WO1995021419A1

WO1995021419A1 - An expert system intervention for smoking cessation

Info

Publication number: WO1995021419A1
Application number: PCT/US1995/001367
Authority: WO
Inventors: James O. Prochaska; Wayne F. Velicer
Original assignee: The Board Of Governors For Higher Education, State Of Rhode Island And Providence Plantations
Priority date: 1994-02-01
Filing date: 1995-02-01
Publication date: 1995-08-10
Also published as: AU1837295A

Abstract

Intervention efficacy is increased when a treatment is maximally matched to the needs of an individual. In the present invention, a computer-based decision making system utilizes client information to produce unique, matched information and interventions. The expert system combines the user matching possible in a clinic-based intervention and the low cost associated with a public health approach. A computer-driven, expert system intervention specifically developed for smoking cessation is described. The invention provides a cost effective, viable, and efficacious means of intervening in a specific problem behavior area.

Description

TITLE

An Expert System Intervention For Smoking Cessation

BACKGROUND AND SUMMARY OF THE INVENTION

To control behavior such as smoking cessation, intervention efficacy can be increased when the treatment is maximally matched to the needs of the user. How to perform the matching in a cost-effective manner presents an important challenge to researchers and service providers. Traditionally, trained experts have been relied upon to guide and deliver interventions. While this approach has often proven effective, it remains prohibitively expensive for a wide variety of applications. Other approaches simply employ the same generic intervention for all clients with no attempt at individualized matching, in what has often been described as the public health approach to intervention.

The present invention embodies an intervention which combines individualized matching of the human-guided intervention with the low cost of the public health approach.

The preferred embodiment of the invention is directed to a system designed to promote smoking cessation. The impact of smoking on public health is enormous. Some 50 million Americans continue to smoke cigarettes despite more than 25 years of health education programs, anti-smoking campaigns, declining social acceptability of smoking, and well established health consequences of smoking. An estimated 390,000 Americans die. each year from diseases caused by smoking, including 115,000 from heart disease and 106,000 from lung cancer. More than one of every six deaths in the United States is caused by smoking. A similar pattern of negative consequences for smoking occurs for most countries in the world. On the positive side, smoking cessation has major and immediate health benefits for men and women of all ages.

An expert system has been defined in a variety of ways. The most general is a software system that mimics the deductive or inductive reasoning of a human expert, Negotia, U.N. (1985) Expert Systems and Fuzzy Systems, CA: Benjamin/Cummings. More specifically, an expert system requires two things: (a) a collection of facts and rules about a field and (b) a way of making inferences from those facts and rules (Negotia, supra). The heuristics and domain theories on which an expert system relies are referred to as surface knowledge. Most expert systems include only surface knowledge but some may also rely on principles and general theories, or deep knowledge, which can become necessary when confronting really difficult problems, Harmon, P., & King, D., (1985), Expert

Systems: Artificial Intelligence in Business , New York, John Wiley & Sons .

There are a variety of advantages attributed to an expert system when compared to a human expert, including ease of documentation, ease of transfer to multiple sites, increased consistency in decision making, increased potential for replicable results, permanence, and low costs. Additionally, considered from a public health perspective, an expert system is logistically compatible with large scale implementation, allowing for potential impact on large segments of the population.

The heart of an expert system is a powerful body of knowledge that accumulates during system building. This knowledge is explicit and organized to simplify decision making, Waterman, D.A., (1986), A Guide to Expert Systems, Redding, MA: Addison-Wesley. It can be partitioned into two fundamental structures: a well developed conceptual model and a body of empirical data to serve as a basis for decision making. The conceptual model embodied in the expert system

(or deep knowledge) of the invention is the transtheoretical model, Prochaska, J.O., & DiClemente, C.C., (1983), Stages and processes of self-change of smoking: Toward an integrative model of change,

Journal of Consulting and Clinical Psychology, 51, 390-395.

Some definitions are more extensive than those mentioned above, requiring that the expert system be able to operate with incomplete data and explain its decision making to the user, Bielawski, L., and Lewand, R., (1988), Expert Systems

Development: Building PC-Based Applications, Boston, MA: Abacus Press. Some suggest that a prerequisite for applying expert, systems technology to some knowledge domain is the existence of human experts in that domain, Johnson, L., & Keravnov, E.T., (1985), Expert Systems Technology: A Guide, Boston, MA: Abacus Press. The Expert System does duplicate a human expert's knowledge. It is also often a requirement that the knowledge base be easily accessed, updated, and modified when new information becomes available. The "knowledge" is the actual intervention message in the written report generated by the Expert System, this is easily modified. The decision making algorithm is based on the application of empirically validated "deep" knowledge. However, the algorithm involves a series of comparisons with normative data from a reference group, which is easily modified as more appropriate reference data become available.

The scope of problems and tasks which have been approached via expert systems is quite varied and continually growing. For example, expert systems have been or are under development to assist managers with complex planning and scheduling tasks, to diagnose diseases, to locate mineral deposits, to configure complex computer hardware, and to aid mechanics in troubleshooting locomotive problems, Harmon, P., & King, D., (1985), Expert Systems: Artificial Intelligence in Business, New

York: John Wiley & Sons. The "message" communicated through each of these systems obviously varies extensively in terms of content. Some systems, such as those that attempt to diagnose disease, attempt to achieve a discrete solution in response to a specific set of circumstances, such as in the case of those that attempt to diagnose disease. Generally, such systems function in the role of a "consultant," with which a human physician would interact. The message provided through this type of system is designed to increase accuracy of diagnoses by drawing on the knowledge of renowned experts within specific medical areas.

Other systems, such as one that guides insurance salesmen through a predetermined, complex sales protocol designed to increase customer purchases, provide a different sort of message, i.e., one that offers multiple options. The benefit of the expert system in this case can be seen in its ability to communicate a large body of complex information about a specific area, essentially teaching it to the user in a reliable, replicable, and systematic fashion. Although such a system would ideally have the ability to interact with the user on several levels (e.g., respond to specific queries, explain reasoning for product choices or means of presentation, etc.), in many ways it could be seen as an on-line, interactive teaching device.

An object of the present invention is to provide an expert system which effectively and positively influences behavioral patterns.

Another object of the present invention is to provide an expert system which is based on information received from a user and prepares a specially tailored report based on that user's response.

It is another object of the invention to monitor on a continual basis the user's progress in achieving the desired behavioral modification.

To achieve these and other objects, an expert system, according to the present invention, broadly comprises a method for processing behavior modification data which comprises: inputting raw data based on an individual's response to a series of questions into a computer;

transforming the raw data into scale scores representing a plurality of variables drawn from a transtheoretical model of change;

assessing the variables to provide an output corresponding to the individual's attitudes and behaviors; matching the information to the individual's stage of change, which stage of change is based on current process of change use and which information is responsive to a balance of pros and cons of the behavior to be corrected; providing at least an initial (Baseline) report comprising a first-time assessment as well as comparisons to a normative sample of individuals who have successfully modified their behavior. In a preferred embodiment, first and second follow-up reports are also generated.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is an execution control diagram of a batch system of the preferred embodiment of the invention;

Fig. 2 is a data flow program overview of the system of Fig. 1;

Fig. 3 is a base line and follow-up data flow diagram of the system of Fig. 1; and.

Fig. 4 is an illustration of an interactive system of an alternative embodiment of the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENT (S)

The process or system of the invention provides individually relevant and appropriate information to aid users in the process of quitting smoking. The primary value of the system is its ability to provide individually tailored advice and concrete behavioral recommendations in response to a brief paper and pencil assessment or other input of a user's and behaviors related to smoking. Previous methods have been based on a generic public health approach, which provides the same recommendations and advice to all smokers, regardless of their current attitudes and behaviors. Additionally, this system has the advantage of providing useful information to all smokers, even those who are not currently considering quitting at the present time. This makes it feasible to communicate the system "message," or deliver the intervention, in a proactive, rather than reactive manner. Most prior smoking cessation interventions have been geared primarily toward those users who are ready to take action. The system of the invention has the advantage of providing useful, helpful information to those less motivated, without alienating these users by making assumptions about their desires or readiness to quit.

The medium of communication utilized by the system is to inter with a computer directly or receive a written report of two .to four pages in length with the information having beenscanned into the computer.

The software associated with all decision making algorithms is completely integrated with word processing software, making the act of report generation relatively simple.

In the following discussion of the preferred embodiment

(batch system) of the invention, reference will be made to various forms used to assess an individual's response, scale scores representing various variables drawn from a transtheoretical model of change, the matching of this information in the reports to an individual's stage of change sensitive to current process or change use in response to a balance of pros and cons of smoking, as well as to potential high risk smoking situations. A report is generated from a first time assessment, provides stage of change relevant information as well as comparisons to a normative sample of successful quitters. Reports are generated from follow-up assessments which add ipsative feedback providing concrete commentary on individual changes since a previous assessment.

The specific questions on the various forms (or screen prompts) and the responses provided to the individuals in the written reports have been previously described in various publications of the inventors. Therefore, the specific format of the questions, the lexicon of paragraphs which form the tailored reports, and the specific format of the reports are not described in detail. The present invention is directed to a process for data, whether identified as a batch or interactive system, which manipulates and processes the information received based on the responses by the individual and provides the output (reports to the individual through the various computer programs). These programs match the type of responses that would be provided are substantially the same as if the individual were receiving private counseling. The invention thus enables, on a very broad basis, individuals to receive specially tailored reports at relatively low cost which heretofore had not been possible.

The message channel employed with the preferred embodiment of the system is a written report, printed on paper, although the same report can be effectively presented through a number of other channels. For example, after a brief, on-line assessment through questions presented on a computer screen, the same individually tailored report could be presented immediately following the assessment on a computer screen. A sound board added to such a system would permit the simultaneous presentation of the material both as a written message and as a spoken message. An implementation such as this is very viable in a setting such as a physician's office, an HMO setting, a school, or a worksite.

In a first level, a user is provided with information in the form of a written report generated by the system. For the preferred embodiment, each user is a smoker. In limiting the scope of the targeted range of behavior associated with the intervention, the knowledge domain necessary to build an effective system is narrowed.

A second level of generality would be a system targeted to a single problem but applicable to all users. For smokers, cessation materials would be presented and for nonsmokers, prevention or passive smoking materials would be presented.

A third level of generality would be a more general purpose system, which involves all users and multiple problems. A multiple risk factor approach can be developed, assuming that there is likely to be some behavior change associated with cancer risk appropriate for everyone. This system could take several alternative approaches. The system could allow the user to identify the primary problem of interest, or the system could determine the "highest" risk behavior, given knowledge of user demographics, and begin there. Problem behavior areas could be dealt with sequentially or simultaneously. The intrinsic qualities of the system intervention make such investigations quite feasible.

Users could also be segmented by environment. The amount and degree of external control that exists in the environment could affect how the system is implemented. For the class of risk behaviors that are considered here, different environments that would effect implementation are school, physician's office, home and work site. A more restrictive environment would permit the expert system to also integrate existing interventions already available in the environment.

Another basis for segmentation could be a key demographic variable such as education level or income. Segmentation based on educational level might require that a system be capable of communicating written messages at several different reading levels. Similarly, were a different message channel used, one would have to be conscious of segmentation based on income that required resources, such as a VCR or a computer in the home, that might not be available to all potential users.

Effects are conceptualized as the impact of the message on the user. Assessment of this impact requires that some sort of message from the user be communicated back to the system. This message is often referred to as feedback. Feedback may be either feedback from a participant to the expert system showing the participants progress or feedback that the expert system gives to the participant in response to answers to the questions. Some systems, such as those that diagnose illness, benefit from feedback from more invasive procedures pertaining to the ultimate accuracy of their conclusions. For example, once diagnoses offered by the expert system are confirmed or disconfirmed, provision of this feedback to the system can influence the probabilities and weightings associated with decision rules. Assessment of the impact on users is also important simply in terms of the efficacy of the expert system. Without some form of feedback from the user, it is impossible to ascertain the effectiveness of the expert system, or to improve its performance.

In the case of intervention with the system of the invention, assessment of effects involves the process and outcome measures described below. The outcome measures include both behavioral and cognitive measures. For example, of interest are the current smoking status and number of cigarettes smoked per day, which reflects current behavior, and the Pros and Cons scales from the decisional balance, which reflects cognitions about the advantages and disadvantages of continuing to smoke.

The feedback channel associated with the system can potentially take a number of forms. With an on-line, interactive system, the user is providing a form of feedback with each response or query communicated to the expert system. Assessment of outcome measures can also be accomplished via an on-line, interactive method.

Feedback can involve the communication of outcome information at some later time. Data of this sort could conceivably take many forms and be communicated to the system in many different ways. Alternative feedback channels include both verbal and direct entry of data. Verbal entry might involve a phone interview where the interviewer enters the data into the computer. The development of CATI (Computer Assisted Telephone Interviewing) systems has increased the efficiency and feasibility associated with large scale interview projects. Direct entry could involve either on-line assessment at a later time, or perhaps use of touch-tone telephone technology as a means of conducting follow-up assessments. It is likely that advances in the area of communication will provide alternative feedback channels.

The feedback channel employed for the system involves user responses to a series of questions printed on a mark sense form which is returned through the mail . Responses to questionnaires are then entered into the system via an optical scanner. This method has benefits both in terms of cost effectiveness, and efficiency with respect to manual labor demands. The integrative nature of the system allows this new information to be linked with data collected previously from the same user, allowing for ipsative comparisons to be made across a series of interventions.

The transtheoretical model represents the deep knowledge basis of the system of the invention. It specifies a series of independent variables, called the processes of change, a temporal ordering called the stages of change, and a series of intervening or outcome measures. The specific application of the measures is dependent on the specific problem behavior. For the core constructs, the items change for different problem areas. Described for the preferred embodiment are measures appropriate for smoking cessation in this section. Measures appropriate for other problem areas such as alcohol abuse, fat reduction in the diet, sun exposure, increasing exercise, and mammography screening are within the scope of the invention.

The intervening or outcome measures include three separate subscales of the Temptation to Smoke Inventory: (1), Positive/Social, (2) Negative/Affective, and (3) Habit/Addictive types of situations.

Also included are the two subscales of the Decisional Balance Inventory, Velicer, W.F., DiClemente, C.C., Prochaska, J.O., & Brandenburg, N, (1985), Decisional Balance Measure For Assessing

And Predicting Smoking Status, Journal of Personality and Social Psychology, 48, 1279-1289: (1) Pros Scale, and (2) Cons Scales. Within the model, behavioral measures are determined by specific problem area. For smoking cessation, these include (1) point prevalence, or current smoking status, Velicer, W.F., Prochaska, J.O., Rossi, J.S., & Snow, M.G., (1992), Assessing Outcome In Smoking Cessation Studies, Psychological Bulletin, 111, 23-41: (2) number of cigarettes currently being smoked; (3) behavior intention, or plans to modify behavior in the next 30 days, (4) continuous abstinence, or no smoking since the last assessment, and (5) previous quit attempts, or the number of times where at least a 24-hour quit attempt occurred in the last six months. The next section describes some of these measures in more detail, and references some of the available empirical evidence.

Stages of Change. The stages of change represents a central concept for the transtheoretical model of behavior change. Similar concepts have been discussed by Brownell, K., Marlatt, G.A., Lichtenstein, E., & Wilson, G.T., (1986), Understanding And Preventing Relapse , American Psychologist, 41, 765-782; Horn, D.A., (1976), A Model For The Study Of Personal Choice Health Behavior, International Journal of health Education, 19, 89-98; Horn, D.A., & Waingrown, S., (1966), Some Dimensions Of A Model For Smoking Behavior Change, American Journal of Public Health, 56, 21. The stages are a central organizing construct, describing when particular processes of change are most profitably employed and when the use of a process might be counter-productive. The pattern of change of the intervening variables across stage is well defined and different measures are more sensitive to change in different stages.

In retrospective, cross-sectional and longitudinal studies of how people quit smoking on their own, evidence was discovered that smokers move through a series of stages of change in their efforts to quit smoking, DiClemente, C.C., & Prochaska, J.O., (1982), Self-Change And Therapy Change Of Smoking Behavior: A Comparison Of Processes of Change In Cessation And Maintenance, Addictive Behaviors, 7, 133-142; Prochaska, J.O., & DiClemente, C.C., (1983), Stages And Processes Of Self-Change Of Smoking: Toward An Integrative Model Of Change, Journal of Consulting and

Clinical Psychology, 51, 390-395; Prochaska, J.O., DiClemente, C.C., Velicer, W.F., Ginpil, S., & Norcross, J.C., (1985), Predicting Change In Smoking Status For Self-Changers, Addictive Behaviors, 10, 395-406.

These stages have been labelled Precontemplation, Contemplation, Preparation, Action, and Maintenance. Precontemplation is a stage in which smokers are not thinking about quitting smoking within the next six months. A six-month time frame was used because it was assumed that this is about as far in the future as most people plan a specific behavior change. Contemplation is the period of time in which smokers are seriously thinking about quitting smoking in the next six months. Preparation has been defined as smokers seriously thinking about quitting smoking in the next month and having made a recent quit attempt. Action is a continuous period ranging from 0 to 6 months after smokers have made the overt change of stopping smoking. Maintenance is defined as the stage beginning six months after Action started and continuing until smoking is terminated as a problem. Inter-stage differences on several variables is described by DiClemente, C.C., Prochaska, J.O., Fairhurst, S.K., Velicer, W.F., Valesquez, M.M., & Rossi, J.S., (1991), The Processes Of Smoking Cessation: An Analysis Of Precontemplation, Contemplation, And Preparation Stages Of Change, Journal of Consulting and Clinical Psychology,

59, 295-304; and intra-stage differences are described by Velicer, W.F., Hughes, S.L., Fava, J.L., Prochaska, J.O., & DiClemente, C. C. (1992), An Empirical Typology Of Users Within The Stages Of Change, Manuscript submitted for publication.

Temptations to Smoke. This questionnaire is either a 20-item (long form) or 9-item (short form) scale devised to measure temptations to smoke across a wide variety of everyday situations. Items are measured using a five-point Likert scale format, with higher scores indicating greater temptations to smoke. The questionnaire is reliable and has been replicated across samples, problems and response formats, DiClemente, C.C., Prochaska, J.O., & Gilbertini, M., (1985), Self-Efficacy And The Stages Of Self-Change Of Smoking, Cognitive Therapy And Research, 9, 181-200; Velicer, W.F., DiClemente, C.C., Rossi, J.S., & Prochaska, J.O., (1990), Relapse Situations And Self-Efficacy: An Integrative Model, Addictive Behaviors, 15, 271-283. The scale includes three correlated subscales: (1) Positive/Social, or the temptation to smoke when having a good time and with other people; (2) Negative/Affective, or the temptation to smoke when feeling lonely or feeling negative emotions; and (3) Habit/Addictive, or the temptation to smoke when feeling a strong craving. The pattern of change in temptations across stages is a gradual decrease from Precontemplation to Maintenance.

Pros and Cons of Smoking. This questionnaire is a 20-item (long form) or 6-item (short form) scale devised to measure the decisional balance between the positive (Pros) and negative (Cons) aspects of smoking cigarettes. Two scales have been identified and labeled the Pros of Smoking and the Cons of Smoking, respectively (Velicer, supra). The two scales have been successful in differentiating between five groups representing stages of change in the quitting process, and have also been successful when employed as predictors of smoking status at a 6-month follow-up (Prochaska, supra).

The Pros scale declines across the stages with the highest scores for Precontemplation and Contemplation followed by regular decrements across the last three stages. The Cons scale is very low in Precontemplation, increases dramatically to equal the Pros in Contemplation, and then declines across the last three stages, always remaining above the Pros scale (Velicer, supra; Prochaska, supra).

Processes of Change. The processes of change is a 40-item (long form) or 20-item (short form) questionnaire which measures the 10 processes described by the transtheoretical model of change, Prochaska, J.O., Velicer, W.F., DiClemente, C.C., & Fava, J.L., (1988), Measuring The Processes Of Change: Applications To The Cessation Of Smoking, Journal of Consulting and

Clinical Psychology, 56, 520-528. The 10 processes are: Consciousness Raising, Self-Reevaluation, Dramatic Relief, Environmental Reevaluation, Social Liberation, Counterconditioning, Stimulus Control, Helping Relationship, Self-Liberation, and Reinforcement Management. The first five are labelled Experiential processes and involve cognitive and emotional activities. The second five are labelled Behavioral Processes and involve primarily behavioral activities or cognitive labelling of behaviors.

Each item is responded to on a five-point Likert scale of frequency of use ranging from (1) never to (5) repeatedly. The validity of this scale for distinguishing successful and unsuccessful users for each of the stages has been demonstrated cross-sectionally Prochaska, J.O., & DiClemente, C.C., (1983), Stages And Processes Of Self-Change Of Smoking: Toward An

Integrative Model Of Change, Journal of Consulting and Clinical

Psychology, 51, 390-395, predictively Prochaska, J.O., DiClemente, C.C., Velicer, W.F., Ginpil, S., & Norcross, J.C.,

( 1985 ) , Predicting Change In Smoking Status For Self-Changers, Addictive

Behaviors , 10 , 395-406 .

The curvi-linear pattern of change for each of the processes across the stages of change is similar. Use of the processes increases, peaks, and then decreases. The processes differ in when the peaks occur, with the more experiential processes peaking early and the more behavioral processes peaking later (Prochaska, et al., 1990) .

Expert System

First, a thorough description of what is required of the user participating in the intervention (which may be thought of as the "input" into the system) is presented, and is followed by some specifics regarding what the user may take away from the intervention (which may be thought of as the "output" generated by the Expert System) . Alternative approaches to implementing a computer-based intervention for smoking cessation are provided by Orlandi, M.A., Dozier, C.E., & Mart a, M.A., (1990) , Computer-Assisted Strategies For Substance Abuse Prevention: Opportunities And Barriers, Journal of Consulting and Clinical Psychology, 58, 425-431 and Currey, S.J., Wagner, E.G., & Grothaus, L.C., (1991), Evaluation Of Intrinsic And Extrinsic

Motivation Interventions With A Self-Help Smoking Cessation Program, Journal of

Consulting and Clinical Psychology, 59, 318-324.

In the batch system a data base is run on a computer and the data base is used to determine what potential users receive survey forms, and paper forms are mailed out to the users. The user answers the questions on the forms. They mail the forms back. That data is then entered into the computer and then a report is printed and mailed to the user. Input

The user is asked to complete a series of questionnaires previously set forth. Recruited users, i.e., a project is advertised in some way and users initiate contact if interested Standardized, individualized, interactive, and personalized self-help programs for smoking cessation, Health Psychology and proactively recruited users, i.e., users are individually contacted and are included in the project unless they refuse service, Prochaska, J.O., Velicer, W.F., DiClemente, C.C., & Fava, J.L., (1988), Measuring The Processes Of Change: Applications To The Cessation Of Smoking , Journal of Consulting and Clinical Psychology, 56, 520-528.

These questionnaires are reproduced on customized, machine-readable forms which require that the respondent "bubble in" circles associated with specific responses. Other assessment modalities (e.g., phone interview responses using a CATI system) have been implemented and are compatible with the expert system as well.

Four items are designed specifically to assess the users' stage of change with respect to quitting smoking. The stages of change include Precontemplation, Contemplation, Preparation, Action, and Maintenance. A user who has taken action by making a quit attempt, but who subsequently returns to smoking is thought of as a "relapser." Although "Relapse" is not technically a stage of change, the expert system responds to users fitting this description with reports which specifically address relapse issues. Change process use is assessed through a refined, twenty-item version of the longer, forty item Processes of Change questionnaire. Using a five-point Likert format, these twenty items assess use of ten processes of change common to the behavior change experience across a number of problem behaviors. These processes have been empirically validated within the contexts of self-change and change initiated within a therapeutic setting or relationship. The processes of change have been described in detail in an earlier section.

Decisional balance, or the Pros and Cons of smoking, are assessed through six five-point Likert format items. These six items comprise two three-item scales which measure how the user assesses the positive and negative aspects of smoking, and represent a shortened version of the Decisional Balance Inventory (Velicer et al., 1985). The user is asked to indicate how important (from "not important" to "extremely important") each statement (item) is with regard to his or her decision to smoke. Examples of items include "Smoking cigarettes relieves tension" (a pro), and "My cigarette smoking bothers other people" (a con). These variables have been found to be related to movement through the stages of change. The relative level of temptation to smoke is assessed via nine five-point Likert formatted items. These nine items form a single scale, which is an indicant of overall temptation level. Additionally, the overall scale can be subdivided into three subscales which provide feedback regarding specific tempting situations (i.e., Habit/Addictive, Positive/Social, Negative/Affective) (Velicer et al., 1990).

Output

Through assessment of these items and calculation of scale scores, a personalized report is generated. The first or baseline report uses only normative comparisons. Follow-up or progress reports use both normative and ipsative information. Each report contains four sections. The first section provides information pertaining to the user's particular Stage of Change. Additionally, this section provides feedback regarding Decisional Balance considerations with respect to the user's decision to smoke. Focusing primarily on the Cons of smoking, this feedback may suggest that the user increase his or her thinking about the negative aspects of smoking, while providing suggestions of specific cons to consider. If the report is a progress report, indicating a follow up assessment has been made, this section provides feedback about the users' progress (or lack of progress) through the stages of change.

The following is an example of the first two paragraphs of the first section from a progress report:

CURRENT STATUS AND STAGE OF CHANGE

Your answers on the last questionnaire show that since you started the

Pathways to Change program, you have moved from the PRECONTEMPLATION stage to the CONTEMPLATION stage of change. This progress indicates that you have started to form a new outlook on your smoking habit and you are thinking about quitting smoking in the next few months. This is a good sign.

However, if you want to continue making progress, you need to increase your awareness of the negative features or CONS of smoking even more. Compared to other contemplators who have been successful in the past, your awareness of the CONS is too low. You still seem to be thinking too much about the positive features or PROS of smoking. In addition, you appear to be having many temptations to smoke. Try to take a close look at the ways you may still be defending your smoking habit. Finding excuses for your smoking will keep you in the contemplation stage too long.

The second section focuses on Process of Change use.

Feedback on the processes of change is stage-based in that within longitudinal studies, differential process use across the stages of change has been linked to successful quitting. There are five to seven processes of change which are relevant to each of the first five stages (Precontemplation, Contemplation, Preparation, Action and Maintenance) and the Relapse "state." (Relapse cannot be reached by a smoker until the first followup. Each report contains feedback on the processes relevant to the users' stage of change. Feedback concerning the user's process use is based on comparison with a normative sample of users in the same stage of change who ultimately became nonsmokers. If the report is a following progress report, the user is provided with ipsative feedback as well.

The following is an example of the feedback on two processes (Environmental Reevaluation and Counterconditioning) from a progress report to a user currently in the Contemplation stage: CURRENT USE OF CHANGE PROCESSES:

Notice How Your Smoking Affects Others: This process concerns understanding the effects that smoking has on other people and on the environment, including the benefits that would result if you were to quit.

Your answers on the last questionnaire show that you have made progress noticing the effect that your smoking has had on others. Moreover, you are thinking sufficiently about the concerns and needs of others as they relate to your smoking. In this area you compare favorably to others who successfully progressed to the action stage of change.

Use Substitute: This process involves finding alternative thoughts and behaviors to substitute when you feel tempted to smoke. Your answers on the last questionnaire show that you have increased your use of alternative thoughts and behaviors when you feel tempted to smoke. However, you need to develop more alternatives or find more substitutes for your smoking. In this area you do not compare favorably to others who successfully progressed to the action stage of change. Feedbacks may recommend either increased use of an under-utilized process or decreased use of an over-utilized process.

The third section of the report provides particular stage specific strategies which have been found useful with regard to moving smokers to the next stage of change. For example, those in the Contemplation stage are encouraged to take a small step toward changing their smoking behavior, like delaying the first cigarette of the day for 30 minutes.

Again, these suggested strategies are based on data collected from successful quitters.

The following example is the third section for a progress report for a user currently in Contemplation:

SUGGESTED STRATEGIES

We have found in our work with smokers like yourself that moving to action with regard to quitting can be a difficult step. Our work with smokers ready for action has shown us that capitalizing on commitment right now requires effort, planning, and support from others. We also know that taking a small step in the quitting process can seem more manageable to smokers, and that many feel breaking down the quitting process into manageable steps is a more realistic approach.

We suggest using the following strategy in the next month: 1. Stop smoking for 24 hours in the next 30 days. We have found that people who manage to quit for a 24 hour period are much more successful in quitting smoking in the long run than users who cannot reach this goal. While this strategy may not seem important, we think it can help you use the PROCESSES OF CHANGE. Not smoking for 24 hours can indicate what it will be like when you eventually quit. This strategy can also point out problems or diffficult situations that you may have when you make your final quit attempt.

The fourth and final section of the report contains information pertaining to the user's temptation to smoke. Specifically, the user is given feedback pertaining to what appears to be the situation in which the user is at most risk with respect to temptation to smoke, and suggestions as to how best to cope in the high risk situation.

The following is an example of the first four paragraphs of the fourth section of a progress report to a person who has relapsed:

HIGH RISK SMOKING SITUATIONS:

As you recall, we asked you to answer some questions about particular situations in which you were most tempted to have a cigarette. Based on your answers, the following pattern emerged:

Many smokers who quit cigarettes start smoking again after a period of time. Since you have relapsed and are now smoking, you may feel frustrated and doubtful about your chances for quitting successfully in the future. We suggest that you do not give more attention to your relapse than you do to your former success. The first step towards making another quit attempt involves having a realistic attitude about your relapse. Think about the progress you have already made as you continue to move towards your goal of becoming a nonsmoker. The second step in getting back on track involves taking a close look at the situation that tempted you to light up that first cigarette after you had quit. Three kinds of situations seem to be most tempting for smokers. One involves social occasions where the focus is on relaxation or celebration. Activities can range from getting together with a friend who smokes to attending parties, weddings or other social events. A second tempting situation for smokers is any event that results in having negative feelings such as anger, frustration, depression or anxiety. A third tempting situation occurs when an ex-smokers is having a physiological craving or strong urge to smoke that may be related to a daily habit. For example, many smokers who light up a cigaretted as soon as they wake up in the morning may associate their smoking with their idea of how they start their day.

We suggest that you think about the situation that led you to start smoking again. Ask yourself these questions:

Who were you with?

Where were you?

What time of day was it?

Was it a special or ordinary occasion?

It is important to emphasize that each report generated by the expert system is tailored to the user based on his or her responses to the assessment instruments described above .

Referring to Figures 1, 2 and 3, the following is a description of the preferred embodiment of the expert system of the invention as a 'batch system' including a detailed description of the required input and potential output associated with the system, the required computer hardware and software, and the basic components of the computer program which supports the intervention.

Computer Hardware

Any state of the art computer hardware can be used to implement the invention. Data can be entered in any fashion as long as it results in an ASCII file which conforms to a designated format.

Computer Programs

The source code for the computer programs are set forth in the accompanying code listing (microfiche). The following is a general discussion of the software.

System Flow Overview

A provider's DBMS shall be the parent task, or controlling procedure, in the execution flow. The Stet DBMS shall select users that are due for intervention processing. It shall build the input file for the Smoking Intervention System, expert system software of the invention, (SIS) and then call the SIS. When the SIS has completed, it shall return control to the Stet DBMS. The Stet DBMS shall then initiate MS-Word to produce the intervention reports. When Word has completed, the Stet DBMS regains control and can perform any close-out processing required. By having the Stet DBMS as the parent task, the provider's Stet DBMS has complete control over the processing flow.

Intervention Periodicity

The periodicity of the interventions is defined by the CPRC as a 90 day cycle. This shall be controlled by the Stet DBMS and validated by the SIS software. Intervention periods of less than 90 days are disallowed. Intervention periods of greater than 90 days are beyond the control of either the provider or the CPRC. This is totally dependent upon the cooperativeness of the users.

The 90 day period shall be defined as the interval between the points at which the assessment forms are generated Therefore, 3 months after the baseline assessment is sent to the users, the second, or 3 month assessment shall be sent. And three months after that, the third assessment form shall be sent. Only 3 intervention reports are customarily provided for each user, however, SIS can process all valid data sent. The provider can control this with the printer control flag in the identifier record. All data will be accepted, analyzed and reported on, but only when the printer control flag is set, will a report be generated - - marked up as follows: SIS: Smoking Intervention System - any embodiment

CPRC: James Prochaska et al.

Provider: Health Care provider, e.g. Johnson & Johnson, Harvard Community Health Plan, etc.

User: Client, patient, smoker who wants to quit. Data Flow Overview

The source code for the system of the preferred embodiment of the invention is set forth in the microfiche appendix entitled Behavioral Modification Program Particularly for Smoking Cessation Batch System, James O. Prochaska et al.

The expert system software is designed to process the data from 1 to N users in any one instantiation.

The provider's DBMS generates one file: the Assessment Data Input File (ADIF). The SIS generates the Assessment Data Output File (ADOF), the Intervention Report Output File (IROF), and the Exception Data Output File (EDOF).

The ADIF contains the name and address records, an identifier record and assessment records (data from the surveys) for each user. The ADOF shall contain the same identifier record as the ADIF along with the completed assessment records and output data from the SIS. The ADOF data is stored by the provider's DBMS for retrieval at the next user assessment. The IROF contains an identifier record similar to the one in the ADIF along with records containing path and file names and output data from the SIS. The IROF is input to MS-WORD, the intervention report generator. The EDOF contains references to the ADIF records that are found to be in error.

Data Flow for Baseline and Follow-up Processing

At all assessment/intervention periods the normative or current input data is copied from the assessment forms, processed and stored by the DBMS, and then passed to SIS via the ADIF. At the first assessment (T₁), the ipsative, or previously derived data does not exist. SIS expects default values in these data fields. SIS processes the current input data, and then stores the current input data and the currently derived data in the ADOF. At time T_n+1, the PDBMS stores the previous input data and previously derived data in the previous data fields in the ADIF, and the current input data in the current data fields. SIS software generates and stores the current input and currently derived data in the ADOF for processing at time T_n+2. This process is repeated at each assessment/intervention period. In this fashion, the previous input and derived data was the current input and derived data from T_n-1. This process was depicted in Fig. 3 SIS Baseline And Follow-up Data Flow.

Error Protocol

The DBMS validates the user responses and ensures that the data is complete. The SIS re-screens the data to prevent inaccurate or incomplete reports from being generated. The initial task is an attempt to clean or correct incomplete or inconsistent responses. Failing that, error messages are generated when invalid states exist in the data. The error messages reflect the data restrictions defined in the file/record layouts section. The error messages have three primary components: user id, a message number and the message text. The message text identifies the item and the error.

Two return codes are supported by SIS; "0" indicates normal system completion, "1" indicates that a non-recoverable system failure has occurred. A return code of "0" does not preclude error messages from being present in the EDOF.

MS-WORD Report Generation

Microsoft Word is used to assemble and generate the reports from the IROF. This requires that the input paragraph files be formatted as true Word ".doc" files. Any report generator can be substituted for MS-WORD, however, the files will have to be converted into a format compatible with that report generator. Additionally, that report generator will have to be able to read and parse the IROF. File/Record Layouts

The files are constructed with one user record and one set of records related to the current assessment. The ADIF shall also contain a second set of records pertaining to the previous assessment. The assessment record sets are organized in ascending order on record number. The basic structure is as follows:

Address Record 01

.

Address Record 04

Identifier Record

Session N, Record 01

Session N-1, Record 01

Session N, Record 02

Session N-1, Record 02

.

Session N, Record M

Session N-1, Record M

All descriptive data references made are in relation to the expert system software. There are four Classes of data: Current Input (CI), Currently Derived (CD), Previous Input

(PI) and Previously Derived (PD). Current Input (CI) fields are provided by the Stet DBMS from the user's profile or assessment form. Previous Input (PI) fields are the CI fields from the previous assessment period. Currently Derived (CD) fields are derived by SIS from the data in the CI fields. Previously Derived fields are the CD fields from the previous assessment. The Stet DBMS shall set CD fields to blank(s) unless otherwise noted.

There are some fields labeled CI/D. These items are input items whose values are changed due to inconsistencies in the responses. The Stet DBMS stores the newly updated values. The Stet DBMS must provide the updated values to SIS for ipsative processing. Each record will have a flag designating if any data in that record was "cleaned" during processing.

Each data field has a Status: Required or Optional. CI data fields are applicable at all time points. CD data fields are generated at all time points. PI and PD data fields are applicable at all time points except baseline. Fields that are not filled shall contain blanks as the default value unless otherwise noted. CI, CD and CI/D data fields become PI, PD and PI/D data fields, respectively, at the next assessment/intervention period. No records are optional.

The assessment items are referenced by their respective page and question numbers.

Greater detail regarding the particular format of the questions asked and the various tailored responses provided in response to those questions are described in the various publications authored by either Prochaska and/or Velicer. The present invention is embodied in the specific handling of the information and providing of the output. Therefore, the specific questions asked and specific responses provided are not described in detail.

Operations Notes

SIS is executed by the call: SIS {DOS path name}. The parameter can be omitted, wherein the default directory, C:\SIS\, is used. The path name in the parameter list cannot exceed 30 characters and must be terminated with a backslash (\). The data files referenced and defined previously have the file names as shown in the table at the right. These files are expected to appear in either the default directory,

or the directory specified in the parameter list, if it Assessment Data Input File: ADIF:DAT is used. The directory is the root directory for the Assessment Data Output File: ADOF.DAT project. The ADIF is the output file of the PDBMS and Intervention Report Output File: IROF.DAT the input file to the expert system SIS. The ADIF can Exception Data Output File: EDOF.DAT contain from 1 to N sets of participant subject data.

The ADOF, IROF and EDOF are output files of the expert Table 5: File Name Cross Reference system. The ADOF and EDOF are input files to the PDBMS, and the IROF is the input file to the report generator. The output files are overwritten each time SIS executes.

There are a variety of system and unit initialization files placed strategically throughout the system. These files are not described here as there are "no user serviceable parts" inside. Only one file is of any interest to the provider. The file, SCTL.DAT, contains the paragraph filename extension ".DOC". If the provider replaces MS-WORD with a different report generator, and needs to change the extension, this change is made in SCTL.DAT. The length is fixed at 4 characters, must begin with "." (if the extension isn't null), and must be left justified.

System Messages

There are two basic classes of messages issued by SIS. System messages are identified by the system id: "SYS STATUS". Participant messages are identified by the participant id. System error messages are considered to be fatal, and the responsible CPRC S/W Engineer must be notified. Participant error messages pertain to circumstances that only affect the individual participant. These will not cause the system to abort, only the processing for that participant will stop.

Messages containing a 0 in the demonstrated positions (xxxx0##0) of the message id are initialization messages. All other messages containing a "0" in the demonstrated position (xxx0###) of the message id are system errors, and are usually fatal. The only exception to this is "PART0202". Though it indicates a system failure, the system can recover. The failure pertains only to the participant whose id is in the message, and only that participant's processing shall fail. The message "SMSG000" indicates a non-recoverable system failure has occurred that is external to and beyond the control of SIS. This message is displayed on the monitor. "SIS Terminated" is displayed just before SIS returns control to its parent task. This message is also displayed on the monitor. The messages are listed in the following figure.

In an alternative embodiment of the invention, as shown in Fig . 4 , the administrator of the system controls all aspects of the invention . The record layout follows and the software is set forth in the enclosed microfiche appendix entitled Behavioral Modification Program Particularly for Smoking Cessation Interactive System, James 0. Prochaska et al .

Purpose

The interactive or schoolbased system runs on a computer and the computer asks the person a question on the screen. The user sits right in front the computer and uses a mouse and answers the question presented on the screen by pressing yes or no or numeric buttons 1, 2, 3, 4 or 5. The computer collects the data and produces the feedback report right on the screen.

This program accepts one set of user responses and generates the intervention report for that user in real time. The system accomplishes its task by reading a variety of initialization data files that contain system control data systems, and this user's individualized input/output file. Input/Output

SCSUBACT.DAT contains the user information, name, age, gender etc. along with the user's responses to the questions, the names of the files containing the text of the components of the intervention report, and the timing data. Additional Output

A copy of the user's data is written to a file on the computer's hard disk and called SCxxxxx.DAT, where xxxxx is replaced by the user's unique id. This file serves as a cautionary backup nd is used for subsequent data analysis.

SCYP Expert System (SCYP)

Subject File Data Record Layouts

The input/output file on the user's personalized disk is called SCSUBACT.DAT

Record 00 is initialized prior to the user's assessment. It is the only record in the file on the subjects personal disk, prior to baseline processing. Records 01 through 24 contain the assessment and expert system data. This includes the raw scores, sums, paragraphs and path names, and timing data. The following tables contain the specific contents of these records. None of the raw scores will be missing, since the users are required to answer all questions as they proceed through the assessment.

Whereas all of the data on the assessment records are managed specifically for data analysis purposes, the data in record 00 is restricted. The fields: Subject Id, Grade, Gender and Birth Date are the only ones that are designed to be available for data analysis. The other fields are specifically related to the operation of SCYP and are subject to change. The definitions table will be helpful in understanding the record formats.

The file is constructed with one subject record and multiple sets of assessment records. The assessment record sets are organized in ascending order on session number. The basic structure is as follows:

Subject record

Session 1 Record 01

Session 1 Record 02

.

Session 1 Record 24 Session 2 Record 01

Session 2 Record 02

.

Session 2 Record 24 Session 3 Record 01 Session 3 Record 02

. . .

Session 3 Record 24

The foregoing description has been limited to a specific embodiment of the invention. It will be apparent, however, that variations and modifications can be made to the invention, with the attainment of some or all of the advantages of the invention. Therefore, it is the object of the appended claims to cover all such variations and modifications as come within the true spirit and scope of the invention.

Having described our invention, what we now claim is:

Claims

1. A method of processing behavior modification data which comprises:

inputting raw data based on an individual's response to a series of questions into a computer;

creating an assessment data input file which includes at least normative assessment data;

assessing the variables to provide information corresponding to the individual's attitudes and behaviors; and matching the information to the individual's stage of change, which stage of change is based on current process of change use and which is responsive to a balance of pros and cons of the behavior to be corrected;

creating an assessment data output file which comprises at least normative assessment data and normative paragraph names for the intervention report;

creating an intervention report output file which contains the paths and file names of feedback paragraphs; generating an individualized intervention report providing feedback on current attitudes and behaviors relating to the behavior to be modified, the report comprising a firsttime assessment as well as comparisons to a normative sample of individuals who have successfully modified their behavior.

2. The method of claim 1 wherein said report is a first assessment report generated at time T₁ and which comprises: generating a second assessment report at time T_n+1, said second report including ipsative feedback to provide concrete commentary on an individual's changes since a previous assessment.

3. The method of claim 1 wherein the stages of change for the initial report comprise precontemplation, contemplation and preparation.

4. The method of claim 3 wherein subsequent assessment reports include the additional stages of action and relapse.

5. The method of claim 2 which includes providing a second assessment report, which second assessment report is based on:

a) storing a time T₁ previous input data (normative) and previous derived data (ipsative) in a previous data field in the ADOF;

b) storing at time T_n+1 current input data and current derived data in a current data field in the ADIF;

c) storing in the ADOF processing at time T_n+2 the current input and current derived data;

d) repeating steps (b) and (c) whereby the previous input and derived data was the current input and derived data from T_n-1.