Joorttul o] Ahtu,rmal ('hihl I's.vchoh,gy. 1"ol. 3. No 2. i~Z5

Multiple Criteria Follow-Up o f Behavior Modification with Families ~ Stephen M. Johnson ~ and Andrew Christensen University of Oregon

ThB report presents term#tatkm and follow-up e,,aluations o f a parent training program based on behavior modiJication principles. Treatment termination outcome data were obtahwd fi~r 22 families, and fidlow-up data were collected on 14 o f these ]amilies 3 and 8 m o n t h s after treatment Results indicated a fairly high level o f ,~uccess at termbtathm on the basis oJ'parent-collected observational data, parent attitude change toward the ('hildren, and parent attitude concerning the process cvtd o u t c o m e o f treatment. Modest levels o f success were evidenced oil the basis o f behavioral data recorded by trahwd observers in the home. Families who participated in fidlow-up demonstrated greater cooperation mid involvement with the treatmettt prograot thatt those who did not participate. In all other respcc'ts, h~wever, these two gr~ntps lt,6,rr similar. F~dh~w-up data on parenl attitude measures demonstrated otahttettattce o f the treatment effects. Parettt observath)nal data were #zcomph'te. hut also showed maintenance in re)How-tip. For the subsample o f 14 ('ases, h o m e observatiott records htdicated a nonsigtliJh.attt de~'lilte hi r viant behavior at terotbtati~tt ] i d l o ~ ' d b.v a ttottsigtti/h'attt hlcrease itt deviant behavior durbtg fidh,w-up. The meatting attd #nplicatiotts o]" these dis('repattt ]htding3' were disr attd contpared with results f r o m other laboratories. In spite o f Cul increasing prolil~.:ration o f behavior modificalion programs for family problems, there have been relatively few follow-up evaluations of the eft'eetiveness of this form o f Ireatmenl. The importance of follow-up studies is widely recognized, bul lhc time and expense required seen~s to have discouraged Mantlscripl rcccwed m final form May 27. 1975. "l'hi.'~research was ';upporled by a Nalkm.'d Instilulr of Mental lleallh (,run[ M H

t9633.

The authors would like to e',,prc~; lhctr ;ff~pfcciulion to Ortn Bolslad. Sheila Eyberg, and Grelchen I.obit:, ",vh~;JS.,;isted in the completion of Various u.spccts of this work. : Requesls l'~r reprinl~ ,dlould be senl t~ Slephen M. J~flmson. Psychology ('linic, University of Oregon. tq,gene. ()rcgon 974(13. 135 ,,'1975 l~lelltl111 l~(lllll~.11111tl ( ' I I I | . ) O I . I I I ( ' H I . ? ~ / Wf~%I l l l h '~.tr~.~tt. N e w Y ( ~ r k . N . Y . 1001].. NO p i l r [ ()J t h i s I ' H I D I I C . ) [ I I ) I I I l l . i V lJ(2 l ( ~ l l l t ~ f ' l l l ( l ~ ( | . S t t ) I I ! d III .l t t ! l l l @ V ~ l l ~.VS|(7111, .~r t r . l n ~ , m l t t e d , in . l i l y ~ l l ; l ] (~.l b y , | l l V iiI(~.iti~, ~,ll(~,C~l(~(ll~. l l l ~ ' ( ] l J l l l l l I L ~ . l l I.)|I(I(I]C(iI~I~IIII[I. ITllCrOfll111111g. f @ C O l C l l n g . (If ()[IT(~IWIS~'. w l t l l o l l l Wrtl. H ' l l J)tTtllli%'~Itlll , ) 1 lht? l ) f l L ~ l l ~ h e l .

136

Johnson and Christensen

that research which would provide the most critical test of arty treatment procedure. Certainly tile most comprehensive follow-up investigations of behavioral interventions in fanfilies have been contributed by Patterson (1974a, 1974b), who has used home ob.scrvation data and daily parent reports on the occurrence of child rel~rral symptonts as outcolne criteria. In tile Patterson program, follow-up probes ~ccurred once a month lbr the first ~ months after termination and once every 2 months thereafter for an additional half year. In some cases, additional intervention was required during the follow-up period and PaIterson (1974a) reported an average per case of 1.92 hours of professional time devoted to such intervention. OI" tile 27 cases treated in the Patterson program, relatively complete follow-up honve observation data were available for 20 cases. Since parent-report data colleclion was begun later in the program, relatively complete data were available for only 14 cases. Both outcome measures revealed a significant drop in child deviant behavior from baseline to termination and, in general, maintenance of that behavior change throughout tile follow-up year. The parent data on occurrence of referral symptom,,; showed a nonsignificant trend toward greater improvetncnt during tile follow-up period. Ferber, Kecley, and Shcmberg (1~74t have recently reported far more discouraging resulls, however, wilh a program modeled after Patterson's work. Altllough syslematic behavioral observ',aion in tile home was conducted during interventiom~, follow-up was done by clinical interview at I month and by a telephone call to parents 1 year after lermination. Observation data was obtained olr only live of the seven families, and follow-up obtained on only four. The authors characterized their results as showing positive shtlrt-term changes for only Ihree t)l" the seven familics, and long-term positive changes for only one. This program, which was limited to 10 treatment sessions, was clearly less comprehensive lhzu~ lhe Patterson program, in terms of both professional time devoted to treatment and the magnitude of evaluation. More optimistic results were reported by Coe and Black (1972) who compared tire results of family operant therapy with the regular outpatient clinic treatment which olherwise would have been provided, and with a combination of the two tJeatmel~t apprt)aches. Four families were given family operant therapy while livc families were Irealetl ill each of the other two groups. Parentcollected data oil their childrcn's desirable arrd undesirable behaviors provided the sole criterion of orllconle. Ill ge~)eral, these data demtmsttated superiority for lhc Iw(~ gr()ttps in which f:ulfily operant treatment was employed, and these thcrapctHic gains were maintained at the 6-m,~.mth follow-up evaluation. Finally, Alexander and I)a~ons (1973) have provided somc very encouraging Iollow-tlp data tm the effects of behavioral intervention in families with delinquenl children. Ill Iris study, 46 families were treated with behavioral intervention and compared with families who received either client-centered therapy, psychodyrlamic family therapy, or m) treatment. At termination, the

Multiple Criteria Follow-Up of Behavior Modification

137

results of these treatment procedures were assessed through observations of family interaction in a discussion session. At that time, the family interaction measures revealed that families who had received behavioral intervention demonstrated greater equality in talk-time, less silence, and more positive interruptions than families in any of the other groups. More impressively, the follow-up data on delinquency recidivism taken up to 18 months after treatment indicated that recidivism was substantially lower for the behavioral intervention group than would have been expected on the basis of the county-wide rate (26% versus 51%). Furthermore, the two comparistm treatment groups and the no-treatment control group achieved an equ',d or greater recidivism rate than the county average. The superiority of the behavioral intervention group was statistically significant in all comparisons. Thus, the limited follow-up data available from three out of four laboratories are encouraging with respect to the long-term beneficial effects of behavioral intervention with families. However, two criticisms Can be made of these followup investigations: In only one study (Patterson, 1974a, 1974b) were multiple criteria employed; and in only one other (Ferber, Keeley, & Shemberg, 1974) were parents asked to report on their feelings about treatment and its effect on their subsequenl evaluation of the treated child. In a recent study from this laboratory (Lobitz, G. K., & Johnson, 1975), it was found that referred children could best be discriminated from a matched group of nonreferred children, not on the basis of their observed behavior, but on the basis of their parents' global attitudes about them. The parent-attitude measure correctly assigned 9 ~ of the sample of 54 children to the appropriate groups while the behavioral measures yielded a considerable degree of false assignment. In view of these findings, multiple criteria, including assessment of parental perceptions, seem warranted. The present report sumlnarizes outcolnr and follow-up results from a family behavior modification treatment program which was modeled in a general way after that described by Patterson and his associates (Patterson, Cobb, & Ray, 1973; Patterson, Ray, & Shaw, 1968; Patterson & Reid, 1973). The termination oulcoule results for the lirst 17 cases seen in this project were reported in ~tetail by Fybcrg and Johnson (1974). The present report summarizes termination results for 22 tamilies and follow-up rcsults at 3 and 8 months after termination for a subsample of 14 t'amilies. Multiple criteria outcome assessments of family behavior and attitude were obtained at each testing period.

METHOD

Sub~cots Twenty-two lhmilies having children reported to exhibit active behavior problems ill the hou]e participated in Ihc treatment program. Tire term active

!.]8

Johnson and Christensen

behavh,r problems was used to relier Io aggressiveness, destructiveness, disobedience, hyperactivity, telnper tantrums, or high-rate activity with annoyance value. Families were not accepted into this program if the problem child was severely retarded, had experienced severe and docunlented brain damage, or exhibited behavior problems which would ordinarily cause him to be labeled autistic or schizophrenic. All bul three of Ihe families had both parents in tile home ~uld all but two of the treated children were boys. The educational level of parents ranged from 8 to 21 years with the nlean of 12.3 years Ibr mothers and 13.4 years for fathers. Income level lot families ranged from $2,000 to $15,000 with a mean of $7,312. The lreated children ranged in age from 4 to 12 with a mean age of

7.85.

Prr~('e(htres These fanlilics were seen at the University of Oregon Psychology Clinic during a ,Z-year period helweerl 1970 and I t~73 by 18 graduate student therapists who saw client fmnilics in permanen! Iwo-pers~n teanls. 3 All families, except 'tll(~.~ on welfare, were required to pay for services on a sliding fee schedulc adjusted t~ ability to pay. Assessment was done prior to i.ntervention, al lernlinalion and al 3 and 8 nlonlhs following termination. Famflies were paid $30 for participating in Ihe second follow-up, but no payment was nlade al ally other assessnlenl period. As indicated earlier, Ihe treatment procedures were generally modeled after those outlined in some detail by Patterson and his associates (Patterson, Cobb, & Ray, 1'073; Pallerson, Ray, & Shaw, 1968; Patterson & Reid, 1973), and a general description of Ille procedures used may be found in Eyberg and Jolmson (I~)74). 4 This prt~granl, which was of all educational nature and based upon Ihe prirlciples of social leamitlg theory and techniques t~l"behavior modification, was limited h~ 12 weekly sessions. The lirst 17 cases were 'also employed for r slndy of Ihe el!eels of a contingency contracting procedure and effects 3Theral)iXls were (;ary Bircllier. Orin Bolsh|d, Joanne Valentine ('lark, Micllael Erickson, Palricia liine~, l)ennis Karpowitz, ('arole Kirkpalrick. Slephen K~pel, Laurie Lerner, Dalia I.e,:lie, (;relchen I.t~bilz, W. ('harie~, l,obiiz. James; McDonald. John Rc~bin.,;on, Jeffrey Slcger, John Vincenl, (;e~fffry While. anti Marie Rering Will. Mark Ackerman. Sheila

I.~yberg. and the I'irsl at~th~r '~erved as .~upervlsors in Ihe initial ),ear. In subseqt~ent years, tiw firsl atHhor SUl'~ervisedall ca.,a,'s. 411 should be n~llctl here !hal Iherapisls were encouraged It~ re,;pond !o and treat any i~ersonal t~r martial problems ,uf the parenl.s which occasion;allyemerged during the cour~ of Iheraf~y. While these effol'ls were less syslemalically programmed than lhe chdd intervenlions. Iherapy linle was devoled Io such problems in st}me easel;. Ill every case. however. Ihe grealer pr~porlit~n tH" Iherapy time was devoled Io interventions concerning the child. Thus, Ihe parent It:lining program wax viewed a:~ a ~el of guidelines for family inlerventi~m which could be I:tih~redand modified to suit the requirements of each family.

Multiple ('riteria Follow-Up of Behavior Modification

139

associalcd wilh order t)f treated problems (Fyherg & Johnson. 1974). Contingency ctmlracling i)rocedures were employed for Ihe last five cases in view of the prior success associated with this procedure. In addition to the l'~)rmal measures of outcome to be outlined here, the therapists m each ca,~ were required to rate the cooperation of the treated fatnily and keep records on the attendance and completion of assigned data by the families. Unless otherwise indicated, all of tile formal outcome measures were obtained prior to intervention, at termination and at 3 and 8 months following termination. While tile lherapists generally arranged for the assessments during each period, they were completely uninvolved in the adnunistralion of tests at termination and follow-up. The tests were administered at termination by a research assistant not otherwise connecled with the treatment, and at follow-up tests were mailed to families over the name of the research director. Furthermore, family members were explicitly told that their evaluations of the therapy process and of the therapists thentselves would not be seen by the therapists. Verbal Rel~Jrt Measures. A form t)f Ihe Becket (1960) Bi-Polar Adjective Checklist was employed to obtain parental description of the treated cluld, s This form has previously been employed by Patterson, Cobb, and Ray (1973) to assess parents' change in perception of their children following treatment. The Therapy Attitude Inventory was constructed specifically for lhe present research to assess parents' satisfaction with the process and outcome of the treatment program and with the therapists themselves. This inventory was administered at termination and at both follow-up periods. Parent Observation Data. The therapists required each parent to record the to-be-treated child's problem behaviors for a I-week baseline period prior to beginning an intervention for them. Parents recorded the frequency and/or dural)on of the behavior of interest for a specified time each day. Recording time 'varied from 89 hour per day for very high-rate or situation-specil~c behaviors to the entire day for lower rate behaviors. During intervention, behavioral recording continued as long us the behavior remained a focus of intervention. Whenever possible, I week of parent observation data was obtained at each follow-up period. I',:xamples t)f typical behaviors subject to parent data collection were compliance h) standing connnands such as gelling up in the morning and off to scho