**3. Results**

#### *3.1. Patient Selection*

In 2020, a total of 637,428 Veterans received care at VA for PTSD, of whom 50.3% (*n* = 320,932) met the criteria for chronic pain. Veterans with PTSD and chronic pain were more likely to be older, female, and African American, relative to those without chronic pain (Table 1). Veterans with chronic pain also displayed higher rates of all psychiatric comorbidities examined, including depressive disorder, substance use disorder, anxiety disorder, bipolar disorder, and psychotic disorder. Veterans with chronic pain were also more likely to be prescribed more antidepressants, opioid analgesics, anticonvulsants, benzodiazepines, antipsychotics, sedative hypnotics, and antimigraine agents.


**Table 1.** Patient characteristics of Veterans with PTSD.

CNS = central nervous system; PTSD = post-traumatic stress disorder.

#### *3.2. Medication Changes*

From 1 January 2021 to 31 December 2021, the number of changes in psychiatric medications was significantly higher for those with chronic pain (mean (M) = 1.8, standard deviation (SD) = 2.0) compared to those without chronic pain (M = 1.6, SD = 1.9) (Z = 38.4, *p* < 0.001). This relationship was also observed with the 2012 sensitivity analysis, where Veterans with chronic pain (M = 2.3, SD = 2.4) had a higher mean number of changes than those without chronic pain (M = 1.9, SD = 2.2) (Z = 58.0, *p* < 0.001). Categorically, 8.9% of Veterans with chronic pain had five or more changes in their psychiatric medication prescriptions compared to 7.2% without chronic pain, and 26.1% of Veterans with chronic pain had three or more changes, compared to 22.7% without chronic pain (Figure 1).

**Figure 1.** Distribution in the number of changes in CNS medications received by Veterans with PTSD, with and without chronic pain, during 2021.

Negative binomial regression was then used to determine whether the relationship between chronic pain and the number of psychiatric medication changes persisted after adjustment for potential confounding factors. The unadjusted IRR was 1.12 (95% CI: 1.11, 1.12), indicating that chronic pain was associated with a 12% greater risk for one additional psychiatric medication change, that is, 12% more likely to have one change than zero changes, 12% more likely to have two changes, relative to one change, etc. After adjustment for important confounding factors including demographics and psychiatric comorbidity, the association between chronic pain and the number of psychiatric medication changes remained significant, and the IRR point estimate was unchanged from the unadjusted model (aIRR = 1.11; 95% CI: 1.10, 1.11; Table 2).

Although not the primary focus of the analysis, several model covariates of note were found to be significantly associated with an elevated risk for changes in psychiatric medications, including female sex, Black or African American race, Charlson Comorbidity Index, recent inpatient hospitalization, and the presence of any examined psychiatric comorbidities. Conversely, covariates associated with a decreased risk for psychiatric medication changes included older age, rural residence, and the presence of at least one psychiatric medication at baseline. The association between chronic pain and an increased risk for the number of psychiatric medication changes was also observed in the 2012 sensitivity analysis (aIIR = 1.16; 95% CI: 1.15, 1.17; Supplemental Table S2).


**Table 2.** Clinical characteristics associated with the number of changes in psychiatric medications as a discrete count.

aIRR = adjusted incidence rate ratio; CI = confidence interval.

#### *3.3. CNS Polytherapy*

The number of concurrent CNS-active medications received during the observation period of 1 January 2021 to 31 December 2021 was M = 2.7 (SD = 1.6) for Veterans with chronic pain compared to M = 2.0 (SD = 1.3) for those without chronic pain (Z = 179.7, *p* < 0.001). Categorically, 12.9% of Veterans with chronic pain and PTSD had five or more concurrent CNS medications versus 4.3% without chronic pain (Figure 2). Differences were also found at four concurrent medications (15.3% with chronic pain versus 8.6% without chronic pain) and at three concurrent medications (23.4% with chronic pain and 19.4% without chronic pain). Cumulatively, 51.6% of Veterans with chronic pain were concurrently prescribed three or more CNS-active medications compared to 32.0% of Veterans without chronic pain.

**Figure 2.** Distribution in the number of concurrent central nervous system (CNS) active medications received by Veterans with PTSD, with and without chronic pain, during 2021.

Negative binomial regression was then used to determine whether the relationship between chronic pain and the number of concurrent CNS medications persisted after adjusting for potential confounding factors. The adjusted IRR was 1.29 (95% CI: 1.28, 1.29; Supplemental Table S3), indicating that chronic pain was associated with a 29% greater risk for having one more CNS-active medication. However, when the number of concurrent CNS medications was restricted to just psychiatric medications, the relationship with chronic pain was substantially diminished (aIRR = 1.03; 95% CI: 1.02, 1.03; Supplemental Table S4).

Similar relationships were also observed with the 2012 sensitivity analysis. Veterans with chronic pain (M = 3.1, SD = 1.7) received more concurrent CNS-active medications than Veterans without chronic pain (M = 2.0, SD = 1.4) (Z = 219.1, *p* < 0.001). The association between chronic pain and the number of concurrent CNS medications was observed in the 2012 sensitivity analysis (aIIR = 1.46; 95% CI: 1.45, 1.46; Supplemental Table S5). As seen in the primary analysis, when the number of concurrent CNS medications was restricted to just psychiatric medications, the relationship with chronic pain was substantially diminished (aIRR = 1.07; 95% CI: 1.07, 1.08; Supplemental Table S6).

#### **4. Discussion**

Our findings, which demonstrate a greater number of psychiatric prescription changes when patients with PTSD also have chronic pain, are consistent with prior work, which demonstrated a magnifying effect of chronic pain on PTSD symptomology [4,5]. We found a moderate effect of a 12% higher risk for each additional medication change across a oneyear timeframe among Veterans with PTSD and chronic pain, compared to those with PTSD alone. This higher likelihood of psychiatric medication changes may indicate instability in the patient's treatment regimen, reflecting a greater symptom burden, instability in symptoms, or greater difficulty in consistently managing symptoms. This finding is also consistent with prior work, which demonstrated an increased number of healthcare visits among Veterans with chronic pain and comorbid PTSD [6]. Veterans with this comorbidity may be seeking, or requiring, a greater number of visits and medication changes in an attempt to treat the heightened symptom load resulting from chronic pain comorbid to PTSD.

Adding to prior work [17], the current findings demonstrate higher rates of CNS polytherapy among Veterans with chronic pain and PTSD, relative to those with only PTSD. The higher rates of CNS-active polytherapy, resulting from the additive effect of psychopharmacologic and analgesic agents, may result, at least in part, from the siloed treatment of these two conditions. Veterans with both conditions may be seen by two different providers, each following a separate set of guidelines [1,33]. Both providers could be prescribing CNS-active medications without knowing the treatment course for the other condition, which could lead to an increased risk for polytherapy in Veterans with chronic pain that is comorbid to PTSD.

As such, providers may benefit from guidelines for treating Veterans with both chronic pain and PTSD. A coordinated cross-specialty treatment plan may result in Veterans having a more stable medication regimen (e.g., fewer medication changes) and lower risks associated with polytherapy [20]. Patients with comorbid chronic pain and PTSD may also benefit from combined behavioral interventions that simultaneously address both PTSD and chronic pain [34]. Because women, rural-dwelling people, and minoritized persons were at greater risk for psychiatric medication changes in our analyses, these populations may stand to benefit most from further research into integrated care for comorbid chronic pain and PTSD.

This study has some limitations. This work includes only Veterans receiving care from VA, so our findings may not generalize to Veterans or other patient populations receiving PTSD care outside of the VA healthcare system. In addition, we were not able to confidently distinguish newly diagnosed PTSD or chronic pain. It is unclear whether the relationships observed in this study would differ between newly diagnosed patients and patients with pre-existing conditions. Another limitation is that certain medications are indicated for both analgesia and psychiatric management (i.e., duloxetine, gabapentin, topiramate). Categorizing these medications as primarily analgesic medications may have impacted the models comparing psychiatric to analgesic prescribing patterns, though this conservative approach was taken to avoid overestimating differences. Finally, the process of prescribing medications in the VA system has been impacted by wide-reaching prescribing initiatives across the past decade, which has significantly reduced overall opioid and benzodiazepine prescriptions [35–37]. As such, to examine whether the patterns we identified reflected ongoing clinical phenomenon, as opposed to being in response to healthcare-system-specific initiatives, we conducted sensitivity analyses across time to determine whether the current findings (2021) remained significant a decade prior (2012). The continued significance of our findings supports the consistency of these findings across time. Our focus was to explore, as a proof of concept, the potential impact of chronic pain on the prescription of psychiatric and other CNS medications. Unfortunately, there is no defined value for what constitutes clinically meaningful in the prescribing metrics examined in this study.

In conclusion, the deleterious impact that chronic pain can have on PTSD symptomatology [4–7] is reflected in the differential prescribing patterns for Veterans with comorbid chronic pain compared to Veterans with PTSD alone. We found that Veterans with comorbid chronic pain and PTSD are at an increased risk for a higher number of medication changes and for receiving CNS-active polytherapy. Providers who treat Veterans with comorbid PTSD and chronic pain may benefit from guidelines to co-manage these conditions, avenues to coordinate care with cross-specialty colleagues, and the development of integrated behavioral interventions that address both conditions.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jcm12144763/s1, Table S1: CNS active medications typically used for psychiatric indications and observed among dispensed medications. Table S2: Clinical characteristics associated with the number of changes in psychiatric medications as a discrete count for year 2012. Table S3: Clinical characteristics associated with CNS polytherapy medications as a discrete count for year 2021. Table S4: Clinical characteristics associated with psychiatric medications in the CNS polytherapy regimen, as a discrete count for year 2021.Table S5: Clinical characteristics

associated with CNS polytherapy medications as a discrete count for year 2012. Table S6: Clinical characteristics associated with psychiatric medications in the CNS polytherapy regimen, as a discrete count for year 2012.

**Author Contributions:** Conceptualization, A.A.P., K.H., E.B.K.T., S.B.N. and B.C.L.; Methodology, A.A.P., K.H., M.A.M., E.B.K.T., K.M., S.B.N. and B.C.L.; Software, K.M. and B.C.L.; Validation, K.H., M.A.M., K.M. and B.C.L.; Formal analysis, A.A.P., K.H., M.A.M., K.M. and B.C.L.; Investigation, A.A.P., K.H. and B.C.L.; Resources, A.A.P., S.B.N. and B.C.L.; Data curation, K.M. and B.C.L.; Writing original draft, A.A.P.; Writing—review & editing, K.H., M.A.M., E.B.K.T., K.M., S.B.N. and B.C.L.; Visualization, A.A.P. and B.C.L.; Supervision, K.H., M.A.M., E.B.K.T. and B.C.L.; Project administration, K.H., E.B.K.T., K.M. and B.C.L.; Funding acquisition, K.H., M.A.M., E.B.K.T., S.B.N. and B.C.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** Funding was provided by the U.S. Department of Veterans Affairs (VA) Office of Rural Health (project number 03857). Visit https://www.ruralhealth.va.gov to learn more. The work reported here was also supported by the U.S. Department of Veterans Affairs Health Services Research and Development (HSR&D) Service through the Center for Access and Delivery Research and Evaluation (CADRE) (CIN 13-412). Pratt's time to develop and write this manuscript was supported by the Department of Veterans Affairs Office of Academic Affiliations Advanced Fellowship Program in Health Services Research, the Center for Access & Delivery Research and Evaluation, Iowa City VA Health Care System. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the U.S. Department of Veterans Affairs or the United States Government.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.
