Key words psychosocial intervention - ART - drop-out predictors - APIM model - risk score for early drop out
Introduction
Assisted reproductive technologies (ART) often represent the method of choice for couples struggling with infertility but also often represent a significant emotional burden [1 ]. Above all, patients generally experience the 14-day waiting time between the embryo transfer and the subsequent pregnancy test (PT) to be particularly emotionally burdensome [2 ], [3 ]. This is reflected in particular in somewhat increased depression and anxiety values as well as more frequently occurring somatic symptoms, as Wischmann et al. [4 ] discovered in comparison to the norm sample.
Until recently, it was assumed that IVF patients – at least those whose health insurance companies covered 100% of the costs – continue their treatment until a pregnancy is achieved or they are advised to end the treatment due to an unfavourable medical prognosis (so-called “active censoring”) [5 ]. According to a current review, however, discontinuing assisted reproductive treatment is triggered primarily through factors relating to the psychological burden (especially personal problems and relationship conflicts) [6 ]. Nearly 30% of couples drop out of treatment early, primarily for psychological reasons [7 ], [8 ], [9 ], [10 ], although there is a good prognosis for a pregnancy as well as cost assumption by the health insurance company [11 ].
Studies throughout Europe found rates of between 17 and 70% for drop-outs from assisted reproductive treatment [7 ], [8 ], [12 ], [13 ], [14 ], [15 ]. Early discontinuation of the treatment is associated with 15% lower pregnancy rates [11 ]. Various studies indicate that a reduction in stress for patients during treatment is useful for increasing adherence to the treatment and thus the prospects of IVF treatment [16 ], [17 ].
In addition to support for patients to help them achieve their goal of parenthood, an optimized treatment programme also increases the financial benefits for the clinic providing the treatment. This is reflected in increased patient compliance (i.e. patient retention) or increased rates of successful pregnancies [5 ]. If stressed patients are offered psychosocial support early on, this could lead to increased health-related behaviour during the treatment, along with lower drop-out rates [18 ]. According to one study [5 ], preventing early termination of the treatment should therefore be a priority for the entire clinic staff.
Psychological stress during the treatment can be buffered by “positive reappraisal coping intervention” (PRCI) in which the situation is redefined and thus seen in a more positive light [5 ]. This short-term coping strategy was specially developed for the waiting time following embryo transfer prior to the pregnancy test during infertility treatment [2 ]. Folkman and Lazarus [19 ] assumed that positive emotions take on an important role for motivating people to face stressful situations and be able to better cope with stress. With the aid of a psychosocial intervention, it may be possible for a reproductive medicine treatment to be continued, despite the ups and downs, and ideally be successfully completed. In an ideal case, patient satisfaction increases, but so do the chances of a pregnancy.
The PRCI has proven itself so far to be acceptable and practicable for women undergoing infertility treatment [2 ]. In one study [3 ], women who read a PRCI card with a positive message twice per day perceived this to be supportive for their coping process in this stressful situation. The patients also considered the positive messages to be helpful in viewing the stressful situation from a more optimistic viewpoint [20 ]. The effect of PRCI on couples was not examined to date and is currently being investigated within the scope of the PACI study (“Positive Ausrichtung bei unerfülltem Kinderwunsch” [Positive adjustment in infertility]).
For some samples, including the sample of this PACI study, there are to some extent significantly high rates of study subjects dropping out of psychosocial interventions. In an investigation by de Klerk [21 ], patients who left a psychosocial intervention study early were compared with those who completed the study. This investigation found no differences between the two groups with regard to the depression and anxiety values which were measured before the IVF treatment. However, it is possible that the women who dropped out early experienced greater stress during the IVF treatment and thus they discontinued the study early.
Another study found significant differences between the various study groups with regard to their drop-out rates [22 ]. While the intervention group with cognitive-behavioural support had a drop-out rate of 16%, 26% dropped out early from the standardised support in the second intervention group. There was a significantly higher drop-out rate of 60% in the control group treated under routine conditions.
Following on this, it should be clarified which mental stressors increase the probability of early drop-outs and how these can be counteracted. With regard to the drop-out rates, we therefore examined the question: “Can psychological predictors for the early termination of a psychological intervention during assisted reproductive treatment be found?” The objective of the study was not to identify predictors for drop-outs from assisted reproductive treatment.
Material and Methods
Sample
The present data from the prospective randomised-controlled pre-post PACI intervention study (RCT) date from the period from August 2017 up to and including April 2018 (interim assessment).
Patient couples who underwent an IVF or ICSI treatment during the above time period in the outpatient infertility unit at the Heidelberg University Hospital of Gynaecology and Obstetrics were invited to take part in the study by the attending physicians.
The inclusion criteria for the sample were consent to participate in the study from IVF and ICSI patients during the above time period, owning a smartphone, and provision of the mobile telephone number. If the partner did not wish to participate, there was an option to be included as an individual person. Exclusion criteria were refusal to participate and/or a lack of German skills and/or lack of a smartphone.
After informed consent was given, computerised randomisation was performed immediately, in pairs, into the intervention group (IG) or the control group (CG). The study participants are pseudonymised by code numbers. The study has two arms and is not blinded for the patients.
Psychological assessment
The validated ScreenIVF questionnaire was selected as an evaluation instrument [18 ]. This questionnaire is based on a prospective Dutch study [23 ] whose objective was to document the emotional risk profile of patients during an assisted reproductive technology measure. In this PACI study, the ScreenIVF was used to identify drop-out predictors and also as a measurement of the effects of the intervention (pre-post effects).
The developers of the ScreenIVF [1 ], [23 ] postulated five “risk factors”: Anxiety, depression, limited social support, helplessness and a lack of acceptance of the situation. The questionnaire has a multiple-choice format with four-point Likert scales. It documents the patientsʼ anxiety with ten items which are based on the Spielberger “State and Trait Anxiety Inventory” [25 ]. The seven items on depression come from the Becks Depression Inventory [26 ]. With the Illness Cognition Questionnaire for IVF patients, helplessness is documented with six items and acceptance with six items with regard to fertility [1 ], [27 ]. The perception of social support is documented using five items from the Inventory for Social Integration [28 ].
Course of the study
Study subject couples who wished to take part (opt-in method) received information material on the study and a brief introduction to the techniques. During this introduction, the methods (creative distraction for the CG and positive re-evaluation for the IG; see below) were briefly explained. In addition, the participants (both partners) received an informed consent form to be signed. The course of the study comprised three measurement time points (see [Fig. 1 ]).
Fig. 1 PACI flow chart.
At measurement time point 1 (T1: pre-intervention) a paper-and-pencil questionnaire was used to document sociodemographic data (age, sex, education level, occupation, duration of the coupleʼs relationship) and fertility-related data (duration of infertility, number of previous children, duration of the ART treatment, suspected cause of the infertility). The ScreenIVF was filled out at T1 using the paper-and-pencil method and at T2 (post-intervention) online by the participants. At T3 (evaluation), four weeks after the pregnancy test, the study subjects were sent a link to the Unipark online survey platform via SMS. Here, they were asked about the efficacy and practicability of the online intervention (see [Fig. 1 ]).
If there was no online participation at T2 and T3, the study subjects received a text message with a reminder about the online participation after seven and then after fourteen days in each case. All online data were saved on the Unipark server in pseudonymised form; otherwise data were stored in encrypted form via the Heidelberg University Hospital. The data were evaluated by the Institute of Medical Psychology Heidelberg.
Intervention
After the couples were randomised and the assisted reproductive technology treatment cycle had started, the computer-aided thirteen-day SMS messaging began in parallel. The IG received the “positive adjustment technique” (PACI) described above [2 ], [29 ]. Starting at the time of the oocyte retrieval, these study subjects received a sentence on their smartphone every midday which was intended to generate a positive attitude, for example “Today Iʼm doing something good for me” or “I am trying to think about the positive things in life more.” By contrast, the CG couples received mental exercises for “creative distraction” on their smartphone every day. For example, one mental exercise involved completing a sequence of numbers.
The study was reviewed and approved by the Ethics Committee of the Medical Faculty of Heidelberg (study S-074/2017). All study details can be viewed at clinicaltrials.org under NCT03118219 and in [29 ].
Data analysis and statistical evaluation
The sociodemographic and fertility-related data as well as the five scales of the ScreenIVF were initially descriptively analysed in this interim assessment.
An early drop-out is defined as not answering the questionnaire at the second and/or the third measurement time point. It was checked whether specific stress indicators (or the cumulative risk score; see below) could be used as possible predictors for a premature drop-out.
Using the Actor–Partner Interdependence Model (APIM) [30 ] it was investigated whether there are connections within a dyad between possible predictors and the drop-out (also taking the influence of the partner on the actor into account). In each case, an APIM was calculated at T1 and T2 for the selected stress indicators. Since this involved distinct dyads, a separate equation was included in the path analytic approach for each part of the dyad and an actor effect (a) and a partner effect (p) was calculated in each case [31 ].
The determination of a risk score as a possible predictor for an early drop-out is derived from [21 ]: Patients are classified as at-risk if they demonstrate clinically relevant values above the respective cut-off level on the five scales of the ScreenIVF and thus have a greater risk of relevant emotional stress [18 ]. The score ranges from 0 – 5 points, whereby 0 means that there is no risk factor, while 5 means that the cut-offs of all five scales are exceeded (see also [24 ]).
The analysis was performed with the statistics software R, version 3.5.0 [32 ] and with the lavaan package, version 0.6.1 [33 ].
Results
Sociodemographic and fertility-related data
In the sample at measurement time point T1 in this interim assessment, there were n = 141 subjects, 60 of whom were men (42.55%) and 81 of whom were women (57.45%) (see [Table 1 ]). The age range was between 27 and 51 years (M = 37.21; S = 4.79).
Table 1 Sociodemographic data.
Frequency
Age distribution
n
%
M
SD
Min
Max
Sex distribution at T1
81
57.45
36.62
4.01
27
44
60
42.55
38.04
5.62
28
51
141
100
37.21
4.79
27
51
Fertility-related data in years at T1
n
M
SD
Min
Max
138
4.88
3.09
1.00
18
138
9.46
4.83
2.00
19
136
2.60
1.96
0.16
9
Does the couple already have children?
n
%
97
70.80
40
29.20
On average, infertility had lasted for 4.88 years and the medical treatment lasted from 2 months to 9 years (M = 2.6 years, SD = 1.96) (see [Table 1 ]).
Despite reminders, some patients in the PACI investigation did not respond to the questionnaire link sent via SMS and were thus counted among the drop-outs defined here. At T2, there were still 104 study subjects in the study, which corresponds to a drop-out rate of 26.24% (see [Table 2 ]). At the time of the evaluation (T3), 90 patients were still in the study, which indicates a drop-out rate of 13.46% and this corresponds to a total drop-out of 36.17% from T1 to T3 (see [Table 2 ]).
Table 2 Sex distribution and drop-out values at T2 and T3.
Frequency
Age distribution
Drop-out
Drop-out at T2
n
%
M
SD
Min
Max
Number of drop-out persons T1 to T2
Drop-out in %
67
64.42
36.19
3.85
27
44
14
17.28
37
35.58
38.14
5.61
29
51
23
38.33
104
100
36.89
4.63
27
51
37
26.24
Drop-out at T3
n
%
M
SD
Min
Max
Number of drop-out persons T1 to T3
Drop-out in %
Number of drop-out persons T2 to T3
Drop-out in %
57
63.33
36.31
3.74
28
44
24
29.63
10
14.93
33
36.67
38.55
5.60
29
51
27
45.00
4
10.81
90
100
37.16
4.63
28
51
51
36.17
14
13.46
It was investigated whether there is a difference between men and women with regard to the likelihood of early drop-out. The probability of dropping out at T2 was significantly higher for men than for women (b = 1.50, SE = 0.57, z = − 2.63, p = 0.008), the odds ratio was 4.50. At T3, a significant difference between women and men was also seen with regard to the probability of a drop-out (b = 1.21, SE = 0.56, z = − 2.15, p = 0.032): For men, the probability was once again significantly higher and the odds ratio was 3.35.
Descriptive statistics of the stress indicators
Overall, the study subjects did not reach the cut-offs of the scales or the clinically abnormal values, on average, in the case of the selected stress indicators. Thus the mean value of the female study subjects for depression at T1 was M = 2.23 (SD = 2.44) and in the case of the male study subjects, it was M = 1.10 (SD = 1.74) (see [Table 3 ]). The cut-off of the short form of the BDI is at four points [26 ]; a score at or above this value corresponds to a mild depressive syndrome.
Table 3 Stress indicators of the study subjects over time, in a gender comparison.
Stress indicator
Sex
Measurement time point
n
M
SD
Min
Max
Depression
f
1
77
2.23
2.44
0
11
2
67
3.06
3.56
0
14
m
1
58
1.10
1.74
0
7
2
37
1.16
1.72
0
7
Anxiety
f
1
75
20.97
5.64
10
33
2
67
21.91
6.43
11
37
m
1
57
18.44
4.63
10
30
2
37
17.78
5.74
10
32
Social support
f
1
78
17.04
3.28
7
20
2
67
16.78
3.28
9
20
m
1
60
17.33
3.32
8
20
2
37
17.73
2.71
11
20
Helplessness
f
1
77
11.58
4.38
6
24
2
67
12.24
4.33
6
24
m
1
59
10.00
3.09
6
17
2
37
9.73
2.95
6
16
Acceptance
f
1
74
14.04
4.05
6
24
2
67
13.00
4.34
6
24
m
1
59
16.00
3.35
10
24
2
37
16.00
4.38
7
24
Risk score
f
1
71
1.38
1.46
0
5
2
67
1.87
1.67
0
5
m
1
54
0.81
1.10
0
5
2
37
0.78
1.42
0
5
For the anxiety score as well (cut-off ≥ 24), the patients were, on average, below the cut-off value at T1 and T2: For women, the mean value at T1 was M = 20.97 (SD = 5.64) and for men, it was M = 18.44 (SD = 4.63). For the other stress indicators, the sample also demonstrates unremarkable values, on average (see [Table 3 ]). The study subjects achieved a risk score on average between 0.78 and 1.87 points.
Stress indicators within the scope of the actor–partner interdependence model (APIM)
In our interim assessment, a significant actor effect at T1 with regard to anxiety could be found: The higher a manʼs anxiety level was, the more likely he also was to drop out early (β = 0.41, p = 0.032). In addition, a significant connection between the drop-out of men and women is seen (β = 0.43, p = 0.020) (see [Fig. 2 ]). At T2, significant partner effects could be found: The higher a manʼs anxiety level was, the greater the likelihood of the woman dropping out early (β = 0.54, p = 0.001). The higher a womanʼs anxiety level was, the lower the likelihood of a drop-out for the man (β = − 0.38, p = 0.014). Here as well, there is a significant connection between the drop-out of men and women overall (β = 0.61, p = 0.061).
Fig. 2 Actor–partner interdependence model for predictor of anxiety at T1.
A significant connection was seen between the social support experienced by men and the drop-out of the women at T1: The more social support the man perceived, the less likely his partner was to drop out early (β = − 0.44, p = 0.016) (see [Fig. 3 ]).
Fig. 3 Actor–partner interdependence model for predictor of social support at T1.
A significant partner effect of the men on the women could be found for the stress indicator of helplessness at T1 (β = − 0.57, p = 0.036): If the man indicates increased helplessness, the probability of the woman dropping out early was significantly lower (see [Fig. 4 ]).
Fig. 4 Actor–partner interdependence model for predictor of helplessness at T1.
No significant actor and partner effects were able to be determined for the cumulative risk score, which is why, for the present sample, it is not suitable as a predictor for an early drop-out.
Discussion
The drop-out rate of a total of 36.17% between the first and last measurement time point shows the need to focus on patientsʼ compliance within psychological interventions in clinical studies.
Domar et al. [22 ] found significantly higher drop-out values in a control group under routine conditions in comparison to a group with cognitive-behavioural interventions. This result allows us to conclude that adherence in studies can be promoted through targeted psychological interventions.
Why do drop-outs take place in the study? Is it primarily the stresses of medical treatment or a lack of efficacy of the psychosocial intervention which the subjects experience? Or is it possibly a combination of the two?
The descriptive observation of these data shows that the sample is, on average, rather non-stressed: In our interim assessment, the study subjects on average do not demonstrate any unusual values with regard to stress indicators. By contrast, other studies were able to show that emotional stressors are the main reasons for female study subjects to drop out early from assisted reproductive treatments [7 ], [14 ]. The most frequent reasons which were given during an investigation of female patients were “I had enough” (66%), “emotional costs” (64%) and “I could not cope with more treatment” (42%) [10 ].
Verhaak et al. [18 ] found no differences in women who dropped out of their study in comparison to those who completed the study. For this, the authors [18 ] had recorded the result of the pregnancy test as well as the depression and anxiety scores of the female study subjects. In our investigation, we did not obtain any information about the reason why study subjects drop out earlier. In the future, the reason for an early drop-out should be documented; this would undoubtedly add value to further drop-out analyses.
Certain coping strategies (such as meaningful coping) make it possible for women and men to better endure stress during a treatment [20 ], [24 ], while other strategies (such as active-avoidant coping) conversely increase the risk of stress. It is worth investigating whether the positive effects of PACI which were found do not demonstrate an effect just in clinical studies (so-called “efficacy”) but also when patients use the method in “everyday life” without any clinical guidance (so-called “effectiveness”).
The results of our study indicate a significant influence of the partner on a drop-out: Men who indicated more anxiety showed a higher drop-out probability. At T2, we found significant partner effects regarding anxiety: The higher the anxiety values, the greater the likelihood for the partner to drop out of the intervention early. The higher the manʼs social support was, the less likely it was that his partner ended the study. By contrast, increased helplessness of the man led to a decreased drop-out risk of the woman. To motivate couples to continue participation in a psychological intervention despite significant emotional stress, fears and needs of the male partner should accordingly be explicitly and actively addressed.
Limitations
In our interim assessment of the PACI study, it was difficult to define a drop-out of the study subjects and differentiate it from a drop-out from the ART treatment (e.g. due to treatment complications, fertilisation failure or non-occurrence of a pregnancy, financial or emotional stress due to ART, switch to a different infertility centre): Documentation as to whether study subjects dropped out only from the clinical intervention study or whether the ART treatment was discontinued at the same time is not available for this investigation.
A portion of the PACI study subjects could not be included after the start of the study. One criterion in this case was embryo transfer that did not take place (e.g. due to “embryonic arrest”). For technical reasons, the telephone number of the study participants could be registered only once and thus participation in the PACI study in a subsequent ART cycle was not possible. The second exclusion criterion was the delayed submission of the informed consent (following egg cell retrieval) which would have allowed participation in the PACI study.
In the PACI study, the recommended “opt-in method” was used for study subject recruitment. By contrast, in an “opt-out method”, all eligible persons are included in the study unless they explicitly state that they do not wish to participate. Junghans et al. [34 ] discovered that study subjects in the opt-in method have a lower risk (44%) of various medical illnesses than study subjects in the opt-out method (60%). Various study findings show that opt-in study subjects performed better overall with regard to clinical indicators. It is suspected that opt-in study subjects can endure the additional demands of a study better [34 ], [35 ]. Overall, the rejection rate of patients with regard to the PACI study can be classified as rather low, according to our interim assessment. However, it may be that, in particular, patients who were less stressed overall registered for the study.
Since the study subject recruitment took place only in one hospital, it remains unclear if the findings can be transferred to couples throughout Germany and Europe and also worldwide. It also remains questionable whether the findings apply to other infertile couples who do not undergo medical treatment at all. Only a certain percentage of couples in Germany undergo diagnostic testing to investigate the reasons for infertility. According to the study by Wippermann [36 ], approx. 63% of couples do not undergo any infertility treatment, although they have received a diagnosis for reduced fertility and frequently are fundamentally ready to use ART. Other estimates [37 ], [38 ], [39 ] assume that up to 30% of couples do not start ART, despite having an indication and despite (proportional) cost coverage by the health insurance company.
Strengths
The innovative intervention via the smartphone can be considered to be a strength of the PACI study. PRCI was previously sent via positive phrases on paper cards [2 ] or as a digital file (via e-mail) [5 ]. The smartphone, which has become an indispensible part of peopleʼs everyday lives nowadays, is a useful medium for receiving a positive message every day and being able to access it at any time, as needed.
Men frequently react in other ways to the diagnosis of infertility than women and the emotional stress can manifest differently [40 ]. Numerous studies have found that men are also psychologically affected by the diagnosis of infertility (e.g. [41 ], [42 ]). However, many studies to date have neglected the emotional influences of fertility problems on men [1 ], [23 ]. Many psychological support measures also focus exclusively on women. If the drop-out values for men and women are compared, men in this study have a significantly higher likelihood of dropping out of the psychosocial intervention early. At T2, the men had a risk of ending the study early which was four and a half times as great as that of women, according to the odds ratio. At T3, men still have a nearly three times greater likelihood of a drop-out. Male infertility patients more frequently indicate that they believe they could manage their feelings very well alone [43 ]. For this reason, men are generally not addressed as the main target group for infertility counselling, although infertility represents a problem for the couple. This phenomenon could be an expression of the fact that, in an ART treatment, men tend to see themselves as an observer (or as the “female patientʼs companion”) and women see themselves more as the patient and as the focus of the treatment [43 ]. The inclusion of the man is therefore a perspective of this investigation which should be assessed positively, because it could also lead to men being able to play a more active part in the reproductive medicine treatment.
Implications for research
Theory and research imply that PACI can have positive effects on the fertility-related quality of life, especially during the phases in which patients experience increased stress and uncertainty (e.g. [44 ], [45 ], [46 ], [47 ]). The PACI program of our study is a “custom-made” intervention for the patients undergoing ART treatment in which all study subjects receive the same support. In the future, it would be of interest to test “tailor-made” interventions, that is, interventions which are individually tailored to study subjects. Many couples presumably feel that text messages are helpful while others prefer optional telephone or chat messages, and yet others do not want to utilise any additional aid at all. There may be a lower drop-out rate during a psychological intervention in the case of couples who receive a tailored intervention than in the case of couples during a general intervention. The different effects of psychosocial face-to-face therapy as compared to online interventions should also be further investigated. Likewise, future research questions should further focus on the different needs and the possibly different effect of PACI (or other psychosocial interventions) on men in comparison to women. From our point of view, the interim assessment of the PACI study shown here clearly indicates that future research in this area should always take the role of the male partner in reproductive medicine into account.
Implications for clinical practice
Certain coping strategies can increase the risk of depression and anxiety [24 ]. Psychological counselling could be helpful for patients to use supportive coping strategies and not use unfavourable strategies (such as active avoidance strategies, for example). The influence of partner effects on the psychological condition, as proven, for example, by APIM [24 ] and which, to some extent – particularly for the male partner – was also found in this interim assessment is a reason for orienting the psychosocial fertility counselling in a targeted manner to the couple (human reproduction takes more than one person) and offering individual counselling as an exception or only upon an explicit request. Overall, the results of our study indicate that couples receiving reproductive medicine treatment are, from a psychological viewpoint, a less vulnerable group, on average (see also [48 ]).
Conclusion
The objective of the study was to identify predictors for the early drop-out of study subjects from psychosocial intervention during ART treatment. In this interim assessment, only individual significant effects with regard to the stress indicators in connection with a drop-out were found. Additional psychological interventions for couples undergoing reproductive medicine treatment should be offered in order to support them and thus optimally utilise the opportunities for medical therapy. Over the long term, this can lead to increased patient satisfaction – and possibly also to higher cumulative pregnancy rates through better compliance with regard to the medical treatment [49 ]. The suggestions for improvement from patients queried in the study by Domar et al. [10 ] refer to this: An offer of psychological support in written form as well as the presence of specialist psychosocial counsellors were the two most frequently mentioned wishes regarding reproductive medicine, from the retrospective viewpoint of drop-outs. The predominant lack of such a psychosocial offering – also and explicitly to the male partner – may represent an important reason why nearly half of all couples in Germany already leave medical treatment after the second treatment cycle as drop-outs, without achieving a pregnancy, although the health insurance companies generally pay for three cycles (proportionally) [49 ]. Overall, the role of the male partner in the reproductive medicine treatment should explicitly be taken into more account [36 ].
Comment
MB, MS, SR, AG, TS, MM, SB, BD and TW have contributed to the design of the study. MS and TW were responsible for obtaining the ethical approval. MS, SR, AG, BD and TW were involved in adapting and compiling the interventions. SR, AG and TS were responsible for maintaining the recruitment process. MM and SB were responsible for maintaining the technical processes of the intervention. MB was responsible for drafting the manuscript. MS planned the statistical analysis. All authors have read, revised, and approved the final manuscript. This study is financed by intramural funding from the Institute of Medical Psychology and the Department of Gynecological Endocrinology and Fertility Disorders, University Womenʼs Hospital Heidelberg.
