Primary Care Patient Engagement With Patient-Reported Outcomes to Assess Osteoarthritis Symptoms

Abstract:

Background

Knee osteoarthritis (OA) is a common chronic condition in the United States, significantly impacting quality of life and health-care costs. Disparities in OA treatment across racial, ethnic, and socioeconomic groups highlight the importance of federally qualified health centers (FQHCs) in managing under-treated populations. This study assessed the feasibility of using a text message-based system to collect patient-reported outcomes (PROs) on knee OA prevalence and symptom severity among FQHC primary care patients.

Methods

A multimethod pilot study invited 223 FQHC patients with knee OA via text to complete a PRO survey on knee pain and function. Nonresponders received text and phone reminders. Researchers evaluated response rates, reasons for noncompletion, and patient perspectives on symptom monitoring.

Results

Overall, 24.7% of patients completed the PRO survey—28% among English speakers and 17% among Spanish speakers. The most common barrier was disconnected phone numbers. Qualitative feedback from both responders and nonresponders showed strong support (92%) for ongoing symptom monitoring. Responders found the survey user-friendly, while nonresponders noted technical and trust-related challenges.

Conclusions

Text message-based PRO collection is a feasible and acceptable method for knee OA symptom tracking in FQHCs. While the approach is scalable and well-received, patient education about the purpose of PROs is essential. Improved communication and accessibility strategies are recommended to increase engagement and fully leverage PROs for managing chronic conditions in low-resource settings.

Introduction

Osteoarthritis (OA) of the knee is one of the most common chronic conditions in the United States, affecting more than 27 million Americans [1]. Additionally, OA is one of the leading causes of disability, and can cause severe pain, functional impairments, and decreased quality of life [2]. Given its widespread impact and chronic progressive nature, knee OA is associated with significant health-care resource utilization and costs [3], and an economic burden estimated to be upwards of $140 billion in the US annually [4]. As the prevalence of OA increases with age [5], the Center for Disease Control estimates that by 2030 25% of US adults (67 million people) will be diagnosed with OA [6].

Despite widespread disease prevalence, significant disparities in the utilization of treatment for OA exist among racial, ethnic, and socioeconomic minority populations [7]. For example, lower levels of total joint arthroplasty utilization have been documented for Black, Hispanic, and Native American patients compared to White patients [8,9].

Primary care clinicians are on the front line in managing patients with knee and hip OA and play a pivotal role in overseeing treatment plans [10] Primary care physicians can guide patients through the variety of OA treatment pathways, including medical management, and as indicated, referral for physical therapy or specialist evaluation, including rheumatology and orthopedic surgery. However, there are currently no evidence-based strategies to assist primary care physicians in OA screening, longitudinal monitoring of OA symptom severity and treatment efficacy, or indications for specialist referrals [11].

Patient-reported outcomes (PROs) offer a standardized method for capturing patients’ assessment of OA symptoms and their impact on daily life [12]. While frequently used in orthopaedic clinics to assess symptom severity and response to intervention, PROs are not routinely used to evaluate OA in the primary care setting [13]. Despite advancements in electronic health records (EHRs) systems which allow increased PRO accessibility via patient portals, collection rates are still only at about 30% [14]. Additionally, many primary care clinics lack this integration of PROs for assessing OA symptom severity [15]. Integrating PROs into routine clinical practice, particularly in underserved settings, could help to improve the identification of knee OA with symptom progression by providing real-time pain and functional assessments. However, the feasibility and acceptability of implementing PRO systems for knee OA in resource-constrained environments, such as federally qualified health centers (FQHCs) have yet to be evaluated.

Text message-based collection of PROs is gaining recognition as the future of PRO collection due to its accessibility, convenience, and higher patient engagement rates compared to email-based systems integrated into EHRs [16]. A randomized control trial by Ziegenfuss et al. found that patients who were initially nonresponsive to PRO collection via email had significantly higher rates of response when given text message reminders [17]. Text messaging allows for real-time, on-the-go interaction, fitting seamlessly into patients’ lives, particularly for those with limited access to email or less familiarity with online patient portals.

This pilot study aimed to evaluate the feasibility and acceptability of reporting PROs via text message in a primary care clinic in an urban FQHC. A multimethod evaluation was utilized to understand patient experiences with PROs and their willingness to use such tools for ongoing knee OA symptom monitoring. This study specifically focused on evaluating the response rate to text-message PRO invitations, identifying barriers to completion, and assessing patient perspectives on the acceptability of regular PRO use. Addressing barriers to patient acceptance and use is foundational in efforts to establish a primary care-based PRO system for OA symptom monitoring and management. Lessons may also be applied in orthopedic settings to improve compliance with Centers for Medicare and Medicaid Services PRO mandates.

Material and methods

The research team obtained institutional review board approval to contact patients from a list that included the names, phone number, and primary language of 223 patients with diagnosed knee pain or knee OA who were receiving care from a primary care clinician at our partner FQHC between January 1, 2022, and December 31, 2022. Of the total, 154 patients identified English as their primary language, and 69 (31%) listed Spanish.

All patients were sent text message invitations in their preferred language with a secure, personalized link to a PRO survey, which included the Knee Osteoarthritis and Injury Outcome score [[18], [19], [20]] validated and standardized measure of knee pain and function used in knee OA clinical care and research and mandated for use in all Medicare patients with knee OA who elect to undergo joint replacement surgery by 2027 [21]. Participant demographics were also collected. The survey included 29 questions and took participants roughly 5-7 minutes to complete. Following the initial invite, patients who did not respond to the survey received one reminder via text message, and one reminder via phone call from a trained research coordinator. The research coordinators resent a text with the survey link during the reminder call, or, if unable to reach the patient, prior to leaving a voicemail.

The demographic profile of PRO completers was generated. Both Knee Osteoarthritis and Injury Outcome score pain and activities of daily living subscores were calculated using the developers’ methods [20] Survey completion rates were calculated overall and by primary language and compared using chi-squared tests. Mean and standard deviation of PRO pain and activities of daily living scores were compared using student t-tests.

All patients from the list, including both survey responders and nonresponders, were invited to participate in a brief 5–10-minute phone interview to evaluate the acceptability of the text-PRO system. Both responders and nonresponders were asked about their willingness to complete PROs on a regular basis to monitor their symptoms. Survey responders were asked to rate the difficulty of completing the survey using their phone, and how well the PRO captured their experience of OA in their daily lives. Nonresponders were asked if they would be willing to complete future surveys on their phone, and why they chose not to complete the survey that they received. Due to the brief nature of the interviews, qualitative data were descriptively and pragmatically summarized on a per question basis. Two investigators reviewed the data, inductively categorized the responses to each question, and then met to discuss.

Results

Among the patients who completed the PRO survey, self-reported sociodemographics include the following: 70.9% female, 29.1% white, 23.6% Black, 10.9% Asian, and 21.8% Spanish language speakers, with a mean age of 58.2 years. Based on available data for patients served at the partner FQHC, 43% identified as Hispanic, Latino/a, or Spanish origin, 28% Black or African American, 9% Asian, and 18% non-Hispanic White [22]; 60.89% were aged 18-64, and 7.28% ≥ 65. This distribution indicates that the patients who completed the PRO survey approximated the diverse racial and ethnic makeup of the FQHC (Table 1).

Table 1.

Sociodemographics of PRO survey completers.

Demographics	Participants
	Mean (SD)
Age	58.2 (13.5)
Biological Sex, N (%)
Male	15 (27.3)
Female	39 (70.9)
Did not report	1 (1.8)
Ethnicity, N (%)
Hispanic or Latino	15 (27.3)
Not Hispanic or Latino	28 (50.9)
Prefer not to disclose or did not report	12 (21.8)
Race, N (%)
Black or African American	13 (23.6)
American Indian	2 (3.6)
Asian	6 (10.9)
White	16 (29.1)
More than one race	3 (5.5)
Prefer not to disclose or did not report	15 (27.3)
Primary Language, N (%)
English	43 (78.2)
Spanish	12 (21.8)

The overall rate of PRO completion was 24.7% (n = 55) with a higher rate observed among English speaking patients (28%, n = 43) as compared to Spanish-speaking patients (17%, n = 12). The difference in response rates did not reach statistical significance (P < .09). The primary reason for noncompletion (N = 168, 75.3%) was the inability of the study team to reach patients by phone (Table 2). Of those who responded, 43.6% completed the survey following the initial text, 9.1% after receiving the initial text and 1 text reminder, and 47.3% following the reminder phone call (Table 3).

Table 2.

Primary reasons for survey noncompletion (n = 168).

Reason for noncompletion	N (168)	% Of noncompleters
Unable to reach participant, left voicemail	62	37%
Unable to reach participant, could not leave voicemail (disconnected/voicemail issues)	36	21%
Reached and agreed to participate but did not complete survey	23	14%
Reached but didn't speak English/Spanish or wrong person	10	6%
Reached but unable to complete survey due to technical difficulties	2	1%
Reached but declined participation:	29	17%
Lack of interest / hung up / no reason given	15	9%
Not currently experiencing knee pain	10	6%
Overwhelmed by other health problems	2	1%
No longer seeing provider at partner FQHC	2	1%
No reason entered	6	4%

Table 3.

Survey completion rates by stage in engagement.

Participants who completed survey…	All participants N (%)	English speaking N (%)	Spanish speaking N (%)
After 1 text	24 (43.6)	16 (37.2)	8 (66.7)
After 2 texts	5 (9.1)	4 (9.3)	1 (8.3)
After the reminder call	26 (47.3)	23 (53.5)	3 (25)
Total	55	43	12

Brief interviews were completed with 48 primary care patients: 29 of the PRO responders (25 English speaking and 4 Spanish speaking) described above, and 19 nonresponders (13 English speaking and 6 Spanish speaking). Among survey responders and nonresponders, 92% said that they would be willing to complete the survey on a regular basis to monitor their symptoms, particularly change over time, and provide clinicians with better information, which they believed could inform treatment recommendations. 93% of PRO survey responders described no challenges completing the survey via cell phone, and 84% of PRO nonresponders also said that they would be willing to complete surveys using their phones.

The main reasons that nonresponders who expressed interest in symptom tracking did not complete the PROs were that they either did not see the text message or ignored it because they did not recognize or trust that the survey was sent from their clinician. Several patients shared technological barriers such as poor cell service or not having their phone set up to receive text messages. Finally, additional participants had medical reasons why they could not complete the survey at the time it was sent out, including struggles with mental health and hand OA which made phone use difficult. A sample of direct patient quotes can be found in Table 4.

Table 4.

Qualitative assessment of text-based PRO system.

Symptom monitoring	Yes [I would complete the survey]. I Think mainly because with the survey you can monitor changes easier than waiting to go to your next appointment that can be months apart. (ID #10, 31-y-old, Latino man, survey responder)
	The symptoms are changing, and the function is changing. It depends on what you are doing in a day or week so it must be monitored regularly. (ID#59, 51-y-old, Asian woman, survey responder)
Providing clinicians with information	Yes. It's something that is obviously impacting me every day, so being able to watch the changes over time would be very helpful to me and my doctor in case things are getting worse or better. (ID #97, 23-y-old, white woman, survey responder)
	I feel like more information is always better. I feel like my knee situation is unique given I've had a lot of procedures done on it. If monitoring the symptoms over time helps with my treatment, then I would want to do this. (ID #73, survey nonresponder, did not provide demographic information)
	More information back and forth is better between me and my doctor so she can make better decisions for my health. (ID# 171, 51-y-old, white women, survey responder)

Discussion

This study evaluated the feasibility and acceptability of collecting PROs via text message among primary care patients at a FQHC. The findings offer insights into the effectiveness of text-based PRO collection, barriers to patient participation, and willingness of primary care patients to engage with symptom monitoring systems.

The overall survey response rate was 24.7% suggesting that text-based PRO collection is feasible in this patient population as the rates are comparable to PRO collection in US health systems via major EHR vendors [23,24]. For context, multiple studies report a 30% PRO response rate through EHR portals among joint replacement patients with knee OA [23,24]. That is not meaningfully higher than what this pilot study observed in the primary care setting. Response rates were, however, higher among English speakers (28%) compared to Spanish speakers (17%), highlighting potential disparities in engagement that could reflect differences in language accessibility, trust in digital health tools, or comfort with technology [25].

Most survey responders (93%) reported no technical difficulty completing the survey on their phones, and many nonresponders (84%) said that they would generally be willing to complete surveys in this way. This may demonstrate improved patient engagement via text messaging compared to more traditional digital health modalities likely due to ubiquitous cell phone ownership across racial, ethnic, and socioeconomic groups [[26], [27], [28]]. A minority of nonresponders did cite technical barriers, such as poor phone service or discomfort with digital platforms, as reasons for not completing the survey. Therefore, future implementations could consider offering both text and phone-based surveys, providing additional support to accommodate patients who face technological challenges, or providing the opportunity for patients to complete the PROs in other settings, such as the clinic waiting room [[29], [30], [31]].

Among nonresponders, the primary barrier to participation was difficulty reaching patients, with 21% of nonresponders having disconnected numbers or unconfigured voicemail boxes. This finding provides important context for this study’s collection rates, as a large percentage of patients likely did not receive the invitation to participate. The 13% of patients who were reached but declined to complete the PROs did so for a variety of reasons. Most notably, 34% (n = 10) stated that they were not currently experiencing knee pain. This highlights that knee OA symptoms vary over time and while many patients find ongoing symptom monitoring with PROs to be valuable, others may not perceive a direct need if their condition is not currently causing significant distress or disability.

Despite this fact, both responders and nonresponders expressed strong support for ongoing OA PRO monitoring, with 92% of all interview participants stating they would be willing to complete PROs regularly to monitor their symptoms and provide clinicians with better information to guide treatment decisions. This aligns with prior literature indicating patients value tools that enhance communication with their health-care providers and offer real-time insight into their health status [32]. It also reinforces the idea that patients are willing and able to complete PROs when they understand that such surveys can help inform their care.

In previous work, this team alerted patients to the value of PROs by sending an email or letter from clinicians notifying them that they would be receiving the PRO and explaining the benefits of completion. This method was associated with significantly higher PRO completion rates [33]. Therefore, future work should include similar strategies to ensure patients recognize the value of PRO completion.

The findings of this pilot study highlight important developments in addressing OA treatment disparities and the opportunity to use PROs across diverse populations and clinical settings. Busy primary care clinics may understandably resist the addition of new screening tools. However, this study shows that direct-to-patient text-message symptom assessments can be automated for patients without increasing clinician workload. Additionally, low-income, and multilingual patients have been underrepresented in PRO collection through EHR systems, especially when collection is implemented via online patient portals [34,35]. When PROs are delivered via accessible platforms, such as text messaging, using the patients’ primary language, diverse populations are willing and able to engage with the PRO reporting process. This finding is in line with previous research indicating that text messaging is an effective tool for increasing appointment attendance and improving medication adherence in low-income populations [36], and that ubiquitous cell phone ownership across racial, ethnic, and socioeconomic groups allows for improved patient engagement via text messaging compared to more traditional digital health modalities [37]. Last, although this work focused on a primary care population, lessons may also be applied in surgical settings, which need to develop effective, sustainable PRO implementation strategies to align with new Centers for Medicare and Medicaid Services mandates regarding PRO collection.

This study is not without limitations. First, the patient list used for recruitment included many patients who were unreachable, which could have led to an underestimation of patient willingness to engage with the PRO system. Second, the generalizability of our findings is limited by the specific study population—patients from a single FQHC. While FQHCs serve diverse populations, findings may not fully apply to other primary care settings with different health-care access patterns. Finally, the study relied on self-reported data to assess patients' experience with PROS. This may be subject to recall bias or variations in patient interpretation. Future work from this group will involve repeating this study using patients with recent or upcoming appointments to validate the response rate, use in clinical care, and longitudinal value of PRO symptom management in OA care.

The study also focused on the collection of PROs, and patient perspectives on their hypothetical use, and did not assess the clinical utility of tracking OA symptoms using PROs. Future research should design and evaluate systems to pair OA symptom assessment with clinician encouragement to address advancing symptoms among patients whose PROs reflect moderate to severe pain or functional limitations. In addition, further research is needed to define strategies to instruct patients on how to use PRO data to initiate conversations with their clinicians when symptoms are affecting their daily life.

Conclusions

This study highlights the potential for using text message-based PRO surveys to assess knee OA symptoms in a diverse primary care population. While response rates were comparable to surgical settings, the willingness of patients to engage with symptom monitoring suggests that with improvements in accessibility and communication, PRO collection via text messaging systems could become a valuable tool for managing OA symptoms in the primary care setting and potentially informing future OA management. Continued efforts to refine PRO implementation and address barriers to participation will be essential to realize their potential in enhancing patient care.

Funding

This research was funded by the American Academy of Hip and Knee Surgeons Foundation for Arthroplasty Research and Education (FARE) grant Grant ID: 2023-05).

CRediT authorship contribution statement

Martha Burla: Writing – review & editing, Writing – original draft, Project administration, Methodology, Investigation, Conceptualization. T. Jacob Selph: Resources, Methodology, Data curation. Finola M. Summerville: Resources, Methodology, Data curation. Rachel Bergman: Writing – review & editing. Isaac Sontag-Milobsky: Writing – review & editing, Writing – original draft. Patricia D. Franklin: Writing – review & editing, Supervision, Resources, Project administration, Investigation, Funding acquisition, Conceptualization. Linda I. Suleiman: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.

Conflicts of interest

T. Jacob Selph Jr. receives Compensation for authorship from Journal of Bone and Joint Surgery.

Linda Suleiman is a paid consultant and receives research support from for Zimmer Biomet; owns Stock or stock options in Corin; and is a Board member/holds committee appointments for AAHKS and RJOS.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2025.101850.