St. Louis-based Ascension wanted to test whether sending patients a personalized text message before a primary care appointment could improve care. The health system started with seven practices, where patients received automated messages from their physicians with personalized care plans pulled from the EHR.
Clinicians said the approach was working, but Ascension wanted more evidence before scaling it. The health system expanded what it calls a proof of concept into a randomized trial across 76 primary care practices in four states, with practices assigned to receive the intervention or continue with usual care.
“What you hear is really important, but what people actually do is most important,” Mitesh Patel, MD, vice president and chief clinical transformation officer at Ascension, told Becker’s.
Dr. Patel has led more than 25 clinical trials, and he now helps decide what gets built out across a system with 91 wholly owned or consolidated hospitals in 36 states and the District of Columbia. He is one of four leaders directing Ascension’s Clinical Innovation Institute. The text-message program, whose results were published online in NEJM Evidence in May 2025, is one example of how Ascension tests new care models before taking them systemwide.
Dr. Patel spoke with Becker’s about what those models have to demonstrate before Ascension scales them and where scaling efforts can fall short.
The Ascension Clinical Innovation Institute generally calls its tests proofs of concept rather than pilots because the goal is to determine early whether an idea will work and should be scaled, Dr. Patel said. If it does not, the system can adjust or change the approach before expanding it.
The process starts with defining the problem, rather than choosing a tool. Dr. Patel said that could mean helping patients access care faster or closer to home, improving the care delivered or extending care into the community.
Ascension then works with clinical, operational, mission, digital and technology teams to determine the process, outcome and experience measures it will use to evaluate the program.
For an idea to move beyond that stage, it has to work clinically, operationally and financially, Dr. Patel said. The system considers whether the model improves patient care, what it requires of clinicians and staff and whether it can be financially sustained as it grows.
Technology can make a program easier to scale, but Dr. Patel said Ascension also considers what will happen when a program becomes part of normal operations. A proof of concept can receive significant attention during its first few months that will not necessarily continue once it is rolled out more broadly.
That is why the system seeks input beyond the people leading an implementation. With the pre-visit text program, Ascension spoke with front desk staff about whether patients were arriving with questions about the messages. Technology teams also provided feedback about the consent process required for patients to receive them.
The trial did not meet its primary outcome of increasing care gaps addressed on the day of the visit, according to the study. Patients who received the messages were more likely to close care gaps within 90 days, an adjusted difference of 5.4 percentage points, and post hoc analyses found an increase in completed appointments and decreases in no-shows and cancellations.
Ascension also used the results to adjust the messages. Each message had been limited to three care gaps drawn from a smaller group of eligible ones. Dr. Patel said the team found that patients received an average of one or two and could manage three without becoming overwhelmed, so Ascension expanded the types of care gaps that could be included while keeping the limit at three.
The pre-visit program is now live across Ascension’s primary care practices. The health system has also expanded the approach to other points in a patient’s care, including after a hospital discharge and after childbirth.
A postpartum program launched last year sends patients personalized messages with a link and phone number to call if they have issues during their first week home. About 10,000 patients receive the messages, and 90% remain enrolled through the full 30 days, according to Dr. Patel.
The broader goal is to create what he calls a digital care plan that can be personalized and delivered at different points in a patient’s care.
Ascension has applied the same approach to preventive care more broadly. A case study published this month in NEJM Catalyst Innovations in Care Delivery reported that a digital outreach campaign run from September 2022 through September 2024 reached 179,711 Medicare patients across 12 states and led to 73,253 appointments scheduled within 14 days of first outreach, a 25.9% scheduling rate.
No single measure determines whether a program will scale, Dr. Patel said. Clinical outcomes, operational burden and financial sustainability are weighed together, with metrics agreed on by leaders and frontline clinicians before a program begins.
For example, a program that improves patient outcomes but requires clinicians and staff to perform significant manual work would be difficult to expand across the health system.
Ascension has also focused on virtual care because it can be scaled across the organization, Dr. Patel said. The system is about halfway through its markets in rolling out same-day and next-day virtual care. Patients with urgent, low-acuity issues such as a headache, cough or congestion can be offered a virtual visit alongside an in-person option.
When scaling digitally enabled care does break down, Dr. Patel said the problem is not always the technology.
“A lot of people, when they think about scaling digitally enabled care, think a lot about the technology,” he said. “The technology is of course important. We need to make sure it is scalable and embedded within workflows. But one of the biggest things we often overlook, and that we have tried to really focus on, is stakeholder alignment and understanding what the real problem is, using unique ways to do that.”
Ascension is testing one way to get that input through an innovation tournament that asks front-line clinicians, staff and nurses to identify problems that could be addressed with an innovative tool. Dr. Patel said the effort has generated hundreds of ideas and helps Ascension identify employees who could be involved in implementing them.
Front-line input also shapes how the tools are designed. When Ascension developed patient-reported outcomes for knee and hip replacement surgery, orthopedic surgeons reviewed mockups before they were added to the EHR.
One surgeon preferred viewing the information as a chart while another preferred a table, so the team added the ability to switch between the two. Another surgeon wanted to give the information to patients, prompting the team to add a print function that creates a PDF.
Dr. Patel said Ascension continues to seek feedback after a tool launches to identify additional changes.
AI is accelerating the development of new technology, but Dr. Patel said it has not changed Ascension’s focus on first identifying the problem it is trying to solve.
He sees opportunities for AI to support clinicians by searching and summarizing information in the medical record. Ascension is also considering how advances in AI-assisted software development could affect the way new products are developed or sourced.
Those uses will require safeguards and clinician oversight, he said.
“Over time, we have to think about how we do this responsibly, putting the right checks and guardrails in place and making sure we have a clinician in the loop who can confirm the feedback,” Dr. Patel said.
Patient feedback is another part of Ascension’s process. The system surveys patients after they complete programs and conducts listening sessions through research protocols.
“Their input is just as important, if not more important, than that of clinicians and staff,” Dr. Patel said. “It really involves understanding the patient journey and how our tools are making an impact on those communities.”
