In late 2024, our first real-world pilots involved a small group of early-access clinics in the Pacific Northwest. Six months later, we asked care coordinators and practice managers at each site a direct question: what had actually changed? We focused on their daily work and patient interactions, rather than dashboards or expectations.
The responses prompted us to revise our view of automated outreach in small clinical practices. Here is what we found, including results that did not fit our original assumptions.
Our starting expectations
We began with a simple hypothesis: sequenced automated outreach after visits would improve follow-up completion, limit care gaps, and reduce reactive calls when patients returned with complications from unresolved issues.
Each outcome occurred to some extent. The underlying process differed from our expectation, and the secondary effects were greater than anticipated.
The completion rate shift
Within our early-access pilot cohort, more post-visit follow-up tasks were completed within 14 days after automated outreach was introduced. Results differed by practice and visit type, but the aggregate change was meaningful. Medication refill confirmations moved less, since patients were already motivated to complete them. Specialist referral confirmations and lab order completions showed a substantially larger change.
The unexpected mechanism was not simply that messages led patients to act, although some did. More of the improvement came from earlier identification of non-responses. Care coordinators received a clear, time-stamped list for personal follow-up calls. Rather than searching full task lists, they began each morning with a focused queue of unresolved follow-ups.
Automation did not replace coordinator work. It gave that work structure.
How coordinator work changed
This finding surprised us most. We had expected total coordinator workload to fall as automation took the first contact. Instead, the work was redistributed.
The total number of outreach interactions assigned to coordinators did not decrease significantly. Their mix did change. Calls to patients who had responded to an automated message took less time, while reactive calls to patients who had not responded to any outreach increased slightly. Those patients were now visible in a structured queue instead of quietly aging out of the task list.
Coordinators said their work felt more purposeful. Their calls focused on patients who needed personal attention, rather than combining routine confirmations with serious non-responses. At a 5-provider family practice, one coordinator called it the difference between "doing outreach" and "doing follow-up." One means contacting everyone. The other means concentrating on those who need you.
The patient experience signal
Patient experience was not an intended pilot measure. We lacked a baseline, and linking such changes to one operational intervention is difficult. Still, coordinators consistently described how patients responded to the messages.
Coordinators repeatedly observed that patients, especially those managing chronic conditions, shared information in automated check-ins that they had not volunteered by phone. After a 7-day post-visit text asked whether a cardiology referral had been scheduled, a patient might reply, "no, I had trouble understanding the discharge paperwork" or "I called and they said the wait is 12 weeks, is that normal?"
This information is clinically useful but otherwise difficult to capture. Patients may not call about a question or concern because it seems too minor for a nurse, or because a business-hours phone tree is difficult to navigate while working. A text with a direct reply option lowers that barrier significantly.
What did not change
Some initial assumptions were unsupported. We had expected fewer urgent same-day callbacks and walk-in visits related to post-visit complications. During the 6-month pilot window, the signal was too unclear for confident conclusions. Whether automated follow-up reduces acute care utilization remains a longer-term question requiring larger cohorts and longer observation than our early pilots provided.
Automation also did not meaningfully improve engagement among the hardest-to-reach patients. Those with no reliable phone, unstable housing, or no SMS opt-in were not reached by automated outreach. Coordinators still needed to contact them directly, often through community health worker partnerships or in-person outreach during scheduled visits. Automated outreach cannot address structural access barriers. It addresses a coordination gap for reachable patients who are not receiving reliable follow-up.
Our six-month assessment
In a small clinic, automated follow-up changes workflow in defined ways. It makes incomplete follow-up tasks more visible, gives coordinators a structured queue, and offers patients a low-friction way to share information they might not report by phone. It also handles routine first contact that coordinators once completed manually for every patient, whether or not personal attention was needed.
It does not replace clinical judgment, remove access barriers, or ensure that hard-to-reach patients are found. A practice expecting automation to manage coordination without a trained coordinator will be disappointed. Adding it to a capable coordinator's workflow, with better information and a focused queue, can produce real operational improvements.
That claim is narrower than the one we may have made when we began building CTPathways. We believe it is also the honest one.
Ready to close care gaps at your practice?
Join our early-access program. Setup takes one afternoon.