“I Felt Like It Was God's Hands Putting the Needles In”: A Qualitative Analysis of the Experience of Acupuncture for Chronic Pain in a Low-Income,Ethnically Diverse,and Medically Underserved Patient Population

Introduction

Chronic pain is an extremely common problem that disrupts the lives of millions of people in the United States. Minority populations in particular experience greater prevalence and worse outcomes of chronic pain. There is a positive association between pain and impairment of physical and psychological functioning, ^{1 –3} lost productivity, ⁴ and lower socioeconomic status. ⁵ Patient-level attributes contributing to undertreatment often involve variation in attitudes, beliefs, and concerns, such as medical mistrust, culturally different expressions of pain behavior, and culturally based opinions regarding the nature of illness and healing. ^6,7

A great deal of evidence now supports the use of acupuncture therapy in the treatment of chronic pain conditions, particularly in osteoarthritis, ^8,9 neck pain, ^10,11 and low back pain. ^{12 –14} Very few studies, however, have examined the impact of acupuncture for pain in a specifically low-income, ethnically diverse, medically underserved population, where the need for more effective pain treatment is perhaps most extreme. Thus, little is known regarding the effectiveness or the process of acupuncture delivery in this setting.

The authors' previous National Center for Complementary and Integrative Health–funded acupuncture trial, ADDOPT (Acupuncture to Decrease Disparities in Outcomes of Pain Treatment) did address this gap in the literature, demonstrating the acceptability and effectiveness of acupuncture treatment for chronic pain in the urban primary care health centers of an urban practice-based research network. This pragmatic effectiveness study completed data collection on a cohort of 227 patients in November 2011. ^15,16 Back pain was the most common enrolled diagnosis (n=135; 59.5%), followed by osteoarthritis (n=37; 16.3%); many patients had multiple conditions (n=36, 15.9%). Participants were older (mean, 54.3 years); many had low income (40.6% had household incomes<$20,000), and 59% were on Medicaid (the government-subsidized insurance available for low-income patients in the United States). More than half were Hispanic (53.6%) and 27.1% were primarily Spanish speaking. About a third (39.1%) were disabled. Mean pain severity (Brief Pain Inventory) and physical health scores (Short-Form 12 instrument) changed significantly over time during the study, with 32.4% of patients experiencing a 30% or greater improvement in pain. This was the first published trial that we know of specifically focused on the effect of acupuncture for pain in low-income, ethnically diverse population.

To understand the experience of acupuncture treatment in patients from this population, a qualitative inquiry with a group of participants in the ADDOPT trial was conducted. The hope is that the findings will help inform the process of wider dissemination of acupuncture services into community-based and underserved settings.

Materials and Methods

The Albert Einstein College of Medicine institutional review board approved this qualitative study. Interviews were conducted during the 12-month period following the end of the trial; the amount of time between the end of the patients' acupuncture experience and the qualitative interview ranged from 6 to 18 months. The average time between completion of the acupuncture experience and the qualitative interview for the individual patients was approximately 11 months in each group. To ensure a sampling of the whole spectrum of acupuncture experience in this trial, two lists of 25 randomly selected patients from the parent trial to approach for qualitative interviews were generated: one from the patients who showed a significant improvement in pain during the trial and a second from the patients who did not. To do this, the pain response was dichotomized by using a 30% change in the Brief Pain Inventory as the cutoff—generally considered a clinically meaningful response in pain studies—and labeling those above that point as having had a significant response. These individuals were contacted sequentially by telephone and invited to participate in an individual qualitative interview. The research coordinator conducted interviews in Spanish or English as preferred by the participant. Those agreeing to be interviewed provided informed consent before the interview began and were provided a $50 stipend for participation. Interviews were audiotaped and then transcribed.

Based on the work of Astin ¹⁷ and Paterson and Britten ¹⁸ and other pertinent integrative medicine literature, ^19,20 probe questions for the interviews were developed to elicit information from participants about their experience of acupuncture during the ADDOPT trial. Constructs such as the concept of self-healing and the healing power of nature, as well as the cultural congruence of complementary and alternative medicine (CAM) therapies and the importance of the patients' experience with CAM served as some of the starting points for the inquiry. Using this set of questions, an initial set of eight patients were interviewed. Three readers then read each of these interviews using a thematic analysis approach. On the basis of this preliminary analysis, several questions were added to the interview guide. This expanded interview guide was then used for the subsequent 29 interviews. Sample questions from the interview guide are provided in Table 1. Interviews were no longer conducted after reaching of the end of the two lists of 25 patients, at which point 20 of 25 responders and 17 of 25 nonresponders had been interviewed, resulting in a total sample of 37 patients. The authors were prepared to interview additional patients if necessary; however, the preliminary data analysis showed that saturation had been reached with this sample, with no new additional themes emerging from the interviews by the end of that analysis.

An inductive thematic analysis strategy, as developed by Braun, ²¹ was used to examine and categorize the data and identify emergent concepts and themes. To develop a set of codes appropriate for the material, three independent readers read a sample of four interviews selected for maximum variation on rich content and independently identified themes. These readers then collaboratively identified the overlapping and unique coding categories and arrived at a tentative consensus set of codes. All three then applied the revised codes to a second sample of three interviews. The data were coded by using Dedoose software (Hermosa Beach, CA), a qualitative data analysis program. The coding of this second set of interviews was then examined for consistency between readers; the degree of consistency was found to be very high by using the Dedoose function to evaluate reliability across analysts.

All the interviews were then coded, with two of the three coders coding all interviews. This stage included a recoding of the original four interviews that had been used to generate the codes. Once the coding was completed, the immersion/crystallization approach ²² was then used to look for emergent patterns in the data; these patterns were grouped into three domains of the acupuncture process. This approach, described by Miller and Crabtree, entails an iterative process in which the researchers first “immerse” themselves in the data by reading portions of it in detail. They then temporarily suspend this process to engage in a reflective analysis to identify and describe patterns that “crystallize” as a result of this immersion. These two processes are then repeated until all the data have been examined and the meaningful patterns and themes extracted and described Once these themes were described, the interviews were re-read to identify any disconfirming data.

Results

The themes grouped naturally into three domains of the acupuncture experience: the decision-making process, the treatment experience, and the effect of acupuncture on health. Interestingly, no specific patterns or themes were unique to the pain responders or non-responders. In the following section, quotes are identified by study number, followed by sex, age in years, and status as responder (R) or non-responder (NR).

Discussion

The themes that emerged in this ethnically diverse, low-income population were very similar to those that have emerged over the past decade of qualitative research on the acupuncture experience. For example, in a qualitative analysis of patients receiving acupuncture for low back pain in the United Kingdom, Hopton et al. ²³ found that important facilitators of acupuncture acceptability included “relaxation, psychological benefit, and reduced reliance on medication” and that the common barriers to acceptability—“needle-related discomfort and financial cost”—did not detract significantly from patients' overall experience. They also found, as was seen in this analysis, that “treatment-related mediators of time, therapeutic alliance, lifestyle advice and the patient's active involvement in recovery” were important factors in acceptability. The current finding that similar factors are at work in mediating the acceptability and effectiveness of acupuncture in a medically underserved patient population potentially less familiar with acupuncture as a practice—along with the authors' previously published quantitative work showing acceptability of the service and effectiveness in reducing pain in this population—suggests a readiness and a need for more access to acupuncture services in this population.

Although much research has focused on the role of expectancy and other cognitive constructs in mediating the experience and outcomes of acupuncture, the current study found—as did Bishop and Lewith in their study ²⁴ —that more psychological and emotional factors, such as hope, optimism, and open-mindedness played a major role in the decision to try acupuncture. The complexity of the processes that go into the decision to try acupuncture—and which potentially may affect the outcomes of acupuncture treatment as well—appear to be similar in the current population to those identified in previous studies.

Given the high prevalence of comorbid depression and anxiety in patients with chronic pain in underserved settings, as well as the well-documented effect of stress on pain experience, another important finding here is that participants reported substantive benefits in mood and a new perspective on how stress influences their pain. This is consistent with qualitative work by Rugg et al. ²⁵ on patients with complex and medically unexplained symptoms that “acupuncture, initially accepted as 'just another referral'—one like many others that had been tried and proved unsuccessful—was valued for the amount of time allotted with a caring practitioner who listened and responded, as well as for the interactive and holistic nature of the sessions.” Patients in that study, as in the current one, reported “changes in their health that spanned physical, psychological, and social dimensions…as well as feelings of greater personal control, calmness, and relaxation.”

The current patients identified access and the cost of care as a major barrier. This is also consistent with previous literature both in the United Kingdom ²⁶ and the United States. ²⁷ Community or group acupuncture—which can be offered at a much lower per patient cost—has been proposed as a possible solution to this issue ^{28 –30} and is in fact the subject of a recently launched Patient-Centered Outcomes Research Institute–funded trial in the authors' network comparing both the outcomes and the experience of patients receiving individual or group acupuncture for chronic pain.

One significant limitation of this study is that because of the need to complete the quantitative analysis before proceeding with the interviews to ensure adequate representation of both responders and non-responders, interviews were conducted at least 6 months and in some case as much as 18 months after the end of the participants' acupuncture treatment; this may have biased the participants' perspective on the experience in some cases.

In conclusion, the experience of receiving acupuncture for chronic pain for patients from a low-income, ethnically diverse population has major elements in common with that of other populations studied to date. New models are needed to address issues of cost and access to care for this patient population.