Risk Communication and Public Health Emergency Responses During COVID-19 Pandemic in Rural Communities in Kenya: A Cross-Sectional Study

Abstract

Background: The COVID-19 pandemic highlighted the crucial role of community preventive behaviors in controlling the virus’ spread. Studies show that people’s risk perceptions and awareness significantly contribute to the containment and prevention of infections by motivating adoption of desired actions and behaviors. This study aimed at assessing the role of risk communication and factors that influenced responses during the COVID-19 pandemic in rural communities in Western Kenya. Methods: A descriptive cross-sectional study was conducted using a quantitative research approach, collecting data from 806 individuals across Kisumu, Vihiga, and Kakamega counties. Descriptive statistics were used to detail the demographic characteristics of the study population, while logistic regression analysis estimated the associations between risk communication and demographic characteristics on COVID-19 vaccine acceptance, compliance with mitigation behaviors, perceived severity, and perceived susceptibility. Results: The results showed that 55% of participants were male and 45% were female, with an average moderate compliance with safety measures (mean = 5.15). A significant portion of participants wore face masks (85.3%), practiced hand hygiene (78.9%), and avoided close contact behaviors (66.6%). Most respondents received information through mass media (86.1%) and health workers (72.9%). Compliance with COVID-19 mitigation measures was highest among those who trusted information from official institutions, health professionals, and mass media, compared to social media, with increased odds of 2.7 times and 2.5 times, respectively. Higher risk perception was significantly associated with older age groups (above 50 years), being male, and working in the private sector. Effective risk communication significantly influenced risk perception, compliance with COVID-19 measures, and vaccination acceptance. Conclusions: The findings suggest that effective risk communication strategies are critical during public health emergencies and hence implications for future public health crises. The results underscore the importance of targeted communication and tailored interventions to improve compliance and vaccine acceptance among different demographic groups, ensuring a more robust public health response during outbreaks and pandemics.

1. Background

The COVID-19 pandemic, which nearly halted global activities in 2020 and has claimed over 7 million lives worldwide [1], underscores the critical need for enhanced international cooperation, emergency preparedness, enhanced global security, surveillance, monitoring, and capacity building for public health emergencies (PHEs) and risk communication [1,2]. The These numbers reflect officially reported deaths; however the actual figures may be higher due to underreporting or differences in surveillance and reporting practices across countries. The International Monetary Fund estimated that the pandemic would cost the global economy approximately USD 13.8 trillion by the end of 2024 [3], in addition to its significant socioeconomic and other broad societal impacts. It is widely regarded as the most severe disaster in living memory by nearly any measure [4]. With the increasing frequency of PHEs, the World Health Organization (WHO) emphasizes the importance of risk communication as a critical area of focus for global investment, collaboration, research, data sharing, and policy development [5,6,7].

In 2015, following the Ebola Virus Disease crisis in West Africa, the WHO established a working group to draft guidelines on building national capacities for communicating health risks during PHEs [8]. Risk communication has been identified as one of the eight core capacities under the International Health Regulations (IHRs of 2005) [9]. And there is a global agenda to advance breakthroughs in risk and crisis communication to support evidence-based campaigns in PHEs [5].

Public health emergencies encompass disease outbreaks (epidemics and pandemics), environmental disasters, and other humanitarian crises and man-made disasters which may be in localized geographical locations or widespread in large areas or across countries and the world in general [10]. Large-scale epidemics and pandemics cause widespread deaths and suffering, disproportionately affecting vulnerable populations, and lead to extensive social, economic, health, and political disruptions, complicating recovery efforts [7]. Therefore, timely and effective responses are critical yet also challenging.

During PHEs, it is essential for people to understand the health risks they face, the nature and sources of these risks, their potential impact at individual and community levels, and the actions they should take to protect their health and that of their communities [8]. Risk communication (RC) involves the real-time exchange of information, advice, and opinions among experts, community leaders, officials, and the public at risk [11]. Effective RC aims to promote health behaviors, such as adherence to public health and social measures (PHSMs), timely screening, and seeking prompt treatment and other mitigation measures [12]. Before the widespread availability of biomedical innovations like COVID-19 tests, vaccines, and treatments, PHSMs were crucial in limiting the virus’s spread [13]. PHSMs help frame risks and risk perception and are integral to managing PHEs beyond just biomedical interventions [12]. Risk perceptions during PHEs and related actions and responses vary and evolve as the situation and circumstances evolve, making imperative an iterative process for effective RC [14,15].

Effective RC influences how people perceive risks and prompts desired actions during emergencies [16,17]. It is a dynamic, interactive, and adaptive process [8,11,18]. However, risk communication models developed in Western countries often fail to resonate with African communities due to contextual differences in risk perception, stereotypes, and other factors like wars, displacement, health literacy, negative colonial medical legacy and practices, and weak healthcare infrastructure [4,19]. Following the WHO’s pandemic declaration in March 2020, African countries quickly implemented countermeasures to curb COVID-19’s spread to mitigate adverse impacts such as increasing global morbidity, mortality, and overwhelmed health systems [20,21]. The WHO Regional Office for Africa’s risk communication and community engagement (RCCE) framework guided many African countries’ responses, with varied outcomes [22]. In May 2022, the Africa Centers for Disease Control and Prevention (Africa CDC), in collaboration with the WHO, USAID, UNICEF, and other partners, established the Public Health Risk Communication and Community Engagement of Practice for Africa (PH-RCCE-CoPA) [22]. Despite these efforts, some African countries faced challenges such as resource shortages, staff constraints, and weak coordination [23].

A study on RCCE strategies for COVID-19 in 13 African countries, including Kenya, identified key RCCE strategies, including risk communication systems, community engagement, public communication, and managing misinformation [23]. Challenges included government distrust, cultural resistance, and misinformation. According to WHO [24], effective risk communication and community engagement are critical components of public health emergency responses, enabling authorities to build trust, encourage protective behaviors, and mitigate misinformation during outbreaks such as COVID-19. In the context of COVID-19, RCCE strategies must be context-specific, inclusive, and responsive to community needs to ensure that individuals and groups are informed, engaged, and empowered to take appropriate health actions [24]. A case study of Kenya’s early COVID-19 response (February 2020–May 2021) found initial significant compliance with non-pharmaceutical measures such as PHSMs due to strict governmental enforcement, shifted to a subsequent decline in compliance once lockdowns were eased because people’s and communities’ perceived risk of COVID-19 infections diminished [25]. However, limited studies focused on the influence of risk communication on risk perception and compliance with preventive behaviors during the COVID-19 pandemic in Kenya. This study aimed at assessing the role of risk communication and factors influencing public health emergency responses during the COVID-19 pandemic in rural communities in Kenya.

2. Materials and Methods

2.1. Research Design

A cross-sectional design using a quantitative research strategy was executed to assess the role of risk communication and factors influencing public health emergency responses during the COVID-19 pandemic in selected rural communities in Kenya. The study was conducted in May–August 2021.

2.2. Study Setting

The study was carried out in Kakamega, Vihiga, and Kisumu counties in western Kenya. Kakamega County covers an area of approximately 3050.3 km². The county has twelve sub-counties, eighty-three locations, two hundred and fifty sub-locations, one hundred eighty-seven Village Units, and four hundred Community Administrative Areas. There are 433,207 households with an average size of 4.3 persons per household and a population of 1,867,579 [26].

Vihiga County lies in the Lake Victoria Basin and covers an area of 531.0 Km². Vihiga County is located around 80 km northwest of Eldoret, around 60 km north of Kisumu, and approximately 350 km west of Nairobi City, the capital city of Kenya. It has a population of 590,013, of which 51.9% are women, while men constitute 48.1%. Sixty-four-point four percent (64.4%) of the total population is under the age of 30 [26]. The county has five administrative sub-counties. The county is further subdivided into 38 locations and 131 sub-locations.

Kisumu County is bordered to the north by Nandi County and to the Northeast by Kericho County. The land area of Kisumu County totals 2085.9 km² [26]. It has a population of 1,155,574, of which women make up 50.1% of Kisumu’s population and men represent 49.9%. Sixty-four percent of the total population is under the age of 25 [27]. Administratively, the county is divided into 7 sub-counties, and these are further divided into 35 wards [27], with an average household size in Kisumu County reported as 4.4 persons per household.

2.3. Study Population and Participants

All community members aged 18 and above were eligible for participation in the study.

2.4. Sample Size Determination

An a priori power analysis was conducted to determine the required sample size (n), calculated based on a desired power level of 80% (1 − β), a significance level of α = 0.05, and a target population effect size of 10%. Based on this analysis, using a one-sample proportion test, the sample size was 779 participants. Conducting an a priori power analysis allows researchers to efficiently manage statistical power before data collection begins [28]. To compensate for potential nonresponse—unit nonresponse (e.g., due to noncontact or refusal)—an additional 10% (78 participants) was added, resulting in a final target sample size of 857. This adjustment is consistent and supported by other studies [29,30,31].

2.5. Sampling Procedure

The counties in Western Kenya were selected using purposive sampling because they were among those with the top 15 counties with the highest COVID-19 prevalence in nationally, and their ranking among the top three leading in prevalence in the western region of Kenya (KNBS) [32]. Proportionate stratified sampling was used to select two sub-counties from each county based on whether they were urban or rural and to select the study subjects from the six sub- counties. Two Wards were selected using simple random sampling from each selected sub-county. A list of households was generated based on the administrative location headed by the Chief in the selected sub-counties. Systematic random sampling was then used to select households in the selected wards. A representative of the eligible study subjects or household heads in the selected households was randomly picked to participate in the study (see

Table 1. Sampling frame.

County	Sampled Sub-County	Sub-County Population	Wards (Population)	Sample	Total Sampled Individuals
Kakamega	Kakamega Central	188,212	Butsotso Central (25,744)	149	219
			Mahiakalu (12,067)	70
	Navakholo	137,165	Bunyala Central (38,407)	142	187
			Ingotse-Matiha (12,091)	45
Kisumu	Kisumu Central	174,145	Kondele (48,004)	93	130
			Milimani (18,902)	37
	Nyando	161,508	Ahero (31,440)	60	121
			Onjiko (30,937)	51
Vihiga	Emuhaya	69,250	Northeast Bunyore (35,908)	36	64
			Central Bunyore (27,316)	28
	Hamisi	148,259	Banja (22,535)	45	85
			Tambua (18,689)	40
Total					806

below).

2.6. Data Sources and Instrument

Data were gathered using a structured questionnaire. The questions were adopted and modified from similar studies [33,34]. In addition, the WHO’s and Kenyan Government Ministry of Health guidelines on COVID-19 infection prevention and control (IPC) were reviewed and additionally used to design the questionnaire [35,36]. The instrument was designed in English and translated into local languages (Luo and Luhya) for use in those languages where necessary during the data collection process, and these were back-translated to check for consistency. The participants who could not understand English were issued with appropriate instruments in local dialects. A pilot study was carried out with 80 people who were chosen from the selected sub-counties but were omitted from the actual study. The reliability of the instrument was evaluated using the interclass correlation coefficients (ICC). The computed interclass correlation coefficients (ICCs) yielded 0.766. An ICC of 0.5 or higher is acceptable, confirming its suitability for the study. Further, data in the questionnaire with a Likert scale were entered in IBM SPSS version 26 and analyzed and a Cronbach coefficient was derived; a correlation coefficient of 0.81 was computed and was considered acceptable. These were to assess the consistency, clarity, and accuracy for making any necessary adjustments to refine the study instrument.

2.7. Variables Definitions

2.7.1. Dependent Variables

The dependent variable for this study was public health emergency responses measured in compliance behaviors in various ways such as adherence to the public health guidelines and mitigation measures, risk perception, and vaccine acceptance.

Compliance behaviors: hand washing, face mask wearing, sanitization, social distancing, cough hygiene/not touching face, and compliance with other WHO guidelines (lock down and travel policies). Participants were asked for their compliance with these behaviors and their responses were categorized as: not at all, rarely, frequently and always. The level of compliance was then classified as high and low accordingly.

Vaccine acceptance: defined as the degree to which individuals accept, question, or refuse vaccination. It is one of the major determinants of the vaccine uptake rate, vaccine hesitancy, and consequently vaccine distribution success.

Risk perception: the subjective judgement that people make about the characteristics and severity of a risk- focused on risk severity and risk susceptibility. This was classified as either low or high based on how severe a participant stated they felt towards potential for COVID-19 infection.

2.7.2. Independent Variable

Risk communication sources: comprise various sources including mass media (radio, television, daily newspapers, social media (X, Facebook), official institutions, and professionals (WHO, CDC, government websites, health professionals such as doctors and nurses etc.).

Socio-demographic characteristics: include age, gender, level of education, county of residence, area of residence, occupation, income level, and compliance behaviors of respondents. The non-employed were individuals who were not salaried in any manner.

2.8. Data Collection Procedure

Administrative approval, and permission were sought and obtained from the administrative offices of the selected sub-counties in the study who each issued a signed permit for the data to be collected in those locations. The approval and permission to conduct the study were obtained from the University Of Eastern Africa, Baraton Institutional Research Ethics Committee (IREC) (UEAB/REC//03/2021), on 29 March 2021, and the National Commission for Science, Technology and Innovation (NACOSTI), License number: NACOSTI/P/21/10100, Nairobi, on 29 April 2021, respectively. Informed consent was sought and obtained from the study participants. The informed consent ensured that the study adhered to ethical considerations in how it was conducted. The study participants were asked about their willingness to participate in the study after being given information on the purpose of, and procedures involved in the study. The respondents were given all the relevant information about the study to be undertaken to allow for voluntary consent without coercion, pressure, or undue enticement. Those who accepted participation signed an informed consent form or pressed a thumb when one was not able to sign. The participants were also informed that their participation in the study was voluntary, and anyone could withdraw from the study at any time without repercussions or any penalty. Six trained research assistants collected the study data for a period of two months from May–August 2021 after the second COVID-19 wave in Kenya. The structured questionnaire was self-administered, but those who were unable to read and write were assisted by the research assistants.

2.9. Data Analysis

The quantitative data collected were cleaned, coded, and entered in the IBM SPSS version 26 program. Data were cleaned and checked for any errors in data entry. Data analysis was performed using descriptive statistics and inferential statistics. Descriptive statistics were used to present the demographic characteristics of the study population and were presented as frequencies, proportions, means, and standard deviations. Logistic regression analysis was used to estimate the association (crude odds ratio) between COVID-19 acceptance, compliance with COVID-19 mitigation behaviors, perceived severity, perceived susceptibility, demographics, and risk communication (p-value < 0.05). p-value ≤ 5% is considered statistically significant at 95% CI. Perceived severity was assessed with five items and a score equal to or greater than 3 was considered high perception and fewer than 3 was considered low perception. Perceived susceptibility was measured using 3 items (a high chance of getting infected, better ways to prevent disease than with vaccine, use of alternative medicine eliminates the need for vaccination), and a score equal to or greater than 2 was considered high susceptibility and less than 2 was considered low susceptibility.

All covariates were added to the model simultaneously. The significance level was set at 0.05. Data are presented in tables, a narrative report, and a graph.

3. Results

In this analysis, the focus was on participants with data on all relevant study variables related to compliance behaviors, and the sample size was n = 806 (100%) participants in the study. No participants had missing information on key demographic data communication, risk perception, compliance behaviors, and vaccine acceptance.

3.1. Demographic Characteristics of the Respondents

As shown in

Table 2. Demographic characteristics of the respondents.

Variable	Frequency (No.)	Percentage (%)
Gender
Male	442	55
Female	364	45
Age
18–30	305	37.8
31–40	264	32.8
41–50	96	11.9
51–60	76	9.4
61–70	45	5.5
>70	20	2.5
Education level
No Education	67	8.0
Primary	106	13.2
Secondary	335	42.9
University	298	36.9
County of residence
Kakamega	406	50.4
Kisumu	251	31.1
Vihiga	149	18.5
Area of residence
Rural	461	57.3
Urban	345	42.8
Occupation
Civil Servant	113	14
Private Sector	96	11.9
Self-Employed	190	23.6
Non-Employed	407	50.5

below, most study participants were men (55.0%), aged 18 to 30 years old (37.8%), and possessed a secondary level of education (42.9%). More participants (57%) lived in rural areas, and slightly over half (50.5%) of the participants were employed. The mean age of the participants was 35.9 (SD = 13.07), and about 70% of the participants were in the age range of 18–40 years, while the modal age group was 18–30 years. Slightly over half (52.6%) of the participants resided in Kakamega County, followed by 33.9% who resided in Kisumu County, while 13.4% of the participants were from Vihiga County.

3.2. Compliance with Safety Measures and Vaccine Acceptance

In terms of compliance with WHO safety measures, on average, the participants complied with the safety measures (mean = 5.15). A majority of the participants wore face masks (85.3%); practiced hand hygiene (78.9%); avoided hand shaking, hugging, and kissing (66.6%); and practiced cough hygiene (66.0%). Compliance with other WHO COVID-19 guidelines consisting of travel and lock-down policies accounted for 62.7%. Social distancing was least adopted (60.2%). Overall, the vaccine acceptance rate was low, at 40% of the participants.

3.3. Risk Communication

Table 3. Risk communication.

Information Source	Yes	No
Social media	549 (68.1)	257 (31.9)
Mass media	694 (86.1)	112 (13.9)
Relatives and friends	422 (52.5)	382 (47.5)
Health workers	638 (72.9)	167 (20.7)
The Internet	517 (64.1)	289 (35.9)
Government officials	502 (62.3)	304 (37.7)
Political leaders	312 (38.7)	494 (61.3)
Print media	505 (62.7)	301 (37.3)

below shows that, on average, each respondent received COVID-19 information from eight information and communication channels (mean = 4.86). Many respondents received information through mass media (86.1%) and health workers (72.9%). Further, 68.1% of the respondents received information via social networks, mainly via Facebook and Twitter (now called X), while those who received information via government sources comprised 62.3%. Political leaders were the least utilized and trusted source of COVID-19 information (38.7%).

3.4. Trusted Sources of Information

Figure 1 below shows that trusted information sources were health care professionals (86.7%), the World Health Organization (86.1%), the Center for Disease Control and Prevention (CDC—Africa regional Office) (85.5%), newspapers (81.1%), local radio and TV news stations (77.3%), Government websites (75.3%), and family and friends (63.9%). Facebook (50.6%) and Twitter (46.4%) were the least trusted sources of information.

Table 4. Risk communication and risk perception, compliance with COVID-19 mitigation measures, and intention to vaccinate against COVID-19.

	Risk Communication Source
	Social Media		Mass Media		Official Institutions and Professionals		OR (95% CI)	p-Value
Perceived susceptibility	Yes	No	Yes	No	Yes	No
High	32 (18.2)	157 (81.8)	177 (92.2)	15 (7.50)	181 (94.3)	11 (5.2)	1.3 (1.17, 5.86)	<0.001
Low	122 (19.9)	492 (80.1)	562 (91.5)	52 (8.5)	582 (95.3)	29 (4.7)	ref
Perceived severity
High	32 (22.5)	110 (72.50)	131 (92.3)	11 (7.7)	630 (95.9)	6 (4.1)	1.6 (1.06, 2.45)	<0.001
Low	125 (18.8)	539 (81.2)	630 (94.9)	56 (8.4)	136 (94.9)	34 (5.1)	ref
Compliance with COVID-19 mitigation measures
Compliant	121 (18.5)	533 (81.5)	597 (91.3)	57 (8.7)	624 (95.4)	30 (4.6)	2.7 (1.47, 3.09)	0.01
Non-Compliant	36 (23.4)	116 (76.5)	143 (93.4)	10 (6.6)	142 (93.4)	10 (6.6)	ref
Intention to vaccinate
YES	47 (23.2)	156 (76.5)	176 (86.7)	27 (13.3)	183 (90.1)	20 (9.9)	1.8 (1.51, 2.18)	0.01
NO	110 (18.2)	493 (81.8)	563 (93.4)	40 (6.6)	583 (96.7)	20 (3.3)	ref

below shows risk communication by trusted source and its role in risk perception, compliance with COVID-19 mitigation measures, and vaccination. Overall, there was significant influence of risk communication on risk perception in terms of risk susceptibility and risk severity (p = 0.001), compliance to COVID-19 mitigation measures (p = 0.001), and vaccination acceptance (p = 0.001). Despite low perceived risk susceptibility and vaccination rates, the results indicate that trusted sources of information significantly influenced risk perception. Specifically, exposure to mass media was associated with an increased risk susceptibility (OR = 1.8, CI = 1.51–2.18; p = 0.001), while information from official institutions and professionals was linked to a decreased perception of risk (OR = 0.8, CI = 0.75–0.97; p = 0.001).

Differences in perceived severity due to mass media and official institutions or professionals as trusted sources of communication were significant (p = 0.001). The percentage of participants who trusted mass media and official institutions and professionals as a source of communication had high perceived severity, representing 82.3% and 82.2%, respectively, compared to only 16.9% who trusted social media as a source of communication. Greater percentages of lower perceived severity were found among participants who trusted social media (79.6%).

Further data reveal that risk communication via official institutions and professionals and mass media raised participants’ perception of risks and compliance with WHO COVID-19 mitigation measures. The proportion of those who complied with the mitigation measures was highest among those who received and trusted information via official institutions and professionals and mass media compared to social media (81.5% and 80.8% vs. 77.1%), resulting in increased odds of 2.7 times and 2.5 times, respectively, for those who received and trusted information from official institutions and professionals and mass media, respectively.

Table 5. Logistic regression analysis of the association between demographic characteristics and risk perception, compliance with WHO public health measures, and vaccination.

		Risk Perception		Compliance with WHO Public Health Measures		Vaccination
		OR (95% CI)	p-Value	OR (95% CI)	p-Value	OR (95% CI)	p-Value
Gender	Male	2.9 (1.9, 3.9)	0.05	1.0 (0.65–1.7)	0.87	1.5 (1.0–2.1)	0.03
	Female	ref
Age	18–30	2.6 (1.6, 4.1)	0.001	2.0 (0.3, 13.1)	0.47	3.5 (0.4, 32.3)	0.27
	31–40	2.8 (2.1, 3.8)	0.001	2.4 (0.2, 6.7)	0.59	4.4 (0.5, 40.2)	0.19
	41–50	2.8 (2.0, 3.7)	0.001	1.9 (0.3, 13.3)	0.42	3.4 (0.36, 32.1)	0.28
	51–60	2.9 (1.9, 3.9)	0.001	1.9 (1.6, 3.5)	0.05	2.6 (0.26, 25.9)	0.40
	61–70	2.9 (2.3, 3.7)	0.001	2.7 (1.3, 4.9)	0.02	4.3 (1.4, 7.4)	0.02
	Above 70	ref		ref		ref
Level of Education	None	2.9 (2.2, 3.7)	0.001	1.2 (1.1, 4.5)	0.001	2.3 (1.2, 4.5)	0.02
	Primary	2.8 (2.2, 3.7)	0.05	1.3 (1.1, 6.2)	0.002	1.7 (1.4, 8.3)	0.08
	Secondary	2.9 (1.9, 4.1)	0.001	0.7 (0.4, 4.4)	0.32	1.3 (1.1, 1.8)	0.01
	Tertiary	ref		ref		ref
County	Kakamega	2.9 (2.4, 3.6)	0.001	0.4 (0.2, 0.7)	0.001	0.6(0.3, 0.9)	0.02
	Kisumu	2.6 (2.3, 3.9)	0.001	0.2 (0.1, 0.6)	0.43	0.9 (0.2, 4.8)	0.28
	Vihiga	ref		ref		ref
Occupation	Civil servant	2.4 (1.4, 4.3)	0.001	1.4 (0.4, 4.3)	0.59	0.6(0.3, 3.4)	0.26
	Private sector	2.9 (2.9, 6.5)	0.001	4.0 (3.0–6.5)	0.05	0.6 (0.2, 0.8)	0.02
	Self employed	2.8 (1.9, 3.9)	0.001	1.1 (0.4, 2.5)	0.92	0.7 (0.3, 8.4)	0.29
	Unemployed	ref		ref

presents the logistic regression analysis that was run to assess any associations between the demographic characteristics and the outcome variables. The study findings revealed that demographic factors played a critical role in influencing participants’ perception of risk and eventual public health action(s). Specifically, examination of the influence of different demographic factors on risk perception compared to respective reference categories shows that the odds for higher risk perception were significantly increased with higher age groups (above 50) (OR = 2.9; 95% CI: 1.9, 3.9), men (OR = 2.9; 95% CI: 1.9, 3.9), and private-sector workers (OR 2.9; 95% CI; 2.9, 6.5). Further, the odds of significantly higher risk perception were 2.9 times greater for those who had no education and also for those who had completed secondary school (OR = 2.9, CI = 2.2, 3.7), (OR = 5.320, CI = 2.9 1.9, 4.1) compared to participants who completed tertiary education.

Notably, participants aged 61–70 years, those with no formal education, those who completed primary school, and private-sector workers were more likely to comply with WHO public health measures. Specifically, the odds ratios indicated that individuals aged 61–70 had an OR of 2.7 (95% CI: 1.3–4.9), while those with no formal education had an OR of 1.2 (95% CI: 1.1–4.5). Participants who completed primary school had an OR of 1.3 (95% CI: 1.1–6.2), and private-sector workers showed the highest likelihood of compliance with an OR of 4.0 (95% CI: 3.0–6.5). These findings suggest a greater likelihood of compliance among these groups compared to individuals over 70 years, those with higher educational attainment, and employees in sectors other than the private. Furthermore, men (OR = 1.5; CI = 1.0–2.1; p = 0.03), motorcycle riders (OR = 5.22; CI = 3.07–12.67; p = 0.03), those aged between 61 and 70 years (OR = 4.3; CI = 1.4, 7.4); p = 0.02), and participants who worked in the private sector (OR = 0.6; CI = (0.2, 0.8); p = 0.02) had a significant association with increased intention for vaccination. Those aged below 60 years, and civil servants were not associated with increased intention for vaccination.

4. Discussion

This study examined the influence of risk communication and demographic characteristics on risk perception and compliance with WHO-recommended COVID-19 mitigation measures in selected rural communities in Western Kenya during the pandemic. Overall, compliance levels were encouraging: 85.3% reported wearing face masks, 78.9% practiced hand hygiene, and 66% adhered to respiratory etiquette by avoiding close physical contact such as shaking hands, hugging, and kissing. These compliance rates were broadly consistent with findings from other studies conducted in the region [37].

Among the public health measures, social distancing recorded the lowest compliance rate (60.2%). Nevertheless, this figure closely mirrors the 59.2% reported in certain parts of Ethiopia, suggesting a potentially broader regional trend [38]. Prior literature highlights several factors that undermine adherence to social distancing, including psychosocial and sociocultural influences (e.g., fear of stigma, misinformation, distrust in authorities, and cognitive dissonance between personal experiences and public messaging) [37]. Also, governmental shortcomings included poor community engagement, inadequate communication, and limited resources or emotional support for some who may be struggling with immense apprehension of the unknown [37]. Furthermore, studies across Africa have emphasized the impact of cultural norms, with some individuals perceiving social distancing practices as socially alienating or contrary to communal values of staying together especially in difficult times [39,40,41].

In relation to sources of information on COVID-19, mass media (86.1%) and health workers (72.9%) emerged as the primary sources for risk communication. Social media platforms, primarily Facebook and X (formerly Twitter) also played a substantial role, with 68.1% of respondents reporting them as sources of information. These findings align with Hailu et al. [37], although their study observed a higher reliance on social media (80%). Similarly, another multi-country study across sub-Saharan Africa reported comparable trends across age, gender, and regional lines [40]. According to Ehigie et al. [42], these patterns underscore the critical need to strategically leverage social media to disseminate accurate information, counter misinformation, and build trust in public health messaging during public health emergencies (PHEs). Effective risk communication should consider not only the information channels but also the evolving perceptions, attitudes, and structural barriers that shape individuals’ behavior and decision-making during the different phases of a pandemic [42].

Despite the widespread use of social media, it was the least trusted medium, with only 50.6% and 46.4% of respondents expressing confidence in information received via Facebook and X, respectively. While these trust levels are not particularly high, probably due to the prevalence of COVID-19 misinformation, they are still significant enough to expose a substantial portion of the population to unreliable information sources [43,44].

Demographic characteristics played a notable role in shaping risk perception. Older age (50 years and above) was associated with nearly threefold higher odds of perceiving COVID-19 as a risk, consistent with other studies in sub-Saharan Africa [33]. This is likely attributable to widespread public awareness of the increased vulnerability of older adults to severe illness and mortality from COVID-19 infections. Interestingly, male participants in this study exhibited higher risk perception than females, an observation that contrasts with findings from other SSA countries where gender differences in risk perception were minimal, suggesting context-specific gender dynamics in public health behavior [45].

An unexpected but compelling finding was the influence of educational attainment on risk perception. Participants with no formal education and those with only secondary education exhibited significantly higher risk perception, −2.9 and 5.32 times greater, respectively–compared to those with tertiary education. While this deviates from typical expectations that higher education correlates with greater awareness and understanding, similar patterns have been reported in a study spanning seven SSA countries [46]. This paradox could reflect a higher sense of vulnerability or lower levels of skepticism among less educated individuals, possibly making them more receptive to health messaging for desired behaviors and actions. Notably, higher risk perception in these groups also translated into higher compliance with mitigation measures. These findings support the hypothesis that risk perception mediates compliance, aligning with results from Ethiopia by Asnakew et al. [39].

Risk communication via trusted sources, particularly mass media, significantly increased both risk perception and compliance. Participants who trusted mass media sources demonstrated the highest compliance (80.8%), with a 2.5-fold increase in odds. This result is consistent with findings from Nigeria, where 60% of respondents indicated that mass media influenced their compliance with public health guidelines, particularly mask-wearing [47]. Other studies confirm the effectiveness of mass media in promoting public health compliance through both direct messaging and indirect pathways such as enhanced public understanding and elevated risk perception [48,49].

These findings underscore the strategic value of mass media in public health campaigns. Mass media channels can enhance outreach, deliver targeted and persuasive messages, foster peer influence, and strengthen community engagement to drive positive behavioral change. For future PHEs, particularly in low-resource settings, timely, credible, and culturally sensitive messaging disseminated through trusted channels will be essential. In this regard, vaccine education efforts should prioritize mass media to ensure the public receives accurate and transparent information about vaccine safety, efficacy, and availability [33,50].

Study Strengths and Limitations

This study has several strengths. First, the present findings are based on a large sample, enhancing representativeness across diverse communities and counties and thereby improving the generalizability of the results. Second, this study employed a priori power analysis to determine the required sample size, enabling investigators to ensure sufficient statistical power prior to data collection. As a result, this study was adequately powered to detect meaningful effect sizes.

Third, this study contributes to showing the importance of contextual impacts in dealing with pandemics and specifically risk communication as the case of this study. Hence, although Kenya’s COVID-19 measures largely aligned with WHO recommendations, important regional differences in public health policies and communication strategies may influence levels of noncompliance. Consequently, variations in compliance behavior are likely shaped by localized campaigns and differing official guidance across regions.

Nonetheless, several limitations should be acknowledged. First, the sample was predominantly drawn from western Kenya and was slightly skewed toward younger individuals and men depending on whoever volunteered to participate, which may limit the broader representativeness of the findings. Second, cross-sectional design limits the ability to draw causal inferences and is susceptible to bias from unmeasured confounding variables, particularly as the pandemic situation evolved. Third, compliance behaviors were self-reported, relying on participants’ ability and willingness to accurately recall and disclose their actions. Such self-reports are prone to social desirability and recall biases, which may affect the accuracy of the data.

Furthermore, risk perception is inherently dynamic and influenced by multiple factors, meaning the findings should be interpreted with caution. Additionally, the purposive selection of the study area, though well informed, systematically excluded other regions of the country, which further limits the generalizability of the results. However, the large sample size, combined with the use of random sampling in participant selection, strengthens the study’s internal validity and statistical power.

5. Conclusions

The study reveals that a majority of the participants effectively adhered to one or more WHO preventive measures to avoid COVID-19 infection. The primary source of trusted information for participants was television, followed by newspapers. Risk communication through official institutions, professionals, and mass media enhanced participants’ risk perception and compliance with COVID-19 mitigation measures.

Demographic factors significantly influenced participants’ risk perception and subsequent public health actions. Higher risk perception was notably associated with older age groups (above 50 years), male private-sector workers, and individuals with lower education levels. In contrast, younger individuals and those with lower education levels exhibited lower adoption of preventive measures.

Evidence suggests that increased risk perception correlates with heightened preventive behavior among the general population. Therefore, targeted risk communication and awareness campaigns are essential to bolster desired public health practices, particularly among demographic segments with lower adoption rates. Focusing on younger individuals and those with less education could improve overall adherence to preventive measures and ensure sustained public health practices throughout the COVID-19 pandemic while supporting those who were compliant with sustaining the behaviors. Governments and health organizations should prioritize investment in rural communication infrastructure. This can be done through continuous engagement with communities to build trust and ensure consistent and accurate dissemination of information and monitoring and evaluation of communication strategies to identify and address gaps in real time. This means understanding people’s changing perceptions and attitudes and the barriers and enablers’ influencing their ability and motivation to adopt and/or sustain positive health behaviors is critical during and after PHEs. Effective coordinated risk communication in rural communities is a dynamic, iterative process requiring multi-sectoral, and multi-dimensional approaches considerate of contextual factors and active community engagement.