Accepted Manuscript Title: Health Literacy and its association with Perception of Teratogenic Risks and Health Behavior during Pregnancy Author: Angela Lupattelli Marta Picinardi Adrienne Einarson Hedvig Nordeng PII: DOI: Reference:
S0738-3991(14)00175-X http://dx.doi.org/doi:10.1016/j.pec.2014.04.014 PEC 4776
To appear in:
Patient Education and Counseling
Received date: Revised date: Accepted date:
14-11-2013 4-4-2014 26-4-2014
Please cite this article as: Lupattelli A, Picinardi M, Einarson A, Nordeng H, Health Literacy and its association with Perception of Teratogenic Risks and Health Behavior during Pregnancy, Patient Education and Counseling (2014), http://dx.doi.org/10.1016/j.pec.2014.04.014 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Health Literacy and its association with Perception of Teratogenic Risks and Health Behavior during Pregnancy
School of Pharmacy, University of Oslo, Oslo, Norway
2
The Motherisk Program, Hospital for Sick Children, Toronto, Canada
3
Norwegian Institute of Public Health, Oslo, Norway
us
cr
1
ip t
Angela Lupattelli1, Marta Picinardi1, Adrienne Einarson2, Hedvig Nordeng1,3
an
Corresponding author at: Angela Lupattelli, School of Pharmacy, Faculty of Mathematics and Natural Sciences, University of Oslo, PO Box 1068 Blindern, 0316 Oslo, Norway. E-mail:
M
[email protected]; phone: +47 41343628; fax: +4722854402
d
Acknowledgments: We thank the Steering Committee of OTIS and ENTIS for reviewing the protocol, all
te
website providers who contributed to the recruitment phase and all study national coordinators (Twigg MJ, Zagorodnikova K, Mårdby AC, Moretti ME, Drozd M, Panchaud A, Hameen-Anttila K, Rieutord A, Gjergja
Ac ce p
Juraski R, Odalovic M, Kennedy D, Rudolf G, Juch H, Passier JLM and Björnsdóttir I). We are grateful to all the participating women who took part in this study. The authors did not report any potential conflict of interest. The study has received support by the Norwegian Research Council (grant no. 216771/F11) and the Foundation for Promotion of Norwegian Pharmacies and the Norwegian Pharmaceutical Society.
1 Page 1 of 24
ABSTRACT Objective: Investigate the association between health literacy and perception of medication risk, beliefs about medications, use and non-adherence to prescribed pharmacotherapy during pregnancy, and whether risk perception and beliefs may mediate an association between health literacy and non-adherence.
ip t
Methods: This multinational, cross-sectional, internet-based study recruited pregnant woman between 1-
cr
Oct-2011 - 29-Feb-2012. Data on maternal socio-demographics, medication use, risk perception, beliefs, and non-adherence were collected via an on-line questionnaire. Health literacy was measured via a self-
us
assessment scale. Mann-Whitney U test, Spearman’s rank correlation, generalized estimating equations and
an
mediation analysis were utilized.
Results: 4,999 pregnant women were included. Low-health literacy women reported higher risk perception
M
for medications, especially penicillins (Rho:-0.216) and swine flu vaccine (Rho:-0.204) and more negative beliefs about medication. Non-adherence ranged from 19.2% (high-health literacy) to 25.0% (low-health
d
literacy). Low-health literacy women were more likely to be non-adherent to pharmacotherapy than their
te
high-level counterparts (OR:1.30; 95% CI:1.02-1.66). Risk perception and beliefs appeared to mediate the
Ac ce p
association between health literacy and non-adherence. Conclusion: Health literacy was significantly associated with maternal health behaviors regarding medication non-adherence.
Practice implications: Clinicians should take time to inquire into their patients’ ability to understand health information, perception and beliefs, in order to promote adherence during pregnancy.
Key words: Health literacy; pregnancy; risk perception; beliefs about medication; drug utilization; nonadherence
2 Page 2 of 24
1.
Introduction
Health literacy is defined as "The degree to which individuals have the capacity to obtain, process, and understand basic health information and services required to make appropriate health decisions" [1,2]. Beyond having poorer health outcomes than their adequate health literacy counterpart [3,4], patients with
ip t
suboptimal health literacy show poorer ability to understand medical instructions and medication labels, and to properly take their medications [5]. In pregnancy, woman’s health status and her understanding of health
cr
information may directly affect maternal-fetal health. Previous research among pregnant subjects with
us
pregestational diabetes has shown that low health literacy is positively associated with unplanned pregnancies, no use of folic acid prior or during pregnancy, and later presentation for prenatal care [6].
an
During pregnancy, most women have a heightened awareness that medications they take may harm their fetus. The label of medications used in pregnancy can also influence the perception of teratogenic risk [7,8]
M
and thereafter shape decision-making regarding whether or not to take a needed medication during pregnancy [9]. Unrealistic elevated risk perception can even influence a woman to terminate a wanted
d
pregnancy [10]. Since inadequate numeracy skills may amplify or otherwise distort perceptions of risk or
Ac ce p
pregnancy.
te
benefit [11], the interplay between health literacy and risk perception may be of particular relevance during
Poor adherence to prescribed pharmacotherapies is a well-known public health concern, which can also occur during pregnancy. Conditions such as diabetes, hypothyroidism, epilepsy, or depression may lead to adverse pregnancy outcomes if not optimally treated [12-15]. Prior research has mostly focused on sociodemographics and life-style factors as an explanation of non-adherence. However, the reasons why women adhere or not to therapy may lie in other factors related to patients’ understanding of health information, perceptions and beliefs. Two recent studies reported that overall, up to 59% of women have poor medication adherence during pregnancy, with women’s beliefs about medication being important determinants of adherence [16,17]. In a recent meta-analysis Zhang et al.[18] found a weak though statistically significant association between health literacy and medication adherence in the general population, generating the
3 Page 3 of 24
hypothesis of a plausible mediator relationship between health literacy and other adherence determinants. To date, no study has investigated such a relationship in the pregnant population. The primary aim of the current study was to explore the role of maternal health literacy in relation to perception of medication risk, beliefs about medications, medication use and non-adherence to prescribed
ip t
pharmacotherapy during pregnancy. The secondary aim was to investigate whether perception of medication risks and beliefs about medications were possible mediators of the association between health literacy and
cr
non-adherence.
Study design and data collection
an
2.1
us
2. Methods
This was a multinational, cross-sectional, internet-based study performed in 18 countries: Australia, Austria,
M
Canada, Croatia, Finland, France, Iceland, Italy, Netherlands, Norway, Poland, Russia, Serbia, Slovenia, Sweden, Switzerland, United Kingdom and United States. Data originating from some South American
d
countries were also collected. Women who were currently pregnant at any gestational week were included in
te
the data analysis. The study subjects were categorized according to the reported country of residency.
Ac ce p
An anonymous on-line questionnaire (http://www.questback.com) was utilized for data collection, accessible for a period of two months in each participating country between 1-Oct-2011 - 29-Feb-2012. The questionnaire was open to the public via utilization of banners on one-four websites per country and/or social networks commonly visited by pregnant women. Websites with sufficiently high number of daily users were selected. Information about recruitment tools utilized and internet penetration rates in each participating country are described in details elsewhere [16]. The questionnaire was first developed in English and Norwegian languages, and then translated into the relevant language(s) for each participating country. The pilot study in four countries elicited no major changes. Collected data were scrutinized for the presence of potential duplicates (based on reported country of residency, socio-demographics, date and time of questionnaire completion) but none were identified. 4 Page 4 of 24
2.2
Variables and definitions
Health literacy was measured by the set of brief screening questions (SBSQ)[19], a self-assessment scale comprising the questions: 1) ‘How often do you have someone help you read hospital materials?’; 2) ‘How confident are you filling out medical forms by yourself?’; 3) ‘How often do you have problems learning
ip t
about your medical condition because of difficulty understanding written information?’ We assigned zero (highest problems with reading or learning/not at all confident in filling out medical forms) to four points
cr
(no problems with reading or learning/extremely confident in filling out medical forms) to the scaled
us
responses for the three questions. We then summed the scores to obtain a 0- or 12-point scale with higher scores indicating higher health literacy level. The SBSQ scale was used as an ordinal variable in the analysis.
an
The SBSQ sum score was thichotomized into low (score 0-5), medium (score 6-9) and high health literacy (score 10-12). The SBSQ has been shown to be effective in detecting inadequate health literacy [19].
M
Subjects were provided with a numeric rating scales ranging from 0 (‘not harmful to the fetus’) to 10 (‘very harmful to the fetus’) and could indicate the perceived risk for 13 agents including medications (i.e.
d
paracetamol, antibiotics (penicillin), antidepressants, thalidomide, swine influenza vaccine and Over-The-
te
Counter (OTC) medicines against nausea), herbal remedies (i.e. ginger and cranberries), food (i.e. blue
Ac ce p
veined/fermented cheese and eggs), radiation (i.e. dental X-ray), alcohol during the first trimester (wine, beer, and spirits) and cigarette smoking. Whenever the agent was not familiar to the subject, the option “unknown substance” could be selected. All exposures, with the exception of thalidomide (risk 10–40%), are considered to have a ≤5% risk of congenital malformation. Other agents assessed as harmful during pregnancy were cigarette smoking, alcohol in the first trimester and blue-veined cheese. Women’s beliefs about medication use in pregnancy were examined via three statements: 1) ‘I have a higher threshold for using medicines when I am pregnant than when I’m not pregnant’ (ST-1); 2) ‘Even though I am ill and could have taken medicines, it is better for the fetus that I refrain from using them’ (ST-2); 3) ‘Pregnant women should preferably use herbal remedies than conventional medicines’ (ST-3). Women could answer ‘strongly agree’, ‘agree’, ‘uncertain’, ‘disagree’, or ‘strongly disagree’ to each of the statements. The three pregnancy-specific statements were used as ordinal variables (trichotomized into agree, 5 Page 5 of 24
disagree, or uncertain) in the analysis. The sum score of the three pregnancy-specific beliefs was computed (range 0-12), with higher score indicating an overall higher level of agreement with the three pregnancyspecific beliefs. The study subjects were asked standardized questions about medication use for specific illnesses and OTC
ip t
medication use during pregnancy. Medications could be reported as free-text entry. A medication was defined as single product containing one or more active ingredients. We coded all recorded medications into
cr
the corresponding Anatomical Therapeutic Chemical (ATC) codes in accordance with the World Health
us
Organization (WHO) ATC index [20]. Medications were coded into the ATC 5th level (i.e. the substance level) whenever possible, otherwise higher levels (i.e. 2nd- 4th) could be utilized. Categorical variables
an
(Yes/No) were employed to measure 1) medication use of at least one medication for any indication during pregnancy; 2) medication use for chronic/long-term disorders; 3) OTC medication use. Iron, mineral
M
supplements, vitamins, as well as herbal remedies and any type of alternative medicine were recorded separately and excluded from the estimation of medication use.
d
Self-reported non-adherence to prescribed medications during pregnancy was measured via the following
te
question: ‘Have you deliberately chosen not to use a medicine prescribed by a doctor because you were
Ac ce p
pregnant’? In affirmative cases, women were asked to report the name(s) of the avoided medication(s) as free-text entry. Avoided medications were coded as described in the measurement of medication use above. Women reporting avoidance during pregnancy of at least one prescribed medication were considered nonadherent.
Maternal characteristics included geographic area of residency (i.e. country and region), age, educational level, mother tongue, employment status, previous children, marital status, folic acid intake, gestational week, multiple pregnancies, smoking status and alcohol use during pregnancy. Assessment of the study external validity was done by comparing socio-demographic and life-style characteristics of the sample with those of the general birthing population in the same country. Reports from National Statistics Bureaus or previous national studies were utilized for this purpose, as described in details elsewhere [16].
6 Page 6 of 24
2.3
Statistical analysis
Descriptive statistics were utilized as appropriate. The risk sum score was calculated for the six listed medication groups. Likewise, the pregnancy-specific belief sum score was computed for ST-1, ST-2 and ST3. To check whether the two sum scores were measuring a single aspect (i.e. unidimensionality) of risk
ip t
attribution and beliefs, respectively, a principal component analysis (PCA) was done. The PCA showed unidimensionality for the medication risk sum score, with exception of thalidomide. This item was therefore
cr
not included in the medication sum score. The medication risk sum score presented adequate reliability with
us
Cronbach’s alpha of 0.73 [21]. The sum score (range 0-50) was also used as ordinal variable (trichotomized into three equal groups: low (score 0-20), medium (score 21-28) and high (score 29-50), with computation
an
of a missing category (26.0% of the sample). The PCA showed unidimensionality for the pregnancy-specific belief sum score. The pregnancy-specific belief sum score (range 0-12) was used as continuous variable.
M
The Mann-Whitney U test was used to identify any correlation between the health literacy groups and: maternal socio-demographics; medication use during pregnancy; non-adherence to prescribed medications
d
during pregnancy; prescribed medication groups avoided during pregnancy. The Spearman’s rank
te
correlation coefficient was used to explore: the correlation within the three SBSQ questions; correlation between the health literacy groups and beliefs about medications. A p-value of
S0738-3991(14)00175-X http://dx.doi.org/doi:10.1016/j.pec.2014.04.014 PEC 4776
To appear in:
Patient Education and Counseling
Received date: Revised date: Accepted date:
14-11-2013 4-4-2014 26-4-2014
Please cite this article as: Lupattelli A, Picinardi M, Einarson A, Nordeng H, Health Literacy and its association with Perception of Teratogenic Risks and Health Behavior during Pregnancy, Patient Education and Counseling (2014), http://dx.doi.org/10.1016/j.pec.2014.04.014 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Health Literacy and its association with Perception of Teratogenic Risks and Health Behavior during Pregnancy
School of Pharmacy, University of Oslo, Oslo, Norway
2
The Motherisk Program, Hospital for Sick Children, Toronto, Canada
3
Norwegian Institute of Public Health, Oslo, Norway
us
cr
1
ip t
Angela Lupattelli1, Marta Picinardi1, Adrienne Einarson2, Hedvig Nordeng1,3
an
Corresponding author at: Angela Lupattelli, School of Pharmacy, Faculty of Mathematics and Natural Sciences, University of Oslo, PO Box 1068 Blindern, 0316 Oslo, Norway. E-mail:
M
[email protected]; phone: +47 41343628; fax: +4722854402
d
Acknowledgments: We thank the Steering Committee of OTIS and ENTIS for reviewing the protocol, all
te
website providers who contributed to the recruitment phase and all study national coordinators (Twigg MJ, Zagorodnikova K, Mårdby AC, Moretti ME, Drozd M, Panchaud A, Hameen-Anttila K, Rieutord A, Gjergja
Ac ce p
Juraski R, Odalovic M, Kennedy D, Rudolf G, Juch H, Passier JLM and Björnsdóttir I). We are grateful to all the participating women who took part in this study. The authors did not report any potential conflict of interest. The study has received support by the Norwegian Research Council (grant no. 216771/F11) and the Foundation for Promotion of Norwegian Pharmacies and the Norwegian Pharmaceutical Society.
1 Page 1 of 24
ABSTRACT Objective: Investigate the association between health literacy and perception of medication risk, beliefs about medications, use and non-adherence to prescribed pharmacotherapy during pregnancy, and whether risk perception and beliefs may mediate an association between health literacy and non-adherence.
ip t
Methods: This multinational, cross-sectional, internet-based study recruited pregnant woman between 1-
cr
Oct-2011 - 29-Feb-2012. Data on maternal socio-demographics, medication use, risk perception, beliefs, and non-adherence were collected via an on-line questionnaire. Health literacy was measured via a self-
us
assessment scale. Mann-Whitney U test, Spearman’s rank correlation, generalized estimating equations and
an
mediation analysis were utilized.
Results: 4,999 pregnant women were included. Low-health literacy women reported higher risk perception
M
for medications, especially penicillins (Rho:-0.216) and swine flu vaccine (Rho:-0.204) and more negative beliefs about medication. Non-adherence ranged from 19.2% (high-health literacy) to 25.0% (low-health
d
literacy). Low-health literacy women were more likely to be non-adherent to pharmacotherapy than their
te
high-level counterparts (OR:1.30; 95% CI:1.02-1.66). Risk perception and beliefs appeared to mediate the
Ac ce p
association between health literacy and non-adherence. Conclusion: Health literacy was significantly associated with maternal health behaviors regarding medication non-adherence.
Practice implications: Clinicians should take time to inquire into their patients’ ability to understand health information, perception and beliefs, in order to promote adherence during pregnancy.
Key words: Health literacy; pregnancy; risk perception; beliefs about medication; drug utilization; nonadherence
2 Page 2 of 24
1.
Introduction
Health literacy is defined as "The degree to which individuals have the capacity to obtain, process, and understand basic health information and services required to make appropriate health decisions" [1,2]. Beyond having poorer health outcomes than their adequate health literacy counterpart [3,4], patients with
ip t
suboptimal health literacy show poorer ability to understand medical instructions and medication labels, and to properly take their medications [5]. In pregnancy, woman’s health status and her understanding of health
cr
information may directly affect maternal-fetal health. Previous research among pregnant subjects with
us
pregestational diabetes has shown that low health literacy is positively associated with unplanned pregnancies, no use of folic acid prior or during pregnancy, and later presentation for prenatal care [6].
an
During pregnancy, most women have a heightened awareness that medications they take may harm their fetus. The label of medications used in pregnancy can also influence the perception of teratogenic risk [7,8]
M
and thereafter shape decision-making regarding whether or not to take a needed medication during pregnancy [9]. Unrealistic elevated risk perception can even influence a woman to terminate a wanted
d
pregnancy [10]. Since inadequate numeracy skills may amplify or otherwise distort perceptions of risk or
Ac ce p
pregnancy.
te
benefit [11], the interplay between health literacy and risk perception may be of particular relevance during
Poor adherence to prescribed pharmacotherapies is a well-known public health concern, which can also occur during pregnancy. Conditions such as diabetes, hypothyroidism, epilepsy, or depression may lead to adverse pregnancy outcomes if not optimally treated [12-15]. Prior research has mostly focused on sociodemographics and life-style factors as an explanation of non-adherence. However, the reasons why women adhere or not to therapy may lie in other factors related to patients’ understanding of health information, perceptions and beliefs. Two recent studies reported that overall, up to 59% of women have poor medication adherence during pregnancy, with women’s beliefs about medication being important determinants of adherence [16,17]. In a recent meta-analysis Zhang et al.[18] found a weak though statistically significant association between health literacy and medication adherence in the general population, generating the
3 Page 3 of 24
hypothesis of a plausible mediator relationship between health literacy and other adherence determinants. To date, no study has investigated such a relationship in the pregnant population. The primary aim of the current study was to explore the role of maternal health literacy in relation to perception of medication risk, beliefs about medications, medication use and non-adherence to prescribed
ip t
pharmacotherapy during pregnancy. The secondary aim was to investigate whether perception of medication risks and beliefs about medications were possible mediators of the association between health literacy and
cr
non-adherence.
Study design and data collection
an
2.1
us
2. Methods
This was a multinational, cross-sectional, internet-based study performed in 18 countries: Australia, Austria,
M
Canada, Croatia, Finland, France, Iceland, Italy, Netherlands, Norway, Poland, Russia, Serbia, Slovenia, Sweden, Switzerland, United Kingdom and United States. Data originating from some South American
d
countries were also collected. Women who were currently pregnant at any gestational week were included in
te
the data analysis. The study subjects were categorized according to the reported country of residency.
Ac ce p
An anonymous on-line questionnaire (http://www.questback.com) was utilized for data collection, accessible for a period of two months in each participating country between 1-Oct-2011 - 29-Feb-2012. The questionnaire was open to the public via utilization of banners on one-four websites per country and/or social networks commonly visited by pregnant women. Websites with sufficiently high number of daily users were selected. Information about recruitment tools utilized and internet penetration rates in each participating country are described in details elsewhere [16]. The questionnaire was first developed in English and Norwegian languages, and then translated into the relevant language(s) for each participating country. The pilot study in four countries elicited no major changes. Collected data were scrutinized for the presence of potential duplicates (based on reported country of residency, socio-demographics, date and time of questionnaire completion) but none were identified. 4 Page 4 of 24
2.2
Variables and definitions
Health literacy was measured by the set of brief screening questions (SBSQ)[19], a self-assessment scale comprising the questions: 1) ‘How often do you have someone help you read hospital materials?’; 2) ‘How confident are you filling out medical forms by yourself?’; 3) ‘How often do you have problems learning
ip t
about your medical condition because of difficulty understanding written information?’ We assigned zero (highest problems with reading or learning/not at all confident in filling out medical forms) to four points
cr
(no problems with reading or learning/extremely confident in filling out medical forms) to the scaled
us
responses for the three questions. We then summed the scores to obtain a 0- or 12-point scale with higher scores indicating higher health literacy level. The SBSQ scale was used as an ordinal variable in the analysis.
an
The SBSQ sum score was thichotomized into low (score 0-5), medium (score 6-9) and high health literacy (score 10-12). The SBSQ has been shown to be effective in detecting inadequate health literacy [19].
M
Subjects were provided with a numeric rating scales ranging from 0 (‘not harmful to the fetus’) to 10 (‘very harmful to the fetus’) and could indicate the perceived risk for 13 agents including medications (i.e.
d
paracetamol, antibiotics (penicillin), antidepressants, thalidomide, swine influenza vaccine and Over-The-
te
Counter (OTC) medicines against nausea), herbal remedies (i.e. ginger and cranberries), food (i.e. blue
Ac ce p
veined/fermented cheese and eggs), radiation (i.e. dental X-ray), alcohol during the first trimester (wine, beer, and spirits) and cigarette smoking. Whenever the agent was not familiar to the subject, the option “unknown substance” could be selected. All exposures, with the exception of thalidomide (risk 10–40%), are considered to have a ≤5% risk of congenital malformation. Other agents assessed as harmful during pregnancy were cigarette smoking, alcohol in the first trimester and blue-veined cheese. Women’s beliefs about medication use in pregnancy were examined via three statements: 1) ‘I have a higher threshold for using medicines when I am pregnant than when I’m not pregnant’ (ST-1); 2) ‘Even though I am ill and could have taken medicines, it is better for the fetus that I refrain from using them’ (ST-2); 3) ‘Pregnant women should preferably use herbal remedies than conventional medicines’ (ST-3). Women could answer ‘strongly agree’, ‘agree’, ‘uncertain’, ‘disagree’, or ‘strongly disagree’ to each of the statements. The three pregnancy-specific statements were used as ordinal variables (trichotomized into agree, 5 Page 5 of 24
disagree, or uncertain) in the analysis. The sum score of the three pregnancy-specific beliefs was computed (range 0-12), with higher score indicating an overall higher level of agreement with the three pregnancyspecific beliefs. The study subjects were asked standardized questions about medication use for specific illnesses and OTC
ip t
medication use during pregnancy. Medications could be reported as free-text entry. A medication was defined as single product containing one or more active ingredients. We coded all recorded medications into
cr
the corresponding Anatomical Therapeutic Chemical (ATC) codes in accordance with the World Health
us
Organization (WHO) ATC index [20]. Medications were coded into the ATC 5th level (i.e. the substance level) whenever possible, otherwise higher levels (i.e. 2nd- 4th) could be utilized. Categorical variables
an
(Yes/No) were employed to measure 1) medication use of at least one medication for any indication during pregnancy; 2) medication use for chronic/long-term disorders; 3) OTC medication use. Iron, mineral
M
supplements, vitamins, as well as herbal remedies and any type of alternative medicine were recorded separately and excluded from the estimation of medication use.
d
Self-reported non-adherence to prescribed medications during pregnancy was measured via the following
te
question: ‘Have you deliberately chosen not to use a medicine prescribed by a doctor because you were
Ac ce p
pregnant’? In affirmative cases, women were asked to report the name(s) of the avoided medication(s) as free-text entry. Avoided medications were coded as described in the measurement of medication use above. Women reporting avoidance during pregnancy of at least one prescribed medication were considered nonadherent.
Maternal characteristics included geographic area of residency (i.e. country and region), age, educational level, mother tongue, employment status, previous children, marital status, folic acid intake, gestational week, multiple pregnancies, smoking status and alcohol use during pregnancy. Assessment of the study external validity was done by comparing socio-demographic and life-style characteristics of the sample with those of the general birthing population in the same country. Reports from National Statistics Bureaus or previous national studies were utilized for this purpose, as described in details elsewhere [16].
6 Page 6 of 24
2.3
Statistical analysis
Descriptive statistics were utilized as appropriate. The risk sum score was calculated for the six listed medication groups. Likewise, the pregnancy-specific belief sum score was computed for ST-1, ST-2 and ST3. To check whether the two sum scores were measuring a single aspect (i.e. unidimensionality) of risk
ip t
attribution and beliefs, respectively, a principal component analysis (PCA) was done. The PCA showed unidimensionality for the medication risk sum score, with exception of thalidomide. This item was therefore
cr
not included in the medication sum score. The medication risk sum score presented adequate reliability with
us
Cronbach’s alpha of 0.73 [21]. The sum score (range 0-50) was also used as ordinal variable (trichotomized into three equal groups: low (score 0-20), medium (score 21-28) and high (score 29-50), with computation
an
of a missing category (26.0% of the sample). The PCA showed unidimensionality for the pregnancy-specific belief sum score. The pregnancy-specific belief sum score (range 0-12) was used as continuous variable.
M
The Mann-Whitney U test was used to identify any correlation between the health literacy groups and: maternal socio-demographics; medication use during pregnancy; non-adherence to prescribed medications
d
during pregnancy; prescribed medication groups avoided during pregnancy. The Spearman’s rank
te
correlation coefficient was used to explore: the correlation within the three SBSQ questions; correlation between the health literacy groups and beliefs about medications. A p-value of
