The definition of risk for women of AMA, similar to that of pregnant women of other ages [31
], had a broader scope than merely medical risks or physical challenges and was influenced by various social and personal characteristics. The findings described above elucidated that this definition also incorporates consideration of the extent of the woman’s support network, her ability to control situations, and whether she had an established relationship, a planned pregnancy, a flexible job or healthy lifestyle and behavior. These findings support the position that approaches to understanding perception of health risks should be comprehensive and broad [8
] and suggest that risk communication will benefit from including women's criteria for defining risk.
As is the case in previous research [32
], pregnancy at AMA was frequently acknowledged by women to be high risk due to an increased risk of genetic abnormalities; however, other medical risks associated with AMA received less consideration. In our study, the "information quests"[32
], in which women sought out a wide range of information, were commonly reported in the preconception period. After becoming pregnant, however, women mostly limited seeking information to avoid increased anxiety and were focused on risk reducing behavior instead. In contrast to findings of Carolan & Nelson (2007) who reported that their Australian participants realized they were at higher risk only after becoming pregnant and through communication with their health care providers, women in our study described risks associated with AMA as "common knowledge in society." This finding is consistent with the finding of another Canadian study [33
Findings elucidated that nulliparous women aged 35
years or older were not a homogenous group in their pregnancy risk assessment. The data described above suggest that the perception of pregnancy risk may be a result of interactions among several factors including physiological and psychological elements, characteristics of the experienced risk, and feedback from health care providers. Despite the fact that our respondents were mostly aware of risks associated with pregnancy at age 35
years or older, the majority of them did not consider themselves to be high risk. This apparent disconnect can be explained by the fact that people typically tend to rate their personal risk lower in comparison to general risk [34
]. This underestimation of personal risk or "unrealistic optimism" [35
], might be related to one's perceived control over a potential risk [34
]. A link has been identified between perceived control and risk perception in previous research [36
]. In our study, those participants who believed they had good control over their physical health also perceived a lower risk for their pregnancy. It was also noted that a woman's perceived control over her health may be important in engaging in various behaviors to maintain the balance between risk and health. To support this point, a relationship between perceived control and health service utilization has been documented in previous research [37
Previous prenatal loss was recently documented as a predictor of depression and anxiety in subsequent pregnancies, independent of other psychosocial and obstetric factors [38
]. Our findings highlighted that having a poor reproductive history contributes to increased perceived risk of pregnancy. Failure to become pregnant in initial attempts or previous loss of pregnancy may have threatened these women's beliefs in their abilities to manage their own health and may lead to increased anxiety and higher risk perception. This statement is supported by Campbell, Dunkel-Schetter, & Peplau (1991) who reported that infertility may be perceived as a threat to an infertile person's life's goals, contributing to a low perception of control [39
]. Based on our data, pregnancy complications could alter risk perception depending on the type of complication, its manageability, and its consequences for a woman's daily life. Alternatively, good physical health and engaging in healthy behaviors and lifestyles were perceived as risk alleviating factors. This finding echoes that of Gerend et al. (2004) who reported that personal health actions from the women's perspectives can reduce their risk.
Concerns about fetal health and well-being, particularly genetic abnormalities, were very common among our participants, as they have been found to be in other studies [16
]. Although having several fetal surveillance tests helped reassure women about their baby's health, as Baillie et al. (2000) have noted, for a number of women, feelings of anxiety may remain throughout the pregnancy [41
]. Hoping for a desirable pregnancy outcome despite having higher perceived risk and anxiety levels was evident in our data. This discrepancy may lead to feelings of uncertainty. Sun et al. (2008), in a qualitative study of women of AMA, reported similar ambivalent and conflicted feelings, characterized by apparent pleasure and hidden fear [42
]. This attitude has been referred as a “jubilant apprehension” by Yuan et al. (2000) that describes feelings of great joy and satisfaction, but also worry about childbirth outcomes [40
Feelings of anxiety were described vaguely by our participants and there were notable variations among participants in expressing and wording their anxiety. This attitude may make the detection of anxiety in these women a challenge. There is evidence that antenatal anxiety is very prevalent and can increase the odds of postnatal depression [43
]. Therefore, identification of women with anxiety is crucial so that effective interventions can be targeted appropriately. Healthcare providers should be aware that some women with high levels of anxiety tend to use different terms to communicate these feelings. Although creating a relaxed environment and establishing a non judgmental communication pattern in prenatal care visits may be beneficial for anxious women in disclosing their actual feelings about potential risks, there is growing evidence suggesting that prenatal screening should also include screening for both depression and anxiety [44
]. We support this perspective and believe that using a reliable screening tool to assess anxiety in pregnant women may be useful to identify women who would benefit from strategies to reduce anxiety. A recent study in Australia demonstrated that the anxiety subscale of the Edinburgh Postnatal Depression Scale might be a reliable measure to screen antenatal anxiety [45
Women experienced a decrease in their perceived risk with advancing gestational weeks, which could be due to women becoming adapted to the state of being pregnant or becoming more positive about the outcome of their pregnancy as it advanced. An uncomplicated pregnancy and favorable screening results may contribute to decreased perception of risk over the course of pregnancy as well. A decline in worry about the baby’s health from early pregnancy to the postpartum period has been reported in previous research [47
Consistent with research in other fields, our results demonstrated that pregnant women's personal or vicarious experience with a risk may increase the psychological availability of the risk and consequently, its perceived probability [48
]. In fact, being familiar with risk, through researching information or indirect experiences, may contribute to women’s risk perception. In the literature, the perceived characteristics of the risk such as prevalence, controllability, preventability, and seriousness along with the availability and representativeness heuristics are important elements in constructing risk perception [35
]. Our findings point to the predictability of risk as a critical characteristic of risk in which predicting the risk and expecting it influenced participants' risk perception. A risk that is expected may be perceived as less risky than an unexpected risk, suggesting that clear communication is essential to help women have realistic expectations about their individual risk. A qualitative study by Patterson (1993) demonstrated that an unexpected shift in health situation or pregnancy outcomes was identified as a high risk condition by pregnant women, while the expected changes were considered as no risk [51
]. One explanation is that anticipating the risk and being prepared to deal with it may increase a woman's perceived control and, consequently, decrease her risk perception.
While discrepancies between pregnant women’s and health care providers’ appraisals of risk have been documented in previous research [52
], what appears to be less emphasized is the influence of health care providers' attitudes towards the risk on pregnant women's risk perception. Several women in this study reported not having any risk communication with their health care providers that focused on age as a risk factor. This has been interpreted by most participants as there not being any serious concern and may imply that pregnant women trust their health care provider’s opinion. In this regard, Heaman et al. (2004) reported that women with or without pregnancy complications rely on their health care providers in assessing risk status. A few participants reported negative risk communication with their health care providers. These participants also had higher anxiety and concern about the well-being of their fetus and pregnancy outcomes. Whether risk communication patterns can increase the anxiety or whether anxiety itself will alter women's interpretation of risk communication is not clear and needs further research. In literature, a link between maternal depression and higher perception of teratogenic risk has been reported [55
In the experience of our participants, any emphasis on the mother’s good health by their care providers was described as reassuring. Conversely, negative messages about age by purely emphasizing pregnancy risks associated with AMA were described as very destructive and challenging, particularly for women with high levels of anxiety. These findings indicate that although offering risk information is part of the risk communication process, the woman’s mental health should be considered to avoid unnecessary stress.
Comparison was a common risk appraisal strategy among our participants. However, women’s evaluations were distinct from those of health professionals in that the participants often compared their risk to a known population such as their family members, friends, or stories from real people, which is entirely different from risks determined by epidemiological studies using population data. This disjuncture stems at least in part from the fact that pregnant women’s understanding of risk is mostly based on their life experiences [31
], suggesting that real stories can alter risk perception. This finding has important implications for practice and public health education. Risk must be tangible in order to be recognized and potentially addressed by women. Interventions to decrease unhealthy behaviors should target this concept to alter risk perception. As Williamson & Weyman (2005) suggested when individuals have less experience or knowledge of risks, the media (e.g., documentary movies) can play an important role in increasing their understanding of those risks [57
One of the strengths of the study was the addition of quantitative measures (i.e., perception of pregnancy risk, anxiety, and perceived control) and the inclusion of a variety of perspectives from women of AMA with varying levels of risk perception, anxiety and perceived control. This diversity in the sample allowed for the documentation of variations in risk appraisal and also the identification of important issues that were common across participants. Another major strength was the contemporaneous exploration of risk perception during the pregnancy. It is important to consider the following limitations when interpreting and applying the results. First, our sample was representative of a middle class and married population; these characteristics may limit the generalization of the results. This is a small study based on 15 interviews in one geographical area in Canada; therefore, the results may not be generalizable to other populations. Finally, we recognize that selection bias may have been introduced as a purposeful sampling method was employed and those who volunteered to participate might have been different from those who did not.