JAN ORIGINAL RESEARCH
Factors influencing osteoporosis preventive behaviours: testing a path
model
Ching-Hsing Hsieh, Chen-Yen Wang, Marilyn McCubbin, Shuqiang Zhang & Jillian Inouye
Accepted for publication 20 December 2007
Correspondence to C.-H. Hsieh:
e-mail: chhsieh@gw.cgit.edu.tw
Ching-Hsing Hsieh PhD RN
Assistant Professor
Nursing Department, Chang Gung Institute
of Technology, Taoyuan, Taiwan
Chen-Yen Wang PhD RN
Associate Professor
School of Nursing and Dental Hygiene,
University of Hawaii at Manoa, Honolulu,
Hawaii, USA
Marilyn McCubbin PhD RN FAAN
Professor Emerita
School of Nursing, University of
Wisconsin – Madison, Wisconsin, USA
Shuqiang Zhang PhD
Associate Professor
College of Education, University of Hawaii at
Manoa, Honolulu, Hawaii, USA
Jillian Inouye PhD RN
Professor, Director of Research
School of Nursing and Dental Hygiene,
University of Hawaii at Manoa, Hawaii, USA
336
HSIEH C.-H., WANG C.-Y., MCCUBBIN M., ZHANG S. & INOUYE J.
(2008) Factors influencing osteoporosis preventive behaviours: testing a path
model. Journal of Advanced Nursing 62(3), 336–345
doi: 10.1111/j.1365-2648.2008.04603.x
Abstract
Title. Factors influencing osteoporosis preventive behaviours: testing a path model.
Aim. This paper is a report of a study to test a model of certain factors influencing
people engaging in osteoporosis preventive behaviours, and to estimate the direct
and indirect effects of personal and social factors on whether people engage in those
behaviours.
Background. Osteoporosis preventive behaviours, including exercise and calcium
intake, help decreasing the risk of developing osteoporosis. Reasons for engaging in
osteoporosis preventive behaviours are complex and influenced by personal and
social factors. Years of education, self-efficacy, knowledge of osteoporosis, social
support and social capital have been indicated to increase people engaging in
osteoporosis preventive behaviours; but age has been shown to decrease those
behaviours.
Methods. The proposed model was developed using Social Cognitive Theory and a
conceptual framework for addressing the social context of health behaviour. A
correlational cross-sectional study was carried out in 2005, using questionnaires and
a convenience sample of 243 participants. Data analysis included descriptive sta-
tistics, correlation techniques and path analysis.
Results. Self-efficacy was a better predictor of engaging in osteoporosis preventive
behaviours than were the other variables. Social capital had a statistically significant
direct and indirect effect on osteoporosis preventive behaviours. The modified path
model showed good fit with the data.
Conclusion. The associations between personal and social factors extend our
knowledge from previous studies and increase our understanding of the complex
relationships among the study variables. The model provides guidance for future
nursing practice, research, and education programs related to osteoporosis pre-
vention.
Keywords: knowledge, nursing, osteoporosis, path analysis, prevention, self-
efficacy, social capital, social support
 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
JAN: ORIGINAL RESEARCH
Introduction
Osteoporosis is recognized as an increasing public health
concern because of its serious consequences such as
fracture and mortality. It affects more than 75 million
people in the United States of America, Europe and Japan,
and causes more than 2Æ5 million fractures in Western
Pacific countries including Australia, China, India, Japan,
New Zealand and Republic of Korea annually (World
Health Organization, 2004). People with osteoporosis are
at a higher risk of developing problems with physical
infirmity and consequent difficulties in carrying out activ-
ities of daily living, and may experience reduced quality of
life due to restricted opportunities to enjoy leisure and
social activities (Kotz et al. 2004). Women with osteopo-
rosis have higher levels of depressive symptoms and a
higher prevalence of depression than women without
osteoporosis (Coelho et al. 1999).
The incidence of osteoporosis increases from 1Æ13% in the
21–30-year age group to 54Æ55% in those aged 80 years and
above in the Taiwanese population (Lin et al. 2001). There
are about 4000 new hip fracture patients in Taiwan each
year, and osteoporosis is the main cause for these fractures
(Tsai 2000). The total economic loss in Taiwan resulting
from hip fractures, including direct and indirect medical and
social costs, was US$ 120 million in 1993 (Tsai 2000).
Prevention is a cost-effective means of managing osteoporosis
(Orces et al. 2003) and can reduce morbidity and economic
cost caused by osteoporosis. If preventive strategies of
osteoporosis are carried out, burdens to the family and
society will decrease.
Nursing professionals should understand the factors that
affect the likelihood of people engaging in osteoporosis
preventive behaviours. After nurses recognize these factors,
they can use their knowledge and skills to provide nursing
care, and help people improve their health and adopt
behaviours to prevent osteoporosis.
Factors influencing whether people engage in osteoporosis
preventive behaviours include age, years of education, self-
efficacy, knowledge of osteoporosis, social capital and
social support (Piaseu et al. 2002, Lindstrom 2003, Resnick
& Nigg 2003, Sorensen et al. 2003). However, relation-
ships among these variables have not been examined
through research. Therefore, the purposes of this study
were to test a model of certain factors influencing people
engaging in osteoporosis preventive behaviours, and to
estimate the direct and indirect effects of personal and
social factors on whether people engage in those behav-
iours.
 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
Factors influencing osteoporosis preventive behaviours
Background
Osteoporosis preventive behaviours
Osteoporosis is a silent disease and is a condition of low bone
mass resulting from an excess of bone consumption over bone
formation. It is a skeletal disorder characterized by compro-
mised bone strength, thereby increasing the risk of bone
fractures (Lappe 2001). Weight-bearing exercises can in-
crease calcaneal broadband ultrasound attenuation, which is
an indirect estimation of bone mineral density (Ay &
Yurtkuran 2005). Babaroutsi et al. (2005) showed that adult
men with a calcium intake exceeding 800 mg/day had
statistically significantly higher bone mass than those con-
suming less calcium. Therefore, in the present study on
osteoporosis preventive behaviours, we only focused on
primary prevention, including calcium intake and exercise.
Development of the model
Development of the model of factors influencing whether
people engage in osteoporosis preventive behaviours (Fig-
ure 1) was based on Social Cognitive Theory (Bandura 1986,
1997, 2004), and the conceptual framework for addressing
the social context of health behaviour described by Sorensen
et al. (2003). Social Cognitive Theory proposes that cognitive
and other personal, environmental and behavioural factors
operate as reciprocal interacting determinants of each other
(Bandura 1986). Social capital is an important component
affecting health behaviour (Sorensen et al. 2003). Therefore,
the factors influencing whether people engage in osteoporosis
preventive behaviours include personal factors (age, years of
education, self-efficacy and knowledge of osteoporosis) and
social factors (social support and social capital) were
included in the study.
Personal factors
Age has been suggested as being one of the most important
factors affecting health behaviours. Older women are more
likely to report frustration with regard to their knowledge
and practices in relation to osteoporosis prevention than
younger women (Hsieh et al. 2001). Age is negatively related
to knowledge of osteoporosis and social support (Chang et al.
2003, Resnick & Nigg 2003). Additionally, level of educa-
tion positively influences social support (Muhlenkamp &
Sayles 1986, Katapodi et al. 2002) and knowledge of osteo-
porosis (Yeh et al. 2002).
Personal knowledge is defined as the understanding of a
new pattern through processing (Sweeney 1994). This
337
C.-H. Hsieh et al.
– Knowledge
Age
+
+
– Calcium Intake
–
Years of Education Social Support
+
+ +
Social Capital
Figure 1 Hypothesized model of factors influencing osteoporosis preventive behaviours.
processing may consist of any combination of human and
environmental interaction (experience), rational intuition,
appraisal, active comprehension and personal judgment
(Sweeney 1994). Personal knowledge is objective because it
makes a person aware of how knowledge affects a situation
(Polanyi 1958). Knowledge of osteoporosis as been reported
to influence calcium intake and exercise indirectly through
self-efficacy in Thai nursing students (Piaseu et al. 2002).
Self-efficacy is defined as people’s judgment of their capa-
bilities to organize and execute courses of action required to
attain designated types of performances (Bandura 1986). Self-
efficacy was statistically significantly correlated with calcium
intake and exercise in a study by Ali and Twibell (1995).
Exercise self-efficacy has been shown to be a positive predictor
of weight-bearing exercise and dietary calcium intake, both of
which help prevent osteoporosis (Wallace 2002). Furthermore,
self-efficacy has been positively related to calcium intake
(Ievers-Landis et al. 2003), and self-efficacy statistically
significantly associated with exercise (Resnick & Nigg 2003).
Social factors
Social support is a form of interpersonal interaction that
provides psychological support or specific aid to help solve a
problem, where the help mainly comes from personal social
connections rather than strangers (Simmons 1994). A study
by Resnick and Nigg (2003) showed that social support is
positively associated with self-efficacy. In addition, social
support statistically significantly predicts knowledge of cal-
cium intake and knowledge of weight-bearing physical
activity (Ievers-Landis et al. 2003).
Social capital refers to the values and beliefs that citizens
share in their everyday dealings, and it is accumulated in the
community through processes of interaction and learning
(Maskell 2000). Social capital also refers to the quantity and
338
+ Self-Efficacy
Calcium intake +
+
+
Self-Efficacy
+ Exercise
+
Exercise
quality of social relationships, such as formal and informal
social connections as well as norms of reciprocity and trust
that exist in a place of community (Kawachi & Berkman
2000). It emphasizes not only primary connections such as
kinship, but also secondary connections such as people’s
social networks and civic engagements (Field et al. 2000).
The impacts of social capital on individual behaviours have
been documented. For example, daily smoking is negatively
associated with social capital (Lindstrom 2003), and social
capital is also negatively correlated with risky sexual behav-
iour, whereas it is positively correlated with protective
behaviour (Crosby et al. 2003). Social capital may affect
social support in certain ways. Therefore, it is important to
conduct further research to examine this statement (Sorensen
et al. 2003).
The study
Aims
The aims of the study were to test a model of factors
influencing whether people engage in osteoporosis preventive
behaviours, and to estimate the direct and indirect effects of
personal and social factors on whether people engage in those
behaviours.
Design
A correlational cross-sectional design was adopted.
Participants
A convenience sample of community-dwelling participants
living in a small agricultural town in Taiwan was recruited.
 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
JAN: ORIGINAL RESEARCH
All participants were aged over 18 years and could speak
Mandarin or Hakka to communicate with the interviewer.
The sample was comprised of Hakka people, who are a
minority ethnic group in Taiwan and have their own Hakka
dialect, dietary customs and cultural values. Hakka people
still maintain traditional dietary habits such as consuming
salty foods, and abundant vegetables and fruits (Liu & Li
2000). The sample size was calculated based on the path
analysis model. There should be a ratio of at least five
participants to each parameter to be estimated in path
analysis (Hatcher 2003). In total, 23 parameters were used in
this study; hence, the minimum sampling size based on this
rule of thumb was 115.
We invited 250 people to participate in this study
through the assistance of six residents who had lived in
the community for over 60 years; of these, 243 (97Æ2%),
123 women and 120 men, accepted the invitation. Seven
people declined because of a lack of interest and time
constraints. Of the 243 people who agreed to participate,
six returned incomplete questionnaires (3Æ5%), mainly
because of time constraints. Missing data were selectively
deleted on a variable-by-variable basis (Polit & Hungler
1995).
Data collection
Data were collected by a trilingual researcher using ques-
tionnaires. Some participants in the study were illiterate, and
so a structured scale and face-to-face interviews were used to
collect data. The data were collected through home visits
from March to September, 2005.
Demographics
Demographic data were collected in part 1 of the question-
naire, including age, gender, years of education, marital sta-
tus, spouse’s years of education and type of chronic disease
diagnosed by a physician. Age and years of education were
the variables examined in this study.
Social support
The Personal Resource Questionnaire 85 – Part 2 (PRQ – Part
2) (Brandt & Weinert 1981) was used to measure social
support. It is a 25-item instrument which uses a seven-point
Likert scale with responses ranging from ‘strongly disagree’
to ‘strongly agree’. A higher score indicates a higher degree of
perceived social support. This scale has previously been
translated into Chinese by Wang (1999) and has been used
previously in Taiwan. The test–retest reliability coefficient at
an interval of 1 week was 0Æ80. Cronbach’s a was 0Æ77 in our
study.
 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
Factors influencing osteoporosis preventive behaviours
Self-efficacy
Horan et al. (1998) developed and evaluated the 12-item
Osteoporosis Self-Efficacy Scale (OSE). Factor analysis of
the responses to the self-efficacy items revealed a logical,
theoretically meaningful two-factor structure, one for phys-
ical activity (six items) and one for calcium intake (six
items). We used the six-item Osteoporosis Self-Efficacy Scale
(OSE)-physical activity to measure self-efficacy for exercise,
and its Cronbach’s a was 0Æ92.The six-item Osteoporosis
Self-Efficacy Scale (OSE)-calcium intake was used to mea-
sure self-efficacy for calcium intake, and its Cronbach’s a
was 0Æ93.
Knowledge of osteoporosis
The Facts on Osteoporosis Quiz (FOOQ) was developed
based on the Self-Care Deficient Theory (Ailinger et al.
2003); it consists of 20 items and was used to measure
knowledge of osteoporosis.. The total number of correct
responses on the FOOQ (ranging from 0 to 20) was cal-
culated as a person’s score. Cronbach’s a was 0Æ78 in this
study.
Social capital
The 31-item Scale of Social Capital (SSC) (Onyx & Bullen
2000) uses a four-point Likert scale with responses ranging
from ‘none’ to ‘very much’. The total score ranges from 0 to
124, with a higher score indicating a higher degree of social
capital. Cronbach’s a was 0Æ84 in this study.
Calcium intake
The Scale of Calcium Intake (SCI) (Lin 1999) was used to
measure the calcium intake of 357 Taiwanese women aged
30–49 years, with acceptable reliability and validity. Con-
tent validity was used and the test–retest reliability coeffi-
cient of the SCI at an interval of 1 week was 0Æ82 among
30 Taiwanese aged over 20 years in our study. In the
present study, the researcher asked the following questions
to an interviewee: In the past week, what calcium-rich food
such as milk, yogurt, cheese, tofu, dry tofu, small dried fish,
kelp, beans and dark-coloured vegetables did you eat? How
many times did you eat that food? What amount did you
eat each time? The names, measurement of calcium-rich
foods and the kind of beans and dark-coloured vegetables
were provided in the scale. Interviewees could use this
reference material to help answer questions after recalling
their food intake over the past week. The researcher
then used the information provided by the scale on the
amount of calcium per unit of food to calculate the calcium
intake per day. The daily calcium intake was calculated as
follows:
339
C.-H. Hsieh et al.

number of times an interviewee had a specific kind of food were illiterate, this was read to them. Anonymity was
in the week  amount of that kind of food consumed maintained via code numbers and protected files.
had each time  amount of calcium contained in one
unit of that kind of food  by 7: Data analysis

Exercise
The 16-item Physical Activity Questionnaire (PAQ) (Liu et al.
2001) for exercise was used to measure exercise behaviour.
The test–retest reliability was analysed by an interclass cor-
relation coefficient (ICC), values were 0Æ77, 0Æ70, and 0Æ80
for renal patients, cancer patients and elderly subjects,
respectively, while that for all subjects was 0Æ74. In the ori-
ginal study, validity was examined in 31 healthy people by
measuring the basal metabolic rate by indirect calorimetry
and multiplying by the physical activity level obtained from
published studies using the double-labelled water method and
also from Food and Agriculture Organization, World Health
Organization and United Nations (Liu et al. 2001).
When collecting data on exercise, the researcher asked the
following questions to interviewees: What exercises did you
do during the last week? And for each exercise you did, how
much time (hours) did you spend on that exercise each time?
Interviewees were asked to use a usual week as reference to
answer the questions. The researcher then used the informa-
tion provided by the PAQ on the number of kilocalories
expended per minute for each kind of exercise to calculate the
interviewee’s total kilocalorie expenditure through exercise a
week.
Ethical considerations
This study was approved by a university institutional review
board. To select participants, we explained the purpose of the
study and the length of the interview to those who might have
an interest in participating. Before prospective participants
gave oral agreement to participate, those who could read
were asked to read an information leaflet. For those who
Data were analysed using LISREL 8 and SPSS 12. Pearson’s
correlations were used to assess relationships between vari-
ables. Observed relationships between study variables helped
to predict an appropriate model (Sonderegger et al. 2004).
Path analysis using regression procedures was used to test and
modify the hypothesized model to specify causal relationships
(Hatcher 2003). Statistical significance was set at P < 0Æ05.
Multicollinearity is a confounder for regression analysis.
Therefore, collinearity diagnostics were performed for each
of the variables in the full model with variance inflation
factors (VIF) and tolerance in this study. Ai et al. (2006)
recommended that a VIF of <3 is an acceptable limit. The
indicators of the model fit included a chi-squared goodness-
of-fit, an adjusted goodness-of-fit index (AGFI), goodness-of-
fit index (GFI), comparative fit index (CFI), normed fit index
(NFI), and non-normed fit index (NNFI). Values of these
indices >0Æ9 indicate an acceptable fit.
Results
Participants’ mean age was 46Æ51 years (SD = 14Æ53;
range = 18–82). Their mean level of education was
11Æ79 years (SD = 4Æ08; range = 0–24). Thirty-five percent
were involved in local community activities such as tai-chi
and a mothers’ support group.
Descriptive analysis of the study variables is presented in
Table 1. Mean daily calcium intake was 352Æ67 mg
(SD = 152Æ28). Over half of the participants (59Æ67%) exer-
cised regularly (more than thrice a week). As for the amount
of energy expended due to exercise, the minimum was 0 kcal
and maximum 6840 kcal (mean = 1430Æ32; SD = 1323Æ08).
The mean social capital score was 76Æ76 (SD = 10Æ38), mean

Table 1 Means, standard deviations (SD ), actual ranges and possible ranges of study variables

Variable Scale Mean SD Actual range Possible range

Calcium intake (mg/day) Scale of calcium intake 352Æ67 152Æ28 26–824 0–unlimited
Exercise (kcal/week) Physical activity questionnaire 1430Æ32 1323Æ08 0–6840 0–unlimited
Social capital (points) Scale of social capital 76Æ76 10Æ38 51–102 31–124
Self-efficacy for calcium intake Osteoporosis self-efficacy scale for 38Æ89 12Æ23 0–60 0–60
(points) calcium intake
Self-efficacy for exercise (points) Osteoporosis self-efficacy scale 33Æ89 14Æ22 0–60 0–60
for physical activity
Social support (points) Personal resource questionnaire 85–Part 2 109Æ76 12Æ69 61–141 25–175
Knowledge (points) Facts on osteoporosis quiz 12Æ07 3Æ01 4–18 0–20

340  2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
JAN: ORIGINAL RESEARCH Factors influencing osteoporosis preventive behaviours

self-efficacy for calcium intake score was 38Æ89 (SD = 12Æ23), self-efficacy for exercise to self-efficacy for calcium intake; the
mean self-efficacy for exercise score was 33Æ89 (SD = 14Æ22), path from social capital to self-efficacy for exercise; and the
mean social support score was 109Æ76 (SD = 12Æ69), and path from age to calcium intake; and three paths be deleted:
mean knowledge of osteoporosis score was 12Æ07 (SD = 3Æ01). the path from age to knowledge, the path from years of
education to social support, and the path from knowledge to
self-efficacy for calcium intake in the path diagram. The final
Relationships among variables
model had an improved v2 (d.f. = 21) of 26Æ99 (P = 0Æ171,
The relationships among age, years of education, knowledge GFI = 0Æ98, AGFI = 0Æ95, NFI = 0Æ96, NNFI = 0Æ98, and
of osteoporosis, self-efficacy for exercise, self-efficacy for CFI = 0Æ99), indicating that the final model fits the data.
calcium intake, social support, and social capital were The results of model testing are presented in Table 3 and
analysed using Pearson’s correlation coefficients (r) (Table 2). Figure 2. In the resulting model, 46% of the total variance in
Self-efficacy for calcium intake was positively correlated with years of education was accounted for by age (b = 0Æ68);
calcium intake. Self-efficacy for exercise and social capital 21% of the total variance in knowledge of osteoporosis was
were positively correlated with exercise. Knowledge of
osteoporosis and social support were positively correlated
with self-efficacy for calcium intake. Knowledge of osteopo-
Table 3 Effects of exogenous variable on endogenous variable
rosis and social support were also positively correlated with
self-efficacy for exercise. Social support and years of educa- Endogenous Exogenous
variable variable b F Total R2 P
tion were positively correlated with knowledge of osteopo-
rosis. Age was negatively correlated with knowledge of CI Age 0Æ18** 30Æ07** 0Æ16** 0Æ001
osteoporosis, social support and years of education. Years of SECa 0Æ34**
education and social capital were positively correlated with Ex SC 0Æ14* 30Æ83** 0Æ25** 0Æ001
SEE 0Æ44**
social support.
SECa SP 0Æ22** 92Æ50** 0Æ46** 0Æ001
SEE 0Æ57**
SEE SC 0Æ25** 13Æ43** 0Æ15** 0Æ001
Measurement model
K 0Æ12*
Based on the correlation matrix, the hypothesized model was SP 0Æ15*
SP Age 0Æ14** 47Æ76** 0Æ29** 0Æ001
tested using path analysis. The hypothesized model revealed a
SC 0Æ51**
statistically significant Chi-square statistic (v2 = 148Æ60,
K Edu 0Æ37** 30Æ07** 0Æ21** 0Æ001
d.f. = 21, P = 0Æ001), and a low CFI of 0Æ78. This suggested SP 0Æ22**
that the hypothesized model inadequately accounted for the Edu Age 0Æ68** 201Æ17** 0Æ46** 0Æ001
input correlation matrix reflecting relationships among the
CI, calcium intake; Ex, exercise; SC, social capital; SECa, self-efficacy
variables in the hypothesized model. To achieve a better for calcium intake; SEE, self-efficacy for exercise; SP, social support;
fit between the model and data, the output of path K, knowledge; Edu, education.
analysis suggested that three paths be added: the path from *P < 0Æ05; **P < 0Æ01.

Table 2 Correlation matrix of all study variables

Variable 1 2 3 4 5 6 7 8 9

1. Age 1Æ00
2. SC 0Æ07 1Æ00
3. Edu 0Æ68*** 0Æ11 1Æ00
4. K 0Æ31*** 0Æ15* 0Æ40*** 1Æ00
5. SP 0Æ18** 0Æ52*** 0Æ17* 0Æ28*** 1Æ00
6. SECa 0Æ02 0Æ35*** 0Æ09 0Æ23*** 0Æ40*** 1Æ00
7. SEE 0Æ05 0Æ33*** 0Æ06 0Æ20** 0Æ31*** 0Æ64*** 1Æ00
8. CI 0Æ19** 0Æ23*** 0Æ14 0Æ13 0Æ26*** 0Æ34*** 0Æ27*** 1Æ00
9. Ex 0Æ12* 0Æ26*** 0Æ09 0Æ01 0Æ19** 0Æ30*** 0Æ49*** 0Æ19** 1Æ00

CI, calcium intake; Ex, exercise; SC, social capital; SECa, self-efficacy for calcium intake; SEE, self-efficacy for exercise; SP, social support;
K, knowledge; K, knowledge; Edu, education.
*P < 0Æ05; **P < 0Æ01; ***P < 0Æ001.

 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd 341
C.-H. Hsieh et al.
Age
R2 = 0·21*
–0·14*
Knowledge
–0·68*
0·22*
0·37*
Years of education Social support
R2 = 0·46* R2 = 0·29*
0·51*
Social capital
Figure 2 Final model of factors influencing osteoporosis preventive behaviours.
accounted for by years of education (b = 0Æ37) and social
support (b = 0Æ22); 29% of the total variance in social
support was accounted for by social capital (b = 0Æ51) and
age (b = 0Æ14); 46% of the total variance in self-efficacy for
calcium intake was accounted for by self-efficacy for exercise
(b = 0Æ57) and social support (b = 0Æ22); 15% of the total
variance in self-efficacy for exercise was accounted for by
knowledge of osteoporosis (b = 0Æ12), social support
(b = 0Æ15), and social capital (b = 0Æ25); 16% of the total
variance in calcium intake was accounted for by age
(b = 0Æ18) and self-efficacy for calcium intake (b = 0.34);
and 25% of the total variance in exercise was accounted for
by social capital (b = 0Æ14) and self-efficacy for exercise
(b = 0Æ44).
Discussion
Study limitations
The study had the following limitations. First, owing to time
constraints, limited human resources and budget, a conve-
nience sample was used with people who were inhabitants of
a particular community. It is recommended that in future
studies, random sampling and different samples from differ-
ent cultural and ethnic backgrounds should be used to test the
generalizability of the results. Second, a cross-sectional design
was used in this study. A longitudinal study with a repeated-
measure design should be used to observe changes over time
342
R2 = 0·16*
–0·18*
Calcium intake
R2 = 0·46*
0·34*
Self–efficacy
calcium intake
0·12*
0·57*
0·22*
Self–efficacy
0·15* exercise R2 = 0·15*
0·25*
0·44*
0·14*
P < 0·05
Exercise
R2 = 0·25*
in personal and social factors affecting osteoporosis pre-
ventive behaviours in the future. Despite these limitations, the
findings contribute to our understanding of the complex
relationships of the factors influencing osteoporosis preven-
tive behaviours.
Discussion of results
Our results showed that the combined components of Social
Cognitive Theory (Bandura 1986, 1997, 2004) and the
conceptual framework for addressing the social context of
health behaviour (Sorensen et al. 2003) explained much of
the variance in osteoporosis preventive behaviour. However,
much unexplained variance remains; therefore, ongoing effort
is needed to identify critical variables that might increase the
explanatory power of the model (Yarcheski & Mahon 1989).
For example, the findings of Walcott-McQuigg and Prohaska
(2001) indicated that motivation influenced the desire and
ability to exercise. Individual perceptions may impede or
increase health promotion behaviour (Ali & Twibell 1995).
These factors should be addressed in future studies.
In this study, social capital was found to have a direct effect
on exercise and a statistically significant indirect effect by
way of social support and self-efficacy for exercise, indicating
that people with high levels of social capital had greater self-
efficacy and were more likely to exercise; however, it is not
yet clear which factors affect social capital and how social
capital can in turn affect behaviour. It is, therefore, suggested
 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
JAN: ORIGINAL RESEARCH
What is already known about this topic
• Age, years of education, knowledge of osteoporosis,
self-efficacy and social support may affect people’s
engagement in osteoporosis preventive behaviours.
• Social capital may affect personal behaviour.
What this paper adds
• Social capital had a statistically significant direct effect
on exercise and indirect effect through social support
and self-efficacy on exercise.
• Social capital indirectly affects calcium intake through
social support, self-efficacy for calcium intake, and self-
efficacy for exercise.
• Self-efficacy directly affects calcium intake and exer-
cise.
• The model of factors influencing osteoporosis pre-
ventive behaviours offers guidance in designing future
nursing interventions, research and public health pro-
grammes.
that a qualitative study be carried out to elucidate the factors
influencing social capital and plausible pathways through
which social capital might influence individual behaviours.
In addition, in this study, people with high social capital
reported high levels of social support, which was addressed in a
study by Sorensen et al. (2003). High levels of social support
were correlated with higher knowledge of osteoporosis, and
with self-efficacy (Ievers-Landis et al. 2003). Participants who
reported high levels of social support also reported high levels
of self-efficacy and knowledge of osteoporosis. Our results
showed that self-efficacy for calcium intake and exercise
directly affected osteoporosis preventive behaviour, indicating
that as self-efficacy improves the chance that a person will
engage in osteoporosis preventive behaviour also improves.
These findings match the results of studies by Ali and Twibell
(1995), Walcott-McQuigg and Prohaska (2001) and Ievers-
Landis et al. (2003). Thus, future applications in nursing
practice need to increase levels of self-efficacy for exercise and
calcium intake and social support for community-dwelling
people, thereby improving their likelihood of engaging in
osteoporosis preventive behaviours.
Our results show that knowledge of osteoporosis had no
statistically significant direct effect on calcium intake and
exercise, indicating that even as knowledge of osteoporosis
increases, the likelihood that a person will engage in
osteoporosis preventive behaviours may remain the same.
 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
Factors influencing osteoporosis preventive behaviours
This result is consistent with those of Terrio and Auld (2002).
The degree to which knowledge affects behaviour is still
controversial. One argument supporting this view is that it is
not necessarily true that individuals who are knowledgeable
about what constitutes healthy behaviour will engage in that
behaviour (Piaseu et al. 2002). Therefore, reconsideration is
needed for whether boosting the knowledge of osteoporosis
should be a goal in preventive programmes.
Orem (1995) proposed that age was one of the factors
influencing self-care behaviour. Her theory was supported by
our results: age had negative direct and indirect effects on
calcium intake, indicating that the older the participants, the
lower their calcium intake, leading to an increased risk for
osteoporosis in older people. Therefore, in providing nursing
care related to increasing osteoporosis preventive behaviours,
age is an important element to be considered.
Conclusion
Social capital, self-efficacy and other variables, as described
by Social Cognitive Theory, and the conceptual framework
for addressing the social context of health behaviours, are
factors influencing osteoporosis preventive behaviours. The
model of factors influencing osteoporosis preventive behav-
iours offers guidance for designing future nursing practice,
research, and education related to osteoporosis prevention. In
future research, attempts should be made to explain the
remaining variance in osteoporosis preventive behaviours not
explained in this study.
Acknowledgements
We greatly appreciate the time that participants devoted to
the study and Dr Chou’s valuable advice about this paper.
Author contributions
CH was responsible for the study conception and design. CH
performed the data collection. CH performed the data
analysis. CH was responsible drafting of the manuscript.
CH made critical revisions to the paper for important
intellectual content. SZ provided statistical expertise. CY,
MM and JI supervised the study.
References
Ai A.L., Peterson C., Bolling S.F. & Rodgers W. (2006) Depression,
faith-based coping, and short-term postoperative global function-
ing in adult and older patients undergoing cardiac surgery. Journal
of Psychosomatic Research 60(1), 21–28.
343
C.-H. Hsieh et al.
Ailinger R.L., Lasus H. & Braun M.A. (2003) Revision of the facts
on osteoporosis quiz. Nursing Research 52(3), 198–201.
Ali N.S. & Twibell R.K. (1995) Health promotion and osteoporosis
prevention among postmenopausal women. Preventive Medicine
24(5), 528–534.
Ay A. & Yurtkuran M. (2005) Influence of aquatic and weight-
bearing exercises on quantitative ultrasound variables in post-
menopausal women. American Journal of Physical Medicine &
Rehabilitation/Association of Academic Physiatrists 84(1), 52–61.
Babaroutsi E., Magkos F., Manios Y. & Sidossis L.S. (2005) Body
mass index, calcium intake, and physical activity affect calcaneal
ultrasound in healthy Greek males in an age-dependent and
parameter-specific manner. Journal of Bone and Mineral Metab-
olism 23(2), 157–166.
Bandura A. (1986) Social Foundations of Thought and Action: A
Social Cognitive Theory. Prentice-Hall, Englewood Cliffs.
Bandura A. (1997) Self-Efficacy: The Exercise of Control. Freeman,
New York.
Bandura A. (2004) Health promotion by social cognitive means.
Health Education & Behavior: The Official Publication of the
Society for Public Health Education 31(2), 143–164.
Brandt P.A. & Weinert C. (1981) The PRQ – a social support mea-
sure. Nursing Research 30(5), 277–280.
Chang S.F., Chen C.M., Chen P.L. & Chung U.L. (2003) Predictors
of community women’s osteoporosis prevention intention – a pilot
study. The Journal of Nursing Research 11(4), 231–240.
Coelho R., Silva C., Maia A., Prata J. & Barros H. (1999) Bone
mineral density and depression: a community study in women.
Journal of Psychosomatic Research 46(1), 29–35.
Crosby R.A., Holtgrave D.R., DiClemente R.J., Wingood G.M. &
Gayle J.A. (2003) Social capital as a predictor of adolescents’
sexual risk behavior: a state-level exploratory study. AIDS and
Behavior 7(3), 245–252.
Field J., Schuller T. & Baron S. (2000) Social capital and human
capital revisited. In Social Capital: Critical Perspectives (Baron S.,
Field J. & Schuller T., eds), Oxford University Press, New York,
pp. 243–263.
Hatcher L. (2003) A Step-by-Step Approach to Using SAS for Factor
Analysis and Structural Equation Modeling. SAS Institute, Carry,
NC.
Horan M.L., Kim K.K., Gendler P., Froman R.D. & Patel M.D.
(1998) Development and evaluation of the Osteoporosis Self-Effi-
cacy Scale. Research in Nursing & Health 21(5), 395–403.
Hsieh C., Novielli K.D., Diamond J.J. & Cheruva D. (2001) Health
beliefs and attitudes toward the prevention of osteoporosis in older
women. Menopause 8(5), 372–376.
Ievers-Landis C.E., Burant C., Drotar D., Morgan L., Trapl E.S. &
Kwoh C.K. (2003) Social support, knowledge, and self-efficacy
as correlates of osteoporosis preventive behaviors among pread-
olescent females. Journal of Pediatric Psychology 28(5),
335–345.
Katapodi M.C., Facione N.C., Miaskowski C., Dodd M.J. & Waters
C. (2002) The influence of social support on breast cancer
screening in a multicultural community sample. Oncology Nursing
Forum 29(5), 845–852.
Kawachi I. & Berkman L. (2000) Social cohesion, social capital, and
health. In Social Epidemiology (Kawachi I., ed.), Oxford Univer-
sity Press, New York, pp. 174–190.
344  2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
Kotz K., Deleger S., Cohen R., Kamigaki A. & Kurata J. (2004)
Osteoporosis and health-related quality-of-life outcomes in the
Alameda County study population. Preventing Chronic Disease
1(1), 05.
Lappe J.M. (2001) Pathophysiology of osteoporosis and fracture. The
Nursing Clinics of North America 36(3), 393–400.
Lin Y.P. (1999) A Study on Osteoporosis Preventive Behavior of
Female Employees in a Workplace. National Taiwan University,
Taipei.
Lin J.D., Chen J.F., Chang H.Y. & Ho C. (2001) Evaluation of bone
mineral density by quantitative ultrasound of bone in 16,862
subjects during routine health examination. The British Journal of
Radiology 74(883), 602–606.
Lindstrom M. (2003) Social capital and the miniaturization of
community among daily and intermittent smokers: a population-
based study. Preventive Medicine 36(2), 177–184.
Liu X.Q. & Li Y.H. (2000) Epidemiological and nutritional research
on prevention of cardiovascular disease in China. The British
Journal of Nutrition, 84(2), 199–203.
Liu B., Woo J., Tang N., Ng K., Ip R. & Yu A. (2001) Assessment of
total energy expenditure in a Chinese population by a physical
activity questionnaire: examination of validity. International
Journal of Food Sciences and Nutrition 52(3), 269–282.
Maskell P. (2000) Social capital, innovation, and competitiveness. In
Social Capital: Critical Perspectives (Baron S., Field J. & Schuller
T., eds), Oxford University Press, New York, pp. 111–123.
Muhlenkamp A.F. & Sayles J.A. (1986) Self-esteem, social
support, and positive health practices. Nursing Research 35(6),
334–338.
Onyx J. & Bullen P. (2000) Measuring social capital in five
communities. The Journal of Applied Behavioral Science 36(1),
23–42.
Orces C.H., Casas C., Lee S., Garci-Cavazos R. & White W. (2003)
Determinants of osteoporosis prevention in low-income Mexican-
American women. Southern Medical Journal 96(5), 458–464.
Orem D.E. (1995) Nursing: Concepts of Practice, 5th edn. St Louis,
Mosby.
Piaseu N., Schepp K. & Belza B. (2002) Causal analysis of exercise
and calcium intake behaviors for osteoporosis prevention among
young women in Thailand. Health Care for Women International
23(4), 364–376.
Polanyi M. (1958) Personal Knowledge: Towards a Postcritical
Philosophy. University of Chicago Press, Chicago.
Polit D.F. & Hungler B.P. (1995) Nursing Research: Principles and
Methods, 5th edn. Lippincott, Philadelphia.
Resnick B. & Nigg C. (2003) Testing a theoretical model of exercise
behavior for older adults. Nursing Research 52(2), 80–88.
Simmons S. (1994) Social networks: their relevance to mental health
nursing. Journal of Advanced Nursing 19(2), 281–289.
Sonderegger R., Barret P. & Creed P. (2004) Models of cultural
adjustment for child and adolescent migrants to Australia: internal
process and situational factors. Journal of Child and Family Studies
13(3), 357–371.
Sorensen G., Emmons K., Hunt M.K., Barbeau E., Goldman R.,
Peterson K., Kuntz K., Stoddard A. & Berkman L. (2003) Model
for incorporating social context in health behavior interventions:
applications for cancer prevention for working-class, multiethnic
populations. Preventive Medicine 37(3), 188–197.
JAN: ORIGINAL RESEARCH
Sweeney N.M. (1994) A concept analysis of personal knowledge:
application to nursing education. Journal of Advanced Nursing
20(5), 917–924.
Terrio K. & Auld G.W. (2002) Osteoporosis knowledge, calcium
intake, and weight-bearing physical activity in three age groups of
women. Journal of Community Health 27(5), 307–320.
Tsai K. (2000) Osteoporosis fracture rate, bone mineral density, and
bone markers in Taiwan, 1999. Endocrinology & Diabetology
14(1-54), 19–23.
Walcott-McQuigg J.A. & Prohaska T.R. (2001) Factors influencing
participation of African American elders in exercise behavior.
Public Health Nursing 18(3), 194–203.
Wallace L.S. (2002) Osteoporosis prevention in college women:
application of the expanded health belief model. American Journal
of Health Behavior 26, 163.
 2008 The Authors. Journal compilation  2008 Blackwell Publishing Ltd
Factors influencing osteoporosis preventive behaviours
Wang H.H. (1999) Predictors of health promotion lifestyle among
three ethnic groups of elderly rural women in Taiwan. Public
Health Nursing 16(5), 321–328.
World Health Organization (2004) WHO Scientific Group of the
Assessment of Osteoporosis at Primary Health Care Level. Retri-
eved from http://www.who.int/chp/topics/Osteoporosis.pdf on 7
December 2007.
Yarcheski A. & Mahon N.E. (1989) A causal model of positive
health practices: the relationship between approach and replica-
tion. Nursing Research 38(2), 88–93.
Yeh H.H., Lin W.S., Chiang C.H. & Yang S.Y. (2002) A study on
knowledge, attitude, and behaviors of osteoporosis among
participants in a health education program. Show-Chwan Medical
Journal 3(2), 55–63.
345