The authors have declared that no competing interests exist.
Conceived and designed the experiments: HM SR MB MS DT IW. Performed the experiments: HM SR MB SB AH HL SP MS DT IW. Analyzed the data: MB RG AK MS ES. Wrote the first draft of the manuscript: HM AK RG ES. Contributed to the writing of the manuscript: HM SR MB SB RG AH AK HL SP MS ES DT IW.
In a randomized controlled trial, Hugh MacPherson and colleagues investigate the effectiveness of acupuncture and counseling compared with usual care alone for the treatment of depression symptoms in primary care settings.
Depression is a significant cause of morbidity. Many patients have communicated an interest in non-pharmacological therapies to their general practitioners. Systematic reviews of acupuncture and counselling for depression in primary care have identified limited evidence. The aim of this study was to evaluate acupuncture versus usual care and counselling versus usual care for patients who continue to experience depression in primary care.
In a randomised controlled trial, 755 patients with depression (Beck Depression Inventory BDI-II score ≥20) were recruited from 27 primary care practices in the North of England. Patients were randomised to one of three arms using a ratio of 2∶2∶1 to acupuncture (302), counselling (302), and usual care alone (151). The primary outcome was the difference in mean Patient Health Questionnaire (PHQ-9) scores at 3 months with secondary analyses over 12 months follow-up. Analysis was by intention-to-treat.
PHQ-9 data were available for 614 patients at 3 months and 572 patients at 12 months. Patients attended a mean of ten sessions for acupuncture and nine sessions for counselling. Compared to usual care, there was a statistically significant reduction in mean PHQ-9 depression scores at 3 months for acupuncture (−2.46, 95% CI −3.72 to −1.21) and counselling (−1.73, 95% CI −3.00 to −0.45), and over 12 months for acupuncture (−1.55, 95% CI −2.41 to −0.70) and counselling (−1.50, 95% CI −2.43 to −0.58). Differences between acupuncture and counselling were not significant. In terms of limitations, the trial was not designed to separate out specific from non-specific effects. No serious treatment-related adverse events were reported.
In this randomised controlled trial of acupuncture and counselling for patients presenting with depression, after having consulted their general practitioner in primary care, both interventions were associated with significantly reduced depression at 3 months when compared to usual care alone.
Controlled-Trials.com
Depression–overwhelming sadness and hopelessness–is responsible for a substantial proportion of the global disease burden and is a major cause of suicide. It affects more than 350 million people worldwide and about one in six people will have an episode of depression during their lifetime. Depression is different from everyday mood fluctuations. For people who are clinically depressed, feelings of severe sadness, anxiety, hopelessness, and worthlessness can last for months and years. Affected individuals lose interest in activities they used to enjoy and sometimes have physical symptoms such as disturbed sleep. Clinicians can diagnose depression and determine its severity by asking patients to complete a questionnaire (for example, the Beck Depression Inventory [BDI-II] or the Patient Health Questionnaire 9 [PHQ-9]) about their feelings and symptoms. The answer to each question is given a score and the total score from the questionnaire (“depression rating scale”) indicates the severity of depression. Antidepressant drugs are usually the front-line treatment for depression in primary care.
Unfortunately, antidepressants don't work for more than half of patients. Moreover, many patients would like to be offered non-pharmacological treatment options for depression such as acupuncture–a therapy originating from China in which fine needles are inserted into the skin at specific points of the body–and counseling–a “talking therapy” that provides patients with a safe, non-judgmental place to express feelings and emotions and that helps them recognize their capacity for growth and fulfillment. However, it is unclear whether either of these treatments is effective in depression. In this pragmatic randomized controlled trial, the researchers investigate the clinical effectiveness of acupuncture or counseling in patients with depression compared to usual care in primary care in northern England. A randomized controlled trial compares outcomes in groups of patients who are assigned to different interventions through the play of chance. A pragmatic trial asks whether the intervention works under real-life conditions. Patient selection reflects routine practice and some aspects of the intervention are left to the discretion of clinician, By contrast, an explanatory trial asks whether an intervention works under ideal conditions and involves a strict protocol for patient selection and treatment.
The researchers recruited 755 patients who had consulted their primary health care provider about depression within the past 5 years and who had a score of more than 20 on the BDI-II–a score that is defined as moderate-to-severe depression on this depression rating scale–at the start of the study. Patients were randomized to receive up to 12 weekly sessions of acupuncture plus usual care (302 patients), up to 12 weekly sessions of counseling plus usual care (302 patients), or usual care alone (151 patients). Both the acupuncture protocol and the counseling protocols allowed for some individualization of treatment. Usual care, including antidepressants, was available according to need and monitored in all three groups. Compared to usual care alone, there was a significant reduction (a reduction unlikely to have occurred by chance) in the average PHQ-9 scores at both 3 and 6 months for both the acupuncture and counseling interventions. The difference between the mean PHQ-9 score for acupuncture and counseling was not significant. At 9 months and 12 months, because of improvements in the PHQ-9 scores in the usual care group, acupuncture and counseling were no longer significantly better than usual care.
These findings suggest that, compared to usual care alone, both acupuncture and counseling when provided alongside usual care provided significant benefits at 3 months in primary care to patients with recurring depression. Because this trial was a pragmatic trial, these findings cannot indicate which aspects of acupuncture and counseling are likely to be most or least beneficial. Nevertheless they do provide an estimate of the overall effects of these complex interventions, an estimate that is of most interest to patients, practitioners, and health care providers. Moreover, because this trial only considers the effect of these interventions on patients with moderate-to-severe depression as classified by the BDI-II; it provides no information about the effectiveness of acupuncture or counseling compared to usual care for patients with mild depression. Importantly, however, these findings suggest that further research into optimal treatment regimens for the treatment of depression with acupuncture and counseling is merited.
Please access these Web sites via the online version of this summary at
The US National Institute of Mental Health provides information on all aspects of
The UK National Health Service Choices website provides detailed information about
The UK charity
More personal stories about depression are available from
MedlinePlus provides links to other resources about
Depression will be the second leading cause of disease burden worldwide by 2020, according to the Global Burden of Disease Study
Acupuncturists commonly treat depression
The primary aim of this study was to determine the clinical effectiveness and cost-effectiveness of short courses of either acupuncture or counselling when compared to usual care for patients with moderate to severe depression, with the secondary aim to compare acupuncture and counselling on the basis that there is structural equivalence regarding time and attention. The health economic results will be reported separately.
Ethical approval was obtained on 21st September 2009 from the York NHS Research Ethics Committee (ref: 09/H1311/75).
The design was that of an open parallel-arm randomised controlled trial with patients randomised by the York Trials Unit to one of three arms using the allocation ratio of 2∶2∶1, respectively: acupuncture plus usual care; counselling plus usual care; and usual care alone, as detailed in the published protocol
Patients who had consulted with depression in primary care within the past 5 years, and were continuing to experience moderate to severe depression were recruited. Identified from general medical practice databases, potential participants were aged 18 or over. Patients returned written consent forms and baseline questionnaires and were eligible if scoring 20 or above on the Beck Depression Inventory-II
Patients allocated to the acupuncture and counselling groups were offered up to 12 sessions usually on a weekly basis. Participating acupuncturists were registered with the British Acupuncture Council with at least 3 years post-qualification experience. An acupuncture treatment protocol was developed
Our primary outcome measure was the PHQ-9
An effect size of 0·39 on the PHQ-9 was sought when comparing either acupuncture versus usual care alone or counselling versus with usual care alone. An allocation ratio of 2∶2∶1 was chosen in order to increase power to detect statistically significant differences between acupuncture and counselling. As a result, a smaller effect size of 0·32 was sought when comparing acupuncture and counselling. Anticipated differences between the treatments were likely to be smaller but not clinically meaningful. With 90% power and a two-sided significance level of 5%, the required group sizes were 204, 204, and 102 participants in the acupuncture, counselling, and usual care alone arms, respectively. Allowing for 20% attrition, the total sample size required was 640 (i.e., groups of 256, 256, 128, respectively).
The primary effectiveness comparisons were between acupuncture plus usual care versus usual care alone, and counselling plus usual care versus usual care alone. A secondary comparison was between acupuncture plus usual care versus counselling plus usual care. The primary outcome was PHQ-9 at 3 months, for which we used analysis of covariance (ANCOVA), with baseline PHQ-9 score as covariate. Multiple imputation by chained regression was used for missing data using treatment group, baseline measures (PHQ-9, BDI-II, SF-36, EQ-5D Anxiety/Depression), and demographics (age and gender). The primary analysis was based on the imputed rather than raw data in order to take account of the profile of non-responders.
To assess the overall clinical impact at 12 months, we evaluated area under the curve (AUC) for PHQ-9 over that period by linear regression, predicting average AUC while controlling for baseline PHQ-9. We also explored PHQ-9 outcomes across all time points in more detail using random intercept linear mixed models with fixed effects for treatment arm, time, and arm by time interaction for each treatment comparison. The models nested time points within patients and controlled for baseline PHQ-9 and potential mediators including patients' prior expectations and preferences regarding the treatments. These were identified by univariate regressions (
All analyses were carried out in Stata (version 12.1) on an intention-to-treat basis. Statistical tests were at a two-sided 0·05 significance level. Analysis of residuals was undertaken for all regression models in order to assess model assumptions.
Patient flow in the trial is presented in
Characteristic | Acupuncture + Usual Care | Counselling + Usual Care | Usual Care | Total |
Mean (SD) | 43·4 (13·24) | 43·5 (13·26) | 43·5 (13·93) | 43·5 (13·37) |
Median (min, max) | 43 (18–86) | 43 (18–93) | 42 (18–89) | 43 (18–93) |
Inter-quartile range | 34–52 | 33–52 | 32–54 | 33–53 |
Missing | – | – | – | – |
Male | 88 (29·1%) | 69 (22·9%) | 44 (29·1%) | 201 (26·6%) |
Female | 214 (70·9%) | 233 (77·1%) | 107 (70·9%) | 554 (73·4%) |
Missing | – | – | – | – |
Mean (SD) | 18·0 (4·69) | 18·0 (3·92) | 18·1 (4·62) | 18·0 (4·37) |
Median (min, max) | 16 (13–54) | 16 (14–48) | 16 (14–54) | 16 (13–54) |
Inter-quartile range | 16–18 | 16–19 | 16–19 | 16–19 |
Missing | 17 (5·6%) | 7 (2·3%) | 7 (4·6%) | 31 (4·1%) |
Full-time education | 13 (4·4%) | 5 (1·7%) | 5 (3·3%) | 23 (3·1%) |
Working full-time | 112 (38·0%) | 107 (36·4%) | 62 (41·3%) | 281 (38·0%) |
Working part-time | 57 (19·3%) | 59 (20·1%) | 28 (18·7%) | 144 (19·5%) |
Unable to work | 38 (12·9%) | 42 (14·3%) | 15 (10·0%) | 95 (12·9%) |
Looking after home | 37 (12·5%) | 32 (10·9%) | 14 (9·3%) | 83 (11·2%) |
Retired | 23 (7·8%) | 30 (10·2%) | 12 (8·0%) | 65 (8·8%) |
Other | 15 (5·1%) | 19 (6·5%) | 14 (9·3%) | 48 (6·5%) |
Missing | 7 (2·3%) | 8 (2·6%) | 1 (0·7%) | 16 (2·1%) |
In last 2 wk | 224 (75·7%) | 235 (78·6%) | 115 (77·7%) | 574 (77·3%) |
Missing | 6 (2·0%) | 3 (1·0%) | 3 (2·0%) | 12 (1·6%) |
Not first major episode | 196 (89·5%) | 217 (93·5%) | 100 (87·7%) | 513 (90·8%) |
Missing | 5 (2·2%) | 3 (1·3%) | 1 (0·9%) | 9 (1·6%) |
4+ previous episodes | 143 (73·0%) | 165 (76·7%) | 81 (82·7%) | 389 (76·4%) |
Missing | 0 (0·0%) | 2 (0·9%) | 2 (2·0%) | 4 (0·8%) |
Mean (SD) | 25·8 (12·69) | 24·9 (11·73) | 24·4 (12·55) | 25·2 (12·28) |
Median (min, max) | 23 (3–79) | 22 (6–71) | 20 (0–78) | 22 (0–79) |
Inter-quartile range | 16–33 | 16–31 | 16–30 | 16–31 |
Missing | 9 (3·0%) | 3 (1·0%) | 4 (2·6%) | 16 (2·1%) |
Depression medication in last 3 mo | 189 (62·6%) | 220 (72·9%) | 110 (72·9%) | 519 (68·7%) |
Missing | 0 (0·0%) | 0 (0·0%) | 0 (0·0%) | 0 (0·0%) |
Analgesic medication in last 3 mo | 147 (48·8%) | 126 (42·3%) | 86 (57·3%) | 359 (47·9%) |
Missing | 1 (0·3%) | 4 (1·3%) | 1 (0·7%) | 6 (0·8%) |
Not anxious/depressed | 8 (2·7%) | 8 (2·7%) | 5 (3·3%) | 21 (2·8%) |
Moderately anxious/depressed | 219 (73·0%) | 221 (73·2%) | 114 (75·5%) | 554 (73·6%) |
Extremely anxious/depressed | 73 (24·3%) | 73 (24·2%) | 32 (21·2%) | 178 (23·6%) |
Missing | 2 (0·7%) | 0 (0·0%) | 0 (0·0%) | 2 (0·3%) |
Mean (SD) | 15·3 (5·33) | 16·6 (5·27) | 16·2 (5·09) | 16·0 (5·29) |
Median (min, max) | 15 (3–27) | 17 (4–27) | 16 (5–27) | 16 (3–27) |
Inter-quartile range | 11–19 | 13–21 | 13–20 | 12–20 |
Missing | 1 (0·33%) | 0 (0·0%) | 0 (0·0%) | 1 (0·1%) |
None (0–4) | 4 (1·3%) | 2 (0·7%) | 0 (0·0%) | 6 (0·8%) |
Mild (5–9) | 44 (14·6%) | 29 (9·6%) | 14 (9·3%) | 87 (11·5%) |
Moderate (10–14) | 97 (32·2%) | 74 (24·5%) | 46 (30·5%) | 217 (28·8%) |
Moderately severe (15–19) | 88 (29·2%) | 96 (31·8%) | 47 (31·1%) | 231 (30·6%) |
Severe (20–27) | 68 (22·6%) | 101 (33·4%) | 44 (29·1%) | 213 (28·3%) |
Mean (SD) | 32·0 (8·54) | 33·3 (9·11) | 31·8 (8·17) | 32·5 (8·72) |
Median (min, max) | 31 (20–57) | 32 (20–60) | 30 (20–56) | 31 (20–60) |
Inter-quartile range | 25–37 | 26–39 | 25–37 | 26–38 |
Missing | – | – | – | – |
Moderate (20–28)2 | 124 (41·1%) | 104 (34·4%) | 56 (37·1%) | 284 (37·6%) |
Severe (29–63) | 178 (58·9%) | 198 (65·6%) | 95 (62·9%) | 471 (62·4%) |
Mean (SD) | 58·8 (27·99) | 58·0 (29·17) | 54·4 (27·83) | 57·6 (28·44) |
Median (min, max) | 62 (0–100) | 62 (0–100) | 51 (0–100) | 52 (0–100) |
Inter-quartile range | 41–84 | 31–84 | 31–74 | 32–84 |
Missing | 1 (0·3%) | 3 (1·0%) | 0 (0·0%) | 4 (0·5%) |
Very ineffective | 10 (3·3%) | 6 (2·0%) | 1 (0·7%) | 17 (2·3%) |
Fairly ineffective | 9 (3·0%) | 11 (3·7%) | 6 (4·0%) | 26 (3·5%) |
Can't decide | 187 (61·9%) | 204 (68·0%) | 103 (68·2%) | 494 (65·6%) |
Fairly effective | 66 (21·9%) | 44 (14·7%) | 29 (19·2%) | 139 (18·5%) |
Very effective | 30 (9·9%) | 35 (11·7%) | 12 (8·0%) | 77 (10·2%) |
Missing | 0 (0·0%) | 2 (0·7%) | 0 (0·0%) | 2 (0·3%) |
Very ineffective | 23 (7·7%) | 18 (6·0%) | 15 (9·9%) | 56 (7·5%) |
Fairly ineffective | 50 (16·7%) | 43 (14·3%) | 21 (13·9%) | 114 (15·2%) |
Can't decide | 101 (33·8%) | 100 (33·2%) | 39 (25·8%) | 240 (32·0%) |
Fairly effective | 95 (31·8%) | 97 (32·2%) | 65 (43·1%) | 257 (34·2%) |
Very effective | 30 (10·0%) | 43 (14·3%) | 11 (7·3%) | 84 (11·2%) |
Missing | 3 (1·0%) | 1 (0·3%) | 0 (0·0%) | 4 (0·5%) |
Very ineffective | 29 (9·6%) | 34 (11·3%) | 6 (4·1%) | 69 (9·2%) |
Fairly ineffective | 60 (19·9%) | 62 (20·7%) | 44 (29·7%) | 166 (22·1%) |
Can't decide | 108 (35·8%) | 92 (30·7%) | 50 (33·8%) | 250 (33·3%) |
Fairly effective | 95 (31·5%) | 97 (32·3%) | 46 (31·1%) | 238 (31·7%) |
Very effective | 10 (3·3%) | 15 (5·0%) | 2 (1·4%) | 27 (3·6%) |
Missing | 0 (0·0%) | 2 (0·7%) | 3 (2·0%) | 5 (0·7%) |
Very ineffective | 10 (3·3%) | 18 (6·0%) | 6 (4·1%) | 34 (4·5%) |
Fairly ineffective | 9 (3·0%) | 43 (14·3%) | 44 (29·7%) | 96 (12·8%) |
Can't decide | 187 (61·9%) | 100 (33·2%) | 50 (33·8%) | 337 (44·9%) |
Fairly effective | 66 (21·9%) | 97 (32·2%) | 46 (31·1%) | 209 (27·8%) |
Very effective | 30 (9·9%) | 43 (14·3%) | 2 (1·4%) | 75 (10·0%) |
Missing | 0 (0·0%) | 1 (0·3%) | 3 (2·0%) | 4 (0·5%) |
Acupuncture | 177 (58·8%) | 171 (57·6%) | 82 (54·7%) | 430 (57·5%) |
Counselling | 55 (18·3%) | 75 (25·3%) | 34 (22·7%) | 164 (21·9%) |
Usual Care | 2 (0·7%) | 7 (2·4%) | 1 (0·7%) | 10 (1·3%) |
No preference | 67 (22·3%) | 44 (14·8%) | 33 (22·0%) | 144 (19·3%) |
Missing | 1 (0·3%) | 5 (1·7%) | 1 (0·7%) | 7 (0·9%) |
Randomised to preferred treatment | 177 (58·8%) | 75 (25·3%) | 1 (0·7%) | 253 (33·8%) |
Randomised to non-preferred treatment | 57 (18·9%) | 178 (59·9%) | 116 (77·3%) | 351 (46·9%) |
No preference | 67 (22·3%) | 44 (14·8%) | 33 (22·0%) | 144 (19·3%) |
Missing | 1 (0·3%) | 5 (1·7%) | 1 (0·7%) | 7 (0·9%) |
Percent given out of non-missing data, percent missing given out of applicable group total.
A BDI-II score of 20 or above was required to be eligible for the trial.
In the acupuncture arm, 266 patients (88.1%) received one or more treatment sessions (mean = 10.3, standard deviation [SD] 3.14) with one of 23 acupuncturists. By the 3-month time point, the average number of sessions received was 8.7 (SD = 3.34), with 133 patients (50.0%) having completed all of their sessions. In the counselling arm, 231 patients (76.5%) received one or more treatment sessions (mean = 9.0, SD = 3.74) with one of 37 therapists. By the 3-month time point, the average number of sessions received was 7.5 (SD = 3.60), with 114 patients (49.4%) having completed all of their sessions. The mean time from randomisation to last treatment was 117 days in both treatment arms (SD = 47.0 and 51.2, respectively), a period that included time to first appointment. Acupuncturists were allocated an average of 13 patients each (range two to 45), and counsellors an average of eight patients each (range one to 27).
In the logbooks, we collected data on acupuncturists' and counsellors' self-reports of intervention protocol violations. We received reports of four cases of violations of the acupuncture protocol (which incidentally did not involve counselling), one of which was deemed a true case involving prescription of a herbal lotion application to reduce swelling and pain. We received seven reports of violations of the counselling protocol, of which two were deemed true cases; one case involved goal setting and in another the counsellor reported being analytical and interpretative beyond the scope of humanistic counselling.
Details of medication use over the period of the trial is reported below, and further details of all three interventions are summarised for acupuncture (
Unadjusted mean scores for the PHQ-9 at all time points are presented in
Acupuncture, black line; counselling, grey line; usual care, dashed line.
Outcome Measures | Acupuncture + Usual Care | Counselling + Usual Care | Usual Care | Total | ||||
Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |||||
PHQ-9 | ||||||||
Baseline | 301 | 15·3 (5·33) | 302 | 16·6 (5·27) | 151 | 16·2 (5·09) | 754 | 16·0 (5·29) |
3 mo | 249 | 9·4 (6·33) | 237 | 10·9 (6·45) | 128 | 12·7 (6·47) | 614 | 10·7 (6·51) |
6 mo | 235 | 9·1 (6·51) | 228 | 10·1 (6·87) | 120 | 12·0 (6·85) | 583 | 10·1 (6·80) |
9 mo | 234 | 9·7 (6·90) | 215 | 10·1 (7·03) | 120 | 11·9 (7·04) | 569 | 10·3 (7·02) |
12 mo | 233 | 9·3 (6·68) | 220 | 10·1 (6·86) | 119 | 11·5 (6·98) | 572 | 10·1 (6·85) |
BDI-II | ||||||||
Baseline | 302 | 32·0 (8·54) | 302 | 33·3 (9·11) | 151 | 31·8 (8·17) | 755 | 32·5 (8·72) |
12 mo | 226 | 20·4 (13·19) | 211 | 21·4 (13·64) | 151 | 23·8 (12·63) | 550 | 21·4 (13·29) |
In the AUC analysis of PHQ-9 scores over 12 months, the benefit of acupuncture and counselling over usual care alone in terms of score reduction seen at 3 months remained when averaging over the 12-month trial period: acupuncture reduced PHQ-9 scores by −1·55 (95% CI −2·41 to −0·70); counselling by −1·50 (95% CI −2·43 to −0·58) (
In secondary analyses, PHQ-9 at 3 months was associated with two factors: expectations of counselling (
Linear mixed modelling showed reduced PHQ-9 depression in the acupuncture and counselling groups as compared to usual care at 3 and 6 months (
Unadjusted mean scores for the BDI-II at baseline and 12 months are shown in
Analysis | Month | Group 1 | Group 2 | Group Difference | ||||||
Mean | SE | Mean | SE | Mean | SE | 95% CI | ||||
Acupuncture | Usual Care | |||||||||
ANCOVA |
3 | 452 | 9·8 | 0·41 | 12·3 | 0·58 | −2·46 | 0·636 | −3·72 to −1·21 | <0·001 |
Mixed |
3 | 372 | 9·8 | 0·27 | 12·1 | 0·38 | −2·29 | 0·475 | −3·22 to −1·36 | – |
6 | 350 | 9·6 | 0·33 | 11·5 | 0·46 | −1·90 | 0·569 | −3·02 to −0·79 | – | |
9 | 348 | 10·2 | 0·39 | 11·1 | 0·54 | −0·83 | 0·671 | −2·15 to 0·49 | – | |
12 | 347 | 9·7 | 0·45 | 10·7 | 0·64 | −0·99 | 0·785 | −2·53 to 0·55 | – | |
AUC |
3–12 | 407 | 10·9 | 0·25 | 12·5 | 0·35 | −1·55 | 0·435 | −2·41 to −0·70 | – |
Counselling | Usual Care | |||||||||
ANCOVA | 3 | 453 | 11·1 | 0·40 | 12·8 | 0·58 | −1·73 | 0·648 | −3·00 to −0·45 | 0·008 |
Mixed | 3 | 362 | 10·9 | 0·28 | 12·8 | 0·38 | −1·83 | 0·477 | −2·76 to −0·90 | – |
6 | 345 | 10·4 | 0·33 | 12·2 | 0·47 | −1·78 | 0·576 | −2·91 to −0·65 | – | |
9 | 332 | 10·5 | 0·40 | 11·8 | 0·55 | −1·26 | 0·688 | −2·61 to 0·08 | – | |
12 | 336 | 10·4 | 0·47 | 11·4 | 0·65 | −1·00 | 0·805 | −2·58 to 0·57 | – | |
AUC | 3–12 | 402 | 11·6 | 0·28 | 13·1 | 0·38 | −1·50 | 0·470 | −2·43 | −0·58 |
Acupuncture | Counselling | |||||||||
ANCOVA | 3 | 603 | 10·0 | 0·41 | 10·8 | 0·40 | −0·76 | 0·514 | −1·77 to 0·25 | 0·140 |
Mixed | 3 | 402 | 9·5 | 0·29 | 9·4 | 0·32 | 0·11 | 0·439 | −0·75 to 0·97 | – |
6 | 371 | 9·6 | 0·35 | 9·1 | 0·38 | 0·45 | 0·527 | −0·58 to 1·49 | – | |
9 | 360 | 10·0 | 0·41 | 9·0 | 0·45 | 0·97 | 0·621 | −0·25 to 2·19 | – | |
12 | 361 | 9·6 | 0·48 | 9·0 | 0·53 | 0·59 | 0·721 | −0·82 to 2·01 | – | |
AUC | 3–12 | 531 | 11·1 | 0·27 | 11·2 | 0·27 | −0·06 | 0·378 | −0·81 to 0·68 | – |
ANCOVA, effect of trial arm on imputed PHQ-9 at 3 months (adjusting for baseline PHQ-9).
Mixed, effect of trial arm on PHQ-9 over 12 months, (mixed effects model, including time, trial arm x time, baseline PHQ-9, baseline expectation of counselling and baseline expectation of allocated treatment. For the comparison between acupuncture and counselling, treatment time by 3 months and practitioner's empathy were additionally included.)
AUC, effect of trial arm on average PHQ-9 depression over 12 months (adjusting for Baseline PHQ-9).
SE, standard error.
The majority (68·7%) of all patients were taking antidepressants at baseline (
Analysis | Month | Group 1 | Group 2 | Group Difference | |||||
Mean | SE | Mean | SE | Mean | SE | 95% CI | |||
Acupuncture | Usual Care | ||||||||
ANCOVA |
12 | 445 | 22·8 | 1·34 | 25·7 | 1·82 | −2·88 | 1·419 | −5·68 to −0·08 |
Counselling | Usual Care | ||||||||
ANCOVA | 12 | 449 | 22·7 | 1·47 | 25·4 | 1·74 | −2·74 | 1·399 | −5·50 to 0·02 |
Acupuncture | Counselling | ||||||||
ANCOVA | 12 | 401 | 22·5 | 0·92 | 21·9 | 1·02 | 0·59 | 1·281 | −1·93 to 3·11 |
ANCOVA, effect of trial arm on imputed BDI-II at 12 months (adjusting for baseline BDI-II, baseline expectation of counselling and baseline expectation of allocated treatment. For the comparison between acupuncture and counselling, treatment time by 3 months and practitioner's empathy were additionally included.)
Prescribed Medication | Acupuncture + Usual Care | Counselling + Usual Care | Usual Care | Total | ||||||||
Percent |
Change from Baseline | Percent | Percent | Change from Baseline | Percent | Change from Baseline | Percent | Change from Baseline | ||||
Baseline | 189 | 62·6% | – | 220 | 72·9% | – | 110 | 72·9% | – | 519 | 68·7% | – |
3 mo | 147 | 60·7% | −0·9% | 155 | 69·2% | −3·7% | 80 | 66·7% | −6·2% | 382 | 65·2% | −3·5% |
6 mo | 126 | 54·6% | −8·0% | 145 | 65·3% | −7·6% | 75 | 63·6% | −9·3% | 346 | 60·6% | −8·1% |
9 mo | 123 | 54·0% | −8·6% | 124 | 60·8% | −12·1% | 74 | 66·1% | −6·8% | 321 | 59·0% | −9·7% |
12 mo | 121 | 52·4% | −10·2% | 124 | 58·2% | −14·7% | 69 | 61·6% | −11·3% | 314 | 56·5% | −12·2% |
Baseline | 147 | 48·8% | – | 126 | 42·3% | – | 86 | 57·3% | – | 359 | 47·9% | – |
3 mo | 73 | 30·8% | −18·0% | 93 | 42·5% | +0·2% | 57 | 48·3% | −9·0% | 223 | 38·9% | −9·0% |
6 mo | 99 | 43·4% | −5·4% | 84 | 38·5% | −3·8% | 52 | 46·4% | −10·9% | 235 | 42·1% | −5·8% |
9 mo | 82 | 36·8% | −12·0% | 71 | 34·8% | −7·5% | 55 | 50·0% | −7·3% | 208 | 38·7% | −9·2% |
12 mo | 84 | 36·8% | −12·0% | 87 | 41·0% | −1·3% | 54 | 49·5% | −7·8% | 225 | 41·0% | −6·9% |
All percentages given out of total number of patients with valid responses.
The number of patients experiencing a serious adverse event (SAE) over the 12 months, as judged by a clinician (IW), was 16 (5·3%) of 302, 26 (8·6%) of 302, and nine (6·0%) of 151 from the acupuncture, counselling, and usual care groups, respectively, of whom nine had more than one SAE (range 2–4). No SAEs, including three deaths, were known to be related to treatment. The number of patients experiencing a non-serious adverse event (NSAE) was 56 (18·5%), 47 (15·6%), and 40 (26·5%), respectively, of whom 17 had more than one NSAE (range 2–4).
In this first randomised trial of acupuncture and counselling for patients continuing to experience depression in primary care, we found statistically significant benefits at 3 months associated with both interventions when provided alongside usual care. This population can be characterised as having experienced recurring bouts of depression (76% having had four or more episodes), with the first episode on average at age 25, some 19 years previously, and 69% were on anti-depressant medication. Moreover significant benefits were observed over the 12-month period as a whole in an AUC analysis. Over time, the usual care group improved such that by 9 and 12 months the benefits of acupuncture and counselling were no longer significantly better than usual care in our primary outcome measure, the PHQ-9. No serious adverse events related to treatment were reported.
This study had a clear and practical research question with an appropriate trial design, namely a pragmatic randomised controlled trial, which modelled closely what would happen if patient referrals to acupuncturists and counsellors were routine. Pragmatic trials have an emphasis on external validity with the intention that findings are generalisable to typical patients and settings. We recruited patients in primary care from those who had consulted with depression and continued to be moderately or severely depressed, thereby excluding patients whose symptoms had been alleviated sufficiently by other treatment. Using a different screening measure (BDI-II) from the one we used for measuring outcome (PHQ-9) had the advantage that we avoided the well-known problems of regression towards the mean that occur when we choose people as being above a cut-off. The natural variation in responses to questionnaires produced a few scores below the cut-off for depression on the PHQ9 measurement, 0.8% of participants being below 5, which would have them classified as not depressed, and 11.5% between 5 and 9, which would be mild depression. This is an inevitable consequence of the difficulty of measuring depression. A single PHQ9 below 5 does not mean that someone is not depressed and a single BDI above 20 does not mean that they are depressed, these scores occurring in the same individuals for six participants. Each of these scores is only an uncertain estimate of the person's position on a continuum.
The attrition between invitation to participate and randomisation at baseline was typical of trials that recruit via primary care databases. Multiple imputation was used to compensate to some degree for the limitations related to the loss of follow-up data. The randomisation into groups provided a control for temporal effects and other factors across all patients that might have influenced outcomes beyond the treatment itself. We provided standardised treatment protocols for the acupuncturists and counsellors that were designed to reflect routine practice, allowing individualisation to match patient variability, while ensuring all practitioners met appropriate standards of qualification and experience. While no attempt was made to standardise usual care, it was however provided to patients in all arms of the trial, with a majority of patients continuing with anti-depressants, and differences between groups in usual care at all time points were minimal (see
Pragmatic trials, however, tell us less about which aspects of the interventions might be most or least beneficial. Such trials are not designed to separate out relative contributions from treatment components that are specific or non-specific to the intervention. With regard to two treatment components in the acupuncture and counselling groups that are often considered “non-specific,” namely session time and quality of therapist attention, we found that specific treatment effects remained when these were accounted for. Despite the limitation that a pragmatic approach has in ascribing outcomes to different treatment components, exacerbated by some loss to follow-up, this design nevertheless provides a useful estimate of overall effects of complex interventions, an estimate of most interest to patients, practitioners and providers.
In a recent Cochrane review of acupuncture for depression, no studies focussed on moderate to severe depression. The only comparator equivalent to the usual care of our study was a wait list comparator, which involved two trials by the same research team
If treatment success is defined as improvement from a depressed PHQ-9 score (≥10) to a non-depressed score (≤9) with an improvement of at least 50%
Further research into the optimal populations and treatment regimens for acupuncture and counselling is merited. The characteristics of patients one recruits to trials of interventions for depression in primary care will depend on the recruitment methods used. Patients recruited at the point of consultation will be different from those recruited via databases, as we did in this trial, and different again from those with depression who are not consulting at all. In the trial we report here, what was important was that patients had consulted in primary care, continued to be depressed, and were seeking other interventions that might help with their depression. A different recruitment method could be used to assess whether the interventions would be more effective for patients referred at the onset of their first episode of depression, rather than after many episodes, as was the case in the present study in which 76% had had four or more previous episodes.
We recruited patients using the Beck Inventory (BDI-II) with scores greater or equal to 20, which is classified by this measure as “moderate to severe.” However other classifications systems, such as DSM, have different categories of severity. What our trial does not tell us is whether acupuncture or counselling performs better or worse for the more mild forms of depression. Nor do we know whether patients not receiving anti-depressant medication (one third of our sample) do better or not, or whether fewer sessions would be sufficient for mild to moderate depression, or more sessions would improve outcomes for those with severe depression (approximately 60% of our sample). For those randomised to acupuncture, it was usually their first experience of this intervention, whereas many randomised to counselling had received this intervention before; however, it is not clear whether this difference impacted on outcomes.
With regard to the acupuncture provision, there is some evidence from the literature that a uniform combination of points can work for depressed patients
A full economic analysis is needed to combine the costs and effects of each treatment option to determine the cost-effective treatment. Although no other trial data are available on the costs and effects of acupuncture for these primary care patients, other trials exist for counselling in primary care and, to guide policy, these should be synthesized into cost-effectiveness estimates.
To our knowledge, our study is the first to rigorously evaluate the clinical and economic impact of acupuncture and counselling for patients in primary care who are representative of those who continue to experience depression in primary care. We have provided evidence that acupuncture versus usual care and counselling versus usual care are both associated with a significant reduction in symptoms of depression in the short to medium term, and are not associated with serious adverse events.
(PDF)
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
We acknowledge the patients, counsellors, acupuncturists, and general medical practitioners; Sylvia Schroer for conducting the pilot study; Mike Bennett and David Smith from York & District Mind; and from the ACUDep research team, Matthew Bailey, Richard Blackwell, Sally Brabyn, Ben Cross, Janet Eldred, David Geddes, Simon Gilbody, Pauline Holloway, Hillary Marshall, Stephen Morley, Karen Overend, Lucy Revell, Val Wadsworth, and Kerry Wheeler. We especially acknowledge Peter Bower who chaired our Independent Steering Group.
analysis of covariance
area under the curve
standard deviation