bs_bs_banner
Journal of Evaluation in Clinical Practice ISSN 1365-2753
Importance of the type of provider seen to begin health care for a new episode low back pain: associations with future utilization and costs Julie M. Fritz PhD PT FAPTA,1,2 Jaewhan Kim PhD3 and Josette Dorius BSN MPH4 1
Professor, Department of Physical Therapy, College of Health, University of Utah, Salt Lake City, UT, USA Associate Dean for Research, College of Health, University of Utah, Salt Lake City, UT, USA 3 Assistant Professor, Department of Family and Preventive Medicine, University of Utah, Salt Lake City, UT, USA 4 Clinical Operations Director, University of Utah Health Plans, Salt Lake City, UT, USA 2
Keywords care pathways, economic analysis, health care utilization, health services research, low back pain Correspondence Dr Julie M. Fritz Department of Physical Therapy College of Health University of Utah 520 Wakara Way Salt Lake City, UT 84108 USA E-mail: [email protected] Accepted for publication: 7 September 2015 doi:10.1111/jep.12464
Abstract Rationale, aims and objective Low back pain (LBP) care can involve many providers. The provider chosen for entry into care may predict future health care utilization and costs. The objective of this study was to explore associations between entry settings and future LBP-related utilization and costs. Methods A retrospective review of claims data identified new entries into health care for LBP. We examined the year after entry to identify utilization outcomes (imaging, surgeon or emergency visits, injections, surgery) and total LBP-related costs. Multivariate models with inverse probability weighting on propensity scores were used to evaluate relationships between utilization and cost outcomes with entry setting. Results 747 patients were identified (mean age = 38.2 (± 10.7) years, 61.2% female). Entry setting was primary care (n = 409, 54.8%), chiropractic (n = 207, 27.7%), physiatry (n = 83, 11.1%) and physical therapy (n = 48, 6.4%). Relative to primary care, entry in physiatry increased risk for radiographs (OR = 3.46, P = 0.001), advanced imaging (OR = 3.38, P < 0.001), injections (OR = 4.91, P < 0.001), surgery (OR = 4.76, P = 0.012) and LBP-related costs (standardized Β = 0.67, P < 0.001). Entry in chiropractic was associated with decreased risk for advanced imaging (OR = 0.21, P = 0.001) or a surgeon visit (OR = 0.13, P = 0.005) and increased episode of care duration (standardized Β = 0.51, P < 0.001). Entry in physical therapy decreased risk of radiographs (OR = 0.39, P = 0.017) and no patient entering in physical therapy had surgery. Conclusions Entry setting for LBP was associated with future health care utilization and costs. Consideration of where patients chose to enter care may be a strategy to improve outcomes and reduce costs.
Introduction Low back pain (LBP) impacts 60–80% of individuals at some in their lives [1,2]. Management imposes a large socio-economic burden on health care systems. Total direct costs in the United States were estimated at over 86 billion dollars in 2005 [3] and costs related to LBP have been increasing at a rate faster than overall health care spending [3,4]. Given the prevalence of LBP it is not surprising that it ranks as the second or third most common symptomatic conditions for which an individual seeks health care [5–7]. An estimated one of every 17 doctor visits are attributable to LBP [6] across several specialties, and LBP is the most common condition encountered in
physical therapy and chiropractic practices [8–10]. Considering the number of providers involved and the myriad management options, it is not surprising that care patterns for LBP are highly variable [11–15]. Fragmented and variable care for LBP contributes to high levels of guideline discordant management, overuse of expensive and invasive procedures, and continual cost escalations without accompanying evidence of improved outcomes [16–18]. Research increasingly points to the importance of early care decisions and guideline adherence in the prognosis of patients with LBP who seek care. For example, ordering a magnetic resonance imaging (MRI) or prescribing opioids within the first weeks is associated with increased risk for persistent symptoms, work disability and high costs [19–23]. Less attention has focused on the
Journal of Evaluation in Clinical Practice 22 (2016) 247–252 © 2015 John Wiley & Sons, Ltd.
247
Entry provider for low back pain care
earliest care decision made by a patient, the type of provider selected to begin care. This decision likely has important implications for the prognosis and costs associated with an episode of LBP care [24]. Numerous provider types including several doctor specialties, physical therapists and chiropractors may serve as the point of entry for an individual with LBP. More research is needed to explore the implications of beginning care with different providers. The purposes of this study were (1) describe the entry providers chosen by individuals with private health insurance for a new LBP consultation, (2) examine differences in patient characteristics based on entry setting and (3) examine associations between entry setting and duration of episode of care, subsequent health care costs and risk for utilization of specific procedures including radiographs, advanced imaging, injections, emergency department or spine surgeon visits, or surgery for LBP.
J.M. Fritz et al.
Patient characteristics and covariates We identified patient characteristics and co-morbid conditions from UUHP enrolment information, the University of Utah Health Care electronic medical record (EMR) and ICD-9 codes for all claims recorded in the year following the entry visit. Patient characteristics included sex, age, zip code of residence and LBP as primary or secondary diagnosis at entry visit. Specific co-morbidities recorded were mental health conditions (296.×, 297.×, 298.×, 300.×, 301.×, 308.×, 309.×, 311.×), smoking status (305.1, V15.82, 649.0×), substance use disorders (291.×, 303.×– 304.×, 305.0, 305.2×–305.9×), chronic pain (338.×) prior lumbar surgery (V45.4, 722.83) and obesity (278.×). We computed the Charlson Co-Morbidity Index (CCI) [25], which was dichotomized as low (≤1) or high (≥2).
Dependent variables
Methods Data and study sample We conducted a retrospective study of new LBP consultations between 1 January 2012 and 31 January 2013 using claims data from the University of Utah Health Plans (UUHP). The UUHP is a non-profit insurer and integrated subsidiary of University of Utah Health Care. We included enrollees with private, employer-based, coverage between the ages of 18 and 60. A new LBP consultation was defined as a provider visit occurring during the inclusion dates associated with a LBP-related ICD-9 code (720.2, 721.3, 722.1, 722.52, 722.73, 722.93, 724.×, 739.3, 739.4, 756.11, 756.12, 846.×, 847.2, 847.3, 847.9) as a primary or secondary diagnosis for whom no charges associated with LBP were received in the prior 90 days. Date of the new consultation was defined as the entry visit. We excluded those not continuously enrolled with UUHP for at least 90 days preceding and 1 year following the entry visit. We excluded patients presenting at the entry visit with an ICD-9 code indicative of a possible non-musculoskeletal cause for back pain including kidney (592.×) or gall bladder stone (574.×), urinary tract infection (599.0), or patients with a red flag condition that may require urgent management including spinal/pelvis fracture (805.×–809.×, 820.×–821.×, 733.13–733.15 or 733.96– 733.98), osteomyelitis (733.×), ankylosing spondylitis (720.0), cauda equina syndrome (344.6×) or any malignant neoplasm (140.×–209.×) diagnosis at the entry visit. Each patient was included only once in the analysis based on the first eligible entry visit using the above criteria.
Independent variables Based on the procedure code and provider associated with the entry visit, we categorized the entry setting as (1) primary care (family medicine, internal medicine, obstetrics/gynaecology), (2) physiatry, (3) chiropractic, (4) physical therapy, (5) spine surgeon (orthopedic or neurosurgeon), (6) emergency department or (7) other doctor specialty (e.g. rheumatologist, neurologist, etc.). The UUHP and state of Utah allow access to these providers without prior authorization requirement. 248
We evaluated a 1-year period following the entry visit. We recorded if any LBP-related charges occurred beyond the entry visit and the duration of the LBP episode of care as the number of days from entry to the last LBP-related charge in the 1-year follow-up. We recorded the occurrence of the following utilization outcomes; (1) radiographs of lumbo-pelvic region, (2) advanced imaging (MRI or computed tomography scan of lumbo-pelvic region, (3) office visit with a spine surgeon (orthopedic or neurosurgeon) beyond the entry visit, (4) surgical procedure (discectomy, laminectomy, fusion or rhizotomy of the lumbosacral region), (5) fluoroscopically guided epidural injection of the lumbar spine or sacroiliac joint) and (6) LBP-related emergency department visit beyond the entry visit. Costs were recorded from allowed costs for all claims associated with a LBP-related ICD-9 code during the year following the entry visit and summed to compute total LBPrelated health care costs. If health care utilization for LBP was identified in the EMR but not in claims data, we imputed the cost as the mean of available claims for the procedure.
Statistical analysis Analyses were performed with settings comprising at least 5% of the sample (primary care, physiatry, chiropractic and physical therapy). We compared patient characteristics and co-morbidities between entry settings using chi square and Kruskal–Wallis tests. Episode of care duration and total LBP-related costs were described as medians or means with 95% confidence interval (CI), respectively. Because we did not randomly assign patients to entry settings, we employed inverse probability weighting using propensity score to control selection bias [26]. Propensity scores were estimated using a multinomial logit regression because the dependent variable was entry provider category with four groups: physical therapy, primary care, physical medicine and chiropractic. In this regression, covariates at baseline (age, gender, LBP primary diagnosis, CCI, smoking status, obesity, chronic pain co-morbidity, substance use disorder, mental health co-morbidity, prior spine surgery, medications and patients’ zip codes were controlled. The inverse probability weighting of each patient was used in the logistic regressions and generalized linear regressions. Utilization outcomes were compared using adjusted odds ratios with 95% CI from multivariate logistic regression. Generalized
© 2015 John Wiley & Sons, Ltd.
J.M. Fritz et al.
Entry provider for low back pain care
linear regressions with gamma distribution and log link function were used to examine duration of care and costs because of the positively skewed distributions [27]. A significance level of P < 0.05 was used.
Results A total of 862 individuals had a new LBP consultation during the study dates, and 747 met all inclusion criteria (Figure 1). Entry visit setting was most commonly primary care (n = 409, 54.8%) followed by chiropractic (n = 207, 27.7%), physiatry (n = 83, 11.1%) and physical therapy (n = 48, 6.4%). Patient characteristics
n = 862 paents age 17-60 with a new LBP consultaon
n = 27 Pregnant at Entry Visit
n=7 Red Flag Condion: 5 – Spine fracture 2 – Neoplasm diagnosis at entry visit
n=6 Possible Non-Musculoskeletal Cause for Back Pain: 6 – urinary tract infecon
n = 75 Alternave Entry Seng: 38 – Emergency department 14 – Spine surgeon 23 – Other (rheumatology, bariatric surgeon etc.)
n = 747 Included in analysis
Figure 1 Identification of the study sample.
and co-morbidities by entry setting are outlined in Table 1. Likelihood that LBP was the primary diagnosis was lower in primary care or chiropractic compared with physical therapy or physiatry (P < 0.001). Patients entering in primary care were more likely to have co-morbid substance use disorders (P = 0.047), and those entering in primary care or physiatry were more likely to have chronic pain co-morbidity (P < 0.001). Outcomes by entry visit setting are described in Table 2. Entry setting was predictive of outcomes in the weighted multivariate models (Table 3 & Table 4). Relative to entry in primary care, entry in physiatry was associated with increased risk for radiograph (OR = 3.46, P = 0.001), advanced imaging (OR = 3.38, P < 0.001), injections (OR = 4.91, P < 0.001), surgery (OR = 4.76, P = 0.012) and total LBP-related health care costs (standardized Β = 0.67, P < 0.001). Entry in chiropractic was associated with decreased risk for advanced imaging (OR = 0.21, P = 0.001) or a surgeon visit (OR = 0.13, P = 0.005), and with increased episode of care duration (standardized Β = 0.51, P < 0.001). Entry in physical therapy was associated with decreased risk of radiographs (OR = 0.39, P = 0.017) and no patient entering in physical therapy had surgery.
Discussion The purpose of this study was to describe the choices made by privately insured patients with LBP about where to begin seeking health care. As expected, we found a wide variety of settings were used for entry, but most common were primary care and chiropractic, with smaller percentages entering with a physiatrist or physical therapist. The percentages in our sample generally conform to national averages [24]. We found the entry setting to predict future health care utilization, costs and the LBP episode of care duration after controlling for patient demographic and co-morbidity variables. Relative to beginning in primary care, entry with a chiropractor or physical therapist was associated with reduced risk for imaging, injections, surgical consultation and surgery, while entry with a physiatrist increased risk for many of these outcomes and overall LBP-related health care costs. Few studies have focused specifically on the choice of entry visit provider as a determinant of the future course of LBP care, but our findings support work that has been done suggesting the future course of care is dependent on the provider with whom a
Table 1 Characteristics and co-morbidities of study sample
Age (mean, SD) Female LBP primary diagnosis* Low co-morbidity (CCI < 2) Smoker Obesity Chronic pain co-morbidity* Substance use disorder* Mental health co-morbidity Prior spine surgery
All patients (n = 747)
Primary care (n = 409)
Chiropractic (n = 207)
Physical medicine (n = 83)
Physical therapy (n = 48)
38.2 (10.7) 61.2% 77.6% 74.8% 6.2% 12.2% 9.4% 3.3% 37.9% 0.8%
38.4 (10.9) 59.4% 75.3% 74.8% 7.8% 13.0% 12.5% 5.1% 39.9% 1.2%
37.5 (10.9) 64.3% 72.0% 72.9% 3.4% 11.1% 2.4% 1.0% 34.3% 0%
39.1 (10.1) 57.8% 95.1% 73.5% 7.2% 13.3% 16.9% 2.4% 39.8% 1.2%
37.7 (9.2) 68.8% 91.7% 85.4% 2.1% 8.3% 0% 0% 33.3% 0%
*Indicates significant differences among groups.
© 2015 John Wiley & Sons, Ltd.
249
Entry provider for low back pain care
J.M. Fritz et al.
Table 2 Unweighted outcomes over 1-year follow-up period based on entry visit setting
Received LBP care after entry visit Episode of care duration (median days, IQR) Radiographs Advanced imaging Emergency department visit Spinal injection Surgeon visit Surgery Total LBP costs (mean, 95% CI)
All patients (n = 747)
Primary care (n = 409)
Chiropractic (n = 207)
Physiatry (n = 83)
Physical therapy (n = 48)
71.9% 56 (0, 247) 32.7% 12.6% 4.4% 9.2% 4.8% 2.4% $1194 ($1043, $1345)
60.1% 20 (0, 211) 30.1% 14.2% 6.1% 8.6% 5.4% 2.2% $1167 ($953, $1382)
91.8% 146 (24, 298) 29.5% 3.4% 1.9% 3.4% 1.0% 0.5% $878 ($664, $1092)
77.1% 56 (1, 241) 62.1% 31.3% 4.6% 31.3% 10.8% 9.6% $2283 ($1665, $2900)
77.1% 44 (5, 192) 16.7% 6.2% 2.1% 2.1% 6.2% 0% $904 ($638, $1171)
IQR, interquartile range; CI, confidence interval.
Table 3 Association of health care utilization outcomes with entry visit setting
Entry visit setting
Radiographs
Advanced imaging
Primary care (reference) Chiropractic Physiatry Physical therapy
1.10 (0.57, 2.14) 3.46 (1.64, 7.29) 0.39 (0.18, 0.84)
0.21 (0.08, 0.50) 3.38 (1.78, 6.42) 0.40 (0.11, 1.46)
Injections
Emergency department visit (after entry)
Spine surgeon visit (after entry)
Surgery
1.14 (0.25, 5.25) 4.91 (2.35, 10.25) 0.27 (0.04, 2.15)
0.42 (0.10, 1.80) 0.60 (0.12, 2.90) 0.36 (0.05, 2.82)
0.13 (0.03, 0.53) 1.43 (0.63, 3.26) 1.35 (0.35, 5.31)
0.18 (0.02, 1.50) 4.76 (1.42, 16.0) No patient had surgery
Statistics represent odds ratios weighted with propensity scores with 95% CI values.
Table 4 Predictors of duration of episode of care for LBP and total LBP-related health care costs over 1-year follow-up period weighted with propensity scores Duration of episode of LBP care
Total LBP-related health care costs
Entry visit setting
Coefficient
P-value
95% CI
Coefficient
P-value
95% CI
Primary care (reference) Chiropractic Physiatry Physical therapy
0.51 0.042 −0.036
Journal of Evaluation in Clinical Practice ISSN 1365-2753
Importance of the type of provider seen to begin health care for a new episode low back pain: associations with future utilization and costs Julie M. Fritz PhD PT FAPTA,1,2 Jaewhan Kim PhD3 and Josette Dorius BSN MPH4 1
Professor, Department of Physical Therapy, College of Health, University of Utah, Salt Lake City, UT, USA Associate Dean for Research, College of Health, University of Utah, Salt Lake City, UT, USA 3 Assistant Professor, Department of Family and Preventive Medicine, University of Utah, Salt Lake City, UT, USA 4 Clinical Operations Director, University of Utah Health Plans, Salt Lake City, UT, USA 2
Keywords care pathways, economic analysis, health care utilization, health services research, low back pain Correspondence Dr Julie M. Fritz Department of Physical Therapy College of Health University of Utah 520 Wakara Way Salt Lake City, UT 84108 USA E-mail: [email protected] Accepted for publication: 7 September 2015 doi:10.1111/jep.12464
Abstract Rationale, aims and objective Low back pain (LBP) care can involve many providers. The provider chosen for entry into care may predict future health care utilization and costs. The objective of this study was to explore associations between entry settings and future LBP-related utilization and costs. Methods A retrospective review of claims data identified new entries into health care for LBP. We examined the year after entry to identify utilization outcomes (imaging, surgeon or emergency visits, injections, surgery) and total LBP-related costs. Multivariate models with inverse probability weighting on propensity scores were used to evaluate relationships between utilization and cost outcomes with entry setting. Results 747 patients were identified (mean age = 38.2 (± 10.7) years, 61.2% female). Entry setting was primary care (n = 409, 54.8%), chiropractic (n = 207, 27.7%), physiatry (n = 83, 11.1%) and physical therapy (n = 48, 6.4%). Relative to primary care, entry in physiatry increased risk for radiographs (OR = 3.46, P = 0.001), advanced imaging (OR = 3.38, P < 0.001), injections (OR = 4.91, P < 0.001), surgery (OR = 4.76, P = 0.012) and LBP-related costs (standardized Β = 0.67, P < 0.001). Entry in chiropractic was associated with decreased risk for advanced imaging (OR = 0.21, P = 0.001) or a surgeon visit (OR = 0.13, P = 0.005) and increased episode of care duration (standardized Β = 0.51, P < 0.001). Entry in physical therapy decreased risk of radiographs (OR = 0.39, P = 0.017) and no patient entering in physical therapy had surgery. Conclusions Entry setting for LBP was associated with future health care utilization and costs. Consideration of where patients chose to enter care may be a strategy to improve outcomes and reduce costs.
Introduction Low back pain (LBP) impacts 60–80% of individuals at some in their lives [1,2]. Management imposes a large socio-economic burden on health care systems. Total direct costs in the United States were estimated at over 86 billion dollars in 2005 [3] and costs related to LBP have been increasing at a rate faster than overall health care spending [3,4]. Given the prevalence of LBP it is not surprising that it ranks as the second or third most common symptomatic conditions for which an individual seeks health care [5–7]. An estimated one of every 17 doctor visits are attributable to LBP [6] across several specialties, and LBP is the most common condition encountered in
physical therapy and chiropractic practices [8–10]. Considering the number of providers involved and the myriad management options, it is not surprising that care patterns for LBP are highly variable [11–15]. Fragmented and variable care for LBP contributes to high levels of guideline discordant management, overuse of expensive and invasive procedures, and continual cost escalations without accompanying evidence of improved outcomes [16–18]. Research increasingly points to the importance of early care decisions and guideline adherence in the prognosis of patients with LBP who seek care. For example, ordering a magnetic resonance imaging (MRI) or prescribing opioids within the first weeks is associated with increased risk for persistent symptoms, work disability and high costs [19–23]. Less attention has focused on the
Journal of Evaluation in Clinical Practice 22 (2016) 247–252 © 2015 John Wiley & Sons, Ltd.
247
Entry provider for low back pain care
earliest care decision made by a patient, the type of provider selected to begin care. This decision likely has important implications for the prognosis and costs associated with an episode of LBP care [24]. Numerous provider types including several doctor specialties, physical therapists and chiropractors may serve as the point of entry for an individual with LBP. More research is needed to explore the implications of beginning care with different providers. The purposes of this study were (1) describe the entry providers chosen by individuals with private health insurance for a new LBP consultation, (2) examine differences in patient characteristics based on entry setting and (3) examine associations between entry setting and duration of episode of care, subsequent health care costs and risk for utilization of specific procedures including radiographs, advanced imaging, injections, emergency department or spine surgeon visits, or surgery for LBP.
J.M. Fritz et al.
Patient characteristics and covariates We identified patient characteristics and co-morbid conditions from UUHP enrolment information, the University of Utah Health Care electronic medical record (EMR) and ICD-9 codes for all claims recorded in the year following the entry visit. Patient characteristics included sex, age, zip code of residence and LBP as primary or secondary diagnosis at entry visit. Specific co-morbidities recorded were mental health conditions (296.×, 297.×, 298.×, 300.×, 301.×, 308.×, 309.×, 311.×), smoking status (305.1, V15.82, 649.0×), substance use disorders (291.×, 303.×– 304.×, 305.0, 305.2×–305.9×), chronic pain (338.×) prior lumbar surgery (V45.4, 722.83) and obesity (278.×). We computed the Charlson Co-Morbidity Index (CCI) [25], which was dichotomized as low (≤1) or high (≥2).
Dependent variables
Methods Data and study sample We conducted a retrospective study of new LBP consultations between 1 January 2012 and 31 January 2013 using claims data from the University of Utah Health Plans (UUHP). The UUHP is a non-profit insurer and integrated subsidiary of University of Utah Health Care. We included enrollees with private, employer-based, coverage between the ages of 18 and 60. A new LBP consultation was defined as a provider visit occurring during the inclusion dates associated with a LBP-related ICD-9 code (720.2, 721.3, 722.1, 722.52, 722.73, 722.93, 724.×, 739.3, 739.4, 756.11, 756.12, 846.×, 847.2, 847.3, 847.9) as a primary or secondary diagnosis for whom no charges associated with LBP were received in the prior 90 days. Date of the new consultation was defined as the entry visit. We excluded those not continuously enrolled with UUHP for at least 90 days preceding and 1 year following the entry visit. We excluded patients presenting at the entry visit with an ICD-9 code indicative of a possible non-musculoskeletal cause for back pain including kidney (592.×) or gall bladder stone (574.×), urinary tract infection (599.0), or patients with a red flag condition that may require urgent management including spinal/pelvis fracture (805.×–809.×, 820.×–821.×, 733.13–733.15 or 733.96– 733.98), osteomyelitis (733.×), ankylosing spondylitis (720.0), cauda equina syndrome (344.6×) or any malignant neoplasm (140.×–209.×) diagnosis at the entry visit. Each patient was included only once in the analysis based on the first eligible entry visit using the above criteria.
Independent variables Based on the procedure code and provider associated with the entry visit, we categorized the entry setting as (1) primary care (family medicine, internal medicine, obstetrics/gynaecology), (2) physiatry, (3) chiropractic, (4) physical therapy, (5) spine surgeon (orthopedic or neurosurgeon), (6) emergency department or (7) other doctor specialty (e.g. rheumatologist, neurologist, etc.). The UUHP and state of Utah allow access to these providers without prior authorization requirement. 248
We evaluated a 1-year period following the entry visit. We recorded if any LBP-related charges occurred beyond the entry visit and the duration of the LBP episode of care as the number of days from entry to the last LBP-related charge in the 1-year follow-up. We recorded the occurrence of the following utilization outcomes; (1) radiographs of lumbo-pelvic region, (2) advanced imaging (MRI or computed tomography scan of lumbo-pelvic region, (3) office visit with a spine surgeon (orthopedic or neurosurgeon) beyond the entry visit, (4) surgical procedure (discectomy, laminectomy, fusion or rhizotomy of the lumbosacral region), (5) fluoroscopically guided epidural injection of the lumbar spine or sacroiliac joint) and (6) LBP-related emergency department visit beyond the entry visit. Costs were recorded from allowed costs for all claims associated with a LBP-related ICD-9 code during the year following the entry visit and summed to compute total LBPrelated health care costs. If health care utilization for LBP was identified in the EMR but not in claims data, we imputed the cost as the mean of available claims for the procedure.
Statistical analysis Analyses were performed with settings comprising at least 5% of the sample (primary care, physiatry, chiropractic and physical therapy). We compared patient characteristics and co-morbidities between entry settings using chi square and Kruskal–Wallis tests. Episode of care duration and total LBP-related costs were described as medians or means with 95% confidence interval (CI), respectively. Because we did not randomly assign patients to entry settings, we employed inverse probability weighting using propensity score to control selection bias [26]. Propensity scores were estimated using a multinomial logit regression because the dependent variable was entry provider category with four groups: physical therapy, primary care, physical medicine and chiropractic. In this regression, covariates at baseline (age, gender, LBP primary diagnosis, CCI, smoking status, obesity, chronic pain co-morbidity, substance use disorder, mental health co-morbidity, prior spine surgery, medications and patients’ zip codes were controlled. The inverse probability weighting of each patient was used in the logistic regressions and generalized linear regressions. Utilization outcomes were compared using adjusted odds ratios with 95% CI from multivariate logistic regression. Generalized
© 2015 John Wiley & Sons, Ltd.
J.M. Fritz et al.
Entry provider for low back pain care
linear regressions with gamma distribution and log link function were used to examine duration of care and costs because of the positively skewed distributions [27]. A significance level of P < 0.05 was used.
Results A total of 862 individuals had a new LBP consultation during the study dates, and 747 met all inclusion criteria (Figure 1). Entry visit setting was most commonly primary care (n = 409, 54.8%) followed by chiropractic (n = 207, 27.7%), physiatry (n = 83, 11.1%) and physical therapy (n = 48, 6.4%). Patient characteristics
n = 862 paents age 17-60 with a new LBP consultaon
n = 27 Pregnant at Entry Visit
n=7 Red Flag Condion: 5 – Spine fracture 2 – Neoplasm diagnosis at entry visit
n=6 Possible Non-Musculoskeletal Cause for Back Pain: 6 – urinary tract infecon
n = 75 Alternave Entry Seng: 38 – Emergency department 14 – Spine surgeon 23 – Other (rheumatology, bariatric surgeon etc.)
n = 747 Included in analysis
Figure 1 Identification of the study sample.
and co-morbidities by entry setting are outlined in Table 1. Likelihood that LBP was the primary diagnosis was lower in primary care or chiropractic compared with physical therapy or physiatry (P < 0.001). Patients entering in primary care were more likely to have co-morbid substance use disorders (P = 0.047), and those entering in primary care or physiatry were more likely to have chronic pain co-morbidity (P < 0.001). Outcomes by entry visit setting are described in Table 2. Entry setting was predictive of outcomes in the weighted multivariate models (Table 3 & Table 4). Relative to entry in primary care, entry in physiatry was associated with increased risk for radiograph (OR = 3.46, P = 0.001), advanced imaging (OR = 3.38, P < 0.001), injections (OR = 4.91, P < 0.001), surgery (OR = 4.76, P = 0.012) and total LBP-related health care costs (standardized Β = 0.67, P < 0.001). Entry in chiropractic was associated with decreased risk for advanced imaging (OR = 0.21, P = 0.001) or a surgeon visit (OR = 0.13, P = 0.005), and with increased episode of care duration (standardized Β = 0.51, P < 0.001). Entry in physical therapy was associated with decreased risk of radiographs (OR = 0.39, P = 0.017) and no patient entering in physical therapy had surgery.
Discussion The purpose of this study was to describe the choices made by privately insured patients with LBP about where to begin seeking health care. As expected, we found a wide variety of settings were used for entry, but most common were primary care and chiropractic, with smaller percentages entering with a physiatrist or physical therapist. The percentages in our sample generally conform to national averages [24]. We found the entry setting to predict future health care utilization, costs and the LBP episode of care duration after controlling for patient demographic and co-morbidity variables. Relative to beginning in primary care, entry with a chiropractor or physical therapist was associated with reduced risk for imaging, injections, surgical consultation and surgery, while entry with a physiatrist increased risk for many of these outcomes and overall LBP-related health care costs. Few studies have focused specifically on the choice of entry visit provider as a determinant of the future course of LBP care, but our findings support work that has been done suggesting the future course of care is dependent on the provider with whom a
Table 1 Characteristics and co-morbidities of study sample
Age (mean, SD) Female LBP primary diagnosis* Low co-morbidity (CCI < 2) Smoker Obesity Chronic pain co-morbidity* Substance use disorder* Mental health co-morbidity Prior spine surgery
All patients (n = 747)
Primary care (n = 409)
Chiropractic (n = 207)
Physical medicine (n = 83)
Physical therapy (n = 48)
38.2 (10.7) 61.2% 77.6% 74.8% 6.2% 12.2% 9.4% 3.3% 37.9% 0.8%
38.4 (10.9) 59.4% 75.3% 74.8% 7.8% 13.0% 12.5% 5.1% 39.9% 1.2%
37.5 (10.9) 64.3% 72.0% 72.9% 3.4% 11.1% 2.4% 1.0% 34.3% 0%
39.1 (10.1) 57.8% 95.1% 73.5% 7.2% 13.3% 16.9% 2.4% 39.8% 1.2%
37.7 (9.2) 68.8% 91.7% 85.4% 2.1% 8.3% 0% 0% 33.3% 0%
*Indicates significant differences among groups.
© 2015 John Wiley & Sons, Ltd.
249
Entry provider for low back pain care
J.M. Fritz et al.
Table 2 Unweighted outcomes over 1-year follow-up period based on entry visit setting
Received LBP care after entry visit Episode of care duration (median days, IQR) Radiographs Advanced imaging Emergency department visit Spinal injection Surgeon visit Surgery Total LBP costs (mean, 95% CI)
All patients (n = 747)
Primary care (n = 409)
Chiropractic (n = 207)
Physiatry (n = 83)
Physical therapy (n = 48)
71.9% 56 (0, 247) 32.7% 12.6% 4.4% 9.2% 4.8% 2.4% $1194 ($1043, $1345)
60.1% 20 (0, 211) 30.1% 14.2% 6.1% 8.6% 5.4% 2.2% $1167 ($953, $1382)
91.8% 146 (24, 298) 29.5% 3.4% 1.9% 3.4% 1.0% 0.5% $878 ($664, $1092)
77.1% 56 (1, 241) 62.1% 31.3% 4.6% 31.3% 10.8% 9.6% $2283 ($1665, $2900)
77.1% 44 (5, 192) 16.7% 6.2% 2.1% 2.1% 6.2% 0% $904 ($638, $1171)
IQR, interquartile range; CI, confidence interval.
Table 3 Association of health care utilization outcomes with entry visit setting
Entry visit setting
Radiographs
Advanced imaging
Primary care (reference) Chiropractic Physiatry Physical therapy
1.10 (0.57, 2.14) 3.46 (1.64, 7.29) 0.39 (0.18, 0.84)
0.21 (0.08, 0.50) 3.38 (1.78, 6.42) 0.40 (0.11, 1.46)
Injections
Emergency department visit (after entry)
Spine surgeon visit (after entry)
Surgery
1.14 (0.25, 5.25) 4.91 (2.35, 10.25) 0.27 (0.04, 2.15)
0.42 (0.10, 1.80) 0.60 (0.12, 2.90) 0.36 (0.05, 2.82)
0.13 (0.03, 0.53) 1.43 (0.63, 3.26) 1.35 (0.35, 5.31)
0.18 (0.02, 1.50) 4.76 (1.42, 16.0) No patient had surgery
Statistics represent odds ratios weighted with propensity scores with 95% CI values.
Table 4 Predictors of duration of episode of care for LBP and total LBP-related health care costs over 1-year follow-up period weighted with propensity scores Duration of episode of LBP care
Total LBP-related health care costs
Entry visit setting
Coefficient
P-value
95% CI
Coefficient
P-value
95% CI
Primary care (reference) Chiropractic Physiatry Physical therapy
0.51 0.042 −0.036
