Introduction Health Information Technology (HIT), in a general sense, is the adoption of technology in order to provide efficient and beneficial

results when catering to patients in the health care industry. Implementing various technologies, most notably Electronic Health Records (EHR) , would allow hospitals and clinics to centralize patient records, such as Electronic Medical Records (EMR) and Personal Health Records (PHR), securely, while allowing them to efficiently access those records to better diagnose, treat, and ideally, provide positive outcomes for current and future patients. According to a Health Services Research study of national attitudes towards HIT, 78% of Americans favor IT in healthcare, 78% believe it can improve quality of care, and 59% believe that it can lower costs in the long run.1 Though this study shows a positive acceptance from the general public, an underlying dilemma that hospitals and clinics face when adopting health IT, however, is the initial cost of implementation, ongoing coordination and maintenance, as well as numerous other costs associated with successfully using the previously mentioned technologies. Literature that concerns Health IT has shown that 92% of the articles reviewed had a positive outlook on the potential benefits that it can provide, and a general consensus has framed health IT as a very important factor in improving the general health of Americans. 2 Even with such optimism, there are many different factors and variables that hospitals will face that may decelerate the process, even with the relatively new legislation attempting to encourage and ensure successful implementation and use of health IT. The United States recession began in 2007, and in effect, there was a great hindrance to

1 (2011), Public Attitudes about Health Information Technology, and Its Relationship to Health Care Quality, Costs, and Privacy.

Health Services Research

shows predominantly positive results. Health Aff (Millwood)2011;30:464-471

the advancement and improvement of the Health Information Technology field,in addition to other aspects of overall hospital advancements. We are now beginning to see an improvement in our economy, and with that, hospitals are more prone to implement information technology into their health care provisions. In 2009, the Obama administration passed the American Recovery and Reinvestment Act (ARRA) in order to stimulate economic activity. The original expected expenditure for this act was $787 billion dollars, but was increased to $840 billion in 2012.3 This bill is intended for the overall well being of the United States economy, but will impact health care in many different ways. Relevant to HIT is the Health Information Technology for Economic and Clinical Health Act (HITECH), which is an unprecedented opportunity for healthcare providers to invest into the technological performance of their industry with the help of federal incentives to do so. The ARRA allocated $22 billion to the HITECH Act, $19.2 billion of which will be directed towards EHRs, and with this money, the government is hoping to not only encourage EHR adoption, but to pave the way with guidelines and certifications so healthcare providers can use technology in such a way that it will significantly improve various aspects of patient care. It is important to note that the U.S. Government sees this act, and the technology it advocates, not as an end to itself, but as a means to improve the overall health of Americans and efficiency of the healthcare system, as well as create more jobs with the resulted demand for HIT workers. Located within the Department of Health and Human Services (DHHS), the Office of the National Coordinator for Health Information Technology (ONCHIT) was created and mandated by the ARRA in order to coordinate HITECH by creating a set of standards for HIT, as well as the incentive program that looks to reward those hospitals that not only adopt EHRs, as already

3 (2009) The Recovery Act, http://www.recovery.gov/About/Pages/The_Act.aspx

mentioned, but be meaningful users of EHRs. According to the DHHS, the Federal Health IT Strategic Plan goals are to use IT in healthcare so improvements can be made on the overall quality of care by reducing errors, making healthcare more affordable by reducing costs, and provide better and thorough information that could be critical to healthcare professionals when diagnosing and treating patients.4 IT is all about information, and health information can be categorized as highly sensitive and critical, and because of this, the Health IT Strategic plan also includes goals to create accuracy, security, and efficiency when collecting, aggregating, and analyzing the information that can further improve patient care. These goals are relatively broad, but efficiently describe the desired outcomes that stem from HITECH. Criteria describing expected outcomes and guidelines, or meaningful uses, within the specified time frame were made by ONCHIT to supplement hospitals and clinics who are making the transition to EHRs and related technologies in accordance with their requirements. 5 The transition to EHRs is outlined into 3 stages that help promote healthcare providers to ultimately be able to capture all health data electronically, as well as have the ability to share the composed information from that data, to patients and healthcare providers. Stage 1 was expected to be met by the end of 2012, and include using Computerized Physician Order Entry (CPOE), making lab reports into EHRs, and sending patient data summaries to concerned physicians and patients. Stage 2, which will range from 2013 to 2015, seeks to make Personal Health Records accessible by respective patients, create an electronic summary record for patients that includes information such as immunization records, and make health alerts available to both physicians

4 (2011) The Federal Health IT Strategic Plan http://healthit.hhs.gov/portal/server.pt?

open=512&objID=1211&parentname=CommunityPage&parentid=2&mode=2
5 Release of the Federal Health IT Strategic Plan http://www.healthit.gov/buzz-blog/from-the-onc-desk/release-federal-health-

strategic-plan/

and patients that are time sensitive or urgent. The final stage, Stage 3, which is to be completed by the end of 2016, is the pinnacle of the previous 2 stages, and has the intended goal of giving physicians access to a centralized and comprehensive source of patient data to aid in clinical decisions, while also giving them the ability to have real time surveillance of patient care. 6 These measures cannot merely function on their own, but have to be meaningful in that they are able to provide sufficient improvements in the quality of care given to patients by reducing errors and increasing accuracy. Incentives authorized by the ARRA, which began to be dispersed in 2011, will only be given to hospitals that are meaningful users of certified Electronic Health Records and are following the strategic plan that ONCHIT created, which makes the meaningful use criteria very critical when adopting Health IT.7

Statement of the Problems In 2009, the New England Journal of Medicine reported that a mere 18% of U.S. physicians were using basic EHR systems.8 This figure increased to 36% in 2011, but the number of hospitals with basic EHR systems increased from only 6% to 12%. 85% of those same hospitals polled admitted that they recognized benefits of EHR, and planned to utilize the incentive program set forth by HITECH, by implementing such technologies.9 These figures represent a yearning for IT in healthcare, but present the initial problem that many healthcare providers are still living in the stone age, so to speak, when it comes to Health IT, and especially EHRs. As with any other business, there are some hospitals that have the budgets and

6CMS EHR Meaningful Use Overview https://www.cms.gov/EHRIncentivePrograms/30_Meaningful_Use.asp 7 Finding The Meaning In Meaningful Use http://www.nejm.org/doi/full/10.1056/NEJMsb1103659#t=article 8 (2009) Stimulating The Adoption of Health IT http://www.nejm.org/doi/full/10.1056/NEJMp0901592 9 HHS News Release http://www.hhs.gov/news/press/2012pres/02/20120217a.html

personnel to either have already begun, or are soon going to begin, creating an EHR system, without the high degree of financial concern that many hospitals possess. Few hospitals began EHR system development before the HITECH Act was passed, and even less had systems in place that would satisfy some of the requirements10. The incentives that can be earned from HITECH, which are dispersed by the Center for Medicare and Medicaid Services, is a reward for those hospitals that invest their own money into EHRs, and then become certified users of said EHRs. It is only after they are certified by one of three accrediting organizations (Certification Commission for Health Information Technology; the Drummond Group; and Infoguard Laboratories) that the provider can begin receiving the medicare and medicaid incentives dispersed by CMS. This is a concern that can be detrimental to HITECHs ultimate goal of efficient, interoperable, healthcare systems across America, because it rewards healthcare providers who are already well-off financially, and are able to implement such technologies, instead of helping those providers that cant afford to aggressively adopt EHRs. This can create a technological divide and accelerate the gap between affluent and struggling healthcare providers.11 If healthcare providers do not meet ONCHITs requirements, they will not receive incentives, but more importantly, they will be subject to penalties and a decrease in Medicare and Medicaid funding each year after Stage 3 is set to be completed. That in itself is pushing many providers to accept federal HIT requirements, because a decrease in funding can cause financial ruin for many of them. Since health IT can be a very expensive expenditure for providers, many of which are still using paper records, becoming a certified EHR user is going to

10 (2009) Implementation of the Federal Health Information Technology Initiative http://www.nejm.org/doi/full/10.1056/

NEJMsr1112158
11 (2011) Hospitals Scramble to Meet Deadlines for Adopting Electronic Health Records http://www.ncbi.nlm.nih.gov/pmc/articles/

PMC3046613/

be a very difficult, but necessary, objective to acquire. The American Hospital Association stated that they, remain concerned that the requirements (from ONCHIT) may be out of reach for many of Americas hospitals.12 This gloom outlook is the product of financial constraints and certification criteria that is almost impossible to meet for many U.S. healthcare providers within the specified timeframe. Healthcare systems will only implement emerging technologies if research on the matter makes a positive correlation between the technology in question, and the efficiency and quality improvements it can create, especially with patient outcomes and potential cost savings in the long run. Statement of Objectives With this study, we will be looking further into the two major problems that we previously discussed, present the degree of impact they have on hospitals, and discuss possible solutions. These dilemmas can be categorized into two areas of study: a. The major gap between early adopters of EHRs and those who still use primitive technology. b. The financial ceiling and time constraints that many hospitals face, along with the lack of impact that HITECH will have on those hospitals to rapidly jumpstart an EHR system. These two problems deal with a healthcare systems current technological and financial position, and how these two factors will affect their ability to satisfy meaningful use requirements within the given timeframe. The Public Library of Science stated that, there is a lack of robust research on the risks to implement these technologies and their cost-effectiveness

12 (2010) American Hospital Association - Special Bulletin: CMS Releases Final Definition of Meaningful Use Of HIT

has yet to be determined, despite being frequently promoted by policy makers and technoenthusiasts as if this was a given.13 Though many are championing HIT, there is still very little research on the matter, causing hospitals to walk on eggshells when venturing into this technological advancement. There are few successful systems to benchmark, so it will be important for healthcare systems looking to be involved in the HITECH incentive program to constantly be up to date, and actively compare and contrast themselves with other healthcare providers in order to plan system projects accordingly. This is why we will be studying and discussing where HIT as a collective is currently positioned throughout U.S. healthcare systems, as well as the projected outcomes that may or may not be enough to consider HITECH a success. Methodology For the first stated problem that we will discuss, we obtained data from the Center for Disease Control. In their study, they surveyed office-based physicians who have basic EHR systems in place, as well as other categories, but we will concentrate on basic EHRs. A basic EHR, according to the CDC, is one that has, patient history and demographics, patient problem list, physician clinical notes, comprehensive list of patient medications and allergies, computerized orders for prescriptions, and the ability to view lab and imaging results electronically.14 The reason for examining office-based physicians is because studies showed that it is those physicians who may have the most resistance to adopt EHRs, at least compared to hospitals and hospital systems.15 Their selection criteria was to include physicians who were, classified as providing

13 (2011) The Impact of ehealth on the Quality and Safety of Healthcare: A Systematic Overview http://www.plosmedicine.org/article/

info%3Adoi%2F10.1371%2Fjournal.pmed.1000387
14

(2011) Center for Disease Control and Prevention: Electronic Health Record Systems and Intent to Apply for Meaningful Use Incentives among Office Based Physicians.Practices: United States, 2010-2011
15 Finding The Meaning In Meaningful Use http://www.nejm.org/doi/full/10.1056/NEJMsb1103659#t=article

direct patient care in office-based practices, including clinicians in community health centers. 16 From 2008, supplemental mail surveys were conducted to discover the EHR systems in place chosen from selected geographic areas. In 2010, the survey was increased five times in sample size in order to allow for state level estimates, with a sample size of 10,401 physicians. The 2010 mail survey was conducted from April through July, and the 2011 survey was conducted from February through June 2011. The statements of differences was based on statistical tests with significance at the p < 0.05 level. With this survey, we were able to see the percentage of officebased physicians within each and every State that had any kind of EHR system, a basic EHR system, and those who were planning to apply for incentives. The second source of data that we will be using comes the Office of the National Coordinator for Health Information Technology report of acute care hospitals that have basic EHR systems. This study is similar to the one conducted by CDC, except that instead of surveying office based physicians, they surveyed acute care hospitals. They used the already mentioned survey by the AHA and had a response rate of 58% and significance testing is based on a p<0.05 as the threshold. The third source of data that we will examine is from the Healthcare Information and Management Systems Society (HIMSS) report titled, Summary of Meaningful Use Readiness.17 With this study, HIMSS sought to discover the amount of hospitals, not officebased physicians (which was the concentration of the study conducted by CDC), that are reported as being ready, or on track to being ready, to have met the first stage of meaningful use. This study was conducted from February 1st 2011 through September 30th 2011, and surveyed

16

(2011) Center for Disease Control and Prevention: Electronic Health Record Systems and Intent to Apply for Meaningful Use Incentives among Office Based Physicians.Practices: United States, 2010-2011
17

(2011) HIMSS Analytics Report Confirms Increase in Hospitals

a total of 778 hospitals. This study was conducted with the same statistical analysis as the previously mentioned study. The fourth source of data that we gathered was from Centers for Medicaid and Medicare Services titled, Combined Medicaid and Medicare Payments by State.18 With this study we will be using the Top States and Bottom States from previous data and comparing it to the incentive payments done by the EHR Incentive Program. We will see the correlation, if any, among the states that have implemented a basic EHR and the incentives that have been paid with the completion of the Stage 1 of meaningful use. First we will concentrate on the study done by the CDC on those office based physicians who have at least a basic EHR system in place, and the second study that surveyed acute care hospitals because those systems that fall in this category are on the right track to become certified according to ONCHIT. Secondly we will then examine the amount of hospitals that have a basic EHR system who will also be on track to becoming certified according HIMSS Analytics. Lastly we will examine the money that has been dispersed thus far from CMS and compare it to our finding of the previous studies. As we had noted in the our Statement of the Problems, much of the literature that we reviewed showed that there were very few healthcare providers that had a basic EHR system in place, and we believed possible reasons for this were high level of constraints that they faced in order to do so, and with our review of the the studies conducted, we sought to see if this holds true. Findings What we discovered when analyzing the data from the AHA, was that there was a large difference between states who had basic EHR systems in place. Referring to Figure 1, we first

18 Combined Medicare and Medicaid Payments by State

http://www.cms.gov/EHRIncentivePrograms/Downloads/Payments_by_state_by_programs.pdf

analyzed the top five states in terms of percentages. The average for the top five states with basic EHR systems was 55.2%, and if this figure was the nationwide average, then it would seem as if the the transition to a nationwide, interoperable EHR system would be attainable. However, these numbers, as you may notice, decline at a steady rate between each state.

The national average is actually only 33.9%, a 26% difference from Wisconsin, who has the highest average. With Figure 2, you can see that the lowest percentage in the U.S. is New Jersey, with a very

very low average of 16.30%, a whopping 43.6% range difference when compared to Wisconsin. This proves the literature that we have reviewed that there is a large gap between healthcare providers who are already in the process of becoming EHR certified, and those who are not. One possible reason that can explain why so few office-based physicians are not motivated to begin implying EHR systems is because the Medicaid and Medicare incentive that are being dispersed will only encourage those who rely heavily on those funds from the Government. According to BMC Health Services research, one component that may affect these same physicians is the lack of personal incentives that they can gain, which with HITECH, are none.19 This may be another reason why these office based physicians are not inclined to begin adopting EHRs. If physicians can only expect to receive incentives after implementing a meaningful use EHR, and those rewards ultimately affect patients that they might not have, then they will be less inclined to begin adopting a system with rigorous standards that will be very costly. These types of physicians are more likely to serve those who have health insurance, and not those who carry
19 Barriers to the Acceptance of Electronic Medical Records by Physician.. http://www.biomedcentral.com/1472-6963/10/

231#B9

Medicaid or Medicare, due to the paperwork and time it takes to cater to such patients. Even though this resistance may be detrimental to the the final goal of a nationwide healthcare system, it is still a reasonable vantage point to hold. We believe that with so many barriers, and so many uncertainties that accompany the adoption of EHRs, those who are not obligated to keep up with meaningful use status will not aggressively attempt to do so. For our next findings, refer to the following figures.

As already mentioned, this data represents those acute care hospitals that reported having a basic EHR system. As with the study of office-based physicians, the large gap between states still holds true. Hawaii is on top of the leaderboard with 66.9%, Utah is at 8.9%. (We are excluding Alaska the low number represents a finding that cannot be seen as reliable) This range of difference, an alarming 58% only reaffirms our study that there are many states that are lagging far behind early adopters. We believe that hospitals are more inclined to apply for the incentive program due to large number of Medicaid and Medicare patients, but the sheer size of most hospitals has made it very difficult to implement EHR systems when compared to officebased physicians. We feel that if hospitals were more leveled when it came to having basic EHR systems, then it would constitute ONCHIT meaningful use stages, but since this is not the case, we believe that the meaningful use deadlines was far too early. Not only will most of these hospitals

continue lagging behind, but those hospitals that are significantly ahead will only continue to expand the gap. The only way for hospitals that are lagging to catch up, would be for them to invest a substantial amount of money. Even though our economy is beginning to show signs of improvement, many hospitals would rather not risk investing large amounts of money because as previously stated, these projects contain high risk, partly due to the relatively small amount of successful projects that they would be able to imitate in order to install successful systems. One solution that we have considered would be for ONCHIT to take this into consideration, and help out those are far behind with money that could accelerate their adoption rates. Our welfare system notes that there are many poor people that need government assistance, and in the same way, ONCHIT should have identified those poor hospitals, and assist them in accelerating the adoption process so that the gap between hospitals who have basic EHRs would not be so significant.

HIMSS analytics reported in 2011 that only 10.03% of the hospitals that they surveyed were expected to be Ready to meet the 1st Stage of the ONCHITs program, and 30.98% who are Most Likely to be ready as well. Even with a relatively high number of hospitals who reported as most likely, it is clear by now that actually being ready is not as easy as it may seem, and most will actually not be ready. Reason for such a high report of Most Likely, is because they are rushing through the processes so that they can be qualified to receive incentives. This approach, however, may result in systems that are not as reliable as those that were built with extensive testing and iteration. The same study reported that 53.47% reported as Not Likely, and 5.53% as No Progress. If you refer to Figure 3, you can see that a total of 59% of hospitals were projected to not reach Stage 1 of meaningful use status. What this tells us that the majority

hospitals are not on correct path to reaching Stage 1, and there should be concern about the timeframe and deadlines that ONCHIT gave because there may not be

sufficient time for those hospitals who do not have basic EHR systems. Finally, we will be looking into the money that was dispersed by CMS. Using the same states that we used in the our second study of acute care hospitals who have basic EHR systems, we analyzed the money that was being dispersed when compared to the Top Five and the Bottom Five. Our finding showed that the average payments given to the Bottom Five states was $55,375 (in millions) and an average payment of $64,042 (in millions) was given to the Top Five States. This is a difference of $8,667 (in millions) dollars that was dispersed. We took into account that some states had a larger population of hospitals, and would likely receive more money, by dividing the amount given to each state by the amount of hospitals that were given money in each state. What we found only reaffirmed our studies. It is clear that those hospitals that already had basic systems ended up receiving the majority of incentive payments.

Conclusion In conclusion, we have found that even though the HITECH Act has good intentions, there are still many obstacles and barriers that must be conquered in order for the Act to be considered a success. The goal of this act is to have a nationwide health system that will supplement every hospital that is participating to better diagnose, treat, and hopefully, obtain positive patient care outcomes. With what we have found, however, this goal seems very out of reach. If there are so many hospitals that do not already have basic EHR systems in place, it will be almost impossible to be able to catch up and become certified meaningful users. As we mentioned, the only way that this Act can be reached is if hospitals who are struggling to keep up are assisted, and if the deadlines to be meaningful users are extended. The only way to accomplish this however, would be for ONCHIT to change their meaningful use criteria, or change the deadline to meet them.

Cut Sections

1. Many factors that have an effect on the cost can include the location, technology, resources, regulations and available funding for each respective hospital, or hospital system 2. According to a 2005 study, Medicare Spending per Beneficiary by Hospital Referral Region, in the Rio Grande Valley was $10,300 to $13,900 (Office, C. o. 2008)This amount was among the few highest in the United States. Also, in 2004 a report shows that in Texas there were about 200-300 physicians per 100,000 residents. This number is significant when compared to other states such as Massachusetts where there were about 400-500 physicians per 100,000 residents. Conclusively, with the expenditures in Health Care being so high, measurement of the recovery of a patient will be the vital variable in determining the relationship 3. Many practices lack of implementation is the cost such as the initial fixed cost of the hardware, software, and technical assistance necessary to install the system, licensing fees, the expense of maintaining the system, and the opportunity cost of the time that health care providers could have spent seeing patients but instead must devote to learning how to use the new system and how to adjust their work practices accordingly.(Hagen, 2008) Uncertainty of the technology has essentially outweighed the benefits of Health Information Technology because of the lack of information in how much exactly the value of IT will be or because of the lack of capital to be able to have such technologies implemented. 5. This concern that hospitals face, is obviously, a money issue. With unlimited resources, any business can implement new and exciting technologies, but since this is far from the truth, healthcare providers must allocate resources wisely. This is very important because not only will improve overall health care, but hospitals that fail to meet HITECH objectives from 2011 to 2015, will not qualify for additional medicaid and medicare funding. In addition to lost incentives, hospitals may then be subject to payment penalties in 2015 and each following year where objectives were not met. 6. The New England Journal of Medicine reported that the majority of studies that seek to fine the degree to which IT affects the quality and safety of patient care, come from four hospitals. Brigham and Women's Hospital in Boston, LDS Hospital in Salt Lake City, Vanderbilt University Medical Center in Nashville, and the Regenstrief Institute in Indianapolis.20 These hospitals have reported relative success, but with long periods of development, and constant improvement cycles, they are significantly ahead of the majority of hospitals. It is interesting to note that even with such promising projects, these hospitals struggled through constant dilemmas, failed systems, and currently, ONCHITs certification process that requires increasingly stringent criterion. The Certification Commission for Health

20 Finding The Meaning In Meaningful Use http://www.nejm.org/doi/full/10.1056/NEJMsb1103659#t=article

Information Technology tested these and other systems, both privately developed and commercially bought, and discovered that only 53% of fatal medication prescriptions were detected by the CPOE systems in place.21 CPOE implementation is a part of Stage one, as previously mentioned, and with such grim figures, the greatest concern facing healthcare systems will not be implementing, even though that aspect is still a large dilemma, but instead will be to employ a system that shows significant improvements in patient care.

http://www.hitechanswers.net/about/about-the-hitech-act-of-2009/ http://www.nejm.org/doi/full/10.1056/NEJMp0912825 http://www.nejm.org/doi/full/10.1056/NEJMsa0900592 http://onlinelibrary.wiley.com/doi/10.1111/j.1475-6773.2010.01233.x/full http://www.recovery.gov/About/Pages/The_Act.aspx http://www.nejm.org/doi/full/10.1056/NEJMp0901592 http://healthit.hhs.gov/portal/server.pt?open=512&objID=1487&mode=2 http://www.healthit.hhs.gov/portal/server.pt/community/ healthit_hhs_gov__regulations_and_guidance/1496 http://www.healthit.gov/providers-professionals/meaningful-use-definition-objectives http://www.hhs.gov/ http://www.hitechanswers.net/ehr-adoption-resources-2012/?wlfrom=%2Fmorehitech%2Fhitcommittees%2F http://blog.softwareadvice.com/articles/medical/ehr-vs-emr-whats-the-difference/

This is from data that surveyed hospitals who had basic ehr systems in place in 2011 The top five States with Basic EHRs Wisconsin 59.9% North Dakota 57.8% Washington 54.6% Oregon 54.5% Utah 49.3% Average 55.24

21 Finding The Meaning In Meaningful Use http://www.nejm.org/doi/full/10.1056/NEJMsb1103659#t=article

The bottom five States with Basic EHRs Nevada 23% District of C.21.2 Mississippi 19.9% South C. 19.5% New Jersey 16.3% Average 19.98% The final source of data that we will be using comes from a report conducted by the New England Journal of Medicine (NEJM) in 2009 that studied healthcare providers and what they believed were barriers to begin implementing EHRs. Their sample size included 3049 hospitals, or 61.3% of acute care hospitals surveyed, and once they omitted Federal hospitals and those located within the District of Columbia, they were left with 2952 hospitals. Using a logistic regression model, they sought to find if the responses that indicated a high level of concern towards the barriers of entry were reason enough for those respective hospitals to not implement EHRs. Statistical significance was based on a two sided p-value less than 0.05.22 HIMSS analytics reported their findings based on a 7 stage model that will measure hospitals on their current position of adopting EHR systems. They call this the EMR Adoption Model where Stage 0 represents a position where none of the required components of having a certified EHR installed. Stage 3 represents the accumulation of meeting the 1st stage of meaningful use according to ONCHIT. Reaching stage 5 in this model will be the equivalent of reaching stage 2 of the ONCHIT incentive program, and naturally, reaching stage 7 will represent having an EHR system that is fully compliant with the overall goals of HITECH. We concentrated on those respective stages in order to better understand the progress that hospitals are making to reach meaningful use status.

22 (2009) Use of Electronic Health Records in U.S. Hospitals. http://www.nejm.org/doi/full/10.1056/NEJMsa0900592#t=articleResults