Digital Outreach Reaches Thousands for Lung Cancer Screening, but Gaps Persist

Lung cancer kills more people in the United States than any other cancer except a handful of others, and it remains the third most commonly diagnosed malignancy in the country. Low-dose computed tomography screening can catch tumors early and reduce mortality among people at high risk, yet the proportion of eligible adults who actually get scanned remains stubbornly low. One of the central bottlenecks is not the scan itself but the step before it: figuring out who qualifies and getting them to care. A new study published in the Journal of General Internal Medicine offers a detailed look at how well a fully automated, digital-first outreach program can accomplish that first critical step across a large, real-world population.

The research, led by Richa Bundy of Advocate Health and Wake Forest University School of Medicine together with colleagues at the University of North Carolina at Chapel Hill, is a secondary analysis of a pragmatic randomized trial conducted at two large academic health systems in North Carolina. The parent trial, registered as NCT04083859, previously demonstrated that a direct-to-patient digital health intervention delivered to patients in their own homes increased lung cancer screening rates. The new analysis asks a different and arguably more fundamental question: how many people does this kind of outreach actually reach, and who gets left behind?

The technical machinery behind the intervention is worth understanding. Patients aged 50 to 77 with a documented history of cigarette smoking received an invitation through their electronic patient portal to use a digital health program known as mPATH-Lung. The program walks users through a series of questions designed to determine whether they meet eligibility criteria for lung cancer screening, then connects those who are eligible with the next steps toward a low-dose CT scan. Because the entire workflow is automated, the approach can in principle be scaled to tens of thousands of patients without adding clinical staff, which is precisely what makes the reach question so important.

The numbers tell a story of impressive scale paired with sobering limits. Of the 26,909 individuals who received an invitation, 3,049, or about 11 percent, completed the lung cancer screening eligibility questions within 16 weeks. That figure represents thousands of people who might otherwise never have been systematically assessed for screening eligibility, since traditional approaches rely on busy clinicians remembering to check smoking history and calculate eligibility during brief visits. At the same time, an 11 percent completion rate means nearly nine out of ten invited patients did not engage, a gap that the authors treat as the central challenge for future iterations.

The engagement funnel breaks down in revealing ways. Within 16 weeks of invitation, 65.5 percent of portal recipients read the message, meaning roughly a third never even opened it. Of those who read the message, 17.1 percent visited the mPATH-Lung landing page, and a smaller fraction went on to complete the full eligibility assessment. Each step in this cascade sheds participants, a pattern familiar to anyone who has studied digital health adoption. The researchers also tested a practical fix: when patients could not be reached through the portal, the system sent text message invitations instead. This fallback expanded the total number of people completing eligibility questions by 604 participants, from 2,445 to 3,409, a meaningful gain achieved simply by adding a second communication channel.

Perhaps the most consequential findings concern who responded and who did not. Reach was higher among women, people aged 50 to 64, individuals with medium or high socioeconomic status, those with commercial insurance, and residents of urban areas. The racial and ethnic disparities were stark: white participants completed the eligibility questions at a rate of 12.8 percent, compared with 8.6 percent for Black participants and 8.0 percent for Hispanic participants. Because lung cancer burden and screening eligibility are not evenly distributed across these groups, a digital strategy that disproportionately engages wealthier, whiter, urban patients risks widening exactly the health disparities that screening programs are meant to narrow.

These patterns echo a broader literature on the digital divide in health care. National survey data show that patient portal use has remained low across the United States, and studies of electronic health tools have repeatedly found that sociodemographic factors such as age, income, education, and race influence who uses them. Older adults, who make up a large share of the lung cancer screening eligible population, are among the least likely to engage with digital health technology. Rural residents, who often face the greatest geographic barriers to accessing screening centers, were also less responsive in this study. The paradox is that digital outreach is attractive precisely because it is cheap and scalable, yet those same qualities can concentrate its benefits among people who are already easiest to reach.

The authors argue that the solution is not to abandon digital outreach but to layer it. Their own data show that adding text messaging to portal invitations meaningfully increased reach, and they suggest that future strategies should continue leveraging multiple communication channels, potentially with cellphone text messaging as a primary strategy rather than a fallback. Text messages reach nearly every mobile phone in the country regardless of whether the owner has activated a patient portal account, and Pew Research Center data confirm that mobile phone ownership is nearly universal across demographic groups. Combining automated digital tools with human navigation, community partnerships, and offline channels could preserve the scalability of the digital approach while closing some of the equity gaps it currently exposes.

There are also technical lessons embedded in the study design. The researchers measured reach using the Area Deprivation Index to characterize neighborhood disadvantage and rural-urban continuum codes to classify geography, and they handled missing data with multiple imputation by chained equations, a standard statistical technique for preserving sample size and reducing bias when information is incomplete. The pragmatic randomized design, conducted within routine clinical operations rather than idealized research conditions, means the findings reflect what health systems can actually expect if they deploy similar programs. That realism is a strength: an intervention that works only under laboratory conditions offers little help to the millions of eligible smokers who pass through American clinics each year.

The stakes of getting this right are enormous. Lung cancer screening with low-dose CT is one of the few interventions proven to reduce death from the leading cause of cancer mortality, yet national data indicate that only a small fraction of eligible adults have ever been scanned. Every percentage point of additional reach translates into thousands of high-risk individuals being assessed and, for those who qualify, connected to a test that can detect tumors while they are still curable. The North Carolina study demonstrates that automated digital outreach can move the needle at population scale, engaging a large and previously unassessed group of patients within weeks. It also demonstrates, with unusual clarity, that technology alone will not deliver equitable access. The next generation of screening programs will need to pair the efficiency of algorithms with the persistence of multiple channels and human support, so that the patients most likely to benefit from early detection are not the ones least likely to hear about it.

Subject of Research: Digital outreach strategies for identifying adults eligible for lung cancer screening

Article Title: Using Digital Outreach to Identify Individuals Eligible for Lung Cancer Screening: A Secondary Analysis of a Randomized Trial

Article References: Bundy, R., Snavely, A. C., Miller, D. P., Reuland, D. S., Ferrari, R. M., Bellinger, C., Randazzo, A., Stradtman, L., Wright, E., Brenner, A. T., & Dharod, A. (2026). Using Digital Outreach to Identify Individuals Eligible for Lung Cancer Screening: A Secondary Analysis of a Randomized Trial. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10895-5

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10895-5

Keywords: lung cancer screening, digital health, patient portal, health disparities, low-dose CT, randomized trial, text messaging outreach, digital divide, cancer prevention, health equity, mPATH, population screening