This week’s edition of the Cavalcade of Risk is up at Colorado Health Insurance Insider.
You are currently browsing the monthly archive for May 2007.
Laurence Baker is a health economist at Stanford’s Center for Health Policy. Much of Mr. Baker’s work has dealt with how HMOs have affected care levels. Today I will briefly review three of Baker’s articles:
HMO Penetration and the Cost of Health Care (AER 1996)
In this paper, Baker and Corts look how HMO market penetration affected health care premiums. I think the article is most pertinent to the health care atmosphere in the late 80s and early 90s when the distinction between HMOs and other insurance plans was starker than it is now.
The authors argue that there are four reasons why increased HMO penetration may affect the cost of traditional insurance.
- Patient Self-Selection: Healthier patients may sort into HMO leaving traditional insurers with more unhealthy patients.
- Physician Selection: HMOs may attract physicians who prefer a more conservative style of medicine.
- Promulgation of Conservative Practice Styles: A higher level of HMO penetration may influence the ‘standard of care’ prevalent in a given metropolitan area.
- Cost shifting. If HMOs are good negotiators, providers may simply accept low margins for HMO patients and charge even higher rates to traditional insurance plans. This argument, however, is illogical if providers are assumed to be profit maximizers.
- Plan characteristics. Traditional insurers may make their plans less generous in order to compete with HMOs.
Using data from 1991, the authors do find that increased HMO market share decreases premiums when the HMOs first enter the market (i.e.: HMO market share is between 0-10%). Additional HMO market penetration (i.e.: HMO market share above 10%), however, is actually found to increase health insurance premium in a local area.
Effect of HMO Market Share on Cancer Screening
The authors posit that HMOs may be more likely to screen patients for cancer since these health plans are more cost-conscious and take a longer-term view of health care. Spillovers effects non-HMO providers may occur where 1) physician practice patterns change due to an increased HMO presence, 2) patients may be more exposed to information regarding cancer screening in areas with high HMO concentration, and 3) areas with a high HMO market share may attract providers who are more likely to screen patients.
The authors use the 1996 Medical Expenditure Panel Survey-Household Component (MEPS-HC) to test their hypothesis. HMO market concentration is measure by segmenting markets into highest, middle two and lowest quartiles, as well as by using a Hirschman-Herfindahl index (HHI). The authors find that an increase in HMO market share increases the probability of breast and cervical cancer screening, but does not affect the propensity for men to get a prostate exam.
Calculating HMO Market Shares
How do Baker and colleagues calculate HMO market share? A paper studying the factors association with mammography screening has an appendix which details how the HMO market shares were calculated. In the paper, the authors find that those who were more likely to be screened were younger, had smaller families, higher education and income, had a recent Pap smear; reported breast problems; lived in an area that had more mammography facilities with reminder systems, areas with higher HMO market shares and higher screen charges.
- Baker; Corts (1996) “HMO Penetration and the Cost of Health Care: Market Discipline or Market Segmentation?” The American Economic Review, Vol. 86, No. 2, (May, 1996), pp. 389-394.
- Baker; Phillips; Haas; Liang; Sonneborn (2004) “The Effect of Area HMO Market Share on Cancer Screening” Health Services Research 39 (6p1), 1751–1772.
- Phillips; Kerlikowske; Baker; Chang, Brown (1998) “Factors associated with women’s adherence to mammography screening guidelines.” HSR, 33:1, pp. 29-53.
Why is it taking so long for the medical industry to adopt information and communication technology? Is it a good thing that it is taking so long? This is the question addressed by an interesting and non-technical NBER working paper by Michael Christensen and Dahlia Remler (“ICT in Chronic Disease Care“)
What is Information and Communication Technology (ICT)
Christensen and Remler divide ICT into 4 categories:
- Technologies that support patient self care and education: These include websites (such as Medco Health) to help people with chronic conditions.
- Patient-provider and provider-provider communication: Email, targeted educational emails, etc.
- Electronic data storage and sharing: EMR
- Technologies that combine all of the above: Examples of this include clinical algorithm engines such as Aetna’s MedQuery.
Will ICT increase health costs?
The answer to this question is not clear. The authors note “In general, while technological improvement can allow us to provide the same level of health at a lower cost, it can potentially also make it worthwhile to spend even more resources on health care to greatly improve health.” Thus, ICT can result in significant efficiency gains to the health care sector without reducing the overall cost.
Why is it so difficult to adopt ICT?
There are many reasons why ICT adoption has been so slow:
- Network Externalities: Items such as EMR have this issue. An investment in EMR increases in value as the number of users increases. Since EMR usage is so fragmented at this point, the positive network externalities in this market are not yet accruing to those who have invested.
- Switching Costs: If a provider (or health plan) chooses a EMR system, and they choose one that becomes a non-dominant system, they may be forced to switch to a new network. Thus, all the fixed costs invested in the original EMR system will have been wasted. A provider can instead decide to wait before purchasing an EMR system in order to see which system emerges as the market leader. Thus, while not adopting EMR early on may cause short term losses, in the long run, it may be a wise decision in the presence of switching costs and thus slow adoption may be an optimal strategy. This issue is treated more thoroughly in the real option pricing theory.
- Consumer Responses Blunted: In regular markets, consumers receive higher quality at lower prices by ‘voting with their feet’ and purchasing only high value items. In the health care sector things are not so simple. There are 4 interlinked markets: health care services, health care insurance, the labor market and political economy. “economy). The result of all these complicated linkages is that market forces from consumers to the health care providers are far more indirect and blunter than are the market forces in most industries. The forces must be transmitted through each stage and each stage brings its own transaction costs.” One example the authors use is trying to persuade doctors to use email. Employers have motivation to demand that doctors supply this service to their employees to make their employees happier. Employees, however, value their job much more than whether their doctor has email or not, so direct employee pressure, while it exists, is likely a weak agent of change in the health care market.
The authors conclude the paper making the following wise pronouncement: “A tradeoff therefore exists between seeking more standardization to allow for greater adoption and less standardization to allow for greater product differentiation.”
Generic drugs dramatically reduce the cost of pharmaceuticals to patients. Can generic medical devices reduce the cost for hospital administrators?
Matthew Holt of The Health Care Blog interviews Richard Kuntz, CEO of Generic Medical Devices. Mr. Kuntz wants to reduce the cost of the implants, surgical implementation, and supplies used in medical care. The company will “…begin developing products that are off-patent, that have proven safety and efficacy, that have existing reimbursement, to reduce health care costs and help save Medicare.”
The full interview is here.
What exactly is Governor Arnold Schwarzenegger proposing in his health care initiative unveiled early this year? Below I briefly summarize his press release, as to what the reforms will entail and then follow with some of my comments.
Individual Mandate
- All individuals must have a minimum level of insurance.
Children
- Children of families whose income falls below 300% of the Federal Poverty Line (the FPL was $19,350 for a family of 4) are eligible for public health insurance. Those living with families below 100% of the FPL will receive Medi-Cal insurance while those living in families with income between 101-300% of the FPL will receive Healthy Families coverage. All people living in California are eligible regardless of their residency status.
- Parents will be responsible for providing health insurance for children in families with income over 300% of the FPL.
Adults
- Adults with incomes below 100% FPL are eligible for Medi-Cal. Adults with incomes between 100% and 250% FPL are eligible for coverage through a state purchasing pool operated by the Managed Risk Medical Insurance Board. Individual insurance premiums increase as follows:
- 100-150%: 3% of gross income
- 151-200%: 4% of gross income
- 201-250%: 6% of gross income
- Adults above 250% FPL will be responsible for purchasing their own health insurance, either on the private market or through work.
- Poor undocumented immigrants will receive care at the county level, through county clinics, UC hospitals, and safety net clinics.
What do I think of the plan? Expanding care to the poor is a desirable social goal. Giving the poor no choice as to what type of care they will receive, however, is undesirable. If Medi-Cal health insurance is of poor quality, the individual will experience a serious income shock if they attempt to leave Medi-Cal and enroll in private insurance and thus their choice of insurance plans is limited to one. Offering individuals the choice of having Medi-Cal/Healthy Families or receiving an equivalent contribution towards another private/employer-provided insurance plan would go a long ways towards increasing competition in this market.
Also, the plan gives a strong disincentive to work for those families which make near the 300% threshold. If a family of 4 were to earn $55,000, they would be eligible for the Healthy Families program with some premium. If they were to earn $60,000 (which is just above the 2005 FPL for a family of 4) then they would lose the Healthy Families coverage and have to find insurance on their own. The gradually increasing price of the Healthy Families program is a wise policy move, but is not sufficient to eliminate distortions due to non-linearities in the Income-Insurance benefit function.
The individual mandate is an interesting case. Is it right to dictate to families where they should be spending their limited resources? On the other hand, families could elect to go without health insurance and if they did fall ill they would be able to fall back on the government safety net. Thus, I do not hold a strong opinion on individual mandates at this time.
The VA has often been held up as an example as to why a centralized health care system works well. An interesting post from The Sentinel Effect looks at a North Carolina’s The News & Observer column (“VA Overselling its medical care“) stating that the VA system results are not as good as they claim.
Do you live in California? Are you interested in health care issues? Then you may be interested in the Cal Health Reform website.
The website has links to various health reform proposals including a summary of Governor Schwarzenegger’s health reform initiative. There is also a brief powerpoint presentation from health economist Jon Gruber of M.I.T. as well as an interesting Myths and Facts page.
A recent paper in the Health Services Research journal (“Hospice…“) looks at whether hospice care reduces hospitalizations for elderly terminally ill patients in nursing homes. In the introduction, authors Pedro Gozalo and Susan Miller state that there are two main implications which result from end-of-life hospitalizations:
“At the patient level, hospitalizations of frail NH [nursing home] residents have been shown to include hazards that negatively affect the quality of life (Creditor 1993) and, in many cases, are inappropriate (Saliba et al. 2000). At the policy level, hospitalizations represent the main component of total health care costs, particularly during the last few months of life. In an recent study using both Medicare and Medicaid claims for NH decedents in the state of Florida in 1999, Miller et al. (2004) found hospital expenditures to account on average for 78 percent of all expenditures in the last month of life among those patients that did not receive hospice and 33 percent among NH residents who had any hospice in the last 30 days of life.“
The main problem in determining whether or not hospice care reduces hospitalization is the issue of selection. Individuals who enroll in hospice care may be relatively ‘healthier’ than those who are hospitalized without hospice care. Also, it could be the case that hospice patients are ones that prefer less aggressive treatment methods. Local market idiosyncrasies can also lead to erroneous conclusions.
Econometrics
To take these selection issues into account the authors use an inverse-probability-of-treatment weighting (IPTW) [see Robin, Hernán and Brumback (2000) for more on IPTW methodology].
- Let H=1 if an individual enters hospice care and H=0 otherwise;
- let Y=1 if the individual is hospitalized and Y=0 otherwise.
- Also, let P(W)=P(H=1|W) which is the probability of being hospitalized conditional on W.
- W is a vector of both patient characteristics (X) and a set of hospice provider characteristics (Z).
Using the IPTW methodology, the authors regress Y on H and X and each observation is weighted by the following term:
- H/P(W) + (1-H)/[1-P(W)]
In other words, if the individual enters a hospice, they receive a weighting of P(W)-1, and if the individual does not enter a hospice the observation is weighted by [1-P(W)]-1. One problem with this method is that it assumes that the vector W adequately models the choice of hospice care and that there is no unobservable sorting into hospice compared to non-hospice care. To reduce the endogeneity of the hospice characteristics (Z), the authors wisely decide to use the characteristics of the hospice nearest to the individual, not the actual hospice chosen.
Results
The authors find that the most important determinants of hospice enrollment are: principal diagnosis of cancer and patient preferences for noncurative care. Further, the paper concludes that nursing homes “that choose to contract with hospices may be less likely to hospitalize their residents, even if [the] hospice was not present.” The authors find that one quarter of the hospice effect on hospitalization is due to patient preferences, but the hospice effect on hospitalization is still strong.
- Gozalo; Miller (2007) “Hospice Enrollment and Evaluation of Its Causal Effect on Hospitalization of Dying Nursing Home Patients” Health Services Research 42:2 (April 2007) pp. 587-610.
- Robin, Hernán and Brumback (2000) “Marginal structural models and causal inference in epidemiology” Epidemiology. Sep;11(5): pp. 550-60.
Does health insurance increase utilization of medical services? Economic theory generally predicts that it will. Health insurance decreases the price individuals pay for medical care and thus the equilibrium quantity of medical care used will increase.
A paper by Buchmueller, Grumbach, Kronick and Kahn (âEffect of Health Insurance on Medical Care Utilization…â?) examines this phenomenon in more detail. The paper is a nice blend reasoning based on both economic theory and empirical evidence.
Adverse Selection: For instance, economic theory predicts that if insurance companies do not have full information regarding the health level of its customers (which is likely) adverse selection will occur. This means that healthy individuals will purchase no (or less generous) health insurance while sicker individuals will purchase (or purchase more generous) health insurance. The authors claim that while this is generally true, it may not hold in the U.S. institutional environment. Most people under 65 receive their health insurance through full-time employment. Since it is less likely that sick people work, some studies (Buchmueller 1995; Stroupe, Kinney and Kniesner 2000; Blumberg and Nicholas 2001; Holahan 2001) find evidence that workers in poorer health are less likely to obtain employer-sponsored coverage.
The Uninsured: A paper by Dubay, Holahan and Cook (Health Affairs 2007) shows that there are 44.6 Americans under 65 [17.5% of this population] who are uninsured. Yet uninsurance does not imply a complete absence of care. âIn areas where there is a well-functioning safety net, the lack of insurance will not mean a complete lack of access to care, so the expansion of coverage will result in smaller changes in utilization than in localities where the uninsured have fewer options. In contrast, in many low-income areas, even persons with Medicaid face access problems because many physicians are not willing to accept Medicaid patients (Fossett and Peterson 1989; Fossett et al. 1992). In such areas where care to publicly insured patients is effectively rationed, increases in utilization from expanded coverage may be limited.â?
Type of Coverage: Due to data limitations, few studies have been able to look at the type of coverage offered to the uninsured. Different expansions of government provided health insurance may lead to different utilization results depending on how generous the insurance plan is. Is the insurance coverage more ‘managed?’ Are there gatekeepers? What specific procedures does the insurance cover? Few studies have answered this question in such a detailed manner.
Supply Side: If the government did decide to provide a more universal type of coverage what would be the impact? Most studies examine the patient demand side and deduce what the patient response would be if they became eligible for government-sponsored or government-subsidized insurance. It is likely, however, that providers will also respond to any changes in the manner in which they are compensated. According to a study of Quebec after Canada, âafter universal coverage was established, physicians shifted away from telephone consultations, which were not reimbursed under the new system, toward office visits, which were reimbursed (Enterline et al. 1973, 1975).â?
The paper goes on to review a number of studies which examine how expanding insurance coverage affected how various groups (i.e.: children and adults) utilize different medical care services (i.e.: outpatient visits, hospitalizations, etc.). One interesting theory which is tested is that health insurance will increase outpatient visits, but because outpatient visits can often catch a disease in its early stages, hospitalizations will decrease. The authors cite a paper by Dafny and Gruber (2000) which finds that â…Medicaid eligibility reduces the rate of avoidable hospitalizations by 3.4% while increasing the overall hospitalization rate.â?
- Buchmueller, Thomas; Grumbach, Kevin; Kronick, Richard; Kahn, James; (2005) “The Effect of Health Insurance on Medical Care Utilization and Implications for Insurance Expansion: A Review of the Literature” Medical Care Research and Review, Vol. 62, No. 1, pp. 3-30.
Robin Hanson has spent the last week blogging on his Overcoming Bias website asking individuals to sign a petitions to redo the RAND Health Insurance Experiment (HIE). For those who do not know what the RAND HIE, Mr. Hanson has two posts (#1 and #2) describing the experiment. I agree with Mr. Hanson when he states “If you remember only one medical study, it should be the RAND health insurance experiment.”
But is a new RAND-style HIE needed? The Overcoming Bias site states that the purpose of a new HIE would be to “…try to make clear the aggregate health effects of variations in aggregate medical spending, variations induced by feasible regimes of quality control, including free patient choice induced by a varying aggregate price.” The proposal is for a study to randomly assign 10,000 people to different insurance types (or to no insurance) and follow them over 10 years. Mr. Hanson claims that this will cost 1/2 billion dollars over the 10 years. This seems reasonable, but possibly a little conservative—it is equivalent to purchasing $5000 worth of health insurance for 10,000 people each year for 10 years. Is it worth it?
As a researcher, I of course would love to be able to have access to new, high-quality data from a randomized controlled trial. On the other hand, I do not like the idea of spending 1/2 billion of taxpayer dollars (or private foundation money). One could be using these funds to feed the poor, provide medical care to the sick, or spend the money researching for new cures.
In the end, I think repeating the RAND HIE for the 21st century will be a worthwhile endeavor, and I have decided to to sign the petition. By conducting this RCT, we can see how much (if at all) health insurance improves health levels. A RAND-styled RCT could separately look at difference forms of care: preventative care, primary care and specialist care. According to a Health Affairs article by Catlin et al. (“National Health Spending in 2005“), government spending reached $902.7 billion in 1995. Of this, $342 billion was spent on Medicare and $313.1 billion was spent on Medicaid. With these astronomical figures at hand, one can see that 1/2 billion dollars over 10 years is only $50 million per year–less than 0.01% of public health care spending per year. If this cost figure is too high for some people, I think a sensible thing to do would be to reduce the sample size to between 7500-8000 people instead of the 10,000 proposed by Mr. Hanson. Nevertheless, Americans need to know what they are getting in exchange for their dollars used to finance Medicare, and Medicaid as well as their own purchases of private health insurance.
Recent comments