Public health reporting and outbreak response: synergies with evolving clinical standards for interoperability

Abstract

Public health needs up-to-date information for surveillance and response. As healthcare application programming interfaces become widely available, a novel data gathering mechanism could provide public health with critical information in a timely fashion to respond to a fast-moving epidemic. In this article, we extrapolate from our experiences using a Fast Healthcare Interoperability Resource-based architecture for infectious disease surveillance for sexually transmitted diseases to its application to gather case information for an outbreak. One of the challenges with a fast-moving outbreak is to accurately assess its demand on healthcare resources, since information specific to comorbidities is often not available. These comorbidities are often associated with poor prognosis and higher resource utilization. If the comorbidity data and other clinical information were readily available to public health workers, they could better address community disruption and manage healthcare resources. The use of FHIR resources available through application programming and filtered through tools such as described herein will give public health the flexibility needed to investigate rapidly emerging disease while protecting patient privacy.

INTRODUCTION

Public health requires both comprehensive and up-to-date information on patients infected with an emerging disease. This case information is used for the causal investigation, contact tracing, quarantine (in some cases), policy decisions on social distancing, and population analytics to project response needs and the future evolution of an outbreak. A great deal of public health surveillance data comes from the healthcare systems, and, in the past, public health has used labor-intensive methods to collect information through paper forms or faxes sent to the public health agencies that are then entered into a health department’s database. This manual process is error-prone and does not yield timely information required by a fast-moving outbreak. However, modern interoperability technologies such as Fast Healthcare Interoperability Resources (FHIR)¹ could substantially support the data requirements of an effective public health response. Due to the widespread adoption of EHRs, most health records are digitized. However, there are some key challenges in using digital data for public health surveillance and response. For digital data to be helpful during the time of an outbreak, it must meet some essential requirements.

• The health data must be available in a digital format within health IT systems

• These data must be interoperable, meaning the health systems must be able to share data on demand

• Shared data must be aggregated, visualized, and analyzed in a manner to be used for public health surveillance and response

• The range of data available must be dynamically definable by public health, to preserve patient privacy through transmission of the minimum required information while supporting detailed case investigation

To these ends, we propose adoption of a public health surveillance information architecture known as PACER (Public Health Automated Case Event Reporting) that collects information through FHIR queries in a way that is robust in its data querying capabilities while being sensitive to privacy and confidentiality-related issues.

BACKGROUND

The US Department of Health and Human Services (HHS) recently proposed a new rule to promote interoperability across different systems.² This new rule requires the healthcare industry to adopt standardized application programming interfaces (APIs), which will allow individuals and organizations to work with interoperable data. The requirement is for electronic health record (EHR) systems to adopt the HL7 Fast Healthcare Interoperability Resources (FHIR) release 4 as a standard for healthcare an interrogatory API data exchange. As APIs for interoperable query-based data exchange become widespread, the FHIR standard can become a cornerstone for a new kind of public health surveillance: cross EHR queries of health systems to identify case information and track outbreaks and to trace their effects on the population. PACER is a healthcare and public health surveillance tool built on FHIR specifications that supports authorization and filtering of queries to a mutually agreed upon standard by providers and public health authorities that is disease-specific and tailorable to outbreak conditions.

INFORMATION ARCHITECTURE

PACER is a federated FHIR query-based public health surveillance system developed as a result of a collaboration between the CDC (Centers for Disease Control and Prevention) and Georgia Tech Research Institute (Figure 1). PACER has been deployed and pilot-tested at the Medical University of South Carolina for public health case reporting of sexually transmitted diseases. PACER leverages not only FHIR protocols, but also the HL7 clinical quality language (CQL) to specify in a reusable way and to collect essential data from EHRs. The architecture is specifically designed to ensure that health systems have control over the data accessed for public health reporting. A federated design propagates queries across regions of health systems, this being practical as responses to such queries are required for EHR certification and to prevent fines under the 21st Century Cures Act to providers for information blocking activities.³ This modular and federated design of PACER, based on mandated federal rule making, allows PACER to play a critical role in data collection during times of a fast-moving outbreak. For the pilot on sexually transmitted diseases, we used electronic laboratory reporting messages coming into a simulated health department as the initial trigger for PACER, which subsequently, using a provider directory, gathered data on diagnosis and treatment to create a complete case report. The query definition is predefined and reviewed by the health system prior to implementation. PACER cannot ask ad hoc questions of the EHR but rather only those query scripts that have been adopted by agreeing parties, requiring consensus by both state and local public health authorities, CDC, and local health systems. Adoption for a new outbreak scenario requires only modification of the CQL query script to gather relevant data elements specific to a certain situation (eg, respiratory disease, chronic conditions). Additionally, the query can be updated at any time as new data requirements emerge with appropriate governance but without other changes to the technology.

Figure 1.

Architecture of PACER.

A PACER server within the clinical provider’s network receives and manages the CQL query. Based on this CQL, PACER queries the health system’s FHIR server and can automatically map between standard and local terminologies. PACER also includes a data-filtering mechanism that ensures provider systems have complete control over the data extracted from the EHR and that it is consistent with case reporting requirements. The filtered information is then sent as a case report to the health department, where it is stored in the health department’s surveillance system. Public health departments may only request and receive reports using preapproved queries, hence avoiding the risk of data overreach or gathering information extraneous to the notifiable condition. This pull-based approach, with collaborative specification of queries in response to emerging information needs, differentiates the PACER architecture from alternative electronic case reporting mechanisms.

Obtaining and analyzing health data has many inherent challenges. These challenges can be magnified when normal clinical and logistical processes are strained as during a pandemic. Without semantic interoperability, data collected from various sources is not easily normalized for population-level surveillance and response. We envision that PACER could be used in 2 phases to help with the outbreak response. PACER is deployed within a health information network inside the institutional firewall so that the healthcare organization can control queries going to the EHR via the API. PACER runs inside of a docker-compose environment. This means that each microservice is enclosed in its own “container environment.” By default, each container can only see other boxes within its network. Only white-listed traffic is allowed into each container, as defined in the docker-compose.yaml file.

In Phase 1 using CQL-based FHIR queries, public health can collect information on demographics and comorbidities of interest (eg, age, sex, diabetes, immunosuppressive treatment or disorders, COPD, asthma, etc.) for those cases confirmed by a positive laboratory report to the public health department from clinical or commercial labs.

In Phase 2, PACER could collect information on available healthcare resources to detect gaps between the requirements of a geographic area and the demand based on the projections above. For this phase of work, public health standards communities such as HL7 as well as EHR vendors will need to expand adoptions of standard around healthcare resource, bed management, and supplies.

DISCUSSION

Public health has, in the past, tended to create and, when offered the opportunity, impose its own standards on the clinical health system. This approach has often resulted in incomplete reporting and noncompliance by health systems. The requirements for EHR systems to make health data available to patients and other healthcare providers through FHIR resources allow public health to create a new kind of data system that eases the heavy lifting on interoperability required of the health system. However, because public health queries typically do not benefit the care of a single patient, but rather serve the needs of populations, it is important for such queries to respect the minimum required information for surveillance. PACER provides an approach to implement this standard in a collaborative way that can evolve without modification to the internal coding of electronic health systems.

FUNDING

The development and conceptualization of PACER was funded by CDC under contract D8954 with Georgia Tech Research Institute.

AUTHOR CONTRIBUTIONS

Each author contributed equally to this article.

The findings and conclusions in this report are of the authors alone and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

CONFLICT OF INTEREST STATEMENT

None declared.