Friday, March 20, 2009

Letter to the Editor II

Dr. Bates, Middleton and I have been asked to respond to a letter in the Washington Post. Our response, including links to the appropriate studies/evidence, is today's blog.

Dear Editor:

As Harvard Medical School faculty and experts in healthcare information technology, we wish to respond to the March 17, 2009 article "Bad Bet on Medical Records" By Stephen B. Soumerai and Sumit R. Majumdar. Our response is below:

"Soumerai and Mejumdar’s critique of electronic health records and the investment being made in them in the U.S. by the Obama administration does not present a balanced view of the evidence. The U.S. already lags behind virtually every other industrialized nation with respect to Health Information Technology (HIT) adoption, especially outside the hospital. The evidence suggests that investing in HIT will pay for itself.

Their first point is that several randomized controlled trials of decision support for one particular area (management of a few chronic conditions) did not show benefit (Article 1, Article 2, Article 3) That is accurate. However, they have not included multiple other studies that show that there is benefit for other conditions like diabetes and heart disease (Article 4, Article 5, Article 6) The data about the benefits of HIT for chronic diseases are more mixed than for other areas, but clear benefits have been demonstrated.

Their next assertion is that health IT does not save money. There are numerous studies showing that it does (Article 7, Article 8, Article 9, Article 10, Article 11). For example, a detailed case study of the cost and quality benefits of EHR at Family Care of Concord, NH found net benefits per clinician per year of $30,324. Another study of hospital-based provider order entry identified net savings of $1.7 million per year from drug dosing guidance, nursing time utilization, and error prevention.

Soumerai and Mejumdar also suggest that information technology makes care less safe. They present data from Children’s in Pittsburgh, which found that children transferred in for special care had an increased mortality rate. They do not mention that this hospital implemented the system poorly (as has been well documented) and made many workflow changes that resulted in delays in care for sick children. Badly implemented software can certainly yield negative results. Other hospitals, including Children’s of Seattle, have implemented exactly the same vendor system, following best practices for implementation, and experienced a trend toward a lower mortality rate.

The authors refer to “modest” error reductions. In fact, the level of medication error reduction with computerization of prescribing seen in multiple studies is over 80%. That is more than modest.

Doing research takes years, years that we do not have if we are to avoid slipping even further behind the rest of the world in this key part of the economy. While we do need research in this area, it should focus on how to improve care, not whether or not to implement electronic health records. That is already clear. Physician and nurse-practitioner teams represent a good idea, but they will be much more efficient if they are supported by the right information technology. Furthermore, health information technology, once implemented, keeps delivering benefits which will grow over time, while approaches like physician-nurse teams require ongoing support.

If we are to deliver high quality care for patients with chronic conditions, electronic records with decision support are needed to help providers track all the many things that need to be done. These records should include tools that enable providers to manage populations of patients with certain conditions like diabetes, and to track their progress. Patients should have tools that allow them to access their records and more actively participate in their care. Finally, we need to provide economic incentives for delivering better care, which will get providers to focus on these issues.

If patients are to have high-quality, safer, lower-cost care, we must move to a digital world in healthcare. Doing so won’t ensure that care gets better by itself, but it is a pivotal step in the right direction.

David Bates MD, MSc is Professor of Medicine at Harvard Medical School, and Professor of Health Policy and Management at the Harvard School of Public Health.

John Halamka MD, MSc is Associate Professor of Medicine, and the Chief Information Officer at Harvard Medical School and Beth Israel Deaconess Medical Center

Blackford Middleton MD, MPH, MSc is Director of Clinical Informatics R&D, and of the Center for Information Technology Leadership, at Partners Healthcare."

4 comments:

JoeBee said...

John,
If memory serves, you have previously offered a very useful analogy involving parachutes. I think you said, 'we dont need a randomized, multi-center, double-blind, placebo-controlled trial to demonstrate the benefit of parachute versus no-parachute for sky-divers.' If I got that right, can you provide a link to the original.

The complexity of decision making and task management associated with tightly-coupled and error-prone processes involved in healthcare make the parachute analogy almost insulting. That said, there seems to be a deep truth there that's worth repeating.

Thanks John.

Dr. Reed D. Gelzer said...
This comment has been removed by the author.
Dr. Reed D. Gelzer said...

Regarding your letter to Health Affairs, I would suggest your confidence in future HIT is justified. While there is good evidence that HIT can improve care, there is also much evidence that HIT does not yet necessarily improve care, meaning that universal HIT implementation would be without merit and remains unsupportable at this time. This is no surprise, all technological improvements are born expensive, clunky, non-standardized and take time to become generally and necessarily reliable and useful.

Some years ago, my wife was the only Internist in a rural county of 45,000 people. As you may imagine, that meant she was on call 24/7/365. This was also pre-cell phone, in the beeper era, when you could be wirelessly notified you were needed, but could not interact without a land line phone. For us, as water sports fans, this meant our boat sat at the dock. Around 1989, we got our first cell phone. It was expensive, clunky, unstandardized but it improved our quality of life greatly and was no detriment to care quality. Perhaps then would have been a great time for a massive expenditure of taxpayer funds to buy every doctor a cell phone, when they cost a $1000 and had a range of about 8 miles (if you happened to be near one of the few towers).

EHRs currently are highly variable in their design, permit documentation actions nobody would execute in a paper record (ex: obliterating original versions of amended records), non-standardized, and Certified against low requirements that will steadily improve over time. (Currently a Certified EHR permits obliteration of the original as above).

To use your analogy, there is no doubt that parachutes are great for skydivers. However, no skydiver goes on the internet and buys anything with the word "parachute" to strap to their back and jump out of planes. They want a parachute that is standardized against rigorous requirements and packed by a known, accountable, and trustworthy source.

HIT can improve, but does not necessarily improve care, HIT cannot necessarily improve our knowledge of care quality as Dr. Persell, et al concluded in Assessing the Validity of National
Quality Measures for Coronary Artery Disease, Using an Electronic Health Record: "Profiling the quality of outpatient CAD care using data from an EHR has significant limitations. Changes in how data are routinely recorded in an EHR
are needed to improve the accuracy of this type of quality measurement. Validity testing in different settings is required."
Arch Intern Med. 2006;166:2272-2277

Furthermore, as we gain experience with HIT, we are discovering a wealth of unintended consequences. See, for example, Dr. Vigoda in "Anesthesia Information Management Systems -Where we are in 2007?" where he notes "Our practice had an unfortunate incident when our AIMS (Note:AIMS = intra-operative EHR) did not record data for over 90 minutes during a case in which there was a serious complication." Also note Dr. McClean et al, "Electronic medical records metadata allow for creation of detailed physician profiles and will likely be used increasingly to discredit physicians during medical malpractice litigation." (J Am
Coll Surg 2008;206:405–411. © 2008 by the American College of Surgeons)

In 2004, while surveying and testing the Compliance support capabilities of EHRs, my associate and I were inquiring a major national vendor regarding how they handled open-item billing (Ex: Claims generated on the ordering of a test or initiation of an encounter record, not on its completion, risking billing for test or service never done.) Unsolicited, a nearby clinician exclaimed that he had a terrible problem with that at their very large Boston facility, with "tens of thousands" of incomplete records and tests that had been billed and paid for but never delivered.

I have been working for advancing health care through information technology for 12+ years and have been active in HL7 EHR Standards development and CCHIT Certification for over 4 years. I can say with complete confidence that HIT, where thoroughly tested and used to best advantage, is beneficial. I point to the MHS CHCS I system (not AHLTA) as an unparalleled accomplishment in CPOE and eRx as well as test results reporting. I also can say with complete confidence that, in time, standardized EHRs certified against rigorous requirements will be worthy of patients' and clinicians' confidence, but that is not where we are today yet. My recent Hopkins Medical Informatics graduate friend who follows her parents' care assiduously recently noted that their caregiver's EHR (major national vendor) had managed to intermix their records, giving an interesting report of her mother's prostate tests and her father's Pap smears. Our own due-diligence testing protocols, published by AHIMA, help illuminate some of the major anomalies in EHRs, but most implementations do not include basic medical records functional testing as part of due diligence, and anomalies like author misrepresentation are discovered after implementation and difficult to repair.

Unless appropriate cautions are counseled, such as in the Health Affairs article, to those hungry for the silver bullet to fix health care (whether inside the Beltway or living in the real world) we will end up shifting our grotesquely inefficient and sometimes dangerous industry from unaccountable paper records to unaccountable electronic records.

As Mr. Reagan said in another context, "Trust, but verify".

Thank you for your continuing championship of HIT Dr. Halamka, but a bit more circumspection is indeed in order.

Thank you for providing this opportunity to comment.

Reed D. Gelzer, MD, MPH, CHCC
Advocates for Documentation Integrity and Compliance
Wallingford, CT

Anonymous said...

Its really very interesting blog..i will always grateful to you for information posted in this blog.
Learn here about sympathy words