International Journal of Epidemiology, 2014, 731–738
doi: 10.1093/ije/dyu057
Advance Access Publication Date: 20 March 2014
Data Resource Profile
Data Resource Profile
Data Resource Profile: The Swedish Panel Study
of Living Conditions of the Oldest Old
(SWEOLD)
Carin Lennartsson1*, Neda Agahi,1 Linda Hols-Salén,1
Susanne Kelfve,1,2 Ingemar Kåreholt,1,3 Olle Lundberg,4
Marti G Parker1 and Mats Thorslund1
1
Aging Research Center (ARC), Karolinska Institutet/Stockholm University, 2Department of Sociology,
Stockholm University, 3Institute of Gerontology, School of Health Sciences, Jönköping University and
4
Center for Health Equity Studies (CHESS), Stockholm University/Karolinska Institutet
*Corresponding author. Aging Research Center (ARC), Karolinska Institutet/Stockholm University, Gävlegatan 16,
SE-11330 Stockholm, Sweden. E-mail: [email protected]
Accepted 21 February 2014
Abstract
As the number and proportion of very old people in the population increase, there is
a need for improved knowledge about their health and living conditions. The SWEOLD
interview surveys are based on random samples of the population aged 77þyears. The
low non-response rates, the inclusion of institutionalized persons and the use of proxy
informants for people unable to be interviewed directly ensure a representative portrayal of this age group in Sweden. SWEOLD began in 1992 and has been repeated in
2002, 2004 and 2011. The survey is based on another national survey, the Swedish
Level of Living Survey (LNU), started in 1968 with 10-year follow-up waves. This longitudinal design provides additional data collected when SWEOLD participants were in
middle age and early old age. The SWEOLD interviews cover a wide range of areas
including health and health behaviour, work history, family, leisure activities and use of
health and social care services. Socio-economic factors include education, previous occupation and available cash margin. Health indicators include symptoms, diseases, mobility and activities of daily living (ADL). In addition to self-reported data, the interview
includes objective tests of lung function, physical function, grip strength and cognition.
The data have been linked to register data, for example for income and mortality follow-ups. Data are available to the scientific community on request. More information
about the study, data access rules and how to apply for data are available at the website
(www.sweold.se).
C The Author 2014; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association
V
731
732
International Journal of Epidemiology, 2014, Vol. 43, No. 3
Key Messages
• SWEOLD provides nationally representative data. The data waves can be studied as sequential cross-sectional surveys
to examine changes in health and well-being over time. Thus, the data can be used to identify age and cohort effects.
• SWEOLD, together with LNU, provide longitudinal interview data over a 40-year period. Individuals can be followed
through several waves from mid life to late life and mortality, thus allowing for a life course perspective.
• Analyses of SWEOLD data have provided valuable insights into health trends among the elderly Swedish population
and into health inequalities between women and men and different socio-economic groups.
Data resource basics
Level of living surveys with a broad multidimensional
approach have been carried out in Sweden since the late
1960s and have been widely used for the production of social statistics and welfare research on the adult population.
By the end of the 20th century, the growing number of
people over the age of 75 years in Sweden, as well as in
other high-income countries, emphasized the need for better data on older people’s living conditions including different aspects of their health. In response to these changes
and an increasing interest in older people’s living conditions and health status, the first SWEdish Panel Study of
Living Conditions of the Oldest OLD (SWEOLD) was carried out in 1992. At that time there were very few nationally representative interview surveys of very old people,
either in Sweden or internationally.
The aim of SWEOLD is to collect relevant, high-quality
and up-to-date data about older adults’ living conditions
and health. The SWEOLD sample was initially designed
based on an existing nationally representative sample of
the total adult population in Sweden—the Swedish Level
of Living Survey (LNU).1 LNU is one of the longestrunning longitudinal multidimensional surveys in the
world. SWEOLD comprises those persons previously
included in the LNU sample who have passed the LNU’s
upper age limit of 75 years and are still living in Sweden at
the time of the survey. LNU began in 1968 and, together
with subsequent survey waves, provides multidimensional,
longitudinal data on individuals over 4 decades.
Both LNU and SWEOLD cover a wide range of topics
relevant to respondents’ living conditions and well-being.
Each wave of both studies is nationally representative,
making the data ideal for repeated cross-sectional analyses
and the results generalizable to the total population. In
addition, combining the surveys provides the opportunity
to study the impact of mid-life conditions and life-course
events on late-life circumstances and health.
Data resource area and population coverage
The basis for LNU was a random sample of 1/1000 of the
Swedish population aged 15 to 75 years. Since the first
survey in 1968, LNU has been conducted in 1974, 1981,
1991, 2000 and most recently in 2010. The sample is obtained using the identification numbers issued to all
Swedish residents by the state. Thus the sample is based on
individuals, not households. The LNU was continued on a
panel data basis, and at each wave a random sample of
young people and immigrants was added to the original
sample, thus ensuring that each wave remained representative of the total adult population. For more information on
LNU see the website (www.sofi.su.se). All LNU/SWEOLD
respondents who have been in the sample in any previous
wave are part of the longitudinal sample. The system of
identification numbers facilitated follow-up and few people have been lost between the interview waves. For example, among those interviewed in SWEOLD 2011, 99%
had been interviewed at least once before, 92% had been
interviewed three times or more and 61% had been interviewed six times or more.
Figure 1 illustrates the sample design of the SWEOLD
surveys and how they are related to the LNU surveys. The
yellow vertical arrows show in which year each LNU survey was conducted and the blue horizontal lines show the
age limits in each survey. The red vertical arrows show the
SWEOLD surveys. As the green line illustrates, the youngest persons interviewed in the 2011 wave of SWEOLD
were 34 years old at the time of the 1968 wave of LNU.
People aged 69 and older in 1968 were too old already in
1974 to be included in the LNU sample. However, if they
were still alive they were included in the SWEOLD sample
in 1992.
Survey frequency
As described above, the SWEOLD sample includes all respondents who had previously been included (but not necessarily interviewed) in the LNU sample, and were living
in Sweden at the time of the interview. The only exception
is SWEOLD 1992 that was restricted to persons who had
previously been interviewed at least once in any of the previous LNU surveys. In 2011, the SWEOLD sample was
complemented by an additional representative sample of
the Swedish population aged 85–99. This additional
International Journal of Epidemiology, 2014, Vol. 43, No. 3
733
Sample design
120
110
100
SWEOLD
1992
90
SWEOLD
2011
SWEOLD SWEOLD
2002
2004
80
Age
Lower age limit 69
Upper age limit 75
70
60
LNU
1981
LNU
1974
LNU
1968
LNU
1991
LNU
2000
Lower age limit 18
Lower age limit 19
LNU
2010
50
40
30
20
10
Lower age limit 15
0
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
Year
Figure 1. The sample design of SWEOLD and LNU
sample allows for more detailed analyses in smaller age
and sex subgroups for those over 85 years of age. Since the
SWEOLD sample is obtained using the Swedish system of
personal identification numbers, it is representative of
the entire older population regardless of living situation
(i.e. type of dwelling did not affect the probability of being
included in the sample). The sex, age and educational
structure of the SWEOLD sample correspond to those of
the general Swedish population. Using the personal identification numbers had the added advantage that we were
able to locate everyone in the sample.
As can be seen in Table 1, we attained a high response
rate in all survey waves. The response rates varied between
84% and 95%. The relatively high response rates were
achieved through intensive and carefully planned fieldwork
that was designed to include very frail and cognitively impaired older persons.
The interviews were carried out in the participants’
homes or institutions where they lived and were based on a
structured questionnaire. In 1992 and 2002 the interviewers used paper questionnaires. Computer-assisted telephone interviewing (CATI) was used in 2004, and in 2011
the interviewers used computer-assisted personal interviewing (CAPI).
The SWEOLD surveys used various interview modes. A
direct face-to-face interview was always preferred. If a respondent was unable to participate or refused to meet face
to face, we offered to conduct the interview by telephone
(telephone interview) or with the support of a close relative
(mixed interview). The final method used was a selfadministered postal questionnaire. Postal questionnaires
were used in SWEOLD 2004 if the interview person had
hearing problems or if the interview person or the indirect
respondent did not want to answer the questions over the
phone. Postal questionnaires were used in SWEOLD 2011
for similar reasons but to a greater extent than in 2004.
The choice to extend the interview modes and include postal questionnaires raised the response rate in SWEOLD
2011 by 6.3 percentage units. The same questionnaire was
used irrespective of the interview mode although the postal
questionnaire was slightly shorter. Using the same structure and wording in the questionnaires independent of
interview mode decreases the risk of systematic differences
in reporting.2 Table 1 gives an overview of the sample with
regard to age, birth cohort, response rates for each interview mode, total response rates and non-response rates for
each survey year.
Indirect or so called proxy interviews were carried
out if the participant was not able to answer questions
by any of the methods described above. The most common reason for an indirect interview was dementia or
frailty. Indirect interviews were conducted with a close
relative, trustee or healthcare personnel. Most of the
indirect interviews were conducted by telephone. In ageing
734
International Journal of Epidemiology, 2014, Vol. 43, No. 3
Table 1. SWEOLD response rates and type of interview by survey year
Total response rate
Face-to-face interviews
Telephone interviews
Postal questionnaire
Indirect interviews
Mixed interviews
Non-response
Sample
1992 Age 77– Birth
cohorts 1892–1915
2002 Age 77–Birth
cohorts 1903–25
2004 Age 69–Birth
cohorts 1904–35
2011a Age 76–Birth
cohorts 1909–34
N
N
N
N
%
%
537
419
34
95.4
78.0
6.3
621
436
56
84.4
70.2
9.0
64
20
26
563
11.9
3.7
4.6d
100.0
82
47
115
736
13.2
7.5
15.6
100.0
%
1180
0b
963c
21
196
0
172
1352
87.3
0
81.6
1.8
16.6
0
12.7
100.0
%
931
538
90
58
187
58
149
1080
86.2
57.8
9.7
6.2
20.1
6.2
13.8
100.0
a
An additional sample of people aged 85 to 99 has been added to the original sample. The additional sample was stratified by gender and age. The female additional sample were born between 1911 and 1920 and the male additional sample were born between 1911 and 1925.
b
SWEOLD 2004 was conducted by telephone, consequently no direct face to face in-person interviews or mixed interviews were performed.
c
An additional 13 postal questionnaires were filled in by an indirect respondent.
d
In 1992 the sample only included persons who had participated in earlier LNU surveys. If the total LNU sample had been used and all of these persons would
have been non-responders this time too, we would have a non-response rate of 11.1%.
research, the use of indirect interviews shows fairly
good validity for easily observed variables, such as physical mobility and function, but the findings for more subjective health variables such as pain, self-rated health and
psychological problems are not as clear.3 For a further
discussion of both practical and ethical problems related to
the use of proxy interviews in SWEOLD, see reference
list.4,5
Table 2 provides information on the distribution of
women and men, age groups and living situations and the
corresponding distribution in the Swedish population at
the time of sample construction. Institutions refer to nursing homes, group homes for people with dementia or different types of assisted living facilities. The threshold into
institutions is determined by an individual’s health and access to informal care as well as policy, all of which are
changeable over time.6 As can be seen in Table 2, the
participants represent the population well in each survey,
perhaps with the exception of 2004 when people in institutional care are underrepresented.
Measures
SWEOLD has a broad multidimensional approach. Like
LNU, we seek to measure individuals’ command over resources in terms of family and social relations, material living conditions (income and wealth), health, education,
working conditions, political life, leisure time activities,
housing conditions, etc. The first time they are interviewed
in LNU, respondents are also asked retrospective questions
regarding childhood conditions (including family socioeconomic position, dissension in the family and health).
To a large extent the SWEOLD and LNU questionnaires
consist of identical questions. However, the SWEOLD
questionnaires are intended to capture key living conditions among those aged 75 and over. For example, questions about functional abilities and ADLs have been
included whereas questions about working life have been
excluded.
SWEOLD and LNU include questions about each of the
respondent’s children. This allows us to consider characteristics of each child in relation to characteristics of the parent. There are questions about geographical proximity,
generational transfers, conflicts and emotional closeness.
Table 3 presents the SWEOLD questionnaire topics and
examples of areas covered by the questions.
SWEOLD includes a number of objective tests to
measure physical and cognitive function. In all face-to-face
interviews, nine simple tests are used to measure physical
functioning, for example picking up a pen from the floor,
lifting one kilogram, reaching toes and rising from a
chair. The tests were selected to cover several aspects of
physical function, for example range of motion, balance and strength. Lung function is measured using the
peak expiratory flow test (PEF). SWEOLD 2011 also
included handgrip strength, measured by a hand-held
dynamometer, and collection of a saliva sample. All tests
are easily administered by the interviewers. These tests provide many possibilities for further multi-disciplinary
research.
Cognitive impairment increases sharply with age. The
Mini-Mental State Examination,7 that is widely used for
diagnostic and research purposes, was considered too time
consuming to be administered in the context of the survey
International Journal of Epidemiology, 2014, Vol. 43, No. 3
735
Table 2. SWEOLD in comparison with the total Swedish population by survey year, sex and age
1992
n
Women
Men
Age-groups
69–70
71–73
74–76
77–79
80–84
85–89
90þ
People aged 80þ
living in institution
Total
325
212
137
233
121
46
69
2002
%
% of total n
population
(31/12/91)
60.5
39.5
d
32.4
43.4
22.5
8.6
15.5
63.0
37.0
32.7
39.8
19.8
7.7
-e
537 100.0
368
253
158
243
151
69
92
2004
%
% of total n
population
(31/12/01)
59.3
40.7
25.4
39.1
24.3
11.1
17.7f
621 100.0
2011
%
% of total na
population
(31/12/03)
Valid % on the % of total
full weighted
population
sampleb
(31/12/10)
62.6
37.6
696
484
59.0
41.0
58.6
41.4
514
417
61.6
38.4
60.7
39.5
29.4
39.1
21.4
10.2
20.0
148
186
198
165
259
141
83
61
12.5
15.8
16.8
14.0
21.9
11.9
7.0
12.6
11.7
17.5
16.4
15.4
21.6
11.7
5.8
18.0f
27c
170
233
175
326
167
4.1
25.8
35.2
22.4
12.6
14.7
8.5
24.0
33.5
22.6
11.7
14.0f
931
100.0
1180 100.0
a
An additional sample of people aged 85 to 99 has been added to the original sample. The additional sample was stratified by gender and age. The female additional sample were born between 1911 and 1920 and the male additional sample were born between 1911 and 1925.
b
The data are weighted with sampling weights (the inverse of the probability that the observation is included) due the over-sampling in the highest age groups.
c
Only respondents age 76 at the time for interview.
d
Persons born 1915 (77 years old) had one-sixth the probability to of being sampled. Therefore the data are weighted with sampling weights.
e
Data not available, since official statistics this year included people in short-term care.
f
Data from the National Board of Health and Welfare.
Table 3. SWEOLD questionnaire topics and examples
Questionnaire modules/areas
Examples
Socio-demographics
Education and occupation
Economic resources
Children
Social relations
Health
Dental health
Health behaviour
Health care utilization
Medicines
Activities of daily living
Leisure-time activities
Political participation
Subjective evaluations
Personality
Objective tests of functiona
Objective tests of cognitiona
Biological samplea
Sex, age, county of residence, household composition, civil status
Years and level of education, main lifetime occupation, spouse’s main lifetime occupation
Income, assets, cash margin, financial problems, private financial transfers
Number of children and characteristics of each child
Social contacts (children, relatives, friends), help and care given and received
Self-rated health, pain, mobility, fatigue, diseases/disorders, height and weight
Functional dental health, problems with dentures, chewing ability
Dietary habits, smoking, alcohol usage
Hospitalization, medical attendance, nurse visits
Number and type of drugs, dosages, handling medicines
IADL (shopping, cooking, cleaning), ADL (bathing, using toilet, eating, dressing, in and out of bed, etc.)
Participation in activities and organizations, religious services
Voting, participating in society
Sense of coherence, estimation of living conditions, loneliness, faith
A short form of the Positive and Negative Affect Schedule (PANAS) (only 2011)
Peak flow test, nine performance tests, hand function tests, hand-grip tests (only 2011), vision tests
Condensed version of the Mini-Mental State Exam, cognitive ADL tests
Saliva (only 2011)
a
The objective tests are described in more detail in the text.
interview. However, a condensed version, as used in the
SWEOLD survey, has been shown to effectively identify
elderly persons with cognitive impairment.8,9 Since 2002
the survey also includes a test of cognition based on the
concept of everyday competence10 and items related to the
use of medicines.11
Interview data have been linked to annual register data
regarding income and date and cause of death.
736
Data resource use
Since 1992, SWEOLD has been used in a variety of publications, ranging from scientific articles to popular science
publications. It has also been used for government investigations, official reports and educational purposes, see website (www.sweold.se) for a complete publications list. The
repeated cross-sectional and longitudinal data can be used
to investigate a broad range of research questions.
Described below are some potential areas of interest to the
epidemiology of ageing that can be explored using
SWEOLD data. The repeated cross-sectional data waves
can be used to examine change over time in population
health and health behaviours. Cross-sectional analyses
have shown that more recent cohorts of older adults report
worse health but manage their daily activities better compared with earlier cohorts.12,13 More recent cohorts also
have more co-existing health and functional problems.14 In
addition, recent cohorts of older adults participate more in
sociocultural and physical activities,15 and are more likely
to drink alcohol compared with earlier cohorts (Kelfve
et al. manuscript submitted for publication).
Various indicators of socio-economic position are available in SWEOLD, making it suitable for research on health
inequalities. Using 1992 data we were among the first to
show that health inequalities between socio-economic
groups can be observed even among the oldest old.16,17
This relationship has subsequently been confirmed18 and
results show that the magnitude of the socio-economic
inequalities in health among older adults in Sweden has remained remarkably stable over time in spite of substantial
social changes.19 (also Fors and Thorslund, manuscript
submitted for publication). Data showed that there are social inequalities in pharmaceutical treatments in old age.20
SWEOLD can also be used to study health diversity in relation to gender. For example, we have shown that older
women generally had more health and dental problems
than men, but there were no gender differences in physician and dental visits21 or in the ability to perform
ADLs.22
Longitudinal data enable studies to explore the impact
of differential exposures, during different life stages, on
health and health inequalities in later life. For instance,
SWEOLD data have been combined with LNU data to
investigate work-related factors,23–25 leisure and social
engagement,26–29 living conditions during childhood30 and
health-related behaviours31 in mid life and later life, in
relation to health and health inequalities in old age.
Information about date of death from 1968 onwards
combined with survey data enable analyses of the impact
of childhood and/or mid-life living conditions and the
probability of surviving into old age.32,33 In addition, data
International Journal of Epidemiology, 2014, Vol. 43, No. 3
can be used to study mortality selection and how this impacts on ageing cohorts.
In addition to studying questions on health, health behaviours and inequalities, the longitudinal design and
broad multidimensional approach of SWEOLD make it
ideal for studying welfare problems and how such problems accumulate over the life course.34,35 SWEOLD data
also allow for studies on intergenerational exchange over
the life course of individuals and families.36 For example,
age-period-cohort analyses show that mobility limitations
are more common in older ages regardless of birth cohort
or time period, whereas differences in toothlessness, abstention from alcohol and smoking patterns are mainly due
to birth cohort, and these cohort differences remain more
or less unchanged over the life span.37–39
Strengths and weaknesses
SWEOLD is a nationally representative study population
of the oldest old in Sweden. We achieved a high response
rate by using mixed interview modes which increased response rates among the very frail and persons living in institutions. These groups tend to be older, less healthy, and
have a higher risk of mortality.5 As 14% of the Swedish
population aged 80 years and over were living in a homecare facility in 2011, it was particularly important for
SWEOLD to include individuals living in nursing homes,
group homes for people with dementia or assisted living
facilities.
SWEOLD follows individuals for more than 40 years,
which allows for the investigation of long-term trajectories
of health and other living conditions as well as the impact
of different exposures over the life course on health and living conditions in later life. Nearly all SWEOLD respondents have participated in at least one earlier data collection
and 6/10 respondents in SWEOLD 2011 have been interviewed in at least three previous survey waves. The linkage
of SWEOLD data to mortality data provides additional information about both the surviving population and those
who died before old age.
SWEOLD covers several key areas of life that are of special importance to older adults. The surveys include, as far
as possible, standardized subjective and objective measures
and each survey wave consists of the same variables and
objective tests. Many questions appear in all SWEOLD
and LNU waves. Hence, SWEOLD provides a rich and
multi-purpose longitudinal database for the scientific community and for use in official statistics.
Limitations include:
i. The sample size. SWEOLD is based on the LNU sample, a random sample of the Swedish population.
Persons older than 75, still alive and living in Sweden
International Journal of Epidemiology, 2014, Vol. 43, No. 3
are included. Due to this selection procedure,
SWEOLD data sometimes lack statistical power for
detailed analyses. The SWEOLD sample of 2011 was
complemented by an additional representative sample
of the oldest population to allow for more statistical
power in analysing those aged over 85.
ii. Intervals of 6–10 years between measurement waves
are too long for estimating trajectories of health and
behaviour change among the oldest old, especially
end-of-life changes. The long intervals between the
interview waves are of less importance when studying
relationships between mid-life circumstances and those
in later life.
iii. In prioritizing a broad multi-dimensional approach
with a minimum of respondent burden, the SWEOLD
interviews were not able to cover all areas in depth.
For example, although we have data about some diseases, and measures of cognition and depression, the
information is rather crude and lacks detail.
Data resource access
The data have been cleaned and new variables and indices
have been generated from the original questions. Wherever
possible, we used standardised measures and tests commonly used in surveys targeting older populations. The
datasets are described and presented in three published
code books available in English on the website (www.
sweold.se) or on request. This documentation includes
variable names, value labels and details on the construction
of variables, as well as frequencies and number of valid
cases for all included variables, minimum value, maximum
value, mean and standard deviation when appropriate.
Information about identical or comparable variables from
the three previous SWEOLD surveys and five previous
LNU surveys are also presented.
We encourage the use of SWEOLD as a resource for social and epidemiological research. Data are available on request and can be provided after approval by the Stockholm
Ethical Review Board. More information about the study
and information about data access, data access rules and
application forms are available at the website (www.
sweold.se).
Funding
SWEOLD 1992 was financed by three different research grants: the
Swedish Council for Social Research (91-0150); the Swedish
Medical Research Council (B92-27Ä-09997-01); and the Swedish
Council for Planning and Coordination of Research (910245:6-5/
86). The 2002 and 2004 studies were financed by the Swedish
Research Council (2001-6651); and the 2011 study was financed by
737
the Swedish Research Council (2009-6183) and the Swedish
Council for Working Life and Social Research (2010-1684).
Conflict of interest: None declared.
References
1. Fritzell J, Lundberg O (eds). Health Inequalities and Welfare
Resources. Bristol, UK: Policy Press, 2007.
2. Dillman DA, Christian LM. Survey mode as a source of instability in responses across surveys. Field Methods 2005;17:30–52.
3. Smith KV, Goldman N. Measuring health status: self-, interviewer, and physician reports of overall health. Journal of Aging
and Health 2011;23:242–66.
4. Lundberg O, Thorslund M. Fieldwork and measurement considerations in surveys of the oldest old. Soc Indicators Res
1996;37:165–87.
5. Kelfve S, Thorslund M, Lennartsson C. Sampling and non-response bias on health-outcomes in surveys of the oldest old. Eur
J Ageing 2013;10: 237–45.
6. National Board of Health and Welfare. (NBHW) Public Health
Report 2009. [Folkhälsorapport 2009] Västerås, Sweden:
National Board of Health and Welfare, 2009.
7. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”: a
practical method for grading the cognitive state of patients for
the clinician. J Psychiatr Res 1975;12:189–98.
8. Braekhus A, Laake K, Engedal K. The Mini-Mental State
Examination: identifying the most efficient variables for detecting cognitive impairment in the elderly. J Am Geriatr Soc
1992;40:1139–45.
9. Parker MG, Gatz M, Thorslund M. Brief cognitive screening in a
field survey of the oldest old. Aging 1996;8:354–59.
10. Diehl M. Everyday competence in later life: current status and
future directions. Gerontologist 1998;38:422–33.
11. Beckman AGK, Parker MG, Thorslund M. Can elderly people
take their medicine? Patient Educ Couns 2005;59:186–91.
12. Parker MG, Ahacic K, Thorslund M. Health changes among
Swedish oldest old: Prevalence rates from 1992 and 2002 show
increasing health problems. J Gerontol A Biol Sci Med Sci
2005;60:1351–55.
13. Fors S, Lennartsson C, Agahi N, Parker MG, Thorslund M.
Intervjustudie om de allra äldstas levnadsvillkor. Äldre har fått
fler hälsoproblem, men klarar vardagen bättre [Interview study
on the living conditions of the very old. Elderly acquire more
health problems, but they manage everyday life better].
Läkartidningen 2013;110:CA33:1403–05.
14. Meinow B, Parker MG, Kåreholt I, Thorslund M. Complex
health problems in the oldest old in Sweden 1992–2002. Eur J
Ageing 2006;3:98–106.
15. Agahi N, Parker MG. Are today’s older people more active than
their predecessors? Participation in leisure-time activities in
Sweden in 1992 and 2002. Ageing and Society 2005;25:925–42.
16. Parker MG, Thorslund M, Lundberg O. Physical function and
social-class among Swedish oldest-old. J Gerontol
1994;49:S196–S201.
17. Thorslund M, Lundberg O. Health and inequalities among the
oldest old. J Aging Health 1994;6:51–69.
738
18. Lennartsson C, Lundberg O. What’s marital status got to do with
it? Gender inequalities in economic resources, health and functional abilities among older adults. In: Fritzell J, Lundberg O (eds).
Health Inequalities and Welfare Resources. Bristol, UK: Policy
Press, 2007.
19. Fors S, Lennartsson C, Lundberg O. Health inequalities among
older adults in Sweden 1991–2002. Eur J Public Health
2008;18:138–43.
20. Haider S, Johnell K, Thorslund M, Fastbom J. Trends in polypharmacy and potential drug-drug interactions across educational groups in elderly patients in Sweden for the period 19922002. Int J Clinical Pharmacol Ther 2007;45:643–53.
21. Schön, P. Gender Matters. Differences and change in disability
and health among our oldest women and men. Department of
Social Work, Stockholm University, 2011.
22. Schön P, Parker M. Sex differences in health in 1992 and 2002
among very old Swedes. J Popul Ageing 2009;1:107–23.
23. Andel R, Kåreholt I, Parker M, Thorslund M, Gatz M.
Complexity of primary lifetime occupation and cognition in
advanced old age. J Aging Health 2007;19:397–415.
24. Andel R, Crowe M, Kareholt I, Wastesson J, Parker MG.
Indicators of job strain at midlife and cognitive functioning in
advanced old age. J Gerontol B Psychol Sci Soc Sci
2011;66:287–91.
25. Parker V, Andel R, Nilsen C, Kareholt I. The association between mid-life socioeconomic position and health after retirement – exploring the role of working conditions. J Aging Health
2013;25:863–81.
26. Agahi N, Parker MG. Leisure activities and mortality: does gender matter? J Aging Health 2008;20:855–71.
27. Lennartsson C, Silverstein M. Does engagement with life enhance survival of elderly people in Sweden? The role of social
and leisure activities. J Gerontol B Psychol Sci Soc Sci
2001;56:S335–42.
28. Kareholt I, Lennartsson C, Gatz M, Parker MG. Baseline leisure
time activity and cognition more than two decades later. Int J
Geriatr Psychiatry 2011;26:65–74.
International Journal of Epidemiology, 2014, Vol. 43, No. 3
29. Lennartsson C. Social ties and health among the very old in
Sweden. Res Aging 1999;21:657–81.
30. Fors S, Lennartsson C, Lundberg O. Childhood living conditions, socioeconomic position in adulthood, and cognition in
later life: exploring the associations. J Gerontol B Psychol Sci
Soc Sci 2009;64:750–57.
31. Agahi N, Shaw BA. Smoking trajectories from midlife to old
age and the development of non-life-threatening health
problems: A 34-year prospective cohort study. Prev Med
2013;57:107–12.
32. Kåreholt I. The long shadow of socioeconomic conditions in
childhood: do they affect class inequalities in mortality. In
Jonsson JO Mills C (eds). Cradle to Grave. Life Course Change
in Modern Sweden. York, UK: Sociology Press, 2001.
33. Fors S, Lennartsson C, Lundberg O. Live long and prosper?
Childhood living conditions, marital status, social class in adulthood and mortality during mid-life: A cohort study. Scand J
Public Health 2011;39:179–86.
34. Heap J, Lennartsson C, Thorslund M. Coexisting disadvantages
across the adult age span: A comparison of older and younger
age groups in the Swedish welfare state. Int J Soc Welfare
2012;22:130–40.
35. Agahi N, Ahacic K, Parker MG. Continuity of leisure participation from middle age to old age. J Gerontol B Psychol SciSoc Sci
2006;61:S340–46.
36. Lennartsson C, Silverstein M, Fritzell J. Time-for-money exchanges between older and younger generations in Swedish families. J Fam Issues 2010;31:189–210.
37. Ahacic K, Kennison FR, Kareholt, I. Changes in sobriety in the
Swedish population over three decades: age, period or cohort effects? Addiction 2012;107:748–55.
38. Ahacic K, Kennison R, Thorslund M. Trends in smoking in
Sweden from 1968 to 2002: age, period, and cohort patterns.
Prev Med 2008;46:558–64.
39. Ahacic K, Parker MG, Thorslund M. Aging in disguise: age,
period and cohort effects in mobility and edentulousness over
three decades. Eur J Ageing 2007;4:83–91.