European Journal of Clinical Pharmacology

, Volume 66, Issue 2, pp 207–214

Perception of risk regarding the use of medications and other exposures during pregnancy


    • Department of Pharmacy, School of PharmacyUniversity of Oslo
    • Division of Mental HealthNorwegian Institute of Public Health
  • Eivind Ystrøm
    • Division of Mental HealthNorwegian Institute of Public Health
  • Adrienne Einarson
    • The Motherisk ProgramThe Hospital for Sick Children
Pharmacoepidemiology and Prescription

DOI: 10.1007/s00228-009-0744-2

Cite this article as:
Nordeng, H., Ystrøm, E. & Einarson, A. Eur J Clin Pharmacol (2010) 66: 207. doi:10.1007/s00228-009-0744-2



Perception of risk may impact a woman’s decision to take a needed drug during pregnancy. There is a paucity of research on this topic in the literature.


(1) To evaluate the perception of risk of 17 commonly used drugs and other substances by pregnant women. (2) To investigate which sources of information regarding exposures during pregnancy were most commonly used by women.


A questionnaire was developed through the University of Oslo’s website for Internet surveys and posted on four Web pages used by pregnant women and mothers, from mid-September 2008 through October 2008. The inclusion criteria included women who were (1) pregnant or 2) a mother of a child less than 5 years old.


A total of 1,793 eligible women completed the questionnaire. Most women overestimated the teratogenic risk associated with all the drugs during pregnancy. Characteristics of the women that were associated with a high perception of risk were primiparity, higher age, higher education, and choosing not to use a drug during pregnancy. More than 80% of the women had used drugs during pregnancy, mostly paracetamol, penicillins and reflux medications. The physician, the product information leaflet and the pharmacist were the three most frequently used sources of information.


Women overestimate the risk of drug use and other exposures during pregnancy. Therefore, it is important for health care providers to use evidence-based information, to reduce unnecessary anxiety, and to ensure safe and appropriate treatment during pregnancy.


Risk perceptionPregnancyMedication


Perception of risk by pregnant and breast-feeding women will impact their decision on whether to use a drug or not especially when using over-the-counter (OTC) medications. Regarding prescribed drugs, a woman’s perception of risk may affect adherence to the physician’s prescription. In addition, since 50% of pregnancies are unplanned [1], women are frequently exposed to drugs prior to awareness of pregnancy, which can cause even more anxiety because of the unintentional nature of the exposure.

Few studies have focused on risk perception among pregnant and breast-feeding women of commonly used drugs. In one study by Sanz et al., the risks associated with 14 specifically listed medications were perceived to be higher than the actual risk for all drugs both by pregnant women (n = 81), non-pregnant women (n = 63), medical students (n = 256) and medical professionals (n = 25). They concluded that the overestimated risk perception among women and health care personnel may lead to abortion of otherwise wanted and healthy infants [2]. Researchers at The Motherisk Program in Toronto, Canada, have also reported that women overestimate the risk for drugs and that evidence-based counselling can lower a pregnant woman’s fear of using prescribed drugs [35]. In one study of 100 women using antidepressants during pregnancy, 87% of the women believed that antidepressants increased the risk of congenital malformations before they were given reassuring information, compared with 12% after counselling [6]. Not all countries provide teratology information services (TIS) to counsel the public [7, 8], and in these countries, such as Norway, pregnant and breastfeeding women have to use other sources of information. Previous studies have documented that besides TIS, the most widely used sources of information about pregnancy-related exposures were physician, pharmacist, midwife, family and friends, books and magazines, and the Internet [911]. In a recently published Swedish study, 84% of pregnant women used the Internet to obtain pregnancy-related information [12]. They did, however, to a lesser degree discuss this information with their midwife.

Regarding foetal exposures other than pharmaceuticals (e.g. food and substances), even less is known regarding how pregnant women perceive the risk of the use of such substances and exposures. In one study, 30 lay persons believed the teratogenic risk of cocaine to be over 50% [13], while the actual risk is probably much less [14]. In a second study, 39 pregnant women exposed to ionizing radiation estimated the risk of major malformation following radiation to be approximately 25% [15]. Evidenced-based counselling reduced risk perception to 17%, which was still substantially higher than the estimated actual risk (i.e. <5%).

The primary aim of this study was to evaluate the perception of risk of well-known drugs and other exposures to which pregnant and breast-feeding women are frequently exposed, as well as to investigate whether perception of risk was associated with socio-demographic characteristics and affected by individual drug use during pregnancy. Our secondary aim was to investigate which sources of information regarding the safety of these exposures are commonly used by pregnant women.

Materials and methods

Study design and population

This study was an anonymous self-completed questionnaire. Inclusion criteria were women who either were (1) currently pregnant or (2) a mother of a child less than 5 years old.

Data collection

A questionnaire was developed and attached to the University of Oslo’s Web site for Internet surveys ( [16]. An invitation to the study was posted on the following four Web pages for pregnant women and mothers: (can be translated as “”), (“”), (“”) and (“”). These Web sites are edited by a midwife and a staff of health care professionals. The questionnaire was accessible during a period of 5 weeks, from 16 September to 25 October 2008.

Measurement of risk perception

To estimate the baseline perception of risk, women were asked the following open-ended question “In a healthy population of pregnant women in a healthy environment, how many infants do you think are born with a congenital malformation?”. The true baseline risk was defined as a risk of malformations at ≤5% [14].

Numeric rating scales ranging from 0 (‘no risk to the foetus’) to 10 (‘foetal malformation following each exposure’) were used to evaluate the perception of teratogenic risk of 17 drugs, foods, chemicals and radiation. The classes of drugs were paracetamol (acetaminophen), penicillins, drugs used for treating heartburn, drugs used to treat nausea and vomiting of pregnancy (NVP), antidepressants, sedatives/anxiolytics and thalidomide. For each medication, the most common trade names were given since these were more likely to be known to the women. The other exposures were herbal drugs, foods [eggs, blue-veined cheese, fermented fish (traditional Norwegian dish)], lifestyle factors (cigarette smoking, alcohol at time of conception, alcohol during the first trimester), x-ray at the dentist, household cleaning chemicals and hair dye.

All exposures, with the exception of thalidomide (risk of congenital malformations 10–40%), were considered to have a ≤5% risk of congenital malformation. The other substances that were considered to be risky in pregnancy were nicotine (risk of growth restriction), alcohol (risk of foetal alcohol defects, lower limit of intake unknown), blue-veined cheese, and fermented fish (generally not recommended to pregnant women due to a small risk of listeria-infection for the foetus).

Socio-demographic and lifestyle variables

The socio-demographic and lifestyle factors included in the questionnaire were age, marital status, parity, education level, occupation at time of conception, geographic area of residency, smoking, and alcohol use prior to and during pregnancy, and pregnancy or motherhood status at the time the questionnaire was completed. If she had delivered a live baby, she was asked how old the infant was and whether she breast-fed. The variables were classified as presented in Table 1.
Table 1

Socio-demographic characteristics (n = 1,793)


Number (%)

Status of the woman


866 (48.3)


927 (51.7)

Age (years)





Marital status


1,695 (94.5)


98 (5.5)



689 (38.4)

 One or more previous children

1,104 (61.6)


 Basic school level

88 (4.9)

 Upper secondary education

390 (21.8)

 Tertiary education, shorta

810 (45.2)

 Tertiary education, longa

421 (23.5)

 Other education

84 (4.7)


 Paid work, not health-related

1,016 (56.7)

 Health-related paid work

426 (23.8)


77 (4.3)


229 (12.8)


45 (2.5)

Smoking status in pregnancy


1,604 (89.4)

 Yes, but less than before

177 (9.9)

 Yes, the same as before

12 (0.7)

Alcohol consumption after known pregnancy


1,626 (90.7)


148 (8.3)

 Can’t remember

19 (1.1)

aTertiary education, short comprises higher education up to 4 years in duration. Tertiary education, long comprises higher education more than 4 years in duration

Medication use during pregnancy and information requirements

To assess the woman’s own use of medication during pregnancy, she was specifically asked whether she had used paracetamol, anti-inflammatory agents (NSAIDs), penicillins, sedatives/anxiolytics, antidepressants, medication used to treat ‘nausea, vomiting of pregnancy’ (NVP), herbal drugs and other miscellaneous drugs. Iron and vitamins and timing of drug use were not included. For each group of medication, the most common trade names were given. If she had used other drugs than those on the list during pregnancy, she was asked to state the name of the drug.

Women were asked about their need for information regarding the use of medicines during pregnancy and which sources of information had been used. Commonly used sources of information were listed: physician, midwife, pharmacy personnel, package information leaflets (PIL) or the Norwegian physician’s desk reference, Internet, personnel in herbal shops, or other. If she had used several sources of information, she was asked whether the information was similar between the sources and what the consequences were for her if the information was conflicting. She was also asked whether she had purposely chosen to avoid using a medication during pregnancy and, if so, why.


The study was approved by the Norwegian Social Science Data Services. All data were handled and stored anonymously.

Data analysis

The mean, median and standard deviation of the perceived risk for each substance were calculated. The percentages of women who perceived correctly the actual risk for each exposure were calculated. Risk sum scores were made for (1) the eight drug classes listed and (2) the eight classes of foods and chemicals. To check whether the scores were measuring a single aspect of risk attribution (i.e. unidimensionality), which is a criterion for valid measurement, a principal component analysis (PCA) was done. The PCA showed unidimensionality for the two sum scores. One exception was the risk score for eggs, which did not fit with the foods and chemicals dimension. This item was therefore not used in the sum score. Subsequently, the sum scores were subjected to reliability testing. The sum scores displayed adequate reliability with Cronbach’s alphas of 0.77 and 0.79 respectively. Cronbach’s alpha is a measure of internal consistency. The value represents the association between the sum score and the latent aspect measured and is therefore the upper boundary of any association with the sum score. An alpha > 0.70 is often regarded as adequate [17].

ANOVA and linear regression analyses were used to examine associations between maternal characteristics and risk perception sum scores. Linear regression analyses were used to examine associations between maternal socio-demographic factors, drug use during pregnancy, and the woman’s risk perception sum scores. Student’s t-test was used to assess differences in risk perception among pregnant women and mothers and associations between risk perception and drug use during pregnancy.


During the 5 weeks the questionnaire was accessible to the women, 1,821 questionnaires were completed and evaluated. Of these, 17 were excluded as the responder did not meet the inclusion criteria, and 11 cases were excluded as they had been completed more than once, resulting in a final sample of 1,793 (98.2%) eligible women who completed the questionnaire.

The socio-demographic characteristics of the study women are presented in Table 1. The socio-demographic characteristics were relatively similar to the general birthing population in Norway as recorded in the Medical Birth Registry of Norway with respect to age (mean 30 years in the general population), parity (41% primiparous), marital status (94% married or cohabiting) and smoking (12% at the end of pregnancy) [18]. The women were geographically similarly distributed as the general Norwegian population with respect to living within the five health regions in Norway (study women versus the general population of women: Region East: 52.0 vs. 50.2%, Region West: 25.1 vs. 25.7%, Region North: 9.3 vs. 9.7%, Middle Region: 8.7 vs. 8.7% and Region South: 4.9 vs. 5.7% [19]). The percentage of women with a tertiary education, however, was higher among the study women than among women aged 25–29 years in the general population (68.7 vs. 49.1%) [20].

Perception of risk

In total, 1,548 women (87.5%) estimated correctly that the general risk of malformation is ≤5%. Only 54 women (3.1%) believed the baseline risk to be over 10%. The risk was evaluated to be highest for cigarette smoking, alcohol use during the first trimester, sedatives/anxiolytics, antidepressants and thalidomide respectively (Fig. 1).
Fig. 1

Median risk score with 95% confidence interval for each of the 17 drugs, foods and chemicals studied. 0 indicated no foetal risk and 10 indicated always foetal malformation

Pregnant women and mothers had a similar perception of risk. There was never more than 0.5 points difference in mean scores for all the 17 substances, and there was no significant difference in mean sum scores for risk perception for drugs or foods for the two groups of women (P = 0.62 and P = 0.16, Student’s t-test).

The adjusted factors significantly associated with a higher risk perception of drugs were primiparity, older age, higher education level and choosing not to use a drug during pregnancy (linear regression analyses, entry method: Entry, P < 0.05 for each of the factors, adjusted R2 0.029). The factors significantly associated with a higher risk perception of foods and chemicals were alcohol use after known pregnancy, cigarette smoking during pregnancy, choosing not to use a drug during pregnancy, and primiparity (linear regression analyses, entry method: Entry, P < 0.05 for each of the factors, adjusted R2 0.021). Sub-analyses revealed an association between the woman’s lifestyle habits and her perception of risk: women who had used alcohol after knowing they were pregnant had significantly lower risk perception scores for alcohol use at time of conception, alcohol use during the first trimester and cigarette smoking during pregnancy compared to women who did not use alcohol after known pregnancy (ANOVA test, P < 0.001). Also, women who smoked during pregnancy had significantly lower risk scores for cigarette smoking (ANOVA test, P < 0.001).

Use of drugs during pregnancy

Most of the women (83.9%) reported having used drugs during pregnancy (Table 2). The most common drugs were paracetamol (acetaminophen), penicillins and heartburn drugs. However, many women (69.4%) also reported that they had chosen not to use a drug because they were pregnant. The most common reason for not using a drug was the fear of harming the unborn child.
Table 2

Use of medication during pregnancy (n = 1,793)


Number (%)


1,200 (66.9)

Drugs against heartburn

500 (27.9)


297 (16.6)

Herbal drugs

186 (10.4)

Drugs against NVP

154 (8.6)


57 (3.2)


30 (1.7)


31 (1.7)

Other drugsa

315 (17.6)

Any drug

1,504 (83.9)

NVP Nausea, vomiting of pregnancy; NSAIDs non-steroidal antiinflammatory drugs

aMostly other anti-infective agents, cold preparations, drugs against asthma and allergy, drugs against endocrine disorders and cardiovascular drugs

Women who had used a specific class of drugs during pregnancy rated the risk of such drug use as less risky than women who had not used these drugs (Table 3). The difference in risk perception was the largest between users and non-users for drugs used for heartburn and the smallest for herbal drugs.
Table 3

Associations between risk perception and use of specific drugs during pregnancy


Mean risk perception scoresa

P value (Student’s t-test)

Ratio of perception scores non-users/users

Non-users of the indicated drugs

Users of the indicated drugs

Drugs against heartburn



< 0.001





< 0.001





< 0.001


Drugs against NVP



< 0.001





< 0.001





< 0.001


Herbal drugs



< 0.001


NVP Nausea, vomiting of pregnancy

aAssessed with numeric rating scales ranging from 0 (no risk to the foetus) to 10 (foetal malformation after every exposure

Need for information about the safety of drugs during pregnancy

In total, 1,373 women (76.6%) stated that they needed information about drug use during their pregnancy. Among these women, the most commonly used sources of information were the physician, the product information leaflet and the pharmacy (Table 4). When several sources were used (n = 1,219), 24.8% reported that the information was frequently conflicting between the sources. In total, 60.0% reported that the information sources were relatively similar (the wording or level of precision differed), and 15.1% reported that information was exactly the same. Among women who reported that the information differed between sources (n = 1,035), the most common consequence was that she chose not to use the drug (16.8%), she trusted one source and disregarded the other/s (13.9%), she consulted a new source of information (11.4%), she became anxious (9.1%), and no consequence (12.0%).
Table 4

Drug information sources used by women (n = 1,373)a

Information source

Number (%)


1,074 (78.2)


684 (49.8)


671 (48.9)


577 (42.0)


362 (26.4)

Family and friends

324 (23.6)

Health food stores

58 (4.2)


35 (2.5)

PIL Package information leaflet, NPDR Norwegian Physician’s Desk Reference

aThe sum exceeds 100% as the woman could list several sources


To the best of our knowledge, this is the largest study assessing women’s perception of risk regarding exposures during pregnancy. We report that despite the fact that most women were able to perceive correctly the general malformation risk, risks associated with the use of drugs during pregnancy were highly overestimated. There are many possible reasons why pregnant women overestimate the risk of drug use during pregnancy. It is likely they have heard about the thalidomide tragedy in the 1960s and thus know about potential teratogenic effects of some drugs and other exposures during pregnancy. The media and Internet may also have an important influence on women’s perception of risk. A search of the Internet with the search engine Google in February 2009 using the words “pregnancy”, “medication” and “danger”, displayed more than 4.6 million hits. Unfortunately, women who read this information are not aware that this is most often not peer-reviewed evidence-based information as anyone can post information on the Internet.

We find that women highly overestimated the risk of certain exposures such as psychotropic drugs, alcohol use and cigarette smoking. The high number of questions about these exposures to teratology information services [21, 22] reflects the public’s special concern about such use during pregnancy. Also, a highly overestimated risk perception of antidepressants has previously been documented [6]. In this study including 100 women taking antidepressants during pregnancy, 87% of the women overestimated the teratogenic risk of these drugs [6].

The fact that alcohol and cigarette smoking are perceived as the most risky during pregnancy probably means that women have a broader perception of malformations or wide interpretation of the concept of malformation, including all types of foetal damage such as growth restriction and mental deficiencies. Due to the difficulties in rating risks, we believe that the relativeness of risk perception may be more interesting than the specific mean risk ratings. For example women perceive smoking as more dangerous than taking antidepressants and taking a dental x-ray as more dangerous than drinking alcohol at conception.

It is noteworthy that in our study we did not find a low perception of risk of herbal drugs, contrary to the often stated claim that pregnant women perceive herbal drugs as safer alternatives to conventional drugs [23, 24]. In fact, drugs used to treat heartburn, NVP and infections were perceived to be safer than herbal drugs. Increasing awareness among the public about potential adverse effects of herbal drugs may have lead to increased reluctance towards using these preparations during pregnancy [23].

As shown in Fig. 1, the women also overestimated the risk associated with exposure to common foods (except eggs) and chemicals. This may indicate a general cautiousness regarding any exposures during pregnancy, but also that many women are unnecessarily anxious about common household exposures during pregnancy. In addition, we found an approximately twice as high perception of teratogenic risk following a dental x-ray as compared to a previous study from the Mother Risk program in Toronto, Canada (50 vs. 25% perceived teratogenic risk) [15]. Our experience is that many dentists do not wish to perform a dental x-ray if a woman is pregnant especially as no national recommendations exist. Reluctancy may also be due to lack of knowledge about different types of diagnostic radiation exposure, threshold effects and how to interpret animal studies with respect to safety during pregnancy [25]. It has previously been documented that physicians also have unrealistically high perception of teratogenic risk associated with abdominal radiography and computer tomography during early pregnancy [26]. In all, 5% of obstetricians and 6% of family physicians would recommend an abortion after computer tomography during early pregnancy, despite no indication for doing so [26].

Several maternal factors were associated with risk perception. Interestingly, there were different factors associated with higher risk perception of drugs vs. foods and chemicals. Whereas older age, primiparity, higher education and avoidance of drug use due to pregnancy were associated with higher risk perception of drugs, lifestyle factors including alcohol use and cigarette smoking were the most important factors associated with risk perception of foods and chemicals. A possible explanation is that women drinking or smoking during pregnancy do so because they perceive that these habits are less dangerous; perhaps because they had a previous child or knew someone who gave birth to a normal healthy child following smoking or drinking in pregnancy. Alternatively, they may report a lower perception of risks because they do not wish to acknowledge that their lifestyle choices may be harmful to their child.

There are several possible consequences of overestimating the risk of drug use during pregnancy. When nearly seven of ten women in this study reported that they had chosen not to use a drug because they feared it was not safe, it is possible that non-compliance may be more common than previously documented. In this study the rates and types of non-compliance could not be determined as we did not specifically ask whether the drug the women chose not to use was prescribed by her physician or not. Avoiding using paracetamol for a headache is very different from avoiding taking prescribed antibiotics or antidepressants. As of today, literature regarding compliance to drug therapy during pregnancy is sparse. In a study among 295 women in the Netherlands, 22% reported non-compliance to drugs during pregnancy, most commonly using a lower daily dose than prescribed [11]. This, despite no evidence to support the theory that a lower dose is safer for the foetus. Non-compliance during pregnancy is a topic that warrants further investigation.

The need for information regarding the use of drugs during pregnancy is high, as 80% of the women reported this need, and should be acknowledged by both health care providers and public health agencies. It is reassuring that 75% stated that the information about drugs and pregnancy they received from different sources was relatively similar. On the other hand, there is still room for improvement as one out of four women reported receiving conflicting information from difference sources. Conflicting information given by health care personnel and drug monographs has previously been documented [22]. In a recent study, advice on drugs during pregnancy given by the Norwegian Drug Information Centres differed with that of the recommendations in the product monographs for almost 50% of the cases that were evaluated (n = 443) [22].

Methodological considerations

The use of the Internet as a tool for data collection appears to be a valid way of sampling data in the current population of young women. According to the Norwegian Statistic Central Bureau, use of the Internet every day or almost every day among women (aged 16 to 74 years) is approximately 67%, and during the last 3 months of 2008, use had increased to 92–98% among women aged 16–44 years [27]. In addition, it may be more comfortable for women to answer sensitive questions truthfully, such as regarding use of alcohol and cigarette smoking during pregnancy in an anonymous questionnaire. Furthermore, using an internet questionnaire is an efficient method of collecting population-based data, thereby reducing the risk for sampling bias. However, we cannot exclude the possibility that the study women differed from the general birthing population in other ways which we could not control for in our analyses.

Also, the results of this study should be interpreted bearing in mind that the women who completed the questionnaire had attained a higher educational level compared to the general population of women aged 25–29 years in Norway (68.7 vs. 49.1% with tertiary education) [20]. As a higher educational level was associated with a higher risk perception, we may have an overestimation of the perception of risk in general. However, the association estimates may be more valid, since association estimates are more robust to sampling bias than prevalence estimates.


The results of this study confirm that women overestimate the risks of drugs and other exposures during pregnancy and that there is a need for evidenced-based information in this field. The way pregnant women perceive teratogenic risk is directly associated with their drug use during pregnancy. Action needs to be taken to make comprehensive, accurate, evidence-based teratology information more available to the public. Several measures to promote accurate risk perception could be as follows: first, establish teratology information services in countries that do not provide them; second, offer high quality information sites on the Internet; and third and foremost, focus on better education of health care providers regarding teratology. Such actions would reduce anxiety and promote safer and more rational use of drugs during pregnancy.

Copyright information

© Springer-Verlag 2009