Abstract
We aimed to determine the association between the duration of widowhood and cognition decline. We compared the decline observed in widowed people compared with married, single, or separated persons using the scores obtained in the cognitive assessment of memory, learning, and visual exploration by adults and older adults in Mexico. The Mexican Health and Aging Study (MHAS) provides the base for this paper. This study is an analysis of the fourth data wave (2015), except for the independent variable: marital status. Marital status was built longitudinally with information from the four surveys (2001, 2003, 2012, and 2015). The sample comprised 6898 adults aged 50 and over. Cognition was assessed with an adapted Cross-Cultural Cognitive Examination (CCCE). Confounders include sociodemographic characteristics (sex, age, schooling, self-perception of economic status, and whether the individual worked or not), multimorbidity, functionality, support networks, and psychological characteristics. Of the total sample, 4094 (59.3%) were women. The mean age was 70.86 years (SD = 7.4). The baseline of the study is 2001. In 2001, 8.7% (n = 600) were widows or widowers. People widowed by 2001 scored −0.158 points in cognition while divorced participants scored - 0.095 points.
Keywords: Widowhood, Cognition, Civil status
Introduction
Life expectancy has increased significantly in recent decades. Science and technology have changed the causes of early death. Medical and pharmacological advances, better hygienic conditions, and the progress achieved in the control of infectious diseases have increased people’s life expectancy. This fact makes aging one of the central concerns of the contemporary world. The population aged 80 or more will likely increase in the coming years (Ardila and Rosselli 2007). This phenomenon occurs in both industrialized countries and developing ones. In Mexico, between 1950 and 2020, the number of people over 60 years of age increased from 2 to 15 million individuals (d’Hyver and Gutierrez Robledo 2014; INEGI 2017). According to data from the National Institute of Statistics and Geography (INEGI), the proportion of death in men and women is different; men die more than women, which results in more women than men in advanced ages (INEGI 2017).
Older adults face different individual, social, and physical situations and adjustments. Some studies have sought to associate several social factors with cognitive skills, mainly social vulnerability (Andrew and Rockwood 2010). Some other investigations have demonstrated an association between cognitive impairment and better health mediated by social support networks, marital status, and participation in social activities (Van Gelder et al. 2006; Waldron 1996; Zunzunegui et al. 2003). Social conditions are crucial for older people. Living alone can have a deleterious effect on cognition (Gibson and Richardson 2017). Also, older people who have an unhappy marriage can present more symptoms of insomnia. In contrast, married people who have a good relationship with their partner, have better physical and mental well-being (Chen et al. 2015).
Cognitive impairment is associated with poor social connections and social disengagement. Little participation in social activities is an example of poor social connections, and little or no visual contact with friends, infrequent attendance to religious services and sporadic participation in social and recreational activities exemplify social disengagement (Bassuk et al. 1999; Zunzunegui et al. 2003). The association may vary by sex; Waldron et al. showed that women who were not married had poorer health (Waldron 1996). Also, widowed people present deficiencies in cognitive performance compared with people with a partner. These deficiencies were already noteworthy two years after the loss of the spouse and got worse with longer durations of widowhood (Biddle et al. 2020).
Marriage seems to have a positive impact on the health of older adults. Various studies have investigated this relationship. Married adults are generally in better health than unmarried adults. Marital status is also related to mortality; there are higher mortality rates in older adults who are not married than in older adults who are married (Bangerter and Waldron 2014; Johnson et al. 2000; Seeman 1996; Waldron 1996). A study investigating the factors associated with cognitive impairment concluded that women had a higher risk but married people a lower one; this result is consistent with some research conducted in the United States (Mejía-Arango et al. 2007). Widowhood can damage the psychological and emotional status of elderly persons and could be associated with a deterioration in their mental health (Monserud and Wong 2015).
A longitudinal study that assessed older men who lived alone for five years reported that they had at least two times more cognitive damage than men who were married or were living with someone during this period (Van Gelder et al. 2006). A study in Colombia looked into the association between the risk of cognitive impairment and age in people over 60 years old. It found that 83.1% of the individuals evaluated had an increased risk of impaired cognition. Among other variables, this problem was more associated with women and individuals who had no spouse or partner (Segura Cardona et al. 2016).
Marital status and mortality are also linked; there are higher mortality rates in older adults who are not married than among married ones (Bangerter et al., 2014; Johnson et al. 2000; Seeman 1996).
The relationship between health status, cognition, and associated factors, as well as the adverse consequences related to old age, have been investigated thoroughly (Carr et al. 2000; Mejía-Arango et al. 2007; Monserud et al., 2015). However, impaired cognition directly related to widowhood has received less attention.
Given the importance of cognitive impairment in the elderly, we aimed to determine the association between widowhood duration and cognition decline; compared with being married, single, or separated using the scores obtained by the cognitive assessment of memory, learning, and visual exploration in adults and older adults in Mexico.
Material and Methods
In this investigation, we examined the population surveyed in the Mexican Health and Aging Study (MHAS). The MHAS is a longitudinal study of adults aged 50 years and over. It is representative of 13 million Mexicans born before 1951 nationwide (MHAS, 2013). The purpose and design of the MHAS have been described elsewhere (Wong et al. 2017, 2015).
The MHAS consists of four data waves. The baseline survey was conducted in 2001; subsequent waves were carried out in 2003, 2012, and 2015, with a total of 15,402; 18,465, and 15,898 individuals surveyed, respectively. The total number of deaths amounted to 4491 by 2015.
A set of questionnaires (sociodemographic, health-related, cognitive performance, functional status, and others) was applied by standardized interviewers trained in the objectives of the survey, the questionnaires, and the face-to-face interviewing process. This study is an analysis of the fourth data wave (2015) except for the independent variable: marital status. Marital status was built longitudinally, with information from the four surveys (2001, 2003, 2012, and 2015). In 2015, 14,779 subjects participated in the survey. Losses in waves 2001, 2003, and 2012 add up to 5848 cases that correspond to people who died (4491) or no longer participated in any wave of the study (1357). Thus, the total number of cases with follow-up of marital status by 2015 was 8931.
We excluded 72 people under 50 years old, 1165 who did not answer one or more questions on the cognitive scales or some explanatory variables, and 691 whose questionnaire was answered by a family surrogate. We also excluded 105 cases of people who answered that they have been widows on more than one occasion. The sample includes people who answered every question on the cognition scales—verbal learning, verbal learning recall, figure copying, recall of copied figure, and visual perception, the final sample included 6898 participants. Figure 1 shows the flow of the sample selection.
Fig. 1.
Flow of the sampling process
Dependent Variable
To assess cognition, the MHAS includes an adapted Cross-Cultural Cognitive Examination (CCCE) (Glosser et al. 1993; Wong et al. 2017). The adapted CCCE evaluates: verbal memory (the person must learn and recall a list of words repeated three times, with scores from 0 to 8 for learn and 0 to 8 for recall), visuo-constructive skills (the interviewee must copy the figure indicated and recall that figure, with scores from 0 to 6 for copy and 0 to 6 for visual memory), and visual scanning (the person highlights the stimulus figure shown, with scores from 0 to 60). The maximum score for the cognitive scales used is 88 points (Mejía-Arango et al. 2015).
The total scores obtained from the five selected scales were added to analyze the cognition in 2015. The maximum possible is 88 points (Mejía-Arango et al. 2015). Next, raw cognition scores were standardized to Z-scores.
Independent Variable
The independent variable was marital status, defined by the condition of widowhood, singleness, divorce or separation, or marriage of the person. The category associated with widowhood was longitudinally constructed using the four waves of the MHAS to detect how long the person had been a widower during the 14 years encompassed by the MHAS, since the first wave in 2001.
There are six categories within the variable of marital status: 1) Widowed more than 11 years (from the study baseline and according to a variable measured in 2001 concerning when widowhood occurred, more than 14 years). 2) Widowed for 4 to 11 years, i.e., those who were widowed by 2003 and until 2015, 3) Widowed for 3 years, i.e., those who were widowed by 2012 and remained so in 2015. We also considered the following categories that were analyzed with the information associated with the wave 2015: 4) single, 5) separated or divorced, and 6) living with a partner.
Confounders
Confounders include sociodemographic characteristics (sex, age, schooling, self-perception of economic status, and whether the individual worked or not), multimorbidity, functionality, support networks, and psychological characteristics (Whitley et al. 2016; Tous and Navarro 1997; Torres 2011; Dorantes-Mendoza et al. 2007).
Participants self-reported chronic diseases diagnosed by a physician or medical personnel: diabetes mellitus, heart attack, hypertension, and stroke. MHAS assessed depressive symptoms using the modified nine-item CES-D (Center for Epidemiological Studies-Depression). Persons reported the presence of nine depression symptoms in the previous week: 1) feeling depressed, 2) feeling that everything he/she did was an effort, 3) feeling that his/her sleep was restless, 4) feeling happy, 5) feeling lonely, 6) feeling that he/she enjoyed life, 7) feeling sad, 8) feeling tired and 9) feeling that he/she had a lot of energy. Items 4, 6, and 9 have a reverse code. There were two possible responses: (yes/no). Five or more responses indicating depressive symptoms meant the individual had a degree of depression (Agudelo-Botero et al. 2018; Aguilar-Navarro et al. 2007).
The Katz Activities of Daily Living scale was used to assess functionality (Katz et al. 1970); participants report whether they needed help to walk around the house, bathe, use the toilet, and get in/out of bed. The number of limitations was also analyzed.
The psychological variables and support networks characteristics include: a feeling of isolation (whether the participants felt isolated or away from other people), independent decision making (whether the individuals make their own decisions or need help by someone else), locus of control, (four questions, scored 1 to 4, measure the variable internal locus of control; scores under the average denote internal locus of control and higher scores, the absence of internal locus of control), having health insurance, and attendance to religious services.
The interviewee expressed satisfaction with life through five statements that allow the person to say whether she would change something in her life or is content with it.
Plan of Analysis
Descriptive analyses of the sample were performed to obtain the distribution of each of the study variables.
Bivariate analyses were performed between the standardized cognition score and each explanatory and confounding variable to understand the difference in cognition according to groups applying analysis of variance (ANOVA), or Pearson correlation.
Our study analyzes the effect of civil status (particularly widowhood and its length) on cognition through multiple linear regression analysis. We used the standardized cognition score as a dependent variable. We controlled for sex, age, schooling, economic perception, and clinical characteristics such as depressive symptoms, hypertension, diabetes mellitus, heart attack, and stroke. We also controlled for variables of functionality in basic activities of daily living, such as difficulties with bathing, walking, eating, going to bed, and using the toilet. Social and personality variables, such as feelings of isolation, internal locus of control, own decision making, having access to health care services, attendance to religious services, and feelings of satisfaction with life, were also controlled for.
The MHAS was approved by the institutional review boards and the ethics committees of the University of Texas Medical Branch, in the United States, and the National Institute of Statistics and Geography (INEGI) and the National Institute of Public Health (INSP) in Mexico. The current analysis was recorded at the National Institute of Geriatrics (DI-PI-006/2017).
Results
Descriptive Results
In our sample, there were 4094 (59.3%) women; of them 2175 (53.1%) had a spouse/partner and 1247 (18%) were widowed. Men accounted for 2804 (40.6%) of the participants; of them 2174 (31.5%) had a spouse/partner and 377 (5.4%) were widowed.
The marital status of the participants was distributed as follows: 8.6% (n = 600) were widowed for more than 11 years, 8.6% (n = 596) were widowed between 4 and 11 years, and 6.2% (n = 428) were widowed for less than 4 years, 3.3% (n = 230) of the respondents were single, 10% (n = 695) were separated or divorced, and 63% (n = 4, 349) had a partner (Table 1).
Table 1.
Socio-demographic and clinical characteristics of the studied sample (2015)
| Sex | Total | ||
|---|---|---|---|
| Women n (%) | Men n (%) | N = 6898 (%) | |
| n = 4094 (59.3%) | n = 2804 (40.6%) | ||
| Marital status* | |||
| Widower more than 11 years | 505 (12.2) | 95 (3.3) | 600 (8.6) |
| Widower from 4 to 11 years | 450 (10.9) | 146 (5.2) | 596 (8.6) |
| Widower less than 4 years | 292 (7.1) | 136 (4.8) | 428 (6.2) |
| Single in 2015 | 174 (4.2) | 56 (1.9) | 230 (3.3) |
| Separated or Divorced in 2015 | 498 (12.1) | 197 (7) | 695 (10) |
| With couple in 2015 | 2175 (53.1) | 2174 (77.5) | 4349 (63) |
| Age | |||
| 50–59 | 361 (8.8) | 22 (0.7) | 383 (5.5) |
| 60–69 | 1760 (42.9) | 1089 (38.8) | 2849 (41.3) |
| 70–79 | 1497 (36.5) | 1271 (45.3) | 2768 (40.1) |
| 80 and more | 476 (11.6) | 422 (15) | 898 (13) |
| Mean (SD) | 69.79 (7.79) | 72.41 (6.56) | 70.86 (7.43) |
| Education | |||
| 0 years | 722 (17.6) | 426 (15.2) | 1148 (16.6) |
| 1–4 years | 1324 (32.3) | 873 (31.1) | 2197 (31.8) |
| 5–9 years | 1678 (40.9) | 1052 (37.5) | 2730 (39.5) |
| More than 10 years | 370 (9) | 453 (16.1) | 823 (11.9) |
| Mean (SD) | 5.04 (5.04) | 5.82 (6.22) | 5.35 (5.56) |
| Socio-economic perception | |||
| Bad | 3110 (75.9) | 2146 (76.5) | 5256 (76.1) |
| Good | 984 (24) | 658 (23.4) | 1642 (23.8) |
| Current job | |||
| No | 3318 (81) | 1529 (54.5) | 4847 (70.2) |
| Yes | 776 (18.9) | 1275 (45.4) | 2051 (29.7) |
| Self-reported diseases | |||
| Hypertension | 1269 (18.5) | 2423 (35.1) | 3692 (53.5%) |
| Diabetes Mellitus | 657 (9.5) | 1179 (17) | 1836 (26.6%) |
| Heart Attack | 160 (2.3) | 157 (2.2) | 317 (4.6%) |
| Stroke | 64 (0.9) | 70 (1) | 134 (1.9%) |
| Other variables | |||
| Depressive symptoms | 667 (9.6) | 1512 (21.9) | 2179 (31.5%) |
| Life Satisfaction | 1187 (17.2) | 1569 (22.7) | 2756 (39.9%) |
| Attendance at religious services | 2252 (32.6) | 3604 (52.2) | 5856 (84.8%) |
| Health Services | 2568 (37.2) | 3847 (55.7) | 6415 (93%) |
| Make decisions on their own | 2243 (32.5) | 3239 (46.9) | 5482 (79.4%) |
| Not feeling isolated | 2293 (33.2) | 3137 (45.4) | 5430 (78.7%9 |
| Cognition (Rank, Mean, SD) | Min 0- max 87 43.22 (17.93) | Min 0- max 87 42.85 (17.29) | Min 0- max 87 43.07 (17.67) |
Source: Compilation based on the MHAS 2015.
The civil status section was built with the databases 2001, 2003, 2012 and 2015
The mean age of participants was 70.86 years (SD = 7.43). The minimum age was 50, and the maximum 113. The mean age of people widowed for more than 11 years was 76.3, those widowed for 4 to 11 years 73.2, and persons widowed for less than 4 years, 72.0. Single participants’ mean age was 72.5 years. Separated or divorced participants were 70.7 years old on average, and respondents with partners were 69.5 years old (data not shown).
Respondents 50 to 59 years old represented 5.5% (n = 383) of the sample, those 60 to 69 years old, 41.3% (n = 2, 849), those 70 to 79 years old, 40.1% (n = 2, 768), those 80 and more 13.0% (n = 898).
The participants perceived their economic situation as bad in 76.1% (n = 5, 256) of cases; while only 23.8% (n = 1, 642) declared to be in a good economic situation.
The years of schooling varied from none to more than 10 years; 16.6% (n = 1, 148) of respondents had no schooling at all, 31.8% (n = 2, 2197) had 1 to 4 years of schooling, 39.5% (n = 2, 730) had 5 to 9 years, and 11.9% (n = 823) had more than 10 years of schooling.
The clinical characteristics of the participants were as follows: 53.5% (n = 3, 692) reported hypertension, 26.6% (n = 1, 836) had diabetes mellitus, 4.6% (n = 317) reported they had had a heart attack, 1.9% (n = 134) had a stroke sometime in their lives, and 31.5% (n = 2, 179) had depressive symptoms (Table 2). As for psychosocial characteristics, 39.9% (n = 2, 756) of respondents had feelings of satisfaction with positive living, 84.8% (n = 5, 856) attended some religious service, 93.0% (n = 6, 415) had some type of health service, 79.4% (n = 5, 482) made personal decisions even though they had help from someone else, and 78.7% (n = 5430) reported no feelings of isolation (Table 1).
Table 2.
Bivariate analysis between cognition score and variables (wave 2015)
| Z-score Mean (SD) | P value | |
|---|---|---|
| Marital status | ||
| Widower more than 11 years | −.419 (.937) | <0.000 |
| Widower from 4 to 11 years | −.193 (.963) | |
| Widower less than 4 years | −.139 (1.01) | |
| Single in 2015 | .037 (1.04) | |
| Separated or Divorced in 2015 | −.009 (.975) | |
| With couple in 2015 | .129 (.970) | |
| Age (years) | ||
| 50–59 | .704 (.894) | <0.000 |
| 60–69 | .333 (.931) | |
| 70–79 | −.137 (.902) | |
| 80 and more | −.780 (.817) | |
| Sex | ||
| Men | −.007 (.966) | 0.39 |
| Women | .028 (1.00) | |
| Socioeconomic perception | ||
| Bad | −.071 (.938) | <0.000 |
| Good | .314 (1.08) | |
| Current job | ||
| Yes | .196 (.981) | <0.000 |
| No | −.054 (.981) | |
| Schooling (years) | ||
| None | −.849 (.710) | <0.000 |
| 1–4 | −.306 (.767) | |
| 5–9 | .357 (.874) | |
| >10 | .988 (.845) | |
| Hypertension | ||
| Yes | −.006 (.975) | 0.015 |
| No | .050 (1.00) | |
| Diabetes Mellitus | ||
| Yes | −.0473 (.934) | 0.001 |
| No | .043 (1.00) | |
| Heart Attack | ||
| Yes | .021 (.990) | 0.629 |
| No | −.006 (.927) | |
| Stroke | ||
| Yes | −.336 (.940) | 0.000 |
| No | .027 (.987) | |
| Depressive symptoms | ||
| Yes | −.203 (.926) | <0.000 |
| No | .123 (.998) | |
| Satisfaction in life | ||
| Yes | .024 (.989) | 0.650 |
| No | .013 (.984) | |
| Attendance at religious services | ||
| Yes | −.001 (.978) | 0.000 |
| No | .125 (1.03) | |
| Health services | ||
| Yes | .036 (.986) | 0.000 |
| No | −.201 (.981) | |
| Make decisions on their own | ||
| Yes | .122 (.985) | <0.000 |
| No | −.375 (.985) | |
| Not to be isolated | ||
| Yes | −.259 (.930) | <0.000 |
| No | .095 (.989) | |
| Internal locus of control | ||
| −0.031 (.012) | 0.009** | |
ANOVA
Correlation
Resource: Self-elaboration using MHAS data 2015
Standardized Cognition Score
The minimum score on cognition was 0, and the maximum score was 88. Scores were standardized to Z-values. None of the evaluated participants obtained the maximum score of 88 (rank min 0 max 87, mean 43.07, SD 17.67).
Bivariate Analysis
The mean cognition score significantly changed (p < 0.0001) depending on the participants’ marital status, age, economic perception, whether the person worked or not, and schooling (Table 2). The score by sex did not show a significant difference (p = 0.39). Regarding clinical characteristics, the mean cognition score changed significantly between individuals with and without depressive symptoms and having a stroke, hypertension, or diabetes. Concerning psychosocial characteristics, the variables with a significant change in the mean score were independent decision-making, attendance to religious services, and feelings of isolation.
Linear Regression
The model was adjusted for age, sex, economic perception, schooling, hypertension, diabetes mellitus, heart attack, stroke, depressive symptoms, feelings of isolation, difficulty to walk, bathe, eat, go to bed, or use the toilet, whether the person worked or not, independent decision-making, internal locus of control, having health services, satisfaction with life, and participation in religious services.
The F-test associated with the linear regression model had a p value of less than 0.0001, indicating that the independent variables in the model jointly explain the cognitive score.
The coefficients of the linear regression are shown in Table 3. The significant variables, with p values of less than 0.05, are displayed in the table. In comparison with individuals with spouse/partner and considering significant categories in the marital status variable: people widowed for more than 11 years had a cognition score 0.158 points lower, whereas for individuals widowed for 4 to 11 years and widowed less than 4 years the score was lower by 0.094 and 0.086 points, respectively and finally, for the separated or divorced participants the score was lower by 0.095 points.
Table 3.
Coefficients associated with the fit of a linear regression between variables and cognition scores
| Cognition Z-score | Coefficient | Standard error | Confidence interval 95% | t | P value | |
|---|---|---|---|---|---|---|
| Marital status (reference = with couple in 2015) | ||||||
| Widower more than 11 years | −0.158 | 0.034 | −0.225 | −0.091 | −4.64 | 0.000 |
| Widower from 4 to 11 years | −0.094 | 0.033 | −0.159 | −0.030 | −2.87 | 0.004 |
| Widower less than 4 years | −0.086 | 0.037 | −0.160 | −0.012 | −2.29 | 0.022 |
| Single in 2015 | −0.098 | 0.050 | −0.197 | 0.000 | −1.95 | 0.051 |
| Separated or Divorced in 2015 | −0.095 | 0.030 | −0.156 | −0.035 | −3.12 | 0.002 |
| Age (reference = 50–59 years) | ||||||
| 60–69 years | −0.299 | 0.040 | −0.380 | −0.219 | −7.33 | <0.01 |
| 70–79 years | −0.590 | 0.042 | −0.672 | −0.507 | −14.00 | <0.01 |
| 80- and more | −1.025 | 0.048 | −1.121 | −0.930 | −21.14 | <0.01 |
| Gender | ||||||
| Woman | 0.093 | 0.020 | 0.052 | 0.134 | 4.48 | 0.000 |
| Good socioeconomic perception | 0.098 | 0.022 | 0.055 | 0.142 | 4.45 | 0.000 |
| Education (reference = without education) | ||||||
| 1–4 years | 0.461 | 0.026 | 0.408 | 0.514 | 17.12 | 0.000 |
| 5–9 years | 1.000 | 0.026 | 0.948 | 1.052 | 37.48 | 0.000 |
| >10 years | 1.543 | 0.035 | 1.473 | 1.613 | 43.14 | 0.000 |
| Without Arterial Hypertension | −0.004 | 0.018 | −0.041 | 0.031 | −0.27 | 0.789 |
| Without Diabetes Mellitus | 0.113 | 0.020 | 0.072 | 0.153 | 5.48 | 0.000 |
| Without Stroke | 0.138 | 0.064 | 0.011 | 0.266 | 2.14 | 0.032 |
| Without Heart Attack | −0.038 | 0.043 | −0.123 | 0.046 | −0.89 | 0.373 |
| Without Depressive Symptoms | 0.064 | 0.021 | 0.021 | 0.106 | 2.98 | 0.003 |
| Without feeling of isolation | 0.096 | 0.023 | 0.049 | 0.142 | 4.05 | 0.000 |
| No difficulty walking (ABVD) | 0.132 | 0.094 | −0.053 | 0.318 | 1.40 | 0.162 |
| Without difficulty to bathe (ABVD) | 0.392 | 0.074 | 0.247 | 0.537 | 5.29 | 0.000 |
| Without difficulty to eat (ABVD) | 0.069 | 0.097 | −0.121 | 0.261 | 0.71 | 0.475 |
| Without difficulty to go to the toilet (ABVD) | 0.024 | 0.106 | −0.183 | 0.233 | 0.23 | 0.815 |
| Without difficulty to go to de bed (ABVD) | 0.060 | 0.085 | −0.106 | 0.227 | 0.71 | 0.478 |
| Working | 0.085 | 0.021 | 0.044 | 0.126 | 4.05 | 0.000 |
| Make decisions on their own | 0.226 | 0.022 | 0.182 | 0.270 | 10.14 | 0.000 |
| Bad internal locus of control | −0.005 | 0.005 | −0.015 | 0.004 | −1.04 | 0.302 |
| Health services | 0.140 | 0.035 | 0.071 | 0.209 | 4.05 | 0.000 |
| Life Satisfaction | −0.030 | 0.018 | −0.067 | 0.006 | −1.63 | 0.103 |
| Attendance at religious services | −0.064 | 0.025 | −0.113 | −0.014 | −2.55 | 0.000 |
Resource: Self-elaboration using MHAS data 2015
The cognition of participants decreased progressively with age. In persons aged 80 or more, the score decreased 1.025 points. In those aged 70 to 79, it was lower by 0.590 points, and for respondents aged 60 to 69, it was lower by 0.299 points, all cases compared with the 50 to 59 age group.
Cognition scores grouped by sex showed that women had standardized scores 0.093 points above those of men.
As for other social and economic variables, the cognition score was higher by 0.098 points for participants who had a positive instead of a negative perception of their economic situation. More years of schooling relate to better cognitive scores compared with people who had no schooling. The score for the respondents with 1 to 4 years of schooling was 0.461 higher. The cognition score of participants with 5 to 9 years of schooling was higher by 1.000 points. More than 10 years of schooling resulted in a better score by 1.543 points.
In terms of clinical conditions, participants without diabetes mellitus reported a better cognition score by 0.113 points. People who never had a stroke scored better by 0.138 points. Respondents with no symptoms of depression or feelings of isolation had a higher score by 0.064 and 0.096 points, respectively. Also, bathing without difficulty resulted in a higher cognition score by 0.392 points.
The cognition score of participants who made their own decisions was higher by 0.226 points. People who were working at the time had a higher score by 0.085 points.
The participants who regularly attended religious services scored less by 0.064 points than those who did not attend religious services.
For individuals with some type of health care services such as public social security, health insurance for government workers, or private health insurance, the score was higher by 0.140 points compared with that of people without any health care service.
A model including the interaction between sex and marital status was fitted, the results showed it was not significant (data not included).
Discussion
We studied the association between widowhood duration and cognition decline. The cognition scores of widowed participants were compared to those of married, single, or separated people using the scores obtained by the cognitive assessment of memory, learning, and visual exploration in adults and older adults in Mexico. This research yielded data that reflects lower cognition scores related to more extended periods of widowhood when compared with individuals with a partner, supporting the relationship between marital status and cognition level.
In recent decades, population aging represents one of the most significant challenges for society and health systems worldwide. Human aging is a complex, multidimensional, multifactorial phenomenon with diverse consequences at the individual, social, economic, and geographical levels (Jones and Higgs 2010; Wong et al. 2015).
For many older adults, the natural process of aging implies an increasing likelihood of vulnerability or weakness that may precede the occurrence of certain diseases. The vulnerability may involve physical and emotional conditions, such as the presence of traumatic events of adverse consequences in the lives of older people (Schaan 2013; Utz et al. 2002, 2012) (Schaan 2013; Utz et al. 2002, 2012). One of these conditions is widowhood. As Holmes & Rahe point out, this event is one of the most shocking situations for an older adult (Holmes and Rahe 1967). Widowhood plunges the aging adult into a state of permanent separation from the spouse that generates intense feelings of loneliness, isolation, and loss of reasons for living. However, despite being a dire situation, not all adults face the loss in the same way, and some put into play important coping mechanisms to deal with the new phase in their lives (Carr et al. 2000; De Vries et al. 2014).
Few studies have delved into the association between widowhood and cognition even if, as Arango reported, married people have better cognition scores (Mejía-Arango et al. 2007). In our study, in accord with Segura Cardona and colleagues (Segura Cardona et al. 2016), we found a statistically significant relationship between marital status and cognition score. In comparison with coupled individuals (married/consensual union), the cognition score of people widowed for more than 11 years decreased more than the score of persons widowed less than 11 years, which, in turn, is lower than the cognition score of people in widowhood for less than three years, using married people as the reference category.
Our results reveal that marital status concerns and affects overall health, particularly the cognition skills of older people. Some literature has shown how widowhood affects the cognition of older people (Liu et al. 2019). Marital status is a cultural topic influenced by the meaning of individual aging in each society. For example, a study that sought to examine the impact of social support on widowers in South Korea suggests that there are salient differences in family dynamics. Men enjoy substantial family support in this culture, which is why a good family relationship with their children is essential (Jeon et al. 2013).
Previous literature (Van Gelder et al. 2006; Waldron 1996; Zunzunegui et al. 2003) supports the assertion that culture is fundamental to determining older adults’ social ties. Hence the importance of analyzing older age from the perspective of culture, history, and beliefs of decades past and present. In this respect, Jeon et al. studied the Korean population. They observed the critical role in aging that culture and society have in that better social connections would be of support during old age, particularly in the case of widowhood (Jeon et al. 2013). Widowhood affects older people more, mainly generating a cognitive decline, which is more marked in people with more years of widowhood (Lyu et al. 2019). Lee et al. (2019) studied 2618 participants from the Health Retirement Study, surprisingly they found no adverse effect on cognitive function. However, they compared previously and recently widowed people who married again. It is possible to assume that having a partner could mediate the effects on cognition.
Wörn et al. (2020) studied 1269 men and women over 65 years old. They found that women could face a temporary decrease in some cognitive dimensions but not in the global cognitive function. Against their expectations, no robust effects were found in men. We found a slightly better cognitive score for women than for men. The reason behind the difference could be a better process of adaptation.
An interesting feature of our study was the analysis of the widowhood status of the participants not only at present but also in recording the transitions and changes between 2001 and 2015. Thus, we took into account the years that individuals were widowed. This follow-up made clear that people who were already widowed at the study baseline, i.e., respondents widowed for more than 10 years, had a lower cognition score, resulting in greater cognitive impairment. The linear regression data revealed that all widowed participants obtained a statistically significant lower value compared with the reference value that corresponded to participants married in 2015. This fact holds true for people who had been widowed for more than 10 years, those widowed for 4 to 11 years, the widowed for 3 years, and also for the persons separated or divorced in the wave 2015. This could mean that the individuals who were married and are now divorced or separated also show greater cognitive impairment, which is similar to that of widowed persons.
Results could be associated with other variables, for example, age. However, the regression model controlled for sociodemographic, clinical, and personality variables, which allowed us to observe notorious differences between men and women, such as the assimilation of widowhood. Variables such as the positive perception of their own economic situation and having health care services affect cognition favorably in older adults. The effect of age is independent, and, as expected, older age entails a more significant negative impact on cognition (Segura Cardona et al. 2016).
Other significant variables of interest were independent decision making and having health insurance. In both cases, cognition scores were better for respondents who made their own decisions and had access to some type of health service. Interestingly, respondents who attended religious services had a 0.058-point lower cognition score than those who did not.
In the model that we present here, we simultaneously included all variables. However, we examined other models: without control variables and adding by blocks sociodemographic, health, and psycho-social variables. We observed that the effect associated with each category of marital status was stronger in a model without control variables, but with a small coefficient of determination, which was drastically increased when sociodemographic variables were added. Additionally, similar estimated coefficients were obtained. This increase in the coefficient of determination was not observed when we separately included the other two blocks. Logically, adding at the same time all blocks the determination coefficient was the largest. We thought that separately reporting each model with the corresponding blocks would be space consuming and that as the coefficient of determination is improved by including all variables, it would be better to include only this model. Additionally, in all models the effects associated with our explanatory variable were in the same direction as those obtained in a model without control variables.
One limitation of this study is the exclusion of some participants because they did not answer every cognitive scale question. It is not possible to dismiss the possibility that the failure to answer reflects some degree of cognitive impairment, thus underestimating the effect. However, the sample size gives us an acceptable confidence level. Another situation is that we analyzed participants starting at the first wave in 2001 and a bias of survival could be present.
Considering that our study is a secondary database analysis, it does not contain information regarding the quality of the partners’ relationship. This missing information could impact the adapting process of older adults to their new lifestyle or on the type and quality of the older person’s support network, which he/she could rely on to face the new situation. However, in the study, we sought to have more information about the network of social support with the question on decision-making, which allowed us to observe whether the participants had some support or made fundamental decisions.
One more limitation of the MHAS database analysis is the long period spanned between the second and third study waves. Despite the high percentage of followed-up respondents, from 2003 to 2012 many data of the adults interviewed in the first surveyCompliance with ethical standards were lost.
Acknowledgments
The MHAS (Mexican Health and Aging Study) is partly sponsored by the National Institutes of Health/National Institute on Aging (grant number NIH R01AG018016) in the United States and the Instituto Nacional de Estadística y Geografía (INEGI) in Mexico. Data files and documentation are of public use and available at www.MHASweb.org.
Funding
The Mexican Health and Aging Study was supported in part by the National Institutes of Health/National Institute on Aging (R01AG018016, R Wong, PI).
Footnotes
Conflicts of Interest/Competing Interests Authors declare no competing interests.
Data availability
Database available upon request.
References
- Agudelo-Botero M, Giraldo-Rodríguez L, Murillo-González JC, Mino-León D, & Cruz-Arenas E (2018). Factors associated with occasional and recurrent falls in Mexican community-dwelling older people. PLoS One, 13(2), e0192926. 10.1371/journal.pone.0192926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Aguilar-Navarro SG, Fuentes-Cantú A, Ávila-Funes JA, & García-Mayo EJ (2007). Validez y confiabilidad del cuestionario del ENASEM para la depresión en adultos mayores. Salud Pública de México, 49(4), 256–262. 10.1590/S0036-36342007000400005. [DOI] [PubMed] [Google Scholar]
- Andrew MK, & Rockwood K (2010). Social vulnerability predicts cognitive decline in a prospective cohort of older Canadians. Alzheimer’s and Dementia, 6(4), 319–325. e1. 10.1016/j.jalz.2009.11.001. [DOI] [PubMed] [Google Scholar]
- Ardila A, & Rosselli M (2007). Neuropsicología Clínica (Primeraed.). México: Manual Moderno. [Google Scholar]
- Bangerter LR, & Waldron VR (2014). Turning points in long distance grandparent–grandchild relation-ships. Journal of Aging Studies, 29, 88–97. 10.1016/j.jaging.2014.01.004. [DOI] [PubMed] [Google Scholar]
- Bassuk SS, Glass TA, & Berkman LF (1999). Social disengagement and incident cognitive decline in community-dwelling elderly persons. Annals of Internal Medicine, 131(3), 165–173. [DOI] [PubMed] [Google Scholar]
- Biddle KD, Jacobs HIL, d’Oleire Uquillas F, Zide BS, Kirn DR, Properzi MR, et al. (2020). Associations of widowhood and β-amyloid with cognitive decline in cognitively. Unimpaired Older Adults. JAMA Network Open, 3(2), e200121. 10.1001/jamanetworkopen.2020.0121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carr D, House JS, Kessler RC, Nesse RM, Sonnega J, & Wortman C (2000). Marital quality and psychological adjustment to widowhood among older adults: A longitudinal analysis. Journals of Gerontology - Series B Psychological Sciences and Social Sciences, 55, 197–207. 10.1093/geronb/55.4.S197. [DOI] [PubMed] [Google Scholar]
- Chen J-H, Waite LJ, & Lauderdale DS (2015). Marriage, relationship quality, and sleep among U.S. older adults. Journal of Health and Social Behavior, 56(3), 356–377. 10.1177/0022146515594631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- De Vries B, Utz R, Caserta M, & Lund D (2014). Friend and family contact and support in early widowhood. Journals of Gerontology - Series B Psychological Sciences and Social Sciences, 69(1), 75–84. 10.1093/geronb/gbt078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- d’Hyver C, Gutierrez Robledo LM, (2014). Geriatria (Tercera ed). Mexico: manual moderno. [Google Scholar]
- Dorantes-Mendoza G, Ávila-Funes JA, Mejía-Arango S, & Gutiérrez-Robledo LM (2007). Factores asociados con la dependencia funcional en los adultos mayores: Un análisis secundario del Estudio Nacional sobre Salud y Envejecimiento en México, 2001. Revista Panamericana de Salud Pública, 22(1), 1–11. 10.1590/S1020-49892007000600001. [DOI] [PubMed] [Google Scholar]
- Gibson AK, & Richardson VE (2017). Living alone with cognitive impairment. American Journal of Alzheimer’s Disease and Other Dementias, 32(1), 56–62. 10.1177/1533317516673154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Glosser G, Wolfe N, Albert ML, Lavine L, Steele JC, Calne DB, & Schoenberg BS (1993). Cross-cultural cognitive examination: Validation of a dementia screening instrument for Neuroepidemiological research. Journal of the American Geriatrics Society, 41(9), 931–939. 10.1111/j.1532-5415.1993.tb06758.x. [DOI] [PubMed] [Google Scholar]
- Holmes TH, & Rahe RH (1967). The social readjustment rating scale. Journal of Psychosomatic Research, 11, 213–218. 10.1016/0022-3999(67)90010-4. [DOI] [PubMed] [Google Scholar]
- Instituto Nacional de Geografía y Estadística (INEGI). (2017). Mujeres y hombres en México 2016. México. Retrieved from http://cedoc.inmujeres.gob.mx/documentos_download/MHM_2016.pdf [Google Scholar]
- Jeon GS, Jang SN, Kim DS, & Cho SI (2013). Widowhood and depressive symptoms among korean elders: The role of social ties. Journals of Gerontology - Series B Psychological Sciences and Social Sciences, 68(6), 963–973. 10.1093/geronb/gbt084. [DOI] [PubMed] [Google Scholar]
- Johnson NJ, Backlund E, Sorlie PD, & Loveless CA (2000). Marital status and mortality: The national longitudinal mortality study. Annals of Epidemiology, 10(4), 224–238. [DOI] [PubMed] [Google Scholar]
- Jones IR, & Higgs PF (2010). The natural, the normal and the normative: Contested terrains in ageing and old age. Social Science and Medicine, 71(8), 1513–1519. 10.1016/j.socscimed.2010.07.022. [DOI] [PubMed] [Google Scholar]
- Katz S, Downs TD, Cash HR, & Grotz RC (1970). Progress in development of the index of ADL. The Gerontologist, 10(1), 20–30. [DOI] [PubMed] [Google Scholar]
- Lee Y, Chi I, & Palinkas AL (2019). Widowhood, leisure activity engagement, and cognitive function among older adults. Aging & Mental Health, 23(6), 771–780. 10.1080/13607863.2018.1450837. [DOI] [PubMed] [Google Scholar]
- Liu H, Zhang Y, Burgard SA, & Needham BL (2019). Marital status and cognitive impairment in the United States: Evidence from the National Health and aging trends study. Annals of Epidemiology, 38, 28–34.e2. 10.1016/j.annepidem.2019.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lyu J, Min J, & Kim G (2019). Trajectories of cognitive decline by widowhood status among Korean older adults. International Journal of Geriatric Psychiatry, 34(11), 1582–1589. 10.1002/gps.5168. [DOI] [PubMed] [Google Scholar]
- Mejía-Arango S, Miguel-Jaimes A, Villa A, Ruiz-Arregui L, & Gutiérrez-Robledo LM (2007). Deterioro cognoscitivo y factores asociados en adultos mayores en México. Salud Publica de Mexico, 49(S4), 475–481. 10.1590/S0036-36342007001000006. [DOI] [PubMed] [Google Scholar]
- Mejía-Arango S, Wong R, & Michaels-Obregón A (2015). Normative and standardized data for cognitive measures in the Mexican health and aging study. Salud Publica de Mexico, 57, s90–s96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Monserud MA, & Wong R (2015). Depressive symptoms among older Mexicans. Research on Aging, 37(8), 856–886. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schaan B (2013). Widowhood and depression among older europeans-the role of gender, caregiving, marital quality, and regional context. Journals of Gerontology - Series B Psychological Sciences and Social Sciences, 68(3), 431–442. 10.1093/geronb/gbt015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Seeman TE (1996). Social ties and health: The benefits of social integration. Annals of Epidemiology, 6(5), 442–451. 10.1016/S1047-2797(96)00095-6. [DOI] [PubMed] [Google Scholar]
- Segura Cardona A, Garzón Duque M, Cardona Arango D, & Segura Cardona A (2016). Riesgo de deterioro cognitivo en personas mayores de las subregiones de Antioquia, Colombia. Revista Brasileira de Estudos de População, 33(3), 613–628. [Google Scholar]
- Torres DIR (2011). El envejecimiento humano activo y saludable, un reto para el anciano, la familia, la sociedad. 30, 354–359. [Google Scholar]
- Tous JM, & Navarro J (1997). Las diferencias individuales en el proceso de envejecimiento humano, 14. [Google Scholar]
- Utz RL, Carr D, Nesse R, & Wortman CB (2002). The effect of widowhood on older Adults’ social participation. The Gerontologist, 42(4), 522–533. 10.1093/geront/42.4.522. [DOI] [PubMed] [Google Scholar]
- Utz RL, Caserta M, & Lund D (2012). Grief, depressive symptoms, and physical health among recently bereaved spouses. Gerontologist, 52(4), 460–471. 10.1093/geront/gnr110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Van Gelder BM, Tijhuis M, Kalmijn S, Giampaoli S, Nissinen A, & Kromhout D (2006). Marital status and living situation during a 5-year period are associated with a subsequent 10-year cognitive decline in older men: The FINE study. Journals of Gerontology - Series B Psychological Sciences and Social Sciences, 61(4), 213–219. 10.1093/geronb/61.4.P213. [DOI] [PubMed] [Google Scholar]
- Waldron I (1996). Marriage protection and marriage selection–prospective evidence for reciprocal effects of marital status and health. Science, 43(I), 113–123. [DOI] [PubMed] [Google Scholar]
- Whitley E, Deary IJ, Ritchie SJ, Batty GD, Kumari M, & Benzeval M (2016). Variations in cognitive abilities across the life course: Cross-sectional evidence from understanding society : The UK household longitudinal study. Intelligence, 59, 39–50. 10.1016/j.intell.2016.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wong R, Michaels-Obregón A, Palloni A, Gutiérrez-Robledo LM, González-González C, López-Ortega M, et al. (2015). Progression of aging in Mexico: The Mexican health and aging study (MHAS) 2012. Salud Publica de Mexico, 57(S1), S79–S89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wong R, Michaels-Obregon A, & Palloni A (2017). Cohort profile: The Mexican health and aging study (MHAS). International Journal of Epidemiology, 46(2), e2–e2. 10.1093/ije/dyu263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wörn J, Comijs H, & Aartsen M (2020). Spousal loss and change in cognitive functioning: An examination of temporal patterns and gender differences. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 75(1), 195–206. 10.1093/geronb/gby104. [DOI] [PubMed] [Google Scholar]
- Zunzunegui M-V, Alvarado BE, Del Ser T, & Otero A (2003). Social networks, social integration, and social engagement determine cognitive decline in community-dwelling Spanish older adults. The Journals of Gerontology Series B, Psychological Sciences and Social Sciences, 58(2), S93–S100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Database available upon request.