Clinician-Level Knowledge and Barriers to Hepatocellular Carcinoma Surveillance

Key Points Question Which clinician-level factors contribute to underuse of hepatocellular carcinoma (HCC) surveillance in patients with cirrhosis? Findings In this survey study of 347 primary care clinicians (PCCs) and gastroenterology and hepatology clinicians across 5 US safety-net health systems, gaps in HCC surveillance knowledge were identified, particularly among PCCs. Perceived barriers and challenges to HCC surveillance use were identified, including health system factors contributing to persisting delays in timely HCC surveillance despite 3 years after the onset of the COVID-19 pandemic. Meaning These findings suggest that improved HCC education, particularly for PCCs, and health system–level interventions must be pursued in parallel to address the complex barriers contributing to suboptimal HCC surveillance.


Introduction
Hepatocellular carcinoma (HCC) is a leading cause of morbidity and mortality globally. 1,2In the US, the American Cancer Society estimates 41 210 new cases of liver cancer will be diagnosed and 29 380 people will die from liver cancer in 2023. 3Progression of underlying chronic liver disease to cirrhosis is the main risk factor for development of HCC, and the incidence of HCC among individuals who have developed cirrhosis exceeds 1.0% per year, a threshold above which current guidelines recommend HCC surveillance. 25][6][7] However, the use of HCC surveillance among patients with cirrhosis remains suboptimal.A recent systematic review and meta-analysis 8 that included 29 studies with a total of 118 799 patients with cirrhosis reported a pooled estimate for surveillance use of 24.0%.Similarly, a recent analysis 9 in a cohort of more than 2000 patients with cirrhosis found the proportion of time covered by surveillance was 24.9%, with only 16% of patients having semiannual surveillance in the year before HCC diagnosis.Low HCC surveillance use in patients with cirrhosis along with data from the National Cancer Institute's Surveillance Epidemiology and End Results database showing overall 5-year survival in patients with HCC less than 30% highlights an urgent need to improve HCC surveillance.
Underlying causes for low HCC surveillance use are complex and multifactorial, and likely reflect patient, clinician, and system-level factors. 10,11Existing studies [12][13][14][15][16][17] evaluating clinician-level factors have identified gaps in clinician knowledge or familiarity with HCC surveillance guidelines as potential contributors to low rates of HCC surveillance.Studies 13,15,17 focusing predominantly on primary care clinicians (PCCs, including family and internal medicine clinicians) have also identified misconceptions and perceived barriers and attitudes toward HCC surveillance as potential contributors to low rates of surveillance use in patients with cirrhosis.2][23][24][25] Hence studies aimed at better understanding contributors to HCC surveillance specifically among safety-net populations have immense potential to drive interventions to improve HCC surveillance and HCC outcomes in patients with cirrhosis.In particular, elucidating clinician-level factors contributing to underuse or delays in implementation of HCC surveillance is critical to identify potentially modifiable factors, opportunities for education, or other targeted inventions to improve HCC surveillance in patients with cirrhosis.
We surveyed PCC and gastroenterology and hepatology clinicians across 5 safety-net health systems in the US to better understand clinician-level factors that may contribute to underuse of HCC surveillance in patients with cirrhosis.

Methods
This survey study was approved by the institutional review boards (IRBs) of each respective institution and health system.We followed the Consensus-Based Checklist for Reporting of Survey Studies (CROSS).A waiver of documentation of informed consent was granted because this survey study was minimal risk and anonymous.

Study Population
We conducted an anonymous, online web-based survey to a convenience sample of PCCs (internal medicine or family practice) and gastroenterology and hepatology clinicians across 5 safety-net health systems in California, Florida, Louisiana, Ohio, and Texas.We included clinicians who identified themselves as attending physicians, advanced practice clinicians, and residents or fellows.

JAMA Network Open | Gastroenterology and Hepatology
Clinician-Level Knowledge and Barriers to HCC Surveillance

Survey Information
The survey was developed using a conceptual model based on Social Cognitive Theory, 26 Theory of Reasoned Action, 27 and Theory of Planned Behavior, 28 which has been previously used and adapted to evaluate clinician practice patterns related to HCC surveillance in cirrhosis 13,15,16 The questions included in the survey (eMethods in Supplement 1) aimed to assess knowledge, attitudes, beliefs, perceived barriers, and COVID-19-related disruptions in HCC surveillance in patients with cirrhosis.
The survey was divided in 6 sections and included a combination of questions and clinical scenarios with multiple choice answers to assess recommended HCC surveillance strategies (eg, when to initiate surveillance, frequency of surveillance, modality to implement surveillance), and a series of statements assessing attitudes, perceptions, and beliefs regarding HCC surveillance using a 4-point Likert Scale.An additional set of questions assessed how the COVID-19 pandemic and transitions in care affected HCC surveillance.Demographic characteristics, such as gender and self-reported race and ethnicity, were also collected.Race and ethnicity data were collected because other studies have suggested that race and ethnicity concordance between patient and clinician may influence a patient's receptiveness of completing recommended testing.Categories were American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or other Pacific Islander, White, and other.
The survey took approximately 15 minutes to complete.Invitations to participate in the survey were distributed via email between March 15 and September 15, 2023, with an embedded survey link to a REDCap-linked survey.After the initial email invitation, reminder emails were sent every 2 weeks for a total of 6 weeks.

Statistical Analysis
Descriptive statistics of the survey respondents used frequency and proportions.Not all survey respondents completed all of the questions in the survey.The denominator used to calculate proportions for each question or category was based on the total number of respondents for each question and excluded those who did not answer the question.The denominator used to calculate the proportions for each question or category are indicated in the first column of each table.HCC surveillance knowledge was assessed with 6 questions that queried the respondent's ability to correctly identify timing of guideline-concordant surveillance and the diagnostic modality with which to perform surveillance (eTable 2 in Supplement 1).One point was given for each correct answer, with a total of 6 possible points for complete HCC surveillance knowledge assessment.Based on the distribution of responses, comparisons of the proportion of respondents answering 5 to 6 questions correctly vs 0 to 4 questions correctly were stratified by clinician characteristics and analyzed with χ 2 methods.Perceived barriers to HCC surveillance based on previous studies were assessed with 8 statements that evaluated factors, such as adequate time during visits to discuss surveillance, language barriers, lack of diagnostic radiology resources, concerns regarding patient out-of-pocket costs, and challenges arranging follow-up testing and/or treatment.The proportion of respondents endorsing 1 or more of these barriers were similarly stratified by patient characteristics and compared using χ 2 methods.Next, we sought to understand whether there were major differences in perceived barriers to HCC surveillance between PCCs (internal medicine and family medicine clinicians) and gastroenterology and hepatology clinicians.Differences in perceived barriers to HCC surveillance as well as beliefs, attitudes, and perceptions toward HCC surveillance were stratified by PCC vs gastroenterology and hepatology clinicians.Similarly, the perceived impact of disruptions associated with the COVID-19 pandemic and HCC surveillance among patients with cirrhosis were compared between PCCs and gastroenterology and hepatology clinicians using χ 2 methods.Statistical analyses were performed using SAS Studio 3.6 on SAS version 9.4 (SAS Institute).Statistical significance was met with a 2-tailed P < .05.Data were analyzed from October to November 2023.

JAMA Network Open | Gastroenterology and Hepatology
Clinician-Level Knowledge and Barriers to HCC Surveillance

Clinician Knowledge of Surveillance
Among respondents who completed the HCC surveillance knowledge questions, 144 of 270 (53.3%) scored at least 5 of 6 correct (Table 1 and eTable 2 in Supplement 1).A higher proportion of gastroenterology and hepatology clinicians had higher surveillance knowledge compared with PCCs (37 of 28 [77.1%]vs 65 of 142 [45.8%];P < .001),and clinicians with less than 10 years of clinical experience had higher knowledge vs those with more than 20 years of experience (84 of 161 [52.2%] vs 29 of 82 [35.4%];P = .046).Clinicians who reported seeing 25 to 49 patients weekly had the greatest HCC surveillance knowledge, which was higher than clinicians who reported seeing 50 or more patients per week (59 of 97 [60.8%] vs 54 of 154 [35.1%];P < .001).No significant differences in HCC surveillance knowledge score were observed by gender or clinician type or training.

Clinician Attitudes Toward Surveillance
Clinician attitudes toward HCC surveillance, stratified by clinician subspecialty, are reported in  Nearly all clinicians, regardless of specialty, did not believe the lack of effective HCC treatments or lack of association with survival contributed to HCC surveillance underuse.Similarly, nearly all clinicians agreed that HCC surveillance is effective at detecting tumors at an early stage and that HCC surveillance is cost-effective in patients with cirrhosis.Finally, nearly all clinicians agreed that better education for PCCs about HCC surveillance is needed.

Impact of COVID-19 on HCC Surveillance
Over 80% of clinicians across both primary care and gastroenterology and hepatology agreed that HCC surveillance was delayed during the pandemic (  5). a Not all respondents answered all questions.
Calculation of proportions were based on the total number of clinicians who answered each question.
The number in parentheses in the first column indicates the total number of respondents for each category.

Discussion
Among PCCs and gastroenterology and hepatology clinicians across 5 safety-net health systems in 5 distinct geographic regions in the US, we identified important potentially modifiable barriers contributing to low HCC surveillance in patients with cirrhosis.First, respondents, particularly PCCs, had suboptimal knowledge regarding the appropriate timing of HCC surveillance as well as appropriate choice of surveillance modality.Specifically, when compared with gastroenterology and hepatology clinicians, PCCs were more likely to cite challenges with accurately identifying patients with cirrhosis as well as feeling not up-to-date with current HCC surveillance guidelines.These observations align with prior studies citing gaps in knowledge contributing to underuse of HCC surveillance, but is unique in focusing on safety-net clinicians and inclusive of both PCCs and gastroenterologists and hepatologists. 13,15,29For example, among 131 PCCs surveyed at a single a Not all respondents answered all questions.
Calculation of proportions were based on the total number of clinicians who answered each question.
The number in parentheses in the first column indicates the total number of respondents for each category.
center health system, Dalton-Fitzgerald 13 noted that 68% of respondents reported feeling that they were not up-to-date with HCC surveillance guidelines.Similarly, among 391 North Carolina PCCs surveyed, only 45% reported ordering HCC surveillance in their patients with cirrhosis, and among those not ordering surveillance, 24% reported being unaware of HCC surveillance recommendations in patients with cirrhosis. 29It is also interesting that survey respondents with less than 10 years of experience had the highest HCC knowledge score, and this could reflect individuals that have more recently completed training and therefore may be more up to date with HCC surveillance guidelines.
While drivers of suboptimal HCC surveillance are multifactorial, incorporating educational interventions or user-friendly decision support tools 30 that can improve HCC surveillance knowledge and disseminate updated HCC guidelines, particularly among PCCs, will be a necessary component of any program to improve HCC surveillance.In fact, nearly 96% of PCCs who responded to our survey agreed that better education about HCC and HCC surveillance are needed for PCCs.While we acknowledge the multifactorial challenges that lead to suboptimal HCC surveillance, effective delivery of education to improve cirrhosis and HCC surveillance knowledge that specifically targets specific gaps or misperceptions is a low-hanging fruit that could improve HCC surveillance use among safety-net health systems.Furthermore, delivery of education to clinicians who use multiple modalities and offered on a recurring basis may be more effective than a 1-time intervention.
However, we acknowledge the challenge of implementing this type of intervention in PCC settings, especially on a recurring basis, given that PCCs have multiple conditions and practice guidelines to stay up to date with across the spectrum of medical care for adults.
In addition to knowledge gaps, 41.6% of respondents in our study reported at least 1 perceived barrier to ordering HCC surveillance, with inadequate time to discuss HCC surveillance during clinical encounters, concerns for patients' out-of-pocket costs, and challenges in arranging follow-up diagnostic testing being the top 3 barriers reported by PCCs.These reported challenges highlight the complex health system level factors that are not unique to PCCs nor unique to safety-net health systems affecting HCC surveillance in patients with cirrhosis. 10In fact, the uniqueness of our study focusing on safety-net populations emphasizes the complex operational challenges that exist in underresourced settings.Programs aiming to improve HCC surveillance particularly in safety-net a Not all respondents answered all questions.
Calculation of proportions were based on the total number of clinicians who answered each question.
The number in parentheses in the first column indicates the total number of respondents for each category.
health systems must also address the health system factors in parallel with clinician and patient specific challenges that exist.We additionally assessed differences in beliefs, attitudes, and perceptions toward HCC surveillance between PCCs and gastroenterology and hepatology clinicians.
Nearly 8% of PCCs believed that current HCC surveillance tools are suboptimal and 55% believe that better data are needed to demonstrate the benefits of HCC surveillance in patients with cirrhosis, both significantly greater than gastroenterology and hepatology clinicians.These observations further emphasize the need for more effective means to share knowledge and disseminate relevant literature and guidelines with PCCs in particular about the outcomes and benefits of HCC surveillance among patients with cirrhosis. 6,31,32ile important clinician-level factors potentially contributing to underuse of HCC surveillance were identified, a unique aspect of our study is the evaluation of COVID-19 pandemic-related disruptions in further exacerbating barriers to timely HCC surveillance among safety-net health systems.Most clinicians reported that patients with cirrhosis often missed appointments and HCC surveillance was delayed due to limitations of in-person visits.4][35] However, even more concerning, among survey respondents in the current study, it is clear that despite being over 3 years post onset of the pandemic, the disruptions in cirrhosis care and HCC surveillance in particular remain, and most respondents do not have an effective mechanism to track and re-engage patients who missed HCC surveillance during the pandemic.This is particularly concerning given that populations who use safety-net health systems have already experienced barriers in timely access to health care even before the onset of the COVID-19 pandemic.[38][39]

Strengths and Limitations
A strength of our study, which focused specifically on safety-net health systems, was that it incorporated survey responses among both PCCs and gastroenterology and hepatology specialists across diverse geographical regions.In addition to assessing knowledge, attitudes, perceived barriers, and beliefs that may affect use of HCC surveillance, the current study is unique in also assessing the COVID-19 pandemic-related impact on HCC surveillance and how those disruptions have persisted into the present day.
However, certain limitations should be acknowledged.Given the survey-based approach, the possibility of recall bias should be considered when interpreting the data.Overall, with the 25.5% response rate, nonresponse bias should be considered as well as social desirability bias, such that respondents answer questions in a socially desirable manner rather than reflecting true practice patterns or beliefs.In addition, some respondents only answered some questions or may have been hesitant to report challenges that were felt to be unique to their health system.We attempted to address some of these concerns by ensuring survey respondents remained anonymous.However, nonresponse and missing data for certain questions may reflect social desirability bias.Another potential limitation is that clinicians who responded to the survey may have been newly employed or may have recently moved from different institutions and thus their responses regarding challenges and barriers, especially as they are associated with health system factors, may be influenced by their experiences at prior places of patient care.Finally, our study collected perspectives of clinicians but does not reflect other challenges associated with surveillance implementation, including patientreported barriers. 40

Conclusions
In this survey study, we identified important gaps in knowledge and perceived challenges and barriers to effective use of HCC surveillance among patients with cirrhosis.Despite these concerning

Table 1 .
Differences in HCC Surveillance Knowledge by Clinician Characteristics lower HCC surveillance knowledge were more likely to report barriers to surveillance (66 of 112 [58.9%] vs 35 of 131 [26.7%];P < .001).Overall, the most commonly reported barriers among all respondents included not having adequate time to discuss HCC surveillance (39 of 187 [20.9%]), concerns about patients' out-ofpocket costs (27 of 187 [14.4%]), and difficulty arranging follow-up diagnostic testing for patients with a positive HCC screening test (17 of 186 [9.1%]).Table 3 compares specific barriers reported by a Not all respondents answered all questions.Calculation of proportions were based on the total number of clinicians who answered each question.The number in parentheses in the first column indicates the total number of respondents for each category.bThe6questions on the HCC surveillance are available in eTable 2 in Supplement 1. with [10.4%];P < .001).Language barriers, shortage of radiology facilities, concerns that patients often do not complete surveillance tests that are ordered, difficulty arranging follow-up diagnostic testing for those with a positive HCC screening test, and difficulty arranging treatment for patient diagnosed with HCC were reported in 10% or less of clinicians and did not differ significantly between PCCs and gastroenterology and hepatology clinicians (Table4).

Table 4
. A higher proportion of PCCs, compared with gastroenerology and hepatology clinicians, believe that current screening tools are suboptimal and miss many HCC (11 of 140 [7.9%]vs 0 of 47

Table 2 .
Differences in Perceived Barriers to HCC Surveillance by Clinician Characteristics P = .048)aswell as endorsing the need for better data to evaluate the benefits of HCC surveillance in patients with cirrhosis (74 of 136 [54.4%] vs 16 of 46 [34.8%];P = .02).Gastroenterology and hepatology clinicians were more likely than PCCs to agree that lack of HCC surveillance poses medical malpractice liability (47 of 48 [97.9%] vs 111 of 137 [81.0%];P = .004).
a Not all respondents answered all questions.bNobarriers reported indicates those who answered no barriers to all questions in Table3.Barriers reported indicated those who answered yes to barriers for any question in Table3.c Including those in training.d Out of 6 correct.JAMA Network Open | Gastroenterology and Hepatology Clinician-Level Knowledge and Barriers to HCC Surveillance JAMA Network Open.2024;7(5):e2411076.doi:10.1001/jamanetworkopen.2024.11076(Reprinted) May 14, 2024 5/13 Downloaded from jamanetwork.comby guest on 05/26/2024 [0%];

Table 5
Furthermore, a large proportion of clinicians felt that COVID-19 pandemic-related delays and barriers associated with HCC surveillance continue to persist at this time, more so among PCCs vs gastroenterology and hepatology clinicians (67 of 135 [49.6%] vs 14 of 45 [31.1%];P = .03)(Table

Table 3 .
Perceived Barriers to Hepatocellular Carcinoma Surveillance by Clinician Specialty

Table 4 .
Beliefs, Attitudes, and Perceptions Toward Hepatocellular Carcinoma Surveillance by Clinician Specialty I do not order liver cancer screening because there are not any effective treatments available (n = 187) Better education for primary care clinicians about liver cancer and liver cancer screening is needed (n = 188)