Menu
Menu
Menu
Menu
Background: There is a paucity of epidemiological data on medication use in pregnancy in Cameroon. Methods: Between March and August 2015, 795 pregnant women attending 8 urban and 12 rural hospitals in Cameroon for antenatal (ANC) or other care were interviewed on first trimester medication use using structured questionnaires. Multivariate logistic regression was used to analyse the association of 18 sociodemographic factors with medication use. Results: A total of 582 (73.2%) women took at least one orthodox (Western) medication during the first trimester, 543 (68.3%) women a non-pregnancy related orthodox medication, and 336 (42.3%)women a pregnancy related orthodox medication. 44% of the women took anti-infectives including antimalarials (33.6%) and antibiotics (20.8%).The other most common medications were analgesics (48.8%) and antianaemias (38.6%). Sulfadoxine/pyrimethamine, contraindicated in the first trimester of pregnancy, was the most commonly used antimalarial(13% of women).0.2% of women reported antiretroviral use. Almost 80% of all orthodox medications consumed by women were purchased from the hospital. 12.8% of the women self-prescribed. Health unit and early gestational age at ANC booking were consistent determinants of prescribing of non-pregnancy related, pregnancy related and anti-infective medications. Illness and opinion on the safety of orthodox medications were determinants of the use of non-pregnancy related medications and anti-infectives. Age and parity were associated only with non-pregnancy related medications. Conclusion: This study has confirmed the observations of studies across Africa indicating the increasing use of medications during pregnancy. This is an indication that access to medicine is improving and more emphasis now must be placed on medication safety systems targeting pregnant women, especially during the first trimester when the risk of teratogenicity is highest.
This cross sectional hospital based medication use survey was conducted during a six month period (March to August 2015) in twenty hospitals across rural and urban settings of the southwest region of Cameroon. Taking into consideration the total number of live births in South West Cameroon for 2013 (12,861 births of approximately 6,687 urban and 6,174 rural), a 6 months data collection period, and a 50% response distribution (worst case scenario), maximum sample size needed was estimated at approximately 374 for both urban and rural strata. We used a two stage cluster sampling technique: in the first step, eligible hospitals were randomly selected and in the second step, all eligible women within the selected hospitals attending during the study period were invited to participate. Both private and government hospitals were eligible to participate if they had an annual delivery rate of over one hundred and two hundred for rural and urban settings respectively. Out of forty-one eligible hospitals (twenty-two urban and nineteen rural), 20 were randomly selected. All pregnant women attending the selected hospitals on the days the researchers were in attendance (registered for antenatal clinics or not) were eligible to participate. To limit recall bias and to target the period within first trimester, only women with a gestation of three to seven months were eligible to participate in the survey. The pregnant women were recruited as they came for antenatal visits or in a small number of cases, for hospital consultation. Out of eight hundred and seventeen eligible women approached, seven hundred and ninety-five agreed to participate in the study (97.3% response rate).The observed distribution of the women into urban (55.2%) and rural (44.8%) settings of residence matched the expected distribution in the general population. Similarly, the proportion of women sampled in each health district within the data collection period was comparable to that of 2013 delivery data. Existing literature on medication use and safety were reviewed to facilitate the design of a questionnaire to be used by interviewers (Additional file 1). The questionnaire was designed to facilitate recall (e.g. the woman had to define the three months of her first trimester (exposure period of interest) prior to completing the section on medication exposure; the section on medication exposure was followed by a section on first trimester illnesses so as to validate data on exposure given (for example, if a woman reported having malaria during the first trimester, one would verify whether she reported taking an antimalarial in the previous section on medication exposure). A picture guide of orthodox and traditional medications was developed to facilitate recall. In Cameroon, patients have individual hospital books which they bring along during hospital visits. Antenatal care (ANC) files kept in the hospitals contain data only on medications prescribed during routine ANC visits. Hence medication data for other hospital visits could only be obtained from hospital books. When available, the data collectors used antenatal files and the hospital books to complement and validate data obtained from interviews. In a sub-study involving 84 participants to evaluate the relevance of using the hospital books (Table 1), we observed that 10.3% of the exposures would have been missed without the use of hospital books. Comparison of source of medication exposure (N=84 Women)a a Excluding those without a hospital book and those that did not take medications according to the hospital book In order to ensure standardized collection of data, eight nurses working in the areas of research and education were trained as data collectors and provided with a guidance note to assist them during data collection. A pilot study enabled the data collectors to feed back to adjust the questionnaire and various aspects of the data collection process. Actual data collection took place from the months of March to August 2015. Using the predesigned questionnaire, the data collectors conducted one-on-one interviews for consented women in private rooms of the hospital to obtain data on first trimester medication exposure. Following the approach of Baraka et al (2014) [10], orthodox medications were grouped as pregnancy related medications and non-pregnancy related medications. Pregnancy related medications were defined as routine medications taken not for ill-health, but to support the health of the mother and the developing fetus. These included anti-anemias, mineral supplements and vitamins. The Anatomical Therapeutic Chemical Classification System of the WHO was used to classify drugs into therapeutic classes. Drugs were also classified according to the old version of United States Food and Drug Administration (FDA) pregnancy risk classification (A, B, C, D or X; see foot note in Fig 6). The FDA classification of each drug was verified from various sources including normal Google search, the Internet Drug Index (RxList) and Drugs .com. Drugs for which no FDA class could be obtained were classified as category “U”. Epi-info 3.1 was used for data entry and cleaning while all data analyses were conducted using SPSS version 22. Prevalence of medication use was determined by dividing the number of women who took at least one medication by the total number of participating women. Differences in the prevalence of medication use within categorical variables were tested using the Pearson Chi-squared test of independence with significance level set at 0.05. Multivariatelogistic regression was used to identify the determinants of medication use. Using backward conditional logistic regression, all the variables were initially included in the model. Then, variables were removed from the model based on significance level set at 0.10, if their removal did not significantly worsen the overall prediction of the model [11]. Variables (all categorical) entered into the model were: health unit (individual primary or secondary/tertiary healthcare facility), setting of hospital (urban/rural), maternal age (13-17 years, 18-25 years, 26-35 years and 36-45 years), marital status, highest level of education attained, living conditions, level of alcohol consumption, gravidity, parity, previous pregnancy termination, gestational age at interview, gestational age at first booking, pregnancy planning, gestational age at pregnancy awareness, opinion on the safety of orthodox medication, safety advice and illness during first trimester. We investigated determinants for general orthodox medication use, pregnancy related medication use, and anti-infectives use. Results were reported as adjusted odds ratios and 95%CI. Determinants of medication use were defined as those variables retained in the final logistic regression model.
