Page 19 - IJPS-4-2

Chirinda W, et al.

a single measure (Mathers and Robine, 1997). This means that conventional measures solely based on mortality, such
as life expectancy, are not sufficient to summarize population health. Progress has been made over the past decades
in developing new summary measures of population health which is called health expectancies (Jagger, Cox, Le Roy
et al., 2006; Saito, Robine, and Crimmins, 2014). Health expectancies can be defined as the average number of years
an individual at a given age can be expected to live in good health based on the prevailing age-specific mortality and
morbidity status (Jagger, Cox, Le Roy et al., 2006). Health expectancies can be derived from various health measures
that vary from objective ones, for example, disability, to subjective ones, such as self-rated health, and well-being. As
such, different conclusions can be reached depending on the health measure used (Alves and Arruda, 2017; Christensen,
Doblhammer, Rau et al., 2009; Doblhammer and Kytir, 2001; Zimmer, Hidajat, and Saito, 2015). Therefore, the first
critical issue in health expectancy research is to define clearly the operational definition of health used (Saito, Robine,
and Crimmins, 2014).
In this present study, a self-rated health measure was used to calculate healthy life expectancies (HLE). Self-rated
health measures, albeit their subjectivity, have been widely accepted as reasonable measures of population health status
(Doblhammer and Kytir, 2001). In fact, their use, initially in sociology (Suchman, Phillips, and Streib, 1958) and later
in medical and epidemiological research (Kaplan and Camacho, 1983), dates back >50 years. Furthermore, self-rated
health measures have been found to be a good predictor of mortality by longitudinal studies (Ardington and Gasealahwe,
2014; Feng, Zhu, Zheng et al., 2016; Idler and Benyamini, 1997; Mossey and Shapiro, 1982) and a good predictor of
health care expenditures (De Salvo et al., 2009). There are some perceptions, though, that self-rated health measures are
too subjective (Jagger, Gillies, Cambois et al., 2010), and questions have been raised about their reliability in developing
areas where people have low awareness of health (Sabatini, 2014). This is due to “health illusions,” a scenario whereby
people “normalize” health deficits due to their low health expectations (Sen, 2002), which is likely the case in low-income
settings. As a result, people may rate their health as good even if they are in a poor disease condition and without basic
health-care facilities. Another view is that older adults can underrate the level of their disability and health challenges
because they may subconsciously rate themselves better compared to their elderly peers (Jagger, Gillies, Cambois et al.,
2010). On the other hand, Jylha’s theoretical framework for self-rated health (Jylhä, 2009) suggests that a cognitive
process is involved in an individual’s rating of their own health status. According to this theory, the cognitive process is
essentially subjective and is influenced by one’s contextual environment (Jylhä, 2009). Therefore, despite its limitations,
self-rated health is a useful global health indicator that can summarize all specific domains into a single health measure
(Crimmins, 2004). The applicability of self-rated health in low-income settings has been established (Burström, 2012), and
its construct validity has been confirmed through its association with socioeconomic status (Subramanian, Subramanyam,
Selvaraj et al., 2009). Such a simple and single measure becomes even more efficient in resource-limited settings, where
it can substitute for detailed composite measures that are more expensive.
Among researchers and policymakers mainly from European countries and the United States, the major interest
has developed over the past years in using health expectancies to monitor population health over time (Jagger, Cox,
Le Roy et al., 2006). Policymakers in these countries have shifted their focus to using health expectancy instead of
life expectancy as a policy tool and primary measure of population health and to monitor health outcomes (Robine,
Romieu, and Cambois, 1999). Furthermore, the indicator is also used for assessing inequalities, planning health care and
social services, and allocating resources (European Health Expectancy Monitoring Unit, 2007). Unfortunately, in Africa,
including South Africa, there is not much awareness of the usefulness of health expectancies for monitoring processes
and policymaking. That is, in spite of the rapid aging processes mentioned above, which call for immediate attention on
the well-being of older people, all that we know is that people are living longer, but we do not know how healthy older
people are in South Africa.
The gender paradox in health is widely well known, that is, women tend to have a longer life yet poorer health than
men. The root causes of the gender paradox are generally attributed to differences in socioeconomic status, genetic and
acquired risks, immune-system responses, hormones, disease patterns and prevention, and health-reporting behaviors
(Crimmins and Saito, 2000; Oksuzyan, Juel, Vaupel, et al., 2008). Nevertheless, a number of studies in Western countries
and in China have persistently showed that the improvement in self-rated health and HLE between older men and older
women largely depends on classification of self-rated health (Doblhammer and Kytir, 2001; Zack, Moriaarty, Ford
et al., 2004; Gu, Dupre, Warner et al., 2009). For example, in Austria, older men and older women witnessed similar
improvements in very good, good, and fair self-rated health and HLE in the 1990s, whereas older women showed greater
improvements if the healthy condition was defined by very good and good (Doblhammer and Kytir, 2001). A study from
the United States demonstrated similar improvements in self-rated health among older men and older women when good
health was categorized as excellent, very good, and good (Zack, Moriaarty, Ford et al., 2004). Using three categories of

International Journal of Population Studies | 2018, Volume 4, Issue 2 13