Depression and Anxiety are Associated with Increased Healthcare Costs and Opioid Use for Patients Undergoing Hip Arthroscopy: Analysis of a Claims Database

Abstract

Purpose:

To determine if opioid use and healthcare costs in the year prior to and following hip arthroscopy for femoroacetabular impingement (FAI) differ between those with or without depression or anxiety.

Methods:

Using the Truven Health Marketscan database, FAI patients who underwent hip arthroscopy between October, 2010 and December, 2015 were identified (CPT codes 29914 (femoroplasty), 29915 (acetabuloplasty), and/or 29916 (labral repair). Patients were excluded if they had incomplete coverage for one year either prior to or following surgery. The number of patients with one or more claims related to depression or anxiety during the year prior to surgery was quantified (ICD-9 codes 296, 298, 300, 309, 311). Healthcare costs in the year prior to and following hip arthroscopy were compared between those with (D/A) or without depression or anxiety (No). We also compared the number of patients in each group who filled a narcotic pain prescription within 180 days prior to surgery as well as >60 or >90 days after hip arthroscopy.

Results:

Depression or anxiety claims were seen in 5,208/14,830 patients (35.1%) prior to surgery. A significantly greater proportion of those with preoperative depression or anxiety filled opioid-related prescriptions both in the six months prior to surgery (36.2% vs. 25.6%, p<0.0001) and both >60 days (31.3% vs. 24.7%, P<0.0001) and >90 days after surgery (29.5% vs. 23.4%, p<0.0001). The group with preoperative depression or anxiety had significantly greater healthcare costs both prior to ( $8775 vs. $5674, p<0.0001) and following surgery ($5287 vs. $3908, p<0.0001).

Conclusion:

Both prior to and following hip arthroscopy, opioid use and healthcare costs were significantly greater for FAI patients with comorbid depression or anxiety.

Level of Evidence:

Level III Retrospective Comparative Therapeutic Study

Introduction

For patients with femoroacetabular impingement (FAI), preoperative pain and symptoms are more closely correlated with the patient’s mental state (e.g., depression and anxiety) than with the size of labral tear or magnitude of bony deformity.[1, 2] This correlation persists after surgery, such that hip arthroscopy patients with comorbid mental health disorders report inferior postoperative outcomes.[3, 4] Not only do mental health disorders negatively impact postoperative pain and function, but depression and anxiety have also been demonstrated to increase the risk of chronic postoperative opioid use in other surgical patient populations.[5-9]

Although a connection between pain and mental health disorders has been established in hip arthroscopy patients, the prevalence of mental health disorders varies greatly between centers, with the prevalence of depression, anxiety, and/or low mental health scores ranging from 13% to 40%.[1, 3, 4] In addition, the role of comorbid mental health disorders and either pre- or postoperative opioid use has not been assessed in this population. As such, the purpose of this study was to determine if opioid use and healthcare costs in the year prior to and following hip arthroscopy differ between FAI patients with or without depression or anxiety. We hypothesized that patients with comorbid mental health disorders would have increased healthcare costs and opioid use both prior to and following surgery.

Methods

Using the Truven Health Marketscan^© database and CPT codes 29914 (femoroplasty), 29915 (acetabuloplasty), and/or 29916 (labral repair), we identified a total of 14,831 patients who underwent hip arthroscopy between October, 2010 and December, 2015. The Truven database contains Marketscan Commercial and Medicare Supplemental claims and encounters and includes more than 135 million unique individuals which represents 20.3% of all non-CMS patients nationwide (CMS = Center for Medicare and Medicaid Services; Data source: Truven Health Copyright© 2012, 2017 Truven Health Analytics Inc.). Patients were excluded if they had incomplete coverage for one year prior to or following the index hip arthroscopy procedure, or a previous hip arthroscopy in the year prior the index surgical procedure. We used +/− 1 year as we mirrored our methods after a previous article of the costs prior to knee arthroplasty.[10]

We then quantified the number of patients with one or more claims related depression or anxiety (ICD-9 codes 296, 298, 300, 309, 311) and collected the total health care costs for each patient in the year prior to and following hip arthroscopy. Because of the previously published associations with depression and anxiety with postoperative outcomes after hip arthroscopy,[3, 11, 12] we targeted these two specific diagnoses and did not include codes for other psychological disorders. Total healthcare costs were defined as the sum of net costs plus any deductibles, co-pays, or co-insurance in the year prior to and following hip arthroscopy in order to get a complete representation of all healthcare costs and utilization. In addition, opioid prescriptions were identified using National Drug Code (NDC) numbers associated with all opioid agonists, opioid partial agonists such as tramadol, and opioid combination drug classes.[13]

Statistical Analyses

Total costs in the year prior to and following hip arthroscopy were log transformed and then compared between those with or without comorbid depression or anxiety using two-tailed two-sample t-tests (Figures 1 and 2). The prevalence of patients with claims related to opioid < 180 days prior to and >90 days following the index hip arthroscopy procedure were compared between those with or without depression or anxiety using chi-square tests. In addition, we performed an analysis of covariance (ANCOVA) to determine if log-transformed postoperative total costs differed between groups after adjusting for preoperative costs, sex, age, and preoperative or postoperative opioid use. Cost data was log transformed as it was not normally distributed. Statistical analyses were performed by experienced biostatisticians (KT, GH), and an alpha-level of 0.05 was considered statistically significant. Throughout the study, the orthopaedic and biostatistics teams met to assess data quality, preliminary findings, and interpretation of the results. These meetings included spot checking of randomly selected lines of data as well as inspection of cost outliers to ensure data quality.

Figure 1.

Proportions of raw postoperative procedural costs.

Figure 2.

Histogram showing the spread of the log transformed postoperative procedural costs. Log transformation was performed as raw postoperative costs were not normally distributed.

Results

Our query of the Truven database identified 14,831 patients that met the inclusion/exclusion criteria. The data quality assessment identified one patient with negative healthcare costs in the year following surgery and this patient was excluded from the analyses resulting in a sample of 14,830 patients (9226 females, 5604 males; mean age = 33.8 years). Depression or anxiety claims were seen in 5,208/14,830 patients (35.1%) prior to surgery and globally in 7,045 (47.5%) patients in either the year preceding or following the index hip arthroscopy event. Those with preoperative claims related to depression or anxiety were significantly more often female than those without comorbid mood disorders.(Table 1) Those with depression or anxiety were also significantly older; however, the mean age difference between groups was 2 years and not likely clinically meaningful finding. A significantly greater proportion of those with preoperative depression or anxiety filled opioid-related prescriptions both prior to and following surgery.(Table 1) Increased postoperative healthcare costs were noted for the group with preoperative depression or anxiety (Table 1) even after adjusting for preoperative procedural costs, sex, age, and preoperative or postoperative opioid use (p= 0.005).

Table 1.

Comparison of demographic factors, opioid use, and total health care costs between FAI patients with or without comorbid preoperative depression or anxiety

	Preoperative Depression or Anxiety?
	No	Yes	p
n	9622 (64.9%)	5208 (35.1%)	-
Age (mean ± SD)	33.1 ± 14.7	35.2 ± 13.1	< 0.0001
Sex (% Female)	57.3%	71.2%	< 0.0001
Opioid Use
< 180 days preoperatively	25.6%	36.2%	< 0.0001
> 90 days postoperatively	23.4%	29.5%	< 0.0001
Total Healthcare Costs (median)
1 year prior to surgery	$5674	$8775	< 0.0001
1 year postoperative	$3908	$5287	< 0.0001

Discussion

The most important finding of the current study was that healthcare costs and opioid use both prior to and following hip arthroscopy were significantly greater for FAI patients with comorbid mental health disorders.

Mental health has been associated with increased pain in FAI and hip arthroscopy patients and other orthopaedic patient populations. Among patients with FAI, rotator cuff tears and osteoarthritis, preoperative pain is more closely correlated with depression and anxiety than with the severity of the tear or degeneration.[1, 2, 14, 15] Mental health also impacts the response to surgery, such that higher depression and anxiety are associated with higher pain and lower function following hip arthroscopy,[3, 4] rotator cuff repair,[16, 17] and total joint arthroplasty.[18, 19]

There may be multiple mechanisms through which mental health disorders influence pain. In the absence of a new injury following surgery, the pathophysiology of persistent postoperative pain includes both a persistent inflammatory response and a neuropathic component.[20] Depression and anxiety are both associated with systemic increases in pro-inflammatory cytokines such as interleukin-6 (IL-6), with increased IL-6 often used as a biomarker of pain.[21] Depression and anxiety may also be the result of neuroplastic changes in response to chronic pain. The disproportionate pain response for FAI patients with smaller labral tears and/or bony deformity may be amplified by central sensitization.[22, 23] Central sensitization, which involves multi-level neuroplastic changes in response to chronic pain, can thereby amplify pain despite the presence of less severe stimuli.[24, 25] Central sensitization has been associated with activation of neural circuits that may also be involved with descending facilitation of pain, which may then result in an increased likelihood of persistent pain and an increased risk of worsening depressive symptoms.[26-28] In addition to inflammatory and neuroplastic mechanisms associated with increased pain, patients with depression and anxiety often demonstrate ineffective pain coping strategies such as pain catastrophizing, whereby they magnify their pain, ruminate on it and feel hopeless or helpless in managing it. Pain catastrophizing is associated with an exaggerated response to actual or anticipated pain,[29] and has been noted to be common amongst FAI patients.[12]

Mental health disorders are associated with increased pre- and postoperative pain but also with increased opioid use.[5-9] More than 35% of hip arthroscopy patients with depression or anxiety in the current study were taking opioids in the six months leading up to surgery. This is consistent with the results recently reported by Westermann et al. that only 55% of hip arthroscopy patients were opioid naïve at the time of surgery.[30] Roughly one quarter of patients without comorbid depression or anxiety continued to use opioid medications more than 90 days following hip arthroscopy which concurs with the findings of Granadillo et al. who also assessed pre- and postoperative opioid use following hip arthroscopy within a large claims database (PearlDiver).[31] In our current study, the prevalence of opioid use more than 90 days after hip arthroscopy was significantly greater for those with comorbid depression or anxiety, and Granadillo et al. also reported that depression was an independent risk factor for prolonged postoperative opioid use. [31]

While depression and anxiety are associated with increased pain, increased pain severity alone does not explain the high rate of postoperative opioid use in this patient group. First, preoperative depression is a better predictor of prolonged postoperative opioid use than severity of postoperative pain.[9] For example, in a broad sample of patients undergoing mastectomy, lumpectomy, thoracotomy, or total joint replacement, each 10-point increase on the preoperative Beck Depression Inventory II was associated with a 42% reduction in the rate of opioid cessation after surgery (P = 0.002). Also, the psychologic aspects of depression may increase the likelihood of opioid dependence. Depression significantly correlates with impulsive behavior, and both depression and impulsive behaviors are more common in those with opioid use than healthy volunteers.[32]

To both improve the likelihood of a successful clinical outcome as well as to lessen the risk of chronic opioid use and healthcare utilization and costs, treatment strategies must be developed to specifically target FAI patients with comorbid mental health diagnoses. Surgeons and healthcare providers must also be cognizant of the increased risk of prolonged opioid use in those with depression or anxiety and counsel patients prior to surgery of this risk. In addition, we must work with our colleagues in pain management, psychology, and behavioral health to lessen the opioid burden for this at-risk subset. By targeting depression and anxiety prior to this elective surgery through tailored treatments, we have the opportunity to dramatically improve postoperative outcomes for patients while also potentially decreasing health care costs.

It was also surprising to see that 12.4% of patients were diagnosed with depression or anxiety in the year following surgery. It is possible that patients with pre-existing depression or anxiety but did not have any claims related to these diagnoses for more than one year before surgery. However, it could also be that some patients were diagnosed after surgery perhaps in response to persistent postoperative pain. Unfortunately, we cannot determine the etiology of the diagnosis for each individual patient, and additional research in this area is warranted. While the etiology of postoperative depression or anxiety remains unknown, clinicians should be cognizant of postoperative depressive symptoms and refer for additional evaluation and care as needed.

This study was not without limitation. First, the analyses were performed with claims data and, as is true with all large database studies, there is an opportunity for coding errors. We attempted to offset this risk through routine meetings between the orthopaedic and biostatistics members of our team to thoroughly assess the quality of the data. Second, we did not specifically include a comparison of reoperations between patients with and without comorbid mental health disorders. Because of the association with increased postoperative pain and inferior outcomes, one may question whether reoperations due to persistent pain may be greater in this group. However, the CPT and ICD-9 codes in the claims database do not allow for laterality of a procedure to be determined (i.e. left vs. right). As such, we would not be able to determine if a secondary hip arthroscopy in the year following the index event was a revision of the initial procedure versus primary hip arthroscopy of the contralateral limb. Third, we opted to use NDC numbers associated with opioid partial agonists consistent with recent investigations at our institution which includes tramadol and other medications of similar class. Fourth, because of the previously published associations with depression and anxiety with postoperative outcomes,[3, 11, 12] we targeted these two specific diagnoses and did not include codes for other psychological disorders. Fifth, we did not have access to data for patients having surgery after 2016. More current data may potentially yield different results. Finally, other factors such as obesity and use of tobacco, alcohol, or illicit drugs could influence either preoperative opioid use and/or persistent postoperative use.[33, 34] We did not assess the roles of these factors on perioperative opioid use in the current study, and there continues to be a need to better understand the collinearity between these factors, mental health disorders, and persistent postoperative opioid use.

Conclusions

Both prior to and following hip arthroscopy, opioid use and healthcare costs were significantly greater for FAI patients with comorbid depression or anxiety.