The Healthcare Analytics Summit is back! Join us live in Salt Lake City, Sept. 13-15. Register Now
Since March, daily, weekly, and monthly data reports have highlighted the devastating effects of COVID-19 on our nation, state by state. Data is central to the public and healthcare ecosystem’s understanding and management of not only coronavirus, but of all future novel viruses. Our ability to quickly and accurately localize, aggregate, segment, and incorporate data into action is greater now than at any time in our history.
At all stages of the pandemic, real-time analytics have impacted decision making – from a primary care system identifying their most vulnerable patients, to a hospital system adjusting their operations to meet care demands or the return of elective procedures, and researchers evaluating infection and transmission rates.
As the pandemic hit, providers knew that the elderly and those most vulnerable because of certain pre-existing conditions were at greatest risk. OneCare Vermont sought to proactively identify those patients who were at the highest risk of serious illness or mortality from COVID-19, and to determine which of its patients with chronic healthcare problems needed care for medical issues other than COVID-19. However, OneCare Vermont’s EMR did not have a mechanism to systematically identify patients with known high-risk factors.
OneCare Vermont turned to Health Catalyst’s Data Operating System platform, which leveraged publicly-available demographics and social data, medical and pharmacy claims and known medical data to filter patients by risk category. With this rapid identification of at-risk patients, providers and care teams used the DOS risk-stratification care coordination tool to conduct patient outreach, including telephone calls and telemedicine virtual visits, ensuring patients receive needed social support and medical care during the pandemic.
Read the full results of OneCare Vermont’s DOS project.
Real-time data played a huge role in Providence St. Joseph Health’s coronavirus response. Ari Robicsek, Chief Medical Analytics Officer at Providence St. Joseph Health, saw an opportunity to enhance the organization’s data architecture while responding to the pandemic in real-time. With 51 hospitals and 1,085 clinics, the analytics team quickly mobilized to prepare and synthesize information about COVID-19 patients throughout their seven-state system.
“During the course of the first six months of the pandemic, the analytics needs of our organization changed rapidly and unpredictably, sometimes from one week to the next,” says Robicsek
Providence created an epidemiological registry tracking data, including patient demographics, relevant diagnoses and clinical features, and the care location where the patient was seen. This allowed them to track the patient journey and visualize the change in syndromic activity by geography.
“The data infrastructure that we developed, including a COVID patient registry, enabled us to flexibly respond to the evolving analytical needs. This doesn’t always mean we were perfect. We made – and learned from – various mistakes, including initially poor predictions of hospital bed demand.”
Providence also launched MyCovidDiary, a clinical research project that uses technology to collect and analyze first-person COVID-19 accounts from thousands of individuals. Using crowdsourcing and Natural Language Processing (NLP), Providence hopes to accelerate the medical identification, understanding, and treatment of this novel disease.
“The pandemic has forced us to upskill our team, especially in natural language processing and geographical information systems, and we’ve developed tools that we think will help our physicians and patients in the future,” Robicsek says.
Controlling the virus’ spread requires answering several data-driven questions, not least of which is whether individuals who previously contracted COVID-19 might be at risk of reinfection. To address the possibility of reinfection, Health Catalyst led a collaborative initiative to analyze on a monthly basis real-time reverse transcription polymerase chain reaction (RT-PCR) data.
The COVID-19 HealthCare Coalition EHR and analytics organizations workgroup completed early analysis using a federated approach, including data from 56,871 patients from EHR and analytics organizations among Health Catalyst COVID-19 National Registry. Based on data distribution trends, it appears the presence of COVID-19 can range from several days to several months. To determine the possibility of reinfection, it is imperative to understand COVID testing effectiveness and best-practices.
During the coming months, as the sample size for the analysis grows, it will be critical to continue to compare the percentage of people exhibiting long-term virus shedding, while taking into consideration the typical false positive rate for SARS-CoV-2 real-time RT-PCR. This will determine if there is evidence of reinfection and its frequency. Health Catalyst is currently completing the next sequence of studies to understand sensitivity and specificity of testing and to determine best practices for testing across its network.
Forward-thinking, real-time data and, more importantly, streamlined data analytics will continue to play a major role in the health care community’s response to coronavirus.
“Technology enables us to accomplish a lot more than we can imagine,” says Chris Hutchins, VP, Chief Data and Analytics Officer for Northwell Health, a recent addition to the Health Catalyst family, who is leading a breakout session at this year’s virtual Healthcare Analytics Summit (HAS20) about Northwell’s experience navigating the coronavirus crisis.
Hutchins recently described the “extraordinary measures [taken] to prepare the health system for many scenarios that involve disruption to health system operations.”
Predictive technology, like Health Catalyst’s COVID Capacity Planning Tool, are critical to ensuring a system’s preparedness. Built on the outstanding Penn Med epidemic model – with additional features, the tool forecasts infections based upon actual county level data and dynamic infection spread rates (Empirical Model) and allows managing, using, and saving scenarios.
Forecasting local COVID-19 demand in the context of a local system capacity sets expectations and informs mitigation strategies, including manage, use, and save scenarios, and bed and ventilator capacity.
“It’s clear from the bedside to the board room, that high-quality data – coupled with high caliber analytics – provides both the evidence we need to change and the confidence to persevere in difficult times,” adds Jason Jones, Chief Data Scientist with Health Catalyst. “COVID-19 has caused unfortunate morbidity and mortality. We hope silver linings persist and include broader gratitude, acceptance of alternate modalities of care, and driving decisions with transparent analytics persist.”
Register for HAS20 to learn more about data’s central role in the fight against COVID in these special breakout sessions: