Abstract
Objective Clinical practice guidelines are not always followed consistently. To better understand potential barriers to the implementation of treatment recommendations in axial spondyloarthritis and ankylosing spondylitis (axSpA/AS), an online survey was conducted.
Methods Email invitations were sent to US rheumatology care providers in January 2023. The questionnaire included 20 questions, with an estimated completion time of 5-7 minutes.
Results One hundred four of 441 (24%) invitees participated, including 80/104 (77%) board-certified rheumatologists and 20/104 (19%) fellows. Survey participants identified UpToDate (85%), treatment guidelines (74%), and colleagues (54%) as relevant sources of knowledge for managing axSpA/AS. Of the participants, 64% and 53% considered themselves to be at least moderately familiar with the American College of Rheumatology/Spondylitis Association of America/Spondyloarthritis Research and Treatment Network (ACR/SAA/SPARTAN) and Assessment of Spondyloarthritis international Society/European Alliance of Associations for Rheumatology (ASAS/EULAR) treatment recommendations for axSpA/AS, respectively. Whereas 69% of participants agreed or strongly agreed that disease activity scores are useful for making treatment decisions in axSpA/AS, only 37% measure patient-reported outcomes (PROs) frequently (≥ 50% of clinic visits) while 82% do so for C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR). PROs are typically recorded during clinic encounters (65%) and CRP/ESR are obtained after the visit (86%). Of the participants, 57% and 47% considered the Bath Ankylosing Spondylitis Disease Activity Index and Ankylosing Spondylitis Disease Activity Score to be at least moderately useful for measuring disease activity in axSpA/AS, respectively; 41% and 37% thought the same about the ASAS 20% improvement criteria and Clinical Disease Activity Index, respectively.
Conclusion Treatment guidelines are an important resource for rheumatologists who manage patients with axSpA/AS. Although there is general agreement that disease activity monitoring is important, the implementation of the respective recommendations is lacking. Reasons may include lack of familiarity and an underdeveloped infrastructure to efficiently collect PROs.
Several treatment recommendations for axial spondyloarthritis (axSpA) and ankylosing spondylitis (AS) have been developed and are regularly updated, including the American College of Rheumatology/Spondylitis Association of America/Spondyloarthritis Research and Treatment Network (ACR/SAA/SPARTAN) recommendations for the treatment of AS and nonradiographic (nr)-axSpA, and the Assessment of Spondyloarthritis international Society/European Alliance of Associations for Rheumatology (ASAS/EULAR) recommendations for axSpA.1,2 The effect of these documents on the care of patients with axSpA/AS is incompletely understood. Studies in multiple areas of medicine have shown that clinical practice guidelines are not followed consistently.3-6 Systematic reviews of barriers to guideline adherence have identified 5 major categories: guideline factors, health professional factors, patient factors, clinical practice (health and organizational) factors, and societal (political and social) factors.7,8 Although a recent systematic review of barriers and facilitators affecting guideline implementation in primary care included 12 systematic literature reviews with 275 original studies,6 research into guideline implementation in axSpA/AS has been comparatively scarce.9
To better understand potential barriers to the implementation of the treatment recommendations for axSpA/AS, we performed a survey among US rheumatology care providers. Here we report the development of the survey and data on disease activity monitoring in clinical practice.
METHODS
Survey questionnaire. An online survey was designed to investigate potential barriers to the implementation of axSpA/AS treatment recommendations, including 5 sections with 4 questions each. The first section collected demographic data, such as training level, practice setting, gender, and approximate number of patients with axSpA/AS seen per week. The other 4 sections asked questions about guideline factors, patient and health professional factors, practice setting factors, and societal factors. The survey was implemented in Research Electronic Data Capture (REDCap).10 Branching logic was used for adaptive questioning when appropriate to collect additional information based on the response to a question. The usability and technical functionality of the survey was tested by sending it to 5 rheumatologists who had agreed to serve as beta testers. Their feedback was taken into account when finalizing the questionnaire. The complete survey is included in the Supplementary Material (available with the online version of this article). Participants were required to answer all questions in a section before moving on to the next; they were not allowed to go back to previous sections after completing a section. The estimated completion time for the final survey with 20 questions was 5-7 minutes.
Study design. Names and email addresses of rheumatology care providers practicing in the New England states (Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont) were identified from a variety of sources, including email lists, scientific publications, and websites of academic medical centers, hospitals, and private practices. In case of missing data, we attempted to extrapolate email addresses from other providers in the same institution with a known email address. We identified a total of 502 names and email addresses. Sixty-one of the 502 invitation emails sent returned a bounce message, suggesting an incorrect or inactive email address. The remaining 441 invitees represented a convenience sample, as we did not have access to a complete listing of all rheumatology care providers in New England.
Survey invitations were sent in January 2023 using REDCap. Each invitee received a unique link to the online questionnaire (ie, closed survey, where each individual could complete the survey only once). Data were collected anonymously. Invitees who had not responded by starting the survey were sent repeat invitations every 3-5 days, for a total of 7 invitations.
As a secondary aim, we investigated the effect of an incentive on the survey response rate by randomizing subjects 1:1 to an intervention vs a control group. In the intervention group, we offered a donation of US $25 to the SAA for every completed questionnaire.11 The emails to the subjects in the incentive group included the statement, “We will donate $25 to the Spondylitis Association of America for every completed questionnaire.” in bold font. The subject line and text of the invitation email were otherwise identical to the control group. To randomize subjects, the 502 candidate study participants were sorted in Microsoft Excel by state and institution, and then assigned in alternating order to the incentive or control group.
Statistics. Data were analyzed using Microsoft Excel and SAS (version 9.4; SAS Institute). Participation rates in the incentive and control groups were compared using the chi-square test. A P value < 0.05 was considered statistically significant. Graphing was done using GraphPad Prism (version 10.0.2; Dotmatics).
Ethical considerations. The study was approved by the Mass General Brigham Institutional Review Board (IRB no. 2020P003836). The invitation email included a fact and information sheet describing the study. Completion of the survey was considered as consent for participation. No personal identifying information was collected. Study design and reporting follow the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) guideline.12
RESULTS
The total number of rheumatology care providers invited to participate in the survey was 441: 219 in the incentive group and 222 in the control group. One hundred four of 441 invitees started the survey, yielding a response rate of 24%. The response rate in the incentive group was 22% (49/219) and in the control group 25% (55/222). There was no statistically significant difference between the groups. Eighty-nine of 104 (86%) participants completed all questions of the survey.
Demographics of the study participants are shown in the Table. Respondents included 80/104 (77%) board-certified rheumatologists, 20/104 (19%) rheumatology fellows, and 3/104 (3%) advanced practice providers (physician assistants or nurse practitioners). Seventy-three of 104 (70%) of the participants worked in academic medical centers, whereas 27 (26%) worked in hospital-based, group, or solo practices. The majority of polled rheumatology care providers (65/104, 63%) indicated that they see < 5 patients with axSpA/AS per week; only 6% see > 10 patients with axSpA/AS per week. Among the board-certified rheumatologists, 34/80 (43%) had 1-10 years of work experience, 23/80 (29%) had 11-30 years of experience, and 23/80 (29%) had ≥ 31 years of experience.
Characteristics of survey respondents (N = 104).
Survey participants identified UpToDate (85%), treatment guidelines (74%), colleagues (54%), and PubMed (46%) as relevant sources of knowledge for managing patients with axSpA/AS. Only 16% refer to textbooks (Figure 1A). Pharmacological therapy was by far the most important reason for consulting axSpA/AS treatment guidelines. Seventy-four percent of survey participants rely on treatment guidelines for questions related to pharmacological therapy, followed by questions about disease activity monitoring (39%), vaccinations (31%), and management of comorbidities (17%; Figure 1B).
Role of treatment guidelines in educating rheumatology care providers about the management of axSpA/AS. (A) Resources consulted for the management of patients with axSpA/AS. (B) Aspects of axSpA/AS management for which treatment guidelines are consulted. (C) Familiarity of survey participants with the ACR/SAA/SPARTAN and ASAS/EULAR treatment recommendations for axSpA/AS. (D) Resources used by study participants to learn about the ACR/SAA/SPARTAN and ASAS/EULAR treatment recommendations. ACR: American College of Rheumatology; AS: ankylosing spondylitis; ASAS: Assessment of Spondyloarthritis international Society; axSpA: axial spondyloarthritis; EULAR: European Alliance of Associations for Rheumatology; PT: physical therapy; SAA: Spondylitis Association of America; SPARTAN: Spondyloarthritis Research and Treatment Network.
Of the participants, 64% and 53% of participants considered themselves to be at least moderately familiar with the ACR/SAA/SPARTAN or the ASAS/EULAR recommendations, respectively (Figure 1C).
Journal articles (55%), board review courses (47%), and the ACR Annual Meeting (43%) were the top 3 resources used by the survey participants to learn about axSpA/AS treatment recommendations (Figure 1D). Social media (6%) and other online resources (5%) were infrequently cited as sources of knowledge.
We then asked survey participants what they thought about the statement, “Disease activity scores provide useful information for making treatment decisions in AS.” Sixty-nine percent either agreed or strongly agreed with this statement, whereas only 8% either disagreed or strongly disagreed (Figure 2A), suggesting that there is strong support for the concept of disease activity monitoring in axSpA/AS.
Disease activity monitoring in clinical practice. (A) Agreement of study participants with the statement, “Disease activity scores provide useful information for making treatment decisions in axSpA/AS.” (B) Frequency of measuring PROs or CRP/ESR in clinical practice. (C) Methods of PRO collection. (D) Timing of CRP/ESR measurements. (E) Type of EMR system used by survey participants. (F) Perceived utility of various disease activity scores for measuring axSpA/AS activity in clinical practice. Numerical values in the figure represent the percentage of participants considering the score to have moderate, high, or very high utility. APP: advanced practice provider; AS: ankylosing spondylitis; ASAS20: Assessment of Spondyloarthritis international Society 20% improvement criteria; ASDAS: Ankylosing Spondylitis Disease Activity Score; axSpA: axial spondyloarthritis; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; CDAI: Clinical Disease Activity Index; CRP: C-reactive protein; DAS28: Disease Activity Score in 28 joints; EMR: electronic medical record; ESR: erythrocyte sedimentation rate; MA: medical assistant; PRO: patient-reported outcome; RAPID3: Routine Assessment of Patient Index Data 3.
The ACR/SAA/SPARTAN guidelines recommend regular-interval use and monitoring of a validated AS disease activity measure and monitoring of C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR).1 Interestingly, only 37% of participants measure patient-reported outcomes (PROs) frequently in their patients with axSpA/AS (defined as ≥ 50% of clinic visits; Figure 2B), and 41% indicated that they do not measure PROs at all. On the other hand, 82% of participants measure CRP or ESR frequently (Figure 2B). Using branching logic, we asked those participants who measure PROs or CRP/ESR when these variables are determined in their practice. PROs are most commonly collected during the clinic visit by a physician or advanced practice provider (physician assistants and nurse practitioners; 63%), and rarely by a medical assistant or nurse (2%). Eighteen percent of survey participants collect PROs at check-in, half of which do so electronically (9%) and the other half using paper forms (9%). The remaining 18% collect PROs prior to the clinic visits (Figure 2C). CRP/ESR values are measured after the clinical visit by the majority of survey participants (86%), whereas only 7% measure CRP/ESR values before the clinical visit so that results are available at the time of the consultation (Figure 2D). Seventy-seven percent of survey participants use the Epic (Epic Systems) electronic medical record (EMR) system in their practice setting; no other single system was used by > 10% of participants (Figure 2E).
When asked about the utility of specific scores to measure disease activity in patients with axSpA/AS, Bath Ankylosing Spondylitis (BASDAI), Ankylosing Spondylitis Disease Activity Score (ASDAS), and Routine Assessment of Patient Index Data 3 (RAPID3) were considered to be at least moderately useful by 57%, 47%, and 44% of participants, respectively.13-15 We included the ASAS 20% improvement criteria (ASAS20), Clinical Disease Activity Index (CDAI), and the Disease Activity Score in 28 joints (DAS28), 3 variables without a role in monitoring disease activity in axSpA/AS, as knowledge controls in this question. ASAS20, CDAI, and DAS28 were considered to be at least moderately useful by about 41%, 37%, and 25%, respectively (Figure 2F). BASDAI and ASDAS were considered to very or extremely useful by 21% and 14%, respectively.
DISCUSSION
We report the results of an online survey among US rheumatologists performed in January 2023. Our findings underscore the central role of axSpA/AS treatment recommendations in educating rheumatologists about axSpA/AS management, especially with regard to pharmacotherapy. Disease activity monitoring using validated scores, recommended by both ACR/SAA/SPARTAN and ASAS/EULAR guidelines, is not commonly performed. Our results advocate for more education and streamlined processes for collecting PROs in order to bolster the adoption of these recommendations.
Treatment guidelines were ranked as the second most commonly used source of information for questions regarding the management of axSpA/AS; only UpToDate was ranked higher. UpToDate is a peer-reviewed evidence-based resource.16 The UpToDate section on axSpA management aligns with the ACR/SAA/SPARTAN and ASAS/EULAR treatment guidelines, giving these guidelines a strong indirect impact. Colleagues were also ranked very high, whereas textbooks appeared to play only a minor role as a resource regarding questions about axSpA/AS management. Education is a major topic when considering the dissemination of treatment guidelines.17 In addition to journal articles, board review courses, the ACR Annual Meeting, and other conferences appear to play important roles for knowledge dissemination, whereas social media and online resources were rarely mentioned by the survey participants. However, we did not ask about specific platforms that may have led to underreporting.18 The sample size was also too small to examine differences between age groups. Our results emphasize the need to better understand physicians’ learning patterns and to incorporate a dissemination plan into updated treatment guidelines.
Whereas 69% of survey participants agreed or strongly agreed that “disease activity scores provide useful information for making treatment decisions in axSpA/AS,” only 37% measure PROs frequently (in 50% of clinic visits or more); 82% measure CRP/ESR frequently. Several factors may explain this discrepancy. Although the US Food and Drug Administration drug labels for AS and nr-axSpA require active disease, no thresholds are defined and documentation of an elevated disease activity score is generally not required for prior authorization by insurance plans in the US (societal factor). Treat-to-target is not recommended in the ACR/SAA/SPARTAN recommendations and there is no clear-cut evidence that treat-to-target is superior to standard care,19,20 which may reduce the enthusiasm of rheumatologists to endorse disease activity monitoring in clinical practice (health professional factor). We included a question on provider knowledge about disease activity monitoring. Although BASDAI, ASDAS, and RAPID3 were correctly recognized by the highest number of study participants as being useful for monitoring disease activity in axSpA/AS, a substantial fraction thought the same about ASAS20 (an outcome measure in axSpA/AS clinical trials), DAS28, and CDAI (disease activity scores for rheumatoid arthritis without a role in monitoring disease activity in axSpA/AS). This does not mean that US rheumatologists use DAS28 or CDAI to monitor patients with axSpA/AS. Rather, these results document a knowledge gap and suggest that education about appropriate tools may improve compliance with the recommendations for disease activity monitoring in axSpA (health professional factor). We also found that the majority of those who measure PROs do so during the clinic visit (63%), taking up valuable clinic time to complete questionnaires, which others may be unwilling to do (practice factor). Only 36% of respondents who check PROs do so before seeing the patient, either prior to the visit or at check-in. The potential of EMRs to expedite data collection before the clinical consultation remains untapped (practice factor).21
Interestingly, the majority of study participants check CRP or ESR routinely, even though the value of these tests in axSpA/AS is less established. This suggests a preference of rheumatology care providers for laboratory tests over PROs for disease activity monitoring (health professional factor). Interestingly, the vast majority of CRP/ESR measurements (86%) are performed after the visit, and therefore results are not available for decision making during the clinic encounter. The lack of laboratory results at the point of care has been discussed as a disadvantage of the ASDAS, which includes CRP or ESR.15,22 In reality, however, physicians appear to accept this delay as they order CRP/ESR tests regardless. Composite disease activity scores that combine PROs with laboratory markers, such as the ASDAS, are thus not necessarily doomed. A recent study in psoriatic arthritis showed that US rheumatologists used CRP testing less frequently than rheumatologists in 5 European countries.23 Differences in testing logistics may partially explain the observed discrepancy. Ideally, however, both PROs and CRP/ESR would be available at the beginning of the clinic visit to inform the decision making while the patient and the rheumatologist are in the same room. This could be achieved with a combination of electronic collection of PROs and point-of-care laboratory testing (practice factor).24-26
Limitations of this study include the restricted participant pool. The total workforce of rheumatologists in the US is approximately 6000.27 We thus sampled approximately 1.5% of US rheumatologists and there was an overrepresentation of academic rheumatologists (70% in this study vs 20% in the overall rheumatology workforce). We cannot exclude that attitudes and practices among rheumatologists in nonacademic practice settings and in geographic locations other than New England differ. However, our data of low frequency of disease activity monitoring in axSpA/AS using BASDAI or ASDAS are consistent with data from France.28,29 Because of the limited sample size, we were also not able to perform subgroup analyses (eg, comparing different practice settings or rheumatologists with various levels of training or experience). A larger study using a national sample frame should be informative and provide additional insight.
Participation rates in email surveys among healthcare providers vary widely, from < 10%30 to > 70%.31 Surveys among small groups or organizations with a shared interest may have higher response rates. The study by Kwan et al, with a response rate of 79%, surveyed members of the ASAS organization, where participation in surveys is a requirement for maintenance of membership privileges.31 A previous study much more similar in design to ours that administered a questionnaire to a convenience sample of 1918 US rheumatologists and tested 2 levels of monetary incentives reported a response rate of 26%, which is comparable to our 24%.32
We limited the number of questions and beta-tested the survey to ensure that the estimated time for completion was only 5 to 7 minutes. Most of the participants (86%) who started the survey completed all sections of the questionnaire, suggesting that length and content of the survey played only a minor role in determining the response rate. A number of strategies has been suggested to increase response rates in email surveys, including personalized email templates, follow-up emails, as well as monetary and nonmonetary incentives.33-35 As a secondary aim, we tested a strategy to improve the response rate by offering a donation to a patient organization for every completed survey, an approach that was used in a previous study but has not been formally tested.11 There was no significant difference in response rates between the incentive group and the control group in our study. However, using REDCap, we were not able to track whether invitees opened and read the invitation email and thus were aware of the incentive.
In summary, the results from this survey support the critical role of treatment recommendations in guiding the management of patients with axSpA/AS. The study showed strong agreement among rheumatology care providers with the concept of monitoring disease activity when managing patients with axSpA/AS. However, validated disease activity scores such as BASDAI and ASDAS are infrequently used in clinical practice. Our data document a knowledge gap regarding appropriate tools for disease activity monitoring in axSpA/AS. In addition, despite the ubiquitous use of EMR systems, the potential of these systems for efficient disease activity monitoring is not realized and requires further study.
ACKNOWLEDGMENT
We thank Drs. Jean Liew, Gregory McDermott, Monica Schwartzman, Derrick Todd, and Richard Zamore for beta testing the survey questionnaire. We thank Dr. Daniel H. Solomon for help with recruitment.
Footnotes
This work was supported by the National Institutes of Health (grant 5R21AR079691).
JE has served as reviewer for UpToDate. The remaining authors declare no conflicts of interest relevant to this article.
- Accepted for publication December 28, 2023.
- Copyright © 2024 by the Journal of Rheumatology
