Harnessing Data for a Healthier Future:

Background

The United States boasts one of the most technologically advanced healthcare ecosystems in the world. Unfortunately, optimal use of these advances has been blunted by a delivery system that is fraught with barriers and inefficiencies. Despite the volume and ubiquity of population and consumer health data, we are challenged with making this data accessible, interoperable, and usable.1 The U.S. spends more of its GDP on healthcare (17.1%) than any other developed nation, nearly double that of the next highest spender² but cannot boast improved health outcomes. Some providers feel that this is due to a combination of factors, including challenges posed by lack of interoperability. We have an illness-based model of healthcare delivery, where reactive care for 50% of the population generates 86% of the total cost in the form of chronic illness management, end-of-life care, and acuity.³

Our ability to manage COVID-19, both from a prevention and a treatment of acuity standpoint, could have been an opportunity to shine, but it exposed some of these kinks in the armor even more so. While many harsh realities were brought to the forefront with the COVID-19 pandemic, we had experienced many of the same challenges, to a lesser degree with the H1N1 pandemic in 2009. These two recent experiences challenged state, local and federal government organizations and moreover, brought emergency departments and hospitals to their knees and impacted the resolve of medical providers on the front line. Throughout the pandemic, providers experienced shortage of diagnostic tests at testing centers to meet the population need and production and supply chain challenges with vaccines. Findings from an Association for Molecular Pathology survey found that more than 70% of U.S. clinical laboratories suffered significant delays related to COVID-19 testing programs.4 These delays were critical to the outcomes of individuals, communities, states and our nation with respect to the pandemic.

Healthcare professionals were challenged with identifying risk factors for severe disease and identifying impending death in those most sick. Sharing information on what treatments worked in different care settings across the nation evolved to this day, it has been a relative trickle compared to the possibilities afforded by the digital age. Healthcare systems failed to harness the power of multi-sourced, multi-formatted raw clinical data because of barriers imposed by standards of nonconformity, data siloes, and care fragmentation. As a result, providers found themselves rationing scarce medical resources to ensure that care for those in rural, urban, and suburban communities was delivered equitably.

During the H1N1 pandemic, Dr. Pinho was running an adult and pediatric medical practice in an upper middle-class neighborhood, which had been designated a state drop site for vaccinations and was fortunate to have received a surplus of vaccines for his patients’ needs. His colleague, an infectious disease doctor, was running a pediatric HIV practice in an urban teaching hospital and hadn’t received a single vaccine for a more at-risk population. The COVID-19 pandemic is studded with similar stories of social, racial and ethnic inequity. In this white paper, we identify three major challenges and opportunities for improvement in order to make future pandemic responses more efficient, fair, and based on solid evidence. The focus of this paper will be to explore how healthcare organizations can tackle challenges associated with data access and accuracy, overcome barriers to data interoperability, and integrate data silos to drive preventive measures, improve outcomes, and promote health equity. One of the most notable preventative tools in the healthcare system’s arsenal is the utilization of high-quality real-world data for efficient and effective decision-making.

Overcoming Data Capture Challenges

Clinical data sourced from electronic health records (EHRs), labs, and health information exchanges (HIEs) contain unique and timely details on patients, such as vital signs, test results, diagnoses, immunizations, and more. However, the non-conformant coding, absence of key normalization, and poor data interoperability have obscured a real time, data-driven, longitudinal view of population health. Lack of interoperability has been particularly devastating for at-risk populations and data shows that infection, hospitalization, complications, and mortality varied significantly by race and ethnicity. This was exacerbated by poor data exchange between episodes of care.

Digital health technology is transforming the way we approach healthcare, and in particular, the adoption of systems to improve data quality and longitudinal health records has the potential to revolutionize patient care and health outcomes. Over the last two years, Availity has been collaborating with the California Department of Public health (CDPH) to identify and improve issues related to data quality, timeliness, and accuracy concerning public health reporting. Availity Fusion is our API, cloud-based technology engine that automates clinical data integration, normalization, and enrichment in real time and at scale through a comprehensive five-step process known as Upcycling Data.™

As part of that collaboration, our team of clinicians and informaticians performed an internal study on a yearlong cohort of 128 million lab reports that showed, among other findings, that race and ethnicity information was not documented in approximately 45% of cases. Through the implementation of the California Physician Reporting System (CPRS), this issue was addressed by identifying instances of poor data capture and returning to the original source for demographic data improvement.

In the pie chart below, one can see that race and ethnicity are “other or unknown” in 45% as reported by laboratories. Census reporting of “other” race is 7%. Census reports of Caucasian race were 46% higher (41% of the population in the laboratory capture and 60% in census capture). As compared to Caucasians, Asians and Blacks are more likely to not have their race documented and are disproportionately represented in the “other or unknown” group. Their actual population numbers are 167% and 96% greater than those reported by labs.

Furthermore, our internal study identified that through outreach to source data providers an increase in the completeness of demographic information by 10% for all demographic variables and an estimated 25% across race and ethnicity categories was achieved. This study emphasized the importance of this information in the context of the COVID-19 pandemic, but also as a critical piece of any modernization strategy at the state or federal government level. Understanding the distribution of the disease across different populations is crucial for effective response and intervention.

While we have made progress in sub-categorizing data by demographic factors, such as race and ethnicity for the COVID-19 pandemic, it is important to also consider workstreams across other reportable diseases like influenza, RSV, and meningitis in order to gain a more comprehensive understanding of the population’s overall health status. Integrating comprehensive vaccination data with clinical documentation from medical providers, hospital settings, and self-testing for all patients is the necessary next step.

Enhancing Patient Care and Public Health Outcomes through Accurate and Standardized Results in Patient Records

The ubiquity of self-administered home antigen tests made it difficult for state and federal agencies to capture and analyze results. Test numbers from outpatient and inpatient labs and wastewater analysis was largely the driver of these insights. The majority of at-home COVID-19 test results were not recorded in patient medical records. However, some at-home tests, such as the BD Veritor™ and the Lucira™ brands, allow for better reporting of results than simple patient interpretation by appending results to a valid LOINC code. This provides the opportunity for these at-home results to be recorded andshared in a standardized manner. While data-sharing and privacy issues will continue to be a perennial challenge, opportunities exist to help augment lab capabilities and public health data in times of increased need.

Pharmacies and states can partner with testing companies through state-level programs, allowing for direct access to results of tests performed by patients within a pharmacy setting or at home. The utilization of point-of-care tests featuring encoded standards that can be deployed in Fast Healthcare Interoperability Resources (FHIR®), in conjunction with the sharing of data via API enabled by the 21st Century Cures Act, facilitates the seamless transfer of information from patients to hospital systems, as well as between health plans and providers, thereby mitigating deficiencies in the continuity of care.

Solving for test result capture, both in the lab or at the point-of-care, is an important first step. Challenges around the limitations of paper-based vaccination records and cards also need to be addressed. One major issue is accessibility – paper records are often not easily accessible to healthcare providers, especially those outside the facility where the vaccinations were administered. Moreover, vaccine registries don’t always connect well to provider offices to relay results. As a practicing physician, Dr. Pinho describes a great information exchange with the vaccine registry. With respect to reciprocal receipt of data, those patients that received their flu shots at a pharmacy, while registry communication likely happened, there wasn’t seamless integration with his practice. Moreover, Dr. Pinho relates that his primary care doctor who is part of a large multispecialty group, did not receive information on his COVID-19 vaccination and relied on his report. Dr. Pinho’s current longitudinal record includes registry information for his county administered vaccine series and a duplicate entry of his self-report to his provider.

One can see how this is a substantial data gap, making it difficult for providers to make informed decisions about patient care. Another issue is perpetuation of data – paper records can become lost, damaged or otherwise unreadable and patients can forget to self-report or dismiss it as irrelevant to their care. Beyond individual care impact it can ultimately influence the effectiveness of vaccination programs and the ability to track the spread of disease and its prevention in a timely fashion.

In contrast, the implementation of electronic transactions of data for vaccination records through FHIR resources has improved interoperability. States including, but not limited to, California, Kentucky, Massachusetts, and New Mexico that have rolled out “COVID-19 SMART Cards,” a digital record of a person’s COVID-19 vaccine history, are able to easily track and monitor the vaccination progress of their residents and share this information with providers and institutions. With the deployment of SMART Cards, nearly 500 million people globally now have digital access to their vaccine records.8 The integration of electronic transactions for the sharing of vaccine data among providers has the potential to enhance real-time data sharing as it pertains to this challenge.

The COVID-19 pandemic has emphasized the importance of enhanced interconnectedness and collaboration among all stakeholders in the healthcare system, such as providers, labs, hospitals, pharmacies, and registries, to ensure maximum benefits for the larger population. Through the electronic sharing of vaccine records we can accurately assess vaccination rates and investigate the connection between vaccines and breakthrough infections in the community. Furthermore, through the integration of provider data, it is possible to assess the probability of hospitalization and mortality in individuals who have tested positive for COVID-19 and have underlying health conditions, as well as measure vaccine efficacy. However, due to the scattered nature of existing data and difficulties with data quality, there are limitations in analyzing these relationships without a holistic solution that focuses on semantic interoperability and comprehensive data sharing.

Moving from Data Silos to Seamless Interoperability

The healthcare industry has been grappling with multiple data silos, which have been a hindrance to effective data analysis and decision-making. With respect to the COVID-19 pandemic, three major sources of clinical data have largely been siloed: COVID-19 lab data, vaccination data, and provider encounter data. Data from these silos viewed in isolation only provides a limited understanding of one part of a patient’s health status and history. For example, COVID-19 lab data may only relay information on a patient’s positive or negative test result. How that information ties into vaccination status and ultimately outcomes, in the case of COVID-19 positivity is not discernable. Similarly, vaccine data might only offer details about the quantity and type of vaccines and booster shots a person has received, but it doesn’t capture positive test results or the severity of disease.

However, when these data silos are shared and integrated, they can paint a much more comprehensive picture of a patient’s health status and history. For example, if COVID-19 lab data is combined with vaccine data, healthcare providers can better understand the efficacy of the vaccines in preventing COVID-19. If provider data is also included, healthcare providers can gain a deeper understanding of the patient’s overall health status and develop more targeted and effective care plans when they do have positive tests. Breaking down clinical data silos is essential for healthcare organizations to improve care coordination, reduce care gaps, and provide better patient care. To achieve this, healthcare data consumers can start by investing in an integrated technology infrastructure that promotes seamless data sharing and interoperability. For public health, this infrastructure should facilitate the collection, analysis, and interoperability of high-quality data from various sources, including vaccine registries, labs, and provider-generated medical charts.