Skip to main content
Download PDF

Real-world impact of acupuncture on analgesics and healthcare resource utilization in breast cancer survivors with pain

BMC Medicine volume 22, Article number: 394 (2024) Cite this article

Abstract

Background

This study evaluated the real-world impact of acupuncture on analgesics and healthcare resource utilization among breast cancer survivors.

Methods

From a United States (US) commercial claims database (25% random sample of IQVIA PharMetrics® Plus for Academics), we selected 18–63 years old malignant breast cancer survivors experiencing pain and ≥ 1 year removed from cancer diagnosis. Using the difference-in-difference technique, annualized changes in analgesics [prevalence, rates of short-term (< 30-day supply) and long-term (≥ 30-day supply) prescription fills] and healthcare resource utilization (healthcare costs, hospitalizations, and emergency department visits) were compared between acupuncture-treated and non-treated patients.

Results

Among 495 (3%) acupuncture-treated patients (median age: 55 years, stage 4: 12%, average 2.5 years post cancer diagnosis), most had commercial health insurance (92%) and experiencing musculoskeletal pain (98%). Twenty-seven percent were receiving antidepressants and 3% completed ≥ 2 long-term prescription fills of opioids. Prevalence of opioid usage reduced from 29 to 19% (P < 0.001) and NSAID usage reduced from 21 to 14% (P = 0.001) post-acupuncture. The relative prevalence of opioid and NSAID use decreased by 20% (P < 0.05) and 19% (P = 0.07), respectively, in the acupuncture-treated group compared to non-treated patients (n = 16,129). However, the reductions were not statistically significant after adjustment for confounding. Patients receiving acupuncture for pain (n = 264, 53%) were found with a relative decrease by 47% and 49% (both P < 0.05) in short-term opioid and NSAID fills compared to those treated for other conditions. High-utilization patients (≥ 10 acupuncture sessions, n = 178, 36%) were observed with a significant reduction in total healthcare costs (P < 0.001) unlike low-utilization patients.

Conclusions

Although adjusted results did not show that patients receiving acupuncture had better outcomes than non-treated patients, exploratory analyses revealed that patients treated specifically for pain used fewer analgesics and those with high acupuncture utilization incurred lower healthcare costs. Further studies are required to examine acupuncture effectiveness in real-world settings.

Peer Review reports

Background

Over two million patients are diagnosed with cancer annually in the United States (US), leading to a projected population of 22 million by 2030 [1]. Cancer patients experience poorer quality of life as a consequence of acute and chronic cancer-related symptoms [2, 3], among which pain is highly disabling and arguably the most feared symptom. Prevalence of pain stands at over 40% across cancer types [4, 5], which can increase to over 70% among metastatic cancer patients [4].

In comparison to nonmalignant pain, cancer pain is multifactorial with varying presentations dependent on the cancer type, stage, and cancer treatment received [6]. Aromatase inhibitors, prescribed for reducing breast cancer recurrence in > 50% of all breast cancer cases [7, 8], cause arthralgias in more than half of the patients [9, 10]. Taxane-related agents can induce neuropathic pain that is difficult to treat and debilitating to quality of life [11]. Although opioids and non-steroidal anti-inflammatory drugs (NSAIDs) remain highly relevant in pain management [12, 13], use of these medications is complicated by adverse events (AEs) such as constipation, gastric ulcers, and medication dependence. Optimizing pain control remains an uphill battle in cancer survivorship and supportive care [14,15,16]. With complex etiologies, significant AEs associated with analgesics use, and improved understanding of psychological and spiritual elements in cancer pain treatment, there has been a paradigm shift towards the integration of conventional pharmacological and complementary non-pharmacological interventions (i.e., integrative medicine) for holistic, safe, and effective pain management [12, 13].

In the recent SIO-ASCO guidelines for pain management, acupuncture received a moderate recommendation, highest among non-pharmacological therapies, for alleviating aromatase inhibitor-induced arthralgias, as well as general and musculoskeletal pain [12]. While published large-scale acupuncture trials have reported significant declines in cancer pain severity [17, 18], it remains unclear if these findings are translatable to real-world practices. The lack of a robust placebo control results in questions regarding the durability of the observed efficacy and criticisms about acupuncture’s identity as a “mega-placebo” [19, 20]. Alongside the dearth of scientific understanding regarding acupuncture among payers, acupuncture utilization is inequitable and largely restricted to patients with private insurance coverage and high disposable income for out-of-pocket expenditure [21]. Novel approaches for assessing the effectiveness and value of acupuncture are needed.

Using a large US commercial claims database, we evaluated the real-world treatment effect of acupuncture for managing cancer pain. Trial-reported benefits of acupuncture for cancer pain are most studied among breast cancer patients who thus serve as the target population for this study [17, 18, 22]. Pain outcomes were determined with analgesics utilization, and we hypothesized that acupuncture treatment had reduced the use of analgesics in comparison to non-treated survivors. The impact on direct medical costs and healthcare resource utilization were also evaluated to investigate the value of acupuncture from the payers’ perspective.

Methods

Data source

This was a retrospective cohort study utilizing a 25% random sample of IQVIA PharMetrics® Plus for Academics (JAN2006-DEC2021), a closed database of fully adjudicated patient-level health plan claims, providing a longitudinal view of inpatient and outpatient services, prescription and office/outpatient administered drugs, costs, and detailed enrollment information. Data contributors are largely commercial health plans, thus making the database representative of the commercially insured US national population for patients under 65 years of age. As data is de-identified in accordance with the US Health Insurance Portability and Accountability Act, institutional review board approval was not required for this study.

Study design

We employed the difference-in-difference (DID) technique, a two-timepoint, pre- and post-intervention analysis, to assess the causal effect of acupuncture exposure with appropriate counterfactual comparisons in a non-randomized setting [23]. In the acupuncture-treated group, the initiation of acupuncture was set as the reference date (i.e., index date) to determine the pre- and post-index periods for comparisons. The index date for non-treated controls was simulated following an observation in this study that patients seek acupuncture after an average of 236 days after a health encounter for pain. The key analysis involved comparing the annualized changes in outcomes, from the pre-index to post-index periods, between acupuncture-treated and non-treated cohorts (Fig. 1).

Fig. 1
figure 1

Illustration of study design and difference-in-difference methodology application

Full size image

Study population and eligibility criteria

Eligible patients were 18–63 years old (to exclude Medicare-switching), at least 1 year removed from malignant breast cancer diagnosis, continuously enrolled in medical and pharmacy plans for 2 years (from 365-day before to 365-day after index date), and with one or more documented musculoskeletal/general/neoplasm-related pain claims during the 365-day pre-index period. Eligible acupuncture-treated patients first received acupuncture after cancer diagnosis and were not found with acupuncture claims during the pre-index period, while non-treated patients must not have any documented acupuncture-related claims. The algorithms for identifying key health encounters (e.g., acupuncture, breast cancer, and pain) were summarized in Additional file 1: Table S1.

Acupuncture exposure

Acupuncture exposure was determined by the presence of one or more claims for acupuncture using CPT-4 codes after cancer diagnosis (Additional file 1: Table 1). One service unit represents a 15-min treatment duration, and for the purpose of the analyses, an acupuncture session is considered as 30 min long [18].

Outcomes

Analgesics of interest [opioids (non-parenteral), NSAIDs, and adjuvant analgesics such as serotonin-norepinephrine reuptake inhibitors (SNRIs), tricyclic antidepressants (TCAs), and gabapentinoids] were selected based on the NCCN clinical practice guidelines for adult cancer pain [13]. We implemented NDC and HCPCS codes to identify prescription fills (Additional file 1: Table S1). Measures of medication utilization include the prevalence of analgesics users and the rate of short-term (< 30-day supply) and long-term (≥ 30-day supply) prescription fills. The primary outcome is the annualized incremental change in the prevalence of opioid users when compared between acupuncture-treated and non-treated cohorts.

Direct medical costs were determined by the all-cause total healthcare costs paid by both patients and payers. Costs were adjusted to 2022 US dollars using the Consumer Price Index’s annual medical care component [24]. Healthcare resource utilization was defined as the rates of all-cause hospitalizations and emergency department (ED) visits.

Covariates

Sociodemographic characteristics of age, gender, and US geographic region were measured at index date. Healthcare payer (commercial, managed Medicare, Medicare advantage/cost, others) and plan [preferred provider organization (PPO), health maintenance organization (HMO), point-of-service (POS), others] types were determined with monthly enrollment characteristics throughout the 2-year study period. Covariates evaluated during the 365-day pre-index period include total pain-related expenditure, pain subtypes (musculoskeletal, general, or neoplasm-related), Charlson’s comorbidity index excluding cancer-related codes as per National Cancer Institute (NCI) recommendations [25,26,27], menopausal symptoms, as well as predictors of severe patient-reported symptoms in cancer such as antidepressant use, depression health encounter, and ≥ 2 long-term opioid prescription fills [28]. Cancer-related characteristics (metastases and exposure to anticancer medications) were identified during the period between cancer diagnosis and index date.

Statistical analysis

Primary analysis

Differences between covariates were tested between acupuncture-treated and non-treated patients using Wilcoxon rank-sum test and Pearson’s chi-squared test for continuous and categorical variables, respectively. Distribution of continuous outcomes (cost outcomes) was examined with histogram plots.

For DID analysis, the model was specified as follows: g(Y) = β0 + β1*[post-index] + β2*[acupuncture group] + β3*[post-index × acupuncture group] + ε, with β3 serving as the main coefficient of interest (i.e., DID estimate) for each outcome Y. Models for prevalence of analgesics use as outcomes were modeled with generalized estimating equations (GEE) for estimating population averaged effects, with logit (prevalence < 10%) or log (prevalence ≥ 10%) link function, Huber-White robust standard errors for controlling for heteroskedasticity, first-order autoregression correlation structure for time-varying outcomes, and binomial distribution. Similarly, with GEE and log link function, count outcomes were examined with negative binomial distribution, and cost data implemented gamma distribution. Changes in outcomes from pre- to post-index were estimated using the exponentials of β1 for non-acupuncture patients, and a linear combination of β1 and β3 for acupuncture patients. We further explored the bivariate relationship between predetermined covariates and outcomes that were found to be significant in unadjusted analyses to evaluate potential sources of significant differences not explained by acupuncture exposure.

For covariate adjustment (covariates are listed in Table 1), we implemented an inverse probability of treatment weighting (IPTW) approach as multivariable adjustment with the GEE-DID models could not converge for various outcomes [29]. The probability of treatment assignment (i.e., propensity scores), conditional on the covariates, was estimated with multivariable logistic regression [30]. We employed restricted cubic smoothing splines with five knots to model the relationship between continuous covariates and log-odds of the exposure [31]. Each case was weighted according to the inverse of the propensity score corresponding to its assigned exposure. We stabilized the weights by multiplying them with the marginal probability of the exposure assignment to correct for very large or small weights that could destabilize the estimated effect [31, 32]. Successful covariate balance between the two groups in the weighted cohort was defined as having standardized mean differences achieving <|0.1| for all covariates [33]. Covariate(s) that could not achieve the ideal balance standard were added to the GEE-DID model after IPTW-weighting.

Table 1 Baseline characteristics of study cohort
Full size table

All analyses were two-tailed, tested at 5% significance level, and presented as ratios or percentage changes accompanying 95% confidence intervals (CI).

Sensitivity analysis

To emulate a randomized controlled trial design, sensitivity analysis comparing acupuncture-treated and non-treated cohorts utilized a 1:1 nearest neighbor matching without replacement. Propensity scores were estimated via logistic regression, adjusting for all covariates listed in Table 1.

Exploratory analysis

Because acupuncture treatment is highly heterogenous in terms of treated conditions [34, 35] and real-world utilization rate [36], we conducted exploratory analyses comparing changes in outcomes between (1) low (< 10 sessions) against high (≥ 10 sessions [18]) acupuncture utilization and (2) acupuncture treatment for pain against treatment for other conditions.

Data extraction, cleaning, IPTW, and propensity score matching (PSM, with R package MatchIt [37]) were executed with R version 4.3.2 [38]. All other analyses were performed on Stata version 16.1.

Results

Cohort characteristics

Our cohort comprised 16,624 eligible breast cancer survivors (median age = 55 years; stage 4 = 7.8%) who averaged 2.5 years (standard deviation, SD = 1.9) after cancer diagnosis. Four hundred ninety-five (3.0%) survivors had received acupuncture within a year of pain-related health encounter. Compared to non-treated survivors, more acupuncture-treated survivors had metastatic cancer, prior taxane chemotherapy exposure, reported more menopausal symptoms, lived in the Western region of the US, and with commercial PPO health plans. Acupuncture-treated patients were also with higher healthcare expenditure during the pre-index period (P < 0.05, Table 1). The selection algorithm for eligible patients is illustrated in Additional file 1: Fig. S1. Cost outcomes, both pre- and post-index, were found with a right-skewed distribution (Additional file 1: Fig. S2).

Acupuncture utilization

Acupuncture-treated patients completed an average of 11.5 (median = 6) acupuncture sessions in a given year. The total annual cost of acupuncture per patient was $919.60 (median = $476.81), paying $183.22 (median = $49.12) out-of-pocket (OOP) on average. Cost per session averaged $84.80 (median = $78.56), with $22.39 (median = $10.17) being OOP expenditure (Additional file 1: Table S2).

Impact of acupuncture on analgesics utilization

Opioid utilization

After acupuncture initiation, the prevalence of opioid users reduced by 36% (95% CI = 0.53 to 0.85, P < 0.01). The DID was estimated to have a prevalence ratio (PR) of 0.80 (95% CI = 0.66 to 0.96, P < 0.05), which indicated that the relative prevalence of opioid use decreased by 20% in the acupuncture-treated group compared to non-treated patients, over the 2-year, pre-index to post-index periods (Table 2, Fig. 2A). However, we did not observe statistically significant DID estimate on opioid prescription fill rates (Table 2, Fig. 2B–C). All predetermined covariates were significantly associated with opioid use in the bivariate analyses, except for biological sex, current or previous tamoxifen exposure, menopausal symptoms, and having two or more 30-day opioid prescription fills (Additional file 1: Table S3).

Table 2 Difference-in-difference analysis comparing changes in outcomes among acupuncture against controls
Full size table
Fig. 2
figure 2

Percentage changes with 95% confidence intervals (pre-index to post-index) in opioid and NSAID utilization outcomes. The percentage change (pre-post) statistics, with 95% confidence intervals, were calculated from the pre-post change ratios from Tables 2 and 3. For example, a pre-post change ratio of 0.64 for opioid users among acupuncture-exposed patients (n = 495, Table 2), the percentage change (pre-post) statistics will be (0.64 − 1) × 100% =  − 36%. Absolute pre-index and post-index outcomes for the subgroups are presented in Table 2, eTable 6, and eTable 8. *P < 0.05 (DID estimate)

Full size image

The characteristics and covariate balance statistics of the IPTW-weighted and PSM-matched cohorts are summarized in eFig. 3 and Additional file 1: Table S4, respectively. After IPTW and PSM, the DID estimates on opioid use prevalence was not significantly different between treated and non-treated patients (Table 2, Fig. 2A–C).

NSAID utilization

Acupuncture treatment was associated with a 32% decline (95% CI = 0.54 to 0.85, P < 0.01) in the prevalence of NSAID users from baseline, with a DID PR estimate of 0.81, albeit not statistically significant (95% CI = 0.65 to 1.01, P > 0.05), compared to non-treated patients (Table 2, Fig. 2D). We did not observe any significant DID estimate on NSAID prescription fill rates or after IPTW and PSM (Table 2, Fig. 2E–F).

Adjuvant analgesics utilization

Acupuncture was not associated with any changes in TCA or SNRI utilization. However, significant increases in prevalence (odds ratio [OR] = 1.44, 95% CI = 1.10 to 1.89, P < 0.01) and long-term prescription fills (rate ratio [RR] = 1.59, 95% CI = 1.21 to 2.08, P < 0.01) of gabapentinoids were found among acupuncture-treated patients. This effect was associated with DID ratio estimate of 1.50 and 1.46, respectively, when compared to non-treated patients (P < 0.01). All predetermined covariates were significantly associated with gabapentinoid use, except biological sex and tamoxifen exposure. Age at breast cancer diagnosis was also significantly associated with rates of long-term prescription fills of gabapentinoids (Additional file 1: Table S3). DID remained statistically significant in the IPTW-weighted model but not after PSM when evaluating odds of gabapentinoid use (Table 2).

Impact of acupuncture on healthcare cost and resource utilization

Breast cancer survivors who initiated acupuncture were found with an average decline in total healthcare cost by 26% (95% CI = 0.61 to 0.89, P < 0.01) and an absolute decrease by $15,513 (95% CI =  − $24,815 to − $6213, P < 0.01). This decrease was not significantly different from non-treated patients. However, DID estimates differ between IPTW and PSM analyses (Table 2).

Exploratory analysis 1: high vs low acupuncture utilization

One hundred seventy-eight (36.0%) patients received ≥ 10 sessions and were characterized by a higher overall healthcare expenditure and increased prior taxane exposure at baseline (Additional file 1: Table S5).

The annualized changes in analgesics and healthcare resource utilization outcomes were not significantly different between low (< 10 sessions) and high (≥ 10 sessions) acupuncture utilization breast cancer survivors. Notably, a significantly larger increase in gabapentinoid utilization was observed among low-utilization patients (P < 0.01) but not among those with higher acupuncture utilization (Table 3, Additional file 1: Table S6).

Table 3 Difference-in-difference exploratory analysis by (1) high vs low acupuncture utilization and (2) acupuncture for pain vs other conditions
Full size table

High-utilization patients, despite spending more on acupuncture sessions (average total cost: $1035 vs $787, P = 0.054), were observed with a significant reduction in total healthcare costs after acupuncture treatment (P < 0.001) unlike low-utilization patients (Table 3, Additional file 1: Table S6). After further stratifying into quintiles based on number of completed acupuncture sessions, the average cost savings in quintiles 1, 2, 4, and 5 were comparable to the PSM-control cohort, except for quintile 3 which comprised patients who completed 5 to 8.5 sessions (Additional file 1: Fig. S4).

Exploratory analysis 2: acupuncture for pain vs other conditions

Two hundred sixty-four (53.3%) patients received acupuncture specifically for pain and, compared to patients treated for other conditions (n = 231), were observed with less neoplasm-related pain diagnoses and more acupuncture sessions (Additional file 1: Table S7).

Patients receiving acupuncture for pain were found with statistically significant DID RR estimates of 0.53 (95% CI = 0.29 to 0.94, P < 0.05) and 0.51 (95% CI = 0.26 to 0.99, P < 0.05), suggesting a decline in short-term opioid and NSAID prescription fills compared to patients using acupuncture for treating other conditions (Table 3, Fig. 2B and F). These DID estimates remained statistically significant after adjusting for neoplasm-related pain diagnoses (P < 0.05, results not presented). On the other hand, patients treated for non-pain conditions were observed with a significantly larger decline in prevalence of TCA users compared to those treated for pain (P < 0.05, Table 3, Additional file 1: Table S8). We also observed a statistically significant increase in gabapentinoid utilization (P < 0.01) and reduction in total healthcare cost (P < 0.001) only among patients treated for pain (Table 3, Additional file 1: Table S8). Further stratified analysis found that gabapentinoid utilization was significantly increased among low-utilization pain-treated survivors (P < 0.05) rather than high-utilization patients (P > 0.05) (Additional file 1: Table S9).

 Discussion

This is the first study that has evaluated the real-world impact of acupuncture treatment for cancer pain. The overall acupuncture utilization for pain is low (< 3%) among commercially insured breast cancer survivors, albeit comparable to previous published data [39]. Across all analyzed subgroups of patients, those treated specifically for pain were found with the largest decline in prevalence and short-term fills of opioids and NSAIDs after acupuncture treatment. High-utilization patients were found with greater healthcare cost savings compared to low-utilization patients, with the latter escalating to the use of gabapentinoids. Taken together, findings confirm that acupuncture reduces pain and that a value-based care model may be most appropriate to facilitate increased acupuncture coverage for pain management among cancer survivors.

Acupuncture analgesia has been widely recognized in multiple clinical guidelines in both cancer and non-cancer patient populations [12, 40, 41] and functionally ascribed to local and systemic neurochemical changes. The role of endogenous opiates in acupuncture analgesia was evident from the lack of analgesic response after naloxone administration and opiate receptors depletion [42, 43]. Local release of adenosines and the activation of adenosine A1 receptors, which serve as inhibitory mediators for neurons, are crucial in inhibiting the transmission of painful stimuli to the anterior cingulate cortex during acupuncture [44]. Clinical evidence for acupuncture in cancer pain is aplenty, with two large, randomized trials specifically described in the SIO-ASCO integrative oncology guidelines for pain management [12]. The PEACE study (n = 360) investigated the efficacy of electroacupuncture and auricular acupuncture compared to waitlist control reported significant pain reduction after 10 weekly acupuncture sessions [18]. Another trial recruited breast cancer patients (n = 226) experiencing aromatase inhibitor-induced joint pain and found that 12 weeks (18 sessions) of acupuncture had relieved pain compared to sham and waitlist controls [17]. Our study contributed by finding a decline in opioid and NSAID real-world medication use between the year before and after acupuncture initiation. Importantly, this improvement is not attributable to non-specific effects of acupuncture as evidenced by the lack of effectiveness observed among patients treated for non-pain conditions. In all, acupuncture is not just a placebo as it has achieved a clinically meaningful impact on patients’ health in the real-world.

Some pain-treated, low-utilization patients were observed with increased gabapentinoid usage. Considering the relative safety of acupuncture across indications [45], we hypothesize that gabapentinoid initiation is confined to non-responders of acupuncture. Relatedly, our assessment of total healthcare expenditure concluded that cost savings after acupuncture initiation were comparable to non-treated controls, other than one-fifth of treated patients who were not frequent users of acupuncture and likely non-responders. Thus, future research should examine ways to predict acupuncture response prior to treatment initiation. Applying the concept of precision medicine, the research into clinical, sociodemographic, environmental, lifestyle, and genetic predictors of treatment response will enable the prescription of acupuncture to patients who will receive maximal health benefits [46]. Adopting a value-based care model by periodically evaluating patient response to acupuncture (e.g., monitoring patient-reported outcomes and ensuring patient adherence) will ensure that only responders should continue with the treatment and prevent unnecessary healthcare expenditure [47]. Furthermore, it is known that acupuncture utilization was more frequent among individuals of female gender, White or Asian racial-ethnic background, and those with higher levels of education and income due to affordability-related barriers [21, 48, 49]. Despite the promises of acupuncture for cancer pain management, the inequitable access to acupuncture for minoritized and marginalized populations will pervade the widening care disparities [21, 50]. Other payers, including Medicaid, Medicare, and HMOs (referring to Additional file 1: Table S3, these payers were observed with higher prescription rates of analgesics), should consider implementing precision medicine and value-based care to enhance cancer survivors’ access to acupuncture treatment for pain and reduce healthcare disparities.

There are several limitations in this study. As with other retrospective studies, unmeasured confounding is unavoidable. Although acupuncture is recommended for all cancer survivors, we have focused only breast cancer survivors to minimize the degree of unmeasured confounding; future studies should investigate in another group of cancer survivors. Even so, not all sociodemographic variables are available, thus we could not adjust for factors such as race/ethnicity and education attainment that are different between acupuncture-treated and non-treated patients. Nevertheless, while these characteristics may impact acupuncture access and utilization, they are unlikely to impact the underlying efficacy [50]. Selection bias is another concern. Naturally, cancer patients seeking integrative therapies such as acupuncture have greater symptom burden and poorer symptom control than their peers [51,52,53]. While we have attempted to control for potential confounders such as antidepressants and chronic opioid use, our PSM-control cohort is observed with greater improvements in ED and hospitalization rates compared to other control and acupuncture-treated patients. Non-pharmacological modalities for pain relief are seldom captured in claims dataset, in vie that they are often out-of-pocket expenditures [39]. Patients seeking these modalities are likely to share similar characteristics to the patients within the acupuncture-treated cohort in our study, thus possibly explaining the lack of associations observed in our analyses after IPTW and PSM. Acupuncture utilization was also highly heterogenous, with only 53% receiving it for pain management. We reason that these limitations have blunted the reported association, and the true effect size may be greater.

Finally, our statistical power and study design are impacted by the low real-world utilization of reimbursed acupuncture treatment. In Medicare Part B, beginning in 2020, acupuncture is claimable for only chronic low back pain [54]. Claims for other pain-related indications can only be obtained from Medicare managed care (i.e., Medicare Part C). The consequence is that SEER-Medicare, the database of choice for cancer-specific questions among older adults, is not suitable as the reliability and completeness of Medicare Part C data remains an area of active research [55, 56]. It is prudent to apply this methodology on SEER-Medicare after data validations to address the effectiveness of acupuncture coverage in older cancer survivors. Regardless, the study’s strength is underscored by its large sample size of over 10,000 patients and a robust pre-post-intervention, DID study design that ensures exposure-outcome temporality and similar index times for comparison of outcomes between acupuncture and control patients, effectively answering an important research question regarding the real-world treatment effect of acupuncture at managing pain in breast cancer survivors.

Conclusion

Although adjusted results did not show that patients receiving acupuncture had better outcomes than non-treated patients, exploratory analyses revealed that patients treated specifically for pain used fewer analgesics and those with high acupuncture utilization incurred lower healthcare costs. Follow-up research should evaluate ways to examine acupuncture effectiveness in real-world settings with pragmatic trials, understand predictors of acupuncture response, validate our findings within SEER-Medicare, and replicate with other groups of cancer survivors (e.g., bone metastases). Clinicians and payers can consider applying value-based care to enhance cancer survivors’ access to acupuncture treatment for pain.

Availability of data and materials

The dataset was available for purchase from IQVIA. We do not own the data and hence are not permitted to share in the original form. The underlying code for this study is not publicly available but may be made available to qualified researchers on reasonable request from the corresponding authors.

Data availability

No datasets were generated or analysed during the current study.

References

  1. Miller KD, Nogueira L, Mariotto AB, et al. Cancer treatment and survivorship statistics, 2019. CA Cancer J Clin. 2019;69(5):363–85.

    Article  PubMed  Google Scholar 

  2. Hamer J, McDonald R, Zhang L, et al. Quality of life (QOL) and symptom burden (SB) in patients with breast cancer. Support Care Cancer. 2017;25(2):409–19.

    Article  PubMed  Google Scholar 

  3. Morrison EJ, Novotny PJ, Sloan JA, et al. Emotional problems, quality of life, and symptom burden in patients with lung cancer. Clin Lung Cancer. 2017;18(5):497–503.

    Article  PubMed  PubMed Central  Google Scholar 

  4. van den Beuken-van Everdingen MH, Hochstenbach LM, Joosten EA, Tjan-Heijnen VC, Janssen DJ. Update on prevalence of pain in patients with cancer: systematic review and meta-analysis. J Pain Symptom Manage. 2016;51(6):1070-1090 e1079.

    Article  PubMed  Google Scholar 

  5. Haenen V, Evenepoel M, De Baerdemaecker T, et al. Pain prevalence and characteristics in survivors of solid cancers: a systematic review and meta-analysis. Support Care Cancer. 2022;31(1):85.

    Article  PubMed  Google Scholar 

  6. Ventafridda V, Saita L, Ripamonti C, De Conno F. WHO guidelines for the use of analgesics in cancer pain. Int J Tissue React. 1985;7(1):93–6.

    CAS  PubMed  Google Scholar 

  7. Heer E, Harper A, Escandor N, Sung H, McCormack V, Fidler-Benaoudia MM. Global burden and trends in premenopausal and postmenopausal breast cancer: a population-based study. Lancet Glob Health. 2020;8(8):e1027–37.

    Article  PubMed  Google Scholar 

  8. Acheampong T, Kehm RD, Terry MB, Argov EL, Tehranifar P. Incidence trends of breast cancer molecular subtypes by age and race/ethnicity in the US from 2010 to 2016. JAMA Netw Open. 2020;3(8): e2013226.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Grigorian N, Baumrucker SJ. Aromatase inhibitor-associated musculoskeletal pain: an overview of pathophysiology and treatment modalities. SAGE Open Med. 2022;10:20503121221078720.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Hershman DL, Shao T, Kushi LH, et al. Early discontinuation and non-adherence to adjuvant hormonal therapy are associated with increased mortality in women with breast cancer. Breast Cancer Res Treat. 2011;126(2):529–37.

    Article  CAS  PubMed  Google Scholar 

  11. Hershman DL, Lacchetti C, Dworkin RH, et al. Prevention and management of chemotherapy-induced peripheral neuropathy in survivors of adult cancers: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol. 2014;32(18):1941–67.

    Article  CAS  PubMed  Google Scholar 

  12. Mao JJ, Ismaila N, Bao T, et al. Integrative medicine for pain management in oncology: society for integrative oncology-ASCO guideline. J Clin Oncol. 2022;40(34):3998–4024.

    Article  PubMed  Google Scholar 

  13. Swarm RA, Paice JA, Anghelescu DL, et al. Adult cancer pain, version 3.2019, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2019;17(8):977–1007.

    Article  CAS  PubMed  Google Scholar 

  14. Colosia A, Njue A, Bajwa Z, et al. The burden of metastatic cancer-induced bone pain: a narrative review. J Pain Res. 2022;15:3399–412.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Broder MS, Gutierrez B, Cherepanov D, Linhares Y. Burden of skeletal-related events in prostate cancer: unmet need in pain improvement. Support Care Cancer. 2015;23(1):237–47.

    Article  CAS  PubMed  Google Scholar 

  16. Hart NH, Crawford-Williams F, Crichton M, et al. Unmet supportive care needs of people with advanced cancer and their caregivers: a systematic scoping review. Crit Rev Oncol Hemat. 2022;176:176.

    Article  Google Scholar 

  17. Hershman DL, Unger JM, Greenlee H, et al. Comparison of acupuncture vs sham acupuncture or waiting list control in the treatment of aromatase inhibitor-related joint pain: a randomized clinical trial. JAMA Netw Open. 2022;5(11).

    Article  PubMed  PubMed Central  Google Scholar 

  18. Mao JJ, Liou KT, Baser RE, et al. Effectiveness of electroacupuncture or auricular acupuncture vs usual care for chronic musculoskeletal pain among cancer survivors the PEACE randomized clinical trial. Jama Oncol. 2021;7(5):720–7.

    Article  PubMed  Google Scholar 

  19. Musial F. Acupuncture for the treatment of pain - a mega-placebo? Front Neurosci-Switz. 2019;13:1110.

    Article  Google Scholar 

  20. Fei YT, Cao HJ, Xia RY, et al. Methodological challenges in design and conduct of randomised controlled trials in acupuncture. BMJ. 2022;376: e064345.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Liou KT, Ashare R, Worster B, et al. SIO-ASCO guideline on integrative medicine for cancer pain management: implications for racial and ethnic pain disparities. JNCI Cancer Spectr. 2023;7(4):pkad042.

    Article  PubMed  PubMed Central  Google Scholar 

  22. He Y, Guo X, May BH, et al. Clinical evidence for association of acupuncture and acupressure with improved cancer pain: a systematic review and meta-analysis. Jama Oncol. 2020;6(2):271–8.

    Article  PubMed  Google Scholar 

  23. Zeldow B, Hatfield LA. Confounding and regression adjustment in difference-in-differences studies. Health Serv Res. 2021;56(5):932–41.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Consumer price index (CPI) databases: U.S. Bureau of Labor Statistics. https://www.bls.gov/cpi/data.htm. Accessed 22 Feb 2024.

  25. National Cancer Institute. SEER-Medicare: comorbidity SAS macros. 2021. https://healthcaredelivery.cancer.gov/seermedicare/considerations/calculation.html. Accessed 16 Jan 2023.

  26. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130–9.

    Article  PubMed  Google Scholar 

  27. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83.

    Article  CAS  PubMed  Google Scholar 

  28. DeClue RW, Antol DD, Casebeer AW, et al. Claims identification of patients with severe cancer-related symptoms. Am J Manag Care. 2020;26(6):e191–7.

    Article  PubMed  Google Scholar 

  29. Ng DQ, Chan D, Acharya MM, Grill JD, Chan A. Research attitude and interest among cancer survivors with or without cognitive impairment. Cancers (Basel). 2023;15(13):3409.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Beal SJ, Kupzyk KA. An introduction to propensity scores: what, when, and how. J Early Adolescence. 2014;34(1):66–92.

    Article  Google Scholar 

  31. Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med. 2015;34(28):3661–79.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Cole SR, Hernan MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol. 2008;168(6):656–64.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res. 2011;46(3):399–424.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Zhang XW, Hou WB, Pu FL, et al. Acupuncture for cancer-related conditions: an overview of systematic reviews. Phytomedicine. 2022;106: 154430.

    Article  PubMed  Google Scholar 

  35. Grant SJ, Kwon K, Naehrig D, Asher R, Lacey J. Characteristics and symptom burden of patients accessing acupuncture services at a cancer hospital. Integr Cancer Ther. 2021;20:15347354211002253.

  36. Highfield ES, Longacre M, Chuang YH, Burgess JF. Does acupuncture treatment affect utilization of other hospital services at an urban safety-net hospital? J Altern Complem Med. 2016;22(4):323–7.

    Article  Google Scholar 

  37. Ho DE, Imai K, King G, Stuart EA. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw. 2011;42(8):1–28.

  38. R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2023. https://www.R-project.org/.

  39. Nahin RL, Rhee A, Stussman B. Use of complementary health approaches overall and for pain management by US adults. Jama-J Am Med Assoc. 2024;331(7):613–5.

    Article  Google Scholar 

  40. Birch S, Bovey M, Robinson N. Acupuncture for chronic primary pain – are UK guidelines now consistent with other countries? Eur J Integr Med. 2021;41: 101257.

    Article  Google Scholar 

  41. Qaseem A, Wilt TJ, McLean RM, et al. Noninvasive treatments for acute, subacute, and chronic low back pain: a clinical practice guideline from the American College of Physicians. Ann Intern Med. 2017;166(7):514–30.

    Article  PubMed  Google Scholar 

  42. Peets JM, Pomeranz B. Cxbk mice deficient in opiate receptors show poor electroacupuncture analgesia. Nature. 1978;273(5664):675–6.

    Article  CAS  PubMed  Google Scholar 

  43. Pomeranz B, Chiu D. Naloxone blockade of acupuncture analgesia: endorphin implicated. Life Sci. 1976;19(11):1757–62.

    Article  CAS  PubMed  Google Scholar 

  44. Goldman N, Chen M, Fujita T, et al. Adenosine A1 receptors mediate local anti-nociceptive effects of acupuncture. Nat Neurosci. 2010;13(7):883–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Baumler P, Zhang W, Stubinger T, Irnich D. Acupuncture-related adverse events: systematic review and meta-analyses of prospective clinical studies. BMJ Open. 2021;11(9): e045961.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Mao JJ. Oncology acupuncture: precision medicine meets patient-centered care. Med Acupunct. 2020;32(6):405–7.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Lagendijk M, Mittendorf E, King TA, Gibbons C, Pusic A, Dominici LS. Incorporating patient-reported outcome measures into breast surgical oncology: advancing toward value-based care. Oncologist. 2020;25(5):384–90.

    Article  PubMed  Google Scholar 

  48. Judson PL, Abdallah R, Xiong Y, Ebbert J, Lancaster JM. Complementary and alternative medicine use in individuals presenting for care at a comprehensive cancer center. Integr Cancer Ther. 2017;16(1):96–103.

    Article  PubMed  Google Scholar 

  49. Austin S, Ramamonjiarivelo Z, Qu H, Ellis-Griffith G. Acupuncture use in the United States: who, where, why, and at what price? Health Mark Q. 2015;32(2):113–28.

    Article  PubMed  Google Scholar 

  50. Baker K, McDonald J, Steel A. Tackling health inequity: a commentary on the potential of acupuncture to improve health outcomes of marginalised populations. Acupunct Med. 2021;39(5):533–7.

    Article  PubMed  Google Scholar 

  51. Huemer M, Pansi A, Hofmann G, et al. Symptom clusters associated with complementary and alternative medicine use by cancer patients: a cross-sectional study. Plos One. 2023;18(12):e0294641.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Wolever RQ, Goel NS, Roberts RS, et al. Integrative medicine patients have high stress, pain, and psychological symptoms. Explore-Ny. 2015;11(4):296–303.

    Article  PubMed  Google Scholar 

  53. Lam CS, Ma CT, Li MCH, et al. The use of traditional, complementary and integrative medicine in Chinese adolescent and young adult patients with cancer: a multicenter cross-sectional study. Eur J Oncol Nurs. 2023;66: 102377.

    Article  PubMed  Google Scholar 

  54. Centers for Medicare & Medicaid Services. Acupuncture for chronic low back pain (CAG-00452N) - decision memo. https://www.cms.gov/medicare-coverage-database/view/ncacal-decision-memo.aspx?proposed=N&NCAId=295. Published 2020. Updated January 21, 2020. Accessed 7 March 2024.

  55. Enewold L, Parsons H, Zhao L, et al. Updated overview of the SEER-Medicare data: enhanced content and applications. J Natl Cancer Inst Monogr. 2020;2020(55):3–13.

    PubMed  PubMed Central  Google Scholar 

  56. Mues KE, Liede A, Liu J, et al. Use of the Medicare database in epidemiologic and health services research: a valuable source of real-world evidence on the older and disabled populations in the US. Clin Epidemiol. 2017;9:267–77.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We acknowledge Chapman University School of Pharmacy for acquiring the data and IQVIA for reviewing this work. We acknowledge that any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the views of IQVIA. We thank Drs Farhad Islami, Chia Jie Tan, and Daniel Wiese for their expert advice during the planning and analysis phases of the study. The authors also wish to acknowledge the support of the Chao Family Comprehensive Cancer Center Biostatistics Shared Resource, supported by the National Cancer Institute of the National Institutes of Health under award number P30CA062203. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Funding

This study was supported by the PhRMA Foundation Predoctoral Fellowship in Value Assessment-Health Outcomes (to Mr. Ng). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Author information

Authors and Affiliations

  1. School of Pharmacy & Pharmaceutical Sciences, University of California Irvine, 802 W Peltason Dr, Irvine, CA, 92697-4625, USA

    Ding Quan Ng & Alexandre Chan

  2. School of Pharmacy, Chapman University, RK 94-206, 9401 Jeronimo Road, Irvine, CA, 92618, USA

    Ding Quan Ng & Yun Wang

  3. College of Korean Medicine, Kyung Hee University, Seoul, South Korea

    Sanghoon Lee

  4. Integrative Medicine Program, Departments of Supportive Care Medicine and Medical Oncology, City of Hope Comprehensive Cancer Center, Duarte, CA, USA

    Richard T. Lee

Authors
  1. Ding Quan Ng
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Sanghoon Lee
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Richard T. Lee
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Yun Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Alexandre Chan
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

All authors read and approved the final manuscript. DQN: conception and design of work; analysis and interpretation of data; drafted and substantially revised manuscript. SL: interpretation of data. RTL: interpretation of data. YW: acquisition and interpretation of data; substantially revised manuscript. AC: interpretation of data; substantially revised manuscript.

Corresponding authors

Correspondence to Yun Wang or Alexandre Chan.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

12916_2024_3626_MOESM1_ESM.docx

Supplementary Material 1: Additional file 1: Table S1 Algorithms for identifying health-related covariates. Table S2 Acupuncture utilization statistics. Table S3 Bivariate association analysis between predetermined covariates and outcomes with significant DID estimates when comparing acupuncture-treated vs non-treated breast cancer survivors. Table S4 Baseline characteristics comparing acupuncture against control patients after propensity score matching. Table S5 Baseline characteristics of high vs low acupuncture utilization patients. Table S6 Difference-in-difference exploratory analysis by high vs low acupuncture utilization. Table S7 Baseline characteristics comparing acupuncture-treated patients for pain vs for other conditions. Table S8 Difference-in-difference exploratory analysis by acupuncture for pain vs other conditions. Table S9 Change in gabapentinoid utilization and total healthcare cost among patients treated with acupuncture for pain, stratified by high vs low acupuncture utilization. Fig. S1 Acupuncture and control cohort selection. Fig. S2 Distribution of annualized total healthcare cost, pre- and post-index. Fig. S3 Standardized mean differences in the unadjusted (original) and weighted (IPTW-weighted) cohorts. Fig. S4 Effect of number of acupuncture sessions (in quintiles) on the change in total all-cause healthcare cost from pre- to post-index.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ng, D.Q., Lee, S., Lee, R.T. et al. Real-world impact of acupuncture on analgesics and healthcare resource utilization in breast cancer survivors with pain. BMC Med 22, 394 (2024). https://doi.org/10.1186/s12916-024-03626-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12916-024-03626-2

Keywords

BMC Medicine

ISSN: 1741-7015

Contact us