Throughout our professional world there has been a push to establish PT as the primary healthcare specialty to address musculoskeletal pain. This objective was expressed broadly in APTA’s Vision 2020, and different models have been proposed or implemented1 in an effort to fulfill our profession’s potential.
To the extent that physical therapy has raised awareness among consumers, treatment for low back pain (LBP) is increasingly being recognized as a service in which we have expertise. This trend can be observed anywhere from op-eds by wellness writers in the New York Times to your significant other’s coworker asking you to check their scoliosis at a holiday party. Sociopolitical issues such as our nation’s evolving views on opioids and healthcare spending have also contributed to more conversation about low back pain, which is now thought to be the single diagnosis most responsible for disability in the US and is predicted to reach epidemic status in the near future.
Despite our aspirations, it may be premature to believe that this is an inflection point for our profession. The current guidelines for primary care physicians recommend plenty of services that PTs can provide (massage, exercise, etc.), yet PT is never explicitly recommended. This may explain why it’s estimated that only 7-13% of patients with non specific LBP get referred to PT in the United States. It’s therefore incumbent on us to prove how utilization of PT services is the best way to meet this societal need. To this end, a recent paper commissioned by the APTQI (Alliance for Physical Therapy Quality and Innovation) concluded that:
adherence to recommended active forms of physical therapy results in lower healthcare costs due to lower likelihood of receiving opioid prescription medications, epidural injections, follow-up advanced imaging, follow-up physician visits, and other health resource use metrics without compromising patient outcomes.2
To prove the value and qualification of physical therapy for the treatment of LBP, this new line of research primarily investigates outcomes, relative cost, and the relevance of evidence-based skilled services. Secondary variables such as intensity and latency of PT services are also of interest, given what they can say about the way PT is utilized. The following is a survey of recent research broken down by each primary criterion.
In utilization studies of populations in the thousands to tens of thousands (collected from employee-sponsored, military, Medicaid, and Medicare insurance databases), outcomes can be inferred from the chronology of procedural codes assigned to patients with LBP diagnosis codes as they move through the medical system. Billing codes that are suggestive of unsatisfactory outcomes from a course of PT include additional physician visits, advanced imaging, surgery, pharmacological interventions, as well as a second referral to PT.
A majority of studies found that initial referral to PT resulted in less incidences of patients seeking additional care for their LBP compared to opioid prescriptions or advanced imaging.1-6 Compared to those receiving no initial treatment, however, patients referred to PT were significantly more likely to receive subsequent advanced imaging, surgery, and injections. Rates of subsequent opioid use between those who were initially given no treatment by their doctor vs. a PT referral were virtually identical. Every study made an effort to control for other variables that could explain their results, including age, physical and mental comorbidities, and concurrent prescription of medication.
Early compared to delayed participation in PT was associated with more favorable results in multiple studies, which suggests PT produces an effect distinct from the natural history of LBP. The study done on a Medicare population found that patients who received 1-2 visits per week for a total of 9-12 visits were the most likely to avoid subsequent LBP-related healthcare spending. A British study1 utilizing a chronicity risk tool to determine PT referral found that a small but clinically meaningful improvement was found in high-risk patients but not medium-risk (low-risk were not referred to PT).
The trends reported in outcomes section correlated to a large extent with cost in these studies. Across the different insurance databases, patients who were diagnosed with LBP and referred to PT incurred larger LBP-related healthcare spending compared to patients who were given no treatment by their primary care physician over periods of up to two years. The results from these larger scale studies likely reflect patients who were not at risk for chronicity or recurrence and attained positive outcomes with just a “wait and sit” approach with or without over the counter NSAIDs. The two largest of these studies,4,6 both authored by the same research group (lead authors John Childs and Julie Fritz), found $1000-2000 less LBP-related healthcare costs over a two-year timeline in the patients who didn’t receive any treatment after visiting their primary care.
Conversely, PT again compared favorably to alternative initial treatments such as advanced imaging, opioids, injections, and surgery. The difference in spending among these subjects was on average of greater magnitude than the difference between the patients who received no treatment or PT (from nearly $5000 with advanced imaging to over $20000 with surgery). Indirect costs were also estimated to be less, with patients who received PT taking 50% less days off work in the British risk-stratification study.
Receiving PT services within two weeks of the receiving the diagnosis of LBP was again associated with more favorable results relative to receiving delayed PT. The highest costs savings were also associated with patients receiving 9-12 PT visits for Medicare Part A spending (44% or nearly $3000 less) and 5-8 in Medicare Part B spending (27% less spending).2
Evidence Based Techniques
Recent research has also dichotomized the billing units from their pool of EMR data in an effort to measure the effect of practice adherent with the physical therapy clinical practice guidelines (CPG) for LBP on outcomes and costs.4,6,7 Units that are considered active such as therapeutic exercise and neuromuscular activity were considered CPG-adherent throughout the course of care, however codes linked to modalities were counted as non-adherent. Manual therapy was considered evidence-based if performed in the acute phase (defined as within 14 days of the start of the episode of care).
Evidence-based care in the four studies on PT utilization that analyzed it as a variable was found to range between 12-40% of total billing units. This is congruent with previous research in which 30% of outpatient ortho PTs self-reported compliance with the guidelines. Patients treated with PT that was considered guideline-adherent consistently received less injections than those who had non-adherent care, and a majority of studies also found significantly less rates of surgery and advanced imaging in the former category as well. No significant difference between the types of treatment was found for subsequent doctor visits or opioid use.
Lower back pain and PT: moving forward
While this body of research doesn’t fully solve the issue of how people with LBP should be managed in primary care, it does offer some insights into when PT might be indicated and the way in which it should be administered:
- Physical therapy is more cost-effective to alternative interventions for non-specific LBP, but isn’t superior to waitful watching among the majority of patients who see their primary care physician for low back pain
- Appropriate referral to PT compared to waitful watching might depend on risk factors for chronicity such as pain intensity, functional limitation, and yellow flags
- The effectiveness of PT is influenced by factors such as the timing of the beginning of service, the dosage of sessions, as well as how active the care is
One of the limitations of this area of research is that it takes such a macro view of the healthcare system that it can’t tell us much about more granular clinical factors that may explain utilization trends. It’s highly possible that there’s a degree of selection bias at play here—a doctor may hypothetically decide to only refer his most difficult patients with LBP to PT, therefore explaining why these patients continue to go on to seek care and incur more costs. Low back pain is not a homogenous diagnosis, and future research should attempt to include analyses such as pain severity and standardized functional limitation (e.g. G-codes) to give a clearer picture on what sort of patient is appropriate for PT.
What is already clear however, given the estimated 25% chronicity rate of LBP in the United States, is that more than just 10% of these patients should be referred into our clinics.