| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||
INVITED COMMENTARY |
Department of Geography, University of Florida, Gainesville.
Gerontologists (Diehl & Willis, 2003
; Golant, 1998; Krause, 1996; Lawton, 1998; Wahl, Scheidt, & Windley, 2003) have often studied how the social and physical contexts of older persons influence their physical and psychological well-being. Theoretical writings by Lawton (1998) early argued that older persons with poorer functional health are more susceptible to environmental effects. More recently, research by epidemiologists (Balfour & Kaplan, 2002; Diez Roux, 2002; Oakes, 2004) has investigated how neighborhood attributes influence individual health outcomes. Establishing a residential context–individual outcome causal relationship is by far one of the most conceptually and statistically challenging types of analyses. Besides the need for a highly sophisticated statistical methodology, the investigator must avoid a variety of conceptual missteps. Most notably, there is the danger of omitting or mismeasuring individual-level or compositional variables that offer the "true" explanation for variations in individual health outcomes or of incompletely or incorrectly measuring the pertinent neighborhood effects. Both these failings can result in erroneous neighborhood statistical effects.
Cagney, Browning, and Wen's article (appearing in this issue) offers a highly competent statistical investigation of how a neighborhood's social organization and socioeconomic status influence the self-rated health outcomes of White and African-American older persons (age 55+) in Chicago. I believe, however, that their analysis does not go far enough to conceptualize and measure the relevant individual-level variables and that their neighborhood variables are "crude proxies ... for actual processes and mechanisms that may link specific characteristics of neighborhoods to the health of the persons who reside in them" (Diez Roux, 2002, p. 517).
EXPLAINING SUBJECTIVE HEALTH: ABSENT PSYCHOLOGICAL ANTECEDENTS
Their task is daunting because self-rated health is a multidimensional construct that measures the direct, indirect, and reciprocal influences of an array of actual and subjectively interpreted physical, functional, emotional and cognitive health conditions with different temporal properties (Bulatao & Anderson, 2004).
Although the authors' individual-level variables (e.g., age, income, education, race, etc.) are probably adequately controlling for the variation in the actual health conditions of older persons, the inclusion of a "self-assessment" health construct magnifies the explanatory task, even as many research studies have reported a strong correlation between the subjective and objective health status of older individuals (George, 1995).
Absent from the models are individual-level variables that might explain why older persons differently appraise and cope with their own actual health conditions and their neighborhood's health-achieving resources and barriers (Lazarus, 1966). Much literature has shown, for example, how the personality and temperament characteristics of older persons influence how they evaluate and cope with stressful events (Lawton, 1998; Ruth & Coleman, 1996). Perceived control, mastery, self-efficacy, optimism, neuroticism, hardiness, and competence are often hypothesized as possible psychological antecedents of subjective health (Diehl & Willis, 2003; Ostir, Ottenbacher, & Markides, 2004). Thus, more optimistic older persons may simply dominate affluent neighborhoods and have greater psychological resources that positively influence their subjective health assessments and enable them to cope more effectively. It is unclear whether controlling for individual and neighborhood income levels—correlated with psychological barriers and resources—eliminates this confounding relationship.
CHALLENGES INTERPRETING THE CONTEXTUAL EFFECTS
There is an inherent theoretical incompleteness and ambiguity as to how the neighborhood ecological indicators, collective efficacy, age concentration, residential stability, population density, and affluence/poverty influence subjective health outcomes.
It is unlikely that the collective efficacy construct will influence a heterogeneous group of older people in the same ways. It does not measure the salience of social resources or supports and unjustifiably assumes that all older residents will interpret the social cohesiveness of their neighborhoods as functionally relevant (Golant, 1984). Collective efficacy also appears conceptually appropriate to explain only partially the variation in self-rated health reported by older persons. A neighborhood's social environment can influence its older occupants' ability to access medical assistance or perform their instrumental activities of daily living (IADLs). It is unclear, however, how this construct would mediate the effects of activities of daily living (ADL), cognitive or chronic health problems, or ineffectual coping styles. It is telling that a spouse's presence or absence has insignificant effects. The spouse is arguably the most important informal caregiver and compared with a neighborhood's social resources is likely to be a far more salient and reliable source of both instrumental and expressive supports. A future analysis should test a potentially important interaction effect—collective efficacy by spousal status. The proposition: neighborhood variations in collective efficacy will affect the self-rated health only of older persons without a spouse (who are, thus, more vulnerable).
Interpreting the effects of "percentage of neighborhood age 55+ population" is difficult for several reasons. First, it indefensibly assumes that old age is a homogeneous stage of life. It is likely, however, that there will be unmeasured neighborhood variations in the presence of age 75+ population concentrations—a group more likely to be at risk of poor health. Second, we know from age-identification theory that has linked residential density with age-identification patterns (Rosow, 1967) that the mere presence of an older (or younger) age-peer group in a neighborhood tells us with little certainty whether its older occupants will attend to or compare themselves positively or negatively with its members. Third, a neighborhood's age density may simply be indirectly measuring the extent to which providers can cost effectively deliver health care and supportive services.
The effects of a neighborhood's poverty (or affluence) concentration are less easily interpreted than might first appear (beyond the well-documented context–population composition problem; Oakes, 2004). Ambiguity exists as to whether a neighborhood's "poverty" or "affluence" is a result of its older or younger population composition (even after controlling for age 55+ neighborhood concentrations). This is important because we know that, in the extreme, the concentration of low-income older adults in a public housing project or neighborhood has never been linked with the same severe social ills and pathologies associated with like concentrations of impoverished younger families with children (Ross, Reynolds, & Geis, 2000).
The residential stability variable is also very hard to interpret. On the one hand, it may reflect a population of moderate- to high-income older persons who have embraced an aging-in-place residential strategy. These self-selected neighborhoods are more likely to report poor individual self-rated health simply because chronologically older persons dominate them. Alternatively, residential stability may point to a population of lower-income older persons who have become habituated or apathetic to their poor living conditions and poor health. Additionally, these neighborhoods may be occupied by a population of very healthy middle- and high-income homeowners in their early middle age who collectively constitute a very different social context than the above two examples. The authors begin to "unpack" these mediating influences by analyzing the interaction effects of residential stability and affluence.
Footnotes
Decision Editor: Charles F. Longino, Jr., PhD
Received for publication September 27, 2004. Accepted for publication October 4, 2004.
References
This article has been cited by other articles:
 |
|
 |
|
  G. C. Smith and G. M. Sylvestre Effects of Neighborhood and Individual Change on the Personal Outcomes of Recent Movers to Low-Income Senior Housing Research on Aging, September 1, 2008; 30(5): 592 - 617. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
