Difference between revisions of "Prescribers' Responses to Alerts During Medication Ordering in the Long Term Care Setting"
m (moved Prescribers’ Responses to Alerts During Medication Ordering in the Long Term Care Setting to Prescribers' Responses to Alerts During Medication Ordering in the Long Term Care Setting: rename title to remove special character)
|(3 intermediate revisions by 3 users not shown)|
|Line 17:||Line 17:|
David Schanding, M.A., M.M.
David Schanding, M.A., M.M.
Latest revision as of 17:49, 11 November 2011
JAMES JUDGE, MD, TERRY S. FIELD, DSC, MARTIN DEFLORIO, RPH, JANE LAPRINO, JILL AUGER, RPH, PAULA ROCHON, MD, MPH, DAVID W. BATES, MD, MSC, JERRY H. GURWITZ, MD, Prescribers’ Responses to Alerts During Medication Ordering in the Long Term Care Setting, Journal of the American Medical Informatics Association, Volume 13, No 4, Jul/Aug 2006
The combination of computerized provider order entry (CPOE) systems and clinical decision support systems (CDSS) are intended to assist practitioners in making sure that medications ordered for their patients are clinically beneficial and that they don’t inadvertently create adverse health effects for the recipient. Clinical decision support systems typically review drug dosage recommendations based on patient characteristics (e.g. age, weight) and drug-drug interactions for patients who take multiple medications (which would be very common in a long term care setting). They also review drug-patient interactions by identifying drug allergies and examining recent lab findings.
These potential adverse events are more commonly associated with older patients. Senior citizens are more prone to renal failure. They are more likely to be dose-sensitive to antipsychotic, anticoagulant, diuretic and antiepileptic medication. Other CPOE/CDSS studies have focused on inpatient and ambulatory settings. Unique characteristics of how the elderly respond to medications appears to call for specialized CPOE/CDSS efforts.
This study took place in an academically-affiliated long term care facility. The CDSS was designed by geriatricians, pharmacists, health services researchers and information system professionals. Presumably this combination of expertise would tailor the CDSS alerts to health aspects specific to seniors. The presumed correct responses to alerts were to cancel or replace an order, change medication dose, and order recommended labs or recommended additional medications. The option of overriding the CDSS recommendation based on additional clinical knowledge was not seen, in this study, as a positive response to the CDSS alert.
Three patient units used CPOE with CDSS and four units used CPOE only. In the CPOE/CDSS component, 9,414 alerts were generated for 47,997 orders. This represented an average of 2.5 alerts per patient per month out of 9 medication orders (28% of orders generated alerts). As will be discussed later, 54% of these alerts were based on limitations of the software and thus not true challenges to the medications prescribed. When comparing CPOE tied to CDSS in the experimental units and CPOE without CDSS in the control units, the only areas where CDSS seemed to consistently alter ordering behavior were in warfarin (blood thinner) management and in prescription of drugs with potential central nervous system side effects.
Discussion: As a segment of the overall population, senior citizens are more prone to adverse drug reactions. They typically take a wider combination of medications due to having multiple health problems. They are sensitive to higher dosages of medications. It is difficult to determine if behavior changes are due to the normal aging process, like with dementia, or are an unintended reaction to a particular medication regimen. CDSS which takes into account the multiplicity of medications ordered and their potential interactions, coupled with known issues related to how seniors handle medications, on the surface, appears to be an excellent idea.
CDSS, while great in theory, need to be sufficiently sophisticated to provide consistently pertinent direction to the provider. The CDSS in this study could not combine dose and strength of medications, therefore inaccurately portrayed the daily dosage of medications given (20 mg BID would not be calculated as a total of 40 mg). When an order was entered, cancelled, and re-entered, the system reacted as if the order was simply doubling the dose, ignoring the cancellation. The CDSS didn’t take standing orders into account. These factors led to 5,132 false alerts, where the order was actually within treatment guidelines, and this represents 54% of the total alerts. The authors acknowledge that the CDSS in this study could best be described as “first generation.” Its ability to assist in correcting warfarin therapy and potential central nervous system involvements makes the potential benefit of CDSS strong. Still, with more than half of alerts determined to be false alerts, the credibility of this CDSS system would be suspect to most practitioners. Clinicians are typically very busy and are annoyed by false alerts which waste time and may brand the entire CDSS as being invalid. Additionally, a clinician’s decision to ignore the recommendation of a CDSS is not necessarily sub-standard medical care. There are often solid reasons to continue with the planned course of treatment.
This study did show that there is efficacy in combining CPOE with CDSS. It also showed that greater sophistication in alerts will need to be addressed to make sure that alerts are relevant the majority of time. Individual patients do not care about percentages as much as they do about their own individual care. A CDSS that corrects a potential central nervous system drug interaction for that individual patient is a valuable CDSS. Presumably as informaticians develop more sophisticated CDSS, they will be adopted and used more consistently across healthcare.
David Schanding, M.A., M.M.