|
Collection of Background Information
on Target Population (Participants and Nonparticipants)
We developed a structured questionnaire to collect nonidentifying
background information on the nonparticipants. Our main purpose
was to determine whether the nonparticipants differed in any significant
way from the participants. Information sought included data on the
identified individual’s age, sex, adaptive and academic functioning,
social and communication skills, and on the education and occupation
of the primary caregiver. This was obtained through direct contact
with a teacher or service worker who knew the individual well and
who was able to provide relevant descriptions without revealing
any information that might identify either individuals or their
families, thus maintaining complete confidentiality. Information
in the same domains of enquiry was collected directly from participants
or their caregivers.
Social strata were computed from educational and occupational data
on the parents or primary caregiver in the household in which the
individual lived or had lived most recently. Following the procedure
described by Hollingshead (19), an occupation score (on a scale
of 1 to 9) and a level of education score (on a scale of 1 to 7)
were combined to provide a social strata score, the range of which
is described on a scale of 1 to 5 (1 = major business or professional;
2 = medium business or minor professional or technical worker; 3
= skilled craftsman, clerical, or sales worker; 4 = machine operator
or semiskilled worker; 5 = unskilled labourer or menial service
worker).
Testing Participants and Screening
Nonparticipants to Confirm MR
Measures of nonverbal intelligence included the Performance Scale
of the Wechsler Adult Intelligence Scale-Revised (WAIS-R) or Wechsler
Intelligence Scale for Children-Revised (WISC-R) and the Merrill-Palmer
Scale of Mental Tests (excluding the verbal items) for less-capable
individuals. The Peabody Picture Vocabulary Test-R (PPVT-R: Form
L) provided an estimate of single word receptive vocabulary. Standard
scores from the WAIS-R or WISC-R and equivalents from the Merrill
Palmer (converted from mental-age scores) were averaged with standard
scores from the Peabody (converted from mental-age equivalents,
where necessary) to provide a composite (verbal and nonverbal) IQ
score.
Background information obtained on nonparticipants
was reviewed independently by 2 of the investigators. Each reviewer
rated the information for adaptive and academic functioning and
social and communication skills, according to criteria for mild-to-profound
MR, based on ICD-10 definitions (20) and following operationally
defined guidelines (1,21). Borderline MR (adaptive and cognitive
functioning in the 76-to-80 IQ range) was defined by 10 criteria
of daily living and functional skills. No participant with MR (IQ
£ 75) was able to drive a motor vehicle; therefore, nonparticipants
reportedly holding a driver’s license were excluded from the
group with MR. Cognitive and adaptive functioning were rated as
shown in Table 1. Agreement between
reviewers was greater than 80%; disagreements were discussed and
a consensus reached.
Results
Target Population
Identification
Research staff were provided with the initials, sex, and birthdate
of 635 individuals considered to have developmental problems (see
Figure 1). Of these, 149 were duplicates
(that is, the same initials were provided from different sources),
and 116 were removed because they did not meet age or residency
criteria. Of the 370 remaining, 204 individuals or their families
indicated an interest in participating (participants); 166 individuals
or their families declined (nonparticipants).
After screening for IQ, 33 of the 204 participants had a composite
IQ greater than 75, leaving 171 with a composite IQ less than or
equal to 75 (these were classified as participants with MR). Background
information was sufficiently comprehensive to determine functioning
level for 136 (82%) of the 166 nonparticipants. Those remaining
(30/166) were randomly assigned to the MR and non-MR groups in the
same ratio as the 136 individuals for whom there was comprehensive
background information. Review of the 166 nonparticipants consequently
identified 84 (51%) individuals with MR (defined as nonparticipants
with MR) and 82 (49%) individuals without MR. Overall, the study
(target) population comprised all 255 individuals aged 14 to 20
years who met criteria for MR (n = 171 participants with MR; n =
84 nonparticipants with MR. See Figure
1).
Participation Rate
Of the 255 individuals with MR, 171 participated, yielding a participation
rate of 67%. Background information was available for all participants
with MR and for 76% of nonparticipants with MR (64/84), which represents
92% of the total group of 255.
Comparison of
Participants with MR and Nonparticipants with MR
Age, Sex, and Level of Functioning. There were no significant differences
between participant and nonparticipant MR groups (Table
2).
Social Strata. There were more missing social
strata data for the nonparticipant MR group (44%, compared with
10% for the participant MR group). Participant and nonparticipant
MR groups differed significantly in social strata. In social strata
1, 2, and 3, there was a greater representation of participants
with MR (55%) than of nonparticipants with MR (24%) (c2 = 4.92,
df 1, P < 0.05).
Prevalence of MR. Based on the most current census data (17), there
were 35 485 young persons between age 14 and 20 years, giving an
overall MR prevalence rate of 7.18/1000 (95%CI, 6.31 to 8.06/1000),
with MMR (IQ = 50 to 75) prevalence estimated at 3.54/1000 and SMR
(IQ < 50) prevalence estimated at 3.64/1000 (Figure
2).
Age-Specific Rates
Figure 2 also shows age-specific
prevalence rates. More individuals were represented at the younger
ages (that is, 14 to 16 years vs 17 to 20 years), but the difference
is not significant.
Age and Functioning Level
There were no significant differences between the number of individuals
with mild and severe MR at each age. The number of identified individuals
with MMR gradually declines with age (from 19% at age 15 years to
11% at age 20 years); no such trend is evident for those with SMR
(15% at age 15 years and 14% at age 20 years). This pattern is likely
related to the end of mandatory schooling at age 16 to 17 years,
when individuals who are more capable outside an academic setting
are less likely to be identified as having MR.
Sex and Functioning Level
Male subjects outnumbered female subjects at all age levels, except
at age 19 years. The overall male-to-female ratio was 1.3:1.0. However,
functional level varied significantly with sex (c2 = 4.66, df 1,
P < 0.05): there were more female subjects with SMR (56% female,
compared with 44% male) and a greater number of male subjects with
MMR (58% male, compared with 42% female).
Social Strata and Functioning Level
Although none of the differences reached significance, individuals
with MMR tended to congregate in the less-advantaged social strata
(that is, 4 and 5, but also 2), and those with SMR in strata 1 and
3. There also was a trend toward more male subjects than female
subjects in social strata 1 to 4 and more female subjects in social
stratum 5.
Discussion
Our prevalence estimate of 3.64/1000 for SMR is similar to rates
reported in earlier studies conducted in Canada and elsewhere. Estimates
for SMR derived from Canada more than 20 years ago include 3.8/1000
(14), 3.7/1000 (15), and a lower 2.2/1000 (16). In their metaanalysis
of prevalence studies conducted between 1960 and 1986, Roeleveld
and others concluded that the rate for SMR in school-age children
was relatively stable—around an average of 3.8/1000 (6). In
a more recent US study carried out in Atlanta, Georgia, the prevalence
of SMR was 3.6/1000 for children aged 10 years in the period 1985
to 1987 (22). Abramowicz reviewed studies conducted before 1960
and reported a median prevalence rate for children with SMR of 3.7/1000
(7). A follow-up study by Richardson and Koller of children in Aberdeen
at age 22 years, born during 1952 to 1954, and initially studied
at age 8 to 10 years, found age-specific prevalence rates of 3.3/1000
at age 8 and 9 years and 2.8/1000 at age 22 years, the drop in rate
being almost exclusively due to mortality (12). Our SMR rates for
age 14 to 20 years ranged from 2.80/1000 to 4.35/1000, according
well with the findings of Richardson and Koller (12). They also
underscore that the prevalence of SMR is relatively stable, not
only across age but also across time.
For individuals with SMR, the disabilities usually necessitate
special supports and services. In epidemiological studies of SMR,
the condition has often been defined and identified according to
the level of service need. Therefore, for SMR ascertainment prevalence
is considered to be a reasonable estimate of true prevalence. Our
finding of an SMR rate similar to that reported in other studies
confirmed to us that our procedure for case identification of MR
was sufficiently comprehensive and inclusive. This was corroborated
by the fact that 23% of our initial referrals were duplicates from
various sources.
Our finding of 3.54/1000 for MMR is in the lower end of the range
established in previous studies (6,11). Low rates raise concerns
about whether ascertainment is incomplete (and underestimates true
prevalence) or whether such rates indeed reflect the true prevalence,
which is lower as a result of such factors as an upward drift in
IQ related to improved environments or policies of integration.
Because we consider our cacase-identification procedure to have
been sufficiently comprehensive (see above), we look to other such
explanations for the low MMR prevalence rate found in our study.
Our MMR prevalence estimate is similar to the lower rates generally
found in Scandinavian studies (6). In Sweden, these low rates are
thought to partly reflect the long-standing tradition of not institutionalizing
individuals with MR unless absolutely necessary. Of those persons
identified with MR in the Swedish studies, most are judged as having
SMR and only 25% as having MMR (23). This is in stark contrast to
figures reported from the US, where the reverse situation prevails:
75% have been identified as mildly retarded and 25% as severely
retarded (23). Murphy and others report a more recent US prevalence
estimate of 8.4/1000 for MMR, which represents 70% of the overall
MR prevalence of 12/1000 (22).
Differing prevalence rates between the US and Sweden have been
attributed to attitudes, practices, social policies, and allocation
of resources. In the Swedish welfare system, service need rather
than IQ or diagnosis determines eligibility for benefits. Sonnander
and others studied a group of over 8000 pupils aged 12 to 13 years
and identified those who met the psychometric criterion for MR but
who were not administratively classified as such (n = 116) (24).
Combining the numbers in this latter group (116/8000 or 14.5 per
1000) with those identified by school personnel as having MR (7.4/1000)
yields a prevalence of 21.9/1000 or 2.19%—close to the expected
or true prevalence estimate (23).
Sweden’s philosophy of including and integrating persons with
MR spread to Ontario and the rest of Canada in the 1960s (25). The
individuals in our study cohort have experienced an education policy
of support and integration throughout their school careers. In Ontario,
children with MR are educated in regular classes wherever possible.
For those with SMR, other options may be available (for example,
self-contained schools or special classes in regular schools). However,
parental preference takes priority, and even the most disabled child
will be accommodated in the regular classroom at the parents’
request. Children with milder disabilities (MMR) are usually found
in mainstream classes, although they may follow a modified program
and be evaluated on this at graduation. If children are having difficulties
in class (for example, in learning or behaviour), they are referred
for remedial support and sometimes for psychological testing. Many
may remain in the remedial category and may never be identified
as having MR. The teacher’s perception of the problem, the
psychological resources available to the school, and a moratorium
on labelling a child as having MMR are key in the extent to which
children will be identified. Our study in Niagara relied on school
personnel flagging those children thought to have MR. Of the children
flagged, 31% were found not to have MR, suggesting that other factors
were affecting teacher selection (we are analyzing these false positives
to identify characteristics that flagged these children as having
MR). We do not have data on the number of children who psychometrically
may have met criteria for MMR but were not flagged by teachers,
presumably because, as has been found in Sweden, their adaptive
behaviour did not distinguish them from their peers.
Our study results are similar to the Swedish studies in the following
other findings: in our ratio of 1.3 male subjects to 1.0 female
subjects, in a greater contribution of SMR to the overall prevalence
rate, and in our finding of more male subjects with MMR and more
female subjects with SMR (26). We conclude that, while prevalence
rates for SMR are relatively stable across geographic locations,
rates for MMR vary—not least because of prevailing philosophies
of care and integration, as well as resources available to implement
these philosophies. What evidence is available suggests that the
low prevalence of MMR in our study is linked to the policies of
integration in Ontario over the past 3 decades—policies that
may have made persons with MMR less visible.
1
| 2 | 3 | 4
|
