Temple Health Takes Maternal Care More Than a Baby Step Further Via Underutilized Data

A patient in Temple Health‘s system missed two prenatal appointments. The reminder letter reached her on schedule, cited the right clinical guidelines, and went straight into a stack of mail she never opened. Healthcare has spent three decades getting the letter right: accurate, timely, evidence-based, generated automatically by systems that took a decade and a data governance program to build correctly. However, it has spent almost none of that effort on why she never opened it.

That type of gap exists in every industry, not just healthcare. An analytics model’s output is usually a pretty pie chart or dashing dashboard, rarely an important intervention. Moreover, data and analytics answer what is true but they rarely answer what changes because it is true. That second question is where much more investment is needed.

Deborah Cancilla, Temple’s Executive Vice President of Data Strategy and Chief Information Officer, pinpointed the challenge, “An accurate analytic model that nobody acts on is just an expensive way of being right after the fact,” she says. “My job isn’t finished when the data is correct. It’s finished when a specific patient, doctor, or administrator does something different because of it.”

Deborah Cancilla, EVP, Data Strategy and CIO, Temple Health

Combining Traditional Analytics and Influencer Methods

Healthcare analytics and influencer platforms are not competing for the same prize, which is precisely why comparing them is useful, and integrating them is crucial.

A dashboard is just telemetry. It reports what the patient has done or not done, such as not making appointments, missing appointments, or otherwise drifting off a care plan. What it rarely does is directly influence action. Influencer platforms, however, as crude as their underlying data often is, close that second loop by design: it acts on attention rather than merely measuring it. Ideally this happens immediately, at the individual level. Healthcare has built extraordinary instruments for the first loop and almost none for the second, Cancilla observes.

Instrumenting Behavioral Change in Healthcare

Closing that second loop as something closer to a personal mandate for Cancilla than just another data and analytics project. LENA, the maternity engagement app is her baby. She conceived it and her data organization gave birth to it. “It’s one of my passion projects, and one I pushed hard for even when no obvious budget existed for it yet,” she says. “There’s value in our data that is regularly unrealized. Delivering this fourth level of care is just one way we’ve been able to drive additional value from our data.”

The name is also the model. LENA stands for Learn, Engage, Nourish, and Achieve, the four pillars Temple built the app’s care logic around: 

• Learn delivers trusted medical guidance calibrated to the patient’s exact week of pregnancy. 
• Engage keeps her connected to her actual care team through appointment reminders and symptom logging rather than generic prompts. 
• Nourish delivers encouragement toward and rewards the specific healthy habits shown to help both mother and baby. 
• Achieve ties clinical milestones, a completed lab draw, a sonogram, a postpartum visit, to real-world rewards, so the app’s incentive structure lines up with the same events a physician is already tracking on the chart.

This patent-pending app is now part of Temple Health’s MAMA Model initiative designed to combat the area’s maternal mortality crisis. Severe maternal morbidity across Pennsylvania rose roughly 40% between 2016 and 2022. Philadelphia County ranks among the state’s highest-burden counties for complications after birth, and black patients experience them at 2.6 times the rate of white patients. Temple’s executive director for the Women & Families campus, Sharon Kurfuerst, has described the broader strategy in similar terms: rather than build first and invite the community in afterward, Temple ran listening sessions before construction began and designed the facility, and the programs inside it, around what it heard. Temple reports 85% enrollment among eligible LENA patients, alongside measurable reductions in emergency room visits and stronger adherence to the postpartum follow-up visit, a checkpoint OB physicians treat as one of the more reliable predictors of longer-term maternal health.

LENA works because of what sits underneath it: identity resolution that recognizes the same patient across a community health worker’s visit, an OB appointment, and an app notification, a data model current enough to update the next nudge as the clinical picture changes, and governance tight enough that a personalized reminder can be trusted with the same clinical data as the chart itself. Then is translates all this into rewarding women to achieve milestones like office visits, testing, and education–the keys to successful engagement. 

Cancilla has described the real work behind LENA as less about building another patient-facing app and more about connecting three data streams (community health worker notes, EHR records, and app engagement) into a single view of where a patient actually stands relative to her own plan of care. “We didn’t set out to build a better reminder system,” she counters. “We set out to make sure the data already sitting in our systems could reach someone in time to matter. And it’s no surprise that narrative messaging outperforms fact-based communication for the majority of patients.”

Behavior as a New Enterprise Asset

Most health systems manage three asset classes deliberately: clinical assets, financial assets, and, increasingly, information assets under a governance program with an executive sponsor and a budget line. Almost none manage a fourth, quieter asset: the organization’s accumulated ability to guide a specific patient toward a beneficial outcome. Call it a behavioral asset. Like any other asset, it depreciates without maintenance, needs an owner, and produces a return only when someone governs it deliberately rather than assembling it campaign by campaign.

Building it requires tools most CIOs already run elsewhere in the organization: identity resolution that follows a patient across encounters and channels, behavioral segmentation that groups patients by readiness to act rather than by risk score alone, and next-best-action logic that turns a static care gap into a specific message delivered through a specific channel at a specific moment. None of this demands new technology. What it demands is treating behavior change as a data product in its own right, owned and measured with the same discipline already applied to a patient’s lab values or a hospital’s balance sheet.

The appointment letter that never got opened wasn’t a data failure or even an operational failure. It failed because nothing downstream of the data was built to notice it failed, and adjust. Every health system or organization in any industry that is sitting on a well-governed data asset and a stalled outcome faces the same opportunity. The data strategy of organizations, perhaps particularly those in healthcare, shouldn’t be judged on how precisely they can predict what a customer or patient will do. Rather the strategy should be judged, and therefore focused on, making change happen.