WBUR

Study: No Cost Savings With Electronic Medical Records

BOSTON — A new Harvard study (PDF) finds that computerized medical records don’t save money or make hospitals more efficient, despite claims that health information technology could generate huge financial returns.

The study has major implications for national health reform. Congress is counting on savings from health IT to help fund the cost of revamping the health care system and the Obama administration plans to use about $30 billion in federal stimulus money to spur the nationwide adoption of health IT, such as electronic medical records.

But Harvard Medical School professor Steffie Woolhandler said an analysis of 4,000 hospitals nationwide found no cost benefit from computerization.

“If people think we’re going to pay for health reform off of savings on hospital computerization, they’re really sorely mistaken,” said Woolhandler, one of the study’s co-authors. “There are no savings we were able to find looking at five years of information.”

The researchers examined data from 2003 to 2007, including from hospitals on the “100 Most Wired List” compiled by Hospital and Health Networks magazine. They analyzed the data for evidence of increased quality, cost savings or improvements in administrative efficiency.

Besides finding no indication that computerization lowers costs, the researchers found that some hospitals actually saw administrative costs rise after computerizing, probably due to the expense of buying and maintaining equipment and paying for technology staffs.

And those increased administrative expenses weren’t just one-time upfront costs of purchasing computer systems; those expenses remained elevated even several years after computerization had taken place.

In addition, the research showed that computerization hasn’t helped hospitals become more efficient, although it may have “modestly improved the quality of care for heart attacks,” according to the study.

One reason computerization may not be improving efficiency and quality of care: many medical software programs are designed primarily to help hospitals with their billing, accounting and registration needs, not their clinical work.

When hospital computer systems are designed with clinical care in mind, Woolhandler said, they can help prevent medication errors, flag a patient’s drug allergies, and ensure that prescriptions are written properly.

“I’m still a believer in computerization; I’m not a total Luddite,” she said. “But I think computerization is very often done wrong. And, certainly from a cost perspective in the U.S., it’s done wrong more than it’s done correctly … so I think we need to be smarter about the way we computerize in this country.”

The study appears online in the American Journal of Medicine.

Please follow our community rules when engaging in comment discussion on wbur.org.
  • Johnye Ballenger, MD

    The patient-doctor relationship has been replaced by the patient-EMR-doctor relationship. The EMR was promoted as a panacea for 21st century healthcare, an opportunity to support and improve the patient-doctor relationship, to facilitate “patient-centered” care. My experience, and view from the PCP front, is that the EMR has been a panacea for insurers, making me a de facto data entry worker for the insurers first, and the patient’s physician second.

    The patient’s history, the patient’s health story, must now conform and fit into predefined fields. In the system we use, a sad example of this is the patient’s problem list which can only accept problems that have an ICD-9 code. No code, no problem, is the message. If the potential benefits of the EMR for patients is to be fully realized, the benefits must extend to include those responsible for the patient’s care. EMRs must become care-centric tools for patient care and the patient-doctor relationship.

    I look forward to reading Steffie Woohandler’s article in the Amer. J of Med.

  • http://www.reliatechconsulting.com Hank Mayeres

    Professor Woodhandler is indeed correct that SOME EMR solutions do indeed force physicians to become documentors and finders of ways to shoe horn assessments into pre-set codes and terminology. However, this is not always the case. Indeed, there is plenty of evidence that systems that are built according to normal clinical flow allow physicians to leave the office much sooner than the previous norm of 8PM. The professor is quite right that true benefits flow from the ways in which the newer systems focus on clinical (physician and nurse) services. These capabilities are the most recent to be added to EMR products.

  • eugene allred

    I am now 70 years old and have seen a lot of doctors and hospitals. I have also taught computer operating systems and database technology around the world for 20 years. I would like to suggest some features of an ideal medical records system (modestly called Planet-Wide Electronic Medical Records Systems (PEMRES) and here are some features I think it should have:
    1. All U.S. (and perhaps world) medical and dental records should be centralized and universally available via internet. I want my medical or dental practitioner, no matter where in the world I am, to have my complete and immediate medical history, even if I arrive unconscious. This is now easily feasible. GOOGLE can do this kind of information management. Banks can do it (thousands of banks use the same data interchange system each day). AMAZON can do it. Hundreds of agencies can do it. Why cannot the worldwide medical services come together like the banks to form a common data collection and interchange system?
    2. The database should contain a complete history of shots, vaccinations, diseases, traumas, allergies, doctors visits, hospital treatments, doctors reports, drug and other prescriptions, x-rays, cat scans, all diagnostic tests, diseases, treatments, blood type, dental records, everything necessary to provide good medical information for future medical practitioners to make the best assessment of treatment that they can.
    3. The important thing is to begin with all new data entry. Old information perhaps should reside with the current medical systems, although it would be advantageous for current large medical database systems to convert to PEMRES.
    4. A new system should be under government control and free (or modestly priced) for all users. The government can (1) provide adequate redundancy of the system; it will be stable no matter what; (2) guarantee perpetual existence of the system; (3) develop an outstanding encryption and control access system; (4) criminalize illegal access. Unfortunately, it cannot guarantee excellent design and implementation.
    5. Technology should make it possible to input almost all medical information directly into PEMRES. Doctors can now easily voice-input directly into a medical database records (Look at Dragon Naturally Speaking 10.1 Medical Edition). Testing and scanning equipment should be able to send results directly to the medical database.
    6. The database should be accessible in almost every language; if I am in China or Japan, I want the medical practitioner to see the information, if possible, in his/her language.
    7. Each medical consumer should have complete access to his/her records and be able to make comments and, perhaps, some edits to the record.
    8. There should be enormous cost savings with this system. I estimate that we could eliminate 3% of total U.S. medical costs ($75 billion annually) if we carefully and cleverly enough designed this system. If created by the government and offered free (or for a reasonable fee) to the world medical services, no special equipment would be needed by any medical provider. All that is needed are fast PC computers, up to date operating systems, large 28-32 inch screens, color printers, ultra fast internet access, wireless communications, excellent microphones and voice to screen technologies are now common, easily managed and cheap (compared to 5 years ago.)
    9. Everything PEMRES needs is ready to go. The required infrastructures are tested and ab-undant. There are now plenty of people skilled enough to put this system together. There are plenty of people skilled enough to use it (if correctly designed). Why such a system should cost more than $200 million to build at this time is totally beyond me. It should be fully tested and operational for that amount. It should take 1 year to design, 1 year to build, 1 year to beta test. It should require the services of fewer than 200 people. It should run at a cost of under $100 million a year. The government is essentially building a web site with a complex data base. How difficult is that these days? (Eugene.l.allred@comcast.net)

  • Martin B. Brilliant

    I think eugene alfred’s PEMRES could be used for centralized filing of insurance claims (though I think medical insurance is an idea whose time has gone). The provider would enter the patient’s ID, the diagnosis and procedure codes and other data. The system would immediately identify the primary payer and the coordination of benefits and come back with the amount payable to the provider. The system could also be used when the patient arrives, before treatment, to determine eligibility. This process could eliminate many sources of error. Knowing the patient’s age and medical history, the system could flag many (but not all) inappropriate procedure codes. It could also eliminate errors in filing with the wrong payer – or if errors occur they could be easily and permanently corrected.

  • armass

    I appreciate that the headline and opening sentence of articles like this are meant to catch your attention, but it is unfortunate that they tend to skew the facts.

    The opening sentence, in particular, makes it sound like researchers have concluded that computerized medical records can’t save money or make the health care system more efficient. However, these researches have really concluded that the currently implemented computerized medical record systems that they studied have not dramatically increased efficiency or reduced cost.

    There’s a big difference between concluding that computerization can’t make the system better and concluding that current systems aren’t making the system better.

  • Cheryl Clymens, Medical Transcriptionist

    I am one of the many medical transcriptionists who will be losing my job as a result of this forced EMR. Forced out of a line of work that I absolutely love and do very well. I don’t wish to answer phones or scan documents or check in patients, that’s not what I do. I AM a medical transcriptionist but by no choice of my own…..but by government control……I have to change careers. Hopefully I can even find a job when mine ends. I catch mistakes on a daily basis that the physicians make when they dictate reports so I’m not so sure how efficient EMR is going to be in catching those mistakes.

  • Pingback: Are Electronic Medical Records a Way to the Triple-Bottom Line? | TriplePundit

Most Popular