Chatbot-interfaced and cognitive-affective barrier-driven messages to improve colposcopy adherence after abnormal Pap test results in underserved urban women: A feasibility pilot study

Abstract

Challenges in ensuring adherence to colposcopy and follow-up recommendations, particularly within underserved communities, hinder the delivery of appropriate care. Informed by our established evidence-based program, we sought to assess the feasibility and acceptability of a novel cognitive-affective intervention delivered through a Chatbot interface, aimed to enhance colposcopy adherence within an urban inner-city population. We developed the evidence-based intervention, CervixChat, to address comprehension of colposcopy’s purpose, human papillomavirus (HPV) understanding, cancer-related fatalistic beliefs, procedural concerns, and disease progression, offered in both English and Spanish. Females aged 21–65, with colposcopy appointments at an urban OBGYN clinic, were invited to participate. Enrolled patients experienced real-time counseling messages tailored via a Chatbot-driven barriers assessment, dispatched via text one week before their scheduled colposcopy. Cognitive-affective measures were assessed at baseline and through a 1-month follow-up. Participants also engaged in a brief post-intervention satisfaction survey and interview to capture their acceptance and feedback on the intervention. The primary endpoints encompassed study adherence (CervixChat response rate and follow-up survey rate) and self-evaluated intervention acceptability, with predefined feasibility benchmarks of at least 70% adherence and 80% satisfaction. Among 48 eligible women scheduled for colposcopies, 27 (56.3%) agreed, consented, and completed baseline assessments. Participants had an average age of 34 years, with 14 (52%) identifying as non-Hispanic White. Of these, 21 (77.8%) engaged with the CervixChat intervention via mobile phones. Impressively, 26 participants (96.3%) attended their diagnostic colposcopy within the specified timeframe. Moreover, 22 (81.5%) completed the follow-up survey and a brief interview. Barriers assessment revealed notable encodings in the Affect and Values/Goals domains, highlighting concerns and understanding around HPV, as well as its impact on body image and sexual matters. Persistent and relatively high intrusive thoughts and lowered risk perceptions regarding cervical cancer were reported over time, unaffected by the intervention. Post-intervention evaluations documented high satisfaction and perceived usefulness, with recommendations for incorporating additional practical and educational content. Our findings underscore the robust satisfaction and practicality of the CervixChat intervention among a diverse underserved population. Moving forward, our next step involves evaluating the intervention’s efficacy through a Sequential Multiple Assignment Randomized Trial (SMART) design. Enhanced by personalized health coaching, we aim to further bolster women’s risk perception, address intrusive thoughts, and streamline resources to effectively improve colposcopy screening attendance.

Our study aimed to increase attendance for in-clinic colposcopy appointments among urban, diverse women following an abnormal Pap test result. We implemented a Chatbot-driven text messaging approach to enhance follow-up rates.

Implications.

Practice: Effective follow-up care and treatment following abnormal Pap test results are crucial in preventing cervical cancer. Mobile communication tools have proven effective in enhancing regular screening practices and could be seamlessly integrated into the healthcare system, aiding providers in improving adherence to follow-up recommendations.

Policy: A pressing requirement exists for heightened resource allocation towards health promotion initiatives that leverage health communication technologies to bolster adherence to abnormal Pap test results.

Research: Subsequent research endeavors could explore the execution of cervical cancer screening promotion initiatives within underserved communities.

Background

Despite the preventability of cervical cancer due to its slow progression, detectable precursors, effective treatments, and human papillomavirus (HPV) vaccination [1], a substantial disparity persists in cervical cancer mortality within urban underserved populations [2]. When diagnosed early and treated appropriately, survival rates approach 100% [3]. The primary screening method involves co-testing with Pap tests and oncogenic HPV types. Following abnormal Pap results, recommended follow-up includes colposcopy with biopsy to confirm high-grade lesions (CIN2+), succeeded by outpatient procedures like loop electrosurgical excision or cold cone biopsy for CIN2+ treatment [4]. Early detection and excision can prevent invasive cervical cancer. However, up to 63% of underserved and racial/ethnic minority women do not adhere to colposcopy after an abnormal Pap test [5–8]. Urban underserved women exhibit consistently lower adherence rates to follow-up recommendations for abnormal tests [8, 9], with higher rates of cervical cancer incidence [10] and mortality [2, 11].

Barriers include inadequate knowledge about HPV and cervical cancer [5, 7, 8, 12–19], misunderstanding the seriousness of an abnormal Pap diagnosis, and anxiety rooted in insufficient knowledge about colposcopy’s purpose. Other obstacles encompass fear of diagnosis and cancer [5, 13, 16, 18–22], practical challenges (e.g. childcare, transportation) [5, 13, 21, 23, 24], and misconceptions (e.g. post-colposcopy fertility concerns) [8, 12, 25]. Poor communication and lack of reminders also lead to reduced adherence [8, 13, 16, 18, 26]. While these barriers are well-identified [5, 7, 8, 12–26], existing interventions fall short in addressing them. Strategies like reminder calls, travel incentives, and community health worker visits have shown success [5, 27–29], but they are costly and burdensome [30–32]. We aim to leverage mobile communication, particularly text messaging (TXT) interventions, as a low-cost solution. TXT-based interventions have improved appointment attendance and cancer prevention and enhanced Pap testing knowledge in African American [33] and Korean American women [34]. Within digital intervention options, Chatbots—automated conversational agents—can simulate patient–nurse interactions, used in hereditary cancer screening [35]. However, TXT or Chatbot interventions for colposcopy follow-up rates after abnormal Pap tests are underexplored. Our qualitative study revealed that underserved Black and Hispanic women face cognitive and emotional barriers, favoring a relevant TXT-based approach [36]. Building on prior telephone outreach efforts [27–29], we aim to pilot an interactive TXT intervention through the Chatbot interface. Our focus is not only on appointment reminders but also on addressing cognitive and emotional barriers to colposcopy adherence among underserved urban women.

Methods

Ethical approval

This research was reviewed and approved by the Thomas Jefferson University’s IRB and Sideny Kimmel Cancer Center’s cancer research committee.

Purpose of the pilot trial

This pilot study’s objective is to evaluate the feasibility and acceptability of a text-based intervention employing a Chatbot interaction interface. This approach has been adapted from our previous telephone-based tailored communication intervention aimed at enhancing colposcopy attendance following abnormal Pap smears, a strategy that demonstrated considerable efficacy particularly within underserved urban populations. For the current pilot, feasibility was gauged through the proportion of participants actively engaging with the intervention and completing the follow-up assessment. To measure acceptability, a brief satisfaction survey was supplemented with post-intervention interview insights. Additionally, cognitive-behavioral pre- and post-measures, along with data from the barrier assessment conducted during the Chatbot interaction, were also reported, serving exploratory purposes.

Our evidence-based telephone tailored communication for cervical cancer risk intervention

Our intervention, evidence-based telephone tailored communication for cervical cancer risk (TC3), was guided by the Cognitive-Social Health Information Processing (C-SHIP) model [37]. Designed for urban, underserved women, TC3 effectively targets five psychosocial barriers to abnormal Pap test follow-up [27–29]. Specifically, TC3 addresses C-SHIP constructs: lower knowledge and risk perceptions, negative beliefs, and expectancies (e.g. pain from colposcopy), interfering affect (e.g. high anxiety, fear), values and goals (e.g. low valuing of preventive health behaviors), and self-regulatory skills (e.g. lack of a repertoire for overcoming and coping with access-related barriers) [29]. The TC3 intervention is specifically devised for integration within clinical settings and is adaptable for administration by both clinic and research personnel. Women who have an upcoming follow-up appointment subsequent to an abnormal Pap test outcome engage in a Cognitive-Affective Barriers Assessment. Trained research staff facilitate this assessment using a computer-assisted telephone interviewing (CATI) system, typically occurring 2–4 weeks prior to the scheduled appointment. Upon analyzing the outcomes of the barriers assessment, the CATI system discerns the two most significant barriers linked to each C-SHIP construct. Subsequently, it generates tailored messages aimed at addressing these identified barriers, resulting in a cumulative total of ten messages. These personalized messages are then verbally conveyed to the patient over the telephone by the designated staff member. For ease of implementation, comprehensive intervention materials, along with an instructive implementation manual, are made accessible through the National Cancer Institute’s repository of Evidence-based Cancer Control Programs (https://ebccp.cancercontrol.cancer.gov/index.do).

Our text-based adapted intervention: CervixChat

The pursuit of enhancing scalability and participant engagement was the impetus behind our intervention’s evolution. This drive, bolstered by insights from our process data, recognized the potential burdens on clinic resources and participants linked to telephone-based interventions. This recognition spurred our exploration of an innovative solution, resulting in the development of the text-based adapted intervention: CervixChat.

While firmly rooted in the C-SHIP model, CervixChat’s advent ushers in a fresh dimension to intervention delivery. This innovation transmutes the theoretical C-SHIP barrier assessment and tailored messaging from TC3 into a streamlined Chatbot-based sequence. This sequence, encompassing 27 barrier assessment questions and corresponding text responses, is efficiently transmitted through a fully automated and script-based system via TXT. This transformation, while preserving fidelity to the C-SHIP model, holds the promise of heightened engagement due to its interactive essence. CervixChat redefines participant interaction and comprehension through the introduction and elaboration of responses. Moreover, this adaptation maintains alignment with the C-SHIP model while embracing the potential benefits inherent in its interactive format.

CervixChat’s structured intervention strategically spans the pivotal timeframe from day 1 to day 11, thoughtfully positioned just before the scheduled colposcopy. This meticulously crafted sequence commences with an initial greeting, followed by a comprehensive barriers assessment, personalized responses, crucial educational content, and a pivotal reminder about the impending colposcopy appointment. Notably, proactive messaging occurs twice during this intervention. The first instance is a welcome message issued upon enrollment (day 1), incorporating a link to access the Chatbot. The participant’s engagement involves clicking the link to initiate the message interaction. This welcome message serves to introduce the participant to the Chatbot’s functionality and usage. Subsequently, a series of 27 questions and corresponding responses unfold based on the participant’s answers. Each interaction transpires at a specific time point and spans approximately 10–15 minutes. For a visual representation of this intervention’s timeline and flow, refer to Fig. 1

Figure 1.

Overview of the CervixChat intervention schedule

The heart of CervixChat’s automation lies in its dynamic identification and delivery of tailored cognitive-behavioral counseling messages. These messages, prompted by participants’ responses to the barrier assessment, form a tailored approach addressing each participant’s unique question topics, enhancing the precision and relevance of educational messaging. See Fig. 2 for sample screenshots of intervention interfaces and Appendix A for intervention content script samples.

Figure 2.

Sample screenshots of the CervixChat user interface for the barriers assessment

Our Chatbot, a cornerstone of our study, functions as a semi-automatized messaging system meticulously optimized for smartphones. The choice of a TXT-delivered app, over a native app, was driven by accessibility and cost considerations. The web-based symptom checker tool ensures data is not stored on users’ devices, aligning with privacy concerns. Built on a foundation of structured questions and tailored responses, the Chatbot’s design stemmed from a collaborative effort blending gynecological expertise, behavioral technology, and data security. Rigorous beta testing further ensured accuracy and ease of use. This effort assured that the Chatbot adhered to medical guidelines, while prioritizing user-friendliness and precise information delivery.

Central to our mission was the Chatbot’s integration into a smartphone-based messaging system. This approach offered accessible and convenient communication, enabling patients to engage at their own pace. Meeting rigorous health data security standards, the Chatbot ensured the safeguarding of sensitive patient information.

The responses, carefully curated to align with established national guidelines, and backed by authoritative resources like ACOG educational materials, underwent meticulous review. Our study team’s exacting scrutiny guaranteed factual precision. Additionally, for situations involving self-regulatory barriers linked to appointment attendance, participants received tailored information, enriching the participant experience with context-specific support.

To ensure cultural and linguistic appropriateness, the CervixChat intervention content underwent a meticulous translation process. Initially, it was forward translated into Spanish by a contracted service provider specializing in Spanish translation. Subsequently, one of the bilingual authors performed a back translation, guaranteeing alignment with cultural nuances and linguistic accuracy.

Participants

Female participants aged 21–65, who had appointments for colposcopy, were enlisted from May 2021 to December 2021 at the colposcopy clinic located in Jefferson Health's center city location. This clinic primarily caters to a socioeconomically disadvantaged demographic. Exclusion criteria involved individuals unable to communicate or read in English or Spanish, those with a history of malignancy, current evidence of invasive cervical carcinoma, other life-threatening medical conditions, pregnancy, prior colposcopy within the past 2 years, age over 65, or exhibiting significant symptoms of cognitive confusion.

Study procedures

Patients with abnormal Pap smears were notified and scheduled for colposcopy appointments. Roughly 2–4 weeks before their initial colposcopy appointments, potential participants were contacted through both mailed and electronic letters via MyChart, along with opt-out information. After expressing interest or not opting out, the research coordinator provided detailed explanations of the study. Upon obtaining informed consent, participants completed a baseline survey encompassing sociodemographic and cognitive-affective measures. A week before their scheduled colposcopies, participants were slated to receive the CervixChat message for the barriers assessment. This assessment involved tailored messages addressing colposcopy’s purpose, HPV comprehension, cancer-related fatalistic beliefs, concerns about procedures, disease progression, and strategies for appointment adherence. This content was provided in both English and Spanish. In instances where participants didn’t respond to the assessment text, CervixChat re-sent the message 2 days before the scheduled colposcopy to encourage participation and message completion, thus preparing them to receive tailored messages. At the 1-month mark post-intervention, participants underwent follow-up through a survey and a brief interview. Compensation of $30 was offered both for completing the baseline survey and for participating in the post-intervention follow-up assessment.

Study outcome and process measures

Feasibility and adherence to initial diagnostic colposcopy appointment

We assessed the feasibility of the intervention by calculating the proportion of the study participants who completed the entire CervixChat intervention interaction at least once and the proportion of the study participants who completed the 1-month follow-up. We stated that we would consider the intervention to be feasible if we achieved at least 70% engagement of intervention usage and if 70% of participants completed the 1-month follow-up survey, a priori. Patients were considered adherent if they attended their initial colposcopy appointment or an appointment that they rescheduled to another date within 3 months of the original appointment date. Three months was used as the interval for timely follow-up based on clinical guidelines and previous studies [38].

Acceptability

The study incorporated three author-constructed satisfaction questions aimed at gauging participants’ recollection of the intervention, their perception of its usefulness, and their overall impression of CervixChat. Participants were prompted to express their agreement with statements using a 5-point Likert scale that ranged from “strongly agree” to “strongly disagree.” The intervention’s acceptability criterion was predefined as achieving at least 80% agreement (strongly agree/agree) across the satisfaction questions.

Additionally, a brief 10-minute telephone interview was conducted to delve into participants’ qualitative feedback and gather suggestions for potential enhancements to the intervention in the future. This interview aimed to capture nuanced insights and pave the way for refining the intervention based on participants’ perspectives.

Demographics and cognitive-affective measures

Participants completed a questionnaire assessing sociodemographic information at baseline and cognitive-affective factors at both baseline and 1-month follow-up. A battery of C-SHIP-based assessment questions used in the current study was developed from our prior work [29, 37]. Knowledge response options were TRUE/FALSE to assess participant understanding of the meaning of an abnormal Pap smear; what a colposcopy examination entails; and what HPV is and its link to cervical cancer (test–retest: ICC = 0.75 in the prior study [29]). Perceived risk for developing cervical cancer was assessed to understand the degree to which participant feel at risk for developing cervical cancer (α_pre = 0.89, α_post = 0.91 in the current study). Expectancies and beliefs was measured to understand participant’s perception of the utility of cervical cancer prevention and treatment (α_pre = 0.87, α_post = 0.88 in the current study). Affect was assessed by the Impact of the Events Scale –Revise (Intrusion subscale) [39] to understand participant’s repeated thoughts such as nightmares, feelings, and thoughts concerning cervical cancer screening(α_pre = 0.85, α_post = 0.86 in the current study). Values and Goals were measured to understand participant’s attitudes about cervical cancer screening and health-related lifestyles (α_pre = 0.90, α_post = 0.88 in the current study). Self-regulatory skills were assessed to understand participant’s confidence and ability to manage all aspects associated with colposcopy (α_pre = 0.89, α_post = 0.90 in the current study). All the measures (excluding Knowledge) were evaluated using a 5-point scale encompassing response options: “extremely,” “very much,” “somewhat,” “slightly,” and “not at all.” Participants were asked to indicate their level of agreement or understanding with each statement, and their responses were rated based on the extent of alignment or comprehension with the specific question.

CervixChat intervention barrier assessment data

Each participant’s responses to the text messaging-delivered 27 barrier questions elicited prior to the colposcopy appointment and assessed by the CervixChat intervention were electronically collected and were used to generate the tailored counseling messages as illustrated in Figs. 1 and 2. We will report the percentage of participants who received a corresponding cognitive-affective message given a moderate-severe barrier response reported.

Statistical analysis

Descriptive statistics were completed for all sociodemographic, cognitive-affective, and adherence outcome variables. Percentages, mean and SD were used when appropriate. Pair sample t-tests were used to explore changes in selected cognitive-affective measures assessed at baseline and 1-month follow-up.

Results

Recruitment and retention process data

A total of 158 patients were assessed for eligibility based on the colposcopy clinical schedule, with 70 individuals excluded for various reasons outlined in Fig. 3. An opt-out letter detailing the study information was sent via both mailing and MyChart inbox to 88 potential eligible candidates. After a 2-week period with no opt-out responses, we initiated outreach calls, successfully engaging 44 potential eligible patients. Out of this group, 16 individuals declined participation, while 32 agreed and completed the informed consent process.

Figure 3.

Consort diagram

However, after consenting, five participants were subsequently excluded. One participant withdrew due to inability to fulfill administration requirements for compensation, another became ineligible due to pregnancy, and three were unable to complete the baseline within the specified timeframe. Consequently, the final analysis included 27 participants who completed the baseline assessment after providing consent.

Descriptive statistics of demographic variables

Among the cohort of 27 participants, the average age was 33.8 years, with 14 (51.8%) identifying as non-Hispanic Whites (10, 37.0% African American). The majority (n = 19, 70.3%) were not married, while 21 (77.7%) were employed. About 14 (51.8%) had received the HPV vaccine, and 13 (48.1%) had experienced an abnormal Pap result as the first time. Furthermore, 15 (55.5%) reported undergoing a Pap Smear within the 1.5 to 2-year timeframe. An overview of these basic demographic details is provided in Table 1.

Table 1.

Participant demographics (n = 27)

	n (%)
Age, mean (SD) and median	33.8 (7.5) and 32
Non-Hispanic white	14 (51.8%)
African American	10 (37.0%)
Education level: below college	8 (27.5%)
Income level: $45K or below	9 (31.0%)
Being single/never married	19 (70.3%)
Employed currently	21 (77.7%)
Received HPV vaccine	14 (51.8%)
Pap smear screening:
Every year	12 (44.4%)
Every 1.5–2 years	15 (55.5%)
First time had an abnormal Pap smear result	13 (48.1%)

Feasibility and adherence to initial diagnostic colposcopy appointment

Out of the total 27 participants, 21 (77.8%) individuals actively engaged with the CervixChat intervention. Impressively, 26 (96.3) participants adhered to the recommended timeframe for attending their diagnostic colposcopy within 3 months. During the one-month follow-up, 22 participants (81.5%) successfully completed both the follow-up survey and a brief interview. Notably, the intervention was utilized in Spanish by one participant within the Spanish-speaking demographic. However, it’s worth noting that among the five participants who did not complete the 1-month follow-up (accounting for a 18.5% attrition rate), three were uncontactable, while two indicated time constraints as the reason for not completing the follow-up process.

Acceptability

Participants exhibited notable satisfaction with the CervixChat intervention, with a substantial 90.0% (n = 20) expressing strong agreement or agreement that they enjoyed receiving messages concerning their appointment. Furthermore, a significant proportion of participants (95.4%, n = 21) reported that the CervixChat intervention facilitated their attendance at the appointment, as indicated in Table 2. An analysis of the responses from participants who participated in a brief interview evaluation (n = 22) revealed recurring themes. Participants consistently characterized CervixChat as “helpful,” serving as a “reminder,” and imparting “informative” content. The intervention was frequently praised for being “clear and concise,” "enhancing accessibility", and deemed “easy to access". Participant interviews also revealed areas for potential future usability improvement: (i) One such area pertains to the navigation process within the intervention. Currently, participants are required to follow a link to access messages, with subsequent clicks needed for progress through the interaction. A refinement could involve streamlining this process, eliminating the need for additional clicks, thus enhancing speed and ease of participant engagement. Moreover, an observation from the interview data indicates that proactively sending all messages might be advantageous. Even if participants are already familiar with the information, the proactive approach could encourage a thorough review of the content. This finding suggests that presenting all messages upfront could lead to enhanced information absorption, contributing to an even more effective participant experience. (ii) Furthermore, participants expressed a desire for enhanced logistical information beyond the colposcopy procedure, particularly regarding parking and parking vouchers. (iii) Additionally, participants recommended the inclusion of supplementary web links within the messages to provide access to further relevant information. (iv) The intervention revealed heightened worry among certain participants due to HPV-related messaging. This unintended effect underscores the delicate nature of conveying risk factors. In the realm of cancer worry, uncertainties surrounding health outcomes often amplify anxiety. Future CervixChat versions will incorporate more empathetic language, personalized support, and pre-testing to address such concerns.

Table 2.

Satisfaction questions (agree and strongly agree) (n = 22)

	Number of patients (%)
Remember about the messages received	22 (100%)
Enjoyed about the messaged received	20 (90.9%)
Messages made attending appointment easier	21 (95.4%)

Cognitive-affective measure pre- and post-test

Regarding participants’ self-reported cognitive-affective measures at baseline and the 1-month post-intervention follow-up, minimal alterations were noted over this period. Overall, participants indicated relatively high levels of knowledge, expectancies/beliefs, values/goals, and self-regulatory skills. The group displayed a moderate risk perception toward cervical cancer. An average score between 1.9 and 2 on the RIES-intrusion subscale indicates a mild to moderate stress response to the event of an abnormal Pap result. For detailed baseline and follow-up measure results, please refer to Table 3.

Table 3.

Cognitive-affective measures (n = 22)

	Baseline	1-month follow-up	P-value
Knowledge (average of % correct answers)	86%	86%	.52
Risk perceptions [1–5, mean (SD)]	2.9 (0.8)	2.9 (0.8)	.47
Expectancies/beliefs [1–5, mean (SD)]	4.4 (0.5)	4.5 (0.5)	.32
Affect [1–4, mean (SD)]	2.0 (0.8)	1.9 (0.7)	.11
Values and goals [1–5, mean (SD)]	4.2 (0.5)	4.3 (0.5)	.37
Self-regulatory [1–5, mean (SD)]	3.8 (0.9)	4.0 (0.8)	.18

CervixChat intervention barriers assessment data

The proportion of participants who reported a moderate or severe barrier during the intervention interaction with the CervixChat intervention for each of the C-SHIP domains was listed in Table 4 and elaborated below.

Table 4.

Cognitive-Affective Barriers Assessment assessed by CervixChat intervention (n = 21)

	Percentage of participants who reported moderate to severe barriers
Understanding colposcopy (5 questions)
How well do you understand the connection between HPV and cervical cancer?	11 (52.4%)
How well do you understand what it means to have an abnormal Pap smear?	11 (52.4%)
How well do you understand what Human Papilloma Virus, or HPV, is or how you would get it?	10 (47.6%)
How well do you understand what a colposcopy is and why it is done?	6 (28.6%)
How well do you understand what a biopsy is and why you may need one when you have your colposcopy?	6 (28.6%)
Beliefs about cervical health (5 questions)
How much do you believe that there is nothing you can do that will change whether or not you get cancer?	7 (33.3%)
How much do you believe that having an abnormal Pap smear means you have cancer?	6 (28.6%)
How much do you believe that a colposcopy will help you?	2 (9.5%)
How sure are you that you will be able to do what your doctor tells you to do after your colposcopy?	2 (9.5%)
How sure are you that you will be able to keep your colposcopy appointment?	0 (0.0%)
Emotions and worry (6 questions)
How much are you worried about passing on HPV to any sexual partners?	15 (71.4%)
How worried are you about having any pain or discomfort from the procedure (colposcopy/biopsy) itself?	13 (61.9%)
How worried are you about the possibility of having cancer?	11 (52.4%)
How worried are you about any treatments you might need after the colposcopy?	10 (47.6%)
How worried are you about having pain or other symptoms after the colposcopy?	8 (38.1%)
How worried are you about having sexual problems after the procedure?	6 (28.6%)
Importance of keeping the appointment (5 questions)
How important is it for you to have a healthy sex life?	21 (100.0%)
How important is it for you to keep a good body image and feel good about yourself?	20 (95.2%)
How important is it for you not to be embarrassed or feel ashamed during the colposcopy?	18 (85.7%)
How important is it for you to be able to become pregnant and have a child?	14 (66.6%)
How important is it for you to keep the colposcopy appointment and follow through with all of your doctor’s recommendations?	1 (4.8%)
Coping skills (6 questions)
Do you have ways to deal with any distress about the colposcopy that has been set up for you?	10 (47.6%)
Do you have ways to pay for of the colposcopy appointment and/or get any referral letters you might need for insurance reasons?	10 (47.6%)
Do you have ways to deal with your other responsibilities, such as getting childcare, eldercare, or coverage at work so that you can keep your colposcopy appointment	3 (14.3%)
Do you have ways to make sure you will remember your colposcopy appointment?	1 (4.8%)
Will transportation to and from the colposcopy appointment be a problem for you?	1 (4.8%)
How much are you worried about not understanding what your doctor is saying because English is not your first language?	1 (4.8%)

Knowledge barriers

Understanding colposcopy: a majority of participants, 52.4% (n = 11), encountered knowledge barriers, primarily regarding the relationship between HPV and cervical cancer, as well as understanding the implications of their abnormal Pap smear results. Almost half of the participants (47.6%, n = 10) indicated a lack of understanding about HPV and its transmission. Overall, 19% (n = 4) of patients reported all five knowledge barriers; 4.8% (n = 1) reported four barriers, and 23.8% (n = 5) reported three barriers.

Expectancies and beliefs

Beliefs about cervical health: In terms of beliefs about cervical health, 33.3% (n = 7) reported a negative belief toward their control over a cancer diagnosis, while 28.6% (n = 6) indicated a positive belief and association between an abnormal Pap smear and the possibility of a cancer diagnosis. Overall, fewer patients reported multiple barriers in this category; only 4.8% (n = 1) of patients reported three barriers, 23.8% (n = 5) reported two barriers, and 19.1% (n = 4) reported one barrier.

Affect

Emotions and worry: Following notification of their initial abnormal Pap smear results, a significant portion, 71.4% (n = 15), expressed fear of transmitting HPV to their sexual partners. Interestingly, 61.9% (n = 13) had concerns about experiencing pain or discomfort during the colposcopy. More than half (52.4%, n = 11) were worried about the possibility of having cancer, and 47.6% (n = 10) were concerned about the potential treatments. Overall, 23.8% (n = 5) reported all six affect barriers, 4.8% (n = 1) reported five barriers, and 9.5% (n = 2) faced four barriers in this category.

Values and goals

Importance of keeping the appointment: Among participants, the significance of maintaining a healthy sex life received unanimous endorsement at 100% (n = 21). Moreover, 95.2% (n = 20) recognized the importance of maintaining a positive body image. Additionally, 85.7% (n = 18) valued not feeling embarrassed during the colposcopy, and 65% (n = 13) acknowledged the importance of their future childbearing ability. Overall, no patient reported more than four barriers for this category; 57.1% (n = 12) reported four barriers, and 33.3% (n = 7) reported three barriers.

Self-regulatory

Coping skills: The most prevalent self-regulatory barriers reported by patients pertained to cost and health insurance-related challenges for the appointment, along with methods for managing associated stress (47.6%, n = 10). Overall, no patient reported more than four barriers for this category; 9.5% (n = 2) reported four barriers, and 4.8% (n = 1) reported three barriers. The majority of patients (33.3%, n = 7) faced one barrier in this category.

Discussion

Despite cervical cancer being largely preventable, women from inner-city underserved communities remain at elevated risk due to disparities in adhering to follow-ups following abnormal cervical cytology results. Leveraging existing research and our prior work, this study demonstrates the feasibility and acceptability of the CervixChat intervention in enhancing adherence to abnormal Pap test follow-ups among minority and medically underserved women within our catchment areas.

Our findings reveal that a majority of participants actively engaged with the intervention, expressing ease of use, and reporting high satisfaction due to its informative nature. This underscores the appropriateness of a Chatbot-interfaced TXT intervention for this demographic. The prevalence of mobile phones makes them an optimal platform for reaching this population. Harnessing mHealth strategies holds substantial potential for mitigating health disparities within underserved communities of color.

Among participants who completed the intervention barriers assessment via Chatbot, our findings highlighted “Distress” and “Values/Goals” as the most prominently endorsed factors, emphasizing their significance within this vulnerable population. This outcome underscores the need to address these aspects thoughtfully. Our findings align with prior research [40] indicating that psychological distress is a pivotal factor contributing to non-adherence following an abnormal Pap test result. It’s plausible that these women encounter challenges in managing the emotional burdens associated with cervical cancer risk and associated procedures.

Notably, concerns regarding the sexually transmitted nature of HPV emerged as a common thread, underlined by the entire sample’s endorsement of a healthy sex life within the Values/Goals category. Feedback from post-intervention interviews accentuated the importance of conveying HPV information in a manner that imparts facts while mitigating anxiety surrounding the infection.

The endorsement of preserving a positive body image and childbearing underscored these themes. Consistent with prior studies, younger women reported fear of loss of reproductive functioning accompanying an abnormal Pap smear result [40]. Increased public awareness about the role of HPV in cervical cancer and the mode of transmission also negatively impacts patients’ feelings about their bodies [41]. Prior studies also showed that 60% of women diagnosed with cervical intraepithelial neoplasia reported a change in their sexual activity due to feeling contaminated. Worries about having any pain or discomfort from the procedure were also highly reported. Some studies have shown that predictor of anxiety before colposcopy was fear of the actual colposcopy examination [42]. Women need to be informed about the potential for discomfort during medical procedures and to be supported by the health care professionals. These potential enhancements can be accomplished by including additional web links to relevant credible websites, as suggested by the participants in the post-intervention interviews.

Our examination of changes in cognitive-affective measures between pre- and post-intervention revealed the presence of persistent moderate intrusive thoughts among women throughout the study duration. This observation resonates with previous studies indicating heightened intrusive thoughts and avoidant coping mechanisms subsequent to abnormal Pap smear results [43, 44]. Offering timely psychological support to individuals with abnormal cervical cytology screenings may prove instrumental in mitigating anxiety and distress, thus potentially enhancing adherence to recommended evaluations.

Moreover, our relatively young sample, with an average age of 34, displayed a moderate level of risk perception regarding cervical cancer. This age group’s perception of risk and its implications for cancer screening behaviors, particularly in cervical cancer, merit deeper investigation. The distinct tendency among younger individuals to underestimate their vulnerability to health issues, including cervical cancer, has far-reaching consequences on proactive health actions. This intricacy becomes crucial to comprehend, especially considering that inadequate risk perception might hinder optimal cervical cancer screening. Our cognitive-affective intervention content didn’t significantly alter participants’ risk perceptions. This underscores the challenge of shifting perceptions through this approach. Therefore, it prompts the need for future innovative health communication strategies to effectively address accurate risk perception related to cervical cancer within the context of a younger demographic.

Despite participants reporting considerable barriers during the CervixChat intervention, with moderately high distress and lowered risk perception from baseline to post-intervention, it’s noteworthy that adherence to colposcopy was notably high, standing at 96% within a 3-month timeframe. Additional research is warranted to investigate adherence behaviors pertaining to long-term follow-ups after a colposcopy. Further, within our study, a subgroup of eight participants refrained from initiating interaction with the Chatbot, constituting a smaller proportion of the cohort. Employing an intent-to-treat strategy enabled us to encompass all participants, including those who didn’t engage with the eHealth intervention. While aligning with real-world conditions, this approach introduces potential variability due to the absence of responses from this subgroup.

The constrained timeframe between an abnormal Pap smear result and a scheduled colposcopy may have impacted participants’ familiarity and engagement with the eHealth intervention. This limitation highlights the intricate nature of intervening effectively within such time constraints while considering potential different personal information-seeking behaviors. Looking ahead, our ongoing research endeavors include the development of a Sequential, Multiple Assignment, Randomized Trial (SMART) design to address this issue. We recognize that individual preferences and habits, such as personal information-seeking behaviors, might influence engagement with interactive eHealth Chatbots. To cater to these dynamics, we’re in the process of implementing a health coaching component, providing personalized counseling for those who don’t respond to or engage with the eHealth intervention. This innovative approach aims to enhance engagement and intervention effectiveness, while also conserving resources by allocating additional support specifically to those requiring it. This line of research reflects our commitment to optimizing participant experiences and outcomes.

Study Limitations

We recognize several limitations in our current study. Firstly, the pilot nature of our design resulted in a small sample size. Secondly, given our study’s urban setting, the generalizability of findings to rural areas is limited. Thirdly, our study observed minimal participation from the Spanish-speaking population. Additionally, the absence of a control arm prevents us from understanding the impact of no intervention or a text reminder-only approach on follow-up rates. Lastly, the lack of data from 19 non-participating women emphasizes the necessity to grasp recruitment challenges within our eHealth study.

Conclusion

In conclusion, our study demonstrated the feasibility and strong acceptability of utilizing a Chatbot interface to deliver theory-based health communication messages via TXT. This approach proved effective in engaging a diverse urban female population and addressing their follow-up care after an abnormal Pap smear. Our findings revealed gaps in knowledge among participants, with the majority lacking understanding of HPV transmission and nearly half being unaware of its association with cervical cancer. Additionally, participants expressed more concern about transmitting HPV to partners than the potential health impact on themselves.

Educating women about the widespread prevalence of HPV infection could potentially enhance comprehension and shift focus toward personal health considerations. Key issues, such as misconceptions about HPV, body image concerns, sexual issues, and procedural discomfort, should be thoughtfully addressed to alleviate psychosocial distress. This mitigation is crucial, as it could substantially impact follow-up behaviors following abnormal cervical cancer screening.

Looking ahead, our findings offer valuable insights that can guide the enhancement and implementation of future cervical cancer screening programs within underserved communities. By tailoring interventions to address knowledge gaps, misconceptions, and psychosocial concerns, we have the potential to create more effective and impactful interventions for these populations.

Appendix A. CervixChat intervention content script samples

C-SHIP barrier domain	Patient response	Tailored CervixChat response
Understanding colposcopy “How well do you understand what it means to have an abnormal Pap smear?”	Not at all	An abnormal Pap smear means that some cells on your cervix could look different from regular cells and these cells could turn into pre cancer or cervix cancer over time. Most of these changes happen very slowly, over years. It is important to get your colposcopy to look for these changes Even if precancerous cells are found there is a simple office procedure that can be done to remove these cells.
Beliefs about Cervical Health “How much do you believe that a colposcopy will help you?”	Not at all	In a colposcopy, the doctor looks directly at your cervix to see if there are any abnormal cells. There are no symptoms that can be felt or seen to know about these abnormal cells. To be sure, your doctor might take a small sample of cells which will be sent to the lab. This is called a biopsy. Both tests will help your doctor decide what the next step will be.
Emotions and Worry “How worried are you about any treatments you might need after the colposcopy?”	Extremely	A colposcopy may be all that is needed, but if needed the treatments after colposcopy are not painful. They take about 15 minutes and are often done in the clinic. There is more than one way to remove the changing cells on your cervix. The doctor will decide which treatment is best for you. Either way, you will be scheduled for a follow-up appointment and your doctor or nurse will call you to let you know what is needed.
Keeping your Appointment “How important is it for you to keep a good body image and feel good about yourself?”	Extremely	Some women might feel like their body has let them down, as if something has gone wrong. In fact, your body is telling your doctor that there are some changes that need attention. Coming in for the exam will help you feel good about taking care of your body.
Getting to your Appointment “Do you have ways to deal with your other responsibilities, such as getting childcare, eldercare, or coverage at work so that you can keep your colposcopy appointment?”	Not at all	If you need a babysitter, try to make plans with people you can trust to watch your child. If your appointment interferes with your work schedule, try to take a personal day or speak with co-workers about ways to cover you. It may be extra hard during COVID-19 to find help for child or eldercare. Look for someone that is in your immediate or close family. It’s OK if you need to reschedule, but make sure you call and let the office know. Also, make sure you set another date for as soon as you can.

Funding

This research was supported by the Provost’s Pilot Clinical Research Award from Thomas Jefferson University, the Sidney Kimmel Cancer Center Support Grant (National Cancer Institute Award 5P30CA056036) and NIH 1R01MD017675-01A1. The CervixChat technical infrastructure was supported by Agility Engine, LLC Company and Mr. Peichung Shih. We thank Ms. Nabaala Claxton and Ms. Rita Smith’s assistance for patient accrual. We thank Ms. Carly West for her data cleaning and analysis assistance.

Conflict of interest statement: the authors declare that they have no conflicts of interest.

Human Rights

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained from all individual participants included in the study.

Welfare of Animals

This article does not contain any studies with animals performed by any of the authors.

Transparency Statements

The study and analysis plan were not formally registered. De-identified data are not available in a public archive. De-identified data will be made available as allowable according to institutional IRB standards by emailing the corresponding author. There is not analytic code associated with the study. Materials used to conduct the study will be made available by emailing the corresponding author.