The ASK study was a pragmatic cluster-randomized trial with the primary aim of evaluating the effectiveness of a personalized PRO-based report containing estimates of likely outcomes of surgery if elected to treat advanced hip and knee osteoarthritis (ClinicalTrials.gov identifier: NCT03102580). A copy of the ASK report is in Multimedia Appendix 1. The study enrolled patients who scheduled an initial visit with one of 36 orthopedic surgeons to discuss knee and hip arthritis treatment options, including total joint replacement surgery. Aim 1 involved report co-design, with patients and clinicians, aim 2 aimed to compare decision quality for patients at sites randomized to receive or not receive the report, and aim 3 included adding a patient coaching intervention (“ASK coaching”) to a randomly assigned subset of study sites. Figure 1 depicts study procedures during aim 3. In this aim, patients completed PROs prior to their initial office-based consultation with a participating orthopedic surgeon. Within the month following this initial consultation, study staff contacted all patients via telephone to invite participation in the study and assess decision quality. Participants receiving care in study sites randomized to coaching were also invited to attend an ASK coaching session at this time.

During the design of the parent study, patient advisers who had recently made decisions regarding the treatment of their hip or knee osteoarthritis were consulted [16]. A subset of those patients expressed that an opportunity to review the content of the ASK report in a group setting would be helpful. The ASK coaching intervention was included in the study design based on this recommendation and, like the ASK report, was developed in collaboration with patient partners and clinical stakeholders.

The research team determined that the coaching intervention would focus on helping participants understand the personal data provided in the ASK report and how to use that data in future conversations with clinicians. The research team then consulted with Citizens for Patient Safety (CPS), an organization that educates patients and clinicians about medical decision-making and navigating the health care system. CPS recommended further helping patients prepare for clinical consultations by encouraging them to be active members of the health care team and helping them develop a list of questions to ask clinicians. They specifically highlighted the Agency for Healthcare Research and Quality’s (AHRQ) “Questions are the Answer” materials [23] as a potentially useful tool to guide the session.

A trained health educator (an individual with a Master of Public Health degree and a certification in health education) then developed the hour-long online group coaching session and an osteoarthritis-tailored patient handout based on the AHRQ materials recommended by CPS. This handout, entitled “I Have a Voice” (Multimedia Appendix 2), was used throughout to help patients reflect on their data and how it might inform future conversations with clinicians. The research team also created a website consolidating additional patient resources from reputable sources, including AHRQ, Arthritis Foundation, American Academy of Orthopaedic Surgeons, and American Association of Hip and Knee Surgeons. After completing this initial design, the research team met with clinical experts, including a patient advocacy leader from CPS, a senior orthopedic surgeon, a physical therapist, a health psychologist, and a group of research coordinators with experience recruiting patients with osteoarthritis. Experts attended a mock coaching session and provided input in a semistructured interview. Specific suggestions were extracted from the interview notes and used to refine the coaching intervention organization, content, and support materials. Experts’ suggestions primarily focused on the best way to communicate sensitive topics during coaching and the framing of patient predictions of likely outcomes if surgery was elected. For the former, clinicians encouraged the use of positive language, such as referring to comorbidities as “areas for possible improvement to optimize surgical outcomes.” For the latter, they suggested emphasizing that a good surgical outcome is one that improves arthritis pain and stiffness, but may not necessarily result in a perfect joint. Experts also offered advice on addressing the needs of nonsurgical patients including the importance of understanding personal data and setting goals regardless of one’s treatment plan.

Mock sessions using the refined intervention were then scheduled with ten patient advisers with hip or knee osteoarthritis recruited through the Arthritis Foundation. Patient advisers attended a session and provided feedback in either a one-on-one interview or written format. Generally, their feedback was positive. Most found reviewing the report during coaching to be helpful because they did not understand the PRO data on their own, and almost all mentioned valuing the supportive environment in which to discuss their concerns. Patient advisers were also enthusiastic about the “I Have a Voice” handout. Small changes to language and emphasis were made based on the patient advisor’s feedback, but the overarching session goals and content did not change.

The coaching intervention was launched for ASK study participants in early May 2020. Each initial session was staffed by two research team members: (1) the health educator who facilitated the session and (2) a research coordinator who gathered feedback from observations and provided technological support to attendees as needed. At the end of the first month of coaching, minor logistical changes were made, including adding answers to frequently asked questions and clarifying the instructions on joining a meeting via telecommunication platform Zoom (Zoom Video Communications). No additional changes were made after the first month. An outline of the final coaching session content is presented in Textbox 1.

All participants in the study’s coaching arm were invited to attend the online, group coaching session at the time of the postconsultation telephone survey. Multiple sessions were scheduled each week at varying times to accommodate patient schedules. Those who agreed to participate were scheduled for a session and both mailed and emailed the presession materials (Figure 2). While all study participants in this aim received a copy of their ASK report, only those who agreed to participate in the coaching session received the other materials. Participants scheduled for a coaching session received a reminder the day before, which included the number of a dedicated research coordinator who they could contact for technical support before or during the session.

Coaching consent and attendance rates were tracked throughout. Of all patients invited to coaching (n=1389), 45% consented to participate, and 32% of the total (N=438) attended a scheduled session. Among those who consented, 21% either did not schedule a session at the time of consent or scheduled a session but canceled, and 9% scheduled a session but did not show. Figure 2 summarizes attendance and cancellation rates. Research coordinators recorded reasons for declining to attend coaching, which included participants feeling they either did not need additional information about the report or their arthritis symptoms or were too busy to attend. Sessions had an average of 3‐4 attendees each, and enrollment extended from May 2020 through February 2022, which overlapped with the peak of the COVID-19 epidemic.

The timing of the quantitative coaching assessments within the broader aim 3 study flow can be found in Figure 1. To understand the profile of patients who attended coaching, sociodemographic information and clinical comorbidity data were obtained from all participants at sites in the study’s coaching arm and were descriptively summarized and compared by coaching attendance status using chi-square and 2-sample t tests that accounted for possible unequal variance. PROs included the Hip disability and Osteoarthritis Outcome Score (HOOS) or Knee injury and Osteoarthritis Outcome Score (KOOS-12) [24,25], osteoarthritis-specific PROs used broadly in research and clinical care, and the Veterans RAND 12-Item Health Survey (VR-12) [26] that assesses global physical and emotional function. Patients also self-reported pain intensity in other joints and lower back, sociodemographic factors, and medical comorbidities. A question to assess confidence in completing health forms screened for health literacy [27]. In late March 2021, roughly 1 year after the launch of ASK coaching, a patient-reported assessment of self-efficacy was added to the postconsultation telephone survey (ie, the time of the coaching invitation) to further understand the profile of those who did or did not choose to attend coaching sessions. This included 4 treatment-related items from the PROMIS (Patient-Reported Outcomes Measurement Information System) v1.0 Self-Efficacy for Managing Medications and Treatments [28] item bank (items SEMMT011, SEMMT012, SEMMT013, and SEMMT014), and was administered to the last 692 participants invited to coaching. Raw scores on this custom short form were converted to T-scores using the Health Measures Scoring Service. The T-scores were compared between those who attended and did not attend coaching using a Mann-Whitney U test as the scores were not normally distributed. A 2-tailed α level of <.05 was considered statistically significant for all inferential analyses.

In addition, to quantitatively assess attendee learning from coaching, a postsession survey was added in late March 2021. It included five self-assessment questions related to perceived change in knowledge and was sent via email after the coaching session. Responses to the self-assessment questions were descriptively summarized.

Data were integrated from multiple sources to qualitatively evaluate the ASK coaching intervention. First, questions related to acceptability and perceived benefits of coaching were included in semistructured interviews conducted in April-July 2021 as part of the parent study; 16 of these 25 interview participants had attended a coaching session. Only coaching-related segments from these 16 interviews were included in the present analysis; findings from other qualitative analyses of these interviews have been previously published [17,18]. Second, paired pre- and postcoaching interviews were conducted for 4 additional attendees in February 2022. These interviews provided finer-grained feedback on how coaching influenced participants’ understanding and potential use of PROs including in conversations with providers. All interviews were completed by a trained interviewer via Zoom and used a semistructured interview guide; they were audio-recorded and transcribed. Third, qualitative data were extracted from optional free-text responses provided by 47 coaching attendees in the postsession survey described above. Refer to Figure 1 for positioning of the qualitative assessments within the broader aim 3 study flow.

We used a qualitative descriptive approach to analysis, which aligned with our pragmatic and focused qualitative aims to learn more about what motivated participants to attend coaching sessions, assess the acceptability of an online group coaching session, and evaluate whether the session met the goals laid out by the research team [29-32]. After independent reading and preliminary coding of the 4 paired pre- and postcoaching interview transcripts, MB and BZS collaboratively developed a deductive codebook based on the qualitative aims. To increase trustworthiness, all interview transcripts and survey comments were dually coded by MB and BZS, with ongoing meetings to discuss emerging findings, iteratively refine the codebook, and resolve discrepant coding.

To ensure informed consent, all participants in the parent study were provided with a detailed consent form outlining the study’s procedures, potential risks and benefits, and their right to withdraw at any time. Participants in the coaching arm received an additional fact sheet with details about the coaching session, as did those who were invited to participate in an interview. Participants were paid US$ 20 for completing the postconsultation telephone survey as part of the parent study and an additional US$ 20 if they participated in an interview. No additional incentives were offered for participating in coaching or completing the postsession survey. All data collected as part of this study were deidentified and stored on a secure server. All study procedures were approved by the University of Massachusetts Medical School institutional review board (H00012297).

Sociodemographic and clinical profiles of coaching attendees versus nonattendees are summarized in Table 1. Compared with nonattendees, patients reporting higher education and greater health literacy were more likely to attend a coaching session when invited (P<.001). In addition, those with Medicare insurance were more likely to attend while those with Medicaid insurance were less likely to attend (P<.001). Medicare and Medicaid are US government–run health insurance programs. Medicaid provides insurance to low-income families and individuals. Medicare provides insurance to senior citizens and some individuals with disabilities. While there was a statistically significant between-group difference for age (P=.002), the mean difference was less than 2 years between those who attended versus those who did not. Furthermore, while there were no statistically significant differences, those who attended coaching were more likely to be female, have at least one comorbidity, and have slightly less preconsultation pain (Table 1). In addition, the subsample of participants with self-efficacy data, self-efficacy for managing treatments was lower in those who attended coaching (n=166; median T-score 48.6, IQR 41.8-57.3) than those who did not (n=526; median T-score 57.3, IQR 42.7-57.3; P=.05).

The postsession self-assessment of learning was completed by 89 coaching attendees (49% of those invited to respond). Approximately, two-thirds of respondents reported greater understanding of their current and projected symptoms. Along with this, three-quarters of participants reported more knowledge of where to find additional osteoarthritis information after coaching (Table 2).

The qualitative evaluation results are organized into participants’ perspectives on motivation for attendance, acceptability of session format, achievement of session goals, and suggestions for improvement. Additional supporting quotes for each subsection are presented in Multimedia Appendix 3.

This paper describes the iterative design and evaluation of an online, group coaching session offered as a companion to a PRO-based decision aid for patients with hip or knee osteoarthritis who had an initial consultation with an orthopedic surgeon. The collaboratively designed session, referred to as ASK coaching, aimed to support patients in understanding their personal health data, including PRO scores, and using the data in future conversations with clinicians. Evaluation findings highlight differences in attendance based on sociodemographic characteristics and provide insight into motivation for attendance, session format acceptability, and achievement of session goals. Implications for potential design refinement, including priorities for future research, are integrated throughout the discussion.

Osteoarthritis of the knees and hips is one of the most common causes of chronic pain and disability in the United States. It currently affects more than 21 million Americans, with increasing prevalence as the population ages [33,34]. Osteoarthritis is a chronic condition for which there is no cure, and treatment options vary based on symptom severity and patient preference [35]. Therefore, patients’ understanding of their symptoms and ability to communicate with providers are particularly important. In today’s clinical practice, patients are accustomed to completing previsit assessments of their demographic, insurance, and health status information. However, receipt of their personal data is new to most patients, as is participating in a group coaching session to support the understanding and use of such data.

While patients routinely completed the previsit PROs, not all of them elected to participate in ASK Coaching, so this paper attempts to understand who this intervention may reach. First, patients reporting higher education and health literacy levels were more likely to join a session. This finding aligns with research showing that patients with more formal education and higher health literacy are more likely to engage in educational programming related to their arthritis care [36], a pattern that can contribute to the exacerbation of inequities if those with the greatest need for support are the least likely to access it. In contrast to high literacy and education, ASK coaching attendees reported lower confidence in managing their treatment as compared to nonattendees, suggesting they may have appropriately self-selected to attend based on their perceived need. Coaching attendees were also less likely to have Medicaid insurance than nonattendees, which may reflect a lack of access to health education or digital interventions among those with lower socioeconomic status [37]. In addition, coaching attendees were slightly older than nonattendees. While this difference in age was likely not clinically meaningful, it provides reassurance that older adults can readily and successfully attend an online session [38]. While there was no significant between-group difference for sex, patients reporting female sex tended to more often attend ASK coaching, which is consistent with existing literature showing women are more likely to seek out health information than men [39,40]. Patients with at least one comorbidity were more likely to attend a coaching session when invited, which may indicate those managing multiple competing health priorities have a greater need for support in preparing for conversations with clinicians. Interestingly, those with slightly lower preconsultation pain were also more likely to attend, which may indicate pain can be a barrier to engaging with supportive resources [41]. Priorities for future research are identifying for whom this intervention is most beneficial, more deeply understanding barriers and facilitators to attendance in subgroups at risk of intervention-generated inequality [37], and refining the intervention design to address those subgroup needs.

A primary motivation for coaching attendance was to gain additional information. Relatedly, the most common reason for not attending was having adequate information. While not all patients may require a resource like ASK coaching to understand PRO data and engage in conversations about their osteoarthritis care, patients also may not initially recognize gaps in their own knowledge. Qualitative findings reinforced patient-perceived benefits regarding gaining information about their PRO data even when they initially reported understanding. Similarly, previous qualitative research about the ASK report highlighted misinterpretations that were not always recognized by participants [18]. Future research could therefore explore how to best communicate the potential value of participation, even to those who feel confident in their knowledge, through co-design of presession materials. A second reason study participants did not join ASK coaching was that they reported feeling too busy. Time is a common barrier when it comes to engaging participants in health promotion efforts [42,43]. While this research was conducted during the COVID-19 pandemic, a particularly challenging time for all, limited patient time supports further refinement of the session length and the development of an asynchronous option to reduce barriers to attendance.

This paper also explores the acceptability and potential effectiveness of coaching patients to use a PRO-based decision report. Among attendees, the online format, with technological support provided as needed, was generally accepted. While some preferred in-person meetings, if possible, online sessions like this one may be easier to implement as they have been shown to be more cost-effective [44], particularly for health systems or case management programs that serve a large geographic area. Most attendees also endorsed the small group nature of ASK coaching. However, qualitative feedback made it clear that this structure could be improved to maximize the session’s benefits by including more unstructured time to ask questions and interact with peers. Given constraints related to clinician availability in many health systems, especially for nonreimbursable educational activities, it is worth noting participants found the nonclinician health educator to be an acceptable professional facilitator. However, adding a peer educator could benefit those interested in peer support, which has been shown to have benefits for other chronic conditions [45,46]. Based on these findings, a variety of design modifications could be evaluated in future research. To encourage participation among those who prefer in-person programs, the intervention design could be refined to improve the experience, including more scalable technological support outside of a research coordinator. Based on feedback about disparate participant experience, groups could also be organized by current treatment plans, such as surgery or nonsurgical treatments, to ensure the discussion is relevant for all. It should be noted, however, that defining groups in this way would limit attendees’ ability to learn from people who were further along in the treatment process, something coaching participants appreciated. Additional options for future intervention refinement include splitting the session into multiple shorter sessions or providing information that could be viewed asynchronously before a moderated discussion, both of which would offer greater opportunities for interaction while addressing time constraints.

With regards to session goals, attendees reported that ASK coaching provided clarification of specific PRO graphs and improved overall report comprehension. Previous publications of patients’ perspectives of the ASK report highlighted challenges with report comprehension, which support the role of coaching [17,18]. The session goal of preparing patients for future conversations with clinicians was also primarily met, with many patients expressing that they felt more ready to speak up and ask questions, finding the “I Have a Voice” handout particularly helpful. These positive findings related to acceptability and preliminary effectiveness were likely a function of the iterative design process, highlighting the value of collaboration with both clinical experts and patients during design to align the intervention to the users’ needs.

Additional implications for design and future research include considerations related to both efficacy and scalability. The timing of the session within the parent study limited patients’ ability to immediately use the information and skills from the session. When the study was designed, the return of PRO data directly to patients was not the norm in clinical care. Thus, the parent study proposed that the ASK report would be distributed in the surgeon’s office enabling each clinician to review the information with their patients. The online coaching session then followed this surgeon consultation and was intended to provide a second, more detailed review of the report. While the design worked well within the established care structure, it meant that participants had a clinical consultation before attending the session. Given that PROs are more commonly included in clinical care today, and participants felt the session helped increase their confidence when communicating with clinicians, it is likely that a greater benefit would have been achieved had the session taken place prior to the initial surgeon consultation. At a minimum, participants reported that receiving information about the “I Have a Voice” handout in advance of that initial appointment would have been helpful, aligning with the principles of decision guidance—supportive elements embedded within or alongside the decision aid for independent use [22]. In addition, providing this patient preparation resource could benefit clinicians who only have a short time for each appointment by ensuring that patients can quickly share details about their arthritis and its impact on their lives [47]. Combining this handout with a prerecorded asynchronous version of the coaching session would likely also improve this intervention’s scalability as it is a one-time cost and does not require additional clinician time or onsite resources. However, this intervention format may not address all patient needs, such as being able to ask questions of a trained facilitator and interacting with other arthritis patients. In future efforts, the research team will propose a real-time return of the ASK report and coaching before the initial surgeon consultation to support patient preparation for that initial treatment discussion. Additional research in collaboration with clinicians and patient experts will further refine the design of coaching and its implementation, including asynchronous or hybrid versions that align with both patient and health system needs.

While the parent study enrolled both English-speaking and Spanish-speaking participants, the coaching session was only offered in English, potentially limiting the generalizability of findings. Also, many participants who chose to attend the session were well-educated about arthritis and treatment options and therefore may have been less likely to benefit from participating. This limits conclusions about the intervention’s effectiveness. While there were no significant racial or ethnic differences in coaching attendance, study participants were largely White, reflecting well-documented racial and ethnic disparities in referrals to orthopedists for joint replacement surgery [10]. Coaching acceptability may differ for racial or ethnic minorities and needs to be examined in future research with more diverse samples, given the potential benefits of participating in an intervention of this nature to reduce inequities in osteoarthritis care. Furthermore, patients with a strong preference for a different session format may have simply chosen to not attend, and beyond asking participants who declined to participate why they were not interested, the research team did not seek more detailed feedback. As such, it is difficult to draw generalizable conclusions about session format acceptability. ASK coaching was also part of a complex health intervention that included the PRO-based report and consultation with the surgeon. This complexity makes it challenging to parse out specific benefits of coaching. Moreover, different components of the quantitative and qualitative evaluation were pragmatically added at various points in this multi-method evaluation, limiting the ability to systematically integrate such data using a mixed methods analysis. Advanced data integration may have been helpful to provide additional insight into which patient subgroups might benefit the most from this intervention. While we designed the intervention for efficiency (eg, group visits and online coach), we did not formally evaluate the intervention’s costs, which could be a barrier to broad implementation in routine care. Finally, this portion of the study took place during the COVID-19 pandemic, which may have influenced patients’ perspectives on the online and group components of the coaching format.

This paper describes the iterative design and evaluation of an online coaching intervention delivered by a health educator to support patients in understanding and using a novel PRO-based report as part of hip or knee osteoarthritis care, including in conversations with clinicians. Among coaching attendees, the acceptability of the session format was overall high, and session goals were achieved. With some modifications related to timing and structure, this intervention appears promising to support patients in using personal health data, including PRO scores, to prepare for conversations with their clinicians. Future research is needed to further refine this intervention and scale implementation, including ensuring reach to diverse patients.