International Journal of Nursing Studies 47 (2010) 1510–1524

Contents lists available at ScienceDirect

International Journal of Nursing Studies

journal homepage: www.elsevier.com/ijns

Evaluation of a conceptual model based on Mishel’s theories of

uncertainty in illness in a sample of Taiwanese parents of children with
cancer: A cross-sectional questionnaire survey
Lin Lin a,*, Chao-Hsing Yeh b, Merle H. Mishel c
a
The University of Texas Health Science Center at Houston, School of Nursing, Department of Integrative Nursing Care, 6901 Bertner Ave., Room 788, Houston, TX
77030, United States
b
School of Nursing, University of Pittsburgh, United States
c
School of Nursing, University of North Carolina at Chapel Hill, United States

A R T I C L E I N F O A B S T R A C T

Article history: Background: The prognoses of childhood cancers have improved over the last few decades.
Received 10 April 2009 Nevertheless, parental uncertainty about the absolute cure and possible relapse pervades
Received in revised form 11 May 2010 the entire illness trajectory. Despite illness-related uncertainty is significantly related to
Accepted 15 May 2010 psychological distress, continual uncertainty may serve as a catalyst for positive
psychological change and personal growth in the context of surviving cancer.
Keywords: Objectives: The purpose of this study was to examine a conceptual model that depicts
Coping coping and growth in Taiwanese parents living with the continual uncertainty about their
Growth through uncertainty
child’s cancer. The conceptual model was guided by Mishel’s theories of Uncertainty in
Parental uncertainty
Illness. The impact of the child’s health status, parents’ education level and perceived
Pediatric cancer
Taiwanese parents
social support on parental uncertainty was analyzed. The mediating effect of coping as
well as the influence of parental uncertainty and parents’ perceived social support on
growth through uncertainty was incorporated in the model testing.
Methods: This study involved a sample of 205 mothers and 96 fathers of 226 children
enrolled in a longitudinal cancer study in Taiwan. This study only analyzed the data
collected at baseline. A cross-sectional design was utilized to examine the relationships
among proposed variables. Parental uncertainty and growth through uncertainty were
measured by the translated questionnaires originally developed by Mishel. Parents’
perceived social support and coping were measured by culturally sensitive instruments
developed in Taiwan.
Results: The full research model and its alternative models fit adequately to the data via
structural equation modeling tests. Parental uncertainty and parents’ perceived social
support were associated with growth through uncertainty which was mediated by coping.
Child’s health status and parents’ perceived social support would significantly predict
parental uncertainty.
Conclusion: This study suggests that parental uncertainty has negative impact on coping
strategies such as interacting with family members while these coping strategies may help
Taiwanese parents gain growth through uncertainty. Coping strategies of searching for
spiritual meaning and increasing religious activities were not significantly influenced by
parental uncertainty in this study. The two coping strategies may be relevant to growth
through uncertainty due to Taiwanese cultural belief. Moreover, the availability of social
support promotes growth through uncertainty by its impact on lowering parental
uncertainty and encouraging more coping. The findings indicate that Taiwanese parents
may gain growth through uncertainty while experiencing their child’s cancer. The research

* Corresponding author. Tel.: +1 713 500 2061.

E-mail address: [email protected] (L. Lin).

0020-7489/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijnurstu.2010.05.009
 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524 1511

model provides possible guidelines for oncology nurses to deliver more culturally
competent health care for Taiwanese parents of children with cancer.
ß 2010 Elsevier Ltd. All rights reserved.

What is already known about the topic? ability. This can happen by parents accepting uncertainty as
the normal rhythm of life and by believing a world without
 With specific ongoing management or coping strategies, absolute certainty and predictability (Mishel, 1990, 1999).
continual uncertainty in chronic illness or in illness with According to Mishel’s (1990) Reconceptualization of Uncer-
a potential for recurrence, may change gradually from a tainty in Illness Theory (RUIT) that deals with specific
threat to a new perspective of life. ongoing management or coping strategies, continual
 Continual uncertainty is a possible catalyst for positive uncertainty in chronic illness or in illness with a potential
psychological change and personal growth in childhood for recurrence may change over time. Through the illness
cancer patients and their families. trajectory, individuals can gradually identify beneficial
 In Western societies, growth through uncertainty for aspects of continual uncertainty and find uncertainty does
parents of children with cancer is a continuous learning not always impact the quality of life. By living with
process throughout the child’s illness trajectory; none uncertainty about the future, people are motivated to make
such study has been done in Taiwanese culture. the most of every day and seize the moment. In Mishel’s
 Few studies have measured growth through uncertainty theory, the process is labeled as growth through uncertainty
quantitatively, and none of them focused on parents of (Mishel, 1990; Mishel and Clayton, 2008).
children with cancer. Continual uncertainty has been found as a possible
catalyst for positive psychological change and personal
What this paper adds growth in childhood cancer patients and their families
(Clarke-Steffen, 1993; Grootenhuis and Last, 1997; Parry,
 The relationship between parental uncertainty and 2003). Parents of children with cancer have described a
positive psychological outcomes in terms of growth number of positive psychological outcomes such as the
through uncertainty is mediated by coping strategies ability to incorporate the reality of an unpredictable course
seen in Taiwanese parents of children with cancer. of treatment and its result into a fresh view that
 The availability of social support may increase growth uncertainty can be a part of the life (Stewart and Mishel,
through uncertainty by lowering parental uncertainty 2000). The reformulation of a new sense of order and the
and encouraging more coping in Taiwanese parents of probabilistic thinking that emerges from the experience of
children with cancer. a childhood cancer corresponds with the process of growth
 Coping strategies such as searching for spiritual meaning through uncertainty addressed in RUIT (Mishel, 1990).
and increasing religious activities may function in the Although several qualitative studies have shown
process of growth through uncertainty in the Taiwanese growth through uncertainty in cancer patients and their
culture. families, only a few studies (Bailey et al., 2004; Mast, 1998;
Porter et al., 2006) have measured growth through
uncertainty quantitatively, and none were focused on
1. Introduction
parents of children with cancer. In the present study,
1.1. Parental uncertainty and growth through uncertainty in parents’ growth through uncertainty was measured
childhood cancer quantitatively using the instrument, Growth Through
Uncertainty Scale (GTUS), which was developed based
The diagnosis and treatment of childhood cancers are on Mishel’s Theories of Uncertainty in Illness (Mishel,
stressful for parents. Uncertainty, the failure to give illness- 1988, 1990; Mishe and Fleury, 1997).
related events a meaning (Mishel, 1981, 1988) has been Moreover, beginning with the publication of the
identified as one of the major constructs in the parents’ Parents’ Perception of Uncertainty Scale (PPUS) based on
psychological experience of their child’s cancer (Santa- Mishel’s theory to examine parent’s experience of uncer-
croce, 2003; Stewart and Mishel, 2000). Because of the tainty during a child’s hospitalization and treatment
improvement of cancer diagnosis and treatment, child- (Mishel, 1983, 1988), some studies on childhood cancers
hood cancers are now viewed as serious conditions that do tested the relationship between parental uncertainty and
not necessarily have a downward trajectory. Nevertheless, negative psychological outcomes such as anxiety and
the illness situations caused by childhood cancers can still depression (Mu et al., 2001, 2002; Stewart and Mishel,
be ambiguous, complex, and unpredictable with unavail- 2000). The present study focused on the relationship
able or inconsistent information (Mishel, 1983; Stewart between parental uncertainty and positive psychological
and Mishel, 2000). Continual uncertainty about the outcomes in terms of growth through uncertainty.
unknown and unknowable causes and outcomes of child-
hood cancers remains a difficult cognitive state for parents 1.2. Continual uncertainty and coping in Taiwanese parents
(Cohen, 1993, 1995). of children with cancer
However, continual uncertainty can be preferable
because it may facilitate patients and their families to The prognoses of childhood cancers have improved
perceive various opportunities and have greater adapt- significantly over the last few decades. About 80% of
1512 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524

children diagnosed with cancer are likely to survive continuous uncertainty may rely on dynamic coping
beyond 5 years in the United States (American Cancer processes to appreciate life’s lessons. Through the pro-
Society, 2009). In Taiwan, the treatment outcomes of cesses, parents may recognize illness as their child’s
childhood cancers are as good as in other developed destiny and learn to live with an uncertain illness
countries for most cancers (Lin, 2001). During a recent time trajectory as it is what life should be (Yeh, 2001b, 2003).
period (2001–2007), the 5-year survival rate was over 70% While developing the Parental Coping Strategy Inventory
for childhood cancers such as Hodgkin lymphoma, (PCSI) for Taiwanese parents of children with cancer by Yeh
retinoblastoma, and Wilms tumor ([Taiwan] Childhood (2001a), the researcher found that searching for alternative
Cancer Foundation, 2008). Although the survival rate spiritual meanings, such as considering that cancer is
constantly improves in Taiwan, parents never feel certain because of the child’s past sin or a tribulation from god, was
about the unpredictable side effects of treatment, illness one of the common coping strategies used by Taiwanese
complications, and possibility of recurrences (Yeh, 2003). parents. Under the beliefs of predetermined fate, Taiwanese
As the reappraisal of continual uncertainty is essential parents would also choose not to predict the illness
to gain psychological growth in Western culture, growth trajectory but maintain an optimistic state of mind by
through uncertainty for parents of children with cancer is believing that there is a way out of everything (Yeh, 2001a,
considered as a learning process throughout the child’s 2003; Yeh et al., 2000). Earlier qualitative studies found
illness trajectory. When parents are able to perceive that positive psychological outcomes in Taiwanese parents of
living with continual uncertainty about the child’s condi- children with cancer (e.g., Chao et al., 2003; Yeh, 2003). The
tion is unavoidable, they may progressively gain psycho- present study further examined the relationship between
logical growth by reappraising continual uncertainty as a Taiwanese parents’ coping strategies listed in PCSI (Yeh,
natural rhythm of life (Mishel and Clayton, 2008; Stewart 2001a) and the positive psychological outcomes in terms of
and Mishel, 2000). The process of growth through growth through uncertainty as measured by the GTUS.
uncertainty then makes the enduring uncertainty become
preferable (Mishel, 1990; Mishel and Clayton, 2008). 2. Conceptual framework
However, the belief of impermanence and the capri-
cious life is deeply embedded in Eastern culture. When an The conceptual framework (see Fig. 1) predicting
unexpected event happens, Taiwanese tend to consider the parents’ positive psychological outcomes was adapted
shifting life as a predestined fate and resign themselves to from Mishel’s (1988) Uncertainty in Illness Theory,
what is inevitable (Hwang, 1977; Shek and Cheung, 1990; combining the concept of growth through uncertainty in
Yeh, 2003). For Taiwanese, disturbances determined by Mishel’s (1990) Reconceptualization of Uncertainty in
fate are lessons assigned to life’s journey; life is normally Illness Theory. Based on Mishel’s (1988, 1990) theories, the
full of uncertainty because no one knows what his or her unknowns about illness-related situations, such as a child’s
life lessons are until different kinds of unexpected events health status (stimuli frame) and the parents’ perceived
have happened. Taiwanese people attempt to cope with social support and education level (structure providers or
the challenges from each life lesson, including clinging to available resources) can influence parents’ perception of
the philosophy of doing nothing but going where life leads uncertainty. On the other hand, social support and coping
(Hwang, 1977). strategies used by Taiwanese parents may facilitate
To justify uncertain causes and outcomes of childhood accepting their child’s cancer as one of life’s lessons.
cancers by searching for culturally relevant explanations Accordingly the parents may gain psychological growth
and solutions is very commonly seen in Taiwanese and other rather than suffer from psychological distress under
Chinese societies (Martinson et al., 1999; Wong and Chan, continual uncertainty in childhood cancers.
2006; Yeh et al., 2000). Yeh (2003) found that ‘‘coming to In this study, the influences of a child’s health status, the
terms’’ is a core category in Taiwanese parents’ response to parents’ education, and perceived social support on parental
childhood cancers. When taking care of children with uncertainty were examined. The mediating effect of
cancer, Taiwanese parents may manage illness-related Taiwanese coping strategies between parental uncertainty
uncertainty by using coping strategies such as learning and growth through uncertainty was tested. The direct and
from health care professionals or increasing their interac- indirect effects of parents’ perceived social support on
tion with family members (Yeh, 2001a, 2003). However, growth through uncertainty were also analyzed.
parents may also accept their child’s cancer by resigning
themselves to fate or relying on supernatural power through
3. Methods
religious activities (Yeh, 2001b, 2003; Yeh et al., 2000).
Uncertain causes and outcomes of childhood cancers 3.1. Participants
can make Taiwanese parents profoundly aware of the
capricious character of life. The confirmation of uncertain This study reports on baseline data of 301 Taiwanese
life and the acceptance of living with continual uncertainty parents of 226 children enrolled in a larger longitudinal
match the concept of growth through uncertainty in study aimed at testing a model of coping process and
Mishel’s (1990) Reconceptualization of Uncertainty in adjustment for children with cancer and their parents.
Illness Theory. However, uncertainty as a natural rhythm Among the 226 children who were diagnosed with cancer,
of life is rooted in the Taiwanese culture. Instead of a baseline data from both parents were available for 75
learning process towards recognition of uncertainty as part children. On the other hand, data were available from
of life, the mechanism of maintaining a positive outlook on mothers only for 130 children and from fathers only for 21
[(Fig._1)TD$IG] L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524 1513

Fig. 1. Theoretical model adapted from the Uncertainty in Illness Theory and the Reconceptualization of Uncertainty in Illness Theory (Mishel, 1988, 1990).

children only. In sum, the data set consists of data from 205 Children and their parents were able to enter the study
mothers and 96 fathers (see Tables 1 and 2). during different periods of cancer treatment (i.e., newly
diagnosed with cancer, during treatment for remission or
3.2. Procedure relapse, or completion of cancer treatment). The entry
criteria included the following: the child was (a) diagnosed
Participants were parents recruited from a major as having cancer and (b) younger than 18 years old; and the
children’s hospital in northern Taiwan. Approval for the parents (a) were willing to permit follow-up phone calls for
study was obtained from the Institutional Review Board of data collection following their child’s discharge from the
the hospital prior to the study. The regulations for hospital and (b) agreed to be contacted for follow-up
protecting confidentiality in the hospital were followed. within a 12-month duration. Subjects were excluded if
parents (a) were unable or unwilling to sign informed
Table 1 consent, (b) planned to move outside of the study area
Demographic characteristics of the parents. before study ends, (c) were emotionally unstable or too
Fathers (N = 96) Mothers (N = 205)
upset to participate, (d) did not have telephone service at
home, or (e) were involved in any other study. Moreover,
Age
Mean (SD) (range) 40.8 (6.0) (23–53) 37.4 (6.1) (24–54)

Years of education Table 2

Mean (SD) (range) 13.0 (2.6) (6–18) 12.5 (2.7) (0–18) Demographic and medical characteristics of the children with cancer.

n % n % Children (N = 226)

Levels of education n %
Primary school or less 1 1.0 3 1.5 Age (years)
Junior high school 12 12.5 29 14.1 <7 91 40.3
High school 39 40.6 108 52.7 7–12 84 37.2
College or above 42 43.8 60 29.3 13–17 51 22.6
Missing 2 2.1 5 2.4
Gender
Religious belief Boys 122 54.0
Folk beliefs 2 2.1 2 1.0 Girls 104 46.0
Buddhist 30 31.3 84 41.0
Taoist 16 16.7 32 15.6 Diagnosis
None 39 40.6 66 32.2 Acute lymphoblastic leukemia (ALL) 106 46.9
Others 7 7.3 15 7.3 Acute non-lymphoblastic leukemia 22 9.7
Missing 2 2.1 6 2.9 Lymphoma 22 9.7
Brain tumor 9 4.0
Socioeconomic status Other cancer 67 29.7
1 (Highest) 2 2.1 4 2.0
2 (Moderately high) 16 16.7 12 5.9 Categorical cancer group
3 (Medium) 37 38.5 42 20.5 Newly diagnosed less than 2 months 66 29.2
4 (Moderately low) 27 28.1 63 30.7 Remission-on treatment 92 40.7
5 (Lowest) 12 12.5 78 38.0 Relapse-on treatment 13 5.8
Missing 2 2.1 6 2.9 Off-treatment less than 2 years 23 10.2
Event free survivors 32 14.2
SD = standard deviation.
1514 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524

families with more than one child having cancer or with a Cronbach’s alpha = .90. Although the PPUS was designed for
child having cancer in a terminal stage were excluded. parents of hospitalized children, data from parents of cancer
To obtain the baseline data used for this study, trained survivors were kept because all participants were recruited
data collectors recruited convenience samples in the in a hospital setting rather than in communities.
participating hospital. The data collectors went to the Perceived social support in this study was measured by
hospital regularly and approached potential eligible subject three subscales of the Parental Coping Strategy Inventory
in the oncology units or outpatient clinics. They verbally (PCSI). The PCSI was developed specifically for Taiwanese
explained the purpose and procedures of the study and after parents of children with cancer (Yeh, 2001a). The items of
written informed consent was obtained, a questionnaire the PCSI were developed based on qualitative interviews
package was distributed for collecting baseline data. All data with parents of children with cancer in Taiwan (Yeh, 2001a;
were collected using self-report questionnaires. Yeh et al., 2000). The construct validity of PCSI was
examined using confirmatory factor analysis (Yeh, 2001a).
3.3. Measures The generalizability of the factor structure was supported by
studies on mothers of children with epilepsy (Yeh, 2001a).
3.3.1. Demographic and background measures The three subscales of the PCSI used as a measure of
perceived social support in this study were the emotion
3.3.1.1. Child’s illness parameters and demographic varia- support subscale, the information support subscale, and
bles. Parents were asked to answer the personal-health the actual support subscale (see Table 3). In these three
questionnaire to state their child’s age and gender, primary subscales, the items ask parents how frequently they
caregiver, and number of children in their household. perceived themselves getting support from spouse,
Illness parameters were obtained from medical charts that families, relatives, friends, health care professionals, and
were reviewed by trained research assistants. The col- other families of children with cancer. Each item is
lected medical information included each child’s type of administrated by using 4-point frequency scales. The total
cancer, treatment status, date of diagnosis and date of score of each subscale is calculated by summing the items,
treatment protocol onset, type of received treatment, with higher scores indicating more available social
history of recurrence, and major laboratory data. support. The three subscales have demonstrated accep-
table psychometric properties (Yeh, 2001a). In the current
3.3.1.2. Parental characteristics. A personal-information study, the internal consistency of the emotion support
questionnaire was completed by parents for collecting data subscale was .87. For the information support subscale, the
on their personal characteristics. Parents’ age, education Cronbach’s alpha was .87 and for the actual support
level, religious belief, marital status, employment status, subscale, the Cronbach’s alpha was .80.
and family income were used for depicting the sample. Coping strategies in this study also were measured by
the Parental Coping Strategy Inventory (PCSI). The original
3.3.2. Variables of model testing
Parental uncertainty in this study was measured by the Table 3
Chinese version of Parental Perception of Uncertainty Scale Descriptive statistics of the measures.
(PPUS) translated by Mu and colleagues (Mu et al., 2001, Measure Mean (n = 301) SD Potential
2002). The original PPUS was developed by Mishel (1983) in range
English to measure parents’ perception of uncertainty that
Parental uncertaintya 80.05 14.94 31–155
influences parents’ response to their child’s illness and Functional statusa 32.21b (n = 300) 4.36 14–42
hospitalization. This 31-item scale employs a 5-point, Likert
Perceived social supportc
scale. After reverse scoring appropriate items, a total score is
Emotion support 12.11 3.13 4–16
calculated by summing up all the items, with higher scores Information support 10.61 3.26 4–16
indicating greater perceived uncertainty. Alpha coefficients Actual support 9.96 3.30 4–16
for the original scale were .86–.93 (Mishel, 1983). Copingc
The PPUS was translated into Chinese using the double Learning 17.28b (n = 300) 2.21 4–20
translation method, and the accuracy of the translation and Maintaining stability 7.20 2.96 4–20
the relevance of the instrument were evaluated separately Maintaining an optimistic 17.09 2.55 4–20
state of mind
by five experts of family studies in Taiwan (Mu et al., 2001,
Searching for spiritual 11.48 3.28 4–20
2002). Construct validity of the Chinese PPUS was tested by meaning
factor analysis, and the result showed a four-factor structure Increasing religious 12.73 3.23 4–20
(i.e., ambiguity, complexity, lack of information, and activities
b
unpredictability) that was consistent with the theoretical Interaction with patient 15.58 (n = 300) 2.90 0–20
Interaction with spouse 15.57b (n = 300) 2.89 0–20
assumption of Mishel’s (1988) Uncertainty in Illness Theory. Interaction with healthy 15.19b (n = 300) 3.97 0–20
Moreover, construct validity was supported by showing a sibling
significant positive correlation with anxiety measured by
Growth through uncertainty 156.63 19.57 39–234
State-Trait Anxiety Inventory (Mu et al., 2001, 2002).
Cronbach’s alphas for the PPUS in Chinese were .79 and SD = Standard deviation.
a
There are significant overall differences on the measures among
.87 in Mu et al. (2001, 2002) studies of the impact of
parents of children in different stages of cancer treatment (p < .01).
uncertainty on parents of children with cancer in Taiwan. For b
There were missing values on the variables.
the current study, the internal consistency was strong with c
The variables were operated as latent variables.
 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524 1515

PCSI consists of 48 items and can be divided into 12 Child’s health status in this study was measured by the
subscales. However, 3 of the 12 subscales were considered revised Functional Status II (FSII-R) (Stein and Jessop, 1990).
as a measure of parents’ perceived social support in this The FSII-R was developed purposely to assess the health
study. One subscale, struggling, was left out in this study status of children with chronic health conditions. The long
because it measures a similar construct as parental version FSII-R contains 43 items, and the short version FSII-R
uncertainty by asking parents about their struggle with contains 14 items. The short version of FSII-R used in this
not knowing how to make medical decisions and not being study consists of items that are suitable to measure child’s
sure if the chosen treatment for their child is right. The health status across the entire age span of childhood (Stein
remaining eight subscales used to measure parents’ coping and Jessop, 1990). The FSII-R has demonstrated excellent
are listed in Table 3. For the scoring of the PCSI, a total score psychometric properties (Stein and Jessop, 1990).
for each subscale is calculated by summing up scores on all The Chinese version of the FSII-R for this study was
the items, with higher scores indicating that parents use offered by the collaborative research team in Taiwan. Each
the coping strategies in the subscale more frequently. For item of the Chinese version of the FSII-R is administered by
parents who have more than one child, the score of using 3-point scales with higher scores indicating better
interaction with families equals (interaction with functional status. For the current study, the internal
patient + interaction with spouse + interaction with consistency was acceptable with Cronbach’s alpha = .76.
healthy sibling)/3; for parents who have only one child,
the score of interaction with families equals (interaction 3.4. Data analyses
with patient + interaction with spouse)/2. Cronbach’s
alphas for these eight subscales ranged between .69 and The current cross-sectional study analyzed the data
.88 in Yeh’s (2001a) study of parents of children with collected at baseline of a longitudinal study. To test the
cancer. In the current study, Cronbach’s alphas for the theoretical relationships among variables proposed in the
subscales ranged between .66 and .91. research model (see Fig. 1), structural equation modeling
Growth through uncertainty in this study was measured (SEM) was conducted to indicate the strength of influence
using the translation of 39-item Growth Through Uncer- among variables by getting an overall fit of model with the
tainty Scale (GTUS) (Mishe and Fleury, 1997). The GTUS data. Goodness-of-fit tests were performed to determine
measures psychological growth as a result of experiencing whether the research model (full model) and its alternative
serious illness through which individuals relinquish their reduced models should be accepted or rejected. In this
old life perspective and construct a new view of life study, models were analyzed using Mplus software version
(Mishel, 1990, 1999). The GTUS is the first instrument 5.0 (Muthen & Muthen, CA: Los Angeles).
designed to measure positive psychological changes and According to the research model, the significance of direct
personal growth through illness-related uncertainty. This and indirect relationships among parental uncertainty,
39-item scale employs a 6-point, Likert scale. After reverse parents’ perceived social support, coping, and growth
scoring appropriate items, a total score is calculated by through uncertainty were tested. Furthermore, the influ-
summing up scores on all the items, with higher scores ences of a child’s health status, the parents’ education and
indicating more psychological growth through uncertainty perceived social support on parental uncertainty were
and changes in life view. Alpha coefficients for the total estimated by putting these variables into the model at the
scale were .94 in a study of men with prostate cancer same time. The mothers’ model and fathers’ model were
(Bailey et al., 2004) and .95 in one study on breast cancer developed and evaluated concurrently; the correlations
survivors (Porter et al., 2006) and .94 in another study on between parents from a same family were incorporated by
breast cancer survivors (Mast, 1998). Construct validity drawing two-way arrows connecting mothers’ model with
was tested by confirmatory factor analysis and the result fathers’ model. Then the two parallel models (mothers versus
showed that 26 of the 39 items consist of four distinct fathers) with the same form were simultaneously fit to the
factors (Mishe and Fleury, 1997). Furthermore, construct data. Model parameters were estimated using maximum
validity was also supported by the negative correlation likelihood methods. Data missing at the variable level were
with the Profile of Mood States Scale (Mast, 1998). ignored (i.e., treated as missing at random [MAR]). All
In this study, the GTUS was translated into Chinese, and observed data were included when fitting the models.
the comparability of content was verified through back- For the purpose of finding a more parsimonious model,
translation procedures; however, the wording was mod- the following model-reduction strategies were used. First,
ified to measure parents’ perception of their child’s cancer the full model (model 1) with all parameters completely
rather than their own disease (see Appendix A). The unrestricted was simultaneously fit for mothers’ and fathers’
readability of the scale was evaluated by three Taiwanese data. All paths in the full model with the estimate to standard
parents locally. Coefficient a for the total scale in Chinese error ratio less than 1.96 in absolute value (i.e., p > .05) were
was .94 according to 248 parents’ baseline data collected in noted; and then in the new ‘‘reduced’’ model (model 2), the
the first year of data collection. Construct validity was path coefficients of these paths were restricted to equal zero,
supported by a significant positive correlation (r = .32, which means these paths were essentially deleted.
p < .01) with the Chinese version of Posttraumatic Growth Then the fit of the reduced model was assessed using
Inventory that had been used to examine personal growth several methods including the Chi-squared difference
among Chinese cancer survivors (Ho et al., 2004; Tedeschi test for comparing the reduced models to the full model
and Calhoun, 1996). For the current study, the internal and the root mean square error of approximation
consistency was strong with Cronbach’s alpha = .95. (RMSEA), along with a 90% confidence interval, the
1516 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524

Comparative Fit Index (CFI), the Tucker-Lewis fit index 4.2. Descriptive statistics of the measures
(TLI), the Akaike Information Criteria (AIC), and the
Bayesian Information Criteria (BIC) for determining the The mean, standard deviation, and potential score range
extent to which the relationships existing in the data are of all the measures are presented in Table 3. Variables with
consistent with those proposed by the models (Chen missing values are found in the data of one participant (see
et al., 2008; Hu and Bentler, 1998, 1999; MacCallum et al., Table 3). The group differences on the measures among
1996). To interpret the Chi-squared difference test, p- parents of children in different stages of cancer treatment
value greater than .05 indicates that the reduced model is (i.e., newly diagnosed with cancer less than 2 months,
not a significantly weaker fit than the full model. For the remission on treatment, relapsed on treatment, off-treat-
RMSEA, the general rule of thumb is that values less than ment less than 2 years, and survivors) are presented in the
.05 indicate close fit, values between .05 and .10 indicate same table. Parental uncertainty and child’s functional
marginal fit, and values greater than .10 indicate poor fit status were significantly different with the p-value of the
(MacCallum et al., 1996). For both CFI and TLI, a value of 1 one-way analysis of variance test (ANOVA) less than .01. All
indicates perfect fit and the general rule of thumb is that other variables did not show a significant difference.
values larger than .9 indicate adequate fit. For using AIC Because SEM can correct for measurement error thus
and BIC as the index for model selection, smaller AIC and providing estimates of relationships among latent variables
BIC indicate better model fit. that are not biased due to unreliable indicators, the value of
After the reduced model that provided adequate fit to measurement error based on reliability and observed
the data was obtained from the methods mentioned above, variance was estimated for each variable indicating a latent
the next step to release more degrees of freedom was to concept (Deshon, 1998). In this study, the measurement
assess whether the path coefficients in the measurement error variances of the three subscales indicating perceived
model (paths from one latent variable to its corresponding social support and the six subscales indicating coping were
subscales) could be restricted to be equivalent between calculated (see Fig. 2). For both mothers and fathers, the
mothers’ model and fathers’ model. This process started percentages assessed as measurement error variances for
with the reduced model (model 2) in which no path emotion support, information support and actual support
coefficients were restricted to be equivalent between subscales were 2–3%. For mothers, the error variances of
mothers’ model and fathers’ model. In this step, the path maintaining stability and searching for spiritual meaning
coefficients of each subscale of parents’ perceived social were about 30% and increasing religious activities was 20%.
support and coping were restricted to be the same between The rest of the three subscales indicating coping were 5–10%.
mothers’ model and fathers’ model. The purpose of this For fathers, the error variances of maintain stability and
step was to examine whether there is a significant searching for spiritual meaning were about 15% and
difference between mothers’ measurement model and increasing religious activities was 10%. The rest of the three
fathers’ measurement model. Then the third model with subscales indicating coping were less than 5%.
restricted path coefficients was compared with the full
model and the fit of the data was assessed using the same 4.3. Model testing and model selection
model selection indices mentioned above.
After model 3 provided adequate fit to the data, the Four models were tested for the fit of the data. For the
final step to restrict the model was to set the path first model, the full model, all parameters completely
coefficients in the theoretical part (the structural model) unrestricted was simultaneously fit for both mothers and
to be equivalent between the mothers’ model and the fathers’ data. In the theoretical part of the model, the path
fathers’ model. The purpose of this step was to examine from education to parental uncertainty was not significant
whether there is a significant difference between the for both parents (p > .05). For the path from perceived
whole mothers’ model and fathers’ model. The fourth social support to parental uncertainty, it was significant for
model was then compared with the full model and the fit mothers (p < .05) but not for fathers (p > .05). In the
of the data was assessed using the same model selection measurement part of the model, the paths from coping to
indices mentioned above. searching for spiritual meaning and to increasing religious
activities were not significant for both parents (p > .05).
However, the model did not show a better fit when the
4. Results
non-significant paths mentioned above were deleted.
4.1. Sample characteristics Based on the theoretical concern, these non-significant
paths were kept in the reduced models (models 2–4).
Table 1 reports the demographic characteristics of the Moreover, the correlation between searching for spiritual
205 mothers and 96 fathers who participated in this study. meaning and increasing religious activities was added into
Table 2 shows the demographic and medical information the model testing according to the suggestion from the
of the 226 children whose parents participated in this modification indices; however, the fit indices did not
study. For the cancer diagnosis, 46.9% of the children were improve much by this procedure.
diagnosed with acute lymphoblastic leukemia, 9.7% were In the first ‘‘reduced’’ model (model 2), the coefficients of
diagnosed with acute non-lymphoblastic leukemia, 9.7% two paths were restricted to equal zero due to the small ratio
were diagnosed with lymphoma, 4% were diagnosed with of estimate to standard error (Est./S.E.) in both mothers’
brain tumor, and 29.7% were diagnosed with other cancers model and fathers’ model. The two paths were from parental
such as neuroblastoma, Wilms tumor, or osteosarcoma. uncertainty to growth through uncertainty and from
[(Fig._2)TD$IG] L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524 1517

Fig. 2. The standard path coefficients of the paths in the final reduced model.

Table 4 social support to parental uncertainty, parental uncertainty

Model comparison for alternative reduced models to the full model.
and parents’ perceived social support to coping, and coping
Model df x2 p-Value to growth through uncertainty.
Model 1: Full model 0 – – The comparisons with the four models described above
Model 2: First reduced model with two 4 3.751 0.44 are presented in Tables 4 and 5. Models 2–4 were compared
path coefficients in the theoretical to the original unrestricted model (model 1) using the Chi-
model equal to zero squared difference test, where the more restrictive model
Model 3: Second reduced model with 11 8.604 0.66
was deemed adequate if the p-value was greater than .05.
path coefficients in the measurement
model equal between mothers and The results showed that all three alternative models (model
fathers 2 to model 4) were not significantly different from the full
Model 4: Final reduced model with path 17 17.042 0.45 model thus succeeding in achieving parsimony (see Table 4).
coefficients in the theoretical model
On the other hand, the fit indices of the four tested models all
equal between mothers and fathers
provided some evidence of marginal fit with RMSEA about
df = Degrees of freedom, x2 = Chi-square test value.
.05, CFI close to .9, and TLI close to .9 (see Table 5). Therefore,
all the four models were considered not rejected by the data.
perceived social support to growth through uncertainty. In Because all three alternative reduced models (models 2–4)
the next step, the second reduced model (model 3) was fit appeared to fit the data adequately and each reduced model
with equivalence restrictions on the path coefficients for the did not show a significantly weaker fit than the full model
paths in the measurement model (i.e., coping to its six (model 1) by chi-squared difference test (p > .05), model 4,
subscales and perceived social support to its three the most parsimonious model, was selected as the final
subscales). Finally, in model 4, in addition to the paths in model due to its smallest values of AIC and BIC (see Table 5).
the measurement model, all the path coefficients for the
paths in the theoretical part of the model (the structural 4.4. Significant direct and indirect effects
model) were set to be equivalent between mothers’ model
and fathers’ model. The paths included child’s health status All the significant results discussed below are based on
to parental uncertainty, parents’ education and perceived the final model (model 4). In the model 4, all unstandar-

Table 5
Model fit statistics for comparing nested alternative models to the full model.

Model CFI TLI AIC BIC RMSEA

Estimate 90% CI

(1) Full model 0.873 0.845 21244.706 21590.180 0.051 0.041,0.060

(2) Reduced model with two paths in the theoretical part equal to zero 0.873 0.848 21240.458 21572.250 0.050 0.041,0.060
(3) Reduced model with selected paths in the measurement part equal 0.875 0.854 21231.311 21539.159 0.049 0.040, 0.058
(4) Final reduced model with all paths in the theoretical part equal 0.873 0.855 21227.749 21515.074 0.049 0.040, 0.058

RMSEA = Root Mean Square Error of Approximation, CFI = Comparative Fit Index, TLI = Tucker Lewis Fit Index, AIC = Akaike Information Criteria,
BIC = Bayesian Information Criteria, 90% CI = 90 percent confidence interval.
1518 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524

dized path coefficients in mothers’ model were set to be The significant relationship between child’s health
equal to the path coefficients in fathers’ model which status and parental uncertainty is consistent with Mishel’s
indicates mothers’ model did not significantly different (1988) theory. The inability to determine the meaning of
from fathers’ model. Standardized estimates of all paths of illness-related circumstances may cause a cognitive state
the final reduced model are shown in Fig. 2. of uncertainty (Mishel, 1988). By perceiving that uncer-
Child’s functional status had a significantly negative tainty may arise from illness-related situations, the level of
effect on parental uncertainty; lower child’s functional parental uncertainty is more likely to change contextually
status was associated with higher parental uncertainty following alteration of an ill child’s medical condition
(p < .01). Parents’ perceived social support had a signifi- rather than to diminish linearly over time. The significant
cantly negative effect on parental uncertainty; lower relationship between child’s functional status and parental
perceived social support was associated with higher uncertainty validates the importance of considering the
parental uncertainty (p < .05). However, parents’ level of child’s illness characteristics when studying parental
education did not significantly associate with parental uncertainty in childhood cancers.
uncertainty (p > .05). According to Mishel (1988, 1997), education as a
Parental uncertainty had no significant direct effect on structure provider can help patients and their families
growth through uncertainty; the significant effect of know where and how to get health information, thus
parental uncertainty on growth through uncertainty was reducing illness-related uncertainty. Although Mishel
fully mediated by coping. Overall, lower parental uncer- (1983) also indicated that education might have influence
tainty was associated with more coping (p < .01), and in on parental uncertainty concerning the hospitalized child,
turn, more coping was associated with more growth there is conflicting evidence about the relationship
through uncertainty (p < .01). between education and uncertainty in Taiwanese parents
Parents’ perceived social support had no significant of children with cancer (Mu et al., 2001, 2002). In the
direct effect on growth through uncertainty; the signifi- current study, the direct effect of parents’ education on
cant effect of parents’ perceived social support on growth parental uncertainty was neither supported by the full
through uncertainty was fully mediated by coping. Overall, model (model 1) nor its reduced models (models 2–4). The
more perceived social support was associated with more non-significance may result from small variance in the
coping (p < .01), and in turn, more coping was associated education level because about 80% of the parents in this
with more psychological growth (p < .01). study had a high school degree or above. Although
education is an important resource assisting patients
5. Discussion and suggestions for future research and families to explain and assign meaning to illness-
related events (Mishel, 1998), more research is needed to
The results of the present study clearly demonstrated clarify how education functions to decrease parental
distinct pathways of growth through uncertainty for uncertainty in childhood cancers in Taiwan.
Taiwanese parents of children with cancer. According to The direct effect of parents’ perceived social support on
Mishel’s theories (1988, 1990), uncertainty in illness is parental uncertainty was supported by the final reduced
more commonly appraised as a danger in Western culture model in which more perceived social support was
by ill persons, caregivers, and parents of ill children. associated with lower parental uncertainty. The research
However, the learning process of re-appraising continuous finding is consistent with Mishel’s (1988) theory; social
uncertainty from an aversive experience to a preferable support can directly influence parental uncertainty by
sense of order, the growth through uncertainty, is observed providing information and thus reduce the ambiguity,
in patients with cancer and in their families (Mishel and complexity, or unpredictability of illness-related events
Clayton, 2008). The life perspective presented in Mishel’s (Mishel, 1983, 1988). Based on the research findings, to
(1997) growth through uncertainty scale matches the interact with someone who can provide health informa-
belief in Eastern culture which uncertainty is well- tion may considerably help Taiwanese parents to decrease
accepted as the natural rhythm of life. As predicted by their uncertainty about childhood cancers. Affective or
our hypotheses, the data evidently indicate that parents in tangible aid perceived as more supportive are also related
Taiwan maintain a positive outlook on uncertainty using to lower levels of parental uncertainty.
culturally relevant coping strategies throughout their To understand the full function of perceived social
child’s illness trajectory. The present study not only shows support in Taiwanese parents of children with cancer, this
that positive psychological outcomes existing in Taiwa- study tested direct relationships between social support
nese parents of children with cancer but also points out the and uncertainty, social support and coping as well as social
vital role of culture differences when the mechanism of support and growth through uncertainty. The results
achieving positive psychological outcomes is assessed. showed that perceived social support did not have a
Based on the theoretical model proposed in the present significant relationship with growth through uncertainty.
study, parental uncertainty is influenced directly by the Instead, perceived social support not only directly related
child’s health status and parents’ education, and perceived to coping but also indirectly influenced coping by its
social support. The research findings indicated that the association with parental uncertainty. Based on the
better a child’s health status and the more parents’ perceived research findings, even though perceived social support
social support were significantly associated with lower does not facilitate growth through uncertainty directly, the
parental uncertainty. However, parents’ education did not availability of coping resources, such as emotion, informa-
show a significant relationship with parental uncertainty. tion, and actual support may facilitate a higher rate of
 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524 1519

using coping strategies, thus promoting parents to gain effects from cancer treatment and the quality of life living
growth through uncertainty. with cancer. However, many studies have found that
The research findings showed that parental uncertainty parental uncertainty never entirely goes away throughout
had no significantly direct effect on growth through the illness trajectory of childhood cancers (Cohen, 1995;
uncertainty, except when mediated by coping. According Cohen and Martison, 1988; Koocher and O’Malley, 1981). It
to Mishel’s (1990) Reconceptualization of Uncertainty in is also not clear if parental uncertainty lessens linearly over
Illness Theory, uncertainty and growth through uncer- time (Stewart and Mishel, 2000).
tainty do not happen concurrently. Psychological growth is Parents of child in different cancer treatment stages
a dynamic process throughout the illness trajectory. While were all included in this study. However, the best way to
examining the cross-sectional relationship between par- describe parents’ psychological changes through an illness
ental uncertainty and growth through uncertainty, the trajectory is to observe them longitudinally and take their
non-significant direct relationship is supported by Mishel’s child’s prognosis into consideration. Longitudinal data
(1990) theory. analysis is the exact approach to assess within-group
By treating coping as one latent variable, the findings changes over time for both parental uncertainty and
indicated that coping fully mediated the relationship growth through uncertainty.
between parental uncertainty and growth through uncer- In this study, child’s age was not put in the model
tainty. Even though parental uncertainty had a signifi- testing. Changing developmental needs at different ages
cantly negative impact on coping, the coping strategies had causes parents to raise specific illness-related concerns.
a positive influence on growth through uncertainty. When Child’s age ranged from less than 1 year old to 18.9 years in
parents reported a higher level of uncertainty in the study, this study. Age was not tested because of its wide range
they felt overwhelmed by the child’s illness-related and the issue that child’s developmental needs do not
situations which they did not understand or could not change linearly with one’s age. This study did not intend to
control. Although the parental uncertainty prevented these generate data to specifically address developmental
parents from coping, parents living with uncertainty could concerns. More developmental variables and parents’
remain positive through coping. The result specified the personal characteristics such as personality traits should
importance of coping in maintaining a positive outlook on be considered in the future studies.
uncertainty for Taiwanese parents.
The study showed that parental uncertainty did not 6. Conclusion
strongly impact the coping strategies of searching for
spiritual meaning and increasing religious activities in the Parental uncertainty is a significant psychological
Taiwanese culture. The two coping strategies were more experience for Taiwanese parents of children with cancer.
likely to be triggered by an awareness of the capricious life In Taiwanese culture, not only the illness-related uncer-
rather than the parental uncertainty about specific medical tainty but also the awareness of capricious life triggered by
concerns measured in the PPUS. Even though these two the child’s cancer may have influence on parents’ coping
coping strategies were not well-explained by parental strategies. Even though many studies have shown sig-
uncertainty about hospitalization experiences of children nificant impacts of uncertainty on negative psychological
with cancer, the two culturally relevant coping strategies outcomes, such as anxiety and depression, this study
may play important roles that promote positive psycho- showed the possibility for Taiwanese parents to remain
logical outcomes. Further research on the cultural dimen- positive while they are trying to cope with the experience
sions of parental uncertainty in childhood cancers may of their children having cancer.
help to better predict Taiwanese parents’ coping strategies This study points out the importance of studying
responding to the illness-related uncertainty. Also, more psychological adaptation in a cultural perspective. Positive
studies are needed to test models that examine the psychological outcomes such as growth through uncer-
function of different coping strategies. tainty may be noted in both Eastern and Western cultures.
Because uncertainty is widely accepted as a natural However, the mechanism for gaining psychological growth
rhythm of life in Eastern culture, growth through through continual uncertainty can be very different. In this
uncertainty may not be a linear learning process study, Taiwanese parents’ coping strategies and the
throughout the illness trajectory for Taiwanese parents. availability of social support played significant roles in
In the current study, growth through uncertainty did not the process of maintaining a positive outlook on uncer-
show a significant difference among parents of children tainty. The findings provide possible guidelines for
in different stages of cancer treatment (p = .98). It is Taiwanese nurses in delivering culturally competent
logical that Taiwanese parents of newly diagnosed health care and thus help parents to manage the transition
children can have a similar level of growth through to live with continual uncertainty in childhood cancers.
uncertainty as parents of children under treatment or Conflict of interest: None declared.
off-treatment do. Role of funding: This research was supported by a grant
On the other hand, parental uncertainty in this study to Dr. Yeh, Chao-Hsing from the National Health Research
was significantly different among parents of children in Institutes, Taiwan (Grant Number: NHRI-EX9397-
different cancer treatment stages. When an ill child has 9302PI).
improved overall health, parents may feel less uncertain Ethical approval: Approval for the study was obtained
about the effectiveness of the treatment and prognosis of from the Institutional Review Board of Chang Gung
the illness. Parents may also worry less about the side Children’s Hospital in the year of 2005.
1520 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524

Acknowledgments research subjects for their participation and the research

team of Dr. Yeh for helping the data collection and
This research was supported by a grant to Dr. Yeh, management. The authors also thank the statistical
Chao-Hsing from the National Health Research Insti- consultation offered by Dr. Mark Weaver at Family
tutes, Taiwan. The authors would like to thank the Health International.

Appendix A. Growth Through Uncertainty Scale (GTUS)

L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524 1521
1522 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524
 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524 1523

References Hu, L., Bentler, P.M., 1998. Fit indices in covariance structure modeling:
sensitive to underparameterization model misspecification. Psycho-
logical Methods 3, 424–453.
American Cancer Society, 2008. Cancer facts and figures. Retrieved March Hu, L., Bentler, P.M., 1999. Cutoff criteria for fit indices in covariance
1, 2009, from http://www.cancer.org/docroot/STT/stt_0.asp. structure analysis: conventional criteria versus new alternatives.
Bailey, D.E., Mishel, M.H., Belyea, M., Stewart, J.L., Mohler, J., 2004. Structural Equation Modeling 6, 1–55.
Uncertainty intervention for watchful waiting in prostate cancer. Hwang, K.K., 1977. The patterns of coping strategies in a Chinese society
Cancer Nursing 27, 339–346. [in Chinese]. Acta Psychologica Taiwanica 19, 61–73.
Chao, C., Chen, S., Wang, C., Wu, Y., Yeh, C., 2003. Psychosocial adjustment Koocher, G.P., O’Malley, J.E., 1981. The Damocles Syndrome: Psychosocial
among pediatric cancer patients and their parents. Psychiatry and Consequences of Surviving Childhood Cancer. McGraw-Hill, New
Clinical Neurosciences 57, 75–81. York.
Chen, F., Curran, P.J., Bollen, K.A., Kirby, J., Paxton, P., 2008. An empirical Lin, K., 2001. Overview of childhood cancer care in Taiwan. In: Proceed-
evaluation of the use of fixed cutoff points in RMSEA test statistic in ings of Asian Pacific workshop for care of children with caner in new
structural equation models. Sociological Methods and Research 36, millennium sponsored by Childhood Cancer Foundation of the ROC,
462–494. Taiwan.
Childhood Cancer Foundation, 2008. Child census. Retrieved March 1, MacCallum, R.C., Browne, M.W., Sugawara, H.M., 1996. Power analysis
2009, from http://www.ccfroc.org.tw/child/child_census.php. and determination of sample size for covariance structure modeling.
Clarke-Steffen, L., 1993. A model of the family transition to living with Psychological Methods 1, 130–149.
childhood cancer. Cancer Practice 1, 285–292. Martinson, I.M., Leavitt, M., Liu, C., Armstrong, V., Hornberger, L., Zhang, J.,
Cohen, M.H., 1993. The unknown and the unknowable: managing sus- Han, X., 1999. Comparison of Chinese and Caucasian families caregiv-
tained uncertainty. Western Journal of Nursing Research 15, 77–96. ing to children with cancer at home: part I. Journal of Pediatric
Cohen, M.H., 1995. The triggers of heightened parental uncertainty in Nursing 14, 99–109.
chronic, life-threatening childhood illness. Qualitative Health Mast, M.E., 1998. Survivors of breast cancer: illness uncertainty, positive
Research 5, 63–77. reappraisal, and emotional distress. Oncology Nursing Forum 25,
Cohen, M.H., Martison, I.M., 1988. Chronic uncertainty: its effect of 555–562.
parental appraisal of a child’s health. Journal of Pediatric Nursing Mishel, M.H., 1981. The measurement of uncertainty in illness. Nursing
3, 89–96. Research 30, 258–263.
Deshon, R.P., 1998. A cautionary note on measurement error corrections Mishel, M.H., 1983. Parents’ perception of uncertainty concerning their
in structural equation models. Psychological Methods 3, 412–423. hospitalized child: reliability and validity of a scale. Nursing Research
Grootenhuis, M.A., Last, B.F., 1997. Adjustment and coping by parents of 32, 324–330.
children with cancer: a review of the literature. Supportive Care in Mishel, M.H., 1988. Uncertainty in illness. Image: Journal of Nursing
Cancer 5, 466–484. Scholarship 20, 225–232.
Ho, S.M.Y., Chan, C.L.W., Ho, R.T.H., 2004. Posttraumatic growth in Chinese Mishel, M.H., 1990. Reconceptualization of the uncertainty in illness
cancer survivors. Psycho-Oncology 13, 377–389. theory. Image: Journal of Nursing Scholarship 22, 256–262.
1524 L. Lin et al. / International Journal of Nursing Studies 47 (2010) 1510–1524

Mishel, M.H., 1997. Uncertainty in acute illness. Annual Review of Nursing Shek, D.T.L., Cheung, C.K., 1990. Locus of coping in a sample of Chinese
Research 15, 57–80. working parents: reliance on self of seeking help from others. Social
Mishel, M.H., 1999. Uncertainty in chronic illness. Annual Review of Behavior and Personality 18, 327–346.
Nursing Research 17, 269–294. Stein, R.E.K., Jessop, D.J., 1990. Functional status II(R): a measure of child
Mishel, M.H., Clayton, M.F., 2008. Theories of uncertainty in illness. In: health status. Medical Care 28, 1041–1055.
Smith, M.J., Liehr, P.R. (Eds.), Middle Range Theory for Nursing. 2nd Stewart, J.L., Mishel, M.H., 2000. Uncertainty in childhood illness: a
ed. Springer Publishing Company, LLC, New York, pp. 55–84. synthesis of the parent and child literature. Scholarly Inquiry for
Mishe, M.H., Fleury, J., 1997. The growth through uncertainty scale. Nursing Practice 14, 299–319.
Unpublished manuscript. The University of North Carolina at Chapel Tedeschi, R.G., Calhoun, L.G., 1996. The posttraumatic growth inventory:
Hill. measuring the positive legacy of trauma. Journal of Traumatic Stress
Mu, P., Ma, F., Hwang, B., Chao, Y., 2002. Families of children with cancer: the 9, 455–471.
impact on anxiety experienced by fathers. Cancer Nursing 25, 66–73. Wong, M.Y.-F., Chan, S.W.-C., 2006. The qualitative experience of Chinese
Mu, P., Ma, F., Ku, S., Shu, H., Hwang, B., Kuo, B.I., 2001. Families of Chinese parents with children diagnosed of cancer. Journal of Clinical Nursing
children with malignancy: the factors impact on mother’s anxiety. 15, 710–717.
Journal of Pediatric Nursing 16, 287–295. Yeh, C., 2001a. Development and testing of parental coping strategy
Parry, C., 2003. Embracing uncertainty: an exploration of the experiences inventory (PCSI) with children with cancer in Taiwan. Journal of
of childhood cancer survivors. Qualitative Health Research 13, 227– Advanced Nursing 36, 78–88.
246. Yeh, C., 2001b. Religious beliefs and practice of Taiwanese parents of
Porter, L.S., Clayton, M.F., Belyea, M., Mishel, M., Gil, K.M., Germino, B.B., pediatric patients with cancer. Cancer Nursing 24, 476–482.
2006. Predicting negative mood state and personal growth in African Yeh, C., 2003. Dynamic coping behaviors and process of parental response
American and Caucasian long term breast cancer survivors. Annals of to child’s cancer. Applied Nursing Research 16, 245–255.
Behavioral Medicine 31, 195–204. Yeh, C., Lee, T., Chen, M., Li, W., 2000. Adaptational process of parents of
Santacroce, S.J., 2003. Parental uncertainty and posttraumatic stress in pediatric oncology patients. Pediatric Hematology and Oncology 17,
serious childhood illness. Journal of Nursing Scholarship 35, 45–51. 119–131.