Subsequent Healthcare Utilization Associated With Early Physical Therapy for New Episodes of Low Back Pain in Older Adults Deven Karvelas, MD University of Michigan 10/04/2015, AAPM&R Annual Assembly

Disclosures and Funding Disclosures – None Funding Source – Agency for Healthcare Research and Quality $10 million ARRA CHOICE award - 1R01HS019222-01 Back Pain Outcomes Using Longitudinal Data (BOLD) PI: Jerry Jarvik, MD, MPH Details: www.backpainproject.org This is a prospective cohort study (Level 1 Evidence) This study was funded by a large grant from the AHRQ and was part of a larger project which was the BOLD project. Further details can be found at the website noted here.

acknowledgements Sean Rundell Janna Friedly Alfred Gellhorn Laura Gold Bryan Comstock Patrick Heagerty Brian Bresnahan David Nerenz Jeffrey Jarvik

Low back pain Background Lifetime incidence between 60% and 85% Accounts for a substantial portion of our annual healthcare expenditures Physical therapy is a commonly-prescribed treatment Optimal timing of physical therapy is not clear Two prior studies (Gellhorn et al and Fritz et al) found early PT was associated with a decrease or no change in subsequent healthcare utilization Low back pain remains a common and important problem and accounting for a substantial portion of our annual healthcare expenditures and physical therapy is a commonly-prescribed for low back pain There is good evidence to support use of physical therapy for chronic low back pain [7-18] But the evidence for physical therapy use for acute episodes of low back pain is mixed [7-18] The optimal timing for prescription of physical therapy is not entirely clear Similar to this study, Two other studies (by Gellhorn et al and Fritz et al) have investigated subsequent healthcare utilization after early physical therapy and found decreases or no change in subsequent healthcare utilization [17,29]

Study objective To estimate the association between initiating early physical therapy following a new visit for an episode of low back pain and subsequent back-pain-specific health care utilization in older adults after adjusting for disease severity, symptom duration and sociodemographic factors. This study was designed to estimate… Our study is unique from the two studies mentioned on the last slide in that we were able to use an number of patient reported outcomes to adjust for important possible confounding variables

Design Design: Prospective cohort study in 3 major integrated healthcare systems (BOLD registry) Patient sample: 4,723 adults, aged 65 and older, presenting to a primary care setting with a new episode of low back pain Intervention: Patients who had early PT (within 28 days of index visit) Comparator: Patients who didn’t have early PT Sub-analysis: Patients with a self-reported symptom duration of less than one month This was a Prospective cohort study using data gathered through the BOLD registry at 3 major integrated healthcare systems (Kaiser-Permanente in Northern California, Henry Ford in Detroit, and Harvard Vanguard Medical Associates in Boston) Our patient sample was 4,723 adults, aged 65 and older, presenting to a primary care setting with a new episode of low back pain, which was defined as no prior visits to a healthcare provider for low back pain in the previous 6 months. We compared patients who had early physical therapy (defined as being initiated within 28 days of their index visit) with those who did not And to look more specifically at early PT use for acute low back pain we performed a sub-analysis of our participants with a self-reported symptom duration of less than one month

Outcomes Primary outcome: Total back-pain-specific relative value units (RVUs), from days 29-365 Secondary outcomes: Overall RVUs for all healthcare use and use of specific health care services including: Imaging (x-ray and MRI or CT) Emergency Department visits Physician visits Physical therapy Spinal injections Spinal surgeries Opioid use Our primary outcome was total back-pain-specific relative value units (RVUs), from days 29-365. Our secondary outcomes were overall RVUs for all healthcare use and use of specific health care services including: Imaging, Emergency Department visits, physician visits, physical therapy treatments, spinal injections, spinal surgeries and opioid use.

Analysis Adjustment performed with generalized linear models All models adjusted for age, sex, race, ethnicity, education, marital status, smoking status, duration of back pain, back- related disability, back pain NRS, leg pain NRS, expectations for recovery, anxiety symptoms, depression symptoms, EQ-5D, study site, and total RVUs for the 365 days prior to the index visit. To adjust for potential confounding variables we used generalized linear models and in addition to baseline sociodemographics, patient reported outcomes were used in adjusting for clinical characteristics such as symptom duration, severity and quality.

Results We started with 5,239 participants in the BOLD registry and after exclusion criteria our initial unadjusted groups included 4,095 participants who did not undergo early PT and 628 participants who did undergo early PT - The two groups were similar at baseline with a few important exceptions. The early PT group had more participants with 0-3 months of symptom duration (66.3% vs 51%) and fewer participants with greater than 1 year of symptom duration (18.8% vs 36.7%) The early PT group also had more participants with leg pain present (73.3 vs 62.3%) So overall the group that did not have early PT had pain that was a little more “chronic” and had less leg pain

Unadjusted RVUs for days 29-365 Results Unadjusted RVUs for days 29-365   Early PT (n=628) No Early PT (n=4095) Mean SD Median Interquartile Range Ratio of means 95% CI p-value Total RVUs 64.7 141.3 25.8 12.6 53.0 50.8 125.3 20.6 8.4 46.4 1.28 1.06 1.54 0.01 Total Spine RVUs 17.0 93.2 5.9 9.5 72.0 0.5 1.79 1.10 2.91 0.02 Total Physical Therapy RVUs 3.7 5.6 1.8 0.0 5.3 1.3 3.6 2.86 2.47 3.31 <0.001 Total Injections RVUs 1.0 4.7 0.6 3.8 1.20 2.68 Total Spine Image RVUs 2.9 7.0 1.1 5.4 1.64 1.33 2.03 Total Surgical RVUs 10.1 81.1 5.1 59.0 1.99 0.97 4.11 0.06 Our unadjusted results did show that the early PT group had greater unadjusted RVUs, this included total spine RVUs circled here as well as total RVU’s and all other secondary RVU outcomes, except for surgical RVUs Unadjusted analysis of total units of healthcare use and percentage of participants using specific healthcare services over the same time period showed a similar pattern, with the early PT group using more of all healthcare services except for Emergency Department visits and spine surgeries (Data not shown)

Adjusted RVUs for days 29-365 Results Adjusted RVUs for days 29-365   Early PT No Early PT Mean SD Ratio of means 95% CI p-value Total RVU 56.1 5.0 52.3 2.1 1.07 0.88 1.31 0.49 Total Spine RVU 12.2 2.6 10.2 1.2 1.19 0.72 1.96 Total Physical Therapy RVU 3.4 0.2 1.3 0.1 2.56 2.17 3.03 <0.001 Total Injections RVU 0.8 0.6 1.33 0.89 2.00 0.17 Total Spine Image RVU 2.5 0.3 1.8 1.37 1.09 1.71 0.01 Total Surgical RVU 5.9 1.9 5.7 1.0 1.03 0.50 2.12 0.94 However, after adjustment there was no statistically significant difference between the two groups for our primary outcome of “total back pain specific RVUs”, as I’ve circled above the ratio of means was 1.19, with a 95% CI of 0.72 to 1.96. The early PT group did have slightly higher RVUs for all secondary RVU outcomes, but these differences were smaller than before adjustment, and only “PT RVUs” and “spine imaging RVUs” remained statistically significant.

Results Acute low back pain sub-analysis: 246 participants with early PT and 1348 without Performed for primary outcome only No statistically significant difference noted Ratio of means 2.01, 95% CI 0.98 to 4.15, P 0.06. Exploratory analysis: Total spine RVUs from day 1-365 Performed to include PT RVUs from first month Similar results to primary analysis Ratio of means 1.13, 95% CI 0.76 to 1.70, P 0.54 Our sub-analysis of participants with acute low back pain included 246 participants who had early PT and 1348 who did not. This analysis was performed for our primary outcome only and showed results similar to our primary analysis, with no statistically significant difference in total spine RVUs noted between groups (ratio of means 2.01, 95% CI of 0.98 to 4.15, P 0.06.) Of note we also performed an exploratory analysis comparing total spine RVUs from day 1-365 (in order to include the physical therapy received in the first month). Results of this were similar to the primary analysis of days 29-365, demonstrating that even with the addition of the RVUs from the physical therapy received during days 1-28, total spine RVUs at the end of one year were not meaningfully different (ratio of means 1.13, 95% CI of 0.76 to 1.70, P 0.54.)

Discussion Early PT was not associated with any meaningful change in subsequent healthcare utilization This was true even when the RVUs for the initial PT were included Early PT specifically for acute LBP also is not associated with any change in subsequent back-pain-specific healthcare utilization In this large prospective cohort study of older adults with low back pain, we found that early PT was not associated with either an increase or decrease in total back-pain-specific RVUs over the course of the following year. This was true even when the RVUs for the initial physical therapy were included. The early PT group was associated with a small but statistically significant increase in total physical therapy use as well as spine imaging. However, the lack of difference seen in total back-pain-specific RVUs implies that these differences were offset by less healthcare utilization in other areas. Based on our analysis of the subgroup of participants with acute low back pain, the use of early PT specifically for acute low back pain also does not appear to significantly increase or decrease total back-pain-specific RVUs

Discussion Rundell, et al: Functional outcomes after early PT in BOLD Modest improvement in the primary outcome (RMDQ) in patients who had early PT (mean RMDQ 1.4 points less) Early PT may lead to modestly improved outcomes without an increase in overall healthcare utilization - In a separate study, also using the BOLD data set, Rundell et al. looked at functional outcomes after early physical therapy [28]. - The primary outcome for his study was the RMDQ and he noted modest improvement in this for those patients who had early physical therapy (mean RMDQ was 1.4 points less) as compared to those who did not - Taken in combination with the results of this study, these findings suggest that prescribing early PT for older adults may lead to modestly improved outcomes without an increase in overall healthcare utilization

Conclusion In older adults presenting for a new episode of LBP, early PT is not associated with any difference in subsequent back- pain-specific healthcare utilization compared to patients not receiving early PT, even for patients with acute LBP. In conclusion, We found that, in a group of older adults presenting for a new episode of low back pain, the use of early PT is not associated with any difference in subsequent back-pain-specific healthcare utilization compared to patients not receiving early PT, even for patients with acute low back pain.