Physician Computer Order Entry

FigureBy now any regular reader of this column is probably aware of the fact that I'm concerned about the coming mandates for physician computer order entry. My basic position is that there are inadequate data to support this hugely costly undertaking and that there will be substantial resistance on the part of practitioners (who resist likely because they also sense that the solution is much worse than the problem to be solved). I know there are some examples in which physician computer order entry was implemented without mass hara-kiri being performed in the doctors' dining room, and I know some largely resident-run hospitals have successfully implemented such programs (where they generally don't have to deal with a community-based medical staff). I also know, however, that Cedars Sinai Medical Center (the hospital of the “stars” in Los Angeles) pulled the plug on an unsuccessful installation of physician computer order entry despite spending tens of millions of dollars (the actual numbers, of course, are disputed and depend on whom you listen to). A physician computer order entry system was linked to a 40% reduction in medication errors but had no effect on adverse drug events Here's another paper that looks at physician computer order entry, this time in a pediatric hospital. The bottom line, errors (very broadly defined) decreased from 5.19 to 3.13 per 1,000 patient days. As is typical, to make the case look compelling, the data need to be reported in relative numbers — a 40 percent decline in errors. Wonderful. Just wonderful. But there was no demonstrated effect on adverse drug events — a whopping 18 that occurred on five wards over a six-year period. The authors are quoted as saying, “Physician computer order entry holds promise as an intervention that may improve patient safety but requires further evaluation of the benefit and costs before widespread adoption.” But the Leapfrog Group obviously knows better. They claim they won't send you another patient after some specified date in the future unless your hospital has physician computer order entry. Worse yet, if, like me, you're in California — this very screwed up state where half the hospitals are already in the red and which has mandated that hospitals spend huge amounts to retrofit themselves to withstand the earthquake of all earthquakes — physician computer order entry (another unfunded mandate) also is required state-wide in the next couple of years. The Effect of Computerized Physician Order Entry on Medication Errors and Adverse Drug Events in Pediatric Inpatients, King WJ, et al, Pediatrics, 2003;112(3):506 BACKGROUND: It has been reported that at least 20 percent of adverse events occurring in hospitalized patients are related to medication use. Children are more vulnerable to such events than adults. Computerized physician order entry has been proposed as a method for reducing inpatient medication errors, and has been recommended by both the Leapfrog Group and the Agency for Healthcare Research & Quality. METHODS: This study examined the effect of implementation of a physician computer order entry system on medication errors (any error involving prescription, dispensing, administration, or monitoring) and adverse drug events (medication error resulting in harm to the patient) at the Children's Hospital of Eastern Ontario. Findings were compared in two wards in which the physician computer order entry system was installed and three wards in which it was not, during three years before and three years after installation. A passive reporting system for notification of medication errors was utilized. RESULTS: The rate of medication errors per 1000 patient days was 4.48 and 4.80 in the intervention and control wards before introduction of the physician computer order entry system; during the subsequent three years, the rate decreased to 3.13 in the intervention wards but was 5.19 in the control wards. Despite this 40 percent reduction in medication errors with the physician computer order entry system, there was no apparent significant effect on adverse drug events. There were only 18 adverse drug events on the five wards during the six years of study. CONCLUSIONS: In this pediatric tertiary care facility, a physician computer order entry system was associated with a 40 percent relative reduction in medication errors but had no significant effect on the occurrence of adverse drug events. Do Physician Guidelines Change Outcomes? I am in favor of physicians being made aware of and encouraged to use evidence-based guidelines. Our 42-hospital system is aggressively monitoring physician care of community-acquired pneumonia (CAP) on the assumption that better process will yield better outcomes. Items being tracked include door-to-antibiotic time, performance of sputum cultures, and blood cultures among “markers” of quality of care. The problem is very straightforward, however. Performance of sputum and blood cultures is unrelated to outcome, and the holy grail of pneumonia care — rapid door-to-antibiotic time —also is not likely to be associated with outcomes (the literature on this topic is very contradictory). What should be measured but is not is the choice of antibiotic. We can give any antibiotic we choose, as long as it is given quickly. The fact is, however, that the single biggest determinant of outcomes is the selection of an antibiotic regimen. The following provocative study was a response to proposed mandated use of a pneumonia protocol. Instead of having the medical staff whine about being forced to do “cookbook medicine,” the authors did something very unusual. They looked at outcomes prior to introduction of the proposed guidelines and concluded that, despite the fact that physicians infrequently ordered the mandated tests (15% had sputum Gram stain ordered, 38% had sputum cultures, and 24% had blood cultures), appropriate patients were admitted, recommended antibiotics were administered in the substantial majority of cases, and outcomes were excellent (90% discharged home, 9% to an extended care facility, and 2.2% died). Care was profitable and physician variability regarding practice was relatively small. The authors suggest that fact- finding should precede decision-making. Shoot, Ready, Aim: Pneumonia Care Quality and Costs in a Community Hospital, Milo LA, et al, Am J Med Qual, 2003;18(5):214 BACKGROUND: Recommendations in guidelines for community-acquired pneumonia (CAP) are variable and not necessarily supported by the available evidence. Nonetheless, adherence to guidelines is often considered a marker of quality of care. METHODS: Prompted by an institutional proposal to mandate adherence to a CAP guideline, based on a decrease in hospital profitability and perceived excessive variation in CAP care, the authors from Summa Health System in Akron, OH, retrospectively examined processes of care for patients with CAP admitted by family physicians and internists, clinical outcomes, and costs of care. Process of care was evaluated for 126 patients, and administrative data were analyzed for 514 patients. RESULTS: Only four percent were potentially suitable for outpatient treatment according to the PORT criteria, and 88 percent fulfilled PORT criteria for highest levels of severity. Only two percent of the charts contained signed pneumonia protocols. From 15 percent to 48 percent of the patients had protocol-recommended blood cultures, sputum cultures, Gram stain and/or pulse oximetry, and most were treated empirically. Levofloxacin was the initial antibiotic used for 80 percent of the patients (in compliance with the IDSA guideline), nearly 90 percent were discharged to home, and the estimated probability of readmission for pneumonia within 30 days was four percent. The mean length of stay was about four days, but chart review suggested that a substantial subset of patients exhibited clinical resolution by day one or two. Median earnings per patient were $760. Compliance with recommended testing would have decreased earnings by 12 percent without providing a clear clinical benefit. Decisions to mandate use of practice guidelines should be preceded by analysis of potential consequences. CONCLUSIONS: Decisions to mandate adherence to clinical practice guidelines should be preceded by an analysis of potential consequences. Tests for Drugs of Abuse It is my personal practice rarely to order panels measuring drugs of abuse. Although psychiatrists seem to have a fondness for these tests when they ask us to perform a medical clearance — our assurance that nothing bad will happen from a medical point of view for the entire psychiatric admission — the literature has repeatedly stressed the lack of value of these tests. The following paper takes a look at this practice from another point of view: Does the patient have a right to decline such a panel and what are our responsibilities if we order one? Although it is likely that most physicians will discard this author's opinion, I'm not so sure that many of his points don't have merit. Should Informed Consent be Required for Laboratory Testing for Drugs of Abuse in Medical Settings?, Warner EA, et al, Am J Med, 2003;115:54 The authors from the University of South Florida in Tampa and Brown University in Rhode Island review issues relating to testing for substance abuse in a medical setting. In the ED, such screening has been recommended by some for trauma patients, and might be useful in young patients with acute myocardial infarction or for the management of certain overdoses (although one study reported that a drug screen influenced immediate management in at most 5% of patients with suspected overdose). The utility of drug testing in other non-emergent situations is even more uncertain, and should be subject to the principles of informed consent. Its putative utility is limited further by the use of arbitrary cut-offs for a positive test, false-positive testing due to cross-reactivity with structurally similar substances, general failure to confirm positives with more definitive testing, and issues relating to proper chain-of-custody procedures. (Limitations related to false-negative tests and lack of laboratory reliability are not discussed.) The potential consequences of a positive drug test in terms of insurability and employability are substantial, and can be particularly heinous when the patient is unaware that testing has occurred. Testing should not be performed without the patient's explicit informed consent, and policies and procedures should be established for ensuring the confidentiality of results. The principles of autonomy dictate that in the absence of the potential for imminent harm to society, the patient, not the physician, should make decisions about testing or treatment for substance abuse. Respect for patient autonomy and confidentiality is required to ensure preservation of trust and a good physician-patient relationship.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.