COVID-19 presented a once-in-a-century challenge to public-health officials around the world, and Mehnaz Mustafa was no exception. As executive director of Healthcare Quality and Informatics at the New Jersey Department of Health, Mehnaz (her preferred name) found herself on the front lines of the pandemic in 2020.
Over the course of the next two and a half years, she and her team engineered a fundamental shift in how New Jersey’s public-health data and analytics systems are designed, configured, and used. The information systems they created helped get “the right data to the right people at the right time,” says Mehnaz, to inform decisions and ultimately deploy a more effective COVID-19 response.
And Mehnaz and her team aren’t stopping there. Their tools are now the cornerstone of a new centralized public-health data and analytics hub that applies the lessons learned during the pandemic to other public-health goals, such as advancing health equity and preventing childhood lead poisoning.
Mehnaz recently sat down with Jessica Kahn, coleader of McKinsey’s US state Medicaid and Human Services data and technology work. The pair talked about New Jersey’s climb up the analytics innovation curve, how the tools created by Mehnaz’s team are contributing to healthier lives for New Jerseyans, and how the state is building resilience for future crises. An edited version of their conversation follows.
Jessica Kahn: Mehnaz, COVID-19 was such an overwhelming global crisis, and at the same time so local and personal in New Jersey. It required a massive response from your agency. How did COVID-19 speed up the pace of innovation, and how did it feel?
Mehnaz Mustafa: COVID-19 had a devastating impact on New Jersey, which, along with New York, was at the epicenter of the pandemic back in March and April of 2020. We had very little information coming to us from regions outside the United States, and at that point—even in countries that were hit before New Jersey—experience was light-years ahead of data.
We identified very quickly that we needed to be transparent with the residents of New Jersey about any information the Department of Health or the state had. So we created a public dashboard with information on cases, hospitalizations, and deaths mainly at the state level. Over time, we have gotten more and more granular, and for cases and deaths in particular now, the dashboard displays data at the county level.
We then got into vaccination efforts, starting off with statewide reporting. Now, you can go into the dashboard and look at vaccination coverage by different age bands for every municipality in the state. We also do a lot of reporting on outbreaks and vaccine coverage in long-term care, in schools, and in institutions of higher education.
We created a dashboard very early on so that everyone could see where we were with the disease progression and how it impacted their communities and themselves personally. We then moved forward, using some of that information to drive decisions that we were making internally for the pandemic response across vaccination, testing, contact tracing, and other response activities.
Jessica Kahn: How did you innovate in your use of data during the pandemic response, particularly in thinking about measuring equity?
Mehnaz Mustafa: I think we have traditionally relied on data that are collected by the department, and this has served us well, because it’s in response to the different programs we have in the department and the services we offer to residents in New Jersey.
But for some of the pandemic response, we also considered data that is external to the department. Census data was a big contributor, and then the CDC’s [US Centers for Disease Control and Prevention] social vulnerability index, because we were trying to solve for issues that most New Jerseyans face without forgetting that there are some communities that have always experienced barriers—particularly barriers to access.
That enabled us to drill down to the municipality and sometimes zip-code level, depending on what we were responding to. For some of the work around vaccinations, we knew that there were issues with some communities accessing vaccination sites. But without data, it was hard to know where those locations were. We did not have vaccine sufficiency at the beginning of the vaccination program. So we asked: Are we allocating resources, particularly vaccines, to the communities where there are true gaps? When we’re setting up sites, are we setting them up in the right places? Not just the megasites that most individuals in New Jersey could access but smaller, community-based vaccination and testing sites. What about the areas where there is low car ownership? Can those community members get to those megasites?
We used a lot of geospatial analysis, and these analyses were routinely updated so we could tell where the gaps were and act accordingly. If we embarked on a certain path, and the very next day or within the week the data told us that we’re probably going down the wrong path, we could course correct. That was remarkable in terms of impact and how resources were allocated. And data was used in every step of the way in the decision making.
Jessica Kahn: Decision making is really the end goal. How did you enable your leadership, your teams, your program colleagues to take good decisions backed by data?
Mehnaz Mustafa: For the pandemic response and especially for the vaccination response, it was important for us to share data in as granular a level as possible with the different partners we have at the local level, so that they could do their work. And we’ve shared those reports with local health officers, with acute care hospitals, with federally qualified health centers, and with a bunch of community and faith-based organizations. Because they are our boots on the ground. And it would not serve any of us if we were asking them to act without any information. It’s really about getting the right data to the right people at the right time.
It’s really about getting the right data to the right people at the right time.
Jessica Kahn: Even getting hold of the right data was a huge challenge for many organizations during the pandemic. Different systems had to talk to each other for the first time that hadn’t previously interacted. Those systems and data sets can be like islands. How did you build bridges between them, fast, to create an infrastructure so that data can inform decisions in real time?
Mehnaz Mustafa: I think the biggest issue we identified very quickly was the need to centralize some of the data and data reporting. There are data that are collected by different parts of the department and different parts of the state. In the pandemic, we had to combine and collate all that information. That’s what created the Department of Health Centralized Data and Analytics Hub. We’ve been growing the work of the hub ever since.
Jessica Kahn: What I find really inspiring and exciting about New Jersey’s journey is how you’re now pivoting to build on that capability and apply it to other areas of public health. I would love to hear more about this.
Mehnaz Mustafa: We are hoping to replicate some of the work done during COVID-19 for other parts of the department. One of the activities or responsibilities of the hub is to pull together lessons learned from the pandemic response and see how we can apply it to things like childhood lead-poisoning prevention, or HIV surveillance, or prevention of maternal and infant mortality.
We have made some huge leaps, through the work of the hub, in how data can be linked between one program and another, both of which are trying to serve the same population—that is, the population of New Jersey.
Jessica Kahn: Building on what you did with the local partners for COVID-19 response, how might that be an example of maintaining those relationships and sharing those kinds of data reports moving forward across other public-health topics?
Mehnaz Mustafa: We’re looking first and foremost at the relationships we built. What the hub has done is create a bridge between people who are solely focused on data analysis and people who are then expected to use that data to drive the programs—the people who visualize data and their intended audience.
For every analysis, we created multiple versions of the same report so that the person the data was shared with understood it is based on their needs. And every day we are identifying other topics and other issues in the department that could benefit from similar reporting structures and from leveraging those exact same relationships. The local health departments that we worked with for the pandemic are the same local health departments we would have conversations with when we’re talking about childhood lead. They’re the ones who are doing the lead inspections and responding to an elevated blood-lead-level case.
I think expanding on some of those built relationships is going to be crucial as we move forward.
Jessica Kahn: One of your next efforts on the horizon is to think about more self-service tools so that those local partners that you mentioned, and even other programs, can help themselves to some of the data analysis and insights that your team is going to create—as opposed to now, where you’re just generating reports and sharing them.
Mehnaz Mustafa: Absolutely. When I think about all the different visualizations we’ve done—for every analysis there was an attachment of 50 different visualizations, because you’re talking to different groups of people. Everyone looks at data and understands data differently. And for every stakeholder group that we engaged with, that same analysis might have been visualized differently. We’ve now learned how to do visualizations for different stakeholder groups.
So the reusability factor of the things that we have done, and done well, is important. Whether we’re talking about building resiliency in long-term care or preventing childhood lead poisoning or preventing maternal and infant deaths—all of those can benefit from some of the data-visualization lessons we’ve learned through the pandemic response.
Many of those stakeholders will remain our stakeholders for other types of work that the department now has to embark upon, and we’re trying to always keep those end users in mind. Part of the job that the hub has been doing is to think about that end user, because that’s where impact happens. And if the data is not presented to the right individuals using the right technology or the right process, you’re sort of dead in the water at the beginning.
It’s all about the intersection of people, processes, and technology. That’s why one of the responsibilities of the hub is to help people upskill. This includes anyone who needs to rely on data to make informed decisions, because how data is analyzed and visualized is ever-changing. We’ve put one cohort through the learning journey, and we’re hoping to do it now across the department.
It’s all about the intersection of people, processes, and technology.
Jessica Kahn: What are your top three goals for the hub moving forward?
Mehnaz Mustafa: I think first and foremost—consistency, transparency, and interoperability.
The biggest contribution of the hub has been connecting data across different programs, making it transparent, and doing it in a consistent manner so that data is treated across the department in a similar fashion.
Building team capabilities across the department in data and analytics is another big function of the hub, and we plan to continue to build on that by delivering dynamic reports to stakeholders. The dynamic aspect of it is particularly important, because all data is not actionable data. And data is ever-changing, ever-evolving, and every time that happens, it’s important to share that with relevant stakeholders.
Governance: the hub cannot actually function without a partnership with the programs that collect that data. It’s their data. They’re the ones responsible for it on behalf of New Jersey residents—whose data it is. And that’s where data privacy and security come in.
And lastly, a really close relationship with our health information technology department. The platform is so important for data to be dynamic and be shared in a dynamic format. So those would be the biggest pulls and functions of the data hub.
