JMIR Publications

ABSTRACT

Background:

As healthcare delivery shifts toward value-based care (VBC), proactive strategies to close preventive care gaps are essential. However, patient engagement remains suboptimal due to logistical, behavioral, and socioeconomic barriers. Traditional outreach methods such as phone calls, emails, and postal mail have long been used, but emerging digital approaches, such as chatbot-based messaging, offer potential advantages in scalability and personalization. Their comparative effectiveness, however, remains underexplored.

Objective:

To evaluate the effectiveness of chatbot outreach compared with traditional communication methods (phone, email, mail, and multichannel) in promoting compliance with and across multiple preventive care domains.

Methods:

Design, Setting, and Participants: This retrospective study evaluated patient outreach campaigns conducted from 2021 to 2023 across an integrated health system in New York. The final analytic sample included 52,364 care gaps from patients eligible for preventive screenings under Healthcare Effectiveness Data Information Set (HEDIS) and Centers for Medicare & Medicaid Services (CMS) criteria. Outreach modalities included automated chatbot Short Message Service (SMS), non-automated phone calls, and organization-led email or mail campaigns. Participant data was enriched with social vulnerability scores to account for community-level disadvantages. Exposure: Outreach method (chatbot, phone, email, mail, or multichannel). The assignment was based on engagement history and operational protocols. Main Outcomes and Measures: The primary outcome was care gap closure or compliance with identified measure gap annually. Logistic regression and chi-square analyses examined associations between outreach method, patient demographics, primary care physician (PCP) relationship, social vulnerability index (SVI), and compliance.

Results:

Phone outreach consistently achieved higher compliance than chatbot or multichannel outreach across most groups and years. Chatbot messages outperformed phone calls only in diabetes care in 2023 (OR=1.81; 95% CI, 1.48–2.21; P < .001). PCP continuity remained a strong predictor of gap closure, especially in primary care (ORs ranged 1.36–2.61; P < .001). Higher SVI quartiles were associated with lower compliance in blood pressure, cancer care, and diabetes care groups. In contrast, primary care outcomes showed higher odds of compliance in the third quartile of SVI, contradicting the typical linear-deprivation narrative. Compared to their counterparts, females, Hispanic or Latino individuals, and Asian patients demonstrated higher odds of compliance in some groups and years.

Conclusions:

Outreach modality is an important, modifiable factor in preventive care adherence. While phone-based outreach remains the most effective overall approach, chatbot-based strategies may have targeted applications, such as in diabetes care, particularly among digitally engaged populations. Segmented, equity-informed outreach strategies that integrate technology, patient preferences, and primary care continuity are essential to achieving high-impact, scalable outcomes in VBC settings.