Physical activity is recommended for adults with all types of arthritis because it can reduce pain and improve physical functioning, mood, and quality of life (1,2). Professional organizations encourage health care professionals to counsel adults with arthritis on physical activity and recommend supervised programs when needed (3,4). Primary care providers (PCPs) frequently treat arthritis (5) and are instrumental in promoting physical activity. Although we previously found that 98.4% of PCPs always or sometimes recommend physical activity to adults with arthritis (6), the content of physical activity counseling may affect its effectiveness (3). Addressing patient concerns (eg, arthritis-specific physical activity barriers such as pain) warrants specific guidance and referrals to safe, supervised programs (3). To build on a previous study, we examined physical activity types and programs recommended among PCPs recommending physical activity to adults with any type of arthritis and compared distributions of characteristics of PCPs recommending programs versus PCPs unaware of them.

We analyzed cross-sectional data from 2018 Porter Novelli DocStyles (https://styles.porternovelli.com/docstyles), an online national market research survey assessing PCP attitudes, patient encounters, and use of medical information resources. Eligible DocStyles participants were family practitioners, internists, obstetrician/gynecologists, and nurse practitioners aged 21 or older, living and practicing in the US, practicing for at least 3 years, treating at least 10 patients weekly, and working at an individual, group, or inpatient practice. From June through August 2018, Porter Novelli invited participants by email to complete the survey from the Sermo Global Medical Panel (www.sermo.com), SurveyHealthcareGlobus (www.surveyhealthcareglobus.com), and WebMD (www.webmd.com). Target quotas (1,000 PCPs, 250 obstetricians/gynecologists, and 250 nurse practitioners) were met by inviting highly responsive participants (defined as completing >75% of any kind of survey [not only DocStyles] in which they had been invited to participate) first from among those not participating in DocStyles 2017. Of 2,582 invited persons, 1,505 completed the survey (response rate, 58.3%) and were compensated $55 to $77 based on number of questions asked. We excluded 116 PCPs not treating adults with arthritis and 23 never recommending physical activity, which resulted in an analytic sample of 1,366. Additional survey details are available elsewhere (6). Although analyses were not subject to Centers for Disease Control and Prevention’s (CDC’s) institutional review board, we followed all Council of American Survey Research Organizations guidelines, and the data set was deidentified.

The 2018 DocStyles Survey included a module with questions about recommendations for CDC-recognized arthritis-appropriate physical activity programs (hereafter “programs”) (7), which have an evidence base for addressing physical activity barriers (8). PCPs treating an average of at least 1 adult with arthritis weekly completed multiple choice questions about physical activity counseling for adults with arthritis, including physical activity types, programs recommended, and reasons for not recommending programs.

We calculated percentages for physical activity type and program variables overall (N = 1,366) and reasons for not recommending programs among PCPs not recommending programs (n = 895). To identify opportunities for promoting program awareness, we generated distributions of PCP characteristics overall (N = 1,366) and for those recommending programs (n = 471) and unaware of programs (n = 710). We generated percentages using SAS version 9.4 (SAS Institute Inc); we performed χ² tests in Excel version 2008 (Microsoft Corp) to assess differences (significant at ɑ = .05) between PCP groups.

PCPs were commonly aged 50 or older (46.2%; 95% CI, 43.5%−48.8%), men (57.5%; 95% CI, 54.8%−60.1%), non-Hispanic White (67.1%; 95% CI, 64.6%−69.6%), and working in a group outpatient practice (67.5%; 95% CI, 65.0%−70.0%) (Table). Most PCPs recommended walking, swimming, or cycling (88.2%; 95% CI, 86.5%−89.9%), stretching (63.8%; 95% CI, 61.3%−66.4%), and physical therapy (60.8%; 95% CI, 58.2%−63.4%) (Figure). Programs were recommended less frequently than physical activity: 34.5% (n = 471) of PCPs recommended 1 or more programs. The most commonly recommended programs were the Arthritis Foundation’s Aquatic Program (18.0%; 95% CI, 16.0%−20.0%), the Arthritis Foundation’s Exercise Program (14.4%; 95% CI, 12.6%−16.3%), and Walk With Ease (13.8%; 95% CI, 12.0%−15.7%) (Figure). Most PCPs did not recommend any programs (65.5%; 95% CI, 63.0%−68.0%); among this group (n = 895), the most commonly reported reasons were being unaware of them (n = 710; 79.3%; 95% CI, 76.7%−82.0%); programs were unavailable in their area (22.5%; 95% CI, 19.7%−25.2%), unaffordable for patients (12.5%; 95% CI, 10.3%−14.7%), or inaccessible to patients (12.2%; 95% CI, 10.0%−14.3%); and believing patients would not attend (10.5%; 95% CI, 8.5%−12.5%).

Overall, 34.5% (95% CI, 32.0%−37.0%) of PCPs reported recommending 1 or more arthritis-appropriate programs (Figure). The distribution of most characteristics did not differ significantly between PCPs recommending physical activity programs and those unaware of physical activity programs, including by age, sex, region, provider type, years practicing, main work setting, number of practitioners in practice, and patient portal availability (Table). Exceptions were race or ethnicity (P < .001), privileges at a teaching hospital (P = .02), average number of patients treated per week (P < .001), and average number of patients with arthritis treated per week (P = .01). Distributions for PCPs recommending versus not recommending programs were significantly different for these same 4 variables.

At least 3 in 5 PCPs recommending physical activity to adults with arthritis recommended low-impact aerobic activities (walking, swimming, or cycling), stretching, or physical therapy. These activities align with professional guidance on optimal activities for most adults with arthritis (2,3), although appropriate activities differ by individual. Still, most PCPs sampled (65.5%) did not recommend programs, with 79.3% of these PCPs unaware of them. Our study demonstrates that the guidance PCPs already consistently offer to patients can be strengthened by recommending programs when needed.

PCPs are important promoters of physical activity (4). Creating a safe, specific, and tailored exercise plan is important for adults with arthritis (2); many are hesitant about physical activity because of misplaced fears about joint damage (9). Additionally, adults with arthritis report the absence of referrals to programs from health care providers as a barrier to exercise; therefore, they are likely to be receptive to program referrals (10).

Adults with arthritis may benefit from greater awareness of safe, arthritis-appropriate, evidence-based physical activity programs. Proven program outcomes include improved physical activity levels, strength, and balance, and reduced pain, fatigue, and stiffness (11). PCPs aware of local resources could be more likely to provide referrals (12). Strategies to promote PCP awareness of physical activity programs include distributing information about program benefits and availability through clinical practice sites, health departments, public health partnerships, continuing medical education, clinical–community linkages, and electronic medical record prompts.

Study strengths include the large sample size and ability to assess counseling for arthritis management. Limitations include using an opportunity sample that was not nationally representative and survey questions that featured a limited list of physical activity types and programs. Future studies might consider examining additional activity and program recommendations.

Strategic focus of public health awareness campaigns promoting arthritis-appropriate physical activity programs to PCPs could increase their reach to adults with arthritis.