Thursday, January 22, 2009

Waste Feared in Digitizing Patient Records: Wall Street Journal

In today's Wall Street Journal an article appeared entitled "Waste Feared in Digitizing Patient Records" by Jacob Goldstein and Jane Zhang.

I would certainly call this title an understatement. Readers of the HC Renewal blog can probably understand why from my many postings on the challenges and the (largely unrealized as of 2009) potential of well done healthcare IT.

President Barack Obama has pledged to "wield technology's wonders to raise health care's quality and lower its cost," but many in the field warn that rushing the process of digitizing patients' records could lead to wasteful spending.

Both President Obama and former President George W. Bush have endorsed the goal of having electronic health records in widespread use by 2014. But the degree of difficulty promises to be high.

A study published last year in the New England Journal of Medicine found that only 4% of U.S. physicians were using "fully functional" electronic health-records systems. What's more, simply installing new IT systems in doctors' offices isn't enough. The systems have to be able to talk to each other, so doctors and hospitals can share information. And doctors and nurses have to learn how to use the systems in ways that help patients.


I would differ from the statement in the article that "many in the field warn that rushing the process of digitizing patients' records could lead to wasteful spending" [i.e., on systems that don't work, require massive remediation, are rejected, fail, etc. - ed.]

I have been writing on these issues for over ten years, and in that time I've only seen a handful of people writing such things publicly. They usually received scorn or were labeled as naysayers - much as economic forecasters and stock market experts the past few years who warned of potential trouble due to risky loans bundled into even riskier government backed securities.

In fact, my own site "Common Examples of Healthcare IT Difficulties" remains nearly unique on the Web in covering health IT problems at a very granular level. (It was originally entitled "Common Examples of Healthcare IT Failure" but I changed the name due to comments from some in the field that the title was too "negative.")

Last week, House Democrats unveiled a fiscal-stimulus bill that includes $20 billion for health-care information technology.

"In some ways I am thrilled, because IT will need federal help," said John Glaser, chief information officer for Partners HealthCare, a large nonprofit hospital system in Boston. "But you can bring in too much money too fast and not only waste it, but set us back."


In what may be a first, a non medical non biomedical informatics-trained (to my knowledge) CIO echoes the sentiments that a relatively small group of informed biomedical informatics, sociology, and other specialists have been saying for years. In fact, Glaser did participate in a site visit by the National Research Council Health IT Report team to Partners Healthcare. The Press Release summary of the NRC report itself entitled "Current Approaches to U.S Healthcare Information Technology are Insufficient" states:

... Current efforts aimed at the nationwide deployment of health care information technology (IT) will not be sufficient to achieve medical leaders' vision of health care in the 21st century and may even set back the cause, says a new report from the National Research Council. The report, based partially on site visits to eight U.S. medical centers considered leaders in the field of health care IT, concludes that greater emphasis should be placed on information technology that provides health care workers and patients with cognitive support [as opposed to just an 'inventory system' of medical information - ed.], such as assistance in decision-making and problem-solving.

Most IT personnel in the healthcare space have spent the past few decades waxing ecstatic about how healthcare IT was a cybernetic miracle to cure healthcare's ills as if by magic, physician and nurse complaints about how the technology slowed them down or created cognitive overload be damned. Unfortunately, it's an unfiltered diet of this propaganda I fear has been ingested by government officials over the past few decades that's led to the pronouncements that billions spent on health IT in 2009 is a sure bet that will, as the President stated, raise healthcare quality and reduce its costs.

What has been missing is a genuine, "political correctness be damned" rationale for why this technology has been disappointing in diffusion and performance even after 40+ years of research and efforts. Then we have this:

The Democrats' bill attempts to address many of the potential pitfalls associated with the rollout, spelling out plans to increase privacy protections, fund regional electronic networks, train technicians and give doctors financial incentives to use electronic records in ways that are likely to benefit patients.

A Republican congressional aide says the minority supports the goal of digitizing health records, but doesn't support spending money on technology that isn't yet developed. The aide adds that Republicans plan to draft amendments to make sure money won't be wasted on a piecemeal system.


I agree with the latter assessment. HIT is still a largely experimental technology. The IT industry uses hospitals, doctor offices and patients as alpha and beta test sites and subjects, unregulated by the FDA or other agency. When HIT fails, there is no central agency to report the failures to, only the vendor. Fixes go into a "queue" for remediation, with priority level decided by the vendor.

The major pitfalls of an attempted large scale national "rollout" can't be addressed when those major pitfalls are rarely found in the literature. Yes, privacy, interoperability and incentives are important. However, IT that causes error, slows clinicians down, confuses and insults them, and is really not useful to them in improving their ability to care for their patients are the most fundamental pitfalls that have impaired this technology's diffusion and acceptance.

The national HIT project in the UK is a quagmire in part for just these reasons (see here).

"If it's too hasty, you can create so many bad experiences that people say...'My data's a mess and my patients are angry,'" Mr. Glaser says.

I would say the major complaint is more akin to "the IT is a mess", not "my data's a mess."

What is at the root of these difficulties?

False assumptions, for starters.

If you attempt to travel to the moon in a hot air balloon under the assumption that since the moon is up and the balloon goes up, therefore the balloon is a viable solution to earth-moon transit, you will just not see Neil Armstrong's footprints in the Sea of Tranquility.


To the moon in a hot air balloon!


The major false assumption in HIT has been that HIT is just a specialized type of MIS (management information system), as opposed to it own subspecialized domain in the IT space.

This assumption and its corollary, that HIT is an MIS that just happens to involve clinicians (as opposed to the reality that HIT is a virtual clinical tool that just happens to involve computers) is at the root of HIT's difficulties. This assumption leads directly to errors in leadership, management, design and implementation of health IT. It leads to strategic, tactical and operational errors. IT personnel should not be leading design, deployment and management of clinical tools, either material or virtual, simply because those tools involve computers.

Let me illustrate by analogy. The recent Airbus that lost engine power due to collision with birds in New York was designed as a fly-by-wire aircraft. The control inputs from the pilots are mediated by computers that in turn control the active aircraft flight surfaces, as opposed to direct electric or hydraulic actuators. These computer systems also provide detailed information and decision support to the pilots. It has been claimed that those automated systems also prevented this NY aircraft from stalling or going out of control through programmed "artificial intelligence."

If this were health IT, the computer personnel would be put in charge of aircraft design, since the computers are now essential to aircraft operation and flight safety, and have the aeronautical engineers and other aeronautical domain experts report to them. The latter would be asked to perform their roles in accordance with what is convenient for the IT personnel, and blamed when they reject poorly conceived and implemented IT that could confuse and overwhelm pilots and cause air accidents.

Yes, this does sound bizarre, but this is what largely goes on in health IT.

In much more detail, the errors that follow from the assumption that "HIT is MIS" include:

  • A striking lack in the HIT industry of the scientific rigor of medicine
  • An HIT track record of success, or lack thereof, at best similar to MIS. See, for example:

Statistics on IT Project Failure Rates. The surveys referenced here provide statistical data regarding IT project failure rates. This topic is not often discussed in the mainstream IT literature. Health care IT failure rates may be even higher due to its greater sociotechnical complexity compared to traditional business IT. link

Pessimism, Computer Failure, and Information Systems Development in the Public Sector. (Public Administration Review 67;5:917-929, Sept/Oct. 2007, Shaun Goldfinch, University of Otago, New Zealand). Cautionary article on IT that should be read by every healthcare executive documenting the widespread nature of IT difficulties and failure, the lack of attention to the issues responsible, and recommending much more critical attitudes towards IT. link to pdf

  • HIT that does not meet clinician needs, slows clinicians down, cognitively overloads them, necessitates ad hoc workarounds that make clinicians' work more complex, and promotes rather than prevents informational errors.
  • Unbacked, often IT marketing driven claims that HIT will ipso facto reduce healthcare costs, reduce FTE's and improve care quality. The latter claims are problematic when scientifically tested, e.g., Electronic Health Record Use and the Quality of Ambulatory Care in the United States” (Arch Intern Med. 2007;167:1400-1405, link to abstract here).
  • Use of tightly controlled and mechanistic approaches and methodologies of MIS more suitable to developing business IT such as accounting and financial systems, but profoundly unsuitable to creating IT that meets the needs of clinicians in actual medical settings due to lack of methodological imagination and agility. In fact, MIS encourages what J. Paul Getty once described as "bureaucratic corpulence, the trap of sterile formalism and unproductive ritual."
  • Paradoxical leadership. Non medical personnel of a technology and/or business focus in charge of HIT design, acquisition, deployment and life cycle, leading to technology modeled after what they know best, and therefore poorly supporting what they know least (clinical medicine and its "hiding in plain sight" complexities).
  • Over reliance on "process" to make up for this deficiency, although a thousand generic workers will always be outperformed by one domain expert who actually knows what they're doing.
  • Paradoxical and intensely dysfunctional reporting of healthcare enablers to healthcare facilitators, such as postdoctoral trained clinician-informatics experts reporting to non-medical, non-science business IT CIO's or other IT personnel with bachelor's level training, and the latter evaluating the performance of the former.
  • Creation of an HIT industry that has created and promotes the myth of HIT cybernetic miracles out of its own interests.
  • Spectacular, expensive, multi-football field sized annual trade shows promoting this technology (e.g., HIMSS).
  • "Cross disciplinary piracy" - a dubious pretense to biomedical leadership qualification and insight by all manner of IT and IT facilitator personnel over seasoned clinicians.
  • Lack of clinician input, and blame placed on hapless clinicians when they reject suboptimal HIT designs.
  • Lack of adequate biomedical informatics input from true informatics experts. Such personnel are often used as "internal consultants" who may or may not be listened to. Common in hospitals is the "Director of Informatics" or "Chief Medical Informatics Officer" role, which usually translates to "Director of Nothing" or "Chief of Nothing" with little or no direct control of resources or personnel. Often used as 'internal salesmen of suboptimal HIT' to other clinicians.
  • Too much input from "medical instamaticists", people who've managed to convince others that doing something with a computer in medicine qualifies them to play critical roles in major HIT projects. Often, these personnel know just enough to be dangerous and disruptive to those who actually know what they're doing.
  • Legions of expensive consultants with the same assumptions.
  • Suppression of reports of HIT difficulties that could - and should - be shared by other organizations as "lessons learned" due to self interest of the involved vendors, consultants and IT personnel.
  • Focus on doing too much at once; e.g., simultaneous or near simultaneous implementation of EHR, CPOE, financial and other ancillary systems that provide little benefit to clinician end users and that drain healthcare's limited resources to the IT sector. This is as opposed to an incremental approach that starts with EHR support for clinicians and highly customized systems for high risk, high revenue areas, to be followed in stages by more comprehensive, integrated IT systems that if done well might have an ROI to the hospital business side.
  • "Big bang" rollouts that attempt massive cultural change in an entire organization instead of incremental change, beginning with a pilot area or areas, then advancing unit by unit, specialty by speciality, subspecialty by subspeciality, learning as one goes to avoid costly errors.

If money is to be massively spent, it would be better spent initially on true R&D. This R&D would strive to end false assumptions and develop well grounded leadership structures, development methods for effective HIT, and ensure proper engagement of experts and end users to facilitate this technology in attaining its true capabilities.

-- SS