Variables in need of explanation
Dataset variables containing survey data are coded according to the response codes in the interview questionnaires. Summary, categorical or dichotomous variables derived from the survey data are labelled in the datasets, and the notes below clarify those that may need further . The Audit dataset contains variables that enable researchers to track the participation of each cohort member through the course of the study.
Codes used in the audit dataset:
This details the source of data gathered for each person in each survey, thus provides a further breakdown of the summary matrix of participation (see Data Available) of how many respondent interviews and proxy informant interviews were collected in each survey.
Variable: in_s*
1 Respondent interview
5 Respondent interview and one proxy interview
6 One proxy interview only
7 Respondent interview and two proxy interviews
8 Two proxy interviews only, no respondent interview
9 Respondent and Proxy mixed interview
10 Proxy interview with interviewer assessment of participant (questionnaire final section)
17 Respondent interview in Survey 1 (Year 0), then not followed-up due to participation in an other study / recruitment from practice 5 (Chesterton Study).
Variable: audit_s*
This explains for each survey, if a respondent was not interviewed, the reason why not.
0 not approached
1 interviewed
2 too ill
3 died before start of current survey
4 died between start and end of current survey
5 moved, not traceable
6 refused
7 refused before visited by contacting study office
8 refused by a relative at any stage
9 refused at telephone contact
10 refused at door
11 permanent refusal
12 never in after many attempts
13 respondent re-joined for current survey after missing at least one previous survey
14 reason no respondent interview not known, but proxy informant interviewed
15 other
16 reason no interview not known
17 transfered to the Chesterton Study
18 only in survey 2
19 dropped because record contained no data
20 proxy re-joined for current survey after missing at least one previous survey
Variable: refused
This explains the reason why an individual was lost to follow-up from the study overall.
Values are coded with the s ame coding values as in the above variable audit_s*
Variable: refstage
This indicates at which stage of the study a respondent first missed interview.
1 Survey 1
2 Survey 2
3 Survey 3
4 Survey 4
5 Survey 5
6 Survey 6
7 Survey 7
8 Survey 8
9 Survey 9
Variable: rejoin
This indicates at which stage of the study a respondent re-joined if they missed any interview(s) but subsequently took part in a later survey.
2 Re-joined in Survey 2
3 Re-joined in Survey 3
4 Re-joined in Survey 4
5 Re-joined in Survey 5
6 Re-joined in Survey 6
9 Not applicable
GP code:
In Survey1year0 this variable indicates with which of the collaborating general practices each participant was registered when first enrolled in the study. In surveys after baseline, this variable indicates with which practice each participant was registered at the time of subsequent interview, with the additional code 8 identifying participants who had changed GP to a practice that was not one of the original partner practices.
1 Arbury Road Surgery
2 Nuffield Road Medical Centre
3 East Barnwell Health Centre
4 124 Newmarket Road (now York S treet M edical P ractice )
5 96 Chesterton Road
6 Lensfields Medical Practice
7 Cornford House Surgery
8 Other
Variable: rres*
(This is the recoded type of house, there are 6 answers on the questionnaires, but recoded to 5 answers in audit table (council residential home and private residential home have been merged together), also some missing values has been checked and populated based on the previous interviews if they did not move.)
-1 Missing
1 House/flat/granny flat
2 Warden controlled
3 Council residential home and private residential home
4 Long stay hospital
5 Other (From Survey 6 onwards other means Nursing home)
7 Not asked
8 Donot know
9 Not applicable
Variable: disab_s*
(This is a summary variable derived from responses to all the Activity of Daily Living (ADL) questions indicating whether or not respondents were disabled and, if so, to what extent.)
0 No disability
1 Disability in instrumental activities of daily living (IADL) only
2 Disability in both instrumental and basic activities of daily living (IADL+ADL)
The codes used in the Survey 1 Year 0 dataset:
Social class was coded according to the Office of Population and Censuses Classification of Occupations, using occupation classifications current in 1987:
I Professional etc occupations
II Managerial and technical occupations
IIIN Skilled non-manual occupations
IIIM Skilled manual occupations
IV Partly-skilled occupations
V Unskilled occupations
These are coded 1, 2, 3, 4, 5, 6 respectively in the Survey1Year0 database.