In a study of 10 North Carolina hospitals [from January 2002 through December 2007], we found that harms remain common, with little evidence of widespread improvement. Further efforts are needed to translate effective safety interventions into routine practice and to monitor health care safety over time.
Unfortunately, I don't believe that the article differentiated between computerized hospitals and paper-based ones. Nor were the subject hospitals selected on the basis of computerization or non-computerization:
We conducted a retrospective study of a stratified random sample of 10 hospitals in North Carolina ...
... All acute care North Carolina hospitals listed in the American Hospital Association (AHA) database except those providing exclusively pediatric, rehabilitation, or psychiatric care were eligible for selection for the study. These hospitals were stratified according to the AHA’s definition of the facility as small, medium, or large; urban or rural; and teaching or nonteaching. The number of hospitals that underwent randomization for inclusion in each stratum reflected the proportion of national discharges from that type of hospital. If an invited hospital declined to participate, another closely matched hospital was randomly invited to participate in its stead.
While the extent of EHR/CPOE and other clinical IT adoption was not measured, some of the hospitals studied would likely have adopted and/or been using clinical IT in various capacities during the study period. The IT might have been expected to contribute to lower error rates over time.
Although that is admittedly speculation, there is a more important point to be made.
Many of the identified errors seem to have little to do with record keeping, but instead with human factors.
From the New York Times article "Study finds No Progress in Safety at Hospitals" covering the NEJM article:
... Landrigan’s study reviewed the records of 2,341 patients admitted to 10 hospitals — in both urban and rural areas and involving large and small and teaching and nonteaching medical centers ... Among the preventable problems that Dr. Landrigan’s team identified were severe bleeding during surgery, serious breathing trouble caused by a procedure being performed incorrectly, a fall that dislocated a patient’s hip and damaged a nerve and vaginal cuts caused by a vacuum extraction device used to help deliver a baby.
The findings were a disappointment but not a surprise, Dr. Landrigan said. Many of the problems were caused by the failure of hospitals to use measures that had been proved to avert mistakes and prevent infections from urinary catheters, ventilators and lines inserted into veins and arteries.
The chart of "harms" in the article, both those deemed "non-preventable" and those deemed "preventable" reads like a textbook of medicine, e.g.:
Cardiac arrest, shock, myocardial ischemia, acute respiratory failure, acute renal failure, hemorrhage, thromboembolic venous event, hematoma, pancreatitis, ileus, stroke or intracerebral hemorrhage, withdrawal symptoms, catheter-related bloodstream infection, urinary tract infection, surgical-site infection, clostridium difficile colitis, surgical anastomosis failure, wound dehiscence, failed procedure, unplanned return to surgery, fetal neonatal complication associated with delivery, hypothermia, pressure ulcer, catheter complication, etc.
It seems unlikely a missing or illegible chart was the cause of many of these adverse events. (In fact, the study was done retrospectively from the patient charts.)
Thus, expectations for major quality improvements in healthcare from the hundreds of billions spent on health IT might be vastly overstated based on false assumptions about the causes of adverse events.
Many of the above adverse events seem not highly amenable to correction via cybernetic information retrieval systems. They might be amenable, though, to improved nursing and ancillary staffing (perhaps I should say 'elimination of understaffing'), decreased work hours, improved CME, and better supervision of trainees ($100's of billions of dollars being spent for IT would surely buy a lot of those items).
What I found striking was this:
“A third of the errors in the intensive care unit disappear when residents work 16 hours or less,” Dr. Landrigan said, although he noted that senior residents often work longer hours.
That being the case, health IT that created heavier workloads and cognitive overload of residents already struggling under sleep deprivation might actually increase the risk of error.
It would seem national healthcare IT may be a 'solution' to the wrong problems in 2010.
We perhaps should better be focusing on human problems not amenable to cybernetic intervention before we start a national medical IT experiment, in the hope that major changes will somehow be effected by the magic of computers.