Determinants of Cervical Cancer Screening Among Women Aged 30–49 Years Old in Four African Countries: A Cross-Sectional Secondary Data Analysis

Abstract

Background

Early-stage cervical cancer screening is essential for providing women with a better chance of receiving effective treatment for precancerous and cancer stages. Delay in cervical cancer screening results in late presentation and cancer metastasis. National-level cervical cancer screening in resource-limited countries was scarce and not well studied in Africa based on national data specifically in Kenya, Cameroon, Nambia, and Zimbabwe.

Objective

To determine the prevalence and determinants of cervical cancer screening among eligible women in Kenya, Cameroon, Nambia, and Zimbabwe.

Methods

This study analyzed demographic and health survey data from Kenya, Cameroon, Nambia, and Zimbabwe. The data were extracted and analyzed by STATA version 15 and further analysis was done. Intraclass correlation coefficient, median odds ratio, and proportional change in variance were calculated to check the appropriateness of multilevel analysis. Variables with P-value < .25 were selected for multivariable multilevel logistic regression analysis. Finally, statistical significance between dependent and independent variables was assessed by odds ratios and 95% confidence intervals.

Results

The prevalence of cervical cancer screening in 4 African countries (Nambia, Kenya, Cameroon, and Zimbabwe) was 23.4 [95%CI: 22.8-24.1]. The determinants identified in this study were women of age 41–50 years [AOR = 1.47; 95% CI 1.24, 1.73], rural residence [AOR = .67; 95% CI .55, .81], women who have their own work [AOR = 1.1; 95% CI 1.0, 1.37], smoking status [AOR = 1.89; 95% CI 1.17, 3.0], age at first birth >=35 [AOR = 5.27; 95% CI 1.29-21.52], condom use [AOR = 1.79; 95% CI 1.46,2.19], husbands having worked [AOR = 1.5; 95% CI 1.08,2.11], rich household wealth [AOR = 1.43; 95% CI 1.13,1.8], and having health insurance [AOR = 2.2; 95% CI 1.8,2.7].

Conclusion

The prevalence of cervical cancer screening in Kenya, Cameroon, Nambia, and Zimbabwe was low as compared to World Health Organization (WHO) recommendations. Age, residence, work status, smoking status, women’s age at first birth, condom use, husbands having work, wealth status, and health insurance were the identified determinants of cervical cancer screening. Programme and policy interventions could address younger, rural residence women, poor wealth status women, women without work, and those who never use health insurance for the uptake of cervical cancer screening.

Keywords

cervical cancer cervical cancer screening cancer cervical intraepithelial neoplasia

Introduction

Cancer is a public health problem worldwide and it is the threat to all nations.^1,2 Cervical cancer is a global public health problem. One of the leading reproductive system infections is cervical cancer, which is caused by the human papilloma virus (HPV). Several studies have reported that sexual activity may increase HPV risk among women, and some medical studies have mentioned that human papilloma virus (HPV) is the primary factor for cervical carcinoma in women.^3-5 Worldwide, cervical cancer is the fourth most common cancer with estimated 604,127 new cases in 2020 and 341831 estimated deaths (90% of all cancer deaths from low- and middle-income countries).^6-8

Cervical cancer is one of the leading causes of cancer death in women in lower and middle-income countries (LMICs), where 83% of new cases and 85% of related deaths from poor resource countries, poor, vulnerable, and disenfranchised women were primarily affected.⁷ Mortality and risk of acquiring cervical cancer are high in low- and middle-income countries.^2,9 Cancer-specific mortality of women is leading in Africa and the lifetime risk and death in lower- and middle-income countries being high as 1.6% and .9% ,respectively.⁴ The burden of cancer in Africa is underestimated and the ability to screen is low due to many reasons. Change in lifestyle results in an increase in the burden of cancer in sub-Saharan Africa, while in the region, it disproportionately results in poor access to health care and poor quality of screening services.¹⁰The burden of cervical cancer in sub-Saharan Africa is estimated at 84%.¹¹

Cervical cancer is the easiest and potentially preventable cancer; it is the cause of maternal morbidity and mortality especially in developing countries. With regular Pap smears and follow up, it is a highly treatable and curable cancer.^3,4,12,13 Cervical cytology is a very specific test when the collection is with quality and reporting by well-trained technicians.⁵ Human papilloma virus types 16 and 18 cause 70%–75% of the overall cervical cancer. For non-communicable diseases, accurate statistics, high-quality surveillance, and monitoring system are important for designing a fundamental action plan.¹⁴ Therefore, studying the magnitude and determinates of cervical cancer is important to design or reshape the existing policy.

Studies revealed the uptake of cervical cancer screening in sub-Saharan Africa was 12.87%;¹⁵ in Botswana, 72% of the women screened for cervical cancer ever by Pap smear;¹⁶ 20.6% of the women in Uganda screened for cervical cancer;¹⁷ in other studies, only 29% screened for cervical cancer;¹⁸ in a study conducted in Kenya, 16.4% of women screened;¹⁹and in Hara region, Ethiopia, 24.5% screened.²⁰ Furthermore, the previous study in Kenya showed that 46% screened;²¹in Guraga zone, Ethiopia, only 3.8% screened;²² in Urban areas, Somali, 32.7% tested;²³ and other national level studies conducted in Kenya indicated that (19.4%) of women screened,²⁴ and only 5% underwent screening.²⁵

Determinants identified in previous literature with the uptake of cervical cancer were age,^19,26-28 residence,^19,29 marital status,^28,30 occupation,^19,29 educational status,^15,19,20 wealth,^17,19,30 knowledge on cervical cancer,^15,17,25,31 age at first sex,^22,28,32 human immune deficiency virus status,^15,25,33 and awareness.^{17,20,22,25-27} Besides, other literature identified awareness on services,^23,27 history of STI,^27,32-34 multiple sexual partners,^20,22,25 contraceptive use,^15,18 provider recommendation, frequency of facility visit and history of vaginal examination,³⁴ binge drinkers, high sugar consumption, and insufficient physical activity.¹⁹

Furthermore, concern about the gender of the health care professional, encouragement by HCPs and perceived quality of CCSS,¹⁸ recommendation by staff was made prior to joining KNH CCC,²¹ perceived susceptibility and awareness about screening locations,¹⁵ perceived susceptibility to cervical cancer and no perceived embarrassment of Pap smear test,²⁴ doctors visit,³⁵ self-perceived susceptibility, having received community health education from health extension workers,²⁷ number of women lifetime sexual partners >1, older age at the first pregnancy, higher number of children born,²⁸ lack of adequate health care facilities, lack of symptoms, not feeling at risk and social stigma,²⁵ negligence, absence of medical problems, fear, economic reasons,³¹ media exposure, had higher household wealth index, employed, insured, and had visited a health facility in 12 months²⁴ were the identified determinants of cervical cancer screening uptake.

World Health Organization (WHO) set the target for cervical cancer elimination by 2030 by implementing a “90%-70%-90%” target, which means by 2030, 90% of girls should be fully vaccinated with the HPV vaccine, 70% of women should be screened, and 90% of those identified with the cervical disease should receive appropriate treatment.⁷ Understanding the status of cervical cancer screening is essential to design a new approach to screening, reshaping, and evaluating the existing methods of screening. To make evidence-based decision techniques, monitoring and understanding the current practice and factor of cervical cancer gives outshining impact to improve maternal health. Therefore, this study aimed to determine the prevalence and determinants of cervical cancer screening in Kenya, Cameroon, Nambia, and Zimbabwe.

Methods and Materials

Design and Study Period

Demographic and health survey was a cross-sectional study conducted every 5 years based on nationally representative areas. This cross-sectional study was conducted in demographic and health surveys of in Kenya, Cameroon, Nambia, and Zimbabwe based on the timeframe of DHS of each country. The demographic and health survey name of each country were as follows: Kenya, KE DHS-2014; Cameroon, CM DHS-2018; Nambia, NM DHS-2013; and Zimbabwe, ZW DHS-2015. The study period for each country was in Kenya from May 7 to October 20, 2014; in Cameroon, from 06/2018–01/2019; in Nambia, from May to September 2013; and in Zimbabwe, from July 6 to December 20, 2015. For detailed information, you can access on demographic health survey database at http://www.dhs program.com. The sample size was calculated from 4 countries of which 3651 in Kenya, 4760 in Cameroon, 5230 in Nambia, and 3791 in Zimbabwe. Generally, the total sample size of a study was 17,432 women. To reach this sample, first we extracted data from DHS data of each country, then we excluded women aged less than 30 and greater than 49, then we excluded missed data, and finally the data with necessary variables were presented as sample size.

Data Source and Population

Demographic and health survey was collected in 2 stages by stratified sampling technique, and it is a nationally representative data which is collected every 5 years.³⁶ Demographic and health survey was collected based on indicators of maternal and child mortality and morbidity. The study includes women who reported for screened or not during the time of data collection, and only one-time screening was included in the study. The study included the self-reported screening status of the women. The author analyzed the 4 African countries (Kenya, Cameroon, Nambia, and Zimbabwe) recent demographic and health survey data with a total weighted sample of 17,432 eligible women. The weighted method and the detail of the population were obtained from http://www.dhs program.com. The survey year and the country of the survey are indicated in Table 1 below. The data were extracted from individual record (IR) files according to the objective of the study and the previous literature. The extracted data were weighted before any analysis for the representativeness of the sample. The authors included 4 countries (Kenya, Cameroon, Nambia, and Zimbabwe), since the outcome variable was only found in this country database. There is no consideration of types of screening in this study. Women’s aged 30 and above living in 4 African countries (Kenya, Cameroon, Nambia, and Zimbabwe) were included in our study since the recommendation for cervical cancer screening in the general population of 30 to 49 years old women according to WHO guidelines. Sampling weights are adjustment factors applied to each case in tabulations to adjust for differences in the probability of selection and interview between cases in a sample. In demographic and health surveys, in most surveys, the sample selected with unequal probability to expand the number of cases available in areas or subgroups for which statistics are needed. In the current study, we used individual weight for women (v005) multiplied by the inverse of the individual response rate for women in the stratum.³⁶ Table 1 showed the numbers of weighted and unweighted data involved in the study in each country.

Table 1.

Showing the Survey Year and Total Weighted Sample for Each Country.

S/No	Country	Survey year	Weighted data	Unweighted data
1	Kenya	2014	3651	5254
2	Cameroon	2018	4760	6677
3	Nambia	2013	5230	10018
4	Zimbabwe	2015	3791	6955
	Total sample size		17,432	28,904

Study Variables

Dependent variable: The outcome variable was cervical cancer screening, where it is a binary variable which was coded as women who screened for cervical cancer; it is coded as “1” for yes, while not screened previously coded as “2” for no. The screening status of the women had not considered the repetition, interval, and any clinical indication; it only categorized the women screened if the women reported being screened during data collection time and not screened when she did not screened.

Independent variable: The independent variables for this study were age, marital status, residence, wealth, educational status, husband’s educational, working status, age at first pregnancy, age at first sex, media exposure, parity, contraceptive use, health insurance, and condom use.

Data Collection Method

Demographic and health surveys implemented interviewer-based structured questionnaires. All reproductive-age women were interviewed by structured questionnaire with face-to-face interview techniques. Reproductive health-related standard questionnaires (in Woman’s Questionnaire) and several variables like socio-economic and demographic information were collected from women and households. For detailed information on the data collection methods, you can access (http://www.dhs.program.com). Among all reproductive-age women who were interviewed during the survey, 17,432 were included in this specific study. There are 4 model questionnaires in demographic and health surveys: A Household Questionnaire, a Woman’s Questionnaire, a Man’s Questionnaire, and a Biomarker Questionnaire. There are also several standardized Questionnaire Modules for countries with an interest in those topics.³⁶ For detailed information about the population, sample size, and questionnaire, access the demographic and health survey dataset (http://www.dhs.program.com). Strobe checklist was used for our study to check the content and guiding of the procedures of cross-sectional study.³⁷ The data which were collected during primary data collection were de-identified of individual profile, and there are no personally identified data in this study.

Variables Measurement

Cervical cancer screened: women who screened for cervical cancer and reported during data collection time of demographic and health survey.

Media exposure: a composite variable of frequency of listening to radio, watching television, and reading a newspaper, in which households were said to have media exposure “if they have exposed to either of listening to radio or watching television or reading a newspaper at least once a week” and no “if did not have exposure to all of the above media sources.”

Wealth status: Categorized as; poor “if the woman was in a poorer and poorest household,” middle and rich “if the woman was in a richer and richest household.” The category was based on demographic and health survey classification of wealth in 5 items and reduced to 3 by classifying poorer and poor to poor, middle wealth as middle and rich, and richest as rich. The descriptions and measurements of dependent and independent variables are presented in Table 2. The measurements and the description of each independent variable are explained in Table 2.

Table 2.

Measurement and Description of Dependent and Independent Variables in the Study.

S/n	Variable	Explanation
1	Cervical cancer screening	Women who screened for cervical cancer and reported during data collection time of demographic and health survey. It is coded “1” for women who have screened for cervical cancer and “2” for women who never screened for cervical cancer in last 5 years preceding the survey.
2	Age group	The current age of the women and re-coded into 2 categories with values of “1” for 30–40, “2” for >=41
3	Marital status	This was the current marital status of women and re-coded in 2 categories with a value of “1” for not in marital union (includes those who were never in union, divorced, widowed, and separated), and “2” for “marital union” (includes those living with a partner and those who are married)
4	Educational status of women	This is the minimum educational level a woman achieved and re-coded into 3 groups with a value of “0” for no education, “1” for primary education, “2” for secondary education, and “3” as higher education
5	Women working status	Re-coded in 2 categories with a value of “2” for not working and “1” for working
6	Husbands working status	Coded in 2 categories with a value of “1” for not working and “2” for working
7	Husbands educational status	This is the minimum educational level husbands achieved and re-coded into 3 groups with a value of “0” for no education, “1” for primary education, “2” for secondary education, and “3” as higher education
8	Wealth status	Categorized as; poor “if woman was in poorer and poorest household in the DHS database” coded as “0,” middle where categorized as middle coded as “1.” and rich “if woman was in richer and richest household” coded as “2”
9	Distance from health facility	The variable distance from health facility is coded as “1” for those considering distance is not big problem and “2” for those considering distance from health facility is big problem
10	Sex of household head	The variable sex of household head is coded as “1” for those who have female household head and “2” for those who have male household head
11	Media exposure	A composite variable obtained by combining whether a respondent reads newspaper/magazine, listening to the radio, and watching television with a value of “1” if a woman was not exposed to at least 1 of the 3 media and “2” if a woman has access/exposure to at least 1 of the 3 media. Those women who had exposure to 1 of them (reading newspaper, listening radio, or watching television) at least once a week are considered as media exposure.³⁸
12	Smoking status	The variable smoking status coded “1” for not smoking and “2” for those who have history of smoking
13	Age at sex	The variable age at sex is coded as “0” for those who have first sex before 18 years, “2” for those who initiated 19–29 years, and “3” for those who initiate sex at 30 and above years
14	Age at first birth	The variable age at first birth is coded as “0” for those who have first birth before 15 years, “2” for those who gave birth at age of 16–25 years, and “3” for those who gave birth at 26–35, for those who gave birth at 36 and above years
15	History of termination pregnancy	The variable history of termination to pregnancy was coded as “1” for those who have no history of termination of pregnancy and “2” for those who have history of termination of pregnancy
16	Contraceptive use	The variable contraceptive use was coded as “1” for those never use contraceptive and “2” for those history to use contraceptive during survey period
17	Health insurance	The variable health insurance coded as “1” for those women who have no health insurance and “2” for those women who have health insurance during survey period
18	Condom use	The variable condom use coded as “1” for the women who never use condom during the survey period and “2” for those use condom last 5 years of the survey
19	Types of residency	The variable place of residence recorded as rural and urban in the dataset was used without change
20	Parity	The variable parity categorized as “0” for <=2, “1” for those parity of 3–5 and “2” for those parity > =6

Statistical Analysis

Data were extracted based on the previous literature and the availability of the variable in the demographic and health survey database. After the data were cleaned, categorized, coded, and weighted, the authors explored the descriptive statistics by using the frequencies and percentages of data and presented by using tables. Before running a multilevel analysis, appropriateness was checked. The appropriateness of multilevel logistic regression was checked by calculating intraclass correlation coefficient (ICC), a proportional change in variance (PCV), and median odds ratio (MOR). The value of ICC greater than 5% to consider a variety of cervical cancer magnitude across the cluster indicated that significant clustering was found, therefore, multilevel logistic regression was a more appropriate method of analysis for the data. The first model was built only with dependent variable, and then the AIC, ICC, and BIC were checked. The second model was built on dependent and community-level factors. The third model was built on dependent variable and individual-level variables. The final model was built with on dependent variable and both community and individual-level variables. The potential correlation was checked between independent variables. The proportional change in variance increased from the null model to the full model (final model), which indicates the variation of cervical cancer screening across the clusters. Four models were built; the null model, (model 1) only dependent variable; model 2, dependent variable and community-level factors; model 3, dependent variable and individual-level variable; and the final model (model 4) was a dependent variable and both the community-level and individual-level factors. The best model was selected by comparing the Akaike information criteria (AIC) level, and the model with the smallest Akaike information criteria (AIC) is the best-fitted model. Therefore, a model with the smallest Akaike information criteria (AIC) value was selected and all interpretations and inferences were made based on this model. After selecting the best-fitted model, bivariable and multivariable multilevel logistic regression was done to determine the uptake and determinants of cervical cancer screening in 4 African countries based on recent demographic and health survey data. Both bivariable and multivariable multilevel analyses were done accordingly. Variable in bivariable logistic analysis with P-value < .25 was selected in multivariable multilevel analysis. Variables with a P value <.25 at bivariable analysis were entered into the multivariable multilevel logistic regression model. Finally, the variable with the P-value ≤.05 is used to declare statistical significance with the dependent variable. The odds ratio with a 95% confidence interval was used to declare the statistical association between the dependent and independent variables.

Results

Background Characteristics of the Study Population

Descriptive statistics showed that the majority (68.5%) of the study participants were categorized in the age category of less than or equal to 40 years old. More than half (51.5%) of the participants were living in rural areas. Sixty-three percent of the women were in marital union. More than one-third (43.3%) of the study participants attended secondary education. Most (67.1%) of the study participants who faced distance from a health facility were a big problem to get health services. More than half (59.7%) of the households were headed by males in the current study. Most (60.8%) of the study participants started their sexual intercourse at the age of 18 and below and the majority (75.8%) of them gave birth to their first child in the age category of 16–25 years old. Nearly half (48.0%) of the study participants were multiparous. Thirty percent of the study participants used condoms for recent sex. Older age women were screened for cervical cancer as compared to younger age women in the current study. The descriptive analysis indicated that women who lived in urban areas were screened more as compared to women who lived in rural areas (29.7% as compared with 17.3%). Nearly fifty percent of college and/or university graduated women were screened for cervical cancer. Of the women who exposed to media, 29.4% were screened for cervical cancer preceding the survey (Table 3).

Table 3.

Background Characteristics of the Study Population With Uptake of Cervical Cancer Screening in Four African Countries in 2021.

Variable	Category	Number (%)	Uptake of cervical cancer screening in percent (%)
Maternal age	<=40	11,941 (68.5%)	21.7
Maternal age	>=41	5491 (31.5%)	27.1
Marital status	In marital union	10,965 (62.9%)	22.4
Marital status	Not in marital union	6467 (37.1%)	25.2
Residence	Rural	8981 (51.5%)	17.3
Residence	Urban	8450 (48.5%)	29.7
Working status	Working	12,620 (72.4%)	24.7
Working status	Not working	4812 (27.8%)	20.1
Educational status	No education	1902 (10.9%)	5.9
	Primary education	6174 (35.4%)	17.3
	Secondary education	7553 (43.3%)	27.6
	Higher	1803 (10.4%)	45.5
Wealth status	Poor	5388 (30.9%)	12.9
	Middle	8670 (49.7%)	18.6
	Rich	3373 (19.4%)	31.9
Smoking status	Yes	2214 (12.7%)	48.0
Smoking status	No	15,218 (87.3%)	31.0
Husbands working status	Working	10,338 (94.3%)	22.9
Husbands working status	Not working	627 (5.7%)	13.0
Husbands educational status	No education	1380 (12.6%)	7.7
	Primary education	3379 (30.8%)	16.6
	Secondary education	4362 (39.8%)	25.8
	Higher (college/university) education	1844 (16.8%)	40.8
Sex of household head	Male	10,401 (59.7%)	20.0
Sex of household head	Female	7030 (40.3%)	27.0
Media exposure	Yes	11,281 (64.7%)	29.4
Media exposure	No	6151 (35.3%)	12.5
Age at first sex	≤18 years	10,501 (60.8%)	18.9
	19–29 years	5340 (30.9%)	30.9
	>=30 years	1427 (8.3%)	28.8
Age at first birth	≤15 years	1385 (7.9%)	13.7
	16–25 years	13,219 (75.8%)	23.2
	26–35 years	1983 (11.4%)	29.9
	≥36 years	845 (4.8%)	47.4
Parity	≤2	5221 (30.0%)	30.7
	3–5	8370 (48.0%)	24.2
	>=6 (Grand multiparty)	3840 (22.0%)	12.0
Distance from health facility	Big problem	11,704 (67.1%)	23.8
Distance from health facility	Not big problem	5728 (32.9%)	34.2
Terminated pregnancy ever	Yes	3250 (18.4%)	20.6
Terminated pregnancy ever	No	14,182 (81.6%)	23.9
Contraceptive use	Yes	12,251 (70.3%)	28.5
Contraceptive use	No	5181 (29.7%)	11.4
Health insurance	Yes	9145 (76.4%)	51.1
Health insurance	No	2833 (23.6%)	25.2
Condom use	Yes	5307 (30.4%)	36.0
Condom use	No	12,125 (69.6%)	20.9

Magnitude of Cervical Cancer Screening

The overall prevalence of cervical cancer screening in 4 African countries was 23.4 [95%CI: 22.8-24.1]. The prevalence of cervical cancer screening in each country is shown in Figure 1.

Figure 1.

The prevalence of cervical cancer screening of each country in 2021.

Random Effect Analysis

The hierarchal nature of the data violates the independent assumptions of the standard binary logistic regression model. Therefore, we fitted a multilevel mixed-effects generalized linear model by assuming that individual specific-effect is uncorrelated with the independent variables. Since there is no significant clustering effect at the household level, we were going to use the two levels, which were individual and community-level analyses. We fitted 4 models, which were the null model containing only the outcome variable, models 1 and 2 containing individual and community-level variables, respectively, and model 3, which contains both individual and community-level variables were fitted. Model 0 (empty model) was fitted without explanatory variables to test random variability in the intercept and to estimate the intraclass correlation coefficient (ICC), MOR, and PCV. Model 1 examined the effects of individual-level characteristics, model 2 examined the effect of community-level variables, and model 3 (full model) examined the effects of both individual and community-level characteristics simultaneously.

Multilevel analysis is necessary because there is a significant clustering of cervical cancer screening in this demographic and health survey data. The intraclass correlation coefficient (ICC) was 13.8%, which means 13.8% of the variability in the magnitude of cervical cancer screening among eligible women were attributed to the clusters. The median odds ratio (MOR) value of the null model 1.9 also indicates the presence of variation in cervical cancer screening between clusters. It means if we randomly select households from different clusters, those households at the cluster with higher cervical cancer screening had a 1.9 times higher chance of screening as compared to their counterparts. As shown in Table 4, model 3 has the smallest Akaike information criteria (AIC = 8596.55) as compared to the random intercept-only model or null model (AIC = 18383.65), model with only community-level factors (AIC = 18093.2), and model with only individual-level factors (AIC = 8695.0) (Table 4). In addition, the proportional change in variance (PCV) increases from 9.44% (null model) to 13.2% (model 3), indicating that mode 3 best explains the variability of cervical cancer screening. Therefore, this model is the best-fitted model for the data because it has the smallest AIC as compared to the rest models. So interpretations and reports were made based on this model (see Table 4).

Table 4.

Model Comparison and Random Effect Analysis Result.

Parameter	Null model 0	Model 1	Model 2	Model 3
ICC	13.8%	12.8%	12.66%	12.3%
Variance	.53 [.43,0.64]	.48 [.39,0.60]	.47 [.37,0.67]	.46 [.35,0.62]
MOR	1.9	1.79	1.77	1.75
PCV	Ref	9.44%	11.32%	13.2%
Model fitness
AIC	18383.65	18093.2	8695.0	8596.55

Determinants of Cervical Cancer Screening

The determinants of cervical cancer screening according to the final model were maternal age, residence, maternal working status, husbands’ occupational status, smoking status, age at first birth, health insurance, and condom use were significantly associated with cervical cancer screening in 4 African countries (Table 5). The odds of cervical cancer screening among the women age group of >=41 years were 1.47 [AOR = 1.47; 95% CI 1.24, 1.73] more likely as compared to the age group of 30–40 years. Women living in rural areas were 33% [AOR = .67; 95% CI .55, .81] less likely to screen for cervical cancer than the women living in urban areas. The women who have their own work were 1.1 [AOR = 1.1; 95% CI 1.00, 1.37] times more likely to screen for cervical cancer as compared with the women with no work. The odds of cervical cancer screening among smokers were 1.89 [AOR = 1.89; 95% CI 1.17, 3.00] times more likely as compared with the non-smoker. Having health insurance were 2.2 [AOR = 2.2; 95% CI 1.80, 2.70] times more screened for cervical cancer than the women who have no health insurance. Women aged at first birth >=35 were 5 [AOR = 5.27; 95% CI 1.29, 21.52] times more likely screened for cervical cancer as compared to women aged less than 15 years. Wealth of the household was another determinant of cervical cancer screening (Table 5).

Table 5.

Determinants of Cervical Cancer Screening Among Eligible Women in Four African Countries in 2021.

Variables	Category	Model 0 (null)	Model 1 AOR (95% CI)	Model 2 AOR (95% CI)	Model 3 AOR (95% CI)
𝛽0-intercept*		−1.33 [−1.40,-1.27]
Maternal age	30–40			1
Maternal age	>=41			1.4 [1.22,1.69]	1.47[1.24,1.73]
Marital status	In marital union			.81 [.66,0.98]	.85 [.70,1.04]
Marital status	Not in marital union			1
Residence	Rural		.46 [.39,0.55]		.67[.55,0.81]
Residence	Urban		1		1
Maternal educational status	No education			1	1
	Primary education			.91 [.59,1.40]	.9 [.60,1.44]
	Secondary education			1.11 [.71,1.73]	1.1 [.70,1.70]
	Higher			1.5 [1.00,2.44]	1.47 [.88,2.45]
Maternal working status	Working			1.17 [1.01,2.15]	1.1[1.00,1.37]
Maternal working status	Not working			1	1
Wealth status	Poor			1	1
	Middle			1.21 [1.00,1.50]	1.1 [.95,1.46]
	Rich			1.71 [1.40,2.11]	1.43[1.13,1.80]
Sex of household head	Male			1	1
Sex of household head	Female			.97 [.83,1.13]	.90 [.84,1.15]
Media exposure	Yes			1	1
Media exposure	No			.73 [.60,0.88]	.7 [.61,0,88]
Husbands educational status	No education			1	1
	Primary education			.79 [.53,1.18]	.8 [.54,1.19]
	Secondary education			.85 [.57,1.28]	.8 [.54,1.26]
	Higher			.88 [.54,1.40]	.89 [.55,1.43]
Husbands occupational status	Working			1.54 [1.10,2.16]	1.5[1.08,2.11]
Husbands occupational status	Not working			1	1
Terminated pregnancy	Yes			1.07 [.89,1.27]	1.0 [.80,1.20]
Terminated pregnancy	No			1	1
Smoking status	Yes			2.07 [1.28,3.37]	1.89[1.17,3.00]
Smoking status	No			1	1
Age at first birth	≤15			1	1
	16–25			.92 [.68,1.25]	.92 [.68,1.26]
	26–34			1.03 [.70,1.53]	1.01 [.66,1.50]
	≥35			5.42 [1.33,21.96]	5.27[1.29,21.52]
Age at first sex	≤18			1	1
	19–29			.93 [.79,1.08]	.9 [.78,1.00]
	30–39			.88 [.69,1.13]	.88 [.68,1.10]
Parity	≤2			1	1
	3–5			.94 [.78,1.13]	.95 [.79,1.14]
	≥6			.74 [.57,0.97]	.76 [.58,1.00]
Contraceptive use	Yes			1.22 [1.00,1.54]	1.2 [.88,1.50]
Contraceptive use	No			1	1
Health insurance	Yes			2.27 [1.88,2.73]	2.2[1.80,2.74]
Health insurance	No			1	1
Condom use	Yes			1.83 [1.49,2.24]	1.79[1.46,2.19]
Condom use	No			1	1
Distance from the health facility	Big problem			.86 [.72,0.99]	.90 [.75,1.07]
Distance from the health facility	Not a big problem			1	1

1 = reference.

Discussion

Cervical cancer is a serious and public health issue all over the world.⁹ Understanding the status of screening and its determinants is important for early intervention and saving the lives of many women. Therefore, this study aimed to determine the prevalence and determinants of cervical cancer screening among women in 4 African countries based on recent demographic and health survey data. The prevalence of cervical cancer screening in this study was 23.4 [95%CI: 22.8, 24.1]. The current finding is higher than the study conducted in Sub-Saharan Africa, 12.87%;¹⁵ Kenya, 16.4%;¹⁹ Uganda, 20.6%;¹⁷ Mozambique, 3%;³⁹ Cameroon, 4%;⁴⁰ Guraga Zone Ethiopia, 3.8%;²² and Tigray region Ethiopia, 5%.²⁵ This could be due to the variation in the study setting and the variation in the access to health facilities across different countries. Specifically, in Sub-Saharan Africa, there was a high burden on health care providers and other contextual factors which may lower the use of cervical cancer screening. Furthermore, the older study period might lower the pooled estimate of cervical cancer screening. In the study conducted in Kenya, there is variation in sample size and it is conducted before the beginning of the sustainable development goal (SDG) which focuses on the reduction of maternal mortality. In the study conducted in Mozambique, the population used was 15–64 years old, which may lower the prevalence of screening uptake since recommended screening services were above 30 years old. Overall, the variation could be due to access to screening services, awareness of the services, availability of the services in health facilities, acceptability of the screening, and perceived risk of the individuals contributing to these variations. Furthermore, variations in behavioral intention to screen for cervical cancer in a different place were different.

However, the current study finding was lower than the study conducted in the urban region, Somali, 32.7%,²³ and Hara region, Ethiopia.²⁰ This could be due to many African countries applying for free of charge cervical cancer screening services. Another possible reason for this variation is the difference in susceptibility, access to the health facility, and cultural and social support for maternal health services. The other possible reason for this variation could be the availability of different testing methods, access to the services, and availability of well-performed providers. Furthermore, the variation could be due to variations in the sociodemographic characteristics of the study participants and variations in the methodology of the study. Overall, the finding implies that context-specific which is inclusive of the community and the health care aspect is very important to improve the screening status of women. This signifies that there is a need to institutionalize the services in those selected countries. On another side, the current finding was comparable with the study conducted in Debre Markos town, Northwest Ethiopia.⁴¹ This could be due to health care policy and the availability of free pocket services in most African country is applicable.

This study identifies the determinants of cervical cancer screening in those in Kenya, Cameroon, Nambia, and Zimbabwe, based on the final model analysis of maternal age, residence, maternal and husband working status, wealth, smoking status, age at first birth, use of health insurance, and condom use were statistically identified determinants. The odds of cervical cancer screening among women’s age >=41 years were 1.47 times more likely screened as compared to the women’s age of <=40 years old. This finding is comparable with the study conducted in Cameroon,⁴⁰Mozambique,³⁹ Kenya,¹⁹ Hawassa, Ethiopia,²⁷ and the study conducted in Rwanda²⁸ which indicated that with the increased age, the odds of cervical cancer screening were increased. This could be due to the mean age of diagnosis of cervical cancer, where the age was between 35 and 44. The other possible explanation may be due to the symptoms of cervical cancer being presented above 30 years old. This gives an issue for the government which intensifies their intervention on lower age group women. The other possible reasons could be due to many country policies and guidelines recommending cervical cancer screening at higher area due to scarcity of resources. This implies that national and international organizations could emphasize younger women in the ways how to improve their screening services.

The odds of cervical cancer screening among rural residents were 33% less likely to screen for cervical cancer than the women who live in urban areas. Similar finding was observed in a study conducted in Kenya¹⁹ and the rural part of Ethiopia²⁹ which showed that women living in rural areas were one of the reasons not screened for cervical cancer. In a previous study conducted in Cameroon, women who lived in rural areas were less likely screened for cervical cancer as compared with women who lived in urban.⁴⁰ This could be due to fare distance from the health facility and low community support which amplify their service use. In addition, this could be due to lake of awareness of risk and the perception of the women regarding cervical cancer was genetic only, which reduces the perceived susceptibility of the women.^42,43 This finding was supported by a systematic review conducted worldwide which showed rural women were more likely to be presented with late-stage cervical cancer.⁴⁴ This implies that working with rural area women regarding cervical cancer screening and brief explanation of the risk regarding cervical cancer may improve the screening services. The odds of cervical cancer screening among women who had their own working were 10% higher than their counterparts. This finding was comparable with the study conducted in Rwanda²⁸ and India,³⁰ which indicated that women who had their work had higher odds of cervical cancer screening than those who did not work. The possible reason for this could be women who had their own work can get money, which covers the cost of the service and transportation.

The highest (rich) wealth of women was 1.43 times more likely to screen for cervical cancer than the women in the poor wealth category. A similar finding was observed in the study conducted in Rwanda,²⁸Cameroon,⁴⁰ Maharashtra, and India³⁰ which indicated that rich women were more likely to get screened. The possible reason could be due to easy access to information and service. This could be due to the easily covered cost of services. Cost is one of the attributes of the low success of cervical cancer screening. This finding supported by a study conducted in South Africa which showed Pap smear use was higher among rich, wealthy status women.⁴⁵

Other determinants of cervical cancer screening were husband occupation. The odds of cervical cancer screening among husbands who have occupation have 1.5 times more likely screened than women’s husbands who have no occupation. Household economy and health care practices are intrinsically related. This could be due to the husband’s support in money or the cost that covers the services intake. Women who smoke cigarettes were 1.5 times more likely to get screened for cervical cancer than women who didn’t smoke. Women who smoke are at high risk of progressive cervical cancer.⁴⁶ The risk of cervical cancer was increased among smokers as compared with non-smokers.⁴⁷ The other prospective cohort study showed that former smokers were more likely to receive screening services than never smokers.⁴⁸ This could be due to the health risk and susceptibility of an individual being increased and the health system could give attention to those population groups. While one study conducted in the United States showed that smoking status and cancer screening have no association,⁴⁹ the possible reasons for this discrepancy were variations in the method of analysis and variable level of measurement.

The odds of cervical cancer screening among women who give their first birth above 35 years were 5.27 times more likely screened than women aged less than equal to 15 years. The World Health Organization recommend screening for cervical cancer at and above 30 years old and every 5 years after 30 years. The possible reason could be due to maternal ability to use services increased with age, and the risk of the individuals increased so the screening status could be increased. The odds of cervical cancer screening were 1.79 times more likely used among condom users as compared with those who never use a condom for recent sexual activity. The possible explanation for this is that women who used a condom could have an awareness of sexually transmitted diseases which increase the seeking behavior of an individual. Women who used health insurance were 2.2 times more likely to take cervical cancer screening as compared with the women who did not use health insurance. A similar finding was observed in the study conducted in California⁵⁰ which showed insurance status is the predictor of cervical cancer screening. Other study conducted in Latin America and Caribbean cities showed women with health insurance having higher odds of cervical cancer screening as compared with women without any health insurance (49). The possible explanation could be due to the service costs covered by health insurance and having health insurance could have some information about health services. The other possible reason could be the women who have health insurance had increased health literacy levels. In some lietratures, the women’s behavior and culture of the community were the factors for cervical cancer screening.^51-54 While behavioral factors were not identified as determinant factors in this study, the author recommended researchers and demographic and health survey experts could assess the behavioral factors of the women.

One of the limitations of this study was that the culture, belief, and behavior of women were not assessed. While the current study only considered the status of screening which did not consider the frequency, the type and the indication of screening were considered as the limitation of the study. Some variables like vaginal examination, history of sexually transmitted disease, perceived susceptibility, doctors’ visit, cost of screening, and drinking status were not found in the demographic and health survey. In addition, the sample size for this specific stud was not calculated since we used all participants in the DHS data collection period which was considered a limitation of the study. As the strength of the study, the study considered both the individual and community-level variables with appropriate methods of analysis.

Conclusion

The prevalence of cervical cancer screening among women aged 30–49 years in 4 African countries was 23.4%, which is low compared with the WHO elimination target of more than 70% of eligible women who should be screened. This showed that cervical cancer was still a less prioritized problem in Africa. Government and non-government organizations could understand the real practice of cervical cancer screening and put their relief map on the increment of cervical cancer screening. In the final model, determinants of cervical cancer screening were identified such as maternal age, residence, maternal and husband’s working status, wealth, smoking status, age at first birth, use of health insurance, and condom use. The government and non-governmental organizations should address the screening service in rural areas. National and international interventions could address rural, lower-age, economically poor, and non-health insurance users. Therefore, policy and program interventions should target women who are younger and less than 40 years, who live in rural areas, women who did not have any work, poor household’s wealth, women who give their first birth at less than 15 years, women who never use health insurance, and women who did not use a condom.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical Approval

A waiver of written informed consent was secured from Demographic and Health Surveys (DHS) program data archivists to download the dataset for this study. After the data were downloaded from the measure DHS website the research of this study will maintain the confidentiality of the data. Since it was based on secondary data, which was publicly available, ethical approval was not required. However, we accessed the data set from the DHS website () through registering or online requesting.

ORCID iD

Melsew Setegn Alie

Appendix

References

WHO . Attaining the Nine Global Noncommunicable Diseases Targets; a Shared Responsibility. Global Status Report on NCDs; 2014.

Karen

. Towards the global elimination of cervical cancer. Papillomavirus Research. 2019;8:100170.

Ferlay

Soerjomataram

Ervik

, et al. GLOBOCAN 2012: Estimated Cancer Incidence, Mortality and Prevalence Worldwide in 2012. World Health Organization; 2012.

WHO . Comprehensive Cervical Cancer Prevention and Control: A Healthier Future for Girls and Women. Geneva, Switzerland: World Health Organization; 2013.

de Sanjosé

Serrano

Castellsagué

, et al. Human Papillomavirus (HPV) and Related Cancers in the Global Alliance for Vaccines and Immunization (GAVI) Countries: A WHO/ICO HPV Information Centre Report. Elsevier; 2012.

Gultekin

Ramirez

Broutet

Hutubessy

. World Health Organization call for action to eliminate cervical cancer globally. Int J Gynecol Cancer. 2020;30(4):426-427.

World health Organization . WHO framework for strengthening and scaling-up of services for the management of invasive cervical cancer. World health organization, 2020.

Singh

Vignat

Lorenzoni

, et al. Global estimates of incidence and mortality of cervical cancer in 2020: A baseline analysis of the WHO Global Cervical Cancer Elimination Initiative. Lancet Global Health. 2023;11(2):e197-e206.

Waggoner

. Cervical cancer. Lancet. 2003;361(9376):2217-2225.

10.

Cohen

Jhingran

Oaknin

Denny

. Cervical cancer. Lancet. 2019;393(10167):169-182.

11.

Ssentongo

Musa

, et al. Prevalence and determinants of cervical cancer screening in five sub-Saharan African countries: A population-based study. Cancer epidemiology. 2021;72:101930.

12.

Boffetta

Parkin

. Cancer in developing countries. CA A Cancer J Clin. 1994;44(2):81-90.

13.

Boulet

Horvath

Berghmans

Bogers

. Human papillomavirus in cervical cancer screening: Important role as biomarker. Cancer Epidemiology and Prevention Biomarkers. 2008;17(4):810-817.

14.

World health Organization . 2008-2013 Action plan for the global strategy for the prevention and control of noncommunicable diseases: Prevent and control cardiovascular diseases, cancers, chronic respiratory diseases and diabetes. World health organization, 2009.

15.

Yimer

Mohammed

Solomon

, et al. Cervical cancer screening uptake in Sub-Saharan Africa: A systematic review and meta-analysis. Publ Health. 2021;195:105-111.

16.

Mingo

Panozzo

DiAngi

, et al. Cervical cancer awareness and screening in Botswana. Int J Gynecol Cancer. 2012;22(4):638-644.

17.

Isabirye

Mbonye

Kwagala

. Predictors of cervical cancer screening uptake in two districts of Central Uganda. PLoS One. 2020;15(12):e0243281.

18.

Osingada

Ninsiima

Chalo

Muliira

Ngabirano

. Determinants of uptake of cervical cancer screening services at a no-cost reproductive health clinic managed by nurse-midwives. Cancer Nurs. 2015;38(3):177-184.

19.

Ng’ang’a

Nyangasi

Nkonge

NG,

et al. Predictors of cervical cancer screening among Kenyan women: Results of a nested case-control study in a nationally representative survey. BMC Publ Health. 2018;18(3):1-10.

20.

Wakwoya

Gemechu

. Prevalence of abnormal cervical lesions and associated factors among women in harar, eastern Ethiopia. Cancer Manag Res. 2020;12:12429.

21.

Njuguna

Ilovi

Muiruri

Mutai

Kinuthia

Njoroge

. Factors influencing cervical cancer screening in a Kenyan health facility: A mixed qualitative and quantitative study. Int J Reprod Contracept Obstet Gynecol. 2017;6(4):1180-1185.

22.

Endalew

Moti

Mohammed

Redi

Wassihun Alemu

. Knowledge and practice of cervical cancer screening and associated factors among reproductive age group women in districts of Gurage zone, Southern Ethiopia. A cross-sectional study. PLoS One. 2020;15(9):e0238869.

23.

Abdikarim

I. K.

Atieno

W. M. C.

Habtu

. Prevalence and associated factors of cervical cancer screening among Somali women in an urban settlement in Kenya. 2017:1-5.

24.

Tiruneh

Chuang

K-Y

Ntenda

PAM

Chuang

Y-C

. Individual-level and community-level determinants of cervical cancer screening among Kenyan women: A multilevel analysis of a Nationwide survey. BMC Wom Health. 2017;17(1):1-14.

25.

Teame

Gebremariam

Kahsay

Berhe

Gebreheat

Gebremariam

. Factors affecting utilization of cervical cancer screening services among women attending public hospitals in Tigray region, Ethiopia, 2018; Case control study. PLoS One. 2019;14(3):e0213546.

26.

Ndikom

Ofi

. Awareness, perception and factors affecting utilization of cervical cancer screening services among women in Ibadan, Nigeria: A qualitative study. Reprod Health. 2012;9(1):1-8.

27.

Azene

. Visual inspection with acetic-acid (VIA) service utilization and associated factors among women in Hawassa city, southern Ethiopia: A community based cross-sectional study. Women’s Midl Health. 2021;7(1):1-9.

28.

Makuza

Nsanzimana

Muhimpundu

Pace

Ntaganira

Riedel

. Prevalence and risk factors for cervical cancer and pre-cancerous lesions in Rwanda. Pan African Medical Journal. 2015;22(1):26.

29.

Gizaw

Teka

Ruddies

, et al. Reasons for not attending cervical cancer screening and associated factors in Rural Ethiopia. Cancer Prev Res. 2020;13(7):593-600.

30.

Nene

Jayant

Arrossi

, et al. Determinants of women’s participation in cervical cancer screening trial, Maharashtra, India. Bull World Health Organ. 2007;85:264-272.

31.

Claeys

Gonzalez

Page

Bello

Temmerman

. Determinants of cervical cancer screening in a poor area: Results of a population-based survey in Rivas, Nicaragua. Trop Med Int Health. 2002;7(11):935-941.

32.

Hailemariam

Yohannes

Aschenaki

Mamaye

Orkaido

Seta

. Prevalence of cervical cancer and associated risk factors among women attending cervical cancer screening and diagnosis center at Yirgalem General Hospital, Southern Ethiopia. J Cancer Sci Ther. 2017;9(11):730-735.

33.

Zena

Elfu

Mulatu

. Prevalence and associated factors of precancerous cervical lesions among women in Ethiopia: A systematic review and meta-analysis. Ethiopian Journal of Health Sciences. 2021;31(1):189-200.

34.

Muluneh

Atnafu

Wassie

. Predictors of cervical cancer screening service utilization among commercial sex workers in Northwest Ethiopia: A case-control study. BMC Wom Health. 2019;19(1):1-9.

35.

Soneji

Fukui

. Socioeconomic determinants of cervical cancer screening in Latin America. Revista Panamericana de Salud Pública. 2013;33:174-182.

36.

Dhs

AAO

Data

. Dhs methodological reports 13. 2014.

37.

Recommendation checklist . STROBE statement—checklist of items that should be included in reports of observational studies. 2007:1-6.

38.

Csace

. Ethiopia demographic and health survey 2016. In: Addis Ababa, Ethiopia, and Rockville. Maryland, USA: CSA and ICF; 2016.

39.

Brandão

Tulsidás

Damasceno

Silva-Matos

Carrilho

Lunet

. Cervical cancer screening uptake in women aged between 15 and 64 years in Mozambique. Eur J Cancer Prev. 2019;28(4):338-343.

40.

Okyere

Duodu

Aduse-Poku

Agbadi

Nutor

. Cervical cancer screening prevalence and its correlates in Cameroon: secondary data analysis of the 2018 demographic and health surveys. BMC Publ Health. 2021;21(1):1-8.

41.

Bante

Getie

Getu

Mulatu

Fenta

. Uptake of pre-cervical cancer screening and associated factors among reproductive age women in Debre Markos town, Northwest Ethiopia, 2017. BMC Publ Health. 2019;19(1):1-9.

42.

Lim

Ojo

. Barriers to utilisation of cervical cancer screening in Sub Sahara Africa: A systematic review. Eur J Cancer Care. 2017;26(1):e12444.

43.

Marashi

Irandoost

Yoosefi Lebni

Soofizad

. Exploring the barriers to Pap smear test in Iranian women: A qualitative study. BMC Wom Health. 2021;21(1):1-12.

44.

Tekalign

Teshome

. Prevalence and determinants of late-stage presentation among cervical cancer patients, a systematic review and meta-analysis. PLoS One. 2022;17(4):e0267571.

45.

Akokuwebe

Idemudia

Lekulo

Motlogeloa

: Determinants and levels of cervical Cancer screening uptake among women of reproductive age in South Africa: Evidence from South Africa Demographic and health survey data, 2016. BMC Publ Health 2021, 21(1):2013.

46.

MacLaughlan

Lachance

Gjelsvik

. Correlation between smoking status and cervical cancer screening: A cross-sectional study. J Low Genit Tract Dis. 2011;15(2):114-119.

47.

Sugawara

Tsuji

Mizoue

, et al. Cigarette smoking and cervical cancer risk: An evaluation based on a systematic review and meta-analysis among Japanese women. Jpn J Clin Oncol. 2019;49(1):77-86.

48.

Eng

David

Ally

Stefanick

Tang

. The association between cigarette smoking, cancer screening, and cancer stage: A prospective study of the women’s health initiative observational cohort. BMJ Open. 2020;10(8):e037945.

49.

Reyes-Ortiz

Velez

Camacho

Ottenbacher

Markides

. Health insurance and cervical cancer screening among older women in Latin American and Caribbean cities. Int J Epidemiol. 2008;37(4):870-878.

50.

Cowburn

Carlson

Lapidus

DeVoe

. Peer reviewed: The association between insurance status and cervical cancer screening in community health centers: exploring the potential of electronic health records for population-level surveillance, 2008-2010. Prev Chronic Dis 2013;10:E173.

51.

Hasahya

Berggren

Sematimba

Nabirye

Kumakech

. Beliefs, perceptions and health-seeking behaviours in relation to cervical cancer: A qualitative study among women in Uganda following completion of an HPV vaccination campaign. Glob Health Action. 2016;9(1):29336.

52.

Wollancho

Amdissa

Bamboro

Wasihun

Tareke

Gizaw

. Determining behavioral intention and its predictors towards cervical cancer screening among women in Gomma district, Jimma, Ethiopia: application of the theory of planned behavior. PLoS One. 2020;15(11):e0238472.

53.

Kamdje

Kalgong

Kamga

Tagne

Amvene

Nangue

. Knowledge and behavior of women on cervical cancer in the northern region of Cameroon. J Anal Oncol. 2017;6(3):125-130.

54.

Acharya Pandey

Karmacharya

. Cervical cancer screening behavior and associated factors among women of Ugrachandi Nala, Kavre, Nepal. European journal of medical research. 2017;22(1):1-9.