Louise I. Schneider

A method and knowledge base for automated inference of patient problems from structured data in an electronic medical record

BACKGROUND Accurate knowledge of a patients medical problems is critical for clinical decision making, quality measurement, research, billing and clinical decision support. Common structured sources of problem information include the patient problem list and billing data; however, these sources are often inaccurate or incomplete. OBJECTIVE To develop and validate methods of automatically inferring patient problems from clinical and billing data, and to provide a knowledge base for inferring problems. STUDY DESIGN AND METHODS We identified 17 target conditions and designed and validated a set of rules for identifying patient problems based on medications, laboratory results, billing codes, and vital signs. A panel of physicians provided input on a preliminary set of rules. Based on this input, we tested candidate rules on a sample of 100,000 patient records to assess their performance compared to gold standard manual chart review. The physician panel selected a final rule for each condition, which was validated on an independent sample of 100,000 records to assess its accuracy. RESULTS Seventeen rules were developed for inferring patient problems. Analysis using a validation set of 100,000 randomly selected patients showed high sensitivity (range: 62.8-100.0%) and positive predictive value (range: 79.8-99.6%) for most rules. Overall, the inference rules performed better than using either the problem list or billing data alone. CONCLUSION We developed and validated a set of rules for inferring patient problems. These rules have a variety of applications, including clinical decision support, care improvement, augmentation of the problem list, and identification of patients for research cohorts.

An electronic health record-based intervention to improve tobacco treatment in primary care: a cluster-randomized controlled trial.

BACKGROUND To improve the documentation and treatment of tobacco use in primary care, we developed and implemented a 3-part electronic health record enhancement: (1)smoking status icons, (2) tobacco treatment reminders, and (3) a Tobacco Smart Form that facilitated the ordering of medication and fax and e-mail counseling referrals. METHODS We performed a cluster-randomized controlled trial of the enhancement in 26 primary care practices between December 19, 2006, and September 30, 2007. The primary outcome was the proportion of documented smokers who made contact with a smoking cessation counselor. Secondary outcomes included coded smoking status documentation and medication prescribing. RESULTS During the 9-month study period, 132 630 patients made 315 962 visits to study practices. Coded documentation of smoking status increased from 37% of patients to 54% (+17%) in intervention practices and from 35% of patients to 46% (+11%) in control practices (P < .001 for the difference in differences). Among the 9589 patients who were documented smokers at the start of the study, more patients in the intervention practices were recorded as nonsmokers by the end of the study (5.3% vs 1.9% in control practices; P < .001). Among 12 207 documented smokers, more patients in the intervention practices made contact with a cessation counselor (3.9% vs 0.3% in control practices; P < .001). Smokers in the intervention practices were no more likely to be prescribed smoking cessation medication (2% vs 2% in control practices; P = .40). CONCLUSION This electronic health record-based intervention improved smoking status documentation and increased counseling assistance to smokers but not the prescription of cessation medication.

Improving completeness of electronic problem lists through clinical decision support: a randomized, controlled trial

Background Accurate clinical problem lists are critical for patient care, clinical decision support, population reporting, quality improvement, and research. However, problem lists are often incomplete or out of date. Objective To determine whether a clinical alerting system, which uses inference rules to notify providers of undocumented problems, improves problem list documentation. Study Design and Methods Inference rules for 17 conditions were constructed and an electronic health record-based intervention was evaluated to improve problem documentation. A cluster randomized trial was conducted of 11 participating clinics affiliated with a large academic medical center, totaling 28 primary care clinical areas, with 14 receiving the intervention and 14 as controls. The intervention was a clinical alert directed to the provider that suggested adding a problem to the electronic problem list based on inference rules. The primary outcome measure was acceptance of the alert. The number of study problems added in each arm as a pre-specified secondary outcome was also assessed. Data were collected during 6-month pre-intervention (11/2009–5/2010) and intervention (5/2010–11/2010) periods. Results 17 043 alerts were presented, of which 41.1% were accepted. In the intervention arm, providers documented significantly more study problems (adjusted OR=3.4, p<0.001), with an absolute difference of 6277 additional problems. In the intervention group, 70.4% of all study problems were added via the problem list alerts. Significant increases in problem notation were observed for 13 of 17 conditions. Conclusion Problem inference alerts significantly increase notation of important patient problems in primary care, which in turn has the potential to facilitate quality improvement. Trial Registration ClinicalTrials.gov: NCT01105923.

Variation in Use of Head Computed Tomography by Emergency Physicians

OBJECTIVE Variation in emergency department head computed tomography (CT) use in patients with atraumatic headaches between hospitals is being measured nationwide. However, the magnitude of interphysician variation within a hospital is currently unknown. We hypothesized that there was significant variation in the rates of physician head CT use, both overall and for patients diagnosed with atraumatic headaches. METHODS This cross-sectional study was conducted in the emergency department of a large urban academic hospital, and institutional review board approval was obtained. All emergency department visits from 2009 were analyzed, and the primary outcome measure was whether or not head CT was performed. Logistic regression was used to control for patient, physician, and visit characteristics potentially associated with head CT ordering. The degree of interphysician variability was tested, both before and after controlling for these variables. RESULTS Of 55,286 emergency department patient encounters, 4919 (8.9%) involved head CT examinations. Unadjusted head CT ordering rates per physician ranged from 4.4% to 16.9% overall and from 15.2% to 61.7% in patients diagnosed with atraumatic headaches, with both rates varying significantly between physicians. Two-fold variation in head CT ordering overall (6.5%-13.5%) and approximately 3-fold variation in head CT ordering for atraumatic headaches (21.2%-60.1%) persisted even after controlling for pertinent variables. CONCLUSION Emergency physicians vary significantly in their use of head CT both overall and in patients with atraumatic headaches. Further studies are needed to identify strategies to reduce interphysician variation in head CT use.

Impact of Provider-led, Technology-enabled Radiology Management Program on Imaging

OBJECTIVE The study objective was to assess the impact of a provider-led, technology-enabled radiology medical management program on high-cost imaging use. METHODS This study was performed in the ambulatory setting of an integrated healthcare system. After negotiating a risk contract with a major commercial payer, we created a physician-led radiology medical management program to help address potentially inappropriate high-cost imaging use. The radiology medical management program was enabled by a computerized physician order entry system with integrated clinical decision support and accountability tools, including (1) mandatory peer-to-peer consultation with radiologists before order completion when test utility was uncertain on the basis of order requisition; (2) quarterly practice pattern variation reports to providers; and (3) academic detailing for targeted outliers. The primary outcome measure was intensity of high-cost imaging, defined as the number of outpatient computed tomography (CT), magnetic resonance imaging (MRI), and nuclear cardiology studies per 1000 patient-months in the payers panel. Chi-square test was used to assess trends. RESULTS In 1.8 million patient-months from January 2004 to December 2009, 50,336 eligible studies were performed (54.1% CT, 40.3% MRI, 5.6% nuclear cardiology). There was a 12.0% sustained reduction in high-cost imaging intensity over the 5-year period (P < .001). The number of CT studies performed decreased from 17.5 per 1000 patient-months to 14.5 (P < .01); nuclear cardiology examinations decreased from 2.4 to 1.4 (P < .01) per 1000 patient-months. The MRI rate remained unchanged at 11 studies per 1000 patient-months. CONCLUSION A provider-led radiology medical management program enabled through health information technology and accountability tools may produce a significant reduction in high-cost imaging use.

Implementation of a CKD Checklist for Primary Care Providers

BACKGROUND AND OBJECTIVES CKD is associated with significant morbidity, mortality, and financial burden. Practice guidelines outlining CKD management exist, but there is limited application of these guidelines. Interventions to improve CKD guideline adherence have been limited. This study evaluated a new CKD checklist (a tool outlining management guidelines for CKD) to determine whether implementation in an academic primary care clinic improved adherence to guidelines. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS During a 1-year period (August 2012-August 2013), a prospective study was conducted among 13 primary care providers (PCPs), four of whom were assigned to use a CKD checklist incorporated into the electronic medical record during visits with patients with CKD stages 1-4. All providers received education regarding CKD guidelines. The intervention and control groups consisted of 105 and 263 patients, respectively. Adherence to CKD management guidelines was measured. RESULTS A random-effects logistic regression analysis was performed to account for intra-group correlation by PCP assignment and adjusted for age and CKD stage. CKD care improved among patients whose PCPs were assigned to the checklist intervention compared with controls. Patients in the CKD checklist group were more likely than controls to have appropriate annual laboratory testing for albuminuria (odds ratio [OR], 7.9; 95% confidence interval [95% CI], 3.6 to 17.2), phosphate (OR, 3.5; 95% CI, 1.5 to 8.3), and parathyroid hormone (OR, 8.1; 95% CI, 4.8 to 13.7) (P<0.001 in all cases). Patients in the CKD checklist group had higher rates of achieving a hemoglobin A1c target<7% (OR, 2.7; 95% CI, 1.4 to 5.1), use of an angiotensin-converting enzyme inhibitor or angiotensin-receptor blocker (OR, 2.1; 95% CI, 1.0 to 4.2), documentation of avoidance of nonsteroidal anti-inflammatory drugs (OR, 41.7; 95% CI, 17.8 to 100.0), and vaccination for annual influenza (OR, 2.1; 95% CI, 1.1 to 4.0) and pneumococcus (OR, 4.7; 95% CI, 2.6 to 8.6) (P<0.001 in all cases). CONCLUSIONS Implementation of a CKD checklist significantly improved adherence to CKD management guidelines and delivery of CKD care.

JOURNAL CLUB: Requiring Clinical Justification to Override Repeat Imaging Decision Support: Impact on CT Use

OBJECTIVE The purpose of this study was to determine the impact of requiring clinical justification to override decision support alerts on repeat use of CT. SUBJECTS AND METHODS This before and after intervention study was conducted at a 793-bed tertiary hospital with computerized physician order entry and clinical decision support systems. When a CT order is placed, decision support alerts the orderer if the patients same body part has undergone CT within the past 90 days. The study cohort included all 28,420 CT orders triggering a repeat alert in 2010. The intervention required clinical justification, selected from a predetermined menu, to override repeat CT decision support alerts to place a CT order; otherwise the order could not be placed and was dropped. The primary outcome, dropped repeat CT orders, was analyzed using three methods: chi-square tests to compare proportions dropped before and after intervention; multiple logistic regression tests to control for orderer, care setting, and patient factors; and statistical process control for temporal trends. RESULTS The repeat CT order drop rate had an absolute increase of 1.4%; 6.1% (682/11,230) before to 7.5% (1290/17,190) after intervention, which was a 23% relative change (7.5 - 6.1)/6.1 × 100 = 23%; p < 0.0001). Orders were dropped more often after intervention (odds ratio, 1.3; 95% CI, 1.1-1.4; p < 0.0001). Statistical control analysis supported the association between the increase in the drop rate with intervention rather than underlying trends. CONCLUSION Adding a requirement for clinical justification to override alerts modestly but significantly improves the impact of repeat CT decision support (23% relative change), with the overall effect of preventing one in 13 repeat CT orders.

Use of a Web-based Risk Appraisal Tool for Assessing Family History and Lifestyle Factors in Primary Care

ABSTRACTBACKGROUNDPrimary care clinicians can play an important role in identifying individuals at increased risk of cancer, but often do not obtain detailed information on family history or lifestyle factors from their patients.OBJECTIVEWe evaluated the feasibility and effectiveness of using a web-based risk appraisal tool in the primary care setting.DESIGNFive primary care practices within an academic care network were assigned to the intervention or control group.PARTICIPANTSWe included 15,495 patients who had a new patient visit or annual exam during an 8-month period in 2010–2011.INTERVENTIONIntervention patients were asked to complete a web-based risk appraisal tool on a laptop computer immediately before their visit. Information on family history of cancer was sent to their electronic health record (EHR) for clinicians to view; if accepted, it populated coded fields and could trigger clinician reminders about colon and breast cancer screening.MAIN MEASURESThe main outcome measure was new documentation of a positive family history of cancer in coded EHR fields. Secondary outcomes included clinician reminders about screening and discussion of family history, lifestyle factors, and screening.KEY RESULTSAmong eligible intervention patients, 2.0 % had new information on family history of cancer entered in the EHR within 30 days after the visit, compared to 0.6 % of eligible control patients (adjusted odds ratio = 4.3, p = 0.03). There were no significant differences in the percent of patients who received moderate or high risk reminders for colon or breast cancer screening.CONCLUSIONSUse of this tool was associated with increased documentation of family history of cancer in the EHR, although the percentage of patients with new family history information was low in both groups. Further research is needed to determine how risk appraisal tools can be integrated with workflow and how they affect screening and health behaviors.

Clinician characteristics and use of novel electronic health record functionality in primary care

BACKGROUND Conventional wisdom holds that older, busier clinicians who see complex patients are less likely to adopt and use novel electronic health record (EHR) functionality. METHODS To compare the characteristics of clinicians who did and did not use novel EHR functionality, we conducted a retrospective analysis of the intervention arm of a randomized trial of new EHR-based tobacco treatment functionality. RESULTS The novel functionality was used by 103 of 207 (50%) clinicians. Staff physicians were more likely than trainees to use the functionality (64% vs 37%; p<0.001). Clinicians who graduated more than 10 years previously were more likely to use the functionality than those who graduated less than 10 years previously (64% vs 42%; p<0.01). Clinicians with higher patient volumes were more likely to use the functionality (lowest quartile of number of patient visits, 25%; 2nd quartile, 38%; 3rd quartile, 65%; highest quartile, 71%; p<0.001). Clinicians who saw patients with more documented problems were more likely to use the functionality (lowest tertile of documented patient problems, 38%; 2nd tertile, 58%; highest tertile, 54%; p=0.04). In multivariable modeling, independent predictors of use were the number of patient visits (OR 1.2 per 100 additional patients; 95% CI 1.1 to 1.4) and number of documented problems (OR 2.9 per average additional problem; 95% CI 1.4 to 6.1). CONCLUSIONS Contrary to conventional wisdom, clinically busier physicians seeing patients with more documented problems were more likely to use novel EHR functionality.

Assessing Strength of Evidence of Appropriate Use Criteria for Diagnostic Imaging Examinations

OBJECTIVE For health information technology tools to fully inform evidence-based decisions, recommendations must be reliably assessed for quality and strength of evidence. We aimed to create an annotation framework for grading recommendations regarding appropriate use of diagnostic imaging examinations. METHODS The annotation framework was created by an expert panel (clinicians in three medical specialties, medical librarians, and biomedical scientists) who developed a process for achieving consensus in assessing recommendations, and evaluated by measuring agreement in grading the strength of evidence for 120 empirically selected recommendations using the Oxford Levels of Evidence. RESULTS Eighty-two percent of recommendations were assigned to Level 5 (expert opinion). Inter-annotator agreement was 0.70 on initial grading (κ = 0.35, 95% CI, 0.23-0.48). After systematic discussion utilizing the annotation framework, agreement increased significantly to 0.97 (κ = 0.88, 95% CI, 0.77-0.99). CONCLUSIONS A novel annotation framework was effective for grading the strength of evidence supporting appropriate use criteria for diagnostic imaging exams.