Chuna Manual Therapy vs Usual Care for Patients With Nonspecific Chronic Neck Pain: A Randomized Clinical Trial.

Importance: The incidence rate of neck pain is increasing worldwide, and the disease is associated with a high social burden. Manual therapy has been widely applied in the treatment of neck pain, but a high-quality, pragmatic randomized clinical trial for this treatment has not been conducted to date. Objective: This study aimed to compare the effectiveness of Chuna manual therapy with that of usual care for patients with chronic neck pain. Design, Setting, and Participants: A multicenter, assessor-blinded, pragmatic, randomized clinical trial was conducted between October 18, 2017, and June 28, 2019. This intention-to-treat analysis included 108 patients with chronic neck pain persisting for at least 3 months; patients were recruited from 5 hospitals in Korea. Interventions: Ten sessions (2 sessions per week for 5 weeks) of Chuna manual therapy or usual care (electrotherapy and oral medication) were conducted. Main Outcomes and Measures: The main outcome was the difference in visual analog scale (VAS) score for chronic neck pain between baseline and 5 weeks after randomization.
Results: This randomized clinical trial recruited 108 patients (mean [SD] age, 38.4 [9.3] years; 73 women [67.6%]). Fifty-four patients were allocated to the Chuna therapy group, and 54 received usual care. At 5 weeks after randomization, manual therapy showed statistically superior results compared with usual care in terms of pain (difference in chronic neck pain VAS, 16.8 mm; 95% CI, 10.1-23.5 mm), function (difference in Neck Disability Index, 8.6%; 95% CI, 4.2%-13.1%), and quality of life (difference in the European Quality of Life-5 Dimension 5 Levels (EQ-5D-5L) scores, -0.07 points; 95% CI, -0.11 to -0.02 points). Regarding the 1-year cumulative values measured using area under the curve analyses, superior outcomes were attained in the manual therapy group in terms of the numerical rating scale for chronic neck pain (1.3 points; 95% CI, 0.5-2.0 points), Neck Disability Index (6.7%; 95% CI, 2.5%-10.9%), Neck Pain Questionnaire (7.4%; 95% CI, 2.3%-12.6%), and EQ-5D-5L scores (-0.03 points; -0.07 to 0.00 points). Conclusions and Relevance: In this randomized clinical trial, for patients with chronic neck pain, Chuna manual therapy was more effective than usual care in terms of pain and functional recovery at 5 weeks and 1 year after randomization. These results support the need to consider recommending manual therapies as primary care treatments for chronic neck pain. Trial Registration: ClinicalTrials.gov identifier: NCT03294785.

Introduction

Neck pain is the second most common musculoskeletal disease after low back pain, and two-thirds of adults experience neck pain at least once in their lifetime.[1,2] Neck pain causes disabilities of varying severity and is responsible for substantial medical expenses and productivity losses.[3] The global burden of neck pain measured in disability-adjusted life-years was estimated at approximately 65.3 million in 2017.[4,5,6]

Diverse treatments exist for neck pain, including conventional guideline-based treatments (eg, medication and physical therapy) and complementary and alternative medicine (CAM) therapies (eg, acupuncture and manual therapy).[7] In the West, CAM is mostly an additional treatment supporting conventional medicine, but in Korea and the rest of East Asia, CAM is often used as the primary treatment modality.

Chuna manual therapy is a type of manual therapy that achieves a therapeutic effect by aiming to create balance in the physiological and pathological conditions of the human body. Typically, a Korean medicine doctor uses their hands or different parts of the body, sometimes with the aid of equipment, to manipulate a specific part of the patient’s body. The current form of Chuna manual therapy is developed based on the traditional Chuna techniques with a theoretical basis in Korean medicine. The therapy has been evolved and developed through integrated use and implementation of Tuina of China, chiropractic and osteopathic manipulations of the US, and Japanese manipulation techniques.[8] To improve the international applicability of Chuna manual therapy, the Korean Society of Chuna Manual Medicine joined the International Federation for Manual/Musculoskeletal Medicine, and several discussions regarding manual therapy techniques were conducted with the American Osteopathic Association through various educational and academic exchanges. In addition, the Jaseng International Conference (2018-2019) was hosted as a forum of information and opinion interchange.[9,10]

According to a Cochrane review[11] of manual therapy for neck pain, manual therapy was more effective than medication for pain and functional recovery in acute and subacute neck pain; for chronic neck pain, it was more effective than transcutaneous electrical nerve stimulation. We performed a systematic review of the East Asian manual therapy Chuna (also known as Tuina) for musculoskeletal disease and found that Chuna was more effective for neck pain than the active control group therapies of traction or drugs.[12] However, these neck pain studies mostly used the Chinese method of Tuina and had limitations, including high risk of bias.

A pragmatic clinical trial refers to a study aimed at providing support for clinical or policy decisions by providing the basis for selecting interventions in real world clinical practice. This approach differs from conventional explanatory trials that aim to determine the efficacy of an intervention.[13] Presently and to our knowledge, no well-designed pragmatic clinical studies have applied Chuna manual therapy to chronic neck pain. Therefore, we aimed to rigorously evaluate whether Chuna manual therapy is more effective and safer than usual care in conventional medicine for nonspecific chronic neck pain and to reflect the results in the Korean Medicine Clinical Practice Guideline for chronic neck pain.

Methods

Study Design and Setting

A parallel, 2-armed, multicenter, assessor-blinded, pragmatic, randomized clinical trial was conducted between October 18, 2017, and June 28, 2019, at 4 hospitals designated as spinal specialty hospitals by the Korean Ministry of Health and Welfare (Jaseng Hospital of Korean Medicine at Gangnam, Bucheon, Daejeon, and Haeundae) and 1 academic Korean medicine hospital (Kyung Hee University Korean Medicine Hospital at Gangdong). For planning and execution, we held multidisciplinary discussions once a month during the study period with experts in clinical practice guidelines, economic evaluation and clinical statistics, and rehabilitation medicine; the results of the effectiveness and economic evaluations will be incorporated in future clinical guidelines.

The trial protocol was approved by the institutional review board of Jaseng Hospital of Korean Medicine. Written informed consent was obtained from all patients. The trial protocol was previously published[14] and is available in Supplement 1. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

Eligibility Criteria

Patients were eligible for the study if they were aged between 19 and 60 years, had chronic neck pain persisting for at least 3 months,[15,16] had a neck pain intensity numerical rating scale (NRS) score greater than or equal to 5 points (range, 0-10 points) in the previous 3 days, and, if radiculopathy was present, showed the same or higher NRS pain score for chronic neck pain as for radicular pain. The exclusion criteria were cervical spine surgery; severe neurological symptoms, such as progressive neurological defect or spinal cord injury; a specific severe comorbidity that could cause chronic neck pain (eg, a tumor or fracture), or using any medications (eg, psychiatric drugs, steroids, or immunosuppressants) that could affect the study results. Additionally, patients who were on medication or had received treatment that could affect pain, such as Chuna manual therapy, physical therapy, or nonsteroidal anti-inflammatory drugs, within the previous week were also excluded, and a minimal 1-week washout period was enforced.

Randomization and Blinding

Stratified block randomization was performed using the clinical trial center as the stratifying factor. Here, random allocation was performed using block randomization to ensure even distribution of the Chuna manual therapy group (Chuna group) and usual care group at each of the 5 participating centers; the randomization schedule was generated using the PROC PLAN procedure in SAS, version 9.4 (SAS Institute). The randomization results were delivered to each center sealed in opaque envelopes. The patient registration coordinator at each center received signed participation consent forms from each patient, opened a sealed envelope in front of patients who met the inclusion criteria, and randomly allocated the patients to the 2 groups in a 1:1 ratio. As the treatments differed in the 2 groups, blinding of the practitioners and investigators was not possible, and so only the outcome assessors were blinded to the interventions.

Interventions

Patients in both the Chuna group and the usual care group underwent 10 total treatment sessions (2 sessions per week for 5 consecutive weeks). Chuna manual therapy was administered by a Korean medicine doctor with at least 3 years of Chuna experience. To standardize the technique and minimize differences between patients, before the start of the trial, the physicians received education regarding Chuna techniques based on Chunauihak (Chuna Manual Medicine, 2.5th edition), which was published by the Korean Society of Chuna Manual Therapy and is currently used as a textbook at Korean medicine colleges.[17] The physicians selected the Chuna technique based on their clinical judgment after examining the patient’s symptoms, physical test results, and imaging data. In Chuna therapy for neck pain, when the therapy was required for applications other than for the cervical area (ie, systemic application), clinical experts of Chuna therapy were consulted for performing Chuna therapy on related areas, such as thoracic vertebrae, lumbar spine, and sacroiliac joints. Accordingly, a list of Chuna manual therapy was prepared and reviewed by clinical experts of Chuna therapy. There was no limit to the number of Chuna techniques that could be used on the Chuna manual therapy list, and the type of Chuna manual therapy used during each session was recorded.

The usual care group was provided with oral medication and physical therapy (electrotherapy). Korea has an integrated medical social security system implemented and managed by the government, which is based on national health insurance covering the entire population and all medical institutions. Medical expenses for most diseases, including neck pain, might be approved for coverage by the insurance. In addition, the Health Insurance Review and Assessment (HIRA) service database contains information on medicines, medical services, and diagnosis.

To accurately reflect the clinical environment in Korea, using national data from the 2014 Korean HIRA Service-National Patient Sample database, we extracted the list of the most frequently used drugs and physical therapy treatments for the treatment categories of nonspecific neck pain, herniated cervical disk, and cervical sprain and provided these data to the medical personnel in charge of the usual care group for reference when prescribing treatment.[18] As the intervention of a pragmatic clinical trial reflects the real world clinical practice and is left to the judgment and discretion of the doctors and the patient’s compliance, the treatment list provided was only presented as a reference. Both the Chuna group and the usual care group did not have limits imposed on the numbers or types of treatment and received equivalent patient-centered care.

As cointerventions, apart from treatment provided until the primary end point (5 weeks after randomization), the rescue drug acetaminophen (up to 4 g per day) was allowed, and its dose was recorded. No further treatments for the purpose of direct neck pain alleviation were allowed up to week 5. There were no restrictions on treatments during the observation period after 5 weeks, and all treatments were recorded.

Outcomes

The primary outcome was the visual analog scale (VAS) score for chronic neck pain over a period of 3 days.[19,20,21] The secondary outcomes were the VAS score for upper limb pain; the NRS score for neck and upper limb pain[22,23]; the Vernon and Mior Neck Disability Index (NDI) score for the functional condition of the neck[24]; the functional scale Northwick Park Neck Pain Questionnaire (NPQ) score for the degree of functional disability due to neck pain[25,26]; Patient Global Impression of Change (PGIC) for overall assessment for comprehensive evaluation of improvement in neck pain and related functional disability[27]; and the European Quality of Life–5 Dimension 5 Levels (EQ-5D-5L)[28,29] and Medical Outcomes Study 12-Item Short-Form General Survey (SF-12) scores to assess participants’ quality of life.[30,31] The Work Productivity and Activity Impairment Questionnaire: Specific Health Problem (WPAI-SPH) score was used to assess loss of productivity,[32] expenses data for economic evaluation (medical costs, time-related costs),[33,34] and drug consumption. The overall study schedule is shown in detail in eTable 1 in Supplement 2, and the details of the outcomes can be found in the previously published trial protocol.[14] These outcomes were measured by blinded investigators; to avoid revealing the group allocation during medical interviews, patients were instructed at baseline and at each consultation that only minimum conversation was allowed apart from what was necessary for the measurement of outcomes.

Statistical Analysis

The sample size was estimated based on clinical trials that have used Chuna manual therapy for patients with chronic neck pain and the VAS score as the primary outcome. The effect sizes of the Chuna group and the usual care group were 1.03 and 1.49, respectively, and the effect size was conservatively hypothesized to be 0.6. Given a level of significance of α = .05, type 2 error β = 0.2, and with test power set at 80%, assuming compliance of 85%, the required sample size was calculated to be 108 patients in total (54 patients per group).[35,36] The effectiveness assessment in this clinical trial was based on differences between the groups in the change in continuous outcomes (NRS, VAS, NDI, NPQ, EQ-5D-5L, SF-12, and WPAI-SPH scores) between the baseline and each time point. An analysis of covariance was performed using the baseline value of each variable, covariant factors showing statistical differences between treatment groups at baseline as covariates, and the group as the fixed factor. To compare the total difference in each outcome within a given time period (treatment period, 5 weeks or total study period, 1 year) in the 2 groups, after random allocation, the areas under the curve (AUCs) were calculated at each time point, and the results were compared using a Student t test.

We compared the ratio of patients who showed a 50% or greater decrease in chronic neck pain indices (VAS and NRS scores) with those at baseline. After random allocation, a Kaplan-Meier survival analysis was performed to measure the time until recovery of chronic neck pain to less than half, and the curves were compared using the log-rank test. A Cox model was used to calculate the hazard ratios to compare the rates of chronic neck pain recovery to less than half. For the hazard ratios, the proportional hazard assumption was tested using Schoenfeld residuals. In this case, the proportional hazard assumption was violated depending on the treatment provided; thus, an additional interaction between group and time was added to the model. The main analysis was intention to treat, and missing values were treated by multiple imputations. For sensitivity analysis, a per protocol analysis (including only patients who completed at least 6 of 10 treatment sessions) was performed, and the last observation carried forward was used for analysis of missing values. All statistical analyses were performed using SAS, version 9.4 (SAS Institute), and significance was defined as a 2-sided P < .05.

Results

Patient Characteristics

This randomized clinical trial recruited 108 patients (mean [SD] age, 38.4 [9.3] years; 73 women [67.6%] and 35 men [32.4%]). The mean (SD) duration of chronic neck pain was 48.6 (41.8) months, and the mean (SD) VAS score was 60.1 (11.9) mm. Patients were enrolled between October 18, 2017, and June 28, 2019, and followed up for 1 year. Patients were allocated to the Chuna (54 patients) and usual care (54 patients) groups. The groups showed no significant differences in baseline characteristics (Table 1). In the usual care group, 5 patients withdrew, and 4 patients disregarded the protocol (<6 treatment sessions, 3 patients; other simultaneous treatment, 1 patient). The final per protocol analysis was performed on 54 patients in the Chuna group and 45 patients in the usual care group (Figure 1).

Table 1.

Baseline Characteristics of Participants by Randomized Group (N = 108)

Characteristic	No. (%)
	Chuna manual therapy (n = 54)	Usual care (n = 54)
Sex
Female	36 (66.7)	37 (68.5)
Male	18 (33.3)	17 (31.5)
Age, mean (SD), y	39.3 (8.2)	37.5 (10.3)
Height, mean (SD), cm	165.1 (7.0)	165.2 (8.5)
Body weight, mean (SD), kg	63.9 (14.6)	63.0 (11.2)
BMI, mean (SD)	23.2 (3.8)	23.0 (3.1)
Pain duration, mean (SD), mo	49.0 (43.4)	48.2 (40.5)
Previous medical usea
Yes	18 (33.3)	22 (40.7)
No	36 (66.7)	32 (59.3)
Major neck pain
Unilateral	17 (31.5)	7 (13.0)
Bilateral	36 (66.7)	47 (87.0)
Unknown	1 (1)	0
Area of major discomfort
Neck	17 (31.5)	19 (35.2)
Trapezius	31 (57.4)	32 (59.3)
Upper-back (between scapula)	4 (7.4)	2 (3.7)
Unknown	2 (3.7)	1 (1.9)
Radiating arm pain
Yes	32 (59.3)	26 (48.1)
No	22 (40.7)	28 (51.9)
Radiating pain area
None	22 (40.7)	28 (51.9)
Upper arm	8 (14.8)	3 (5.6)
Lower arm	10 (18.5)	12 (22.2)
Both	14 (25.9)	11 (20.4)
Major radiating pain
None	22 (40.7)	28 (51.9)
Unilateral	24 (44.4)	12 (22.2)
Bilateral	8 (14.8)	13 (24.1)
Unknown	0	1 (1.9)
Pain type (current)
Continuous	29 (53.7)	33 (61.1)
Fluctuating	25 (46.3)	20 (37)
Unknown	0	1 (1.9)
Sensory deficiency
Yes	5 (9.3)	6 (11.1)
No	49 (90.7)	48 (88.9)
Motor weakness
Yes	5 (9.3)	3 (5.6)
No	49 (90.7)	51 (94.4)
Straight neckb
Yes	41 (75.9)	43 (79.6)
No	13 (24.1)	11 (20.4)
Cervical disk space narrowingb
Yes	21 (38.9)	16 (29.6)
No	33 (61.1)	38 (70.4)
Disk degenerationb
Yes	13 (24.1)	11 (20.4)
No	41 (75.9)	43 (79.6)
Experience of Chuna manual therapy
Yes	30 (55.6)	24 (44.4)
No	24 (44.4)	30 (55.6)
Preferred intervention
Chuna manual therapy	40 (74.1)	40 (74.1)
Usual care	4 (7.4)	4 (7.4)
No preference	10 (18.5)	10 (18.5)
Credibility and expectancy of improvement with treatment, mean (SD), pointsc	7.0 (1.4)	6.0 (1.8)
VAS, mean (SD), mmd
Neck	59.5 (13.1)	60.6 (10.6)
Arm	33.3 (26.5)	28.1 (26.4)
NRS, mean (SD), pointse
Neck	5.9 (1.2)	6.2 (0.9)
Arm	3.3 (2.7)	2.9 (2.7)
NPQ, mean (SD), %f	38.4 (12.9)	36.8 (11.5)
NDI, mean (SD), %g	33.0 (11.6)	32.3 (10.6)
EQ-5D-5L score, mean (SD), pointsh	0.76 (0.11)	0.77 (0.11)
SF-12 score, mean (SD), pointsi
Physical component summary	41.7 (6.1)	44.6 (6.5)
Mental component summary	49.4 (10.0)	48.1 (9.9)
WPAI-SHP, %j	53.1 (18.1)	49.0 (18.7)

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); EQ-5D-5L, European Quality of Life–5 Dimension 5 Levels; NDI, Neck Disability Index; NPQ, Northwick Park Questionnaire; NRS, numerical rating scale; SF-12, Medical Outcomes Study 12-Item Short-Form General Survey; VAS, visual analog scale; WPAI-SHP, Work Productivity and Activity Impairment Questionnaire: Specific Health Problem.

Indicates any medical intervention sought by the patient within the last 3 mo for the purpose of alleviating neck pain.

Diagnosed by a radiology consultant after radiography.

Measured on a 9-point Likert scale.

Measured by having patients indicate their own pain level on a line, from 0 (no pain) to 100 (most severe pain imaginable), in millimeters.

Measured by having patients report their own pain level as a number from 0 (no pain) to 10 (most severe pain imaginable).

Calculated as a percentage, where higher scores indicate more severe pain and disability.

Calculated as a percentage, where higher scores indicate more severe disability.

Calculated by converting patient responses to a scale from −0.066 (lowest quality of life) to 1 (highest quality of life).

Calculated by converting patient responses to a scale from 0 (lowest quality of life) to 100 (highest quality of life).

Calculated as a percentage, evaluating the overall work impairment due to neck pain during the last week. For patients who were unemployed or not working in the last week, the impairment in daily living was rated.

Figure 1.

Study Flowchart

NRS indicates numerical rating scale.

Prespecified Outcome Measures

In the analysis by time point, at the primary end point of week 5, compared with the usual care group, the Chuna group showed significantly superior effectiveness for all pain outcomes (difference in chronic neck pain VAS, 16.8 mm; 95% CI, 10.1-23.5 mm; NRS score, 1.6 points; 95% CI, 0.9-2.3 points; NPQ, 9.0%; 95% CI, 4.3%-13.7%; P < .001), function (difference in NDI, 8.6%; 95% CI, 4.2%-13.1%), and quality of life outcomes apart from the Mental Component Summary of the SF-12 (difference in the EQ-5D-5L scores, −0.07 points; 95% CI, −0.11 to −0.02 points; Physical Component Summary, −2.6 points; 95% CI, −4.9 to −0.3 points; Mental Component Summary, −0.8 points; 95% CI, −5.0 to 3.4 points). At the 1-year follow-up, the Chuna group continued to show superior effectiveness for the chronic neck pain NRS (1.3 points; 95% CI, 0.5-2.0 points), NPQ (7.4%; 95% CI, 2.3%-12.6%), NDI (6.7%; 95% CI, 2.5%-10.9%), PGIC (−0.8 points; 95% CI, −1.3 to −0.4 points), WPAI-SPH scores (12.5%; 95% CI, 4.3%-20.8%), and EQ-5D-5L scores (−0.03 points; −0.07 to 0.00 points) (Table 2).

Table 2.

Primary and Secondary Outcomes According to Treatment and Time Since Randomization

Variable	Mean (95% CI)		Difference in change in score (95% CI)a
	Chuna manual therapy	Usual care
Week 5 postrandomization
VAS, mmb
Neck pain	26.1 (21.3 to 30.8)	43.3 (38.3 to 48.3)	16.8 (10.1 to 23.5)c
Arm pain	14.7 (10.5 to 18.9)	26.2 (20 to 32.5)	14.0 (8.0 to 19.9)c
NRS,b points
Neck pain	2.9 (2.4 to 3.3)	4.6 (4.1 to 5.1)	1.6 (0.9 to 2.3)c
Arm pain	1.6 (1.2 to 2)	2.8 (2.1 to 3.5)	1.4 (0.8 to 2.1)c
NPQ,d %	21.4 (17.8 to 25)	29.5 (25.4 to 33.5)	9.0 (4.3 to 13.7)c
NDI,d %	17.0 (13.9 to 20.2)	25.3 (21.5 to 29.1)	8.6 (4.2 to 13.1)c
EQ-5D-5L,b points	0.86 (0.84 to 0.88)	0.8 (0.76 to 0.84)	−0.07 (−0.11 to −0.02)e
SF-12,b points
PCS	47.7 (46.2 to 49.1)	46.1 (44.1 to 48)	−2.6 (−4.9 to −0.3)f
MCS	51.4 (48.4 to 54.3)	49.9 (46.5 to 53.3)	−0.8 (−5.0 to 3.4)
PGIC,b points	2.3 (2.0 to 2.5)	3.2 (2.9 to 3.4)	−0.9 (−1.2 to −0.5)c
WPAI-SHP,d %	29.3 (24.6 to 34.0)	44.4 (38.6 to 50.1)	16.1 (8.7 to 23.5)c
Month 3 postrandomization
NRS, points
Neck pain	3.5 (3.1 to 4.0)	4.3 (3.8 to 4.8)	0.7 (0.0 to 1.4)
Arm pain	1.6 (1.1 to 2.1)	2.2 (1.6 to 2.9)	0.8 (0.2 to 1.5)f
NPQ, %	20 (16.8 to 23.2)	25.7 (22.1 to 29.2)	6.2 (1.6 to 10.8)e
NDI, %	15.7 (12.9 to 18.6)	21 (18.2 to 23.8)	5.5 (1.6 to 9.3)e
EQ-5D-5L, points	0.86 (0.84 to 0.88)	0.84 (0.81 to 0.87)	−0.02 (−0.06 to 0.01)
SF-12, points
PCS	48.8 (47.1 to 50.5)	45.7 (44.1 to 47.4)	−4.1 (−6.4 to −1.8)c
MCS	54.3 (52.0 to 56.6)	52.5 (50.0 to 55.0)	−1.4 (−4.6 to 1.8)
PGIC, points	2.7 (2.4 to 3.0)	3.4 (3.1 to 3.7)	−0.7 (−1.1 to −0.3)c
WPAI-SHP, %	33.7 (28.6 to 38.8)	42 (36.9 to 47.0)	8.7 (1.4 to 16.0)f
Month 6 postrandomization
NRS, points
Neck pain	3.4 (2.9 to 3.9)	4.3 (3.8 to 4.8)	0.9 (0.1 to 1.6)f
Arm pain	2.0 (1.3 to 2.6)	2.1 (1.4 to 2.8)	0.3 (−0.5 to 1.2)
NPQ, %	19.5 (16.0 to 23.0)	24.8 (20.9 to 28.6)	5.8 (0.7 to 10.9)f
NDI, %	13.7 (11.0 to 16.4)	19.4 (16.3 to 22.5)	5.9 (1.8 to 9.9)e
EQ-5D-5L, points	0.87 (0.84 to 0.89)	0.86 (0.83 to 0.88)	−0.01 (−0.05 to 0.02)
SF-12, points
PCS	49.0 (47.4 to 50.5)	47.3 (45.7 to 48.9)	−2.3 (−4.6 to −0.1)f
MCS	53.4 (51.6 to 55.3)	53.0 (50.6 to 55.5)	0.0 (−2.9 to 2.8)
PGIC, points	2.7 (2.4 to 3.0)	3.4 (3.1 to 3.7)	−0.7 (−1.1 to −0.3)e
WPAI-SHP, %	31.3 (25.7 to 37)	40.7 (35.4 to 46.1)	10.2 (2.3 to 18.0)f
Month 9 postrandomization
NRS, points
Neck pain	3.4 (2.9 to 4)	4.0 (3.4 to 4.5)	0.4 (−0.3 to 1.2)
Arm pain	1.5 (1.0 to 2.1)	1.6 (1.0 to 2.2)	0.2 (−0.5 to 0.9)
NPQ, %	18.7 (15.0 to 22.3)	24.8 (21.6 to 28.0)	6.6 (1.8 to 11.4)e
NDI, %	14.4 (11.4 to 17.3)	18.6 (15.9 to 21.2)	4.4 (0.6 to 8.2)f
EQ-5D-5L, points	0.87 (0.84 to 0.89)	0.84 (0.82 to 0.87)	−0.03 (−0.06 to 0.01)
SF-12, points
PCS	48.2 (46.6 to 49.8)	46.9 (45.2 to 48.6)	−1.7 (−4.1 to 0.7)
MCS	55.0 (53.0 to 56.9)	52.8 (50.5 to 55.1)	−1.7 (−4.4 to 1.0)
PGIC, points	2.6 (2.3 to 2.9)	3.2 (2.9 to 3.5)	−0.6 (−1.0 to −0.2)e
WPAI-SHP, %	31.5 (26.0 to 37.0)	38.9 (33.1 to 44.6)	8.3 (0.3 to 16.3)f
Month 12 postrandomization
NRS, points
Neck pain	2.7 (2.2 to 3.2)	4.0 (3.4 to 4.6)	1.3 (0.5 to 2.0)c
Arm pain	1.5 (1.0 to 2.0)	1.8 (1.2 to 2.4)	0.4 (−0.4 to 1.2)
NPQ, %	17.3 (13.8 to 20.8)	24.3 (20.4 to 28.3)	7.4 (2.3 to 12.6)e
NDI, %	12.4 (9.7 to 15.1)	19.0 (15.7 to 22.2)	6.7 (2.5 to 10.9)e
EQ-5D-5L, points	0.87 (0.85 to 0.89)	0.84 (0.81 to 0.87)	−0.03 (−0.07 to 0.00)
SF-12, points
PCS	48.4 (46.6 to 50.2)	46.6 (44.9 to 48.4)	−2.3 (−4.9 to 0.3)
MCS	53.5 (51.3 to 55.8)	52.3 (49.6 to 54.9)	−0.9 (−4.2 to 2.5)
PGIC, points	2.5 (2.2 to 2.8)	3.4 (3 to 3.7)	−0.8 (−1.3 to −0.4)c
WPAI-SHP, %	29.4 (23.7 to 35.1)	40.8 (34.6 to 46.9)	12.5 (4.3 to 20.8)e

Abbreviations: EQ-5D-5L, European Quality of Life–5 Dimension 5 Levels; MCS, Mental Component Summary; NDI, Neck Disability Index; NPQ, Northwick Park Questionnaire; NRS, numeric rating scale; PCS, Physical Component Summary; PGIC, Patient Global Impression of Change; SF-12, the Medical Outcomes Study 12-Item Short-Form Health Survey; VAS, visual analog scale; WPAI-SHP, Work Productivity and Activity Impairment Questionnaire: Specific Health Problem.

The effectiveness outcomes were assessed as the decrease since baseline, and the differences between the 2 groups were analyzed using analysis of covariance, adjusting for the baseline value, except PGIC. The primary end point was week 5 postrandomization. Missing values were filled in by multiple imputation. The estimates for each of the groups and the differences in the changes between the 2 groups at each time point are displayed, together with the 95% CIs. The difference between the 2 groups was analyzed by performing an independent t test on the end point values.

VAS, NRS, EQ-5D-5L, SF-12, and PGIC were measured as point scores. PGIC was assessed on a scale from 1 (“improved”) to 7 (“worsened”), where a lower score indicates more improvement.

P < .001.

NPQ, NDI, and WPAI-SHP were calculated as percentages.

P < .01.

P < .05.

In the analysis of the cumulative values of each outcome at 1 year based on the AUCs, the Chuna group showed significantly superior outcomes for the chronic neck pain NRS (−46.3 points; 95% CI, −73.4 to −19.2 points), NPQ (−302.7%; 95% CI, −496.9% to −108.5%), NDI (−278.9%; 95% CI, −438.6% to −119.1%), and EQ-5D-5L (1.24 points; 95% CI, 0.01 to 2.47 points) scores. However, there were no differences between the groups in terms of the radiating pain NRS (−20.0 points; 95% CI, −52.7 to 12.8 points) or the SF-12 scores (Figure 2; eTable 4 in Supplement 2).

Figure 2.

Changes in Outcomes Over Time and Areas Under the Curves (AUCs)

Change in score for A, neck pain numerical rating scale (NRS); B, Neck Pain Questionnaire (NPQ); C, Neck Disability Index (NDI); and D, European Quality of Life–5 Dimension 5 Levels (EQ-5D-5L); during the 52 weeks postrandomization. The dots show the mean scores, and error bars show the 95% CIs. Missing values were imputed using multiple imputation. The AUCs were obtained by applying the trapezoidal rule to the Chuna manual therapy and usual care group data. The differences in AUCs between the groups were analyzed using independent t tests. AUCs indicates areas under the curves.

In the survival analysis, where recovery was defined as a 50% or greater decrease of chronic neck pain during the 5-week intervention period compared with chronic neck pain at baseline, the Chuna group showed a more rapid rate of recovery (median survival time for recovery: 4 weeks; 95% CI, 3-5 weeks; P < .001, log-rank test), and this difference was maintained at the 1-year follow-up (median survival time for recovery: 5 weeks; 95% CI, 4-13 weeks; P = .008, log-rank test). The absolute percent difference in patients who had recovered at week 5 was 46.3% (95% CI, 28.0%-64.6%), which was smaller at the 1-year follow-up (22.2%; 95% CI, 2.9%-41.6%). The hazard ratio, determined using a univariate Cox model, with recovery at 1 year as the end point was 3.27 (95% CI, 1.65-6.49) favoring the Chuna group (Figure 3).

Figure 3.

Cumulative Incidence Curves of Recovery by Group

Recovery was defined as a 50% or greater decrease in neck pain compared to that at baseline. During the intervention period, panel A shows the results for measurement of the visual analog scale (VAS) score for neck pain. During the subsequent 1-y period, panel B shows the results for measurement of the numerical rating scale (NRS) score for neck pain. The cumulative incidence curves for recovery events were obtained for each group. The median time to recovery (up to 5 weeks postrandomization) was 4 weeks postrandomization in the Chuna group (95% CI, 3-5 weeks); this was not evaluated for the usual care group. The median time to recovery up to 1 year was 5 weeks postrandomization in the Chuna group (95% CI, 4-13 weeks) and 26 weeks postrandomization in the usual care group (95% CI, 26-52 weeks). Shaded areas indicate 95% CI.

Sensitivity analysis for the principles of analysis and missing values showed similar results (eTables 5-7 in Supplement 2).

Discussion

The most clinically important key goals in treatment decisions for chronic neck pain are the extent of recovery; whether recovery is maintained 1-year postrandomization; and, given that recovery may occur spontaneously, if 2 treatment methods both show recovery in the 1 year after randomization, whether the rate of recovery is faster or cumulative pain or function is improved. Here, Chuna manual therapy showed positive results for these key goals compared with usual care in terms of improved pain relief, function, and quality of life outcomes at week 5 and 1 year after randomization and a faster rate of recovery resulting in better cumulative values after 1 year. At the primary end point of week 5 after randomization, the mean difference in the pain index was 1.8 out of 10; the mean difference in the functional index, represented by the NDI, was 8.98%. These differences can be considered both statistically and clinically significant (minimal clinically important difference for chronic neck pain: NRS = 1.5-2.5 points, NDI = 7.5 points).[37]

Manual therapies have been directly compared with other active treatments for chronic neck pain. Hoving et al[38] assessed the administration of manual therapy, physical therapy, or continued care by a general practitioner 1 to 2 times per week for 6 weeks and reported that manual therapy showed superior effectiveness on success rates and pain intensity, based on PGIC, at 7 weeks. Specifically, manual therapy showed a slightly larger difference for general practitioner care, consisting of medication and self-care advice, than did physical therapy, consisting mostly of exercise therapy. At 1 year after randomization, the superiority of manual therapy gradually decreased, and there was no statistical difference.[39] However, because the rate of recovery was faster and treatment was less costly, manual therapy was concluded to be more cost-effective.[40] Conversely, another randomized clinical trial of nonspecific neck pain assessed the administration of manual therapy or physical therapy 1 to 2 times per week for 6 weeks and reported no significant differences at either the 5-week or 1-year postrandomization follow-up and concluded that manual therapy was not cost-effective.[41,42] However, given that physical therapy included both exercise therapy and some types of manual therapy, such as manual traction and massage, this cannot be considered a strict comparison between manual therapy and other treatments.

Comparing the previous 2 studies with ours, the pain intensity was similar, at approximately 6 points (out of 10) in all 3 studies; however, although the other studies included patients with chronic pain, they mostly involved patients with subacute pain (2 weeks to 3 months), whereas in our study, all patients had chronic neck pain and a mean disease duration of 4 years. Thus, it is important that the outcomes in the Chuna group showed better improvement than those in the usual care group. In terms of baseline characteristics of these patient groups, more than half of the patients had no medical usage within the last 3 months, possibly because these patients had stopped seeking treatment after experiencing no particular improvement from previous usual care. In this regard, when we investigated the credibility and expectancy of improvement, the usual care group showed lower expectations than the Chuna group, and in both groups, more than 70% of patients showed higher preference for Chuna manual therapy. This could be a factor in the positive outcomes of the Chuna group. However, we performed numerous subgroup analyses (eFigure in Supplement 2), and the Chuna group showed superior effects to the usual care group irrespective of use of medical services, pain patterns (continuous, fluctuations), level of expectations, preference for Chuna, and previous experience of Chuna.

Chuna manual therapy was directly based on overall evaluation of the patient’s condition, considering both musculoskeletal symptoms and their overall health status and the results of neurological testing and radiological examinations. The time spent directly on treatment was approximately 10 minutes, shorter than that for other manual therapies, and the method was safe and efficient. In a recent systematic review on the safety of manual therapy, various adverse events were reported, including severe injuries, such as blood vessel dissection, spinal cord injury, and disk rupture.[43] In the Chuna group, moderate or worse adverse events were experienced by 2 of 54 patients, similar to the usual care group, and there were no severe adverse events (eTable 8 in Supplement 2).

Strengths and Limitations

Our study had certain limitations. Most importantly, it was not possible to blind the patients and practitioners from the data. As the major outcomes for chronic neck pain are all subjective variables, this is a serious problem that cannot be resolved. Additionally, the follow-up rate in the usual care group (91%) was slightly lower than that in the Chuna group (100%). If differences in the distribution could have greatly affected treatment preference, there are clearly inherent limitations, although the sensitivity analysis for adjusting credibility and expectancy of improvement with treatment showed robust results. Another limitation included issues regarding the provision of National Health Interview Survey to the usual care group. The intention in this study was to conduct usual care as accurately as possible by providing real world clinical reference data, but this may have potentially limited the treatment by the physician involved in the case, affecting the provision of optimal treatment for the patients’ neck pain. Specifically, there was an error in 1 question in the SF-12 in the initial case report form and, although several intuitive and statistical methods were applied to try to resolve this error, it remains a matter of concern (eTable 9 in Supplement 2).

Strengths of this investigation included the multicenter, assessor-blinded, pragmatic randomized clinical trial design. A pragmatic clinical trial is a study that provides information on decisions for clinical treatments, and the interventions reflect real world clinical practice with an aim at enhancing the external validity. We aimed at producing the most pragmatic design that could be applied to the real world. In the Chuna group, the techniques and sites of treatment were not restricted, and the physicians were instructed to record the actual treatment details. Although the patients had chronic neck pain, most patients received therapy that involved lumbar and thoracic spine techniques, indicating that the physicians considered overall skeletal and postural correction important (eTable 10 in Supplement 2).

In the usual care group, physicians were provided with actual National Health Insurance Service claims data for chronic neck pain and took full responsibility to decide the type of treatment; the actual distribution of chronic neck pain medications and types of physical therapy were similar to those in the 2014 claims data (eTables 11 and 12 in Supplement 2). In addition, in order to provide clinical physicians with various information on Chuna therapy for patients experiencing neck pain and usual care treatment, this study also provided reports on the results obtained by subgroup analysis. Through the provided results, it could be determined whether Chuna therapy was more effective than usual therapy for various clinical conditions of patients with neck pain. In addition, the fact that the effectiveness of Chuna therapy did not change according to the preference, experience, or expectation of the clinician could be useful for physicians; thus, this information is presented in detail in eTable 8 in Supplement 2.

Conclusions

This pragmatic randomized clinical trial showed that Chuna manual therapy is a safe and effective treatment strategy. In the chronic pain disease guidelines, a nondrug treatment strategy is recommended first; particularly, manual therapy is of substantial importance, as it is a nondrug treatment strategy.[44] We anticipate that these results will be reflected in determining the level of evidence and recommendation grades in the up-to-date Korean Medicine Clinical Practice Guideline for neck pain and that, after research on the effectiveness of Chuna manual therapy for lower back pain,[44] which has been used to support the introduction of Chuna manual therapy to the National Health Insurance Service, this study will also contribute to policy decisions regarding neck pain.