Professional Documents
Culture Documents
The Herth Hope Index (HHI) and Related Factors in The Japanese General Urban Population
The Herth Hope Index (HHI) and Related Factors in The Japanese General Urban Population
In this study we sought to demonstrate the reliability and validity of the Japanese Herth Hope
Index (HHI) 12, a psychometric instrument for measuring hope, in a general urban population,and to
investigate factors correlated with the HHI score. Anonymous written surveys were collected from
255 male and female residents livingin Tokyo "N" ward and Saitama "T" township aged 20-69. The
Cronbach a coefficientwas 0.89, and confirmatory factor analysis showed three-dimensionality. We
also observed hope to strongly correlate with having reasons for living and benefit-finding, confirm-
ing the reliabilityand validity of the Japanese HHI scale. Mean HHI score was 35.5. Stratified multi
ple linear regression analysis on factors which correlated with the HHI score showed HHI scores to
correlate positive with age, the presence of a spouse, and sufficient psychosocial support. We also
found that younger subjects receiving sufficient social support enjoyed HHI scores equivalent to
older subjects, while women's higher HHI scores appeared attributable to extensive social support
networks. No correlation was observed with adverse experiences or history of illness. This study
shows that HHI scores in the general urban populationare not unexpectedly high compared with pre
viously studied subjects with serious disease, that advanced age correlates with higher HHI scores,
and that receipt of social support may mediate positive influences on the HHI score.
Key words : Herth Hope Index (HHI) score, general urban population, demographic attribute, stressful
experience, social support
Department of Health Sociology,Division of Health and Nursing Sciences, TokyoUniversity Graduate School of
Medicine
32 Jpn J Health & Human Ecology 2007;73 (1)
da, 2005). Hope can be conceived of as a coping 2006), as well as other individuals suffering
strategy or the psychosocial internal resources nec- chronic difficulties such as the elderly (Koizumi et
essary to maintain a positive outlook in the face of al., 1999), certain categories of workers (Konishi
adversity and stress (Farran et al., 1995 ; Herth, and Esaki, 2004), and family members providing
1992 ; Snyder et al.,1991) . Dufault and Martocchio palliative home care (Chapman and Pepler, 1998).
(1985) define hope as "a multi-dimensional Factors found to contribute to the HHI score
dynamic life force characterized by a confident yet include sex, income, marital status, length of illness
uncertain expectation of achieving good, which to (Herth, 1992 ; Konishi and Esaki, 2004), specific
the hoping person, is realistically possible and per illness diagnosis (Herth, 1990), pain and other
sonally significant." Academic research on the the forms of physical suffering (Chen, 2003 ; Herth,
ory of hope has occurred mostly outside of Japan 1992 ; Hirano et al., 2006 ; Lin et al., 2003), uncer
since the concept was first proposed and codified tainty (Wonghongkul,2000), illness-related wor
in early 1990 (Farran et al.,1995) . ries and frustrations, social difficulties, psychoso
Several scales have been developed to measure cial support, and happiness (Hirano et al., 2006).
hope, of which the simplest is the Herth Hope Meanwhile, high HHI scores have been found to
Index (HHI). Chinese, Japanese, Thai, and other correlate with high quality of life (Herth, 2000) and
versions of the scale have been developed based low despair, somatization, and loss of control (Chap-
on the original English version. There are two apa- man and Pepler, 1998). It is expected that results
nese versions as well, developed by Koizumi et al. such as these will impact/influence/contribute to
(1999) and Yamakiand Yamazaki (2003). The HHI clinical patient care.
is a 12 question version of the longer 30 item These examples illustrate how research on hope
Herth Hope Scale (HHS), developed by Herth to and HHI has been limited to minority populations
measure hope based on Dufault and Martocchio's such as individuals confronted with severe illness
model of hope (Dufault and Martocchio, 1985). or other stressful situations. No study of hope in a
Dufault and Martocchio's model of hope postulates general population has yet been undertaken. We
two spheres within which to measure hope, ener- feel that evaluating hope in a general population is
alized and particularized, and six dimensions, con- particularly important in Japan today given the
textual, affective, cognitive, behavioral, affiliative, endemic "loss of hope" that is often felt to currently
and temporal. In Herth's version, these six dimen exist in Japan, which has underscored the impor
sions are combined into three subscales each tance of hope for the population at large (Yamada,
involving two concepts "inner sense of temporal 2005).
ity and future," "inner positive readiness and For these reasons, we first sought to verify the
expectancy," and "interconnectedness with self and reliability and validity of the Japanese version of
others." The validity and reliability of these the Herth Hope Index 12 scale, then used this psy
subscales have been verified (Herth, 1992). chometric instrument to investigate HHIs score as
HHI-related research has focused on individuals well as correlated factors in a general urban
which chronic disease such as HIV/AIDS (Herth, population.
1990), cancer (Chen, 2003 ; Herth, 1990 ; Herth,
2000 ; Lin et al., 2003 ; Wonghongkul,2000), my-
otrophic lateral sclerosis (ALS) (Hirano et al.,
Jpn J Health & Human Ecology 33
considered to indicate informed consent to partici- Index (GFI) : 0.928 ; comparative fit index (CFI)
pate in the study. 0.937 ; and root mean square error of approxima
tion (RMSEA) : 0.073 (results shown in Fig. 1).
‡V Results
GFI and CFI are used to assess model fitness,
1, HHI Factor Structure and Concurrent while RMSEA is used to assess stability (Yama
Validity moto and Onodera, 2002).
The a coefficient for the HHI was 0.89. One fac- To study concurrent validity,we performed a par
tor with an eigenvalue of 1 or above was obtained tial correlation analysis between the HHI score
from exploratory factor analysis (principal factor and several potentially analogous concepts, while
analysis, promax solution) of the 12 HHI questions. controlling for gender and age (results shown in
Eigenvalue contribution was 45.1%. Next, we per- Table 1). Survey items found to have a strong sig
formed a confirmatory factor analysis on a model nificant correlation with HHI score (p<0.001)
consisting of the three factors proposed by Herth. were, in descending order, reasons for living (r=
The results of this analysis were goodness of fit 0.729), benefit-finding (r=0.507), dreams (r=
36 Jpn J Health & Human Ecology 2007;73 (1)
Notes
1) Excluding invalid responses
2) Adjusted for age and gender
3) * : p<0.05, * * * :p<0.001
Table 2 HHI score gender and age distributions 3. Factors Related to HHI
Table 3 ANCOVA of HHI scores for patient attributes, stressful experiences, and social support
Notes
1) SE : Standard Error
2) Adjusted for age and gender
3) * * : p<0.01, * * ; p<O.001
these factors were treated as binary variables in In Model 2, we included social support networks
multiple linear regression. Also, presence of chil- in the analysis in addition to factors used in Model
dren was highly correlated with marital status and 1. We found a correlation between higher HHI
was potentially multicolinear, so this factor was scores and increasing subject age (/3 =0.134, p=
excluded from analysis. 0.043). HHI scores were also extremely high in
In Model 1, the independent variables employed subjects reporting sufficient sources of support
were demographic attributes and a history of compared to those with no such sources of support
stressful experiences. Female sex (,6=-0.162, p= (3 = 0.442, p<0.001). Incorporating the number
0.020) and the presence of a spouse (/3 =0.285, p of sources of support into the model was sufficient
<0.001) were correlated with significantly higher to eliminate the significant correlation of the other
HHI scores. two social support variables. HHI scores were con-
38 Jpn J Health & Human Ecology 2007;73(1)Table 4 A stratified
Jpn J Health & Human Ecology 39
sistently high for those married regardless of Herth's three factors and proved goodness of fit.
social support variables (a =0.237, p<0.001). The Japanese version of the HHI 12 is composed
Finally,correlations between HHI scores and gen- of the same three factors as the original instru-
der disappeared in Model 2. ment and demonstrates factorial validity.
In evaluating concurrent validity, we found that
‡W Discussion
the HHI has corellates strongly with the presence
1, Reliability and Validity of the Japanese of reasons for living, with r value at 0.729, followed
The HHI a coefficient was 0.89, demonstrating psychological comfort, PHCS, and satisfaction with
high internal consistency. social roles. This demonstrates concurrent validity
To study validity we performed exploratory fac- in this study of the HHI, which consists of the
tor analysis on the 12 HHI questions, which dem three factors described above, and suggests that
onstrated unidimensionality. Previous HHI the HHI is a measure of coping strategies or the
research has used exploratory analysis to elucidate internal psychosocial resources necessary to main
factor structure In 1992 Herth identified the tain a positive outlook in the face of adversity and
rality and future," "inner positive readiness and Drawing upon previous research into the con
expectancy," and "interconnectedness with self and cept of reasons for living, Nomura (2005) identified
others" in a study of adult men and women with component factors of "meaning and value found in
chronic disease (n=172) ; Koizumi et al. (1999) living" and"introspective, positive feelings about liv-
identified the two factors of "emotion and human ing" in an elderly populace. That these factors are
interaction accompanying the process of hope" and in turn reminiscent of hope as defined here lends
"confidence in hope support to the conclusions of our study.
, assertive preparations and
expectations" in a study of the elderly in Japan (n=
87) ; and Konishi and Esaki (2004) identified the 2, HHI Scores and Related Factors
three factors of "positive awareness," "give and The mean HHI score for the 255 valid responses
take between self and others," and "pessimistic received was 35.5± 5.6. In previous research study-
awareness" in a study involving adult laborers (n= ing persons with serious illness (Chen, 2003 ; Her
1,909). As has been previously observed (Koizumi th, 1992 ; Hirano et al., 2006 ; Lin et al., 2003;
et al., 1999 ; Konishi and Esaki, 2004), differences Wonghongkul, 2000), physical and psychological
in the number and nature of factors influencing the discomfort or suffering resulting from illness has
HHI may be attributable to differences in culture, been found to negatively influence HHI scores.
subject background, and even translational nuance. Therefore, it would be anticipated that our sample,
In this study we used confirmatory instead of which involved a more general population sample,
exploratory factor analysis because of this including healthy individuals, would tend to demon
method's ability to prove theoretical hypotheses strate higher HHI scores than those found in stud
(Yamamoto and Onodera, 2002). GFI, CFI, and ies of persons with serious illness. Accordingly,the
RMSEA were greater than 0.9, greater than 0.9, scores reported in this study are not unexpectedly
and lower than 0.08 respectively, demonstrating high when evaluated in light of the previously
good fitness characteristics. We used a model with reported mean scores of 34.3±1.6 for US cancer
40 Jpn J Health & Human Ecology 2007;73 (1)
patients (n115) (Herth, 2000), 39.0•¬4.3 for US women has the provision of social support been
terminal patients (n=20) (Herth,1990), 37.3±5.3 shown to result in higher life satisfaction (Kin et
(n=226) (Chen, 2003) and 32.5±4.2 (n=233) (Lin al., 1999). This previous work suggests that clear
et al., 2003) in studies of Taiwanese cancer positive benefits of social support may exist
patients, 41.6 •¬ 5.4 for Thai breast cancer patients
women that may not exist for men, a conclusion
(n=71) (Wonghongkul, 2000), 37.8±7.0 for eld- supported by our study as well.
erly Japanese (n=87) (Koizumi et al., 1999), 33.9± Model 1 showed no correlation between HHI
5.1 for adult laborers (n=1,909) (Konishi and score and age, but a significant correlation did
Esaki, 2004), and 32.16•¬7.68 for ALS patients on appear in Model 2. Previous research has shown
invasive mechanical ventilation (n=157) (Hirano et no consistent pattern of correlation between age
al., 2006). While we did not investigate the cause and hope (Herth,1992 ; Konishi and Esaki, 2004
of this finding, we can postulate that it may demon Wonghongkul,2000) ; this study showed a signifi
strate the ability of individuals in adverse situa cant correlation only when social support was fac
tions to adapt to and cope with illness-related diffi tored in. In our sample of the general urban popula
culty through the maintenance of hope. tion, younger subjects showed a tendency towards
•@In order to investigate factors correlated with lower HHI scores except when they enjoyed suffi
the HHI score, we first performed a preliminary cient social networks-in these cases their HHI
bivariate analysis of factors thought to affect the scores were as high as older subjects. Reasons
HHI score-demographic attributes, stress, and that advanced age was associated with higher HHI
social support-with the HHI score, after which are thought to include freedom from the burdens
we performed a stratified multiple linear regres- of work and childcare and increased latitude in
sion analysis. The final model showed that higher terms of finances and time.
HHI scores were significantly correlated with Marital status showed a strongly significant cor-
higher age, presence of a spouse, and sufficient relation with HHI score regardless of age or
psychosocial support. gender. This result is in agreement with previous
•@In Model 1, women were found to have signifi research (Herth,1992 ; Konishi and Esaki, 2004).
cantly higher HHI scores than men. However, in We also found no significant correlations
Model 2 (which incorporated social support) we between HHI score and illness or adversity. Previ-
observed a significant score correlation with the ous research involving patients with severe dis
number of people offering social support, while the ease has shown that lower levels of HHI are corre-
correlation with gender disappeared. This result lated with illness-related experiences and distress
explains the higher HHI scores seen in women, (Chen, 2003 ; Herth, 1992 ; Hirano et al., 2006;
who receive greater support, than men. Higher Lin et al., 2003 ; Wonghongkul, 2000). This is
HHI scores have been previously reported in likely because severe illnesses create suffering
women (Konishi and Esaki, 2004), and our data that may persist intermittently up until the time of
suggest this outcome may be mediated by social the HHI survey itself. In contrast, the survey ques
support parameters. Other research has shown tion asked whether such adversity had been experi
that women enjoy superior social support net- enced in the past three years, a relatively long
works because for women these networks are period that allowed for the suffering associated
broader (Yoshii et al., 2005) and because only for with the experience to dissipate. This may explain
Jpn J Health & Human Ecology 41
the lack of a correlation found between illness, future by studying a more diverse subject sample
adversity, and physical limitations and the HHI which also includes individuals in rural areas.
score in the current study.
V Conclusion
We investigated the relationship between social
support and HHIs score through the ANCOVA Our results confirm the reliability and validity of
test. While a significant correlation was found the Japanese version of the Herth Hope Index
between the HHI score and the number of people (HHI) 12 scale by demonstrating a satisfactory
to whom subjects gave psychosocial support, and Cronbach a coefficient and strong positive correla
the number of people whom they trusted, these tions of HHI scores with both reasons for living
correlations disappeared in stratified multiple lin- and benefit-finding. We have shown that HHI
ear regression analysis when the number of people scores in the general urban population are not
offering psychosocial support was factored in. This unexpectedly high compared to the subjects of pre
suggests that subjects who gave abundant support vious research with serious diseases. We have also
to others and who maintained more than two trust- shown that higher age correlates with higher HHI
ing relationships enjoyed networks which gave scores, and that the receipt of social support
them plentiful support as well, resulting in higher appears to have a mediating effect on HHI scores.
HHI scores. Hope is cultivated through relation-
ships with others (Farran et al., 1995 ; Herth, Acknowledgments
1992), and the positive correlation between social We would like to express our gratitude to the
support and hope has been well established (Herth, respondents who took the time to participate in
1990 ; Herth, 2000 ; Hirano et al., 2006). This our survey, as well as to our fellow graduate stu-
study supports this conclusion and is the first dents and other individuals whose invaluable assis
research to elucidate a mediating effect between tance made this study possible.
Yuko HIRANO, Mayumi SAKITA, Yoshihiko YAMAZAKI, Kaoru KAWAI, Miho SATO
Jpn J Health & Human Ecology,73(1)31-42, 2007
和文抄録
頼 性 お よ び 妥 当 性 を 示 した 上 で,HHI得 点 とそ の 関 連 要 因 に つ い て 明 らか に す る こ とで あ る。 東 京 都N区
デ ィ ン グ と強 い 正 の 相 関 が 認 め られ た こ とか ら,信 頼 性 と構 成 概 念 妥 当 性 は概 ね 示 さ れ た と言 え る.HHI
十 分 に得 られ て い る こ とに よ っ てHHI得 点 が 男 性 よ り高 くな っ て い る可 能 性 が 示 され た.逆 境 経 験 と病 い
経 験 とは 有 意 な 関 連 は 見 られ な か っ た.以 上 よ り,重 篤 な 病 い を もつ 人 々 を 対 象 と した 先 行 研 究 と比 べ る
と,本 研 究 で は,一 般 住 民 のHHI得 点 は 決 して 高 い 値 で は な い こ と,HHI得 点 の 特 有 の 関 連 要 因 と して
年 齢 が あ げ られ る こ と,ソ ー シ ャ ル サ ポ ー トの 受 領 はHHI得 点 に 対 して 媒 介 効 果 を も つ こ とが 明 らか に
な っ た.