Surgeon- and Practice-Directed Interventions to Reduce Overtreatment of Low-Risk Thyroid Cancer

Key Points

Question

What barriers and facilitators exist for implementing interventions to reduce overtreatment of low-risk thyroid cancer?

Findings

In this qualitative study, participating surgeons discussed 3 key barriers to successful integration of interventions: (1) the complexity of thyroid cancer; (2) the perception that interventions were appropriate for other practices but not one’s own; and (3) surgeons’ professional identity impacting practice change.

Meaning

This study suggests that an acceptable and effective intervention for reducing overtreatment of low-risk thyroid cancer may take different forms, provided it considers both behavior change and the end user.

Abstract

Importance

Overtreatment of low-risk thyroid cancer (LRTC) exists worldwide, with many patients unnecessarily undergoing total thyroidectomy. Effective surgeon- and practice-directed interventions to reduce overtreatment are needed.

Objective

To identify barriers to and facilitators of implementing surgeon- and practice-directed interventions aimed at reducing total thyroidectomy for LRTC.

Design, Setting, and Participants

This qualitative study, conducted between June 1, 2024, and February 28, 2025, used a focused ethnography that included semistructured interviews of 19 thyroid surgeons and site visits to 7 community and academic practices where thyroidectomy is performed.

Main Outcomes and Measures

Participants ranked 6 potential interventions for reducing overtreatment of LRTC and identified implementation barriers and facilitators. Verbatim transcripts were analyzed using thematic analysis to identify the most acceptable aspects of the interventions. Ethnographic field notes provided context for interview results.

Results

The 19 participants were predominantly male (13 [68.4%]), with a mean (SD) age of 46.6 (8.2) years. Participants’ opinions on the usefulness of the proposed interventions differed. However, when asked to explain their intervention rankings, participants’ reasoning for why they preferred particular interventions followed similar patterns. These patterns led to 3 key themes for the development of an effective intervention: (1) thyroid cancer is complex (a one-size-fits-all approach would not allow for nuanced decision-making); (2) although participants recognized the value of an intervention, they believed interventions were not appropriate for their own practice; and (3) surgeon professional identity can impact practice change, resulting in either reluctance toward practice change or motivation to change. Barriers to implementation observed during site visits were associated with differences in the clinic environment (eg, completing clinical demands, physical infrastructure) and across staffing and staff interactions (eg, presence of trainees, patient-surgeon interactions).

Conclusions and Relevance

In this qualitative study of thyroid surgeons, key factors were identified for the development of an intervention for reducing overtreatment of LRTC. Participants’ assessment of proposed interventions indicated a successful intervention must be seen as useful, surgeons may be motivated through performance feedback, and the intervention must fit well into the existing surgical workflow. Results demonstrated that an acceptable and effective intervention in this setting may take many forms as long as the intervention considered both behavior change and the end user.

This qualitative study of thyroid surgeons examined barriers to and facilitators of implementing surgeon- and practice-directed interventions aimed at reducing total thyroidectomy for low-risk thyroid cancer.

Introduction

Although 10 years have elapsed since the American Thyroid Association (ATA) guidelines initially endorsed less-extensive management with thyroid lobectomy or active surveillance for small, low-risk thyroid cancer (LRTC),¹ most patients with LRTC in the US continue to receive the most extensive treatment: total thyroidectomy.^2,3 Individuals with LRTC treated with lobectomy or total thyroidectomy have similar long-term oncologic outcomes,^4,5,6,7,8 but overtreatment with total thyroidectomy can negatively impact patient-reported quality of life and cause unnecessary financial burden.^9,10,11 To decrease overtreatment, interventions are needed to discourage unnecessary use of total thyroidectomy for LRTC.

When discussing the extent of their treatment, patients place substantial trust in surgeons, whose recommendations can independently contribute to overuse of total thyroidectomy.^11,12 Therefore, interventions aimed at surgeons and their practices are crucial. Previous interventions for LRTC have focused on treatment-related factors^13,14,15 or shared decision-making in the patient-surgeon interaction,^16,17,18 with minimal broadscale effects. Additional understanding is needed about the implementation of surgeon- and practice-directed interventions, including barriers and facilitators for successful implementation.

Within this context, this study aimed to identify barriers and facilitators for implementing surgeon- and practice-directed interventions directed toward reducing overtreatment for patients with LRTC via assessment of 6 potential interventions and practice observation. The interventions, identified through a systematic process,¹⁹ included designs centered on (1) performance feedback, (2) consensus-building skills, (3) clinical algorithms, (4) education materials for patients and families, (5) patient-physician communication skills, and (6) disease-related education for surgeons. Understanding surgeons’ views of potential interventions prior to adaptation for the thyroid cancer context is crucial for ensuring success and lasting practice change.

Methods

Study Design

For this qualitative study, we designed a focused ethnography to understand what type of intervention would be most useful in clinical settings. We conducted semistructured interviews to understand and describe surgeon perspectives and conducted observational site visits to understand clinical flow. Interviews incorporated a ranking exercise to evaluate surgeons’ attitudes and beliefs about 6 potential interventions to reduce overtreatment for LRTC. This study was approved by the University of Michigan Medicine institutional review board. Participants were informed about study objectives and provided written or electronic consent prior to participation. This study followed the Consolidated Criteria for Reporting Qualitative Research (COREQ) guideline.²⁰

Interview Participants and Recruitment

Purposive sampling was used to recruit surgeons with a broad range of practice types, specialties, annual thyroidectomy volume, and geographic locations in Michigan. Surgeons were eligible to participate in interviews if they were actively practicing and performed thyroidectomy for thyroid cancer. Participants were recruited via flyers that were distributed at 3 professional conferences. Sample size was informed by the principles of information power.²¹ No study team member involved in data collection had existing relationships with participants.

Data Collection

Data were collected through semistructured interviews and ethnographic observations at outpatient clinics in Michigan. The semistructured interview guide was developed by study team members (C.B.J., M.E.B., and S.C.P.) to gauge participant opinions about 6 interventions. Participants were provided with definitions of each intervention, asked to rank their top 3 interventions, explain their selections, and identify any potential barriers to implementation (Table 1). Participants were also asked about their clinical practices and how they make treatment decisions, including processes (eg, referring physicians, practice culture, patient preference) that impact decision-making.

Table 1. Intervention Definitions and Ranking.

Intervention	Definition	Ranked in top 3 interventions, No. (%) (N = 19)
Performance feedback	A process that involves monitoring and providing individualized feedback on performance to surgeons, usually in the form of systematic reports on individual- or practice-level data. The feedback aims to motivate and direct improvements in clinical practice, particularly through the influence of peer comparisons on surgeon behavior.	12 (63.2)
Clinical algorithms	Diagrams or flowcharts that provide a clear, step-by-step guide designed to aid surgeons in determining the most appropriate treatments based on current best practices. These algorithms are designed to simplify guidelines and be easily integrated within clinical routines, such as educational posters to prompt recommended practices.	11 (57.9)
Consensus-building skills	A technique used to establish agreement among surgeons and other health care professionals on medical practice or guidelines. It uses local discussions, needs assessments, and structured methods to reach consensus on care pathways or appropriate actions.	11 (57.9)
Education materials for patients and families	Tailored resources developed to improve patient knowledge of their health condition and treatment options. These materials are designed to be accessible and comprehensive to facilitate informed patient discussions and decision-making in collaboration with their surgeon.	11 (57.9)
Patient-physician communication skills	An educational process designed specifically for surgeons to enhance patient comprehension of risk-benefit information and increase patient engagement in treatment decision-making. These approaches emphasize the importance of providing patients with the information needed to make well-informed treatment choices in collaboration with their surgeon.	11 (57.9)
Disease-related education	An approach to improve health care outcomes by educating surgeons using tailored strategies to improve knowledge of guideline updates and expected outcomes. It includes active learning from experts and opinion leaders to address health care disagreements, create expectations of adherence, and facilitate patient education.	7 (36.8)

Interviews were conducted virtually via Zoom or in person by 3 qualitative experts (2 women: S.E.B. and M.E.B; 1 man: J.D.T) between July 1, 2024, and January 31, 2025. Interviews lasted 45 to 60 minutes and were recorded on encrypted devices, deidentified, and transcribed verbatim for analysis. Participants provided verbal informed consent to be recorded and for quotes to be used in dissemination. All participants completed a demographic survey after the interview and self-identified their sex and race and ethnicity. Participants were asked to self-identify their annual thyroidectomy volume compared with other surgeons to capture surgeon self-concept and perception of their practice. Interviews were conducted prior to the release of 2025 ATA guidelines.²² Participants did not provide feedback on transcripts or findings.

Sites for ethnographic observation were selected to maximize representation of different regions of the state, urban or rural status, practice type (community vs academic), and surgeon and facility total thyroidectomy volume. Visits occurred between June 1, 2024, and February 28, 2025. Field notes were taken using a structured field note guide (ie, description of activities, participants, setting, workflow, and interpersonal interactions witnessed) during site visits to provide context for participant interview responses. At each site, research team members took notes on the facility setting and workflow, spoke with clinical staff, and observed interactions between patients and clinicians if possible.

Statistical Analysis

A team-based analytical approach involving regular discussion by all team members was used for data analysis of the semistructured interviews.²³ Deidentified transcripts were coded in MAXQDA, version 24.11.1, a qualitative data management software (VERBI Software GmbH). At the beginning of the data analysis process, 3 separate interview transcripts were coded by team members with a background in endocrine surgery (C.B.J. and S.X.) and qualitative methods (S.E.B.). This process resulted in the creation of inductively derived codes and the refinement of code definitions. Remaining transcripts were divided and individually coded. Coders met regularly with the rest of the project team to discuss patterns, identify themes, and resolve discrepancies through discussion. Participant feedback on each intervention was coded for barriers and facilitators specific to each intervention. Higher level thematic analysis was also conducted to identify cross-cutting themes.²⁴ Because clear consensus could not be achieved among participants about which intervention strategy was perceived to be most effective, we instead aimed to identify which aspects of each intervention were most appealing to participants. Field notes from site visits were used to triangulate interview data.

Results

A total of 19 surgeons (mean [SD] age, 46.6 [8.2] years; 13 men [68.4%] and 6 women [31.6%]; 3 Asian [15.8%], 1 Middle Eastern or North African [5.3%], and 15 White [78.9%]; 12 academically affiliated [63.2%]) participated in the semistructured interview (Table 2). Nine of 19 surgeons (47.4%) said they have a much higher volume of thyroidectomies than other surgeons. Seven site visits were conducted and included university-affiliated (n = 3), community (n = 2), private (n = 1), and cancer center (n = 1) clinics. Interviews conducted during site visits were conducted before or after participants’ clinic sessions (n = 10). Additional interviews were conducted during conference meetings or via Zoom (n = 9).

Table 2. Participant Demographics.

Characteristic	Total, No. (%) (N = 19)
Age, mean (SD), y	46.6 (8.2)
Sex
Female	6 (31.6)
Male	13 (68.4)
Race and ethnicity
Asian	3 (15.8)
Middle Eastern or North African	1 (5.3)
White	15 (78.9)
Specialty
General surgery	5 (26.3)
Endocrine surgery	10 (52.6)
Otolaryngology	4 (21.1)
Self-reported total thyroidectomy volume compared with average surgeon (n = 18)
Much higher	9 (50.0)
Higher	2 (11.1)
About the same	3 (16.7)
Lower	4 (22.2)
Much lower	0
No response	1

All participants were asked to rank the 6 potential interventions (Table 1) and explain their reasoning for their top and bottom 3 selections. Participants differed greatly in their opinion on which interventions were most useful to their practices. Performance feedback was the most popular intervention strategy (ranked in the top 3 interventions by 12 participants), followed by clinical algorithms (n = 11), consensus-building skills (n = 11), education materials for patients and families (n = 11), patient-physician communication skills (n = 11), and disease-related education (n = 7). Participants provided support for all 6 interventions and recognized their usefulness given a variety of clinical circumstances and patient populations.

Interview Results

Higher-level thematic qualitative analysis revealed similar themes in participants’ reasoning for preference of particular interventions. We report these themes to better explain how participants consider implementing practice change and to identify desirable features of a potential intervention. Three key themes were identified (Table 3): (1) thyroid cancer is too complex for one-size-fits-all approaches; (2) interventions were “not for them,” or that, although participants recognized the value of an intervention, they believed it was not appropriate for their own practice; and (3) surgeon professional identity can impact practice change.

Table 3. Illustrative Quotes.

Theme	Subtheme	Illustrative interview quote (participant identifier and specialty)
Thyroid cancer is complex	Generalized patient education may give patients things to worry about that are not relevant to their condition	[Patient education is] a double-edged sword. On one hand, I think it’s often very helpful and I think that it helps patients be engaged and activated and they thought more about these things before they come in. You can have a more guided conversation about what they’re actually worried about…but on the other hand I think that sometimes patients will fixate on things. (Surgeon 1; endocrine surgeon)
	Thyroid cancer is not algorithmic	I really feel like it’s almost impossible not to approach thyroid cancer as individualized or personalized medicine. So, algorithms help, however…it’s amazing how…every patient is different. (Surgeon 9; endocrine surgeon)
	Performance feedback may not be valuable if the underlying data are not robust	Performance feedback is interesting, I would have to understand better how that performance is gathered ’cause it’s always that age-old expression of garbage in, garbage out. (Surgeon 16; otolaryngologist)
	Performance feedback must not be oversimplified and must include clinical nuances	Sometimes performance metrics, if it’s only looking at the size of the nodule, but that’s not picking up on the fact that the patient has a CHEK2 mutation or the patient’s grandma died of thyroid cancer, right? I think that those are different factors that can’t be captured. (Surgeon 1; endocrine surgeon)
Interventions are not for them	Interventions are not as relevant for specialty-trained surgeons	I know most of thyroid surgeries are done by general surgeons in this country, but I hope most endocrine surgeons have enough education to be up to date with their knowledge and skills. (Surgeon 9; endocrine surgeon)
	Disease-related education can benefit lower-volume sites	I can read what [literature] comes through…. It’s…almost never about thyroid cancer…. Since I’m a general surgeon in a smaller town, I’m doing hernias and gallbladders and cancer cases just as much as I’m doing thyroid cases. In fact, probably even more…and so it would be useful for me to get updates periodically. (Surgeon 5; general surgeon)
	Interventions must include other roles beyond the surgeon in order to meaningfully shift practice change	As an endocrine surgeon, I should be up to speed on the ATA [American Thyroid Association] guidelines. But a lot of…primary care docs who make the referrals, sometimes it’s the radiologist, the IR [interventional radiology] docs, potentially may not be aware. (Surgeon 6; endocrine surgeon)
Surgeon identity impacts practice change	Surgeons may be reluctant to change practices if they believe they are already consistently using best practices	Maybe experienced surgeons will say, “Oh, I don’t need to carry a pocket card. I already know this.” I mean, some kickback or pushback, from that type of experience, ego kind of base. (Surgeon 17; general surgeon)
	Surgeons believe that communication is a skill that does not need to be taught because it is naturally learned over time	I’ve seen many different styles [laughs] of patient-physician communication, but in the end, I think in general, we do a good job at that. And I don’t really see that being a barrier to…practice pattern changes. (Surgeon 14; endocrine surgeon)
	Comparative data may foster a competitive drive in surgeons	We look at our data compared to everybody else’s data and there’s some competitiveness to that. So I think that that would be a very motivating factor. (Surgeon 17; general surgeon)
	Not all surgeons are motivated by performance feedback	I find that feedback doesn’t help me much…. I just don’t get encouraged by the feedback. I don’t know that it incites change in me. Maybe that’s not good on my part,…maybe I need to be a little bit more open-minded…but the performance feedback, I always take with a grain of salt. (Surgeon 7; endocrine surgeon)

Theme 1

When discussing interventions for practice change, participants commonly reported that the “nuance” of LRTC and its treatment are challenges to designing effective interventions (surgeon 7; endocrine surgeon) (Table 3). Participants stated the difficulty of a “one-size-fits-all” approach (surgeon 11; endocrine surgeon), as only “very specific” (surgeon 16; otolaryngologist) interventions could support practice change.

Although participants often discussed the need for reliable and accurate information for patients, patient education materials could be “a double-edged sword” (surgeon 1; endocrine surgeon), helping to guide conversations but also causing unnecessary worry (Table 3). Other participants believed resources should be “tailored specifically” for patient populations and education levels and focus on “what’s actually applicable” (surgeon 16; otolaryngologist) to the patient’s thyroid cancer.

When asked about clinical algorithms, many participants thought thyroid cancer, and its treatment options, were not straightforward enough (Table 3). Both clinical and nonclinical factors can mean patients “don’t fit into a neat little flow diagram...because there could be some exceptions that maybe aren’t captured” (surgeon 17; general surgeon). Surgeon 14 (endocrine surgeon) described, “It’s not the case probably that everyone with a 3-cm cancer should have a lobe…that everyone with a 5-cm cancer absolutely needs a total.” Referencing nonclinical patient factors, surgeon 13 (endocrine surgeon) explained, “There are so many things in what we do that just fall off traditional algorithms…. When you have a patient in front of you who’s crying, who wants you to take it out, I mean, I’m gonna take it out, right? ...An algorithm’s not gonna allow you to have a holistic view of the patient.”

Some participants also raised concerns about surgeon audit and performance feedback interventions (Table 3). Participants worried that feedback might be of limited value if underlying data were not robust or if reported metrics failed to capture clinically important contextual factors for surgical decision-making, as commonly used metrics may omit key risk modifiers and patient context. Participants suggested that audit and feedback interventions should account for the clinical nuances and contextual factors that influence surgical decision-making for LRTC, rather than relying solely on simplified or incomplete metrics.

Theme 2

Despite broad acceptance of all 6 interventions, most participants felt they would be more useful for other practices than for their own (Table 3). For example, 1 surgeon anticipated “differential impact depending on…what proportion of somebody’s practice is actually endocrine surgery” (surgeon 3; otolaryngologist), noting that capacity and willingness to change practice depend on thyroidectomy volume. During site visits, clinical staff frequently noted clinics were short on time and overbooked. Field notes indicated that surgeons who believe time is a limited resource may be less likely to want to adopt a practice change that disrupts business as usual.

Participants who considered themselves to have a high volume of thyroidectomies described the interventions as less applicable to their own practice because they believed they were already delivering guideline-concordant care (Table 3). However, these participants and some specialty-trained participants acknowledged that any of the interventions could be more relevant for surgeons with a broader scope of practice such as, “...general surgery practices where there’s people doing…a host of procedures, including thyroidectomy” (surgeon 14; endocrine surgeon). General surgeon participants with a lower volume of thyroidectomies agreed that the interventions would be better suited to general surgery settings. However, many also described the interventions as potentially valuable resources that could strengthen their thyroid cancer care, particularly through accessible, clinically relevant updates.

Despite the volume of patients with endocrine disorders, participants emphasized that practice change is also influenced by other factors, including referring clinicians (eg, primary care physicians, radiologists, and endocrinologists) (Table 3). This suggests that interventions may need to extend beyond surgeons to meaningfully shift practice patterns.

Theme 3

Many participants described the role of surgeons’ professional identity in implementing practice change (Table 3). When professional identity was explicitly discussed as a barrier to practice change, participants emphasized reluctance to changing practices perceived as routine or mastered. When asked about challenges to using clinical algorithms, 1 surgeon anticipated that experienced surgeons might resist decision-support tools. A surgeon during a site visit expressed a similar sentiment when he described that most surgeons think they are doing the right thing or recommending the right treatment, and surgeons may not see the value in practice change if they believe they are consistently delivering quality care.

Participants described communication as an area where surgeons hold strong and sometimes inflexible beliefs, viewing communicating with patients and other care clinicians as something, “a lot of us get good at…over the years” (surgeon 8; endocrine surgeon) (Table 3). These accounts suggest internalized beliefs about competence can limit perceived need for change.

Conversely, several participants noted that surgeon identity can facilitate practice change when it aligns with openness to learning and motivation from comparative data (Table 3). Specifically, participants explained that “there’s some competitiveness” among surgeons (surgeon 17; general surgeon), which can be fostered by audit and performance feedback. Consistent with this observation, performance feedback reports were most commonly ranked among the top interventions.

Participants noted variation in how surgeons respond to change. One surgeon remarked that “a lot of surgeons are pretty stubborn…everyone will take it a little bit differently” (surgeon 8; endocrine surgeon) (Table 3). Overall, participants’ accounts suggest that professional identity, particularly perceptions of competence and openness to feedback, can either impede or support the uptake of practice change interventions.

Observed Barriers to Practice Change at Site Visits

Observation hours at site visits ranged from 2.5 to 6.0 hours (mean [SD], 4.1 [1.5] hours), totaling approximately 43 hours of data collection across the 7 sites. Potential barriers to practice change differed across the visited locations, reflecting differences in clinic environment, workflow, completing clinical demands, patient populations, and surgeon beliefs (Table 4). The clinic environment and workflow shaped how care was delivered at each location, including some situations in which surgeons saw patients across multiple clinics. Clinicians also faced competing demands that varied by site. For example, 1 participant noted that the business of his parathyroid practice left no time to develop thyroid cancer–specific programming. Variations in the physical infrastructure of the clinic environment also impacted clinical workflow, including 1 site that lacked computers or screens in examination rooms, limiting feasibility of any interventions with a virtual component at that site. At another site, variability in access to point-of-care ultrasonography led to patients being scheduled for both total thyroidectomy and lobectomy pending repeat imaging.

Table 4. Field Note Observations and Interpretation for Intervention Development.

Domain	Field note observations	Potential barriers to practice change	Recommendations for intervention development
Clinic environment	Physical differences in facilities can lead to variations in the extent of multidisciplinary interactions (for example, the presence of multidisciplinary or shared workspaces), potentially affecting coordination and consistency of practice change. Use of technology to deliver care differed. Some sites used telehealth appointments while others did not. One site had no computers or screens in the examination rooms to facilitate one-on-one conversations.	Sites with fewer opportunities for multidisciplinary interaction may have less capacity for consensus-building skills. Variations in availability and use of technology might limit the usability of patient education or share decision-making tools that require a virtual component.	Developed interventions should be adaptable to different clinical environments, regardless of facility resources or structure of the clinical environment.
Clinical workflow	The clinic environment and workflow shaped how care was delivered at each location, including some situations where surgeons saw patients across multiple clinics in a single day. At the busiest sites, there were limited opportunities for patient-surgeon interactions (often only 1 or 2 visits before surgery).	Surgeons may have to travel to multiple sites during a day. Also, surgeons at different sites will have different degrees of opportunity to interact with patients prior to surgery.	Surgeons working across multiple clinics could require an intervention that could travel with them (eg, it is easier to carry a pocket card vs hanging a poster on the wall). An intervention that required more than 1 interaction with patients would not be effective at some sites.
Competing clinical demands	Specialty clinics were often described as extremely busy, sometimes overbooked, with substantial administrative burden around scheduling. Clinicians across sites frequently mentioned being overbooked and that it was easy to get behind schedule during a clinic.	An intervention that requires significant practice change or that disrupts existing workflow will be less acceptable because of the heavy workload already experienced by surgeons and other clinicians.	Interventions must be compatible with existing workflow and not contribute additional burden on top of existing clinical demands.
Patient population	Patient population characteristics differed greatly between facilities due to factors such as geography, type of insurance accepted at each facility, and whether or not a facility was a specialty clinic.	Because thyroid patients are not a homogenous population, interventions that are patient directed must be appropriate for a range of patient backgrounds and disease characteristics.	Patient-directed interventions must be adaptable to the needs of a facility’s patient population.
Surgeon beliefs	Observations showed that clinician beliefs about patients can influence how treatment options are discussed, for example, if patients are perceived to have poor health literacy. Surgeons also had different beliefs about the efficacy of potential interventions. For example, 1 surgeon believed that “quality surgery is more important than quality decision-making.” Another noted that most surgeons believe they “are doing the right thing” when it comes to surgical options, so it is challenging to get them to consider practice change.	Surgeons may not buy in to interventions that they do not perceive to be efficacious. For example, they may not distribute educational materials to their patients if they do not believe the materials will be valuable to the patient, use shared decision-making tools if they do not value shared decision-making, or be open to practice change if they do not believe there is a need for change.	Surgeon beliefs about interventions will greatly impact their buy-in and use of provided materials. Surgeon engagement and buy-in should be considered early in the intervention development process, especially if interventions are surgeon directed.

Second, differences in workflow or clinical demands could affect intervention implementation or integration (Table 4). Trainee presence varied by site, and specialty clinics were often described as busy with substantial administrative burden. Patient populations also differed by geography and the types of insurance accepted at each facility, and the research team observed how clinician beliefs about patients influenced discussion. For example, a surgeon at one site believed their patient population had low health literacy and were more likely than patients in other locations to desire total thyroidectomies due to fear of cancer. At the busiest sites, limited patient-surgeon interactions (often only 1 or 2 visits before surgery) further reduced opportunities to discuss treatment options and reinforce practice changes.

Discussion

This qualitative study used a focused ethnography approach to identify potential barriers to practice change and to assess participants’ attitudes toward 6 interventions intended to decrease use of total thyroidectomy for LRTC, as recommended by 2025 ATA guidelines.²² Although participants’ endorsement of various interventions varied, analysis of participants’ attitudes and beliefs about the validity and potential impact of the interventions revealed 3 key cross-cutting themes to be considered in the implementation of any intervention directed to decrease overtreatment of LRTC: (1) thyroid cancer is too complex for one-size-fits-all approaches; (2) interventions were “not for them,” or, only appropriate for other practices; and (3) surgeon professional identity can impact practice change. When considered within the context of the site visit observations, these themes highlight key areas to address when developing surgeon- and practice-directed interventions for reducing overtreatment of LRTC.

Qualitative results revealed the importance of surgeon readiness for change and perceived importance of an intervention for improving outcomes. Some participants were quick to dismiss certain intervention types (eg, clinical algorithm) as ineffective for LRTC due to the complexity and nuance of decision-making and were unlikely to adopt an intervention if they believed they were already providing the best care. Due to the fact that surgeons have been shown to be most likely to adopt practice change interventions they personally believe will lead to improved patient outcomes,^25,26,27 potential interventions to reduce overtreatment of LRTC need to highlight the benefit of thyroid lobectomy on long-term patient outcomes to be perceived as worth any potential disruption as the change is adopted.^28,29 However, as disease-related education for surgeons was the least likely intervention type to be rated in the top 3 interventions, additional barriers to integrating this type of education may exist. Participant feedback and existing literature on surgeon continuing education indicate that the more directed and relevant such education is perceived to be, the more likely it is to be valued.³⁰

Another major barrier to implementation of the interventions was that some participants did not see interventions as appropriate for their practice, even if they saw the benefit for other surgeons. Personal prioritization of practice change is a common characteristic of successful interventions targeting surgeon decision-making around procedures,^31,32 and this study’s results showed that readiness for change can be complicated by the way in which ego can motivate and, alternatively, inhibit clinician-focused practice changes. Previous research on surgeon identity has shown that, because of the importance of professional identity for surgeons, some may be less resistant to practice change if they are able to compare their abilities to those of their peers.^33,34 Study participants noted that some surgeons might feel competitive if they were aware of their peers’ total thyroidectomy rates through performance feedback mechanisms, and performance feedback was the intervention most likely selected in a participant’s top 3 interventions. For these reasons, a successful intervention might address a surgeon’s belief that interventions would benefit only other clinicians by including a component that compares that surgeon’s rates of thyroid surgery with rates of their peers working at other facilities. Providing feedback reports with an appropriate performance comparator has been shown to encourage clinicians to identify discrepancies between their current practice and desirable practice,³⁵ with surgeons especially receptive if the reference group is considered highly relevant and comparable.³⁶

Participant responses, as well as the variations observed during site visits, showed the importance of any intervention fitting into the surgeon’s existing workflow. Previous studies have shown that a lack of fit between intervention and practice setting is one of the most often-cited barriers to implementation.³⁷ To ensure that an intervention fits well into a variety of practice settings, any intervention should be developed with regular, ongoing participation and feedback from relevant stakeholders (ie, surgeons who perform thyroidectomy in a variety of clinical settings). In addition, the support for consensus-building skills as an intervention shows a recognition that there are upstream factors that may influence surgeon decision-making (including expectations of referring physicians). Examining upstream factors at a facility may help identify additional stakeholders who should be involved in decision-making. Previous interventions in health care settings have shown that engagement of targeted users of an intervention in the design phase of implementation research is critical to both feasibility and relevance of practice change.^38,39,40 If surgeons provide feedback through the design and initial implementation of an intervention, they will be able to ensure that the resulting product is as compatible as possible with both the priorities and workflow of the desired end users.

Limitations

This study has several limitations. Our sample included a diverse cohort of clinics in terms of practice setting, but interviews mostly included academically affiliated thyroid surgeons, which may limit generalizability to other settings. In addition, participants ranked potential intervention designs without actual prototypes, which may have impacted feedback. Our interview guide was also grounded in literature about clinician practice change in general and factors associated with total thyroidectomy overuse but did not specifically incorporate theory-based factors associated with clinical decision-making. Nonetheless, the results provide critical information about preferred intervention features and new data about barriers to implementation in different clinical settings. Although we asked surgeons about their self-perception of how their annual rate of thyroidectomy compared with other surgeons, we did not capture actual rates, which limited the ability to compare the results with published criteria for high-volume and low-volume surgeons. Finally, the clinical observations, while crucial for providing context for interview data, did not specifically assess site readiness for practice change. Future studies would benefit from closely following the implementation of a practice change intervention in different clinical settings.

Conclusions

In this qualitative study of 19 thyroid surgeons, 6 different interventions to reduce overuse of total thyroidectomy for LRTC were assessed. No single intervention was strongly favored, but analysis showed that any intervention would be influenced by surgeons’ perception of the fit with LRTC complexity, perception of the appropriateness of an intervention for their own facility, and surgeon identity. These results suggest that an acceptable and effective intervention may take different forms, provided it considers both behavior change and the needs of end users. Next steps in tool development could include a stakeholder-driven process to create an intervention that incorporates these findings. Future research could then focus on assessing the acceptability of intervention prototypes among surgeons, with the goal of sustainable surgeon behavior change around the treatment of LRTC.

Supplement.

Data Sharing Statement