Yi-Wen Shih

Date of Award

Summer 2023

Document Type

Open Access Dissertation


Health Services and Policy Management

First Advisor

Peiyin Hung


Chronic noncancer pain (CNCP) is defined as persistent pain lasting for more than three months due to conditions other than malignancy. This condition affects one in five American adults annually. The prescribing of opioids to individuals experiencing CNCP has been recognized as a critical contributor to the ongoing opioid crisis. In 2016, the Centers for Disease Control and Prevention (CDC) recommended Complementary and Integrative Health (CIH) as a nonpharmacologic treatment to be used among adults with CNCP. Despite the potential benefits of CIH for pain relief, there may be a discrepancy between individuals with CNCP and their insurance coverage for CIH, with CNCP being more prevalent among Medicaid beneficiaries, while private insurance is more likely to provide coverage for CIH. Additionally, individuals’ pain management strategies may vary based on their pain severities; those with severe pain conditions may have greater opioid use. To better target those with CNCP and opioid use, the National Pain Strategy introduced the concept of high-impact chronic pain, which focuses on the extent to which chronic pain interferes with individuals' daily lives. However, the relationships between the use of CIH among CNCP populations, by pain impact level, remain unclear. Therefore, this study investigated (1) pain-related CIH use by insurance type, (2) the associations between opioid and CIH use overall, and (3) the differential associations by pain impact level among adults with CNCP during the period of 2010-2017.

This retrospective cross-sectional study utilized data from Medical Expenditure Panel Survey Panels 15-21 (2010-2017), which collected health services utilization and expenditures among national representative noninstitutionalized US adults. The exposure variable in Aim 1 was individual’s insurance coverage. The main outcome in Aim 1 and main exposure in Aims 2 and 3 was pain-related CIH use (hereafter, CIH use) – a dichotomous variable, defined as at least on CIH visit to any of the CIH providers (i.e., chiropractors, acupuncturists, massage therapists, homeopathic/naturopathic/herbalists, and other alternative/complementary care providers) during the study period. In addition, individual’s opioid use was measured by two approaches: 1) ever using opioids and 2) number of opioid prescriptions as the main outcomes in Aims 2 and 3. Rao-Scott Chi-Square tests were used to assess bivariate relationships between categorical variables; survey-weighted logistic regressions were used in Aim 1 and opioid use measured as ever used opioids in Aims 2 and 3. The survey-weighted zero-inflated negative binomial regressions were used to investigate relationships between number of opioid prescriptions and pain-related CIH use in Aims 2 and 3. To detect whether different version of ICD codes impact the results, sensitivity analyses of comparing results from Panels 15-19 and from Panels 15-21 were conducted. This study further excluded those with CIH use at the beginning to explore associations between changes in individuals’ opioid use and their changes in CIH use by round. All models were adjusted for individual-level sociodemographic and health care utilization factors guided by Andersen’s behavioral health model, accounting for a MEPS complex survey design.

During 2010-2017, adults with CNCP varied between 30 million in Panel 15 (2010-2011) and 21 million in Panel 21 (2016-2017), with a peak at 34 million in Panel 19 (2014-2015). In 2010-2011, 60.2% of this population had private insurance (18 million), followed by those with only public insurance (30.3%, 9 million), and uninsured (9.5%, 2 million). Around 21.7% of adults with CNCP used CIH for their pain with increasing use from 19.4% in 2010-2011 to 25.3% in 2016-2017. In addition, CIH users tended to be non-Hispanic White, adults aged 35-44, held graduate or professional degrees, had a high-income level, and had any private insurance, and experienced lower or moderate pain impact levels (hereafter, pain level). Across all years, compared to adults with any private insurance, adults with only public insurance had 33% lower odds of CIH use (aOR: 0.67, 95% CI [0.54, 0.82], p

Between 2010 and 2017, adults with CNCP had increased use of CIH for their pain conditions, with one in four adults of CNCP had CIH visits in 2017. Individuals with only public insurance were less likely to use CIH than private insured. In addition, CIH use was associated with a reduced number of opioid prescriptions among adults with CNCP, especially among those with low pain levels. This study’s findings suggest that increasing public insurance coverage of CIH use among adults with CNCP may close the uptake gaps of CIH for pain management; early CIH interventions among CNCP adults with opioid use at different pain levels may ameliorate their opioid use and pain.

Available for download on Sunday, August 31, 2025