Family dysfunction and cognitive decline in aging: the “Health, Wellbeing, and Aging” (SABE) longitudinal population-based study

ABSTRACT. Stress during aging is not uncommon and dysfunctional family relationships are important sources of stress in the elderly. Considering the potential stressor that family dysfunction represents, it is questioned whether prolonged exposure to dysfunctional family arrangements can contribute to cognitive decline in aging. Objective: To verify whether family dysfunction is a predictive factor of cognitive decline in aging. Methods: Secondary study with analysis of existing data from the longitudinal, population-based study “Health, Wellbeing and Aging” (SABE). Data from 791 elderly people from two cohorts of the SABE study between 2006 and 2015 were analyzed. Family dysfunction was assessed using the Apgar family instrument, while cognitive performance was assessed using the Mini-Mental State Examination (MMSE), verbal fluency (animals) and digit length in reverse order. Cognitive decline was measured by the difference in scores in the period between 2006 and 2015. Results: Approximately 10% of the sample had family dysfunction. The familial Apgar score was not associated with decline on MMSE (p=0.732), verbal fluency (p=0.852) and digit span scores (p=0.718). Scores related to cognition and family functionality, such as age, education, living alone, depression and family Apgar, do not explain cognitive decline. Conclusion: The findings indicate that family functioning is not associated with cognitive decline in community-dwelling elderly. New studies will be needed to analyze the qualitative characteristics of family relationships in the cognitive performance of the elderly.


INTRODUCTION
I mpaired family functioning, i.e., the ability of the family to meet and harmonize its essential functions with the identity and tendency of its members and respond appropriately to hazards and opportunities within its social context is associated with declines in independence, autonomy and quality of life of older individuals [1][2][3] .Older adults from dysfunctional families have a 5-fold greater risk of presenting depressive symptoms compared to those living in functional families 4 .Family dysfunction is also associated with pain, falls and low quality of life 2,5 .
Roughly 10% of older adults have dysfunctional families, a phenomenon associated with individuals who are male, subject to ill-treatment, living alone and dependent for basic and instrumental activities of daily living (ADLs), pain, poor self-rated health, number of falls and osteoporosis 6,7 .
Emotional dysregulation stemming from tense, stressful family relationships, with consequent social isolation, is one of the main outcomes associated with family dysfunction, representing a factor that negatively impacts cognitive health 3,8 .Social interactions, either within the family or with other social actors, promote active aging by allowing the involvement of older adults in activities which represent cognitive challenges and mental well-being.Conversely, social isolation, or lack of social interaction, can be a risk factor for cognitive decline 8 .
The available evidence associating family functioning with cognitive health is scant.However, some studies reveal that indirect measures related to family arrangements, such as marital status, frequency of contact, support for instrumental ADLs, affective expression and conflicting family relationships, are associated with worse cognitive performance.A longitudinal study found that divorced or widowed older adults had greater risk of decline in memory performance and orientation over a 7-year follow-up compared with married individuals 9 .Another study showed that being married was associated with greater satisfaction with life and lower risk of developing mild cognitive impairment or dementia 10 .A population-based study of Chinese older immigrants found that participants with conflicted family relationships had poorer overall cognition and memory performance over a 2-year period than those with unobligated ambivalent relationships 11 .However, these studies centered on the relationship between family typology and interaction, limiting interpretation of the influence of family functioning structure on cognitive performance.
Cross-sectional data from our research group revealed that the poorer the family support, the lower the cognitive performance (global cognition, verbal fluency and digit span) 12 .However, these findings were drawn from a cross-sectional analysis, precluding any meaningful interpretation of the impact of family functioning on the probability of cognitive decline over time 12 .
The objective of the present study was to determine the prevalence of family dysfunction and its association with cognitive decline in aging.The study hypothesis holds that the rate of cognitive decline is associated with poor family functioning, even after controlling for confounding factors such as age, gender, education and depressive symptoms.
Identifying groups vulnerable to cognitive decline can help inform future investigations on risk and protective factors for successful aging and also provide a basis for systematic assessment of family functioning in professional practice, with consequent interventions preventing cognitive decline in older people.

Study design, setting and participants
The present study drew on data from the Brazilian "Health, Wellbeing And Aging" (SABE) study performed in the city of São Paulo, the country's most populous metropolis and home to the largest absolute number of older adults, a group that is highly diverse due to immigration and internal migration 13 .Since 2000, a total of 3,257 Brazilians aged ≥60 years living in the urban area of São Paulo city, comprising four cohorts (A, B, C and D), were included in the study.This population was selected by probability sampling of households registered in census sectors.The households visited were chosen randomly using cluster sampling.SABE participants are reassessed every five years, at which point new cohorts of individuals aged 60-64 years are included (Cohort B in 2006,  Cohort C in 2010, Cohort D in 2015).All data were obtained by applying a face-to-face survey at the participants' homes, conducted by trained interviewers and health professionals (Figure 1).Institutionalized older adults were not included in the SABE study.
Given that family dysfunction was the variable of primary interest in the present study and that assessment of the participants' family functioning was only included from 2006, the population universe of the present investigation comprised participants from cohorts A06 and B06 (nA06+nB06=n=1,413; Figure 1).
Of the 1,413 records of older adults (≥60 years) from cohorts A06 and B06, exclusions were made according to the following criteria:  2).Application of the criteria led to the exclusion of 620 patients, giving a final sample of 793 patients (Figure 2).
The study was approved by the Ethics Committee for Research in Humans of School of Public Health of the University of São Paulo (#COEP/23/10), São Paulo, Brazil.All participants provided informed written consent and information on their identity was anonymized.

Cognitive performance
Cognitive decline was measured based on the difference in test scores between the baseline assessment in 2006 and final assessment in 2015.Score difference (delta) was used for the following cognitive tests: • MMSE -comprising question items grouped into seven categories: temporal orientation, spatial orientation, immediate memory, attention and calculus, delayed recall (long-term memory), language and visuoconstructive ability, adapted for use in Brazil. 14Cut-off points varied according to the educational level of the respondent (≤20 points for illiterate subjects; ≤25 points for individuals with 1-4 years of education; ≤26 points for 5-8 years; ≤28 points for 9-11 years; and ≤29 points for individuals with >11 years of formal education) 14,15 .• Verbal f luency (animals) -assesses the semantic component of verbal f luenc y.Participants must say as many names of animals as they can within 1 minute.The total number of animals produced represents the score on the test.Scores below the cut-off values according to education indicate a cognitive deficit (cut-off score ≤9 for illiterate subjects; ≤12 for 1-7 years of education; ≤13 for >7 years of education) 14,15 .• Backward digit span 16 -measures working memory performance and, in the SABE study, consists of repeating five random digits in reverse order read aloud by the examiner at a rate of one1 per second.Each digit repeated correctly scores 1, giving a maximum score of 5 16,17 .
The cognitive decline variable for each of the tests used classifies the participant into two categories:

Family support
The family Apgar 18 , a valid instrument for Brazilian older adults 19 , evaluated the family functionality components of adaptability, partnership, growth, Affection, and resolve using five questions to reflect participants' views on the functional status of their family.The participants rated their satisfaction with family partnership, empathy, support, care, and interaction for each question on a scale from 0 (never) to 4 (always).The total score ranges from 0 to 20 points and is calculated as the sum of points on each question.Scores ≤12 indicate family dysfunction.

Covariates
The following variables were analyzed:

RESULTS
The sample predominantly comprised participants who were female, low-educated, white, married and with a low-to-middle monthly income (Table 1).
Regarding health, most participants had hypertension, cardiovascular disease and diabetes.The rate of smoking and alcohol misuse was low.Depressive symptoms were low overall, as was the percentage of frail individuals, in the evaluation of the frailty syndrome.Most respondents rated their health as fair or excellent (Table 1).For cognitive performance, participant scores were within cut-off values (Table 1).The majority of the participants lived with others and family dysfunction was detected in around 10% (Table 1).With the exception of gender and self-rated health, no significant differences were found between participants that remained in the study until 2015 and those who did not (losses) (Table 1).

Cognitive decline over a ten-year period and family functioning
There was a significant decline in mean score for the MMSE (p<0.001),verbal fluency (p=0.002) and digit span tests (p<0.001; Figure 3A).The magnitude of these declines on the MMSE (p=0.732),verbal fluency (p=0.852) and digit span (p=0.718),however, proved similar for individuals with and without family dysfunction (Figure 3B, 3C and 3D).

Association between family functioning and cognitive performance
In the multivariate linear regression analysis, no influence of Family Apgar score on cognitive decline on the MMSE, verbal fluency or digit span tests was evident (Table 2).

DISCUSSION
In the present study, whose objective was to determine the prevalence of family dysfunction and its association with cognitive decline in aging, only 10% of the sample had family dysfunction.These findings corroborate the results of other studies, reporting prevalence rates of family dysfunction ranging by 9-10% in the samples investigated 21,22 .
In aging, family dysfunction is characterized primarily by the interaction between low social support and impaired family bonds 23 .Moreover, occupational, economic and functional challenges can compromise family harmony during aging 22 .Social support, such as financial assistance, transport and help with domestic chores, are essential for the well-being of older people 24 .In the present study, most of the participants were classified as having a moderate economic level, perhaps explaining, albeit in part, the low rate of family dysfunction.
Family support should be bi-directional, i.e. older adults are supported by family members and, in turn, help with activities within the family structure (e.g.taking care of grandchildren, household tasks, informal work activity to boost income).Close contact with the older person allows family members to pinpoint physical, functional or emotional changes quickly 25 .
In the present study, most of the participants lived with other family members, another factor possibly explaining the low rate of family dysfunction.
Recent studies suggest that caring for an elderly person can have negative impacts on the caregiver's physical and mental health, especially when the care   is related to some form of dementia, which can lead to high levels of stress, depression, anxiety, pain, fatigue and insomnia, which can even present negative effects in the future for the elderly 25 .However, in the study, a low rate of family dysfunction was observed, which may suggest that caring for the elderly can have positive effects in some families.A review that aimed to identify the positive aspects of caring for someone with dementia listed effects such as emotional satisfaction, a sense of purpose and better quality of life 26 .However, contradicting the main hypothesis of the present study, no effect of family functioning on cognitive decline was detected.Although significant declines in scores on the MMSE, verbal fluency and digit span tests were seen over the follow-up period, these declines proved similar for individuals with and without family dysfunction.In addition, the family Apgar score had no influence on the odds of cognitive decline over the ten-year follow-up, i.e. according to participant perceptions, family functioning (satisfaction with partner, empathy, support, care and interaction with family members) had no influence on the risk of cognitive decline over ten years.
Echoing these findings, a previous cross-sectional study of 2,052 older Brazilians (mean age=70.8years) failed to observe a significant association between family dysfunction, as measured by the Apgar, and cognitive performance 27 .Also, in a longitudinal study conducted in Japan (mean age of participants around 73 years), the authors found that family support had no influence on cognitive performance 27 .However, the support of neighbors and friends was associated with better cognitive performance, highlighting that this type of support can be a protective factor against cognitive decline, neutralizing the influence of family relationships 28 .
However, the results of the cross-sectional Population Study of Chinese Elderly in Chicago, involving a sample of Chinese immigrants, revealed that detached, commanding, conflicted, and tight-knit relationships were associated with poorer performance in episodic memory, working memory and MMSE 11 .However, unlike the present study, the authors analyzed family relationships rather than family functioning, which potentially explains the disparities in findings.In a cross-sectional analysis of the SABE cohort which included only older adults aged 60-64 years, functional family support, but not structural support (number of individuals living with older adult), was found to be associated with higher scores on the MMSE, verbal fluency and digit span tests 12 .Moreover, the authors found that, unlike receiving social support, providing the community with support services was associated with better cognitive performance 12 .It is important to point out, however, that the population in question was younger than that of the present study, and participants with low cognitive performance were not excluded.The cross-sectional design precludes analysis of cause and effect, allowing the risk of reverse causality in which cognitive impairment may have already existed in the sample of older individuals, thereby explaining the dysfunctional family relationship.Cognitive impairment can significantly impact functional independence of older adults and of the family unit.The present study hypothesis holds that family dysfunction leads to cognitive decline.Thus, excluding individuals with low cognitive performance in 2006 reduced the likelihood that family burnout was already present in the population study.
The absence of an association in this study suggests that the protective role of family relationships on the cognitive health of older individuals might only be significant in young-old.Cognitive reserves decrease in older age, yet with synergic factors, such as social or family interactions, this decline produces little or no effect 29 .In addition, the present study was conducted in a cohort of older Brazilians with average income.Consequently, the sociodemographic and cultural characteristics of Latino populations can explain why family dysfunction did not represent a risk factor for cognitive decline in this study.Factors moderating this relationship, such as resilience stemming from a positive outlook on life and its challenges, might also attenuate the negative effect of family dysfunction.Corroborating this theory, a Brazilian study showed that most of the older adults investigated (n=84.9%)had a high score on an instrument measuring resilience and a positive association with cognitive performance 30 .This lends credence to the hypothesis that the absence of a significant association between family dysfunction and cognition in the present study participants is explained by the resilience and perception this group has to overcome adversities arising in their lives, along with the negative impact of family dysfunction over time.
The inconsistency in the findings of the present study versus data from the literature indicates the need for further longitudinal studies encompassing both quantitative and qualitative data.Future studies can analyze not only family functioning but also the characteristics defining the relationships built in living with the older adults, as well as the moderating role of resilience, coping and spirituality in the relationship between family functioning and cognitive decline.Identifying the factors which preserve cognitive reserve in aging is a crucial element for promoting healthy, active aging.

Study limitations
Given the study entailed secondary analysis of existing data, some methodological limits in interpreting the results should be noted.The first of these limitations relates to the cognitive assessment.The use of specific tests for assessing abilities such as information processing speed, attention, executive function and short-term memory, which decline with age, may yield different results to those reported.An assessment of functional capacity for ADLs in all participants would also add to the analyses carried out, enabling individuals to be classified according to cognitive impairment as opposed to only performance decline over time.The aim of the SABE study was to identify the living and health conditions of the older population in the city of São Paulo, explaining the availability of more global cognitive data such as that obtained using the MMSE and verbal fluency tests.Although studies that analyze existing data represent an innovative cost-effective approach, drawbacks in terms of measuring the variables of interest of the secondary study may introduce bias and, hence, influence results.Another important point involves losses over the long follow-up period.In view of the low difference in cognitive performance between 2006 and 2015, the reduction in sample size may have influenced the power of statistical tests to identify significant differences.
Notwithstanding these limitations, the present study is the first to analyze the influence of family relationships on cognitive decline using longitudinal data from a probability sample of community-dwelling older adults.The SABE study is the only longitudinal population-based study in São Paulo city which has been running for more than 20 years.The methodological rigor with which the data are collected and organized confer greater robustness, relevance and originality to the evidence reported.
In conclusion, although family dysfunction is prevalent in aging and global cognitive performance, and verbal fluency and working memory declined over the ten-year follow-up, the present study did not find an association between cognitive decline and family functioning.The influence of factors moderating the relationship between family functioning and cognitive performance should be explored in future studies to determine which factors enhance and which impair cognition during aging.

Figure 1 .Figure 2 .
Figure 1.Graphic representation of the cohorts of the "Health, Wellbeing and Aging" (SABE) study, carried out in the city of São Paulo in the years 2000, 2006, 2010 and 2015, with a projection for 2020 13 .
Abbreviation: MMSE, Mini-Mental State Examination.Note: Changes in cognitive score means from 2006 to 2015.

Figure 3 .
Figure 3. Mean scores of Mini-Mental State Examination, verbal fluency and digit span in individuals with functional and dysfunctional families.

Table 1 .
Participant characteristics, health status, cognitive performance and family functioning.

Table 2 .
Logistic regression coefficients between decline on cognitive tests and Apgar scores, adjusted for covariates (age, education, living alone, Geriatric Depression Scale score).Abbreviations: GDS, Geriatric Depression Scale; MMSE, Mini-Mental State Examination; OR, odds ratio; CI, Confidence Interval.Notes: *p≤0.05 indicates statistical significance.