Table 1. Different ways of expressing the same data about risk. The use of risk Example An unwanted event which may or may not occur.[28–30] Non-Dutch women and women aged 35-39 years have a risk of perinatal morbidity and mortality. The cause of an unwanted event which may or may not occur. [28–30] Maternal non-Dutch ethnicity and aged 35-39 years are risk factors for morbidity and mortality. The statistical expectation value of an unwanted event which may or may not occur.[28–30] The risk of fetal and neonatal mortality for non– Dutch, 35-39 year old women is higher than for Dutch, 35-39 year old women. The probability of an unwanted event which may or may not occur, presented as:[28–30] − 0.21%; − Two times higher than Dutch woman in the same age category; − One additional perinatal death for every 1000 births among Dutch woman in the same age category.[31,32] Absolute risk; Relative risk; Attributable risk. Furthermore, risk is generally overestimated,[33] due to the framing of risks.[34,35] Studies show that the language used to express probabilities is interpreted in different ways.[36,37] However, verbal descriptors are more sensitive to manipulation of context and framing,[38] generally leading to an overestimation of risks compared to numerical information.[39,40] Morgan et al. (2020) argue that the term ‘risk’ should be replaced with terms such as chance, ‘likelihood’, or ‘probability of harms’.[41] Visual formats help increase the understanding of numerical risk information.[42–44] Besides being generally overestimated, risk is also subjective in its acceptability.[45] Generally, a risk of less than0.1% is considered acceptable, essentially conceived as zero risk. [28,46–50] This threshold is arbitrary, resulting from legal and regulatory considerations, and can be traced back to the determination of ‘safety’ levels in carcinogenicity testing. [51,52] Also, risk is historically shaped. What constitutes as a risk today may not necessarily be viewed in the same light tomorrow.[53] This is apparent in the changes in the list of medical indications for referral between primary midwife-led and obstetrician-led care in the hospital.[54] The number of indications in the list has almost doubled between 1987 and 2003, from 81 indications to 143 indications. Based on advanced knowledge some indications were added, such as infectious diseases, and others were removed, such as pelvic abnormalities. Notably, some emerging risks, such as maternal age,[55] were previously removed from the list.[54] Thus, the understanding of risk is ambiguous, meaning different things to different people at different times. The role of professional beliefs and interests in risk selection The variety in interpretation of risk is influenced by professionals’ beliefs regarding risks associated with pregnancy and birth as well as concerns with protecting professional interests. 164 7 CHAPTER 7