- Email: [email protected]
Automated Dispensing Cabinet Alert Improves Compliance with Obtaining Blood Cultures Before Antibiotic Administration for Patients Admitted with Pneumonia
The Joint Commission Journal on Quality and Patient Safety Information Technology
Automated Dispensing Cabinet Alert Improves Compliance with Obtaining Blood Cultures Before Antibiotic Administration for Patients Admitted with Pneumonia Rishi Sikka, MD; Rolla Sweis, PharmD, MA, BCPS; Carleen Kaucky, RN, BSN; Erik Kulstad, MD, MS
T
he US Centers for Medicare & Medicaid Services (CMS) has identified patients hospitalized with pneumonia as an important target for quality improvement.1 Despite disagreement regarding the utility of blood cultures for patients admitted with pneumonia,2–5 CMS and The Joint Commission Pneumonia National Hospital Inpatient Quality Measures include PN-3b, “Blood cultures performed in the emergency department prior to initial antibiotic received in hospital.”1,6,7 Blood cultures may be ordered on the basis of clinical discretion on pneumonia patients admitted to the floor or telemetry. It is generally acknowledged that the yield of clinically useful information is greater if the culture is collected before antibiotics are administered.8,9 Failure to check blood cultures prior to antibiotic administration may affect the growth of any bacteria and prevent a culture from becoming positive, further reducing the yield of a test estimated by some to require as many as 100 patients to identify 1 patient infected with an organism resistant to empiric therapy.3 Consequently, the CMS quality goal specifically measures the number of pneumonia patients whose initial ED blood culture specimen was collected prior to first hospital dose of antibiotics.1 Despite the importance that CMS ascribes to blood culture collection prior to antibiotic administration, there is little published information on a process or intervention to improve compliance with the corresponding quality measure. As these measures receive greater scrutiny by both payers and the public, it becomes important to study, establish, and disseminate proven recommendations that may be adapted in a variety of settings to improve the quality of care. This article reports a study undertaken to measure the impact of an automated dispensing cabinet (ADC) alert on improving compliance with blood cultures prior to antibiotics for patients admitted with pneumonia.
224
May 2012
Article-at-a-Glance Background: A Centers for Medicare & Medicaid Services
(CMS) pneumonia quality measures with particular impact on the emergency department (ED) is blood cultures prior to antibiotic administration for patients admitted with pneumonia. A study was conducted to measure the impact of an automated dispensing cabinet (ADC) alert on improving compliance with the quality measure of obtaining blood cultures prior to giving antibiotics for patients admitted with pneumonia and who have blood cultures ordered. Methods: The pre-post study involved ED adult patients with an admitting diagnosis of pneumonia from October 2007 through September 2008. The intervention consisted of a series of questions in the ED medication ADC regarding blood culture orders and antibiotic administration. Patients with an admitting diagnosis of pneumonia were identified through a search of the ED electronic health record (EHR). The proportion of patients in whom blood cultures were obtained prior to antibiotic administration in the pre- (October 2007–March 2008) and postintervention (April 2008–September 2008) periods were compared. The chi-square test was used to test for statistical significance. Results: Some 951 patients with pneumonia were identified during the study period, 426 pre- and 525 postintervention. Compliance with obtaining blood cultures prior to antibiotic administration was 84% (205/245, 95% confidence interval [CI]: 79%–88%) and 95% (275/291, 95% CI: 92%–97%) in the pre and postintervention periods, respectively (p <. 001). Conclusions: In this population of patients with pneumonia, a series of questions in an ADC improved compliance with the quality measure regarding the obtaining of blood cultures prior to administering antibiotics to patients in whom blood cultures are requested.
Volume 38 Number 5
Copyright 2012 © The Joint Commission
The Joint Commission Journal on Quality and Patient Safety Question Flow on Withdrawing Pneumonia Antibiotics from Automated Dispensing Cabinet (ADC)
Yes
(No additional messaging— medication dispensed)
Were blood cultures drawn?
Yes No Were blood cultures ordered?
No
Have the physician order them and make sure the cultures are drawn before you give the antibiotic.
(No additional messaging— medication dispensed)
Figure 1. These questions, which appeared for all antibiotics commonly administered in the emergency department for pneumonia, all required a response by the nurse, and antibiotics were not dispensed until all the questions were answered.
Methods STUDY DESIGN, SETTING, AND POPULATION This pre-post study involved all ED adult patients (age > 18 years) with a primary admitting diagnosis of pneumonia from October 2007 through September 2008. The study setting was a tertiary care, suburban community hospital that has approximately 85,000 ED visits annually and 700 inpatient beds. All patients with a physician-entered primary admitting diagnosis of pneumonia in the electronic health record (EHR) were identified through a search of the EHR. The study was approved by the hospital’s Institutional Review Board.
INTERVENTION The intervention consisted of a series of questions in the ED medication ADC regarding blood culture orders and antibiotic administration. These questions, which appeared for all antibiotics commonly administered in this ED for pneumonia (Figure 1, above), all required a response by the nurse, and antibiotics were not dispensed until all the questions were answered. If the user did not answer all of the questions in the series, medications could not be dispensed, which ensured a compliance of 100% with respect to answering the questions. The sequence of activities for a particular nurse and the place of the intervention are detailed in Figure 2 (page 226). The cascade of action is initiated with a physician written order for antibiotics for a patient with suspected pneumonia. The nurse takes the order to the ADC, selects the patient, and orders the antibiotics. If the antibiotics ordered include medications typically given for pneumonia in our ED, the nurse must answer the se-
May 2012
ries of questions in the ADC. If cultures are ordered and the nurse answers appropriately, the medications are dispensed and can be administered to the patient. However, if cultures are not ordered, the nurse notifies and confirms the intent with phlebotomist and physician and orders the cultures as needed. The rollout for the intervention consisted of education of ED nurses and phlebotomists at regularly scheduled staff meetings. The education, which addressed the purpose of the alerts, the steps involved, and the intervention’s importance, was supplemented by in-person demonstration of the alerts at the ADC by the ED pharmacists. Compliance Reports. On a regular basis, compliance reports were generated from the ADC (Appendix 1, available in online article). These patient-level reports included the specific responses to the alerts input by the nursing staff. The ADC patient reports were cross-referenced retrospectively by the ED outcomes nurse against the list of patients with blood cultures ordered for pneumonia. In instances in which there was noncompliance with the timing of blood cultures and incorrect responses to the ADC alert, the outcomes nurse placed a letter in the nurse’s mailbox detailing the noncompliance (Figure 3, page 227). This was often followed up by in-person, repeat education of the ADC alert by the ED pharmacist [R.S.]; the session included a review of the purpose of the alert and the steps in the process.
DATA COLLECTION AND PROCESSING A trained registered nurse (RN) [C.K.] used a closed-ended abstraction form to collect the time of (1) patient arrival, (2)
Volume 38 Number 5
Copyright 2012 © The Joint Commission
225
The Joint Commission Journal on Quality and Patient Safety Nursing Flow and Intervention Context Physician orders antibiotics for pneumonia patient.
If cultures are ordered, the nurse removes the drug and administers the medication(s).
If cultures are not ordered, the nurse notifies the technician/physician.
Nurse takes orders to ADC.
Nurse answers the questions.
Nurse orders the cultures and then the technician draws the blood cultures.
Nurse selects patient and selects drug(s) based on order.
A series of questions regarding blood cultures are asked.
Nurse administers the antibiotics.
Figure 2. As shown in the sequence of activities for a particular nurse and the place of the intervention, action is initiated with a physician written order for antibiotics for patients with suspected pneumonia. ADC, automated dispensing cabinet.
blood culture collection, and (3) first antibiotic administration from the EHR. The EHR featured explicit fields required by the phlebotomist or nurse to document the occurrence and timing of laboratory and blood culture collection, as well as specific fields for nursing documentation of the timing and route of all medications. There were no data missing from these fields during the study period.
OUTCOME MEASURES We compared the proportion of patients in whom blood cultures were obtained before antibiotic administration in the preintervention (October 2007–March 2008) and postintervention (April 2008–September 2008) periods. The chi-square test was used to test for statistical significance.
Results During the study period, a total of 951 patients with pneumonia were identified—426 in the preintervention period and 525 in the postintervention period. Blood cultures were ordered on 58% (245/426, 95% confidence interval [CI]: 53%–62%) of patients in the preintervention period and on 55% (291/525, 95% CI: 51%–60%) of patients in the postintervention period. Compliance with obtaining blood cultures prior to administering antibiotics was 84% (205/245, 95% CI: 79%–88%) in the preintervention period and 95% (275/291, 95% CI: 92%–97%) in the postintervention period (p < .001).
Discussion In this population of patients admitted with pneumonia, a series of questions in an ADC significantly improved compliance of blood cultures prior to antibiotics. In contrast, previous investigations of computer reminders have reported mixed results, with 226
May 2012
a review of 28 studies concluding that point-of-care computer reminders generally achieve small to modest improvements in provider behavior; no specific reminders or contextual features described were significantly associated with effect magnitude.10 To the best of our knowledge, as confirmed by discussion with the manufacturer of the ADC, this is the first published report of using an ADC base alert for the purpose of driving compliance with a CMS quality measure. Our ED had already exhibited a high baseline level of compliance with the obtaining of blood cultures prior to administration of antibiotics. Still, a significant rate of improvement was observed after our intervention. Previous interventions to improve compliance with pneumonia guidelines focused primarily on intensity of education and feedback.11–13 In contrast, what we describe herein is primarily a systems-based intervention that prompts changes in behavior without additional personnel or educational resources. This type of intervention may readily overcome the personnel and organizational barriers that often present themselves when attempts are made to change a complex care process entailing fragmentation of care.12 Since implementing our intervention, we have continued to achieve high success rates. For example, for 2010 and 2011, an average of 99% and 100%, respectively, of cases met the goal of obtaining blood cultures before antibiotics. The roots of the intervention arose from problems identified in the blood culture collection process, which were identified during a series of focused interviews with physicians, nurses, and phlebotomists. In this ED, the vast majority of blood cultures are drawn by phlebotomists and not by the nursing staff. This often led to the involvement of three separate parties, sometimes not communicating with one another, in the execution of the work processes represented by the pertinent quality measure—physi-
Volume 38 Number 5
Copyright 2012 © The Joint Commission
The Joint Commission Journal on Quality and Patient Safety Sample Letter Detailing Noncompliance Blood Culture Process with Automated Dispensing Cabinet (ADC) Use Date: ________________________ Nurse’s name: ______________________ Medical record #: ____________________ Performance/Behavior Identified You did not properly identify if this patient had blood cultures drawn prior to administering antibiotics through the ADC. The process is: –Identify the physician order with the correct patient. –Put patient information into the ADC. –Answer the questions provided. 1. Were blood cultures ordered? YES NO If no, ensure they get ordered. If yes, continue to the next question. 2. Were blood cultures drawn–blank space to enter response type in: YES or NO –Remove the antibiotics from the ADC. Please review the above process and let me know if you have any questions regarding it. Thank you, Outcomes Coordinator
Figure 3. This sample letter details a nurse’s noncompliance with the timing of blood cultures and incorrect responses to the automated dispensing cabinet alert.
cians ordering blood cultures, nurses administering antibiotics, and phlebotomists drawing blood cultures. The lack of coordination between these groups was believed to be at the core of the episodes of noncompliance, and the intervention was designed to create a hard stop in the patient care process to force communication and coordination among all members of the clinical team. The effort needed to implement our system was confined to the cost of programming and testing the alert within the ADC. Because this type of alert had never been attempted before at our institution, the setup work involved several hours of programming and testing time with an engineer and clinical pharmacist. The initial version of the alert included one question that was answered with a simple “yes” or “no,” which could be clicked by the nurse. Within a few days of initial testing, the response was changed to require the nurse to type an answer so that the alert was not just bypassed without thought and effort. To the best of our knowledge, this is the first study to establish the effect of a systems-based change to improve compliance with blood cultures before antibiotics. However, for an analogous quality measure regarding blood cultures prior to antibi-
May 2012
otics in patients with sepsis, the Institute for Healthcare Improvement has recommended placing prompts in locations near antibiotic storage areas to query staff regarding whether blood cultures have been drawn14 (the effect of this recommended action on compliance is not known). Given the success of our intervention, we have subsequently adopted this ADC alert for several additional indications throughout our institution, as follows: 1. Prompts for blood cultures before any antibiotic administration, including sepsis 2. Instructions for dilution of specific medications 3. Instructions for specific infusion rates 4. Prompt for International Normalized Ratio (INR) check prior to warfarin administration 5. Prompt for timing of anticoagulant administration followed epidural placement/removal
Limitations The study entailed a retrospective review of patients with a primary admitting diagnosis of pneumonia. Individuals with a secondary admitting diagnosis of pneumonia could not be captured in this analysis and were subsequently not included. The impact of additional patients with a secondary admitting diagnosis of pneumonia is unknown and difficult to quantify. The analysis did not examine the impact of patient demographic variables, day of week, time of year, or ED overcrowding on the impact of compliance with blood cultures before antibiotics. Although we were unable to specifically measure or identify differences in patient characteristics between the before and after groups, no large-scale changes or events occurred in our facility that resulted in dramatic changes in our patient population, volume, or general severity of illness. It is also worthwhile to recognize that the intervention consisted of components in addition to the ADC alert, including a formal education program and regular feedback on noncompliant cases and subsequent individual coaching. Although it is not possible to tease out the effect of any one component, they all are probably necessary for the successful implementation and maintenance of an ADC alert with an impact on patient care. Still, the core of the intervention remains an alert and series of mandatory questions prior to the dispensing of medication. We did not perform any formal analysis or survey to evaluate nursing opinion on performing the additional steps required in our intervention. Informal questioning of a number of full-time ED nurses suggested a range of opinions, varying from the extreme of no perceived additional impact on nursing time to the
Volume 38 Number 5
Copyright 2012 © The Joint Commission
227
The Joint Commission Journal on Quality and Patient Safety other extreme of an impact that is more noticeable during periods of increased time pressure. Although our ED, like many other large EDs that we are aware of, has a fairly robust nursing turnover rate, no significant changes in the turnover rate were evident during, before, or after our study. However, it is worth noting that additional outcomes data collected during the study period did not reveal any change in compliance with timeliness of antibiotic treatment (within four hours of arrival) for patients with pneumonia. The attempt to improve compliance with a CMS measure likely would have required, as noted, increased communication among the various providers involved in patient care, which may in itself have entailed an increase in time demands. However, we did not attempt to measure the time taken for the discrete steps in the overall process for the administration of antibiotics. Despite the intention of the intervention to ensure the acquisition of blood for cultures rather than to determine whether blood cultures were indicated, the possibility remains that increased attention to the issue of blood cultures may have resulted in a change in ordering pattern.
Summary In this study, a series of questions in an ADC for antibiotics commonly administered for pneumonia was associated with improved compliance with the quality measure to obtain blood cultures prior to administering antibiotics to patients in whom blood cultures are requested. J The work described in this article was presented at the Society of Academic Emergency Medicine annual meeting, New Orleans, May 2009.
Online-Only Content
8
See the online version of this article for
Appendix 1. Sample Automated Dispensing Cabinet (ADC) Compliance Report
References 1. Centers for Medicare & Medicaid Services, QualityNet. Specifications Manual, Version 4.1: Discharges 07/01/2012 to 12/31/2012: Section 2.3—Pneumonia (PN), Pneumonia National Hospital Inpatient Quality Measures. Accessed Mar 15, 2012. http://qualitynet.org/dcs/ContentServer?c=Page&pagename= QnetPublic%2FPage%2FQnetTier4&cid=1228771525863. 2. Walls RM, Resnick J. The Joint Commission on Accreditation of Healthcare Organizations and Center for Medicare and Medicaid Services community-acquired pneumonia initiative: What went wrong? Ann Emerg Med. 2005;46(5):409–411. 3. Kennedy M, et al. Do emergency department blood cultures change practice in patients with pneumonia? Ann Emerg Med. 2005;46(5):393–400. 4. Ramanujam P, Rathlev NK. Blood cultures do not change management in hospitalized patients with community-acquired pneumonia. Acad Emerg Med. 2006;13(7):740–745. 5. Campbell SG, et al. The contribution of blood cultures to the clinical management of adult patients admitted to the hospital with community-acquired pneumonia: A prospective observational study. Chest. 2003;123(4):1142–1150. 6. The Joint Commission. Pneumonia Measures. Nov 2, 2011. Accessed Mar 15, 2012. http://www.jointcommission.org/pneumonia/. 7. US Department of Health & Human Services, Agency for Healthcare Research and Quality, National Quality Measures Clearinghouse. Pneumonia: Percent of patients whose initial emergency room blood culture specimen was collected prior to first hospital dose of antibiotics. Accessed Mar 15, 2012. http://qualitymeasures.ahrq.gov/content.aspx?id=27403. 8. Mandell LA, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis. 2007 Mar 1;44 Suppl 2:S27–72. 9. Metersky ML, et al. Predicting bacteremia in patients with community-acquired pneumonia. Am J Respir Crit Care Med. 2004 Feb 1;169(3):342–347. 10. Shojania KG, et al. The effects of on-screen, point of care computer reminders on processes and outcomes of care. Cochrane Database Syst Rev. 2009 Jul 8;(3):CD001096. 11. Chu LA, et al. Improving the quality of care for patients with pneumonia in very small hospitals. Arch Intern Med. 2003 Feb 10;163(3):326–332. 12. Halm EA, et al. Limited impact of a multicenter intervention to improve the quality and efficiency of pneumonia care. Chest. 2004;126(1):100–107. 13. Yealy DM, et al. Effect of increasing the intensity of implementing pneumonia guidelines: A randomized, controlled trial. Ann Intern Med. 2005 Dec 20;143(12):881–894. 14. Institute for Healthcare Improvement. Blood Cultures Obtained Prior to Antibiotic Administration. Accessed Mar 15, 2012. http://www.ihi.org /knowledge/Pages/Changes/BloodCulturesObtainedPriortoAntibiotic Administration.aspx.
Rishi Sikka, MD, is Vice President of Clinical Transformation, Advocate Health Care, Oak Brook, Illinois; and Clinical Associate Professor, University of Illinois, Chicago. Rolla Sweis, PharmD, MA, BCPS, is Interim Director of Pharmacy, Advocate Christ Medical Center and Hope Childrens Hospital, and Pharmacy Residency Program Director, Oak Lawn, Illinois; and Adjunct Clinical Professor, University of Illinois, Chicago. Carleen Kaucky, RN, BSN, is Clinical Outcomes Coordinator and Performance Improvement Liaison, Advocate Christ Medical Center. Erik Kulstad, MD, MS, is Research Director, Department of Emergency Medicine, Advocate Christ Medical Center; and Clinical Associate Professor, University of Illinois, Chicago. Please address correspondence to Rishi Sikka, [email protected].
228
May 2012
Volume 38 Number 5
Copyright 2012 © The Joint Commission
The Joint Commission Journal on Quality and Patient Safety Online-Only Content
8
Appendix 1. Sample Automated Dispensing Cabinet (ADC) Compliance Report
This sample patient-level compliance report, generated from the ADC, includes the specific responses to the alerts input by the nursing staff.
AP1
May 2012
Volume 38 Number 5
Copyright 2012 © The Joint Commission
Automated Dispensing Cabinet Alert Improves Compliance with Obtaining Blood Cultures Before Antibiotic Administration for Patients Admitted with Pneumonia Rishi Sikka, MD; Rolla Sweis, PharmD, MA, BCPS; Carleen Kaucky, RN, BSN; Erik Kulstad, MD, MS
T
he US Centers for Medicare & Medicaid Services (CMS) has identified patients hospitalized with pneumonia as an important target for quality improvement.1 Despite disagreement regarding the utility of blood cultures for patients admitted with pneumonia,2–5 CMS and The Joint Commission Pneumonia National Hospital Inpatient Quality Measures include PN-3b, “Blood cultures performed in the emergency department prior to initial antibiotic received in hospital.”1,6,7 Blood cultures may be ordered on the basis of clinical discretion on pneumonia patients admitted to the floor or telemetry. It is generally acknowledged that the yield of clinically useful information is greater if the culture is collected before antibiotics are administered.8,9 Failure to check blood cultures prior to antibiotic administration may affect the growth of any bacteria and prevent a culture from becoming positive, further reducing the yield of a test estimated by some to require as many as 100 patients to identify 1 patient infected with an organism resistant to empiric therapy.3 Consequently, the CMS quality goal specifically measures the number of pneumonia patients whose initial ED blood culture specimen was collected prior to first hospital dose of antibiotics.1 Despite the importance that CMS ascribes to blood culture collection prior to antibiotic administration, there is little published information on a process or intervention to improve compliance with the corresponding quality measure. As these measures receive greater scrutiny by both payers and the public, it becomes important to study, establish, and disseminate proven recommendations that may be adapted in a variety of settings to improve the quality of care. This article reports a study undertaken to measure the impact of an automated dispensing cabinet (ADC) alert on improving compliance with blood cultures prior to antibiotics for patients admitted with pneumonia.
224
May 2012
Article-at-a-Glance Background: A Centers for Medicare & Medicaid Services
(CMS) pneumonia quality measures with particular impact on the emergency department (ED) is blood cultures prior to antibiotic administration for patients admitted with pneumonia. A study was conducted to measure the impact of an automated dispensing cabinet (ADC) alert on improving compliance with the quality measure of obtaining blood cultures prior to giving antibiotics for patients admitted with pneumonia and who have blood cultures ordered. Methods: The pre-post study involved ED adult patients with an admitting diagnosis of pneumonia from October 2007 through September 2008. The intervention consisted of a series of questions in the ED medication ADC regarding blood culture orders and antibiotic administration. Patients with an admitting diagnosis of pneumonia were identified through a search of the ED electronic health record (EHR). The proportion of patients in whom blood cultures were obtained prior to antibiotic administration in the pre- (October 2007–March 2008) and postintervention (April 2008–September 2008) periods were compared. The chi-square test was used to test for statistical significance. Results: Some 951 patients with pneumonia were identified during the study period, 426 pre- and 525 postintervention. Compliance with obtaining blood cultures prior to antibiotic administration was 84% (205/245, 95% confidence interval [CI]: 79%–88%) and 95% (275/291, 95% CI: 92%–97%) in the pre and postintervention periods, respectively (p <. 001). Conclusions: In this population of patients with pneumonia, a series of questions in an ADC improved compliance with the quality measure regarding the obtaining of blood cultures prior to administering antibiotics to patients in whom blood cultures are requested.
Volume 38 Number 5
Copyright 2012 © The Joint Commission
The Joint Commission Journal on Quality and Patient Safety Question Flow on Withdrawing Pneumonia Antibiotics from Automated Dispensing Cabinet (ADC)
Yes
(No additional messaging— medication dispensed)
Were blood cultures drawn?
Yes No Were blood cultures ordered?
No
Have the physician order them and make sure the cultures are drawn before you give the antibiotic.
(No additional messaging— medication dispensed)
Figure 1. These questions, which appeared for all antibiotics commonly administered in the emergency department for pneumonia, all required a response by the nurse, and antibiotics were not dispensed until all the questions were answered.
Methods STUDY DESIGN, SETTING, AND POPULATION This pre-post study involved all ED adult patients (age > 18 years) with a primary admitting diagnosis of pneumonia from October 2007 through September 2008. The study setting was a tertiary care, suburban community hospital that has approximately 85,000 ED visits annually and 700 inpatient beds. All patients with a physician-entered primary admitting diagnosis of pneumonia in the electronic health record (EHR) were identified through a search of the EHR. The study was approved by the hospital’s Institutional Review Board.
INTERVENTION The intervention consisted of a series of questions in the ED medication ADC regarding blood culture orders and antibiotic administration. These questions, which appeared for all antibiotics commonly administered in this ED for pneumonia (Figure 1, above), all required a response by the nurse, and antibiotics were not dispensed until all the questions were answered. If the user did not answer all of the questions in the series, medications could not be dispensed, which ensured a compliance of 100% with respect to answering the questions. The sequence of activities for a particular nurse and the place of the intervention are detailed in Figure 2 (page 226). The cascade of action is initiated with a physician written order for antibiotics for a patient with suspected pneumonia. The nurse takes the order to the ADC, selects the patient, and orders the antibiotics. If the antibiotics ordered include medications typically given for pneumonia in our ED, the nurse must answer the se-
May 2012
ries of questions in the ADC. If cultures are ordered and the nurse answers appropriately, the medications are dispensed and can be administered to the patient. However, if cultures are not ordered, the nurse notifies and confirms the intent with phlebotomist and physician and orders the cultures as needed. The rollout for the intervention consisted of education of ED nurses and phlebotomists at regularly scheduled staff meetings. The education, which addressed the purpose of the alerts, the steps involved, and the intervention’s importance, was supplemented by in-person demonstration of the alerts at the ADC by the ED pharmacists. Compliance Reports. On a regular basis, compliance reports were generated from the ADC (Appendix 1, available in online article). These patient-level reports included the specific responses to the alerts input by the nursing staff. The ADC patient reports were cross-referenced retrospectively by the ED outcomes nurse against the list of patients with blood cultures ordered for pneumonia. In instances in which there was noncompliance with the timing of blood cultures and incorrect responses to the ADC alert, the outcomes nurse placed a letter in the nurse’s mailbox detailing the noncompliance (Figure 3, page 227). This was often followed up by in-person, repeat education of the ADC alert by the ED pharmacist [R.S.]; the session included a review of the purpose of the alert and the steps in the process.
DATA COLLECTION AND PROCESSING A trained registered nurse (RN) [C.K.] used a closed-ended abstraction form to collect the time of (1) patient arrival, (2)
Volume 38 Number 5
Copyright 2012 © The Joint Commission
225
The Joint Commission Journal on Quality and Patient Safety Nursing Flow and Intervention Context Physician orders antibiotics for pneumonia patient.
If cultures are ordered, the nurse removes the drug and administers the medication(s).
If cultures are not ordered, the nurse notifies the technician/physician.
Nurse takes orders to ADC.
Nurse answers the questions.
Nurse orders the cultures and then the technician draws the blood cultures.
Nurse selects patient and selects drug(s) based on order.
A series of questions regarding blood cultures are asked.
Nurse administers the antibiotics.
Figure 2. As shown in the sequence of activities for a particular nurse and the place of the intervention, action is initiated with a physician written order for antibiotics for patients with suspected pneumonia. ADC, automated dispensing cabinet.
blood culture collection, and (3) first antibiotic administration from the EHR. The EHR featured explicit fields required by the phlebotomist or nurse to document the occurrence and timing of laboratory and blood culture collection, as well as specific fields for nursing documentation of the timing and route of all medications. There were no data missing from these fields during the study period.
OUTCOME MEASURES We compared the proportion of patients in whom blood cultures were obtained before antibiotic administration in the preintervention (October 2007–March 2008) and postintervention (April 2008–September 2008) periods. The chi-square test was used to test for statistical significance.
Results During the study period, a total of 951 patients with pneumonia were identified—426 in the preintervention period and 525 in the postintervention period. Blood cultures were ordered on 58% (245/426, 95% confidence interval [CI]: 53%–62%) of patients in the preintervention period and on 55% (291/525, 95% CI: 51%–60%) of patients in the postintervention period. Compliance with obtaining blood cultures prior to administering antibiotics was 84% (205/245, 95% CI: 79%–88%) in the preintervention period and 95% (275/291, 95% CI: 92%–97%) in the postintervention period (p < .001).
Discussion In this population of patients admitted with pneumonia, a series of questions in an ADC significantly improved compliance of blood cultures prior to antibiotics. In contrast, previous investigations of computer reminders have reported mixed results, with 226
May 2012
a review of 28 studies concluding that point-of-care computer reminders generally achieve small to modest improvements in provider behavior; no specific reminders or contextual features described were significantly associated with effect magnitude.10 To the best of our knowledge, as confirmed by discussion with the manufacturer of the ADC, this is the first published report of using an ADC base alert for the purpose of driving compliance with a CMS quality measure. Our ED had already exhibited a high baseline level of compliance with the obtaining of blood cultures prior to administration of antibiotics. Still, a significant rate of improvement was observed after our intervention. Previous interventions to improve compliance with pneumonia guidelines focused primarily on intensity of education and feedback.11–13 In contrast, what we describe herein is primarily a systems-based intervention that prompts changes in behavior without additional personnel or educational resources. This type of intervention may readily overcome the personnel and organizational barriers that often present themselves when attempts are made to change a complex care process entailing fragmentation of care.12 Since implementing our intervention, we have continued to achieve high success rates. For example, for 2010 and 2011, an average of 99% and 100%, respectively, of cases met the goal of obtaining blood cultures before antibiotics. The roots of the intervention arose from problems identified in the blood culture collection process, which were identified during a series of focused interviews with physicians, nurses, and phlebotomists. In this ED, the vast majority of blood cultures are drawn by phlebotomists and not by the nursing staff. This often led to the involvement of three separate parties, sometimes not communicating with one another, in the execution of the work processes represented by the pertinent quality measure—physi-
Volume 38 Number 5
Copyright 2012 © The Joint Commission
The Joint Commission Journal on Quality and Patient Safety Sample Letter Detailing Noncompliance Blood Culture Process with Automated Dispensing Cabinet (ADC) Use Date: ________________________ Nurse’s name: ______________________ Medical record #: ____________________ Performance/Behavior Identified You did not properly identify if this patient had blood cultures drawn prior to administering antibiotics through the ADC. The process is: –Identify the physician order with the correct patient. –Put patient information into the ADC. –Answer the questions provided. 1. Were blood cultures ordered? YES NO If no, ensure they get ordered. If yes, continue to the next question. 2. Were blood cultures drawn–blank space to enter response type in: YES or NO –Remove the antibiotics from the ADC. Please review the above process and let me know if you have any questions regarding it. Thank you, Outcomes Coordinator
Figure 3. This sample letter details a nurse’s noncompliance with the timing of blood cultures and incorrect responses to the automated dispensing cabinet alert.
cians ordering blood cultures, nurses administering antibiotics, and phlebotomists drawing blood cultures. The lack of coordination between these groups was believed to be at the core of the episodes of noncompliance, and the intervention was designed to create a hard stop in the patient care process to force communication and coordination among all members of the clinical team. The effort needed to implement our system was confined to the cost of programming and testing the alert within the ADC. Because this type of alert had never been attempted before at our institution, the setup work involved several hours of programming and testing time with an engineer and clinical pharmacist. The initial version of the alert included one question that was answered with a simple “yes” or “no,” which could be clicked by the nurse. Within a few days of initial testing, the response was changed to require the nurse to type an answer so that the alert was not just bypassed without thought and effort. To the best of our knowledge, this is the first study to establish the effect of a systems-based change to improve compliance with blood cultures before antibiotics. However, for an analogous quality measure regarding blood cultures prior to antibi-
May 2012
otics in patients with sepsis, the Institute for Healthcare Improvement has recommended placing prompts in locations near antibiotic storage areas to query staff regarding whether blood cultures have been drawn14 (the effect of this recommended action on compliance is not known). Given the success of our intervention, we have subsequently adopted this ADC alert for several additional indications throughout our institution, as follows: 1. Prompts for blood cultures before any antibiotic administration, including sepsis 2. Instructions for dilution of specific medications 3. Instructions for specific infusion rates 4. Prompt for International Normalized Ratio (INR) check prior to warfarin administration 5. Prompt for timing of anticoagulant administration followed epidural placement/removal
Limitations The study entailed a retrospective review of patients with a primary admitting diagnosis of pneumonia. Individuals with a secondary admitting diagnosis of pneumonia could not be captured in this analysis and were subsequently not included. The impact of additional patients with a secondary admitting diagnosis of pneumonia is unknown and difficult to quantify. The analysis did not examine the impact of patient demographic variables, day of week, time of year, or ED overcrowding on the impact of compliance with blood cultures before antibiotics. Although we were unable to specifically measure or identify differences in patient characteristics between the before and after groups, no large-scale changes or events occurred in our facility that resulted in dramatic changes in our patient population, volume, or general severity of illness. It is also worthwhile to recognize that the intervention consisted of components in addition to the ADC alert, including a formal education program and regular feedback on noncompliant cases and subsequent individual coaching. Although it is not possible to tease out the effect of any one component, they all are probably necessary for the successful implementation and maintenance of an ADC alert with an impact on patient care. Still, the core of the intervention remains an alert and series of mandatory questions prior to the dispensing of medication. We did not perform any formal analysis or survey to evaluate nursing opinion on performing the additional steps required in our intervention. Informal questioning of a number of full-time ED nurses suggested a range of opinions, varying from the extreme of no perceived additional impact on nursing time to the
Volume 38 Number 5
Copyright 2012 © The Joint Commission
227
The Joint Commission Journal on Quality and Patient Safety other extreme of an impact that is more noticeable during periods of increased time pressure. Although our ED, like many other large EDs that we are aware of, has a fairly robust nursing turnover rate, no significant changes in the turnover rate were evident during, before, or after our study. However, it is worth noting that additional outcomes data collected during the study period did not reveal any change in compliance with timeliness of antibiotic treatment (within four hours of arrival) for patients with pneumonia. The attempt to improve compliance with a CMS measure likely would have required, as noted, increased communication among the various providers involved in patient care, which may in itself have entailed an increase in time demands. However, we did not attempt to measure the time taken for the discrete steps in the overall process for the administration of antibiotics. Despite the intention of the intervention to ensure the acquisition of blood for cultures rather than to determine whether blood cultures were indicated, the possibility remains that increased attention to the issue of blood cultures may have resulted in a change in ordering pattern.
Summary In this study, a series of questions in an ADC for antibiotics commonly administered for pneumonia was associated with improved compliance with the quality measure to obtain blood cultures prior to administering antibiotics to patients in whom blood cultures are requested. J The work described in this article was presented at the Society of Academic Emergency Medicine annual meeting, New Orleans, May 2009.
Online-Only Content
8
See the online version of this article for
Appendix 1. Sample Automated Dispensing Cabinet (ADC) Compliance Report
References 1. Centers for Medicare & Medicaid Services, QualityNet. Specifications Manual, Version 4.1: Discharges 07/01/2012 to 12/31/2012: Section 2.3—Pneumonia (PN), Pneumonia National Hospital Inpatient Quality Measures. Accessed Mar 15, 2012. http://qualitynet.org/dcs/ContentServer?c=Page&pagename= QnetPublic%2FPage%2FQnetTier4&cid=1228771525863. 2. Walls RM, Resnick J. The Joint Commission on Accreditation of Healthcare Organizations and Center for Medicare and Medicaid Services community-acquired pneumonia initiative: What went wrong? Ann Emerg Med. 2005;46(5):409–411. 3. Kennedy M, et al. Do emergency department blood cultures change practice in patients with pneumonia? Ann Emerg Med. 2005;46(5):393–400. 4. Ramanujam P, Rathlev NK. Blood cultures do not change management in hospitalized patients with community-acquired pneumonia. Acad Emerg Med. 2006;13(7):740–745. 5. Campbell SG, et al. The contribution of blood cultures to the clinical management of adult patients admitted to the hospital with community-acquired pneumonia: A prospective observational study. Chest. 2003;123(4):1142–1150. 6. The Joint Commission. Pneumonia Measures. Nov 2, 2011. Accessed Mar 15, 2012. http://www.jointcommission.org/pneumonia/. 7. US Department of Health & Human Services, Agency for Healthcare Research and Quality, National Quality Measures Clearinghouse. Pneumonia: Percent of patients whose initial emergency room blood culture specimen was collected prior to first hospital dose of antibiotics. Accessed Mar 15, 2012. http://qualitymeasures.ahrq.gov/content.aspx?id=27403. 8. Mandell LA, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis. 2007 Mar 1;44 Suppl 2:S27–72. 9. Metersky ML, et al. Predicting bacteremia in patients with community-acquired pneumonia. Am J Respir Crit Care Med. 2004 Feb 1;169(3):342–347. 10. Shojania KG, et al. The effects of on-screen, point of care computer reminders on processes and outcomes of care. Cochrane Database Syst Rev. 2009 Jul 8;(3):CD001096. 11. Chu LA, et al. Improving the quality of care for patients with pneumonia in very small hospitals. Arch Intern Med. 2003 Feb 10;163(3):326–332. 12. Halm EA, et al. Limited impact of a multicenter intervention to improve the quality and efficiency of pneumonia care. Chest. 2004;126(1):100–107. 13. Yealy DM, et al. Effect of increasing the intensity of implementing pneumonia guidelines: A randomized, controlled trial. Ann Intern Med. 2005 Dec 20;143(12):881–894. 14. Institute for Healthcare Improvement. Blood Cultures Obtained Prior to Antibiotic Administration. Accessed Mar 15, 2012. http://www.ihi.org /knowledge/Pages/Changes/BloodCulturesObtainedPriortoAntibiotic Administration.aspx.
Rishi Sikka, MD, is Vice President of Clinical Transformation, Advocate Health Care, Oak Brook, Illinois; and Clinical Associate Professor, University of Illinois, Chicago. Rolla Sweis, PharmD, MA, BCPS, is Interim Director of Pharmacy, Advocate Christ Medical Center and Hope Childrens Hospital, and Pharmacy Residency Program Director, Oak Lawn, Illinois; and Adjunct Clinical Professor, University of Illinois, Chicago. Carleen Kaucky, RN, BSN, is Clinical Outcomes Coordinator and Performance Improvement Liaison, Advocate Christ Medical Center. Erik Kulstad, MD, MS, is Research Director, Department of Emergency Medicine, Advocate Christ Medical Center; and Clinical Associate Professor, University of Illinois, Chicago. Please address correspondence to Rishi Sikka, [email protected].
228
May 2012
Volume 38 Number 5
Copyright 2012 © The Joint Commission
The Joint Commission Journal on Quality and Patient Safety Online-Only Content
8
Appendix 1. Sample Automated Dispensing Cabinet (ADC) Compliance Report
This sample patient-level compliance report, generated from the ADC, includes the specific responses to the alerts input by the nursing staff.
AP1
May 2012
Volume 38 Number 5
Copyright 2012 © The Joint Commission