Page 73 - PhD GT

the elements of the alcohol dependence syndrome as described by Edwards and Gross (1976). Each element was operationalised into a minimum of one item in their scale. One of the items in the scale refers to amnesias, described after Chick (1980a) as being “not conceptually part of dependence” but included because it was thought it “might be a useful marker of tolerance and good at spreading the responses of the populations tested” (Raistrick et al. 1983 p. 93). Small samples were initially used to generate the correct language and terms for the scale which was taken through seven pilot stages (Raistrick et al. 1983 p. 90). The result is described as the scale which tests all elements of the alcohol dependence syndrome in self completion format (Stockwell et al. 1994; Davidson 1987). Following its development using three samples: a sample of hospital workers described as “regular drinkers” (Raistrick et al. 1983 p. 90), a psychiatric patients sample and a specialist clinical sample, and concurrent clinical ratings by three experienced practitioners, it was further validated in three separate studies. Concurrent validity was examined by comparison with other measures of related problems in one study, by comparison with the SADQ (Stockwell et al. 1979) in a second study and by comparison with an interview schedule, the Edinburgh Alcohol Dependence Scale (EADS) (Chick 1980a) in a third study (Davidson and Raistrick 1986). Factorial analysis of responses to the scale in three groups of patients at geographically separate locations (Leeds, Omagh and Derry) supported the view that the SADD was a unidimensional scale. One main factor accounting for 44% of the variance was found which represented all the items but one, that which referred to attempts to control drinking by complete abstinence (Davidson et al. 1989).
The value of the SADD was tested with reference to five properties stated at the outset as being desirable for scales to be used in a clinical population: i) that it was usable in such a population, ii) that it measured present state dependence, iii) that it was sensitive to the full range of dependence iv) that it was sensitive to change over time and v) that it was culturally neutral. Its construct validity suggested it was suitable for use in a help seeking population as did its brevity and the ease of its use. Tests of concurrent validity showed it capable of measuring present state dependence, a criterion useful for making clinical decisions. The authors pointed to increasing evidence for the ability of the dependence construct to predict moderation drinking at low levels of severity across a variety of measurement methods (Saunders and Kershaw 1979; Polich et al. 1981) and the introduction of the suitability of briefer methods of treatments for this population (Edwards et al. 1977b) as the basis for developing a scale which would discriminate well between all levels of severity. The SADD was developed for the purpose of assessing the whole range of dependence severity with greater emphasis than in other available scales on the lower levels of severity. Its sensitivity across the whole range of dependence was demonstrated in the reported validation
61