Reducing Health Care Inefficiencies

Can Health IT save us from ourselves?  See the Yahoo Article on another assessment of health care spending in the U.S. and how much money is wasted in service delivery costs.  According to this article, about 1/2 of the spending of the U.S. on health care is wasted in inefficient use of resources.  Now, if you read the article, you will note that the top 8 items on their list only add up to about $600 billion, where the claim is that $1.2 trillion is lost (and you would think that a bunch of accountants could add, so maybe the journalists misread the fine print on the analysis), but even if you accept that much as the cost of inefficiency, that is more than it costs for the Medicare program in the U.S.

One of the big items on the list is inefficiencies with insurers who “magically deny” claims or otherwise require far too much in order for a provider to get paid appropriately.  I find it interesting that this remains on the list of problems.  In 1996, HIPAA was originally passed by Congress.  Part of HIPAA was to mandate that, through regulation, standards be developed for the electronic transfer of information between insurers and providers of health care, including claims.  The regulations eventually required that all or substantially all providers be able to submit claims electronically, which, one would expect, would be more efficient than the manual processing of paper claim forms.

So, if the auditors suggest that we still are wasting $200 billion per year on inefficient data exchanges with insurers, perhaps this deserves more focus.

Getting paid by insurers happens at the end of the process of service delivery to patients by providers.  At the beginning, patients present to the doctor’s office with a problem, see a Medical Assistant or Nurse for preliminary weights and measures (like blood pressure and weight, etc.), see the physician, CRNP or physician’s assistant, who may then refer the patient to another provider, write a prescription, make other suggestions to the patient, require that the patient get lab work to rule out certain causes, and so on.  At the conclusion of the visit, the physician will document, diagnose, and generate a financial transaction that must be processed and submitted to an insurer for payment.

The patient then will see other providers, the lab, the pharmacy, and perhaps come back for a follow-up visit with the physician.  All of the steps in the process involve data transfer between several information systems, often housed in several different facilities, with different standards and different purposes.  A key for a physician to get paid at all is to have accurate insurance information about the patient.  Surprisingly, patients are not necessarily the best source of this information.  However, insurers are apparently no better at knowing this on average.  Otherwise, it would follow that we would already have regional databases or a national database of eligibility data available for all providers.  I assert this because the standards for eligibility data have been around for a fair amount of time in the form of the ANSI X12 standard, but still there is a fair amount of lost dollars in the claims processing area of health care.

Perhaps this is so because providers want to get paid but insurers don’t have a good reason to pay them.  Insurers do benefit from holding onto capital to accrue interest on it.  The longer an insurer can do this, the more interest on the investment they collect, which goes straight to their bottom line.  ARRA’s incentive system requires that physicians meaningfully use health IT and participate in some form of a health information exchange.  But there is no comparable set of incentives for insurers to participate in HIE’s, or to incentive providers.

For example, this could be achieved by insurers preferring providers with health IT in place compared to those that don’t.  Another example would be for insurers to pay incentives to providers for a higher degree of clinical outcomes (only possible if the providers can produce useful and independently verifiable data such as lab information, which is really only possible through the use of an HIE).  The market may figure this out on its own, but I honestly doubt it.  Perhaps the feds will pick up on this market failure and intervene to start improving efficiencies in this area in either health reform now or in ARRA part II in the next several years.

Obama & Health Care IT

President Obama’s plan (published here) (the “Plan”) describes a multi-part approach to expanding the amount of health insurance available to those without insurance while attempting to reduce the costs of providing health care to Americans.  A portion of this plan involves the expansion of health information technology to help reduce the costs of administering health care.  On page 9 of the Plan, paper medical records are identified as a health care expense which can be reduced through records computerization.  The Plan cites a study by the RAND group (published here) (“RAND”) that indicates that the processing of paper claims costs twice as much as processing electronic claims.

Estimated Savings and Quality Improvements by Adoption of Health IT

The RAND group suggests that fully implemented health IT would save the nation approximately $42 billion annually, and would cost the nation’s health care system approximately $7.6 billion to implement.  RAND at 3.  According to their review of the literature on health IT adoption, approximately 20% of providers in 2005 had adopted an information system (which may have several meanings from patient reminder systems to clinical decision support).  RAND at 20-21.  Full implementation of health IT would require a substantial number of providers to convert to regular use in order for the total savings identified by RAND to be realized.  RAND estimated that in 2005 there were approximately 442,000 providers in the U.S.; this suggests that about 353,000 providers would need to convert from paper to electronic systems before the full savings to the health system would be realized.  RAND at 20.

Areas of savings in the outpatient setting noted include: transcription, chart pulls, laboratory tests, drug utilization, and radiology.  RAND at 21.  Areas of savings in the inpatient setting noted include: reduction of unproductive nursing time, laboratory testing, drug utilization, chart pulls and paper chart maintenance, and reduction of length of stay in the hospital.  RAND at 36.  Savings on the inpatient side account for approximately 2/3rds of the total savings, and the largest area of annual savings is tied to the reduction in the length of stay of patients as a result of the adoption of health IT.  Id. This overall cost savings is based on adoption of health records by virtually all health care providers in a 15 year period; the total savings to the health system during that time would total about $627 billion.  Id.

The Plan also discusses increasing the quality of health care delivered to all patients through the implementation of disease management programs (which are driven by health data of individual patients to monitor progress and outcomes), and the “realignment” of provider reimbursement with quality outcomes.  Plan at 7.  Realignment typically occurs when health insurance plans pay not for the total visits billed by a provider, but based on some kind of quality measure that tracks how well patients are doing in managing their health condition.  This is also driven by the availability of reliable health outcomes data (for example, the hemoglobin a1c test results of patients with diabetes over time, and the percentage that report a result under the “normal” or expected value).

The Trouble with Adoption of Health IT

Adopting health IT systems, however, is no small feat.  Systems have been available to the health care infrastructure for a substantial period of time (Centricity, a health information system now owned by General Electric, was originally developed by Medicalogic in the mid-80’s and became popular in the 1990s).  See Article.  In 2000, Medicalogic had penetrated the practices of about 12,000 physicians in the U.S., or around 3% of the total market, and was described then as the market leader in electronic medical records (which perhaps a total of 10% of the market had adopted a system by that time).  Using RAND’s analysis, five years passed and 20% of physicians had adopted some form of health IT.

If market penetration is to double every five years, by 2010, 40% of physicians should be using a health IT system, and by 2015, 80% should have adopted such a system.  (Admittedly, this assertion is weak because there is not sufficient data in this article to support this assertion.  In addition, adoption rates tend to follow a parabolic rather than a linear pattern, so that larger numbers of adopters join the crowd as time progresses.  But, dear reader, please feel free to comment with specifics to help improve the quality of this article!)

The New England Journal of Medicine, with a likely more restrictive definition of health IT, found that less than 13% had adopted such a system as of 2008, based on their sample of 2,758 physicians.  Article here.  An article in the Journal of Evaluation of Clinical practice reported that about 18% of the practices it surveyed (847 in total) had an electronic health record in use in 2008.  Article here.  (“JECP”)  As RAND had pointed out in its own literature search conducted in 2005, the definitions of health IT vary widely across the empirical surveys conducted, so an accurate estimate of market penetration is hard to come by.  However, it does appear that the number of practices that have adopted general health IT is not significantly higher than in 2005.

An interesting article suggested that some of the problem with health IT adoption may be regional – that some regions of the country tend to have a slower adoption rate of technology in general, which would tend to slow down the adoption of health IT in those areas.  Article here.  The JECP survey also indicated that specialty practices and smaller practices tend to be slower to adopt health IT as compared to their primary care provider counterparts.  Access to adequate capital to fund health IT purchases is an obvious reason for not implementing such systems.  Id. I would also posit that the adoption of health IT does not generally distinguish health care providers in the market of health care delivery (physicians don’t advertise that they have a health record system).  It would be interesting if patients could receive information on average health outcomes by physician when researching who they want to use for medical services (only possible if health IT is widely adopted and there is general consent to the publication of such data, which today is putting the cart before the horse).

There is, therefore, a market failure in that, if we accept that health IT reduces medical costs or improves outcomes over time, the market has not made a concerted effort to adopt this technology.  The Plan puts forward capital to help implement records and has an incentives component that rewards improved health outcomes.  Time will tell if these investments and market changes will actually reduce health care costs in the U.S.