Thursday, June 20, 2013

An example of chi-square analysis with SAS: "The association between drinking quantity and alcohol abuse or dependence among middle adult drinkers with and without major depression"



1. Introduction

Alcohol misuse by adults is an important public health concern with significant consequences such as the detriment of the drinker’s health, personal relationships, and social standing. On the other hand, small amounts of alcohol can provide some health benefits and a drink once in a while is a socially accepted custom. It follows that it is not easy to predict behaviors such as alcohol abuse or alcohol dependence except when the above serious consequences emerge. Furthermore, adults experiencing depression seem to be more likely to drink than those without depression. This may be because those with depression drink to soothe the painful condition caused by the depression itself or, alternately, because they are sensitive to alcohol abuse or dependence at low level of drinking quantity compared to individual without depression.

1.1 Research Questions


  1. Is drinking level associated with the experience of alcohol abuse or dependence. 
  2. Is the association between drinking and alcohol abuse or dependence similar for individuals with and without major depression.

2 Methods


2.1 Sample


  • Middle Adults (age 35 to 59) who reported daily or nearly daily drinked in the past year (n=1498) were drawn from the first wave of the National Epidemiologic Study of Alcohol and Related Conditions (NESARC).
  • NESARC is a nationally representative sample of non-institutionalized adults in the U.S.

2.2 Measures


  • Major depression was assessed using the NIAAA, Alcohol Use and Associated Disabilities Interview Schedule - DSM-IV (AUDADIS-IV)
  • A diagnosis of DSM-IV alcohol abuse requires that a person show a maladaptive pattern of alcohol use leading to clinically significant impairment or distress; it requires at least one of four specified abuse criteria.
  • A diagnosis of DSM-IV alcohol dependence requires that a person meet at least three of seven specified dependence criteria.
  • Current drinking was evaluated through quantity (“How many drinks did you USUALLY have on days when you drank during the last 12 months?”)

3 Results


3.1 Univariate


  • Daily, or nearly daily, middle adult drinkers, drinked an average of 3.3 drinks per day (s.d. 2.5)
  • A total of 28% of daily, or nearly daily, middle adult drinkers met criteria for DSM-IV alcohol abuse or dependence
  • A total of 20.2% met criteria for major depression at some point in their life










3.2 Bivariate

  • Chi-square analysis showed that among daily, or nearly daily, middle adult drinkers the number of drinks (categorical explanatory) is positively and significantly associated with past year DSM-IV alcohol abuse or dependence (categorical response). In fact, corresponding to 1 drink of any alcohol usually consumed (USQUAN=1) we have a total of 5.9% of daily, or nearly daily, middle adult drinkers meeting the criteria for alcohol abuse or dependence; for 2 drinks (USQUAN=2) we have a total of 16.8%; for 3 drinks (USQUAN=3) we have a total of 35.5%; for 8 drinks (USQUAN=8) we have a total of 57.8%. X2 = 256.6, 3 df, p < .0001. That is the more drinks daily, or nearly daily, drinkers drink, the more they are likely to be affected by alcohol abuse or dependence.
  • Moreover, Bonferroni method showed that among daily, or nearly daily, middle adult drinkers the number of drinks is positively and significantly associated with past year DSM-IV alcohol abuse or dependence for each pair of drinking quantity levels with an overall (or family-wise) 95% confidence level, i.e. with an overall type I error rate of 5% (6 comparisons , Bonferroni Adjustment = 0.05 / 6 = 0.008).
  • Daily, or nearly daily, adult drinkers with major depression were significantly more likely to meet the criteria for alcohol abuse or dependence (42%) than those without major depression (24.4%) X2= 37.1, 1 df, p < .0001.




 The following six comparisons are necessary to apply Bonferroni method (Bonferroni Adjustment = 0.05 / 6 = 0.008).














3.3 Moderation


Major depression did not moderate the association between drinking quantity and alcohol abuse or dependence among middle adult drinkers, i.e. drinking quantity is positively and significantly associated with alcohol abuse or dependence for those with major depression (corresponding to 1 drink of any alcohol usually consumed, i.e. USQUAN=1, we have a total of 14.9% of daily, or nearly daily, middle adult drinkers meeting the criteria for alcohol abuse or dependence; for USQUAN=2 have a total of 25.0%; for USQUAN=3 have a total of 50%; for USQUAN=8 have a total of 73.3%) and without major depression (for USQUAN=1 have a total of 3.5%; for USQUAN=2 have a total of 15.2%; for USQUAN=3 have a total of 31.9%; for USQUAN=8 have a total of 52.9%). Moreover, among daily, or nearly daily, middle adult drinkers at each level of drinking the probability of alcohol abuse or dependence was higher among those with major depression than those without major depression.










4 Discussion


4.1 What might the results mean?


  • Drinking quantity is positively and significantly associated with alcohol abuse or dependence among middle adult drinkers both in case they had a major depression at some point in their life and in case they had not.
  • Individuals with major depression seem to be more sensitive to alcohol abuse or dependence across a range of drinking quantity levels.

4.2 Strengths


Results are based on a large nationally representative sample of U.S. middle adults daily, or nearly daily, drinkers.

4.3 Limitations


  • The present findings are based on data collected by observing middle adults daily, or nearly daily, drinkers with and without major depression and they do not show at which drinking quantity levels or drinking frequencies alcohol abuse or dependence emerges, i.e. in case of middle adults drinking less than every day or nearly every day (e.g. 3 to 4 times a week, 2 times a week, etc.)
  • Moreover these findings are based on data collected observing middle adults daily, or nearly daily, drinkers in the past year and do not show drinking quantity levels or drinking frequencies of previous years.

4.4 Recommended Future Research


Further research is needed to determine whether current alcohol abuse or depression in middle adult drinkers may be correlated to high levels of drinking quantity or drinking frequency when such subjects were younger.

5 SAS Source code


                     page 1 / 2




 


                     page 2 / 2