TY - JOUR
AU1 - Niemi,, Kristin
AU2 - Geary,, Siobhan
AU3 - Quinn,, Barbara
AU4 - Larrabee,, Mark
AU5 - Brown,, Kevin
AB - Abstract Purpose. The ability of an automated system used to identify pneumonia and heart failure (HF) patients in real time and prompt clinical interventions was evaluated. Summary. An automated system evaluated all adult patients with an emergency department (ED) visit or hospital admission for possible pneumonia in real time for 31 days. Two quality-indicator alerts were sent to appropriate clinicians for possible interventions. The system evaluated all hospitalized adult patients for HF for 30 days. A list of possible HF patients printed every 12 hours and was used for possible interventions. Pneumonia and HF identification accuracy was assessed by comparison with discharge diagnosis. Compliance with quality indicators was assessed using three pneumonia and three HF indicators. The effect was measured by comparing compliance data and the number of monthly quality indicators in the top decile seven months before and after implementation. There were 3053 ED visits and 986 inpatient admissions during the pneumonia study. The system sensitivity for pneumonia ED identification was 89% and the specificity was 86%. The sensitivity for pneumonia admissions was 92% and the specificity was 90%. There were 1037 inpatient admissions during the HF study. The sensitivity for HF identification was 94% and the specificity was 90%. Sixty-seven percent of the six indicators studied increased the percentage of months in the top decile after implementation. The average increase was 26%. Conclusion. An automated system effectively identified pneumonia and HF patients in real time. The system prompted interventions, which helped increase compliance with national quality indicators. Automation, Compliance, Computers, Decision making, Heart failure, Pneumonia, Protocols The introduction of pay-for-performance based on compliance with quality indicators has prompted studies investigating the correlation between compliance and outcomes. 1,–4 To date, the results of these studies have been inconclusive, and the clinical validity of some quality indicators has been questioned. 5,–7 While the debate continues, most health care organizations have already implemented systems to assist in compliance with the quality indicators for the Centers for Medicare and Medicaid Services (CMS) in order to maximize reimbursement. 4 Concurrent with the development of pay-for-performance is the development of electronic clinical decision support (CDS). One CDS focus has been medication safety, incorporating a patient’s age, weight, and renal function as part of CDS rules. 8 CDS dosing rules applied to a specific patient identify medication doses or dosage intervals that are outside of a defined safe range. 9 CDS is also used to identify and decrease the incidence of adverse drug events. This use of CDS has been promoted by the Agency for Healthcare Research and Quality. 10 CDS has also been used to assess quality through retrospective analysis of patients with the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) core measure diagnoses by comparing electronic medical record review, manual patient chart review, and hybrid (automated followed by manual) review. 11,–13 Examples of hospitals using CDS to assist in meeting quality indicators in real time have been limited to single electronic triggers. 14 CMS has included heart failure (HF) and pneumonia in the current core measures and has defined specific quality indicators for each disease. 15 Despite the existence of best practice guidelines and financial incentives to encourage compliance with CMS quality indicators, many patients do not receive optimal treatment. 1,16 The December 2007 national averages for hospital compliance with left-ventricular-function assessment, angiotensin-converting-enzyme (ACE) inhibitor or angiotensin-II-receptor blocker (ARB) use for left-ventricular-systolic dysfunction (LVSD), and provision of patient discharge instructions ranged from 65% to 85%. 17 The national averages for hospital compliance with indicators related to pneumococcal vaccination status, receipt of initial antibiotic by appropriate patients within four hours of hospital arrival, and appropriate initial antibiotic selection ranged from 73% to 85%. 17 Real-time identification of patients for whom the guidelines apply is one of the barriers to accomplishing full compliance with best practices. 11 Traditional strategies, such as education and the use of standardized order sets that change prescribing behavior, require significant implementation time and effort. Real-time manual chart review is a labor-intensive process. 10 An automated CDS tool capable of identifying patients with possible diagnoses, rapidly responding to best practice changes, and efficiently directing patient-specific review has the potential to increase efficiency and improve quality of care. At Sutter Medical Center Sacramento, we created an automated CDS system with two goals: (1) to identify core measure patients in real time, and (2) to provide alerts to the appropriate clinician with sufficient time to allow for intervention when performance measures were not being met. The system is called the core measure manager (CMM). The purpose of this study was to evaluate the ability of the CMM to identify pneumonia and HF patients in real time, to prompt clinical interventions, and to improve compliance with selected national quality indicators. Development of the CMM The development of the CMM included developing rules to identify possible core measure patients, identifying the information systems in which documentation of core measure compliance is electronically found at our facility, creating alerts when time-sensitive core measure criteria were not being met, and determining what format is most useful to the intervening clinician. Our tertiary care facility does not yet have an electronic medical record. Instead, there are multiple information systems including a pharmacy information system (PharmNet, Cerner Corporation, Kansas City, MO); automated dispensing machines (ADMs) (Pyxis, Cardinal Health, Dublin, OH); a patient admission, transfer, discharge, and billing system (Med Series 4, Siemens, Berlin, Germany); a radiology order system (Misys-Rad, Misys Healthcare Systems, Raleigh, NC); a laboratory management system (Misys-Lab, Misys Healthcare Systems); and a heart catheterization laboratory hemodynamic monitoring system (MACLab, GE Healthcare, Waukesha, WI). Health Level Seven feeds (i.e., a streamlined way to deliver messages) were required to link data from the existing systems into a CMM server. The rules and alerts were executed on the data in the CMM server (Figure 1). Rules were developed to define text to serve as triggers for identification of possible pneumonia patients (Table 1) and HF patients (Table 2) based on point systems. Threshold point values indicating that a patient had a possible diagnosis of pneumonia or HF were established. Rules and point values were determined by consensus among appropriate clinicians based on clinical experience and chart reviews. Rules and point values were further refined based on patient cases during the CMM development stage. In order to minimize the number of false-positive alerts for possible pneumonia patients, a higher point value was chosen to trigger an alert at the 20-hour time point compared with the point value chosen to trigger an alert at 3 hours after admission. The HF threshold was not time dependent and therefore could be triggered at any time during admission. Throughout the development of the CMM, clinicians who used the alerts or lists were consulted to determine how and where alerts should be sent (Table 3). Operation of the CMM The CMM continually monitors data from each independent information system. All transactions are stored on the CMM server, and the rules for pneumonia and HF run continuously. When a patient accumulates enough points to indicate that he or she may have pneumonia, the patient’s therapy is evaluated based on ADM transactions or pharmacy-information-system orders. If no transactions for CMS-recommended antibiotics are detected, the CMM prints an alert to the patient’s nursing unit. When a patient accumulates enough points to indicate that he or she may have HF, the patient’s name appears on an HF update report printed every 12 hours. The CMM automatically prints alerts or update lists to the appropriate clinician at the designated time and location. Three specific alerts were developed for pneumonia. The first alert is the 3-hour alert and occurs when a patient admitted to the emergency department (ED) accumulates 6 or more points and is therefore identified as a possible pneumonia patient. If this occurs 3 to 4 hours after patient arrival at the facility, the CMM evaluates ADM transactions and the pharmacy information system to determine whether an antibiotic removal or antibiotic order has occurred. If no removal is detected, a 3-hour alert prints in the ED at 3 hours after admission. This alert gives the clinician 1 hour to confirm or rule out pneumonia as a possible diagnosis and treat appropriately. The alert is suppressed if an antibiotic has been removed or ordered. The next alert, the 20-hour alert, occurs when an admitted patient accumulates 12 points or more. The CMM then compares the antibiotic orders in the pharmacy information system to the CMS-recommended antibiotics. 8 If the antibiotics are not consistent with CMS recommendations, a 20-hour alert prints in the main pharmacy. The alert is suppressed if appropriate antibiotics have been prescribed. The pharmacist has 4 hours to contact the treating physician to discuss the diagnosis and initiate intervention if appropriate. Upon admission, the CMM sends the last alert to the printer on the patient’s nursing unit, indicating when the pneumococcal vaccine was previously dispensed during any prior admission. This assists in documenting vaccination history if the patient is unsure. Every 12 hours, the CMM prints a list of patients who have accumulated 20 points or more within the previous 12 hours based on the HF identification table. The list includes each patient’s location and demographic information, chief complaint, diagnosis, and the details of the patient core measure point values and point total. Only patients reaching 20 points for the first time print on the update list to avoid redundancy. Clinicians use this list to review patients who may need further discussion with the physician in order to assure appropriate indicators have been met and documented. Evaluation of the CMM Study population Pneumonia Patients 18 years and older with an ED visit, hospital admission, or both between October 1, 2006, and October 31, 2006, were included. Patients with orders for comfort care, those who developed pneumonia more than 24 hours after admission, and patients with a white blood cell (WBC) count of ≤4 × 1000/mm3 were excluded. HF Patients 18 years and older admitted between February 11, 2007, and March 12, 2007, were included. No patients were excluded. The institutional review board approved this study. Study design The study design had three segments: the CMM identification of pneumonia patients, the CMM identification of HF patients, and the impact on compliance with selected CMS quality indicators after CMM implementation. The pneumonia identification portion had two postadmission time-sensitive points—3 and 20 hours after admission. HF identification by the CMM could occur at any time during the patient admission. Pneumonia identification All patients identified by the CMM with a possible diagnosis of pneumonia (a score of 6 points or greater on the pneumonia patient-identification table) 3 hours after hospital arrival, 20 hours after hospital arrival (a score of 12 points or greater), or at both time points were recorded. The number of alerts printed was also recorded. The 3-hour alert focused on patients admitted to the ED. Not all patients admitted to the ED were necessarily admitted as inpatients. Therefore, the 3-hour time point included all visits to the ED, while the 20-hour time point included all in-patients. The list of patients who the CMM identified at each time point was compared to all patients with a primary ICD-9-CM discharge diagnosis of pneumonia during the study period. The charts of all patients with an ICD-9-CM diagnosis of pneumonia who were not identified by the CMM were reviewed. Patients with an ICD-9-CM discharge diagnosis of pneumonia and orders for comfort care who developed pneumonia more than 24 hours after hospital admission or who had a WBC count of ≤4 × 1000/mm3 were excluded from the study. The CMM alerts were processed by appropriate staff, but specific responses to the alerts were not recorded. HF identification All patients identified by the CMM with possible HF at any time during the hospitalization were recorded. The CMM update list was used by appropriate staff, but specific responses were not used for this study. When discharge coding was completed, patients with an ICD-9-CM discharge diagnosis of HF were compared to patients identified by the CMM. The charts of all patients with an ICD-9-CM discharge diagnosis of HF who were not identified by the CMM were reviewed. Impact of compliance with the quality indicators The overall effectiveness of the CMM on compliance with initial antibiotic administration within four hours of admission, correct antibiotic selection, pneumococcal vaccination status, left-ventricular-function assessment, ACE or ARB for LVSD, and discharge instructions was measured in two ways. Compliance with these indicators seven months pre- and post-CMM implementation was compared. In addition, the number of months these indicators were in the top national decile during the seven months before and after CMM implementation was compared using the hospital compare database. Statistical analysis Sensitivity, specificity, and overall accuracy were computed based on the CMM’s ability to identify pneumonia patients 3 and 20 hours after admission and HF patients at any time during admission compared to ICD-9-CM discharge diagnosis. A test of independent proportions was used to compare the compliance with quality indicators seven months before implementation of the pneumonia and HF sections of the CMM with our performance during the seven months after implementation. The CMM’s effect on compliance was measured by comparing the number of months our hospital achieved top decile performance for seven months preimplementation and postimplementation. Results Pneumonia alerts at 3 hours There were 3053 patients seen in the ED during the study period. Forty-seven were identified at discharge as having pneumonia by ICD-9-CM diagnosis codes. Of the 47 patients, 37 met inclusion criteria for evaluation. Ten patients were excluded—2 patients had orders for comfort care, 5 patients with pneumonia were diagnosed more than 24 hours after admission, and 3 patients had a WBC count of ≤4 × 1000/mm3. The CMM identified 33 of 37 of these patients and correctly identified 2570 of the 3016 patients without pneumonia. The CMM incorrectly identified 436 patients as possibly having pneumonia when they did not. The sensitivity, specificity, and overall accuracy of the CMM in detecting patients with pneumonia at 3 hours in the ED were 0.89 (95% confidence interval [CI], 0.79–0.99), 0.86 (95% CI, 0.84–0.87), and 0.86 (95% CI, 0.84–0.87), respectively. Pneumonia alerts at 20 hours There were 986 inpatient admissions during the study period. A total of 37 patients met inclusion criteria. The CMM identified 34 of the 37 pneumonia patients and correctly identified 846 of the 939 patients without pneumonia. The CMM incorrectly identified 93 patients as having pneumonia when they did not. The sensitivity, specificity, and overall accuracy of the CMM in detecting patients with pneumonia at 20 hours were 0.92 (95% CI, 0.83–1.00), 0.90 (95% CI, 0.88–0.92), and 0.90 (95% CI, 0.88–0.92), respectively. The CMM failed to identify 2 patients with an ICD-9-CM discharge diagnosis of pneumonia at both 3 and 20 hours. One patient had a normal WBC count and initial diagnosis of chronic obstructive pulmonary disease. The other patient was a direct admission from another hospital with an initial cardiac or neurologic differential diagnosis. Two additional patients were not detected at 3 hours. One of those patients was a direct admission from a physician’s office, and the other patient was involved with an equipment malfunction. An additional patient who was not detected at 20 hours had a complex differential diagnosis. HF Of the 1037 patients included in the study population, 94 had HF as documented by ICD-9-CM discharge diagnosis. The CMM correctly identified 88 of these and correctly identified 849 of the 943 patients without HF. The CMM incorrectly identified 10% (94 of 943 patients) as having HF when they did not. The sensitivity, specificity, and overall accuracy of the CMM in detecting patients with HF were 0.94 (95% CI, 0.89–0.99), 0.90 (95% CI, 0.88–0.92), and 0.90 (95% CI, 0.89–0.92), respectively. The 6 patients who the CMM failed to identify did not have enough points to indicate HF during the study period hospitalization. HF was mentioned in each of these patient’s electronic history and physical examination dictation. These electronic documents were not accessible to the CMM server. Quality-indicator compliance Compliance with the indicators for antibiotic administration within four hours (83% versus 87%) and pneumonia vaccination status documentation (82% versus 92%) both increased during the period following implementation of the CMM, while compliance with indicators for appropriate pneumonia antibiotic selection slightly decreased after implementation (93% versus 92%); however, none of these changes in compliance were statistically significant. After the CMM implementation, compliance with left-ventricular-function assessment (98% versus 100%) and ACE or ARB initiation (95% versus 98%) increased slightly, but the differences were not statistically significant. There was a statistically significant increase in compliance with the provision of discharge instructions (84% versus 95%, p < 0.01). For 67% (4 out of 6) of the indicators, the number of months that our institution was in the top decile for compliance increased after implementation. The number of indicator months in the top decile increased by 26% (11 of 42 months). A comparison of the number of months that our institution was in the top decile before and after implementation of the CMM using hospital compare data is provided in Figure 2. The largest improvement in compliance was seen in the provision of HF discharge instructions. No improvement was seen in the indicators for administration within 4 hours of admission or administration of correct antibiotics within 24 hours of admission. The number of months in the top decile was not decreased for any indicator after implementation of the CMM. Discussion Our use of CDS to identify possible patient diagnoses by creating rules applied to discrete electronic data fields was based on several observations and assumptions. One observation was that this support service is needed because the presenting complaint of the patient may not state the underlying diagnosis. For example, while developing the rules for pneumonia, we found that patients, especially in the elderly population, can present to the ED with chief complaints nonindicative of pneumonia, such as falls or weakness. We observed that hospital staff would be more receptive to innovation that did not require any additional input outside of normal workflow. Therefore, the CMM operates automatically. One assumption was that high sensitivity for CMM identification is very important because the program operates as a safety net for patients who do not present with a clear diagnosis. After using the CMM output, clinicians stated that if a patient is identified by the CMM, even if the diagnosis is not correct, the patient benefits from further clinical review, so high specificity is not essential. Another assumption was the importance of eliminating alerts if the indicator was met in order to avoid alert fatigue. While developing the point-value alert thresholds, input from staff concerning alert number and frequency was monitored and changes were made to keep the sensitivity high and the specificity acceptable to the clinicians using them. Further adjustment and evaluation of specific point values are necessary to determine if the number of false positives can be reduced further without loss of sensitivity. We also assumed that ICD-9-CM coding was the standard for diagnosis. It is possible that additional retrospective manual chart review could provide alternative diagnoses. Two of the patient exclusion criteria from the pneumonia study were supported by applying the exclusion criteria from the CMS reporting rules for quality-indicator compliance. Since the completion of this study, the pneumonia identification table has been modified to include points for patients with a WBC count of ≤4 × 1000/mm3. This change should address the three patients who were not identified by pneumonia alerts during the study. The fact that the change in our internal compliance response either stayed the same or increased slightly reflects the difficulty and importance of increasing compliance and that missing indicators on individual patients is important. The fact that there was a 26% increase in the number of indicator months in the top decile demonstrates that small increases in performance can have a large impact when compared nationally. One limitation of this study was that we did not document specific responses to each alert. Multiple factors influenced compliance with indicators during the study period, including staff education, preprinted orders, and physician awareness and physician disagreement with CMS guidelines. Each of these factors could have resulted in intervention or lack of intervention. Anecdotal responses from pharmacy staff, nursing staff, and physicians that clearly indicated an intervention as a direct result of CMM output have provided positive support for CMM alerts. Further study should assist in output modification. The lack of access to the electronic history and physical examination information and the lack of electronic recording of vital signs at our facility are limitations of CMM design. Future addition of this information could increase both the sensitivity and specificity of the results. Another limitation of this study is the use of hospital compare data to measure effectiveness. The national data also represent multiple factors. Our results do not show a direct cause and effect relationship between CMM implementation and improved performance. This study was designed to demonstrate how an automated system such as the CMM can assist in the effort to improve performance. Our assumption was that every hospital is addressing compliance using multiple approaches; therefore, the significant increase in postimplementation top decile performance does represent the impact of the CMM. Although this study was limited to documenting the ability of the CMM to identify pneumonia and HF patients, other uses for the CMM have been implemented and others are planned. The basic format is easily adapted to meet additional quality goals. The CMM has been adapted to acute myocardial infarction patients at our institution, and identification of adverse drug events by the CMM is under development. Previously published reports show inconclusive results correlating compliance with quality indicators and improved outcomes. 1,2 This may be indicative and supportive of the basic premise by the Institute of Medicine (IOM) that health care is staffed by good people trying to do good work. 17 The focus on financial incentives or disincentives may not be what is needed. Instead, as also supported by IOM, system innovation and redesign can improve care. 18 The integration of CDS into health care systems and the acceptance of computer guidance by clinicians are under investigation. 19,20 Determining the most effective and inclusive approach will require further research. CDS systems that are capable of streamlining the excessive amount of knowledge and data available to clinicians to allow for effective patient-specific decisions are essential tools for improving patient outcomes. Conclusion An automated system effectively identified pneumonia and HF patients in real time. The system prompted interventions, which helped increase compliance with national quality indicators. Table 1. Clinical Decision-Support Triggers for Identification of Patients with Possible Pneumonia Trigger Location Trigger Points Assigneda aAn alert for possible pneumonia is triggered at 3 hours for patients who are admitted to the emergency department and accumulate =6 points; an alert is triggered at 20 hours for patients who have been admitted to the hospital and accumulate =12 points. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cCBCM = complete blood count with manual differential, CBCA = complete blood count with automated differential, CBCND = complete blood count with no differential, BC = blood culture. dThe core measure manager (CMM) was programmed to identify any antibiotic (other than azithromycin) included in section 8:12–8:13.99 of the American Hospital Formulary Service (AHFS) classification system. eThe CMM was programmed to assign points based on key phrases typically found in chest radiology reports. Reports were considered positive, likely, possible, or negative based on review of key phrases. Age (yr) on admission, transfer, discharge, and billing system >60 1 >75 2 <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: pneumonia, pnuemonia, pnemmonia, pna 20 Text including any of the following: shortness of breath or SOB, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, respiratory, cough, weakness, weak, chest pain, back pain, lung pain, body pain, pain all over, hurts all over, nausea, vomiting, vomit, vommit, vomitt, cold, flu, congestion, wheezing, phlegm, conscious 1 for each trigger Text including any of the following: decreased oxygen saturation, fever, elevated temp, increased temp, temperature 2 for each trigger Text including any of the following: motor vehicle accident or MVA, trauma, fx, fracture 0, overriding all other points assigned Laboratory test orders on laboratory management system CBCM, CBCA, CBCND, BCc 1 for each trigger Respiratory sputum, Res 4 Laboratory test results on laboratory management system White blood cell count of >11 × 1000/mm3 3 Positive blood culture (filtered) 2 Medication ordered on pharmacy information system Azithromycin 2 Any other antibioticd 1 for each trigger Medication removal from automated dispensing machine Azithromycin 2 Any other antibioticd 1 for each trigger Radiology order on radiology order system Chest exam 2 Radiology report for chest examination on radiology order systeme Positive 20 Likely 3 Possible 2 Negative or no idea 0 Trigger Location Trigger Points Assigneda aAn alert for possible pneumonia is triggered at 3 hours for patients who are admitted to the emergency department and accumulate =6 points; an alert is triggered at 20 hours for patients who have been admitted to the hospital and accumulate =12 points. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cCBCM = complete blood count with manual differential, CBCA = complete blood count with automated differential, CBCND = complete blood count with no differential, BC = blood culture. dThe core measure manager (CMM) was programmed to identify any antibiotic (other than azithromycin) included in section 8:12–8:13.99 of the American Hospital Formulary Service (AHFS) classification system. eThe CMM was programmed to assign points based on key phrases typically found in chest radiology reports. Reports were considered positive, likely, possible, or negative based on review of key phrases. Age (yr) on admission, transfer, discharge, and billing system >60 1 >75 2 <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: pneumonia, pnuemonia, pnemmonia, pna 20 Text including any of the following: shortness of breath or SOB, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, respiratory, cough, weakness, weak, chest pain, back pain, lung pain, body pain, pain all over, hurts all over, nausea, vomiting, vomit, vommit, vomitt, cold, flu, congestion, wheezing, phlegm, conscious 1 for each trigger Text including any of the following: decreased oxygen saturation, fever, elevated temp, increased temp, temperature 2 for each trigger Text including any of the following: motor vehicle accident or MVA, trauma, fx, fracture 0, overriding all other points assigned Laboratory test orders on laboratory management system CBCM, CBCA, CBCND, BCc 1 for each trigger Respiratory sputum, Res 4 Laboratory test results on laboratory management system White blood cell count of >11 × 1000/mm3 3 Positive blood culture (filtered) 2 Medication ordered on pharmacy information system Azithromycin 2 Any other antibioticd 1 for each trigger Medication removal from automated dispensing machine Azithromycin 2 Any other antibioticd 1 for each trigger Radiology order on radiology order system Chest exam 2 Radiology report for chest examination on radiology order systeme Positive 20 Likely 3 Possible 2 Negative or no idea 0 Table 1. Clinical Decision-Support Triggers for Identification of Patients with Possible Pneumonia Trigger Location Trigger Points Assigneda aAn alert for possible pneumonia is triggered at 3 hours for patients who are admitted to the emergency department and accumulate =6 points; an alert is triggered at 20 hours for patients who have been admitted to the hospital and accumulate =12 points. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cCBCM = complete blood count with manual differential, CBCA = complete blood count with automated differential, CBCND = complete blood count with no differential, BC = blood culture. dThe core measure manager (CMM) was programmed to identify any antibiotic (other than azithromycin) included in section 8:12–8:13.99 of the American Hospital Formulary Service (AHFS) classification system. eThe CMM was programmed to assign points based on key phrases typically found in chest radiology reports. Reports were considered positive, likely, possible, or negative based on review of key phrases. Age (yr) on admission, transfer, discharge, and billing system >60 1 >75 2 <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: pneumonia, pnuemonia, pnemmonia, pna 20 Text including any of the following: shortness of breath or SOB, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, respiratory, cough, weakness, weak, chest pain, back pain, lung pain, body pain, pain all over, hurts all over, nausea, vomiting, vomit, vommit, vomitt, cold, flu, congestion, wheezing, phlegm, conscious 1 for each trigger Text including any of the following: decreased oxygen saturation, fever, elevated temp, increased temp, temperature 2 for each trigger Text including any of the following: motor vehicle accident or MVA, trauma, fx, fracture 0, overriding all other points assigned Laboratory test orders on laboratory management system CBCM, CBCA, CBCND, BCc 1 for each trigger Respiratory sputum, Res 4 Laboratory test results on laboratory management system White blood cell count of >11 × 1000/mm3 3 Positive blood culture (filtered) 2 Medication ordered on pharmacy information system Azithromycin 2 Any other antibioticd 1 for each trigger Medication removal from automated dispensing machine Azithromycin 2 Any other antibioticd 1 for each trigger Radiology order on radiology order system Chest exam 2 Radiology report for chest examination on radiology order systeme Positive 20 Likely 3 Possible 2 Negative or no idea 0 Trigger Location Trigger Points Assigneda aAn alert for possible pneumonia is triggered at 3 hours for patients who are admitted to the emergency department and accumulate =6 points; an alert is triggered at 20 hours for patients who have been admitted to the hospital and accumulate =12 points. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cCBCM = complete blood count with manual differential, CBCA = complete blood count with automated differential, CBCND = complete blood count with no differential, BC = blood culture. dThe core measure manager (CMM) was programmed to identify any antibiotic (other than azithromycin) included in section 8:12–8:13.99 of the American Hospital Formulary Service (AHFS) classification system. eThe CMM was programmed to assign points based on key phrases typically found in chest radiology reports. Reports were considered positive, likely, possible, or negative based on review of key phrases. Age (yr) on admission, transfer, discharge, and billing system >60 1 >75 2 <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: pneumonia, pnuemonia, pnemmonia, pna 20 Text including any of the following: shortness of breath or SOB, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, respiratory, cough, weakness, weak, chest pain, back pain, lung pain, body pain, pain all over, hurts all over, nausea, vomiting, vomit, vommit, vomitt, cold, flu, congestion, wheezing, phlegm, conscious 1 for each trigger Text including any of the following: decreased oxygen saturation, fever, elevated temp, increased temp, temperature 2 for each trigger Text including any of the following: motor vehicle accident or MVA, trauma, fx, fracture 0, overriding all other points assigned Laboratory test orders on laboratory management system CBCM, CBCA, CBCND, BCc 1 for each trigger Respiratory sputum, Res 4 Laboratory test results on laboratory management system White blood cell count of >11 × 1000/mm3 3 Positive blood culture (filtered) 2 Medication ordered on pharmacy information system Azithromycin 2 Any other antibioticd 1 for each trigger Medication removal from automated dispensing machine Azithromycin 2 Any other antibioticd 1 for each trigger Radiology order on radiology order system Chest exam 2 Radiology report for chest examination on radiology order systeme Positive 20 Likely 3 Possible 2 Negative or no idea 0 Table 2. Clinical Decision-Support Triggers for Identification of Patients with Heart Failure Trigger Location Trigger Points Assigneda aEvery 12 hours, the core measure manager prints a list of patients who have accumulated 20 or more points within the previous 12 hours. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cACE = angiotensin-converting-enzyme, ARB = angiotensin-receptor blocker. Age (yr) on admission, transfer, discharge, and billing system <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: heartfailure, heart failure, congestive heart, CHF (congestive heart failure), pulmonary edema 20 Text including any of the following: shortness of breath, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, difficult breathing, short of breath or SOB, respiratory, dyspnea 4 Text including any of the following: fatigue, leg swelling, extremity swelling 2 Combination of age on admission, transfer, discharge, and billing system and N-terminal prohormone brain natriuretic peptide (pro-BNP) test results on laboratory data management system Age of <50 yr and pro-BNP concentration of >450 pg/mL, age of 50–75 yr and pro-BNP of >900 pg/mL, or age of >75 yr and pro-BNP of >1800 pg/mL 10 Laboratory test results on laboratory management system Serum sodium concentration of <130 meq/L 5 Medication orders on pharmacy information systemc Nesiritide 20 Carvedilol 10 Furosemide (injectable) 7 Spironolactone 5 Torsemide (oral) 3 Any combination of furosemide (injectable), ACE inhibitor, and metolazone; or furosemide (injectable), ARB, and metolazone 20 Any 3-drug combination of ACE inhibitor, diuretic, and β-blocker; ARB, diuretic, and β-blocker; digoxin, diuretic, and ARB; or digoxin, diuretic, and ACE inhibitor 20 Any 4-drug combination of hydralazine, isosorbide, ARB, and diuretic; or hydralazine, isosorbide, ACE inhibitor, and diuretic 20 Any 2-drug combination of diuretic and ACE inhibitor; diuretic and ARB; diuretic andβ-blocker; digoxin and ACE inhibitor; digoxin and ARB; or hydralazine and oral nitrate 10 Medication removal from automated dispensing machine Spironolactone 3 Radiology reports on radiology order system Text including any of the following: biventricular devices, BiV, 3 lead system identified on xray, BiV pacemaker 3 leads, Bi- V, ICD, internal cardiac defibrillator 20 Pulmonary edema 10 Text including any of the following: pacemaker, cardiomegaly, pleural effusion, echocardiogram, ejection fraction or EF < 55% 5 each Trigger Location Trigger Points Assigneda aEvery 12 hours, the core measure manager prints a list of patients who have accumulated 20 or more points within the previous 12 hours. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cACE = angiotensin-converting-enzyme, ARB = angiotensin-receptor blocker. Age (yr) on admission, transfer, discharge, and billing system <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: heartfailure, heart failure, congestive heart, CHF (congestive heart failure), pulmonary edema 20 Text including any of the following: shortness of breath, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, difficult breathing, short of breath or SOB, respiratory, dyspnea 4 Text including any of the following: fatigue, leg swelling, extremity swelling 2 Combination of age on admission, transfer, discharge, and billing system and N-terminal prohormone brain natriuretic peptide (pro-BNP) test results on laboratory data management system Age of <50 yr and pro-BNP concentration of >450 pg/mL, age of 50–75 yr and pro-BNP of >900 pg/mL, or age of >75 yr and pro-BNP of >1800 pg/mL 10 Laboratory test results on laboratory management system Serum sodium concentration of <130 meq/L 5 Medication orders on pharmacy information systemc Nesiritide 20 Carvedilol 10 Furosemide (injectable) 7 Spironolactone 5 Torsemide (oral) 3 Any combination of furosemide (injectable), ACE inhibitor, and metolazone; or furosemide (injectable), ARB, and metolazone 20 Any 3-drug combination of ACE inhibitor, diuretic, and β-blocker; ARB, diuretic, and β-blocker; digoxin, diuretic, and ARB; or digoxin, diuretic, and ACE inhibitor 20 Any 4-drug combination of hydralazine, isosorbide, ARB, and diuretic; or hydralazine, isosorbide, ACE inhibitor, and diuretic 20 Any 2-drug combination of diuretic and ACE inhibitor; diuretic and ARB; diuretic andβ-blocker; digoxin and ACE inhibitor; digoxin and ARB; or hydralazine and oral nitrate 10 Medication removal from automated dispensing machine Spironolactone 3 Radiology reports on radiology order system Text including any of the following: biventricular devices, BiV, 3 lead system identified on xray, BiV pacemaker 3 leads, Bi- V, ICD, internal cardiac defibrillator 20 Pulmonary edema 10 Text including any of the following: pacemaker, cardiomegaly, pleural effusion, echocardiogram, ejection fraction or EF < 55% 5 each Table 2. Clinical Decision-Support Triggers for Identification of Patients with Heart Failure Trigger Location Trigger Points Assigneda aEvery 12 hours, the core measure manager prints a list of patients who have accumulated 20 or more points within the previous 12 hours. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cACE = angiotensin-converting-enzyme, ARB = angiotensin-receptor blocker. Age (yr) on admission, transfer, discharge, and billing system <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: heartfailure, heart failure, congestive heart, CHF (congestive heart failure), pulmonary edema 20 Text including any of the following: shortness of breath, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, difficult breathing, short of breath or SOB, respiratory, dyspnea 4 Text including any of the following: fatigue, leg swelling, extremity swelling 2 Combination of age on admission, transfer, discharge, and billing system and N-terminal prohormone brain natriuretic peptide (pro-BNP) test results on laboratory data management system Age of <50 yr and pro-BNP concentration of >450 pg/mL, age of 50–75 yr and pro-BNP of >900 pg/mL, or age of >75 yr and pro-BNP of >1800 pg/mL 10 Laboratory test results on laboratory management system Serum sodium concentration of <130 meq/L 5 Medication orders on pharmacy information systemc Nesiritide 20 Carvedilol 10 Furosemide (injectable) 7 Spironolactone 5 Torsemide (oral) 3 Any combination of furosemide (injectable), ACE inhibitor, and metolazone; or furosemide (injectable), ARB, and metolazone 20 Any 3-drug combination of ACE inhibitor, diuretic, and β-blocker; ARB, diuretic, and β-blocker; digoxin, diuretic, and ARB; or digoxin, diuretic, and ACE inhibitor 20 Any 4-drug combination of hydralazine, isosorbide, ARB, and diuretic; or hydralazine, isosorbide, ACE inhibitor, and diuretic 20 Any 2-drug combination of diuretic and ACE inhibitor; diuretic and ARB; diuretic andβ-blocker; digoxin and ACE inhibitor; digoxin and ARB; or hydralazine and oral nitrate 10 Medication removal from automated dispensing machine Spironolactone 3 Radiology reports on radiology order system Text including any of the following: biventricular devices, BiV, 3 lead system identified on xray, BiV pacemaker 3 leads, Bi- V, ICD, internal cardiac defibrillator 20 Pulmonary edema 10 Text including any of the following: pacemaker, cardiomegaly, pleural effusion, echocardiogram, ejection fraction or EF < 55% 5 each Trigger Location Trigger Points Assigneda aEvery 12 hours, the core measure manager prints a list of patients who have accumulated 20 or more points within the previous 12 hours. bText for chief complaint is entered manually into system; common spelling errors and abbreviations were included intentionally. cACE = angiotensin-converting-enzyme, ARB = angiotensin-receptor blocker. Age (yr) on admission, transfer, discharge, and billing system <18 0, overriding all other points assigned Chief complaint on admission, transfer, discharge, and billing systemb Text including any of the following: heartfailure, heart failure, congestive heart, CHF (congestive heart failure), pulmonary edema 20 Text including any of the following: shortness of breath, difficulty breathing, can’t breathe, trouble breathing, hard time breathing, difficult breathing, short of breath or SOB, respiratory, dyspnea 4 Text including any of the following: fatigue, leg swelling, extremity swelling 2 Combination of age on admission, transfer, discharge, and billing system and N-terminal prohormone brain natriuretic peptide (pro-BNP) test results on laboratory data management system Age of <50 yr and pro-BNP concentration of >450 pg/mL, age of 50–75 yr and pro-BNP of >900 pg/mL, or age of >75 yr and pro-BNP of >1800 pg/mL 10 Laboratory test results on laboratory management system Serum sodium concentration of <130 meq/L 5 Medication orders on pharmacy information systemc Nesiritide 20 Carvedilol 10 Furosemide (injectable) 7 Spironolactone 5 Torsemide (oral) 3 Any combination of furosemide (injectable), ACE inhibitor, and metolazone; or furosemide (injectable), ARB, and metolazone 20 Any 3-drug combination of ACE inhibitor, diuretic, and β-blocker; ARB, diuretic, and β-blocker; digoxin, diuretic, and ARB; or digoxin, diuretic, and ACE inhibitor 20 Any 4-drug combination of hydralazine, isosorbide, ARB, and diuretic; or hydralazine, isosorbide, ACE inhibitor, and diuretic 20 Any 2-drug combination of diuretic and ACE inhibitor; diuretic and ARB; diuretic andβ-blocker; digoxin and ACE inhibitor; digoxin and ARB; or hydralazine and oral nitrate 10 Medication removal from automated dispensing machine Spironolactone 3 Radiology reports on radiology order system Text including any of the following: biventricular devices, BiV, 3 lead system identified on xray, BiV pacemaker 3 leads, Bi- V, ICD, internal cardiac defibrillator 20 Pulmonary edema 10 Text including any of the following: pacemaker, cardiomegaly, pleural effusion, echocardiogram, ejection fraction or EF < 55% 5 each Table 3. Core Measure Manager (CMM) Output for Pneumonia and Heart Failurea Quality Indicator by Diagnosis Output Name Method Threshold Score Location aICU = intensive care unit, CMS = Centers for Medicare and Medicaid Services. beICU = electronic intensive care unit; the surveillance of ICU patients occurred through remote monitoring and communication with bedside staff. Pneumonia     Antibiotic within 4 hr 3-hr alert Pneumonia patient-identification table ≥6 points Emergency department printer     Antibiotic selection ICU and non-ICU after 24 hr 20-hr alert Pneumonia patient-identification table, CMS antibiotic guidelines, PharmNet active orders ≥12 points Main pharmacy printer     Pneumococcal vaccination Previous Pneumovax dispense alert Pneumonia patient-identification table, Pyxis dispense per CMM database Previous dispense Adult nursing stations, e-mail to eICUb     Nonspecific PNA (pneumonia) update list Pneumonia patient-identification table ≥12 points Quality-assurance clinical nurse, prints twice daily Heart failure     Nonspecific Heart failure update list Heart failure table ≥20 points Quality-assurance clinical nurse, prints twice daily Quality Indicator by Diagnosis Output Name Method Threshold Score Location aICU = intensive care unit, CMS = Centers for Medicare and Medicaid Services. beICU = electronic intensive care unit; the surveillance of ICU patients occurred through remote monitoring and communication with bedside staff. Pneumonia     Antibiotic within 4 hr 3-hr alert Pneumonia patient-identification table ≥6 points Emergency department printer     Antibiotic selection ICU and non-ICU after 24 hr 20-hr alert Pneumonia patient-identification table, CMS antibiotic guidelines, PharmNet active orders ≥12 points Main pharmacy printer     Pneumococcal vaccination Previous Pneumovax dispense alert Pneumonia patient-identification table, Pyxis dispense per CMM database Previous dispense Adult nursing stations, e-mail to eICUb     Nonspecific PNA (pneumonia) update list Pneumonia patient-identification table ≥12 points Quality-assurance clinical nurse, prints twice daily Heart failure     Nonspecific Heart failure update list Heart failure table ≥20 points Quality-assurance clinical nurse, prints twice daily Table 3. Core Measure Manager (CMM) Output for Pneumonia and Heart Failurea Quality Indicator by Diagnosis Output Name Method Threshold Score Location aICU = intensive care unit, CMS = Centers for Medicare and Medicaid Services. beICU = electronic intensive care unit; the surveillance of ICU patients occurred through remote monitoring and communication with bedside staff. Pneumonia     Antibiotic within 4 hr 3-hr alert Pneumonia patient-identification table ≥6 points Emergency department printer     Antibiotic selection ICU and non-ICU after 24 hr 20-hr alert Pneumonia patient-identification table, CMS antibiotic guidelines, PharmNet active orders ≥12 points Main pharmacy printer     Pneumococcal vaccination Previous Pneumovax dispense alert Pneumonia patient-identification table, Pyxis dispense per CMM database Previous dispense Adult nursing stations, e-mail to eICUb     Nonspecific PNA (pneumonia) update list Pneumonia patient-identification table ≥12 points Quality-assurance clinical nurse, prints twice daily Heart failure     Nonspecific Heart failure update list Heart failure table ≥20 points Quality-assurance clinical nurse, prints twice daily Quality Indicator by Diagnosis Output Name Method Threshold Score Location aICU = intensive care unit, CMS = Centers for Medicare and Medicaid Services. beICU = electronic intensive care unit; the surveillance of ICU patients occurred through remote monitoring and communication with bedside staff. Pneumonia     Antibiotic within 4 hr 3-hr alert Pneumonia patient-identification table ≥6 points Emergency department printer     Antibiotic selection ICU and non-ICU after 24 hr 20-hr alert Pneumonia patient-identification table, CMS antibiotic guidelines, PharmNet active orders ≥12 points Main pharmacy printer     Pneumococcal vaccination Previous Pneumovax dispense alert Pneumonia patient-identification table, Pyxis dispense per CMM database Previous dispense Adult nursing stations, e-mail to eICUb     Nonspecific PNA (pneumonia) update list Pneumonia patient-identification table ≥12 points Quality-assurance clinical nurse, prints twice daily Heart failure     Nonspecific Heart failure update list Heart failure table ≥20 points Quality-assurance clinical nurse, prints twice daily Figure 1. Open in new tabDownload slide Design for linkage of patient data from existing information systems to the core measure manager (CMM) server. Figure 1. Open in new tabDownload slide Design for linkage of patient data from existing information systems to the core measure manager (CMM) server. Figure 2. Open in new tabDownload slide Comparison of months in top decile nationally for specific quality indicators before and after implementation of the core measure manager (CMM) system. Percentages indicate compliance for the indicator during the respective study periods. Figure 2. Open in new tabDownload slide Comparison of months in top decile nationally for specific quality indicators before and after implementation of the core measure manager (CMM) system. Percentages indicate compliance for the indicator during the respective study periods. References 1 Fonarow GC, Abraham WT, Albert NM et al. Association between performance measures and clinical outcomes for patients hospitalized with heart failure. JAMA . 2007 ; 297 : 61 –70. Crossref Search ADS PubMed 2 Glickman SW, Ou FS, DeLong ER et al. Pay for performance, quality of care, and outcomes in acute myocardial infarction. JAMA . 2007 ; 297 : 2373 –80. Crossref Search ADS PubMed 3 Peterson LA, LeChauncy DW, Urech BA et al. Does pay-for-performance improve the quality of health care? Ann Intern Med . 2006 ; 145 : 265 –72. Crossref Search ADS PubMed 4 Johnson PE. Introduction. Update on pharmaceutical reimbursement and implications for health-system pharmacists. Am J Health-Syst Pharm. 2008 ; 65 (suppl 1): S2 –3. [Erratum, Am J Health-Syst Pharm. 2008; 65:286.] 5 Timmermans S, Mauck A. The promises and pitfalls of evidence-based medicine. Health Aff . 2005 ; 24 : 18 –28. Crossref Search ADS 6 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 : 409 –11. Crossref Search ADS PubMed 7 Metersky ML, Sweeney TA, Getzow MB et al. Antibiotic timing and diagnostic uncertainty in Medicare patients with pneumonia: is it reasonable to expect all patients to receive antibiotics within 4 hours? Chest . 2006 ; 130 : 16 –21. Crossref Search ADS PubMed 8 American Medical Informatics Association. About AMIA. Strategic initiatives and programs. www.amia.org/inside/initiatives/cds/ (accessed 2008 Nov 21). 9 Reichley RM. Implementing a commercial rule base as a medication order safety net. J Am Med Inform Assoc . 2005 ; 12 : 383 –9. Crossref Search ADS PubMed 10 U.S. Department of Health and Human Services. Agency for Healthcare Research and Quality. Reducing and preventing adverse drug events to decrease hospital costs. www.ahrq.gov/qual/aderia/aderia.htm (accessed 2008 Feb 22). 11 Persell SD, Wright JM, Thompson JA et al. Assessing the validity of national quality measures for coronary artery disease using an electronic health record. Arch Intern Med . 2006 ; 166 : 2272 –7. Crossref Search ADS PubMed 12 Baker DW, Persell SD, Thompson JA et al. Automated review of electronic health records to assess quality of care for outpatients with heart failure. Ann Intern Med . 2007 ; 146 : 270 –7. Crossref Search ADS PubMed 13 Udris EM, Au DH, McDonnell MB et al. Comparing methods to identify general internal medicine clinic patients with chronic heart failure. Am Heart J . 2001 ; 142 : 1003 –9. Crossref Search ADS PubMed 14 Bailey TC, Noirot LA, Blickensderfer A et al. An intervention to improve secondary prevention of coronary heart disease. Arch Intern Med . 2007 ; 167 : 586 –90. Crossref Search ADS PubMed 15 Joint Commission. Performance measurement initiatives. Current specification manual for national hospital quality measures. www.jointcommission.org/performancemeasurement/current+nhqm+manual.htm (accessed 2008 Nov 24). 16 Niederman MS, Mandell LA, Anzueto A et al. Guidelines for the management of adults with community-acquired pneumonia. Diagnosis, assessment of severity, antimicrobial therapy, and prevention. Am J Respir Crit Care Med . 2001 ; 163 : 1730 –54. Crossref Search ADS PubMed 17 Kohn LT, Corrigan JM, Donaldson MS, eds. To err is human: building a safer health system. Institute of Medicine. Washington, DC: National Academy Press; 2000 . 18 Institute of Medicine. Crossing the quality chasm: a new health system for the 21st century. Washington, DC: National Academy Press; 2001 . 19 Weed LL. New connections between medical knowledge and patient care. BMJ . 1997 ; 315 : 231 –5. Crossref Search ADS PubMed 20 Weed LL, Weed L. Opening the black box of clinical judgment—an overview. Interview by Abi Berger. BMJ . 1999 ; 319 : 1279 . Crossref Search ADS PubMed Author notes Carol Parise, John Mesic, and the staff at Sutter Medical Center Sacramento are acknowledged for their support and assistance. The program described in this article was a recipient of an ASHP Best Practices Award in 2007. The authors have declared no potential conflicts of interest. Copyright © 2009, American Society of Health-System Pharmacists, Inc. All rights reserved.
TI - Implementation and evaluation of electronic clinical decision support for compliance with pneumonia and heart failure quality indicators
JF - American Journal of Health-System Pharmacy
DO - 10.2146/ajhp080143
DA - 2009-02-15
UR - https://www.deepdyve.com/lp/oxford-university-press/implementation-and-evaluation-of-electronic-clinical-decision-support-m2bx0WYAqn
SP - 389
VL - 66
IS - 4
DP - DeepDyve
ER -