Developing Methods of Repurposing Electronic Health Record Data for Identification of Older Adults at Risk of Unintentional Falls

by Adam Baus, PhD, MA, MPH; Keith Zullig, PhD, FASHA; Dustin Long, PhD; Charles Mullett, MD, PhD; Cecil Pollard, MA; Henry Taylor, MD, MPH; and Jeffrey Coben, MD

Abstract

Nationally, nearly 40 percent of community-dwelling adults age 65 and older fall at least once a year, making unintentional falls the leading cause of both fatal and nonfatal injuries among this age group. Addressing this public health problem in primary care offers promise. However, challenges in incorporating fall risk screening into primary care result in a problem of missed opportunities for screening, counseling, intervention, and ultimately prevention. Given these barriers, this study examines the potential for the innovative use of routinely collected electronic health record data to provide enhanced clinical decision support in busy, often resource-thin primary care environments. Using de-identified data from a sample of West Virginia primary care centers, we find that it is both feasible and worthwhile to repurpose routinely collected data for the purpose of identification of older adults at risk of falls. Searching of both free-text and semistructured data was particularly valuable.

Key words: electronic health records; unintentional falls; older adults; West Virginia

Introduction

Unintentional falls among older adults are a complex, formidable public health problem both nationally and in West Virginia. Falls often result in moderate to severe injuries such as head trauma and fractures while increasing the risk of early death.1 Recent information from the US Preventive Services Task Force (USPSTF) highlights that nearly 40 percent of community-dwelling adults age 65 and older fall at least once a year, making unintentional falls the leading cause of both fatal and nonfatal injuries among this age group.2, 3 Unintentional falls accounted for more than 70 percent of emergency department visits among persons age 65 and older in 2010.4 In 2012, there were 2.4 million nonfatal emergency department visits due to falls among older adults, with approximately 722,000 of those events resulting in hospitalization.5 Further, recent research highlights an increased prevalence of falls among older adults.6 This problem is especially relevant in West Virginia, in which the population is aging faster on average than the rest of the nation.7, 8 Further, poor health outcomes and complications following falls are exacerbated by various comorbidities prevalent among older adults.9 Direct medical costs associated with these injuries were approximately $19.2 billion in 2000,10 increased to approximately $30 billion in 2012,11 and are projected to reach $43.8 billion by 2020.12

This nonexperimental retrospective study examines the utility of importing electronic health record (EHR) data into an external clinical information system to systematically identify older patients at risk of falls among select West Virginia primary care centers. Data are from an EHR that was certified by the Certification Commission for Health Information Technology (CCHIT). Previous research has identified common limitations of EHRs in the area of functionality necessary for analysis and research because they are instead designed primarily to support patient care.13–17 Given this limitation, this research repurposes EHR data for fall risk identification, paying particular attention to the determination of the value added by data gathered from various areas of the medical record, including free-text notes. The outcome of interest is the development of methods to repurpose EHR data to identify this particular at-risk patient population. This expanded use of EHR data increases the opportunity to transform data collected at the time of patient care into knowledge that can be applied to better target services and intervention to patients in need, inform healthcare decisions, and bolster practice-based research.18 Further, this approach offers the advantage of moving from an acute model of patient-by-patient screening to one of a planned, population-based model of data-driven clinical decision support for fall risk identification. This study was classified as non–human subjects research by the West Virginia University Office of Research Integrity and Compliance (protocol number 1402217616) because it involves secondary data that do not include information protected under the Health Insurance Portability and Accountability Act (HIPAA).

Background

Detecting community-dwelling older adults at risk of falling poses a serious challenge. The timed Get-Up-and-Go test is the gold-standard assessment recommended by the USPSTF for determining fall risk.19 However, this test is best considered within a larger battery of assessments to more definitively measure physical function20, 21 and depends on clinicians’ use of standard procedures and equipment.22 Although the test can be completed in less than a minute, this additional task can be challenging to incorporate into brief office visits given the complex health needs of older patients.23 Nationally, screening for fall risk is completed only 30 to 37 percent of the time.24

Given the need for efficient, systematic primary care screening for fall risk, exploring the use of EHR data to identify patients at risk is warranted. EHRs have the potential to be valuable tools for health outcomes research in primary care25–29 and a critical component in the reduction of preventable deaths through increased adherence to preventive services.30 However, EHRs are primarily designed to support patient-level care and often lack population-level reporting and health analytics features essential to public health efforts.31–35 Moving EHR data to an external system allows for more in-depth querying of the data, data transparency in that key data (i.e., patient diagnoses, demographics, vital signs, laboratory results, and services) can be queried for coding consistency and completeness and for analysis of free-text or narrative data. Analysis of free-text or narrative data is of particular interest because of the potential for essential information to be found in these locations and not in the coded areas of the EHR data.36–41

While repurposing EHR data for the identification of patients at risk of some chronic health conditions has been explored,42–44 to date no known studies have examined the use of EHR data for identification of older adults at risk of falls. Given this gap in knowledge, this study examines the utility of importing EHR data into an external clinical information system to systematically identify older patients at risk of falls, incorporating methods for determining the accuracy and completeness of the data, or internal validity. Considering the tendency for important information to be entered into EHRs in free-text or narrative form as opposed to standardized data entry,45–50 a secondary aim of this study is to use natural language processing methods to assess the potential for and value of finding information related to fall risk in free-text or narrative data in the EHR. Criteria used to identify fall risk reflect current fall prevention guidelines presented in a systematic review of current USPSTF guidelines and a meta-analysis of fall risk factors among community-dwelling older adults.51, 52

Methods

This study is a nonexperimental retrospective analysis of de-identified EHR data from two primary care center organizations, representing nine physical locations, excluding school-based health centers and dental clinics, partnering with the West Virginia University Office of Health Services Research. Data were gathered using extract, transform, and load (ETL) methodology.53 Appropriate data were selected and collected for analytical processing using SAP Business Objects.54 This software is linked with the EHR and allows for querying and data export. Transformation of the data files was performed with a Microsoft Access–based clinical information system.55 This tool is open-source, public domain software shown effective in previous research analyzing EHR data for diagnostic coding56 and identifying patients at risk of diabetes.57 Data were de-identified using the Safe Harbor method of data de-identification.58 De-identified data were loaded from the Microsoft Access–based clinical information system into JMP Pro version 11.0, which served as the common data repository for analysis. The EHR data were examined for completeness and accuracy, which are measures of internal validity, through descriptive analyses in which JMP Pro was used to calculate percentages of missing, out-of-range, and questionable results for each data element.59 Natural language processing techniques, based on manual evaluation, were used to examine the potential for obtaining value-added information from free-text or narrative data in the medical record. This was an iterative process in which a series of search terms were successively refined to improve their case-finding ability. With the use of the string handling functions in Microsoft Access Visual Basic for Applications (VBA), pertinent clinical narrative elements suggesting fall risk were identified, extracted, and coded into the same database format as the coded EHR data to retain continuity of the database structure to help ensure that the information could be presented in a way that would be suitable for use by clinicians and researchers.60 Value added in locating data throughout various parts of the medical record (i.e., structured, semistructured, and free text) was determined through descriptive analyses examining the percentage of cases missed when accounting for International Classification of Diseases, Ninth Revision (ICD-9) or Current Procedural Terminology (CPT) coding alone. Further, the chi-square test of independence was used to examine the relationship between variables.

Measures

Three categories of modifiable risk factors are associated with falls among older adults: biological, behavioral, and environmental.61 The primary risk criteria for falls included in this study are biological and behavioral because these data elements are intrinsic to the patient and therefore more apt to be gleaned from EHR data. Criteria used to identify fall risk reflect current fall prevention guidelines presented in a systematic review of current USPSTF guidelines and a meta-analysis of fall risk factors among community-dwelling older adults.62, 63 Key variables of interest were as follows: age greater than or equal to 65 years; female gender; gait or balance impairment; history of falls; fear of falling; vision impairment; hearing impairment; diagnosis of Parkinson’s disease; dizziness/vertigo; cognitive impairment; use of a walking aid or device; current prescription for a sedative medication; current prescription for an antiepileptic medication; current prescription for an antihypertensive medication; and polypharmacy (currently taking four or more medications). Appendix A lists the priority factors, the locations in which the data were found, and the coding used to locate the data.

This core set of variables was expanded to include a set of secondary variables based on a literature review of potential fall risk factors. Expanded factors or variables of interest were as follows: race, ethnicity, insurance status, fall assessment, fall guidance, hypertension, hypotension, dementia, osteoporosis, muscle weakness, rheumatoid arthritis, type 1 diabetes, type 2 diabetes, diabetic retinopathy, diabetic neuropathy, epilepsy, height, weight, body mass index (BMI), systolic blood pressure, and diastolic blood pressure. Insurance status was included to account for potential differences among patient groups. Appendix B lists the expanded set of variables, the locations in which the data were found, and the coding used to locate the data.

Results

The data sets from the two primary care organizations included in this study comprised nine unique locations excluding school-based health centers and dental clinics. Among these nine sites, 50,433 unique patients were identified. Of these, 43,531 patients (86.3 percent) were determined to be active on the basis of having at least one documented office visit, service, or laboratory test within three years of the date of data extraction (March 31, 2014). Among the 43,531 active patients, 3,933 patients (9.03 percent) were age 65 and older (see Table 1). This finding is slightly lower than the Uniform Data System result for patients 65 and older for 2013, which is 12.8 percent, yet is comparable to national results, with 7.0 percent of the national patient population age 65 and older.64

Table 2 provides demographic data for the 3,933 patients age 65 and older. While statewide data for patients were sought for comparability, only gender statistics were available for the patient population age 65 years and older (87.9 percent female; 12.1 percent male).65 Table 2 details patient demographics by age categories, gender, race, ethnicity, and health insurance information. Patients tended to be age 65 to 74 years (62.1 percent) with a mean age of 73.5 years; female (61.3 percent); white (95.7 percent); not Hispanic/Latino (99.1 percent); and insured under Medicare (63.1 percent). Data completeness and quality were strong in that all demographic data were coded consistently, which is likely attributable to standardization in the EHR data selections upon data entry; that is, there were no missing data (i.e., empty cells) across these metrics, only 0.1 percent of patients refused to report race or had race marked as unreported, and only 0.2 percent of patients refused to report ethnicity or had ethnicity marked as unreported.

Table 3 provides data on physical characteristics and vital signs of the 3,933 patients age 65 and older. These data include patient height, weight, BMI, and systolic and diastolic blood pressures. In general, patients tended to be overweight with relatively controlled blood pressure. However, a chi-square test of independence was performed to further examine the relation between age and BMI. The relation between these variables was significant, χ2(1, N = 3,607) = 127.3, p < .0001. Patients age 65 to 84 years were more likely to be overweight or obese than patients 85 years and older. A check on data completeness and quality revealed some issues, with 8.0 percent of patients 65 and older having no documented height in their medical records, 2.9 percent having no documented weight, and 1.3 percent having no documented systolic or diastolic blood pressure readings. The majority of patients with these data missing were in the 65- to 84-year age range (90.1 percent height, 91.1 percent weight, 91.1 percent BMI, 89.3 percent systolic, 89.8 percent diastolic).

Priority health conditions relating to unintentional falls were identified in a stepwise process using data from multiple areas of the EHR in order to build a data set as complete as possible. These areas are (1) ICD-9 coding; (2) Medcin findings, which are semistructured data; (3) free-text notes; and (4) vital signs related to both high and low blood pressure diagnoses. Table 4 provides data on the value added by searching multiple areas of the medical record data. ICD-9 coding alone missed a minimum of 1.2 percent of cases (diabetes type 2) to a maximum of 98.1 percent of cases (vision impairment), with a median of 39.8 percent of cases missed across all conditions. Looking to multiple areas of the EHR data to identify patients with priority health conditions offers a clear advantage in case finding. Noteworthy, fear of falling, which is one of the priority fall risk metrics, was identified in only 1 patient record (0.02 percent) across all search methods. Likewise, use of a walking aid was identified in only 6 patient records (0.1 percent). Those instances were identified using free-text notes as opposed to coded information. Appendix A provides information on the specific text string used to search for this key word.

In sum, 238 instances of falls were documented among patients age 65 and older. These falls were documented across 133 unique patients. Falls range from a minimum of one documented fall among 80 patients (60.1 percent) to a maximum of 16 documented falls in one patient (0.7 percent), with a median of one documented fall. Free-text information was especially important in the identification of patients with a history of falls, with 33.8 percent of all cases added through free-text notes. Even with this expanded search method, however, only 133 patients (3.4 percent) had an indication in their medical records of having had an unintentional fall at some point in the past. This is likely a low estimate because one of three adults aged 65 and older nationwide experiences a fall each year, yet less than half of these individuals talk with their healthcare providers about falling.66 Free-text searches were also developed to identify falls using the derivations “slip,” “trip,” and “stumble.” Only 1 patient record (0.02 percent) had an indication of having stumbled. This notation, however, included no mention of a fall and therefore added no value to the case-finding process. No patient records were identified through searches on the words “slip” or “trip.” Appendix B provides information on the specific text strings used to search for these key words.

A recent systematic review of current USPSTF guidelines and a meta-analysis of fall risk factors among community-dwelling older adults67, 68 highlight sedatives, antiepileptic medications, and antihypertensive medications as associated with increased risk of unintentional falls. Further, polypharmacy, defined as currently taking four or more medications,69, 70 is also highlighted as associated with increased risk of unintentional falls. Table 5 provides counts and percentages of active patients age 65 and older that were identified as having current prescriptions for these priority medications or polypharmacy. Data on medications were found in the medications portion of the EHR data only. Polypharmacy was identified in 85 percent of patients 65 and older.

Documented fall risk assessments were identified using data from multiple areas of the EHR. These areas are (1) CPT coding, (2) Medcin findings, and (3) free-text notes. Table 6 provides information on the counts of patients with documented fall risk assessments according to each search method, the numbers of patients added in each consecutive data step, the total unduplicated counts, and the prevalence among patients 65 and older. Noteworthy, only 23 patients (0.6 percent) have documentation in their medical records of a fall risk assessment at some point in the past. CPT coding alone missed 26.1 percent of all fall risk assessments. The value added by free-text notes alone is 13.0 percent of all assessments. Further, only two patient records (0.05 percent) indicated that the patient received anticipatory fall guidance at any time. Both of those instances were located in semistructured Medcin findings. Neither of these patients had documentation of a fall. Appendix B provides information on the specific text strings used to search for these metrics.

Discussion

This study supports the development of a novel methodology for repurposing EHR data to identify older patients at risk of falls for the purpose of early identification of risk and efforts toward prevention. Further, findings from this study draw attention to the need for increased emphasis on fall prevention during routine office visits. Among the 3,933 patients age 65 and older, only 133 patients (3.4 percent) had indication in their medical records of having had an unintentional fall at some point in the past. Searching the free-text data was vital to finding even this low number of patients because 33.8 percent of them were identified using free-text searches. Given the national statistic that falls occur among approximately 40 percent of adults 65 and older,71 we can be confident that falls are underreported and/or underdocumented in this sample. Likewise, fall risk assessments were sparse, with only 23 patients (0.6 percent) having documentation in their medical records of a fall risk assessment at some point in the past. As with falls, fall risk assessments in the EHR data were largely found in semistructured and free-text data. Searching the CPT coding alone missed 26.1 percent of all fall risk assessments. While this study is based on one EHR system only, the results suggest that thorough accounting for multiple data types when searching for clinical information is important to ensure quality data for population health management, quality of care improvement, and practice-based research. Further, this study points to the need for EHRs to be developed in such a way that a comprehensive set of fall risk metrics can be consistently tracked and reported. A planned approach to systems development would support efforts in quality of care improvement and practice-based research.

This study draws attention to a multifaceted problem with the identification of falls in this sample of outpatient clinics. While low documentation of falls is an issue, this problem is combined with documentation practices that make it difficult to retrieve data that have been recorded. This research highlights a complex problem deserving of targeted quality improvement efforts and practice-based research. Although the Physician Quality Reporting System and the National Quality Forum have focused some attention on reporting of data and benchmarking regarding unintentional falls, the more commonly measured health conditions and metrics, such as diabetes, hypertension, vital signs, and patient demographics, were by far more commonly documented among this sample of clinics. While duration of EHR use may be a factor, all clinics in this study have used EHRs for at least six years.

One primary limitation of this study is that purposive sampling was used to identify primary care organizations for inclusion, thereby decreasing the generalizability of the findings. Second, this study focuses on intrinsic (biologic or behavioral) fall risk factors and not extrinsic, environmental risk factors because of the type of data available through the EHR. Combining data made available from EHRs with data sources offering extrinsic information would be beneficial. Third, this study is subject to limitations in the documentation of EHR data such as miscoding, missing data on falls, and gaps in data due to limited sharing of information from hospitals, physical rehabilitation centers, and other care locations where information on falls may have been recorded.

Conclusion

This expanded use of EHR data demonstrates an opportunity to transform data collected at the time of patient care into knowledge that can be applied to better target services and interventions to patients in need, inform healthcare decisions, and bolster practice-based research.72 Further, this approach offers the advantage of moving from an acute model of patient-by-patient screening to one of a planned, population-based model of data-driven clinical decision support for the identification of fall risk. The strength of this study is its practical importance to public health: it facilitates the identification of a sector of the patient population at increased risk for falls in a way that is efficient and data-driven, taking into account the healthcare demands of primary care. For EHR data to be most useful not only for identification of the risk of falls but for identification of any health condition or injury, issues of data quality, format, and accessibility need to be addressed.73 Recognizing the limits of EHR data and developing steps or interventions to improve those data are paramount, not only for health informatics purposes, but for the improvement of patient care and outcomes.

Two additional research efforts are underway in the use of EHR data for fall risk identification. First, research is being completed in the development of a validated EHR data-driven model for identifying older adults at risk of falls. This research builds on that presented in this article, aiming for application of methods to repurpose EHR data for fall risk identification and intervention in the clinical setting. Second, qualitative research into the context of fall risk screening practices in West Virginia primary care centers is underway to better understand facilitators and barriers to screening, EHR documentation practices, and the potential for EHRs to be used as clinical decision support for fall risk identification and ultimately prevention of falls.

 

Acknowledgments

Special thanks to Mary Swim, business research analyst with the West Virginia University School of Public Health Office of Health Services Research, for her support and guidance in developing the Visual Basic for Applications programming necessary for the free-text data analysis in this study. Special thanks as well to Traci Jarrett, PhD, MPH, of the West Virginia University School of Public Health Prevention Research Center and the West Virginia Clinical and Translational Science Institute and a visiting scholar at the University of Kentucky, for her support and encouragement throughout the research process.

Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under award number U54GM104942. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

 

Conflict of Interest

The authors declare no competing financial interests.

 

Adam Baus, PhD, MA, MPH, is the assistant director of the Office of Health Services Research at West Virginia University School of Public Health in Morgantown, WV.

Keith Zullig, PhD, FASHA, is the chair of the Department of Social and Behavioral Health Sciences at West Virginia University School of Public Health in Morgantown, WV.

Dustin Long, PhD, is an assistant professor in the Department of Biostatistics at West Virginia University School of Public Health in Morgantown, WV.

Charles Mullett, MD, PhD, is an associate professor and section chief of Pediatric Critical Care in the Department of Pediatrics at West Virginia University School of Medicine in Morgantown, WV.

Cecil Pollard, MA, is the director of the Office of Health Services Research at West Virginia University School of Public Health in Morgantown, WV.

Henry Taylor, MD, MPH, is a senior associate in health policy and management at Johns Hopkins Bloomberg School of Public Health in Baltimore, MD.

Jeffrey Coben, MD, is the associate vice president for clinical innovations at West Virginia University School of Public Health in Morgantown, WV.

 

Notes

  1. Centers for Disease Control and Prevention. “Falls among Older Adults: An Overview.” 2014. http://www.cdc.gov/homeandrecreationalsafety/falls/adultfalls.html.
  2. Ibid.
  3. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in Older Adults: A Systematic Evidence Review for the U.S. Preventive Task Force.” Annals of Internal Medicine (2010): 815–25.
  4. Villaveces, A., R. Mutter, P. Owens, and M. Barrett. Causes of Injuries Treated in the Emergency Department, 2010 (Statistical Brief No. 156). Rockville, MD: Agency for Healthcare Research and Quality, 2013. Available at http://www.hcup-us.ahrq.gov/reports/statbriefs/sb156.pdf.
  5. Centers for Disease Control and Prevention. “Falls among Older Adults: An Overview.”
  6. Cigolle, C., J. Ha, L. Min, P. Lee, T. Gure, N. Alexander, and C. Blaum. “The Epidemiologic Data on Falls, 1998–2010: More Older Americans Report Falling.” Journal of the American Medical Association 175, no. 3 (2015): 443–45.
  7. Christiadi. Why West Virginia Population Is Aging Faster Than the U.S. Morgantown, WV: Bureau of Business and Economic Research, West Virginia University, College of Business and Economics, 2010.
  8. US Census Bureau. State & County QuickFacts: West Virginia. 2012. http://quickfacts.census.gov/qfd/states/54000.html.
  9. Carpenter, C., M. Scheatzle, J. D’Antonio, P. Ricci, and J. Coben. “Identification of Fall Risk Factors in Older Adult Emergency Department Patients.” Academic Emergency Medicine 16, no. 3 (2009): 211–19.
  10. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in OlderAdults: A Systematic Evidence Review for the U.S. Preventive Task Force.”
  11. Centers for Disease Control and Prevention. “Falls among Older Adults: An Overview.”
  12. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in OlderAdults: A Systematic Evidence Review for the U.S. Preventive Task Force.”
  13. Dean, B., J. Lam, J. Natoli, Q. Butler, D. Aguilar, and R. Nordyke. “Use of Electronic Medical Records for Health Outcomes Research: A Literature Review.” Medical Care Research and Review 66, no. 6 (2009): 611–38.
  14. Benin, A., et al. “Validity of Using an Electronic Medical Record for Assessing Quality of Care in an Outpatient Setting.” Medical Care 43, no. 7 (2005): 691–98.
  15. Kukafka, R., et al. “Redesigning Electronic Health Record Systems to Support Public Health.” Journal of Biomedical Informatics 40, no. 4 (2007): 398–409.
  16. Terry, A., V. Chevendra, A. Think, M. Stewart, J. Marshall, and S. Cejic. “Using Your Electronic Medical Record for Research: A Primer for Avoiding Pitfalls.” Family Practice 27 (2010): 121–26.
  17. Tolar, M., and E. Balka. “Beyond Individual Patient Care: Enhanced Use of EHR Data in a Primary Care Setting.” International Perspectives in Health Informatics 164 (2011): 143–47.
  18. Okun, S., et al. Making the Case for Continuous Learning from Routinely Collected Data. Washington, DC: Institute of Medicine of the National Academies, 2013.
  19. Moyer, V. “Prevention of Falls in Community-Dwelling Older Adults: US Preventive Services Task Force Recommendation Statement.” Annals of Internal Medicine 157 (2012): 197–204.
  20. Lindsay, R., E. L. James, and S. Kippen. “The Timed Up and Go Test: Unable to Predict Falls on the Acute Medical Ward.” Australian Journal of Physiotherapy 50 (2004): 249–51.
  21. Piva, S. R., G. K. Fitzgerald, J. J. Irrgang, F. Bouzubar, and T. W. Starz. “Get Up and Go Test in Patients with Knee Osteoarthritis.” Archives of Physical Medicine and Rehabilitation 85 (2004): 284–89.
  22. Siggeirsdóttir, K., J. Akranes, H. Jónsson Jr., and S. Iwarsson. “The Timed ‘Up & Go’ Is Dependent on Chair Type.” Clinical Rehabilitation (2002): 609–16.
  23. Boyd, C. M., J. Darer, C. Boult, L. P. Fried, L. Boult, and A. W. Wu. “Clinical Practice Guidelines and Quality of Care for Older Patients with Multiple Comorbid Diseases: Implications for Pay for Performance.” Journal of the American Medical Association 294, no. 6 (2005): 716–24.
  24. Hayden, S., et al. Strengths and Needs Assessment of Older Adults in the Denver Metro Area: Technical Report. Boulder, CO: National Research Center, 2004.
  25. Dean, B., J. Lam, J. Natoli, Q. Butler, D. Aguilar, and R. Nordyke. “Use of Electronic Medical Records for Health Outcomes Research: A Literature Review.”
  26. Ethredge, L. “Creating a High-Performance System for Comparative Effectiveness Research.” Health Affairs 29, no. 10 (2010): 1761–67.
  27. Hanna, K., S. Anderson, and S. Maddox. Think Research: Using Electronic Medical Records to Bridge Patient Care and Research. Washington, DC: Center for Accelerating Medical Solutions, 2005.
  28. Weiner, M. G., J. Lyman, S. Murphy, and M. Weiner. “Electronic Health Records: High-Quality Electronic Data for Higher-Quality Clinical Research.” Informatics in Primary Care 15 (2007): 121–27.
  29. de Lusigna, S., and C. van Weel. “The Use of Routinely Collected Computer Data for Research in Primary Care: Opportunities and Challenges.” Family Practice 23, no. 2 (2005): 253–63.
  30. Farley, T., M. Dalal, F. Mostashari, and T. Frieden. “Deaths Preventable in the US by Improvements in the Use of Clinical Preventive Services.” American Journal of Preventive Medicine 38, no. 6 (2010): 600–609.
  31. Dean, B., J. Lam, J. Natoli, Q. Butler, D. Aguilar, and R. Nordyke. “Use of Electronic Medical Records for Health Outcomes Research: A Literature Review.”
  32. Benin, A., et al. “Validity of Using an Electronic Medical Record for Assessing Quality of Care in an Outpatient Setting.”
  33. Kukafka, R., et al. “Redesigning Electronic Health Record Systems to Support Public Health.”
  34. Terry, A., V. Chevendra, A. Think, M. Stewart, J. Marshall, and S. Cejic. “Using Your Electronic Medical Record for Research: A Primer for Avoiding Pitfalls.”
  35. Tolar, M., and E. Balka. “Beyond Individual Patient Care: Enhanced Use of EHR Data in a Primary Care Setting.”
  36. Chen, E., G. Hripcsak, and C. Friedman. “Disseminating Natural Language Processed Clinical Narratives.” American Medical Informatics Association Symposium Proceedings (2006): 126–30.
  37. Friedman, C., and G. Hripcsak. “Natural Language Processing and Its Future in Medicine.” Academic Medicine 74, no. 8 (1999): 890–95.
  38. Gerbier, S., et al. “Evaluation of Natural Language Processing from Emergency Department Computerized Medical Records for Intra-hospital Syndromic Surveillance.” BioMed Central Public Health 11, no. 50 (2011).
  39. Ware, H., C. Mullett, and V. Jagannathan. “Natural Language Processing Framework to Assess Clinical Conditions.” Journal of the American Medical Informatics Association 16, no. 4 (2009): 585–89.
  40. Botsis, T., G. Hartvigsen, F. Chen, and C. Weng. “Secondary Use of EHR: Data Quality Issues and Informatics Opportunities.” Summit on Translational Bioinformatics (2010): 1–5.
  41. Hayrinen, K., K. Saranto, and P. Nykanen. “Definition, Structure, Content, Use and Impacts of Electronic Health Records: A Review of the Research Literature.” International Journal of Medical Informatics 77, no. 5 (2008): 291–304.
  42. Terry, A., V. Chevendra, A. Think, M. Stewart, J. Marshall, and S. Cejic. “Using Your Electronic Medical Record for Research: A Primer for Avoiding Pitfalls.”
  43. Hanna, K., S. Anderson, and S. Maddox. Think Research: Using Electronic Medical Records to Bridge Patient Care and Research.
  44. Baus, A., G. Wood, C. Pollard, B. Summerfield, and E. White. “Registry-based Diabetes Risk Detection Schema for the Systematic Identification of Patients at Risk for Diabetes in West Virginia Primary Care Centers.” Perspectives in Health Information Management (Fall 2013).
  45. Chen, E., G. Hripcsak, and C. Friedman. “Disseminating Natural Language Processed Clinical Narratives.”
  46. Friedman, C., and G. Hripcsak. “Natural Language Processing and Its Future in Medicine.”
  47. Gerbier, S., et al. “Evaluation of Natural Language Processing from Emergency Department Computerized Medical Records for Intra-hospital Syndromic Surveillance.”
  48. Ware, H., C. Mullett, and V. Jagannathan. “Natural Language Processing Framework to Assess Clinical Conditions.”
  49. Botsis, T., G. Hartvigsen, F. Chen, and C. Weng. “Secondary Use of EHR: Data Quality Issues and Informatics Opportunities.”
  50. Hayrinen, K., K. Saranto, and P. Nykanen. “Definition, Structure, Content, Use and Impacts of Electronic Health Records: A Review of the Research Literature.”
  51. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in OlderAdults: A Systematic Evidence Review for the U.S. Preventive Task Force.”
  52. Deandrea, S., E. Lucenteforte, F. Bravi, R. Foscht, C. La Vecchta, and E. Negri. “Risk Factors for Falls in Community-dwelling Older People: A Systematic Review and Meta-analysis.” Epidemiology 21 (2010): 658–68.
  53. Business Intelligence Insider. “Data Integration Techniques (ETL and Data Federation).” May 31, 2011. Available at http://bi-insider.com/portfolio/data-integration-techniques-etl-and-data-federation/.
  54. SAP Business Objects. “SAP Business Management Software Solutions, Applications and Services.” 2013. http://www.sap.com/index.epx.
  55. West Virginia University Office of Health Services Research. “Chronic Disease Electronic Management System (CDEMS).” Available at http://publichealth.hsc.wvu.edu/ohsr/services/chronic-disease-electronic-management-system-cdems/.
  56. Baus, A., M. Hendryx, and C. Pollard. “Identifying Patients with Hypertension: A Case for Auditing Electronic Health Record Data.” Perspectives in Health Information Management (Spring 2012).
  57. Hayrinen, K., K. Saranto, and P. Nykanen. “Definition, Structure, Content, Use and Impacts of Electronic Health Records: A Review of the Research Literature.”
  58. US Department of Health and Human Services. “Health Information Privacy.” 2014. http://www.hhs.gov/ocr/privacy/hipaa/understanding/coveredentities/De-identification/guidance.html.
  59. Chan, K., J. Fowles, and J. Weiner. “Electronic Health Records and the Reliability and Validity of Quality Measures: A Review of the Literature.” Medical Care Research and Review 67, no. 5 (2010): 503–27.
  60. Chen, E., G. Hripcsak, and C. Friedman. “Disseminating Natural Language Processed Clinical Narratives.”
  61. Stevens, J. A., and R. J. Schuster. The Stopping Elderly Accidents, Deaths, & Injuries (STEADI) Tool Kit for Health Care Providers. Centers for Disease Control and Prevention, 2013.
  62. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in Older Adults: A Systematic Evidence Review for the U.S. Preventive Task Force.”
  63. Deandrea, S., E. Lucenteforte, F. Bravi, R. Foscht, C. La Vecchta, and E. Negri. “Risk Factors for Falls in Community-dwelling Older People: A Systematic Review and Meta-analysis.”
  64. US Department of Health and Human Services, Health Resources and Services Administration. “2013 Health Center Data: West Virginia Program Grantee Data.” 2013. Available at http://bphc.hrsa.gov/uds/datacenter.aspx?year=2013&state=WV.
  65. Ibid.
  66. Centers for Disease Control and Prevention. “Falls among Older Adults: An Overview.”
  67. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in Older Adults: A Systematic Evidence Review for the U.S. Preventive Task Force.”
  68. Deandrea, S., E. Lucenteforte, F. Bravi, R. Foscht, C. La Vecchta, and E. Negri. “Risk Factors for Falls in Community-dwelling Older People: A Systematic Review and Meta-analysis.”
  69. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in Older Adults: A Systematic Evidence Review for the U.S. Preventive Task Force.”
  70. Deandrea, S., E. Lucenteforte, F. Bravi, R. Foscht, C. La Vecchta, and E. Negri. “Risk Factors for Falls in Community-dwelling Ol der People: A Systematic Review and Meta-analysis.”
  71. Michael, Y. L., E. P. Whitlock, J. S. Kin, R. Fu, E. A. O’Connor, and R. Gold. “Primary Care–Relevant Interventions to Prevent Falling in Older Adults: A Systematic Evidence Review for the U.S. Preventive Task Force.”
  72. Okun, S., et al. Making the Case for Continuous Learning from Routinely Collected Data.
  73. Mendes, R., and P. Rodrigues. “Main Barriers for Quality Data Collection in EHR: A Review.” Health Informatics 2011: International Conference on Health Informatics (2011): 451–54.

Printer friendly version of this article.

Adam Baus, PhD, MA, MPH; Keith Zullig, PhD, FASHA; Dustin Long, PhD; Charles Mullett, MD, PhD; Cecil Pollard, MA; Henry Taylor, MD, MPH; and Jeffrey Coben, MD. “Developing Methods of Repurposing Electronic Health Record Data for Identification of Older Adults at Risk of Unintentional Falls.” Perspectives in Health Information Management (Spring 2016): 1-32.

Leave a Reply