Human Reproduction, Vol.29, No.2 pp. 293– 302, 2014
Advanced Access publication on November 27, 2013 doi:10.1093/humrep/det418
ORIGINAL ARTICLE Psychology and counselling
Are patients at risk for psychological
maladjustment during fertility
treatment less willing to comply with
treatment? Results from the Portuguese
validation of the SCREENIVF
V. Lopes 1, M.C. Canavarro1, C.M. Verhaak2, J. Boivin 3, and S. Gameiro3,*
1
Faculty of Psychology and Educational Sciences, University of Coimbra, Rua do Colégio Novo, Coimbra, Apartado 6153, Portugal 2Department
of Medical Psychology, Radboud University Medical Center, Center, PO Box 9101, Nijmegen 6500 HB, The Netherlands 3Cardiff Fertility
Studies Research Group, School of Psychology, Cardiff University, Tower Building, Park Place, CF10 3AT Cardiff, Wales, UK
*Correspondence address. E-mail: [email protected]
Submitted on April 23, 2013; resubmitted on September 18, 2013; accepted on September 26, 2013
study question: Do patients at risk for psychological maladjustment during fertility treatment present lower intentions to comply with
recommended treatment than patients not at risk?
summary answer: Patients at risk of psychological maladjustment present similar high intentions to comply with recommended fertility
treatment to those not at risk but their intentions are conditioned by the degree of control they perceive over their fertility and its treatment and
their capacity to accept a future without biological children.
what is known already: Infertile couples refer to the psychological burden of treatment as one of the most important reasons for
withdrawal from recommended treatment. The SCREENIVF can be used before treatment to screen patients at risk for psychological maladjustment by assessing five risk factors: anxiety, depression, helplessness and lack of acceptance cognitions and social support.
study design, size, duration: Cross-sectional study. First, we investigated the psychometric properties of the Portuguese version of
the SCREENIVF. Secondly, we investigated associations between risk for psychological maladjustment and intentions to comply with treatment.
participants/ materials, setting, methods: Two hundred and ninety-one women and 92 men undergoing any stage of
fertility treatment at Portuguese infertility clinics were recruited online or in the clinical setting (55% response rate). Participants completed questionnaires that assessed their emotional adjustment, quality of life and compliance intentions.
main results and role of chance: The confirmatory factor analysis for the SCREENIVF indicated good fit [x 2 ¼ 188.50,
P , 0.001; comparative fit index ¼ 0.97; root-mean-square error of approximation ¼ 0.06 (90% CI 0.05– 0.07)] and all dimensions were reliable
(a ≥ 0.70, except depression for men: a ¼ 0.66). Fifty-two percent of women and 30% of men were at risk for maladjustment. Women and men
at risk and not at risk for maladjustment reported similar intentions to comply with treatment (P . 0.05). Cognitive risk factors moderated
negative associations found between distress and compliance intentions. Higher anxiety was associated with lower compliance intentions for
patients with lower helplessness cognitions (b ¼ 20.45, P ¼ 0.01) and men with higher acceptance cognitions (b ¼ 20.60; P ¼ 0.03), but
not for patients with higher helplessness cognitions (b ¼ 0.25, P ¼ 0.13) and men with lower acceptance cognitions (b ¼ 0.38; P ¼ 0.21).
Higher depression was associated with lower compliance intentions for patients with higher helplessness cognitions (b ¼ 20.33, P ¼ 0.02),
but not for patients with lower helplessness cognitions (b ¼ 0.19, P ¼ 0.30).
limitations, reasons for caution: Few men participated and thus only medium-to-large effect sizes could be detected for them.
Forty-eight percent of participants were recruited online and this could have resulted in higher rates of patients at risk.
wider implications of the findings: The SCREENIVF is not useful to identify patients at risk for non-compliance. However, the
clinic staff should be aware that patients who score high on helplessness cognitions and low on acceptance may need additional decisional aid to
make autonomous and satisfying decisions about uptake of treatment. The Portuguese version of the SCREENIVF is valid and reliable and can be
used with women undergoing any type of fertility treatment.
& The Author 2013. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: [email protected]
294
Lopes et al.
study funding/ competing interest(s): S.G. received a postdoctoral fellowship from the Portuguese Foundation for Science
and Technology (FCT-SFRH/BPD/63063/2009). There are no conflicts of interest to declare.
Key words: psychology / infertility / compliance / medically assisted reproduction
Introduction
Chances of conceiving are around 72% if infertile couples are willing to
undergo repeated fertility treatment cycles (Brandes et al., 2010).
However, the proportion of patients who fail to comply with recommended treatment is around 15% for intrauterine insemination (IUI)
(Goverde et al., 2000) and 22% for assisted reproductive treatment
(ART) (Gameiro et al., 2013). Patients refer to the psychological
burden of fertility and its treatment as one of the main reasons to discontinue treatment (Brandes et al., 2009; Gameiro et al., 2012). Factors that
increase patients’ vulnerability for psychological maladjustment during
treatment may thus also affect patients’ willingness to comply with
treatment.
The early identification of patients at risk for psychological maladjustment is important because it enables fertility staff to offer additional care
to these patients in order to prevent such problems and ease their experience of treatment. The SCREENIVF (Verhaak et al., 2010) is the
first screening tool specific for women in fertility care. It identifies
women at risk for maladjustment during treatment by assessing them
on five factors that were identified in prospective research as risk
factors for emotional problems after unsuccessful ART (Verhaak et al.,
2005a,b, 2010). These are pretreatment distress in terms of anxiety
and depression, helplessness cognitions in relation to infertility, lower acceptance cognitions about infertility and a childlessness life style and lack
of perceived social support. Studies showed that the SCREENIVF is an
acceptable instrument to identify women at risk for psychological maladjustment (Verhaak et al., 2010) and that its use in the clinic context
is feasible. These data suggest that the SCREENIVF may be an important
tool for female patients (Van Dongen et al., 2012). However, only its
Dutch version has been validated (Verhaak et al., 2010).To use the
SCREENIVF in other countries its psychometric properties need to be
investigated with the populations that use these countries’ fertility
care. In addition, the tool was also not validated for men or patients
undergoing other types of fertility treatment either than ART.
Compliance refers to the uptake of all fertility treatments recommended by the medical team, as long as there is ability to cover treatment
costs (Boivin et al., 2012). From the patients’ point of view, noncompliance, represents giving up the goal of biological parenthood. For
clinics it translates in lower success rates (Gameiro et al., 2012).
The literature shows that ‘emotional distress’ is the most cited reason
(22%) for non-compliance (Brandes et al., 2009). In addition, a recent
systematic review of patients’ stated reasons for non-compliance
showed that most reasons vary across treatment type but that the psychological burden of treatment is cited in all types and stages of treatment
(Gameiro et al., 2012). If psychological distress is associated with noncompliance with treatment, it may be expected that patients with
higher vulnerability for psychological distress during treatment will be
less likely to comply.
Because psychological distress is associated with non-compliance, we
hypothesized that all SCREENIVF risk factors (anxiety, depression,
helplessness, and lack of acceptance cognitions and social support)
would be negatively associated with compliance intentions. In addition,
we hypothesized that helplessness and acceptance would moderate
the associations between distress (anxiety and depression) and compliance. Helplessness refers to a sense of lack of control about infertility and
is associated with perceptions of self-inefficacy (Seligman, 1975). Thus,
more distressed patients with more helplessness cognitions may be
less able to continue treatment than more distressed patients with less
helplessness cognitions. On the contrary, acceptance of a childfree lifestyle facilitates disengagement from the goal of parenthood (Daniluk,
2001). Thus, more distressed patients who have a higher acceptance
of a childless life style may be less willing to continue treatment than
more distressed patients with lower acceptance.
When compared with infertile men, infertile women express higher
need and identification with the parenthood role, report higher
infertility-related distress (Chachamovich et al., 2009; Slade et al.,
2007), and tend to be more proactive in the pursuit of treatment
options (Jordan and Revenson, 1999). These data suggest that women
may be more willing to undergo the necessary treatment to achieve parenthood. Thus, we predicted that gender would moderate the associations between risk factors (and its predicted interactions) and
compliance, whereby these would be stronger or only found for men.
The present cross-sectional study had two main goals. The first was to
investigate the psychometric properties of the Portuguese version of the
SCREENIVF. The second was to investigate the relationship between vulnerability to psychological maladjustment and compliance by looking at
patients’ intentions to comply with treatment. We investigated if patients
identified by the SCREENIVF as being at risk for maladjustment would
report lower compliance intentions than patients identified as not
being at risk; and the relationship between the five SCREENIVF risk
factors and patients’ compliance intentions.
Materials and Methods
Procedures
The study was approved by the Ethics Committee of the Coimbra University
Hospitals.
Participants were recruited online and at the clinical setting between
January 2011 and February 2012. Inclusion criteria were being adult and
undergoing fertility diagnosis or treatment at a fertility clinic in Portugal.
The Portuguese Government reimburses a maximum of three ART cycles
per couple, whether on public or private clinics. Similar treatments are provided and similar conditions are required in public and private clinics: individuals need to be over 18 years old, be in an heterosexual relationship and be
married (or in a similar situation) for more than 2 years. Online recruitment
was done through a web-based survey that was divulged in the APFertilidade
website, the main patient advocacy group in Portugal. A Facebook cause was
also created and divulged among all APFertilidade Facebook friends. The clinic
setting consisted of a large university-based hospital where individuals were
consecutively invited to participate in the study. While participants recruited
295
Psychological adjustment and treatment compliance
online could only fill the questionnaire online, participants recruited at the
clinic could choose between filling it online or on paper. In the latter case, participants were given the survey in an envelope and instructed to complete it at
the clinic while waiting for their appointment or at home and return it to the
clinic in a pre-addressed sealed envelope. All participants signed an informed
consent and confidentiality was guaranteed. Figure 1 of Supplementary data
presents the sample collection flowchart. A total of 222 questionnaires were
submitted online but nine duplicates (same email address provided) were
excluded. At the clinical setting a total of 478 patients were invited to participate but only 233 filled and returned questionnaires (response rate 49%). In
total 446 questionnaires were delivered. From these, only those where 80%
of the SCREENIVF was filled were retained and 12 were excluded because
they were identified as outliers (based on age, relationship duration or
time trying to conceive).
Materials
Socio-demographic information included gender, age, relationship duration,
university education (no, yes) and area of residence (urban, rural). Fertility
information (self-reported) included duration of infertility, number of previous fertility treatments, parity (0, .1) and current stage of treatment (diagnostic examination, medication/injections, IUI, waiting list for ART, ART).
Risk for psychological maladjustment was assessed with the SCREENIVF
tool (Original Version: Verhaak et al., 2010; Portuguese Version: Gameiro
and Canavarro, 2011). The SCREENIVF is composed of 34 items organized
in 5 dimensions that assess risk for maladjustment. Anxiety was assessed with
10 items (e.g. ‘I get very nervous and worried when thinking about my current
troubles’) from the short version of the Spielberg State and Trait Anxiety Inventory (Spielberger, 1983). Depression was assessed with seven items (e.g.
‘I feel sad’) from the Beck Depression Inventory, version for patients of
general practitioners (Beck et al., 1997). Helplessness (six items, e.g. ‘My fertility problems control my life’) and acceptance cognitions (six items, e.g.
‘I can accept my fertility problems’) were assessed with items from the
Illness Cognition Questionnaire for IVF patients (Evers et al., 2001;
Verhaak et al., 2005b). Social support was assessed with five items (e.g.
‘When I feel sad there is always someone I can talk to’) derived from the Inventory of Social Involvement (Van Dam-Baggen and Kraaimaat, 1992). The
original version of the SCREENIVF exhibited excellent reliability in all scales
(Cronbach’s alphas between 0.82 and 0.92) (Verhaak et al., 2010). Based
on patients’ scores on the five risk factors, the tool classifies patients as ‘at
risk’ or ‘not at risk’ for emotional maladjustment. The SCREENIVF correctly
identified 69% of the total of patients who presented clinical significant emotional difficulties and 77% of those who did not (Verhaak et al., 2010). To
develop the Portuguese version of the SCREENIVF, we followed Humbleton’s recommendations for adapting tests (Humbleton et al., 2005). To classify individuals as at risk, we followed the procedures described by Verhaak
et al. (2010). The cut-off score for depression was four or higher, which is
in line with previous studies (Beck et al., 1997; Verhaak et al., 2010). For
anxiety, helplessness and acceptance cognitions and social support, scores
were based on one standard deviation above or below the sample mean
scores. Thus, cut-off score for anxiety was 27 or above; for helplessness cognitions was 15 or above; for acceptance cognitions was 11 or below and for
social support was 13 or below. In each of the five risk factors, if patients
scored above/below the cut-off point, they were assigned a score of 1 (at
risk); otherwise, their score was 0 (not at risk). Patients are classified as at
risk if they are at risk in at least one of the five risk factors.
Quality of life (QoL) was assessed with the Fertility Quality of Life tool
(FertiQol; Original Version: Boivin et al., 2011; Portuguese Version:
Gameiro and Canavarro, 2010). The core FertiQol is composed of 24
items organized in four QoL domains: emotional (six items, e.g. ‘Do your
fertility problems cause feelings of jealousy and resentment?’), mind-body
(six items, e.g. ‘Are your attention and concentration impaired by thoughts
of infertility?’), relational (six items, e.g. ‘Have fertility problems strengthened
your commitment to your partner?’) and social (six items, e.g. ‘Do you feel
social pressure on you to have (or have more) children?’). The FertiQoL
treatment module is composed of 10 items organized in 2 domains: environment (6 items, e.g. ‘Are you satisfied with your interactions with fertility
medical staff?’), and treatment tolerability (4 items, e.g. ‘Are you bothered
by the effect of the treatment on your daily or work-related activities?’).
Total scores range from 0 to 100, with higher scores indicating higher
QoL. The Portuguese version of the FertiQoL has good psychometric characteristics (Melo et al., 2012). In the present sample Cronbach’s alpha values
varied from 0.71 to 0.90.
Compliance intentions were assessed with the FertiQoL persistence scale
(Boivin et al., 2011). This scale is composed of six item (e.g. ‘How often do
you consider withdrawal from treatment?’), with a 5-point Likert answering
scale ranging from 1 (Never) to 5 (Always), assessing patient’s motivation to
persist in treatment. Scores vary from 6 to 30, with higher scores indicating
stronger intentions to do more treatments. In the present sample Cronbach’s alpha value was 0.77.
Data analysis
Socio-demographic and clinic characteristics
We compared the group of patients recruited online with the group of
patients recruited at clinic setting regarding their socio-demographic and
clinic characteristics (Fisher’s exact test, Student’s t-test, Crammer’s V),
their wellbeing (multivariate analysis of variance) and compliance intentions
(Student’s t-test).
Psychometric properties of the SCREENIVF
The psychometric properties investigated were construct validity and reliability.
Construct validity was first examined with confirmatory factor analysis
(CFA). The CFA was performed using AMOS, version 17.0, to test the structure of the SCREENIVF tool. The model was a first-order model with five
latent variables that correspond to the five risk factors for emotional maladjustment. For each of the dimensions of the SCREENIVF, three parcels
(i.e. combination of individual items) were generated by randomly combining
the items of that dimension (Little et al., 2002), a statistical procedure that is
known to ameliorate model fit and to produce less bias in the estimation of
structural parameters (Bandalos, 2002). To assess model fit different parameters commonly used were analysed: x 2, the comparative fit index (CFI)
and the root-mean-square error of approximation (RMSEA) (Byrne,
2010). A model is considered to have very good fit if the x 2 value is nonsignificant (P . 0.05), the CFI is .0.95 and the RMSEA is ,0.06 (Hu and
Bentler, 1998). To investigate if the SCREENIVF can be used with different
groups of patients its measurement (associations of observed scores to the
latent variables) and structural (associations of latent variables with each
other) invariance were tested across gender and treatment type (pre-ART
versus ART). Invariance in the model occurs when the x 2 difference
between the tested models is non-significant (Byrne, 2010) or the CFI difference is , 0.01 (Cheung and Rensvold, 2002). Secondly, associations
between the five SCREENIVF risk factors and between these and the QoL
domains were investigated. The largest associations between the different
risk factors were expected between anxiety and depression and the lowest
between social support and the remaining risk factors. In addition, higher
anxiety, depression and helplessness and lower acceptance and social
support were expected to be significantly associated with lower QoL. The
largest correlations were expected between anxiety and depression and
the emotional and mind-body QoL domains, between helplessness and acceptance cognitions and the emotional QoL domain, between social
support and the relational and social QoL domains. The lowest correlations
were expected between the risk factors and the environment and tolerability
to treatment domains.
296
The reliability (internal consistency) of the SCREENIVF was investigated
using Cronbach’s alpha and by analyzing the correlation between each item
and its specific dimension.
Relationship between vulnerability to psychological distress and compliance to
fertility treatment
Analyses involving compliance intentions were limited to a subgroup of the
total sample composed by 295 patients (213 women and 82 men) who
were undergoing treatment at a public clinic and had done fewer than
three IVF/ICSI cycles, thus assuring that patients met the legal criteria to
access governmental funding for treatment in Portugal and eliminating financial confounders.
To investigate if patients identified by the SCREENIVF as at risk for emotional maladjustment reported lower intentions to comply with treatment
than patients not at risk we used univariate analysis of variance (ANOVA).
Finally, to investigate how the SCREENIVF risk factors were associated with
intentions to comply with treatment, one hierarchical linear regression was
performed. First, any socio-demographic or clinic characteristics that were
associated with patients’ compliance intentions were entered in the first
step of the model along with the five SCREENIVF dimensions and gender.
This allows controlling for individual variation in patients’ background and fertility characteristics. In step two the interaction products between the five risk
factors and gender were entered. In step three the four interaction terms
between cognitions (helplessness and acceptance) and emotional adjustment
(anxiety and depression) were entered. Finally, in the fourth step, three-way
interactions of gender, cognitions and emotional adjustment were entered.
Continuous variables were transformed into z-scores to avoid multicolinearity
problems in the interaction products (Baron and Kenny, 1986).
Results
Participants
Sample characteristics and SCREENIVF results are presented in Table I.
The final sample was composed of 291 women and 92 men. From the
total of participants, 25% were couples. Men formed a larger proportion
of recruits at the clinic than online (37 versus 10%, x 2 ¼ 37.94,
P , 0.001). Both women and men were in their early 30s. Individuals
had been with their partners for an average of 7 years. Women [33.24
(3.62) versus 32.92 (3.54), t ¼ 20.75, P ¼ 0.46] and men [35.18
(4.05) versus 33.78 (4.15), t ¼ 21.31, P ¼ 0.20] recruited at the clinical
setting were not significantly older than women and men recruited
online. Individuals recruited at the clinic context were less likely to
attend college or university (33 versus 65%, x 2 ¼ 39.59, P , 0.001)
and to live in urban areas (44 versus 87%, x 2 ¼ 39.59, P , 0.001) than
individuals recruited online. The majority of participants did not have children (89.7%) and they had been trying to get pregnant for an average of
four years, having done on average 0.43 IUI and 0.90 IVF treatment
cycles. Participants recruited at the clinical context were more likely to
be at less advanced stages of treatment (waiting for or undergoing
ART: 40 versus 57%, x 2 ¼ 77.70, P , 0.01) and had done significantly
less IVF treatments [0.56 (0.90) versus 1.26 (1.61), t ¼ 5.11,
P , 0.001] than individuals recruited online.
Individuals recruited online showed less QoL in all domains than
individuals recruited at clinic context (F(4,370) ¼ 10.172; h 2 ¼ 0.099;
P , 0.001; Pillai’s Trace ¼ 0.099): emotional (F(1,376) ¼ 27.21; P , 0.01),
mind-body (F(1,376) ¼ 28.43; P , 0.01), relational (F(1,373) ¼ 4.08;
P , 0.05) and social (F(1,376) ¼ 38.07; P , 0.01) domains. The two
groups did not differ in their compliance intentions [24.43 (4.34)
versus 24.28 (4.06), t ¼ 0.296; P ¼ 0.768)].
Lopes et al.
Table I Mean (SD) or frequencies (%) for sample
characteristics and SCREENIVF results (n 5 383).
Socio-demographic
Gender, n (%)
Female
Male
Age (years), mean (SD)
291 (76.0)
92 (24.0)
33.50 (3.8)
Female
33.06 (3.57)
Male
34.90 (4.09)
Relationship duration (years), mean (SD)
7.07 (3.33)
College or university education, n (%)
No
197 (52.0)
Yes
182 (48.0)
Residence zone, n (%)
Urban
244 (64.2)
Rural
136 (35.8)
Clinic
Infertility duration (years), mean (SD)
4.30 (2.51)
Number of previous treatments, mean (SD)
IUI
0.43 (1.03)
IVF/ICSI
0.90 (1.38)
Children, n (%)
No
341 (89.7)
Yes
39 (10.3)
Current stage of treatment, n (%)
Diagnostic testing
101 (26.9)
Medication/injections
68 (18.1)
IUI
24 (6.4)
Waiting to start IVF/ICSI
IVF/ICSI
SCREENIVF, At risk
Anxiety
Depression
Helplessness cognitions
56 (14.9)
126 (33.6)
n (%)
70 (18.4)
106 (28.1)
83 (21.7)
Acceptance cognitions
70 (18.3)
Social support
70 (18.3)
At risk in one or more dimensions
Men
28 (30.4)
Women
152 (52.2)
Total
180 (47.0)
A total of 180 (47%) individuals were identified as at risk for psychological maladjustment. These included 52% of the women and 30% of
the men who participated in the study.
Psychometric properties of the SCREENIVF
Construct validity
Figure 1 presents standardized estimates for measurement and structural
paths of the tested model, separately for women and men. The x 2 value
of the model was significant (x 2(80) ¼ 188.50; P , 0.001). The index
values were very good and good, respectively: CFI ¼ 0.97; RMSEA ¼
297
Psychological adjustment and treatment compliance
Figure 1 Standardized regression weights of factor loading. Note: E, error, P, parcel, C, standardized regression weights for women; F, standardized
regression weights for men.
0.06 [confidence interval (CI) 90% 0.05 –0.07]; (Hu and Bentler, 1998).
All the standardized factor loadings of the item parcels into their correspondent latent construct were statically significant (P , 0.001) and
.0.75. The five risk factors were all statistically related. Standardized regression weights suggest large associations between anxiety, depression
and helplessness and acceptance cognitions; and medium associations
between social support and the other four risk factors. The larger association was between depression and anxiety (0.84) and the smaller
between helplessness cognitions and social support (0.31). Supplementary data, Table SI presents results of the test of the measurement and
structural invariance of the SCREENIVF. Results indicate that the
SCREENIVF was invariant across treatment types. Significant structural
variance was observed for gender. While in general the correlation coefficients between helplessness and acceptance cognitions and anxiety and
depression were large (r . 0.50) for women, for men they were medium
(0.30 , r , 0.50). In addition, the correlation coefficient between helplessness and social support was medium for women but small for men.
Associations between the SCREENIVF risk factors and between these
and the FertiQoL domains of QoL are presented in Table II and were as
expected. The only exception was that no significant association was
found between social support and tolerability to treatment.
Reliability
Supplementary data, Table SII shows mean and standard deviation values
for each item, item-subscales correlations, Cronbach’s alpha values if the
item was deleted and Cronbach’s alpha values for each subscale (for
women and men separately). All item-subscale correlation exceeded
the value of 0.40, indicating that all items adequately represent the
concept that each subscale measures (Cohen, 1992). Cronbach’s
alpha varied from 0.85 (depression) to 0.93 (acceptance cognitions)
for women and between 0.66 (depression) and 0.91 (social support)
for men.
Discriminant validity
The CFA showed that the structural relations of the SCREENIVF were
different for women and men. Therefore, all analysis of variance based
on the SCREENIVF classification of individuals as at risk or not were
run separately for each gender and are presented in Table III. Results
showed that women and men identified by the SCREENIVF as at risk
for maladjustment reported lower QoL than women (F(4, 280) ¼
50.93, h 2 ¼ 0.42, P , 0.001; Pillai’s Trace ¼ 0.42) and men (F(4, 85) ¼
7.47, h 2 ¼ 0.26, P , 0.001; Pillai’s Trace ¼ 0.26) not at risk across all
domains of QoL (emotional, mind-body, relational and social).
Relationship between vulnerability
to psychological distress and compliance
to fertility treatment
The CFA showed that the structural relations of the SCREENIVF varied
with gender. Therefore, the ANOVA of intentions to comply with treatment, which had the SCREENIVF classification of individuals as at risk or
not as the between subject factor, was run separately for men and
women. Results are presented in Table III. No significant group differences were found between women and men at risk and not at risk for
maladjustment.
298
Lopes et al.
Table II Descriptive statistics and correlations between the SCREENIVF risk factors and the FertiQoL tool domains of QoL.
SCREENIVF
.....................................................................................................................................
Mean (SD)
Anxiety
Depression
Helplessness cognitions
Acceptance cognitions
Social support
.............................................................................................................................................................................................
SCREENIVF
Anxiety
Depression
20.87 (6.28)
1
2.51 (3.09)
0.70***
Helplessness cognitions
11.10 (4.29)
0.62***
1
0.65***
Acceptance cognitions
15.68 (4.68)
20.58***
20.51***
20.55***
1
Social support
16.48 (3.70)
20.37***
20.39***
20.27***
0.32***
1
1
FERTIQOL
Emotional
62.13 (21.41)
20.69***
20.69***
20.75***
0.71***
0.32***
Mind-body
66.17 (23.67)
20.63***
20.69***
20.74***
0.61***
0.28***
Relational
78.82 (15.59)
20.45***
20.48***
20.34***
0.35***
0.42***
Social
68.04 (20.20)
20.53***
20.61***
20.67***
0.53***
0.33***
Environment
62.28 (17.53)
20.26***
20.25***
20.14***
0.24***
0.15**
Tolerability
68.63 (19.86)
20.34***
20.31***
20.30*
0.34***
0.06
**P , 0.01, ***P , 0.001.
Table III Differences between women and men classified as at risk and not at risk regarding QoL and intentions to comply
with treatment.
Women
.....................................................................
At risk
(n 5 148)
....................
Mean
SD
Not at risk
(n 5 137)
SD
....................................................................
At risk
(n 5 27)
....................
Mean
Men
....................
F
h2p
Mean
0.38
63.89
SD
Not at risk
(n 5 63)
....................
Mean
SD
h2p
F
.............................................................................................................................................................................................
QoL
Emotional
45.56
18.52
71.66
14.69
171.90***
14.62
79.63
15.10
20.93***
0.19
Mind-body
49.12
22.10
76.12
16.44
135.26***
0.32
70.31
21.35
82.62
15.04
9.74**
0.10
Relational
71.39
17.17
85.43
11.05
66.25***
0.19
73.77
15.10
84.06
11.65
12.28**
0.12
Social
54.53
20.24
76.34
14.85
106.18***
0.27
70.96
16.30
80.22
13.03
8.20**
0.09
At risk
(n ¼ 104)
Mean
Intentions to comply with treatment
23.98
SD
4.16
Not at risk
(n ¼ 101)
Mean
24.84
SD
4.32
At risk
(n ¼ 25)
F
2.11
h2p
Mean
0.01
23.47
SD
3.81
Not at risk
(n ¼ 50)
Mean
24.46
SD
3.96
h2p
F
1.07
0.01
Note: F, F values, h2p, partial eta-squared.
**P , 0.01, ***P , 0.001.
Table IV presents results from the hierarchical regression investigating
predictors of patients’ intentions to comply with treatment. Age was
negatively associated with patients’ intentions to comply with treatment.
Significant effects were found for the interaction between helplessness
cognitions and anxiety and the interaction between helplessness cognitions and depression. Post hoc analysis for these interactions showed
that for patients with low helplessness cognitions, higher anxiety was
associated with lower intentions to comply with treatment
(b ¼ 20.45, P ¼ 0.01). However, for patients with high helplessness
cognitions, no significant association was found between anxiety and
intentions to comply with treatment (b ¼ 0.25, P ¼ 0.13). Moreover,
the results showed that for patients with high cognitions of helplessness,
higher depression was associated with lower intentions to comply with
recommended treatment (b ¼ 20.33, P ¼ 0.02). No association
was observed for patients with low helplessness cognitions (b ¼ 0.19,
P ¼ 0.30).
A significant effect for the three-way interaction between gender, acceptance cognitions and anxiety was also found. Post hoc analysis showed
that for men the interaction between acceptance cognitions and anxiety
was significant (b ¼ 20.53; P ¼ 0.02), but for women it was not
299
Psychological adjustment and treatment compliance
Table IV Hierarchic regression for intentions to comply with treatment (n 5 274).
Predictors
B
b
SE
DF
Adj R 2
DR 2
P
.............................................................................................................................................................................................
Step 1
2.97
0.05
0.07
0.005
Age
20.24
0.07
20.22
0.001
Gender
20.03
0.74
20.00
0.963
Anxiety
20.42
0.45
20.10
0.350
Depression
20.30
0.50
20.07
0.547
Helplessness cognitions
0.43
0.45
0.10
0.343
Acceptance cognitions
0.13
0.37
0.03
0.720
Social support
0.06
0.31
0.01
Step 2
0.851
0.41
0.04
0.01
0.840
Gender × anxiety
20.05
0.94
20.01
0.960
Gender × depression
20.55
1.13
20.05
0.628
Gender × helplessness cognitions
Gender × acceptance cognitions
Gender × social support
0.29
0.84
0.03
0.735
20.62
0.90
20.07
0.491
0.48
0.66
0.06
Step 3
Helplessness cognitions × anxiety
Helplessness cognitions × depression
Acceptance cognitions × anxiety
Acceptance cognitions × depression
Gender × helplessness cognitions × depression
Gender × acceptance cognitions × anxiety
Gender × acceptance cognitions × depression
0.04
0.02
0.303
1.47
0.54
0.44
0.007
21.09
0.46
20.44
0.019
0.41
0.54
0.10
0.456
20.51
0.56
20.16
0.361
21.82
1.05
20.19
1.40
0.94
0.14
0.138
22.47
1.04
20.30
0.018
2.90
1.51
0.26
0.056
Step 4
Gender × helplessness cognitions × anxiety
0.464
1.22
2.64
2
0.07
0.04
0.035
0.082
2
Note: B, b-values; SE, standard error; b, beta values; DF, F change; Adj R , Adjusted R squared; DR , R squared change.
Bold values indicate P , 0.05.
(b ¼ 0.10; P ¼ 0.46). For men with high levels of acceptance cognitions,
higher anxiety was associated with lower intentions to comply with treatment (b ¼ 20.60; P ¼ 0.03). On the other hand, for men with low
levels of acceptance cognitions anxiety was not associated with their
intentions to comply with treatment (b ¼ 0.38; P ¼ 0.21).
Discussion
The importance of screening patients at the start of treatment in order to
provide tailored care is currently accepted. The SCREENIVF tool has
proved valid for the effect and its use feasible in daily routine care in
the Netherlands. The Portuguese version of the SCREENIVF demonstrated construct validity and reliability in a sample of men and women
undergoing infertility diagnosis or treatment at Portuguese clinics. As
predicted, risk factors for psychological distress are associated with compliance intentions. More specifically, associations between patients’ distress (anxiety and depression) and their willingness to comply with
treatment are conditioned by the degree of control patients perceive
in relation to fertility and its treatments and their capacity to accept a
future without biological children.
Results from the CFA indicated that the measurement model of the
SCREENIVF has a good fit in our sample data. The CFA showed that
the five risk factors assessed by this instrument are independent but
significantly associated. In addition, associations with the different
domains of QoL were as expected, suggesting that the risk factors do
capture differences in the functioning status of the individuals. Internal
consistency analysis showed that, in general, the subscales that assess
the five risk factors are reliable for men and women. Measurement invariance was ascertained across gender and treatment type, indicating that
the SCREENIVF items contribute equally to the assessment of each
risk factor for all patients. This supports the use of single cut-off scores
for classifying patients as at risk or not for psychological maladjustment
across the different risk factors, regardless of gender and treatment
stage. The structure of the SCREENIVF varied across gender, reflecting
a stronger link between cognitions and distress for women than men (see
Fig. 1). The implications of this for the classification of men as at risk or not
for psychological distress need to be further investigated. Overall, the
results show that the SCREENIVF is a valid and reliable tool to assess
risk factors for psychological maladjustment to infertility treatment.
Although the SCREENIVF was firstly developed to screen women
entering ART, the fact that it is invariant across types of treatment suggests that it can be applied to all women, regardless of the treatment
they are undergoing (i.e. from diagnosis to ART). Because different
studies carried out in Portugal have shown that patients undergoing fertility treatment report similar experiences to patients in other European
countries (Galhardo et al., 2011; Moura-Ramos et al., 2012) we can
300
expect the screening capacity of the SCREENIVF to be similar to that of
the original Dutch version (Verhaak et al., 2010). This means that we can
expect it to be able to identify women at risk for maladjustment.
However, results of the CFA showed structural variance across
gender, reflecting that the way in which the five risk factors associate is
different for women and men. This finding suggests that its predictive capacity may be different for men and therefore needs further investigation.
On the one side, it is known that women and men report similar experiences and difficulties across a single cycle of IVF/ICSI (Boivin et al., 1998)
and this would support the use of the SCREENIVF to identify men at risk
for maladjustment. On the other side, it is clear that women have stronger emotional reactions to the diagnosis of infertility (Wischmann et al.,
2009) and to treatment failure (Slade et al., 1997) and this could suggest
that the SCREENIVF may result in too many false positives.
Comparing with the original study of the SCREENIVF (Verhaak et al.,
2010) the cut-off scores obtained with our sample were either equal or,
in the case of anxiety, helplessness and social support more conservative
(3, 1 and 2 points difference, respectively). Our results also show that
more patients scored above the cut-off scores (i.e. classified as at risk)
in terms of anxiety and depression (10 and 11% versus 18 and 28%,
respectively) and overall (34 versus 52% for women in the SCREENIVF
validation study; 40 versus 52% for women and 20, 26 versus 30% for
men in the study by van Dongen and colleagues). In these studies all
participants were recruited in clinics while 48% of our sample was
recruited online. Online samples score higher on distress (Boivin et al.,
2011), especially if participants come from patient advocacy groups
and this can explain the differences found. Further research is advisable
to investigate if different cut-off scores should be defined based on
exclusive clinical samples.
Contrary to what was predicted, women and men at risk for psychological maladjustment were equally willing to comply with treatment
as those not at risk and no associations between risk factors and compliance intentions were found. On average patients reported high
intentions to comply with treatment (24 on a scale from 0 to 30).
Additionally, only 6.5% of patients stated that they were not or not
at all likely to undergo (more) recommended treatment (data not
shown). However, research shows that noncompliance is a real phenomenon (Gameiro et al., 2013). These data suggest that both distressed and
non-distressed patients may be overconfident about their ability to
comply with treatment, as already observed in patients with other
medical conditions that require demanding treatment such as breast
cancer (Güth et al., 2012).
As predicted, helplessness and acceptance cognitions moderated the
association between distress and compliance intentions. More precisely,
results showed that depressed patients are less willing to do more treatment only when they perceive lower control (i.e. higher helplessness),
and anxious men are less willing to comply with treatment only when
they report higher acceptance of infertility. However, an unexpected
result showed that more anxious patients are less willing to comply
with treatment only if they perceive higher control (i.e. lower helplessness). A previous study showed that 80– 85% of patients find it very difficult to end treatment when they experience the feeling of becoming
overwhelmed by their desire for a child (Rauprich et al., 2011). It is difficult to disentangle lack of acceptance of infertility and perceived helplessness (in our sample the correlation between the two was large) but both
reflect a sense of lack of purpose in life without children. Overall these
results suggest that negative cognitions about infertility and childlessness
Lopes et al.
may threaten patients’ capacity to make autonomous decisions when
balancing their present wellbeing with the hypothesized rewards from
parenthood.
These results highlight the importance of providing clear information
about treatment success rates, managing patient’s expectations about
treatment and helping them to consider in advance all possible outcomes.
The clinical staff can also have an important role in promoting patients’ personal efficacy and mastery regarding treatment by involving patients in the
treatment process and all associated decision-making (Israel et al., 1994;
Perkins and Zimmerman, 1995)and by addressing their concerns about
the treatment procedures (Pedro et al., 2013).
This study presented methodological limitations that must be considered. First, the sample included a small number of men and only medium
to large effect sizes could be detected for them (P , 0.05, power ¼
0.80, effect size ¼ 0.33) (Faul et al., 2007). Secondly, 48% of participants
were recruited online and differences between participants recruited
online and in clinical setting were observed. Although the literature indicates that data collected through online recruitment is valid (Lieberman,
2008) one would need to determine whether the differences observed
may have implications for the definition of the SCREENIVF cut-off
scores. Thirdly, we did not investigate the predictive value of the Portuguese version of the SCREENIVF. Thus, and although we can expect its
screening capacity to be similar to that of the original Dutch version
(Verhaak et al., 2010), longitudinal research for this purpose is still
required. Fourthly, although we investigated risk factors for women
and men, we did not investigate how the psychopathological vulnerability
of one couple’s member may affect the other member and the couples’
compliance behavior, which should also be the subject of future research.
Finally, we did not investigate actual behavior, but only compliance intentions. However, it should be noted that intentions are identified as the
core psychological predictor of behavior, explaining on average, 28%
of the variance in future behavior (Sheeran, 2002).
To conclude, the Portuguese version of the SCREENIVF proved to be
valid and reliable, meaning that it can be used with women undergoing
any type of fertility treatment. Further investigation is needed to fully
attest its usefulness with men. The use of the SCREENIVF is feasible in
routine care and may be of advantage for patients and fertility staff
alike. For patients, the feedback provided by the SCREENIVF may increase their awareness about their (or their partner’s) vulnerabilities.
Patients that recognize their risk profile may be more willing to seek professional help or accept it when offered (Van Dongen et al., 2012). They
may also be more willing to seek other services, for instance, patients
without social support may join patient support groups or online
forums. Even if patients choose not to seek help, the clinic staff should
tailor the care provided for at risk patients and they now have an integrated model to help them tackling patient vulnerabilities that contribute
to increase the burden of ART treatment (Boivin et al., 2012). Because
this study showed that the SCREENIVF can be used with patients undergoing any type of treatment it may be useful to help fertility staff to target
patients’ vulnerabilities from the moment they enter the clinic. This can
result in a better overall treatment experience and a better patient preparation for the more demanding types of treatment, such as ART. Although the SCREENIVF is not useful to identify patients at risk for
non-compliance, the clinic staff should be aware that patients who
score high on helplessness and low on acceptance may need additional
decisional aid to make autonomous and satisfying decisions about
uptake of treatment.
Psychological adjustment and treatment compliance
Supplementary data
Supplementary data are available at http://humrep.oxfordjournals.org/.
Authors’ roles
V.L., M. C. C., C.M.V., J. B. and S.G. gave substantial contributions to the
conception and design of this study. V. L. was involved in the acquisition of
the data. V. L. and S.G. were involved in the interpretation of data. All
authors were involved in revising this paper.
Funding
S.G. received a postdoctoral fellowship from the Portuguese Foundation
for Science and Technology (FCT-SFRH/BPD/63063/2009).
Conflict of interest
None declared.
References
Bandalos D. The effects of item parceling on goodness-of-fit and parameter
estimate bias in structural equation modeling. Educ Psychol Pap Publ 2002;
9:78 – 102.
Baron R, Kenny D. The moderator-mediator variable distinction in social
psychological research: conceptual, strategic, and statistical
considerations. J Pers Soc Psychol 1986;51:1173 – 1182.
Beck AT, Guth D, Steer RA, Ball R. Screening for major depression disorders
in medical inpatients with the beck depression inventory for primary care.
Behav Res Ther 1997;35:785 – 791.
Boivin J, Andersson L, Skoog-Svanberg A, Hjelmstedt A, Collins A, Bergh T.
Psychological reactions during in-vitro fertilization: similar response patter
in husbands and wives. Hum Reprod 1998;13:3262 – 3267.
Boivin J, Takefman J, Braverman A. The fertility quality of life (FertiQoL) tool:
development and general psychometric properties. Hum Reprod 2011;
26:2084– 2091.
Boivin J, Domar A, Shapiro D, Wischmann T, Fauser B, Verhaak C. Tackling
burden in ART: An integrated approach for medical staff. Hum Reprod
2012;0:1 – 10.
Brandes M, van der Steen J, Bokdam S, Hamilton C, de Bruin J, Nelen W,
Kremer J. When and why do subfertile couples discontinue their fertility
care? A longitudinal cohort study in a secondary care subfertility
population. Hum Reprod 2009;24:3127 –3135.
Brandes M, Hamilton C, de Bruin J, Nelen W, Kremer J. The relative
contribution of IVF to the total ongoing pregnancy rate in a subfertile
cohort. Hum Reprod 2010;25:118 – 126.
Byrne B. Structural Equation Modeling with AMOS: Basic Concepts, Applications
and Programming, New York: Routledge, 2010.
Chachamovich J, Chachamovich E, Fleck MP, Cordova FP, Passos E.
Congruence of quality of life among infertile men and women: findings
from a couple-based study. Hum Reprod 2009;24:2151– 2157.
Cheung G, Rensvold R. Evaluating goodness-of-fit indexes for testing
measurement invariance. Struct Equ Modeling 2002;9:233 – 255.
Cohen J. A power primer. Psychol Bull 1992;112:155 – 159.
301
Daniluk J. Reconstructing their lives: a longitudinal, qualitative analysis of the
transition to biological childlessness for infertile couples. J Counsel Dev
2001;79:439 – 449.
Evers AW, Kraaimaat FW, Lankveld W, Jongen PJ, Biijlsma JWJ. Beyond
unfavourable thinking: the illness cognition questionnaire for chronic
diseases. J Counsult Clin Psychol 2001;69:1026 – 1036.
Faul F, Erdfelder E, Lang A-G, Buchner A. G*power 3: a flexible statistical
power analysis program for the social, behavioral, and biomedical
sciences. Behav Res Methods 2007;39:175– 191.
Galhardo A, Pinto-Gouveia J, Cunha M, Matos M. The impact of shame and
self-judgment on psychopathology in infertile patients. Hum Reprod 2011;
26:2408– 2414.
Gameiro S, Boivin J, Peronace LA, Verhaak C. Why do patients discontinue
fertility treatment? A systematic review of reasons and predictors of
discontinuation in fertility treatment. Hum Reprod Update 2012;18:652–669.
Gameiro S, Verhaak CM, Kremer JAM, Boivin J. Why we should talk about
compliance with assisted reproductive technologies (ART): a systematic
review and meta-analysis of ART compliance rates. Hum Reprod Update
2013;19:124 – 135.
Goverde A, McDonnell J, Vermeiden J, Schats R, Rutten F, Schoemaker J.
Intrauterine insemination or in-vitro fertilisation in idiopathic subfertility
and male subfertility: a randomised trial and cost effectiveness analysis.
Lancet 2000;355:13 –18.
Güth U, Myrick ME, Kilic N, Eppenberger-Castori S, Schmid SM. Compliance
and persistence of endocrine adjuvant breast cancer therapy. Breast Cancer
Res Treat 2012;131:491– 499.
Hu L, Bentler P. Fit indices in covariance structure modeling: sensitivity to
underparameterized model misspecification. Psychol Methods 1998;
3:424 – 453.
Humbleton R, Merenda P, Spielberg C. Adapting Educational and Psychological
Tests for Cross-Cultural Assessment. New Jersey: Lawrence Erlbaum
Associates, Inc., Publishers, 2005.
Israel B, Checkoway B, Schulz A, Zimmerman M. Health education and
community empowerment: conceptualizing and measuring perceptions
of individual, organizational, and community control. Health Educ Q
1994;21:149 – 170.
Jordan C, Revenson T. Gender differences in coping with infertility: a
meta-analysis. J Behav Med 1999;22:341– 358.
Lieberman D. Evaluation of the stability and validity of participant samples
recruited over the internet. Cyberpsychol Behav 2008;11:743 – 745.
Little T, Cunningham W, Shahar G, Widaman K. To parcel or not to parcel:
exploring the question, weighing the merits. Struct Equat Modeling 2002;
9:151 – 173.
Melo C, Gameiro S, Canavarro MC. Does the FertiQoL assess quality of life?
Results from the validation of the Portuguese version of the FertiQoL. Hum
Reprod 2012;27:268 – 273.
Moura-Ramos M, Gameiro S, Canavarro MC, Soares I. Assessing infertility
stress: re-examining the factor structure of the fertility problem
inventory. Hum Reprod 2012;27:496 –505.
Pedro J, Canavarro MC, Boivin J, Gameiro S. Positive experiences of
patient-centred care are associated with intentions to comply with
fertility treatment: findings from the validation of the portuguese version
of the PCQ-infertility tool. Hum Reprod 2013;28:2462 –2472.
Perkins D, Zimmerman M. Empowerment theory, research, and application.
Am J Commun Psychol 1995;23:569 – 579.
Rauprich O, Berns E, Vollmann J. Information provision and decision-making
in assisted reproduction treatment: results from a survey in Germany. Hum
Reprod 2011;26:2382 – 2391.
Seligman M. Helplessness: on Depression, Development and Death. New York:
W. H. Freeman and Company, 1975.
Sheeran P. Intention-behavior relations: a conceptual and empirical review.
Eur Rev Soc Psychol 2002;12:1 – 36.
302
Slade P, Emery J, Lieberman B. A prospective, longitudinal study of emotions
and relationships in in-vitro fertilization treatment. Hum Reprod 1997;
12:183 – 190.
Slade P, O’Neill C, Simpson AJ, Lashen H. The relationship between
perceived stigma, disclosure patterns, support and distress in new
attendees at an infertility clinic. Hum Reprod 2007;22:2309 – 2317.
Spielberger C. Manual for the State-Trait Anxiety Scale. Palo Alto: Consulting
Psychologists Press, 1983.
van Dam-Baggen R, Kraaimaat F. De inventarisatielijst sociale betrokkenheid
(ISB): een zelfbeoordelingslijst om sociale steun te meten [The
inventarisation inventory to measure sociale integration: a self report
inventory to assess social support]. Gedragstherapie 1992;25:27 – 45.
Van Dongen AJC, Kremer JAM, Van Sluisveld N, Verhaak CM, Nelen WL.
Feasibility of screening patients for emotional risk factors before in vitro
Lopes et al.
fertilization in daily clinical practice: a process evaluation. Hum Reprod
2012;27:3493 – 3501.
Verhaak C, Smeenk J, Evers A, van Minnen A, Kraaimaat FW. Predicting
emotional response to unsuccessful fertility treatment: a prospective
study. Journal of Behavioral Medicine 2005a;28:181 – 190.
Verhaak C, Smeenk J, van Minnen A, Kremer J, Kraaimaat FW. A longitudinal,
prospective study on emotional adjustment before, during and after
consecutive fertility treatment cycles. Hum Reprod 2005b;20:2253– 2260.
Verhaak C, Lintsen A, Evers A, Braat D. Who is at risk of emotional problems
and how do you know? Screening of women going for IVF treatment. Hum
Reprod 2010;25:1234 – 1240.
Wischmann T, Scherg H, Strowitzki T, Verres R. Psychosocial characteristics
of women and men attending infertility counselling. Hum Reprod 2009;
24:378 – 385.