2015 Keeneland Conference Session 3B

Data, Electronic Health Records, & Capacity Building

Room: Blackberry Lilly
Wednesday, April 22, 2015, 10:45 AM to 12:15 PM

Moderator: Valerie A. Yeager, DrPH


Jonathon Leider, PhD

Characterizing the Informatics Capacities and Needs of Local Health Departments in a Post-Affordable Care Act Landscape

Co-Investigator(s): Gulzar Shah, PhD; Brian Castrucci, MA; Karmen Williams, MPH; Akrati Gupta, MPH; & James Sprague, MD

Background: With dramatic gains in the proportion of physician offices using Electronic Health Records (EHRs), public health departments are looking at a profound and vast source of information on the health status and health behaviors of the people they serve. While EHRs are arguably the largest “new” source of data, many other types of administrative and other data exchange occur across the nation’s 2,800 health departments. The ability to access and make use of these myriad data sources is not at all uniform. Research Objective: Characterize informatics capacities and needs of local health departments Data Sets and Sources: This project draws on interview data from 50 leaders of local health departments across the United States Study Design: This project employed a key-informant based study design, where 50 leaders from local health departments across the country were interviewed about current practices, capacities, and needs in the realm of public health informatics in fall 2014. Participants were purposively selected by their LHD’s size and informatics uptake based on response to the 2013 Profile from the National Association of County and City Health Officials.  Analysis: Data were coded thematically and independently in batches by two researchers.  After each batch of interviews was coded, coding was compared, differences were resolved iteratively, and all interviews were re-coded using the consensus definitions.  Qualitative data analysis focused on major drivers of informatics uptake, or lack of uptake. Analysis paid special attention to the notion of modifiable versus non-modifiable characteristics that affected uptake of more sophisticated systems or practices.   Principal Findings: Preliminary results show wide variability in the types of systems LHDs utilize, as well as the availability of formal data collection and management technologies. This appears somewhat to be a function of size, but at least as important from participant perspectives are statutory responsibilities, local county context, and interaction with the state. Interoperability is the exception, not the rule. Many LHD leaders said that even with new systems in development, interoperability is not a prospective value, although the lack thereof is a significant constraint. The majority of interview participants who have EHRs have had them for fewer than five years, and many fewer than one year. The majority of leaders from small LHDs said they didn’t think they would be able to analyze most new electronic health information that might be accessible in the future due to challenges with human capital. Conclusions: Even with significant financial and other operational constraints, leaders interviewed as part of this project talked about a reasonably bright, if uncertain, future for public health informatics. Uptake of more advanced systems and analytic strategies appears contingent on funding, training, and good relationships with the state health agencies to ensure appropriate data that collected by public and private organizations are available to protect and improve population health. Implications: Uptake of informatics systems is limited by many non-modifiable constraints. However, leaders can cross-train existing staff and advocate for greater support from SHAs.


Gulzar Shah, PhD, MS, BS

Characteristics of Local Health Departments Associated with Their Implementation of Electronic Health Records and Other Informatics Systems

Co-Investigator(s): Jonathon P. Leider, PhD; Brian Castrucci, MA; Karmen Williams, MSPH, MA; & Huabin Luo, PhD

Background: Information technology and information systems (IT/IS) play a critical role in the daily operation of local health departments (LHDs). Assessing LHDs’ informatics capacities is important, especially within the context of broader, system-level health reform efforts.  Research Objective:  This study assesses a nationally representative sample of LHDs’ level of adoption of information systems, technology, and the factors associated with adoption/implementation. Specifically, five areas of public health informatics were examined: electronic health records (EHRs), health information exchange (HIE), immunization registry (IR), electronic disease reporting system (EDRS), and electronic lab reporting (ELR). Data Sets and Sources:  Data from NACCHO’s 2013 National Profile of LHDs was used. Descriptive statistics and multinomial logistic regression were performed for the five implementation-oriented outcome variables of interest, with three levels of implementation.  Independent variables included infrastructural capacity, financial capacity, and other characteristics theoretically associated with informatics capacity. Study Design: This study uses a cross-sectional survey research design.   Principal Findings:  Thirteen percent of LHDs had implemented HIEs. About 22 % had implemented EHRs, 47% ELR, 72.2% EDRS, and 82% had implemented Immunization Registry.  Significant determinants of health informatics adoption included provision of greater number of clinical services, greater per capita public health expenditures, having health information system specialists on staff, having larger population size, having decentralized governance system, having one and more local boards of health, and having top executive with greater number of years in the job. Conclusions:  The capacity of LHDs to use real-time, local data and information is critical. Many LHDs do not have this capacity. This may be due to lack of specialized staff, availability of data systems, or a host of other political or organizational constraints.  This is especially the case for smaller jurisdictions.  Cross-jurisdictional sharing or regionalization of some informatics and surveillance functions may be a reasonable approach to address these shortfalls. Implications for Public Health Practice and Policy: A combination of investment in public health informatics infrastructure, additional training of new informatics staff and existing epidemiologists, and better integration with healthcare systems is needed to augment LHD informatics capacity and ensure governmental public health can meet the information needs of the 21st century.


Erika Martin, PhD, MPH

Evaluating the Quality, Usability, and Fitness of Open Data for Public Health Research

Co-Investigator(s): Jennie Law, MPA; Weijia Ran; Natalie Helbig, PhD, MPA; & Guthrie Birkhead, MD, MPH

Background: Thousands of government datasets have been released on open data platforms that are publicly accessible, available in non-proprietary formats, free of charge, and with unlimited use and distribution rights. Although these resources provide new opportunities for public health research and practice, the extent to which open health data are usable, fit, and of high quality for health research is unknown.   Research Objective: This study aims to evaluate the health data available on open data platforms, the alignment of data with best practices and usability for researchers, and differences across platforms.   Data Sets and Sources: A stratified simple random sample of 5% of Healthdata.gov data objects (federal platform, N=75), 25% of Health Data NY objects (New York State (NYS) platform, N=71), and all NYC Open Data objects in the health category (New York City (NYC) platform, N=37) was reviewed using a coding instrument. Data objects are the units of analysis.   Study Design: A systematic review of health-related data objects on three open data platforms (federal, NYS, NYC) was performed by adapting the Institute of Medicine standards for systematic literature reviews. Two raters used a 99-item coding guide to assess multiple dimensions of intrinsic and contextual data quality, metadata quality, and platform usability.  The coding guide was developed based on an interdisciplinary literature review and key informant interviews, and refined through pilot testing.   Analysis: Items were aggregated into summary indices for intrinsic data quality, contextual data quality, and adherence to the Dublin Core Metadata Standards and the five-star open data deployment scheme; all scales are from 0 to 1. ANOVAs tested for differences across platforms.   Principal Findings: Less than half of data objects were traditional tabular datasets; other objects were charts (14.7%), documents about the data (12.0%), maps (10.9%), query tools (7.7%), application programming interfaces (1.6%), or not viewable in any browser (7.7%). All objects met the “one-star” criteria of availability on the web and most met the “three-star” criteria of availability in non-proprietary formats (federal=47.3%, NYS=90.1%, NYC=79.0%), but fewer met the “five-star” criteria of being hyperlinked to other data (federal=12.1%, NYS=77.5%, NYC=0%). The NYS site had the highest intrinsic data quality index (federal=0.419, NYS=0.670, NYC=0.290, p<0.001), contextual data quality index (federal=0.42, NYS=0.82, NYC=0.43, p<0.001), and Dublin Core Metadata Standards index (federal=0.546, NYS=0.890, NYC=0.702, p<0.001). Many objects, particularly on the federal site, appear to be aimed at software developers rather than researchers.   Conclusions: Despite directives to release open data, government agencies have little guidance on how to make data usable for different user communities including health researchers. Although all platforms have areas of improvement to better meet researchers’ needs, New York’s Health Data NY platform scores high on many dimensions and could serve as a model for other sites. Sustained effort on releasing and improving these data is important for ensuring the public health community uses these data, thereby increasing the value of these data.   Implications for Public Health Practice and Delivery: Open health data are likely to become an important source of public health information.


Dawn Marie Jacobson, MD, MPH

The Impact of Health Information Technology Investments on Public Health E-reporting

Co-Investigator(s): Deborah Porterfield, MD, MPH (presenting); Benjamin Yarnoff, PhD; & Suzanne Ryan-Ibarra, MPH

Background: Major financial investments for electronic data sharing within the healthcare sector have been implemented since 2010.  The Beacon Community Grant Program provided funds to local organizations to create real-time electronic data sharing of patient health data. It is unknown how the Beacon Community lead organizations collaborated with the local health departments (LHDs) within their designated region to share resources or support LHD data system development. Research Objective:  Our objective is to determine whether LHDs located within Beacon Communities developed more robust data system capacity and efficient reporting processes than LHDs that were not located in Beacon Communities.  Data Sets and Sources:  Pre-exposure data are from the 2010 NACCHO Profile, the Area Health Resource File, and CMS provider data. Post-exposure data are from the 2013 NACCHO Profile, the Area Health Resource File, CMS provider data, and a primary LHD survey designed by the research team. Study Design:  The study is a mixed-methods, quasi-experimental design with LHDs as the unit of analysis. Phase 1 used baseline data and propensity scoring methods to match LHDs located within Beacon Communities (exposed sample, n=80) with LHDs located outside of Beacon Communities but within the same state (non-exposed, n=80). Variation at the state level was controlled through within–state matching of exposed and non-exposed LHDs. Phase 2 utilized key informant interviews from 6 LHDs in the exposed group to determine partnerships, resource sharing, LHD information technology (IT) development, and LHD electronic reporting capabilities that emerged during the Beacon Community grant timeframe. Phase 3 is in progress and uses an online survey of the matched LHD sample (n=160) to examine the association of being located in a Beacon Community region with changes in the capabilities of collecting, sharing, and reporting electronic public health data. Analysis:  Variables that assess area level demographics, LHD organizational characteristics, IT infrastructure, and timeliness of LHD reporting processes were selected. For variables collected in both Phase 1 (baseline) and Phase 3 (follow-up) of the study, we will build a quasi-experimental model that accounts for changes over time in area-level population characteristics and IT factors as well as other unobservable differences between the exposed and unexposed LHDs. Analyses will be carried out using cross-tabulations, and logistic and linear regression methods. Principal Findings:  Baseline analyses demonstrated that LHDs located in Beacon Communities are more likely than the rest of the LHD sample to have a Board of Health that sets LHD priorities, to employ a Chief Information Officer, and to conduct population-level chronic disease prevention activities. They are less likely to have a health professional leading the department, to conduct chronic disease screenings, or to provide primary care services. Results from the Phase 2 interviews confirmed that new partnerships, shared resources, and IT development occurred between LHDs and the designated Beacon Community Grant lead organizations. Conclusions:  Preliminary results suggest that LHDs located within Beacon Communities received IT development support through innovative partnerships and shared resources created through the Beacon Community grant program. The impact on LHD electronic data sharing and timeliness of LHD reporting will be further assessed upon completion of the Phase 3 LHD survey. Implications for Public Health Practice and Policy: Best practices for partnerships and resource sharing between the clinical care and public health systems for real-time data system development and efficient reporting are needed. Lessons learned could be applied to advocate for future LHD specific grant funding and/or opportunities to receive incentive dollars through designation as a “provider” in the CMS Meaningful Use program.