Journal of Autism and Developmental Disorders, Vol. 14, No. 3, 1984
Suppression of Pica by Overcorrection and Physical
Restraint: A Comparative Analysis 1
Nirbhay N. Singh 2 and Leon W. Bakker
University of Canterbury, Christchurch
Overcorrection and physical restraint procedures have been shown to be
effective in controlling certain classes o f maladaptive behavior in mentally
retarded persons. In the present study, an alternating treatments design was
used to measure the differential effects o f overcorrection and physical
restraint procedures in the treatment o f pica. Changes in collateral
behaviors were also monitored. Each occurrence o f pica was f o l l o w e d by
either an overcorrection procedure or a physical restraint procedure.
Although both procedures reduced the occurrence o f pica and had a similar
effect on the occurrence o f collateral behaviors, physical restraint was
clinically more effective in terms o f immediate response reduction.
Pica, the i n g e s t i o n o f n o n n u t r i t i v e o r i n e d i b l e o b j e c t s , is c o m m o n l y o b served in m e n t a l l y r e t a r d e d p e r s o n s . A l t h o u g h p i c a is also o b s e r v e d in n o r m a l i n f a n t s up to the age o f 12 to 18 m o n t h s ( B a l t r o p , 1966), it tends to persist with m e n t a l l y r e t a r d e d a n d d e v e l o p m e n t a l l y d e l a y e d c h i l d r e n a n d
a d u l t s . F o r e x a m p l e , Singh a n d W i n t o n (1982) f o u n d 8 % o f an institut i o n a l i z e d p o p u l a t i o n o f 598 m e n t a l l y r e t a r d e d p e r s o n s e n g a g e d in pica.
F u r t h e r m o r e , a b o u t 26~ o f an i n s t i t u t i o n a l i z e d p o p u l a t i o n o f 991 m e n t a l l y
r e t a r d e d a d u l t s have been r e p o r t e d to i n d u l g e in pica ( D a n f o r d & H u b e r ,
1982).
~The experimental protocol for this study was reviewed and approved by the Ethics Committee
of the North Canterbury Hospital Board. This research was supported in part by Grants No.
10 and No. 82-59 from the New Zealand Labor Department. The authors are grateful to Dr.
Jim Marshall, Peter Bootsma, and the ward staff of Templeton Hospital and Training School
for their generous assistance. A special thanks to Lyonne Dalley, Jane Guillen, and Robin
Phillips, who assisted with the manuscript preparation.
2Address all correspondence to Nirbhay Singh, Department of Psychology, University of
Canterbury, Christchurch, New Zealand.
331
0162-3257/84/0900-0331 $03.50/0 ~ 1984 Plenum Publishing Corporation
332
Singh and Bakker
The suppression of pica is of some importance since it may lead to
lead poisoning (Lourie & Millican, 1969; Snowdon, 1977) and medical complications such as intestinal obstruction, constipation, and nutritional
anemia (Kanner, 1962). A number of behavioral techniques have been used
to suppress pica in normal and mentally retarded persons. For example, differential reinforcement, verbal reprimand, time-out, discrimination training, and screening procedures have been used (Ausman, Ball, & Alexander,
1974; Madden, Russo, & Cataldo, 1980; Singh & Winton, 1984).
Two other procedures, physical restraint and overcorrection, have
been used to suppress pica. The physical restraint procedure requires the
subject's arms to be held for a brief period contingent on a maladaptive
response. In one study, Bucher, Reykdal, and Albin (1976) used verbal
reprimand ("No") and a 30-sec physical restraint to control pica by two
mentally retarded children. However, only partial suppression was
achieved, and the differential effects of verbal reprimand and physical
restraint were not assessed. These findings were extended by Winton and
Singh (1983), who showed that physical restraint alone was effective in controlling pica and that a 10-sec duration was more effective than 30 sec with
one subject and 3 sec with another.
Overcorrection procedures are designed not only to suppress maladaptive behavior but also to teach individuals appropriate alternative behavior.
Foxx and Martin (1975) used an overcorrection procedure to treat pica that
required the subject to spit out or throw away the inedible object, engage in
oral hygiene training followed by personal hygiene, and tidy the floor and
empty the trash can. Variations of this procedure have been used by
Matson, Stephens, and Smith (1978) and Mulick, Barbour, Schroeder, and
Rojahn (1980) in the treatment of pica.
The relative effects of these procedures on stereotypic behaviors has
recently been investigated, with one study showing overcorrection and
physical restraint to be equally effective (Shapiro, Barrett, & Ollendick,
1980), and the other showing them to be differentially effective across subjects (Ollendick, Shapiro, & Barrett, 1981). However, the overcorrection
procedure used in the treatment of stereotyped behavior (i.e., verbal warning and manual guidance in appropriate tasks) is topographically different
from the Foxx and Martin (1975) overcorrection procedure for treating
pica. Thus, the findings from the two comparative studies (Ollendick et al.,
1981; Shapiro et al., 1980) cannot be generalized to the treatment of pica.
The present study compared the relative effects of overcorrection and
physical restraint on pica and collateral behaviors of two profoundly
mentally retarded girls.
Suppre~ion of Pica
333
METHOD
Subjects
The subjects were two girls from an institution for the mentally retarded. Both were profoundly retarded on the basis of the A A M D criteria
(Grossman, 1977) and had receptive but only minimal expressive language.
The two girls exhibited high rates of maladaptive stereotypic and selfstimulatory behavior, including pica. They exhibited minimal prosocial
behavior. Subject 1 was a 20-year-old Polynesian who had been institutionalized for 7 years. She was the second twin of a monozygous pair. Subject 2 was 21 years old and had been institutionalized for 12 years. The
etiology of both subjects was not known. Their IQs were below 20, and
their social age on the Vineland Social Maturity Scale was less than 12
months. Both had a long history of ingesting nonnutritive substances, including stones, cigarette butts, remains of food off the floor, bits of string,
grass, and other materials. Neither subject had been tested for pica-related
lead poisoning. Regular ward treatment for pica through differential reinforcement and punishment procedures had little effect on the behavior of
both girls. This was probably due to the inconsistent application of these
treatments by rostered staff.
Settings
Observation and treatment sessions were scheduled in two settings: a
sunroom inside the subjects' regular residential ward, and outside the ward
(e.g., playground and lawn area).
Behavior Observed
Pica and three collateral behaviors were observed throughout the
study.
Pica was defined as an inedible or nonnutritive substance either
touching the subject's lips or being placed in the mouth.
Picking and handling was defined as touching, picking up, a n d / o r
holding an inedible or nonnutritive substance. Pica, not picking and handling, was recorded if the item was subsequently brought into contact with
the lips or placed in the mouth. Picking up and handling as deviant behavior
334
Singh and Bakker
was distinguished from other picking up and handling of objects by the type
of substance picked up. This category was coded only when those substances used for pica were picked up.
Stereotypy was defined as repetitive complex finger movements, body
movements, or rocking.
Social behavior was defined as smiling, appropriate speech or
laughter, appropriate toy play, and interaction with other residents and
staff.
Recording and Interobserver Agreement
Five observers with extensive experience from earlier studies on pica
(e.g., Singh & Winton, 1984; Winton & Singh, 1983) were given additional
training before participating in this study. These observers had no experience or training in behavior modification, were naive to the experimental procedures, and were not informed of the experimental hypothesis. Data
were collected by two observers, one per subject, randomly assigned on a
daily basis. A third observer was also randomly assigned during about 25~
of the sessions (for each subject) for reliability checks. Two sessions were
conducted daily, one in each setting per subject. Behaviors were observed in
90 10-sec periods. The total free response time across baseline and experimental sessions was held constant at 15 minutes; i.e., time spent in treatment was not recorded.
Interobserver agreement was computed by dividing the number of
agreements on the occurrence of each target behavior, on an interval-byinterval basis, by the sum of the agreements and disagreements, and
multiplying by 100. An agreement was defined as both observers recording
an occurrence of the same target response during the same interval. The
mean interobserver agreements (with ranges in parentheses) for Subject 1
were p i c a - 9 7 %
(92-100), picking and h a n d l i n g - - 8 5 % (82-89),
s t e r e o t y p y - 9 4 % (86-98), and social b e h a v i o r - 8 1 % (75-91). For Subject 2
they were p i c a - 9 1 % (82-96), picking and h a n d l i n g - 8 4 % (79-91),
s t e r e o t y p y - 9 5 % (90-100), and social b e h a v i o r - 8 3 % (78-89).
Experimental Design and Procedures
An alternating treatments design (Barlow & Hayes, 1979) was used to
assess the effects of overcorrection and physical restraint on pica and collateral behaviors. Following baseline observations, overcorrection and
physical restraint procedures were used in two settings (sunroom, outside),
Suppression of Pica
335
with each procedure being randomly allocated to a setting each day for each
subject. Subsequently, only the more effective treatment for each subject
was implemented in both settings.
The three experimental phases were as follows:
Baseline. Data were collected on the naturally occurring rate of pica
and collateral behaviors for 5 consecutive days, twice a day, per subject. No
programmed contingencies for any of the behaviors were in effect during
this phase except that, for medical and ethical reasons, the ward staff
removed any inedible substances from the subject's mouth when they were
observed.
Alternating Treatments. The alternating treatments phase was in effect for 10 days for both subjects. Each of the two treatments was randomly
assigned to one of the two settings on a daily basis. Each occurrence of pica
was followed by either a 10-sec physical restraint (see Winton & Singh,
1983) or the Foxx and Martin (1975) overcorrection procedure. During
physical restraint, the subject was required to spit out or throw away the inedible object, and her arms were restrained at the side of her body for 10
sec. No verbal reprimand was used. Release from physical restraint was
contingent on a period of nondisruptive behavior for 10 sec. During overcorrection, the subject was required to spit out or throw away the inedible
object, undertake oral hygiene, tidy the area in the vicinity of the subject,
pick up trash and empty the trash can, and engage in personal hygiene training. The entire sequence took about 15 minutes (see Mulick et al., 1980),
and the maximum number of overcorrection treatments was limited to eight
per session (i.e., 2 hours excluding observation time of 15 minutes).
Physical Restraint. The more effective of the two procedures, physical
restraint, was used in both settings in this phase with both subjects. The
physical restraint contingency was the same as in the previous phase.
RESULTS
Figures 1 and 2 show the daily percent of intervals of pica and collateral behaviors across all phases and settings for Subject 1 and Subject 2.
The means for these behaviors during each phase are given in Table I. The
data from the alternating treatments phase are presented according to the
type o f treatment. As the physical restraint treatment proved more effective
with both subjects, this treatment was employed in the final phase in both
settings with the two subjects.
Subject 1. During baseline, pica occurred at a slightly higher rate in
the sunroom than outside. This reduced to similar low levels in both settings
336
Singh and Bakker
Table 1. Mean Percent of Intervals and Observed Behaviors Across Experimental Conditions
Alternating treatments"
Setting and behavior
Subject 1
Sunroom
Pica
Picking and
Stereotypy
Social
Outside
Pica
Picking and
Stereotypy
Social
Subject 2
Sunroom
Pica
Picking and
Stereotypy
Social
Outside
Pica
Picking and
Stereotypy
Social
handling
handling
handling
handling
Baseline
Physical
restraint
Overcorrection
Physical
restraint
32.2
33.4
42.8
7.4
6.4
21.9
37.3
9.6
9.9
15.8
36. I
8.4
1.3
7.9
37.5
1.3
20.4
27.4
40.4
4.2
10.8
12.2
83.6
3.2
83.8
14.0
35.4
.6
1.4
5.8
38.2
13.5
10.8
7.5
72.6
5.0
25.8
10.0
55.1
.4
.9
5.0
74.5
7.8
4.3
4.0
49.1
1.0
"For ease of presentation, the data on alternating treatments are presented with the s u n r o o m
data. In practice, the two treatments were r a n d o m l y presented either in the s u n r o o m or outside on a daily basis.
during the alternating treatments phase under both the overcorrection and
physical restraint contingencies. However, suppression was more rapid and
complete under physical restraint. In the final phase, pica was maintained at
very low levels by the physical restraint contingency.
Picking and handling decreased during the alternating treatments
phase under both treatments, with a greater reduction being evident under
the overcorrection contingency for pica. The final phase resulted in further
reductions in both settings under the physical restraint contingency for pica.
Stereotypic behavior remained relatively stable across phases. Changes in
social behavior were not consistent; in the sunroom, social behavior
occurred less in the final phase than in the baseline, while social behavior increased in the other setting.
Subject 2. During baseline, pica occurred more frequently in the outside setting than in the sunroom. This was due to this subject's particular
craving for substances usually found in the outside setting, such as grass
and stones. As for Subject 1, both treatments reduced pica, with physical
restraint being more effective.
Suppression of Pica
337
BASELINE ALTERNATING
TREATMENTS
4o
30
g
E
20
i S~m•
,
9 Restra.lt
Outside/ u O v e r t
i t . ~l,
PHYSICAL RESTRAINT
9 SLfYL~I~
(~Hslde
otrt2clto~
w.
~x_..., (
10
0
i
i
~
z
(..9
E
i
i
\
4O
r'c
r
i
10
m
I--
5
I
I
I
i
i
i
20
25
30
(J
rr
3O
m
20
I0
0
i
5
~
15
DAYS
Fig. 1. Rate of pica and collateral behaviors for Subject 1 across
experimental conditions.
PERCENT OCCURRENCE
P
SOCIAL
STEREOTYPY
PICKING
PICA
0
.
.
.
.
.
.
.
.
.
.
o
W
9
g
rn
r-
am 83
&
rm
<
E
Ez
m~
- ~~Z
m
o
Om
-~
I',2
2
9
"U
rm
U~
-.t
2IJ
Z
-.-I
:r
Suppression of Pica
339
Picking and handling was reduced under both treatment contingencies
for pica and reached even lower levels during the final phase under physical
restraint. No consistent changes in stereotypic behavior were noted,
although a slight decrease was observed in the sunroom and an increase in
the outside setting. Social behavior increased only slightly across both treatment phases in both settings.
DISCUSSION
The results showed that while both treatments suppressed pica,
physical restraint was clinically more effective than overcorrection with
both subjects. In terms of the mean response rate, the difference between
the two procedures was small for Subject 1. Although the difference was
much larger for Subject 2, there was a greater variability in the daily data of
this subject during the alternating treatments phase. Nonetheless, there was
a clear quantitative difference between the two procedures for both
subjects.
In contrast to the present study, previous comparative studies found
physical restraint and overcorrection to be either equally effective (Shapiro
et al., 1980) or differentially effective across subjects (Ollendick et al.,
1981). However, the overcorrection procedure for pica used in the present
study is not directly comparable with the overcorrection procedure for
stereotypy used by Ollendick et al. (1981) and Shapiro et al (1980). One of
the hazards of using a single term, such as overcorrection, for a combination of procedures with different components is that it may lead to false
comparisons of its efficacy across studies.
In terms of collateral behaviors, both treatments appeared to affect
them equally. Picking and handling decreased under both treatments for
pica, and stereotypy remained at about the same level. Only minor gains
were made in social behavior but these were distributed equally across both
treatments. Thus, while physical restraint was clinically more effective than
overcorrection in the reduction of the target behavior, both treatments
affected the collateral behaviors in the same way and to the same extent.
In practical terms, the physical restraint procedure may have certain
advantages over the overcorrection procedure employed in this study. It
takes a lot less staff-training time and the procedure is much briefer. Thus,
it is more likely to be used by primary care staff in large, understaffed institutions. Furthermore, since the procedure does not require the use of
special equipment, it can be implemented more easily and more
systematically in applied settings (e.g., classrooms, workshops) or even in
the community (e.g., supermarkets, buses). In terms of social validity, all
340
Singh and Bakker
therapists reported a preference for physical restraint when compared to the
overcorrection procedure.
Both subjects reacted to the two treatments in similar ways. They
struggled and resisted treatment for the first two or three sessions when
physical restraint was the intervention in effect. The subjects showed more
passive resistance to the overcorrection procedure, particularly with the first
component (i.e., picking up trash and emptying the trash can). While most
resistance was evident during the initial treatment sessions, some passive
resistance was observed during most overcorrection sessions.
In sum, the study showed that physical restraint was clinically more
effective than the Foxx and Martin (1975) overcorrection procedure for
treating pica by two mentally retarded persons. This study further confirms
the efficacy of brief response-contingent physical restraint for controlling
the maladaptive behaviors of mentally retarded persons.
REFERENCES
Ausman, J., Ball, T. S., & Alexander, D. (1974). Behavior therapy of pica in a profoundly
retarded adolescent. Mental Retardation, 12, 16-18.
Bahrop, D. (1966). The prevalence of pica. American Journal o f Diseases o f Children, 112,
116-123.
Barlow, D. H., & Hayes, S. C. (1979). Alternating treatments design: One strategy for comparing the effects of two treatments in a single subject. Journal o f Applied Behavior
AnaO'sis, 12, 199-210.
Bucher, B., Reykdal, B., & Albin, J. (1976). Brief restraint to control pica in retarded children.
Journal o f Behavior Therapy and Experimental Psychiato,, 7, 137-140.
Danford, E. E., & Huber, A. M. (1982). Pica among mentally retarded adults. American Journal o f Mental Deficiency, 87, 141-146.
Foxx, R. M., & Martin, E. D. (1975). Treatment of scavenging behavior (coprophagy and pica)
by overcorrection. Behaviour Research and Therapy, 13, 153-162.
Grossman, H. D. (1977). Manual on terminology and classification in mental retardation.
Washington, D.C.: American Association on Mental Deficiency.
Kanner, C. (1962). Child psychiatry. Springfield, Illinois: Charles C Thomas.
Lourie, R. S., & Millican, F. K. (1969). Pica. In J. G. Howells (Ed.), Modern perspectives
in international psychiatry (pp. 455-470). Edinburgh: Oliver & Boyd.
Madden, N. A., Russo, D. C., & Cataldo, M. F. (1980). Behavioral treatment of pica in
children with lead poisoning. Child Behavior Therapy, 2(4), 67-81.
Matson, J. L., Stephens, R. L., & Smith, C. (1978). Treatment of self-injurious behavior with
overcorrection. Journal o f Mental Deficiency Research, 22, 175-178.
Mulick, J. A., Barbour, R., Schroeder, S. R., & Rojahn, J. (1980). Overcorrection of pica
in two profoundly retarded adults: Analysis of setting effects, stimulus and response
generalization. Applied Research in Mental Retardation, 1, 241-252.
Ollendick, T. H., Shapiro, E. S., & Barrett, R. P. (1981). Reducing stereotypic behaviors: An
analysis of treatment procedures utilizing an alternating treatments design. Behavior
Therapy, 12, 570-577.
Shapiro, E. S., Barrett, R. P., & Ollendick, T. H. (1980). A comparison of physical restraint
and positive practice overcorrection in treating stereotypic behavior. Behavior Therapy,
11, 227-233.
Suppression of Pica
341
Singh, N. N., & Winton, A. S. W. (1982). Pica in institutionalized mentally retarded persons.
Unpublished data, University of Canterbury.
Singh, N. N., & Winton, A. S. W. (1984). Effects of a screening procedure on pica and
collateral behaviors. Journal o f Behavior Therapy and E_~:perimental Psychiatr.v, 15,
59-65.
Snowdon, C. T. (1977). A nutritional basis for lead pica. Ph.vsiology and Behavior,
18, 885-893.
Winton, A. S. W., & Singh, N. N. (1983). Suppression of pica using brief-duration physical
restraint. Journal o f Mental Deficiency Research, 27, 93-103.