Patients’ use of physical therapy for lower back pain: A qualitative study

Abstract

Objective:

The purpose of this study was to identify factors that influence a patient’s decision to use physical therapy (PT) services for a low back pain (LBP) complaint.

Methods:

Semi-structured qualitative phone interviews were conducted with patients who were offered an early outpatient PT visit secondary to patients’ primary appointment for LBP with a non-operative sports medicine specialist physician. Interviews were recorded, transcribed, and analyzed to identify themes using an iterative process.

Results:

Forty participants were interviewed; 20 accepted early PT services, and 20 did not. Patients’ decisions were influenced by perceived provider training, costs, doctor recommendations, wait times, symptoms, and a desire for a diagnosis. Patients preferred the care of non-operative sports medicine doctors over physical therapists for LBP due to their beliefs that favored doctors’ diagnosis and management of LBP. Patients perceived exercise as an effective treatment for back pain. Physical therapists were viewed as an adjunct service, despite positive comments about PT and the belief that exercise is one of the most effective treatments for LBP.

Conclusion:

Barriers including costs, patient preferences, and knowledge about physical therapists limited patients’ use of PT. Value-based care strategies aimed at improving the management of LBP increasingly promote the early use of PT. For these strategies to be effective, it is critical that patient perceptions and the influence of barriers on PT use are further understood. This study highlights the need to promote confidence in physical therapists’ expertise in the management and diagnosis of lower back pain.

1. Introduction

The use of physical therapy (PT) in patients suffering from lower back pain (LBP) is driven by a variety of factors that are not well understood (Aguirre et al., 2019; Sharpe et al., 2020; Thackeray et al., 2017). PT use is recommended by evidence-based guidelines for LBP (Childs et al., 2015; Chou et al., 2017; Qaseem et al., 2017). Despite the evidence base supporting PT, PT is utilized at low rates in the United States which likely contributes to poorer clinical outcomes and including increased downstream medical utilization (Garrity et al., 2020; Sharpe et al., 2020; Thackeray et al., 2017; Zheng et al., 2017).

In some countries, the reasons for low use rates of PT involve a complex relationship between patients’ decisions, knowledge, and barriers. Several studies have identified factors that influence patients’ decisions to use PT, including patients’ ability to navigate the health system and attitudes toward PT (Aguirre et al., 2019; Bath et al., 2016; Huber et al., 2019; Khoja et al., 2020; Kirby et al., 2014; Oosman et al., 2019; Sibbritt et al., 2016; Thackeray et al., 2017; Zoorob et al., 2017). Patients appear to make health care choices based on their preferences, values, perceived likelihood of success, and the severity of their condition (Fotaki et al., 2005; Kirby et al., 2014; Klojgaard et al., 2014; Modi et al., 2014; Sibbritt et al., 2016; Sweeting et al., 2011). In Australian women with back pain, patients’ provider decision was influenced by patients’ familiarity with treatments and practitioners, social recommendations, distance to the clinic, pain severity, and the perceived qualifications of the providers (Kirby et al., 2014; Sibbritt et al., 2016).

Most of the studies evaluating patients’ decisions towards PT have been performed in research settings using discrete choice or survey methods. Little is known about how patients in routine clinical settings decide to use PT for LBP. Patients in the United States are somewhat independent in their care choices relative to other countries, but they are presented with a complex health system that has limited coordination between providers, systems, and payers (Lipsitz, 2012; Penm et al., 2017). In other countries, including the United Kingdom, PT use is required before advancing to the care of surgeons or specialist physicians, limiting the influence of patient choices (Sharpe et al., 2020). In the United States, patients appear to play a pivotal role in driving health care use. Thus health care systems must understand patient decision-making processes to direct care more effectively (Sharpe et al., 2020).

This study was part of a quality improvement initiative to improve patient use of the RapidAccess program for LBP in the *** health system. The initiative was designed to reduce barriers to PT care for patients with LBP who call for a visit with a physical medicine and rehabilitation physician. The RapidAccess program requires the call center staff to identify the patient based on their diagnosis and insurance, screen the patient using a guide, offer the program, and document their decision (Magel et al., 2018). As part of the clinical flow, the clinical staff provides a PT appointment within 72 h of their scheduling call (Magel et al., 2018). Only 31% of patients accept the program’s PT visit (Magel et al., 2018). This rate is problematic as many patients are not benefiting from guideline concordant PT treatment, but it is consistent with use rates observed in other systems (Aguirre et al., 2019; Dennis et al., 2018; Freburger et al., 2018; Maldonado and Greenland, 1993; Salt et al., 2016; Sandstrom, 2017; Sharpe et al., 2020; Thackeray et al., 2017; Zheng et al., 2017). The purpose of this study was to explore the factors that influence patient decisions to use PT in a routine clinical setting.

2. Methods

We conducted a qualitative study using semi-structured interviews with patients who were offered an outpatient PT appointment at the *** health system through the RapidAccess program (Magel et al., 2018). We examined factors that influence patient decisions to use PT care, expanding upon previous research (Aguirre et al., 2019; Bath et al., 2016; Huber et al., 2019; Khoja et al., 2020; Kirby et al., 2014; Oosman et al., 2019; Sibbritt et al., 2016; Thackeray et al., 2017; Zoorob et al., 2017). The *** Institutional Review Board approved this study. All participants provided consent using a consent cover letter through REDCap before participation and were made aware they could leave the study at any time. This study adhered to the Consolidated Criteria for Reporting Qualitative (COREQ) checklist (Tong et al., 2007). Survey data was collected in REDCap, and interview data were stored using encrypted storage devices. Patients who were not offered participation in the RapidAccess program and non-English speaking patients were excluded since the interviewer only spoke English and interpretation services were unavailable.

The cohort included patients ages 18–70 in the RapidAccess program. Inclusion in the RapidAccess program required patients to have a primary complaint of LBP with or without sciatica and had called the *** outpatient orthopaedic center to schedule an appointment with a physical medicine and rehabilitation physician (also known as physiatrists or non-operative sports medicine doctors) using commercial insurance. The RapidAccess program does not track patients’ prior health care use. Some patients were seen by primary care providers, including primary care physicians (also known as family practice doctors or physicians and general practitioners) before interacting with the RapidAccess program. The RapidAccess program is fully described elsewhere (Magel et al., 2018). The program adds a secondary/tertiary PT care pathway for patients seeking a visit with a physical medicine and rehabilitation physician.

Interviews were completed until 20 patients who accepted the RapidAccess program and 20 who declined participated through opportunity sampling. Recruitment goals were based on Ando’s definition of coding saturation of around 12–20 participants (Ando et al., 2014; Hennink et al., 2017). Attendance to the PT or physical medicine and rehabilitation physician appointment was not required before the interview was conducted. We attempted to interview patients within 72 h of their call center interaction to reduce the influence of other health care interactions on patients’ perceptions. Patients were identified through reports of the RapidAccess PT program from ***’s Electronic Data Warehouse and recruited via email and phone calls.

2.1. Data collection

Crabtree and Miller’s iterative process informed data collection for conducting clinical qualitative studies (Clark-Cutaia et al., 2019; Crabtree et al., 1995). Primary data collection involved single semi-structured individual telephone interviews. The interviewer conducted the interviews in a private office, and participants were not required to be alone, but they were required to answer questions independently. Guided by Andersen’s Behavioral Model of Health Services Use, background literature, and the expertise of the team members, we designed the semi-structured interview guide to elicit patients’ preferences and opinions about providers and treatments for LBP focused on PT use described in Fig. 1 (Andersen, 2008; Babitsch et al., 2012). Patient input was not included during the development of the interview guide. Andersen’s model organizes factors into predisposing, enabling, and need factors (Andersen, 2008; Babitsch et al., 2012; Bradley et al., 2002). Predisposing factors include demographics, social, and health beliefs. Enabling factors are the organizational and financial conditions that enable care use. Need factors include patients’ and providers’ perceived need for health care.

Fig. 1.

Andersen’s behavioral model of health services use.

For this manuscript, we use the terms decision or decide to describe the process patients go through to use PT care, recognizing that the word “choice” is often used to describe the same processes. In each step of this project for methodological rigor, we included at least one team member directly involved with the RapidAccess program (**, **, **) and one team member who was not directly involved in the program (**, **, **).

The interview guide was designed with input from all team members. ** has extensive experience in qualitative research, ** has been trained in qualitative research, and the other researchers have clinical research expertise (**, **, **, and **). The guide was pilot tested with four participants to allow for guide modification for any unexpected themes. The pilot-tested participants were included in the final analysis as only minor grammatical and clarification changes were made to the interview guide. The interview guide (Appendix A) started with broad questions transitioning to more specific questions, eventually using examples of factors from the pilot testing and past research as probes (Kirby et al., 2014). Using rank-based questions, the interviews concluded by asking the participants’ thoughts about treatments and different providers (primary care, physical therapists, physical medicine and rehabilitation physicians). These included the rank-based question, “can you please rank the most important factors that influenced your decisions for your back-pain care.” The other rank-based questions included: the provider the patient saw first, the provider they believe is best to diagnose their back pain, and what they think is the best treatment for their back pain. These questions were included to further explore patients’ perspectives about their care options for LBP and evoke discussion.

The primary investigator, **, conducted all interviews and was the primary coder. ** is a male clinical/academic physical therapist and was a Ph.D. candidate during the interviews. Participants were made aware during the consent process that the interviewer was a physical therapist for transparency.

2.2. Data synthesis and analysis

The data analysis occurred in 4 steps using Crabtree and Miller’s iterative framework (Crabtree et al., 1995). Throughout this process, the research team searched for themes and used previous steps to inform later stages similar to the Grounded Theory Approach (Fitzpatrick and Boulton, 1996; Gupta et al., 2017). During the interviews ** took field notes that drove later analyses. Interview transcripts were transcribed using NVivo’s online transcription service and were assessed for accuracy by a research assistant and the primary author. All interviews were deidentified before analysis.

The research team then searched for “meaningful units or segments” in the text during the coding and theme identification process (Crabtree et al., 1995). To reduce the influence of the positionality of the clinical team members, we included at least one research team member in each step of the process who was not a clinician, ** and **.

The codebook was developed in two stages with collaboration between **, **, and **. The first stage used inductive reasoning to establish an initial codebook from the first 12 interviews, with 6 interviews from each group. In the second stage, the codebook was refined to consolidate codes and eliminate or add necessary codes. The new codebook was then applied to another 8 cases examining for redundancy. The rest of the codes were defined and agreed upon between the 2 primary coders. Specific question-based codes were coded once each time the specific question was asked. We considered any codes not connected to question-based codes, a global code, and coded them each time a relevant statement was found. Throughout this process, themes were identified by examining for pertinent patterns in the data.

The primary author (**) coded all interviews in groups of 10, the secondary coder (**) evaluated all codes involved in the first and second stages. Each stage was not complete until there was coding agreement between coders. In the final stage, the potential themes identified throughout the process were discussed and the most pertinent themes were determined by **, **, and **. After these themes were identified, the entire study team was provided with the coding and theme findings to provide feedback. Theme pertinence was determined based on each theme’s expected impact and relevance to the patient. In our final analysis step, to maximize the rigor of our analysis process and representation of the patients’ thoughts, we conducted a member checking process. Each participant was emailed a draft of the paper and given 2 weeks to provide input.

2.3. Role of the funding sources

This study was funded by the University of Utah’s Imagine Perfect Care seed funding program. JS was supported by the Promotion of Doctoral Studies I and II scholarships from the Foundation for Physical Therapy Research. JS was also funded by the Durham Center of Innovation to Accelerate Discovery and Practice Transformation (CIN 13–410) and VA Quality Scholars Program (OAA #AF-3Q-05–2019-C) at the Durham VA Health Care System during the dissemination of this research. The funders played no role in the design, conduct, or reporting of this study.

3. Results

Among patients offered RapidAccess PT, 188 individuals were invited to participate in the current study. Forty-four (24%) agreed to participate and completed the initial surveys. Forty participants (21%) completed an interview between July 15, 2019, and February 25, 2020. After contacting participants for input on the manuscript, we received several replies the paper but no requests for changes. Interviews lasted an average (±standard deviation) of 40 ± 23 min.

3.1. Sample characteristics

The sample’s mean age was 45 ± 13.3 years old (Table 1). The cohort was predominantly White (n = 34, 85%), non-Hispanic (n = 38, 95%), college/postgraduate educated (n = 29, 73%), working full time (n = 25, 63%), and married or living with a partner (n = 26, 65%). Participants were interviewed after their scheduling call on average 11.2 ± 4.7 days in the accepted group and 13 ± 5.1 days in the did not use group.

Table 1.

Demographic characteristics by group.

Demographic	Accepted n (%)	Did not accept or did not remember n (%)
Age (Standard Deviation)	43.6 (14.9)	46.4 (11.4)
Race (%)
White	17 (85)	17 (85)
Asian	0 (0)	3 (15)
Other	3 (15)	0 (0)
Hispanic (%)
Hispanic	2 (10)	1 (5)
Did not answer	0 (0)	1 (5)
Non-Hispanic	18 (90)	18 (90)
Education (%)
Less than high school	0 (0)	1 (5)
High school diploma or GED	6 (30)	4 (20)
Four year college degree	5 (25)	9 (45)
Some post-graduate beyond four year degree	3 (15)	2 (10)
Postgraduate	6 (30)	4 (20)
Work status (%)
Full time	14 (70)	12 (60)
Student	2 (10)	2 (10)
Part time	0 (0)	2 (10)
Modified Duty	0 (0)	1 (5)
Other health	0 (0)	1 (5)
Back permanent	0 (0)	1 (5)
Unable to work due to back	0 (0)	1 (5)
Unemployed but seeking work	2 (10)	0 (0)
Home Maker	1 (5)	0 (0)
Retired	1 (5)	0 (0)
Marital status (%)
Divorced	3 (15)	3 (15)
Married	8 (40)	11 (55)
Live with Partner	4 (20)	3 (15)
Never been Married	5 (25)	3 (15)
Patient recall of the program (%)
Understood	18 (90)	10 (50)
Unclear	1 (5)	3 (15)
Did not remember	1 (5)	7 (35)

Forty percent of patients either did not know or were unsure that they were asked by the call center about the RapidAccess PT program, 10% among those in the accepted RapidAccess group, and 50% among those who declined RapidAccess. This finding was unexpected because the call center is required to use a defined script and document if the patient was offered the program. These patients were included in the general analyses, but they were excluded from Table 3 as these patients were fundamentally different from the patients who made an active decision to use or not use PT. Patients’ difficulty remembering or understanding the RapidAccess PT program was a clear barrier to PT care. The research team included these patients in the analyses because they likely represent a significant portion of health care users. There was one patient who did not remember being asked about the RapidAccess program, but they still used PT through RapidAccess. They were included in the accepted group.

Table 3.

Patient preferences.

Rank-based questions	Yes n (%)	No n (%)
“Of everything we discussed, what was most important when making your decision”	Yes (%)	No (%)
Provider training	3 (16)	2 (20)
Costs	3 (16)	3 (30)
Recommendation	2 (11)	1 (10)
Time	1 (5)	1 (10)
Symptoms	3 (16)	1 (10)
Diagnosis	2 (11)	0 (0)
Past Experiences	2 (11)	0 (0)
Reputation	0 (0)	2 (20)
Trust	2 (11)	0 (0)
Social	1 (6)	0 (0)
“Who is the best provider to diagnose your back pain?”
Physical Medicine and Rehabilitation, Sports medicine and Orthopaedics	16 (84)	10 (91)
PT	2 (11)	0 (0)
Primary Care	1 (5)	1 (9)
“Who was the first provider you saw for your back pain?”
Physical Medicine and Rehabilitation, Sports medicine and Orthopaedics	12 (40)	5 (36)
Primary Care	7 (23)	7 (50)
Chiropractor	4 (13)	2 (14)
Physical Therapist	4 (13)	0 (0)
Urgent or Emergency Care	1 (3)	0 (0)
Surgery	2 (7)	0 (0)
“What type of treatment do you think would work for you back pain?”
Exercise	13 (72)	8 (80)
Injection	2 (11)	0 (0)
Medication	2 (11)	1 (10)
Surgery	1 (5)	1 (10)

Notes: This chart only includes patients who were aware of the program. The percentages are the frequency each response is selected divided by the total number of responses. Participants could select none, one, or more than one.

3.2. Identified themes

Coding analyses showed 3.6% disagreement between coders (** and **) and did not result in substantial changes to theme identification. All conflicts were resolved with discussions until there was 100% agreement using the codebook described in Table 2. Saturation occurred at 15–20 participants per group—consistent with our recruitment plan (Ando et al., 2014). Based on our unique patient population and limited past research on patient PT use decision-making processes, the research team determined all themes were relevant and are reported below.

Table 2.

Code definitions.

Code	Definition
Treatment Choice
Location	Driving or travel distance or location of the clinic
Social	Social advice or influence on decision (can include health professionals outside of professional capacity)
Time	Days to appointment from scheduling
Costs	Any direct treatment costs including copay, insurance, deductibles. Can include costs to the health system
Symptoms	Pain, functional limitations, or other symptoms driving their decision
Prior level of function	Goal or mention of getting back to their prior level of function
High	High functioning, athletes, manual labors, physically demanding jobs, patients capable of walking 10,000 steps
Normal	Sedentary jobs, less than 6000 steps a day, walking at most for recreation
Self-management	Reports trying to self-manage symptoms (ice, heat, massage, exercise, etc.)
Diagnosis	Desire for a diagnosis
Pathoanatomical	Pathoanatomical diagnosis (possibly not directly related to their choice)
Patients’ Familiarity-Relationship with the Health System	Patient’s relationship with the health system
Patients’ perception of provider training	Includes credentials (ex MD vs PA), years of training, specialization, experience with patients with back pain
Trust	Comfort level with their provider, office, or health system
Past experiences	Any experiences, positive or negative
Reputation of the health system	Reputation of the health system from online, social, etc.
Provider referral or recommendation	Recommendation from any medical provider previously seen regarding other providers or treatment
Question Based Codes
Primary reason for patient’s choice	Primary factor driving the patient’s choices in the health system
Choice first	The provider patient first saw for back pain
Best to diagnose	The provider the patient identifies as best to give a diagnosis for their back pain
Best treatment	The treatment that the patient believes will best resolve their back pain

Seven global themes were identified and organized by the possible patient-level impact of each theme. The themes were: symptoms, diagnosis, patients’ perception of providers’ training, recommendations, time, costs, and familiarity-relationship with the health system. Subthemes under familiarity-relationship included past experiences, reputation, trust, and social influences. Figs. 2 and 3 show coding trees, including the number of patients who mentioned each theme. Nonspecific mentions of physicians (doctors) without specialty are reported as physicians.

Fig. 2. Coding tree examining patient choice.

Notes: n= is the number of participants who made statements that fit the corresponding code. These codes are not based on specific and direct question. Each code under the main question is organized going left to right and top to bottom by the number or participants mentioning each.

Fig. 3. Coding tree for question-based codes.

Notes: These themes are based on direct questions asking each topic (e.g., primary reason for making choices in the health system), Codes are organizing top to bottom by the number or participants mentioning each.

3.3. Symptoms

Patient symptoms were frequently mentioned during the interviews and often overlapped with the time theme. For example, a patient who accepted RapidAccess cited their symptoms and functional limitations as driving them towards any providers they could see the fastest with the fewest barriers, without preference for a provider type. Functional limitations drove some patients to access care exemplified by the following statements: "[I] just want to be healthy,” “start enjoying life more,” and “have less back pain, so [I] can go out [and] enjoy [my grandkids]."

3.4. Diagnosis

Participants regularly desired an understanding of their diagnosis, channeling their care decisions towards physicians. One participant was particularly adamant that they receive diagnostic imaging for their back pain:

I really wanted someone to take a look at my back and determine that there’s something going on - more than [physical] therapy [can do] …. I want someone to X-ray it or do whatever scans they do, to find out what’s going on. Because I’m concerned that there’s something more than just a muscle thing going on.

Patients preferred physical medicine and rehabilitation physicians to diagnose their back pain over physical therapists. Statements including, “the therapy folks, they are not seeing the underlying process that either caused injury or disease,” and that physical therapists will work with patients "... once a diagnosis has been made by someone of merit [a physical medicine and rehabilitation physician]," were examples of patient thoughts regarding physical therapists’ ability to diagnose back pain.

3.5. Patients’ perception of providers’ training

Participants declining or not remembering being offered RapidAccess often indicated that a provider’s training drove their decision-making and was associated with the quality of care they would receive. Credentials were particularly important to one patient who expressed frustration with “being exposed to the shotgun method too many times” by primary care providers and physical therapists, which led them to start to “question credentials.” “The shotgun method” was described as receiving a nonspecific cluster of treatments. This patient preferred the care of an orthopaedic “spine or back specialist” physician from a prominent health system. Another patient sought care from physical medicine and rehabilitation physicians because of the “nature of their work,” indicating that they are medically trained, treat orthopaedic conditions, and are more qualified to diagnose and treat back pain than other providers.

3.6. Recommendations

Physician recommendations influenced patients’ care use, including downstream PT use. One participant stated, “the primary care physician obviously knows what she’s talking about,” leading them to follow the recommendations of their provider and see a physical therapist.

3.7. Time

Several patients were concerned with using care quickly to avoid further pain and disability, as expressed by a participant who did not use RapidAccess PT:

If the back pain is related to the nerve damage or the tingling that I’m feeling, I know the sooner you get treated and get started... the better off you’re going to end up... rather than waiting.

One participant, who did not recall being offered PT, stated “the most important (factor) by far was the urgency” due to their high levels of pain driving them towards emergency care. This participant mentioned that they “would have definitely taken that PT appointment” if their symptoms were “less urgent” and they had been aware they were offered RapidAccess.

3.8. Costs

Patients often mentioned their perceived costs associated with treatments as a barrier to care. For example, copays deterred patients from PT because they expected to have several visits with a physical therapist, which is typical of PT care in the US. In the United States, copays are a payment that patients make in addition to insurance reimbursement. Copays are often the same amount for physical therapists and specialist visits. Physical therapists provide a specialized service, but the specialist copay structure is designed based on physician treatment schedules. Our qualitative findings show that copay associated costs deter patients from initiating physical therapy care. One patient avoided PT, stating:

Yeah, I think it’s expensive, [that] is probably one of the biggest factors. Honestly, cause if it was like, you know - what are copays now, like, 30 dollars, 35 dollars for an appointment. What if they prescribed the 8 appointments?

3.9. Patients’ familiarity-relationship with the health system: past experiences, reputation, trust, and social influences

Although not as commonly mentioned as the other themes the influence of patients’ familiarity-relationship with the health system appeared to be influential for some patients. This theme included the sub-themes of the reputation of the health system, patients’ past experiences with health care, trust in the providers, and the social influences of friends and family. Each of these factors was frequently mentioned when participants were discussing their general decisions in the health system.

3.10. Question-based codes: the provider the patient saw first, the provider who is best to diagnose back pain, and the best treatment for their back pain

Patients most frequently saw physical medicine and rehabilitation physicians or primary care as their entry point in the health system. Despite the patient preference for the care of a physician (doctor), patients overwhelmingly indicated that exercise was the best treatment for LBP. Patients who did not accept RapidAccess PT often chose primary care as their first provider and physical medicine and rehabilitation physicians as the best to diagnose their pain. Patients who accepted RapidAccess PT were more diverse in their first provider choice. The participants who did not accept RapidAccess PT prioritized provider training, health system reputation, and costs. The PT participants were also more diverse in the factors that influenced their health decisions. Table 3 describes the question-based codes between the patient groups, excluding those who did not recall being asked about RapidAccess by the call center.

4. Discussion

This study demonstrates the intricacy and complexity of the patient decision-making process for LBP care. We found that patient decisions often develop and change over time, consistent with past research (Aguirre et al., 2019; Bath et al., 2018; Kirby et al., 2014; Oosman et al., 2019; Sibbritt et al., 2016; Thackeray et al., 2017; Zoorob et al., 2017). The factors that influence patients’ decisions include patients’ knowledge or beliefs, past providers’ medical recommendations, patients’ symptoms or functional limitations, costs, and location. Patients did not appear to avoid PT; instead, they preferred the care of a physical medicine and rehabilitation physician whom they perceived as a “spine or back specialist.” This preference corresponded with patients’ perception that physical medicine and rehabilitation physicians are the “highest qualified” providers to diagnose and treat LBP, likely due to their ability to order imaging (Sharpe et al., 2020).

We expanded on past research that Andersen’s predisposing, need, and enabling factors influence patients’ decisions and their ability to use health care for their back pain (Babitsch et al., 2012). When participants were asked to rank the most important factors influencing their decision the enabling and predisposing factors were mentioned more often than need factors. This finding suggests that patients’ ability to use care and their preferences drive health care utilization and access more than their symptoms, functional limitations, or perceived need. As shown in prior studies, enabling factors, including costs and copays, were significant barriers to PT use (Aguirre et al., 2019; Bath et al., 2016; Huber et al., 2019; Kirby et al., 2014; Oosman et al., 2019; Sharpe et al., 2020; Zoorob et al., 2017). In the United States, patients are frequently charged a copay each time they see a specialist, including physical therapists. As it is common for patients to see physical therapists over several visits, their out-of-pocket costs can be significantly more than a physical medicine and rehabilitation physician whom they may see only a few times. Need factors, including patients’ symptoms and diagnosis, were less but still influential, consistent with other research showing that worse symptoms are associated with health-seeking behavior (Sibbritt et al., 2016).

There appeared to be some differences between the patients who accepted PT and patients who did not or who were unaware that they were offered RapidAccess. PT participants saw a variety of providers as their first contact in the health system and cited a variety of factors that led to their decision. Patients who accepted early PT frequently mentioned that high levels of symptoms prompted them to seek PT care. This finding contradicts past research suggesting that high levels of symptoms drive patients towards medical providers over non-medical (e.g., physical therapists) providers (Sibbritt et al., 2016). Patients who did not accept or were unaware of the program often emphasized that their expected outcomes and desire for a diagnosis by a physician drove their health care decisions. This group repeatedly reported following a more conventional care pathway for LBP using the care of a primary care physician often followed by a specialist.

Several barriers appeared to influence patient decisions in those who did not accept or did not recall being offered RapidAccess PT. These included their health care knowledge, costs, wait times, and their ability to navigate the health system. Each factor has been shown to reduce the use of PT in patients with pelvic floor or back pain diagnoses, homeless populations, and underserved populations (Aguirre et al., 2019; Bath et al., 2016; Huber et al., 2019; Oosman et al., 2019; Zoorob et al., 2017). The patients who were unaware of being offered the program may have had more difficulty navigating health systems leading to possible delays in care. Delayed care has been associated with unmet health needs, financial burdens, and increased use of surgeries and hospitalizations (Aguirre et al., 2019; AHRQ, 2016; Fritz et al., 2015; Sharpe et al., 2020; Thackeray et al., 2017). Our findings thus highlight the need to study patients and barriers to care, including patients’ ability to navigate the health system, to improve access to care and outcomes.

Patients’ belief that physical therapists cannot diagnose back pain was a distinct barrier to treatment. Interestingly, patients believed that exercise was an effective treatment for LBP but preferred other providers over physical therapists despite mostly positive comments about PT. These decisions demonstrate a disconnect between patients’ desire to receive a pathoanatomical diagnosis for their back pain and their belief that exercise is an effective treatment. Participants’ desire for a diagnosis may be related to a need for reassurance about the origin of their symptoms consistent with past research (Holt et al., 2015; Traeger et al., 2015). Current evidence shows that pathoanatomical diagnoses are poorly associated with symptoms and outcomes in non-complicated mechanical LBP (George et al., 2020; Tonosu et al., 2017). Patients seem unaware of the poor correlation between imaging findings and symptomology in LBP. Patient education about physical therapists’ ability to diagnose and manage LBP pain as exercise experts could increase the use of PT since patients already believe in exercise treatment. Programs like the RapidAccess PT program could increase the use of PT by increasing patient interaction with physical therapists as frontline providers, thus improving patient knowledge through experience.

Past studies have shown that the location of clinics, transportation, and social network recommendations influence the use of care, but our participants rarely mentioned these as significant factors (Aguirre et al., 2019; Bath et al., 2016; Huber et al., 2019; Kirby et al., 2014; Oosman et al., 2019; Zoorob et al., 2017). The participants were mostly white, non-Hispanic, educated, and employed full-time. This lack of diversity may explain why barriers in other patient groups, including clinic location, discrimination, transportation, or insurance, were not as influential in the interviewed patients (Aguirre et al., 2019; Bath et al., 2016; Huber et al., 2019; Oosman et al., 2019; Zoorob et al., 2017).

4.1. Limitations

Our study has several important limitations. First, a portion of patients in the RapidAccess PT program (40%) did not remember being offered the program suggesting a communication issue between the call center and patients. Patients’ inability to remember the program could be due to their difficulty navigating the scheduling system. This communication issue limited our ability to examine patients’ decisions in this group, but we included them in our analyses since they represented an important group of healthcare users. Health care systems could address poor patient recall by improving patient information exchange through greater investment in scheduling staff and patient educational programs. Poor patient recall and difficulty navigating the health care system are likely not unique to this cohort. Second, we expanded our analysis after pilot testing to include all patient decisions made for back pain related to their most recent episode of care. The decision-making process was more complicated and protracted than we expected. Some participants had several episodes of LBP or were not clear about who they decided on as their first provider. Third, we attempted to interview patients as rapidly as possible after their scheduling call but were unable to interview some patients before they saw a health care provider, which may have influenced the patients’ answers to the interview questions. Fourth, our interviews, analysis, and this report have all been influenced by the primary investigator’s (**) positionality. His positionality includes his clinical background and the positionality of the other authors. Lastly, we did not obtain patient input in the design of this study.

4.2. Conclusions

To our knowledge, this is the first qualitative study examining patient decisions towards PT for LBP in a clinical setting in the United States. Patient decisions towards PT for back pain were influenced by a variety of factors, including patients’ ability to navigate barriers and understand the health system. Similar to past research, we found that lack of patient knowledge of and experience with physical therapists, including PTs ability to treat and diagnose pain, were significant barriers to PT use (Aguirre et al., 2019; Bath et al., 2016; Huber et al., 2019; Kirby et al., 2014; Oosman et al., 2019; Zoorob et al., 2017). Often, patients overcame cost or wait time barriers to schedule with a physical medicine and rehabilitation physician, whom they perceived as a “spine or back specialist” providing high-quality care. Physical therapists were viewed as providing adjunct services that occur after a patient has a visit with a physician for a diagnosis and treatment plan. If patient perceptions about physical therapists do not change, the use rate of PT will likely continue to be low in the United States. Physical therapists should educate patients and the public about their ability to diagnose and treat LBP. As our participants were part of a unique program, future work needs to explore how patients make decisions and what influences their use of PT in typical clinical settings. Increased understanding of patient decisions will help health systems and physical therapists reduce barriers and direct care more effectively towards PT, thus improving outcomes and health care value.

APPENDIX A. Semi-structured interview guide

General introduction

May I please speak with Mr./Mrs./Ms./Dr. (insert participant last name)?

If not available: leave a message for the participant to call back at their earliest convenience.

My name is (name) and I’m calling from the ‘patient choice study’ at the *** that you recently scheduled an interview for. Does this time still work well for you to complete the interview?

If not, when would be a good time for me to call you back?

As a reminder, we are interviewing you to help us better understand why patients choose certain treatments over others for their lower back pain. This interview is designed to take 30–60 min to complete, although it can go longer. Before I begin, do I have your permission to audiorecord our conversation? We will record the conversation so that we can transcribe it later. We want to make sure we don’t miss anything.

If you need to take a break or have any questions at any time, please let me know. You may leave the interview at any time or keep any information out of the report and study.

Do you have any questions before we begin the interview? [start audio recorder after answering questions]. Start recorder(s)

This is interview [insert ID#] on [insert date].

Please remember that everything you share with me will be kept strictly confidential. None of your information will be shared with any of your health care providers. Please know that there are no right or wrong answers to these questions; we are simply interested in your thoughts. We plan to share our research findings to improve patient care for lower back pain and musculoskeletal conditions. Your name will not be associated with any of your responses. We may use a sentence or two from the interview in presentations and publications. You will have the option to look this over before we use a quote if you choose.

First, we want to make sure you are eligible for the study.

1. Did you recently call the *** Orthopaedic center for a physical medicine and rehabilitation (PM&R) physician, also known as a physiatrist (Physical Medicine and Rehabilitation physician), appointment for back pain?

2. Were you asked if you would like to have an appointment with a physical therapist while waiting for your physiatrist (Physical Medicine and Rehabilitation physician) appointment, known as a RapidAccess appointment?

3. Which type of medical care for your back pain have you received at this point? (Can jump to Reasons for their choice at any time)

   1. How long have you had back pain?
   2. Who was the first provider you saw for your back pain?
   3. What treatment suggestions did they make for you, including other providers/medical care?
   4. Did they refer you to any other providers?
   5. Did you see a second provider?
   6. Any others?
   7. Do you have or plan for any upcoming appointments with a health care provider for your back pain?
   8. Have you seen any other providers or plan on seeing any other providers for your back pain? If yes, what kind of care?

4. Are you planning or do you have any other future appointments scheduled?

Now, I am going to ask you some questions specific to the types of care you have received. We want to explore how you decided to access (insert the healthcare providers they saw here that they described already), your back pain.

Reasons for their choice

1. You were involved in a program called RapidAccess where you were offered an early visit to physical therapy while waiting for your physiatry appointment.

   1. Did you feel like anything was confusing or unclear about the program when you were talking to the call center staff?
   2. Was there any aspect of what the call center staff said or did that influence your decision about care?
   3. Do you want to add anything about the phone conversation you had with the call center staff that may have influenced your decision?

2. Can you tell me why you chose to go to your **insert appropriate medical provider** first/second?

3. What influence did past experiences in the medical system with treatments for your back pain have on your decision?

   1. Past experiences

      1. Who do you typically see first for your lower back pain when you have an increase in symptoms?
      2. How do your experiences with other musculoskeletal (bone/ muscle/joint) problems and health care influence your decision?
      3. Did any other health care providers influence your decision of the provider(s) you decided to see, possibly a primary care physician you saw first?

1. Referrals?

   • iv.
     How did the ease or difficulty of receiving treatment influence your decision, including the location and availability of appointments?

1. Time, pain, costs, etc.

   • v.
     How did the wait time for the physiatry appointment influence your choice of using PT or not?

1. Wait time for physical therapy?

   • b.
     Did your friends and family, including on social media, influence your decision?

     1. Please describe their influence.
     2. Was it on social media, in person, etc.?
     3. Did you have any information that you knew before from either schooling, training, or other experiences that influenced your decisions?
   • c.
     Treatment types - There are many different options for lower back pain treatment. I am going to ask you a couple of questions regarding your perceptions of the treatments available.

     1. What type of treatment do you think would work for your back pain?
     2. How would seeing a physiatrist (Physical Medicine and Rehabilitation physician) help?

1. How would medication help? (Can ask ‘would medication help?”)

2. Injections?

3. Surgery?

   • iii.
     Primary care?
   • iv.
     Surgeon?
   • v.
     What type of benefits would you get from going to a physical therapist?
   • vi.
     Do you think seeing another provider (chiropractor, massage, acupuncture, etc.) would help? Who would help and how would they help?
4. Many people have a goal(s) or an outcome they are seeking from treatment. What were/are you most hoping to achieve from treatment for your back pain?

   1. Pain reduction?
   2. Functional improvement?
5. Did copays or costs of medical care influence your decision?

   1. How did insurance costs or copays influence the choice of provider you saw first?
   2. How did financial factors influence your choice of provider?
6. Did your health influence your decision of who to see?

   1. General

      1. Self-Assessed Health Status - In general, how would you rate your health?

1. Excellent, very good, good, fair, or poor

   • ii.
     Do you have other health conditions that change how you manage your back pain?
   • iii.
     Do you exercise regularly?
   • iv.
     Do you smoke? If so, how much?

     • b.
       Did your level of symptoms or limitations with activities have any influence on your choice of health care provider?

       1. Did a desire for a diagnosis (reassurance) influence your decision? If so how or how not?
       2. Who is the best provider to diagnose your back pain?
       3. Did a desire for reassurance that everything was OK influence the provider you chose to see? If yes how so?
       4. What did you want from the treatment for your back pain?
       5. Did that influence the provider you chose?

Conclusion

Is there anything else you would like me to know?

The main points that I picked up from our discussion are (outline the findings)

Rank the importance of what influenced your decision from most to least.

Is there anything I missed?

Thank you for participating. The information you provided will be very helpful in improving patient care for lower back pain and musculoskeletal conditions. All information will be confidential, and if you want to look over anything before it is published, I can provide you the necessary contact information. Thank you and enjoy the rest of your day.

Demographics and questionnaires on REDCap

If they have not completed the questionnaires on REDCap, please remember to complete the questionnaires on REDCap.

1. Basic demographics using the Pew Research Demographics questions minus the political questions.

2. Oswestry Disability Index.

3. Verbal Analogue Scale for pain: Subject is asked: On a scale of 0–10, with 0 being no pain and 10 being the worst pain imaginable, what you rate your current pain?

4. PEG: Three-Item Scale Assessing Pain Intensity and Interference

If needed:

Probes

General: Try to capture reasons or thought processes behind patient choices in the health system regarding their back pain.

• Use:

• Please say more about…

• Describe …

• What is it about…

• What caused you to …

• Examples:

• Reasons for their choice

• Past experiences with medical or physical therapy treatment

  • Context and organization of health care system
• Beliefs about treatment effectiveness

  • Beliefs about what they need

• Insurance and financial influences on their choice.

• Health habits - do they exercise, smoke, drink etc.