Thus, I did a double-cringe when I first read the abstract for the following article:
Frederiksen, Farver-Vestergaard, Gronhoj Skovgard, Ingerslev, and Zacharaie (2015) Efficacy of psychosocial interventions for psychological and pregnancy outcomes in infertile women and men: a systematic review and meta-analysis--- which can be found here.
In this study, the authors conducted a meta-analysis of 39 different studies. The studies looked at the effect of psychosocial interventions on mental-health symptom reduction and clinical pregnancy rates in research subjects undergoing infertility treatment. A meta-analysis uses other clinical studies found in the literature as its source of data. Studies were only included in this meta-analysis if they studied infertile participants, presented data regarding some sort of psychosocial intervention program (such as cognitive-behavior therapy, group therapy, etc.), measured outcome variables such as stress, distress and pregnancy outcome, both before and after the psychosocial intervention occurred, and used a quantitative research method. The data for the 39 studies were analyzed for a number of variables. Interestingly, they found that women who had received any type of psychosocial intervention were twice as likely to become pregnant than women who had not received psychosocial treatment. Although they found that the psychosocial interventions were effective at decreasing both symptoms of anxiety and depression, only a reduction in anxiety levels was significantly correlated with an increased pregnancy rate.
On the face of it, this is a quite dramatic finding. The authors theorized that increased anxiety levels could have a negative effect on the uterine environment, and could thus negatively impact pregnancy rates. However, they also reported a number of limitations to their study that might also explain their results. For instance, there was no distinction made as to type of infertility diagnosis or type of infertility treatment received among the different studies. Obviously, the level of severity of the infertility diagnoses involved among the different studies would have a large impact on pregnancy rates, as would the types of medical treatments offered. Secondly, the authors of the meta-analysis had no access to the pregnancy outcomes for women who dropped out of the studies. Without this information, it is hard to obtain a complete picture of what really happened. In addition, the authors had to contend with "publication bias"--meaning that usually only studies with statistically significant results are likely to be published. Although information that disproves a hypothesis might be just as important from a scientific perspective, it often is not considered as desirable for publication, leading to the "drawer effect"--it gets left in a desk drawer and never sees the light of day again. For this reason, looking merely at published research can give a skewed perspective on an issue. The authors did attempt to find unpublished studies to include in their meta-analysis, but as you can imagine, this is a difficult task. Thus, the results of their meta-analysis might be skewed in a positive direction towards the positive effects of psychosocial intervention.
As a mental health clinician, I would love to believe that the psychosocial interventions I offer would double the pregnancy rates among my clients. Not only would it be great for business, but I would love it if I could have that kind of positive impact on helping people solve such a difficult problem. However, I have my doubts. To me, I think the key problem with this study is its inability to distinguish between the types of infertility diagnoses among research subjects. Imagine Patient A, who has a mild but correctible type of infertility diagnosis. She receives the appropriate medical care along with a psychosocial intervention. The psychosocial intervention helps her to better cope with her situation, but as she learns more about her diagnosis and her medical treatment commences, she feels less anxious because she is justifiably optimistic about her prognosis. And as things go according to plan, she becomes pregnant, and there is a nice, happy ending. Now imagine Patient B. Patient B has an unexplained infertility diagnosis with some indications of a larger, more serious problem (let's say her AMH and FSH levels are borderline, for example). Patient B might not get assigned a psychosocial intervention treatment due to luck of the draw. As her infertility treatment progresses, it doesn't go well, with failed cycles, and her prognosis looks poor. Naturally, she's not going to get pregnant during the study, because she has an underlying medical problem which prevents it, and she's also going to understandably get more anxious and distressed as she realizes what her situation truly entails. But we can be pretty sure that her anxiety isn't at the root of her problems getting pregnant. As you can see from these two examples, diagnosis and situation makes all the difference--but the study isn't able to pick that up.
Thus, I don't think this study really can tell us with any certainty that decreased anxiety levels are associated with increased pregnancy rates. I believe that psychosocial interventions, such as therapy, relaxation training, or support groups can be helpful during infertility treatment, in that they can provide support and coping tools during a difficult time. I think that is an incredibly valuable thing in and of itself. But I don't think that the data support the idea that a psychosocial intervention is going to increase pregnancy rates.
I also think we can all still be justified in getting annoyed the next time someone tells us to "just relax".
Thank you as always for reading, and I look forward to any comments and questions you may have!
In this study, the authors conducted a meta-analysis of 39 different studies. The studies looked at the effect of psychosocial interventions on mental-health symptom reduction and clinical pregnancy rates in research subjects undergoing infertility treatment. A meta-analysis uses other clinical studies found in the literature as its source of data. Studies were only included in this meta-analysis if they studied infertile participants, presented data regarding some sort of psychosocial intervention program (such as cognitive-behavior therapy, group therapy, etc.), measured outcome variables such as stress, distress and pregnancy outcome, both before and after the psychosocial intervention occurred, and used a quantitative research method. The data for the 39 studies were analyzed for a number of variables. Interestingly, they found that women who had received any type of psychosocial intervention were twice as likely to become pregnant than women who had not received psychosocial treatment. Although they found that the psychosocial interventions were effective at decreasing both symptoms of anxiety and depression, only a reduction in anxiety levels was significantly correlated with an increased pregnancy rate.
On the face of it, this is a quite dramatic finding. The authors theorized that increased anxiety levels could have a negative effect on the uterine environment, and could thus negatively impact pregnancy rates. However, they also reported a number of limitations to their study that might also explain their results. For instance, there was no distinction made as to type of infertility diagnosis or type of infertility treatment received among the different studies. Obviously, the level of severity of the infertility diagnoses involved among the different studies would have a large impact on pregnancy rates, as would the types of medical treatments offered. Secondly, the authors of the meta-analysis had no access to the pregnancy outcomes for women who dropped out of the studies. Without this information, it is hard to obtain a complete picture of what really happened. In addition, the authors had to contend with "publication bias"--meaning that usually only studies with statistically significant results are likely to be published. Although information that disproves a hypothesis might be just as important from a scientific perspective, it often is not considered as desirable for publication, leading to the "drawer effect"--it gets left in a desk drawer and never sees the light of day again. For this reason, looking merely at published research can give a skewed perspective on an issue. The authors did attempt to find unpublished studies to include in their meta-analysis, but as you can imagine, this is a difficult task. Thus, the results of their meta-analysis might be skewed in a positive direction towards the positive effects of psychosocial intervention.
As a mental health clinician, I would love to believe that the psychosocial interventions I offer would double the pregnancy rates among my clients. Not only would it be great for business, but I would love it if I could have that kind of positive impact on helping people solve such a difficult problem. However, I have my doubts. To me, I think the key problem with this study is its inability to distinguish between the types of infertility diagnoses among research subjects. Imagine Patient A, who has a mild but correctible type of infertility diagnosis. She receives the appropriate medical care along with a psychosocial intervention. The psychosocial intervention helps her to better cope with her situation, but as she learns more about her diagnosis and her medical treatment commences, she feels less anxious because she is justifiably optimistic about her prognosis. And as things go according to plan, she becomes pregnant, and there is a nice, happy ending. Now imagine Patient B. Patient B has an unexplained infertility diagnosis with some indications of a larger, more serious problem (let's say her AMH and FSH levels are borderline, for example). Patient B might not get assigned a psychosocial intervention treatment due to luck of the draw. As her infertility treatment progresses, it doesn't go well, with failed cycles, and her prognosis looks poor. Naturally, she's not going to get pregnant during the study, because she has an underlying medical problem which prevents it, and she's also going to understandably get more anxious and distressed as she realizes what her situation truly entails. But we can be pretty sure that her anxiety isn't at the root of her problems getting pregnant. As you can see from these two examples, diagnosis and situation makes all the difference--but the study isn't able to pick that up.
Thus, I don't think this study really can tell us with any certainty that decreased anxiety levels are associated with increased pregnancy rates. I believe that psychosocial interventions, such as therapy, relaxation training, or support groups can be helpful during infertility treatment, in that they can provide support and coping tools during a difficult time. I think that is an incredibly valuable thing in and of itself. But I don't think that the data support the idea that a psychosocial intervention is going to increase pregnancy rates.
I also think we can all still be justified in getting annoyed the next time someone tells us to "just relax".
Thank you as always for reading, and I look forward to any comments and questions you may have!