Challenges in using electronic health record data for CER: experience of 4 learning organizations and solutions applied.

OBJECTIVE: To document the strengths and challenges of using electronic health records (EHRs) for comparative effectiveness research (CER). METHODS: A replicated case study of comparative effectiveness in hypertension treatment was conducted across 4 health systems, with instructions to extract data and document problems encountered using a specified list of required data elements. Researchers at each health system documented successes and challenges, and suggested solutions for addressing challenges. RESULTS: Data challenges fell into 5 categories: missing data, erroneous data, uninterpretable data, inconsistencies among providers and over time, and data stored in noncoded text notes. Suggested strategies to address these issues include data validation steps, use of surrogate markers, natural language processing, and statistical techniques. DISCUSSION: A number of EHR issues can hamper the extraction of valid data for cross-health system comparative effectiveness studies. Our case example cautions against a blind reliance on EHR data as a single definitive data source. Nevertheless, EHR data are superior to administrative or claims data alone, and are cheaper and timelier than clinical trials or manual chart reviews. All 4 participating health systems are pursuing pathways to more effectively use EHR data for CER.A partnership between clinicians, researchers, and information technology specialists is encouraged as a way to capitalize on the wealth of information contained in the EHR. Future developments in both technology and care delivery hold promise for improvement in the ability to use EHR data for CER.

Google Street View Images as Predictors of Patient Health Outcomes, 2017-2019.

Collecting neighborhood data can both be time- and resource-intensive, especially across broad geographies. In this study, we leveraged 1.4 million publicly available Google Street View (GSV) images from Utah to construct indicators of the neighborhood built environment and evaluate their associations with 2017-2019 health outcomes of approximately one-third of the population living in Utah. The use of electronic medical records allows for the assessment of associations between neighborhood characteristics and individual-level health outcomes while controlling for predisposing factors, which distinguishes this study from previous GSV studies that were ecological in nature. Among 938,085 adult patients, we found that individuals living in communities in the highest tertiles of green streets and non-single-family homes have 10-27% lower diabetes, uncontrolled diabetes, hypertension, and obesity, but higher substance use disorders-controlling for age, White race, Hispanic ethnicity, religion, marital status, health insurance, and area deprivation index. Conversely, the presence of visible utility wires overhead was associated with 5-10% more diabetes, uncontrolled diabetes, hypertension, obesity, and substance use disorders. Our study found that non-single-family and green streets were related to a lower prevalence of chronic conditions, while visible utility wires and single-lane roads were connected with a higher burden of chronic conditions. These contextual characteristics can better help healthcare organizations understand the drivers of their patients' health by further considering patients' residential environments, which present both risks and resources.

Screening for Social Determinants of Health During Primary Care and Emergency Department Encounters.

This cross-sectional study explores characteristics of patients who are screened and who screen positive for social determinants of health (SDOH) needs in different clinical settings within a large integrated health system.

Reducing Health Care Costs Through Patient Targeting: Risk Adjustment Modeling to Predict Patients Remaining High Cost.

CONTEXT: The transition to population health management has changed the healthcare landscape to identify high risk, high cost patients. Various measures of patient risk have attempted to identify likely candidates for care management programs. Pre-screening patients for outreach has often required several years of data. Intermountain Healthcare relied on cost-ranking algorithms which had limited predictive ability. A new risk-adjusted algorithm shows improvements in predicting patients' future cost status to facilitate identifying patient eligibility for care management. CASE DESCRIPTION: A retrospective cohort study design was used to evaluate high-cost patient status for two of the next three years. Modeling was developed using logistic regression and tested against other decision tree methods. Key variables included those readily available in electronic health records supplemented by additional clinical data and estimates of socio-economic status. FINDINGS: The risk-adjusted modeling correctly identified 79.0% of patients ranking among the top 15% of costs in one of the next three years. In addition, it correctly estimated 48.1% of the patients in the top 15% cost group in two of the next three years. This method identified patients with higher medical costs and more comorbid conditions than previous cost-ranking methods. MAJOR THEMES: This approach improves the predictive accuracy of identifying high cost patients in the future and increases the sensitivity of identifying at-risk patients. It also shortened data requirements to identify eligibility criteria for case management interventions. CONCLUSION: Risk-adjustment modeling may improve management programs' interface with patients thus decreasing costs. This method may be generalized to other healthcare settings.

A Framework for Aligning Data from Multiple Institutions to Conduct Meaningful Analytics.

INTRODUCTION: Health systems can be supported by collaborative networks focused on data sharing and comparative analytics to identify and rapidly disseminate promising care practices. Standardized data collection, quality assessment, and cleansing is a necessary process to facilitate meaningful analytics for operations, quality improvement, and research. We developed a framework for aligning data from health care delivery systems using the High Value Healthcare Collaborative central registry. FRAMEWORK: The centralized data registry model allows for multiple layers of data quality assessment. Our framework uses an iterative approach, starting with clear specifications, maintaining ongoing dialogue with diverse stakeholders, and regular checkpoints to assess data conformance, completeness, and plausibility. LESSONS LEARNED: We found that an iterative communication process is critical for a central registry to ensure: 1) clarity of data specifications, 2) appropriate data quality, and 3) thorough understanding of data source, purpose, and context. Engaging teams from all participating institutions and incorporating diverse stakeholders of clinicians, information technologists, data analysts, operations managers, and health services researchers in all decision making processes supports development of high quality datasets for comparative analytics across multiple institutions. CONCLUSION: A standard data specification and submission process alone does not guarantee aligned data for a collaborative registry. Implementing an iterative data quality improvement framework with extensive communication proved to be effective for aligning data from multiple institutions to support meaningful analytics.

A Retrospective Analysis of Precision Medicine Outcomes in Patients With Advanced Cancer Reveals Improved Progression-Free Survival Without Increased Health Care Costs.

PURPOSE: The advent of genomic diagnostic technologies such as next-generation sequencing has recently enabled the use of genomic information to guide targeted treatment in patients with cancer, an approach known as precision medicine. However, clinical outcomes, including survival and the cost of health care associated with precision cancer medicine, have been challenging to measure and remain largely unreported. PATIENTS AND METHODS: We conducted a matched cohort study of 72 patients with metastatic cancer of diverse subtypes in the setting of a large, integrated health care delivery system. We analyzed the outcomes of 36 patients who received genomic testing and targeted therapy (precision cancer medicine) between July 1, 2013, and January 31, 2015, compared with 36 historical control patients who received standard chemotherapy (n = 29) or best supportive care (n = 7). RESULTS: The average progression-free survival was 22.9 weeks for the precision medicine group and 12.0 weeks for the control group ( P = .002) with a hazard ratio of 0.47 (95% CI, 0.29 to 0.75) when matching on age, sex, histologic diagnosis, and previous lines of treatment. In a subset analysis of patients who received all care within the Intermountain Healthcare system (n = 44), per patient charges per week were $4,665 in the precision treatment group and $5,000 in the control group ( P = .126). CONCLUSION: These findings suggest that precision cancer medicine may improve survival for patients with refractory cancer without increasing health care costs. Although the results of this study warrant further validation, this precision medicine approach may be a viable option for patients with advanced cancer.

Using Electronic Health Record Data to Identify Prostate Cancer Patients That May Qualify for Active Surveillance.

INTRODUCTION: The introduction of the protein-specific antigen (PSA) test in care means that prostate cancer (PCa) is being detected earlier and more frequently. The result of increased screening using PSA, digital rectal examination and awareness of prostate was an increase in the number of men with low risk cancers. Active surveillance has become a viable alternative to immediate treatment with surgery, radiation and other forms of localized treatment. Evidence suggests that there is no significant difference in mortality rates between AS and surgery. In addition, patients may potentially delay other complications associated with surgery, radiation or deprivation therapy. METHODS: This quality improvement study using a delivery system science framework describes the electronic identification of AS candidates given an evidence-based identification criteria based upon national guidelines and clinical judgement. The study population (n=649) was drawn from health records of all patients who received a prostate biopsy (n=1731) at Intermountain Healthcare from 1/1/2013 to 12/31/2014. Automated and manual abstraction was performed. Receiver operating characteristic (ROC) analysis was used to compare AS-eligible patients to the actual treatment received to identify potential care improvement opportunities. Among those with complete data, 24.7% of this population (n=160) met "AS-eligible" criteria. 39.1% of the population had not received surgery, radiation or androgen deprivation therapy and were considered as being treated using an AS approach. 9% of AS-eligible patients did not receive AS; 27% of patients who did not meet AS-eligible criteria received AS. Estimated guideline adherence measured using area under the curve was 0.70 (95% CI: 0.66-0.73). Modest variation in criteria parameters for identifying AS-eligible patients did not significantly change estimated adherence levels. CONCLUSION: Implementation of evidence-based criteria for detection of AS candidates is feasible using electronic health record data and provides a reasonable basis for delivery system evaluation of practice patterns and for quality improvement.

Introduction of an Area Deprivation Index Measuring Patient Socioeconomic Status in an Integrated Health System: Implications for Population Health.

INTRODUCTION: Intermountain Healthcare is a fully integrated delivery system based in Salt Lake City, Utah. As a learning healthcare system with a mission of performance excellence, it became apparent that population health management and our efforts to move towards shared accountability would require additional patient-centric metrics in order to provide the right care to the right patients at the right time. Several European countries have adopted social deprivation indices in measuring the impact that social determinants can have on health. Such indices provide a geographic, area-based measure of how socioeconomically deprived residents of that area are on average. Intermountain's approach was to identify a proxy measure that did not require front-line data collection and could be standardized for our patient population, leading us to the area deprivation index or ADI. This paper describes the specifications and calculation of an ADI for the state of Utah. Results are presented along with introduction of three use cases demonstrating the potential for application of an ADI in quality improvement in a learning healthcare system. CASE DESCRIPTION: The Utah ADI shows promise in providing a proxy for patient-reported measures reflecting key socio-economic indicators useful for tailoring patient interventions to improve health care delivery and patient outcomes. Strengths of this approach include a consistent standardized measurement of social determinants, use of more granular block group level measures and a limited data capture burden for front-line teams. While the methodology is generalizable to other communities, results of this index are limited to block groups within the state of Utah and will differ from national calculations or calculations for other states. The use of composite measures to evaluate individual characteristics must also be approached with care. Other limitations with the use of U.S. Census data include use of estimates and missing data. CONCLUSION: Initial applications in three meaningfully different areas of an integrated health system provide initial evidence of its broad applicability in addressing the impact of social determinants on health. The variation in socio-economic status by quintile also has potential for clinical significance, though more research is needed to link variation in ADI with variation in health outcomes overall and by disease type.

Variable specimen handling affects hormone receptor test results in women with breast cancer: a large multihospital retrospective study.

CONTEXT: Intermountain Healthcare hospitals use a single, standardized laboratory and automated testing process for estrogen receptor/progesterone receptor (ER/PR) tests to minimize testing errors. OBJECTIVES: To test the (1) variability in ER/PR negativity among hospitals and (2) association between specimen handling conditions and ER/PR negativity. DESIGN: Retrospective study of women who had breast cancer surgery at 7 Intermountain hospitals and ER/PR tests ordered between 1997 and 2003. Data were extracted from cancer registry. Frequency of ER/PR negativity was calculated for each surgery day and compared among hospitals and between 2 groups: regular (specimens obtained Sunday through Thursday, more likely to be tested within 24 hours of surgery) and prolonged (specimens obtained on Friday and Saturday, more likely to be tested more than 24 hours after surgery) specimen handling conditions. RESULTS: Five thousand seventy-seven women were tested for ER/PR. The frequency of ER and PR negativity was 20.9% and 27.9%, respectively. It increased with each day of the week for both ER (P = .03) and PR (P = .059) and tended to be higher for prolonged specimens for ER (23.6% versus 20.4%; P = .03) and for PR (30.1% versus 27.4%; P = .11) compared with regular specimens. After controlling for age and tumor size, both ER (P = .02) and PR (P = .02) negativity was significantly different among the hospitals and was associated with prolonged specimens for ER (P = .04) but not for PR (P = .09). CONCLUSIONS: Estrogen receptor and PR negativity remained highly variable among hospitals despite use of a single laboratory and tended to be significantly associated with prolonged specimen handling. More studies are needed to confirm these findings.