Sex differences in the association of youth body mass index to adult health-related quality of life: the physical activity longitudinal study.

Herman, Katya M. ; Hopman, Wilma M. ; Craig, Cora L. 等

Obesity leads to increased risk of numerous chronic diseases and conditions, including early mortality; (1) however, almost 60% of Canadian adults and over 1 in 4 young people aged 2-18 years are overweight/obese. (2)

Health-related quality of life (HRQL) refers to perceived physical and mental health over time. (3) Research examining the links between obesity and HRQL in healthy general population samples has only recently begun to evolve. In adults, obesity is often associated with significant impairments in HRQL, with greater impairments associated with greater degrees of obesity, and weight loss leading to improved HRQL. (4) Yet, some studies have shown a positive association between mental HRQL and weight status. (5-7) More limited research exists linking youth obesity to poorer youth HRQL, including effects on self-reported health, self-esteem, and physical and social health and functioning. (8-10)

The long-term effects of youth obesity on adult HRQL are unclear; however, over 90% of pediatricians believe that youth overweight could affect future HRQL. (11) In a previous investigation of Canadian youth followed up 2 decades later, we reported that overweight youth had greater odds of scoring at or above the norm on several aspects of HRQL in adulthood, particularly the mental aspects, compared to healthy weight youth. (12) To our knowledge, the study, using dichotomized HRQL outcomes based on Canadian normative values, was the first to explore the potential long-term effects of youth body mass index (BMI) on HRQL in adulthood. The current objectives are to explore the relationship between youth BMI and adult physical and mental HRQL using continuous HRQL outcomes, examining differences by sex.

METHODS

Sample

Data were from the Physical Activity Longitudinal Study (PALS), (13) the 22-year follow-up of a 20% subsample of participants ages [greater than or equal to] 7 from the nationally representative 1981 Canada Fitness Survey (CFS; N=23,397). (14) Detailed information regarding the conceptual design, methods and ethics approval of PALS has previously been published. (13) The final sample included 2,511 participants (51.2% of eligible individuals); the subsample for this analysis included 281 individuals (139 males, 142 females) ages 7-18 in 1981, who had BMI data at baseline and HRQL data in 2002-2004 (ages 29-41). This represents 34.4% of the 817 participants aged 7-18 at baseline.

Measures

Baseline height and weight were measured in 1981. At follow-up (2002-04), all data were self-reported via questionnaire. BMI was calculated as weight (kg)/height ([m.sup.2]). Youth overweight and obesity were categorized according to the International Obesity Task Force (IOTF) age- and gender-specific cut-offs, (15) associated with the predicted WHO-defined adult BMI cut-offs of 25 (overweight) and 30 (obese) kg/[m.sup.2] at age 18. (16)

Multidimensional HRQL was assessed using the self-administered Medical Outcomes Study Short Form 36 (SF-36), comprised of 36 questions, assessing 8 physical and mental health domains: physical function (PF), role physical (RP), bodily pain (BP), general health perceptions (GH), vitality (VT), social functioning (SF), role emotional (RE) and mental health (MH). (17) Scores range from 0-100, with higher scores representing better health and less pain. Domains are also weighted and summed to calculate physical and mental component summaries (PCS and MCS), standardized to a mean of 50, with scores above and below 50 representing better and worse than average function, respectively. (18) A 5-point difference (domain scores) and 2-3 point difference (summary scores) are considered clinically relevant. (17,18)

Statistical analysis

The distributions of 5 of 8 SF-36 outcomes were negatively skewed, with many participants achieving the maximum score of 100 on the 8 domains: GH, VT and MH appeared normally distributed, and PCS and MCS deviated only slightly from normal. Spearman correlations between youth BMI (ages 7-18) and adult SF-36 scores were calculated. To account for the effect of age on youth BMI, the latter was regressed on age, and the residuals used in analysis to represent age-adjusted youth BMI. Differences in mean adult SF-36 scores by youth weight status were then assessed by independent samples t-tests and Mann Whitney U tests, adjusted for multiple comparisons (Holm method). Linear regression analysis was performed for GH, VT, MH, PCS and MCS. Initially, the association between adult BMI and the SF-36 outcomes was assessed; youth BMI was then tested both as a continuous and dichotomous predictor. Models were adjusted for age, adult BMI and physical activity, marital and smoking status, education level (university degree), household income (</>$60,000), and number of comorbidities (0-7). All analyses were performed separately for males and females.

[FIGURE 1 OMITTED]

RESULTS

Descriptive characteristics of the study participants are reported in Table 1. At baseline, approximately 9% of participants were overweight (12 male, 12 female) or obese (1 female); 22 years later, approximately 58% of males and 37% of females were classified as overweight/obese. Over 80% of overweight youth remained overweight/obese as adults, while about 36% of healthy weight youth became overweight/obese.

In females, youth BMI was significantly negatively correlated with PCS, and significantly positively correlated with RE, MH and MCS (Table 2). No significant correlations existed for males.

Figure 1 (a & b) illustrates the mean SF-36 scores by sex for healthy weight vs. overweight/obese youth. Both male and female overweight youth scored higher on RE (all scored the maximum of 100), and overweight females scored higher on MH and MCS, compared to their healthy weight counterparts; however, these differences did not remain statistically significant following adjustment for multiple comparisons.

An initial adjusted regression analysis assessed the association of adult BMI with the 5 normally distributed SF-36 scores. Overweight/obese males scored significantly lower on PCS (p<0.01) and overweight/obese females scored significantly lower on GH (p<0.05) compared to their healthy weight adult counterparts (data not shown).

Tables 3 and 4 show the linear regression results, by sex, for youth BMI and adult SF-36 scores, with and without adjustment for adult BMI. No significant predictions were observed in the male participants (Table 3). Prior to adjustment for adult BMI, overweight/obese female youth scored 10 points higher on both VT and MH and 9.1 points higher on MCS in adulthood, compared to healthy weight female youth (Table 4a, Models A), and a 1 unit increase in youth BMI led to 1.2 and 1.1 point increases in adult MH and MCS, respectively (Table 4b, Models A). After adjustment for adult BMI, on average, overweight/obese female youth scored 16, 13.4, and 12.7 points higher on GH, VT and MH in adulthood, respectively, and 10.9 points higher on MCS, compared to healthy weight female youth (Table 4a, Models B); a 1 unit increase in youth BMI led to 1.7, 1.5 and 1.4 point increases in adult VT, MH and MCS, respectively (Table 4b, Models B).

DISCUSSION

To our knowledge, the PALS study is the first to examine the long-term effects of youth overweight on adult HRQL. Building on earlier findings, (12) we found significant positive associations between youth BMI and subsequent adult HRQL 22 years later in females, but no association in males. Past research has shown that youth overweight/obesity can affect physical growth and development into adulthood, (19) and is linked to worse self-reported health,9 increased likelihood of physical limitations, (8,9) decreased psychosocial functioning and self-esteem, (8,20) and increased incidence of bullying. (21) However, we found no long-term negative implications on adult HRQL, at least at moderate levels of youth overweight.

In a previous analysis comparing the PALS sample against Canadian SF-36 normative values, overweight youth were more likely than healthy weight youth to score at/above the norm on both MH and MCS, with positive associations also found for GH, SF and RE; however, results for GH in particular implicated adult BMI in a moderating role, with the relative increase in BMI from youth to adulthood likely the key factor involved, rather than youth BMI itself exerting an independent effect. (12)

In the present sex-stratified analysis, a higher youth BMI predicted higher absolute scores on RE, VT, MH and MCS in adulthood, but only in females. Regression results are presented both with and without adjustment for adult BMI, for reasons discussed at length previously. (12) Briefly, as this variable may not be a true confounder (22,23) but may instead (or also) be a moderator in the pathway from youth BMI to adult HRQL, the reversal paradox may be invoked whereby the association between two variables can be reversed or enhanced when another variable is statistically controlled. (23,24) Indeed, the relationship of youth BMI with GH in females became significant only upon inclusion of adult BMI in the regression model. This inclusion also somewhat strengthened the positive relationship of youth BMI to VT, MH and MCS. As adult BMI was also found to be a significant negative predictor of GH, it is likely that the relationships may be partly (VT, MH and MCS) or completely (GH) due to the relative change in BMI from youth to adulthood, rather than to a purely independent effect of youth BMI on adult HRQL. (12)

In adults, several studies have reported that overweight and obesity were related to poorer physically oriented SF-36 outcomes. (5,6,25-28) In this analysis, current adult BMI was a significant negative predictor of GH (females) and PCS (males) only. While tracking of BMI in the PALS sample is high, with over 80% of overweight youth becoming overweight/obese adults, the majority of overweight/ obese adults were not overweight in their youth. (29) Thus, current adult weight status may be more important to physical HRQL than past youth weight status.

Previously reported associations of adult obesity with the mentally oriented SF-36 outcomes have been weaker and less clear than those found for the physical domains and PCS, (6,25,26,28) including some studies which found a positive association between adult obesity and the SF-36 mental domains; (5-7,27) as such, our findings of a positive association of youth BMI to the adult RE, MH and MCS may not be completely unusual.

While potential sex differences have been identified in previous studies, these have been inconsistent. One study showed that obese men ages 16-34 were more likely to rate both their physical and mental HRQL lower than healthy weight men, compared to women who did not report worse mental HRQL. (28) Another found that after controlling for physical ill-health, overweight and obesity were associated with better mental health in women, but not in men. (7) Another reported that while both obese men and women showed worse physical functioning than healthy weight persons, obesity was associated with better mental HRQL in males only. (5)

As very few overweight youth became healthy weight adults (Table 1), relative "weight loss" over time is unlikely to be driving the positive association of youth overweight with adult mental HRQL in females. In a previous analysis of longitudinal BMI status change, PALS participants who were overweight/obese both as youth and adults had significantly greater odds of scoring at/above the norm on both MH and MCS, while relative "weight gain" over time led to decreased odds of scoring at/above the norm on GH. (12) As well, in a logistic regression stratified by adult BMI categories, youth BMI appeared to be a significant positive predictor of scoring at/above the norm on RE, MH and MCS in both healthy weight and overweight/obese adults. (12) The present analysis confirms a positive association of youth BMI to adult mental HRQL in terms of absolute scores in females, which seems at least partly independent of adult BMI or the change in weight between the two time points.

The main strengths of this research are the 22-year length of follow-up from childhood into adulthood in a representative Canadian population sample, the use of measured BMI data at baseline, and the use of the valid and widely-accepted SF-36 instrument.

Study limitations include the small sample size and high attrition, a common problem in longitudinal studies spanning several decades. BMI as a measure of overweight is inherently limited in its lack of information regarding body composition, and lack of sensitivity to youth maturational age; however, it remains the most widely investigated and easily used indicator of weight-related health risk on a population level. Moreover, the effects of age and maturation were diminished by our use of IOTF cut-offs for youth BMI categorization, age-adjusting youth BMI to examine correlations, and adjusting for age in regressions, all within a sex-stratified analysis. Finally, while the prevalence of youth overweight in the PALS sample is representative of Canada in 1981, the low absolute number of overweight/obese youth (only 1 obese child) may have impacted results. As our results are based on a primarily healthy weight sample, caution should be used when extrapolating findings to overweight/obese groups due to the small sample in these groups. The prevalence of youth overweight/obesity has nearly tripled since that time, (2) which may yet lead to negative consequences on HRQL over the long term.

In conclusion, these results suggest that moderate levels of youth overweight may not lead to increased risk of poorer HRQL in adulthood, and may be associated with better mentally-oriented HRQL in females. However, as youth overweight has known immediate health consequences, (19) and tracks strongly into adulthood, (29) where it is associated with substantially increased risk of a host of other biomedical conditions, (1) it is not an issue that can be ignored. It is also possible that greater and more negative HRQL consequences might result from the higher degrees of youth obesity present in today's population. More research is needed on larger cohorts with a higher prevalence and degree of youth obesity to further assess the long-term impact on physical and mental health, including further exploration of sex differences, possible critical age ranges, and the role of BMI status changes over time.

Acknowledgements: K.M. Herman was supported by a Canadian Institutes of Health Research Doctoral Research Award. PALS was funded by the Social Sciences and Humanities Research Council of Canada and the Canadian Institutes of Health Research (Strategic Joint Initiative Grant on Society, Culture and the Health of Canadians II Grant No. 839.2000-1032; C.L. Craig, L. Gauvin). The 1981 CFS was supported by Fitness and Amateur Sport (now the Healthy Living Program of the Public Health Agency of Canada). The authors thank Peter Katzmarzyk for his contribution to the conceptualization of the paper and direction provided during early stages of this work.

Conflict of Interest: None to declare.

Received: April 21, 2010

Accepted: July 23, 2010

REFERENCES

(1.) Haslam DW, James WP. Obesity. Lancet 2005;366(9492):1197-209.

(2.) Statistics Canada. Health Indicators, June 2006. Ottawa, ON: Government of Canada, 2006: Catalogue No. 82-221-XIE. Available at: www.statcan.ca/english/freepub/82-21-XIE/82-21-XIE2006001.htm (Accessed April 6, 2010).

(3.) Centers for Disease Control and Prevention. Measuring Healthy Days: Population Assessment of Health-Related Quality of Life. Atlanta, GA: CDC, 2000.

(4.) Kolotkin RL, Meter K, Williams GR. Quality of life and obesity. Obes Rev 2001;2(4):219-29.

(5.) Lopez-Garcia E, Banegas Banegas JR, Gutierrez-Fisac JL, Perez-Regadera AG, Ganan LD, Rodriguez-Artalejo F. Relation between body weight and health-related quality of life among the elderly in Spain. Int J Obes Relat Metab Disord 2003;27(6):701-9.

(6.) Hopman WM, Berger C, Joseph L, Barr SI, Gao Y, Prior JC, et al. The association between body mass index and health-related quality of life: Data from CaMos, a stratified population study. Qual Life Res 2007;16(10):1595-603.

(7.) Jorm AF, Korten AE, Christensen H, Jacomb PA, Rodgers B, Parslow RA. Association of obesity with anxiety, depression and emotional well-being: A community survey. Aust N Z J Public Health 2003;27(4):434-40.

(8.) Friedlander SL, Larkin EK, Rosen CL, Palermo TM, Redline S. Decreased quality of life associated with obesity in school-aged children. Arch Pediatr Adolesc Med 2003;157(12):1206-11.

(9.) Swallen KC, Reither EN, Haas SA, Meier AM. Overweight, obesity, and health-related quality of life among adolescents: The National Longitudinal Study of Adolescent Health. Pediatrics 2005;115(2):340-47.

(10.) Williams J, Wake M, Hesketh K, Maher E, Waters E. Health-related quality of life of overweight and obese children. JAMA 2005;293(1):70-76.

(11.) Story MT, Neumark-Stzainer DR, Sherwood NE, Holt K, Sofka D, Trowbridge FL, et al. Management of child and adolescent obesity: Attitudes, barriers, skills, and training needs among health care professionals. Pediatrics 2002;110(1 Pt 2):210-14.

(12.) Herman KM, Hopman WM, Craig CL. Are youth BMI and physical activity associated with better or worse than expected health-related quality of life in adulthood? The Physical Activity Longitudinal Study. Qual Life Res 2010;19(3):339-49.

(13.) Craig CL, Gauvin L, Cragg S, Katzmarzyk PT, Stephens T, Russell SJ, et al. Towards a social epidemiological perspective on physical activity and health: The aims, design, and methods of the Physical Activity Longitudinal Study (PALS). JPhys Act Health 2005;2(3):272-84.

(14.) Stephens T, Craig CL, Ferris BF. Adult physical activity in Canada: Findings from the Canada Fitness Survey I. Can J Public Health 1986;77(4):285-90.

(15.) Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: International survey. BMJ 2000;320(7244):1240-43.

(16.) World Health Organization. Obesity. Preventing and Managing the Global Epidemic. Report of a WHO Consultation on Obesity. Geneva: World Health Organization, 1998.

(17.) Ware JE, Snow KK, Kosinski M, Gandek B. SF-36 Health Survey Manual and Interpretation Guide. Boston, MA: The Health Institute, New England Medical Centre, 1993.

(18.) Ware JE, Kosinski M, Keller SD. SF-36 Physical and Mental Health Summary Scales: A User's Manual. Boston: The Health Institute, New England Medical Centre, 1994.

(19.) Must A, Strauss RS. Risks and consequences of childhood and adolescent obesity. Int J Obes Relat Metab Disord 1999;23(Suppl 2):S2-S11.

(20.) Wang F, Veugelers PJ. Self-esteem and cognitive development in the era of the childhood obesity epidemic. Obes Rev 2008;9(6):615-23.

(21.) Janssen I, Craig WM, Boyce WF, Pickett W. Associations between overweight and obesity with bullying behaviors in school-aged children. Pediatrics 2004;113(5):1187-94.

(22.) Lucas A, Fewtrell MS, Cole TJ. Fetal origins of adult disease-the hypothesis revisited. BMJ 1999;319(7204):245-49.

(23.) Tu YK, West R, Ellison GT, Gilthorpe MS. Why evidence for the fetal origins of adult disease might be a statistical artifact: The "reversal paradox" for the relation between birth weight and blood pressure in later life. Am J Epidemiol 2005;161(1):27-32.

(24.) Tu YK, Gunnell D, Gilthorpe MS. Simpson's Paradox, Lord's Paradox, and Suppression Effects are the same phenomenon--The reversal paradox. Emerg Themes Epidemiol 2008;5:2.

(25.) Doll HA, Petersen SE, Stewart-Brown SL. Obesity and physical and emotional well-being: Associations between body mass index, chronic illness, and the physical and mental components of the SF-36 questionnaire. Obes Res 2000;8(2):160-70.

(26.) Ni Mhurchu C, Bennett D, Lin R, Hackett M, Jull A, Rodgers A. Obesity and health-related quality of life: Results from a weight loss trial. N Z Med J 2004;117(1207):U1211.

(27.) Huang IC, Frangakis C, Wu AW. The relationship of excess body weight and health-related quality of life: Evidence from a population study in Taiwan. Int J Obes (Lond) 2006;30(8):1250-59.

(28.) Larsson U, Karlsson J, Sullivan M. Impact of overweight and obesity on health-related quality of life--A Swedish population study. Int J Obes Relat

Metab Disord 2002;26(3):417-24.

(29.) Herman KM, Craig CL, Gauvin L, Katzmarzyk PT. Tracking of obesity and physical activity from childhood to adulthood: The Physical Activity Longitudinal Study. Int J Pediatr Obes 2008;4(4):281-88.

Katya M. Herman, PhD, [1] Wilma M. Hopman, MA, [2,3] Cora L. Craig, MSc [4,5]

Author Affiliations

[1.] School of Kinesiology and Health Studies, Queen's University, Kingston, ON

[2.] Clinical Research Centre, Kingston General Hospital, Kingston, ON

[3.] Department of Community Health and Epidemiology, Queen's University, Kingston, ON

[4.] Canadian Fitness & Lifestyle Research Institute, Ottawa, ON

[5.] School of Public Health, University of Sydney, Sydney, NSW, Australia

Correspondence: Katya M. Herman, PhD, School of Kinesiology & Health Studies, Queen's University, Kingston, ON K7L 3N6, Tel: 613-533-2666, Fax: 613-533-2009, E-mail: 5ch3@queensu.ca

Table 1. Sample Characteristics at Baseline (1981) and
22-year Follow-up (2002-04)--The Physical Activity
Longitudinal Study

Characteristic                                   Males       Females
                                                (N=139)      (N=142)

Baseline 1981: Ages 7-18 years                 12.7 (3.4)   12.2 (3.5)
  BMI: Healthy Weight                          91.4%        90.8%
    Overweight/Obese *                          8.6%         9.2%
  PA (kcal/kg/day)                              4.9 (5.5)    4.4 (4.9)
  PA [greater than or equal to]
     1.5 kcal/kg/day                           80.8%        67.0%
  PA [greater than or equal to]
     3.0 kcal/kg/day                           57.7%        46.4%

Follow-up 2002-2004: Ages 29-41 years          35.4 (3.4)   34.9 (3.6)
  BMI (kg/[m.sup.2])                           26.0 (3.4)   25.3 (5.2)
    Underweight (<18.5)                         1.4%         1.5%
    Healthy Weight (18.5-24.9)                 40.6%        61.2%
    Overweight (25.0-29.9)                     47.1%        22.4%
    Obese ([greater than or equal to]30.0)     10.9%        14.9%
  PA (kcal/kg/day)                              4.2 (5.0)    3.0 (3.6)
    PA [greater than or equal to]
      1.5 kcal/kg/day                          63.7%        60.2%
    PA [greater than or equal to]
      3.0 kcal/kg/day                          42.0%        31.2%
  Current Smoker                               19.0%        21.3%
  Self-Reported Health: Very Good/Excellent    48.5%        45.1%
  Married                                      71.7%        82.3%
  Education: [greater than or equal to]1
     University Degree                         43.1%        43.6%
  Total Household Income: [greater than or
     equal to]$60,000                          63.8%        68.5%
  Comorbidities: [greater than or equal to]1   46.8%        62.0%
  Longitudinal BMI change (1981-2002-2004):
    Healthy Weight--Healthy Weight             41.2%        59.1%
    Healthy Weight--Overweight/Obese           50.7%        32.6%
    Overweight/Obese--Healthy Weight            0.0%         3.0%
    Overweight/Obese--Overweight/Obese          8.1%         5.3%

Values are means (SD) or percentages of total sample.

* There is only 1 obese child (F); 12 overweight M, 12 overweight F.

Table 2. Spearman Correlations Between Youth BMI ([dagger]) (1981)
and Adult SF-36 Scores (2002-04)--The Physical
Activity Longitudinal Study

             Male (n=139)         Female (n=142)

SF-36       r       p-value       r       p-value

PF       -0.02      0.784      -0.03      0.777
RP        0.06      0.507      -0.08      0.355
BP        0.09      0.320      -0.09      0.301
GH       -0.02      0.815      -0.08      0.339
VT       -0.07      0.409       0.14      0.090
SF        0.00      0.976       0.14      0.095
RE        0.10      0.231       0.19 *    0.031
MH        0.09      0.319       0.22 **   0.008
PCS      -0.04      0.647      -0.21 *    0.013
MCS       0.02      0.781       0.26 **   0.002

([dagger]) Age-adjusted residuals from regression of BMI on Age.
PF=physical functioning; RP=role physical; BP=bodily pain;
GH=general health; VT=vitality; SF=social functioning; RE=role
emotional; MH=mental health; PCS=physical component summary;
MCS=mental component summary

* p<0.05, ** p<0.01.

Table 3. Adjusted Linear Regression of Adult SF-36
Outcomes (2002-04) on Youth BMI (1981)--The Physical
Activity Longitudinal Study--Males

SF-36     (A) [beta] (95% CI)   p-value   (B) [beta] (95% CI)   p-value
Outcome

a) Youth BMI (Overweight/Obese vs. Healthy Weight)
GH        2.7 (-9.0,14.3)       0.650     5.4(-6.3, 17.1)       0.362
VT        -3.4 (-16.4,9.5)      0.602     -1.8(-15.0,11.5)      0.791
MH        7.1 (-4.0,18.3)       0.210     7.3(-4.3, 18.8)       0.214
PCs       -1.5 (-7.0, 4.0)      0.599     0.3(-5.2, 5.7)        0.924
MCS       5.0 (-2.8, 12.7)      0.604     4.7(-3.3, 12.6)       0.246

b) Youth BMI (Continuous)
GH        0.0 (-1.5, 1.5)       0.987     0.5(-1.12.1)          0.517
VT        -0.8 (-2.5, 0.8)      0.319     -0.6(-2.3,1.2)        0.518
MH        0.5 (-1.0, 2.0)       0.503     0.5(-1.0, 2.1)        0.515
PCs       -0.1 (-0.8, 0.6)      0.743     0.2(-0.5,1.0)         0.525
MCS       0.4 (-0.6,1.4)        0.466     0.3(-0.8,1.4)         0.571

GH=general health; VT=vitality; MH=mental health; PCS=physical
component summary; MCS=mental component summary.

"A" models adjusted for age, adult physical activity, household
income, education, married, current smoker, # comorbidities; "B"
models adjusted for adult BMI, in addition to previous factors.

Table 4. Adjusted Linear Regression of Adult SF-36
Outcomes (2002-04) on Youth BMI (1981)--The
Physical Activity Longitudinal Study--Females

SF-36        (A) [beta]                     (B) [beta]
Outcome       (95% CI)        p-value        (95% CI)        p-value

a) Youth BMI (Overweight/Obese vs. Healthy Weight)
GH        9.4 (-1.8, 20.5)      0.098    16.0 (3.7, 28.4)      0.011
VT        10.0 (0.4, 19.5)      0.041    13.4 (2.8, 24.1)      0.014
MH        10.0 (1.0, 19.0)      0.029    12.7 (2.6, 22.8)      0.014
PCS        1.6 (-3.6, 6.7)      0.604     2.4 (-3.5, 8.3)      0.429
MCS        9.1 (2.2, 16.0)      0.010    10.9 (3.0, 18.8)      0.007

b) Youth BMI (Continuous)
GH         0.1 (-1.3, 1.4)      0.929     1.2 (-0.5, 2.8)      0.162
VT         0.8 (-0.4, 2.0)      0.173      1.7 (0.3, 3.1)      0.017
MH          1.2 (0.2, 2.3)      0.027      1.5 (0.2, 2.8)      0.024
PCS       -0.3 (-0.9, 0.3)      0.290    -0.3 (-1.1, 0.4)      0.396
MCS         1.1 (0.3, 1.9)      0.010      1.4 (0.4, 2.4)      0.008

GH=general health; VT=vitality; MH=mental health; PCS=physical component
summary; MCS=mental component summary.

"A" models adjusted for age, adult physical activity, household income,
education, married, current smoker, # comorbidities; "B" models adjusted
for adult BMI, in addition to previous factors.