prescribed by their physicians.
6. Being in a position to choose how to spend their money would prompt consumers to demand better information to make better choices.
7. Word of these favorable trends would spread, causing more employers and more consumers to embrace these programs.
8. And, finally, consumers controlling more of the money and demanding better information would prompt new demands for health services delivery that is more accountable, more efficient, more convenient, and of better quality that what we have had in the past.

Five years later, all of these predictions are being realized:

1. Patient behavior is changing and people are being more cautious about needless use of services.
2. Consumers are more compliant with treatment regimens, especially those with chronic conditions who are high utilizers of services.
3. The rate of increase in health care costs is down substantially for people and groups in these plans.
4. The demand for information, transparent prices, and patient support services is high.
5. The adoption rate in the benefits market is sizzling.
6. The transformation of service delivery is beginning, though still very formative. Early indicators include the growth of retail clinics, concierge medicine practices, and medical tourism.

Not all of the evidence agrees with these conclusions. But the counter evidence tends to be based on three surveys that are of questionable value. Let’s look at those first.

NEGATIVE REPORTS 

KFF/HRET Annual Survey of Employer Health Benefits

This is an annual telephone survey of some 2,000 employers. It has been conducted every year for some 20 years, and is very comprehensive. In fact, its comprehensiveness may be a weakness. It asks some 400 detailed questions of respondents, testing the limits of telephone survey research.  The respondents must be extremely knowledgeable about their benefit programs and articulate enough to answer questions verbally. Sample sizes are also a problem for some of the market segments it surveys. While it interviews 321 employers with more than 5,000 employees (out of a universe of 8,114 such firms), it interviews only 113 companies with 3-9 employees, out of a universe of 2 million firms of that size. 

The enrollment information collected by KFF/HRET is questionable to start with. Its 2005 survey found 2.4 million workers in CDHPs (0.8 million in HSAs and 1.6 million in HRAs), growing to only 2.7 million in 2006 (1.4 million in HSAs and 1.3 million in HRAs), and 3.8 million in 2007 (with 1.9 million in each type). This is counting workers only, not family members and not those with non-group coverage, so total enrollment in early 2007 would equal some 11 million, which is consistent with what most industry watchers estimate. But still, the survey pretends a level of precision that is unwarranted. It is questionable, for instance, that HRA enrollment dipped in 2006 and then rose again in 2007.  

Some of the more detailed information contained within the report can be spun in a negative or positive light, depending on the point of view of the reader. The 2006 survey found that 19 percent of workers with a choice of plan chose the CDHP. Is that high or low? It can be characterized either way. One critic emphasized that 39 percent of the people in a CDHP were offered no other choice, but neglected to mention that much higher percentages of people in HMOs and PPOs were offered no other choice of plan. [2] 

KFF Survey of Enrollees in Consumer-Directed Health Plans

The Kaiser Family Foundation also published a survey in November 2006 that takes a closer look at the experience of people in Consumer Driven Plans. It found generally that people in CDHPs are more cost conscious and more likely to use information tools than others, but are also less happy with their plan and more likely to skip needed care. However there are some major problems with this survey. 

For one thing, it interviewed only 272 individuals with CDHP, and 715 in a “control group.” But 21 percent of the control group said they had deductibles that were high enough to qualify as a CDHP and 14 percent said they had a “personal savings account that they can use for health related expenses.” There is no explanation why these people were included in the control group rather than the CDHP group. Eighty-two percent of those with a CDHP said they were in a “health savings account” (rather than an HRA), even though population-wide the enrollment in the two programs is about equal. All of  the control group but only 78 percent of the CDHP had employer-sponsored coverage, which alone could account for the difference in satisfaction. (We are happier when someone else pays the bills.) Respondents were not asked if they had a choice of plan, which could be important when they are later asked whether they would switch plans if they could. Also, the people in the CDHP group were far more likely to have been in their plan for one year or less (48 percent vs. 20 percent), so the natural sorting process had not worked itself out yet. 

There are a host of other problems with the survey, but the problems of the small sample size could have been overcome if the researchers had made sure they were testing people in similar situations. They did not, so the results are questionable. [3] 

EBRI/Commonwealth Survey

Even more problematic is “The 2nd Annual EBRI/Commonwealth Fund Consumerism in Health Care Survey, 2006: Early Experience With High-Deductible and Consumer-Driven Health Plans” that was conducted jointly by the Employee Benefits Research Institute and the Commonwealth Fund in December 2006. This is an online survey of 3,000 people drawn from an online database of Internet users who have agreed to participate in online surveys. In their initial sampling they found only 21 people who were enrolled in a CDHP. So they went back and “oversampled” to get 150 or so respondents. This is a survey whose sole purpose was to look at the experience people were having with CDHPs, so it is hard to imagine why they would be satisfied with such a small number of people with relevant experience. Interestingly, in their “chartpack” of results, they never mention the actual numbers of respondents, just the percentages. Reliable research will always tell the reader what the “n” (number of respondents) equals. 

Not surprisingly, the results are terribly skewed. It estimates the number of people enrolled nationally at just 1.3 million in 2006. Even the KFF/HRET survey cited above reported 2.7 million workers (6.75 million people) enrolled in 2006. And contrary to most other reports, the survey report finds less satisfaction, more “missed care,” less impact on the uninsured, and less access to information than “comprehensive care” provides. But given how unreliable the simple count of enrollees is, there is no reason to think the rest of this survey is any better. [4] 

These three surveys are the basis for virtually all of the negative reports and articles that have been written about consumer driven health care in the past year. Reporters, and even academics rarely look any further, and they end up quoting each other to beef up their references, even though it is all based on just three very flawed public opinion surveys. 

Completely ignored is a wealth of other, more reliable information that is not hard to find. In fact, there are surveys that are more comprehensive and vendor reports that are more accurate in terms of enrollment, utilization changes and cost effects than these limited public opinion surveys could ever be.  

POSITIVE REPORTS 

Far more convincing is the empirical evidence coming from employers and health plans about the actual changes in consumer behavior. Public opinion surveys may be easily skewed to achieve the results the surveyors are hoping for. They rely on a very small number of respondents who may or may not be representative of the wider population, and may or may not be well informed about their own health care situation and the terminology used to describe alternative plan designs. 

The reports from health plans and employers know with precision whether an individual is enrolled in an HRA or an HSA. They can actually count the number of services used before and after enrollment. They can compare the experience of CDHP enrollees with a large population of people in more traditional forms of coverage, and adjust the effects for differences in demographics and health status. 

Enrollment Trends

Let’s begin with the available information on enrollment. 

America’s Health Insurance Plans (AHIP) is the trade association of the insurance industry. Not all of the insurance plans and HMOs belong to it, but the vast majority do. Every year since March 2005 it has surveyed its members about HSA enrollment. It does not ask about HRAs or other programs that might be included in a definition of “consumer driven health care.” These companies do not have to guess. They know precisely how many of their members are enrolled in HSA programs versus HRA, FSA or any other incarnation.  

AHIP’s nose count found 1 million enrolled in HSA programs as of March 2005, increasing to 3.2 million in January 2006 and 4.5 million in January 2007. These are three times as many people in HSAs alone as EBRI counted for both HSAs and HRAs.  

AHIP also found that people enrolled in HSA programs have widespread access to preventive services, disease management programs, and information and patient support tools. The vast majority have access to account information online (93 percent of all HSA enrollees), health education information (99 percent), physician-specific information (97 percent percent), hospital-specific quality information (86 percent), and health care cost information (88 percent). The companies offer coverage of disease management for diabetes (91 percent), coronary artery disease (90 percent), congestive heart failure (89 percent), and asthma (87 percent). Other programs aren’t as widely offered, but are growing.[5] 

More recent surveys find CDHPs have continued to grow rapidly in 2007. United Benefits Advisors (UBA) surveyed 10,000 employers (five times the number surveyed by KFF/HRET) and found that 56 percent more companies offered CDHPs in 2007 than in 2006, and 76 percent more people were enrolled. It also reported that this growth is concentrated in the 25–100 employee group market, the very market segment KFF/HRET was most likely to miss. [6]  

Recent press reports suggest that growth in HSAs is likely to double between January 2007 and January 2008 to 10 million. One report cited interviews of a number of benefit consultants who attribute the growth to a greater number of employees offering the programs and a greater proportion of employees enrolling in them. [7] 

Cost Trends

The growth in enrollment is fueled largely by favorable cost trends. The UBA survey cited above found that the cost of CDHPs went up just 2.7 percent in 2006, compared with 7.2 percent for all health plans. This finding is supported by many other reports: 

• Deloitte reports that cost trend for CDHPs in 2006 was 2.6 percent, as opposed to 7.4 percent for HMOs, 7.5 percent for PPOs, 7.3 percent for POS, and 6.6 percent for traditional indemnity coverage. [8]
• Cigna reports an overall trend of 10.3 percent in 2005, but only 4.8 percent for its HRA products and minus 1.2 percent for its HSAs. [9]
• An updated report from Cigna (October 2007) found that medical trend for its CDHP enrollees was less than half the trend for its PPO and HMO enrollees, even though out-of-pocket costs were similar for the two groups. [10]
• Minneapolis-based HealthPartners reported in October 2007 that medical costs for its CDHP enrollees were 4.4 percent lower than for people in traditional coverage, even after adjusting for health status. [11]
• In the non-group market, eHealthInsurance reported that premium costs for HSAs dropped 17 percent for individuals and 4.6 percent for families from 2004 to 2005. [12]
• Aetna reported on four years of experience with HRAs and found a 1 percent annual increase for full-replacement employers and 6.7 percent for employers that offered them as an option. [13]

Clearly something important is happening here. The same phenomenon is being reported by many different and independent sources. The cause is not a mystery. It comes from very favorable utilization changes.  

Utilization Trends

Enrollment is going up and costs are stabilizing because Consumer Driven Health Plans are doing exactly what they promised to do – change patient behavior.

UnitedHealth Group has recently reported that people in CDHPs are –

• Far more likely to see a doctor for diabetes (73 percent vs. 54 percent) and 16 percent more likely to receive HbA1c tests if they have diabetes.
• 22 percent more likely to have lipid tests if they have coronary artery disease.
• 6 percent more likely to use ACE inhibitors, 41 percent more likely to get creatinine tests and 26 percent more likely to receive potassium tests if they have congestive heart failure.
• 16 percent more likely to get cervical and prostate screening.
• 10 percent more likely to get cholesterol screening. [14]

The Blue Cross Blue Shield Association reported in 2006 that people with HSAs are more likely to –

• Use nurse hotlines (10 percent vs. 6 percent)
• Participate in wellness programs (20 percent vs. 8 percent)
• Use provider information tools (39 percent vs. 10 percent)
• Use Rx cost & comparison tools (42 percent vs. 19 percent)
• Use Web site based coverage information (53 percent vs. 32 percent)  [15]

A more recent report from the Blue Cross Blue Shield Association confirms these findings. Jennifer Vachon, Executive Director of the Marketing and Consulting Services Group at the Association, summed up the new survey by saying, “These findings show us that CDHPs are beginning to deliver on their promise. Our survey shows that CDHPs empower consumers and help them become more engaged in their health care decisions.”

Some of the information provided includes the following:

HSA enrollees are much more likely to research health information, including:

• Doctor quality: 20 percent of HSA enrollees; 14 percent of non-CDHP enrollees
• Doctor costs: 14 percent HSAs; 4 percent non-CDHPs
• Hospital quality: 12 percent HSAs; 7 percent non-CDHPs
• Hospital costs: 10 percent HSAs; 3 percent non-CDHPs
• Insurance information: 25 percent HSAs; 17 percent non-CDHPs

HSA enrollees are much more likely to plan and save for future health care expenses:

• Track health care expenses: 63 percent of HSAs; 43 percent of non-CDHPs
• Estimate future health care expenses: 38 percent of HSAs; 19 percent of non-CDHPs
• Save for future health care expenses: 47 percent of HSAs; 18 percent of non-CDHPs

HSA enrollees are much more likely to participate in wellness programs:

• Smoking Cessation: 20 percent of HSAs; 6 percent of non CDHPs
• Stress Management: 22 percent of HSAs; 8 percent of non-CDHPs
• Nutrition Programs: 27 percent of HSAs; 12 percent of non-CDHPs
• Exercise Programs: 29 percent of HSAs; 12 percent of non-CDHPs

HSA enrollees are no more likely to forego care due to cost:

• Did Not Go To Doctor: 18 percent of HSAs; 18 percent of non-CDHPs
• Delayed Treatment: 17 percent of HSAs; 17 percent of non-CDHPs
• Delayed Prescription: 15 percent of HSAs; 15 percent of non-CDHPs

The survey was conducted by Knowledge Networks of 3,000 people enrolled in Blue and non-Blue CDHPs and non-CDHPs. [16]

Cigna studied the experience of 38,211 “Choice Fund” (including both HSAs and HRAs) enrollees and compared it with the experience of 231,680 people enrolled in their PPO and HMO products. It found the Choice Fund enrollees had 11 percent lower costs for pharmaceuticals, 24 percent lower for inpatient care, and 10.7 percent lower for outpatient care. It found these savings were not the result of healthier enrollment. It also found that Choice Fund enrollees were 12 percent more likely to use preventive care and that, “Choice Fund members are more compliant with medications that manage ongoing conditions, and more discerning in their use of medications with over-the-counter alternatives.”  [17] 

These findings were confirmed by Cigna in October 2007 in a follow-up report that said, “First-year member preventive visits increased and second-year member visits remained significantly higher than those among traditional plan members (and) use of maintenance medications that support chronic conditions increased while costs decreased.”  [18] 

McKinsey & Company reports that people in CDHPs are:

• More likely to comply with treatments than people in traditional plans  (36 percent vs. 27 percent for diabetes, and 51 percent vs. 31 percent for HBP)
• 25 percent more likely to engage in healthy behaviors and 30 percent more likely to get an annual physical. [19]

A study in the Journal of the American Medical Association (March 14, 2007) found that people in CDHPs have 10 percent fewer emergency room visits overall and 25 percent fewer repeat visits, almost entirely for non-severe conditions. “Our study showed that for most members, the high-deductible plan seemed to work as intended,” said Frank Wharam, MD, MPH, research fellow in the Department of Ambulatory Care and Prevention at the Harvard Medical School and the study’s lead author. “Patients went to the emergency room less frequently for non-emergency conditions.”  [20] 

Now, it must be said that most of these reports are also limited. They measure only what is measured and do not report on other indicators. They don’t always control for differences in population characteristics and they often don’t present their methodology. These are not randomized control trials, but neither are they opinion surveys.

In fact, they are not aimed at public policy at all. This is market research. It is how business, not government, solves problems. Public policy research tries to identify issues on which there is broad agreement. It is looking for 50 percent-plus support of an idea that Congress or state legislators will be comfortable voting on. Once enacted into law a new idea or program will rarely be repealed or substantially changed. The citizenry is stuck with it, for better or worse. So, it had better have widespread support beforehand. 

Markets work very differently. A new product or idea does not need widespread support. It simply needs enough buyers to be potentially profitable. The first iteration is a test. People buy it because they like new things and they think it will improve their lives in some manner. Business looks at how well the new product is doing, what are the strengths and weaknesses, what works and doesn’t work, and revises it to eliminate the problems and build on the strengths.

The second iteration – Version 2.0 – is almost always better and cheaper than the first. Version 2.0 will appeal to a broader population and it, too, will be measured and improved. And so it goes until it reaches market maturity, growth levels off and people wonder how they ever did without it in the old days.

 WHERE IS ACADEMIA? 

With the exception of the JAMA article above, none of this information is peer-reviewed, scholarly research. Unfortunately, such research is almost always behind the times and doomed to be looking in a rear-view mirror. Typically, it takes two or three years for academic research to be conducted and published. In a fluid market, three years is ancient history. 

One example is the special edition on consumer driven plans published by the journal Health Services Research in June 2004. [21] This was a mere six months after HSAs became law so the information on Consumer Driven Plans was confined to a few companies that had pioneered the Health Reimbursement Arrangement concept. A few of the articles looked at the program Humana had first offered in 2001, but Humana had already discovered the problems with that design and had revised its product by the time the papers were published. 

Another example would be a literature review by Melinda Beeuwkes Buntin and colleagues at the RAND Corporation published in Health Affairs in October 2006. Buntin does not attempt to offer original research, nor does she evaluate the quality of the available literature. Instead, she provides a reasonably balanced, though uncritical, synopsis of what has been published, including many, though not all, of the studies reported on above.  

Unfortunately, this results in speculative work from 1996 being given equal weight to current reports from the field. Of course, in 1996 the original “Medical Savings Account” law (the predecessor of current approaches), had not even been written, let alone enacted. Projections from the time would have been entirely hypothetical and were aimed at influencing Congressional deliberations. 

Even more disturbing is the continued reliance on the RAND Health Insurance Experiment (HIE) from 1971-1982. The HIE was a powerful research project that actually informed much of the interest in Consumer Driven Health Care today. But its limitations are profound. The federal HMO Act hadn’t been passed when the project was designed, PPOs didn’t exist, communications tools were primitive, and it predated much of the treatment protocols and medications that are standard today. It is disturbing that academics still reach back 30 years to predict what effects these programs will have on today’s market, especially when there is an enormous real-life experiment involving many millions of Americans taking place before their eyes. [22] 

Academic versus Market Analysis

In her conclusion, Buntin calls (naturally enough for a researcher) for more research, especially research that “adopts rigorous analytic techniques and methods that will produce reliable and generalizable conclusions.”  

Once again, this call underscores the differences between academic, public policy, and market research. [23] “Generalizable” to what? The entire population? The 50 percent-plus-1 of the population needed to pass new laws? Or some market segment, such as single mothers with high school educations living in the Midwest?  

Marketeers know that different population segments behave differently and want different products. No one product is right for everybody. The trick is to differentiate products so that every market segment is happy. Not everyone likes cell phones, so land lines and CB radios continue to be available on the market. But cell phones are still an extremely successful innovation.  

Similarly, not everyone will like Consumer Driven Health Care. It takes some effort to make choices and many people are passive. They will prefer HMO coverage. How big is the population that prefers each approach? It is impossible to say, at least for another 10 years or so, once the newer approach begins to reach maturity. By then there will be even newer products available and some small number of “early adopters” will try out the newer products to see how they work. 

People in business are entirely comfortable with a fluid environment like this. Academics and bureaucrats are not. They want programs that are universal and unchanging. Social Security today is very much like Social Security was 70 years ago, and if some people aren’t pleased with it, too bad. Most people like it fine and minorities of voters don’t matter.  

Meanwhile, we are in the midst of a transformation in American health care. Not everything about consumer directed health care will succeed, but the overwhelming preponderance of the evidence says it is working exactly as it was intended to work. Policy-makers and academics who ignore or deny this development are missing out on the most significant change in health care of their lifetimes.

ADDENDUM

The data below was not included in the study as it was originally published in November 2007 by Consumers for Health Care Choices.  

The Mercer consulting firm has issued a report on a new survey it conducted on employer benefits. It finds that the increases in benefits costs has declined substantially since 2002, and leveled off at around 6 percent annually – still double inflation. Smaller employers (under 200 workers) are terminating coverage. The percentage providing benefits has dropped from 69 percent in 2001 to 61 percent in 2007. The remaining employers have slowed trend through the use of health management and employee consumerism.

CDHPs (both HSAs and HRAs) have grown substantially from 1 percent of covered workers in 2005 to 3 percent in 2006 to 5 percent in 2007, and are likely to grow further in 2008. Employers with fewer than 500 workers increased their offering of CDHPs from 5 percent in 2006 to 7 percent in 2007, to a projected 11 percent in 2008. For large employers (500–20,000 workers), it has gone from 11 percent in 2006, to 14 percent in 2007 and 18 percent in 2008. For jumbos (over 20,000 workers) it was 37 percent in 2006, 41 percent in 2006, and 43 percent in 2008.

The survey says, “Evidence that the (CDHP) plans are cost-effective is accumulating. CDHPs delivered substantially lower cost per employee than either PPOs or HMOs in 2007. CDHP cost averaged $5,970 per employee, compared to $7,120 for HMOs and $7,352 for PPOs. Of the two types of CDHPs, HSA-based plans were less expensive than HRA-based plans ($5,679 compared to $6,224).”

Trend between 2006 and 2007 was also favorable to CDHPs, suggesting it is not just a one-time saving, but an ongoing advantage. The trend for CDHPs was 3.5 percent, compared to 7.6 percent for HMOs, 6.1 percent for PPOs, and 11.2 percent for POS plans.

Mercer says this cost difference is not due only to higher deductibles, but because the “account feature encourages more careful health spending.” Comparing CDHPs to high-deductible PPOs still shows a savings of $700 per employee per year.

When given a choice of plan, a greater percentage of employees are choosing the CDHP, up from 21 percent in 2005 to 29 percent in 2007.

The survey also asked about employer support of reform proposals. Only 23 percent support a “pay or play” mandate on employers and only the same number (23 percent) support an individual mandate. Over twice as many oppose these ideas.

SOURCE:
http://www.mercer.com/referencecontent.jhtml?idContent=1287790

Endnotes

Greg Scandlen is the president and founder of Consumers for Health Care Choices, a non-partisan, non-profit membership organization aimed at empowering consumers in the health care system, and a senior fellow for the Georgia Public Policy Foundation, an independent think tank that proposes practical, market-oriented approaches to public policy to improve the lives of Georgians. Nothing written here is to be construed as necessarily reflecting the views of the Georgia Public Policy Foundation or as an attempt to aid or hinder the passage of any bill before the U.S. Congress or the Georgia Legislature.

© Georgia Public Policy Foundation (March 17, 2008). Permission to reprint in whole or in part is hereby granted, provided the author and his affiliations are cited.