Skip to main content

Table 5 Logistic models for two year dementia

From: The descriptive epidemiology of delirium symptoms in a large population-based cohort study: results from the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS)

Ā 

N

Missing data

OR

LCI

UCI

P

Delirium symptom clusters

Ā Ā Ā Ā Ā Ā 

Inattention

1347

37

1.90

0.77

4.69

0.16

Acute change

1149

235

7.63

3.47

16.75

<0.01

Fluctuation

1347

37

6.84

3.67

12.77

<0.01

Drowsiness

1347

37

4.83

2.50

9.35

<0.01

Judged delirium*

1149

235

4.44

0.78

25.26

0.09

Delirium: final model

1140

244

Ā Ā Ā Ā 

ā€ƒDelirium

Ā Ā 

8.82

2.76

28.2

<0.01

ā€ƒAge (per year)

Ā Ā 

1.11

1.08

1.14

<0.01

ā€ƒSex (women versus men)

Ā Ā 

0.96

0.61

1.50

0.85

ā€ƒIllness severity

Ā Ā Ā Ā Ā Ā 

ā€ƒā€ƒNone

Ā Ā 

[Ref]

Ā Ā Ā 

ā€ƒā€ƒMild

Ā Ā 

1.66

0.57

4.79

0.35

ā€ƒā€ƒModerate

Ā Ā 

1.41

0.31

6.37

0.66

ā€ƒā€ƒSevere

Ā Ā 

(omitted)**

Ā Ā Ā 

Subsyndromal delirium: final model

1140

244

Ā Ā Ā Ā 

ā€ƒSubsyndromal delirium

Ā Ā 

4.31

2.41

7.73

<0.01

ā€ƒAge (per year)

Ā Ā 

1.10

1.07

1.14

<0.01

ā€ƒSex (women versus men)

Ā Ā 

0.94

0.60

1.47

0.78

ā€ƒIllness severity

Ā Ā Ā Ā Ā Ā 

ā€ƒā€ƒNone

Ā Ā 

[Ref]

Ā Ā Ā 

ā€ƒā€ƒMild

Ā Ā 

1.02

0.35

2.95

0.98

ā€ƒā€ƒModerate

Ā Ā 

1.54

0.41

5.77

0.52

ā€ƒā€ƒSevere

Ā Ā 

(omitted)**

Ā Ā Ā 
  1. OR odds ratio, LCI UCI 95% lower and upper confidence intervals respectively.
  2. This table shows logistic regression models for dementia at two year follow-up.
  3. ā€œDelirium symptom clustersā€ shows individual symptom clusters, and their association with dementia (adjusted for age, sex, baseline dementia and illness severity), i.e. the independent effects of each symptom cluster.
  4. *ā€˜Judged deliriumā€™ refers to the overall impression of the interviewer that a participant had delirium.
  5. ā€œDelirium: final modelā€ refers to the full model for full syndromal delirium and the same adjusted covariates.
  6. ā€œSubsyndromal delirium: final modelā€ describes the full model for subsyndromal delirium and the same adjusted covariates.
  7. **The effect of severe illness was not estimated due to no participants being assigned to this category.
  8. Missing data. These arise through non-responses in interviewed participants or respondents. There were also losses to follow-up that did not include death (nā€‰=ā€‰267). Estimated models are based on complete cases, i.e. where all covariates are complete for the sample.