The Challenge

2630 followers

How can we harness data and information for the health of communities? read the brief

Entry

Camden Health Explorer

Leveraging hospital data with insights from the Obama campaign, we’ll develop an interactive dashboard with real-time healthcare enrollment, cost, and outcomes metrics for use by providers, policymakers, journalists, researchers, and the public.
Each year, hospital care costs for Camden, New Jersey’s 77,344 residents reach more than $100 million.  Of this amount, a full 30 percent of costs stem from just 1 percent of patients.  Many of these patients cycle through the hospital EDs and inpatient wards, arrive at emergency rooms regularly for easily treatable conditions, or seek treatment for advanced conditions that could have been prevented. This pattern is not unique to Camden but is common nationwide.  For patients, families, and communities, it comes at a very high personal and financial cost.

To address this problem, the Camden Coalition of Healthcare Providers (CCHP) developed a citywide Health Information Exchange (HIE) that facilitates communication and data sharing between the three hospitals and the outpatient community in Camden. At the heart of the Camden Coalition’s work is its focus on “hotspotting”—the identification of patients who are heavy users of the healthcare system in a timely manner, providing targeted interventions to address their needs and change their utilization patterns. Each day, the Coalition sees a census of who is currently hospitalized, and Coalition staff begin to reach out to high utilizing patients to coordinate the critical period following discharge and the transition back to primary care. The Camden Coalition also works with community groups to provide preventive care and health education.  

In the past, we have relied on developing our own data tools and solutions because we found the available data and analytics tools inadequate and expensive. As the Coalition’s data becomes more diverse and complex, however, we are facing growing hurdles in uniting disparate data sets for rapid deployment and creating tools for better access and analysis.

In order to provide stakeholders with real-time data, help community actors assume a greater role in this collaborative effort, and continue to improve the hotspotting system, the Camden Coalition is partnering with BlueLabs—an analytics, data, and technology company formed by senior leaders of the Obama for America analytics team.  Together, we plan to create the Camden Health Explorer: a detailed, anonymized, and real-time health metrics dashboard about healthcare enrollment, utilization, costs, and outcomes in Camden. BlueLabs’ deep data-integration and predictive modeling skills are ideally suited to take the hotspotting project to the next level.  Funding from the Knight foundation would be crucial to our ability to standardize and enhance the functionality of this ongoing effort.


Metrics and ACO Progress

We will develop and release anonymous health metrics broken into patient segments, location, and time series. Data will be published in three phases, with each phase concentrating on data that facilitates better-informed decisionmaking within a distinct health policy area.  The first phase will focus on providing metrics that find hotspots in healthcare utilization and enrollment, supporting tactical decisions for programs that coordinate care and enroll patients. The second phase will dig deeper into data about the insurance status of current patients, allowing enrollment organizations to focus their resources on the populations that will benefit most from outreach. The third phase will leverage labs data for health organizations to identify populations where additional chronic condition testing and treatment is needed, supporting community needs assessments and exploration into areas ripe for health care programming.

As the Camden Medicaid provider community transitions to an Accountable Care Organization (ACO) model, it will be critical to maintain tactical intelligence about the change in health outcomes against historical baseline health metrics, both as a topline for the entire Camden community and for subgroups within it. We will contribute to the process of building sound statistical risk baselines, and as those baselines are rolled out, the relevant Camden health metrics will be published as relative gain/loss against the baseline. This will help members of the ACO make appropriate data-informed course adjustments. By presenting this data publicly and in real time, we will help develop a true culture of accountability.

Camden Health Explorer will draw on a diverse group of advisors and collaborators to guide metric-development and dashboard visualization, including Dr. Joseph Doyle, a MIT Health Economist; Dr. Joel Cantor of Rutgers Center for State Health Policy; Dan Gorenstein of NPR Marketplace’s Health Desk; and Edward Livingston, Executive Director of Camden Churches Organized for People.


Dashboard Interface

The dashboard interface will be designed to highlight health metrics in a way that facilitates data discovery and better tactical decisions. It will present the health metrics on a map, in time series, and as crosstabs, allowing users to find a variety of data stories using the best visualization methods to answer their question.

Data Topics

Data will be sourced from CCHP’s individual-level HIE data and citywide hospital claims database. This data will be aggregated and published in granular buckets by time series, neighborhood, insurance status, diagnosis history, language spoken, and socioeconomic indicators. Based on the potential for programmatic impact, we will release data in three topic area phases:


1. Utilization and Hotspotting

The Camden Coalition is known nationally for using data to identify and target super-utilizers and developing effective clinical interventions that help super-utilizers better manage their diseases and connect to effective primary care. The proposed project will allow us to publish data that provides more insight about health care utilization in Camden with an emphasis on costs, admissions, readmissions, high-utilizers, and patients who visit multiple emergency departments. For example, using hospital claims data, we will build cost curves for each patient segment, allowing CCHP and its partners to target resources towards patients with the steepest cost trajectories.


2. Enrollment and the Affordable Care Act

As enrollment begins in the health insurance exchanges and Medicaid providers switch to the Accountable Care Organization (ACO) payment model, we will begin by publishing aggregate data profiles on the insurance status of admitted patients. This will help navigators and other enrollment groups to more efficiently target communities that would see immediate benefit from being enrolled in Medicaid or coverage through the exchange.


3. Outcomes

For the last half century, lab results have been a cornerstone of patient care decisions, and HIEs provide an opportunity to utilize this quantitative data stream at the community level. Based on lab data, we will develop and present on metrics that indicate the successful management of chronic diseases, especially diabetes and HIV. This will help to highlight any disparities in disease management and early detection, and allow efficient allocation of community resources.


Impact beyond Camden

One of the core values of the Camden Coalition is its commitment to being completely open-source, sharing everything we’ve learned to free data and get it into the hands of those who need it to inform and guide their work. The hotspotting message resonates with diverse audiences across the nation. From our work here in Camden all the way up to the federal-level, we have encountered increasing demand for tools that better activate data. Funding from the Knight foundation would be crucial to our ability to standardize and enhance the functionality of our ongoing efforts.

It is important to us that communities beyond Camden will benefit from our Health Explorer.  The dashboard application and the data transformations necessary to get from individual-level patient and claims data to actionable metrics will be open sourced, allowing other communities to use our project with their datasets or to adopt the common indicators we develop.
Finally, even without our code being adopted by other communities, by publishing metrics based on HIE data, we will highlight super-utilization and effective interventions, encourage ACOs to adopt transparent goalposts, and demonstrate the opportunity to free data from operational databases in order to better the health of communities.


Project Advisors

Our project advisors represent each type of user that we anticipate benefitting from this project. They will help us to define metrics that will have the most impact on their work, and a dashboard interface that is flexible enough to support their needs. They will also help publicise the tool through their networks, and encourage other communities to replicate our successes.


Journalist Advisor: NPR Marketplace’s Dan Gorenstein

Dan Gorenstein is the senior reporter for Marketplace’s Health Desk, covering the business of healthcare. Dan is currently spending 18 months embedded in Camden covering the rollout of the Affordable Care Act.


Academic Advisor: MIT Health Economist Joseph Doyle (PhD)

Joseph Doyle (PhD, University of Chicago) is the Erwin H. Schell Professor of Management and a Professor of Applied Economics at the MIT Sloan School of Management. Joe’s focus is on conducting empirical research using microeconomic theory and modern empirical methods with large scale datasets and natural experiments to answer policy questions, particularly around health provider quality.


Academic Advisor: University of Pennsylvania Social Policy Professor Dennis Culhane (PhD)

Dennis Culhane (PhD, Boston College) is the Dana and Andrew Stone Professor of Social Policy at the University of Pennsylvania and Co-Director of the Actionable Intelligence for Social Policy initiative, a MacArthur-funded project to promote the development of integrated database systems by state and local governments for policy analysis and systems reform. Dennis is nationally recognized for his research around homelessness and housing, and is a leader in the integration of administrative data for research.


Policy Advisor: Joel Cantor (ScD), Director of Rutgers Center for State Health Policy

Joel C. Cantor (Sc.D., Johns Hopkins University) is the Director of the Center for State Health Policy, the Distinguished Professor of Public Policy at Rutgers University, and frequently serves as an advisor on health policy matters to New Jersey state government. Dr. Cantor’s research focuses on issues of health care coverage, financing, and delivery.


Public Advisor: Edward Livingston, Executive Director of Camden Churches Organized for People

Ed Livingston is the Executive Director of Camden Churches Organized for People (CCOP), a PICO-member, faith-based community organization comprised of congregations working together to transform conditions for individuals and families in the city of Camden. CCOP brings individuals from different racial and ethnic backgrounds, religious traditions, and neighborhoods together to address the serious concerns facing families in Camden including healthcare, public safety, housing, economic development, and unemployment.
 

Answers to Knight followup Questions

1.       Who is working on the project? Who are your partners?

Camden Health Explorer will be created through a partnership between the Camden Coalition of Healthcare Providers (CCHP) and BlueLabs.  The Camden Coalition is based in Camden, NJ and works with healthcare providers, hospitals, and physician practices to improve the care and coordination of healthcare for Camden residents while decreasing costs. BlueLabs is an analytics, data, and technology company formed by senior members of the Obama for America analytics team. It works with a diverse range of nonprofits to harness the power of data and have a positive impact in the world.

 

2.       How do you know there is demand for this project?

From the City of Camden to the federal-level, we have encountered demand for hotspotting and health exploration tools.  We have built relationships with stakeholders interested in partnering in efforts to reduce hospitalizations, increase uptake of Medicaid expansion, and promote primary care in the community.  Locally, there is demand from community health organizations, policy advocates, hospital and health system executives, accountable care organizations, journalists and quality improvement specialists.  We at the Camden Coalition need to effectively mash up the disparate data so that we can provide a more holistic picture of the health of our community and assess the impact of our ongoing efforts.

 

3.       How is your project different from what already exists?

There are several community-wide health data portals that provide a variety of health metrics collected from Departments of Health. Additionally, most HIE software has reporting tools built in for high level administrators to view confidential health metrics. However, we believe this would be the first comprehensive open dataset directly linked to HIE and claims data, providing actionable metrics to all stakeholders in near real time.


4.       How will the data or information you use or create be made open?

The primary goal of this project is to take data previously locked away in narrow silos, enhance it with metrics to help inform decision making, and make it more accessible to a broad audience that includes journalists, researchers, policymakers, and the public itself. This project will help render obsolete the outdated model of sending data queries to IT department and waiting weeks (if permission is even obtained) for an answer that invariably misinterprets the question. Camden Health Explorer will serve as a core platform that can be developed and modified for future applications as new data sources and different policy questions arise, presenting unique opportunities for crowdsourcing down the road.


5.       What will you make or do in this project?

We will roll up individual level HIE and medical claims data into open anonymized aggregates by patient segment, geography, and topic area (enrollment, hotspots, and outcomes). We will publish that data to an interactive data exploration tool that allows users to see that data on a map and in charts, graphs, and crosstabs.  The metrics themselves, and all of the data aggregation and display code will be open source.
 

6.       How can others learn from/build on what you do?

The Camden Coalition’s hotspotting approach serves as a model nationwide for innovative approaches to improving care delivery, and Camden Health Explorer will continue to expand upon the organization’s open source values. Converting private HIE and claims data into metrics that can be used to better target healthcare enrollment, measure consumption and quality of healthcare delivery, and provide surveillance of key outcomes at the community-level would represent a unique and unparalleled mechanism for using medical data to enable data-informed decision making. With more and more communities across the country approaching the Camden Coalition for technical assistance, this project with the Knight Foundation will help build out cutting edge analytical tools that can be replicated and adopted by others.

 

7.       How much do you think it will cost?

We estimate that the project will cost $432,000. This includes software development for rolling up metrics, data deidentification, API development, and frontend development, and analyst time to make sure we are presenting the most informative metrics rolled up in the best ways. This is also includes budget for external consulting on visual design and data privacy, and a modest budget for CCHP staff to provide support to this project.

  

8.       How would you use News Challenge funds?

The Knight Foundation funding would be used to provide BlueLabs staff software development and analytics time, allow CCHP to spend time focusing on outwardly facing metrics development, and fund the hiring of an outside designer and a data privacy consultant.

 

Describe your project in one sentence.
Camden Health Metrics Explorer will provide granular real-time metrics about health and healthcare in the Camden community, empowering policymakers and care providers to effectively deploy community-level resources.
Who is the audience for this project? How does it meet their needs?
The audience for Camden Health Metrics Explorer will include the Camden Coalition of Healthcare Providers’ clinical teams and organizational staff, as well as its broad member network composed of hospitals, primary care offices, social service organizations, and public representation. Through real-time metrics and dashboards, these tools will allow the Camden community as well as the funders of the innovation work being done in Camden (Centers for Medicare and Medicaid Services, private foundations, health insurers, state government) to effectively assess impact at the community level as the Camden Coalition seeks to improve quality and bend the cost curve.
What does success look like?
Success for Camden Health Metrics Explorer will be the development and visualization of relevant, high-quality community-level metrics that can be used by a broad array of stakeholders, including care teams, community organizations, policy makers, and ultimately the public, to better inform the work being done on the ground in Camden. As the Camden Coalition becomes a citywide ACO, success for this project will pivot on Camden Health Metrics Explorer providing real-time tracking of health care consumption as Camden looks to improve quality and bend the cost curve.
Your Location
Camden, New Jersey is a city of nine square miles located along the Delaware River, directly across from Philadelphia. Camden is perennially designated among the poorest and most dangerous cities in the United States. While negative headlines continue to dominate how Camden is portrayed, the work of the Camden Coalition of Healthcare Providers and its network of community partners have made the city uniquely poised to be the first in the country to bend the cost curve.

Comments

Join the conversation and post a comment.

Jean-Ezra Yeung

September 25, 2013, 19:01PM
Erek, good to see your team competing in this challenge and well done for making it this far. I think the visualizations are an important tool but can you team think of specific use cases that would not only impact medical-decision making, but take it one step further to reduce cost, utilization and/or improve quality? If you could illustrate that with the capabilities of your tool, I think that would help your team hit a home run.

Yvette Nuñez

September 22, 2013, 17:23PM
Given the great need in Camden, this data could be used to impact policies beyond health. Good luck.

Jeremy Albisser

September 22, 2013, 12:12PM
We have developed a system that provides these types of indicator mapping and visualization tools to (Accountable Care) Integrated Health Service Providers in Ontario. We would be very interested in partnering with a few US Accountable Care organizations to localize our data transformations and develop a Metrics explorer specifically for US ACOs.

Rebekah Monson

September 22, 2013, 09:58AM
The amount of data utilized and real-time aspects of this project would be amazing, and I can see this tool making a big and fast impact in public health policy. I'd love a little more information on how the data might be used by the community as mentioned in the Outcomes section.

Erek Dyskant

September 22, 2013, 13:10PM
Immediately, data we publish will be used by care providers, community health workers, and navigators to locate communities where their efforts will have the most impact, both at an absolute scale and, where relevant, relative to ACO risk profiles. For example, if flu-related Emergency Department visits spike among individuals in a certain community, health workers could target those areas for vaccination. Or, if a provider sees that re-admission rates are especially high among individuals with COPD and heart disease, they can recommend those individuals for post-discharge support.

Erek Dyskant

September 22, 2013, 13:10PM
Labs-based metrics, like those described in the Outcomes section provide a uniquely quantitive view of diagnosis and care at the community level, which can be used to identify treatment gaps and target community-level interventions. Targeted interventions could include data-driven deployment of mobile screening resources, or increased engagement of individuals who are members of an identified at-risk community and living with a specific condition. This part of the project is inspired by the Grant Colfax's work on mapping HIV viral load in San Francisco (http://www.princeton.edu/cbli/student-projects-1/PPT-Colfax-HIV-Prevention-and-Mapping-Community-Viral-Load.pdf).

Jeremy Albisser

September 27, 2013, 09:33AM
Actually lab-based metrics have little to do with most of community care (a dozen or so of the 200+ different community care functional areas would have a lab component) - in order to manage patients in the community, standardized assessment, follow-up and discharge tools are typically used for the bulk of community care for outcomes measurement.

Megan Campbell Smith

September 22, 2013, 00:24AM
Do I understand correctly that this project aims to convert private HIE data (owned by ACO) into public data that reports, in real time, the efficacy of quality care delivery? For real?! You'd really disclose cost and outcomes for every case?! That would be AMAZING!

StuartAdam Eisenstadt

September 22, 2013, 02:42AM
I would like to have clarification on that also on those very same topics; actually doing all of this in real time (or close to) would be very impressive.

Megan Campbell Smith

September 22, 2013, 10:22AM
Furthermore, IMHO, this kind of HIE project is the only real scientific work I'm seeming on the quality care front. You know, observation for observation's sake. I've seen a lot of ACOs, they are all fascinating businesses, but none truly exists until the data prove it. I'd like to see a data design/journalist partner in there to study meaningful ways to represent that data.

Erek Dyskant

September 22, 2013, 10:47AM
Thanks for your enthusiasm! Yes, that is absolutely our goal. Metrics about enrollment, utilization, and labs will be reported in near real-time. Claims are reconciled after care, so costs will lag based on the hospitals' claims processing and reporting schedules. In order to maintain patient privacy, metrics will be reported at the detailed aggregated level.

Erek Dyskant

October 02, 2013, 17:04PM
Thanks for your suggestion of including a journalist partner! We've added an advisory group including Dan Gorenstein, Health Journalist at NPR's Marketplace, who will help us develop metrics and visualizations that are helpful for reporting on health and healthcare in Camden.

Arijit Sengupta

September 21, 2013, 21:57PM
Could you find a way to provide actionable insights to the patients themselves so that we could enable a bottoms up disruptive innovation approach? While improving the visualization of information available to decision-makers is definitely valuable, given all the politics involved, it may be easier to transform healthcare by driving informed consumption. What do you think?

Erek Dyskant

September 22, 2013, 11:51AM
A large audience of this project is people who are already having daily 1:1 contact with Camden residents and patients. Success includes those individuals using this tool to make data-informed decisions about the work that they are doing, within discretion they're currently given. For example, a navigator could see that enrollment is especially low among Spanish-speaking individuals in a particular neighborhood and reach out to members of that community about getting enrolled.

Even without policy outcomes that involve "politics", using data generated from patient care to inform community health initiatives will have a positive and significant impact.
Login
Close
Login to News Challenge
 
or