Development of the Eating Choices Index (ECI)

Development of the Eating Choices Index (ECI)

GK Pot1, CJ Prynne1, M Richards2, AM Stephen1

1MRC Human Nutrition Research, Cambridge UK

2MRC Unit for Lifelong Health and Ageing, London UK

Background

• Capture diet in one measure

• Numerous diet quality tools:

• Healthy Eating Index (HEI)

• Mediterranean Diet Score (MDS), etc.

• Mostly targeted at meeting recommendations in terms of nutrients or consumption of specific types of food

• Existing diet indices not aimed at dietary behaviour

Objective

• To develop an score that

• discriminates between overall healthy and unhealthy dietary behaviour

• can be applied to dietary assessments of various types (including diet diaries and repeated 24HR)

• can also be used as a short questionnaire to categorise individuals in large surveys

Elements of Eating Choices Index (ECI)

1. Frequency of Breakfast consumption

• Important indicator of healthy diet

• Often not assessed in FFQs

2. Type of milk

• Consumption of low-fat milk associated with overall dietary intake lower in fat and higher in fibre

3. Type of bread

• Associated with improved dietary quality

4. Fruit intake

• Important component of healthy diet

• Easier to quantify than vegetable intake

MRC National Survey of Health and Development 1946 British Birth Cohort

1946 Birth

36 4 BF M 43 53

= Dietary Assessment

63

2007-1

1

n=2256

n=1253

N=16,695

n=5632

http://www.nshd.mrc.ac.uk/default.aspx

Components of Eating Choices Index (ECI)

Factor Factor name Category 1 Category 2 Category 3 Category 4 Category 5

1 Frequency of

Breakfast

No days Some days

but not all

days

All days

Score 1 3 5

2 Type of milk

consumed

Whole (min

3.5%) only

Some whole,

some semi-

skimmed (1.7%)

1.Semi-

skimmed only

Some semi-

skimmed, some

skimmed

Skimmed

only (0.1%)

Score 1 2 3 4 5

3 Type of bread

consumed

White only Some white,

some brown or

granary

1.Brown or

granary only

2. Some

white, some

wholemeal

Some brown or

granary, some

wholemeal

Wholemeal

only

Score 1 2 3 4 5

4 Fruit intake

portions (per day)

0 >0 <1 1 >1 ≤ 2 ≥3

Score 1 2 3 4 5

Higher score, healthier dietary behaviour

Methods

• Correlations and quartiles of ECI score with indicator nutrients including macronutrients, and fibre, vitamin C, iron, calcium and folate

• Tracking over time (n=1253)

• Exploration of subgroup analyses of

• Low and high social class (manual vs. non-manual)

• BMI (based on WHO cut-offs)

• Four fruits vs. total fruits

Components of ECI

0 2 4 6 8 10 12 14

ECI

Fruit factor

Bread factor

Milk factor

Breakfast factor

ECI score

Women

Men

ECI correlates well with macronutrients

r CHO

All 0.25

Men 0.26

Women 0.22

r Protein

All 0.18

Men 0.13

Women 0.21

34

35

36

37

38

39

40

41

42

43

q1 q2 q3 q4 q1 q2 q3 q4 q1 q2 q3 q4

ALL men women

Fat

(%

FE

)

Fat (%food energy)

r Fat

All -0.22

Men -0.13

Women -0.31

Similar for protein

35

36

37

38

39

40

41

42

43

44

45

q1 q2 q3 q4 q1 q2 q3 q4 q1 q2 q3 q4

ALL men women

Carb

oh

yd

rate

(%

FE

)

Carbohydrate (% food energy)

0

2

4

6

8

10

12

14

16

18

20

q1 q2 q3 q4 q1 q2 q3 q4 q1 q2 q3 q4

ALL men women

Fibre: Total NSP (g/d)

ECI correlates well with fibre and micronutrients

All Men Women

NSP 0.51 0.53 0.54

Vitamin C 0.37 0.37 0.37

Iron 0.29 0.35 0.35

Calcium 0.24 0.25 0.32

Folate 0.21 0.21 0.35

ECI & Tracking over time

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

82 89 99

All

82 ECI >=15

82 ECI 13-15

82 ECI 10-12

82 ECI <=9

n=1253

ECI and socio-economic status

P trend <0.001

In 1989, 43y

n=432 n=565 n=492 n=615

Missing data n=152

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

89 ECI <=9 89 ECI 10-12 89 ECI 13-15 89 ECI >=15

Manual

Non manual

ECI and BMI

P trend <0.001

In 1989, 43y

Missing data n=13

n=615 n=432 n=565 n=492

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

ECI <=9 ECI 10-12 ECI 13-15 ECI >=15

obese (>=30.0)

overweight (25-29.9)

normal weight (18.5-24.9)

underweight (<18.5)

Four fruits vs. total fruit

• 64% of total fruit intake from four fruits: apples, pears, bananas, and oranges

• Correlations with indicator nutrients very similar

Nutrient r ECI four fruits r ECI total fruit

Fat (% FE) -0.22* -0.22*

Protein (% FE) 0.15* 0.18*

CHO (% FE) 0.29* 0.25*

NSP # 0.51* 0.51*

Vitamin C # 0.32* 0.37*

Iron 0.29* 0.29*

Calcium 0.25* 0.24*

Folate 0.23* 0.21*

# data transformed, * p< 0.001

Discussion

• Generalizability

• Comparison to other ‘Diet quality’ scores

• Tracking over time:

• Changes in food availability

• Dietary recommendations

• Ageing of population

• Total fruit consumption vs. four fruits

Conclusion

• We have developed a score that

• Discriminates between overall healthy and unhealthy behaviour

• Can be applied to dietary assessments of various types

• Can also be used as a short questionnaire to categorise individuals in large surveys

• Next

• Validate as short questionnaire

• Relate to health outcomes longitudinally

• Examine in other population groups

Acknowledgements

• NSHD cohort members

• NSHD Interviewers & Researchers

• MRC Unit for Lifelong Health and Ageing

[email protected]

MRC Human Nutrition Research

Cambridge, UK

Thank you for your attention