Introduction
In China, approximately two in five elders meet screen-positive depression (Li et al., Reference Li, Zhang, Shao, Qi and Tian2014). Although cognitive behavioural therapy (CBT) is an effective treatment for late-life depression (Jayasekara et al., Reference Jayasekara, Procter, Harrison, Skelton, Hampel, Draper and Deuter2015), CBT for elders in rural China is rare due to a shortage of counsellors and treatment manuals. Furthermore, older adults may have limited education and their depressive symptoms are often associated with or expressed as somatic problems (Yin et al., Reference Yin, Yuan and Qu2020). Adapting CBT and training lay health workers as counsellors may fill this gap.
This pilot study aimed to test whether an adapted CBT-based intervention delivered by lay health workers had good adherence for older adults and could significantly reduce depressive symptoms.
Method
First, we recruited, trained and selected ten rural primary care health workers (village doctors, see Supplementary material), who had almost no experience in psychological therapies, as lay counsellors in Mianzhu, Sichuan province, China. We then screened elders’ mental health and conducted a pilot, parallel group, one-to-one randomized, usual-care controlled trial with masked outcome assessment to improve depression symptoms in the villages where the counsellors worked.
Participants
Undergraduates, majoring in psychology and fluent in dialect, screened adults aged 60 years or older after informed consent (n = 724; see Fig. S1 in Supplementary material).
Screen tools were the Center for Epidemiologic Studies Depression Scale (CES-D; range 0–60; details in Supplementary material), Geriatric Depression Scale (GDS; range 0–30) and the Mini-Mental State Examination (MMSE; range 0–30).
The enrolment criteria for the trial were: (a) 60 years old or older; (b) CES-D score >15; (c) GDS score >9; and (d) MMSE score >17. The exclusion criteria were: (a) lifetime psychosis; (b) significant suicide risk; and (c) poor physical condition (e.g. cancer, dementia) or impairment in communication. We used two depression scales because parallel screening could reduce the false-positive rate.
Those who met all the criteria were invited to take part in the trial. After signing new consent forms, participants were randomly assigned to the CBT arm or care as usual (CAU) arm.
CBT intervention
Therapist training, session format, supervision and a qualitative evaluation of therapy is detailed in the Supplementary material and a previous paper (Tang et al., Reference Tang, Yang, Tang, Yang, Zhang and Wang2015). Some adaptions from traditional CBT were that: (1) counsellors were suggested or encouraged to flexibly replace jargon in CBT with dialectics, proverbs and other attuned words (e.g. ‘labeling’ to ‘wear tinted colored glasses’, ‘homework’ to ‘mind prescription’); (2) older adults decided the place (home/clinic); (3) each session started with a physical health examination to address somatic complaints; (4) behavioural activations were tailored to improve adherence in daily life (e.g. going to temple regularly with neighbours, filling in the diary of activities by family members or using symbols if older adults were unable to write).
Health workers delivered the eight, 45-minute, weekly individual therapies.
A research assistant, with a master’s in psychology, monitored the counsellors’ fidelity to the manual by reviewing 25% of recordings. A licensed psychologist held a weekly, face-to-face group supervision and provided individual urgent supervision.
CAU control
Usual care was door-to-door chronic disease management for all older adults or out-patient services in clinics. Counsellors were encouraged to provide psychosocial education and referrals for CAU participants with severe mental problems.
Crisis intervention
Counsellors were trained to assess suicide risk in each CBT session and routine visit. Counsellors should immediately initiate a crisis intervention for those reporting suicidal ideation or behaviour, and inform their family members and research assistants.
Outcomes and assessments
The primary outcome was adherence to intervention (completion of sessions) and the changes in GDS scores over eight weeks. The secondary outcome was changes in anxiety and social relationships, assessed using the Self-rating Anxiety Scale (SAS; range 20–100) and domain 3 of the World Health Organization Quality of Life-BREF (WHOQOL-BREF; range 4–20), respectively.
Assessors, selected from interviewers in screen procedures, assessed outcomes at the baseline, week 4 and week 8 (or after the eighth session).
Sample size
The sample size was 50 because of convention and estimation for a future large-scale trial (see Supplementary material).
Randomization and masking
The participants were stratified by the treatment site in a one-to-one allocation using random block sizes. Only the assessors and data analysts were masked to the arm assignments (see Supplementary material).
Data analyses
Simple chi-squared test, Fisher’s exact test and generalized t-test for unequal variances were used to compare baseline characteristics. Mixed-effect linear regressions with treatment, time (weeks since baseline), and treatment×time interaction examined the effects of treatment. Missing scores were imputed using the ‘last observation carried forward’ method. All analyses followed the intention-to-treat principle. The significance level was .05. The software used was Stata 15.0.
Results
Characteristics of elders
We randomized 50 participants (age: mean 70.5 ± 5.6 years, range 64–90; female: 24): 24 to the CBT arm and 26 to the CAU arm (see Fig. S1 in the Supplementary material). The imbalanced numbers across the two arms were because a couple was enrolled, and they shared the same ID number during the randomization to avoid interfering with each other. The two arms were comparable on demographic characteristics and the GDS scores at baseline (p = .38; Tables S1 and S2 in Supplementary material). However, the CBT arm were more anxious (t = 2.02, p = .05) and had lower social relationships (t = –2.23, p = .03).
There were no difference in baseline demographics and GDS scores between those with complete (n = 49) vs incomplete data (n = 1, week 8, in the CAU arm).
Delivery of CBT
The intervention started on 7 December 2014 and ended on 17 February 2015. In the CBT arm, 19 of 24 participants completed eight sessions. There was no difference in baseline GDS scores between the three who refused or missed initially and the others who received at least one session [13.33 (SD 3.21) vs 13.86 (3.21), p = .81]. The difference in GDS scores was not significant between the five who withdrew and the retention group (14.00 ± 2.55 vs 13.74 ± 3.72, p = .88).
Effect on outcomes
Over 8-week follow-up, the difference-in-difference between the CBT arm and the CAU arm was –2.08 GDS score points (95% CI, –5.03 to 0.87). The proportion of participants with GDS scores >9 was higher in the CBT arm than in the CAU arm (9 out of 24 vs 3 out of 26; χ2 = 4.61, p = .032). The absolute benefit was 25.5%. The number needed to treat (NNT) was 3.92. The effect size (crude RR) was 3.25 (95% CI, 1.00 to 10.61).
The mixed-effect regression indicated a decrease in GDS scores over time in the CBT arm compared with the CAU (group×time interaction: β = –0.79, 95% CI, –1.39 to –0.19; Table 1).
Table 1. Mixed models to test the effect of cognitive behavioural therapy (CBT) intervention on severity of depressive symptoms, anxiety symptoms and social relationships (n = 50)
GDS, Geriatric Depression Scale; SAS, Self-Rating Anxiety Scale; WHOQOL-BREF, World Health Organization Quality of Life-BREF assessment; CI, confidence level. All regressions were adjusted for the elders’ sex (male/female), age (64–69/70–79/80–90), education (unschooled/primary school/middle school or above), marital status (married/single, divorced or widowed), and the number of self-reported chronic diseases.
However, only time fixed effect on SAS scores was significant, indicating a similar decrease over time between the two arms. The between-group effect on anxiety and social relationships was not at a statistically significant level (Table 1).
Adverse events
No significant increases in suicidal ideation (Table S3 in Supplementary material), suicidal behaviors, hospitalization or death occurred.
Discussion
As one of the pioneer trials to improve rural elders’ depressive symptoms, we adapted CBT to match the older adults’ needs, education levels and linguistic habits, and health workers’ clinical expertise. Cultural attunement was achieved by bicultural counsellors through balancing clinical flexibility and adherence to the treatment manual. The retention rate demonstrated good adherence. The significant changes in the GDS score revealed potential effectiveness to reduce depressive symptoms over eight sessions.
Our study adds evidence to prevent late-life depression in low- or middle- income areas by remaining both cognitive and behavioural strategies of CBT. A trial in India proved that problem-solving therapy and behavioural treatment for insomnia provided by lay counselors could prevent major depression among elders with sub-syndromal depressive symptoms. Another Chinese team also developed a behaviour activation-based intervention for late-life depression that was delivered by postgraduate nursing students (Xie et al., Reference Xie, He, Ding, Pan, Zhang, Zhou and Iennaco2017). As our counsellors were from the local context, our intervention design may be culturally sensitive and scale-up easily. Moreover, Xie et al. (Reference Xie, He, Ding, Pan, Zhang, Zhou and Iennaco2017) did not test the significance of the true intervention effect using methods for repetition data.
The change in anxiety and social relationships was non-significant. The reason for this may be that the CBT arm scored worse at baseline and our intervention was targeted for depression.
There were limitations. First, we used self-report scales to screen and assess the outcome. The therapy process quality was not assessed using a structured form. An insufficient sample size led to the results being under-powered. Finally, the effect size of the intervention was small, and modifications of the manual were needed.
Conclusions
The culture-adapted CBT conducted by trained lay health workers had good adherence and was temporarily effective for treating older adults with screen-positive depression in rural China.
Acknowledgements
None.
Financial support
This work was supported by the National Social Science Fund of China (Zhiyong Qu, grant number 16BSH052).
Conflicts of interest
None.
Ethics statement
This study was approved by the Review Board of the School of Social Development and Public Policy, Beijing Normal University. Oral and written informed consent was obtained from older adults before screening and randomization. Trained interviewers read consent forms and asked for their voluntary decision with voice recorded.
Supplementary material
To view supplementary material for this article, please visit: https://doi.org/10.1017/S1352465820000818
Comments
No Comments have been published for this article.