OBJECTIVE:To identify factors contributing to changes on quality, productivity, and safety outcomes during a large commercial electronic health record (EHR) implementation and to guide future research.METHODS:We conducted a mixed-methods study assessing the impact of a commercial EHR implementation. The method consisted of a quantitative longitudinal evaluation followed by qualitative semi-structured, in-depth interviews with clinical employees from the same implementation. Fourteen interviews were recorded and transcribed. Three authors independently coded interview narratives and via consensus identified factors contributing to changes on 15 outcomes of quality, productivity, and safety.RESULTS:We identified 14 factors that potentially affected the outcomes previously monitored. Our findings demonstrate that several factors related to the implementation (e.g., incomplete data migration), partially related (e.g., intentional decrease in volume of work), and not related (e.g., health insurance changes) may affect outcomes in different ways.DISCUSSION:This is the first study to investigate factors contributing to changes on a broad set of quality, productivity, and safety outcomes during an EHR implementation guided by the results of a large longitudinal evaluation. The diversity of factors identified indicates that the need for organizational adaptation to take full advantage of new technologies is as important for health care as it is for other services sectors.CONCLUSIONS:We recommend continuous identification and monitoring of these factors in future evaluations to hopefully increase our understanding of the full impact of health information technology interventions.
CONTEXT:In existence for nearly 25 years, the Healthcare Systems Research Network (HCSRN) is an established and sustainable network of health care systems that serves as a "real world" laboratory to enable the integration of research findings into practice. The objective of this paper is to demonstrate how the HCSRN serves as an ideal environment for studying dissemination and implementation of evidence-based practices into health care systems through the example of developing a multi-site study on the implementation of evidence-based precision medicine practices. CASE DESCRIPTION:The "Implementing Universal Lynch Syndrome Screening (IMPULSS)" study (NIH R01CA211723) involves seven HCSRN health care systems and two external health care systems. The IMPULSS study will describe and explain organizational variability around Lynch syndrome (LS) screening to identify which factors in different organizational contexts are important for successful implementation of LS screening programs and will create a toolkit to facilitate organizational decision making around implementation and improvement of precision medicine programs in health care systems. MAJOR THEMES:The strengths of the HCSRN that facilitate D&I research include: 1) a culture of collaboration, 2) standardization of data and processes across systems, and 3) researchers embedded in diverse health care systems. We describe how these strengths contributed to developing the IMPULSS study. CONCLUSION:Given the importance of conducting research in real world settings to improve patient outcomes, the unique strengths of the HCSRN are of vital importance. The IMPULSS study is one case example of how the strengths of the HCSRN make it an excellent environment for research on implementing evidence-based precision medicine practices in health care systems.
The Cancer Research Network (CRN) is a consortium of 12 research groups, each affiliated with a nonprofit integrated health care delivery system, that was first funded in 1998. The overall goal of the CRN is to support and facilitate collaborative cancer research within its component delivery systems. This paper describes the CRN's 20-year experience and evolution. The network combined its members' scientific capabilities and data resources to create an infrastructure that has ultimately supported over 275 projects. Insights about the strengths and limitations of electronic health data for research, approaches to optimizing multidisciplinary collaboration, and the role of a health services research infrastructure to complement traditional clinical trials and large observational datasets are described, along with recommendations for other research consortia.
Objective: To assess whether implementation of age-dependent therapeutic targets for high hemoglobin A1c (HbA1c) changed clinicians’ ordering of diabetes medications for older adults.Background: In 2016, Kaiser Permanente Southern California (KPSC) changed the therapeutic targets for alerting clinicians about high HbA1c results in the electronic health record, KP HealthConnect (KPHC). Previously, all HbA1c results ≥7.0 percent were flagged as high in adult patients with diabetes. Starting in 2016, HbA1c therapeutic targets were relaxed to <7.5 percent for patients age 65 to 75, and to <8.0 percent for patients over age 75 to reduce treatment intensity and adverse events.Methods: This retrospective analysis used logistic regression models to calculate the change in odds of a medication change following an HbA1c result after age-dependent HbA1c flags were introduced.Results: The odds of medication change decreased among patients whose HbA1c targets were relaxed: Odds Ratio (OR) 0.72 (95 percent CI 0.67–0.76) for patients age 65–75 and HbA1c 7.0 percent–7.5 percent; OR 0.72 (95 percent CI 0.65–0.80) for patients over age 75 and HbA1c 7.0 percent–7.5 percent; and OR 0.67 (95 percent CI 0.61–0.75) for patients over age 75 and HbA1c 7.5 percent–8.0 percent. In the age and HbA1c ranges for which the alerts did not change, the odds of medication change generally increased or stayed the same. There was little evidence of medication de-intensification in any group.Conclusions: These findings suggest that the change in therapeutic targets was associated with a reduction in medication intensification among older adults with diabetes.
BACKGROUND:Appendicitis is a common surgical emergency in children, yet diagnosis can be challenging. An electronic health record (EHR) based, clinical decision support (CDS) system called Appy CDS was designed to help guide management of pediatric patients with acute abdominal pain within the Health Care Systems Research Network (HCSRN). OBJECTIVES:To describe the development and implementation of a clinical decision support tool (Appy CDS) built independently but synergistically at two large HCSRN affiliated health systems using well-established platforms, and to assess the tool's Triage component, aiming to identify pediatric patients at increased risk for appendicitis. RESULTS:Despite differences by site in design and implementation, such as the use of alerts, incorporating gestalt, and other workflow variations across sites, using simple screening questions and automated exclusions, both systems were able to identify a population with similar appendicitis rates (11.8 percent and 10.6 percent), where use of the full Appy CDS would be indicated. DISCUSSION:These 2 HCSRN sites designed Appy CDS to capture a population at risk for appendicitis and deliver CDS to that population while remaining locally relevant and adhering to organizational preferences. Despite different approaches to point-of-care CDS, the sites have identified similar cohorts with nearly identical background rates of appendicitis. NEXT STEPS:The full Appy CDS tool, providing personalized risk assessment and tailored recommendations, is undergoing evaluation as part of a pragmatic cluster randomized trial aiming to reduce reliance on advanced diagnostic imaging. The novel approaches to CDS we present could serve as the basis for future ED interventions.
BACKGROUND:Reducing hospital-acquired pressure ulcers (PUs) in intensive care units (ICUs) has emerged as an important quality metric for health systems internationally. Limited work has been done to characterize the profile of PUs in the ICU using observational data from the electronic health record (EHR). Consequently, there are limited EHR-based prognostic tools for determining a patient's risk of PU development, with most institutions relying on nurse-calculated risk scores such as the Braden score to identify high-risk patients.METHODS AND RESULTS:Using EHR data from 50,851 admissions in a tertiary ICU (MIMIC-III), we show that the prevalence of PUs at stage 2 or above is 7.8 percent. For the 1,690 admissions where a PU was recorded on day 2 or beyond, we evaluated the prognostic value of the Braden score measured within the first 24 hours. A high-risk Braden score (<=12) had precision 0.09 and recall 0.50 for the future development of a PU. We trained a range of machine learning algorithms using demographic parameters, diagnosis codes, laboratory values and vitals available from the EHR within the first 24 hours. A weighted linear regression model showed precision 0.09 and recall 0.71 for future PU development. Classifier performance was not improved by integrating Braden score elements into the model.CONCLUSION:We demonstrate that an EHR-based model can outperform the Braden score as a screening tool for PUs. This may be a useful tool for automatic risk stratification early in an admission, helping to guide quality protocols in the ICU, including the allocation and timing of prophylactic interventions.
Introduction: The symptom burden faced by long-term head and neck cancer survivors is not well understood. In addition, the accuracy of clinical data sources for symptom ascertainment is not clear.Objective: To 1) describe the prevalence of symptoms in 5-year survivors of head and neck cancer, and 2) to evaluate agreement between symptoms obtained via self-report and symptoms obtained from clinical data sources.Methods: We recruited 5-year survivors of head and neck cancer enrolled at Kaiser Permanente Washington (n = 54). Symptoms were assessed using the MD Anderson Symptom Inventory head and neck cancer module. For each symptom, we assessed the agreement of the patient’s survey response (“gold standard”) with the 1) medical chart and 2) administrative health care claims data. We computed the sensitivity, specificity, positive predictive value (PPV), and negative predictive value, along with their 95 percent confidence intervals, for each clinical data source.Results: Eighty percent of patients responded. Nearly all participants (95 percent) reported experiencing at least one symptom from the MDASI-HN, and 93 percent reported two or more symptoms. Among patients reporting a given symptom, there was generally no evidence of the symptom from either clinical data source (i.e., sensitivity was generally no greater than 40 percent). The specificity and PPV of the clinical data sources were generally higher than the sensitivity.Conclusion: Relying only on medical chart review and/or administrative health data would substantially underestimate symptom burden in long-term head and neck cancer survivors.
BACKGROUND:Sharing of research data derived from health system records supports the rigor and reproducibility of primary research and can accelerate research progress through secondary use. But public sharing of such data can create risk of re-identifying individuals, exposing sensitive health information.METHOD:We describe a framework for assessing re-identification risk that includes: identifying data elements in a research dataset that overlap with external data sources, identifying small classes of records defined by unique combinations of those data elements, and considering the pattern of population overlap between the research dataset and an external source. We also describe alternative strategies for mitigating risk when the external data source can or cannot be directly examined.RESULTS:We illustrate this framework using the example of a large database used to develop and validate models predicting suicidal behavior after an outpatient visit. We identify elements in the research dataset that might create risk and propose a specific risk mitigation strategy: deleting indicators for health system (a proxy for state of residence) and visit year.DISCUSSION:Researchers holding health system data must balance the public health value of data sharing against the duty to protect the privacy of health system members. Specific steps can provide a useful estimate of re-identification risk and point to effective risk mitigation strategies.
Examples of fully integrated machine learning models that drive clinical care are rare. Despite major advances in the development of methodologies that outperform clinical experts and growing prominence of machine learning in mainstream medical literature, major challenges remain. At Duke Health, we are in our fourth year developing, piloting, and implementing machine learning technologies in clinical care. To advance the translation of machine learning into clinical care, health system leaders must address barriers to progress and make strategic investments necessary to bring health care into a new digital age. Machine learning can improve clinical workflows in subtle ways that are distinct from how statistics has shaped medicine. However, most machine learning research occurs in siloes, and there are important, unresolved questions about how to retrain and validate models post-deployment. Academic medical centers that cultivate and value transdisciplinary collaboration are ideally suited to integrate machine learning in clinical care. Along with fostering collaborative environments, health system leaders must invest in developing new capabilities within the workforce and technology infrastructure beyond standard electronic health records. Now is the opportunity to break down barriers and achieve scalable growth in the number of high-impact collaborations between clinical researchers and machine learning experts to transform clinical care.
Introduction: The use of electronic health records (EHR) data in research to inform recruitment and outcomes is considered a critical element for pragmatic studies. However, there is a lack of research on the availability of substance use disorder (SUD) treatment data in the EHR to inform research.Methods: This study recruited providers who used an EHR for patient care and whose facilities were affiliated with the National Institute on Drug Abuse’s National Drug Abuse Treatment Clinical Trials Network (NIDA CTN). Data about providers’ use of an EHR and other methods to support and document clinical tasks for Substance use screening, Brief Intervention, and Referral to Treatment (SBIRT) were collected.Results: Participants (n = 26) were from facilities across the country (South 46.2%, West 23.1%, Midwest 19.2 percent, Northeast 11.5 percent), representing 26 different health systems/facilities at various settings: primary care (30.8 percent), ambulatory other/specialty (26.9 percent), mixed setting (11.5 percent), hospital outpatient (11.5 percent), emergency department (7.7 percent), inpatient (3.8 percent), and other (7.7 percent). Validated tools were rarely used for substance use screen and SUD assessment. Structured and unstructured EHR fields were commonly used to document SBIRT. The following tasks had high proportions of using unstructured EHR fields: substance use screen, treatment exploration, brief intervention, referral, and follow-up.Conclusion: This study is the first of its kind to investigate the documentation of SBIRT in the EHR outside of unique settings (e.g., Veterans Health Administration). While results are descriptive, they emphasize the importance of developing EHR features to collect structured data for SBIRT to improve health care quality evaluation and SUD research.
Background: Despite the importance of characterizing colonoscopy indication for quality monitoring and cancer screening program evaluation, there is no standard approach to documenting colonoscopy indication in medical records.Methods: We applied two algorithms in three health care systems to assign colonoscopy indication to persons 50–89 years old who received a colonoscopy during 2010–2013. Both algorithms used standard procedure, diagnostic, and laboratory codes. One algorithm, the KPNC algorithm, used a hierarchical approach to classify exam indication into: diagnostic, surveillance, or screening; whereas the other, the SEARCH algorithm, used a logistic regression-based algorithm to provide the probability that colonoscopy was performed for screening. Gold standard assessment of indication was from medical records abstraction.Results: There were 1,796 colonoscopy exams included in analyses; age and racial/ethnic distributions of participants differed across health care systems. The KPNC algorithm’s sensitivities and specificities for screening indication ranged from 0.78–0.82 and 0.78–0.91, respectively; sensitivities and specificities for diagnostic indication ranged from 0.78–0.89 and 0.74–0.82, respectively. The KPNC algorithm had poor sensitivities (ranging from 0.11–0.67) and high specificities for surveillance exams. The Area Under the Curve (AUC) of the SEARCH algorithm for screening indication ranged from 0.76–0.84 across health care systems. For screening indication, the KPNC algorithm obtained higher specificities than the SEARCH algorithm at the same sensitivity.Conclusion: Despite standardized implementation of these indication algorithms across three health care systems, the capture of colonoscopy indication data was imperfect. Thus, we recommend that standard, systematic documentation of colonoscopy indication should be added to medical records to ensure efficient and accurate data capture.
Background: Stable angina patients have difficulty understanding the tradeoffs between treatment alternatives. In this analysis, we assessed treatment planning conversations for stable angina to determine whether inadequate health literacy acts as a barrier to communication that may partially explain this difficulty.Methods: We conducted a descriptive analysis of patient questionnaire data from the PCI Choice Trial. The main outcomes were the responses to the Decisional Conflict Scale and the proportion of correct responses to knowledge questions about stable angina. We also conducted a qualitative analysis on recordings of patient-clinician discussions about treatment planning. The recordings were coded with the OPTION12 instrument for shared decision-making. Two analysts independently assessed the number and types of patient questions and expressions of preferences.Results: Patient engagement did not differ by health literacy level and was generally low for all patients with respect to OPTION12 scores and the number of questions related to clinical aspects of treatment. Patients with inadequate health literacy had significantly higher decisional conflict. However, the proportion of knowledge questions answered correctly did not differ significantly by health literacy level.Conclusions: Patients with inadequate health literacy had greater decisional conflict but no difference in knowledge compared to patients with adequate health literacy. Inadequate health literacy may act as a barrier to communication, but gaps were found in patient engagement and knowledge for patients of all health literacy levels. The recorded patient-clinician encounters and the health literacy measure were valuable resources for conducting research on care delivery.
INTRODUCTION:In aggregate, existing data quality (DQ) checks are currently represented in heterogeneous formats, making it difficult to compare, categorize, and index checks. This study contributes a data element-function conceptual model to facilitate the categorization and indexing of DQ checks and explores the feasibility of leveraging natural language processing (NLP) for scalable acquisition of knowledge of common data elements and functions from DQ checks narratives. METHODS:The model defines a "data element", the primary focus of the check, and a "function", the qualitative or quantitative measure over a data element. We applied NLP techniques to extract both from 172 checks for Observational Health Data Sciences and Informatics (OHDSI) and 3,434 checks for Kaiser Permanente's Center for Effectiveness and Safety Research (CESR). RESULTS:The model was able to classify all checks. A total of 751 unique data elements and 24 unique functions were extracted. The top five frequent data element-function pairings for OHDSI were Person-Count (55 checks), Insurance-Distribution (17), Medication-Count (16), Condition-Count (14), and Observations-Count (13); for CESR, they were Medication-Variable Type (175), Medication-Missing (172), Medication-Existence (152), Medication-Count (127), and Socioeconomic Factors-Variable Type (114). CONCLUSIONS:This study shows the efficacy of the data element-function conceptual model for classifying DQ checks, demonstrates early promise of NLP-assisted knowledge acquisition, and reveals the great heterogeneity in the focus in DQ checks, confirming variation in intrinsic checks and use-case specific "fitness-for-use" checks.
CONTEXT:Sustaining electronic health data networks and maximizing return on federal investment in their development is essential for achieving national data insight goals for transforming health care. However, crossing the business model chasm from grant funding to self-sustaining viability is challenging. CASE DESCRIPTION:This paper presents lessons learned in seeking the sustainability of the Scalable Architecture for Federated Translational Inquiries Network (SAFTINet), and electronic health data network involving over 50 primary care practices in three states. SAFTINet was developed with funding from the Agency for Healthcare Research and Quality to create a multi-state network for comparative effectiveness research (CER) involving safety-net patients. METHODS:Three analyses were performed: (1) a product gap analysis of alternative data sources; (2) a Strengths-Weaknesses-Opportunities-Threat (SWOT) analysis of SAFTINet in the context of competing alternatives; and (3) a customer discovery process involving approximately 150 SAFTINet stakeholders to identify SAFTINet's sustaining value proposition for health services researchers, clinical data partners, and policy makers. FINDINGS:The results of this business model analysis informed SAFTINet's sustainability strategy. The fundamental high-level product needs were similar between the three primary customer segments: credible data, efficient and easy to use, and relevance to their daily work or 'jobs to be done'. However, how these benefits needed to be minimally demonstrated varied by customer such that different supporting evidence was required. MAJOR THEMES:The SAFTINet experience illustrates that commercialization-readiness and business model methods can be used to identify multi-sided value propositions for sustaining electronic health data networks and their data capabilities as drivers of health care transformation.
Context: Enhanced recovery after surgery (ERAS) aims to improve surgical outcomes by integrating evidence-based practices across preoperative, intraoperative, and postoperative care. Data in electronic medical records (EMRs) provide insight on how ERAS is implemented and its impact on surgical outcomes. Because ERAS is a multimodal pathway provided by multiple physicians and health care providers over time, identifying ERAS cases in EMRs is not a trivial task. To better understand how EMRs can be used to study ERAS, we describe our experience with using current methodologies and the development and rationale of a new method for retrospectively identifying ERAS cases in EMRs. Case Description: Using EMR data from surgical departments at the University of North Carolina at Chapel Hill, we first identified ERAS cases using a protocol-based method, using basic information including the date of ERAS implementation, surgical procedure and date, and primary surgeon. We further examined two operational flags in the EMRs, a nursing order and a case request for OR order. Wide variation between the methods compelled us to consult with ERAS surgical staff and explore the EMRs to develop a more refined method for identifying ERAS cases. Method: We developed a two-step method, with the first step based on the protocol definition and the second step based on an ERAS-specific medication definition. To test our method, we randomly sampled 150 general, gynecological, and urologic surgeries performed between January 1, 2016 and March 30, 2017. Surgical cases were classified as ERAS or not using the protocol definition, nursing order, case request for OR order, and our two-step method. To assess the accuracy of each method, two independent reviewers assessed the charts to determine whether cases were ERAS. Findings: Of the 150 charts reviewed, 74 were ERAS cases. The protocol only method and nursing order flag performed similarly, correctly identifying 74 percent and 73 percent of true ERAS cases, respectively. The case request for OR order flag performed less well, correctly identifying only 44 percent of the true ERAS cases. Our two-step method performed well, correctly identifying 98 percent of true ERAS cases. Conclusion: ERAS pathways are complex, making study of them from EMRs difficult. Current strategies for doing so are relatively easy to implement, but unreliable. We have developed a reproducible and observable ERAS computational phenotype that identifies ERAS cases reliably. This is a step forward in using the richness of EMR data to study ERAS implementation, efficacy, and how they can contribute to surgical care improvement.
BACKGROUND:Since the inception of Accountable Care Organizations (ACOs), many have acknowledged the potential synergy between ACOs and health information technology (IT) in meeting quality and cost goals.OBJECTIVE:We conducted a systematic review of the literature in order to describe what research has been conducted at the intersection of health IT and ACOs and identify directions for future research.METHODS:We identified empirical studies discussing the use of health IT via PubMed search with subsequent snowball reference review. The type of health IT, how health IT was included in the study, use of theory, population, and findings were extracted from each study.RESULTS:Our search resulted in 32 studies describing the intersection of health IT and ACOs, mainly in the form of electronic health records and health information exchange. Studies were divided into three streams by purpose; those that considered health IT as a factor for ACO participation, health IT use by current ACOs, and ACO performance as a function of health IT capabilities. Although most studies found a positive association between health IT and ACO participation, studies that address the performance of ACOs in terms of their health IT capabilities show more mixed results.CONCLUSIONS:In order to better understand this emerging relationship between health IT and ACO performance, we propose future research should consider more quasi-experimental studies, the use of theory, and merging health, quality, cost, and health IT use data across ACO member organizations.
CONTEXT:Integrating electronic health records (EHR) with other sources of administrative data is key to identifying factors affecting the long-term health of traditionally underserved populations, such as individuals involved in the justice system. Linking existing administrative data from multiple sources overcomes many of the limitations of traditional prospective studies of population health, but the linking process assumes high levels of data quality and consistency within administrative data. Studies of EHR, unlike other types of administrative data, have provided guidance to evaluate the utility of big data for population health research.CASE DESCRIPTION:Here, an established EHR data quality framework was applied to identify and describe the potential shortcomings of administrative juvenile justice system data collected by one of four case management systems (CMSs) across 12 counties in a Midwest state. The CMS data were reviewed for logical inconsistencies and compared along the data quality dimensions of plausibility and completeness.MAJOR THEMES:After applying the data quality framework, several patterns of logical inconsistencies within the data were identified. To resolve these inconsistencies, recommendations regarding data entry, review, and extraction are offered.CONCLUSION:The recommendations related to achieving quality justice system data can be applied to future efforts to link administrative databases from multiple sources. Increasing trust in administrative data quality related to vulnerable populations ultimately improves knowledge of pressing public health concerns.
BACKGROUND:Tailored care management requires effectively segmenting heterogeneous populations into actionable subgroups. Using patient reported data may help identify groups with care needs not revealed in traditional clinical data.METHODS:We conducted retrospective segmentation analyses of 9,617 Kaiser Permanente Colorado members age 65 or older at risk for high utilization due to advanced illness and geriatric issues who had completed a Medicare Health Risk Assessment (HRA) between 2014 and 2017. We separately applied clustering methods and latent class analyses (LCA) to HRA variables to identify groups of individuals with actionable profiles that may inform care management. HRA variables reflected self-reported quality of life, mood, activities of daily living (ADL), urinary incontinence, falls, living situation, isolation, financial constraints, and advance directives. We described groups by demographic, utilization, and clinical characteristics.RESULTS:Cluster analyses produced a 14-cluster solution and LCA produced an 8-class solution reflecting groups with identifiable care needs. Example groups included: frail individuals with memory impairment less likely to live independently, those with poor physical and mental well-being and ADL limitations, those with ADL limitations but good mental and physical well-being, and those with few health or other limitations differentiated by age, presence or absence of a documented advance directive, and tobacco use.CONCLUSIONS:Segmenting populations with complex care needs into meaningful subgroups can inform tailored care management. We found groups produced through cluster methods to be more intuitive, but both methods produced actionable information. Applying these methods to patient-reported data may make care more efficient and patient-centered.
Suicide is a preventable public health problem. Zero Suicide (ZS) is a suicide prevention framework currently being evaluated by Mental Health Research Network investigators embedded in six Health Care Systems Research Network (HCSRN) member health systems implementing ZS. This paper describes ongoing collaboration to develop population-based process improvement metrics for use in, and comparison across, these and other health systems. Robust process improvement metrics are sorely needed by the hundreds of health systems across the country preparing to implement their own best practices in suicide care. Here we articulate three examples of challenges in using health system data to assess suicide prevention activities, each in ascending order of complexity: 1) Mapping and reconciling different versions of suicide risk assessment instruments across health systems; 2) Deciding what should count as adequate suicide prevention follow-up care and how to count it in different health systems with different care processes; and 3) Trying to determine whether a safety planning discussion took place between a clinician and a patient, and if so, what actually happened. To develop broadly applicable metrics, we have advocated for standardization of care processes and their documentation, encouraged standardized screening tools and urged they be recorded as discrete electronic health record (EHR) variables, and engaged with our clinical partners and health system data architects to identify all relevant care processes and the ways they are recorded in the EHR so we are not systematically missing important data. Serving as embedded research partners in our local ZS implementation teams has facilitated this work.
Introduction: Priorities Wizard is an electronic health record-linked, web-based clinical decision support (CDS) system designed and implemented at multiple Health Care Systems Research Network (HCSRN) sites to support high quality outpatient chronic disease and preventive care. The CDS system (a) identifies patients who could substantially benefit from evidence-based actions; (b) presents prioritized evidence-based treatment options to both patient and clinician at the point of care; and (c) facilitates efficient ordering of recommended medications, referrals or procedures.Methods: The CDS system extracts relevant data from electronic health records (EHRs), processes the data using Web-based clinical decision support algorithms, and displays the CDS output seamlessly on the EHR screen for use by the clinician and patient. Through a series of National Institutes of Health-funded projects led by HealthPartners Institute and the HealthPartners Center for Chronic Care Innovation and HCSRN partners, Priorities Wizard has been evaluated in cluster-randomized trials and expanded to include over 20 clinical domains.Results: Cluster-randomized trials show that this CDS system significantly improved glucose and blood pressure control in diabetes patients, reduced 10-year cardiovascular (CV) risk in high-CV risk adults without diabetes, improved management of smoking in dental patients, and improved high blood pressure identification and management in adolescents. CDS output was used at 71–77 percent of targeted visits, 85–98 percent of clinicians were satisfied with the CDS system, and 94 percent reported they would recommend it to colleagues.Conclusions: Recently developed EHR-linked, Web-based CDS systems have significantly improved chronic disease care outcomes and have high use rates and primary care clinician satisfaction.
