Three Approaches to Maintenance Therapy for Chronic Insomnia in Older Adults ABSTRACT Insomnia is nearly twice as common among older adults as it is the general population. This is of significant clinical concern as insomnia is a risk factor for new onset and recurrent psychiatric and medical illness. Taken together, the prevalence and consequences of insomnia in older adults suggests that insomnia should not go untreated. This clinical imperative is further underscored by 1) the reconceptualization of Insomnia within the DSM-5 and ICSD-3 as a disorder (vs. a symptom of other disorders) and 2) the findings that targeted treatment of sleep continuity disturbance may produce clinical gains for medical and psychiatric disorders that occur comorbidly with insomnia. Thus, at present, the question is not whether to treat but how to best treat the disorder in general, and specifically in the context of older adults. While CBT-I is the treatment of choice, the medical treatment of insomnia remains the primary alternative for those for whom CBT-I is not indicated, preferable, or available. Of the available medical treatments, the best studied strategies are benzodiazepines and benzodiazepine receptor agonists (BZRAs). In both cases, treatment is typically accomplished with either nightly or intermittent dosing. In the case of nightly dosing (QHS), BZRAs have been found to be safe and efficacious for periods of up to a year. Less clear is whether such efficacy can be maintained in the context of maintenance therapy (over the course of years or decades). In the case of intermittent dosing (IDS), the reduced usage approach is thought to extend the efficacy and safety “half-life” of pharmacotherapy, but at a cost: little or no treatment effects on non-medication nights. In order to address this issue, we have proposed that behavioral principles be applied to pharmacotherapy to determine whether clinical gains obtained with standard therapy can be maintained with partial reinforcement (nightly pill use where some of the pills contain zolpidem and some are placebos). Building upon the findings from our prior study with partial reinforcement, we propose to assess a low frequency approach to dosing as compared to nightly medication use in older adults. The study will be conducted in three phases. In Phase 1, all subjects receive 5mg zolpidem nightly for one month and are assessed for treatment response. In Phase 2, treatment responders will be randomized to one of four maintenance conditions for three months: Nightly medication use (QHS); one of two low frequency partial reinforcement conditions (1 or 3 active doses per week and the rest placebos); or a low frequency IDS condition (1 to 3 active doses per week). Phase 3 will be an extension period to assess, over 9 months, the long-term durability of the approaches. The outcomes for the study will be: rate of relapse, latency to relapse, average sleep continuity, number and severity of medical symptoms function d...