Deciding on surveillance colonoscopy: a multi-phase study of a conversation aid for older adults

Abstract

Background

Surveillance colonoscopy in older adults with a history of colon polyps presents a complex decision, as procedural risks increase and potential benefits of colorectal cancer prevention decrease with age. Despite this, no formal guidance exists to support decision-making. We aimed to develop, refine, and pilot-test a one-page conversation aid to facilitate shared decision-making (SDM) between older adults and primary care providers (PCPs) regarding surveillance colonoscopy.

Methods

We conducted a multi-phase study incorporating both qualitative and quantitative methods to develop, refine, and pilot-test a conversation aid. In Phase 1, we iteratively refined a draft conversation aid through semi-structured interviews with older adults aged ≥ 70. In Phase 2, we surveyed older adults, PCPs, and gastroenterologists to assess acceptability and implementation preferences. In Phase 3, we pilot-tested the conversation aid during routine primary care visits. We assessed SDM using the Observer OPTION-5 tool and applied thematic analysis and descriptive statistics.

Results

Phase 1: Ten older adults participated in interviews and suggested minor improvements to the conversation aid, including the addition of illustrations. Phase 2: Surveys of 200 older adults, 25 PCPs, and 25 gastroenterologists showed that the conversation aid was clear, useful, and appropriately timed for discussions about repeat colonoscopy. Phase 3: The conversation aid was pilot-tested by five PCPs in clinical encounters with 17 patients. It was particularly helpful for patients who were undecided about continuing surveillance. The average OPTION-5 score reflected moderate levels of SDM, with variation across encounters. PCPs supported flexible use based on individual patient preferences and endorsed multiple modes of delivery.

Conclusions

The conversation aid was feasible and acceptable for use in primary care. It shows promise for supporting individualized decisions about surveillance colonoscopy in older adults. Future research should evaluate its impact on decision quality and integration into clinical workflows.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12911-026-03431-2.

Background

Surveillance colonoscopy is commonly recommended for patients with a history of colon polyps and is the most frequent indication for colonoscopy among U.S. adults ≥ 65 in the US [1, 2]. While surveillance colonoscopy is appropriate in healthy individuals, understanding the risks, trade-offs, and other complexities of pursuing surveillance colonoscopy is especially important in older adults with competing health demands, declining function, and limited life expectancy [3]. Risks of surveillance colonoscopy increase with age and include bleeding, perforation, aspiration pneumonia, arrhythmias, and heart attacks, which are significantly heightened in those with co-morbid illness [4, 5]. It is estimated that 5.6 million adults > 75 undergo surveillance colonoscopy annually [6], even after age 75 and when in poor health [7, 8]. Currently, there are no explicit guidelines for when to stop surveillance [1, 3].

There is a critical gap in the communication of risks, potential benefits, indications and alternatives around surveillance colonoscopy in older adults [9–11]. Focus groups among older adults with polyps have identified that they want to have a larger role in decision-making around surveillance and currently have an insufficient understanding of the magnitude of relative benefits and risks of colonoscopy [11]. The majority of older adults have an exaggerated sense of the malignant potential of polyps, greater than 10-fold higher than reality [11, 12]. In addition, primary care providers (PCPs) want more guidance on decision-making around surveillance [9, 13]. Shared decision-making (SDM) and the use of conversation aids have shown promise in colorectal cancer screening and other contexts involving preference-sensitive decisions, offering structured, evidence-based approaches to engage patients in meaningful dialogue [14, 15]. Specifically, decision aids designed for older adults have increased knowledge, risk perception, and participation in SDM, and decreased decisional conflict [16].

We developed an encounter conversation aid designed to support SDM between older adults with a history of polyps and their PCPs regarding surveillance colonoscopy. We conducted a multi-phased study incorporating qualitative and quantitative methods to develop, refine, and pilot-test a one-page conversation aid to facilitate SDM between older adults and PCPs regarding surveillance colonoscopy.

Methods

Design

All methods and results are reported according to the COnsolidated criteria for REporting Qualitative research (CORE-Q) checklist criteria for qualitative interviews and the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) for online surveys [17, 18]. To address Aim 1, we conducted a cross-sectional qualitative exploration of older adults’ perceptions of the usability and acceptability of an encounter conversation aid. We used one-on-one semi-structured telephone interviews. Based on user feedback on content and format, we iteratively refined and enhanced the encounter conversation aid to create an acceptable, usable prototype. To address Aim 2, we conducted two different online surveys with older adults in the general population and PCPs and gastroenterologists. To address Aim 3, we piloted the implementation of the encounter conversation aid among older adults with a history of polyps in a primary care clinic and assessed the patient and PCP experience. Figure 1 provides a broad overview of our study timeline and activities.

Fig. 1.

Study timeline overview

Intervention

We developed a tabular one-page encounter conversation aid to help patients and their physicians make an informed decision about whether to stop or continue surveillance colonoscopy (see Appendix 1). This encounter conversation aid was organized around patients’ frequently asked questions about surveillance colonoscopy. It synthesized evidence from focus groups [11, 19, 20] and surveys of older adults with polyps, with data from large trials on CRC incidence [21–23], mortality [24], and colonoscopy complications [4, 5, 25–27].

Setting and participants

Aim 1

The research team conducted Aim 1 activities with participants remotely using telephone and Zoom video conferencing. We included English-speaking adults aged ≥ 70 with prior colon polyps due for surveillance colonoscopy within 24 months at an academic medical center in rural New England, Dartmouth-Hitchcock Medical Center in Lebanon, New Hampshire. Discussions about stopping surveillance colonoscopy often begin several years before cessation, which prior research suggests commonly occurs around age 75. Including adults in their early 70s allowed us to capture perspectives at a stage when anticipatory decision-making is clinically relevant. We excluded patients with a personal history of colorectal cancer, inflammatory bowel disease, a genetic colorectal cancer syndrome such as familial adenomatous polyposis syndrome or hereditary nonpolyposis colorectal cancer syndrome (Lynch), cognitive impairment and/or inability to consent. Patients must also have had the availability and access to a phone for the one-on-one telephone interview.

Aim 2

The research team conducted Aim 2 activities with participants throughout the United States remotely. We used a quota-based convenience sampling strategy managed by Qualtrics and Sermo. For the online patient survey, we included any individual 65 years or older able to complete an online survey in English (target of n = 200). We used a pre-programmed quota to achieve a 50/50 split between males and females for the survey. For the online health professional survey, we aimed to recruit 20 physicians, half PCPs and half gastroenterologists who see older adults. A target of 20 physicians was selected based on pragmatic considerations and consistent with a survey-based study of evaluating a colorectal cancer screening decision aid [28].

Aim 3

For Aim 3, we included (1) PCPs in Internal Medicine at Dartmouth-Hitchcock Medical Center who regularly care for older adult patients and (2) their older adult patients (65+) who had a history of adenomatous colon polyps and were being seen to discuss repeat colonoscopy. We included adults aged ≥ 65 years to reflect real-world practice, where PCPs may use the conversation aid both to support decisions to discontinue surveillance and to affirm decisions to continue surveillance among slightly younger older adults.

Procedure

Aim 1

The research team developed a semi-structured interview guide designed using Morville’s user-experience honeycomb framework, which addresses the usability, usefulness, credibility, desirability, value, findability, and accessibility of the encounter conversation aid (Appendix 2) [29]. The interview guide also included self-reported demographic questions. The interview guide was internally pilot-tested by the research team prior to conducting interviews with participants.

We recruited participants using a combination of convenience sampling approaches to enhance the diversity of perspectives and increase generalizability.

1. Dartmouth-Hitchcock Health at Lebanon Endoscopy Database.

   • This was our primary method of recruitment due to COVID-19 restrictions. The PI (AC) used the endoscopy database at Dartmouth-Hitchcock Medical Center in Lebanon to identify and purposively recruit eligible patients. We attempted sampling evenly across age groups (70–74, 75–79, 80–84 years old). A research team member contacted eligible patients by telephone using an informational script to recruit patients.

• 2.
  Community-based recruitment.

  • A research team member contacted local senior centers in the Upper Valley area in New Hampshire and Vermont with an email template to distribute to center members with information about the study and to contact the research team to participate.

All patients who agreed to participate in the interview were mailed a hard copy of the encounter conversation aid by a research team member for review before the scheduled interview. A female experienced research team member (DS) obtained verbal consent and conducted and recorded one-on-one telephone interviews (45 min maximum) with consented eligible participants. We conducted the interviews until we reached data saturation, which was determined by the research team during regular discussions, as no new themes emerged from the data [14]. Recordings were transcribed using HIPAA-compliant procedures. No repeat interviews were carried out with patient participants. Field notes were not collected during the interviews. Interviews were recorded and transcribed verbatim.

Aim 2

The research team engaged with staff at Qualtrics (Qualtrics, Provo, UT) for the public survey and Sermo (Sermo, Cambridge, MA) for the physician survey, which both managed the recruitment of survey participants and survey data collection. The research team internally developed the patient and physician surveys, which focused on assessing the acceptability of the conversation aids and how they could be used in clinical practice (Appendices 3–4). Once the survey quotas were filled, the aggregate, anonymous survey data were given to the research team by Qualtrics and Sermo. Any requested changes to the conversation aid, agreed upon by the research team, were subsequently made prior to conducting Aim 3.

Aim 3

The study team presented the pilot study opportunity to the General Internal Medicine and Geriatrics Department at Dartmouth-Hitchcock Medical Center. The Principal Investigator (AC) sent emails inviting PCPs to join the study. Those who were interested and available were then consented to participate. Using an effective brief training method, the research team met them virtually to conduct training on SDM, how to use the conversation aid with their eligible patients, and other relevant study-related procedures [30]. Research team members, male and female medical students at Dartmouth-Hitchcock Health with pre-approved electronic health record (EHR) access, identified eligible patients seen by the participating PCPs using the EHR. Research team members confirmed patient eligibility with PCPs and contacted patients by phone to recruit them to the study. Patients who provided verbal consent were sent an electronic copy of the encounter conversation aid to review before their upcoming appointment with their PCP and a link to a pre-appointment patient survey in Qualtrics to collect self-reported patient demographics. At the eligible appointment, a research team member or participating PCP recorded the conversation regarding repeat colonoscopy and when they used the encounter conversation aid.

Post-appointment, the research team contacted eligible patients by phone to participate in a short interview about their experience with the encounter conversation aid and to report the patient’s perception of SDM. The research team internally developed and piloted this interview guide, designed to take no more than 30 min to conduct and to evaluate the acceptability and utility of the encounter conversation aid (Appendix 5). After patient recruitment was completed, the research team invited participating PCPs to participate in a post-study interview to evaluate their experience using the encounter conversation aid and their interest in using it in the future. The research team internally developed and piloted the post-study PCP interview guide designed to take no more than 15 min to conduct (Appendix 6). No repeat interviews were carried out with patient or PCP participants. Field notes were not collected during the interviews. They were recorded and transcribed verbatim.

Data management and analysis

All data collected between Aims 1–3 were safely uploaded and stored using Dartmouth’s HIPAA-compliant storage platform, Dartmouth Sharepoint. Any transcripts from recordings were not returned to participants for comment or correction as they were transcribed verbatim by a human translator from an established online translation company, Rev.com.

Aim 1

Two independent coders (DS and AC) coded the interview transcripts using Morville’s user-experience ‘honeycomb’ framework and inductive coding using ATLAS.ti qualitative coding software [31]. The coders discussed and agreed upon major and minor themes with an arbitrator present (M-AD).

Aim 2

The Principal Investigator (AC) computed summary statistics for patient participant characteristics using Microsoft Excel. Survey responses were presented as N(%). For the open-ended question posed to PCPs and gastroenterologists, a research team member (DS) conducted a qualitative thematic analysis using ATLAS.ti.

Aim 3

A research team member (DS) conducted descriptive statistics for the patient pre-appointment survey using SPSS quantitative analysis software. Health literacy was assessed using Chew’s validated 3-item screening tool, with each item scored on a 5-point Likert scale and summed to produce a total score ranging from 0 to 12. Based on established thresholds, patients were categorized as having adequate (0–4), marginal (5–6), or inadequate (7–12) health literacy [32]. Two coders (AC and DS) inductively coded the patient post-appointment interviews and PCP post-study interview transcripts using ATLAS.ti. A purposive sample of each interview type was selected for pilot coding. Once agreement regarding coding was achieved, AC and DS independently coded the rest of the interview transcripts. Major and minor themes were discussed between the coders and agreed upon as a broader research team.

The appointment interview recordings were analyzed using Observer OPTION-5, an observer-based measure of SDM [33]. Two researchers (AD and SX) piloted rating a purposive sample of the recordings. The researchers reviewed the results of the pilot and resolved discrepancies with an arbitrator (DS), who is experienced in SDM research. The researchers then independently rated the rest of the recordings. The research team then discussed and agreed upon the results of the ratings with the coders present. Interrater reliability was measured between the two independent raters using Cohen’s kappa [34].

Results

We reviewed and discussed the results for Aims 1–3 with study stakeholders, participating PCPs, and other relevant clinical staff who offered advice about the best presentation of results given the length and nature of this study.

Aim 1

We completed Aim 1 between June and October 2020. We interviewed 10 adults with a history of colon polyps. The mean age was 77.2 years (range 70–83). Four participants were female, nine were White, eight attained at least some college education, three had Medicaid, and nine reported adequate health literacy. The interviews averaged 31 minutes long (range 16–53 minutes). Results are reported by Morville’s framework in Appendix 7. Analysis of Aim 1 interviews identified four primary themes related to the encounter conversation aid: findability and usability, perceived usefulness, credibility and value, and accessibility for older adults. Participants offered minor suggestions to improve the encounter conversation aid’s content: adding illustrations and emphasizing the preparation for colonoscopies. We adapted the conversation aid based on feedback and added illustrations (Fig. 2).

”I think it’s good. And if people had other questions following it, they can ask their doctor.” - ID 05 (Female, 74 years old, Adequate health literacy).

“Well, I think the pictures would be good, yeah.” - ID 09 (Male, 79 years old, Adequate health literacy)

Fig. 2.

Updated two-sided conversation aid (front and back) used for Aim 3

Aim 2

Public survey. In Fall 2021, 200 patients completed the survey. The majority were 65–74 years old (71%), non-Hispanic White (89.5%) with some college education or more (67%). 84% had adequate literacy, and 7.5% prior experience with conversation aids (Appendix 8). Most wanted to receive the conversation aid before the appointment (62.5%), followed by during the appointment or at multiple time points. Respondents wanted to receive it in person (44.5%), by mail (31.5%), electronically (37.5%), or a combination of ways. Most wanted to review it with their PCP (53%), followed by a colonoscopist, family or medical assistant, or a combination of people. See Appendix 9.

Physician survey. In Fall 2021, there were 25 PCPs and 25 gastroenterologists from across the US. Most (68%) were moderately or extremely familiar with conversation aids. 82% were somewhat or very willing to use it in practice. They suggested sending the conversation aid through the patient portal (48%) or through mail (36%) before the appointment, use a paper-based version during the appointment (76%), and give the patient a copy after the appointment for consideration. See Appendix 10. Qualitative analysis of the free-text responses showed that most participants said that using the conversation would be helpful during a routine or annual wellness visit and generally covered their usual discussions of risks regarding repeat colonoscopy. They noted that the conversation aid would be most useful for patients who are undecided about repeat colonoscopy.

“For people on the fence about whether to get one, I think this would be especially useful.” - PCP in Texas

“It would fit well within usual care, provide some numbers to back up qualitative statements.” - Gastroenterologist in Oregon

Aim 3

Aim 3 was conducted between August 2022 and November 2023 and includes results from a pre-appointment patient demographics survey, quantitative observation of SDM during the consultations where the conversation aid was used, and qualitative analysis of patients’ perspectives post-appointment and PCPs’ perspectives post-study. See Table 1 for patient demographics of those who took the survey, had a consultation recorded, and completed a post-appointment interview.

Table 1.

Aim 3 patient characteristics (n = 17)

	n (%)
Age range
70–74	6 (35)
75–79	7 (41)
80–84	4 (24)
Preferred gender
Male	8 (47)
Female	9 (53)
Race
White	16 (94)
Prefer not to respond	1 (6)
Education
High school graduate	1 (6)
Some college	2 (12)
Associates or bachelors degree	6 (35)
Graduate degree	7 (41)
Other	1 (6)
Health insurance
Medicare with Medicaid	1 (6)
Medicare with supplemental	15 (88)
Other	1 (6)
Health literacy
Adequate	13 (76)
Marginal	3 (18)
Inadequate	1 (6)
Considering repeat colonoscopy
Yes	12 (71)
No	5 (29)

Five internal medicine PCPs participated in the pilot, three female (60%) and two male (40%). The mean time since medical school graduation was 28.2 (range 12–38) years.

From 17 recordings of appointments between eligible patients and participating PCPs, the average OPTION-5 rating was 8.3 (range 1.5–14) with a weighted Cohen’s kappa of 54.7%, suggesting a weak level of agreement [34].

Findings from qualitative interviews with the patients

Patient interviews lasted 8 min on average (range 5–15 min). We were not able to conduct a formal analysis of collaboRATE given the inconsistency in responses due to the qualitative nature of the interviews [35]. However, interview participants noted high levels of SDM based on their responses.

“Oh, she made the effort, she really did.” - ID 24 (Female, 74 years old, Adequate health literacy)

Patients noted that there was a high level of comfortability of their PCP using the conversation aid with them and that the conversation aid was delivered and used with them at the right time.

“He seemed very comfortable.” - ID 07 (Male, 78 years old, Adequate health literacy)

“Yeah, it was [good timing]. I would be due for a colonoscopy later this year.” - ID 10 (Male, 81 years old, Adequate health literacy)

However, most patients did not find that using the conversation aid with their PCP affected their decision to pursue a repeat colonoscopy, as they already felt informed and had their minds made up. They felt that the conversation aid might be more useful for patients who are on the fence about whether or not to pursue a repeat colonoscopy or who are considering their first repeat colonoscopy.

“I know I have to have them. I know I’ve had polyps. I know I have problems. […] Well, [it - the conversation aid - would be helpful] for someone who doesn’t necessarily have problems in the past, might have problems now and doesn’t really know ” - ID 24 (Female, 74 years old, Adequate health literacy)

“I was already going to do it.” - ID 08 (Male, 81 years old, Adequate health literacy)

Overall, patients recommended that the conversation aid be used in clinical practice in the future. They suggested the patient portal as a helpful way to deliver the conversation aid, but they emphasized the importance of using a paper copy in person during the consultation. If there are additional colon cancer screening options in the future, they agreed that adding that content to the conversation aid would be helpful.

“I do [think conversation aids should be used with a patient like themselves], because there are many…who do not understand.” - ID 24 (Female, 74 years old, Adequate health literacy)

“I like the portal, and I do use the portal all the time.” - ID 20 (Female, 71 years old, Adequate health literacy)

“I think the paper, getting it all in front of you is probably a good way to communicate all the points that you want to make.” - ID 08 (Male, 81 years old, Adequate health literacy)

Findings from qualitative interviews with the PCPs

PCP interviews lasted 18 minutes on average (range 10–35 minutes). PCPs varied their use of the conversation aid with patients depending on their previous knowledge of their patients and patient reactions.

“Yeah, there were some patients who really wanted to hear every little bit of it. And others who were like fine, fine, fine…Yeah, I know.” - PCP #2 (Female, > 35 years since medical school graduation)

“Being a PCP and knowing their preferences…yeah, I would go through it a little bit quicker. […] So I would do less in depth for people who I knew what their preferences were already.” - PCP #5 (Female, 20 years since medical school)

Most participating PCPs indicated they are interested in using the conversation aid in the future and recommend that other colleagues use it with their patients. The PCPs suggested using the patient portal and sending the conversation aid ahead of time as delivery strategies to support future implementation and efficiency of the SDM process.

“If you had them (conversation aids) in the rooms and you’d have them in folders, I would definitely grab it now and then if it was appropriate. […] We have a lot of resources in the rooms, and so this is something I think people would find beneficial.[…] I think it’d be useful if the patient had seen it ahead of time.” - PCP #5 (Female, 20 years since medical school)

“If they have an annual wellness visit coming up when we know we’re going to touch on this […] I think some patients would review it ahead of time. […] A lot of my patients prefer to get things over myDH (patient portal), and most people who have myDH activated do prefer to get information that way.” - PCP #4 (Female, 12 years since medical school)

Discussion

This study demonstrates the early promise of a novel conversation aid to facilitate SDM around surveillance colonoscopy in older adults with a history of colon polyps, a preference-sensitive decision often complicated by competing health risks, limited life expectancy, and uncertainty about the benefit of continued screening. In Phase 1, older adults provided feedback that led to important refinements in the aid’s content and format, including the addition of illustrations. In Phase 2, national survey results showed that both patients and PCPs found the aid clear, useful, and appropriately timed, with preferences for receiving it before appointments and reviewing it during clinical visits. In Phase 3, pilot testing in primary care visits showed that the aid was particularly helpful for patients unsure about pursuing further colonoscopy. PCPs adapted its use to individual patient needs, and SDM scores reflected moderate engagement. Overall, the conversation aid was viewed as feasible to implement and valuable for promoting informed decisions. Across study phases, findings were broadly consistent: older adults and PCPs preferred receiving the conversation aid before or during clinical visits, emphasized flexible use tailored to patient preferences, and viewed the aid as most helpful for patients experiencing decisional uncertainty.

These findings build upon and extend prior literature highlighting the utility of decision aids for colorectal cancer screening, which have shown improvements in patient knowledge, engagement, and reduced decisional conflict, especially in older populations [14–16]. Our results are consistent with Schoenborn et al., who found that older adults preferred nuanced discussions about screening cessation, especially when framed by individual health priorities [19].

Compared to earlier decision support tools, which have focused predominantly on average-risk screening decisions, the conversation aid in this study addresses a significant gap in supporting surveillance decisions among polyp-bearing older adults [14]. While many patients had already made up their minds about repeat colonoscopy, those who were uncertain found the conversation aid especially beneficial, echoing findings from Scalia et al., who emphasized that decision aids are most effective when patients experience decisional conflict [15]. The study’s findings also resonate with research by Calderwood et al., which underscores a disconnect between provider uncertainty and patient misperceptions of polyp malignancy risk [11, 12]. Our conversation aid may help close this gap by offering evidence-based, accessible information in a format that promotes individualized discussion.

The multi-phased, user-centered design approach is a notable strength, contributing to the conversation aid’s usability and relevance. By incorporating patient and PCP feedback iteratively, the conversation aid reflects real-world needs and preferences. The estimate of mortality from colonoscopy for older adults used in the conversation aid was based on the best available evidence at the time. We acknowledge that the two most relevant references were from the early 2010s [5, 27], and while one focused on the indication of screening [27], the other included a mix of screening and diagnostic indications [5]. Additional research should be undertaken to better understand the attributable risk of colonoscopy by age, co-morbidities, and colonoscopy indication. In the meantime, the conversation aid is a starting point for communicating potential harms of colonoscopy, including mortality, as well as the uncertainty around these estimates for an individual person. The pilot sample was small, predominantly white, highly educated, and included many patients who had already made decisions regarding surveillance colonoscopy. This may have limited the observed impact of the conversation aid on decision outcomes; however, the tool appeared most beneficial for patients experiencing decisional uncertainty. Furthermore, the use of the OPTION-5 tool to evaluate SDM, alongside qualitative feedback, offers a comprehensive perspective on how the conversation aid influenced decision-making dynamics in clinical encounters. However, there was a weak level of agreement between the independent raters, and only a small number of recordings in total and per participating PCP. We also did not assess patient knowledge before or after use of the conversation aid, which represents a minor limitation; however, prior evidence demonstrates that decision aids for older adults reliably improve knowledge and informed decision-making [16].

Future work should focus on the potential impact of this conversation aid on surveillance and other health outcomes in broader populations. Although participants described the conversation aid as clear and useful, acceptability alone does not imply impact on decisional quality or health outcomes. This study represents an early, feasibility-focused step, rather than an evaluation of effects on decisions, surveillance utilization, or clinical outcomes. We acknowledge the many competing priorities for older patients and their PCPs during clinical encounters, which are often time-limited. Integrating the conversation aid into EHR and patient portals to streamline delivery and usage, especially for older adults engaged in telehealth may be useful. Personalization of the conversation aid based on polyp history, age, and comorbidity profile, using a dynamic interface that automatically pulls data from the EHR, could enhance clinical utility and efficiency. The conversation aid also offers the opportunity to incorporate emerging non-invasive alternatives for colorectal cancer prevention in patients with a history of polyps, such as fecal immunochemical tests (FIT) [23–25]. Importantly, future efforts should also consider how best to identify and engage patients who are undecided about continuing surveillance, as they appear to derive the most benefit from the conversation aid. Since the tool is not intended for first-time screening decisions, refining its application for this specific subgroup could increase its overall effectiveness in clinical practice.

Conclusions

This multi-phase study explored the development and implementation of a conversation aid to support SDM around surveillance colonoscopy among older adults with a history of polyps. The conversation aid was consistently described as clear, helpful, and well-timed, particularly for patients facing decisional uncertainty. PCPs found it adaptable to their workflows and responsive to patient needs. As colon polyp surveillance becomes more individualized in an aging population, this conversation aid offers a practical, evidence-informed approach to facilitate informed, values-based decisions. Its integration into routine clinical practice could enhance the quality of decision-making and promote more patient-centered care.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

DS: Conceptualization, Methodology, Formal Analysis, Investigation, Resources, Writing - Original Draft, Writing - Review & Editing, Visualization, Project Administration; ARD: Formal Analysis, Investigation, Writing - Review & Editing; SX: Formal Analysis, Investigation, Writing - Review & Editing; PRR: Investigation, Writing - Review & Editing; MAD: Conceptualization, Writing - Review & Editing; SKL: Conceptualization, Writing - Review & Editing; AHC: Conceptualization, Methodology, Formal Analysis, Investigation, Resources, Writing - Original Draft, Writing - Review & Editing, Supervision, Project Administration, Funding Acquisition. All authors read and approved the final manuscript.

Funding

We acknowledge funding and resources provided by NCI P30CA023108 and the Dartmouth Cancer Center Hopeman Foundation Fund (Calderwood).

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to human subjects approval restrictions, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethics approval for this study was provided by Dartmouth Health’s Institutional Review Board on June 23rd, 2020, and all participants in the study consented or provided verbal consent to participate. Our study complied with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.