The postcard intervention against depression among community-dwelling older adults: study protocol for a randomized controlled trial

Background: Depression in older adults deteriorates quality of life and increases morbidity, mortality, and medical expenses. Medicine and social policy should work together to decrease this burden. Existing prevention studies are often based on time-consuming psychotherapies, which therefore are not feasible for a wide application at the community level. Postcard interventions have been shown to be effective for patients after hospitalization for major depression, drug overdose, or self-harm. This paper describes the protocol of a pragmatic, randomized controlled trial designed to examine the efficacy of a postcard intervention for depression among community-dwelling individuals aged 65 years or older. Methods/Design: This is a pragmatic, non-blinded, parallel comparison, randomized controlled trial using Zelen's design in a community setting. Participants will include community-dwelling older adults (aged 65 years or older) with limited social support (indicated by eating meals alone) and with symptoms of depression (scoring 4 or higher on the 15-item Geriatric Depression Scale (GDS)). The intervention will consist of sending postcards with handwritten messages and seasonal reports from a historical city to participants once a month for eight consecutive months. Self-addressed, stamped envelopes will be enclosed to facilitate non-obligatory replies. Primary outcomes will be changes in the GDS scores that are administered to all elderly inhabitants of the community every year as part of annual health checks. Secondary outcomes include quality of life as measured by a visual analogue scale, and self-rated basic and advanced activities of daily living. We will also examine the subjective sense of effectiveness of the intervention, recollection of the number of intervention mailings received, and the number of mailed replies as the index of the acceptability of the postcard intervention. The time x group interaction for two consecutive years will be analyzed using a generalized linear mixed model. To detect an effect size of 0.5 at alpha error of 0.05 and statistical power of 0.80, 63 participants per group are required. Based on an estimated consent and dropout rate of 70%, a total of 180 subjects will be recruited.