Extra Method & Context Flashcards

(20 cards)

1
Q

Recruitment years for UK Biobank baseline?

A

2006 – 2010, with dietary recalls added 2009-2012.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Final censor date used in follow-up?

A

31 December 2020 (≈ 10 y max follow-up).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Age window of baseline cohort?

A

40 – 69 years.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Exact HbA1c definition for pre-diabetes?

A

5.7 % – 6.4 % (39-47 mmol/mol).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

How many WebQ recalls could each participant complete?

A

Up to five 24-h recalls; mean of all used to lower measurement error.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Scoring range for every PDI index?

A

18 (min) – 90 (max).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Primary composite outcome ICD-10 codes captured which three diagnoses?

A

Coronary heart disease, stroke, and atrial fibrillation.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Two-step screening before analysis?

A

Remove baseline CVD/cancer, then drop invalid diet recalls, T1D, extreme intake.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Extreme‐energy cutoff values (men / women)?

A

< 600 or > 4 200 kcal (men); < 500 or > 3 600 kcal (women).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Covariate capturing neighbourhood deprivation?

A

Townsend Deprivation Index (TDI).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Difference between Model 1 and Model 2 in Cox analysis?

A

Model 1: age + sex only; Model 2: full multivariable block incl. BMI, TDI, lifestyle, clinical history (plus disease duration in diabetes).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Software noted for mediation (‘mediate’ macro) runs in which package?

A

SAS (version 9.4).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Proportional-hazards assumption — how assessed?

A

Authors state they checked before Cox modelling (method not detailed).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Sensitivity test that tackles reverse causality?

A

Excluded events occurring within 2 years of baseline; results unchanged.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Key subgroup where hPDI benefit disappeared?

A

People ≥ 60 y or with family history of diabetes.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Main dietary component explaining uPDI harm in diabetes?

A

Low whole-grain intake (≈ 36 % of excess risk).

17
Q

Biomarker panel logged before diet recalls — why a limitation?

A

Temporal mismatch can cloud causality in mediation paths.

18
Q

Attrition percentage from initial at-risk pools?

A

About 70 % (pre-DM) and 75 % (DM) excluded before final sample.

19
Q

Free-text keyword authors use for “unhealthful plant foods”?

A

“Ultra-processed, sugary, or refined plant foods.”

20
Q

UK Biobank ethical approval reference?

A

North West MREC 11/NW/0382.