Smartphone-based ecological momentary assessment (EMA) of alcohol and cannabis use in older adults with and without HIV infection

Abstract

Introduction: Ecological momentary assessment (EMA) has been used to characterize substance use among adult populations; however, little is known about the validity of EMA and the patterns and predictors of substance use among older adults with and without HIV infection.

Methods: Thirty-five (22 HIV-positive, 13 HIV-negative) older adults aged 50-74 were assessed for 14days and completed up to four smartphone-based surveys per day.

Results: Participants completed an average of 89.5% of possible EMA surveys. EMA self-reported alcohol and cannabis use were significantly positively correlated with laboratory-assessed, self-reported days of alcohol (r=0.52, p=0.002) and cannabis (r=0.61, p<0.001) used and quantity of alcohol (r=0.42, p=0.013) and cannabis (r=0.41, p=0.016) used in the 30days prior to baseline assessment. In a subset of 15 alcohol or cannabis users, preliminary analyses of the effects of mood and pain on alcohol or cannabis use showed: 1) greater anxious mood predicted substance use at the next EMA survey (OR=1.737, p=0.023), 2) greater happiness predicted substance use later in the day (OR=1.383, p<0.001), and 3) higher pain level predicted substance use earlier in the day (OR=0.901, p=0.005).

Conclusions: Findings demonstrate that EMA-measured alcohol and cannabis use has convergent validity among older adults with and without HIV infection. Preliminary results showing predictors of substance use highlight the importance of gathering EMA data to examine daily variability and time-dependent antecedents of substance use among this population.

Keywords: Aging; Ambulatory assessment; HIV/AIDS; Proximal risk factors; Real-world setting; Substance use.

Figure 1

Associations between EMA self-reported substance use and retrospective self-reported substance use. Y-axes represent EMA self-reports and X-axes represent laboratory self-reports.

Figure 2

Observed proportions of substance use by time of day. Error bar denotes 95% confidence interval for the mean. Time of day: four EMA surveys per day. Periods 1–4 refer to the morning, midday, afternoon, and evening in a day, respectively.

Figure 3

The predicted proportion of substance use by time of day based on happiness rating. The figure shows that the higher happiness is associated with a greater increase of use throughout the day.

Figure 4

The predicted proportion of substance use by time of day based on pain levels. The figure shows that higher pain levels are associated with a higher odds of substance use in the morning that stay constant or decrease throughout the day, whereas lower pain levels are associated with lower odds of substance use in the morning that increase throughout the day.