To investigate the 20-year relationship between anxiety, depression and pain medication use. A total of 521 individuals reporting chronic pain from the National Survey of Midlife Development in the USA (MIDUS) study. Structural equation modeling of 20-year longitudinal survey data. Over 20 years, a bidirectional relationship between depression and anxiety in individuals with chronic pain was indicated. Pain medication utilization predicted later use at 10 years. Pain medication use was not strongly related to later anxiety; however, heightened anxiety was associated with later use. Depression and anxiety show an extensive long-term bidirectional relationship. While there was little indication of a relationship between pain medication use and later negative mood, anxiety was associated with subsequent pain medication use.