State variations of risky drinking among cancer survivors and individuals with no cancer history: NHIS 1997-2010

Background: Although moderate consumption of alcohol has been shown to impart some benefits to cancer survivors, risky drinking can be harmful and has not been studied in cancer survivors compared with individuals with no cancer history (non-survivors). We examined state-level prevalence of risky drinking among cancer survivors compared to non-survivors to help determine where limited resources could be utilized.

Methods: Pooled data from the 1997-2010 National Health Interview Survey were used to determine state-level prevalence of risky drinking among cancer survivors and non-survivors. Risky drinking was defined as ten or more drinks/week in men and seven or more drinks/week in women; or five or more drinks at one sitting, one or more times/year for both men and women. Two-tailed t-tests were used to compare each state for prevalence of risky drinking stratified by cancer survivor status. Data were adjusted for survey weights and analysis was performed using SUDAAN v10.

Results: Our sample included 15,802 cancer survivors and 226,649 non-survivors. Overall, cancer survivors were less likely to be risky drinkers than non-survivors 28.4% (SE=0.5) vs. 37.7% (SE=0.2); p<0.001, respectively. However, the exception to this was found in Louisiana where there was a non-significantly higher percentage of risky drinking in cancer survivors than in non-survivors (34.0% [SE=2.5] vs. 31.1% [SE=2.1], respectively; p=0.76). For cancer survivors the highest prevalence of risky drinking was found in Utah (37.9%; SE=5.9) and Kentucky (37.8%; SE=3.2) and, for those without cancer, in Idaho (48.1; SE=3.1) and Utah (47.6; SE=1.5). The lowest prevalence was seen in Arkansas (19.5%; SE=4.9) and Nebraska (19.7%; SE=3.9) for cancer survivors and Delaware (24.5%; SE=2.1) and Rhode Island (26.5%; SE=2.3) for non-survivors.

Conclusion: Overall, the prevalence of risky drinking is exceedingly high, and much higher in non-survivors than in cancer survivors. As there are serious repercussions for risky drinking, whether a cancer survivor or not, targeted interventions to reduce this high risk behavior are warranted, particularly in Utah for everyone, Kentucky for cancer survivors, and Idaho and Iowa for non-survivors. Further research into cancer survivors in Louisiana is needed to determine why cancer survivors have higher risky drinking behaviors than non-survivors.