Research That Matters (January 17 - 20, 2008)

Purpose: Problem alcohol use among older adults is a significant but poorly understood public health issue. Often described as an “invisible epidemic”, at-risk alcohol use is projected to increase in the future due to increased numbers of older adults and attitudinal differences regarding alcohol by the baby-boom generation. Alcohol misuse is particularly hazardous for older adults because of age related alcohol sensitivity and medication interactions. Alcohol misuse in elders is associated with cognitive problems, falls, and suicidality. Problem alcohol use often goes undetected in this population, due to DSM-IV criteria that may not apply well to aging individuals. NIAAA guidelines suggest that consumption patterns acceptable for younger populations are inappropriate for older adults. Our aim was to use items that reflect alcohol consumption as well as DSM-IV alcohol criteria to discern classes of problem users in older members of the general population. A second goal was to ascertain risk of psychiatric comorbidity and current health status in these subclasses. Methods: Latent Class Analysis (LCA) was used to characterize older (60+) past-year alcohol users (n=4,672) who participated in the National Epidemiological Survey of Alcohol and Related Conditions (NESARC). LCA is a technique that assigns a probable class membership based on patterns of symptoms (“response profiles”). It is considered a person-centered analysis, detecting homogeneous subgroups more subtly than simple diagnostic thresholds. Dichotomous indicators used in the LCA included 11 DSM-IV abuse/dependence criteria and NIAAA quantity/frequency guidelines for older adults. Multinomial models examined the association between latent class membership and past 12-month diagnosis of Major Depression, anxiety disorder, and personality disorder. Linear regression analyzed differences in norm-based physical health status. Results: Using statistical measures of model fit, a three class solution best fit the data. Individuals were classified as severe problem (SP) drinkers (n=25; <1%), moderate problem (MP) drinkers (n=243; 5%) or nonproblem (NP) drinkers (n=4403; 94%). SP's displayed the highest probability of endorsing abuse/dependence criteria, while MP's displayed low probability of endorsing criteria with the exception of the dependence criteria of desire to cut down/efforts to quit and the abuse criteria of hazardous use. MP's were characterized by higher probability of binge use in the past year compared to SP's but lower endorsement probabilities for DSM-IV alcohol criteria. Compared with NP's, SP's had higher odds of Major Depression (OR=10.51; p<.001) and anxiety disorders (OR=3.80; p<.05) when adjusting for sociodemographic differences (race, gender, marital status, income, education and age). MP's showed greater risk of Major Depression (OR=1.92; p<.05) and personality disorders (OR=1.86; p<.05). Adjusting for sociodemographic differences, SP's reported significantly poorer physical health compared to NP's (t=-2.85, p<.01) but there was no significant difference between MP's and NP's. Conclusions and Implications: Our findings reaffirm evidence of comorbidity between Major Depression and problem alcohol use in older adults. We also found that severe problem drinkers reported worsened health status compared with nonproblem drinkers, but for moderate problem drinkers, self-rated health status was not significantly different from nonproblem drinkers. Assessment of depressed older adults should include both screening of alcohol consumption and DSM-IV criteria.