Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The Association of Levels of and Decline in Grip Strength in Old Age with Trajectories of Life Course Occupational Position

The Association of Levels of and Decline in Grip Strength in Old Age with Trajectories of Life Course Occupational Position

  • Hannes Kröger, 
  • Johan Fritzell, 
  • Rasmus Hoffmann



The study of the influence of life course occupational position (OP) on health in old age demands analysis of time patterns in both OP and health. We study associations between life course time patterns of OP and decline in grip strength in old age.


We analyze 5 waves from the Survey of Health Ageing and Retirement in Europe (n = 5108, ages 65–90). We use a pattern-mixture latent growth model to predict the level and decline in grip strength in old age by trajectory of life course OP. We extend and generalize the structured regression approach to establish the explanatory power of different life course models for both the level and decline of grip strength.


Grip strength declined linearly by 0.70 kg (95% CI -0.74;-0.66) for men and 0.42 kg (95% CI -0.45;-0.39) for women per year. The level of men’s grip strength can best be explained by a critical period during midlife, with those exposed to low OP during this period having 1.67 kg (95% CI -2.33;-1.00) less grip strength. These differences remain constant over age. For women, no association between OP and levels of or decline in grip strength was found.


Men’s OP in midlife seems to be a critical period for the level of grip strength in old age. Inequalities remain constant over age. The integration of the structured regression approach and latent growth modelling offers new possibilities for life course epidemiology.


Life course research in social epidemiology has often considered either time-patterns of socio-economic position (SEP) as predictors for static measurements of physical function, or static measurements of SEP as determinants of levels and change in physical function in old age. In the first type of study, the objective is often to identify the influence of different trajectories of SEP on health in later life. Many studies try to distinguish between models of accumulation, critical periods, or social mobility. While some authors have argued that this distinction might never be possible[1], other approaches have been developed that treat discriminating between models as an empirical problem[26]. The accumulation model proposes that in each period in the life course risk factors can influence health, and that resulting health inequalities accumulate over time. The critical or sensitive period models stress the importance of a particular time window in the life course (often childhood) as the main or even sole period in which SEP exacts an influence on health later in life[7]. Whereas the stronger version of this model assumes irreversibility, the softer version, often referred to as sensitive period model, note that later events can modify the effects of the earlier exposure. Alternatively, social mobility models stress possible health effects of both inter- and intra-generational upwards or downwards social mobility. These studies take a dynamic perspective on SEP or other risk factors, but predominantly a static perspective on the health outcome [812]. The second type of studies often use indicators from a certain point in time, or time-constant indicators of SEP, to predict trajectories of health[1315] Moreover, few of these studies focus on health at old age. If the developments are investigated with respect to health inequalities, three general scenarios can be expected. The age-as-leveler hypothesis predicts decreasing inequalities with higher age due to selective mortality, lack of further exposure to poor working conditions, and the overriding influence of biological aging[16,17]. Proponents of cumulative (dis)advantage would expect existing inequalities to increase as certain factors like health behavior or living conditions continue to work as drivers of health inequality in old age[18]. Lastly, it is possible that health inequalities remain largely stable throughout old age.

We endeavor to combine these two perspectives to investigate the association between trajectories in occupational position (OP) throughout the life course and levels and decline of grip strength. This combined approach yields a more comprehensive picture of the interplay between trajectories of OP and physical function in old age[19]. We estimate the association between levels of and decline in grip strength with life course OP for all possible trajectories separately, and test patterns in these trajectories according to the established models of accumulation, critical periods, or social mobility. For this purpose, we generalize the structured regression approach[20] to the framework of structural equation modeling (SEM) to be compatible with the prediction of differences in intercept and slope of a latent growth model (LGM).

Grip strength has become a popular indicator of physical functioning in surveys. It is both indicative of overall muscle and physical functioning [21]. Physical functioning in old age is an important prerequisite for independence, quality of life and for avoiding comorbidities due to inhibited mobility [22]. Grip strength is objectively measured, avoiding biases that might arise in self-reports. It is further predictive of mortality, showing that it is related to health status more generally. Further, its measurement has no relevant floor or ceiling effects. This means that improvement or worsening of the indicator is possible at almost all levels of the measurement. This is especially important if individual decline is to be measured and a great advantage over scales of physical functioning like activities of daily living (ADL) in surveys which do have floor and ceiling effects.

Our study contributes to life course epidemiology: firstly, by integrating theories on life course models of exposure and change in health inequalities; secondly, by providing empirical results from a large European dataset; and thirdly, by illustrating a novel combination of the structured regression approach with LGM.

Materials and Methods

The analysis is based on anonymized secondary data. The SHARE survey is subject to ethical approval of Ethics Council of the Max-Planck-Society for the Advancement of Science.

The Survey of Health Ageing and Retirement in Europe consists of data on health and socioeconomic variables of non-institutionalized individuals aged 50 and older across 20 European countries[23]. We use waves 1–5, collected bi-annually between 2004 and 2013. In the third wave, retrospective life course data was collected on OP from childhood to old age[24].

The total sample of the SHARE respondents was restricted in the following way. Firstly, only participants of wave three who answered the life history questionnaire were retained. Secondly, all those who reported never having been employed, or who had missing information on childhood or adulthood OP, were excluded. The third restriction was that all individuals had to be between 65 and 90 (birth cohorts 1922 to 1938) and no longer working during the period of observation. The resulting sample consists of 3067 men and 2041 women in 13 countries taking part in wave three (Austria, Germany, Sweden, Netherlands, Spain, Italy, France, Denmark, Greece, Switzerland, Belgium, Czech Republic, Poland).


We dichotomized OP, separating the group of elementary occupation (ISCO-88 major group 9) from all others. Workers in this group perform mostly simple and routine tasks, in some cases with considerable physical effort[25]. We defined three life course periods: Childhood (0–15), young adulthood (16–35), and midlife (36–64). The cut-off between young adulthood and midlife was chosen at 35, because grip strength peaks around the mid-thirties[26,27]. The indicator of exposure to low OP for childhood is the OP of the main breadwinner at age ten. For young adulthood and midlife periods the indicator refers to their own occupational status. We coded as ‘exposed to low OP’ all individuals who had worked in an elementary occupation for at least half of the years in which they reported an occupational status.

The highest grip strength measurement of two measurements per hand (dynamometer type: Smedley, S Dynamometer, TTM, Tokyo, 100kg)[28] is used as an indicator of grip strength, measured in all five waves. Grip strength has been shown to be a predictor of disability[22,2931], morbidity[21,29,32], and mortality [3336], and a correlate of other aspects of physical aging like frailty[37]. Furthermore, different dimensions of SEP predict grip strength[3840], making it a useful indicator for grip strength in old age.

Grip strength is not observed for all individuals at each wave. We divided the sample into five missing value patterns (Table 1) and applied a pattern-mixture (PM) model [41] to correct for possible bias in the estimation of the LGM due to health related drop-outs or item-non-response[42]. This includes a correction for those who drop out due to death and those who are no longer able to perform the grip strength measurement. The means of the intercept and slope, and their association with life course occupational status, are estimated separately for these five groups. We also reran the analyzes using the more restrictive missing at random (MAR) assumption. The conclusions in our paper are not affected by the decision to adopt a PM over a MAR approach. We report the PM results in the main paper, because they rely on less strict assumptions than the MAR approach. We document the core results of MAR in Tables C, P and Q in S1 Appendix.

Table 1. Missing value patterns and participation in waves in the sample–Frequency (%).

In our model we let the initial level of grip strength (intercept) and the decline in grip strength (slope) be correlated. Positive values of the correlation indicate that those with higher initial levels of grip strength have a stronger decline, while negative values mean that initially high levels of grip strength are associated with slower decline. This means that all parameter estimates and comparisons of models in our analyses take this potential association into account and can be interpreted as after accounting for possible correlation of initial level and subsequent decline in grip strength.

All data preparation and summary statistics were conducted using Stata 14.1 with user written extensions [43,44]. The whole code necessary to replicate the analyses is available in S1 Code.

Statistical analysis

We proceed in three steps in our analysis. First, we estimate the overall level of and decline in grip strength using LGM. We use month-specific age at each wave as an individually varying time-point[45]. The LGM is defined in the following way: (1)

The index i refers to the observed individuals. y is the vector of the five observed measures of grip strength, Λ is the vector of constraints, identifying intercept and slope. ϵi represents the vector of individual specific errors in grip strength. ηi contains the values of each individual on the intercept and slope parameter, μη holds the means of the intercept and slope, representing level and decline in physical functioning, and ζi is the individual deviation from the mean of the intercept and slope, representing the variability in level and decline of physical functioning.

Second, we estimate the association between life course OP trajectories and both level and decline of grip strength. In our model we include dummy variables for each of the countries and dummies for three year cohorts. The coefficients of these dummies reflect differences in initial level and decline of grip strength between cohorts and countries. Therefore, our estimates can be interpreted as differences in level and decline of grip strength between different trajectories of occupational position, after adjusting for differences between countries and cohorts.

Third, we use the structured regression approach (SRA), which was developed to distinguish between patterns of life course exposure according to the theories of accumulation, critical and sensitive periods, and social mobility[20]. We generalize the original approach so that it can be applied in a SEM framework for the prediction of level and decline of grip strength.

The intercept (level) and slope (decline) of grip strength are predicted by exposure in the three periods of life course and all their interactions (Xi) and controlled for differences between countries and cohorts (Ci): (2)

In the SRA, a saturated model is defined as consisting of the freely estimated effects of all periods of exposure (and all their possible interactions) on grip strength. This estimates different levels of the outcome variable for every possible trajectory of OP, yielding maximum explanatory power. If a significant association with trajectories of OP can be found in the saturated model, a set of restrictions corresponding to life course models is applied to the saturated model, and the relative model fit is assessed. For testing constraints on the coefficients predicting both levels of and decline in grip strength, we use Wald tests, which can be easily implemented into SEM, instead of the original F-test. Following a further development of the SRA, we compare the Akaike information criterion (AIC) of those models that show a p-value of over 0.1[46]. (For more technical details of the models, see S1 Appendix).


Table 2 reports the sample statistics for men and women. In the first step, we estimated LGMs for men and women to determine the shape of decline of grip strength[45]. We compared the model fit of three specifications of the growth trajectory (linear, quadratic, linear-semi-parametric). Fig 1 shows the fit of predictions of the three model specifications against the actually observed decline (locally weighted scatterplot smoothing), and Table 3 shows the BIC for the respective models. A linear decline models the trajectory of grip strength in our data set well. For men, we can observe an average grip strength of 37.54 kg (95% CI 37.28;37.80) at age 75, and an average estimated decline of 0.70 kg (95% CI -0.74;-0.66) between the ages of 65 and 90.

Fig 1. Model fit of three specifications of the slope in latent growth model of grip strength for men and women.

Note: Figure shows the predictions of three specifications compared to the observed trajectory of grip strength estimated by locally weighted scatterplot smoothing (LOESS).

Table 3. Model fit (BIC) of three specifications of the slope for the latent growth model of grip strength.

At the age of 75, women have an average grip strength of 23.60 kg (95% CI 23.38;23.82). During the period of observation, the annual decrease is estimated to be 0.42 kg (95% CI -0.45;-0.39). The correlation between intercept and decline in grip strength is weak for both men (0.04; 95% CI -0.09;0.17) and women (-0.11; 95% CI -0.28; 0.07), and not statistically significant (for more detailed results see Table B in S1 Appendix). This means that the initial level of grip strength is not associated with rate of decline for men or women.

In the second step, we use life course trajectories of OP to predict the average level and the annual decline of grip strength. Fig 2 shows these predictions. The figure uses a 1 for exposure to low OP and a 0 for no exposure. Tables 4 and 5 report the respective predicted intercept and slope parameters for each of the trajectories of OP. Trajectory (101) is not plotted due to the low number of cases that make the prediction unreliable. For men, we can see that the lines run mostly parallel, although there are slight differences between the levels. A slight exception is trajectory (010), those who are exposed to low OP only in young adulthood. They are predicted to have a higher decline in physical health. From Table 4 we can see that men who achieve intergenerational upward mobility (100) have, on average, the highest grip strength, followed by those with intra-generational mobility (110) and those who are never exposed (000). The lowest levels of grip strength are found in those of opposite occupational trajectory: the inter-generationally downwardly mobile (011), the continuously exposed (111), and those who are intra-generationally downwardly mobile (001). The difference to the three highest groups is about 2 kg, which is substantial as it translates into a difference of 3 years’ decline in physical function.

Fig 2. Model implied predictions of developments of grip strength by OP trajectory for men and women.

Note: Trajectory (101) was not plotted, because the number of observations was too small to yield reliable predictions. Trajectories are described by 1 for exposure to low OP, and 0 for no exposure. The first digit indicates the status for childhood, the second for early adulthood and the third digit represents midlife. For example, 000 means always is high OP, 111 always in low OP, 001 represents downward social mobility in adulthood.

Table 4. Predictions of intercept and slope of grip strength by life course OP pattern (kg)–Men.

Table 5. Predictions of intercept and slope of grip strength by life course OP pattern (kg)–Women.

In the lower part of Fig 2 and in Table 5 we can see that both slopes and levels of grip strength are much closer together for women than for men. This indicates only small differences related to life course OP. The only trajectories that have notably higher levels of grip strength are (110) and (010), containing those who are intra-generationally upwardly mobile and those who are exposed only in young adulthood, respectively. Note, however, that these categories contain only few women.

In the third step of our analysis, we conducted Wald tests on the restrictions of the coefficients that reflect the different life-course models. The technical description of these restrictions can be found in Table A of S1 Appendix. For the level of grip strength for men, the best model is a critical period in midlife showing the highest p-value and lowest AIC (Tables 6 and 7). Based on this model, those who are in a low OP during midlife are predicted to have 1.67 kg (95% CI -2.33;-1.00) less grip strength in old age. Table 6 shows further that there is no significant association between life course patterns of OP and the slope of grip strength for men in the model, as the null model cannot be rejected. That means that decline in health does not differ systematically across trajectories of OP. There is no increase or decrease in differences; health inequalities remain constant.

For women, the application of the life course model tests confirms the impression given by Fig 2. Both the levels of and decline in physical health are not related systematically to women’s OP trajectory (Table 6).

To check the sensitivity of our results to the coding of the exposure to low OP variable, we reran the analyses, with exposure defined as a blue-collar occupation (ISCO major groups 6–9) versus all white-collar occupations yielding very similar results (Tables D-G in S1 Appendix). In addition, we also reran our models to include those individuals with missing information for 1 or 2 periods. The results remain stable (Tables L–O in S1 Appendix). As additional sensitivity checks, we controlled separately for height and weight, which also did not change our results (Tables H-K in S1 Appendix).


In this study we investigated the influence of low OP over the life course on levels and decline in grip strength at old age. For men, the results of the generalized structured regression approach suggest that exposure to low OP is especially harmful during midlife. For women, no relevant association to life course OP could be found.

For men, the results stand in contrast to other studies which stress the importance of early life exposure to low OP and different dimensions of health[15,4751], because taking the different trajectories into account did not reveal a relevant influence of parental OP during childhood on grip strength in old age. However, it should be noted that there are also other studies which do not find an association of childhood OP and grip strength[52].

We found that the difference in physical function between those exposed during peak working age and those unexposed was already established at the age of 65, and our analysis provided no evidence for any of the life course theories suggesting convergence (age-as-leveler) or divergence (cumulative disadvantage) after that age. One possible explanation might be that exposure to low OP reduces maximum attainable strength (functional reserve)[53], but does not affect the rate of decline in old age. An argument against this proposition is that studies have shown that maximum grip strength is usually reached before the age of 35[26], which is before the critical period of exposure. Consequently, it seems more likely that there is differential decline in grip strength during the period of later working life, favoring those who are not exposed. To the extent that our indicator of OP capture more directly working life experience our results then could actually be seen as being in line with the cumulative disadvantage since those earlier exposed are now retired from the labour market and therefore not directly exposed. More importantly is to stress that our findings with similar decline from different levels of grip strength imply that inequalities in physical function continues also at advanced old age. Both in research and practice, it is consequently important to highlight the heterogeneity in old-age health and the impact of social stratification.

Previous studies for adult and old-age health have found mixed support for the gender difference in influence of OP on physical function; some find a similar association for men and women[39,54], some find weaker or no association for women[38], as in our study. The clear gender difference found in our study stresses the importance of gender-specific analysis, because both trajectories and their health consequences might be different for men and women[55].

One explanation for the lack of association between life course OP and grip strength in old age for women is that the women in the cohorts under investigation who are continuously active on the labor market represent a health-selected population. When the male breadwinner model is dominant, women with health limitations are more likely to drop out of employment. This might suppress a possible association between their life course OP and health in old age. Accordingly, a study adopting the household as the unit to define OP, what is normally called the dominance approach, might have led to a different finding for women [56,57].

Strengths and limitations

In our study we combined a dynamic view on OP over the life course with a dynamic view on inequalities in physical function in old age. This gives a more comprehensive perspective on differences in aging by jointly addressing life course models of accumulation, critical period, and social mobility with theories on the development of health inequalities such as age-as-leveler and cumulative advantage. We demonstrated that the structured regression approach can be generalized to a SEM framework, allowing more flexible tests of life course models. We made use of a large representative data set that collected grip strength as an objective indicator of grip strength over five points in time. This allowed us to combine life course information on OP with estimated of the trajectories in grip strength which is rarer due to the restricted number of data sets containing information on both aspects. We took differences in likelihood of drop-out due to poor health (and death) and low grip strength into account by modeling separately for five patterns of missing values. Our study demonstrates the need and the potential to integrate different strands of theory on socially stratified processes of aging with the appropriate methods developed in different fields of longitudinal and life course research.

A recent study proposed an alternative strategy of establishing the explanatory power of life course models[58] for which an integration with LGM might be useful in future research.

Despite these strengths there are several limitations to our study. Occupational position reflects only one dimension of socioeconomic position (SEP) of individuals. Usually, income, education, and wealth are treated as other important dimensions of SEP. It has been shown that these different dimension can have different impact on health and health inequalities in the life course [59,60] and on the trajectories of different health indicators, including functional limitations [61]. Therefore, our results should not be generalized to all aspects of SEP. Instead further research could look at trajectories in other dimensions. The limitations are that education is usually time-constant at a certain age, early in the life of individuals. Wealth on the other hand is by definition the outcome of a cumulative process, increasing for most individuals throughout the life course. Classical upward or downward mobility patterns therefore do not apply to this dimension. Last, income could be analyzed in a similar fashion, but here data availability is the problem as information about income over the whole life course (including parental income) is still hard to acquire in combination with old age health outcomes. Improvements in survey data and cohort studies might allow replications of our analyses with trajectories of income as the determinant of physical functioning in old age.

The dichotomization is on the one hand necessary to ensure that the trajectories do not become overly complex. On the other hand using a dichotomous indicator will hide a lot of variation within the categories. However, we showed that a different coding into blue-collar and white-collar workers did not change the results. Comparative work with alternative indicators of OP should be added in the future. It should be further noted that by definition the use of an occupational indicator limits the analyses to the employed population. Alternative indicators of SEP like trajectories of household income could be used in future studies to get estimates for the non-employed as well.

The use of retrospective data has several advantages and disadvantages for our study. Despite the fact that SHARE uses a life grid approach that is designed to maximize accuracy in remembering occurrence and temporal ordering of events in the life course[24,62], retrospective data faces the problem of incorrect recall of occupational status[6365]. In contrast to many prospective cohort studies, we have continuous yearly information on occupational status from age 16 onwards. That allows us to average occupational status during adulthood and midlife respectively, which should reduce measurement error and recall bias. Another advantage of retrospective data is that there are no drop-outs during the observation of OP in the life course[6668]. We face the problem of increasing selectivity only to a small degree, as there are only 2 waves in which drop-out can occur before OP is measured for the whole life course. Still, it needs to be acknowledged that we might have a positively-selected sample with regard to (decline in) health if those who dropped out before answering the retrospective questionnaire in wave 3 have lower health overall, or a stronger health decline. This might lead to an underestimation of the association between life course occupational position with grip strength in our study. Additionally, our sample only includes survivors to old age. This means that mortality in early and midlife can lead to a reduction in health inequalities that we do not observe. Thus, the results only apply to those who survived until the age of 65 and should not be generalized to younger ages. A further problem might arise if childhood OP has a higher degree of error than midlife OP. In this case the relative strength of the association of childhood OP with grip strength might be underestimated compared to the association of midlife OP with grip strength.

One possible concern with using grip strength related to an occupational indicator is that workers in elementary occupations have physically more demanding jobs, which could lead to a training effect of their muscles. After retirement they could therefore have a higher decline as they no longer engage in physical labor, which would reflect detraining, and not decline in grip strength. We do not believe that this constitutes a source of bias. Firstly, the literature on detraining shows that detraining happens very quickly, i.e. training effects vanish almost completely after half a year, often earlier, depending on the training treatment[6971]. The sample consists of those who are no longer working, and therefore any possible detraining will, for the most part, have happened before the observation begins. Second, there is little evidence that suggests that blue-collar workers or workers who have a high physical workload have higher grip strength than white-collar workers[26,7275], a result that we can replicate with our data for those individuals below 65 who are still active on the labor market. Finally, it is important to note that the results cannot be interpreted as causal estimates in the counter-factual sense. It is possible that there are common factors for both trajectories in occupational position and level and development of grip strength in old age (e.g. physiological dispositions acquired very early in life). The results should be regarded as associations representing the total effect of life course OP on grip strength, including direct and mediated associations, and possible selection effects during the life course.


Combining a dynamic perspective on both life course OP and grip strength in old age provides a good view on the pattern of health inequalities in old age. For men, mid-life exposure to low OP correlates with decline in grip strength. No further convergence or divergence could be found during old age. As grip strength is a reliable indicator of other aspects of physical functioning we would expect to see similar results in the association of life course OP and other indicators of physical functioning. Furthermore, it is important to look at socioeconomic trajectories of men and women separately. Our extension of the structured regression approach to SEM and LGM can be used in future research in life course epidemiology.

Supporting Information

S1 Appendix. Technical details on models and results of sensitivity analyses.


S1 Code. Code to replicate all analyses conducted in the study.



The authors would like to thank two anonymous reviewers for their helpful comments on an earlier version of the paper.

Author Contributions

Analyzed the data: HK. Wrote the paper: HK RH JF.


  1. 1. Hallqvist J, Lynch J, Bartley M, Lang T, Blane D. Can we disentangle life course processes of accumulation, critical period and social mobility? An analysis of disadvantaged socio-economic positions and myocardial infarction in the Stockholm Heart Epidemiology Program. Soc Sci Med. 2004;58.
  2. 2. Otero-Rodríguez A, León-Muñoz LM, Banegas JR, Guallar-Castillón P, Rodríguez-Artalejo F, Regidor E. Life-course socioeconomic position and change in quality of life among older adults: evidence for the role of a critical period, accumulation of exposure and social mobility. J Epidemiol Community Health. 2011;65: 964–971. pmid:20974837
  3. 3. Heraclides A, Brunner E. Social mobility and social accumulation across the life course in relation to adult overweight and obesity: the Whitehall II study. J Epidemiol Community Health. 2010;64: 714–719. pmid:19737739
  4. 4. Niedzwiedz CL, Katikireddi SV, Pell JP, Mitchell R. Life course socio-economic position and quality of life in adulthood: a systematic review of life course models. BMC Public Health. 2012;12: 628. pmid:22873945
  5. 5. Galobardes B, Lynch JW, Smith GD. Childhood Socioeconomic Circumstances and Cause-specific Mortality in Adulthood: Systematic Review and Interpretation. Epidemiol Rev. 2004;26: 7–21. pmid:15234944
  6. 6. Fritzell J. Life course inequalities: generations and social class. In: Fritzell J, Lundberg O, editors. Health Inequalities and Welfare Resources: Continuity and Change in Sweden. Bristol: Policy Press; 2007. pp. 67–86.
  7. 7. Ben-Shlomo Y, Kuh D. A life course approach to chronic disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives. Int J Epidemiol. 2002;31: 285–293. pmid:11980781
  8. 8. Kapetanakis VV, Rudnicka AR, Wathern AK, Lennon L, Papacosta O, Cook DG, et al. Adiposity in Early, Middle and Later Adult Life and Cardiometabolic Risk Markers in Later Life; Findings from the British Regional Heart Study. PLOS ONE. 2014;9.
  9. 9. Muniz de Quadros L de C, Quevedo L de A, Motta S, Carraro A, Ribeiro FG, Horta BL, et al. Social Mobility and Mental Disorders at 30 Years of Age in Participants of the 1982 Cohort, Pelotas, Rio Grande Do Sul—RS. PLOS ONE. 2015;10.
  10. 10. Park MH, Sovio U, Viner RM, Hardy RJ, Kinra S. Overweight in Childhood, Adolescence and Adulthood and Cardiovascular Risk in Later Life: Pooled Analysis of Three British Birth Cohorts. PLOS ONE. 2013;8.
  11. 11. Sovio U, Giambartolomei C, Kinra S, Bowen L, Dudbridge F, Nitsch D, et al. Early and current socio-economic position and cardiometabolic risk factors in the Indian Migration Study. Eur J Prev Cardiol. 2013;20: 844–853. pmid:22514214
  12. 12. Gustafsson PE, Persson M, Hammarström A. Socio-economic disadvantage and body mass over the life course in women and men: results from the Northern Swedish Cohort. Eur J Public Health. 2011; ckr061.
  13. 13. Sacker A, Worts D, McDonough P. Social influences on trajectories of self-rated health: evidence from Britain, Germany, Denmark and the USA. J Epidemiol Community Health. 2009;65: 130–136. pmid:19996360
  14. 14. Østbye T, Malhotra R, Landerman LR. Body mass trajectories through adulthood: results from the National Longitudinal Survey of Youth 1979 Cohort (1981–2006). Int J Epidemiol. 2011;40: 240–250. pmid:20819785
  15. 15. Haas S. Trajectories of functional health: the “long arm”of childhood health and socioeconomic factors. Soc Sci Med. 2008;66: 849–861. pmid:18158208
  16. 16. Hoffmann R. Illness, not age, is the leveler of social mortality differences in old age. J Gerontol B Psychol Sci Soc Sci. 2011;66: 374–379. pmid:21430089
  17. 17. Liang J, Bennett J, Krause N, Kobayashi E, Kim H, Brown JW, et al. Old Age Mortality in Japan Does the Socioeconomic Gradient Interact With Gender and Age? J Gerontol B Psychol Sci Soc Sci. 2002;57: S294–S307. pmid:12198109
  18. 18. Willson AE, Shuey KM, Glen H. Elder J. Cumulative Advantage Processes as Mechanisms of Inequality in Life Course Health. Am J Sociol. 2007;112: 1886–1924.
  19. 19. McDonough P, Sacker A, Wiggins RD. Time on my side? Life course trajectories of poverty and health. Soc Sci Med. 2005;61: 1795–1808. pmid:16029777
  20. 20. Mishra G, Nitsch D, Black S, Stavola BD, Kuh D, Hardy R. A structured approach to modelling the effects of binary exposure variables over the life course. Int J Epidemiol. 2009;38: 528–537. pmid:19028777
  21. 21. Cooper R, Kuh D, Cooper C, Gale CR, Lawlor DA, Matthews F, et al. Objective measures of physical capability and subsequent health: a systematic review. Age Ageing. 2011;40: 14–23. pmid:20843964
  22. 22. Vermeulen J, Neyens JC, van Rossum E, Spreeuwenberg MD, de Witte LP. Predicting ADL disability in community-dwelling elderly people using physical frailty indicators: a systematic review. BMC Geriatr. 2011;11: 33. pmid:21722355
  23. 23. Börsch-Supan A, Brandt M, Hunkler C, Kneip T, Korbmacher J, Malter F, et al. Data Resource Profile: The Survey of Health, Ageing and Retirement in Europe (SHARE). Int J Epidemiol. 2013;42: 992–1001. pmid:23778574
  24. 24. Börsch-Supan A, Schröder M. Retrospective Data Collection in the Survey of Health, Ageing and Retirement in Europe. SHARELIFE Methodol. 2011;5.
  25. 25. International Labour Office. International standard classification of occupations: ISCO-88. Geneva: International Labour Office; 1990.
  26. 26. Günther CM, Bürger A, Rickert M, Crispin A, Schulz CU. Grip Strength in Healthy Caucasian Adults: Reference Values. J Hand Surg. 2008;33: 558–565.
  27. 27. Dodds RM, Syddall HE, Cooper R, Benzeval M, Deary IJ, Dennison EM, et al. Grip Strength across the Life Course: Normative Data from Twelve British Studies. PLoS ONE. 2014;9: e113637. pmid:25474696
  28. 28. Andersen-Ranberg K, Petersen I, Frederiksen H, Mackenbach JP, Christensen K. Cross-national differences in grip strength among 50+ year-old Europeans: results from the SHARE study. Eur J Ageing. 2009;6: 227–236.
  29. 29. Kuh D, Bassey EJ, Butterworth S, Hardy R, Wadsworth MEJ, Team and the MS. Grip Strength, Postural Control, and Functional Leg Power in a Representative Cohort of British Men and Women: Associations With Physical Activity, Health Status, and Socioeconomic Conditions. J Gerontol A Biol Sci Med Sci. 2005;60: 224–231. pmid:15814867
  30. 30. Rantanen T, Guralnik JM, Foley D, et al. Midlife hand grip strength as a predictor of old age disability. JAMA. 1999;281: 558–560. pmid:10022113
  31. 31. den Ouden MEM, Schuurmans MJ, Arts IEMA, van der Schouw YT. Physical performance characteristics related to disability in older persons: A systematic review. Maturitas. 2011;69: 208–219. pmid:21596497
  32. 32. Nybo H, Gaist D, Jeune B, McGue M, Vaupel JW, Christensen K. Functional Status and Self-Rated Health in 2,262 Nonagenarians: The Danish 1905 Cohort Survey. J Am Geriatr Soc. 2001;49: 601–609. pmid:11380754
  33. 33. Gale CR, Martyn CN, Cooper C, Sayer AA. Grip strength, body composition, and mortality. Int J Epidemiol. 2007;36: 228–235. pmid:17056604
  34. 34. Sasaki H, Kasagi F, Yamada M, Fujita S. Grip strength predicts cause-specific mortality in middle-aged and elderly persons. Am J Med. 2007;120: 337–342. pmid:17398228
  35. 35. Martin-Ruiz C, von Zglinicki M. A life course approach to biomarkers of ageing. In: Kuh D, Cooper R, Hardy R, Richards M, Ben-Shlomo Y, editors. A Life Course Approach to Healthy Ageing. New York: Oxford University Press; 2014. pp. 177–186.
  36. 36. Cooper R, Kuh D, Hardy R, Group MR. Objectively measured physical capability levels and mortality: systematic review and meta-analysis. BMJ. 2010;341: c4467. pmid:20829298
  37. 37. Syddall H, Cooper C, Martin F, Briggs R, Sayer AA. Is grip strength a useful single marker of frailty? Age Ageing. 2003;32: 650–656. pmid:14600007
  38. 38. Hairi FM, Mackenbach JP, Andersen-Ranberg K, Avendano M. Does socio-economic status predict grip strength in older Europeans? Results from the SHARE study in non-institutionalised men and women aged 50+. J Epidemiol Community Health. 2010;64: 829–837. pmid:19884112
  39. 39. Hurst L, Stafford M, Cooper R, Hardy R, Richards M, Kuh D. Lifetime Socioeconomic Inequalities in Physical and Cognitive Aging. Am J Public Health. 2013;103: 1641–1648. pmid:23865666
  40. 40. Russo A, Onder G, Cesari M, Zamboni V, Barillaro C, Capoluongo E, et al. Lifetime occupation and physical function: a prospective cohort study on persons aged 80 years and older living in a community. Occup Environ Med. 2006;63: 438–442. pmid:16782827
  41. 41. Enders CK. Missing not at random models for latent growth curve analyses. Psychol Methods. 2011;16: 1–16. pmid:21381816
  42. 42. Frederiksen H, Hjelmborg J, Mortensen J, Mcgue M, Vaupel JW, Christensen K. Age trajectories of grip strength: cross-sectional and longitudinal data among 8,342 Danes aged 46 to 102. Ann Epidemiol. 2006;16: 554–562. pmid:16406245
  43. 43. Kröger H. Newspell—Easy management of complex spell data. Stata J. 2015;15: 155–172.
  44. 44. Jann B. Making regression tables simplified. Stata J. 2007;7: 227.
  45. 45. Sterba SK. Fitting Nonlinear Latent Growth Curve Models With Individually Varying Time Points. Struct Equ Model Multidiscip J. 2014;21: 630–647.
  46. 46. Mishra GD, Chiesa F, Goodman A, Stavola BD, Koupil I. Socio-economic position over the life course and all-cause, and circulatory diseases mortality at age 50–87 years: results from a Swedish birth cohort. Eur J Epidemiol. 2013;28: 139–147. pmid:23435736
  47. 47. Kuh D, Bassey J, Hardy R, Sayer AA, Wadsworth M, Cooper C. Birth Weight, Childhood Size, and Muscle Strength in Adult Life: Evidence from a Birth Cohort Study. Am J Epidemiol. 2002;156: 627–633. pmid:12244031
  48. 48. Sayer AA, Syddall HE, Gilbody HJ, Dennison EM, Cooper C. Does Sarcopenia Originate in Early Life? Findings From the Hertfordshire Cohort Study. J Gerontol A Biol Sci Med Sci. 2004;59: M930–M934. pmid:15472158
  49. 49. Hayward MD, Gorman BK. The long arm of childhood: The influence of early-life social conditions on men’s mortality. Demography. 2004;41: 87–107. pmid:15074126
  50. 50. Case A, Fertig A, Paxson C. The lasting impact of childhood health and circumstance. J Health Econ. 2005;24: 365–389. pmid:15721050
  51. 51. Starr JM, Deary IJ. Socio-economic position predicts grip strength and its decline between 79 and 87 years: the Lothian Birth Cohort 1921. Age Ageing. 2011;40: 749–752. pmid:21705769
  52. 52. Birnie K, Cooper R, Martin RM, Kuh D, Sayer AA, Alvarado BE, et al. Childhood Socioeconomic Position and Objectively Measured Physical Capability Levels in Adulthood: A Systematic Review and Meta-Analysis. PLoS ONE. 2011;6: e15564. pmid:21297868
  53. 53. Kuh D, Richards M, Cooper R, Hardy R, Ben-Shlomo Y. Life course epidemiology, ageing research and maturing cohort studies: a dynamic combination for understanding healthy ageing. In: Kuh D, Cooper R, Hardy R, Richards M, Ben-Shlomo Y, editors. A Life Course Approach to Healthy Ageing. New York: Oxford University Press; 2014. pp. 3–15.
  54. 54. Murray ET, Hardy R, Strand BH, Cooper R, Guralnik JM, Kuh D. Gender and Life Course Occupational Social Class Differences in Trajectories of Functional Limitations in Midlife: Findings From the 1946 British Birth Cohort. J Gerontol A Biol Sci Med Sci. 2011;66A: 1350–1359.
  55. 55. Corna LM. A life course perspective on socioeconomic inequalities in health: A critical review of conceptual frameworks. Adv Life Course Res. 2013;18: 150–159. pmid:24796266
  56. 56. Erikson R. Social Class of Men, Women and Families. Sociology. 1984;18: 500–514.
  57. 57. Blane D, Berney L, Montgomery SM. Domestic labour, paid employment and women’s health: analysis of life course data. Soc Sci Med. 2001;52: 959–965. pmid:11234868
  58. 58. Smith ADAC, Heron J, Mishra G, Gilthorpe MS, Ben-Shlomo Y, Tilling K. Model Selection of the Effect of Binary Exposures over the Life Course. Epidemiology. 2015;26: 719–726. pmid:26172863
  59. 59. Bartley M. Health Inequality: An Introduction to Theories, Concepts and Methods [Internet]. {Polity Press}; 2004. Available:
  60. 60. Torssander J, Erikson R. Stratification and Mortality—A Comparison of Education, Class, Status, and Income. Eur Sociol Rev. 2010;26: 465–474.
  61. 61. Kok A, Aartsen M, Deeg D, Huisman M. Socioeconomic inequalities in a 16-year longitudinal measurement of Successful Aging. J Epidemiol Community Health. 2016;[forthcoming].
  62. 62. Blane D, Netuveli G, Stone J. The development of life course epidemiology. Rev DÉpidémiologie Santé Publique. 2007;55: 31–38.
  63. 63. Solga H. Longitudinal Surveys and the Study of Occupational Mobility: Panel and Retrospective Design in Comparison. Qual Quant. 2001;35: 291–309.
  64. 64. Dex S. The Reliability of Recall Data: a Literature Review. Bull Méthodologie Sociol. 1995;49: 58–89.
  65. 65. Havari E, Mazzonna F. Can We Trust Older People’s Statements on Their Childhood Circumstances? Evidence from SHARELIFE. Eur J Popul. 2015;31: 233–257.
  66. 66. Diggle P, Farewell D, Henderson R. Analysis of longitudinal data with drop-out: objectives, assumptions and a proposal. J R Stat Soc Ser C Appl Stat. 2007;56: 499–550.
  67. 67. Hogan JW, Roy J, Korkontzelou C. Handling drop-out in longitudinal studies. Stat Med. 2004;23: 1455–1497. pmid:15116353
  68. 68. Jones AM, Koolman X, Rice N. Health-related non-response in the British Household Panel Survey and European Community Household Panel: using inverse-probability-weighted estimators in non-linear models. J R Stat Soc Ser A Stat Soc. 2006;169: 543–569.
  69. 69. Fatouros IG, Kambas A, Katrabasas I, Nikolaidis K, Chatzinikolaou A, Leontsini D, et al. Strength training and detraining effects on muscular strength, anaerobic power, and mobility of inactive older men are intensity dependent. Br J Sports Med. 2005;39: 776–780. pmid:16183776
  70. 70. Alomari MA, Mekary RA, Welsch MA. Rapid vascular modifications to localized rhythmic handgrip training and detraining. Eur J Appl Physiol. 2010;109: 803–809. pmid:20225082
  71. 71. Toraman NF. Short term and long term detraining: is there any difference between young-old and old people? Br J Sports Med. 2005;39: 561–564. pmid:16046344
  72. 72. Nygård C-H, Luopajärvi T, Suurnäkki T, Ilmarinen J. Muscle strength and muscle endurance of middle-aged women and men associated to type, duration and intensity of muscular load at work. Int Arch Occup Environ Health. 1988;60: 291–297. pmid:3372036
  73. 73. Schibye B, Hansen AF, Søgaard K, Christensen H. Aerobic power and muscle strength among young and elderly workers with and without physically demanding work tasks. Appl Ergon. 2001;32: 425–431. pmid:11534787
  74. 74. Era P, Lyyra AL, Viitasalo JT, Heikkinen E. Determinants of isometric muscle strength in men of different ages. Eur J Appl Physiol. 1992;64: 84–91.
  75. 75. Clement FJ. Longitudinal and Cross-Sectional Assessments of Age Changes in Physical Strength as Related to Sex, Social Class, and Mental Ability. J Gerontol. 1974;29: 423–429. pmid:4833758