We conducted electronic searches of the following databases: Ageline, Biblioteca Virtual em Saúde/Centro Latino-Americano e do Caribe de Informação em Ciências da Saúde (BVS-Bireme), Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane, Embase, PsycArticles, PSYCINFO, Pubmed, Scopus, and Web of Science. Initially, articles published (without specifying starting date) up until July 2020 were searched, with a new update in February 2021. The new search string differed from the original search, with the inclusion of “self-reported health” in the search strategy, providing a new search string: Aged OR “Aged, 80 and over” OR “Old People” OR “Old Adults” AND Neuroticism OR Personality AND “Self-reported health” OR “Self-Rated Health”. When available in the databases, filters were applied to only select papers published in the selected languages (English, Portuguese, and Spanish). Additionally, hand searches were conducted using the included sources.

All primary studies, including cohort studies, analyzed the association between neuroticism and self-reported health in community-dwelling older adults aged 60 years and older. Mean age was greater than or equal to 60 years (this was a deviation from protocol). Only studies published up to 2021 were eligible. Studies that involved residents of longterm care facilities or hospitalized patients were excluded because the instruments were tailored specifically for community-dwelling older adults.

Data from the final list of eligible studies were extracted and listed in a spreadsheet (Microsoft Excel 2011) by 2 of the reviewers. Another reviewer cross-checked the information for accuracy. Any discrepancies were resolved by discussion and consensus among the reviewers. The main categories of extracted data were characteristics of the study (first author, year, type, country), characteristics of the participants (sample size, sex, age), methodological details (instruments used), and biopsychosocial factors. The data extraction strategies were pilot tested for four studies and refined throughout the process. Whenever necessary, first authors were contacted by email for requesting further information. Results were graphically reported with tables containing the narrative that described the body of retrieved literature. Outcomes were reported in three sections considering the aims of the review as following:

1. 1. Focus of studies;
2. 2. Definitions and measures of neuroticism and self-reported health;
3. 3. Associations with biopsychosocial factors.

Studies considering the association between neuroticism and self-reported health in community-dwelling older adults make up a small proportion of the gerontology and psychology literature. Of all included studies, 47% were published between 2008 and 2017, with publication gaps in 1993 – 2009, 2009 – 2012, and 2015 – 2016.

This review found 11 cross-sectional studies and 4 studies with longitudinal design. The characteristics of the included studies are shown in Table 1.

TABLE 1.
Characteristics of the included studies (n = 15).

TABLE 1.
Characteristics of the included studies (n = 15).

Note: Medicare PCC: Medicare Primary and Consumer-Directed Care Demonstration; BLSA: Baltimore Longitudinal Study of Aging; HRS: Health and Retirement Study; ELSA: English Longitudinal Study of Ageing; NHATS: National Health and Aging Trends Study; FIBRA: Frailty in Brazilian Older Adults; DIA: From Disability to Activity: The Challenge of Aging; NEO PI-R: Revised NEO Personality Inventory; SIDP-IV: Structured Interview for DSM-IV Personality; NEO-FFI: NEO Five Factor Inventory; EPQ-R: Eysenck Personality Questionnaire-Revised; EPI: Eysenck Personality Inventory; MIDI: Midlife Developmental Inventory Personality scale; SF-36: 36-item Short Form Health Survey.

All studies were published in the English language. Regarding the countries of origin, 9 studies were from the USA,5,12,15,21,27,28,29,30,31 1 was from Australia,10 1 was from the United Kingdom,7 1 was from Ireland,11 1 was from Israel,32 1 was from Brazil and Portugal,33 and 1 was from Germany.34 The countries of origin of the studies are shown in Table 1.

Eight studies involved older adults only (60 years and over), but the age range used to define this population varied across studies7,15,21,28,29,30,31,35 or was not specified.10 Jerant et al.28 recruited people living in the community aged 40 years and over, whereas Ward29 included non-institutionalized individuals aged 50 years and over recruited from the Health and Retirement Study. Gale et al.7 used data from the UK Biobank and selected individuals aged 37 – 73 years living in the community. Some studies did not specify the age range and the authors used the term “old people” to describe the target population.10 None of the studies focused on a specific gender, selecting women and men. Study sample size ranged from 266 to 321 456 people.

Most of the studies (n = 10) drew on secondary data from larger studies, while 3 collected data based on a convenience sample,28 the electoral roll,10 or listed phone numbers.30 The analyses of the study by Aiken-Morgan et al.,21 for instance, were obtained from the Baltimore Study of Black Aging (BSBA), whereas the study by Burke et al.11 used data from the Technology Research for Independent Living (TRIL) study. Ward29 collected data from the Health and Retirement Study (HRS), and Gale et al.7 performed their analyses based on the UK Biobank; Wettstein et al.35 used data from the Interdisciplinary Longitudinal Study of Adult Development (ILSE). The study by Elran-Barak et al.32 drew on data from the Israel Diabetes and Cognitive Decline study, the only investigation considering older adults with specific diseases. Lastly, Mitchell et al.31 employed data from the Health and Retirement Study, while Duberstein et al.27 and Chapman et al.12 and both used the same sample from outpatient facilities.

Most of the studies included (n = 8) in the review sought to explore the relationship between personality and self-reported health, albeit subjective or objective. Noteworthy30 is a study that examined personality pathology and perceived health and a study7 that investigated the association between neuroticism and mortality and the influence of self-reported health on this relationship. Personality as a moderator of self-reported health in older adults with diabetes and of the effects of stressful life events31 was featured in 13.3% of the included articles.32

Biopsychosocial factors played a role in the association between personality and self-reported health.11,28,29,33 Only one study11 failed to find an association between neuroticism and self-reported health on multivariate analysis.

Firstly, according to the five-factor model (Big 5) of Costa & McCrae,14 personality is hierarchically organized, with 5 higher order dimensions or factors that can be divided into lower-level components or facets. At the top of the hierarchy, personality can be described along the following dimensions: neuroticism (N), extraversion (E), openness to experience (O), agreeableness (A), and conscientiousness (C). Neuroticism is characterized by distress, chronic negative emotionality, and vulnerability to stress; extraversion has the hallmarks of an optimistic outlook and a sociable and active lifestyle; openness to experience manifests as a willingness to consider new ideas and experiences and esthetic and/or intellectual sensitivities; agreeableness is characterized by trust, altruism, and friendship; and conscientiousness reflects reliability, diligence, and leaning toward action.36

Whereas the “Big 5” assesses the 5 dimensions, the Eysenck Personality Questionnaire-Revised (EPQ-R)37 is based on 3 similar dimensions of personality traits, namely: psychoticism (P), extraversion (E), and neuroticism (N). However, an updated EPQ-R, named the Eysenck Personality Inventory (EPI) and, also developed in 1964, is based on a personality concept that identifies extraversion and neuroticism as 2 primary independent factors comprising a global personality construct which measures 2 independent invasive dimensions of personality.38 Besides these instruments, the Midlife Development Inventory (MIDI) developed by Lachman & Weaver39 is a 26-item scale in which respondents are asked to indicate how well each describes them on a scale of 1 (not at all) to 4 (a lot). The inventory contains 4 items for extraversion, 5 for agreeableness, 5 for conscientiousness, 4 for neuroticism, and 7 for openness.

Regarding other measures, 7 studies used the NEO Five-Factor Inventory,7,12,27,28,32,33,34 2 employed the EPQ-R,7,10 3 used the Revised NEO Personality Inventory,5,30 1 used the EPI,11 and 3 studies used the MIDI.16,29,31 The instruments used to assess the neuroticism personality trait are outlined in Table 1.

Second, self-reported health can be defined as “... a summary statement about the way in which numerous aspects of health, both subjective and objective, are combined within the perceptual framework of the individual respondent.’’1 There are often implicit contextual frameworks that shape the ways in which people evaluate their own health and these are the most important in empirical research. These frameworks are difficult to measure directly in quantitative studies. Usually, they are indirectly represented by indicators such as age, nationality, or ethnic group. The framework of evaluation may also be shaped by individual characteristics such as an optimistic or pessimistic general disposition.1

As outlined earlier, self-reported health can be measured using a general question and its derivatives. Nevertheless, although self-reported health can predict some negative health outcomes such as mortality,40 it largely comprises 2 domains: physical and mental health,8,41 which can contribute unequally to these outcomes. Of the selected studies, 4 chose to use the question: “In general, how would you rate your overall health?”, 3 studies used a question derived from the general question (How would you rate your overall health?), and 1 study used 3 questions to measure self-reported health. Three articles were more specific regarding the questions probing self-reported health, eg, “In the last 12 months, would you describe your general health as excellent, very good, good, fair, or poor?” or “How satisfied are you with your health?” or “In general, compared to other people your age, would you say your health is excellent, very good, good, fair, or poor?”

Lastly, 5 studies used predefined self-reported health scales, namely the SF-36, the EQ-5D, and the Health Problem Checklist. For more details, see Table 1.

The studies showed that, besides the association between self-reported health and neuroticism, some biopsychosocial factors can contribute to self-reported health and neuroticism. The associations of biopsychosocial factors with self-reported health and neuroticism in the reviewed articles are illustrated in Figure 2.

FIGURE 2.
Framework with biopsychosocial factors.

Of the 15 included studies, only 5 assessed health indicators using instruments. For example, the study by Burke et al.11 analyzed comorbidities using the Charlson Comorbidity Index (CCI), showing a mean risk of death in the sample of 1.92 (SD = 2.07). The study by Duberstein et al.27 quantified diseases using the Cumulative Illness Rating Scale and the severity of depressive symptoms with the Hamilton Depression Rating Scale, detecting a significant association between neuroticism and poor perceived health, over and above the effects of depression and disease burden. In contrast to the study by Duberstein et al.,27 the investigation by Chapman et al.12 found that neuroticism was associated with perceived health controlled by disease burden and demographic variables, but not by depression. Participants in the study by Aikan-Morgan et al.21 showed that high neuroticism was associated with increased reports of general and cardiovascular health problems.

Only 2 studies addressed chronic diseases in the statistical models. Based on the collected data, physical health factors were grouped into arthritis, asthma, chronic obstructive pulmonary disease (COPD), coronary heart disease, heart failure, depression, diabetes, hypertension, stroke, and cancer. Among the older adults assessed in the study by Jerant et al.,28 the most prevalent noncommunicable disease (NCD) was arthritis (56.1%), followed by (46.5%) and diabetes (41.4%). It is noteworthy that older adults who had higher self-reported health also informed diabetes, whereas a poor rating was associated with arthritis, depression, and high neuroticism. The study by Löckenhoff et al.5 using data from the Medicare Primary and Consumer-Directed Care Demonstration (PCC) and the Baltimore Longitudinal Study of Aging (BLSA), grouped health conditions (hypertension, diabetes, cancer, stroke, and coronary heart disease) and depression. Depression was the most prevalent condition in the Medicare PCC sample (9.3%), where the number of health conditions had a mean of 1.7 (SD = 1.1). In the study, the effect of neuroticism was moderated by health conditions. Lastly, Jorm et al.10 did not report the prevalence of health factors. Nonetheless, the study showed that arthritis was correlated with neuroticism in women (r = 0.19).

In addition to chronic NCDs, Elran-Barak et al.32 showed that low handgrip strength, slow gait speed, and poor general cognitive functioning were associated with poor subjective health only among individuals with high neuroticism. Similarly, Cachioni et al.33 found that Brazilian and Portuguese older adults with low neuroticism self-reported their health as good or very good. Regarding the psychosocial factors assessed, analyses included sense of control,29 social support,11 and stressful life events.31 Ward29 found that 25.1% of the sample self-reported their health as fair and/or poor, and that positive and negative affect correlated strongly with fair and/or poor health, and functional limitations correlated with sense of control. Burke et al.11 found that self-reported health correlated with the CCI (r = -0.33, p < 0.01), followed by instrumental activities of daily living (r = 0.33, p < 0.01), CES-D (r = -0.29, p < 0.01), the Pittsburg Sleep Score (r = -0.20, p < 0.01), social support from relatives (r = 0.13, p < 0.01) and friends (r = 0.17, p < 0.01), and neuroticism (r = -0.24, p < 0.01). Finally, Mitchell et al.31 found that neuroticism was directly associated with higher levels of depressive symptoms and worse stressful life events. Only one study had mortality as an outcome. Gale et al.7 found that, after adjustment for sex and age, there was an increase in all-cause mortality in study participants who had higher neuroticism (HR = 1.06, 95%CI 1.03 – 1.09, p < 0.001).

Correspondence data Meire Cachioni – Arlindo Bettio, 1000 – Vila Guaraciaba – CEP: 03828-000 – São Paulo (SP), Brazil. E-mail: meirec@usp.br