What comes next in healthcare reform?

The passage of House Bill 3590 was billed as a momentous event, the first moderately successful effort at universal healthcare in the United States. Unfortunately the passage of House Bill 3590 also sets the United States on a perilous economic path. The reason for this concern is that House Bill 3590 at its core by itself an incomplete bill. Opponents continually cite the $935-940 billion dollar price tag associated with the bill as economically crippling whereas proponents counter that the most recent CBO report estimates a $118-138 billion dollar reduction in the deficit over the next decade and another $1-1.2 trillion dollar reduction over the next decade after that. However, the CBO uses wildly optimistic assumptions and some mathematical incongruities (common in government accounting) to come to those conclusions. Even then despite the assumptions used, to suggest any level of accuracy in a prediction 20 years from now on an issue as dynamic as healthcare is foolishness regardless of whether the analysis yields a positive or negative economic prediction. For those that place their faith in these types of predictions, go down to the local automobile dealership and ask to see the latest model of the flying car.

Originally the primary objective behind healthcare reform was an intelligent and noble one, albeit with a misguided methodology (attempting to do so much in one single omni-bill was rather silly for the complicated and problematic issue that is healthcare). However, two elements influenced the original objective eventually resulting in House Bill 3590, which was the same bill passed by the Senate months earlier. First, it appears that the Obama Administration and a large number of Democrats in Congress arrived at the conclusion that due to the length and ferocity of the debate surrounding healthcare that to not pass anything would be viewed as an unrecoverable failure. Whether justified or not, not passing any type of bill related to healthcare reform would have resulted in a severe handicapping of President Obama’s remaining term, which would have made it difficult to undertake any future domestic policies. The very contemplation of such a situation is sad because the Republicans in Congress did nothing of significance to aid in the generation of a positive bill and appropriate healthcare reform. Therefore, there was desperation to pass something, regardless of whether or not it was proper to do so.

Second, the proper foundations were not established before addressing the issue of the uninsured. The 40+ million uninsured individuals in the United States has always been the sexy hot-button issue when citing the problems in the healthcare system. However, that problem cannot be solved in a long-term economically stable manner without addressing other problems first. The entire issue reminds one of the scene from Apollo 13 where Jack Swigert (Kevin Bacon) raised concern about coming in too shallow and Houston not informing them and Jim Lovell (Tom Hanks) replies that there are a thousand things that have to happen in order and Jack was talking about number 692 when they were only on number 8. Dealing with the uninsured was addressed much earlier than it should have been. Although it is not too late these earlier problems must now be addressed, otherwise all of the good feelings that have resonated among the supporters of House Bill 3590 will eventually give way in the face of a much bigger problem in healthcare than the uninsured. Below are the five most important problems in healthcare.

Problem – No permanent solution to Medicare reimbursement payments that currently do not increases appropriately;

Why this is Important – Since mid 2008 a number of medical institutions including the Mayo Clinic have turned away individuals on Medicare because the amount of money they receive through Medicare for treatment is not enough to cover costs associated in treatment. Thus the institution must either lose money or refuses care. Clearly if medical costs continue to rise especially those paid for by Medicare, which is a strong possibility due to the increasing poor health of seniors and those soon becoming seniors, more and more medical institutions will have to turn away Medicare patients.

Discussion of a Solution – Congress has tried many times to create a permanent solution, but has run into problems with ensuring adherence to Pay-go statutes. The fact is that there are too many seniors that require the services of Medicare on Medicare. Therefore, it is difficult to raise reimbursement rates without increasing the deficit or reducing costs (which would indirectly increases reimbursement rates). This new law institutes another temporary increase in reimbursement rates to cover the next couple of years, but nothing different or substantial. It also attempts to explore the possibility of bundling Medicare payments as a means to accomplish the indirect increase methodology by shifting the cost burden onto the medical institution and away from the government. For more information regarding why this problem persists see a previous post on healthcare here.

http://bastionofreason.blogspot.com/2009/11/addressing-problems-in-health-care.html

The attempted use of Medicare bundle payments aim to change the complexion of how medical services are reimbursed. Instead of receiving payment for each service rendered, hospitals are going to be paid a single lump sum payment for each individual. The idea is that the hospital will maximize intelligent care and be able to treat the patient for a much lower cost because the focus will be reducing wasteful and unnecessary tests and treatments because the total profit/loss will be based on outcome. In this payment environment running a battery of tests just because the patient wants them or just for the heck of it will be detrimental to the hospital/physician because each test is no longer reimbursed individually, the costs of those tests will come from the lump sum payment. Thus under the bundle system high volume will lose money not make money.

However, whether such a method works is entirely dependent on the size of the bundle payment. The Mayo Clinic is probably the most efficient healthcare provider in the country maybe even the world, and over the last couple of years it has been one of the medical institutions that has been forced to start turning away Medicare patients due to loss of funds through treatment. Basically Medicare reimbursement payments were so low that the most cost-effective treatment center in the country could not break even, let alone make money, when treating Medicare patients. Such a result demonstrates that using the Mayo Clinic for a reimbursement floor would be practical. If the Mayo Clinic cannot make money from the bundle payment, the payment is too low. However, the Mayo Clinic cannot be used as an inflexible floor because it is the best of the best due to its organization and personnel, which cannot be emulated across the country. Thus these bundles need to be higher than the floor because even high-quality medical institutions are not 100% accurate and a single miss could lose a lot of money.

Another potential problem with bundled payments is that their profitability is dependent on patient behavior. It is not the fault of the medical institution if a patient comes in with what originally appears to be a single-occurrence condition and after further examination that patient is diagnosed with an expensive and chronic condition. What is the medical institution supposed to do in such a situation, say ‘sorry you have such a detrimental condition, but we’ll lose money treating you, so we’re going to have to cut you loose…’? In addition, the patient may avoid medical attention for a small problem that later evolves into a much larger and more costly problem that the medical institution has to absorb. Thus to avoid these realities a medium system will have to be generated between 100% bundle and 100% fee for service if reimbursement rates are going to be sufficient enough that medical institutions can still afford to treat Medicare patients, but also avoid waste.

Problem – Lack of Primary Care Physicians;

Why this is Important – A lack of primary care physicians is a huge problem because the chief reason most economists believe insuring the uninsured is a win for cost control and the economy is that those uninsured individuals will now use their insurance for preventative care to reduce the probability that they develop higher cost ailments. However, if there are a lack of primary care physicians those newly insured individuals will have to go to the emergency room for preventative care, which is not the job of the ER and would create a greater bottleneck effect further reducing the efficiency of ERs and probably costing lives. A number of initiatives in the new law depend on preventative measures to reduce costs; however, with a lack of primary care physicians those preventative measures will become less probable. Thus without an increase in primary care physicians insuring the uninsured will actually increase total medical costs and the overall deficit and those that are already insured will have a higher expectation of cost due to a reduced ability to undertake preventative measures under a physician’s direct supervision.

Discussion of a Solution – The American Academy of Family Physicians predicts a shortfall of 40,000 primary care physicians by 2020.1 The new law attempts to combat that shortage by increasing funding for training programs (usually funded through Medicare funding), more scholarship and loans for those entering primary care or committing to primary care out of medical school or loan forgiveness for those that start a practice into a ‘high-need’ region. Unfortunately a vast majority of this new funding only appears to affect those entering medical school, which implies that any effort from this bill to increase the number of primary care physicians will be delayed by at least 6-10 years (due to completion of medical school and residency). Additionally the bill encourages medical schools to open new residence spots to increase the overall enrollment number because a number of talented individuals are still unable to enter medical school due to enrollment caps.

In addition to the delay the funding alone will probably not be enough to wean new students away from the monetary and time rewards that come from specialization (this contention assumes that the new funding does not end up violating the Stark Rules). Physicians that specialize in a particular field are paid more money and have to work fewer hours in a more homogenous environment; a little help with student loans is probably not going to drive most of them from a given specialty to primary care, which at its core is a more arduous job. In fact it is not common for the pay differential to be in the range of 100 – 250% more per year. Therefore, other attempts need to be made to increase the number of primary care physicians beyond simple salary concerns. The problem is in a capitalistic market it is difficult to influence such a change without triggering some form of lawsuit.

One possible means to create a specialist shift is to create statistical approval limits on medical care. Basically if a medical procedure does not aid in the diagnosis of a perceived condition in a statistically significant manner then the government or private insurance company has the option of not covering the procedure. The application of such a rule would reduce the number of unnecessary diagnostic tests and procedures, which in turn could reduce the overall salaries of the specialists that carryout those tests further closing the gap between primary care physicians as well as reduce overall medical care costs. However, the chief problem with this idea is fighting the suggestion that such statistical restrictions promote ‘rationing’ of medical care and the simple fact that the general public is not experienced in using statistics to make decisions and would probably not understand why one test is not covered by insurance as opposed to a similar test.

Due to the time required to train new physicians as well as any turn-around time to reverse the downward trend of medical school students selecting specialized fields over general practice field a more immediate strategy needs to be applied. The new law focuses on the application of ‘medical homes’ to reduce the influence of the current and short-term future shortfall. A ‘medical home’ is a facility that incorporates teams of physicians, nurses and other specialized personnel to provide general primary care services largely to patients suffering from chronic conditions. Under the umbrella of a ‘medical home’ physicians can combine their patient load and their resources in an attempt to expand hours and general preventative communication to hopefully increase the probability of lowering costs while maintaining at least the same level of care, if not improving it. Early test models of ‘medical homes’ work well enough, but there is no real information regarding how scalable they are or how they work in more isolated rural communities. Overall the trend delineating the reduction in primary care physicians did not just start a few years ago, it has been happening for quite some time. This was a problem that should have been addressed at least 5 years ago, sadly the late start means some significant growing pains.

Problem – Emergency Room Streamlining and Reform;

Why this is Important – The second element in increasing the probability that insuring the uninsured actually lowers healthcare costs is ER reform. Under this law soon 32 million individuals will have the ability to visit the ER without being stuck with the bill, so these individuals will more than likely visit the ER more often and sooner than they would normally. Quick side note – Most people incorrectly believe that an uninsured individual receives free medical care when going to an ER. This is not correct. The Emergency Medical Treatment and Active Labor Act of 1986 only obligates ERs to care for individuals regardless of ability to pay. These uninsured individuals that receive care from an ER still receive a bill for the services rendered. If they are unable to pay the bill then their credit score is negatively affected and if the hospital/physician so desires they can be sued for the amount. This billing is why individuals do not go to the ER for every little thing that may be wrong with them.

Most medical experts believe more frequent primary care physician or ER visits to be a good thing because most potential detrimental conditions will be caught at an earlier stage of development making them cost less to treat. However, if generic ER input and output streams remain the same, these more frequent ER visits by more individuals will significantly increase wait time and procedural costs to the ER either forcing an increase in medical costs or more ERs going out-of-business. The coverage breadth represented by this law most closely models that implemented by Massachusetts in 2006. In response to its ‘universal healthcare’ program, Massachusetts’s ERs have receive a much greater input and are having significant problems coping with the increase. It is reasonable to presume that such is the fate of other ERs across the nation unless steps are taken to neutralize this problem. Therefore, an essential element that must be addressed is streamlining both the input and output elements in an ER to reduce the influence of overcrowding as much as possible.

Discussion of a Solution – Addressing ER reform stems from one of two possible strategies, reducing the input or hastening the output. With regards to the input the shortage of primary care physicians is especially important in the issue of ER reform. Recall that the uninsured typically wait too long to go the ER to receive medical treatment largely due to the fact that they do not have insurance. However, even if these individuals were given insurance if there were still a lack of primary care physicians then there would only be a very small shift in the probability that these individuals would not go the ER because of the previously noted time discrepancy. In fact there may be reason to believe that insuring the uninsured without any change in the number of available primary care physicians will increase ER overcrowding because conditions can be treated in less advanced stages over a situation where the individual does not have insurance for a lower cost against the insurance company. Such a contention is supported by the results from Massachusetts.

A second possibility for reducing the rate of input is to integrate an electronic health record system into ERs. Such a system offers the potential to create a more effective triage system, which can reduce wait times as well as assuage tempers from other patients who feel they are getting short-changed. Any reduction by ER staff in having to deal with individuals they are not currently registering or treating can be better directed at those activities. The real solution behind altering the rate and total amount of the input in an ER has to do with many outside components that also need to be addressed in the medical world like lack of primary care physicians, streamlining medical records and diagnostic procedures and increasing application of prevention methods for patients.

The second component in overcrowding has to do with output speed. One of the reasons for the increase in ER crowding is a decrease in turnover due to a lack of available nurses. One of the only silver linings that may have come from the recession is an increased interest in nursing because of job availability. Nurses typically have to do more work than primary care or emergency care physicians, get paid even less and have to have a significant portion of the education that is required of a physician. With these conditions it is not surprising that the occupational field of nursing is prone to shortages. Regardless of any other factor, processing time is negatively influenced if there is no one to conduct the processing and aid in the in-patient care regardless of what technology is available or what the patients are suffering from.

Another important factor for the lack of turnover is the lack of beds for those that need to remain in the hospital for further observation and/or treatment. Obviously not everyone that comes into the ER has a condition that allows for a same day discharge. In fact common sense would imply that such a reality should not be the case for a number of individuals visiting the ER, especially the elderly. For example suppose someone is rushed to the ER after getting into an automobile accident, odds are that individual will need to stay at least one night in the in-patient unit. However, if there are no available beds in the in-patient unit that individual will have to stay in the ER or be moved somewhere else after the initial round of treatment, which will reduce the rate of recovery and contribute to over-crowding.

Problem – Organization and maintenance of medical records;

Why this is Important – With 32+ million new individuals acquiring health insurance, one can anticipate those individuals will make use of that insurance. For those that are healthy and were trying to save money by not having health insurance, their introduction into or the updating of their medical records takes a backseat to records representing individuals that were denied health insurance due to pre-existing conditions. It is reasonable to anticipate that these individuals will see physicians and hospital stays at a much higher rate than the more healthy individuals thus these individuals need to have organized medical records to ensure efficient administration of proper medical procedures. Most pundits have presumed that electronic health records (EHRs) will be incorporated into the healthcare infrastructure to accomplish this goal.

Discussion of a Solution – Despite the funds earmarked for EHRs from the American Recovery and Reinvestment Act of 2009 there are still questions regarding their mass implementation. There are two problems that surround EHR adoption. The first problem is the informational disconnect between those that provide the service of upgrading hospitals to an EHR system and those same hospitals. For example one of the largest healthcare software installation companies is EPIC. EPIC claims that over 90 million individuals are affected by their system based on the number of hospitals that have received its services. The concern is that such a statement seems to run counter to the findings that only approximately 17% of U.S. physicians use a minimally functional or a comprehensive electronic records system.2 So what is the difference in criteria between those in the academic community versus those in industry to classify a hospital as having a viable electronic healthcare record system? Until a universal standard between both parties and even the government itself is developed it will be difficult to identify what hospitals have what type of systems from a statistical perspective which will limit the rate of installation severely limiting the ability of hospitals to share information, which is one of the chief advantages cited by those that support the mass adoption of EHRs.

The second problem is that most physicians still do not use electronic record systems in full or even at the least simple basic EHRs. As mentioned above total level of EHR installation seems to depend on the definition used for ‘installation’. One of the most pervasive studies of installation through hospital surveys concluded that only 1.5% of hospitals have a comprehensive electronic-records system (all clinical units use it), an additional 7.6% have a basic system (at least one clinical unit), but in total only 17% of hospitals have the basic computerized provider-order entry for medications.2 A second study concluded that approximately 16.5% (16.41%) of all physicians use the most basic form of EHRs.3 Both studies concluded that a majority of that installation had occurred in large hospitals that employ at least 50 physicians.

The reason behind this choice needs to be identified because with the level of incentive money both the federal government ($44,000 per physician)3 and private institutions ($40,000 - $60,000 per physician)3 are providing, is money still the principle obstacle? It makes sense to presume that money still is the principle obstacle based on the installation pattern from the above studies. Larger hospitals are going to have greater financial resources available to pay for installation as well as save more time and money from the installation. However, this blog has made the argument before and still maintains that the lack of a cohesive easily assessable information network regarding the installation of EHRs is also a meaningful obstacle. Without an easy to understand pro-con comparison information portal that hospital administrators and physicians can reference, these individuals have to do more leg-work and research on their own adding to the opportunity costs and training time required for the installation and maintenance of an EHR system, especially for smaller practices that only include 1-5 physicians. There is no reason for the federal government to continue to wait on establishing this database as an independent objective auditor and evaluator of EHR services provided by private industry.

System instillation has become easier in the last few years due to widespread software licensing reducing costs and less rigid data entry parameters.3 Also the evolution of EHR systems with regards to additional features like bill creation, electronic prescribing and coding functions have made EHR systems more attractive relative to their costs.3 However, the evolution of EHR systems need to expand to include a larger diagnostic slant to aid in efficient record keeping regarding what has been attempted and recommendations for further treatment or diagnosis. These types of records will be critical to any successful attempt to bundle payments to quality of care because if an accurate record of care does not exist how can one be judged on value or quality of care offered? It is true that widespread adoption of a new currently unknown tracking strategy may replace EHRs, but so much hope and hype has been placed in EHR systems for streamlining care, it is unlikely that any new idea would become widespread in a reasonable amount of time.

This rationality is also why implementation to smaller networks of physicians and hospitals must occur. If bundling becomes the principle manner in which Medicare reimburses physicians smaller practices that do not possess at least a basic EHR system, if not the complete systems, will be placed at a significant disadvantage and may have to turn away Medicare patients creating ‘lack of treatment pockets’ in various communities, especially those in rural areas away from major hospitals. Other issues that need to be addressed are patient privacy (especially with the prevalence of computer hackers in modern times) and the possibility of some form of adaptor system that can ease communication between two different types of EHR systems. It is rational to conclude that not all hospitals will have the same EHR system (due to competition each system will have its own nuances and features), thus when moving information from one system to another a protocol must be in place to ensure the fewest number of errors in the transfer as possible.

Problem – Exponential increases in overall healthcare costs;

Why this is Important – The importance of this problem is self-explanatory. If healthcare costs continue to increase at their current rate, the healthcare system in the United States will collapse.

Discussion of a Solution – At the start of the discussion regarding healthcare reform the goal of reducing the overall costs in the future to a more manageable level was one of the primary goals. Unfortunately, strong ideas that have a high probability of providing cost controls have all but vanished from the newly passed bill. The new law attempts to reduce costs through three different methods. However, a critical analysis of these three methods does not instill much confidence that any of them will be overly successful, especially to the level that is required.

The first method thought to control costs in the future is the previously discussed bundle Medicare payments. As discussed above the idea is to shift the focus on patient care, but the bundle idea almost tries to suggest that a large amount of the waste in the medical care system is due to greedy doctors just trying to pad the bill. In reality these bundles seem to have almost as much of a chance at bankrupting the medical care facility than reducing overall healthcare costs. Even if successful, bundle payment will create significant treatment variance for patients and their respective hospitals as some patients will be quite cheap for the hospital to treat whereas others will be cripplingly expensive. Such variance may increase the probability that certain patients are refused care. For the bundle system to reduce long-term costs without significant disruption in care they will need to be carefully organized and highly classified with regards to what conditions are being treated and what new conditions can arise from the escalation of those conditions.

The second method thought to control costs in the future is the frequently mentioned health insurance exchange. Simple market principles drive the confidence in the ability of the exchange to reduce costs. The mindset is that one of the major reasons healthcare costs are so high is the lack of competition between insurance providers. In most states no more than 3 major insurance providers exist, thus if you do not like the premium price that you receive from Company A there is little you can do, but go without insurance. Thus, proponents believe that creating a new marketplace for insurance companies will expand the level of competition and force insurance companies to cater to its potential customers. Also with additional companies in the mix consumers will have more options available to identify the best available program for them, which will make their coverage more efficient more than likely reducing costs.

Unfortunately proponents only focus on the generic competitive marketplace when making these claims. Insurance is not a generic competitive marketplace. In the generic marketplace a supplier can make mistakes and still have the ability to recover because those mistakes are not crippling. In the insurance marketplace a bit of bad luck for both the customer and the company (a previously healthy individual develops cancer) can significantly impact the bottom line of the company. This feature of the insurance market makes it very difficult for new providers to emerge from scratch without significant backing. Here is another instance where the public option would have been a valuable tool because it could guarantee stability and longevity unlike any new private companies that emerge. Also with the new law stripping the ability of the insurance companies to reject the sickest individuals (not denial of care due to pre-existing conditions) the margin of error is further strained for the company.

Proponents seem to envision a world where 10+ insurance companies aggressively compete within a single state resulting in lower costs and maximization of care efficiency. Sadly this is not very probable as if companies cannot retain a significant amount of no/low cost customers they will go out-of-business. Under the current law it appears that the only way that multiple companies will be able to survive in a competitive environment is unified cooperation or collusion. Basically these companies decide as a group to evenly divide up all of the high cost customers so that no company is saddled with their unbalanced costs. However, such a design is highly improbable as remember capitalism is about competition not cooperation, if company A is on the ropes and needs company B to take some of its losses is company B really going to help? Overall one merely needs to look at the outcome of similar competitive measures in Vermont in the 70’s, where a multitude of insurance companies existed before competition whittled the number to single digits.

The third strategy for lowering medical costs is largely associated with the new tax on ‘Cadillac’ insurance plans. The idea behind the tax is to place a preemptive penalty on prospective waste. The general structure of the ‘Cadillac’ insurance plan is one with a high premium, but almost no deductible or co-pay thus encouraging use of medical services regardless of whether or not they are needed or statistically relevant. In addition, most of the expensive premiums that fund these plans are paid by employers, which receive a tax exemption further reducing their burden. However, the tax aims to forcibly change wasteful behavior through taxing insurance companies for plans that exceed $27,500 for a family or $10,200 for a single individual.

One of the chief problems is that this tax does not go into effect until 2018. In addition, the general target is too small because of an overestimation of waste through its use. Waste in the system is not confined to only plans that exceed the tax boundary. While it is true that there is a higher probability of waste in these plans, realistically most of the waste in medical treatment is volumetric spread-out over millions of insurance plans not concentrated in a few hundred super-plans. Also the tax could create a new twist on malpractice suits. Suppose that a patient elects not a receive procedure A that could predict severely detrimental condition A in order to avoid being taxed (because it makes rational sense that although the insurance company is liable for paying the tax, the cost of that tax will be passed on to the company or individual that holds the policy); later that individual develops detrimental condition A. It is highly likely that the patient will sue the physician for not detecting detrimental condition A. Can the physician cite tax avoidance as a meaningful defense?

As previously discussed limiting insurance coverage to diagnostic or therapeutic treatments that have demonstrated empirical statistical significance is another way to significantly reduce healthcare costs. Unfortunately although this method has been discussed in cost controls there are a lot of social barriers to applying such a method, especially with all of the confusion and misinformation concerning most of the more popular or mainstream techniques that would be removed from coverage under such a proposal. Sadly the most powerful weapon in using this technique would have been in the public option. The public option could impose the statistical restrictions where private insurance could choose to do the same or not use any statistical information to dictate coverage. Those that wanted to pay less, but have some popular, but ineffective treatment options not covered would go with the public option whereas those that wanted those procedures covered could sign up with the appropriate private insurer.

Another way to focus on waste reduction is to use preventative measures as a prerequisite for more advanced treatments. For example if an individual wants to receive treatment x they would need to demonstrate that they took an appropriate preventative measure like had a physical in the last 18 months. This method avoids the luck problem with the ‘Cadillac’ tax where an individual takes the proper precautions and still develops a chronic and expensive condition where the tax would further increase those costs because it would only apply to conditions that have a generalized cause and can be detected with a reasonable probability through general preventative measures.

Overall these are 5 important issues that must continue to be discussed in the coming years. To simply come to the conclusion that healthcare is solved now that this new law has passed would be a recipe for failure. Although the new law attempts to address some of these issues, none of the solutions are ironclad and most are still experimental. Steps that should be taken in the very near future are: 1) working with highly efficient organizations like the Mayo Clinic to define a general bundle pay scale that would reduce costs, but also allow for forgiveness of a small number of mistakes to avoid significant losses in treatment; this scale also must increase overall physician reimbursement; 2) generating an information portal so hospitals and physicians can more easily identify whether or not an existing EHR system would result in a more efficient/low cost practice; 3) have a serious discussion regarding the usefulness of incorporating statistical significance when covering a given medical procedure and how much money eliminating those statistically insignificant tests would save; 4) create a general reform structure for ERs and begin testing its application in ERs across the country;

Healthcare is a dynamic beast that must be carefully observed and analyzed so proper alterations can be made within the appropriate time guidelines. Otherwise, costs will continue to increase threatening the general stability and welfare of the United States and her citizens.

==
1. http://www.sciencedaily.com/releases/2008/06/080617111826.htm

2. Jha, Ashish, et, Al. “Use of Electronic Health Records in U.S. Hospitals.” New England Journal of Medicine. 2009. 360:16 1628-1638.

3. Shea, Steven and Hripcsak, George. “Accelerating the Use of Electronic Health Records in Physician Practices.” New England Journal of Medicine. 2010. 362:3 192-195.