Effect of Computerized Physician Order Entry with Clinical Decision Support on Adverse Drug Events in the Long-term Care

1. Effect of Computerized Physician Order Entry with Clinical Decision Support on Adverse Drug Events in the Long-term Care Setting Jerry H. Gurwitz, M.D. Chief, Division of Geriatric Medicine University of Massachusetts Medical School Executive Director Meyers Primary Care Institute Worcester, Massachusetts Dr. Atkinson, Dr. Lietman, members of the Society, and guests. I would like to express my appreciation for this truly special honor. I would also like to give special thanks to Bill Abrams who I consider a mentor, a colleague, and a friend. Over the years, Bill has been a major force in stimulating my interest in the field of Geriatric Clinical Pharmacology. As much as any other single individual, Bill was responsible for the establishment of the Merck/AFAR Fellowship in Geriatric Clinical Pharmacology which I had the opportunity to benefit from and which has supported the training of 18 young investigators since its inception in 1988. As many of you are aware, the Society yesterday awarded the first William B. Abrams Award in Geriatric Clinical Pharmacology. This award will continue to honor Bill for many years to come. I’d like to take this opportunity to thank Bill for his boundless and unselfish efforts on behalf of our Society and on behalf of the field of Clinical Pharmacology. This morning, I will be discussing the topic of Geriatric Pharmacotherapy and I’d like to begin with a quote from a major figure in the history of modern medicine- William Withering.Dr. Atkinson, Dr. Lietman, members of the Society, and guests. I would like to express my appreciation for this truly special honor. I would also like to give special thanks to Bill Abrams who I consider a mentor, a colleague, and a friend. Over the years, Bill has been a major force in stimulating my interest in the field of Geriatric Clinical Pharmacology. As much as any other single individual, Bill was responsible for the establishment of the Merck/AFAR Fellowship in Geriatric Clinical Pharmacology which I had the opportunity to benefit from and which has supported the training of 18 young investigators since its inception in 1988. As many of you are aware, the Society yesterday awarded the first William B. Abrams Award in Geriatric Clinical Pharmacology. This award will continue to honor Bill for many years to come. I’d like to take this opportunity to thank Bill for his boundless and unselfish efforts on behalf of our Society and on behalf of the field of Clinical Pharmacology. This morning, I will be discussing the topic of Geriatric Pharmacotherapy and I’d like to begin with a quote from a major figure in the history of modern medicine- William Withering.

2. Co-investigators Terry S. Field Paula Rochon James Judge Leslie Harrold Monica Lee Kathleen White Jane LaPrino Janet Erramuspe-Mainard Martin DeFlorio Linda Gavendo Chaim Bell David Bates

3. Disclosure Statement The research reported during this presentation was supported by grants from the Agency for Healthcare Research and Quality. The investigators retained full independence in the conduct of this research. Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug

4. Adverse Drug Events

5. Introduction Adverse drug events (ADEs) occur frequently among nursing home residents, and preventable adverse drug events are most commonly associated with errors in medication ordering and monitoring. Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug

6. Incidence of ADEs in Two Large Academic LTC Facilities Adverse drug events About 10 ADEs per 100 resident-months Preventable adverse drug events About 4 preventable ADEs per 100 resident-months In other words, if you have 100 residents, 2 events per monthApproximately half of the events we found were judged preventable - high rateIn other words, if you have 100 residents, 2 events per monthApproximately half of the events we found were judged preventable - high rate

7. Error Stage for Preventable ADEs

8. What is the right approach?A systems-based approach

9. Computerized Clinical Decision Support System (CDSS) High-severity drug interactions Potentially problematic laboratory test results Early identification of adverse drug effects through increased monitoring Recommendations regarding geriatric-appropriate dosing Recommendations for prophylactic measures Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules  determining whether ADE required explicit link to drug  included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules determining whether ADE required explicit link to drug included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)

10. Purpose The purpose of this study was to evaluate the efficacy of computerized physician order entry with clinical decision support for preventing ADEs in the long-term care setting. Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug

11. Methods Study conducted in two large academic long-term care facilities Total of 1229 beds Total of 29 resident care units were randomized All units had existing CPOE Units randomized to having the CDSS or not Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules  determining whether ADE required explicit link to drug  included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules determining whether ADE required explicit link to drug included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)

12. CPOE with Clinical Decision Support

13. Methods Drug-related incidents were detected using multiple methods: Review of long-term care facility records in monthly segments Computer-generated signals Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules  determining whether ADE required explicit link to drug  included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules determining whether ADE required explicit link to drug included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)

14. Computer Generated Signals Abnormal laboratory results Elevated INRs, high potassium levels Medications (antidotes) Vitamin K, sodium polystyrene sulfonate Abnormal drug levels Phenytoin Digoxin Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules  determining whether ADE required explicit link to drug  included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules determining whether ADE required explicit link to drug included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)

15. Methods Chart reviews were performed by trained clinical pharmacist investigators Incidents were classified independently by two physician reviewers: adverse drug event severity preventability Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules  determining whether ADE required explicit link to drug  included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules determining whether ADE required explicit link to drug included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)

16. Results

17. Effect of CPOE with CDS on ADE Rates

18. Effect of CPOE with CDS on Preventable ADE Rates

19. Conclusion Use of CPOE with this particular computerized clinical decision support system was not found to reduce the occurrence of ADEs in the long-term care setting. Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug

20. Why?… The limits of a first-generation system Lack of specificity of alerts – alert burden Need to increase scope of system to address a broader range of ADEs Need to integrate more clinical information into the clinical decision support system Setting the bar too high: ADEs vs errors Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug Study establishes the extent and seriousness of the problem clarifies the potential for prevention lays the ground work for determining what directions needed in designing interventions We believe this is a very conservative estimate - tip of the iceberg, because: used medical record review - limits in quality of nursing home records strictly information entered or reported by providers high end homes rigid about definition of ADE - had to be strong evidence linking symptoms to drug

22. Computerized Clinical Decision Support System (CDSS) Warnings to reconsider specific drug orders Recommendations for laboratory monitoring Alerts to monitor closely for selected adverse drug effects Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules  determining whether ADE required explicit link to drug  included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)Every six weeks, an investigator reviewed charts on all patients looking for: changes in medications and discontinuations specific medications which might be used to treat an ade unusual lab values changes in symptoms and events which might be linked to an ade - lethargy, confusion, bleeding, falls, GI problems hospitalizations and ER visits Log books left at each nursing station - very few events reported that way Classification process - no explicit rules determining whether ADE required explicit link to drug included rating of confidence on 6 point scale - only used those with upper levels assigning level of severity range: fatal, life-threatening, serious, significant examples: neuropsych event - if oversedation significant, if delirium serious; hemorrhaghic event - if a bleed wignificant, if required transfusion or hospitalization serious; falls always serious determining preventability range: no error, definitely not preventable, probably not preventable probably preventable, definitely preventable ex of preventable: resident has a documented prior allergy to a drug; resident is given the drug and has an ADE - called “preventable” because allergy was known (even if not easily found in the chart)

23. Computer on Wheels -“COW”