Community Covid Checkup

Leveraging our unique experience in community-based disaster medicine (https://pubmed.ncbi.nlm.nih.gov/26030400/) in Red Hook, Brooklyn, we activated a medical relief effort in the beginning of the COVID19 pandemic in New York. Emerging from this 600 person pilot program, we have now adapted our protocol into an app prototype, which includes a risk-stratification algorithm that triggers tailored recommendations for geography-specific and language-appropriate medical, mental health, and social work resources accessible regardless of insurance status. Our app further communicates the most up-to-date education regarding COVID19 symptoms and utility of testing. Importantly, our tool is being deployed both online, and being communicated by phone to tech-limited senior citizens who are in the highest risk category. We are an interdisciplinary team including programmers from Cornell Tech, doctors, scientists, the Red Hook Initiative community organization, New York City Council, and Health & Hospitals – the largest municipal hospital system in the USA.

New Yorkers, especially minorities, underinsured, and lacking a primary doctor, are sick or dying in their homes without support. The cause is both COVID19 itself, and decompensation of other chronic conditions in a segregated, destabilized healthcare system. The pattern of cases and fatalities in New York underscores the uneven distribution of this challenge. Communities in the Bronx, Queens, and Brooklyn were disproportionately affected. The reasons are clear: even before COVID19, these communities were chronically underinsured, lacked healthcare access, and had a high burden of poorly managed chronic medical conditions. During COVID19, this high concentration of medically fragile individuals, many of whom could not afford to stop working to isolate, were the basis of the highest fatality rate in the country. Future peaks of COVID19 cases will surely be sparked in these communities. 

The essential challenge is the disconnect between a symptomatic or concerned individual and clear information, medical advice, testing applicability, and local resources in a language-appropriate, free-of-charge, accessible format. Thus, our project provides an interactive data-driven tool to prevent illnesses from escalating in community settings, acting pre-hospital to unburden the healthcare system. It answers the question: what should I do right now, where I am, in my situation?

With Red Hook facing COVID19, a phone-based survey was deployed to assess for COVID19 symptoms, underlying chronic conditions, essential medications, and other indicators of medical fragility throughout the pandemic peak. An automated scoring algorithm was developed to flag at-risk individuals based on their responses. Individuals with a high-risk score received follow-up to coordinate medical, mental health, and case management. This community-centric intervention was designed to anticipate deteriorating medical conditions and preventing costly, fatal in-hospital emergencies. 

We have programmed a web app that automates our community response protocol, and facilitates expansion into other at-risk communities. Community Covid Checkup is a web application that integrates the screening survey, risk-stratification algorithm, and follow-up procedures into a simple web tool. The tool connects high-risk individuals to proven, local, language-appropriate services regardless of insurance – other online tools only make generic recommendations leaving the user without an action plan. Low-risk individuals will receive the best available up-to-date advice for how to take action in the future if their condition changes. Furthermore, information regarding the appropriateness of testing, with a recommendation utility linking to testing and case-tracing via the NYC Health & Hospitals (H&H) System, will be implemented as H&H deploys their new testing program.    

The pilot phase of our work just in Red Hook included over 8,000 phone calls, with hundreds screened, and over a hundred flagged for essential follow-up services. The number of individuals flagged as at-risk are consistent with estimates of COVID19 cases and associated complication rates. We expect further improvement of our web app functionality and broadening of scope into other communities will allow us to communicate with tens of thousands more individuals. Depending on the severity of the future COVID19 peaks, we would therefore expect to identify and support the key subset of future at-risk individuals, who may likewise number in the thousands. 

Importantly, beyond flagging and supporting currently at-risk individuals, our most far-reaching impact will be informing the majority healthy population about what to expect, convey essential up-to-date information about best-practice regarding COVID19, and alleviate uncertainty about what local resources are available.

The above interventions serve to put out fires before they start in our vulnerable communities, stopping the exponential rise of infections, and spike in fatalities.  

While a slew of stakeholders are working fervently to address the challenges posed by COVID-19, tech innovators have a crucial role to play. Solving the problem of uneven distribution of COVID-19’s impact in NYC, primarily among minority communities, is imperative for strengthening the collective resilience of New York. Through the deployment of a web-app that allows for free community-wide risk assessment, education, and resource referral, our project directly addresses the issues of slowing and tracking the spread of COVID-19 at a low cost, and solidifies a response infrastructure that will aid vulnerable communities in future health emergencies and pandemics.

Clear messaging and simple tools that help vulnerable individuals navigate the rapidly changing, complex COVID19 environment have been elusive. Lost in a noise of conflicting online information, media messaging, incoherent government programs, and disconnected healthcare resources is a simple, free, and adaptable mechanism to connect the best available information and resources to the communities at the local level.  

Community Covid Checkup has built such a tool informed by our unique and thorough experience, including with disaster medicine during Hurricane Sandy, and early activation to pilot a medical relief effort before, during, and after the COVID19 peak in Red Hook, Brooklyn. Our program facilitates both web-based and human-to-human contact to have the broadest scope. This free tool integrates screening and algorithmic risk-stratification to trigger reflex recommendations, based on an individual’s risk score. 

Crucially, and differentiating our tool from all other available tools (e.g. Apple, JHU), the recommendation function links individuals to hyper-local, proven, language-appropriate healthcare services that do not require insurance. Generic recommendations like “see a doctor” are not specific and actionable, introducing a barrier to care. Our data shows that many individuals do not have a primary doctor or insurance in Red Hook. Furthermore, more detailed recommendations concerning getting tested locally are often unclear.  

Our unique partnership between community-focused non-profits, clinicians, scientists, and data-driven resources has created a mechanism to keep a hand on the pulse of our most vulnerable communities. We are eager to share our successful effort in Red Hook with other communities.  

The core technology is a web-based application that presents the user with a survey, available in 5 languages. The 10 questions are based on iterations of the survey deployed during the initial peak of COVID19 in Red Hook, Brooklyn. Each question corresponds to a numerical weight, which scales with the importance of the question towards predicting an individual’s health risk. The answers to the questions are themselves also weighted, similarly corresponding to risk severity, and are multiplied by the question weight. Upon completion of the survey, the risk engine computes a score, which conveys the degree of urgency to the user. Furthermore, depending on the combination of answers a user chooses, one or more of 5 “bins” are highlighted. These bins link to a recommendation page, which presents actionable information appropriate for the user’s condition. If the individual is low-risk, then up-to-date educational material is conveyed, e.g. signs/symptoms to watch for.  Importantly, depending on what language and zip code the user enters, the recommendation pages communicate information specific to their location in the appropriate language. We have piloted this web app to be specific for recommendations local to Red Hook; however, it is easily adaptable to other locations and/or languages. We expect to further include a formal link to testing and contact-tracing services provided by Health & Hospitals NYC. Finally, anonymized data will be pooled for summary statistics to help the medical community follow emerging trends in future peaks.   

Community Covid Checkup activated in the early phase of COVID19 in New York, beginning in mid-March 2020. Our multifaceted team includes Cornell Tech, CIV:LAB, medical professionals including NYC Health & Hospitals, community organizations like the Red Hook Initiative, and the city council office of Carlos Menchaca. By mid-April, 8,306 calls were made to Red Hook residents. 590 individuals were screened and scored using our survey. Our data shows 10% of these individuals had COVID19 symptoms, with 56% having an at-risk pre-existing condition, 39% requiring essential medical therapies, and 12% lacking a primary care doctor. Ultimately 51 individuals received medical follow up, including direct calls from doctors, 2 were connected to emergency services, and 96 received mental health and/or social work follow up. Otherwise low-risk individuals who were screened on these calls received educational materials and references to help them navigate COVID19 conditions.

A prototype of the novel web app emerging from this effort, and additional information, is available here:

http://covid19-redhook.cornelltech.io/

Tackling COVID19 requires a fundamental shift in the way the healthcare system engages the pandemic. Thus far, the approach has been reactive, rather than proactive, where infectivity escalates uncontrolled in the community before landing at the emergency room door. We expect that future COVID19 peaks will present unique challenges, and we have specifically designed our tools and procedures to be adaptable and meet these challenges. As with our approach during the pilot phase of our project tackling the first peak of COVID19 in New York, the core of our mission is to engage multiple stakeholders collaboratively, developing our solution into a tool that empowers the community with resilience and preparedness. This addresses the fundamental threat of COVID19, which is that it first sparks in the community, then escalates into a devastating fire. Our solution is a simple, accessible, adaptable measure to put out the fire at its source.

Towards adaptation and collaboration, data from this effort will be submitted for peer review to the medical literature, with the further goal of medical providers using this application to provide guidance to their patients in these uncertain times. Feedback from users and the medical community will surely inform our efforts towards adapting to changing conditions. If communities in other locations outside New York become hot spots for COVID19, our effort may be best served focusing on tailoring our recommendation protocols to these locations. Given the adaptability of our web app, if new symptoms become associated with COVID19, we can adapt our screening tool in real-time. Funding will help maintain the activity of our current volunteers, but importantly, may attract the skills and resources of others who become interested in helping us develop our tools in the future. For example, other computer scientists, community organizations, or healthcare providers may provide valuable support to adapt our efforts. Already, we have been contacted by organizations in the Bronx and Brooklyn, with affiliations to Mount Sinai, Columbia, and Montefiore Medical Centers. 

By its simple, free, and easily scalable nature, our web app removes many barriers to deployment in at-risk communities. Anyone with an internet-enabled device can access the app, and for those who are not tech-enabled, we call registered phone numbers to run the survey and convey the information by phone. We have previously focused on Red Hook for our pilot, calling 8,306 residents, screening and scoring 590 individuals, with 51 receiving required medical follow up, 2 connected to emergency services, and 96 receiving mental health and/or social work follow up. The number of individuals flagged as at-risk are consistent with estimates of COVID19 cases and associated complicated illness. We expect further improvement of our web app functionality and broadening of scope into other communities will allow us to communicate with tens of thousands more individuals. Depending on the severity of the future COVID19 peaks, we would therefore expect to identify and support the key subset of future at-risk individuals, who may likewise number in the thousands.  We were previously limited to call-based surveying only, but now our web-based app packages all functions we were running locally into a tool accessible to anyone. The scope is only limited by how many communities want to create their own unique recommendation pages for hyper-local resources. We hope in the next year to have community partners in all 5 boroughs of New York, and in five years we hope communities will utilize the app’s functions nationwide. 

In the next year, our goal is to have communities in all 5 boroughs of New York utilize the app, with customized recommendation pages for local clinical resources, each translated into the appropriate languages. For example, more languages are spoken per geographic area in Queens than anywhere else in the world. Our Health & Hospitals partners at Queens Hospital, or at Maimonides in the Bronx, can help customize the clinical recommendations and languages for their locations. Furthermore, in the next year we will integrate recommendations for testing and contact-tracing services provided by H&H into the app. 

In the next five years, we hope to expand nationwide. We already have partners in New Jersey’s more rural communities who are likewise interested in using our app for outreach to their population. Crucially, as medical science advances our understanding of COVID19, we can modify the nature of the questions, answers, weights, and recommendations to adapt to new conditions and emerging trends.  

Thus far, all efforts have been volunteer and pro bono. In the next year, funding would support dedicated, continued development of the web app, and deployment in Red Hook and other communities, including with partners in Queens and the South Bronx, to anticipate future COVID19 peaks that occur during reopening and winter months. While all our volunteer efforts since early March have made a big impact, to make our impact sustained, increase in scale, and achieve higher levels of complexity, we will need to support the computer scientists, web administrators, and community volunteers financially – especially since many of them are currently struggling in this weak economy.  

In the next five years, funding would help recruit new individuals to the effort, with skill sets helping to facilitate optimization for our nation's diverse communities with variable resources and changing conditions.

Securing funding for our otherwise volunteer effort would effectively remove many of the barriers to our personnel continuing to tackle these challenges, allowing them to dedicate their time to necessary optimizations and scaling. Furthermore, funding would support continued phone-based deployment of this tool for those with limited internet or technical capabilities, many of whom are at-risk seniors most likely to be impacted by COVID19. For example, currently our phone calls are facilitated by HubDialer, a communication tool which requires a paid subscription.

2 fulltime developers, 3-4 part-time admin staffers, 20-person call center, 20 medical professionals

Our effort draws on the combined experience of community organizations, local government, clinicians, and medical and computer scientists. This includes Cornell Tech, CIV:LAB, medical professionals including NYC Health & Hospitals, the Red Hook Initiative community center, and the city council office of Carlos Menchaca. Such a partnership between academic and public stakeholders has maintained the mission of effective, data-driven, low cost community outreach that has proven to be effective. It is proven by experience, as this is not the first “perfect storm” weathered by Red Hook. In 2012, Hurricane Sandy devastated this community, with a massively disorganized and insufficient government response. In the wake of the storm, a large-scale community-led medical response effort emerged, resulting in hundreds of individuals being monitored and supported during this crisis, with disastrous outcomes and fatalities averted (https://pubmed.ncbi.nlm.nih.gov/26030400/). Finally, with a partner like Health & Hospitals – the largest municipal hospital system in the country – we have strong ties to a comprehensive clinical knowledge and resource base linking to our web app. 

We currently partner with Cornell Tech, CIV:LAB, medical professionals including NYC Health & Hospitals, the Red Hook Initiative community center, and the city council office of Carlos Menchaca. In the pilot phase of our project, team members had weekly digital meetings to discuss progress and overcome challenges in real-time. Future internal assessments will likewise occur. Regarding distinct milestones and reporting, we reassessed our progress on a monthly basis by generating summary statics for the following parameters:

- number of individuals initiating a survey

- number completing a screen

- number flagged as at-risk: high vs. medium vs. low-risk

- distribution of symptoms, risk factors, and resiliency profiles for those at-risk

- number of flagged at-risk individuals connected to follow up services

- languages most frequently utilized

- geotagging for emerging at-risk hot spot locations

As this was an initiative that grew out of the pandemic, we have been operating under our 501(c)(3) non-profit and therefore our business model to date is focused on leveraging government and philanthropic funding.

While we will not look at any resident/consumer based models, we will look to create a model leveraging other potential revenue streams such as digital marketing, data and analytics, and lead generation for medical services.

We need support and expertise to take this to the next level. To date this has been a community approach to solving our local problems but our solution can be used in neighborhoods around NYC and the country, creating the infrastructure to respond and communicate locally for future pandemics - especially those health related. After reviewing the Solve program, it was clear that there was clear synergy in goals and approach to health impact.

We aim to establish partnerships with stakeholders that represent the areas checked above. 

We would like to partner with Tech companies that could help continue to build the product and capacity; health care providers and community based organizations that can help the product scale; academic institutions to help fully analyze data; additional institutions who might have need to leverage the infrastructure to provide access and other resources to communities in need, especially those in underserved communities.

Our project directly addresses the criteria listen in The Elevate Prize for Health Security, making us eligible/qualified.

Our project directly addresses the criteria listen in The People's Prize, making us eligible/qualified.