You are here

  • Home
  • Archive
  • Volume 14, Issue 7
  • Demographic, socioeconomic and health determinants of depressive symptoms in adults 50 years and older from Mexico: a secondary data longitudinal analysis from the Mexican Health and Aging Study

Article Text

Article menu

Download PDFPDF

Geriatric medicine
Original research
Demographic, socioeconomic and health determinants of depressive symptoms in adults 50 years and older from Mexico: a secondary data longitudinal analysis from the Mexican Health and Aging Study
  1. Liliana Giraldo-Rodríguez1,
  2. Sara Torres-Castro2,
  3. Paloma Arlet Roa-Rojas3,
  4. Teresa Alvarez-Cisneros4
  1. 1 Demographic and Determinants of Health Research, Instituto Nacional de Geriatria, Mexico City, Mexico
  2. 2 Demography and Determinants of Health Research, Instituto Nacional de Geriatria, Mexico City, Mexico
  3. 3 Clinical Epidemiology Research, Instituto Nacional de Geriatria, Mexico City, Mexico
  4. 4 Clinical Epidemiology Research, Instituto Nacional de Geriatria, Cuernavaca, Mexico
  1. Correspondence to Dr Teresa Alvarez-Cisneros; teresaalvarez.c2{at}gmail.com

Abstract

Background Depression in ageing adults is a public health problem. Worldwide studies have identified social and health risk factors for depressive symptoms. However, little is known about their longitudinal determinants in Mexico.

Objectives and setting To find the prevalence of depressive symptoms and their longitudinal individual and contextual risk factors in Mexican adults aged 50 and older.

Design Secondary data of 6460 persons aged 50 years and older from the Mexican Health and Aging Study were analysed using a ‘between-within’ panel data analysis approach.

Results The prevalence of depressive symptoms increased from 35% in 2003 to 38% in 2015. The significantly longitudinal factors associated with these symptoms were getting older (OR 1.02, 95% CI 1.01 to 1.03), being a woman (OR 2.39, 95% CI 2.16 to 2.64), less time spent in formal education (0 years and less than 6 years OR 1.52, 95% CI 1.32 to 1.75 and OR 1.33, 95% CI 1.19 to 1.50, respectively), lower net worth (OR 1.13, 95% CI 1.08 to 1.17), being recently unemployed (OR 1.25, 95% CI 1.10 to 1.25), increased (OR 1.17, 95% CI 1.10 to 1.25) or increasing number (OR 1.23, 95% CI 1.15 to 1.31) of chronic conditions, poor (OR 4.68, 95% CI 4.26 to 5.15) or worsened (OR 1.71, 95% CI 1.61 to 1.81) self-rated health and having impairments on instrumental activities of daily living (IADLs) (OR 2.94 95% CI 2.35 to 3.67) or a new IADL impairment (OR 1.67, 95% CI 1.48 to 1.89), as well as having impairments on ADLs (OR 1.51, 95% CI 1.23 to 1.86) or a new ADL impairment (OR 1.34, 95% CI 1.21 to 1.48).

Conclusions The prevalence of depressive symptoms in Mexican adults aged 50 and older is high. Our findings show that they are longitudinally associated with the individual’s demographic, socioeconomic, health and disability characteristics. Efforts in public policy should focus on preventing chronic conditions and disability, as well as fighting inequalities to reduce the prevalence of depressive symptoms.

  • mental health
  • aging
  • epidemiologic studies
  • geriatric medicine
  • depression & mood disorders

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study. Data are public and available at https://www.mhasweb.org/Home/index.aspx.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjopen-2023-075035

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    STRENGTHS AND LIMITATIONS OF THIS STUDY

    • Analysis results from a longitudinal, nationally representative sample.

    • Methods included a between-within longitudinal panel, which is helpful in controlling individual heterogeneity.

    • Although it is a nationally representative sample, sociocultural differences between regions within the country are not considered.

    • We are unable to control for the duration of health conditions or other social or personal characteristics not included in the survey.

    Introduction

    The WHO defines depression as a ‘common mental disorder, characterised by persistent sadness and loss of interest in activities that a person normally enjoy, accompanied by an inability to carry out daily activities, for at least 2 weeks: loss of energy; a change in appetite; sleeping more or less; anxiety; reduced concentration; indecisiveness; restlessness; feelings of worthlessness, guilt or hopelessness; and thoughts of self-harm or suicide’.1 Depression is considered one of the most prevalent neuropsychiatric conditions among older adults2 and its prevalence increases with age.3 The worldwide prevalence of depression in older adults is between 7% and 70%, depending on the diagnostic method or measurement scale. This is relevant as most clinical studies focus on the prevalence of major depressive disorders while most epidemiological studies use screening scales such as the CESD used in this study.4–6 Using such scales, the WHO has calculated that depression represents 5.7% of years lived with disability in this age group.1 In Mexico, studies using epidemiological data estimated a prevalence of depression in older adults between 13.2% and 35.6% while the prevalence of depressive symptoms in younger adults was estimated to be 23%.7–11 In comparison, studies conducted among groups of older adults in selected cities or in clinical settings have found a higher prevalence.11–13 García-Peña et al found the following U-shaped prevalences of depressive symptoms in Mexican older adults: 20% in the 60–64 age group, 25.6% in the 75–79 age group and 16% in the 90+ age group.14

    As a result of this high prevalence and its negative effects on health, depression has been an important public health problem in Mexico since 1999.15 It is multifactorial and it has been suggested to be a result of the interaction of biological, psychological and social vulnerabilities in older adults.16 Some of the biological vulnerabilities in older age groups include the lowering of monoamines in the central nervous system, as well as neuroendocrine alterations, neuroanatomic modifications and maladjustment in the interactions of the immune and neuroendocrine systems. The higher prevalence of multimorbidity and functional limitations in this age group are other biological factors associated with depression.17–20 In addition, some psychosocial factors play a significant role in the onset of depressive symptoms. Recent studies have found a significant correlation between financial hardship, lower educational attainment, being a victim of child abuse or elder abuse, adverse life events, poor social support and family struggles with older age depression.21–24 Furthermore, a study in Mexico showed that depressive symptoms in older adults are more associated with mobility limitations rather than age.25

    Significant attention has been given to the effects of having depression and a higher prevalence of disability and dependency in older adults.23 24 26 27 Furthermore, depression is associated with higher rates of mortality and suicide.28 A recent meta-analysis including 293 studies and 1 813 733 participants found that the relative risk of mortality among depressed individuals was 1.64 (95% CI 1.56 to 1.76). It was higher than among those without depression and such risk was reduced to 1.52 once adjusted for publication bias.28 These results did not change when patients had specific illnesses.

    To date, most longitudinal studies have focused on analysing depressive symptoms and their relationship with chronic conditions,27 cognitive impairment29 and disabilities.30 Most of these studies have been conducted in high-income countries. However, it is known that the prevalence of depressive symptoms is increasing particularly in middle-income and low-income countries. Most of the evidence from middle-income countries comes from studies conducted in China and Thailand.31 32 There is an accelerated ageing process in Mexico, and depression has deleterious effects in older age. Therefore, this study aims to find the overall prevalence of depressive symptoms, as well as the longitudinal individual and contextual risk factors of these symptoms in Mexican adults aged 50 years and older. Until now, no other study has used a nationally representative sample to calculate the prevalence of clinically significant depressive symptoms and how it has changed over time. Also, no other study had focused on the determinants of depression in Mexican older adults.

    Methods

    Research design

    Data and sample

    The sample was obtained from four waves of the Mexican Health and Aging Study (MHAS) 2001 (n=15 186), 2003 (n=14 250), 2012 (n=18 465) and 2015 (n=15 988). MHAS is a prospective study of ageing focusing on the impact of disease on health, function and mortality, following the framework of the US Health and Retirement Study.33 It includes a representative randomised sample of adults over 50 years old from Mexico, with a representative refresher sample of 6259 individuals of those born between 1952 and 1961, added in 2012. The cumulative reported deaths from 2001 to 2012 was 3288. Using survey weights that are specific for each wave of the study, data from 6460 individuals in 2001, 6279 individuals in 2003, 6032 individuals in 2012 and 6315 individuals in 2015 were analysed. Sample changes within each wave, as well as all complete responses from that year, were included in each cross-sectional analysis. However, for the panel data analyses, we considered the entire sample as between-within random effects models adjusted for missing data by including the fixed effect of individual change.

    The MHAS is a collaboration between the University of Texas Medical Branch, the Mexico’s National Institute of Statistics and Geography (INEGI, by its Spanish acronym), the Mexico’s National Institute of Geriatrics, the Mexico’s National Institute of Public Health, the Columbia University and the University of California Los Angeles. It is partly supported by the National Institutes of Health (NIH) (grant number NIH RO1AG018016) and the INEGI. Participants and their families are aware of the relevance of this study. They are encouraged to participate in every visit since findings are useful for improving care and policy recommendations.

    Patient and public involvement

    This is a secondary data analysis research. Therefore, patients were not involved in the development of the research question and outcome measures.

    Measures

    Depressive symptoms

    Depressive symptoms were measured using the CESD-9, an adaptation of the CESD-20 scale for Anglo-Saxon participants. This scale has been previously validated for use in older adults.34 35 Its equivalent, CESD-9, is also validated for use in the Mexican population with a calculated sensitivity of 80.7% and a specificity of 68.7%.36 In line with previous research methodologies employed to determine the prevalence of depressive symptoms, a threshold of 5 or higher was applied to assess the prevalence of clinically significant depressive symptoms as a binary variable.37

    Predictors of depressive symptoms

    The demographic factors considered in the analyses included three age categories (50–59 years old, 60–69 years old and 70+ years old) and gender (male or female). Level of education (divided into four categories), being employed or not and net worth in quintiles (highest, medium high, medium, medium low and lowest) constituted the socioeconomic factors. Assessment of physical health characteristics included respondent’s body mass index (BMI) calculated as body weight (kg) divided by squared height (m2) and divided into three categories (normal, overweight and obese); the number of chronic conditions (diabetes, arthritis, hypertension, heart conditions, lung disease and cancer) and self-rated health (excellent and very good, good and poor) as a proxy of objective health status. Lastly, because of their relevant association with depression, instrumental activities of daily living (IADL) disabilities and ADL disabilities were also included as control variables. Both variables were used as dichotomic, and those having one impairment were categorised as having a disability. The same types of variables were included in the longitudinal analyses while the operationalidation differed, age, net worth and BMI were included as continuous variables. This is because the within effect in the analyses examines the change within the same individual so variables should preferably be either dichotomic or continuous.

    Analysis strategy

    Two main analyses were performed. The first part included a series of cross-sectional results from each wave of the study. The second part included the longitudinal analyses focusing on analysing data from all participants from all waves. For the cross-sectional analyses on the prevalence estimates of depressive symptoms, weighted data for each year were included. This was followed by bivariate analyses on the association of depressive symptoms and the control variables, using Fisher’s F as an exact significance test.

    The longitudinal analyses were performed using panel data analysis. In this type of analysis, fixed-effects models are commonly acknowledged to offer a superior advantage compared with random-effects models, as they effectively account for all level 2 characteristics (whether measured or unmeasured). However, the effect of variables that do not change with time cannot be estimated. The between-within method or ‘hybrid model’ used in these analyses includes the between and within effect in a random-effects model.38 Consequently, the random effects are given by:

    Embedded Image

    Where i denotes subjects and t occasions. The ‘hybrid effect’ or between-within effect provides the within effect by including the time-demeaned variable effect in a random effects model.

    Embedded Image

    Where B1 provides the within effect by including the time-demeaned change of each of the variables that have more than one value over time. This method allows the inclusion of the between-subjects effect (as an independent measure design) and the within-subject effect (repeated measures design) in the same analysis. Thus, causal inference is facilitated by considering fixed effects and time-varying effects. This enhances the internal validity of the study and reduces confounding effects. The within-subject effect helps examine how individuals change over time: within-subject variations are recorded, examining the dynamics of individual experiences. It is adequate for studying processes that occur over extended periods of time. This also helps to control heterogeneity by including time-demeaned variant characteristics such as net worth or employment from each wave. This is relevant as the effects of the average net worth and average number of chronic conditions were measured, as well as the decreasing net worth or increasing number of chronic conditions over time. By controlling heterogeneity, the effect of attrition was minimised, mitigating biases associated with missing data. Also, the entire sample was used to avoid losing information. Four different models that include demographic (age and gender), socioeconomic (level of education, employment and net worth) and health and disability status (BMI, the number of chronic conditions, and self-rated health, and IADL and ADL disability) characteristics were conducted.

    Results

    Descriptive analysis

    In this section, results from the descriptive analyses are presented. Figure 1 shows the sample size and the prevalence of clinically significant depressive symptoms among adults older than 50 years in each of the four waves of the study.

    Figure 1
    Figure 1

    Flow chart of the sample size and prevalence of clinically significant depressive symptoms in waves of 2001, 2003, 2012 and 2015 of the MHAS study. Prevalence was calculated with survey-weighted data. MHAS, Mexican Health and Aging Study. Clinically significant depressive symptoms were calculated using the Center for Epidemiological Studies-Depression Scale (CESD) cuttoff of 5.

    No difference was found from 2001 to 2003, as the prevalence remained to be 35%. Nine years later (in the wave of 2012), the proportion of older adults reporting depressive symptoms was 36%. By 2015, this prevalence increased to 38%. As observed in table 1, the proportion of respondents in the younger age group decreased over time while the proportion of those aged 70+ years increased. Across all waves of the study, there was a higher proportion of female respondents. The proportion of individuals without formal education decreased over time, and the proportion of respondents with 7 or more years of formal education increased. The proportion of individuals working decreased with age while the proportion of those in the lowest net worth category increased. With regard to health characteristics, the proportion of respondents with normal weight increased while the proportion of overweight and obese individuals decreased. The highest proportion of respondents with no chronic conditions occurred in the wave of 2012 while the lowest proportion was in 2015, where the prevalence of one, two or three or more conditions was the highest. The percentage of individuals with excellent and very good health remained low throughout the years. Most respondents reported having fair self-rated health in all waves. The prevalence of IADL disabilities was reported by half of the respondents in 2001 but when the refresher wave of 2012 was introduced, 80% reported having no IADL disability. This occurred similarly with ADL disabilities. Half of the respondents reported one or more impairments in the first wave. Then, the proportion decreased to 23% in 2012 with the refresher sample. However, 50% of respondents reported having at least one ADL impairment 3 years later.

    View this table:
    Table 1

    Descriptive characteristics of participants in each wave

    Findings presented in table 2 show that there were no significant age differences between individuals with clinically significant depressive symptoms and those without them in 2001. However, with regard to gender, a significantly higher proportion of women (70% vs 49%, compared with men) reported significant symptoms of depression. A higher proportion of respondents with symptoms of depression belonged to the schooling groups and were not working. In this baseline analysis, no significant differences in net worth were found between the group without depressive symptoms and the group with these symptoms.

    View this table:
    Table 2

    Demographic, socioeconomic and health characteristics respondents without and with depressive symptoms in 2001

    With respect to the association between health characteristics and depressive symptoms, our analyses did not show significant differences with regard to BMI. However, a higher number of chronic conditions was more prevalent among those with depressive symptoms. This group also reported poorer health: 24% reported poor health and only 1% reported excellent and very good health. In comparison, 6% reported poor health, and 8% reported excellent health in the group without depressive symptoms. Respondents with depressive symptoms did not report a significantly higher prevalence of difference in IADL impairments when compared with the group without depressive symptoms. Nevertheless, the group with depressive symptoms reported a significantly higher prevalence of ADL impairments (15% vs 4% in the group without depressive symptoms), as observed in the prevalence of ADL impairments between groups.

    Longitudinal determinants of depressive symptoms

    Results from the ‘between-within’ models are summarised in table 3. Four models were performed to demonstrate how the inclusion of socioeconomic, health and disability characteristics modify the significance of other variables on depressive symptoms. Model 1 considers only age and gender. Results show that higher average age and increasing age raise the OR for clinically significant depressive symptoms (OR 1.03, p<0.001 and 1.01, p<0.001, respectively). Also, women had significantly higher odds (OR 3.25, p<0.001). With the inclusion of socioeconomic conditions in model 2, the OR decreased but remained significant for the three variables. Respondents belonging to the categories of 6 years, 1–5 years and no years of schooling (OR 1.70, OR 2.44 and OR 2.97, p<0.001, respectively) reported significantly higher odds of depressive symptoms when compared with those with seven or more years of schooling. Regarding employment status and net worth, respondents who did not have a job (OR 1.51, p<0.001) and who lost their job during the study (OR 1.34, p<0.001) reported a significantly higher prevalence of depressive symptoms than those having a job. On the other hand, only those with lower average net worth reported depressive symptoms, which was significantly higher than those within the higher net worth categories (OR 1.21, p<0.001).

    View this table:
    Table 3

    Longitudinal determinants of depression in older adults from Mexico

    In model 3, health characteristics such as BMI, the number of chronic conditions and self-rated health were included. The fact of belonging to older age groups reduced the odds of depressive symptoms once the health characteristics were controlled (OR 0.99, p<0.001) while the effect of increasing age remained almost unchanged (OR 1.03, p<0.001). The higher odds of depressive symptoms among women decreased but remained significant (OR 2.25, p<0.001). Respondents belonging to the 6 years of schooling group did not report significantly higher depressive symptoms than those with seven or more years of schooling. The OR of those in the two lower schooling categories decreased to 1.30 and 1.54, p<0.001, respectively. The effect of having no job became non-significant. At the same time, the effect of losing a job decreased but remained significant (OR 1.25, p<0.001). The effect of belonging to a lower net worth category also decreased but remained significant (OR 1.16, p<0.001). The inclusion of health characteristics showed that having a higher average BMI significantly decreased the OR of depressive symptoms (OR 0.98, p<0.001). It also revealed that a higher average (OR 1.21, p<0.001) and increasing number (OR 1.26, p<0.001) of chronic conditions and a lower self-rated health (OR 4.97, p<0.001), as well as the worsening of self-rated health (OR 1.68, p<0.001) significantly increased the OR of depressive symptoms.

    Lastly, the effects of at least one IADL or ADL impairment were added to model 4. Results from model 3 remained practically unchanged with getting older, being a woman, having a schooling of 0 years and less than 6 years, belonging to a lower net worth quintile, losing a job, having a higher number of chronic conditions or increasing the number of chronic conditions. Also, having a poor self-rated health and worsened self-rated health significantly increased the OR of clinically significant depressive symptoms. Both were in the groups with IADL disabilities (OR 2.94, p<0.001), having a new IADL impairment (OR 1.67, p<0.001), with ADL impairments (OR 1.51, p<0.001) and having a new ADL impairment (OR 1.34, p<0.001), which significantly increased the OR for depressive symptoms in this group of Mexican older adults.

    Discussion

    Main findings and comparison to other studies

    Results from this study suggest that the prevalence of depression in older adults from Mexico in 2001 was 35% and remained the same by 2003. In 2012 and 2015, the prevalence increased by 1% and 3%, respectively. This increase could be a result of the major economic crisis or major personal events. However, as the population became older, chronic conditions, multimorbidity, and IADL and ADL impairments became more prevalent. These factors are longitudinally associated with a likelihood of depressive symptoms, as our findings suggest. While several studies have focused on the risk factors for depressive symptoms in older adults, this is the first longitudinal study to analyse the longitudinal within and between effects of each time-changing characteristic associated with depressive symptoms. With this method, the effects of increasing age, worsened net worth, increasing number of chronic conditions, worsened self-rated health or increasing number of IADL or ADL impairments could be analysed. This is relevant as it highlights the importance of interventions in already disadvantaged individuals.

    Our results showed that once demographic, socioeconomic, health and disability characteristics were controlled, there were no meaningful cross-sectional age differences between the group without depressive symptoms and the one with depressive symptoms. Instead, there was a longitudinally positive association between increasing age and their likelihood. This means that, while there are no significant differences in the prevalence of depressive symptoms between generations, getting older does increase the likelihood of depressive symptoms within the same generation. This is consistent with studies where, regardless of the method used, the prevalence of depressive symptoms among older adults increased with age.39 Additionally, our results showed that women were significantly more likely to report depressive symptoms cross-sectionally and longitudinally. Epidemiological studies have reported that recurrence of depressive symptoms is at least twice as common in women as in men,40 considering gender as a risk factor for depression. This is in line with the 2.39 OR of significant depressive symptoms for women aged 50 years and older, reported in this study. The prevalence of depressive symptoms in women is associated with hormonal and genetic factors, pre-existing anxiety, women’s specific socialisation and coupling styles, traumatic life experiences, and social conditions and roles, among others. Also, loneliness acts as a confounder factor for depression, mostly for women of older ages. Studies have found significant associations between loneliness and depression,41 considering that loneliness may be experienced as a state of being alone or as the feeling of being alone, despite living with other people.

    Individuals with less time spent in formal education, with lower net worth, and those who were not working were also more likely to report depressive symptoms. These characteristics increased the likelihood of depressive symptoms, despite controlling health and disability. This has been explored previously by Xie et al,39–42 addressing the association between depression and socioeconomic factors. Results from these studies show that individuals with lower income and lower educational attainment were more likely to be depressed. This outcome was associated with high psychiatric morbidity, greater degree of disability/more disabilities and poorer access to medical care.43 A more recent systematic review and meta-analysis by Errazuriz et al showed that a lower human development index and a higher gender inequality were associated with a higher prevalence of depression in Latin American countries.44 It is important to highlight that once these socioeconomic variables were included, the likelihood of depressive symptoms decreased from OR 3.25 to OR 2.56 for women. Thus, part of the higher likelihood of depression among women can be explained by the socioeconomic disadvantage (lower schooling, lower net worth and less employment opportunities). This acts as a stressor, affecting mental and physical health status and triggering depressive symptoms in women.45

    Regarding the health characteristics associated with a higher likelihood of depressive symptoms, the fully adjusted longitudinal model highlighted that those with higher BMI were less likely to have depressive symptoms, but that an increasing weight made no difference. Previous studies have suggested a bidirectional association between overweight and obesity and depressive symptoms.46 47 However, our findings are not consistent with this. One possible explanation is that multimorbidity is controlled and one of the main hypotheses for the association between obesity and depression is the multimorbidity/depression association with other chronic conditions.48 This could also be a result of the age of individuals in our study. Most studies focusing on the association between obesity and depression include younger cohorts.

    With regard to the effect of multimorbidity, a higher number of chronic conditions were more prevalent among those with depressive symptoms. This relationship is longitudinal, as the odds of depressive symptoms increased significantly among older adults with higher number of chronic conditions, as well as among those with an increasing number of chronic conditions. These findings are in line with multiple studies. Berkman et al found that older adults with physical illnesses and disabilities reported many somatic complaints. All complaints were part of the depressive symptomatology, altering their score on the CESD scale.49 Recently, Zuh et al 50 evidenced that a higher chronic disease burden is a consistent risk factor for unfavourable depression trajectories in middle and old-age individuals. Furthermore, findings from Birk et al suggest that coronary artery disease, chronic arthritis and stroke are associated with increased probabilities of depression.18 Older people with clinically substantial and persistent depressive symptoms should be evaluated for both chronic conditions and risk factors.51

    Similarly, a higher percentage of respondents from the group with depressive symptoms also reported being in poorer health (24% vs 1%, who reported excellent and very good health). In addition, the longitudinal analyses showed that belonging to the group with lower self-rated health was associated with increased odds of depressive symptoms (OR 4.68). The worsening of self-rated health also increased these odds (OR 1.71) in the fully adjusted model. This association has been shown to occur cross-sectionally and longitudinally.51 52 Goldney et al 53 suggested that a poorer self-rated health could mediate the relationship between multimorbidity and depressive symptoms. However, our study shows that these two variables have an independent effect. Respondents with multiple conditions had higher odds of suffering from depressive symptoms when controlling for self-rated health. Requesting self-rated health for patients at risk for or diagnosed with depression has been shown to be relevant. It is not only related to a higher prevalence of depressive symptoms but also to poor long-term prognosis of depression outcomes and difficult treatment.52

    Results regarding the effects of disability on depressive symptoms show a strong association between functional disability and depressive symptoms, as it has often been reported in older adults (19, 26 and 51). This is attributed to the disabling effects of depression, as well as to the depressive effects of physical health-related disability, which reinforce each other over time.18 52–54 While cross-sectional analyses from 2001 did not show a significant difference in the prevalence of IADL impairments between the groups with and without depressive symptoms, findings from the longitudinal analyses showed that having one or more IADL impairments and having a new IADL impairment were associated with higher odds of depressive symptoms. Considering the effects of ADL impairments, the group with depressive symptoms reported a significantly higher prevalence of ADL impairments (4% vs 15%). Also, having a higher number of ADL impairments or an increasing number of impairments was also associated with higher odds of depressive symptoms in the longitudinal analyses.

    Strengths and weaknesses of the study

    The two main strengths of the study are the sample and analysis approach. The longitudinal perspective allows tracking changes over time, providing a comprehensive understanding of patterns and possible causes of depressive symptoms. On the other hand, a nationally representative survey ensures a diverse and inclusive sample, enhancing the generalisability of findings to the broader population. The richness of data enabled the exploration of various potential contributors to depressive symptoms, from demographic, socioeconomic and health characteristics. It also allowed for thorough control of confounding factors.

    The analysis strategy of using between-within panel data analysis enabled the exploration of individual and contextual factors influencing depressive symptoms. By employing panel data analysis, it was possible to make robust claims about causality.

    The main limitations of the study include survey biases. While using a nationally representative sample is a strength of the study, conditioning is a panel bias that can impact the outcome. Long-term participation in the study may influence participants’ responses and attitudes.55 Also, despite being a nationally representative sample, it does not portray the regional and sociocultural variations of the country and how such variations impact depressive symptoms. This is relevant because Mexico is one of the most culturally diverse countries and this is not reflected in the results. Lastly, the variables included in the model are those included in the survey. This means that other relevant factors contributing to depressive symptoms were not analysed. However, recognising differences in the results of the between and within effects helped recognise heterogeneity, a characteristic in the group that was not considered.

    Implications of findings

    Findings from this research underscore the need for comprehensive public policies targeting mental health in older populations. Tailoring interventions to address specific risk factors, such as unemployment and health conditions, could enhance overall well-being. Given the association between lower educational attainment and depression, this study suggests the importance of educational initiatives aimed at empowering individuals with knowledge and resources to mitigate mental health risks. Strategies to prevent job loss and support reemployment, especially for older individuals, could serve as crucial components of mental health interventions, which should be aligned with the identified risk factor of job loss. This study emphasises the significance of holistic healthcare, incorporating both physical and mental health considerations. Integration of mental health assessments into routine health check-ups for older adults could facilitate early identification and intervention. Tailoring community programmes to address the specific needs of older individuals, including those with chronic conditions or impairments, can contribute to fostering social support networks and reducing the risk of depressive symptoms. Policy-makers should consider the allocation of resources toward mental health services for older adults, considering the identified factors contributing to depressive symptoms. This may include funding for mental health screenings and targeted interventions. Launching public awareness campaigns to destigmatise mental health issues among older adults and promote the importance of seeking help could encourage early intervention and support.

    In summary, the findings highlight the need for a multifaceted approach, combining healthcare, social and policy interventions to promote mental health and well-being in the ageing population.

    Future research

    Future research should focus on using more tailored surveys from different regions in Mexico. As acknowledged in the limitations of the study, despite this being a nationally representative sample, there are differences in the demographic, socioeconomic and health characteristics between areas. It should also focus on studying the impact of policies targeting each of the areas influencing the prevalence of depressive symptoms in older adults.

    Conclusions

    Using a nationally representative study of ageing, a prevalence of 38% of Mexican older adults with depressive symptoms was found in this study. Such prevalence has increased by 3% in the past decade. This may be because of the higher prevalence of multimorbidity, disabilities and worsening of socioeconomic conditions experienced by this age group. The fully adjusted model, which included demographic, socioeconomic, health and disability characteristics, highlights that an increasing age, being a woman, having fewer years of education, having lower net worth, losing a job, having a higher and increasing number of chronic conditions, having poorer and decreasing self-rated health, and having IADL or ADL impairments s increase significantly the likelihood of having depressive symptoms in this population. This is the first longitudinal study to address depressive symptoms as the result of a combination of demographic, socioeconomic and health factors. This highlights the importance of holistic health and public policy interventions to improve the well-being of older adults.

    The corresponding author has the right to grant on behalf of all authors and does grant on behalf of all authors, a worldwide licence to the publishers and its licences in perpetuity, in all forms, formats and media (whether known now or created in the future), to (1) publish, reproduce, distribute, display and store the contribution, (2) translate the contribution into other languages, create adaptations, reprints, include within collections and create summaries, extracts and/or abstracts of the contribution, (3) create any other derivative work(s) based on the contribution, (4) to exploit all subsidiary rights in the contribution, (5) the inclusion of electronic links from the contribution to third party material wherever it may be located and (6) licence any third party to do any or all of the above.

    Data availability statement

    Data sharing not applicable as no datasets generated and/or analysed for this study. Data are public and available at https://www.mhasweb.org/Home/index.aspx.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This is a secondary data analysis of the MHAS (Mexican Health and Aging Study), which was sponsored by the National Institutes of Health/National Institute on Aging. Ethics approval was granted by the National Institute of Statistics and Geography (INEGI) in Mexico and by the NIH in the USA with the grant number NIH RO1AG018016.

    References

    1. ↵
      1. WHO
      . Depression and other common mental disorders global health estimates. 2017.
    2. ↵
      2. McCall WV ,
      3. Kintziger KW
      . Late life depression. Psychiatr Clin North Am 2013;36:S0193-953X(13)00093-2. doi:10.1016/j.psc.2013.08.011
    3. ↵
      2. Stordal E ,
      3. Mykletun A ,
      4. Dahl AA
      . The association between age and depression in the general population: a multivariate examination. Acta Psychiatr Scand 2003;107:132–41. doi:10.1034/j.1600-0447.2003.02056.x
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      2. Barua A ,
      3. Ghosh MK ,
      4. Kar N , et al
      . Prevalence of depressive disorders in the elderly. Ann Saudi Med 2011;31:620–4. doi:10.4103/0256-4947.87100
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      2. Blazer DG
      . Depression in late life: review and commentary. J Gerontol A Biol Sci Med Sci 2003;58:249–65. doi:10.1093/gerona/58.3.m249
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      2. Guerra M ,
      3. Ferri CP ,
      4. Sosa AL , et al
      . Late-life depression in Peru, Mexico and Venezuela: the 10/66 population-based study. Br J Psychiatry 2009;195:510–5. doi:10.1192/bjp.bp.109.064055
      OpenUrlAbstract/FREE Full Text
    7. ↵
      2. Alvarez-Cisneros T ,
      3. Roa-Rojas P ,
      4. Garcia-Peña C
      . Longitudinal relationship of diabetes and depressive symptoms in older adults from Mexico: A secondary data analysis. BMJ Open Diabetes Res Care 2020;8:8–2:e001789. doi:10.1136/bmjdrc-2020-001789
      OpenUrl
    8. ↵
      2. Gallegos-Carrillo K ,
      3. García-Peña C ,
      4. Mudgal J , et al
      . Role of depressive symptoms and comorbid chronic disease on health-related quality of life among community-dwelling older adults. J Psychosom Res 2009;66:127–35. doi:10.1016/j.jpsychores.2008.07.007
      OpenUrlCrossRefPubMed
    9. ↵
      2. Fernández-Niño JA ,
      3. Manrique-Espinoza BS ,
      4. Bojorquez-Chapela I , et al
      . Income inequality, socioeconomic deprivation and depressive symptoms among older adults in Mexico. PLoS ONE 2014;9:e108127. doi:10.1371/journal.pone.0108127
    10. ↵
      2. García-Pérez A ,
      3. González-Aragón Pineda AE ,
      4. Sandoval-Bonilla BA , et al
      . Prevalence and factors associated with the depressive symptoms in rural and urban Mexican older adults: evidence from the Mexican health and aging study 2018. Salud Publica Mex 2022;64:367–76. doi:10.21149/13340
      OpenUrl
    11. ↵
      2. Internacional Raffaelli O ,
      3. Andrade D ,
      4. Cristina F , et al
      . Related differences in depressive symptoms and health behaviors among. Mex Youth Interam J Psychol 2013;47:83–90. Available: http://www.redalyc.org/articulo.oa?id=28426980010
      OpenUrl
    12. ↵
      2. Martínez-Mendoza JA ,
      3. Martínez-Ordaz VA ,
      4. Esquivel-Molina CG , et al
      . Prevalencia de Depresión Y Factores de Riesgo en El Adulto Mayor Hospitalizado. Rev Med Inst Mex Seguro Soc 2007;45:21–8.
      OpenUrlPubMed
    13. ↵
      2. Pando Moreno M ,
      3. Aranda Beltrán C ,
      4. Alfaro Alfaro N , et al
      . Prevalencia de Depresion en Adultos Mayores en Poblacion Urbana. Rev Esp Geriatr Gerontol 2001;36:140–4. doi:10.1016/S0211-139X(01)74704-X
      OpenUrl
    14. ↵
      2. García-Peña C ,
      3. Wagner FA ,
      4. Sánchez-Garcia S , et al
      . Depressive symptoms among older adults in Mexico city. J Gen Intern Med 2008;23:1973–80. doi:10.1007/s11606-008-0799-2
      OpenUrlCrossRefPubMed
    15. ↵
      2. Wagner FA ,
      3. Gallo JJ ,
      4. Delva J
      . Depression in late life: A hidden public health problem for Mexico? Salud Publica Mex 1999;41:189–202. doi:10.1590/s0036-36341999000300007
      OpenUrlCrossRefPubMed
    16. ↵
      2. Fiske A ,
      3. Wetherell JL ,
      4. Gatz M
      . Depression in older adults. Annu Rev Clin Psychol 2009;5:363–89. doi:10.1146/annurev.clinpsy.032408.153621
      OpenUrlCrossRefPubMed
    17. ↵
      2. Parajuli J ,
      3. Berish D ,
      4. Jao YL
      . Chronic conditions and depressive symptoms in older adults: the mediating role of functional limitations. Aging Ment Health 2021;25:243–9. doi:10.1080/13607863.2019.1693971
      OpenUrl
    18. ↵
      2. Birk JL ,
      3. Kronish IM ,
      4. Moise N , et al
      . Depression and Multimorbidity: considering temporal characteristics of the associations between depression and multiple chronic diseases. Health Psychol 2019;38:802–11. doi:10.1037/hea0000737
      OpenUrlCrossRefPubMed
    19. ↵
      2. Jiang J ,
      3. Tang Z ,
      4. Futatsuka M
      . The impact of ADL disability on depression symptoms in a community of Beijing elderly, China. Environ Health Prev Med 2002;7:199–204. doi:10.1007/BF02898005
      OpenUrlCrossRefPubMed
    20. ↵
      2. Spijker J ,
      3. Graaf R ,
      4. Bijl RV , et al
      . Functional disability and depression in the general population. results from the Netherlands mental health survey and incidence study (NEMESIS). Acta Psychiatr Scand 2004;110:208–14. doi:10.1111/j.1600-0447.2004.00335.x
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      2. Qiu Q ,
      3. Qian S ,
      4. Li J , et al
      . Risk factors for depressive symptoms among older Chinese adults: A meta-analysis. J Affect Disord 2020;277:341–6. doi:10.1016/j.jad.2020.08.036
      OpenUrl
    22. ↵
      2. Rosenquist JN ,
      3. Fowler JH ,
      4. Christakis NA
      . Social network determinants of depression. Mol Psychiatry 2011;16:273–81. doi:10.1038/mp.2010.13
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      2. Agudelo-Cifuentes M-C ,
      3. Cardona D ,
      4. Segura A , et al
      . Vulneración a La Integridad Física de Personas Mayores: Prevalencia del Maltrato Y Factores Asociados. Agora USB 2020;20:129–39. doi:10.21500/16578031.4568
      OpenUrl
    24. ↵
      2. Gupta S ,
      3. Bélanger E ,
      4. Phillips SP
      . Low socioeconomic status but resilient: panacea or double trouble? John Henryism in the International IMIAS study of older adults. J Cross Cult Gerontol 2019;34:15–24. doi:10.1007/s10823-018-9362-8
      OpenUrl
    25. ↵
      2. Picazzo-Palencia E
      . Depression and mobility among older adults in Mexico. Hisp Health Care Int 2016;14:94–8. doi:10.1177/1540415316650846
      OpenUrlCrossRefPubMed
    26. ↵
      2. Friedrich MJ
      . Depression is the leading cause of disability around the world. JAMA 2017;317:1517. doi:10.1001/jama.2017.3826
      OpenUrlCrossRefPubMed
    27. ↵
      2. de Lima Filho BF ,
      3. da Nóbrega Dias V ,
      4. Carlos AG , et al
      . Factors related to depressive symptoms in older adult patients with type 2 diabetes mellitus. Exp Gerontol 2019;117:72–5. doi:10.1016/j.exger.2018.11.006
      OpenUrl
    28. ↵
      2. Cuijpers P ,
      3. Vogelzangs N ,
      4. Twisk J , et al
      . Comprehensive meta-analysis of excess mortality in depression in the general community versus patients with specific illnesses. AJP 2014;171:453–62. doi:10.1176/appi.ajp.2013.13030325
      OpenUrl
    29. ↵
      2. Jiang C ,
      3. Hu C ,
      4. Li D
      . Linking chronic diseases, cognitive impairment and depressive symptoms in Chinese older adults: A three-wave population-based longitudinal study. J Affect Disord 2023;324:496–501. doi:10.1016/j.jad.2022.12.150
      OpenUrl
    30. ↵
      2. Chen CM ,
      3. Su YY ,
      4. Lan TY , et al
      . Health-related services use and the onset of functional disability: 10 year follow-up study. Arch Gerontol Geriatr 2014;58:356–63. doi:10.1016/j.archger.2013.11.011
      OpenUrl
    31. ↵
      2. Liu H ,
      3. Fan X ,
      4. Luo H , et al
      . n.d. Comparison of depressive symptoms and its influencing factors among the elderly in urban and rural areas: evidence from the China health and retirement longitudinal study (CHARLS). IJERPH 18:3886. doi:10.3390/ijerph18083886
    32. ↵
      2. Pengpid S ,
      3. Peltzer K ,
      4. Anantanasuwong D
      . Prevalence and determinants of incident and persistent depressive symptoms among middle-aged and older adults in Thailand: prospective cohort study. BJPsych Open 2023;9:e99. doi:10.1192/bjo.2023.72
    33. ↵
      2. Wong R ,
      3. Michaels-Obregon A ,
      4. Palloni A
      . Cohort profile: the Mexican health and aging study (MHAS). Int J Epidemiol 2017;46:e2. doi:10.1093/ije/dyu263
    34. ↵
      2. Karim J ,
      3. Weisz R ,
      4. Bibi Z , et al
      . Validation of the eight-item center for epidemiologic studies depression scale (CES-D) among older adults. Curr Psychol 2015;34:681–92. doi:10.1007/s12144-014-9281-y
      OpenUrlCrossRef
    35. ↵
      2. Van de Velde S ,
      3. Levecque K ,
      4. Bracke P
      . Measurement equivalence of the CES-D 8 in the general population in Belgium: a gender perspective. Arch Public Health 2009;67. doi:10.1186/0778-7367-67-1-15
    36. ↵
      2. Aguilar-Navarro SG ,
      3. Fuentes-Cantú A ,
      4. Avila-Funes JA , et al
      . Validity and reliability of the screening questionnaire for geriatric depression used in the Mexican health and age study. Salud Publica Mex 2007;49:256–62. doi:10.1590/s0036-36342007000400005
      OpenUrlCrossRefPubMed
    37. ↵
      2. Torres JM ,
      3. Wong R
      . Childhood poverty and depressive symptoms for older adults in Mexico: a life-course analysis. J Cross Cult Gerontol 2013;28:317–37. doi:10.1007/s10823-013-9198-1
      OpenUrlCrossRefPubMed
    38. ↵
      2. Allison PD
      . Fixed effects regression models. 2455 Teller Road,Thousand Oaks, California, 91320, United States of America: Sage, 2009. Available: https://methods.sagepub.com/book/fixed-effects-regression-models
    39. ↵
      2. Xie Y ,
      3. Ma M ,
      4. Wu W , et al
      . Factors associated with depressive symptoms among the elderly in China: structural equation model. Int Psychogeriatr 2021;33:157–67. doi:10.1017/S1041610220001398
      OpenUrl
    40. ↵
      2. Urbina Torija JR ,
      3. Flores Mayor JM ,
      4. García Salazar MP , et al
      . Síntomas Depresivos en Personas Mayores. Prevalencia Y Factores Asociados. Gac Sanit 2007;21:37–42. doi:10.1157/13099119
      OpenUrlPubMed
    41. ↵
      2. Cacioppo JT ,
      3. Hawkley LC ,
      4. Crawford LE , et al
      . Loneliness and health: potential mechanisms. Psychosom Med 2002;64:407–17. doi:10.1097/00006842-200205000-00005
      OpenUrlAbstract/FREE Full Text
    42. ↵
      2. Lorant V ,
      3. Croux C ,
      4. Weich S , et al
      . Depression and socio-economic risk factors: 7-year longitudinal population study. Br J Psychiatry 2007;190:293–8. doi:10.1192/bjp.bp.105.020040
      OpenUrlAbstract/FREE Full Text
    43. ↵
      2. Lorant V ,
      3. Deliège D ,
      4. Eaton W , et al
      . Socioeconomic inequalities in depression: A meta-analysis. Am J Epidemiol 2003;157:98–112. doi:10.1093/aje/kwf182
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      2. Errazuriz A ,
      3. Avello-Vega D ,
      4. Ramirez-Mahaluf JP , et al
      . Prevalence of depressive disorder in the adult population of Latin America: a systematic review and meta-analysis. Lancet Reg Health Am 2023;26:100587. doi:10.1016/j.lana.2023.100587
    45. ↵
      2. Poleshuck E ,
      3. Wittink M ,
      4. Crean HF , et al
      . A comparative effectiveness trial of two patient-centered interventions for women with unmet social needs: personalized support for progress and enhanced screening and referral. J Womens Health (Larchmt) 2020;29:242–52. doi:10.1089/jwh.2018.7640
      OpenUrl
    46. ↵
      2. Werrij MQ ,
      3. Mulkens S ,
      4. Hospers HJ , et al
      . Overweight and obesity: the significance of a depressed mood. Patient Educ Couns 2006;62:126–31. doi:10.1016/j.pec.2005.06.016
      OpenUrlCrossRefPubMed
    47. ↵
      2. Hung C-F ,
      3. Rivera M ,
      4. Craddock N , et al
      . Relationship between obesity and the risk of clinically significant depression: Mendelian Randomisation study. Br J Psychiatry 2014;205:24–8. doi:10.1192/bjp.bp.113.130419
      OpenUrlAbstract/FREE Full Text
    48. ↵
      2. Kostka J ,
      3. Borowiak E ,
      4. Kostka T
      . Nutritional status and quality of life in different populations of older people in Poland. Eur J Clin Nutr 2014;68:1210–5. doi:10.1038/ejcn.2014.172
      OpenUrl
    49. ↵
      2. Berkman LF ,
      3. Berkman CS ,
      4. Kasl S , et al
      . Depressive symptoms in relation to physical health and functioning in the elderly. Am J Epidemiol 1986;124:372–88. doi:10.1093/oxfordjournals.aje.a114408
      OpenUrlPubMedWeb of Science
    50. ↵
      2. Zhu J ,
      3. Xiang Y-B ,
      4. Cai H , et al
      . Associations of obesity and weight change with physical and mental impairments in elderly Chinese people. Maturitas 2018;108:77–83. doi:10.1016/j.maturitas.2017.11.015
      OpenUrl
    51. ↵
      2. Atlantis E ,
      3. Fahey P ,
      4. Foster J
      . Collaborative care for comorbid depression and diabetes: a systematic review and meta-analysis. BMJ Open 2014;4:e004706. doi:10.1136/bmjopen-2013-004706
    52. ↵
      2. Ambresin G ,
      3. Chondros P ,
      4. Dowrick C , et al
      . Self-rated health and long-term prognosis of depression. Ann Fam Med 2014;12:57–65. doi:10.1370/afm.1562
      OpenUrlAbstract/FREE Full Text
    53. ↵
      2. Goldney RD ,
      3. Eckert KA ,
      4. Hawthorne G , et al
      . Changes in the prevalence of major depression in an Australian community sample between 1998 and 2008. Aust N Z J Psychiatry 2010;44:901–10. doi:10.3109/00048674.2010.490520
      OpenUrlCrossRefPubMed
    54. ↵
      2. Ormel J ,
      3. Rijsdijk FV ,
      4. Sullivan M , et al
      . Temporal and reciprocal relationship between IADL/ADL disability and depressive symptoms in late life. J Gerontol B Psychol Sci Soc Sci 2002;57:P338–47. doi:10.1093/geronb/57.4.p338
      OpenUrlCrossRef
    55. ↵
      2. Bach RL
      . A methodological framework for the analysis of panel conditioning effects. measurement error in longitudinal data. 2021;19–42. doi:10.1093/oso/9780198859987.001.0001

    Footnotes

    • Contributors The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. TA-C conceived of the presented idea and is the author responsible for the overall content. TA-C and MLG-R developed the theory, downloaded the dataset and performed most statistical analyses. ST-C and PAR-R contributed to the interpretation of results, writing the manuscript and providing feedback for the writing of the manuscript. All authors provided ideas for the revision and final version of the manuscript.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. However, funding for the publication of this article was provided by the Instituto Nacional de Geriatría in Mexico City.

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.