Physical activity is the centerpiece of many strategies to prevent and reduce chronic disease and improve health. The US Department of Health and Human Services (HHS) sets Physical Activity Guidelines for Americans based on the current body of evidence.  Adults are encouraged to engage in 2 hours and 30 minutes of moderate intensity physical activity per week; children and adolescents are encouraged to participate in 60 minutes of physical activity daily.

Suppose that we work for HHS, which is interested in understanding how body size and physical activity during different periods in adolescence may impact health. We are tasked with critically evaluating a recent study that examined the association between adolescent body size, physical activity, and risk of breast cancer (Oh et al. Interaction between early life body size and physical activity on risk of breast cancer. Int J Cancer 2015; 137:571-581). 

This investigation was a prospective cohort study of female nurses aged 25-42 who were free of breast cancer at enrollment in 1989. In 1997, participants were asked to report their childhood and adolescent body size (at ages 5, 10, and 20) using a 9-level figure drawing (1 being very thin and 9 being overweight). Participants have been followed-up since 1989 for incidence of breast cancer.

1. Based on the description in the introductory paragraph, which potential confounding variables were controlled through the use of restriction?
2. The following table shows some baseline characteristics of the study population according to average childhood body size. The characteristics listed are considered possible risk factor for breast cancer.

Average Body Size at ages 5-10 years

Characteristic

1

1.5-2

2.5-3

3.5-4

4.5+

Age (years)

42.5

41.8

42.0

42.3

42.5

Birth weight >=3.9 kg, %

10.0

11.5

13.9

16.2

18.7

Age at menarche

12.8

12.6

12.3

12.1

12.0

Total physical activity at ages 12-17  (MET-hr/wk)

56.3

54.5

51.9

48.4

46.0

Age at first birth (years)  (among parous women)

26.3

26.5

26.5

26.5

26.5

Family history of breast cancer, %

9.6

9.7

9.3

9.3

9.5

State which variables are potential confounders and which variables are not potential confounders according to these data. Give the rationale for your answer.

3- In the multivariate regression, investigators compare participants with body size=1 (index group) to participants with body size ? 4.5 (reference group). Table 1 shows the crude and adjusted risk ratios. The age-adjusted RR controlled only for age. The MV-adjusted (multivariable-adjusted) RR controlled for many potential confounders at the same time.

Crude and adjusted RR for the association between adolescent body size and breast cancer

RR for breast cancer

95% CI

Crude

1.79

1.28-2.19

Age-adjusted

1.64

1.31-2.06

MV-adjusted

1.64

1.29-2.03

Which of the following is the best interpretation of the crude RR?

• In these data, women with adolescent body size? 4.5 had 1.79 times the risk of breast cancer compared to those with adolescent body size=1 over the study period.
• In these data, women with adolescent body size=1 had 1.79 times the risk of breast cancer compared to those with adolescent body size? 4.5 over the study period, adjusting for numerous potential confounders.
• In these data, women with adolescent body size=1 had 1.79 times the crude risk of breast cancer compared to those with adolescent body size? 4.5 over the study period.

Which of the following is the best interpretation of the age-adjusted RR?

• In these data, women with adolescent body size=1 had 1.64 times the crude risk of breast cancer compared to those with adolescent body size? 4.5 over the study period.
• In these data, women with adolescent body size=1 had 1.64 times the risk of breast cancer compared to those with adolescent body size? 4.5 over the study period, adjusting for age.
• In these data, women with adolescent body size? 4.5 had 1.64 times the risk of breast cancer compared to those with adolescent body size=1 over the study period, adjusting for age.
• In these data, women with adolescent body size=1 had 1.64 times the risk of breast cancer compared to those with adolescent body size? 4.5 over the study period, adjusting for numerous potential confounders.

State in words your interpretation of the 95% confidence interval for the MV-adjusted RR.

Based on the 95% CI alone, is the MV-adjusted RR statistically significant?

What is the magnitude of confounding by this collection of potential confounders?

Is this collection of potential confounders actually confounding the association between adolescent body size and risk of breast cancer?