As specified by the Family Smoking Prevention and Tobacco Control Act of 2009 (FSPTCA),1 the evaluation of a modified risk tobacco product (MRTP) with regard to the public health standard concerns, in part, an evaluation of the product with regard to its tendency to promote the following:
• Initiation and continuation of its regular use
• Switching to its use and cessation of the consumption of more harmful tobacco products (e.g., aid in cessation of use of conventional cigarettes)
• Dual use (use of the MRTP concurrent with continued use of an existing harmful form of tobacco use such as smoking conventional cigarettes)
• Relapse back to more harmful tobacco use (e.g., resume smoking conventional cigarettes after an extended period of abstinence)
All of these outcomes can be logically related to the reinforcing value of the MRTP (how rewarding it is).
1 Family Smoking Prevention and Tobacco Control Act of 2009, Public Law 111-31, 123 Stat. 1776 (June 22, 2009)
The chief reason for testing reinforcement value in the laboratory setting is that measures yielded by such testing show a good correspondence to a product’s addiction potential in real-world use (Haney and Spealman, 2008). Specifically, drugs that have a positive subjective evaluation and are self-administered in laboratory tasks are ones that tend to be used and abused recreationally in real-world use (Comer et al., 2008; Haney, 2009).
The reinforcement value of an agent (e.g., a specific drug, such as nicotine) or a product (i.e., a drug[s] provided via a particular delivery system, such as smokeless tobacco or cigarettes) can be gauged through animal research; however, in the present situation, animal research on reinforcement value does not appear optimal. First, animal research is especially warranted when the product poses significant immediate health risks. However, to the extent that an MRTP has been adequately screened in preclinical work, it seems that the MRTP could be safely used in laboratory assessments of reinforcement value or self-administration (where toxic effects of possible prolonged dual use would not pertain). Second, because of the difficulty in modeling certain kinds of delivery systems with particular tobacco products (e.g., snus), human research may present the most externally valid research option. Third, human research methods afford an array of research paradigms that should yield meaningful assessment of MRTP reinforcement potential. Finally, human research requires less extrapolation because of a lack of interspecies differences, which can be substantial in terms of nicotine reinforcement (Rogers et al., 2009).
Key Considerations for Reinforcement and Self-Administration Studies
Almost by definition, an addictive agent must support self-administration. Moreover, there is a long history of research that shows a rough correspondence between the reinforcement capacity of an agent in the laboratory setting and its abuse potential in real-world contexts (Comer et al., 2008; Haney, 2009). Reinforcement is generally defined as the capacity of an agent to sustain self-administration. Therefore, one meaningful step in assessing the ability of an MRTP to support self-administration in real-world contexts is to determine whether it supports self-administration in laboratory or controlled settings.
Evaluating reinforcement is complicated by several factors, one of which is a continuum of reinforcement potency. Therefore, methods must capture the reinforcement potential of a product relative to other products or agents to provide meaningful comparisons. In theory, a desirable MRTP should be somewhat more reinforcing than nicotine replacement therapies (NRTs), but perhaps less reinforcing than conventional cigarettes (at least among current smokers who have demonstrated considerable
susceptibility to cigarette reinforcement). The relative value of products will be affected by the dose of product tested. Doses may reflect what is considered a meaningful dose in terms of real-world use, they may be based upon brief ad libitum use, or they may be established via dose banding methods. Ideally, an MRTP would be sufficiently reinforcing so as to attract smokers away from conventional cigarettes but not encourage the widespread dependent use of the product by individuals who were previously nonusers or who would have quit smoking. NRTs represent a meaningful lower bound of reinforcement magnitude because they tend not to support addictive or dependent use (Shiffman et al., 2008a). Further, there appear to be interactions between specific products and individuals, such that individuals differ in terms of the hierarchy of reinforcement potential across products (Perkins, 2009). The determinants of such individual differences in product-relative reinforcement are unknown but no doubt reflect multiple influences such as prior experience (because reinforcement changes with exposure), genetic factors, and social influences. Thus, the level of reinforcement value may lie more in the type of research participant than in the type of product.
Reinforcement and Self-Administration Methods
Likelihood of initiation, as well as maintenance or persistence of use, can be studied across multiple types of studies ranging from laboratory studies, to randomized controlled trials (RCTs), to population-based cohort studies. Different methodological principles and standards apply to each type of study. As in all research, research methods are determined in part by the question(s) being addressed. In the case of the evaluation of an MRTP, the core questions in this area involve the extent to which the product will attract and support heavy self-administration and abusive use. Several relevant experimental contexts can be used in the effort to determine the self-administration and use or abuse potential of an MRTP:
1. Subjective evaluation of the product both initially, and with repeated exposure or use in laboratory contexts relative to appropriate comparison products
2. Acute self-administration in laboratory contexts (only reflecting use within laboratory sessions), relative to appropriate comparison products
3. Use in extended residence facilities
4. Natural environment contexts where long-term use can be studied in real-world contexts, via
a. long-term use in RCTs,
b. cross-sectional survey studies, and
c. longitudinal cohort studies.
Additionally, methodological approaches must be tailored to each research context. Unless otherwise specified, these considerations apply to both acute laboratory and residential stay experiments.
Size and Nature of the Sample
Recruited participants must permit appropriate inferences regarding the populations and questions to be addressed. No standard sample size can be specified confidently for the studies described in this section. Each study must be powered consistent with the study questions posed and the comparison products used. Some guidance on power might be gleaned from studies in which high- and low-preference products are evaluated (e.g., conventional cigarettes and NRT products [Johnson et al., 2004; Perkins et al., 2004a, 2009]). Clearly the nature of the sample will differ with regard to the particular research question posed.
Relative Reinforcement Value in Regular Smokers One question of key importance is the extent to which an MRTP is reinforcing among current heavy smokers. This would be relevant to the extent to which the product would be used heavily enough by smokers to serve as a cessation aid or a long-term substitute with regard to smoking conventional cigarettes. A very high reinforcement value in smokers of conventional cigarettes would suggest the product could serve as a cessation aid or long-term substitute to conventional cigarettes and could also present a meaningful risk of initiation of use among nonsmokers or ex-smokers. The use of a population of current smokers has the advantage of ensuring that the tested population is sensitive to nicotine reward (Carter and Griffiths, 2009). If current smokers are used, the researcher should ensure that the research participants have no strong desire to quit, so the findings relate to smoking behavior in regular smokers and not quitting behavior (Perkins et al., 1997). About 45 percent of current smokers attempt to quit each year (CDC, 2009), and including such smokers in the sample might not only produce greater within-cell error, but also distort outcomes systematically. Such smokers, for instance, might be especially willing to self-administer a perceived safer alternative to smoking conventional cigarettes and more likely to try to avoid smoking conventional cigarettes. Thus, their self-administration data might not validly reflect the actual reinforcement value of the product.
Relative Reinforcement Value in Nonsmokers, Ex-Smokers, and Adolescents Testing an MRTP among nonsmokers would provide some evidence of attractiveness and reinforcement potential in people who are
essentially nicotine naïve.2 If multiple sessions are used, the research could yield some evidence on how much drug experience might be needed to show an increase in reward value. To increase the likelihood that the tested population comprises at-risk individuals, some selection factors could be used such as high levels of impulsivity, extreme delay discounting (Bickel et al., 2010), use of other abused drugs, or risk haplotypes for tobacco dependence (Weiss et al., 2008). Also, because there may be a relation between age and reaction to nicotine and vulnerability to dependence (Weiss et al., 2008), it may be important to use relatively young individuals in such research. Adolescents might be optimal, but research methods and oversight would have to be appropriate for their participation. Adolescents who have experimented with smoking might constitute a particularly high-risk population with high public health significance. Finally, the use of ex-smokers would suggest the potential reinforcement value of MRTP use in this population, which has demonstrated sensitivity to nicotine reinforcement. Of course, inclusion of ex-smokers would require a careful assessment of the risks and benefits of participation.
In addition, because reinforcement from nicotine or tobacco can vary with gender, age, tobacco experience, and other factors, the researchers should ensure that such dimensions are appropriately represented or controlled (e.g., used for blocking or as exclusion criteria) in the sample to the extent that it is compatible with the question addressed.
Characterization of the Sample
A comprehensive characterization of the sample is important because it defines the population to which the conclusions may be most directly related. It also permits tests of the interaction of person factors with MRTP liking or use—factors that appear to modulate product reinforcing value. Variables that may be important to measure, based on prior research on tobacco reinforcement, are gender, age, ethnicity, educational and socioeconomic status, tobacco and nicotine use history (including peak tobacco use levels, prior quitting history, age of initial use, and use histories of different tobacco and nicotine products), expectations about the effects of the products to be tested, tobacco or nicotine dependence, blood or breath levels of tobacco or nicotine exposure, health and mental health status and history, and use of psychoactive products including psychiatric medications. These variables are important because they have been related to nicotine dependence, tobacco self-administration, and ability to control tobacco use.
2 Non-tobacco users are defined as those who have never smoked more than 10 cigarettes and who have never used any other form of tobacco
In terms of tobacco dependence assessment, the Fagerstrom Test for Nicotine Dependence or one of the new multifactorial dependence assessments (the Nicotine Dependence Syndrome Scale [Shiffman and Sayette, 2005; Shiffman et al., 2004] or the Wisconsin Inventory of Smoking Dependence Motives [Smith et al., 2010]) appear to provide more accurate appraisal of dependence than do the Diagnostic and Statistical Manual of Mental Disorders criteria (Hughes et al., 2011). In addition, researchers should ensure that the dependence instrument used is one that is appropriate to the population in question. For instance, there is concern that some dependence instruments may not be appropriate for young or light smokers, so researchers should use an instrument validated with such populations (Colby et al., 2000).
Standardization of Pre-Session Experiences
Investigators should ensure that research participants have similar experiences prior to experimental sessions. Standard durations of abstinence from, or controlled use of, nicotine, caffeine, and other psychoactive agents or products before sessions is needed so subjects enter sessions at similar motivational states. Deprivation tends to significantly increase motivation to use tobacco and its self-administration (Fant et al., 1995; Perkins et al., 1994a; Zinser et al., 1999). Studies designed to test maximal motivation would impose a period of deprivation, such as overnight deprivation, which could be tested with a carbon monoxide (CO) test in the case of deprivation of combustible products. Another approach would be to impose a modest but standard level of deprivation (e.g., 1-2 hours) to model a motivational state that would typically occur throughout the day. The most comprehensive approach to assessing self-administration would be to test products across a variety of deprivation levels. Deprivation prior to clinical studies may add complications in data interpretation. An alternative, although more costly and time consuming, is observation of ad libitum self-administration so that the response measured reflects real use.
It is probably not a concern if subjects take their normal prescription medication, including psychiatric medication, on the days of sessions or measurement. This is because the main outcome data will be relative preference for, or use of, the tested products, and this presumably will not be differentially affected by chronic use of psychiatric medications.
It is important that subjects have similar expectations about the experiment and what it entails (e.g., the nature of the tested products) unless manipulation of expectations is an explicit element of the study design (because expectations can significantly affect response to a tobacco product [Perkins et al., 2010]). One possible strategy is to provide subjects with considerable superfluous information, which may reduce disparities in expectations
(Griffiths et al., 2003). Finally, to the extent that measures are complex (e.g., with certain types of cognitive performance tasks) it is important that practice effects be reduced by pre-session task familiarization.
Reinforcement and Self-Administration Measures
Biochemical measures of tobacco or nicotine exposure are important because they reflect prior self-administration intensity or tolerance, and therefore they should serve as useful covariates for laboratory-based self-administration. The appropriate measure could be CO level for cigarette smokers, or nicotine or cotinine levels (from blood, saliva, or urine) in other types of nicotine or tobacco users (those using smokeless tobacco or NRT). In particular, acute blood nicotine absorption profiles in response to both single and repeated use of products is a relevant component in assessing the addictive potential of MRTPs. Cotinine might be preferred to CO and nicotine because of its longer half-life. This could be extremely useful if long-term abstinence is imposed prior to experimental sessions or if subjects have engaged in only infrequent use of a nicotine or tobacco product. Also, if a noncombustible MRTP is studied, CO levels during or after the experiment will not provide measures of effective dosing. Therefore, to obtain a true baseline for such later measures, either nicotine or cotinine should be measured at baseline. In deciding between assessing cotinine versus nicotine, if the intent is to study effective self-dosing acutely (over minutes or 1-3 hours), then nicotine is the measure of choice, while cotinine would be the measure of choice if the effects of dosing over an extended time period (many hours or days) are targeted. The best predictor of plasma cotinine may be measurement of urine cotinine corrected for creatinine concentration (Benowitz et al., 2009). Finally, the investigator might wish to measure both nicotine and 3-hydroxy-cotinine in order to estimate nicotine metabolism (Schnoll et al., 2009). However, cotinine may be a poor choice for dual-use studies because it can reflect nicotine from multiple sources.
Selection of a biochemical assay depends upon the particular experiment, the questions posed, and the nature of the product. If relatively sensitive determination of nicotine receipt is sought, then it would be necessary to measure venous or arterial nicotine levels (typically via a venous catheter) and to obtain multiple measures over time to determine boost peak (peak baseline level) and area under the curve (see Benowitz  for calculation).
Imaging methods such as positron emission tomography or functional magnetic resonance imaging could be used to further characterize
the addiction potential of MRTPs. There is increasing evidence that particular neurotransmitter systems and associated brain regions are critically involved in the motivational processing of nicotine cues and nicotine anticipation: e.g., the dorsal striatum, nucleus accumbens, and anterior cingulate cortex (Gloria et al., 2009; McClernon et al., 2009). Therefore, amongst experienced MRTP users, MRTP cues or anticipation of MRTP delivery would be expected to activate such brain regions. However, at present there is little evidence that such measures possess the sensitivity to yield accurate rank-orderings of the addictive potential of different products or delivery systems.
Nature of the Comparison Stimuli
The selection of products or stimuli to be compared should be determined by the goals of the experiment and the need to obtain a sufficient number of comparators to permit an informative interpretive context. However, as discussed elsewhere, it seems that use of both conventional cigarettes (when smokers or ex-smokers are used as subjects) and NRT would be informative, because these represent products with very high versus modest reinforcement value. The study by Kotlyar et al. (2007) reveals how MRTPs can be meaningfully compared with NRTs on the basis of subjective evaluation and effective nicotine delivery.
It may be important to compare the product with nonpharmacologic stimuli as a means of providing a generally meaningful anchor point for the comparison of the pharmacologic products (including the MRTP). For instance, nicotine or tobacco products might be compared with pictorial stimuli (e.g., the International Affective Picture System), attractive music, compounds that stimulate taste buds, or money (Perkins et al., 1997). It is especially important to use nonpharmacologic stimuli as comparison stimuli (e.g., money) when using nonsmokers as subjects because it would be important to compare the MRTP with a stimulus of meaningful reinforcing value.
If the study is using current smokers as subjects, it would be informative to use the subject’s own or preferred brand of cigarettes, as this could represent an optimally reinforcing product against which to compare the MRTP. However, another strategy would be to use cigarettes with a range of known nicotine contents, which would provide a range of reinforcement value against which the MRTP could be compared.
One standard method of evaluating reinforcement value is to use an operant self-administration paradigm in which some sort of instrumental
response (key presses for instance) is executed to “earn” doses of the product. How hard an individual is willing to work for a dose is related to the addictive potential of the product. For example, the subject might be given the opportunity to earn either puffs of a conventional cigarette, inhalations from a nicotine inhaler, or doses from an MRTP. Such operant paradigms permit collection of many different sorts of measures, such as: (1) response rates including peak response rates for each type of product; (2) relative response rates on concurrent schedules (Perkins et al., 1997); and (3) demand elasticity for each type of product (the extent to which responding is affected by increasing the response requirement or dose). The last index may be especially useful because it permits meaningful interproduct (or interstimulus) comparisons on the basis of demand curves (Johnson and Bickel, 2006), in essence permitting more direct inferences regarding reinforcement magnitude.
Timing and Exposure Parameters
Experiments aimed at characterizing reinforcement value could present MRTPs and other products in diverse ways. The mode of presentation should be dictated by the experimental paradigm used, as well as the research question. In acute dose-effect comparison studies conducted in laboratories settings, presentation of discrete doses of products or stimuli should be counterbalanced, controlling for amount and order of delivery. In self-administration studies or behavioral economic studies, the researcher could use progressive ratio schedules in separate sessions for each product or concurrent schedules (e.g., comparing each product with monetary payment) or could test products individually across different response requirements to construct demand curves. In either acute dose-effect studies or self-administration studies, relatively standard doses with cigarettes can be achieved either with puff duration signals or with devices that control puff volume mechanically (Perkins et al., 1997). Timing signals might be the best way to manage dose parameters with products such as smokeless tobacco or NRT (Shiffman et al., 2003).
There are many things to consider in setting up and interpreting such experiments. One concern is how much experience or exposure to permit in the experiment. There is certainly evidence that preference or reinforcing value changes over exposure. This could occur because of tolerance to aversive effects, sensitization, familiarity (learning how to self-dose), development of dependence, and so forth. Thus, the researcher must structure the study so the person’s experience prior to the study and the exposure during the study are designed to match the experimental goals. An important principle, however, is that the best estimate of the ultimate reinforcement potential of an MRTP may be obtained after fairly extensive use.
Another concern is the interdose interval and amount of exposure (dose) to the products. Different delivery systems may deliver different doses of nicotine and doses with different pharmacodynamics. The investigator must consider whether standard dosing or exposure parameters do not “unlevel” the playing field for the various products (e.g., creating toxic effects or different levels of withdrawal for one product versus another). Investigators may also want to mimic extreme use, because some users may overuse the product. Interdose intervals should be determined based on the anticipated pharmacodynamics of the tested products.
Because the ordering of stimuli or products might affect the response (such as when an earlier product might satiate the subject, thereby reducing his or her motivation to self-administer additional nicotine), it is especially important to counterbalance stimulus presentations in acute dose-effect comparison studies so order effects are not inextricably confounded with stimulus effects. In essence, great care must be taken to ensure that exposures to products relatively late in the exposure sequence are meaningful. To the extent that earlier exposures result in high nicotine levels, reduced withdrawal, or priming effects, the subject’s motivational state is altered and therefore the subject’s responses are not similarly meaningful across the sequence. One strategy that could be used to address this is to have subjects “earn” dosings during a session but not consume them until after the session (Perkins et al., 1997). This may not be appropriate where delay would distort the motivational value of exposure. There is evidence that immediate versus delayed access to addictive agents or products makes a substantial difference in motivational and evaluative response (Gloria et al., 2009; Sayette and Hufford, 1994).
Another concern with timing of the experimental sessions is to ensure the anticipated end of the experimental session does not bias subjects’ responses. For instance, if one of the measures of product evaluation is instrumental to secure a dose of the product or amount of money needed to purchase a dose of the product from the subject, these measures could be distorted if the subject knows that he or she will shortly be released from the session and have ready access to nicotine or tobacco. Therefore, a postsession waiting period (which might range from 30 to 90 minutes) is often imposed so the only prospect of imminent tobacco receipt is that which will occur in the session (Perkins et al., 1999).
Additionally, with some procedures such as instrumental self-administration (behavioral economic strategies) or with unusual controlled dosing procedures, it may be desirable to allow the subjects some practice with the procedure so learning or familiarization effects are not confounded with changes in reinforcement value that develop with drug use experience.
In most self-administration experiments it would probably be important to determine the efficiency of self-administration, meaning the relation between self-administration and effective drug delivery of doses consumed (measured by biochemical indices of product receipt, such as CO and nicotine). This would allow one to distinguish gross self-administration behaviors from effective drug delivery. This distinction, for instance, might be relevant to questions about whether compensation occurs because of use of an MRTP. For instance, use of an MRTP might decrease the number of conventional cigarettes that a person smokes. However, this does not necessarily mean the person is actually exposed to less smoke or takes in less nicotine (Benowitz et al.  provides a compensation determination formula for cigarettes with known machine determined yields). Multiple measures are available to assess self-administration behavior so as to capture effective delivery more accurately (Rose et al., 2003; Strasser et al., 2007). This could be done by the use of especially sensitive assessments of self-administration. One example of this is the use of smoking topography measures that permit assessment of puff duration, inhalation force, and so on via force or flow transducers (Strasser et al., 2009). Video cameras and monitors have also been used to assess puff number and duration (Benowitz et al., 2006).
Finally, one could indirectly infer the effective dose by repeatedly measuring physiological responses that are acutely sensitive to nicotine dose and rise-time effects (e.g., nicotine-induced tachycardia or skin temperature effects [Benowitz et al., 2006; Perkins et al., 1994b]) and deriving peak and area under the curve indices.
Reinforcement and Self-Administration Study Designs
Acute Dose-Effect Comparison Studies
This approach has been labeled as a standard with regard to human abuse liability drug testing, because of the correspondence between subjective ratings of drug effects and real-world abuse potential (Carter and Griffiths, 2009). This sort of research is faster and more economical to conduct than human self-administration studies. In this research, appropriate subject groups are given discrete agent or product exposures and asked to rate them on validated scales. These are generally placebo-controlled, blinded, within-subject crossover designs. However, the apparent differences among some tobacco products (snus versus conventional cigarettes or e-cigarettes) may compromise the ability to achieve true placebo control or blinding. Each product, though, could have a placebo preparation, which should control for some expectancy effects. Ideally, subjects should be allowed to rate a variety of dose levels or exposures to the products
to obtain a more comprehensive index of product effects. In addition, it would be important in at least a subset of studies to test at multiple intervals postexposure to ensure the pharmacodynamics of response are characterized for each product. This is important in part because pharmacodynamics may greatly affect reinforcement value and abuse potential (Dewit et al., 1993; Mumford et al., 1995). While acute dose-effect comparison studies are often conducted on closed or residential wards when using illicit drugs, this seems unnecessary for the type of research discussed because the tested products will not be significantly intoxicating, the product would not be a controlled substance, and biochemical and self-report means can be used to determine intersession use.
Measures for Use in Acute Dose-Effect Paradigms Certainly researchers would collect self-report measures of subjective responses to the MRTP and other products, either in anticipation of receipt of the product (after the subject has some familiarity with it) or following its effects. There are well-characterized scales that permit the assessment of a variety of relevant rating dimensions (e.g., the Duke Cigarette Evaluation Scale and the Duke Sensory Questionnaire [Benowitz et al., 2006; Rose et al., 1999; Westman et al., 1996]; also cf. [Kotlyar et al., 2007]), including physical and affective reactions to the rated products (Benowitz et al., 2006). There is substantial evidence attesting to the validity of such self-report assessments. For instance, similar items have been shown to be sensitive to degree of drug deprivation (Carter and Tiffany, 1999; Sayette et al., 2003; Zinser et al., 1999) and have been shown to be sensitive to the actual nicotine content of cigarettes (Benowitz et al., 2006; Rose et al., 1999). However, they are not consistently strongly related to actual self-administration (Hughes et al., 1996; Perkins et al., 1997), leading to suggestions that self-administration and subjective ratings capture different facets of reinforcement value.
The short form of the Addiction Research Center Inventory is a self-report measure that has been used most extensively to index subjective reactions to nonnicotine drug effects (Jasinski, 1977). This measure contains the Morphine-Benzedrine Group scale, which purportedly measures euphoria (Bigelow, 1991; Foltin and Fischman, 1991; Jasinski, 1977). While this scale appears to reflect subjective evaluations of multiple drugs of abuse, it is unclear at present whether it is ideal for measuring nicotine reinforcement.
Other measures could be incorporated into acute dose-effect comparison studies. For instance, 24-hour retrospective recall of reinforcement would reveal the extent to which postexposure processing alters the memorial representation of incentive properties (Carter and Griffiths, 2009). It is important to use exactly the same questions in those recall tests
as used in earlier tests to ensure that change is due to passing of time, not altered assessment formats. Also, the multiple-choice procedure can be used to monetize the worth of additional product doses or exposures at the end of sessions to provide additional data on reinforcement value (Griffiths et al., 1993). A study by Hatsukami et al. (2011) shows how subjective evaluation measures can be paired with tobacco product use measures and product choice measures to enhance the validity of the subjective evaluation measures.
Behavioral Economic Self-Administration Studies
When addictive agents are self-administered in the laboratory context, there is a meaningful relation between laboratory assessed self-administration on the one hand, and clinical evidence of dependence and drug motivation on the other hand (Bickel and Madden, 1999a; Madden and Bickel, 1999; Perkins et al., 2004b; Piasecki et al., 2010). If a contingency is established between the receipt of an agent or product on the one hand, and execution of an instrumental response (e.g., access to MRTP dosing and pressing a lever) on the other hand, then instrumental responses for the agent or product would constitute a key indication of reinforcement potency.
In an acute laboratory setting subjects could work for products across several different contexts: under differing levels of tobacco withdrawal, with different response requirements, and using different instrumental paradigms (progressive ratio schedules for individual products, concurrent schedules for the MRTP versus conventional cigarettes and/or money; and with varying response requirements to generate demand curves). Product exposures could be controlled with smoking topography equipment for cigarettes, while the investigator might have to rely upon duration of use (e.g., duration of oral exposure to smokeless tobacco) and number of self-administrations (e.g., nicotine nasal spray, gum) for noncigarette products (Perkins et al., 2004b; Shiffman et al., 2003). Effective exposure could be indexed by biochemical indices for all products.
Measures Gathered in Behavioral Economic Self-Administration Studies The key measure would certainly be counts of the instrumental response, but it could also include biochemical measures of nicotine or smoke exposure, subjective product evaluations, and withdrawal symptoms. That is, the study could not only assess self-administration under various conditions, but also gather data on perceived reinforcement value and the ability of the product to alleviate withdrawal symptoms (combining the goals of both acute dose-effect studies and behavioral economic studies).
Other self-administration studies could be conducted that do not rely upon instrumental self-administration methods. For instance, there is substantial evidence that tobacco withdrawal plays a major role in spurring relapse back to tobacco use, which may occur because smokers try to escape aversive withdrawal symptoms or because withdrawal enhances the incentive value of smoking cues (Baker et al., 2004a, 2004b; Piasecki et al., 2003). Therefore, researchers might explore the extent to which the MRTP, used either ad libitum or under controlled dosing, can ameliorate withdrawal symptoms caused by discontinuation of smoking of conventional cigarettes. Acceptable methods for such studies have been well developed (Shiffman et al., 2003; Welsch et al., 1999). In such research, heavy smokers of conventional cigarettes could be withdrawn from tobacco for an extended period of time and then permitted to use an MRTP. A well-characterized withdrawal scale (Hughes and Hatsukami, 1986; Hughes et al., 1991; Welsch et al., 1999) could be used to measure the extent to which use of the MRTP versus a placebo or other comparison product (e.g., NRT) reduces withdrawal. Such data would be relevant to the notion that an MRTP could substitute for conventional cigarettes and thereby perhaps reduce their use.
In addition to measures of hedonic and evaluative responses, researchers might also gather measures of product impact on other measures such as cognitive performance (attention, memory) and psychomotor performance. Some individuals may use nicotine to enhance their cognitive performance (Heishman et al., 2010; Kleykamp et al., 2011), and such measures could index this source of reinforcement, especially for selected populations such as persons with schizophrenia or attention deficit disorder. Such data would be relevant to the question of whether the MRTP might substitute for conventional tobacco products in such populations.
Finally, while human drug discrimination paradigms can be highly informative in the evaluation of new products (Carter and Griffiths, 2009), they seem less germane to the current questions of interest because the goal is not to compare different types of agents or drugs but instead to compare different nicotine delivery systems.
Analyses Analyses for most of the studies described in this section should be fairly straightforward. For instance, repeated measures of analyses of variance could be used to identify significant main effects associated with the various types of products or stimuli used, and product by repeated measures interactions could be used to determine if products differ in their patterns of change over repeated exposures. Moreover, analyses could be conducted with repeated exposures within sessions crossed with days (or sessions) in order to examine if changes
within sessions vary as a function of number of days of exposure or some feature of days (e.g., amount of deprivation preceding a day). Instead of analysis of variance, growth curve modeling (e.g., via hierarchical linear modeling) could be used to estimate intercepts and trajectories and to model days as second-level variables. Appropriate covariates might include gender, starting CO or nicotine level, and dependence. In addition, interaction terms could test whether effects differ significantly as a function of any special subpopulations (e.g., those high versus low in dependence). In all such analyses, the normal analytic considerations pertain such as examining and adjusting scores for distributional deviations, missingness, and autocorrelation.
The analysis of behavioral economic data presents special challenges. For some outcomes such as evaluation of demand elasticity, special formulas are required to model the relation between cost and response (Hursh and Silberberg, 2008; Murphy et al., 2011). Demand elasticity refers to the extent to which work for a substance (e.g., an MRTP or conventional tobacco product) is sensitive to price or work requirements to obtain the substance (e.g., the extent to which self-administration decreases with increased cost). Presumably, the more reinforcing a substance is, the less its self-administration is affected by increased cost. The determination of formal demand curves from self-administration data can be costly in terms of time and resources. Easier to implement strategies are available that may allow for more efficient determination of the relative reinforcement value of different substances: e.g., hypothetical purchase tasks (Murphy et al., 2009). In addition, measures such as peak-response rate and breakpoint are related to the economic measures of maximal output and elasticity of demand and could also be used (Bickel and Madden, 1999b).
A principal message of the research literature on drug reinforcement value is that no single approach to assessing reinforcement value provides a comprehensive index of value and that using a variety of approaches conveys superior information about relative reinforcing value of pharmacologic agents or products and factors that influence their value.
Therefore, an overarching observation is that a comprehensive assessment of product motivational value includes studies that examine reinforcement value in different relevant populations, with different paradigms, with multiple comparison stimuli and products, and with different types of outcome measures. Specifically, the comprehensive evaluation of the reinforcement value of an MRTP may examine reinforcement value as per the five categories described below.
1. Subject populations: Examination of reinforcement value in daily smokers of conventional cigarettes who range in level of tobacco dependence and in beginning smokers (especially young smokers) may be necessary. Other potentially useful populations would be daily smokers interested in cessation, smokeless tobacco users, and nonsmokers.
2. Experimental paradigm: Collection of data on subjective evaluations of the MRTP in acute dose-effect comparison studies and in behavioral economic self-administration studies testing over multiple days and extended sessions is necessary. Use of behavioral economic paradigms would permit more informative indices of interstimulus reinforcement value as it could be denominated on the basis of a standard behavioral response. Moreover, some self-administration paradigms may not only examine reactions to, and self-administration of, the MRTP relative to other products, but also examine the ability of the product to quell tobacco withdrawal (especially urges) and to reduce motivation to smoke conventional cigarettes because of preloading with the MRTP. Important in all of these paradigms is the modeling of change over repeated exposure occasions because this could reflect development of increased reinforcement value owing to tolerance to aversive effects or dependence development.
3. Comparison stimuli and products: Examination of subjects’ reactions to the MRTP relative to conventional cigarettes and acute forms of NRT (nicotine nasal spray, gum, lozenge, or inhaler) may be necessary. It would also be quite informative to conduct evaluations in which preference for each product could be monetized, at least via a multiple-choice procedure (Griffiths et al., 1993).
4. Outcome measures: It may be necessary to include measures of self-administration, biochemical indices of effective dosing, and self-report of preference and psychoactive effects. Other measures such as withdrawal severity may be used to explore effects such as withdrawal suppression; it may be efficient to also include putative biomarkers of disease risk (Hatsukami et al., 2006).
5. Data interpretation: This may be one of the most challenging aspects of the assessment of liability for adverse effects on the public’s health. There is no clear outcome that signals whether the MRTP has the “right” level of reinforcement potential in order to supplant smoking conventional cigarettes but yet not be so reinforcing that its availability poses additional significant threats to the public health. Presumably it will be more reinforcing than NRTs, because NRTs are not sufficiently reinforcing to support even prescribed levels of use (Lam et al., 2005; Liu et al., 2001; Shiffman et al., 2008b; Vogt et al., 2008). But, the MRTP presumably should not be as reinforcing as smoking conventional cigarettes. So, roughly speaking, an MRTP should be intermediate in reinforcement magnitude. Of course, decisions about optimal reinforcement magnitude depend on
other factors such as the product’s delivery of toxicants (a product that results in little toxicant exposure would present little risk even if being highly reinforcing) and the results of other research efforts (data from RCTs).
As noted in the Introduction, the evaluation of an MRTP with regard to the public health standard concerns such factors as (1) how heavily it is used, (2) the extent to which its use directly exposes individuals to toxicants, (3) the effect of its use on the consumption of conventional tobacco products, (4) how conjoint use of the MRTP plus conventional tobacco affects health, and (5) how its use affects the initiation of use of conventional tobacco products and relapse back to use of such products (e.g., resumption of smoking conventional cigarettes after an extended period of abstinence).
Some of these issues can be explored via RCTs. In particular, the RCT may be a highly efficient means of examining such related issues as (1) acceptability and use of the MRTP; (2) the ability of the MRTP to increase cessation in users of conventional tobacco products, either by enhancing total cessation or by reducing use of such products; and (3) the likelihood that MRTP availability will lead to dual use. An RCT could also, in theory, produce evidence on such topics as (1) whether and how much individuals use an MRTP after they have used it to help them quit use of conventional tobacco products, (2) changes in perception of the MRTP with its continued use, and (3) the MRTP’s ability to suppress tobacco withdrawal symptoms. The last effect would increase the likelihood that the MRTP would serve as an effective cessation aid.
Key Considerations for the Use of Randomized Clinical Trials
An RCT that tests the potential public health impact of an MRTP requires decisions about key issues that will affect the validity and relevance of the resulting data. One issue is the balancing of internal versus external validity. This issue has implications for multiple aspects of the experimental design and methods, such as how heavy the assessment burden should be, whether subjects are asked to pay for the MRTP at some point in the trial, and how to maintain subject involvement in the trial. Therefore, the challenge is to ensure the real-world relevance of the work, while maintaining enough internal validity (experimental control) so strong inferences can be made. Other major decisions concern the nature of the specific comparison products to be tested in the research (see
discussion below), whether and how to implement blinding procedures, the nature of the outcomes to be assessed, and the nature of the population to be recruited.
It is important to recognize that no single RCT can address all of the important issues that pertain to the possible public health impact of an MRTP. Therefore, it may be necessary to conduct two or more RCTs in order to address the major questions that exist. For instance, it would seem desirable for one RCT to emphasize internal validity, while another might be designed to emphasize external validity (real-world relevance). Moreover, it may be economical of time and other resources (burden and risk to the individuals who would participate in an RCT) that an RCT be launched only after there is some evidence from laboratory studies that the MRTP (1) has a significantly favorable toxicant profile, (2) is sufficiently reinforcing or nonaversive so as to permit a reasonable level of use by smokers, and (3) is not so reinforcing (or addictive) so as to lead to high levels of use by nonsmoking youth. The suggestion for prior laboratory studies is made despite the fact that clear-cut criteria do not presently exist that would allow definitive determinations with regards to the above issues. A key question for both laboratory studies and RCTs is how data or outcomes can be interpreted so as to have optimal meaning or relevance with regard to public health impact. That is, what patterns of use and effects (e.g., impact on smoking cessation) would suggest a net positive versus harmful effect?
Design and Overarching Methods Considerations
The trial design should reflect the questions targeted. If the major question is how MRTP availability affects future use of conventional tobacco products, the design should contrast a condition where some subjects are randomly assigned to use the MRTP and others a placebo or a comparison product (Robinson et al., 2000). For instance, a meaningful comparison condition would be the provision of acute administration NRT products that have strong patient acceptance and use (e.g., perhaps newer acute NRT products that show relatively high rates of patient acceptance). A highly acceptable and efficacious NRT would be a good benchmark for MRTP evaluation. Such products show modest levels of smoker acceptance and use, tend not to substitute effectively for conventional tobacco use (e.g., smoking) among individuals not making quit attempts (a large portion of the smoker population has not switched from conventional tobacco products to NRT as a form of long-term use), and pose little risk of addictive or dependent use. Presumably, if an MRTP has promise to attract individuals away from use of conventional tobacco products, it should be somewhat more reinforcing than NRT, promoting greater sustained
use and substituting for conventional tobacco use more effectively than NRT. The value of the use of an NRT as an MRTP comparison product is apparent in a study by Kotlyar et al. (2007). Other criteria could also be forwarded, such as withdrawal from an MRTP should not be as severe as that arising from withdrawal from conventional tobacco products. In addition, NRT makes a meaningful comparison because it is a potential marketplace competitor with the MRTP, meaning that most forms are widely available over the counter. Presumably an MRTP would achieve meaningful use only if it were more appealing than NRT. Thus, NRT would appear to be a more meaningful comparison product than a prescription cessation aid (e.g., varenicline, bupropion) because the latter aids would not be available competitors for chronic use. Although NRT would constitute a meaningful comparison product in an RCT, interpretation of MRTP effects, and estimation of its potential risks and benefits, would be a challenging task (see the “Inference” section below).
Therefore, a reasonable design would be one in which subjects are randomly assigned to either the MRTP or to the comparison product(s) (ideally both a placebo and NRT) with blocking on intention to quit or interest in quitting. Ideally, more than one RCT should be conducted, with trials constituting both efficacy and effectiveness trials. Thus, the former would recruit highly motivated subjects, be double blind, entail fairly heavy assessment with compensation for adherence, and include other features designed to reduce error and encourage high use of the tested products (sustained provision of free products). The effectiveness studies might recruit “all comers,” use open label product, use relatively brief nonburdensome assessments, and provide products in a manner that more closely resembles real-world use.
An alternative to such a traditional RCT design would be one in which multiple products were tested in full factorial or fractional factorial design (Collins et al., 2011). This would permit the simultaneous and efficient testing of multiple comparison products and also testing of the interactions among such products. In addition, a crossover design could be used in which participants alternate the tobacco products that they use over standard cycles of use (Hatsukami et al., 2009).
At least some of the RCTs should permit extended use of the MRTP. This is because the impact or acceptability of a product might change with time. For instance, there is evidence that nicotine nasal spray use increases when individuals learn to use it properly (Blondal et al., 1997; Fiore et al., 2008; Sutherland et al., 1992). In addition, over time, other factors might encourage changes in use (e.g., secular events such as tax increases on conventional tobacco products, development of dependence). Also, some patterns of use, such as dual use, might be transitional stages that ultimately convert to more stable use patterns. For instance, there is considerable
evidence that chronic conjoint use of an NRT while smoking increases subsequent smoking cessation attempts and success (Carpenter et al., 2004; Chan et al., 2011). Finally, relapse back to smoking occurs at meaningful levels even after a year of cigarette abstinence (Hawkins et al., 2010; Heffner et al., 2010; Hughes et al., 2008); it seems important, therefore, to study MRTP effects up to the point where significant relapse risk has passed. It is possible, in fact, that quitting with the use of an MRTP results in higher than normal relapse because the continued use of the product primes continued or resurgent motivation to resume conventional tobacco use (Shaham et al., 1996, 2003). Any duration recommended for an MRTP RCT would be somewhat arbitrary. But, because the rate of relapse tends to drop to between 2 percent and 4 percent per year after 2 years of abstinence (Krall et al., 2002), a minimum 2-year duration seems advisable. This would suggest that observed cessation rates for conventional tobacco products observed at study end would be fairly stable.
Another important concern is whether the study involves an explicit quit date for those expressing interest in quitting. Setting a quit date for all subjects to make a cessation attempt would probably constitute the most sensitive test of the ability of an MRTP to boost cessation success in a given attempt. If the goal of the RCT is more focused on internal than external validity, and where subjects are motivated to make quit attempts, the investigator could encourage subjects to select a quit date so assessments could be concentrated around this date. This would increase the sensitive measurement of important factors such as quitting-related withdrawal symptoms. However, this design feature would probably not resemble real-world MRTP use, where many people might use an MRTP without intending (at least initially) to make a cessation attempt. Therefore, designs that permit long-term use without formal quit attempts and with individuals not motivated to quit would possess greater external validity. Such designs should certainly be used in at least one or more of the RCTs.
The Consolidated Standards of Reporting Trials (CONSORT) reporting recommendations for clinical trials emphasize the importance of explicitly identifying primary and secondary outcomes on an a priori basis. (The CONSORT 2010 checklist is presented in Table 4-1.) Primary outcomes should be few in number and explicit. It seems that a primary outcome should be percentage of smokers of conventional cigarettes (or another conventional tobacco product, depending on the goal of the study) who are abstinent at critical endpoints (e.g., 1 and 2 years post-study initiation). Other outcomes could include percentage of smokers who engage in dual use, amount of smoking of conventional cigarettes by those who engage in dual use, use rates and use prevalence of the MRTP, attitudes and perceptions of the MRTP (in particular, perceptions of relative
|Checklist Item||Reported on Page No.|
|Title and abstract|
|1a||Identification as a randomized trial in the title||___________|
|1b||Structured summary of trial design, methods, results, and conclusions (for specific guidance, see CONSORT for abstracts)||___________|
|Background and objectives||2a||Scientific background and explanation of rationale||___________|
|2b||Specific objectives or hypotheses||___________|
|Trial design||3a||Description of trial design (such as parallel, factorial) including allocation ratio||___________|
|3b||Important changes to methods after trial commencement (such as eligibility criteria), with reasons||___________|
|Participants||4a||Eligibility criteria for participants||___________|
|4b||Settings and locations where the data were collected||___________|
|Interventions||5||The interventions for each group with sufficient details to allow replication, including how and when they were actually administered||___________|
|Outcomes||6a||Completely defined prespecified primary and secondary outcome measures, including how and when they were assessed||___________|
|6b||Any changes to trial outcomes after the trial commenced, with reasons||___________|
|Checklist Item||Reported on Page No.|
|Sample size||7a||How sample size was determined||___________|
|7b||When applicable, explanation of any interim analyses and stopping guidelines||___________|
|8a||Method used to generate the random allocation sequence||___________|
|8b||Type of randomization; details of any restriction (such as blocking and block size)||___________|
|9||Mechanism used to implement the random allocation sequence (such as sequentially numbered containers), describing any steps taken to conceal the sequence until interventions were assigned||___________|
|Implementation||10||Who generated the random allocation sequence, who enrolled participants, and who assigned participants to interventions||___________|
|Blinding||11a||If done, who was blinded after assignment to interventions (for example, participants, care providers, those assessing outcomes) and how||___________|
|11b||If relevant, description of the similarity of interventions||___________|
|Statistical methods||12a||Statistical methods used to compare groups for primary and secondary outcomes||___________|
|12b||Methods for additional analyses, such as subgroup analyses and adjusted analyses||___________|
|Participant flow (a diagram is strongly recommended)||13a||For each group, the numbers of participants who were randomly assigned, received intended treatment, and were analyzed for the primary outcome||___________|
|13b||For each group, losses and exclusions after randomization, together with reasons||___________|
|Recruitment||14a||Dates defining the periods of recruitment and follow-up||___________|
|14b||Why the trial ended or was stopped||___________|
|Baseline data||15||A table showing baseline demographic and clinical characteristics for each group||___________|
|Numbers analyzed||16||For each group, number of participants (denominator) included in each analysis and whether the analysis was by original assigned groups||___________|
|Outcomes and estimation||17a||For each primary and secondary outcome, results for each group, and the estimated effect size and its precision (such as 95% confidence interval)||___________|
|17b||For binary outcomes, presentation of both absolute and relative effect sizes is recommended||___________|
|Ancillary analyses||18||Results of any other analyses performed, including subgroup analyses and adjusted analyses, distinguishing prespecified from exploratory||___________|
|Harms||19||All important harms or unintended effects in each group (for specific guidance see CONSORT for harms)||___________|
|Checklist Item||Reported on Page No.|
|Limitations||20||Trial limitations, addressing sources of potential bias, imprecision, and, if relevant, multiplicity of analyses||___________|
|Generalizability||21||Generalizability (external validity, applicability) of the trial findings||___________|
|Interpretation||22||Interpretation consistent with results, balancing benefits and harms, and considering other relevant evidence||___________|
|Registration||23||Registration number and name of trial registry||___________|
|Protocol||24||Where the full trial protocol can be accessed, if available||___________|
|Funding||25||Sources of funding and other support (such as supply of drugs), role of funders||___________|
health risks, addictiveness, liking of the MRTP, and value in curbing use of conventional tobacco products), motivation and plans to quit smoking among those continuing to do so, self-efficacy estimates of ability to quit with and without the MRTP, severity of the withdrawal syndrome in any quit attempts, incidence of quit attempts, nicotine dependence, and quitting self-efficacy. If the study is a postmarketing study, investigators could also inquire about MRTP use in the subjects’ social networks.
Finally, based upon the results of basic research on toxicant exposure or other sources, it might be warranted to include physical health assessments of the participants to determine if MRTP use is associated with changes in toxicants or with other biomarkers of relevant disease processes (e.g., pulmonary function tests).
Randomized Clinical Trial Design and Methods
Nature of the Sample
Perhaps the major question that exists is whether the product will help participants quit use of conventional tobacco products, either resulting in complete abstinence from any tobacco product, including the MRTP, or by the MRTP serving as a long-term substitute. Therefore, chronic smokers of conventional cigarettes (the most harmful conventional tobacco product) are an important target population for an RCT. Today there are more questions than in the past about what constitutes a current smoker because today’s smokers are smoking significantly less than in the recent past (CDC, 2005; CDC and National Center for Health Statistics, 2008; Pierce et al., 2011).
Because the results of an RCT should be broadly applicable to today’s smokers, the sample should comprise smokers who smoke relatively little (e.g., daily smokers who smoke at least two cigarettes/day) and very heavily (e.g., with no upper limit on daily smoking). The participation of light smokers would be dependent, of course, on the determination that their participation did not pose an unacceptable health risk (e.g., from nicotine toxicity). The participation of very light smokers is warranted for several reasons: (1) they perceive themselves to be at a reduced disease risk and (2) they appear to differ from other smokers in their motives for smoking (Piper et al., 2004). In addition, if the trial is designed to yield data on population-based effects of MRTP availability, then the sample should comprise both those willing and unwilling to quit. The latter population is appropriate because use of the product could encourage smoking reductions or quitting in those not initially wanting to do so, just as NRT encourages quitting in previously unmotivated individuals (Chan et al., 2011; Schuurmans et al., 2004; Stead and Lancaster, 2007). Moreover,
those who are not initially interested in quitting smoking might be more likely to engage in long-term MRTP use than would others, or engage in dual use (conventional cigarette smoking plus MRTP use), because they might use the MRTP but have little desire to quit smoking. Either of those outcomes would have public health relevance. Therefore, at least one or more of the RCTs conducted should comprise subjects with a range of intentions or motivations to quit use of conventional tobacco products. Finally, while it might seem difficult to attract smokers into a clinical trial who do not wish to quit, in fact, many such smokers are willing to participate in order to try a new product that might be safer than conventional tobacco or that might allow them to reduce their smoking (Carpenter et al., 2004).
One topic that could be addressed in an RCT is the extent to which the MRTP aids cessation or substitution by young or adolescent smokers. To address this, adolescent or young adult smokers could be recruited into the research either in a main study or a study focused on this topic. As with this and other research, an attempt should be made to recruit a representative sample with regard to gender, ethnicity, and socioeconomic status.
Smokeless tobacco users tend not to respond to NRT medications in the same way as cigarette smokers (Fiore et al., 2008). Therefore, if it is deemed important to study MRTP effects in smokeless tobacco users, then it would be important that the trial is adequately powered so as to permit inferences about smokeless users per se.
Characterization of the Sample
As discussed in the section on reinforcement and self-administration studies, a comprehensive characterization of the sample is important because it defines the population to which the conclusions may be most directly related. It also permits tests of the interaction of person factors with MRTP effects. Variables that should be measured are gender, age, ethnicity, educational and socioeconomic status, history of tobacco and nicotine use (including peak tobacco use levels, prior quitting history, age of initial use, and use histories of different tobacco and nicotine products), expectations about the effects of the products or agents to be tested, motivation to quit using tobacco, self-efficacy regarding ability to quit, tobacco or nicotine dependence, blood or breath levels of tobacco or nicotine exposure, health and mental health status and history, other co-addictions (alcohol, narcotics, etc.), and use of psychoactive products including psychiatric medications. The last factor is important because not only may it signal mental health history, but also because some psychiatric medications are effective smoking cessation agents (e.g., bupropion,
nortriptyline) and should be detected for that reason. In addition, measures should also target environmental factors that relate to tobacco cessation success; these include home and work smoking policies and whether the subject lives with a smoker (Bolt et al., 2009). These variables are important because they have been related to nicotine dependence, tobacco self-administration, and ability to control tobacco use.
In terms of tobacco dependence assessment, the Fagerstrom Test for Nicotine Dependence or one of the new multifactorial dependence assessments (the Nicotine Dependence Syndrome Scale [Shiffman and Sayette, 2005; Shiffman et al., 2004] or the Wisconsin Inventory of Smoking Dependence Motives [Smith et al., 2010]) appear to provide more accurate appraisal of dependence than do the Diagnostic and Statistical Manual of Mental Disorders criteria (Hughes et al., 2011).
Subject Recruitment and Randomization
Subjects could be recruited via media announcements or via smoker identification methods used at primary care clinics. The former tends to be more appropriate for efficacy studies where highly motivated subjects are targeted, while the latter tends to be more appropriate for effectiveness studies because the primary care recruitment does not focus on “treatment seekers.” Care must be taken to ensure that recruitment and screening do not set up expectancies among subjects that bias the findings (e.g., expectations that would not be present in real-world use).
To obtain a large sample, studies might have to be conducted at multiple sites. All sites must be adequately described and methods should be adopted that ensure that recruitment, screening, and research and treatment methods be uniform across sites. Further, poolability analyses should be conducted to determine the consistency of findings across sites.
In terms of sample size, it must be set to detect effects in the primary outcome(s) that would be of public health significance. There is no effect size that has been accepted as having clear public health significance for an outcome such as smoking cessation. One approach would be to test whether an MRTP enhances long-term outcomes to a similar degree as over-the-counter cessation medications, which tend to approximately double 6-month abstinence rates (Fiore et al., 2008). However, research grants in this area are often powered to detect effect sizes in which the experimental intervention increases long-term (e.g., 6-month) cessation rates by at least 10 percent (e.g., 10 percent in controls and 20 percent in experimental subjects). A Cochrane report suggested that a cessation increment of 6 percent could be of public health or clinical significance (Lancaster and Stead, 2005). Because the effects of an MRTP would occur on a population-wide basis with use by many thousands of individuals, it
seems prudent to power an RCT to detect relatively small effects. Therefore, consistent with the Cochrane report, it would make sense to power at least one of the RCTs to detect an effect (increment in cessation) of 5-6 percent or greater.
Of course, an RCT permits the collection of information on multiple outcomes, even if many are secondary. If there are especially important secondary outcomes, these too must be considered in setting recruitment goals. For instance, it may be highly efficient to collect data on disease biomarkers or surrogates over the extended use of MRTPs during the trial, although the validity of such biomarkers would need to be considered in such decisions (Hatsukami et al., 2006). Some biomarkers and surrogates may be expected to show changes over the course of a lengthy clinical trial follow-up lasting over a year (e.g., exposure biomarkers or surrogate endpoints like endothelial dysfunction). Biomarkers and surrogates are discussed further in Chapter 3.
The randomization process should follow CONSORT recommendations (Schulz et al., 2010). If multiple sites are used, then randomization should be balanced within sites. Also, the method of randomization should ensure blinding (at least blinding from staff and assessors to the extent possible). Moreover, blocking within each site should be used for factors that might powerfully influence outcomes (e.g., whether or not the research participant has an intention to quit).
Randomized Clinical Trial Measures
As discussed in the section on reinforcement and self-administration measures, biochemical measures of tobacco or nicotine exposure should be collected because they reflect prior self-administration intensity or tolerance and are often related to likelihood of future cessation (al’Absi et al., 2004; Powell et al., 2010). The appropriate measure could be CO level for cigarette smokers, but it could be nicotine or cotinine levels (from blood, saliva, or urine) in other sorts of nicotine or tobacco users (those using smokeless tobacco or NRT). In particular, acute blood nicotine absorption profiles in response to both single and repeated use of products is a meaningful component in assessing the addictive potential of MRTPs. In an RCT where acute effects of self-dosing are not targeted, cotinine may be preferred over nicotine levels, because its longer half-life should provide a more accurate index of chronic consumption. This would be especially important if light smokers are included in the sample. Also, if a noncombustible MRTP is studied, CO levels during or after the experiment will not provide measures of effective dosing. Therefore, to
obtain a true baseline for such later measures, either nicotine or cotinine should be measured at baseline. Measurement of urine cotinine corrected for creatinine concentration may be the best predictor of plasma cotinine (Benowitz et al., 2009). Finally, the investigator might wish to measure 3’-hydroxycotinine in order to estimate nicotine metabolism (Schnoll et al., 2009), which might predict heavy product use and the long-term substitution of the MRTP for smoking versus smoking cessation per se.
Ideally baseline data should be collected via computerized data acquisition systems that ensure complete data recording and detection of out-of-range values. Baseline measures should be taken of all those variables that are to be used as outcomes, moderators, covariates, or to characterize the sample, such as smoking rate, use of all tobacco products, biochemical measures of heaviness of tobacco use, nicotine dependence, socioeconomic and educational status, withdrawal symptoms, affect, mental health history, physical health status and perceived health status, medication use, aversive events (e.g., due to nicotine toxicity), smoking history (e.g., age of first smoking/daily smoking, longest prior abstinence from a quit attempt, prior use of quitting aids), quitting self-efficacy, perceptions of the MRTP, motivation to quit, home and work smoking policies and restrictions, and alcohol use. Investigators should use psychometrically sound instruments and should routinely report psychometric data for their own sample (e.g., coefficient alpha). Also, to the extent possible investigators should use commonly used instruments to enhance assessment of comparability of the recruited sample with samples used in previous research.
Assessment During the Cessation Trial
The key assessment targets include use of both conventional cigarettes and the MRTP. Such use data can be gathered from a variety of means, such as interactive voice response (IVR) assessments via subjects’ cell and landline phones, by mailed questionnaires, or by Internet assessment. If a targeted quit day has been set (e.g., in the context of an efficacy study), then assessments could be concentrated around this time. Otherwise, assessments could occur at intervals of sufficient frequency to permit accurate recall. There is evidence that subjects can complete smoking calendars accurately over 3- to 6-month intervals (Piper et al., 2009), with calendars capturing whether or not smoking occurred on a particular day (i.e., a binary measure of smoking, not a specific amount smoked) over the past 6 months. Assessment of number
of cigarettes smoked/day over the past week only would allow for the estimation of current smoking heaviness (and this would also permit point prevalence assessment for the past week). It seems likely that subjects could supply similar information with regard to MRTP use (with estimates of amount of use per day being captured only for recent days [past week]).
Ideally, periodic ecological momentary assessment data (perhaps captured via IVR calls) could be used to assess heaviness of use of both conventional tobacco and the MRTP. These could target use of both products over the past 24 hours and could occur every other week, or even monthly in an extended study, without constituting an undue burden. Recent clinical trials on smoking cessation have used IVR calls with follow-up durations of a year or more (Brendryen and Kraft, 2008; Reid et al., 2007). In an efficacy study such assessment strategies could be maintained for many months, but they might require compensation in order to obtain high completion rates. Such assessments could also track quit attempts, withdrawal symptoms, self-efficacy, and aversive events.
It is sometimes acceptable not to collect biochemical confirmation of tobacco use status for follow-up outcome assessment, especially in effectiveness studies where there has been little interpersonal contact between the research staff and subjects (Hughes et al., 2003). However, in any study involving extended and multiple assessment contacts, and where degree of product use is important (not just binary measures of use such as targeted in point-prevalence assessments), it would be important to collect biochemical indices of use. Both urine cotinine and CO should be collected, with care taken to collect self-report information on use of any product (e.g., NRT) that could affect levels of biochemical indices of exposure. Therefore, for any RCT it seems highly desirable to schedule in-person visits every 6 months for biological sample collection (Smith et al., 2009). At such in-person visits, researchers could collect additional self-report point-prevalence (past week) data not only on conventional tobacco and MRTP use, but also on such secondary outcomes as MRTP attitudes and liking, motivation to quit, tobacco dependence, changes in important environmental variables (e.g., smokers in the home), and MRTP use in subjects’ social networks (if study is postmarketing).
Throughout the trial, the investigators should track all events that need to be reported in CONSORT event diagrams: numbers of individuals contacting the research program and assessed for eligibility, number excluded and reasons for exclusion, number who declined participation during the induction process and when and why they declined, number assigned to each experimental condition, amount of experimental intervention and assessment received, number who formally discontinued participation and reasons for discontinuation, number lost to follow-up
(unable to contact), and number analyzed and reasons for any departures from intent-to-treat principles. All data should be reported for the entire sample and with respect to treatment condition for measures collected after random assignment. Of course, CONSORT reporting recommendations will no doubt change over time, and researchers should ensure that their methods reflect the most current standards.
Finally, good experimental design standards demand that aside from the manipulation of the independent variable(s), all procedures in the study, including types and intensity of assessments, be equivalent across all experimental conditions.
Selecting and Delivering the Tobacco Products
Some questions concern the method for product provision, the need for a placebo control, and the need for product blinding. With regard to the method for making the tested products available to subjects, two sets of questions can be distinguished with regard to how an MRTP might affect cessation. One question is, does optimal use of the MRTP help a smoker quit smoking, and if so, how effective is it, and how does it compare in this regard to other widely available cessation aids such as NRTs? This is the sort of question addressed in an efficacy trial, which is designed to gauge intervention effectiveness under near-ideal circumstances. A second question is whether MRTP availability per se affects the likelihood of future cessation. This second question is concerned with a real-world effectiveness issue: under conditions of real-world use (or near real-world use), where many individuals may not even use the MRTP or attempt to quit using a conventional tobacco product, how does MRTP availability affect outcomes? If it is deemed important to determine the effectiveness of the product in a formal, structured quit attempt relative to cessation aids such as NRTs, then it should be offered with considerable support for its use. This would entail free product use for the duration of the trial and perhaps training in use, encouragement of use, and perhaps prompting of use. Such a trial would show how effective the MRTP could be in boosting cessation rates (of conventional tobacco products) under ideal conditions. It might even make sense to offer the MRTP in conjunction with adjuvant interventions that are readily available in real-world use, such as quit-line counseling (Miller and Sedivy, 2009; Smith et al., 2009; Tinkelman et al., 2007).
However, it seems that the former question (optimal MRTP efficacy as a cessation aid) is of somewhat less interest than questions that would focus more on real-world use and effects. That is, it seems most relevant to determine if MRTP availability to a group of individuals exerts a net effect on the future use of conventional tobacco products across the population of potential users. The goal of external validity would be served by
providing little support for MRTP use (i.e., providing the MRTP at least nominal cost, providing no more use information than would be provided by package instructions). In addition, the MRTP would be offered by itself with no provision of adjuvant therapy or encouragement for its use. However, it might be that this approach would provide even less use support than would occur in real life, where a person’s social network for instance, might encourage use and provide information.
A related consideration is that the RCTs will probably be used to address multiple questions (even if only one or two are deemed primary). For instance, not only is it of interest to determine if the MRTP affects future use of conventional tobacco, but also it is important to obtain additional information on the health risks that might attend chronic and unsupervised use, or the extent to which MRTP use affects tobacco withdrawal symptoms. Unless a meaningful portion of the sample uses the MRTP regularly, then no inferences can be made about such topics (Does heavy use increase liking? How does heavy real-world use affect nicotine and toxicant exposure?). Therefore, it seems that a good compromise strategy is to conduct at least one efficacy trial and one effectiveness trial. The effectiveness trial could perhaps start out with free use in the early stages of the trial to ensure some initial trial of the product, and then weaning the subjects off supported use, with their eventual request of the product reducing their supported payments by some meaningful amount.
With regard to the issue of placebo control, it seems as though use of a placebo would be desirable in an efficacy trial but not in the effectiveness trial. The reasons that it would be desirable in the efficacy trial are that (1) there is a history of very strong and persistent responses to placebo tobacco products (e.g., cigarettes that contain no nicotine [Perkins et al., 2008]) and (2) even if the MRTP were compared with another “active” product that contained nicotine (e.g., NRT), this would not control for effects of novelty and “newness” that might accompany the provision of a new or less familiar nicotine delivery system. (If the research occurred in a postmarketing context, then this could affect the need for a placebo control.) If a placebo were used, the research should be double blind. However, subjects would not be blind to the product they were using in an effectiveness study. In any study, to the extent possible, the staff collecting assessment and outcome data should be blinded to treatment assignment or product use. Steps to ensure this and quality assurance measures should be described. Also, if placebo control is used, then data should be collected on subjects’ beliefs about the product they were given.
Other intervention procedures should be similar to those used in any well-designed RCT evaluating the use and effectiveness of a cessation aid. Subjects should receive enough product to permit optimal dosing, they should be given instructions for product use that fit the nature of the RCT
(efficacy versus effectiveness), they should be given clear information on health risks and how to spot adverse reactions or effects, they should be given a way to communicate about health concerns and get professional advice, and they should have their use of the products tracked in multiple ways (e.g., “pill counts,” self-report, ecological momentary assessment self-report, medication recording devices). Finally, it would be important that the subjects not be given clear messages about the possible or targeted effects of the products because this could produce biases in subsequent ratings or behaviors (e.g., disappointment, placebo effects, and so on). Perhaps the subjects could merely be told that the MRTP is being evaluated to determine how much people will use it, how it might affect their use of other forms of tobacco, and their attitudes about it.
There may be instances where cluster assignment of participants may be warranted (e.g., where communities or schools are assigned to various products). This would permit assessment of product effects within larger social units (spread of use within peer groups) and also permit assessment of environmental impacts (community cardiac events or bronchitis incidence).
Long-Term Follow-up Methods
Certain methods have been shown in prior research to boost trial participation and adherence:
1. Clear information early on about the assessment burden
2. Timely payment for assessment information and visits
3. Ability of a subject to reschedule assessments
4. Use of brief, clear questionnaires
5. Use of the same assessor over time to promote the development of a personal relationship
6. Collection of information via multiple contact routes (multiple phone numbers, e-mail addresses, home and work addresses, and collateral informants) to facilitate long-term contact
7. Regular inquiries about the subject possibly moving and likely future addresses
8. Explicit permission for a subject to skip a follow-up contact with the understanding that s/he may resume participation at some future point in time
These methods should be adopted in an effort to reduce attrition and boost ascertainment rates. With such methods it may be possible to track clinical trial participants over several years.
As suggested above, tracking of outcomes should occur via multiple
routes: phone calls, mailed questionnaires, Internet questionnaires, and in-person visits. In general, use of multiple data collection routes yields more comprehensive data and higher ascertainment rates. For instance, in-person visits could be made at a periodicity of 6 months to obtain calendar data on smoking and MRTP use (and biochemical samples or physical health tests as needed), but at that periodicity, fine-grained use data (how many cigarettes or MRTP doses were consumed each day) could be obtained only for the past week. Therefore, interval sampling methods using cell phone calls, perhaps on a monthly basis, could provide information on intervening product use and symptoms.
Important elements of an analytic report include
1. as per CONSORT recommendations, primary and secondary outcomes specified a priori;
2. a description of any significant protocol deviations;
3. a complete CONSORT diagram;
4. adherence to intent-to-treat analytic principles and description of exact subject counts included in each analysis;
5. use of experiment-wise error correction, except where primary hypotheses are tested or outcomes important to subject welfare are being evaluated;
6. evaluation of covariates to determine their ability to reduce type II error; and
7. reporting of all adverse events and their relation to MRTP use described.
In addition, the analysis plan should examine relations of MRTP use to outcomes, perhaps with use of formal mediation analytic strategies (MacKinnon et al., 2002; Piper et al., 2008). As with some pharmaceutical products, there may be particular patterns of use that are especially beneficial or harmful; such patterns may be most identifiable through the use of real-time assessments of use patterns, perhaps via electronic monitoring strategies (Cramer et al., 1990; Matsui et al., 1992; Metry, 1999).
At the minimum, it is critical that RCTs analyze the following in order to comprehensively capture the effect of MRTP availability on public health and to support later modeling of such effects:
1. Use of the MRTP
2. Relations of MRTP availability (treatment assignment) and use with measures of use of conventional tobacco products (e.g., cigarette
smoking), with use reflected in both binary and continuous measures (abstinence versus smoking rate data; dual-use rate versus smoking rate data)
3. Relations of MRTP use with occurrence of quit attempts and duration of abstinence achieved in such attempts, and whether MRTP use reduces quit attempts with other sorts of cessation aids (e.g., there may be no net effect on smoking cessation per se, only a shift in type of quitting, as in use of the MRTP versus NRT)
4. Effect of MRTP use on withdrawal and craving during quit attempts and when individuals reduce their use of conventional tobacco products
5. Nicotine dependence with regard to use of both the MRTP as well as conventional tobacco products
6. Changes in perception of conventional tobacco products and of the MRTP as a function of MRTP use over time (e.g., liking, addictiveness, safety)
7. Quitting self-efficacy and quitting intentions in response to use of both conventional tobacco products and the MRTP. Such outcomes should be measured both at discrete endpoints (e.g., abstinence rates at 6-month visits) as well as via ecological momentary assessments that generate data for intensive longitudinal data analysis (e.g., assessment of smoking over time with MRTP use serving as a time varying covariate in growth curve models).
Interpretations of the obtained data need to be synthesized in order to attain a comprehensive assessment of the potential public health impact of approving a product as an MRTP. RCTs can yield data on the use of the MRTP over time on the proportion of people who use it and how heavily they use it, the extent to which it produces or sustains nicotine dependence, and the extent to which it reduces use of conventional tobacco products (e.g., smoking) or reduces use of cessation aids. Data from all relevant measures must be integrated, for instance, taking into account not only the size of the effects of the MRTP on important outcomes but also the prevalence of use and safety findings. For instance, if the product is unappealing and infrequently used, then its potential for a positive public health impact is reduced even if it can boost smoking cessation success. Evaluation of the effects of MRTP will be an iterative process, because information gained from postmarket observations may inform or correct assumptions for laboratory and preclinical investigations. In addition, such synthesis may take into account projected costs to the user and society (e.g., via health care impacts). By supplying data on the outcomes noted above (heaviness of use, duration of use, impact on smoking), RCTs
should yield evidence that would be useful for modeling of population based health and economic impacts. Models can account for and potentially predict the effect of marketing an MRTP on initiation, cessation, or relapse. Simulation models that use mathematical formulas need to account for population dynamics, because initiation and cessation rates can depend on demographic differences and social behaviors.
The synthesis of all of this information will be challenging because it involves explicit or implicit weightings of the various possible outcomes. No well-defined cut scores are available for gauging benefit, and interrelations of variables may be complex. For instance, an MRTP should be compared with one or more NRTs in RCTs (Kotlyar et al., 2007); however, note that the MRTP need not necessarily be “better” or even equivalent to the NRT in order to exert a public health benefit. An MRTP that is inferior to NRTs (more toxicants, less effective at boosting cessation of smoking conventional cigarettes) could still exert a net public health benefit if its modest effects were additive, meaning they occurred on top of those of NRTs. For example, while not being very effective at helping smokers quit when used as a sole product, it is possible that the combination of NRT plus the MRTP yields additive (or even positive synergistic) effects on smoking cessation when in combination. This is entirely possible because combinations of NRT medications are more effective than single medications (Fiore et al., 2008; Piper et al., 2009; Smith et al., 2009). Another possibility is that dual use reduces the rate of cigarette use and exposure to toxicants and therefore results in a net benefit to both individual and public health. Conversely, the net public health impact of the MRTP may be compromised to the extent that it reduced use of NRTs that ultimately led to smoking cessation. Or, the MRTP might benefit a different population of smokers than do NRTs. Ideally, an experimental design should permit the testing of a broad range of MRTPs and MRTP effects.
al’Absi, M., D. Hatsukami, G. L. Davis, and L. E. Wittmers. 2004. Prospective examination of effects of smoking abstinence on cortisol and withdrawal symptoms as predictors of early smoking relapse. Drug and Alcohol Dependence 73(3):267-278.
Baker, T. B., T. H. Brandon, and L. Chassin. 2004a. Motivational influences on cigarette smoking. Annual Review of Psychology 55:463-491.
Baker, T. B., M. E. Piper, D. E. McCarthy, M. R. Majeskie, and M. C. Fiore. 2004b. Addiction motivation reformulated: An affective processing model of negative reinforcement. Psychological Review 111(1):33-51.
Benowitz, N. L., P. Jacob, and B. Herrera. 2006. Nicotine intake and dose response when smoking reduced-nicotine content cigarettes. Clinical Pharmacology and Therapeutics 80(6):703-714.
Benowitz, N. L., K. M. Dains, D. Dempsey, B. Herrera, L. Yu, and P. Jacob. 2009. Urine nicotine metabolite concentrations in relation to plasma cotinine during low-level nicotine exposure. Nicotine & Tobacco Research 11(8):954-960.
Bickel, W. K., and G. J. Madden. 1999a. The behavioral economics of smoking. Chicago: University of Chicago Press.
Bickel, W. K., and G. J. Madden. 1999b. A comparison of measures of relative reinforcing efficacy and behavioral economics: Cigarettes and money in smokers. Behavioural Pharmacology 10(6-7):627-637.
Bickel, W. K., R. Yi, E. T. Mueller, B. A. Jones, and D. R. Christensen. 2010. The behavioral economics of drug dependence: Towards the consilience of economics and behavioral neuroscience. In Behavioral neuroscience of drug addiction. Vol. 3, Current topics in behavioral neurosciences, edited by D. W. Self, and J. K. Staley Gottschalk. Berlin, Germany: Springer Berlin Heidelberg. Pp. 319-341.
Bigelow, G. E. 1991. Human drug abuse liability assessment: Opioids and analgesics. British Journal of Addiction 86(12):1615-1628.
Blondal, T., M. Franzon, and A. Westin. 1997. A double-blind randomized trial of nicotine nasal spray as an aid in smoking cessation. European Respiratory Journal 10(7):1585-1590.
Bolt, D. M., M. E. Piper, D. E. McCarthy, S. J. Japuntich, M. C. Fiore, S. S. Smith, and T. B. Baker. 2009. The Wisconsin Predicting Patients’ Relapse questionnaire. Nicotine & Tobacco Research 11(5):481-492.
Brendryen, H., and P. Kraft. 2008. Happy ending: A randomized controlled trial of a digital multi-media smoking cessation intervention. Addiction 103(3):478-484.
Carpenter, M. J., J. R. Hughes, L. J. Solomon, and P. W. Callas. 2004. Both smoking reduction with nicotine replacement therapy and motivational advice increase future cessation among smokers unmotivated to quit. Journal of Consulting and Clinical Psychology 72(3):371-381.
Carter, B. L., and S. T. Tiffany. 1999. Cue-reactivity and the future of addiction research. Addiction 94(3):349-351.
Carter, L. P., and R. R. Griffiths. 2009. Principles of laboratory assessment of drug abuse liability and implications for clinical development. Drug and Alcohol Dependence 105(Suppl. 1):S14-S25.
CDC (Centers for Disease Control and Prevention). 2005. Annual smoking-attributable mortality, years of potential life lost, and productivity losses—United States, 1997-2001. Morbidity and Mortality Weekly Report 54(25):625-628.
CDC. 2009. Cigarette smoking among adults and trends in smoking cessation—United States, 2008. Morbidity and Mortality Weekly Report 58(44):1227-1232.
CDC and National Center for Health Statistics. 2008. National Health and Nutrition Examination Survey 2007-2008. http://www.cdc.gov/nchs/nhanes.htm (accessed October 5, 2010).
Chan, S. S., D. Y. Leung, A. S. Abdullah, V. T. Wong, A. J. Hedley, and T. H. Lam. 2011. A randomized controlled trial of a smoking reduction plus nicotine replacement therapy intervention for smokers not willing to quit smoking. Addiction 106(6):1155-1163.
Colby, S. M., S. T. Tiffany, S. Shiffman, and R. S. Niaura. 2000. Measuring nicotine dependence among youth: A review of available approaches and instruments. Drug and Alcohol Dependence 59:23-39.
Collins, L. M., T. B. Baker, R. J. Mermelstein, M. E. Piper, D. E. Jorenby, S. S. Smith, B. A. Christiansen, T. R. Schlam, J. W. Cook, and M. C. Fiore. 2011. The multiphase optimization strategy for engineering effective tobacco use interventions. Annals of Behavioral Medicine 41(2):208-226.
Comer, S. D., J. B. Ashworth, R. W. Foltin, C. E. Johanson, J. P. Zacny, and S. L. Walsh. 2008. The role of human drug self-administration procedures in the development of medications. Drug and Alcohol Dependence 96(1-2):1-15.
Cramer, J. A., V. L. Ouellette, and R. H. Mattson. 1990. Effect of microelectronic observation of compliance. Epilepsia 31:617-618.
Dewit, H., S. Dudish, and J. Ambre. 1993. Subjective and behavioral effects of diazepam depend on its rate of onset. Psychopharmacology 112(2-3):324-330.
Fant, R. V., K. J. Schuh, and M. L. Stitzer. 1995. Response to smoking as a function of prior smoking amounts. Psychopharmacology 119(4):385-390.
Fiore, M. C., C. R. Jaen, T. B. Baker, W. C. Bailey, N. Benowitz, S. J. Curry, S. F. Dorfman, E. S. Froelicher, M. G. Goldstein, C. G. Healton, P. N. Henderson, R. B. Heyman, H. K. Koh, T. E. Kottke, H. A. Lando, R. E. Mecklenburg, R. J. Mermelstein, P. D. Mullen, C. T. Orleans, L. Robinson, M. Stitzer, A. C. Tommasello, L. Villejo, and M. E. Wewers. 2008. Treating tobacco use and dependence: 2008 update. Rockville, MD: U.S. Department of Health and Human Services, U.S. Public Health Service.
Foltin, R. W., and M. W. Fischman. 1991. Assessment of abuse liability of stimulant drugs in humans: A methodological survey. Drug and Alcohol Dependence 28(1):3-48.
Gloria, R., L. Angelos, H. S. Schaefer, J. M. Davis, M. Majeskie, B. S. Richmond, J. J. Curtin, R. J. Davidson, and T. B. Baker. 2009. An fMRI investigation of the impact of withdrawal on regional brain activity during nicotine anticipation. Psychophysiology 46(4):681-693.
Griffiths, R. R., J. R. Troisi, II, K. Silvermian, and G. K. Miumford. 1993. Multiple-choice procedure: An efficient approach for investigating drug reinforcement in humans. Behavioural Pharmacology 4(1):3-14.
Griffiths, R. R., G. E. Bigelow, and N. A. Ator. 2003. Principles of initial experimental drug abuse liability assessment in humans. Drug and Alcohol Dependence 70(3):S41-S54.
Haney, M. 2009. Self-administration of cocaine, cannabis and heroin in the human laboratory: Benefits and pitfalls. Addiction Biology 14(1):9-21.
Haney, M., and R. Spealman. 2008. Controversies in translational research: Drug self-administration. Psychopharmacology 199(3):403-419.
Hatsukami, D. K., N. L. Benowitz, S. I. Rennard, C. Oncken, and S. S. Hecht. 2006. Biomarkers to assess the utility of potential reduced exposure tobacco products. Nicotine & Tobacco Research 8(2):169-191.
Hatsukami, D. K., K. Hanson, A. Briggs, M. Parascandola, J. M. Genkinger, R. O’Connor, and P. G. Shields. 2009. Clinical trials methods for evaluation of potential reduced exposure products. Cancer Epidemiology, Biomarkers & Prevention 18(12):3143-3195.
Hatsukami, D. K., J. Jensen, A. Anderson, B. Broadbent, S. Allen, Y. Zhang, and H. Severson. 2011. Oral tobacco products: Preference and effects among smokers. Drug and Alcohol Dependence 118(2-3):230-236.
Hawkins, J., W. Hollingworth, and R. Campbell. 2010. Long-term smoking relapse: A study using the British household panel survey. Nicotine & Tobacco Research 12(12):1228-1235.
Heffner, J. L., T. C. Lee, C. Arteaga, and R. M. Anthenelli. 2010. Predictors of post-treatment relapse to smoking in successful quitters: Pooled data from two phase III varenicline trials. Drug and Alcohol Dependence 109(1-3):120-125.
Heishman, S. J., B. A. Kleykamp, and E. G. Singleton. 2010. Meta-analysis of the acute effects of nicotine and smoking on human performance. Psychopharmacology 210(4):453-469.
Hughes, J. R., and D. Hatsukami. 1986. Signs and symptoms of tobacco withdrawal. Archives of General Psychiatry 43(3):289-294.
Hughes, J. R., S. W. Gust, K. Skoog, R. M. Keenan, and J. W. Fenwick. 1991. Symptoms of tobacco withdrawal—a replication and extension. Archives of General Psychiatry 48(1):52-59.
Hughes, J. R., A. H. Oliveto, and S. Y. Terry. 1996. Saying versus doing and other methodological issues in the study of human drug self-administration. Experimental and Clinical Psychopharmacology 4(2):234-238.
Hughes, J. R., J. P. Keely, R. S. Niaura, D. J. Ossip-Klein, R. L. Richmond, and G. E. Swan. 2003. Measures of abstinence in clinical trials: Issues and recommendations. Nicotine & Tobacco Research 5(1):13-25.
Hughes, J. R., E. N. Peters, and S. Naud. 2008. Relapse to smoking after 1 year of abstinence: A meta-analysis. Addictive Behaviors 33(12):1516-1520.
Hughes, J. R., T. Baker, N. Breslau, L. Covey, and S. Shiffman. 2011. Applicability of DSM criteria to nicotine dependence. Addiction 106(5):894-895.
Hursh, S. R., and A. Silberberg. 2008. Economic demand and essential value. Psychological Review 115(1):186-198.
Jasinski, D. 1977. Assessment of the abuse potentiality of morphine-like drugs (methods used in man). In Handbook of experimental pharmacology. Vol. 45, edited by W. Martin. Heidelberg: Springer. Pp. 197-258.
Johnson, M. W., and W. K. Bickel. 2006. Replacing relative reinforcing efficacy with behavioral economic demand curves. Journal of the Experimental Analysis of Behavior 85(1):73-93.
Johnson, M. W., W. K. Bickel, and A. P. Kirshenbaum. 2004. Substitutes for tobacco smoking: A behavioral economic analysis of nicotine gum, denicotinized cigarettes, and nicotine-containing cigarettes. Drug and Alcohol Dependence 74(3):253-264.
Kleykamp, B. A., J. M. Jennings, and T. Eissenberg. 2011. Effects of transdermal nicotine and concurrent smoking on cognitive performance in tobacco-abstinent smokers. Experimental and Clinical Psychopharmacology 19(1):75-84.
Kotlyar, M., M. I. Mendoza-Baumgart, Z. Li, P. R. Pentel, B. C. Barnett, R. M. Feuer, E. A. Smith, and D. K. Hatsukami. 2007. Nicotine pharmacokinetics and subjective effects of three potential reduced exposure products, moist snuff and nicotine lozenge. Tobacco Control 16(2):138.
Krall, E. A., A. J. Garvey, and R. I. Garcia. 2002. Smoking relapse after 2 years of abstinence: Findings from the VA normative aging study. Nicotine & Tobacco Research 4:95-100.
Lam, T. H., A. S. M. Abdullah, S. S. C. Chan, A. J. Hedley, and Hong Kong Council on Smoking and Health Smoking Cessation Health Centre Steering Group. 2005. Adherence to nicotine replacement therapy versus quitting smoking among Chinese smokers: A preliminary investigation. Psychopharmacology 177(4):400-408.
Lancaster, T., and L. F. Stead. 2005. Individual behavioural counselling for smoking cessation. Cochrane Database of Systematic Reviews (2):CD001292.
Liu, H. H., C. E. Golin, L. G. Miller, R. D. Hays, K. Beck, S. Sanandaji, J. Christian, T. Maldonado, D. Duran, A. H. Kaplan, and N. S. Wenger. 2001. A comparison study of multiple measures of adherence to HIV protease inhibitors. Annals of Internal Medicine 134(10):968-977.
MacKinnon, D. P., C. M. Lockwood, J. M. Hoffman, S. G. West, and V. Sheets. 2002. A comparison of methods to test mediation and other intervening variable effects. Psychological Methods 7(1):83-104.
Madden, G. J., and W. K. Bickel. 1999. Abstinence and price effects on demand for cigarettes: A behavioral-economic analysis. Addiction 94(4):577-588.
Matsui, D., C. Hermann, M. Braudo, S. Ito, N. Olivieri, and G. Koren. 1992. Clinical use of the medication event monitoring system: A new window into pediatric compliance. Clinical Pharacology and Therapeutics 52:102-103.
McClernon, F. J., R. V. Kozink, A. M. Lutz, and J. E. Rose. 2009. 24-h smoking abstinence potentiates fMRI-bold activation to smoking cues in cerebral cortex and dorsal striatum. Psychopharmacology 204(1):25-35.
Metry, J. M. 1999. Measuring compliance in clinical trials and ambulatory care. In Drug regimen compliance: Issues in clinical trials and patient management. 1st ed, edited by J.-M. Metry, and U. A. Meyer. Chichester: John Wiley & Sons. Pp. 1-21.
Miller, C. L., and V. Sedivy. 2009. Using a quitline plus low-cost NRT to help disadvantaged smokers quit. Tobacco Control 18(2):144-149.
Mumford, G. K., S. M. Evans, J. C. Fleishaker, and R. R. Griffiths. 1995. Alprazolam absorption kinetics affects abuse liability. Clinical Pharmacology and Therapeutics 57(3):356-365.
Murphy, J. G., J. MacKillop, J. R. Skidmore, and A. A. Pederson. 2009. Reliability and validity of a demand curve measure of alcohol reinforcement. Experimental and Clinical Psychopharmacology 17(6):396-404.
Murphy, J. G., J. MacKillop, J. W. Tidey, L. A. Brazil, and S. M. Colby. 2011. Validity of a demand curve measure of nicotine reinforcement with adolescent smokers. Drug and Alcohol Dependence 113(2-3):207-214.
Perkins, K. A. 2009. Acute responses to nicotine and smoking: Implications for prevention and treatment of smoking in lower SES women. Drug and Alcohol Dependence 104(Suppl. 1):S79-S86.
Perkins, K. A., L. H. Epstein, J. Grobe, and C. Fonte. 1994a. Tobacco abstinence, smoking cues, and the reinforcing value of smoking. Pharmacology Biochemistry and Behavior 47(1):107-112.
Perkins, K. A., J. E. Grobe, C. Fonte, J. Goettler, A. R. Caggiula, W. A. Reynolds, R. L. Stiller, A. Scierka, and R. G. Jacob. 1994b. Chronic and acute tolerance to subjective, behavioral and cardiovascular effects of nicotine in humans. Journal of Pharmacology and Experimental Therapeutics 270(2):628-638.
Perkins, K. A., J. Grobe, and C. Fonte. 1997. Influence of acute smoking exposure on the subsequent reinforcing value of smoking. Experimental and Clinical Psychopharmacology 5(3):277-285.
Perkins, K. A., E. Donny, and A. R. Caggiula. 1999. Sex differences in nicotine effects and self-administration: Review of human and animal evidence. Nicotine & Tobacco Research 1(4):301-315.
Perkins, K. A., L. Jacobs, L. Clark, C. A. Conklin, M. Sayette, and A. Wilson. 2004a. Instructions about nicotine dose influence acute responses to nasal spray. Nicotine & Tobacco Research 6(6):1051-1060.
Perkins, K. A., L. Jacobs, M. Sayette, and A. Caggiula. 2004b. The influence of instructions and nicotine dose on the subjective and reinforcing effects of smoking. Experimental and Clinical Psychopharmacology 12(2):91-101.
Perkins, K. A., M. Ciccocioppo, C. A. Conklin, M. E. Milanak, A. Grottenthaler, and M. A. Sayette. 2008. Mood influences on acute smoking responses are independent of nicotine intake and dose expectancy. Journal of Abnormal Psychology 117(1):79-93.
Perkins, K. A., S. B. Coddington, J. L. Karelitz, C. Jetton, J. A. Scott, A. S. Wilson, and C. Lerman. 2009. Variability in initial nicotine sensitivity due to sex, history of other drug use, and parental smoking. Drug and Alcohol Dependence 99(1-3):47-57.
Perkins, K. A., J. L. Karelitz, C. A. Conklin, M. A. Sayette, and G. E. Giedgowd. 2010. Acute negative affect relief from smoking depends on the affect situation and measure but not on nicotine. Biological Psychiatry 67(8):707-714.
Piasecki, T. M., D. E. Jorenby, S. S. Smith, M. C. Fiore, and T. B. Baker. 2003. Smoking withdrawal dynamics: II. Improved tests of withdrawal-relapse relations. Journal of Abnormal Psychology 112(1):14-27.
Piasecki, T. M., M. E. Piper, and T. B. Baker. 2010. Refining the tobacco dependence phenotype using the Wisconsin Inventory of Smoking Dependence Motives: II. Evidence from a laboratory self-administration assay. Journal of Abnormal Psychology 119(3):513-523.
Pierce, J. P., K. Messer, M. M. White, D. W. Cowling, and D. P. Thomas. 2011. Prevalence of heavy smoking in California and the United States, 1965-2007. JAMA 305(11):1106-1112.
Piper, M. E., T. M. Piasecki, E. B. Federman, D. M. Bolt, S. S. Smith, M. C. Fiore, and T. B. Baker. 2004. A multiple motives approach to tobacco dependence: The Wisconsin Inventory of Smoking Dependence Motives (WISDM-68). Journal of Consulting and Clinical Psychology 72(2):139.
Piper, M. E., E. B. Federman, D. E. McCarthy, D. M. Bolt, S. S. Smith, M. C. Fiore, and T. B. Baker. 2008. Using mediational models to explore the nature of tobacco motivation and tobacco treatment effects. Journal of Abnormal Psychology 117(1):94-105.
Piper, M. E., S. S. Smith, T. R. Schlam, M. C. Fiore, D. E. Jorenby, D. Fraser, and T. B. Baker. 2009. A randomized placebo-controlled clinical trial of 5 smoking cessation pharmacotherapies. Archives of General Psychiatry 66(11):1253-1262.
Powell, J., L. Dawkins, R. West, and A. Pickering. 2010. Relapse to smoking during unaided cessation: Clinical, cognitive and motivational predictors. Psychopharmacology 212(4):537-549.
Reid, R. D., A. L. Pipe, B. Quinlan, and J. Oda. 2007. Interactive voice response telephony to promote smoking cessation in patients with heart disease: A pilot study. Patient Education and Counseling 66(3):319-326.
Robinson, M. L., E. J. Houtsmuller, E. T. Moolchan, and W. B. Pickworth. 2000. Placebo cigarettes in smoking research. Experimental and Clinical Psychopharmacology 8(3):326-332.
Rogers, S., T. Gould, and T. Baker. 2009. Mouse models and the genetics of nicotine dependence. In Phenotypes and endophenotypes: Foundations for genetic studies of nicotine use and dependence. Tobacco Control Monograph No. 20. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute.
Rose, J. E., E. C. Westman, F. M. Behm, M. P. Johnson, and J. S. Goldberg. 1999. Blockade of smoking satisfaction using the peripheral nicotinic antagonist trimethaphan. Pharmacology Biochemistry and Behavior 62(1):165-172.
Rose, J. E., F. M. Behm, E. C. Westman, J. E. Bates, and A. Salley. 2003. Pharmacologic and sensorimotor components of satiation in cigarette smoking. Pharmacology, Biochemistry and Behavior 76(2):243-250.
Sayette, M. A., and M. R. Hufford. 1994. Effects of cue exposure and deprivation on cognitive resources in smokers. Journal of Abnormal Psychology 103(4):812-818.
Sayette, M. A., C. S. Martin, J. G. Hull, J. M. Wertz, and M. A. Perrott. 2003. Effects of nicotine deprivation on craving response covariation in smokers. Journal of Abnormal Psychology 112(1):110-118.
Schnoll, R. A., F. Patterson, E. P. Wileyto, R. F. Tyndale, N. Benowitz, and C. Lerman. 2009. Nicotine metabolic rate predicts successful smoking cessation with transdermal nicotine: A validation study. Pharmacology Biochemistry and Behavior 92(1):6-11.
Schulz, K. F., D. G. Altman, and D. Moher. 2010. CONSORT 2010 statement: Updated guidelines for reporting parallel group randomized trials. Annals of Internal Medicine 152(11):726-732.
Schuurmans, M. M., A. H. Diacon, X. van Biljon, and C. T. Bolliger. 2004. Effect of pre-treatment with nicotine patch on withdrawal symptoms and abstinence rates in smokers subsequently quitting with the nicotine patch: A randomized controlled trial. Addiction 99(5):634-640.
Shaham, Y., H. Rajabi, and J. Stewart. 1996. Relapse to heroin-seeking in rats under opioid maintenance: The effects of stress, heroin priming, and withdrawal. The Journal of Neuroscience 16(5):1957-1963.
Shaham, Y., U. Shalev, L. Lu, H. De Wit, and J. Stewart. 2003. The reinstatement model of drug relapse: History, methodology and major findings. Psychopharmacology 168(1-2):3-20.
Shiffman, S., and M. A. Sayette. 2005. Validation of the nicotine dependence syndrome scale (NDSS): A criterion-group design contrasting chippers and regular smokers. Drug and Alcohol Dependence 79(1):45-52.
Shiffman, S., W. G. Shadel, R. Niaura, M. A. Khayrallah, D. E. Jorenby, C. F. Ryan, and C. L. Ferguson. 2003. Efficacy of acute administration of nicotine gum in relief of cue-provoked cigarette craving. Psychopharmacology 166(4):343-350.
Shiffman, S., A. Waters, and M. Hickcox. 2004. The nicotine dependence syndrome scale: A multidimensional measure of nicotine dependence. Nicotine & Tobacco Research 6(2):327-348.
Shiffman, S., S. G. Ferguson, J. Rohay, and J. G. Gitchell. 2008a. Perceived safety and efficacy of nicotine replacement therapies among US smokers and ex-smokers: Relationship with use and compliance. Addiction 103(8):1371-1378.
Shiffman, S., C. T. Sweeney, S. G. Ferguson, M. A. Sembower, and J. G. Gitchell. 2008b. Relationship between adherence to daily nicotine patch use and treatment efficacy: Secondary analysis of a 10-week randomized, double-blind, placebo-controlled clinical trial simulating over-the-counter use in adult smokers. Clinical Therapeutics 30(10):1852-1858.
Smith, S. S., D. E. McCarthy, S. J. Japuntich, B. Christiansen, M. E. Piper, D. E. Jorenby, D. L. Fraser, M. C. Fiore, T. B. Baker, and T. C. Jackson. 2009. Comparative effectiveness of 5 smoking cessation pharmacotherapies in primary care clinics. Archives of Internal Medicine 169(22):2148-2155.
Smith, S. S., M. E. Piper, D. M. Bolt, M. C. Fiore, D. W. Wetter, P. M. Cinciripini, and T. B. Baker. 2010. Development of the brief Wisconsin Inventory of Smoking Dependence Motives. Nicotine & Tobacco Research 12(5):489-499.
Stead, L. F., and T. Lancaster. 2007. Interventions to reduce harm from continued tobacco use. Cochrane Database of Systematic Reviews (3):CD005231.
Strasser, A. A., C. Lerman, P. M. Sanborn, W. B. Pickworth, and E. A. Feldman. 2007. New lower nicotine cigarettes can produce compensatory smoking and increased carbon monoxide exposure. Drug and Alcohol Dependence 86(2-3):294-300.
Strasser, A. A., K. Z. Tang, P. M. Sanborn, J. Y. Zhou, and L. T. Kozlowski. 2009. Behavioral filter vent blocking on the first cigarette of the day predicts which smokers of light cigarettes will increase smoke exposure from blocked vents. Experimental and Clinical Psychopharmacology 17(6):405-412.
Sutherland, G., J. A. Stapleton, M. A. Russell, M. J. Jarvis, P. Hajek, M. Belcher, and C. Feyerabend. 1992. Randomised controlled trial of nasal nicotine spray in smoking cessation. Lancet 340(8815):324-329.
Tinkelman, D., S. M. Wilson, J. Willett, and C. T. Sweeney. 2007. Offering free NRT through a tobacco quitline: Impact on utilisation and quit rates. Tobacco Control 16(Suppl. 1):i42-i46.
Vogt, F., S. Hall, and T. M. Marteau. 2008. Understanding why smokers do not want to use nicotine dependence medications to stop smoking: Qualitative and quantitative studies. Nicotine & Tobacco Research 10(8):1405.
Weiss, R. B., T. B. Baker, D. S. Cannon, A. von Niederhausern, D. M. Dunn, N. Matsunami, N. A. Singh, L. Baird, H. Coon, W. M. McMahon, M. E. Piper, M. C. Fiore, M. B. Scholand, J. E. Connett, R. E. Kanner, L. C. Gahring, S. W. Rogers, J. R. Hoidal, and M. F. Leppert. 2008. A candidate gene approach identifies the CHRNA5-A3-B4 region as a risk factor for age-dependent nicotine addiction. PLoS Genetics 4(7):e1000125.
Welsch, S. K., S. S. Smith, D. W. Wetter, D. E. Jorenby, M. C. Fiore, and T. B. Baker. 1999. Development and validation of the Wisconsin Smoking Withdrawal Scale. Experimental and Clinical Psychopharmacology 7(4):354-361.
Westman, E. C., F. M. Behm, and J. E. Rose. 1996. Dissociating the nicotine and airway sensory effects of smoking. Pharmacology, Biochemistry and Behavior 53(2):309-315.
Zinser, M. C., M. C. Fiore, R. J. Davidson, and T. B. Baker. 1999. Manipulating smoking motivation: Impact on an electrophysiological index of approach motivation. Journal of Abnormal Psychology 108(2):240-254.