CoChat
Sign Up
Cited by 967Health disparities and outcomesAging, Health, and DisabilityPhysical Activity and Health

The relationship between physical and mental health: A mediation analysis

Open access
Ohrnberger, Julius, Fichera, Eleonora, Sutton, Matt·2017·Social Science & Medicine
DOI 10.1016/j.socscimed.2017.11.008
Open this paper in CoChat — free accountOpen PDF

Abstract

There is a strong link between mental health and physical health, but little is known about the pathways from one to the other. We analyse the direct and indirect effects of past mental health on present physical health and past physical health on present mental health using lifestyle choices and social capital in a mediation framework. We use data on 10,693 individuals aged 50 years and over from six waves (2002-2012) of the English Longitudinal Study of Ageing. Mental health is measured by the Centre for Epidemiological Studies Depression Scale (CES) and physical health by the Activities of Daily Living (ADL). We find significant direct and indirect effects for both forms of health, with indirect effects explaining 10% of the effect of past mental health on physical health and 8% of the effect of past physical health on mental health. Physical activity is the largest contributor to the indirect effects. There are stronger indirect effects for males in mental health (9.9%) and for older age groups in mental health (13.6%) and in physical health (12.6%). Health policies aiming at changing physical and mental health need to consider not only the direct cross-effects but also the indirect cross-effects between mental health and physical health.

AI snapshot

Written from the paper’s full text.

TL;DR

Using six waves of the English Longitudinal Study of Ageing, this paper finds significant direct and indirect cross-effects between past mental and physical health, with lifestyle choices and social capital mediating about 10% and 8% of the total effects. Physical activity is the largest mediator, and indirect effects vary by age and gender.

Key findings

  • Past mental health has significant direct and indirect effects on present physical health, and past physical health has significant direct and indirect effects on present mental health. The total indirect effect accounts for 9.7% of the effect of past mental health on physical health and 8% of the effect of past physical health on mental health.
  • Physical activity is the largest contributor to the indirect effects: it explains 77.27% of the total indirect effect of past mental health on physical health, all of the total indirect effect of past physical health on mental health, 7.55% of the total effect of past mental health on physical health, and 8% of the total effect of past physical health on mental health.
  • Social interaction explains 13.63% of the total indirect effect of past mental health on physical health. In the full-sample mediation, the indirect cigarette effect is not statistically significant for either direction, though better baseline mental health is associated with smoking about 16% fewer cigarettes on average.
  • Indirect effects are heterogeneous: stronger indirect effects are reported for males in mental health (9.9%) and for older age groups in mental health (13.6%) and physical health (12.6%). For 50–60 year olds, cigarette smoking explains 41.7% of the total indirect effect and 5.4% of the total effect of past mental health on physical health.

Methods

  • Data: six waves (2002–2012) of the English Longitudinal Study of Ageing, covering 10,693 individuals aged 50 years and over, with 29,944 observations.
  • Measures: physical health via the six-item Activities of Daily Living index (0–6, higher = better); mental health via the eight-item Center for Epidemiological Studies Depression scale (0–8, higher = better); mediators are lagged physical activity, cigarette consumption, and a social interaction index; controls include demographics, education, retirement, insurance, income, region, and year.
  • Analysis: random effects regression models estimate direct effects; mediation is estimated by total differentiation/product-of-coefficients, with bootstrapped direct, indirect, and total effects; subgroup analyses are by gender and age quartiles (50–60, 60–70, 70–80, 80+), with robustness checks for smoking addiction stock, two-period lagged mediators, balanced panel, and attrition.

Limitations

  • The paper states that only three mediators are considered, whereas other factors such as socio-economic status, dietary choices, or medical care may be important; therefore the estimates are lower bounds for mediator effects and upper bounds for direct effects.
  • The paper states that further studies are needed to consider a broader set of mediators and to examine specific health interventions and long-term cross-effects.
  • Evident from the text: the analysis relies on observational ELSA data and assumes conditional exogeneity of lifestyle choices and social interaction, random effects, and no reverse causality between physical and mental health, which may limit causal interpretation.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.socscimed.2017.11.008

Cites

55 of 55 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 42 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    The indirect effect accounts for about 10% of the total effect in both models.

    “The indirect effect accounts for about 10% of the total effect in both models, mainly explained by past physical activity.”
    Quote in the paperSubstantial claim
  • Findingown data

    Physical activity is the largest contributor to the indirect effects.

    “Physical activity is the largest contributor to the indirect effects.”
    Quote in the paperSubstantial claim
  • Findingown data

    Estimation by subgroups shows heterogeneous effects by age and gender.

    “Estimation by subgroups shows heterogeneous effects by age and gender.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Better baseline mental health increases the intensity of physical activity on average by about 10% in the full sample.

    “We find that better baseline mental health increases the intensity of physical activity on average by about 10% in the full sample in Table 2 (3).”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.
    • The authors state this as an assumption: The indirect effects of past mental health on current physical health and of past physical health on current mental health are mediated through lagged effects of physical activity, social interactions, and cigarette consumption.
    • The authors state this as an assumption: The errors v_it and ρ_it in equations (1) and (2) are uncorrelated.
    • The authors state this as an assumption: The composite error term v_it = μ_i + ε_it has a time-invariant individual component μ_i and a time-varying error ε_it.
    • The authors state this as an assumption: Choices of health investments and consumption are maximised over health, requiring that errors are not serially correlated and choices are independent of previous choices.
    • The authors state this as an assumption: RE rules out correlation between the regressors and the unmeasured variability across individuals.
  • Findingown data

    The baseline value of mental health has a statistically significant and positive direct effect on present physical health.

    “The baseline value of mental health has a statistically significant and positive direct effect on present physical health.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.
    • The authors state this as an assumption: The indirect effects of past mental health on current physical health and of past physical health on current mental health are mediated through lagged effects of physical activity, social interactions, and cigarette consumption.
    • The authors state this as an assumption: The errors v_it and ρ_it in equations (1) and (2) are uncorrelated.
    • The authors state this as an assumption: The composite error term v_it = μ_i + ε_it has a time-invariant individual component μ_i and a time-varying error ε_it.
    • The authors state this as an assumption: Choices of health investments and consumption are maximised over health, requiring that errors are not serially correlated and choices are independent of previous choices.
    • The authors state this as an assumption: RE rules out correlation between the regressors and the unmeasured variability across individuals.
  • May combine two claims. 20
  • Classified as probably not a substantial claim. 19

  • May combine two claims.

    Findingown data

    Both present physical and mental health are explained by significant direct and indirect effects of past mental and physical health, respectively.

    “To preface our results, we find that both present physical and mental health are explained by direct and indirect effects of past mental and physical health, respectively.”
    • Also stated in: § Abstract · plain
      “We find significant direct and indirect effects for both forms of health, with indirect effects explaining 10% of the effect of past mental health on physical health and 8% of the effect of past physical health on mental health.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Past physical activity and social interactions improve physical health.

    “Past physical activity and social interactions improve physical health.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.
    • The authors state this as an assumption: The indirect effects of past mental health on current physical health and of past physical health on current mental health are mediated through lagged effects of physical activity, social interactions, and cigarette consumption.
    • The authors state this as an assumption: The errors v_it and ρ_it in equations (1) and (2) are uncorrelated.
    • The authors state this as an assumption: The composite error term v_it = μ_i + ε_it has a time-invariant individual component μ_i and a time-varying error ε_it.
    • The authors state this as an assumption: Choices of health investments and consumption are maximised over health, requiring that errors are not serially correlated and choices are independent of previous choices.
    • The authors state this as an assumption: RE rules out correlation between the regressors and the unmeasured variability across individuals.

  • May combine two claims.

    Findingown data

    The coefficient of baseline physical health is positive and statistically significant in all specifications.

    “For the physical health models, the coefficient of baseline physical health is positive and statistically significant in all specifications.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.
    • The authors state this as an assumption: The indirect effects of past mental health on current physical health and of past physical health on current mental health are mediated through lagged effects of physical activity, social interactions, and cigarette consumption.
    • The authors state this as an assumption: The errors v_it and ρ_it in equations (1) and (2) are uncorrelated.
    • The authors state this as an assumption: The composite error term v_it = μ_i + ε_it has a time-invariant individual component μ_i and a time-varying error ε_it.
    • The authors state this as an assumption: Choices of health investments and consumption are maximised over health, requiring that errors are not serially correlated and choices are independent of previous choices.
    • The authors state this as an assumption: RE rules out correlation between the regressors and the unmeasured variability across individuals.

  • May combine two claims.

    Findingown data

    The share of the total indirect effect in the total effect is lowest for 60-70 year olds (7.9%) and highest for 50-60 year olds (12.9%).

    “The share of the total indirect effect in the total effect is lowest for 60-70 year olds (7.9%) and highest for 50-60 year olds (12.9%).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The mediation models are not confounded by unmeasured variables.

  • May combine two claims.

    Findingown data

    Only physical activity mediates the effect of baseline physical health on current mental health, with a positive association.

    “Only physical activity mediates the effect of baseline physical health on current mental health, with a positive association.”
    • Also stated in: § 5.3. Estimates of the mediating effects · plain
      “The sub-sample analysis by age groups and gender on mental health, presented in table A7, is consistent with the full sample analysis in finding that only physical activity mediates the effect of baseline physical health with current mental health.”
    • Also stated in: § 6. Discussion & conclusion · plain
      “The relationship of past physical health with present mental health is only mediated by past physical activity.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The mediation models are not confounded by unmeasured variables.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The coefficient of baseline mental health on smoking is slightly stronger for females with -0.237 than for males with -0.205.

    “The coefficient is slightly stronger for females with -0.237 than it is for males with -0.205.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.
    • The authors state this as an assumption: The indirect effects of past mental health on current physical health and of past physical health on current mental health are mediated through lagged effects of physical activity, social interactions, and cigarette consumption.
    • The authors state this as an assumption: The errors v_it and ρ_it in equations (1) and (2) are uncorrelated.
    • The authors state this as an assumption: The composite error term v_it = μ_i + ε_it has a time-invariant individual component μ_i and a time-varying error ε_it.
    • The authors state this as an assumption: Choices of health investments and consumption are maximised over health, requiring that errors are not serially correlated and choices are independent of previous choices.
    • The authors state this as an assumption: RE rules out correlation between the regressors and the unmeasured variability across individuals.

  • May combine two claims.

    Findingown data

    The total effect of baseline physical health on current mental health is positive and increases present mental health by about 3.5%.

    “The total effect is positive and increases present mental health by about 3.5%.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The mediation models are not confounded by unmeasured variables.

  • May combine two claims.

    Findingown data

    Physical activity is important throughout the life-cycle from 50+, but is stronger for higher age groups.

    “Physical health activity is important throughout the life-cycle from 50+, but is stronger for higher age groups.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The mediation models are not confounded by unmeasured variables.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Approximately 55% of the burden of disease among people aged 60 years and over in high-income countries is avoidable by changes in lifestyles.

    “A report by the Ageing Research Group (Melzer et al., 2012) finds that approximately 55% of the burden of disease amongst the population aged 60 years and over is avoidable by changes in lifestyles in high income countries.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Melzer et al., 2012

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Social isolation, loneliness and exclusion are strong risk factors for ill health and impose a high risk on the health of older populations.

    “These are strong risk factors for ill health and impose a high risk on the health of older populations that could be easily avoided with low-cost interventions (Steptoe et al., 2012).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Steptoe et al., 2012 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    On the full population, lifestyle factors explain up to 70% of the burden of disease.

    “On the full population, lifestyle factors explain up to 70% of the burden of disease (WHO, 2009).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WHO, 2009

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Physical activity is negatively associated with depression/anxiety disorders and poor physical health outcomes.

    “Several studies have found that physical activity is negatively associated with depression/anxiety disorders and poor physical health outcomes (Gerber and Puehse 2009;De Mello et al., 2013;Durstine et al., 2013;Wang et al., 2014;Hegberg and Tone, 2015).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Gerber and Puehse 2009

      Not checked in the free run.

    • cites De Mello et al., 2013 source

      Not checked in the free run.

    • cites Durstine et al., 2013 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Systematic reviews find strong evidence for positive effects of exercising on both mental and physical health outcomes for older people.

    “Systematic reviews find strong evidence for positive effects of exercising on both mental and physical health outcomes for older people (Forbes et al., 2008;Clegg et al., 2012).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Forbes et al., 2008 source

      Not checked in the free run.

    • cites Clegg et al., 2012 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductnone given

    The findings are robust to the timing of the mediators and how the stock of addiction is modelled.

    “The findings are robust to the timing of the mediators and how the stock of addiction is modelled.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Tobacco use explains 17.9% of all deaths.

    “Tobacco use explains 17.9% of all deaths and physical inactivity explains 7.7% of all deaths (WHO, 2009).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WHO, 2009

      Not checked in the free run.

  • May combine two claims.

    Study design and conductreasoning

    Using a baseline value that pre-dates the mediators mitigates concerns over reverse-causality and captures earlier life investments in health and health endowments.

    “Using a baseline value that pre-dates the mediators mitigates concerns over reverse-causality and captures earlier life investments in health and health endowments.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Study design and conductnone given

    Bootstrapping takes both the longitudinal structure and the random effect error assumption into account.

    “Bootstrapping takes both the longitudinal structure and the random effect error assumption into account.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: The composite error term v_it = μ_i + ε_it has a time-invariant individual component μ_i and a time-varying error ε_it.
    • The authors state this as an assumption: RE rules out correlation between the regressors and the unmeasured variability across individuals.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Physical inactivity explains 7.7% of all deaths.

    “Tobacco use explains 17.9% of all deaths and physical inactivity explains 7.7% of all deaths (WHO, 2009).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WHO, 2009

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Social capital is important in increasing the risks of social isolation, loneliness and exclusion of older populations, which lead to higher risks of mortality.

    “Social capital is similarly important in increasing the risks of social isolation, loneliness and exclusion of older populations, which lead to higher risks of mortality (Holt-Lunstad et al., 2012;Melzer et al., 2012;Steptoe et al., 2012).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Holt-Lunstad et al., 2012 source

      Not checked in the free run.

    • cites Melzer et al., 2012

      Not checked in the free run.

    • cites Steptoe et al., 2012 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Reverse-causality between smoking and mental health is likely to occur because smoking rates are twice as high among adults with depression or anxiety disorders.

    “Reverse-causality is likely to occur as smoking rates are twice as high among adults with depression or anxiety disorders (RCP, 2013).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites RCP, 2013

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The estimates are lower bounds of the indirect effects between physical and mental health and upper bounds of their direct effects.

    “For these reasons, we consider our estimates to be lower bounds of the indirect effects between physical and mental health and upper bounds of their direct effects.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Other potential mediators such as stress and sleep are captured directly by the measure of mental health.
    • Rests on an assumption: Social interactions are critical production factors in health.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Mental health is measured using the validated eight-item version of CES-D developed by Radloff (1977).

    “We measure mental health using the validated eight-item version of the Center for Epidemiological Studies-Depression (CES-D) developed by Radloff (1977).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Radloff (1977) source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Other factors such as obesity, high blood pressure and alcohol consumption are widely used in the literature as health outcome measures.

    “Thirdly, other factors such as obesity, high blood pressure and alcohol consumption are widely used in the literature as health outcome measures (Banks et al., 2006).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Banks et al., 2006 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Smoking is the leading cause of preventable death in the UK.

    “Smoking is the leading cause of preventable death in the UK (Office for National Statistics UK, 2016).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Office for National Statistics UK, 2016

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Alcohol consumption and dietary choices together explain only 4.1% of deaths.

    “For example, alcohol consumption and dietary choices together explain only 4.1% of deaths (WHO, 2009).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WHO, 2009

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    The ELSA started including questions about dietary choices only in the third wave and the phrasing of these questions changed from wave 5 onwards.

    “For example, the ELSA started including questions about dietary choices only in the third wave and the phrasing of these questions changed from wave 5 onwards.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Potential interrelations of mental and physical health are accounted for by the inclusion of baseline health values.

    “Potential interrelations of mental and physical health are accounted for by the inclusion of baseline health values.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductcitation

    The social interaction index is similar to Steptoe et al. (2012) and reflects changes in social isolation and loneliness.

    “The measure is similar to Steptoe et al. (2012) and also reflects changes in social isolation and loneliness.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Steptoe et al. (2012) source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Previous studies have found a strong positive relationship between social interaction and mental health.

    “Previous studies have found a strong positive relationship between social interaction and mental health (Dour et al., 2014;Bekele et al., 2015).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Dour et al., 2014 source

      Not checked in the free run.

    • cites Bekele et al., 2015 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The models control for year and regional effects.

    “We control in all models for year and regional effects.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Better physical and mental health status leads to more physical activity, which in return has a positive association with better mental health and physical health.

    “Better physical and mental health status therefore leads to more physical activity which in return has a positive association with better mental health and physical health.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The mediation models are not confounded by unmeasured variables.

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    A limitation of this paper is that only a set of three mediators is considered, whereas other factors such as socio-economic status, dietary choices or medical care may be important.

    “A limitation of this paper is that only a set of three mediators is considered whereas other factors such as socio-economic status, dietary choices or medical care may be important.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Gerber and Puehse (2009) found a positive association between mental health and physical activity.

    “Gerber and Puehse (2009) have found a similar positive association between mental health and physical activity.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Gerber and Puehse (2009)

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The study can only capture a lower bound estimate for the mediator effects and an upper bound for the direct effects.

    “Whilst we justify the choice of mediators with a conceptual framework, existing literature, key policy relevance and data availability, we note that we can only capture a lower bound estimate for the mediator effects and an upper bound for the direct effects.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Social interactions favour good health behaviours and vice versa.

    “An explanation for this effect is that social interactions favour good health behaviours and vice versa (Umberson et al., 2010).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Umberson et al., 2010

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The positive association of social ties with mental health has previously been found.

    “The positive association of social ties with mental health has previously been found (Umberson and Montez 2010).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Umberson and Montez 2010

      Not checked in the free run.

36 of 36

  • The paper relies on this without support.

    Little is known about the potential pathways through which mental health affects physical health and vice versa.

    “However, little is known about the potential pathways through which mental health affects physical health and vice versa (i.e.so-called “indirect effects”).”

  • The paper relies on this without support.

    No existing mediation study has examined the relationship between physical and mental health and the pathways considered in this paper.

    “Many psychological, epidemiologic, and economic studies have used mediation analysis, but none in the context of the relationship between physical and mental health and the pathways that we consider.”

  • The authors state this as an assumption.

    Conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.

    “We assume that, conditional on baseline health, lifestyle choices and social interaction are exogenous to mental and physical health.”

  • The authors state this as an assumption.

    The indirect effects of past mental health on current physical health and of past physical health on current mental health are mediated through lagged effects of physical activity, social interactions, and cigarette consumption.

    “We assume that the indirect effects of past mental health on current physical health and of past physical health on current mental health are mediated through lagged effects of the individual health investments, including past physical activity ( -PhysAct it 1 ), past social interactions ( -SC Index it 1 ), and past cigarette consumption ( -CIG it 1 ).”

  • The authors state this as an assumption.

    The errors v_it and ρ_it in equations (1) and (2) are uncorrelated.

    “We assume that the errors v it and ρ it in equations ( 1) and ( 2) are uncorrelated.”

  • The authors state this as an assumption.

    The composite error term v_it = μ_i + ε_it has a time-invariant individual component μ_i and a time-varying error ε_it.

    “We account for unobserved heterogeneity by including random effects (RE) in all equations, assuming a composite error term = + v μ ε it i it with timeinvariant individual component μ i and time-varying error ε it .”

  • The authors state this as an assumption.

    Choices of health investments and consumption are maximised over health, requiring that errors are not serially correlated and choices are independent of previous choices.

    “This is an important assumption for our model where choices of health investments and consumption are maximised over health requiring that errors are not serially correlated and choices are independent of previous choices.”

  • The paper relies on this without support.

    Other potential mediators such as stress and sleep are captured directly by the measure of mental health.

    “Secondly, other potential mediators such as stress and sleep are captured directly by our measure of mental health.”

  • The paper relies on this without support.

    Social interactions are critical production factors in health.

    “Finally, social interactions are critical production factors in health.”

  • The paper relies on this without support.

    The ELSA is a representative sample of the English population aged 50 years and older.

    “The study is a representative sample of English population aged 50 years and older, collecting objective and subjective data on physical and mental health, and on socio-economic status.”

  • The paper relies on this without support.

    The effect of health lags is captured by the mediators.

    “It differs from dynamic models with lagged health because the effect of health lags is captured by the mediators.”

  • The authors state this as an assumption.

    RE rules out correlation between the regressors and the unmeasured variability across individuals.

    “RE rules out correlation between the regressors and the unmeasured variability across individuals.”

  • The paper relies on this without support.

    The mediation models are not confounded by unmeasured variables.

    “The mediation analysis on physical health shows that the effect of baseline mental health on present physical health is significantly mediated by past physical activity and by past social interaction.”

  • The paper relies on this without support.

    The temporal ordering of baseline health, mediators, and outcomes is correctly specified.

    “For models (3)-(5) the time-variant covariates take on (t-1)-lagged values (e.g.Age (t-1), Age Squared (t-1), Higher Education (t-1), Household Size (t-1), Private Insurance (t-1), Retired (t-1), Log-equivalised household income (t-1)).”

  • The paper relies on this without support.

    The ELSA sample is representative of the older English population.

    “We use six waves of the ELSA (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012), and investigate potential mediating effects on the full sample size and heterogeneous effects by age-quartiles and gender.”
15 of 15

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Ohrnberger and colleagues used a mediation model and reported that the relations between psychological and physical health are mediated by lifestyle factors, including increased physical activity and social connections.

    “Ohrnberger and colleagues used a mediation model and reported that the relations between psychological and physical health are mediated by lifestyle factors, including increased physical activity and social connections.”

    From Cancer, now what? A cross-sectional study examining physical symptoms, subjective well-being, and psychological flexibility (2023) · cites this paper as Ohrnberger et al., 2017

    The Results (§5.3) report that physical activity and social interaction mediate the effect of past mental health on present physical health. The Discussion (§6) says the reverse relationship—from past physical health to present mental health—is mediated only by physical activity. The claim therefore overgeneralizes the role of social connections across the relationship in both directions. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • These findings support large-scale studies that highlight direct and indirect links between physical and psychological wellness.

    “These findings support large-scale studies that highlight direct and indirect links between physical and psychological wellness (see Ohrnberger et al., 2017).”

    From Cancer, now what? A cross-sectional study examining physical symptoms, subjective well-being, and psychological flexibility (2023) · cites this paper as Ohrnberger et al., 2017

  • SNS addiction (e.g., preoccupation with social media, excessive use) is associated with lower well-being and depression, which can predict worse physical health.

    “SNS addiction (e.g., preoccupation with social media, excessive use) is associated with lower well-being and depression,14,58 which can predict worse physical health.59”

    From Social Media Use and Its Link to Physical Health Indicators (2022) · cites this paper as [59]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Social-interaction shares imply more than the stated maximum indirect share

    Minor

    For ages 70-80, a pathway accounting for 1.3% of the indirect effect and 0.19% of the total effect implies that the indirect effect is about 14.6% of the total. That exceeds the stated maximum of 12.9% across age groups, even allowing for the printed rounding.

1 of 1

Cites

Flagged first, then by citations

4 of the 59 references couldn’t be matched to a known paper.

Research works printed but not linked: 1

  1. Umberson, 2010, Social relationships and health: a flashpoint for helath policy, J. Health Soc. Behav., 51, 554

Cited by

Sorted by citations

Related

Authors
Ohrnberger, J., Fichera, E., & Sutton, M.
Affiliations
  • University of Manchester (GB)
  • University of Bath (GB)
Corresponding author
Ohrnberger, Julius
ORCID
Venue
Social Science & Medicine
Publisher
Elsevier BV
Published
2017-12
Volume
195
Pages
42-49
ISSN
0277-9536 (ISSN-L)
DOI
10.1016/j.socscimed.2017.11.008
OpenAlex
W2766487143
Semantic Scholar
ffc0fc37cb36639a67f3600d87f92e6ae76fbf8f
Cited by
967
References
59
Citations by year
2026: 70 · 2025: 130 · 2024: 173 · 2023: 150 · 2022: 142 · 2021: 128 · 2020: 123 · 2019: 38 · 2018: 13
FWCI
52.79
Type
journal article
Language
en
Open access
Licence
Locations
  • Social Science & Medicine (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Research Explorer (The University of Manchester) (publishedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
Funders
National Institute for Health and Care Research (NF-SI-0515-10080)
Topics
  • Social Sciences › Health › Health disparities and outcomes
  • Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Aging, Health, and Disability
  • Health Sciences › Medicine › Physiology › Physical Activity and Health
Keywords
mental health, physical health, mediation analysis, physical activity, lifestyle choices, social capital, activities of daily living, depression, older adults, gender differences, age differences, health policy
MeSH terms
Aged, Female, United Kingdom, Health Status, Humans, Longitudinal Studies, Male, Mental Health, Middle Aged

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-08.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-08.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review