Using Computerized Provider Order Entry and Clinical Decision Support to Improve Prescribing in Patients With Decreased GFR
2010; Elsevier BV; Volume: 56; Issue: 5 Linguagem: Inglês
10.1053/j.ajkd.2010.09.006
ISSN1523-6838
AutoresWilliam Galanter, Jessica Moja, Bruce L. Lambert,
Tópico(s)Patient Safety and Medication Errors
ResumoRelated Article, p. 832Adverse drug events due to medication errors are associated with increased mortality, length of stay, and cost.1Kohn L.T. Corrigan J.M. Donaldson M.S. To err is human: building a safer health system. National Academy Press, Washington, DC1999Google Scholar, 2Classen D.C. Pestotnik S.L. Evans R.S. Lloyd J.F. Burke J.P. Adverse drug events in hospitalized patients Excess length of stay, extra costs, and attributable mortality.JAMA. 1997; 277: 301-306Crossref PubMed Google Scholar Studies have demonstrated the effectiveness of health information technology, specifically computerized physician order entry3Bates D.W. Teich J.M. Lee J. et al.The impact of computerized physician order entry on medication error prevention.J Am Med Inform Assoc. 1999; 6: 313-321Crossref PubMed Scopus (997) Google Scholar, 4Bates D.W. Leape L.L. Cullen D.J. et al.Effect of computerized physician order entry and a team intervention on prevention of serious medication errors.JAMA. 1998; 280: 1311-1316Crossref PubMed Scopus (1650) Google Scholar and clinical decision support5Ammenwerth E. Schnell-Inderst P. Machan C. Siebert U. The effect of electronic prescribing on medication errors and adverse drug events: a systematic review.J Am Med Inform Assoc. 2008; 15: 585-600Crossref PubMed Scopus (486) Google Scholar in reducing the incidence and severity of medication errors. The Federal Government is now offering financial incentives to promote the use of computerized physician order entry and clinical decision support as a means of improving care and decreasing errors.6Blumenthal D. Tavenner M. The "meaningful use" regulation for electronic health records.N Engl J Med. 2010; ([published online ahead of print July 13]) (doi:10.1056/NEJMp1006114)Google ScholarThe rate of medication errors among hospitalized patients with decreased glomerular filtration rate (GFR) is high, between 20% and 50%.7Oppenheim M.I. Vidal C. Velasco F.T. et al.Impact of a computerized alert during physician order entry on medication dosing in patients with renal impairment.Proc AIMA Symp. 2002; : 577-581Google Scholar, 8Markota N.P. Markota I. Tomic M. Zelenika A. Inappropriate drug dosages adjustments in patients with renal impairment.J Nephrol. 2009; 22: 497-501PubMed Google Scholar, 9Sellier E. Colombet I. Sabatier B. et al.Effect of alerts for drug dosage adjustment in inpatients with renal insufficiency.J Am Med Inform Assoc. 2009; 16: 203-210Crossref PubMed Scopus (44) Google Scholar, 10Sheen S.S. Choi J.E. Park R.W. Kim E.Y. Lee Y.H. Kang U.G. Overdose rate of drugs requiring renal dose adjustment: data analysis of 4 years prescriptions at a tertiary teaching hospital.J Gen Intern Med. 2007; 23: 423-428Crossref Scopus (26) Google Scholar, 11Bertsche T. Fleischer M. Pfaff J. Encke J. Czock D. Haefeli W.E. Pro-active provision of drug information as a technique to address overdosing in intensive-care patients with renal insufficiency.Eur J Clin Pharmacol. 2009; 65: 823-829Crossref PubMed Scopus (18) Google Scholar, 12Salomon L. Deray G. Jaudon M.C. et al.Medication misuse in hospitalized patients with renal impairment.Int J Qual Health Care. 2003; 15: 331-335Crossref PubMed Scopus (76) Google Scholar In a recent study, almost half of patients admitted with decreased GFR had either potential or actual adverse drug events.13Hug B.L. Witkowski D.J. Sox C.M. et al.Occurrence of adverse, often preventable, events in community hospitals involving nephrotoxic drugs or those excreted by the kidney.Kidney Int. 2009; 76: 1192-1198Crossref PubMed Scopus (74) Google Scholar Given the high rate of adverse drug events among patients with decreased GFR, and the demonstrated ability of computerized physician order entry and clinical decision support to reduce adverse drug events, there is a great opportunity to improve patient care by applying these tools in this context. In this issue of the American Journal of Kidney Diseases, McCoy et al demonstrate that computerized physician order entry and clinical decision support can improve the quality of care for patients with decreased GFR.14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google ScholarMcCoy and colleagues questioned whether computerized physician order entry with clinical decision support could improve the quality of medication management for hospitalized patients with acute kidney injury (AKI). A study of clinical decision support for AKI has previously been reported by Rind et al15Rind D.M. Safran C. Phillips R.S. et al.Effect of computer-based alerts on the treatment outcomes of hospitalized patients.Arch Intern Med. 1994; 154: 1511-1517Crossref PubMed Scopus (238) Google Scholar using clinical decision support that delivered e-mail–type messages to clinicians when a patient with AKI was prescribed a nephrotoxic medication. The study by Rind et al demonstrated an improvement in the timeliness of the response of clinicians to the potentially dangerous situation.The clinical decision support employed in the study by McCoy et al14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar focused on medication use in patients with AKI, defined as an increase in creatinine of 0.5 mg/dL (44.2 μmol/L) in 48 hours.14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar The approach was novel in a few respects. First, it used a multimodal clinical decision support with 3 ways of informing clinicians about the AKI and medications of concern: (1) passive alerts which could be viewed in the electronic medical record during computerized physician order entry, (2) alerts printed on electronic medical record–generated rounding reports,16Van Eaton E.G. Horvath K.D. Lober W.B. Rossini A.J. Pellegrini C.A. A randomized, controlled trial evaluating the impact of a computerized rounding and sign-out system on continuity of care and resident work hours.J Am Coll Surg. 2005; 200: 538-545Abstract Full Text Full Text PDF PubMed Scopus (220) Google Scholar and (3) interruptive alerts demanding action at the end of an ordering session. Previous studies suggest that passive alerts are not as useful as interruptive alerts,17Balcezak T.J. Krumholz H.M. Getnick G.S. Vaccarino V. Lin Z.Q. Cadman E.C. Utilization and effectiveness of a weight-based heparin nomogram at a large academic medical center.Am J Manag Care. 2000; 6: 329-338PubMed Google Scholar, 18Chrischilles E.A. Fulda T.R. Byrns P.J. Winckler S.C. Rupp M.T. Chui M.A. The role of pharmacy computer systems in preventing medication errors.J Am Pharm Assoc (Wash). 2002; 42: 439-448Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar but passive alerts are less annoying to providers than interruptive alerts. The idea behind multimodal alerts was to give clinicians the opportunity to view the passive alerts and take appropriate actions prior to being interrupted.Second, clinical decision support attempted to minimize interruptive alerts for patients with severely decreased GFR at the time of hospital admission (defined as creatinine clearance less than 30 mL/min [.5 mL/s]) as well as patients "flagged" by physicians as receiving dialysis. The rationale was to avoid alerts for patients in whom the threshold increase in serum creatinine represented clinically insignificant changes in GFR, as in these cases the alerts could be a nuisance.The authors demonstrated that the computerized physician order entry–based interruptive alerts improved the rate of order modification of potentially nephrotoxic medications or medications cleared by the kidneys from 35% to 56% during the first 24 hours after the detection of AKI. This finding is important as this is an inexpensive intervention with important benefit to patients, but, as found by other studies in this field,7Oppenheim M.I. Vidal C. Velasco F.T. et al.Impact of a computerized alert during physician order entry on medication dosing in patients with renal impairment.Proc AIMA Symp. 2002; : 577-581Google Scholar, 19Chertow G.M. Lee J. Kuperman G.J. et al.Guided medication dosing for inpatients with renal insufficiency.JAMA. 2001; 286: 2839-2844Crossref PubMed Scopus (345) Google Scholar, 20Galanter W.L. Didomenico R.J. Polikaitis A. A trial of automated decision support alerts for contraindicated medications using computerized physician order entry.J Am Med Inform Assoc. 2005; 12: 269-274Crossref PubMed Scopus (112) Google Scholar the effectiveness was limited by nonadherence. The passive alerts as a whole demonstrated no benefit.Nonadherence with clinical decision support alerts has hindered many efforts to improve prescribing in patients with decreased GFR. Previous studies have demonstrated the effectiveness of clinical decision support in improving initial dosing and medication selection in hospitalized patients with decreased GFR. Chertow et al19Chertow G.M. Lee J. Kuperman G.J. et al.Guided medication dosing for inpatients with renal insufficiency.JAMA. 2001; 286: 2839-2844Crossref PubMed Scopus (345) Google Scholar used a computerized physician order entry user interface designed to give the appropriate medication dose and frequency based on the GFR as the highlighted default menu item during computerized physician order entry. This intervention improved appropriate ordering, but neither the dose nor frequency was correct more than two-thirds of the time after the intervention. Oppenheim et al7Oppenheim M.I. Vidal C. Velasco F.T. et al.Impact of a computerized alert during physician order entry on medication dosing in patients with renal impairment.Proc AIMA Symp. 2002; : 577-581Google Scholar provided ordering clinicians with alerts at completion of a computerized entry for a dose that was excessive based on the patient's level of GFR. Roughly half of the ordering clinicians ignored the alerts. Galanter et al20Galanter W.L. Didomenico R.J. Polikaitis A. A trial of automated decision support alerts for contraindicated medications using computerized physician order entry.J Am Med Inform Assoc. 2005; 12: 269-274Crossref PubMed Scopus (112) Google Scholar implemented a system to warn against ordering contraindicated medications in patients according to the level of GFR. As with the other studies described, the intervention was efficacious, but limited by nonadherence.Nonadherence is not necessarily an inappropriate action, though well-designed clinical decision support should produce alerts for which nonadherence is more likely associated with a medication error than adherence. Distinguishing between clinically appropriate and inappropriate disregard for an alert is a measurement problem as it requires careful chart review and clinical judgment, which are expensive and time-consuming. Due to the difficulty in measuring the appropriateness of an alert and actions taken after the alert, the compliance to alerts is typically measured, not the appropriateness of the clinical action. Research into clinical decision support, clinical logic design, user interface, and human factors needs to provide solutions for nonadherence to be applied to future clinical decision support systems.The exclusion of patients receiving dialysis was an important element of the clinical decision support design by McCoy et al to decrease nuisance created by unnecessary alerts.14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar Anecdotal evidence from our own institution suggests that imperfect exclusion of patients receiving dialysis generates significant numbers of nuisance alerts. In the present study, though the total number of patients receiving dialysis was not reported, only 7% of the patients for whom alerts were generated had received dialysis. If a dialysis session is treated as a procedure that is charted in the electronic medical record as an actionable piece of data, clinician input would not be necessary to "flag" the patient. This will cause a decrease in nuisance alerts, but of equal or greater importance will be the ability of clinical decision support safety systems which have been shut off to begin alerting again in patients whose GFR improves or in patients who receive kidney transplants. Clinicians and the clinical decision support should be on high alert for new episodes of AKI in such patients. The clinical decision support designed in this study would not be able to deal with an episode of AKI that occurred after initial identification as a dialysis patient.In the study of McCoy et al,14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar the threshold increase in serum creatinine of 0.5 mg/dL (44.2 μmol/L) for the definition of AKI was the same as that used by Rind et al,15Rind D.M. Safran C. Phillips R.S. et al.Effect of computer-based alerts on the treatment outcomes of hospitalized patients.Arch Intern Med. 1994; 154: 1511-1517Crossref PubMed Scopus (238) Google Scholar and slightly higher than the Acute Kidney Injury Network (AKIN)21Mehta R. Kellum J. Shah S. et al.Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury.Crit Care. 2007; 11: R31Crossref PubMed Scopus (5123) Google Scholar criterion based on absolute creatinine values alone. One of the problems with definitions based on absolute changes in creatinine is that decrease in GFR is not proportional to the creatinine, but rather to the inverse of the creatinine.22Cockcroft D.W. Gault M.H. Prediction of creatinine clearance from serum creatinine.Nephron. 1976; 16: 31-41Crossref PubMed Scopus (12988) Google Scholar Thus, an absolute increase is a sensitive marker with high-baseline GFR but a more nonspecific marker with lower-baseline GFR.An ideal system should be able to alert on a decrease in GFR from 25 mL/min to 10 mL/min (0.42 mL/s to 0.17 mL/s), without producing an alert on a trivial change from 25 mL/min to 20 mL/min (0.42 mL/s to 0.33 mL/s) which may be associated with an absolute creatinine increase of roughly 0.5 mg/dL (44.2 μmol/L) in a 65-year-old man. Use of estimated GFR rather than serum creatinine may be helpful for this purpose.23Stevens L.A. Levey A.S. Measured GFR as a confirmatory test for estimated GFR.J Am Soc Nephrol. 2009; 20: 2305-2313Crossref PubMed Scopus (396) Google Scholar As proposed in the AKIN criteria,21Mehta R. Kellum J. Shah S. et al.Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury.Crit Care. 2007; 11: R31Crossref PubMed Scopus (5123) Google Scholar an additional definition which looks at the percent increase in creatinine could be helpful in patients with more severely decreased baseline GFR, but who still may be experiencing a clinically important deterioration, potentially from exposure to a nephrotoxic agent. The criteria used by AKIN stage 3 of a 1.5-fold increase21Mehta R. Kellum J. Shah S. et al.Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury.Crit Care. 2007; 11: R31Crossref PubMed Scopus (5123) Google Scholar in creatinine could be a good starting point for further study and would work in the example provided above.The refinements studied by McCoy et al,14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar namely, using both passive and interruptive alerts and allowing clinicians to stop alerts for patients receiving dialysis, should improve our ability to design clinical decision support that helps clinicians care for hospitalized patients with AKI. Like previously published studies, the intervention reduced but did not eradicate the medication errors, and effectiveness appeared to be related to prescriber nonadherence with alerts. Given the high prevalence of decreased GFR in hospitalized patients and the vulnerability of these patients to harmful adverse drug events, the need for continued improvement of clinical decision support systems to reduce medication errors in these patients and to identify and avoid potential nephrotoxins should continue to be a high priority for research and practice. Related Article, p. 832 Related Article, p. 832 Related Article, p. 832 Adverse drug events due to medication errors are associated with increased mortality, length of stay, and cost.1Kohn L.T. Corrigan J.M. Donaldson M.S. To err is human: building a safer health system. National Academy Press, Washington, DC1999Google Scholar, 2Classen D.C. Pestotnik S.L. Evans R.S. Lloyd J.F. Burke J.P. Adverse drug events in hospitalized patients Excess length of stay, extra costs, and attributable mortality.JAMA. 1997; 277: 301-306Crossref PubMed Google Scholar Studies have demonstrated the effectiveness of health information technology, specifically computerized physician order entry3Bates D.W. Teich J.M. Lee J. et al.The impact of computerized physician order entry on medication error prevention.J Am Med Inform Assoc. 1999; 6: 313-321Crossref PubMed Scopus (997) Google Scholar, 4Bates D.W. Leape L.L. Cullen D.J. et al.Effect of computerized physician order entry and a team intervention on prevention of serious medication errors.JAMA. 1998; 280: 1311-1316Crossref PubMed Scopus (1650) Google Scholar and clinical decision support5Ammenwerth E. Schnell-Inderst P. Machan C. Siebert U. The effect of electronic prescribing on medication errors and adverse drug events: a systematic review.J Am Med Inform Assoc. 2008; 15: 585-600Crossref PubMed Scopus (486) Google Scholar in reducing the incidence and severity of medication errors. The Federal Government is now offering financial incentives to promote the use of computerized physician order entry and clinical decision support as a means of improving care and decreasing errors.6Blumenthal D. Tavenner M. The "meaningful use" regulation for electronic health records.N Engl J Med. 2010; ([published online ahead of print July 13]) (doi:10.1056/NEJMp1006114)Google Scholar The rate of medication errors among hospitalized patients with decreased glomerular filtration rate (GFR) is high, between 20% and 50%.7Oppenheim M.I. Vidal C. Velasco F.T. et al.Impact of a computerized alert during physician order entry on medication dosing in patients with renal impairment.Proc AIMA Symp. 2002; : 577-581Google Scholar, 8Markota N.P. Markota I. Tomic M. Zelenika A. Inappropriate drug dosages adjustments in patients with renal impairment.J Nephrol. 2009; 22: 497-501PubMed Google Scholar, 9Sellier E. Colombet I. Sabatier B. et al.Effect of alerts for drug dosage adjustment in inpatients with renal insufficiency.J Am Med Inform Assoc. 2009; 16: 203-210Crossref PubMed Scopus (44) Google Scholar, 10Sheen S.S. Choi J.E. Park R.W. Kim E.Y. Lee Y.H. Kang U.G. Overdose rate of drugs requiring renal dose adjustment: data analysis of 4 years prescriptions at a tertiary teaching hospital.J Gen Intern Med. 2007; 23: 423-428Crossref Scopus (26) Google Scholar, 11Bertsche T. Fleischer M. Pfaff J. Encke J. Czock D. Haefeli W.E. Pro-active provision of drug information as a technique to address overdosing in intensive-care patients with renal insufficiency.Eur J Clin Pharmacol. 2009; 65: 823-829Crossref PubMed Scopus (18) Google Scholar, 12Salomon L. Deray G. Jaudon M.C. et al.Medication misuse in hospitalized patients with renal impairment.Int J Qual Health Care. 2003; 15: 331-335Crossref PubMed Scopus (76) Google Scholar In a recent study, almost half of patients admitted with decreased GFR had either potential or actual adverse drug events.13Hug B.L. Witkowski D.J. Sox C.M. et al.Occurrence of adverse, often preventable, events in community hospitals involving nephrotoxic drugs or those excreted by the kidney.Kidney Int. 2009; 76: 1192-1198Crossref PubMed Scopus (74) Google Scholar Given the high rate of adverse drug events among patients with decreased GFR, and the demonstrated ability of computerized physician order entry and clinical decision support to reduce adverse drug events, there is a great opportunity to improve patient care by applying these tools in this context. In this issue of the American Journal of Kidney Diseases, McCoy et al demonstrate that computerized physician order entry and clinical decision support can improve the quality of care for patients with decreased GFR.14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar McCoy and colleagues questioned whether computerized physician order entry with clinical decision support could improve the quality of medication management for hospitalized patients with acute kidney injury (AKI). A study of clinical decision support for AKI has previously been reported by Rind et al15Rind D.M. Safran C. Phillips R.S. et al.Effect of computer-based alerts on the treatment outcomes of hospitalized patients.Arch Intern Med. 1994; 154: 1511-1517Crossref PubMed Scopus (238) Google Scholar using clinical decision support that delivered e-mail–type messages to clinicians when a patient with AKI was prescribed a nephrotoxic medication. The study by Rind et al demonstrated an improvement in the timeliness of the response of clinicians to the potentially dangerous situation. The clinical decision support employed in the study by McCoy et al14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar focused on medication use in patients with AKI, defined as an increase in creatinine of 0.5 mg/dL (44.2 μmol/L) in 48 hours.14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar The approach was novel in a few respects. First, it used a multimodal clinical decision support with 3 ways of informing clinicians about the AKI and medications of concern: (1) passive alerts which could be viewed in the electronic medical record during computerized physician order entry, (2) alerts printed on electronic medical record–generated rounding reports,16Van Eaton E.G. Horvath K.D. Lober W.B. Rossini A.J. Pellegrini C.A. A randomized, controlled trial evaluating the impact of a computerized rounding and sign-out system on continuity of care and resident work hours.J Am Coll Surg. 2005; 200: 538-545Abstract Full Text Full Text PDF PubMed Scopus (220) Google Scholar and (3) interruptive alerts demanding action at the end of an ordering session. Previous studies suggest that passive alerts are not as useful as interruptive alerts,17Balcezak T.J. Krumholz H.M. Getnick G.S. Vaccarino V. Lin Z.Q. Cadman E.C. Utilization and effectiveness of a weight-based heparin nomogram at a large academic medical center.Am J Manag Care. 2000; 6: 329-338PubMed Google Scholar, 18Chrischilles E.A. Fulda T.R. Byrns P.J. Winckler S.C. Rupp M.T. Chui M.A. The role of pharmacy computer systems in preventing medication errors.J Am Pharm Assoc (Wash). 2002; 42: 439-448Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar but passive alerts are less annoying to providers than interruptive alerts. The idea behind multimodal alerts was to give clinicians the opportunity to view the passive alerts and take appropriate actions prior to being interrupted. Second, clinical decision support attempted to minimize interruptive alerts for patients with severely decreased GFR at the time of hospital admission (defined as creatinine clearance less than 30 mL/min [.5 mL/s]) as well as patients "flagged" by physicians as receiving dialysis. The rationale was to avoid alerts for patients in whom the threshold increase in serum creatinine represented clinically insignificant changes in GFR, as in these cases the alerts could be a nuisance. The authors demonstrated that the computerized physician order entry–based interruptive alerts improved the rate of order modification of potentially nephrotoxic medications or medications cleared by the kidneys from 35% to 56% during the first 24 hours after the detection of AKI. This finding is important as this is an inexpensive intervention with important benefit to patients, but, as found by other studies in this field,7Oppenheim M.I. Vidal C. Velasco F.T. et al.Impact of a computerized alert during physician order entry on medication dosing in patients with renal impairment.Proc AIMA Symp. 2002; : 577-581Google Scholar, 19Chertow G.M. Lee J. Kuperman G.J. et al.Guided medication dosing for inpatients with renal insufficiency.JAMA. 2001; 286: 2839-2844Crossref PubMed Scopus (345) Google Scholar, 20Galanter W.L. Didomenico R.J. Polikaitis A. A trial of automated decision support alerts for contraindicated medications using computerized physician order entry.J Am Med Inform Assoc. 2005; 12: 269-274Crossref PubMed Scopus (112) Google Scholar the effectiveness was limited by nonadherence. The passive alerts as a whole demonstrated no benefit. Nonadherence with clinical decision support alerts has hindered many efforts to improve prescribing in patients with decreased GFR. Previous studies have demonstrated the effectiveness of clinical decision support in improving initial dosing and medication selection in hospitalized patients with decreased GFR. Chertow et al19Chertow G.M. Lee J. Kuperman G.J. et al.Guided medication dosing for inpatients with renal insufficiency.JAMA. 2001; 286: 2839-2844Crossref PubMed Scopus (345) Google Scholar used a computerized physician order entry user interface designed to give the appropriate medication dose and frequency based on the GFR as the highlighted default menu item during computerized physician order entry. This intervention improved appropriate ordering, but neither the dose nor frequency was correct more than two-thirds of the time after the intervention. Oppenheim et al7Oppenheim M.I. Vidal C. Velasco F.T. et al.Impact of a computerized alert during physician order entry on medication dosing in patients with renal impairment.Proc AIMA Symp. 2002; : 577-581Google Scholar provided ordering clinicians with alerts at completion of a computerized entry for a dose that was excessive based on the patient's level of GFR. Roughly half of the ordering clinicians ignored the alerts. Galanter et al20Galanter W.L. Didomenico R.J. Polikaitis A. A trial of automated decision support alerts for contraindicated medications using computerized physician order entry.J Am Med Inform Assoc. 2005; 12: 269-274Crossref PubMed Scopus (112) Google Scholar implemented a system to warn against ordering contraindicated medications in patients according to the level of GFR. As with the other studies described, the intervention was efficacious, but limited by nonadherence. Nonadherence is not necessarily an inappropriate action, though well-designed clinical decision support should produce alerts for which nonadherence is more likely associated with a medication error than adherence. Distinguishing between clinically appropriate and inappropriate disregard for an alert is a measurement problem as it requires careful chart review and clinical judgment, which are expensive and time-consuming. Due to the difficulty in measuring the appropriateness of an alert and actions taken after the alert, the compliance to alerts is typically measured, not the appropriateness of the clinical action. Research into clinical decision support, clinical logic design, user interface, and human factors needs to provide solutions for nonadherence to be applied to future clinical decision support systems. The exclusion of patients receiving dialysis was an important element of the clinical decision support design by McCoy et al to decrease nuisance created by unnecessary alerts.14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar Anecdotal evidence from our own institution suggests that imperfect exclusion of patients receiving dialysis generates significant numbers of nuisance alerts. In the present study, though the total number of patients receiving dialysis was not reported, only 7% of the patients for whom alerts were generated had received dialysis. If a dialysis session is treated as a procedure that is charted in the electronic medical record as an actionable piece of data, clinician input would not be necessary to "flag" the patient. This will cause a decrease in nuisance alerts, but of equal or greater importance will be the ability of clinical decision support safety systems which have been shut off to begin alerting again in patients whose GFR improves or in patients who receive kidney transplants. Clinicians and the clinical decision support should be on high alert for new episodes of AKI in such patients. The clinical decision support designed in this study would not be able to deal with an episode of AKI that occurred after initial identification as a dialysis patient. In the study of McCoy et al,14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar the threshold increase in serum creatinine of 0.5 mg/dL (44.2 μmol/L) for the definition of AKI was the same as that used by Rind et al,15Rind D.M. Safran C. Phillips R.S. et al.Effect of computer-based alerts on the treatment outcomes of hospitalized patients.Arch Intern Med. 1994; 154: 1511-1517Crossref PubMed Scopus (238) Google Scholar and slightly higher than the Acute Kidney Injury Network (AKIN)21Mehta R. Kellum J. Shah S. et al.Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury.Crit Care. 2007; 11: R31Crossref PubMed Scopus (5123) Google Scholar criterion based on absolute creatinine values alone. One of the problems with definitions based on absolute changes in creatinine is that decrease in GFR is not proportional to the creatinine, but rather to the inverse of the creatinine.22Cockcroft D.W. Gault M.H. Prediction of creatinine clearance from serum creatinine.Nephron. 1976; 16: 31-41Crossref PubMed Scopus (12988) Google Scholar Thus, an absolute increase is a sensitive marker with high-baseline GFR but a more nonspecific marker with lower-baseline GFR. An ideal system should be able to alert on a decrease in GFR from 25 mL/min to 10 mL/min (0.42 mL/s to 0.17 mL/s), without producing an alert on a trivial change from 25 mL/min to 20 mL/min (0.42 mL/s to 0.33 mL/s) which may be associated with an absolute creatinine increase of roughly 0.5 mg/dL (44.2 μmol/L) in a 65-year-old man. Use of estimated GFR rather than serum creatinine may be helpful for this purpose.23Stevens L.A. Levey A.S. Measured GFR as a confirmatory test for estimated GFR.J Am Soc Nephrol. 2009; 20: 2305-2313Crossref PubMed Scopus (396) Google Scholar As proposed in the AKIN criteria,21Mehta R. Kellum J. Shah S. et al.Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury.Crit Care. 2007; 11: R31Crossref PubMed Scopus (5123) Google Scholar an additional definition which looks at the percent increase in creatinine could be helpful in patients with more severely decreased baseline GFR, but who still may be experiencing a clinically important deterioration, potentially from exposure to a nephrotoxic agent. The criteria used by AKIN stage 3 of a 1.5-fold increase21Mehta R. Kellum J. Shah S. et al.Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury.Crit Care. 2007; 11: R31Crossref PubMed Scopus (5123) Google Scholar in creatinine could be a good starting point for further study and would work in the example provided above. The refinements studied by McCoy et al,14McCoy A.B. Waitman L.R. Gadd C.S. et al.A computerized provider order entry intervention for medication safety during acute kidney injury: a quality improvement report.Am J Kidney Dis. 2010; 56: 832-841Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar namely, using both passive and interruptive alerts and allowing clinicians to stop alerts for patients receiving dialysis, should improve our ability to design clinical decision support that helps clinicians care for hospitalized patients with AKI. Like previously published studies, the intervention reduced but did not eradicate the medication errors, and effectiveness appeared to be related to prescriber nonadherence with alerts. Given the high prevalence of decreased GFR in hospitalized patients and the vulnerability of these patients to harmful adverse drug events, the need for continued improvement of clinical decision support systems to reduce medication errors in these patients and to identify and avoid potential nephrotoxins should continue to be a high priority for research and practice. The content of this editorial is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality. Support: Drs Galanter and Lambert are supported by grant number U18HS016973 from the Agency for Healthcare Research and Quality , US Department of Health & Human Services . Financial Disclosure: Dr Lambert is the Principal Investigator for a study on medication safety at the University of Illinois at Chicago Center for Education and Research on Therapuetics, that involves using clinical decision support to link lab and pharmacy data systems. The study is funded by the Agency for Healthcare Research and Quality , US Department of Health & Human Services and receives in kind support from Cerner Corporation. A Computerized Provider Order Entry Intervention for Medication Safety During Acute Kidney Injury: A Quality Improvement ReportAmerican Journal of Kidney DiseasesVol. 56Issue 5PreviewFrequently, prescribers fail to account for changing kidney function when prescribing medications. We evaluated the use of a computerized provider order entry intervention to improve medication management during acute kidney injury. Full-Text PDF
Referência(s)