Docstoc

and Physical Activity

Document Sample
and Physical Activity Powered By Docstoc
					Outcomes of a Field Trial to                                                                      Improve
Children's Dietary Patterns
and Physical Activity
The Child and Adolescent Trial
for Cardiovascular Health (CATCH)
Russell V. Luepker, MD; Cheryl L. Perry, PhD; Sonja M. McKinlay, PhD; Philip R. Nader, MD; Guy S. Parcel, PhD;
Elaine J. Stone, PhD, MPH; Larry S. Webber, PhD; John P. Elder, PhD; Henry A. Feldman, PhD;
Carolyn C. Johnson, PhD; Steve H. Kelder, PhD, MPH; Margaret Wu, PhD; for the CATCH Collaborative Group

   Objective.\p=m-\Toassess the outcomes of health behavior interventions, focusing              CARDIOVASCULAR disease (CVD)
on  the elementary school environment, classroom curricula, and home programs,                   is the leading cause of death in the United
for the primary prevention of cardiovascular disease.                                            States and most other developed coun¬
                                                                                                 tries.1·2 Epidemiological data link CVD
   Design.\p=m-\Arandomized, controlled field trial at four sites with 56 intervention and       to the lifestyles of a population, includ¬
40 control elementary schools. Outcomes were assessed using prerandomization
measures (fall 1991) and follow-up measures (spring 1994).
                                                                                                 ing cigarette smoking, sedentary behav¬
                                                                                                 ior, overconsumption of food, and dietary
   Participants.\p=m-\Atotal of 5106 initially third-grade students from ethnically              intakes that are high in saturated fat
diverse backgrounds in public schools located in California, Louisiana, Minnesota,               and sodium.2·3 These behaviors, learned
and Texas.                                                                                       in childhood and adolescence and estab¬
   Intervention.\p=m-\Twenty-eightschools participated in a third-grade through fifth\x=req-\    lished by young adulthood, form the ba¬
grade intervention including school food service modifications, enhanced physical                sis of the insidious development of ath¬
education (PE), and classroom health curricula. Twenty-eight additional schools                  erosclerosis.4"8 Because health behaviors
received these components plus family education.                                                 are imbedded in the social fabric of

   Main Outcome Measures.\p=m-\Atthe school level, the two primary end points were               American life, populationwide preven¬
                                                                                                 tion strategies aimed at young people
changes in the fat content of food service lunch offerings and the amount of                     and complemented by interventions for
moderate-to-vigorous physical activity in the PE programs. At the level of the indi-             those at elevated risk are widely rec¬
vidual student, serum cholesterol change was the primary end point and was used                  ommended.2·9·10
for power calculations for the study. Individual level secondary end points included               Schools, with their trained staff, their
psychosocial factors, recall measures of eating and physical activity patterns, and              accessto most children and adolescents,
other physiologic measures.                                                                      and organizational structures, policies,
   Results.\p=m-\Inintervention school lunches, the percentage of energy intake from
fat fell significantly more (from 38.7% to 31.9%) than in control lunches (from 38.9%
to 36.2%)(P<.001 ). The intensity of physical activity in PE classes during the Child
                                                                                                   From the Division of Epidemiology, University of
and Adolescent Trial for Cardiovascular Health (CATCH) intervention increased                    Minnesota School of Public Health, Minneapolis (Drs
significantly in the intervention schools compared with the control schools (P<.02).             Luepker and Perry); New England Research Institutes,
                                                                                                 Watertown, Mass (Drs McKinlay and Feldman); De-
Self-reported daily energy intake from fat among students in the intervention                    partment of Pediatrics, Community Pediatrics Division,
schools was significantly reduced (from 32.7% to 30.3%) compared with that among                 University of California at San Diego, La Jolla (Drs
                                                                                                 Nader and Elder); Center for Health Promotion Re-
students in the control schools (from 32.6% to 32.2%) (P<.001). Intervention stu-                search and Development, University of Texas Health
dents reported significantly more daily vigorous activity than controls (58.6 minutes            Science Center, Houston (Drs Parcel and Kelder);
vs 46.5 minutes; P<.003). Blood pressure, body size, and cholesterol measures                    Project Office, National Heart, Lung, and Blood Insti-
                                                                                                 tute, Bethesda, Md (Drs Stone and Wu); and Tulane
did not differ significantly between treatment groups. No evidence of deleterious                University School of Public Health and Tropical Medi-
effects of this intervention on growth or development was observed.                              cine, New Orleans, La (Drs Webber and Johnson).
                                                                                                    A complete list of the members of the CATCH Col-
   Conclusion.\p=m-\TheCATCH intervention was able to modify the fat content of                  laborative Group appears at the end of this article.
school lunches, increase moderate-to-vigorous physical activity in PE, and improve                  Reprint requests to Division of Epidemiology,
                                                                                                 University of Minnesota School of Public Health, 1300
eating and physical activity behaviors in children during 3 school years.                        S Second St, Suite 300, Minneapolis, MN 55455
                                                                      (JAMA. 1996;275:768-776)   (Dr Luepker).


                                      Downloaded from jama.ama-assn.org by guest on May 7, 2011
and environments that can support             Design                                        The CATCH Intervention     Program
healthy behavior, are ideal sites for
prevention programs.11 Schools have              The schools were randomized to ei¬            The CATCH interventions consisted
traditionally provided information and        ther intervention (56 schools [14 per field   of school-based (school food service, PE,
knowledge concerning hygiene and              center]) or control status (40 schools [10    classroom curricula) and family-based
health. However, health programs based        per field center]). Randomization oc¬         (home curricula, family fun nights) com¬
on this premise showed little evidence        curred after all baseline measurements        ponents.22 School food service changes
of influencing behaviors.12 Beginning in      were     completed. The intervention          and PE enhancement were continuous
the late 1970s, a series of school health     schools were further randomized into          programs during the 3 school years. The
promotion programs based on behavior          two equal subgroups. One group re¬            classroom and home curricula were
change models from social psychology          ceived a school-based program consist¬        implemented by classroom teachers over
rather than traditional educational           ing of school food service modifications,     a fixed time period during each school
models were tested.13 These programs,         PE interventions, and the CATCH cur¬          year, and addressed eating habits
which were concerned with preventing          ricula (28 schools [seven per field cen¬      (grades 3 through 5), physical activity
 cigarette smoking and promoting health¬      ter]); the other group received the same      (grades 4 and 5), and cigarette smoking
ful eating or physical activity patterns,     school-based program plus a family-           (grade 5 only). Below is an overview of
primarily with innovative classroom           based program (28 schools [seven per          the intervention components, with de¬
 curricula, demonstrated success in chang¬    field center]). The CATCH interventions       tails provided elsewhere.22·26
ing short-term behavior.10·1317 Longer-       began in the 1991-1992 school year and           Eat Smart was the food service in¬
term behavioral outcomes were noted           continued as students progressed              tervention, the goal of which was to pro¬
for programs that were multiyear, that        through grade 5 (1993-1994). The con¬         vide children with tasty meals that were
included environmental components,            trol group received the usual health cur¬     lower in total fat (to 30% of energy),
and that elicited parent and community        ricula, PE, and food service programs,        saturated fat (to 10% of energy), and
support.18"21                                 but none of the CATCH interventions.          sodium (600 to 1000 mg per serving),
   The Child and Adolescent Trial for         The design of the study is described in       while maintaining recommended levels
Cardiovascular Health (CATCH)22 was           detail elsewhere.25,26                        of essential nutrients and child partici¬
designed to augment the research ofthe           The primary study comparison was           pation in the school meal programs.27·28
1980s in CVD prevention among young           between the 56 intervention and 40 con¬       Food service personnel participated in a
people by using a sophisticated research      trol schools with respect to changes from     1-day training session at the beginning
design involving a large number of            baseline (fall 1991) to the end of the        of each school year. Monthly follow-up
schools, a multicomponent behavioral          intervention period. There were two pri¬      visits to the schools and booster ses¬
health intervention over three grades,        mary end points at the school level by        sions provided further information, help
and children of diverse communities. We       which the effectiveness of the CATCH          in planning, and support.27·28
report herein the major outcome data          food service and PE interventions was            The CATCH PE intervention sought
from the CATCH interventions.                 assessed. These were reducing the total       to increase the amount of enjoyable
                                              fat content of the food served to 30% of      MVPA during PE classes at school to
METHODS                                       total energy intake and the sodium con¬       40% of the PE class.29"31 The PE spe¬
                                              tent to 600 to 1000 mg per serving and        cialists and teachers had 1 to IV2 days of
Subjects                                      increasing the amount of PE time that         training each school year.
   The CATCH interventions involved           students spent in moderate-to-vigorous           The classroom curricula included the
four field centers (the University of Cali¬   physical activity (MVP A) to 40% of class     Adventures of Hearty Heart and
fornia at San Diego, La Jolla; Univer¬        time. At the level of the individual stu¬     Friends, Go for Healths, and Go for
sity of Minnesota, Minneapolis; Univer¬       dent, serum cholesterol concentration         Health-5, for grades 3,4, and 5, respec¬
sity of Texas at Houston; and Tulane          was the primary end point. It was the         tively.32·33 They consisted of 15, 24, and
University School of Public Health and        basis of sample size determination, with       16 lessons during 5, 12, and 8 weeks in
Tropical Medicine, New Orleans, La), a        expected differences of 0.13 mmol/L (5.1      grades 3, 4, and 5, respectively. Each
coordinating center (New England Re¬          mg/dL) between intervention and con¬          lesson was 30 to 40 minutes in length.
search Institute, Watertown, Mass), and       trol schools at follow-up. Secondary end      The curricula targeted specific psycho-
the National Heart, Lung, and Blood           points included changes in psychosocial       social factors and involved skills devel¬
Institute Project Office, Bethesda, Md.       factors, self-reported dietary fat and so¬    opment focused on eating behaviors and
   Recruitment of schools was based on        dium intake and time engaged in MVPA,         physical activity patterns.34 Addition¬
their distance from one of the four study     and systolic blood pressure. The sec¬         ally, F.A.C.T.S. for Five, a four-session
centers, their ethnic diversity, their food   ondary study comparison, the addition         tobacco use prevention curriculum, was
service's potential for intervention, and     of the home/family component, was             implemented in grade 5. Classroom
their commitment to offering at least 90      addressed by examining differences in         teachers attended 1 to IV2 days of train¬
minutes a week of physical education           secondary end points between the two         ing each year to learn how to implement
(PE) and to participating in a 3-year         intervention conditions.                      the curricula. Standardized protocols
study.85 Schools also were required to           Measurements of school- and student-       were used at all sites to ensure compa¬
cooperate with random assignment to            level outcomes were made at the begin¬       rable implementation.34
treatment or control status.                   ning of third grade (fall 1991). Follow-        The home curriculum involved activ¬
   The study involved 5106 initially third-    up measurements were made in spring          ity packets that complemented the class¬
grade students (mean age 8.76 years at         1994. A substantial portion of stu¬          room curricula.16 These packets were
baseline) in 96 public schools from 12         dents left the district prior to the end      sent home with the students and re¬
school districts24 who agreed to provide       of the study but were tracked within          quired adult participation to complete.
a blood sample at baseline. Those who          a 100-mile radius and underwent mea¬         There were 19 activity packets over the
participated in the blood assessment at        surements to enable analysis accord¬          course of 3 school years. Score cards to
follow-up formed the primary cohort for       ing to the intention to treat prin¬           record points for completing home ac¬
student-level study findings.                 ciple.25                                      tivities were used for giving small re-


                              Downloaded from jama.ama-assn.org by guest on May 7, 2011
wards to encourage family participation.      school at both baseline and follow-up.45        lograms divided by the square of height
In addition, during grades 3 and 4, stu¬      A nonqualified food record45 was com¬           in meters.
dents invited their family members to a       pleted by students on the previous day             Aerobic fitness was assessed in all
"family fun night" to culminate the class¬    and was used as a prompt for the in¬            students using a group-administered
room activities. The fun nights consisted     terviewer who conducted the 24-hour             timed 9-minute run. The test-retest re¬
of dance performances by the students,        recall. The data were directly entered          liability coefficient of the 9-minute run
food booths with healthy snacks, distri¬      into a laptop computer during the in¬           with third-grade students was 0.90. Stan¬
bution of recipes, and games. These fun       terview, and the Nutrition Coding Cen¬          dards of performance for youth are well
nights followed a standardized protocol       ter database was used for evaluation.41         established, and correlations with more
across all sites.35                           This method has previously been shown           complex measures of fitness were esti¬
                                              to be reliable and valid.43·46                  mated during the pilot phase.47·50
Measures                                         The Self-administered Physical Activ¬
                                              ity Checklist (SAPAC) was developed and         Statistical Methods
   Program Implementation.—Process
measures were developed to monitor            validated during CATCH.47 It was ad¬               The statistical approach for analyzing
implementation fidelity of the interven¬      ministered in grade 5 only to assess type,      CATCH data has been described else¬
tions as well as potential confounding        duration, and intensity of selected leisure     where in detail.25·51"53 All end points were
influences and policies within the schools.   time physical activities, television watch¬     analyzed by the single-stage, mixed-
These have been described in detail else¬     ing, and video games.47 Children reported       model strategy to take advantage of the
where.80'34"39                                the number of minutes they had spent            special design feature of CATCH,
  Outcome Measures at the School              during the previous day in various com¬         namely, that randomization was per¬
Level.—The Eat Smart program was as¬          mon physical activities and selected sed¬       formed at the school level while most of
sessed at baseline, in the spring of grade    entary activities.48                            the outcome data were collected by re¬
4, and at follow-up by trained observers         Outcome Physiological Measures.—A            peated measurements on individuals.
who collected recipes, menus of meals         detailed description of the physiological       Analytical models for the various
offered, and vendor product information       measures in CATCH has been previously           CATCH end points took into account
for 5 consecutive days from food service      published,24 and these measures are             the measurement level, frequency, and
personnel and records of student partici¬     briefly summarized below. All of these          nature of each outcome variable.
pation in school meals. The reliability and   measures were performed at baseline and            Individual-level measures were ana¬
quality control procedures of these meth¬     the fifth-grade follow-up.                      lyzed by mixed-model analysis of covari-
ods are described in detail elsewhere.28·40      Nonfasting venipuncture samples were         ance (ANCOVA) with the follow-up value
Recipe and menu analysis was carried          drawn for    lipid determination. The se¬       as the dependent variable, the CATCH
out using the Nutrition Data System of        rum samples were sent to a central labo¬        intervention group as the independent
the University of Minnesota Nutrition         ratory (Miriam Hospital, Providence, RI)        variable, and the baseline value as a co-
Coding Center.41 These analyses provide       for analysis of total cholesterol, with high-   variate. End points analyzed in this man¬
the specific micronutrient and macronu-       density lipoprotein cholesterol and apo-        ner included all physiological measures
trient content of foods served.               lipoprotein measured in a 45% randomly          (n=4019 paired measures on cohort sub¬
   The CATCH PE intervention was as¬          selected subsample. The laboratory suc¬         jects), the 24-hour food recall (n=1182),
sessed by an instrument called the Sys¬       cessfully participated in the Centers for       and the 9-minute run (n=2995). In each
tem for Observing Fitness Instruction         Disease Control and Prevention Lipid            case, the ANCOVA was controlled for
Time (SOFIT).42 During each of the six        Standardization Program.49                      sex, race, CATCH field site, and the ran¬
semesters of the study, every school was         Arm size was measured for selection of       dom effect of school within site and in¬
visited twice by trained observers who        the appropriate cuff size for measurement       tervention group. Additional covariates
used SOFIT to observe the type and            of blood pressure while the participant         were included as appropriate: age, height,
intensity of the children's activities and    was seated for 5 minutes in a quiet room.       and body mass index as predictors of risk
the behaviors of the PE specialists or        Five recordings of systolic and diastolic       factors and weather conditions as a pre¬
teachers in PE classes. The measure           pressure and pulse rate were taken at           dictor of distance in the 9-minute run.
provided a quantitative measure of les¬       1-minute intervals (Dinamap Automated           Each end point was also tested for inter¬
son length and type of physical activity      Device, model 8100XT, Critikron, Ine,           actions ofthe CATCH intervention group
during classes. Training and quality con¬     Tampa, Fla). The average ofthe last three       with site, sex, and race.
trol procedures have been described           readings was used for outcome analyses.            School-level measures were analyzed
elsewhere.29                                  The intraclass correlation coefficients for     by repeated measures ANCOVA with
   Outcome Measures at the Individual         Dinamap readings 3,4, and 5 in the base¬        the CATCH intervention group as the
Level.—The Health Behavior Question¬          line and follow-up data ranged from 0.54        independent variable. Compound sym¬
naire (HBQ) was a class-administered 45-      (systolic blood pressure, third grade) to       metry was assumed, ie, equal correla¬
minute instrument designed to evaluate        0.88 (heart rate, fifth grade).                 tion among repeated measures on a given
factors associated with diet, exercise, and      For each student, triceps and subscapu-      school. Dependent variables analyzed in
smoking at baseline, in the spring of         lar skinfolds were measured using Lange         this manner were school menu data
grades 3 and 4, and at follow-up. It in¬      calipers, height was measured using a           (n 96 schools x 3 semesters 288),
                                                                                                 =                                =


cluded psychosocial data on dietary knowl¬    portable stadiometer (Perspective En¬           school-mean responses to the HBQ
edge, intentions, usual food choices, so¬     terprises, Ine, Kalamazoo, Mich), and           (n 96 schools x 4 semesters 384), and
                                                                                                 =


cial reinforcement and support, and self-     weight was measured using a balance             PE class observations using the SOFIT
                                                                                                                              -




efficacy. The reliability and validity of     scale. Skinfold thickness was measured          (n 96 schools x 6 semesters x 1.97 av¬
                                                                                                 =


the instrument, assessed during the pilot     three times at each site, with intraclass       erage observation days X 1.85 average
phase of CATCH and with the baseline          correlation coefficients exceeding 0.97.        lessons observed 2092). For the pres¬
                                                                                                                 =



data, were found to be adequate.43'44         Height was measured to the nearest 0.1          ent analysis, the HBQ was treated as a
   A 24-Hour Dietary Recall measured          cm, weight to the nearest 0.1 kg, and           cross-sectional characterization of the
total daily food and nutrient intake in a     skinfolds to the nearest millimeter. Body       school rather than a longitudinal mea¬
random subsample of 30 students per           mass index was defined as weight in ki-         sure of an individual's behavior; HBQ



                                     Downloaded from jama.ama-assn.org by guest on May 7, 2011
Table 1.—Number of Children in the Child and Adolescent Trial for Cardiovascular Health   Longitudinal Cohort at Baseline, by Race, Sex, and Site,   1991   Through   1994

                              White                       African American                        Hispanic                            Other

   Site              Boys              Girls           Boys             Girls              Boys              Girls            Boys             Girls                  Total
California             507                                               52                 107                                 56                                    1379
                       510              390             174              188                                                    10                                    1299
Minnesota              609
Texas                                   279              77              100                222              227                                 10                   1191
Total                 1894             1636             313              361                                                                    101                   5106



responses were therefore aggregated by                 baseline were in CATCH schools and 28%                        compared with the control school lunches
school within each measurement period,                 (1455) were not in CATCH schools. Par¬                        (P<.01). The sodium content of the foods
including noncohort students. The mixed                ticipation for the fifth-grade blood mea¬                     provided rose in both intervention and
model for SOFIT included nested ran¬                   sures among students in the CATCH                             control schools, although the differences
dom effects for observation days within                schools was 90% (3297 of 3651) and for                        were not significant. A statistically sig¬
semester and lessons within observa¬                   those who migrated out 50% (722 of 1455),                     nificant increase in potassium (adjusted
tion days. The SOFIT analysis was ad¬                  for a final participation rate of 79% (4019                   for food quantity) was seen in the inter¬
ditionally controlled for the location of              of 5106). There were no significant base¬                     vention school lunches (from 330.0 mg/
the lesson (indoors or outdoors) and the               line differences in conditions for any of                     1000 MJ to 357.9 mg/1000 MJ) compared
specialty of the teacher (classroom or                 the primary or secondary end points be¬                       with the control schools (325.3 mg/1000 MJ
PE) as fixed effects.                                  tween participants and those who were                         to 326.6 mg/1000 MJ) (P<.01).
   Each end point was also tested for                  lost to follow-up. There were also no sig¬                       The average PE lesson length did not
interactions of the CATCH interven¬                    nificant differences in follow-up partici¬                    change significantly, remaining at about
tion group with site. We used SAS pro¬                 pation by intervention assignment or sex,                     30 minutes for both groups (Table 3). The
cedures      for mixed-model analysis                  although African-American students and                        intensity of the physical activity, how¬
throughout.52·53 The values presented                  students in California were more likely to                    ever, increased significantly more in in¬
herein were not adjusted for multiple                  have dropped out.                                             tervention compared with control schools,
comparisons   or multiple outcome mea¬                    Participation in the intervention pro¬                     as shown by higher energy expenditures
sures, since no appropriate denomina¬                  grams was consistently high across the                        during the second, fourth, and fifth se¬
tor for the type I error rate was fea¬                 56 intervention schools, as reported else¬                    mesters and marginally higher energy
sible, given the multiple research                     where.34"36 Training sessions were at¬                        expenditures during the third and follow-
questions addressed in the CATCH.                      tended by 86% of school food service                          up semesters. The time spent at higher
                                                       cooks, 94% of the PE specialists, and                         levels of activity was equivalent in the
Safety                                                 86% of the teachers. Over 90% of the                          two groups at baseline but increased sig¬
   The safety aspects of the CATCH pro¬                school food service guidelines were met                       nificantly in later semesters (Figure). The
grams were monitored by an indepen¬                    during the Eat Smart intervention, and                        curves of the intervention and reference
dent Data and Safety Monitoring Board,                 over 80% of the specified CATCH PE                            conditions diverged significantly, as in¬
which met periodically to review the data.             activities were implemented. Classroom                        dicated by repeated-measures analysis,
Quality control procedures were imple¬                 observations revealed that more than                          for both MVPA (F=2.71, d/=5,1979,
mented throughout the study and moni¬                  88% of the curriculum sessions were                           P=.02) and vigorous activity (F=2.35,
tored by the Quality Control Subcommit¬                completed without modification. Teach¬                        d/=5,1979, P=.04).
tee of the CATCH Steering Committee.                   ers reported completing over 90% of all
                                                       of the activities in the four curricula.                      Student-Level Outcomes
RESULTS                                                More than 70% of the families partici¬
                                                       pated in some home curricula during                              Response scores for dietary knowledge,
Participation and Implementation                                                                                     dietary intentions, and self-reported food
Fidelity                                               each grade level, and over 63% attended                       choice changes on the HBQ were signifi¬
                                                       a family fun night. All of the schools held
   Among the 96 schools measured at                    family fun nights in each of the third and                    cantly greater for the intervention schools
baseline, there were no significant dif¬                                                                             at follow-up (Table 4). Perceived social
ferences in the study conditions for all               fourth grades and implemented over 90%                        reinforcement for healthful eating pat¬
                                                       of the specified activities for these nights.
relevant variables, including environmen¬                                                                            terns also was significantly higher in the
tal, behavioral, psychosocial, and risk fac¬                                                                         intervention groups. The self-efficacy
tor data, ensuring equivalency among                    School-Level Outcomes                                        measures for both diet and physical ac¬
groups. All 96 schools maintained their                    Student participation in school lunch                     tivity, although significantly higher in the
participation in their allocated treatment              programs was between 70% and 75% on                          intervention groups at the end of the first
condition during the 3-year study period.               average throughout the study in all                          year of intervention, showed no signifi¬
   For the student-level measurements,                  schools. Nutrient analysis for lunch menus                   cant difference at follow-up. Self-reported
the CATCH cohort was defined as those                   for intervention and control schools is                      positive social support for physical activ¬
for whom lipid assessment was completed                 shown in Table 2. The primary school-                        ity differed between the ends of the third
at baseline and follow-up (n=4019). The                 level end point, the percentage of energy                    and fourth grades only. The school-plus-
CATCH students measured at baseline                     intake from total fat in the meals, was                      family intervention group compared with
included 60.4% (n=5106) ofthe third-grade               significantly reduced in the intervention                    the school-only intervention group had
students enrolled. There were no signifi¬               school lunches (from 38.7% to 31.9%) com¬                    greater positive changes only for dietary
cant differences by site, sex, or ethnic                pared with the control school lunches                        knowledge.
group among those who did and did not                   (from 38.9% to 36.2%) (P<.001). Energy                         The 24-Hour Food Recall showed in¬
participate. Table 1 presents baseline                  intake from saturated fat and total en¬                      creased total daily energy intake among
characteristics. At the fifth-grade follow-             ergy intake were also significantly re¬                      children in both intervention and control
up, 3651 (72%) of the 5106 students at                  duced in the intervention school lunches                     schools with their aging, but this increase

                                      Downloaded from jama.ama-assn.org by guest on May 7, 2011
Table 2.—School Lunch Menu Analysis at Three Time Points for Child and Adolescent Trial for Cardiovas¬
cular Health Intervention and Control Groups, 1991 Through 1994*
                                                                                                                                                      o   Intervention       ·   Control
      Value Measured                           Group                Baseline                  Interim                 Follow-up               60 -
Total energy content, MJ                           C                2.97(0.04)             3.04(0.04)                 3.12(0.04)
                                                   I                3.01 (0.04)            2.93 (0.04)                2.86 (0.04)             50
_P=.05_P<.001                                                                                                                         £UJ
                                                                                                                                     -JCO
Energy from total fat,       %                     C                38.9   (0.5)           36.2   (0.5)               36.2   (0.5)   •fi T. 40
                                                                    38.7   (0.4)           32.5   (0.4)               31.9   (0.4)                                       Moderate-to-Vlgorous
                                                                                             P<.001                    P-C.001
                                                                                                                                     ffs 30
Energy from      saturated fat, %                  C                15.1(0.3)              13.6(0.3)                  13.7(0.3)      •i=
                                                                                                                                      O 0)
                                                                    14.8(0.2)              12.1(0.2)                  12.0(0.2)      <
                                                                                              P=.02                     P=.007
                                                                                                                                     «a 20
Cholesterol content, mg                            C                80.3 (2.4)             75.2 (2.4)                 83.2   (2.4)   CO   <
                                                                    77.7 (2.0)             72.3 (2.0)                 74.9   (2.0)            10                                      Vigorous
                                                                                              P=.95                     P=.17
Sodium content,        mg/MJ                       C                386(7)                 415(7)                     421(7)
                                                                    377 (6)                401 (6)                    423 (6)                                    3          4
                                                                                              P=.64                     P=.34                                   Semester
Potassium content,          mg/MJ                  C                325    (5)             333    (5)                 327    (5)
                                                                    331    (4)             350    (4)                 357 (4)
                                                                                              P=.18                    P=.004        Moderate-to-vigorous and vigorous physical activity
                                                                                                                                     observed during Child and Adolescent Trial for Car¬
   *C indicates control; and I, intervention. Data for baseline, Interim, and follow-up are adjusted means (SE) from                 diovascular Health (CATCH) physical education
                                                                                                                                     classes at six time points, 1991 through 1994. The
repeated-measures analysis of variance, adjusted for site and school random effect. values compare C with I,
adjusting for baseline difference. The school + family intervention group did not differ from the school-only group                  CATCH intervention, introduced during semester 2,
for any end point (P>.20). 1 MJ=239 kcal. 1 mg/MJ=4.184 mg/1000 kcal.                                                                increased the percentage of time spent In moderate-
                                                                                                                                     to-vigorous and vigorous activity as measured by the
                                                                                                                                     System for Observing Fitness Instruction Time class¬
Table 3.—Observations of Physical Education Classes at Six Time Points for Child and Adolescent Trial for                            room observation system. Intervention and control
Cardiovascular Health Intervention and Control Groups, 1991 Through 1994*                                                            curves diverged significantly according to repeated-
                                                                                                                                     measures analysis of variance with the class session
                                                                                    Semester                                         as the unit of analysis: for moderate-to-vigorous ac¬

       Value Measured                 Baseline
                                                                                                                                     tivity, F=2.17, d/=5,1979, P=.02;forvigorous activity,
                                                                                                                                     F=2.35, c*=5,1979, P=.04. Analysis was controlled
Lesson      length,   min                                                                                                            for CATCH site, the location of the lesson, the spe¬
  C                                   29.9(0.8)        29.6(0.8)      29.0(0.8)     29.4(0.8)           29.3(0.8)      30.0(0.8)     cialty of the teacher, and random variation among
                                      28.9 (0.7)       30.0 (0.7)     28.4 (0.7)    29.8 (0.7)          29.5 (0.7)     30.1 (0.7)    schools and weeks of observation.
                                                          .22               .74        .21                   .27             .33
Energy expenditure, kJ/kg
  C                                    9.1              8.9               9.1
                                                                                                                                     significantly higher in the intervention
                                           (0.3)            (0.3)           (0.3)    9.2 (0.3)             9.5(0.3)      9.6(0.3)    schools (intervention mean=58.6 minutes
                                       9.2(0.2)        10.2(0.2)        9.8(0.2)    10.3(0.2)             10.4(0.2)    10.3(0.2)     vs control mean=46.5 minutes; P<.003).
                                                         .002               .08        .005                  .04             .08        Total blood cholesterol concentration
Energy expenditure rate,                                                                                                             declined in the intervention schools (from
  kJ/kg per h                                                                                                                        4.39 to 4.35 mmoUL [169.9 to 168.3
       C                              18.6(0.3)        18.3(0.3)      18.8(0.3)     19.1(0.3)           20.0(0.3)      19.6(0.3)
       I                              19.1(0.3)        20.7(0.3)      20.7(0.2)     21.0(0.2)           21.4(0.2)      20.8(0.2)
                                                                                                                                     mg/dL]) as well as in the control schools
                                                         <.001              .01        .01                   .14             .25
                                                                                                                                     (from 4.41 to 4.38 mmol/L [170.7 to 169.5
                                                                                                                                     mg/dL])(Table 6). Apolipoprotein fell
   *C Indicates control; and I, intervention. Data for baseline and semesters 2 through 6 are adjusted means (SE)                    among children in the intervention
from  repeated-measures analysis of variance, adjusted for site, class location, teacher's specialty, and random                     schools (from 2.33 to 2.31 mmol/L [90.2
effects of school and observation week. values compare C with I, adjusting for baseline difference. The school
+ family intervention group did not differ from the school-only group for any end point (P>.50). 1 kj=0.239 kcal. 1 kJ               to 89.4 mg/dL]) and the control schools
per h=0.00398 kcal per min.                                                                                                          (from 2.31 to 2.30 mmol/L [89.4 to 89.0
                                                                                                                                     mg/dL]). The differences were not sta¬
was        greater in the control schools (Table                     age of energy from protein and carbohy¬                         tistically significant.
5). Fat intake              was   significantly reduced              drates marginally increased in the inter¬                          Measures of body size, including
among children in intervention schools at                            vention schools compared with the                               height, weight, body mass index, and
follow-up (from 32.7% to 30.3% of energy                             controls (P=.07 and P=.06, respectively).                       skinfolds, did not differ between the in¬
comsumed) compared with children in                                  Sodium consumption, adjusted for energy                         tervention and control groups at base¬
control schools (from 32.6% to 32.2% of                              intake, also marginally increased among                         line or follow-up (Table 6). Growth was
energy) (P<.01). Much of this difference                             intervention students (P=.06). Intake of                        within the normal limits of the expected
came   from the decline in the intake of                             vitamins and other micronutrients re¬                           patterns for this age period. Similarly,
saturated fatty acids in the intervention                            mained at recommended levels in the in¬                         pulse rate and systolic and diastolic blood
schools (from 12.7% to 11.4% of energy)                              tervention group (data not shown).                              pressure were not significantly differ¬
but changing little in children in the con¬                              The SAPAC was administered at the                           ent among the groups.
trol schools (from 12.5% to 12.1%) (P<.01).                          fifth-grade follow-up only.47 Total min¬                           The 9-minute distance run by students
Polyunsaturated and monounsaturated                                  utes of reported daily physical activity,                       increased with age, in intervention
fat intake were also significantly reduced                           measured by this method, were not sig¬                          schools by 100 yards and in control
in the intervention schools compared with                            nificantly different between the inter¬                         schools by 84 yards. The difference be¬
the controls (P<.02). Dietary cholesterol                            vention and control schools (interven¬                          tween conditions was not significant.
was significantly reduced among children                             tion mean=145.5 minutes vs control
in the intervention group (from 223 mg to                            mean=154.8 minutes). Vigorous physi¬                            COMMENT
206 mg) compared with the controls (from                             cal activity during which the student                              The CATCH is the largest and most
218 mg to 225 mg) (P<.05). The percent-                              reported breathing hard, however, was                           rigorous school-based health promotion

                                                        Downloaded from jama.ama-assn.org by guest on May 7, 2011
Table 4.—Results of the Health Behavior Questionnaire at Four Time Points for Child and Adolescent Trial              eating patterns of students in the in¬
for Cardiovascular Health Intervention and Control Groups, 1991 Through 1994*                                         tervention schools, with intervention
                                                                                       Semester                       students significantly decreasing their
                                                                                                                      total fat, saturated fat, and cholesterol
              Scalef                      Group         Baseline
                                                                                                                      intakes. No intervention effect, however,
Dietary knowledge                                         4.0   (0.3)    5.5   (0.3)    6.9 (0.3)       7.6 (0.3)     was noted on the students' sodium in¬
  (±14; a=.78)                                            4.8   (0.2)    9.4   (0.2)   10.0(0.2)       10.0(0.2)      take. Results from the activity recalls
                                                                         <.001           <.001          <001          are concordant with the increases in
                                            s~            4.7 (0.3)      9.1 (0.3)       9.5(0.3)       9.7 (0.3)     MVPA in PE, with significantly higher
                                            SF~           4.8 (0.3)      9.7(0.3)      10.5(0.3)       10.3(0.3)      levels of self-reported vigorous activity
                                                           >.10            .05            .002            .06         reported by students in the interven¬
Dietary intention                           C             1.4(0.3)      -2.1   (0.3)    -1.2(0.3)      -0.8   (0.3)   tion schools. Finally, for blood pressure,
  (±13; a=.76)                                                           6.2             4.8 (0.2)                    no significant differences were found be¬
                                                          1.8(0.2)             (0.2)                    1.1   (0.2)
                                                                         <.001           <.001          <.001         tween the intervention and control stu¬
Usual food choice                                         0.8            0.7             0.0           -1.1 (0.3)
                                                                                                                      dents. Review of the secondary end
                                                              (0.3)          (0.3)           (0.3)
  (±14; a-76)                                             0.8 (0.2)      4.2 (0.2)       2.9 (0.2)      0.2   (0.2)   points, then, suggests significant and
                                                                         <.001
                                                                                                                      consistent changes in targeted psycho-
                                                                                         <.001          <.001
Food choice, social reinforcement
                                                                                                                      social factors, lower fat intake, and in¬
                                                         -1.0(0.4)       0.6   (0.4)     0.0   (0.4)   -2.0   (0.4)
  (±21;ct=.89)
                                                                                                                      creased physical activity among children
                                                        -0.4  (0.3)      6.8   (0.3)     S.7   (0.3)    3.3   (0.3)   in the intervention schools, but no de¬
                                                                         <.001           <.001          <.001
                                                                                                                      tectable changes in risk factors.
Food   choice, friend reinforcement                     -1.5(0.1)       -1.5(0.1)      -1.9(0.1)       -2.8(0.1)         The changes in the children's eating
  (±7; a=.83)                                                                          -0.4(0.1)
                                                        -1.3(0.1)       -0.0(0.1)                      -2.2(0.1)      patterns are notable. National goals for
                                                                         <.001           <.001            .04
                                                                                                                      consumption of fat and cholesterol of
Food choice, parent reinforcement
  (±7; ««,74)
                                                          0.8(0.1)       1.4(0.1)        1.2(0.1)      -0.9(0.1)      approximately 30% of energy intake and
                                                          1.0(0.1)       2.9(0.1)        2.8(0.1)       2.0(0.1)      200 mg, respectively, were met during
                                                                         <.001           <.001          <.001         a 3-year program.54·55 As in the school
Food choice, teacher reinforcement                      -0.2  (0.2)      0.7 (0.2)       0.8 (0.2)      0.0 (0.2)     lunch offerings, these changes were ac¬
  (±7; o=.87)                                             0.0(0.1)       3.9(0.1)        4.2(0.1)       3.4(0.1)      complished without loss of energy from
                                                                         <.001           <.001          <.001         protein or carbohydrates in the chil¬
Dietself-efficacy                                         5.5 (0.2)      6.2 (0.2)       6.3 (0.2)      5.8   (0.2)
                                                                                                                      dren's diet. Data on height and weight
  (±15;cc=.87)                                            5.9 (0.2)      7.6 (0.2)       6.7 (0.2)      5.6
                                                                                                                      also suggest that this recommended eat¬
                                                                                                              (0.2)
                                                                         <.001             .89            .09         ing pattern did not inhibit growth. Simi¬
                                                                                                                      lar results are shown in The Dietary
Physical activity, positive support                       5.2(0.1)       5.6(0.1)      -6.0(0.1)       -5.7(0.1)      Intervention Study in Children (DISC),56
  (±11;a=67)                                              5.3(0.1)       6.4(0.1)       6.4(0.1)        5.7(0.1)      which focused on lowering cholesterol
                                                                         <.001             .11            .72
                                                                                                                      in youth with elevated blood cholesterol
Physical activity, negative support                       3.1 (0.1)      3.9(0.1)      -4.3(0.1)       -4.7(0.1)      levels in intensive one-on-one interven¬
  (±7; cc=.56)                                            3.5(0.1)       4.3(0.1)       4.7(0.1)        4.9(0.1)      tions outside the school setting.
                                                                           .59             .81            .43            The lack of changes in cholesterol lev¬
Physical activity, self-efficacy            C             2.1 (0.1)      2.5(0.1)        2.7(0.1)       2.8(0.1)      els and other physiological risk factors,
  (±5; oc=.69)                               "            2.2(0.1)       2.9(0.1)        3.0(0.1)       2.7(0.1)      although disappointing, is not surpris¬
                                            P~                            .006             .04            .07         ing, given several other observations.
                                                                                                                      The magnitude of the changes in food
  *C indicates control; I, intervention; S, school Intervention alone; and SF, school + family intervention.
  tWith numerical range and Cronbach reliability coefficient a. Scales are based on a series of questions, scored     intake and physical activity across the
+1 for correct or healthy response, -1 for incorrect or less healthy response. Data for baseline and semesters 2,
4, and 6 are adjusted means (SE) from repeated-measures analysis of variance on school-aggregated data, adjusted
                                                                                                                      population, while statistically significant,
for site and school random effect. values compare C with I (all scales) or S with SF (dietary knowledge only),        was modest. The observed changes in
adjusting for   baseline difference.                                                                                  dietary fat and cholesterol intakes, when
                                                                                                                      analyzed by the Keys equation, which
 field trial that has been implemented to                 from saturated fats. Positive results                       relates fat intake to blood cholesterol,
 date. The CATCH maintained the in¬                       were also observed for CATCH PE; the                        predict changes of a magnitude similar
 volvement of 96 public elementary                        percentage of PE class time devoted to                      to those observed herein.57 This is also
 schools, with considerable ethnic and                    MVPA significantly increased in the in¬                     a period of important growth and de¬
 geographic diversity, using a multicom-                  tervention schools compared with the                        velopment for these youths, with the
 ponent intervention. The results of the                  control schools, surpassing the study and                   onset of puberty occurring for many dur¬
 trial should be considered in terms of its               Year 2000 goals of 50% of class time.11                     ing the study.58 The effect of diet and
effectiveness for primary prevention and                  At the individual level, however, the                       exercise changes such as those described
its potential as a model for school-based                 decreases in serum cholesterol levels                       herein may be obscured by these more
health promotion.                                         among students in the intervention                          substantial developmental changes. As
    The CATCH was successful in meet¬                     schools compared with control students,                     suggested in the recently released DISC
ing two of the three primary outcomes                     while in the hypothesized direction, were                   study,56 it may be more important for
of the study. At the school level, caf¬                    not significant.                                           this age group to demonstrate the abil¬
eterias in the intervention schools were                        Postulated
                                                                         changes in secondary end                     ity of the program to modify nutrition
able to significantly modify their lunch                   points were also found as a result of                      and physical activity behaviors in ways
offerings to approach the national rec¬                   CATCH. Seven of the 11 scales of the                        leading to lifetime health habits rather
ommendations of 30% total fat energy                      HBQ were significantly higher for the                       than to reduce immediate physiological
intake and 10% saturated fat energy in¬                   intervention group at follow-up. The food                   risk levels. Such modest modifications
take, with much of the reduction coming                   recall data revealed changes in the daily                   in health habits throughout the popu-


                                       Downloaded from jama.ama-assn.org by guest on May 7, 2011
Table 5.—Total Daily Intakes (24-Hour Recall) at Baseline and        Follow-up for Child and Adolescent Trial for       ables than the negligible rate of change
Cardiovascular Health Intervention and Control Groups*                                                                  that had been anticipated from previous
      Value Measured                 Group          Baseline         Follow-up            Changef
                                                                                                                        studies and national surveys. These find¬
Total energy intake, MJ                             1.49(0.13)       9.08(0.15)         0.60(0.15)            .01
                                                                                                                        ings reduced the magnitude of differ¬
                                                                                                                        ences attributable to the program. This
                                                   8.55(0.10)        8.68(0.12)         0.14(0.12)
                                                                                                                        observation has been noted in several
Energy intake,    %
  Total fat                                        32.6 (0.3)        32.2   (0.3)       -0.5  (0.4)           .001
                                                                                                                        prominent population-based trials.60 A
                                                   32.7 (0.3)        30.3   (0.3)       -2.3  (0.3)                     willingness to participate in a random¬
  Saturated fat                                    12.5(0.1)          12.1 (0.2)        -0.4 (0.2)                      ized trial of school health promotion pro¬
                                                                                                              .005
                                                   12.7(0.1)          11.4(0.1)          -1.1 (0.1)                     grams may indicate a greater interest in
                      fat
                                                                                                                        this area and a desire to make changes.
  Polyunsaturated                                   5.9(0.1)            6.1 (0.1)         0.3(0.1)
                                                    5.7(0.1)            5.6(0.1)        -0.1 (0.1)                      Despite controlled publicity and other
  Monounsaturated fat                              11.9(0.1)          11.6(0.1)         -0.3 (0.2)
                                                                                                                        safeguards instituted by the CATCH,
                                                                                                              .02       some contamination by centralized food
                                                   11.9(0.1)          11.1 (0.1)        -0.8(0.1)
                                                  53.7 (0.4)          54.7 (0.4)
                                                                                                                        services, PE staff, and transfers of staff
  Carbohydrate                                                                            1.0(0.5)            .06       within the districts was possible (even
                                                  53.9 (0.3)          56.1 (0.3)          2.2 (0.4)
                                                                                                                        though not detected by CATCH staff),
  Protein                                          14.9(0.2)          14.3(0.2)         -0.5 (0.2)                      since the study involved existing infra¬
                                                   14.6(0.1)          14.8(0.1)           0.1 (0.2)                     structures. Some of the changes noted
Cholesterol intake, mg                             218(7)              225 (8)              5(8)                        in the control schools over time, along
                                                   223 (6)             206 (6)           -15(6)                         with other public health efforts, could
Sodium intake, mg                                 3042 (62)          3168(65)             184(69)                       reflect these observations.
                                                                                                              .61
                                                  2929(46)           3107(48)             138(57)                          The CATCH study had several limi¬
Sodium   intake, mg/MJ                             362 (5)             354 (5)             -2(5)                        tations. The first was the statistical power
                                                                                                              .06       of the study.25 The targeted end point for
                                       I           347 (4)             364 (4)             12(4)
                                                                                                                        power calculations was a differential be¬
   *Restricted to cohort substudy students with paired data. C indicates control (40 schools, 473 students); and I,     tween conditions of 0.13 mmol/L (5.1
intervention (56 schools, 709 students).
   tData for baseline and follow-up are unadjusted means (SE). Change estimate (SE) is the excess of the group's        mg/dL) of cholesterol, which may not have
follow-up mean over the pooled baseline mean, after adjustment by analysis of variance for baseline value, site, sex,   been attainable with a population inter¬
race, and school random effect. Pvalue compares C with I. The school + family intervention group did not differ from
the school-only group for any end point (P>.30). 1 MJ=239 kcal. 1 mg/MJ=4.184 mg/1000 kcal.                             vention, or even with children at high
                                                                                                                        risk.56 This could not be determined at
lation may produce considerable public                       gram may also need to be more exten¬                       the outset of the CATCH, and it limited
health effects. However, the impact of                       sive or more intensive in implementa¬                      our ability to detect any physiological ef¬
such school-based interventions on sub¬                      tion than was possible in the CATCH. It                    fects. The second limitation was partici¬
sequent CVD incidence is uncertain and                       is likely that to realize more substantial                 pation rates. At baseline 60.4% of the
can only be determined by long-term                          modifications in health habits and                         third-grade populations in the 96 schools
follow-up studies.                                           changes in physiological risk factors,                     agreed to participate and provided ad¬
    The sodium in school lunches and re¬                     family involvement will be necessary.                      equate blood samples. Because partici¬
ported sodium consumption were not                              The CATCH demonstrated that a                           pation required written (active) consent
changed as a result of the CATCH in¬                         school-based program involving school                      by the child and parent, moderate rates
tervention. Further analyses will reveal                     food service, PE, classroom curricula,                     of participation could be expected, espe¬
whether foods that are higher in fat were                    and family programs can be successfully                    cially since the evaluation involved blood
replaced by those higher in sodium, es¬                      implemented in diverse populations in                      samples from the children. Multiple meth¬
pecially in the school lunches.                              four areas of the country. There were no                   ods to improve participation were imple¬
    The secondary comparison in the                          significant inconsistencies by site, sex,                  mented, including several notices and tele¬
study between the school-based pro¬                          or ethnic group with respect to the pri¬                   phone calls to parents and individual and
grams and the school-plus-family pro¬                        mary and secondary end points, which                       classroom rewards for participation. For¬
grams revealed only a modest benefit in                      suggests that the results reported were                    tunately, "there were no differences by
dietary knowledge by adding the family                       found across all of the schools and stu¬                   treatment condition, site, sex, or ethnic
component. Previous reports have in¬                         dents who participated in the CATCH                        group in participation at baseline. The
dicated that family involvement in health                    intervention program. The CATCH pro¬                       third limitation was the amount of inter¬
promotion programs is difficult due to                       grams, involving a total of about 20 ex¬                   vention that was feasible in public elemen¬
low levels of parent participation.59 The                    tra hours per year of class time, were                     tary schools. The classroom curricula in¬
CATCH was able to obtain some par¬                           implemented with a high degree of com¬                     volved classroom time, food preparation,
ticipation by more than 70% of the par¬                      pliance by food service staff, PE spe¬                     and skill development activities that were
ents during each of the 3 years of the                       cialists, and regular classroom teachers,                  limited by costs, staff time, and compet¬
intervention. This participation, how¬                       based on direct observations by CATCH                      ing classroom instructional requirements.
 ever, was limited to working with their                     evaluation staff. This was achieved with                   The students had 3-month summer va¬
 child on at least one of five to eight                      carefully prepared protocols and 1 to 2                    cations, during which they were not
 activity packets each year and to at¬                       days of training each year. These find¬                    reached and recidivism could occur. More
 tending family fun nights.35 An earlier                     ings suggest that CATCH interventions                      intensive family programs were not pos¬
 study using home curricula with third-                      were compatible with the needs and                         sible, as the home curricula were already
 grade children demonstrated significant                     structures of schools and school staff                     seen as unique. Staff training time was
 short-term dietary improvement among                        and could be implemented successfully                      limited by the school districts and bud¬
 children whose families participated, but                   across diverse schools and students.                       gets for substitutes, and 1 to 2 days per
 the implementation methods were more                           The control schools were observed to                    year were found to be appropriate for the
 intensive than in the CATCH.16 These                        change at a greater positive rate with                     tasks asked of the school staff and yielded
 findings suggest that a family-based pro-                   respect to diet and physical activity vari-                high attendance rates across sites. Fi-

                                                Downloaded from jama.ama-assn.org by guest on May 7, 2011
Table 6.—Selected Risk Factor Variables at Baseline and Follow-up for Child and Adolescent Trial for Cardiovascular Health Intervention and Control                       Groups*
       Risk Factor Variable                     Group                Paired                 Baseline                    Follow-up                   Changet
Height,    m                                                          1648                 1.325(0.001)               1.463(0.002)                0.139(0.001)
                                                                                           1.327(0.001)               1.465(0.002)                0.139(0.001)
Body   mass    index, kg/m2                                           1627                 17.68(0.08)                19.74(0.10)                  2.06 (0.05)
                                                                      2332                 17.58(0.06)                19.66(0.09)                  2.13(0.04)
Triceps skinfold,   mm                                                1649
                                                                      2365
                                                                                             12.5(0.1)
                                                                                             12.4(0.1)
                                                                                                                        15.3(0.2)
                                                                                                                        15.2(0.1)
                                                                                                                                                    2.8(0
                                                                                                                                                           2) J
                                                                                                                                                    2.9 (0.2)
                                                                                                                                                           -
                                                                                                                                                                 1
Subscapular skinfold,    mm                                           1649                     1.4(0.1)                 11.2(0.2)                   2.8(0.2)
                                                                      2364                    8.2(0.1)                  11.0(0.2)                   2.9(0.1)
Heart rate, beats per min                                             1648                   86.9 (0.2)                 82.9 (0.3)                 -4.1 (0.4)
                                                                      2363                   87.0 (0.2)                 82.9 (0.2)                 -4.1 (0.3)
Systolic   blood pressure,    mm   Hg                                 1647                 104.8(0.2)                 109.8(0.2)                    4.8 (0.2)
                                                                      2363                 105.2(0.2)                 110.0(0.2)                    4.8 (0.2)
Diastolic blood pressure,     mm   Hg                                 1647                   53.5 (0.2)                 55.6 (0.2)                  2.1 (0.2)
                                                                      2363                   53.5(0.1)                  56.0(0.1)                   2.4 (0.2)
Total cholesterol, mmol/L     [mg/dL]                                 1625                   4.41 (0.02)                4.38 (0.02)               -0.02 (0.02)
                                                                                          [170.7(0.8)]               [169.5 (0.8)]                [-0.8 (0.8)]
                                                                      2311                   4.39(0.01)                 4.36 (0.01)               -0.03 (0.02)
                                                                                          [169.9 (0.4)]              [168.7 (0.4)]                [-1.2(0.8)]
HDL cholesterol, mmol/L       [mg/dL]                                  707                   1.33(0.01)                 1.27 (0.01)               -0.06(0.01)
                                                                                            [51.5(0.4)]                [49.1 (0.4)]               [-2.3 (0.4)]
                                                                      1035                   1.34(0.01)                 1.27 (0.01)               -0.07(0.01)
                                                                                            [51.9(0.4)]                [49.1 (0.4)]               [-2.7 (0.4)]
Apolipoprotein B, mmol/L [mg/dL]                                       706                   2.31 (0.02)                2.30 (0.02)               -0.01 (0.02)
                                                                                            [89.4 (0.8)]               [89.0 (0.8)]               [-0.4 (0.8)]
                                                                      1033                   2.33 (0.02)                2.31 (0.02)               -0.03(0.01)
                                                                                            [90.2 (0.8)]               [89.4 (0.8)]               [-1.2(0.4)]
   *Restricted to cohort students with paired data. C indicates control (40 schools, 1653 students); I, intervention (56 schools, 2366 students); and HDL, high-density lipoprotein.
   tData for baseline and follow-up are unadjusted means (SE). Change estimate (SE) is the excess of the group's follow-up mean over the pooled baseline mean, after adjustment
for baseline value, site, sex, race, and school random effect. Height, body mass index, and triceps and subscapular skinfolds were additionally adjusted for age; heart rate,
blood pressure, and lipids were additionally adjusted for height and body mass index. The value compares C with I. The school + family intervention group did not differ from
the school-only group for any end point (P>.10).

nally, the CATCH intervention did not                        ported in this publication included the following:
                                                                University of California, San Diego, Depart-
                                                                                                                           Edmundson, PhD (coinvestigator); Jack Wilmore,
                                                                                                                           PhD (coinvestigator).
encompass community members, mass                            ment of Pediatrics, Community Pediatrics Divi-                   Tulane University, School of Public Health and
media, grocery stores, parks and recre¬                      sion, La Jolla, Calif.\p=m-\PhilipNader, MD (principal        Tropical Medicine, New Orleans, La.\p=m-\Larry
ation facilities, or other resources that                    investigator); John Elder, PhD (co-principal inves-           Webber, PhD (principal investigator); Carolyn
might be important in the maintenance of                     tigator); Thom McKenzie, PhD (coinvestigator);
                                                             Kathryn Bachman, MS, EdS (project director);
                                                                                                                           Johnson, PhD (project director); Theresa Nicklas,
                                                                                                                           DrPH, MPH, LDN (investigator); Vincent An-
children's healthy behaviors.20
   The results of the CATCH interven¬
                                                             Shelia Broyles, PhD; Ellen Busch, RD; Steve                   thony; Nancy Baker, MS, RD; Kathryn Barnwell;
                                                             Danna; Todd Galati, MA; Karen Haye; Christine                 Stacy Belou; Gerald Berenson, MD; Susan Bonura,
tion lead to several interpretations and                     Hayes, MA; Michael McGreevy; B. J. Williston,                 MPA; Keith Bordelon; Sandra Cameron, RN; Ann
recommendations. The CATCH inter¬                            MEd; Michelle Zive, MS, RD.                                   Clesi, MEd; Lisa Crochet, MEd; Amy Cunningham,
vention demonstrates that the policies                          University of Minnesota, Division of Epidemi-              MS, RD; Didara Franklin; Anwarul Haque, MBBS;
                                                             ology, School of Public Health, Minneapolis,                  David Harsha, PhD; Javed Joy, MPH; Saundra M.
and practices of schools can be changed                      Minn.\p=m-\CherylPerry, PhD (principal investiga-             Hunter, PhD; Daniel Kuras; Pam Lambie; Ann
without substantial new school resources                     tor); Leslie Lytle, RD, PhD (co-principal investi-            Layman, MA; Sharon Little-Christian, MEd;
and time. The changes observed posi¬                         gator); Russell Luepker, MD (co-principal investi-            Sheryl Pedersen, MPH; Jacqueline Reeds-Epping,
tively affected health behaviors and met                     gator); Beth Davidann, MPH (project coordinator);             MEd; Rochelle Rice, MPH; Kelly Romero, MEd;
national recommendations.11·54·55·61 These
                                                             Pat Brothen; Virginia Dahlstrom, MA; Maxine                   Cynthia Pitcher-Smith; Patricia Strikmiller, MS;
                                                             Dammen; Sally Ehlinger, PhD; Tawny Greene;                    Marion White, MS, RD.
changes, when spread across the entire                       Barbara Hann; Jean Heberle; Tom Hofflander,                      Coordinating Center, New England Research
 school-based population, have the po¬                       MEd; Colleen Kelder; Pat Kelliher; Therese Kunz;              Institutes, Watertown, Mass.\p=m-\SonjaMcKinlay,
tential to produce long-term cardiovas¬                      Bonnie Manning; Donna McDuffie; Todd Morrow;                  PhD (principal investigator); Stavroula Osganian,
 cular health benefits. Moreover, these                      Margie Miller; Jennifer Mrosala; Gretchen New-                MD, MPH (co-principal investigator and project
                                                             man, MA; Mellanie Pusateri; Margaret Reinhardt,               director); Henry Feldman, PhD (senior statisti-
 school-based programs were introduced                       RD, MPH; Renee Sieving, MSN; Jennifer Smisson;                cian); Herman Mitchell, PhD (co-principal investi-
 with minimal but concentrated and ef¬                       Mary Smyth; Pat Snyder, MA, RD; Mike Stau-                    gator); Susan Budman, MSW (field coordinator);
 fective training of existing school per¬                    facker, MA; Janelle Traut; Tracy Wick.                        Patty Connell; Michael Koehler; Paul Mitchell, MS;
 sonnel and modest follow-up support                            University of Texas, Centerfor Health Promotion            Christine Kannler; Glenn Rennie; Debbie Sellers,
                                                             Research and Development, University of Texas                 PhD; Moira Walsh; Minhua Yang, MS.
from CATCH intervention staff. The                           Health Science Center, Houston.\p=m-\GuyParcel, PhD             Frances Stern Nutrition Center/New England
CATCH, therefore, provides an impor¬                         (principal investigator); Steve Kelder, PhD, MPH              Medical Center Hospital, Boston, Mass.\p=m-\Johanna
tant model of a school-based health pro¬                     (co-principal investigator); Deanna Montgomery,               Dwyer, DSc, RD (co-principal investigator and di-
motion program for the primary pre¬                          PhD, RD (coinvestigator); Milton Nichaman, MD,                rector); Mary Kay Ebzery, MS, RD; Anne Garceau,
                                                             ScD (coinvestigator); Wendell Taylor, PhD, MPH                MS, RD; Lynn Hewes, MS, RD; Caitlin Hosmer,
vention of CVD that should be feasible                       (coinvestigator); Kathleen (Wambsgans) Cook, MEd              MS, RD; Debra Raizman, MPH, RD.
and effective for America's schools.                         (project director); Elvira Barrera, SpMFT, LPC;                  The Miriam Hospital Lipoprotein Analysis
                                                             Larry Berry; Janet Carbonneau, MEd, RD; Karen                 Laboratory, Providence, RI.\p=m-\LindaBausserman,
  The research reported in this publication was              Chemycz, MA; Peter Cribb, MEd; Sandra Evans,                  PhD (laboratory director).
supported by the National Heart, Lung, and Blood             MPH; Rebecca Gordon; Jenifer Gwinn; Susan Luton,                Project Office, National Heart, Lung, and Blood
Institute, Bethesda, Md (U01-HL-39880, U01-HL\x=req-\        MA; Barbara Scaife, MS, RD; Sue Sharkey; Sharon               Institute, Bethesda, Md.\p=m-\ElaineStone, PhD,
39906, U01-HL-39852, U01-HL-39927, and U01-HL\x=req-\        Snider, MPH; Sema Spigner; Kay Wilson, MPH, RD;               MPH (project administrator); Marguerite Evans,
39870)                                                       Sherry Woods, EdD.                                            MS, RD (deputy administrator); Margaret Wu,
   Participants in the CATCH Collaborative Re-                  University of Texas, Department of Kinesiol-               PhD; Jeffrey Cutler, MD, MPH.
search  Group and contributors to the work re-                ogy and Health     Education, Austin.\p=m-\Elizabeth           Data and Safety Monitoring Committee.\p=m-\



                                         Downloaded from jama.ama-assn.org by guest on May 7, 2011
Ronald Lauer, MD (chairperson); Thomas Coates,               and the Class of 1989Study. Am J Public Health.              of food coding for nutrient intake studies. J Am
PhD; William Haskell, PhD; C. Anderson Johnson,              1992;82:1210-1216.                                           Diet Assoc. 1988;88:1263-1267.
PhD; Ronald Prineas, MD, PhD; Linda Van Horn,                21. Kelder SH, Perry CL, Lytle L, Klepp K-I. Com-            42. McKenzie TL, Nader PR. SOFIT: system for
PhD, RD; Joel Verter, PhD.                                   munity-wide youth nutrition education: long-term             observing fitness instruction time. J Teaching Phys
                                                             outcomes of the Minnesota Heart Health Program.              Educ. 1991;11:195-205.
References                                                   Health Educ Res. 1995;10:119-131.                            43. Parcel G, Edmundson E, Perry C, et al. Mea-
1. Tunstall-Pedoe H, Kuulasmaa K, Amouyel P,                 22. Perry CL, Stone EJ, Parcel GS, et al. School\x=req-\     surement of self-efficacy for diet-related behaviors
Arveiler D, Rajakangas AM, Pajak A. Myocardial               based cardiovascular health promotion: the Child             of elementary school children. J Sch Health. 1995;
infarction and coronary deaths in the World Health           and Adolescent Trial for Cardiovascular Health               65:23-27.
Organization MONICA Project. Circulation. 1994;              (CATCH). J Sch Health. 1990;60:406-413.                      44. Edmundson E, Parcel GS, Perry CL, et al. Third
90:583-612.                                                  23. Lytle LA, Johnson CC, Bachman K, et al. Re-              grade intervention effects on psychosocial deter-
2. National Heart, Lung, and Blood Institute Re-             cruitment strategies and issues for school-based             minants of CVD risk behavior. J Health Promo-
port of the Task Force on Research in Epidemiol-             health promotion: experiences from CATCH. J Sch              tion. In press.
ogy and Prevention of Cardiovascular Diseases.               Health. 1994;64:405-409.                                     45. Lytle LA, Nichaman MZ, Obarzanek E, et al,
Washington, DC: US Dept of Health and Human                  24. Webber LS, Osganian V, Luepker RV, et al, for            for the CATCH Collaborative Group. Validation of
Services, National Institutes of Health, National            the CATCH Study Group. Cardiovascular risk fac-              24-hour recalls assisted by food records in third\x=req-\
Heart, Lung, and Blood Institute; 1994.                      tors among third grade children in four regions of           grade children. J Am Diet Assoc. 1993;93:1431-1436.
3. Allan R. Is coronary heart disease a lifestyle            the United States: the CATCH study. Am J Epi-                46. Smiciklas-Wright HA, Mitchell DC, Norton LD,
disorder? a review of psychologic and behavioral             demiol. 1995;141:428-439.                                    Derr JA. Interviewer reliability of nutrient intake
factors, I. Cardiovasc Rev Rep. 1992;13:13-53 (non\x=req-\   25. Zucker DM, Lakatos E, Webber LS, et al. Sta-             data from 24-hour recalls collected using the Min-
sequential).                                                 tistical design of the Child and Adolescent Trial for        nesota Nutrition Data System (NDS). J Am Diet
4. Webber LS, Srinivasan SR, Wattingney WA, Be-              Cardiovascular Health (CATCH): implications ofclus-          Assoc. 1991;91:28A. Abstract.
renson GS. Tracking of serum lipids and lipoproteins         ter randomization. Control Clin Trials. 1995;16:96\x=req-\   47. Sallis J, Strikmiller P, Harsha D, et al. Valida-
from childhood to adulthood: the Bogalusa Heart              118.                                                         tion of interviewer- and self-administered physical
Study. Am J Epidemiol. 1991;133:884-899.                     26. Perry CL, Parcel GS, Stone EJ, et al. The Child          activity checklists for fifth grade students. Med Sci
5. Clarke WR, Schrott HG, Leaverton PE, Connor               and Adolescent Trial for Cardiovascular Health               Sports Exerc. In press.
WE, Lauer RM. Tracking of blood lipids and blood             (CATCH): overview of the intervention program                48. Ainsworth BE, Haskell WL, Leon AS, et al.
pressures in school age children: the Muscatine              and evaluation methods. Cardiovasc Risk Factors.             Compendium of physical activities: classification of
Study. Circulation. 1978;58:626-634.                         1992;2:36-44.                                                energy costs of human physical activities. Med Sci
6. Kelder SH, Perry CL, Klepp K-I. Longitudinal              27. Nicklas TA, Stone E, Montgomery D, et al. An             Sports Exerc. 1993;25:71-80.
tracking of adolescent smoking, physical activity,           overview of the CATCH Eat Smart school nutri-                49. Hainline A, Karon JM, Winn CL, Gill JB. Ac-
and food choice behaviors. Am J Public Health.               tion program: a model for the future. J Health               curacy and comparability of long-term measure-
1994;84:1121-1126.                                           Educ. 1994;25:299-307.                                       ments of cholesterol. Clin Chem. 1986;32:611-615.
7. Newman WP III, Freedman DS, Voors AW, et                  28. Osganian SK, Ebzery MK, Montgomery DH.                   50. Turley KR, Wilmore JH, Simons-Morton B, Wil-
al. Relation of serum lipoprotein levels and systolic        Changes in the nutrient content of school lunches:           liston BJ, Epping JR, Dahlstrom G. The reliability
blood pressure to early atherosclerosis: the Boga-           results from the CATCH Eat Smart food service                and validity of the 9-minute run in third grade
lusa Heart Study. N Engl J Med. 1986;314:138-144.            intervention. Prev Med. In press.                            children. Pediatr Exerc Sci. 1994;6:178-187.
8. Kannel WE, Dawber TR. Atherosclerosis as a                29. McKenzie TL, Feldman H, Woods SE, et al.                 51. Koepsell TD, Martin DC, Diehr PH, et al. Data
pediatric problem. J Pediatr. 1972;80:544-554.               Children's activity levels and lesson context during         analysis and sample size issues in evaluations of
9. Lenfant C. Improving the health of America's              third grade physical education. Res Q Exerc Sport.           community based health promotion and disease pre-
youth: the NHLBI perspective. J Health Educ.                 1995;66:184-193.                                             vention programs: a mixed model analysis of vari-
1995;26:6-8.                                                 30. McKenzie TL, Strikmiller PK, Stone EJ, et al.            ance approach. J Clin Epidemiol. 1991;44:701-713.
10. Stone EJ, Perry CL, Luepker RV. Synthesis of             CATCH: physical activity process evaluation in a             52. SAS/STAT Software: Changes and Enhance-
cardiovascular behavior research for youth health            multicenter trial. Health Educ Q. 1994;21(suppl 2):          ments, Release 6.07. Cary, NC: SAS Institute Inc;
promotion. Health Educ Q. 1989;16:155-169.                   S73-S89.                                                      1992:289-368. SAS technical report P-229.
11. US Public Health Service. Healthy People 2000:           31.Siedentop D. Developing Teaching Skills in                53. Murray DM, Wolfinger RD. Analysis issues in
National Health Promotion and Disease Preven-                Physical Education. Mountain View, Calif: May-               the evaluation of community trials: progress to-
tion Objectives. Washington, DC: US Dept of Health           field; 1991.                                                 ward solutions in SAS/STAT MIXED. J Commu-
and Human Services; 1990. DHHS publication (PHS)             32. Perry CL, Mullis R, Maile M. Modifying eating            nity Psychol. 1994:140-154.
91-50212.                                                    behavior of children: a pilot intervention study.            54. Food and Nutrition Board. Recommended
12. Lloyd DM, Alexander HM, Callcott R, et al.               J Sch Health. 1985;55:399-402.                               Dietary Allowances. 10th ed. Washington, DC:
Cigarette smoking and drug use in schoolchildren,            33. Parcel GS, Simons-Morton BG, O'Hara NM,                  National Academy of Sciences; 1989.
III: evaluation of a smoking prevention education            Baranowski T, Wilson BS. School promotion of diet            55. National Cholesterol Education Program. Re-
programme. Int J Epidemiol. 1983;12:51-58.                   and physical activity: impact on learning outcomes           port of the Expert Panel on Blood Cholesterol Lev-
13. Evans RI. Smoking in children: developing a              and self-reported behavior. Health Educ Q. 1989;             els in Children and Adolescents. Bethesda, Md: US
social psychological strategy of deterrence. Prev            16:181-199.                                                  Dept of Health and Human Services, Public Health
Med. 1976;5:122-127.                                         34. Edmundson EW, Luton SC, McGraw SA, et al.                Service, National Institutes of Health, National
14. Luepker RV, Johnson CA, Murray DM,                       CATCH: classroom process evaluation in a multi-              Heart, Lung, and Blood Institute; 1991. NIH pub-
Pechacek TF. Prevention of cigarette smoking:                center trial. Health Educ Q. 1994;21(suppl 2):S27\x=req-\    lication 91-2732.
three-year follow-up of an education program for             S50.                                                         56. The Writing Group for the DISC Collaborative
youth. J Behav Med. 1983;6:53-62.                            35. Johnson CC, Osganian SK, Budman SB, et al.                Research Group. Efficacy and safety of lowering
15. Stone EJ, Baranowski T, Sallis JF, Cutler JA.            CATCH: family process evaluation in a multicenter            dietary intake of fat and cholesterol in children with
Review ofbehavioral research for cardiopulmonary             trial. Health Educ Q. 1994;21(suppl 2):S91-S106.             elevated low-density lipoprotein cholesterol: the
health: emphasis on youth, gender, and ethnicity.            36. Stone EJ, McGraw SA, Osganian SK, Elder JP,              Dietary Intervention Study in Children (DISC).
J Health Educ. March/April 1995;26(suppl):S9-S17.            eds. Process evaluation in the multicenter Child             JAMA. 1995;273:1429-1435.
16. Perry C, Luepker R, Murray D, et al. Parent              and Adolescent Trial for Cardiovascular Health               57. Keys A, Anderson JT, Grande F. Serum
involvement with children's health promotion: the            (CATCH). Health Educ Q. 1994;21(suppl 2, theme               cholesterol response to changes in the diet, IV:
Minnesota Home Team. Am J Public Health. 1988;               issue):1-143.                                                particular saturated fatty acids in the diet. Me-
79:1156-1160.                                                37. Raizman DJ, Montgomery DH, Osganian V, et                 tabolism. 1965;14:776-787.
17. Nader PR, Sallis JF, Broyles S, et al. Family\x=req-\    al. CATCH: food service program process evalua-              58. Berenson GS, Srinivasan SR, Cresanta JL, Fos-
based cardiovascular risk reduction among Mexican\x=req-\    tion in a multicenter trial. Health Educ Q. 1994;21          ter TA, Webber LS. Dynamic changes of serum li-
Americans and Euro-Americans. Fam Community                  (suppl 2):S51-S71.                                           poproteins in children during adolescence and sexual
 Health. 1992;15:57-74.                                      38. McGraw SA, Stone EJ, Osganian SK, et al.                  maturation. Am J Epidemiol. 1981;113:157-170.
 18. Flynn BS, Worden JK, Secker-Walker RH, Bad-             Design of process evaluation within the Child and             59. Crockett SJ, Mullis RM, Perry CL. Parent nu-
 ger GJ, Geller BM, Costanza MC. Prevention of               Adolescent Trial for Cardiovascular Health                   trition education: a conceptual model. J Sch Health.
 cigarette smoking through mass media interven-              (CATCH). Health Educ Q. 1994;21(suppl 2):S5-S26.             1988;58:53-57.
 tion and school programs. Am J Public Health.               39. Elder JP, McGraw SA, Stone EJ, et al. CATCH:             60. Susser M. The tribulation of trials\p=m-\interven-
1992;82:827-834.                                             process evaluation of environmental factors and              tion in communities. Am J Public Health. 1995;85:
19. Walter HJ, Vaugham RD, Wynder EL. Pri-                   programs. Health Educ Q. 1994;(suppl 2):S107-S127.           156-158.
mary prevention of cancer among children: changes            40. Ebzery MK, Montgomery DH, Evans MA, et                   61. Council on Physical Education for Children,
in cigarette smoking and diet after six years of             al. School meal data collection and documentation            National Association for Sport and Physical Edu-
intervention. J Natl Cancer Inst. 1989;81:995-999.           methods in a multisite study. Sch Food Serv Res              cation. Guidelines for Elementary School Physical
20. Perry CL, Kelder SH, Murray DM, Klepp K-I.               Rev. In press.                                               Education. Reston, Va: National Association for
Community-wide smoking prevention: long-term                 41. Feskanich D, Buzzard IM, Welch BT, et al.                Sport and Physical Education; 1994.
outcomes of the Minnesota Heart Health Program               Comparison of a computerized and a manual method

                                                  Downloaded from jama.ama-assn.org by guest on May 7, 2011

				
DOCUMENT INFO
hkksew3563rd hkksew3563rd http://
About