In a paper on habit mechanisms in smoking, Hunt and Rlatarazzo(3) noted that the resumption of smoking or the relapse rate over time for a group of people who had successfully completed a treatment program and given up smoking conforms to a negatively accelerated, typical "extinction" curve such as those presented in Fig. 1. Their original figure was based on some 17 studies. The curves in Fig. 1 are based on some 84 studies.The curve should be considered as illustrative, or typical, rather than mathematically exact, as not all the studies surveyed utilized the same time periods in reporting the follow-up information and many of them gave data only for a single point in time, which of course makes individual curve fitting impossible. We arbitrarily selected 3, 6, and 12 months and established each point by averaging all the studies that reported for any of the three times. We also have constructed individual curves for each of those studies that reported data over comparable periods, the average of these studies, and the average of all studies that reported for only one of these points, etc. All the curves are remarkably similar.They are marked by a steep decline during the first 3 months, a subsequent gradual levelling off, and by an asymptotic level well above zero. Originally, we had hoped to use the differences in slope between individual curves as a differential criterion to evaluate various treatment techniques, but the lack of uniformity in reported follow-up data, plus the common sources of contamination in the different techniques themselves as noted by Bernstein('), suggest that the field is not ready as yet for sophisticated evaluative methods.While we are not yet ready to comment on slope differences between treatment techniques, the universality of the curve characteristics permits two general comments, one practical and one theoretical. On the practical side, the high incidence of relapse or "backsliding" during the first 3 months would indicate that, even disregarding those failures who never do manage to stop smoking despite treatment, treatment methods as presently practiced make only a moderate impression on those who do manage to stop. Either the usual treatment period is too brief or the methods too inefficient to produce a lasting effect. Obviously, the majority of "patients" need some further supportive or booster treatment during the first 6 months after successful completion of therapy.More interesting to us are the theoretical implications of the curve in becoming asymptotic before it reaches zero. Why is it that roughly 20% of the Ss treated never return to smoking? If this is attributable to some personality and/or physiological characteristics, then this group of people certainly deserves further study. The steepness of the slope of the curve and its relatively sudden leveling off above zero suggest that they may be a discrete and easily identifiable group. We are more inclined, however, t o believe that bwo kinds of learning may be involved, one of which is susceptible to the usual r...

No abstract

No abstract

PROBLEMWe have been studying relapse rates after treatment for smoking, alcoholism, and heroin addiction(33* 34, 35, 36* 3 7 ) . Starting with a group of people who have completed treatment successfully and are abstemious a t its termination, we find that many relapse quickly into their previous pattern of nonabstemious behavior, which results in a negatively accelerated curve over time that breaks fairly sharply a t the 3-to &month point and asymptotes well above zero with a residue roughly of about 25% long-time abstainers. The relative uniformity of such curves across the different addictions studied has theoretical implications, but remembering that such curves are composites obtained by averaging many different treatment methods in many different studies, it is apparent that they also offer a source of evaluative comparison to assess the success of any individual treatment. As we have suggested(33+ 35) such evaluation would enable us to concentrate our developmental efforts oil those treatments that show above-average promise, e.g., standard deviation above the norm, and to neglect the less successful ones.The present study offers such an evaluation of methods to modify smoking behavior.