The Changing
Pattern
of Primary
Pediatric
Care:
Update
for One Community
Klaus
J. Roghmann,
PhD,
Robert
A. Hoekelman,
MD, and
Thomas
K. Mclnerny,
MD
From the Department of Pediatrics, University of Rochester School of Medicine, Rochester, New York
ABSTRACT. The changing pattern of pediatric practice
in the Rochester, New York, community between the late
1960s and the late 1970s is described, and some
extrapo-lations for the next decade are discussed. The child
population shrunk from 243,000 in 1970 to 192,000 in
1980 and is expected to decrease further to 160,000 in
1990, possibly stabilizing at 140,000 by 2000. The number of pediatric beds as well as occupancy rates declined, but the number of full-time equivalent practicing pediatni-cians increased slightly. One third of them are now prac-ticing out of neighborhood health centers or health main-tenance organizations. Problems of manpower shortage and inadequate access to care for the inner city residents have long since disappeared. Utilization rates by race and socioeconomic area are similar once the children enter
the care system. For “well child” care, however, there
may still be lower utilization for blacks, especially for
older children. The high “market penetration” for child
health services by pediatricians, and the high proportion ofwell child visits (40%) among all visits, may be atypical
for the nation as a whole, but is probably indicative of
what pediatric care elsewhere will be in the future. Fewer children, and less acute care per child, will allow peclia-tricians to focus increasingly on preventive, developmen-tal, and psychosocial needs. Pediatrics 1984;73:363-374;
community pediatrics, primary care, utilization rates.
There is a history of medical care research in the
Rochester, New York community that dates back
to the days of Earl L. Koos. His classic study, The
Health of Regionville’ was done while he chaired the Department of Sociology at the University of Rochester. Regionville, an hour’s drive south of Rochester, is now linked by an interstate highway to the “metropolis” in the north. It was restudied a few years ago to update Koos’ classic work.2
Received for publication March 16, 1983; accepted July 26, 1983. Reprint requests to (K.J.R.) Department of Pediatrics, Univer-sity of Rochester School of Medicine, 601 Elmwood Aye, Roch-ester, NY 14642.
PEDIATRICS (ISSN 0031 4005). Copyright © 1984 by the
American Academy of Pediatrics.
The system of pediatric are of the community
has been explored by several studies conducted over
a 10-year period (1966-1976) under the direction of Haggerty. A summary of the earliest findings was published in i975. Most of the problems of access to and availability of pediatric care for children were solved in the early 1970s through a number of patient care programs largely directed and evalu-ated by the Department of Pediatrics. As a result, our research orientation has changed from prob-lems in community pediatrics to those of general pediatrics with special emphasis on collaborative
research with practicing pediatricians.4 In 1977 an
ambulatory care survey involving nine pediatric
practices5’6 was undertaken with a focus on the “new morbidity.” In 1979 this work was extended to 30 practices. The continuity of our research orientation in the tradition of community child health studies, in spite of the shift from equity of care in a given child population to the nature of primary pediatric care, will become evident in the description of this latest project. We will examine the recent changes in the population served, de-scribe the practice patterns ofpediatnicians in 1979, estimate the volume of pediatric care, compute
utilization rates by sociodemographic
characteris-tics to link our findings with the earlier studies, and, finally, describe visits by diagnostic category.
POPULATION
AND
PRACTICE
PATTERN
Changing
Child Population,
1970-1980
The early period of the Rochester Child Health Studies (1966-1972) was characterized by a rapidly growing child population and a pediatric manpower
shortage, which gave rise to problems in access and
availability to pediatric care. The “baby boom” and
the migration of minority families from the South
dra-Age (yr) on Census Census Net Expected Change Due
April 1 by 1970 1980 Change 1980 Census to Migration
Group .
Year Live
Births*
Infants
<1 13,076t 10,027 -3,049 1979 9,677 +350
1 12,709t 8,923 -3,786 1978 9,142 -219
Preschool
2 12,614t 9,220t -3,394 1977 9,447 -227
3 12,661t 8,683t -3,978 1976 8,966 -283
4 13,554t 8,726t -4,828 1975 9,010 -284
5 14,055 8,758 -5,297 1974 9,087 -329
Elementary
School
6 14,225 8,701 -5,524 1973 9,100 -399
7 14,089t 8,808t -5,281 1972 9,700 -892
8 13,892t 10,235t -3,657 1971 11,271 -1,036
9 14,432t 12,054t -2,378 1970 13,275 -1,221
Middle
School
10 14,148 11,976t -2,172 1969 13,066 -1,090
11 14,148t 11,640t -2,508 1968 12,699 -1,059
12 14,148t 11,554t -2,594 1967 12,605 -1,051
13 14,148t 11,330t -2,818 1966 12,361 -1,031
High School
14 13,733 12,129 -1,604 1965 12,863 -734
15 13,270 13,059 -211 1964 13,306 -247
16 12,823 13,094 +271 1963 12,857 +237
17 12,169 12,911 +742 1962 12,820 +91
Totals 243,894 191,828 -52,066 201,252 -9,424
Av age 8.52 yr 9.20 yr
* Live births 1980, 10,037; 1981, 10,188.
t Intrapolated from grouped data.
matically without a corresponding increase of pe-diatnicians. Both trends were reversed in the early
1970s and led to a rapid solution of the local
“med-ical care crisis”7 for children. The previously ob-served widespread inappropriate use of the area’s emergency rooms for acute care of underserved inner city populations largely disappeared.
Populations of large countries change only
slowly, hiding the rapid changes that specific age
cohorts as well as locally defined populations un-dergo. The child population of Monroe County, consisting of the City of Rochester and its
sun-rounding suburbs, is an example of such rapid
change; it declined by more than 52,000 or 21%
from 1970 to 1980 as a result of two
factors-migration and smaller birth cohorts. A
demonstra-tion of how these two factors affected the popula-tion served by Monroe County pediatricians is
pre-sented in Table 1. Each age cohort in 1970 had
between 12,000 and 14,000 children, the largest
groups were of elementary and middle school age.
The dramatic decrease of the birth rate slowly replaced these large school age cohorts with small ones of only 9,000, resulting in a top heavy age
distribution for the child population by 1980. The
TABLE 1. Changing Child Population in 1980
high school cohorts were still 13,000 strong, but preschool and early elementary school cohorts av-eraged less than 9,000. The comparison between specific cohorts in 1970 and 1980 (Table 1 “Net Change”) shows reductions of up to 5,000 or 40%
for single elementary school age cohorts. The effect
due to migration is best recognized by comparing the number of live births, as recorded during the preceding 18 years by the county’s Vital Statistic Office, with the number of children of these cohorts still in the community at the time ofthe 1980 census (Table 1 “Change Due to Migration”). Of the total decline between 1970 and 1980 in the child popu-lation (52,000), approximately 18% (9,400) was due
to migration and approximately 82% (42,600) to
the declining birth rate. Furthermore, the child
population in 1980 was “older” than in 1970; the
average age had increased from 8.53 years to 9.20
years.
The 1980s will show a continuation of this trend
in the community. The annual number of
new-borns, after decreasing to less than 9,000 in the
mid-1970s, has increased to about 10,000 and will
remain that high through the 1980s due to the
number of “baby boom” children now having
1970-dren of their own (echo boom). Even these cohorts, however, are not large enough to replace the 12,000-to 13,000-sized cohorts now in the high schools. By 1990 the total child population will have declined
by another 30,000, that is, from 190,000 now to
160,000 in 1990, with approximately 20,000 of these
lost because of the lower birth rate and 10,000
because of migration.
The 1990s will not yet bring a stabilization of the child population. By 1995 the number of live births will again be down to 9,000 per year, as the echo boom will have passed, and by the year 2000 annual live births will decrease to 8,000 as the small cohorts of the 1970s enter their childbearing period. The number of children in the community will stabilize at 140,000 by the turn ofthe century, approximately
100,000 (43%) less than in 1970.
Socioeconomic and health indicators for the
corn-rnunity showed an improvement over the decade of
the 1970s. The number of infant deaths, which
averaged 241/year in 1967 to 1969, declined to 109 in 1981. Infant mortality was 18.8/1,000 for 1967 to 1969, but was down to 10.7/1,000 in the county by 1981, reaching 14.5/1,000 in the city and 7.7/
1,000 in the suburbs. (Secondary analysis of vital
statistics files by authors. No separate statistics by
surburbs and city are published.) Racial differences
in infant mortality persisted even when controlling for area of residence: In the city, the infant
mortal-ity among whites was 12.3/1,000 compared with
18.0/1,000 among blacks; in the suburbs, infant
mortality among whites was 7.4/1,000 compared
with 12.0/1000 among blacks.
The birth rate among blacks has shown the same
downward trend as that among whites. In 1981
there were fewer black births in the community
than there were in 1970, although the proportion
of black births among all births increased from
14.5% in 1970 to 18.3% in 1981. The number of
children on Medicaid declined from about 30,000 in 1970 to approximately 20,000 in 1980. (Personal communication from Public Relations Office,
De-partment of Social Services, Monroe County. No
precise statistics by age are kept.)
Pediatric
Practices
in the Community
Pediatric practice in the Rochester community
also changed during the 1970s. In the 1960s, private
practice was the exclusive mode of primary
pediat-nc care delivery. It followed the pattern of solo
practices in established city neighborhoods and new
group practices in the rapidly growing suburbs.
Inner city areas, with increasing minority
popula-tions, were underserved. Their residents relied
heavily on public facilities, largely hospital
emer-gency rooms, for acute care. Two factors changed
this pattern. First, neighborhood health centers
were established in underserved areas and provided
primary care for minority groups through
multi-specialty teams or family practice groups. Second, the cost explosion in the health care sector led to efforts for cost containment through several
pre-payment plans. By mid-1979, there were 12 solo
practitioners, 38 group practitioners (two- to
five-pediatnican partnerships), and 24 health center
practitioners providing primary pediatric care in
the community. One third of the health center
pediatricians worked in suburban, health
mainte-nance organization-structured group practices; two thirds worked in urban neighborhood health center (NHC) multispecialty teams.
Primary pediatric care in hospitals has, for the
most part, disappeared. Outpatient clinics have
been converted to group practices working out of professional buildings located on hospital grounds. Two small continuity clinics, each about the size of a solo practice are still operating to train residents
in primary pediatric care.
The number of emergency room (ER) visits to
the seven acute care hospitals has stayed at the
same level of about 210,000 visits per year over the last decade, but the proportion of child visits has continued to drop. Nearly 30% of ER visits in 1972 were by children; by 1982, this had fallen to 23% at
the two largest hospitals. Most of the visits were
considered to be appropriate by the pediatric direc-tors of the ERs.
Pediatric beds were reduced in all Rochester
hos-pitals except at the University Medical Center. The
pediatric bed occupancy rate for all pediatric beds, reported to the Health System Agency, decreased
from 70.9% in 1970 to 54.9% in 1981, with large
variations between hospitals.
All these figures indicate a marked improvement
in the medical care situation for children.
Rochester Ambulatory Medical Care Survey, 1979
In 1979, we conducted the, to our knowledge,
largest ambulatory medical care survey ever
de-signed for dense sampling of pediatricians and
chil-dren in one community. As in 1977, the major focus
was on mental health problems known to the
pe-diatrician, but all visits were carefully studied to
obtain accurate prevalence rates. The sample
selec-tion is described under Appendix.
RESULTS BY PRACTICE
The 30 pediatricians reported 21,537 visits by
18,181 different children during the study
(un-weighted sample). This sample represents
TABLE 2. Practice Patterns for Participating Pediatricians, Rochester Ambulatory Care Medical Survey, 1979 Solo
Practice
Group Practice
Health Centers
NHC HMO
Total
Un- Weighted*
weighted
No. of physicians 6 15 6 3 30
Av age of MD (yr) 53.8 46.4 50.2 42.0 48.2 48.0
Av time since Board certification (yr) 11.7 6.9 13.5 2.3 8.7 8.3
Reported workload:
Office visits/wk 113.33 111.80 82.17 86.67 103.67 103.92
Phone calls/wk 56.50 57.80 32.50 41.67 50.77 50.97
Newborns/mo 8.17 10.20 11.50 12.67 10.30 10.55
Sessions/wk:
Mornings 4.22 4.33 3.00 3.50 3.96 3.92
Afternoons 3.67 4.00 3.67 2.67 3.73 3.78
Evenings Total
!di
8.06 8.33
_Q
6.67 6.50P4Z
7.76P
7.75Observed workload:
Office visits/2 mo 824.50 773.67 491.33 679.00 717.90 725.15
No. for well-child care 348.10 284.00 178.02 161.99 263.43 261.07
No. by black patients 32.30 29.46 189.33 46.34 63.69 76.68
No. by Medicaid patients 30.45 10.80 184.18 17.65 50.09 62.84
% well child visits 42.2 36.7 36.2 23.9 36.7 36.0
% by black patients 3.9 3.8 38.5 6.8 8.9 . 10.6
% by Medicaid 3.7 1.4 37.5 2.6 7.0 8.7
Av age of patients 6.71 6.23 5.18 6.21 6.12 6.07
Av SEA of patients 2.42 2.03 3.27 2.55 2.41 2.43
Children seen/2 mo 668.67 662.07 410.33 592.00 606.03 611.83
Children with mental health problems, 1st 33.83 27.07 33.17 35.33 30.47 31.15
visit
Children with mental health problems, re- 4.00 1.20 .83 .67 1.63 1.56
peat visit
% prevalence of mental health problems 5.1 4.1 8.1 6.0 5.0 5.1
% incidence of mental health problems 0.6 0.2 0.2 0.1 0.3 0.3
* Weights varied from .8108 for solo practitioners to 1.6211 for health center practitioners recruited during the first
half of the year. Weighting inflates the number of visits and children seen slightly, but does not bias use of standard errors in a major way. For an estimate of the total annual volume of visits, multiply average office visits by 30 (number
of physicians in study) times 6 (each reporting on 1/6 of the year) times 2.467 (inverse of sample fraction .4054).
Abbreviations used are: NHC, neighborhood health center; HMO, health maintenance organization; SEA,
socioeco-nomic area.
County. Of these children, 964 had an emotional,
behavioral, or school problem known to the
pedia-tnician, a 5.30% period prevalence rate of known
mental health problems during the time children
could have come for repeat visits. Most of these
children (914 of the 964) were reported as having a
problem at the first visit during the 2-month
re-porting period (point prevalence of 5.03%), but some (50) were only reported or recognized as such at a repeat visit. After weighting, the number of children increased to 18,350.7, 934.6 of whom had a mental health problem reported at their first visit (point prevalence of 5.09%). The standard error for this rate of .61%. No further mention ofthis finding will be made here; rather, we will focus on pediatric
practices and on children in the community. A
detailed analysis of the visits made to pediatricians by children with mental health problems has been reported elsewhere.8
There were major differences as to age, training,
and work load among the participating
pediatni-cians (Table 2). Their average age was 48, with solo
practitioners the oldest (54 years) and HMO health
center practitioners the youngest (42 years). HMO
pediatricians had been Board certified in pediatrics for an average of only 2 years; solo practitioners, on the other hand, had been certified for an average of nearly 12 years. The study pediatricians in gen-eral worked about 8 half days per week in offices,
with morning and afternoon sessions being almost
equally represented. Two practitioners (one from a
solo practice, one from an HMO) reported having
evening sessions. Health center practitioners were
more flexible in terms of working part time if
needed or allocating half days to administrative duties. The figures shown in each column of Table 2 are unweighted, except for the right-most column, which shows weighted results for all pediatricians
in the community, adjusting for oversampling solo
practitioners and undensampling health center
pe-diatnicians.
visits per week and 50 telephone consultations per
week. Pediatricians in private practice were
signifi-cantly busier, reporting 110 visits per week corn-pared with 80 to 90 visits per week by pediatricians in health centers. Health center pediatricians, how-ever, took more newborns into their practice: health center practitioners reported 12 to 13 newborns as new patients per month, compared with eight per
month for solo practitioners and ten per month for
group practitioners. If we use the reported number of newborns per month to estimate the total num-ber of newborns accepted as new patients by
pedia-tricians during 1979, the estimate (9,370) is about
97% of all live births (9,677) by residents of the
county. (The estimate is the number of newborns
per month in all 30 practices times 12 months times the inverse of the sampling fraction, ie, 316.5
x
12x
(74/30) = 9,370.) The remaining newborns arecared for by family practitioners or by pediatric residents at university hospital-affiliated
continu-ity clinics.
The characteristics of the children seen also dif-fered by practice type. Black patients (10.6% of the sample of visits) were seen mostly at the neighbor-hood health centers. Medicaid patients (8.7% of the sample), too, were concentrated in the urban neigh-borhood health centers, where they comprised more than one third of all patients. Both black patients and Medicaid patients are underrepresented in this
study in relation to the total child population
(15.2% black, about 11% Medicaid). This
underrep-resentation is especially marked for older age
groups. Up to age 1 year, the proportion of blacks in the study was 11.2%; for ages 2 and 3 years, the proportion was 10.1%; it remained between 7% and
8% for ages 4 to 15 years; and after age 16 years
there were hardly any blacks among the pediatri-cians’ patients. The average age of all patients seen
was 6.1 years, ranging from a low of 5.2 years for
the neighborhood health center pediatricians to a
high of 6.7 years for the solo practitioners. The socioeconomic index (reversed range: 1 “highest,” 5
“lowest”) for the census tracts in which the patients
lived also showed considerable variation. The high-est value (2.03) was observed for the suburban group practitioners, the lowest value (3.27) for the neighborhood health center practitioners.
Pediatricians in private practices devoted a
higher proportion of their time to well child care: 42% of visits to solo practitioners were devoted to well child visits compared with 24% to HMO
pedia-tricians. The HMO employs one pediatric nurse
practitioner for every two pediatricians on the staff, allowing pediatricians to focus on children with acute problems.
Practices, however, were only the entry point to the study of children and their visits. We will
present data on the medical complaints or reasons
for the visits, the diagnostic impressions, and the
“continuity of care” provided in terms of knowledge
of the patient; demographic variables such as age, sex, race, and ethnicity will be key to the analysis. Household/family structure, Medicaid enrollment, and census tract information are used to describe the socioeconomic context in which the children lived. We will first take the community perspective and link this study to earlier projects that estimated volume of services and utilization rates for children. We will then take the practitioner’s perspective and examine continuity of care and the types of pnob-lems presented.
VOLUME
OF PEDIATRIC
OFFICE
VISITS
AND
AVERAGE
UTILIZATION
RATES
In our earlier community studies, we estimated total volume of services both on the basis of pro-vider information and on household interview data,
cross-validating the information from one source
with that of the other. Total hospital-based visits could be obtained from service statistics (ER visits,
outpatient visits); the number of private practice
visits was estimated from the results of a survey of
practitioners about their weekly work loads.6’9
Knowing the pediatrician’s average number of of-fice and home visits plus telephone consultations
per week, the number of weeks worked per year,
and the number of full-time equivalent
practition-ers, allows one to calculate the total volume of
pediatric services rendered to the known child
pop-ulation of the county.
We had observed a decline in the average number
of office visits by individual children to
pediatni-cians- from 1.82 visits per child in 1967 to 1.71
visits per child in 1969 to 1.64 visits per child in
1971. Using a similar estimation technique for 1979
gives a total volume of 296,696 visits, which
amounts to an average rate of 1.55 visits per child
for 1979. (The estimate for 1979 is obtained by
multiplying the 2-month reported number of visits
by 6 to obtain the 12-month volume, and
multiply-ing this total with the inverse of the sampling
fraction for pediatricians.) Because of a somewhat
older cHild population in 1979 compared with that in 1971, this rate appears similar to the 1.64 visits
per child reported 8 years earlier. The use of
non-pediatricians, however, may have declined. No
re-cent data on the use of nonpediatnician physicians
by children are available for the Rochester area.
The reported work load for office visits to private practice pediatricians (113/week) in this 1979 study
is somewhat lower than the reported work load for
1975 (124/week). The national average for
with earlier studies, the number of phone consul-tations, reported as 50/week, decreased. Up to 125
phone consultations had previously been
ne-ported.’1’3
The rates reported above were for all children up
to age 17 years and have to be specified as to
subgroup. For example, the younger the patient, the higher the utilization rate. Visit rates by age are shown in Table 3. Children from outside the county and children already 18 years or olden were excluded from these calculations as census data for 1980 for the county served as denominator. For infants and 1 year olds, the rate was 4.90 visits pen year; for preschoolers 2 to 5 years of age the rate was 2.09
visits per year. In the 15- to 17-year age group, the
rate was only .50 visits pen year. These rates for our community are much higher than correspond-ing national data. The National Ambulatory
Med-ical Care Survey (NAMCS) for 1975, for which a
breakdown by pediatric age groups was published,’4 reported rates only half that large (less than 2 years,
2.44 visits; 2 to 5 years, 1.05). NAMCS also showed
that nationally in the preschool years, only 54% of physician visits were to pediatricians. The Roch-ester utilization rates for office visits to
pediatni-cians were even higher than the total of office visits to all types of physicians (4.32 for those less than age 2 years, 2.06 for those 2 to 5 years old) reported by NAMCS. In short, the figures for Rochester show very high average rates of pediatrician office visits and thus indicate a very high “market share” by this specialty of all physician office visits made by children. To the extent that other areas of the country have not yet reached a similar high level, one can argue for an additional need for pediatni-cians.
There are no recent household survey data on
physician utilization available for Rochester. This makes it impossible to quantify the pediatrician’s
market share beyond 1975. Also, we cannot give a
recent distribution of visits over children; some
children will have had no pediatrician contact, oth-ers will have had many times the average rate.9
Given our sampling of children through
pediatni-cians’ offices, we are more likely to sample high utilizers than low utilizers. For example, sampling children through physicians’ offices, leads to oven-sampling of chronically sick children, as a subset of high utilizers. Only a community household sam-ple would provide an unbiased sample of children.
TABLE 3.
Visits by Sociodemographic Characteristics-Only Children Residing in Mon-roe County, NY, and Less Than 18 Years of Age; Weighted AnalysisVisits Reported No. of Children Rate/yr* () Patient, Seen Before? Days Since Last Visit Utiliza-tion Rate/yr No Yes
Total 20,047 191,828 1.55 20.4 79.6 131.82 2.77
Age group (yr) 0-1 2-5 6-10 11-14 15-17 6,279 5,009 4,734 2,699 1,326 18,950 35,387 51,774 46,653 39,064 4.90 2.09 1.35 0.86 0.50 14.8 85.2 21.5 78.5 24.1 75.9 22.5 73.5 25.0 75.0 47.37 113.86 178.23 239.79 261.83 7.71 3.21 2.04 1.52 1.39 Sex Female Male 9,589 10,456 93,673 98,155 1.52 1.58 19.9 80.1 20.8 79.2 132.47 131.53 2.76 2.78 Race White Black Asian American Indian 17,374 2,276 368 24 154,920 29,186 1,719 345 1.66 1.15 3.17 1.03 20.3 79.7 20.8 79.2 22.2 77.8 16.7 83.3 133.38 124.66 115.21 105.72 2.74 2.93 3.17 3.45 Hispanic No Yes Unknown 19,080 739 222 184,761 7,067 . .. 1.53 1.55 ... 20.1 79.9 22.5 77.5 39.6 60.4 132.97 110.64 122.70 2.74 3.30 2.97 Socioeconomic area 1 Highest 2 High 3 Middle 4 Low SLowest 4,437 7,315 5,988 1,433 847 33,537 71,385 60,917 17,839 8,155 1.96 1.52 1.45 1.19 1.59 25.7 74.3 17.8 82.2 18.9 81.1 22.8 77.2 21.6 78.4 154.31 132.37 123.90 105.64 117.32 2.37 2.76 2.95 3.46 3.11
Variations
in Utilization
Rates
Acknowledging this bias in our child sample
(nonusers completely missed, high users overrep-resented), it is still useful to compute utilization rates both for all children in the community and
for children in this sample to check for differences by sociodemographic characteristics in addition to those already presented for age (Table 3). The rates for all children are computed by dividing visits in the sample oven the known number of children from census data. For rates for our sample, a different
technique is used. The pediatricians were asked
whether the child was their patient or some other doctor’s patient. If the child was their patient, they were asked whether this was the first visit of the child. If not, what was the date of the previous visit? For children who were their patients and had been seen before, the average interval between
vis-its is a substitute measure for a utilization rate.
The less clustering of visits around illness episodes,
and the more well-child visits are involved, the
more valid is such a substitute. An average visit interval of 3 months translates, of course, into a
utilization rate of four visits per year. (Total
months [days] per year divided by visit interval in
months [days] equals the utilization rate.) The more cases there are in a category and entered into the computation of an average interval, the better the estimate of the yearly rate for that category.
We made use of this relationship by computing the
average number of days since the last visit (Table 3) for various subgroups and estimating utilization rates for such subgroups. Table 3, therefore,
pre-sents two estimates of utilization rates-one for all
children in the community and one for the children in the sample (more precisely, for those children in
the sample for whom we know the date of the
previous visit).
Age is the best predictor of utilization rates both for the total population and for our sample. Infants through the first year of life in our sample averaged
7.71 visits to the pediatrician; teenagers aged 15 to
17 averaged only 1.39 such visits. Boys (2.78 visits for our sample, 1.58 for the total population) had slightly more visits per year than girls (2.76 for our sample, 1.52 for the population). Black patients
(2.93 visits per year) had significantly higher
utili-zation rate than white patients (2.74 visits per year) in our sample but a much lower rate than whites for the population (1.15 visits compared with 1.66
visits). Hispanics had higher than average rates
both for our sample (3.30 visits pen year) and the
population (1.55 visits). In our sample,
socioeco-nomic area (SEA) was negatively related to
utili-zation: patients from the highest SEAs had the
lowest (2.37 visits) rate, patients from low SEAs
had the highest (3.46 visits per year) rate. An
op-posite pattern is seen for the population rates.
Whenever such opposite patterns are observed for
our sample and for the total population, the best interpretation is that once in the system, rates are high, but barriers may still exist as to getting into the system of pediatric care. Further multivaniate analysis is needed to adjust these figures for age, illness, and other possibly confounding variables, but the evidence so far shows that once a child is in the system, there appears to be no sociodemo-graphic barrier to utilization in this community.
Differences may, however, persist in terms of
get-ting into the system. Only household surveys and ambulatory care surveys including nonpediatnician providers can answer that question.
An analysis of visits by sociomedical character-istics of the child is shown in Table 4. In terms of family structure, the impact was mixed: Children with only the mother at home had a slightly higher utilization rate once in the system than those with both parents at home; on a population basis, how-ever, their utilization was lower. Medicaid patients had higher utilization rates than non-Medicaid pa-tients once they were in the system, but lower rates than the population as a whole. Patients at
neigh-bonhood health centers and HMOs were more
fre-quent users, once in the system, than those of solo
and group practitioners. (Population rates by
pro-vider could not be computed because no census data on practice enrollments are available.)
Reason for visit, of course, was strongly related to utilization in our sample. As expected, those being seen for well child care had the “lowest” rate
(2.26 visits), coming closest to the population rate
for all children in the community (1.55 visits). Low utilization rates (2.24 visits) were also computed for those seen because of accidents. Children with an acute episode of illness were the highest current utilizers (3.49 visits per year), due to return visits required for acute cane and the resulting short av-erage interval between visits. Children with chronic physical conditions and those being seen for
coun-seling, surprisingly, did not differ much from the
average pattern. Whenever the reason for a visit is
associated with a clustering of visits, the short
duration between repeat visits leads to high esti-mates of utilization rates. Those in need of cane have, at that time, the highest utilization, but only
a small proportion of the population is in need of
care at any time.
Variations
in Continuity
of Care
Continuity of care can only be present if the child
is the pediatrician’s own patient, and only if the
TABLE 4. Visits by Sociomedical Ch 18 Years of Age; Weighted Analysis
aractenistics -Only Children Residin g in Monroe Coun ty, NY, and Less Than
Visits Reported
No. of Rate/yr*
Children #{216}Patient, Seen Before? No Yes Days Since Last Visit Utiliza-tion Rate/yr
Total 20,047 191,828 1.55 20.4 79.6 131.95 2.77
Family structure Both parents Mother onlyt Father onlyt Stepparent Unknown 16,205 2,231 97 827 687 155,083 1.68 31,979 1.03 4,766 0.30 --1: 19.4 80.6 19.6 80.4 27.8 72.2 22.5 77.5 42.4 57.6 130.37 128.35 236.10 166.19 128.83 2.80 2.84 1.55 2.20 2.83 Medicaid No Yes 18,210 1,837 171,828 1.57 20,000 1.36 20.1 79.9 22.9 77.1 134.75 104.33 2.71 3.50 Practice type Solo Group
Neighborhood health center Health maintenance organization
3,696 11,027 3,841 1,483 Not applicable (NA) NA NA NA 6.8 93.2 22.4 77.6 21.1 78.9 36.9 63.1 136.41 142.91 99.01 120.45 2.68 2.55 3.69 3.03 Reason Preventive Accident Acute Chronic Emotional/counseling Other 7,414 1,032 11,046 436 93 26 NA NA NA NA NA NA 8.1 91.9 37.3 62.7 26.8 73.2 28.2 71.8 9.7 90.3 15.4 84.6 161.68 162.95 104.59 122.21 146.22 136.71 2.26 2.24 3.49 2.99 2.50 2.67 Continuity
Not own, 1st visit Not own, repeat visit Own, 1st visit Own, 1-4 times Own, 5+ times Own since birth
1,441 1,850 791 1,900 4,105 9,960 NA NA NA NA NA NA
100.0 . ..
100.0 . ..
100.0 . ..
... 100.0
. .. 100.0 .. . 100.0
... . .. ... 172.71 164.75 110.57 NA NA NA 2.11 2.22 3.30
See text for computation.
t The 36,745 children living in nonmarried families were divided proportionately as to one-parent households with
male or female head.
:1:Not reported in census; visits for these categories combined with “Both parents” for rate calculation.
albeit condensed (see Tables 3 and 4), is a require-ment for estimating utilization rates for our sample of children. A detailed breakdown of our continuity index by sociodemographic and sociomedical cate-gonies is provided in Tables 5 and 6. Patients were first classified as to whether they were the
physi-cian’s own patient and whether they had been seen
before. If the patient was the physician’s own, we checked how often he or she had been seen
before-either never, one to four times, five or more times,
or if the physician had known the patient since birth. The breakdown according to this continuity
index is presented in row percentages. The
break-down of the visits by sociodemognaphic and
so-ciomedical categories is presented in column pen-centages.
To some degree, continuity of care is a function
of the age of the child. The older the patient, the
greater the likelihood of him or hen having been
seen before, even if not as the physician’s own
patient, and the less the likelihood of him or her
having been known by the physician since birth. Sex of the child is unrelated to continuity of care, as would be expected. Black and Asian patients, because of their use of neighborhood health centers, are more likely to be new patients and less likely to have been known since birth; there is no evidence,
however, that they lack continuity of care. This is
even more so in the case for Hispanics.
Most of the differences in “continuity” appear to be related to the pediatrician’s type of practice. Solo practitioners had the greatest number of pa-tients they could call their own, with only 4.2% of
the patients seen by them not being their own
patients; 62.7% of the patients seen by them had been known to them since birth. In our community, pediatricians practicing in HMOs ranked lowest on our measure of continuity. Some 23% of their
pa-tients were not their own, and only 22% had been
TABLE 5.
Continuit NY, and Less Than 1y of Care by
8 Years of Age
Sociodemographic Charactenistics-O ; Weighted Analysis
nly Children Residing in Monroe County,
No. of Children
Column
Percents .
Not Own Patient
Never Seen
Row Percents
. Own Patient
New Seen 1-4 Seen 5+ Known
Seen Before Patient Times Times Since
Before Birth
Total 20,047 100.0 7.2 8.9 3.7 8.9 20.6 50.7
Age group (yr)
0-1 6,279 31.3 6.7 5.2 2.5 4.6 3.6 77.4
2-5 5,009 25.0 6.7 9.6 4.8 9.7 17.1 52.1
6-10 4,734 23.6 7.5 11.2 4.4 10.7 27.8 38.4
11-14 2,699 13.5 7.8 10.1 3.9 12.5 40.4 25.3
15-17 1,326 6.6 8.2 13.0 3.1 11.9 46.1 17.7
Sex
Female 9,589 47.8 7.0 8.6 3.8 9.6 20.1 50.9
Male 10,458 52.2 7.3 9.2 3.6 8.2 21.0 50.6
Race
White 17,379 86.7 7.1 9.4 3.2 8.1 21.0 51.2
Black 2,276 11.4 8.2 5.0 7.7 15.6 16.4 47.2
Asian 368 1.8 6.5 6.8 7.7 11.3 19.0 48.7
American Indian 24 .1 8.0 8.0 .0 4.0 36.0 44.0
Hispanic
No 19,086 95.2 7.0 9.1 3.5 8.6 20.6 51.2
Yes 739 3.7 9.2 2.7 9.5 18.0 19.8 40.6
Unknown 222 1.1 15.3 18.6 4.9 7.7 17.5 36.1
Socioeconomic area
1 Highest 4,437 22.1 8.2 13.7 3.5 7.7 23.6 43.4
2 High 7,315 36.5 6.2 8.6 2.4 7.3 21.7 53.8
3 Middle 5,988 29.9 6.8 7.7 4.2 9.7 18.8 52.9
4 Low 1,433 7.1 10.0 3.9 8.1 15.5 13.7 48.8
5 Lowest 874 4.4 9.4 4.8 7.9 15.9 17.3 44.7
many of their partners’ patients and had therefore many patients that were not their own.
Reason for visit was also strongly related to con-tinuity of care. Of the patients seen by physicians other than their own, 35% came to that physician because of an accident (5% of all visits); physicians, however, rarely provided preventive care to patients
other than their own. A somewhat disappointing
finding was the low level of continuity of cane
provided for children with chronic illnesses. Very
often, these patients (22%) were not seen by their
own doctors, and only 26% had been known to their
doctor since birth.
There was no evidence for a lack of continuity
due to Medicaid status. Medicaid patients were no
more likely to be seen by others than by their own
physician. When seen by their own physician,
how-ever, they were more likely to be new patients on to
have been seen only one to four times previously
rather than five or more times previously on since
birth.
Children living with both parents had the
great-est continuity of care; this was probably due,
how-ever, to the type of practice rather than the nature ofcare rendered (continuity v noncontinuity).
Chil-dren from incomplete families on those living with
stepparents were more likely to be patients in
NHCs and to be new patients on only known to the physician from a few previous visits. Staffing of
neighborhood health centers, age of children with
stepparents, and greaten mobility of incomplete
families probably contributed most to this pattern of “noncontinuity” of cane.
Visits by Diagnostic
Category
Approximately 39% of all visits were for preven-tive care, with a strong relationship of this category to both age and sex of the patient. For infants, about 55% of all visits were for well child care; for
teenagers 15 to 17 years of age, the proportion was
only about 30%. Girls had somewhat more visits
for preventive care than did boys.
Respiratory tract infections and ear infections
were the two largest acute care categories, both
accounting for approximately 12% of all visits and
both more frequent in the preschool years. Otitis
media peaked (18%) in the 2- to 5-year age group.
Pharyngitis accounted for about 5% of all visits,
TABLE 6. Continuity of Care by Sociomedical
Less Than 18 Years of Age; Weighted Analysis
Characteristics-Only Children Residing in Monroe County NY, and
No. of Children
Column Percents
Row Percents
.
Not Own Patient Own Patient.
Never Seen
Seen Before
Before
New Seen 1- Seen 5+
Patient 4 Times Times
Known Since Birth
Total 20,047 100.0 7.2 8.9 3.7 8.9 20.6 50.7
Family structure
Both parents 16,205 80.8 6.7 9.3 2.9 7.5 19.9 53.7
Mother only 2,231 11.1 6.1 5.6 7.4 15.2 21.2 44.0
Father only 97 0.5 8.2 8.2 9.2 14.3 29.6 30.6
Stepparent 827 4.1 5.7 9.6 6.8 15.2 34.1 28.7
Unknown 687 3.4 24.6 9.2 8.9 14.8 14.8 27.6
Medicaid
No 18,210 90.8 7.0 9.2 3.3 8.2 20.9 51.3
Yes 1,837 9.2 8.0 5.4 8.7 17.3 16.4 44.2
Practice type
Solo 3,696 18.4 3.1 1.1 2.7 6.1 24.3 62.7
Group 11,027 55.0 6.4 13.3 2.7 5.6 19.1 53.0
Neighborhood health center 3,841 19.2 6.8 5.3 8.1 19.9 18.7 41.2
Health maintenance organiza- 1,483 7.4 22.7 8.5 5.7 18.2 22.6 22.3
tion
Reason
Preventive 7,414 37.0 0.9 0.8 5.9 8.5 17.1 66.7
Accident 1,032 5.1 18.1 16.5 2.3 10.2 24.9 28.0
Acute 11,046 55.1 10.6 13.5 2.3 8.9 21.6 43.0
Chronic 436 2.2 4.0 18.2 5.4 8.8 37.5 26.1
Emotional/counseling 93 0.5 1.1 4.5 4.5 19.1 36.0 34.8
Other 26 0.1 3.6 7.1 3.6 7.1 35.7 42.9
these conditions was in the 6- to 10-year age group.
Eczema and other dermatitides represented nearly
2% of all visits and were more prevalent among boys. The details for all the other major diagnostic
categories are shown in Table 7. Note that the
category “other symptoms and conditions” becomes
the largest category in the later teens. In the first few years of life, the 18 diagnostic categories listed
account for more than 90% of all diagnoses made
in ambulatory care.
DISCUSSION
Our first study of primary pediatric cane in one
community was published in 1978.6 This second
study covers a later and larger sample. Both are
relevant for designing curricula for the training of
new pediatricians and for anticipation of national
tends and manpower needs. A special paper will
specify these implications by comparing our
corn-munity findings with national figures from
NAMCS’4 and with those from the pediatric prac-tice study conducted by the University of Southern
California School of Medicine in i977.’ However,
we also wanted to address questions that go beyond a statistical profile of pediatric practice. Are the
achievements in equity of health care for children
so solidly incorporated in the health care system
that we need not worry about them anymore? Will
the recent federal cuts in funding of child health
services reverse the positive trends of the last 15
years? Rapid advances in neonatology and genetics
practices have made such high levels of maternal
and child health possible that the “traditional” morbidity and mortality still encountered will
in-creasingly be due to “social causes” such as poverty
or lack of care due to maldistnibution of manpower and facilities. What can still be done to manage pediatric care more efficiently and more effectively to reach all children in the community in need of care?
The data used for the above analysis came from
census and vital statistics files and from an
ambu-latory medical care survey. Although a household
survey and the inclusion of physicians other than
pediatricians in our ambulatory care survey, and in settings other than physicians’ offices would have been desirable, our analysis technique did permit
us to assess utilization rates and continuity of care.
The kinds of patients seen and the types of
prob-lems presented to practicing pediatricians are good
indicators of child health in any community. The
TABLE 7. Major Diagnos tic Cate gonies (%) by Age a nd Sex -All C hildren Less Than 1 8 Years of Age; Unweighted Analysis
Category ICDA* Girls byAge (yr) Boys by Age (yr) Total Total
0-1 2-5 6-10 11-14 15-17 0-1 2-5 6-10 11-14 15-17 Girls Boys
Well child . . . 57.9 35.2 29.7 33.0 28.8 53.4 32.3 29.1 31.9 30.7 39.5 37.6 38.5
Upper respiratory 465.0 7.0 7.9 5.4 5.6 4.8 7.2 7.0 4.4 4.2 4.9 6.5 5.9 6.2
tract infection
Bronchitis 466.0 1.3 2.6 2.5 1.5 2.3 1.6 2.9 2.4 1.7 1.5 2.0 2.1 2.1
Pneumonia 486.0 0.3 1.1 1.3 1.0 0.6 0.4 1.5 1.3 0.8 1.0 0.9 1.0 0.9
Asthma 493.0 0.2 1.1 0.6 0.8 1.1 0.4 1.1 1.7 1.3 1.5 0.7 1.1 0.9
Hayfever 507.0 0.0 0.8 1.1 1.6 1.4 0.1 0.8 1.7 2.3 1.9 0.8 1.1 1.0
Respiratory symp- 783.0 0.4 1.3 0.7 1.4 1.4 0.3 1.2 1.7 1.7 1.5 0.9 1.1 1.0
toms
Conjunctivitis 360.0 0.6 0.7 0.5 0.4 0.6 0.6 0.4 0.7 0.5 0.3 0.6 0.5 0.6
Otitis 381.0 12.9 18.0 12.4 5.8 4.4 14.9 18.2 11.8 5.9 3.4 12.4 13.0 12.7
Pharyngitis 462.0 0.7 4.5 8.0 7.9 6.4 0.7 4.7 8.2 5.8 5.5 4.9 4.6 4.7
Tonsillitis 463.0 Q.1 0.7 1.3 0.7 0.9 0.1 1.1 1.0 0.7 0.9 0.7 0.7 0.7
Laryngitis 464.0 0.8 1.4 0.6 0.4 0.0 1.1 1.8 0.7 0.3 0.0 0.8 1.0 0.9
Strep throat 034.0 0.1 1.4 3.1 1.3 2.5 0.4 1.7 2.5 1.4 1.0 1.5 1.4 1.5
Impetigo 684.0 0.3 1.1 0.8 0.3 0.3 0.3 0.7 0.5 0.5 0.3 0.6 0.5 0.5
Eczema 692.0 1.1 0.7 1.4 1.8 1.4 1.3 1.4 2.8 3.0 2.4 1.3 2.0 1.7
Viral disease 079.0 1.7 3.3 4.5 3.5 4.7 2.3 2.2 4.5 3.0 2.9 3.3 3.0 3.1
Diarrhea 009.0 2.2 1.1 0.6 0.5 0.5 2.1 1.1 0.9 0.7 0.3 1.1 1.3 1.2
Gastrointestinal 785.0 0.4 0.4 1.1 2.0 1.4 0.1 0.5 0.8 0.7 1.4 0.9 0.5 0.7
symptoms
General symptoms 788.0 2.2 2.1 1.5 0.6 1.4 2.9 2.7 1.0 1.1 1.0 1.7 2.0 1.9
Other symptoms and . . . 9.9 14.5 22.5 29.8 35.2 9.8 16.7 22.1 32.6 37.5 18.9 19.5 19.2
conditions
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
No. of children 2,456 2,179 2,090 1,243 640 2,770 2,326 2,318 1,321 587 8,608 9,322 17,930
SInternational Classification of Diseases Adapted.
The pattern of primary pediatric care has
changed remarkably over the last decade, as the
practice arrangements in HMOs and neighborhood
health centers document. There are clear effects of
such arrangements on continuity of care, due in
part to the large number of young pediatricians employed; the observed pattern, however, is not that different from continuity care delivered in private group practices. The “place” of care receives the patient’s loyalty as much as the individual physician, a change in attitude that reflects both the greater mobility of patients and providers and the trust of patients in the physician’s role rather than in the person occupying that role. Medicaid
and prepayment plans have contributed to this
change in loyalty toward “centers,” “plans,” or “in-stitutions”and away from the individual physician.
The distribution of care provided to children was
only indirectly assessed. Race, socioeconomic standing, and family structure all appear to affect access to care, bqt once care is established, differ-ences due to these factors disappear. There is evi-dence of differential access to certain sources of care in relation to certain demographic variables, but no evidence for different amounts of care once
contact with the source is made. On the basis of
our data, it is difficult to continue to make a strong argument for the existence of excessive inequities
in the provision of care in our community due to social determinants.
Although improvements in child health and ac-cess to child health care undoubtedly have occurred, pediatricians cannot claim all the credit. The
de-creasing birth rate, in the presence of an ample
supply of health care providers, is probably the
major contributing factor. Delayed child beaning,
smaller families, and better financial situations of
parents also contributed. These trends will
proba-bly continue in the foreseeable future. As the costs of children increase and compete with other costly family needs such as housing and transportation,
two-income families will become the rule and day
care needs will increase.
The recent reductions in federal medical and
social programs for children will shift more of the costs of childrearing to parents. The health of
chil-dren will probably not be directly affected, but
family size may. The future family will be small.
Parents will have only two options-to reduce the
number of children they have or to reduce the
amount they spend on on each of their children,
morbid-ity and mortality will increase if ve fail to protect children from the stressful effects of a competitive
economic world. Community child health studies
will remain a needed method to monitor child
health and pediatric cane.
ACKNOWLEDGMENTS
This analysis is based on data from a study supported by contract 278-78-0038(DB) from the National Institute of Mental Health, Rockville, MD.
istics. When studying prevalence of symptoms and
treat-ments that do not correlate with socioeconomic
charac-tenistics, weight factors need not be applied as they will have no effect on the results. In general, we recommend
using weight factors when analyzing practices and
esti-mating community rates, but not when analyzing visits
or patients for their medical characteristics. Weighting leads to minor differences in prevalence rates, but these
differences are small when compared with the standard
errors involved. Standard errors are relatively large when practices are the unit of analysis and still considerable for the cluster samples of visits and children. Further details of the methodology are reported elsewhere.8
REFERENCES
APPENDIX
Sample Selection
A stratified random sample of 30 pediatricians was
selected from the 74 practitioners in the community. The two continuity clinics were not included in this sample, but studied in a separate project with the same instru-ments.’6 The pediatricians were asked to report for an assigned 2-month period throughout 1979 on all office visits made to them. Inpatient visits, telephone consul-tations, and visits to nurses or nurse practitioners in the
offices were excluded. The main objective was to gain
information on the mental health status, as known to
pediatricians, of a 10% sample of all children seen by pediatricians. The Rochester community is unique in that the majority of children are cared for by pediatricians.
An earlier community survey (1975)’ had shown that
96% of all children had a regular doctor or a “regular” place of care. For those with a regular doctor, 74% had a pediatrician. Regular places of care such as health centers or outpatient clinics were mostly staffed by pediatricians, although family practice groups also served some city neighborhoods. By 1979, the number of general practi-tioners in the community had declined further, and access to pediatricians had increased; it was reasonable, there-fore, to assume that an even larger proportion of the child population than in 1975 was cared for by pediatricians.
An effort was made to represent equally all months of the year and all types of pediatric practice, ie, solo prac-tice, group practice, and health center practice. Pediatri-cians who refused to participate were replaced with back-up pediatricians from each type. After approaching 41 pediatricians, afinal sample of 30 (30 to 74 = 40.5%) was recruited, with some oversampling of solo practitioners (6 of 12 = 50%) and some undersampling of health center
practitioners (3 of 8 = 37.5% of the surburban HMO
practitioners, 6 of 16 = 37.5% for the urban NHC prac-titioners). The group practices were sampled at a rate (15 of 38 = 39.5%) closely approximating the overall rate of 40.5%. When estimating the patient population by social characteristics, weight factors should be applied to cor-rect for these known differences in sampling fractions by practice type because patient populations differ signifi-cantly by practice type in terms of economic
character-1. Koos EL: The Health of Regionville. New York, Hafner Publishing Co, 1954
2. Kunitz SJ, Sorensen AA, Cashman SB: Changing health
care opinions in Regionville, 1946-1973. Med Care 1975; 13:549
3. Haggerty BA, Roghmann KJ, Pless lB (eds): Child Health and the Community. New York, John Wiley & Sons, 1975 4. Hoekelman RA, Sutherland SA, Mclnerny TA, et al:
Col-laborative research between a department of pediatrics and its clinical faculty. Clin Pediatr 1979;18:623
5. Goldberg ID, Regier DA, Mclnerny TK, et al: The role of the pediatrician in the delivery of mental health services to children. Pediatrics 1979;63:898
6. Mclnerney TK, Roghmann KJ, Sutherland SA: Primary
pediatric care in one community. Pediatrics 1978;61:389
7. Roghmann KJ: Looking for the medical care crisis in utili-zation data. Inquiry 1974;11:282
8. Goldberg ID, Roghmann KJ, Mclnerney TK, et al: Mental health problems among children seen in pediatric practice: Prevalence and management. Pediatrics 1984;73:278
9. Roghmann KJ, Haggerty RI: Measuring the use of health services by household interviews: A comparison of proce-dures used in three child health surveys.
mt
J Epidemiol 1974;1:7110. Glandon GL, Shapiro Rd (eds): Profile of Medical Practice
1980: Decade of the 1970s-The Changing Profile of Medical
Practice. Monroe, WI: American Medical Assoc, 1980 11. Hessel SJ, Haggerty TM: General pediatrics: A study of
practice in the mid-60s. J Pediatr 1968;73:271
12. Hercules C, Charney E: Availability and attentiveness: Are these compatible in pediatric practice? Clin Pediatr 1969; 8:381
13. Breese BB, Disney FA, Talpey W: The nature of a small pediatric group practice. Pediatrics 1966;38:264
14. National Center for Health Statistics. Ambulatory Care
Uti-lization Pattern of Children and Young Adults: National Ambulatory Medical Care Survey, United States 1975.
Wash-ington, DC, Dept of Health and Human Services, PHS
Series 13, No. 39, 1978
15. Hoekelman RA, Starfield B, McCormick M, et al: A profile of pediatric practice in the United States. Am J Dis Child
1983;137:1057
16. Sommerfelt A, Roghmann K, Hoekelman R: An educational process study of pediatric residency training, in Lipkin M, Boufford J, Froom J (eds): Primary Care Research in 1982. New York, NYU Medical Center, 1982
17. Roghmann KJ: Codebook for the 1975 Community Child