A menu of potential reinforcers in a methadone maintenance program

Journal of Substance Abuse Treatment, Vol. 11, No. 5, pp. 425-431, 1994 Copyright 0 1994 Elsevier Science Ltd Printed in the USA. All rights reserved OiW-5412/94 $6.00 + .OO

0740-5472(94)00053-O

ARTICLE

A Menu of Potential Reinforcers in a Methadone Maintenance Program JOY M. SCHMITZ, PhD,

HOWARD RHOADES, PhD, AND JOHN GRABOWSKI, PhD

Department of Psychiatry and Behavioral Sciences, The University of Texas Medical School at Houston and Substance Abuse Research Center

Abstract- This study demonstrates the use of paired comparisons and interval scaling techniques for m-ring the relative priority of program privileges available at a methadone maintenance clinic. Fifteen methadone program privileges were combined in all possible pairs (N = 105) on a reinforcer menu and administered to a group of 12 methadone patients and a second group of counselors (IV = 4). Data were converted to interval scales using the law of comparative judgment to form a quantitative continuum from least to most preferable. Free methadone, free dental service, and more take-homes were ranked highest in both groups; however, patients showed less differentiation in their preference for these privileges. Dose decreases were least preferred. Results are discussed in terms of their clinical applicability in identifying privileges for potential use in modifVing the behavior of drug abusers. The method of paired comparisons has excellent psychometric properties and may offer some advantages over other response scale formats. Keywords-methadone maintenance; reinforcer; contingency; paired comparisons; interval scaling; behavioral-pharmacological interventions.

INTRODUCTION

rangements to influence target behaviors, including counseling attendance (Stitzer & Bigelow, 1976), opiate use (Hall, Bass, Hargreaves, & Loeb, 1979; Milby, Garrett, English, Fritschi, & Clarke, 1978; Stitzer, Bigelow, & Liebson, 1980), benzodiazepine use (Stitzer, Bigelow, & Liebson, 1979; Stitzer, Bigelow, Liebson, & Hawthorne, 1982), alcohol use (Liebson, Tommasello, & Bigelow, 1978), and other illicit polydrug use (Iguchi, Stitzer, Bigelow, & Liebson, 1988; Magura, Casriel, Goldsmith, Strug, & Lipton, 1988; Stitzer, Iguchi, & Felch, 1992). The establishment of an effective contingency management program is based on the identification and availability of potent incentives. Questionnaire survey methods have been used to’ gather information about the desirability and potential utility of program privi-

CONTINGENCY REINFORCEMENT PROCEDURES have been used in clinical settings as a means to increase desirable target behaviors. Outpatient methadone maintenance programs provide a suitable context for the application of these behavioral procedures. Methadone programs consist of many potential reinforcers or privileges that have been used effectively in contingent ar-

Preparation of this manuscript was supported by Grant DA 06143-01 from the National Institute on Drug Abuse to John Grabowski. The authors thank the staff of the Addictive Behaviors Clinic and the Substance Abuse Research Center for their support and cooperation. Requests for reprints should be addressed to Dr. Joy M. Schmitz, University of Texas Mental Sciences Institute, 1300Moursund, Houston, TX 77030.

Received April 23, 1992; Revised April 25, 1994; Accepted May 16, 1994. 425

426

JN.

leges in a methadone maintenance treatment setting. To date, the published survey studies (Stitzer 8z Bigelow, 1978; Yen, 1974) have utilized rank ordering assessment procedures in which clients are asked to list a number of privileges in order of importance or relative desirability. With this method, the subject’s judgment is used as a direct report of the value of the privilege on a linear subjective continuum of desirability. Although the method of rank ordering has the practical advantages of being straightforward, easy to construct, and rapidly administered and scored, the theoretical and methodological shortcomings of this method have been reported (Guilford, 1954). An alternative method of ranking items is the paired comparison approach (Guilford, 1954). This method involves the comparison of each stimulus (or survey item) with every other stimulus. From the number of times each stimulus is chosen, the stimuli can be rank ordered with more precise information than if all of the stimuli were rank ordered without the forced comparison. The method of paired comparisons is based on traditional psychophysical laws for evaluating stimuli on a given dimension. Psychometric studies have found the paired comparisons method to have advantages over rank ordering in terms of reliability, validity, and logical consistency (Lohaus, 1990; Penner, Homant, & Rokeach, 1968; Sutherland et al., 1989). Ordinal scaling provides useful information about the overall rank of objects but offers very little information concerning the relative distances between and among measured items. Mapping these comparisons onto an interval scale is considered to be a more meaningful and useful method for analyzing ranked data. When objects are assigned a value on a linear scale with interval properties, the relative size .of the distances between items has meaning and corresponds to differences in the amounts of the dimension of interest (e.g., desirability). The law of comparative judgment (Thurstone, 1927a, 1927b) provides the conceptual and statistical mechanism for generating interval scales. Since its formulation in the 192Os, there have been many applications of interval scaling using the law of comparative judgment (e.g., Davis &McLean, 1988; deSoete 8z Winsberg, 1993; Krus, 1986; Morse & Morse, 1988; Thurstone, 1928).

Schmitz

et al.

To demonstrate the methodology and utility of the paired comparison procedure and interval scaling, a “reinforcer menu” of clinic privileges was constructed and administered to a group of methadone maintenance patients. In previous rank ordering survey studies (Stitzer & Bigelow, 1978; Yen, 1974), the opportunity to earn medication take-home privileges was ranked as most desirable, whereas items such as use of recreational facilities in the center and a monthly party received lower rankings. These privileges, plus others that have not been previously evaluated, were included in the present menu. It was expected that the general ordering of privileges from highly ranked (e.g., take-home medication doses) to low ranked (e.g., monthly party) would be replicated in our independent sample of methadone maintenance clients. Unlike earlier survey studies, however, the present assessment method was expected to yield interesting and potentially useful information about the positioning of privileges on a quantitative dimension of perceived desirability. Finally, because behavioral contingency programs are often designed and implemented by treatment staff, we were interested in determining whether they have accurate perceptions of the reinforcers rated as important by the clients. METHOD Subjects Subjects were 12 clients who were attending an outpatient drug abuse treatment clinic at the University of Texas Department of Psychiatry and Behavioral Sciences Mental Sciences Institute at Houston. This sample was randomly drawn and represented 50% of the patients receiving methadone at the clinic. The 12 clients (4 females, 8 males) had an average age of 38.1 years (SD = 3.56) and averaged 10.9 years of education (SD = 1.86). They were predominantly White (16% Black, 8% Hispanic), and 50% were employed. Treatment characteristics of the surveyed clients are shown in Table 1. A second sample of subjects consisted of 4 counselors (1 male, 3 females). All of the counselors were in contact with the methadone clients on a weekly basis.

TABLE 1 Treatment Characteristics of Study Participants (N = 12) Variable Methadone dose (mg) No. of weekly take-homes Duration of prior methadone treatment (months) Duration of methadone treatment in present program (months)

M

SD

Range

66.60 3.50 69.60 12.36

25.07 1.24 42.53 5.04

30-l 00 o-4 12-l 56 5-20

427

Methadone Program Privileges

The background treatment setting was one in which behavioral-pharmacological interventions were prominent components. Patients were maintained on a treatment plan that provided dosage change (up or down) and take-home doses contingent on urine samples free of illicit drugs. Questionnaire Items appearing on the reinforcer menu are shown in Figure 1. Six of the 15 items selected for inclusion on the reinforcer menu were consistent with those found on other reinforcer questionnaires (Stitzer & Bigelow, 1978; Yen, 1974). Additional items, generated via informal interviews with clients and treatment counselors, were selected on the basis of their relevance to the treatment setting and practical considerations. A multiple paired comparison, forced-choice format (Guilford, 1954) was used to design the reinforcer survey. Items were arranged in pairs so that each item was paired with every other one, resulting in n (n - 1)/2 = 105 comparison pairs. The order of the items in the

Patients

pairs was varied so that each item appeared equally as first or second, to control for space error. No item was presented in two successive pairs. These procedures for arranging pairs are consistent with the guidelines recommended by Ross (1934). Subjects were instructed to indicate which item of each pair they preferred. Counselors completed the same survey; however, they were instructed to respond according to how they perceived the clients would respond. Procedure The reinforcer menu was completed by all subjects during a l-week period. The research assistant provided verbal explanation and instructions upon administration of the surveys. All surveys were completed independently and remained anonymous. Determining Scale Values The paired comparison data were converted into scales based on the law of comparative judgment to subjec-

counsekrs

1. Adjmtownbvd* 2. Doaedeueawe 3. Doclshcrewm 4. Eams1operweek* 5. Fewwcwudng-* 6. Free-ticket8 7. Freedmtalwvice 8. Freemehdone 9. Fmebmpaws 10. Frwmovbjmstm 11. Jobasawanw 12. kdkatbnatdffemlttlme 13. MonthlyPatty* 14. Mcre~aelMbn8* 15. Moretake-hmms*

FIGURE 1. Interval scalings of preference of privileges by patlents and counselors.

428

J.M. Schmitz et al.

tive values (e.g., Coombs, 1967; Thurstone, 1927a, 1927b). To construct interval scales, a matrix was formed from each set of data showing the proportion of subjects that preferred each privilege in comparison to each of the other privileges. The data from the patient group are presented in Table 2, with each cell containing the proportion of patients who judged the column privilege to be preferable to the row privilege. It is assumed that each item would be judged greater than itself half of the time, so .50 goes in each diagonal of the table. Next, each proportion was translated into its equivalent normal deviate by looking in a table of areas under the normal curve. The sum of normal deviates and the average for each column was obtained. To make the scale values all positive, an arbitrary origin was located at the privilege that was judged as least preferred. For both groups, this privilege was dose decreases (Item 2). Figure 1 illustrates the interval scaling of items obtained from each set of data. The numerical code is the alphabetical order of privileges.

wmw0wc9rrwc9w0m-

r-wnlw@Jw~cnwwwmc9* . . . . . . . .

.

.

.

.

.

.

-l-9wwwrrm0-0-wwl~w-rNol~wLoozlLobelol~ . . . . . .

.

.

.

.

.

.

.

.

.

.

’ . ‘0

GG . .

.

wwmc9w-owww

0

ooc9wwbmo-. . . .

0

w‘0

25:.

.

owrwww

5:

7

2

mrtl-00

.v- . . . . .

?

‘0

RESULTS

The scale values obtained from patient’s data indicate that free methadone (Item 8) was the most preferred item, followed by free dental service (Item 7) and more take-home doses (Item 15). As mentioned before, the distinguishing feature of this scaling methodology is that it permits interpretation of the differences between ordered items. For instance, the scale values obtained from patient’s data show a relatively small distance between free dental service and more take-home doses (1 scale unit), suggesting that these two privileges are perceived as being very similar in preference. Similarly, only two units separate the privilege of earning $10 per week (Item 4) and medication at a different time (Item 12). The greatest interval between items was found on the lowest ranked end of the continuum, where dose decreases were least preferred to all other privileges. Other lower ranked privileges were less differentiated. For example, free baseball tickets (Item 6) and free movie passes (Item 10) had nearly the same scale value as did more counseling sessions (Item 14) and monthly party (Item 13). In contrast to patients, counselors ranked more take-home medication (Item 15) higher than free dental service (Item 7), with the number of intervals between these rankings being greater than that found in patient rankings. Most of the items on the counselor’s scale were ranked lower in relation to their ranked position on the patient’s scale. One exception was dose increases (Item 3), which assumed a higher position on the scale for counselors. For counselors, monthly party (Item 13) and more counseling sessions (Item 14) were in close proximity to the lowest ranked privilege, but patients showed greater relative spacing among these items.

22 . . 0000

:g

5:

‘00

8

‘000

0wwwm0r00wwwwww wwoonJwulwmwo--oNw . . . . . .

.

WOC7lCuwmQ) .

.

.

.

.

.

.

.

.

wc9-mow(3

l-wcrlwww~ .

77

F

mm0c7wwww c7~wmwr00

Fj5j

’

000

%a .

.

.

:

. .

‘0

.

womwowwwwww 0wwwwul0r00w

. 0

.

’ ’ . .

.

Metkadone Program Privileges

The greater sensitivity of interval scaling is also seen in the correlations between patient and counselor-rated privileges. The Spearman rank correlation was r = .88 (a < .00002), and the Pearson product-moment correlation was r = .92 (p < .OOOOl).Both samples identified free methadone (Item 8) as the most preferred privilege and dose decreases (Item 2) as the least preferred privilege. DISCUSSION

The paired comparisons approach used in the present study represents a departure from traditional scaling methods used to identify potential reinforcers in methadone treatment settings. Evaluative studies on the methodological effects of various response scale methods have found the paired comparisons method to be superior to rank-ordering procedures and rating scales in terms of reliability and validity (Cohen & VanTassel, 1978; Lohaus, 1990; Lohaus & Trautner, 1989; Reynolds & Jolly, 1980). The systematic comparison of all possible pairs is viewed as a more thorough method of ordinal measurement (Nunnally, 1978). Despite its long and distinguished history in psychological assessment and survey research, the method of paired comparisons has been used less extensively than other response scales. Perhaps the primary objection to the method of paired comparisons is that it can be time and labor consuming for the judges, as well as the investigator. This is particularly true when the number of stimuli to judge is large. However, Lohaus (1990) compared the paired comparisons procedure and rank ordering of 20 items (190 pairs) and found that subjects preferred the former task. Apparently having to compare multiple objects nearly simultaneously, as in the rank order procedure, appears to be more complex than having to make multiple comparisons of two objects at a time. Methods to reduce the number of pairs of stimuli, without serious loss in the accuracy of the results, have been proposed (e.g., Olson, 1978). Computer programs for calculating, storing, and analyzing scale values from paired comparisons data are available (e.g., Molina, Melia, 8z Sanmartin, 1993). As shown here, interval scaling models can be easily applied to data from this type of measurement. Using interval scales, it is possible to evaluate the relative position of ranked items, as well as interpret distances between items. For example, our data indicate that, for patients, the privileges of receiving free dental services and more take-home doses were essentially similar in preference, as were the privileges of earning money per week and receiving medication at a different time. In practice, this finding suggests that items in close proximity on the scale might be interchangeable or effective substitutes for each other in a contingency management arrangement. This type of

429

information is lost when simple ordinal scaling methods are used. The present findings replicate and extend those of other researchers (i.e., Stitzer & Bigelow, 1978; Yen, 1974) in that privileges such as earning money, receiving take-home methadone, and self-adjustment of medication dose were ranked as much more desirable than dose decreases, more counseling sessions, and monthly parties. Treatment staff involved in the design of contingency management programs did identify the privileges most preferred among patients. The privilege of receiving free methadone from the clinic, an item not included in previous reinforcer menus, ranked higher than receiving more take-homes. Patients in the present program were paying for their methadone on a weekly basis ($20). These treatment conditions may have increased the relative desirability of free methadone. A comparison of free methadone to increased take-homes has not been experimentally evaluated; however, the reinforcing efficacy of monetary consequences has been demonstrated. Stitzer et al. (1980) found that cash rewards ($15) proved to be more effective than take-home privileges and a dose self-regulation option when offered within a contingent reinforcement procedure to promote reductions in drug use. In the present study, patients’ valuation of free methadone in terms of its monetary amount ($20) is consistent with the lower preference for the privilege of receiving $10 per week. In an ongoing follow-up study, we have equated these two privileges by including the item of receiving $20 per week on the reinforcer menu. The relatively high ranking of free dental service was a somewhat unexpected finding. Previous studies have not included this privilege on reinforcer menus. Empirical studies are needed to determine whether the availability of the spectrum of health and social care services can be used effectively as contingent reinforcers in a methadone treatment setting. It should be recognized that there is considerable variability among methadone programs in such vital areas as structure, staffing patterns, dosing procedures, paying arrangements, professional orientation, and availability of ancillary services. With regard to the generalizability of these findings to other samples, it is possible that the relative rankings of methadone program privileges are influenced by the treatment environment. The treatment context of the present study was one in which take-home privileges were strongly established as a reinforcer, with most patients having a history of earning substantial take-home opportunities. In other clinical settings where there is less emphasis on manipulating reinforcers and more emphasis on psychotherapeutic relationship development, a different set of rankings might be expected. The reinforcer menu we have developed offers a new way to compare the positioning of potential reinforcers across

J.M. Schmitt at al.

programs and populations. In our work with cocainedependent patients, for example, we have begun to utilize the reinforcer menu to compare the rankings of program preferences across drug-abusing patient populations (Grabowski, Rhoades, Elk, Schmitz, & Creson, 1993). The data collection and scaling procedure demonstrated here can be used as the first step in discovering potential reinforcers for use in behavioral treatment programs for drug abusers. This scaling method can be used in future research to answer specific questions regarding the hierarchy of preferences and to test hypothesized differences between items within a sample or across different samples. Another important area that could be explored is the issue of stability. Knowing that priorities change over time would be useful information for the clinician involved in the development of a contingency management treatment program. The critical question, however, concerns the practical implications of the scale points. Preference, as stated on the survey, provides an indication of the relative potency of each program privilege ranked against each of the others. Actual effectiveness remains an empirical clinical issue. The application of contingency management procedures in outpatient methadone maintenance settings has received favorable support over the years (e.g., Baldridge, McCormack, Thompson, Zarrow, &Primm, 1974; Stitzer et al., 1977; Stitzer et al., 1980). Because the successful application of these procedures depends on the identification of highly valued reinforcers, assessment studies such as the present one appear to be important in the development and evaluation of contingency management programs. Ultimately, the usefulness of a reinforcer menu is dependent on the effectiveness of the privileges in modifying the target behavior.

REFERENCES Baldridge, P., McCormack, M., Thompson, L., Zarrow, A., & Primm, B.J. (1974). Providing incentives to successful methadone patients. New York State Journal of Medicine, 74, 11l-l 14. Cohen, AS., &Van Tassel, E. (1978). A comparison of partial and complete paired comparisons in sociometric measurement of preschool groups. Applied Psychological Measurement, 2, 31-40. Coombs, C.H. (1967). Thurstone’s measurement of social values revisited forty years later. Journal of Personality and Social Psychology, 6, 85-91. Davis, T.M., & McLean, J.E. (1988). Simplifying ranking tasks in survey research: A method and example. Psychological Reports, 62, 987-992. deSoete, G., & Winsberg, S. (1993). A Thurstonian pairwise choice model with univariate and multivariate spline transformations. Psychometrika, 58, 233-256. Grabowski, J., Rhoades, H., Elk, R., Schmitz, J., & Creson, D. (1993). Clinic wide and individualized behavioral interventions in drug dependence treatment. In L. Onken, J. Blame, & J. Boren

(Eds.), Behavioral treatments for drug abuse and dependence (NIDA Research Monograph 137, pp. 37-72). Washington, DC: Department of Health and Human Services. Guilford, J.P. (1954). Psychometric methods (2nd ed.). New York: McGraw-Hill. Hall, S.M., Bass, A., Hargreaves, W.A., & Loeb, P. (1979). Contingency management and information feedback in outpatient heroin detoxification. Behavior Therapy, 10, 443-451. Iguchi, M.Y., Stitzer, M.L., Bigelow, G.E. & Liebson, LA. (1988). Contingency management in methadone maintenance: Effects of reinforcing and aversive consequences on illicit polydrug use. Drug and Alcohol Dependence, 22, 1-7. Krus, D.J. (1986). Russian cannibals: The case of the downed Korean airliner. Psychological Reports, 59, 3-9. Liebson, I.A., Tommasello, A., & Bigelow, GE. (1978). A behavioral treatment of alcoholic methadone patients. Annals of Internal Medicine, 89, 342-344. Lohaus, A. (1990). Response scales in developmental psychology: Empirical results of a comparative study. Journal of Genetic Psychology, 151, 473-481. Lohaus, A., & Trautner, H.M. (1989). Information integration by children: The identification of rules by an alternative method. Genetic, Social, and General Psychology Monographs, 115, 329-347.

Magura, S., Casriel, C., Goldsmith, D.S., Strug, D.L., & Lipton, D.S. (1988). Contingency contracting with polydrug-abusing methadone patients. Addictive Behaviors, 13, 113-l 18. Milby, J.B., Garrett, C., English, C., Fritschi, O., & Clarke, C. (1978). Take-home methadone: Contingency effects on drugseeking and productivity of narcotic addicts. Addictive Behaviors, 3, 315-320. Molina, J.G., Melia, J.L., & Sanmartin, J. (1993). A BASIC program for calculating scale values using four unidimensional scaling methods in random subsamples. Educational and Psychological Measurement, 53, 103-105. Morse, J.M., & Morse, R.M. (1988). Cultural variation in the inference of pain. Journal of Cross Cultural Psychology, 19, 232-242. Nunnally, J.C. (1978). Psychometric theory (2nd Ed.). New York: McGraw-Hill. Olson, M.A. (1978). An applicatory matrix sampling to the method of pair comparisons. Journal of Educational Measurement, 15, 49-52. Penner, L., Homant, R., Br Rokeach, M. (1968). Comparison of rank-order and paired-comparison methods for measuring value systems. Perceptual and Motor Skills, 27, 417-418. Reynolds, T.J., & Jolly, J.P. (1980). Measuring personal values: An evaluation of alternative methods. Journal of Marketing Research, 17, 531-536. Ross, R.T. (1934). Optimum orders for the presentation of pairs in the method of paired comparisons. Journal of Educational Psychology, 25, 375-382. Stitzer, M.L., & Bigelow, G.E. (1978). Contingency management in a methadone maintenance program: Availability of reinforcers. International Journal of the Addictions, 13, 737-746. Stitier, M., Bigelow, GE., Lawrence, C., Cohen, J., D’Lugoff, B., & Hawthorne, J. (1977). Medication take-homes as a reinforcer in a methadone maintenance program. Addictive Behaviors, 2, 9-14. Stitzer, M.L., Bigelow, G.E., & Liebson, I. (1979). Reducing benzodiazepine self-administration with contingent reinforcement. Addictive Behaviors, 4, 245-252. Stitzer, M.L., Bigelow, G.E., & Liebson, I. (1980). Reducing drug use among methadone maintenance clients: Contingent reinforcement for morphine-free urines. Addictive Behaviors, 5, 333-340. Stitzer, M.L., Bigelow, G.E., Liebson, I.A., & Hawthorne, J.W. (1982). Contingent reinforcement of benzodiazepine-free urines:

Me&adone

Program Privileges

Evaluation of a drug abuse treatment intervention. Journal of Applied Behavior Analysis, 15, 493-503. Stitxer, M.L., Iguchi, M.Y., 8t Felch, L.J. (1992). Contingent take home incentive: Effects on drug use of methadone maintenance patients. Journal of Consulting and Clinical Psychology, 60, 921-934. Sutherland, H.J., Lockwood, GA., Minkin, S., Tritchler, D.L., Till, J.E., & Llewellyn-Thomas, H.A. (1989). Measuring satisfaction with health care: A comparison of single with paired rating strategies. Social Science and Medicine, 28, 53-58.

431 Thurstone, L.L. (1927a). A law of comparative judgment. Psychological Review, 34, 273-286. Thurstone, L.L. (1927b). The method of paired comparisons for social values. Journal of Abnormal and Social Psychology, 21, 384-400. Thurstone, L.L. (1928). An experimental study of nationality preferences. Journal of General Psychology, 1,405-425. Yen, S. (1974). Availability of activity reinforcers in a drug abuse clinic: Apreliminaryreport. PsychologicalReports, 34,1021-1022.