Verifying Eligibility for Lung Cancer Screening via TiMeToAct Text Message Intervention

Lung cancer screening (LCS) with low-dose CT (LDCT) reduces lung cancer mortality but suffers from low uptake, ranging from 16.4% to 19.6% in 2022.^1,2 Barriers to LCS adoption include missing or inaccurate smoking status data in electronic health records (EHRs), perpetuating inefficiencies in identifying screen-eligible individuals.³ Addressing inaccurate smoking history in EHRs is critical to increasing LCS referral and uptake. Automated bidirectional text messaging is a promising tool for health care delivery enhancement^4,5 that may facilitate identifying screen-eligible individuals; however, to our knowledge, its potential to promote LCS has yet to be investigated. We aimed to evaluate the feasibility of using text messages to confirm pack-year and duration of smoking history when unavailable in the EHR to identify LCS-eligible individuals.

Methods

The primary focus of the TiMeToAct (Text Messaging To Advance Lung Cancer Screening) intervention was to measure response rates to bidirectional text messages to verify LCS-eligibility using US Preventive Services Task Force 2021 criteria: age 50 to 80 years, with a 20-pack-year smoking history, currently smoking or quit within 15 years.⁶ This initiative was undertaken as a quality improvement initiative (QII) from May 2023 to August 2023. It was exempt from the University of Rochester Medical Center Institutional Review Board oversight based on guidelines for defining human participant research. Using EHRs, the Clinical and Translational Science Institute identified 63,527 individuals aged 50 to 80 years documented as having a smoking history in the 12 months before the QII. From this group, the research coordinator randomly selected 200 individuals, consisting of 163 White, 37 Black, and 40 Hispanic participants. Individuals received 5 text messages to confirm their pack year and smoking duration history, and LCS eligibility. Text messages were sent in English or Spanish based on the individual’s preferred language in the EHR. We developed culturally and linguistically inclusive text messages by translating them into Spanish, excluding cancer from the messages because the word can evoke fear and pain in some cultures,⁷ and creating an easy-to-understand questionnaire. An introductory message was followed by affirmative responses, which prompted 3 additional messages to confirm smoking history and eligibility for lung cancer screening (Fig 1). After the first round of text messages, 4 responders had already undergone LCS, after which a rolling enrollment using an LDCT Current Procedural Terminology code excluded individuals who had undergone LDCT in the prior 12 months. If interested, LCS-eligible individuals were offered referral to the centralized LCS program, which conducts shared decision-making. The LCS coordinator contacted individuals to schedule an LCS visit. Individuals who declined were encouraged to discuss LCS with their primary care physician, and those who had undergone LDCT were encouraged to continue annual LCS. Individuals who responded Stop after the first message received no further text messages.

Figure 1.

TiMeToAct text messaging intervention messages.

Results

Of the 200 individuals receiving text messages, 105 (52%) were female, 163 (81%) were White (including Hispanic and non-Hispanic White), and 37 (18.5%) identified as Black. Thirty-seven (18.5%) White and Black individuals identified as Hispanic. Most (97.5%) preferred English and had their smoking status as currently smoking (97%). Table 1 illustrates the demographic comparison between any responders (89 of 200, 45%), including those responding Yes to complete eligibility (34 of 89, 38%) or Stop after the first message (55 of 89, 62%) and nonresponders, those who did not respond to the first text messages (111 of 200, 55%). Comparing the rates of those who responded Yes to nonresponders, they were higher among female individuals, Black individuals, and those younger than 70 years. Rates of nonresponders were higher among male individuals, White individuals, Hispanic individuals, and those older than 70 years. Table 2 illustrates the demographic characteristics of those who responded Yes and eligible individuals for LCS. Of 34 (17%) responders, 25 (12.5%) completed the eligibility assessment questions; 9 (26%) were eligible for LCS, and 16 (48%) did not meet US Preventive Services Task Force eligibility criteria because of a < 20 pack-year history or having quit > 15 years ago.

Table 1.

Comparison Between Responders and Nonresponders

Characteristics	Responders 89 of 200		Nonresponders 111 of 200
Sex	‘YES’ (34 of 89)	‘STOP’ (55 of 89)
Male	14 (41%)	27 (50%)	55 (50%)
Female	20 (59%)	28 (50%)	56 (50%)
Age, y
50-60	16 (47%)	29 (53%)	54 (49%)
61-70	17 (50%)	17 (31%)	39 (35%)
71-80	1 (3%)	9 (16%)	18 (16%)
Race
Whitea	25 (73%)	46 (84%)	92 (82%)
Black	9 (27%)	9 (16%)	19 (17%)
Ethnicity
Hispanicb	6 (18%)	8 (14%)	25 (23%)
Not Hispanic	28 (82%)	47 (86%)	86 (77%)

^aIncludes Hispanic White individuals.

^bIncludes Hispanic White and Black individuals.

Table 2.

Demographic Characteristics of Responders and Eligible LCS Individuals^a

Characteristics	No. (%)
Preferred language of the participants
•English	196 (98)
•Spanish	4 (2)
No. of participants who responded Yes to complete eligibility	34 (17)
•Eligible participants	9 (26)
•Ineligible participants	16 (48)
•Did not complete eligibility assessment	9 (26)
No. of participants who responded Stop after the first message	55 (27)
No. of participants who did not respond	111 (55)
Responders (Yes to complete eligibility, n = 34)
Sex
Males	14 (41)
•Eligible for LCS	4 (44)
Females	20 (58)
•Eligible for LCS	5 (56)
Race
White—non-Hispanic	19 (56)
•Eligible for LCS	6 (67)
White—Hispanic	6 (18)
•Eligible for LCS	1 (11)
Black	9 (28)
•Eligible for LCS	2 (22)
Age, y
50-60	16 (47)
•Eligible for LCS	5 (56)
61-70	17 (50)
•Eligible for LCS	3 (33)
71-80	1 (3)
•Eligible for LCS	1 (11)
Preferred language
English	34 (100)
•Eligible for LCS	9 (100)
Insurance status
Medicare	11 (32)
•Eligible for LCS	3 (33.3)
Medicaid	7 (21)
•Eligible for LCS	3 (33.3)
Private	15 (44)
•Eligible for LCS	3 (33.3)
No coverage	1 (3)
•Eligible for LCS	0

LCS = lung cancer screening.

^aOf 200 individuals (No.) between 50 and 80 years old, randomly selected to receive text messages.

Discussion

Our QII demonstrated the feasibility of bidirectional text messaging to mitigate a significant barrier to LCS: difficulty identifying screen-eligible individuals via EHRs because of inadequate or missing smoking history. Text messaging facilitated the referral of LCS-eligible individuals who may not have been referred otherwise to the LCS program. Furthermore, text messaging provided an opportunity to remind individuals who had enrolled in LCS to continue annual screening. We had an appropriate representation of Black and Latino individuals. Only about one-third replied STOP to stop receiving text messages despite concerns that health care-related texts may be bothersome to individuals. In addition to the difficulty obtaining accurate smoking history in the EHR, lack of referral to LCS, social stigma related to smoking, low health literacy, and social determinants of health are barriers to LCS implementation and uptake.⁸ Integrating digital medicine into traditional health care systems is a cost-effective and scalable tool with the potential to overcome these barriers and revolutionize how medicine is practiced.⁹ Particularly, digital tools allow us to connect with remote and socioeconomically disadvantaged populations, improve the doctor-patient relationship and shared decision-making, and ultimately, improve clinic attendance and routine follow-up. Text messaging has been widely used across the cancer spectrum, from prevention and early detection to supportive care.^10,11 This technology has been shown to improve screening for many types of cancer but has yet to be studied in LCS.⁷ The potential for text messages to encourage LCS uptake may be even more significant among hard-to-reach, socioeconomically disadvantaged, and uninsured populations disproportionally impacted by cancer.^8,10,11

Male individuals, White individuals, Hispanic individuals, and those over 70 years of age appear less likely to engage with text messages, but these results should not be interpreted as “not being interested in LCS.” Familiarity with mobile technologies may explain lower engagement among older individuals, and we may have sent messages in English to Spanish-speaking individuals. Moreover, Hispanic persons perceive multiple barriers to LCS, which were not addressed by our text messages.⁷

Most participants identified as currently smoking. The first eligibility question asked whether the individual had smoked in the last 15 years. Although this should have captured individuals who formerly smoked but quit within 15 years, expanding the EHR search to include individuals who ever smoked may better identify those who formerly smoked who can receive text messages to assess LCS eligibility.

This QII provided a platform to improve LCS eligibility assessment, increase awareness of LCS availability and benefits, and remind those who had undergone LCS about the importance of annual adherence. However, substantial work is needed to improve the response rate. To engage a broader audience and promote social equity, we must explore further factors such as education, age, socioeconomic status, and health literacy.^8,12 Hence, future work will focus on enhancing our questionnaire to personalize the screening material, creating an even more patient-centered tool,¹² sending additional messages given the low rate of STOP requests, with a proposed recurrence cycle every 3 or 6 months. Reminders for annual LCS adherence also should be introduced, and widespread dissemination and connection to supplementary messages should be promoted to assist in navigating the complexities of LCS.

Our QII had several limitations. It was a single-arm pilot, and we had limited attempts at reaching individuals. There is also the potential for confounding results because of the day of the week the text messages were sent. Education level, health literacy, and digital health literacy were not analyzed.

This QII highlights an opportunity to study the impact of increased text messaging attempts to confirm LCS eligibility and refer individuals to LCS programs.

Funding/Support

The QII described in this publication was supported by the University of Rochester Clinical and Translational Science Institute award number UL1 TR002001 from the National Center for Advancing Translational Sciences of the National Institutes of Health.

Financial/Nonfinancial Disclosures

The authors have reported to CHEST the following: M. P. R. has NIH/NCI funding to study lung cancer screening. None declared (F. A. C., G. G., A. C.-I., F. C.-B., K. F., A. P. C., D. H. A.).