public health – PLOS Currents Outbreaks http://currents.plos.org/outbreaks Wed, 07 Nov 2018 23:18:12 +0000 en-US hourly 1 https://wordpress.org/?v=4.5.3 Enhancing Ebola Virus Disease Surveillance and Prevention in Counties Without Confirmed Cases in Rural Liberia: Experiences from Sinoe County During the Flare-up in Monrovia, April to June, 2016 http://currents.plos.org/outbreaks/article/enhancing-ebola-virus-disease-surveillance-and-prevention-in-counties-without-confirmed-cases-in-rural-liberia-experiences-from-sinoe-county-during-the-flare-up-in-monrovia-april-to-june-2016/ http://currents.plos.org/outbreaks/article/enhancing-ebola-virus-disease-surveillance-and-prevention-in-counties-without-confirmed-cases-in-rural-liberia-experiences-from-sinoe-county-during-the-flare-up-in-monrovia-april-to-june-2016/#respond Thu, 09 Nov 2017 12:30:10 +0000 http://currents.plos.org/outbreaks/?post_type=article&p=76680 Introduction: During the flare-ups of Ebola virus disease (EVD) in Liberia, Sinoe County reactivated the multi-sectorial EVD control strategy in order to be ready to respond to the eventual reintroduction of cases. This paper describes the impacts of the interventions implemented in Sinoe County during the last flare-up in Monrovia, from April 1 to June 9, 2016, using the resources provided during the original outbreak that ended a year ago.

Methods: We conducted a descriptive study to describe the key interventions implemented in Sinoe County, the capacity available, the implications for the reactivation of the multi-sectoral EVD control strategy, and the results of the same. We also conducted a cross-sectional study to analyze the impact of the interventions on the surveillance and on infection prevention and control (IPC).

Results: The attrition of the staff trained during the original outbreak was low, and most of the supplies, equipment, and infrastructure from the original outbreak remained available. With an additional USD 1755, improvements were observed in the IPC indicators of triage, which increased from a mean of 60% at the first assessment to 77% (P=0.002). Additionally, personnel/staff training improved from 78% to 89% (P=0.04). The percentage of EVD death alerts per expected deaths investigated increased from 26% to 63% (P<0.0001).

Discussion: The low attrition of the trained staff and the availability of most supplies, equipment, and infrastructure made the reactivation of the multi-sectoral EVD control strategy fast and affordable. The improvement of the EVD surveillance was possibly affected by the community engagement activities, awareness and mentoring of the health workers, and improved availability of clinicians in the facilities during the flare-up. The community engagement may contribute to the report of community-based events, specifically community deaths. The mentoring of the staff during the supportive supervisions also contributed to improve the IPC indicators.

]]>

Introduction

The Ebola virus disease (EVD) outbreak started in Guinea in 20131,2, and as of June 10, 2016, 28,616 cases had been registered, with 11,310 deaths32. After the end of the original outbreak, in the three most affected countries in 20152, specifically, in Liberia in May, Sierra Leone in November, and Guinea in December, different flare-ups were reported. The biggest flare-up was in Guinea, which occurred from February 274 to June 1, 2016, with 10 reported cases and seven deaths3, while the smallest one was in Sierra Leone, from January 14 to March 17, 2016, with two reported cases and one death5.

Flare-ups may occur due to importation, reintroduction of the virus from an animal reservoir, a missed chain of transmission, or reemergence of the virus from a survivor4,6,7,8,9,10, and can be easily detected when EVD surveillance, including community-based surveillance and laboratory capacity, is established11. EVD flare-ups can also be controlled on time when a multi-sectorial EVD control strategy is implemented effectively. This strategy involves different committees, including clinical case management, surveillance, laboratory, logistic, behavioral and social interventions, psychosocial support, coordination, and others11.

Liberia reported three flare-ups after the initial declaration of “disease-free” status on May 9, 20152,12,13, the first one being from June 29 to September 3, 2015 in Margibi County14, which occurred after the re-emergence of the virus from a survivor through sexual contact12,14, and the second one being in Duport road, Monrovia, from November 24, 201515,16 to January 14, 2016, which started from a pregnant Ebola survivor who became infectious when her immune system weakened due to the pregnancy15. The two flare-ups were detected through a postmortem swab tested for EVD14,15. The last flare-up, which occurred from April 1 to June 9, 2016, was imported from Guinea3. All the flare-ups were detected early, during the 90 days of enhanced EVD surveillance recommended after the end of the outbreak, which includes the swabbing of all dead bodies for EVD laboratory investigation17. The active Incident Management System (IMS) for coordination, a temporary field-based Emergency Operation Center (EOC), implementation of the rapid response plan developed to respond to eventual flare-ups, and the presence of experienced staff trained during the original outbreak contributed to early containment of the flare-ups14.

Sinoe County reported 22 confirmed EVD cases and 11 deaths during the original outbreak, which was controlled with the isolation of the cases18, establishment of Ebola task forces, training of the staff, and other strategies19,20.The last confirmed case died in December 2014 and no flare-up was reported in the county. However, during the flare-ups in Liberia and neighboring countries, Sinoe County reactivated the multi-sectorial EVD control strategy in order to be ready to respond to the eventual reintroduction of cases.

This paper describes the impacts of the interventions implemented in Sinoe County during the last flare-up in Monrovia, using the resources provided during the original outbreak, in order to be ready to respond to the eventual importation of cases.

Methods

Setting

Sinoe County, one of the southeastern counties in rural Liberia, is divided into 10 health districts, four of which have a history of EVD positive cases reported during the original outbreak (Fig. 1)21, and 305 communities. The capital city, Greenville, is located at about 150 miles from the capital of Liberia, Monrovia. The population is dispersed, with 104,932 inhabitants and a density of 27 people per square mile22. It is difficult to reach many communities on account of forests, rivers, swamps, and hills, and the average distance from community to healthcare facility is 6.6 km21,23. However, in each of the communities, there is a community health volunteer who is responsible to provide voluntary health services such as referrals, health awareness, treatment of simple malaria, diarrhea, common cold, and pneumonia in children within the age range of 2months to 5 years. During the original outbreak, when Sinoe County reported confirmed cases of EVD, 250 of the community volunteers were trained in contact tracing and deployed immediately in communities with contacts, including Poly town community where they detected a secondary chain of transmission. The county is also served by two medical doctors, 18 physician assistants, and 67 nurses, and it has 35 health facilities, including one referral hospital with a capacity of 100 beds, and 34 clinics21.

Map of Sinoe County, Liberia, showing the number of Ebola virus disease cases reported per district during the original outbreak in 2014

Fig. 1: Map of Sinoe County, Liberia, showing the number of Ebola virus disease cases reported per district during the original outbreak in 2014.

Study design

We conducted a descriptive study to describe the key interventions implemented in the county from April 1 to June 9, 2016, the capacity available, the implications for the reactivation of the multi-sectoral EVD control strategy, and the results of the same. We also conducted a cross-sectional study to analyze the impact of the interventions on surveillance and infection, prevention, and control (IPC).

Data analysis

We entered the data into MicrosoftTM Excel and used MedCalc® Statistical Software version 17.224 for IPC and EVD surveillance data analysis. To determine statistical significance, we performed Student’s t-test for the IPC assessment data and the chi-squared test for the EVD surveillance data. We calculated the number of expected deaths using the crude death rate in Liberia, of 8.8 deaths/1000 population/year25,26.

Key interventions

We reactivated the different committees involved in the EVD control activities, as recommended by the WHO (Fig. 2)11.

Fig 2

Fig. 2: The organizational structure of the Ebola virus disease control activities highlighting the committee activated in Sinoe County, Liberia, during the flare-up of 2016.

Coordination

The county health team (CHT) developed an emergency plan to prepare for responding to the eventual importation of cases from Monrovia using the resources provided during the original outbreak. When the end of the flare-up was declared, we analyzed the level of implementation of the plan as well as the costs involved.

The coordination committee conducted meetings with county, district, and community stakeholders, partners, and other line ministries for coordination, awareness, and advocacy, in order to mobilize resources.

Logistics, human resources and, Infection Prevention and Control assessments

We analyzed the human resources database to assess how many Rapid Response Team (RRT) members, contact tracers, and burial team members trained during the original outbreak were available in the county during the last flare-up.

Using the minimum standards assessment tool developed during the original outbreak (Appendix: S1 File), we conducted the first integrated assessment of the logistic capacity available at 30 (88%) health facilities in the county, including the referral hospital, in April 2016. However, five (14%) of the 35 health facilities in county were inaccessible.

After the initial assessment we replenished the supplies, mentored the health care workers, and conducted the second assessment from the end of May to October, 2016 to verify the changes. The assessments were conducted through direct observation, interviews of the healthcare workers, and perusal of the documents available (Appendix: S1 File).

We also assessed the 10 most important check points connecting Sinoe County and other counties to verify the knowledge of the staff and availability of IPC supplies.

Clinical case investigation, surveillance, and laboratory

We conducted supportive supervisions at 30 (88%) health facilities in Sinoe County and held weekly meetings with district health officers and district surveillance officers to reinforce the triage of all the patients, the use of the EVD outbreak case definitions, and to analyze the EVD surveillance situation in the county. Besides using the WHO case definition for the investigation of cases (Fig. 3)27 we collected swabs of all the dead bodies, regardless of the cause of death, to perform real-time quantitative reverse transcription PCR (qRT-PCR) before and during the flare-up, as part of the 90 days of enhanced surveillance implemented after the end of each EVD outbreak.

Fig 3

Fig. 3: The World Health Organization’s case definition of Ebola virus disease used during the outbreaks and used in Sinoe County before and during the flare-up in Liberia, from April to June 2016.

The specimens were collected by trained clinicians and lab staff in all the health facilities in the county and were transported to the regional labs in Liberia with the capacity to perform the qRT-PCR for EVD. We perused all the lab records to quantify the numbers of specimens collected before and during the flare-up.

Contact Tracing

We conducted a rapid assessment of availability of contact listing, follow-up, reporting and monitoring forms, as well as the availability of staff with experience in contact tracing supervision. The partners were mobilized to provide logistic support for contact tracers and supervisors as well as food and water for the contacts.

Behavioral and social interventions

We disseminated EVD prevention messages through radio talk shows with county authorities and traditional leaders at a local radio station. We also conducted community meetings in high-risk communities, churches, mosques, and funeral and healing homes to increase awareness and to encourage the reporting of community deaths to the health facilities.

Ethical considerations

Ethical approval was not required to implement the activities, as they were part of the public health interventions of the Ministry of Health to respond to outbreaks in Liberia. We did not use any confidential data and did not disclose any unauthorized names in our report.

Results

Coordination

The county’s task force for EVD was activated and was responsible to ensure that the preventive measures were implemented at all levels and that any suspected case was promptly reported. As the resources were already available from the original outbreak, the three-month plan cost an additional USD 1755 besides the budget for the routine activities. The dissemination of EVD prevention messages to churches, mosques, households, meetings in high-risk communities, and funeral and healing homes was the most expensive activity, at about USD 400. This was mostly attributable to purchasing fuel for the activities and paying daily subsistence allowance (DSA) for the staff (Table 1).

Table 1

Table 1: List of the main activities implemented routinely and during the flare-up of Ebola virus disease prevention and surveillance in Sinoe County, Liberia, in 2016.

Logistics, human resources, and Infection Prevention and Control (IPC) assessments

At the first assessment, we included 30 health facilities (29 clinics and one hospital) in our analysis of the IPC indicators. From this analysis, we excluded the indicators that were not applicable to clinics (Appendix: S1 File), according to the tool used. Personnel/staff training was the group of indicators with the highest scores. The mean of the four individual indicators in this group was 78% [standard deviation (SD) = 11%] including 90% of the health facilities with staff trained in IPC by the Ministry of Health and Social Welfare (MOHSW) in a training called “keep safe keep serving.” Additionally, 63% had staff meeting the criteria outlined in the MOHSW’s Essential Package of Health Services (EPHS), the minimum skills required to work in the facilities (Table 2). On the other hand, the group of four indicators for triage had a mean of 60% (SD = 12%), and the other three indicators assessing the facilities in terms of having an appropriate isolation space ready to receive cases had a mean of 52% (SD = 9%). The structure of the community care center in the Karquekpo community, with a capacity of 12 beds, and the ETU in Greenville district, with a capacity of 60 beds, remained intact, but these facilities required equipment and supplies to start receiving patients. Out of the facilities assessed, 27 (90%) had IPC supplies that would last them for one month. These supplies were provided during the original outbreak.

*The mean percentage was calculated for each group of indicators; †P-value was calculated for each group of indicators using a paired t-test and was considered statistically significant when ≤ 0.05; ‡n = number of the health facilities where the indicator was assessed. In total, 17 health facilities (n = 17) were assessed and reassessed with the same tool, while 13 health facilities were reassessed with a new tool that excluded some indicators; §SD = standard deviation; it was calculated based on a sample for the group of indicators; ¶Indicator was not included in the new tool used for the reassessment of 13 clinics; **Indicator was not included in the new tool for reassessment but was assessed in two clinics where the tool was used.

Table 2: Performance of the Infection Prevention and Control indicators in Sinoe County at the beginning (baseline) and at end (reassessment) of the flare-up in Monrovia, Liberia, in 2016. *The mean percentage was calculated for each group of indicators; †P-value was calculated for each group of indicators using a paired t-test and was considered statistically significant when ≤ 0.05; ‡n = number of the health facilities where the indicator was assessed. In total, 17 health facilities (n = 17) were assessed and reassessed with the same tool, while 13 health facilities were reassessed with a new tool that excluded some indicators; §SD = standard deviation; it was calculated based on a sample for the group of indicators; ¶Indicator was not included in the new tool used for the reassessment of 13 clinics; **Indicator was not included in the new tool for reassessment but was assessed in two clinics where the tool was used.

In the second assessment, significant improvements were observed on the indicators of triage (an increase from 60% during the first assessment to 77% on reassessment; P = 0.002) and personnel/staff training (an increase from 78% to 89%; P = 0.04), while other indicators did not exhibit significant improvements.

The lab supplies for the oral swab of dead bodies and whole blood for live alert investigation were available in all the facilities (100%) during the first and second assessments.

The county had three vehicles that were available to be used in case of an outbreak, including one ambulance for the transportation of cases.

Of the 21 members of the RRT trained during the original outbreak, 19 (90.5%) remained present in the county. All 250 (100%) community volunteers trained for contact tracing remained in their communities, and they were available to resume the task in case of need. Of the 14 members of the two burial teams trained during the previous outbreak, 12 (86%) remained present in the county.

All 10 most important checkpoints connecting Sinoe with other counties had a bucket for handwashing and thermometers, although some were not functional. After mentoring the staff and replacing the thermometers, nine (90%) check points reactivated the monitoring of temperature and handwashing for people crossing these locations.

Clinical case investigation, surveillance, and laboratory

The percentage of EVD death alerts investigated, including the oral swabs collected and sent to the regional lab in Liberia, increased from 26% of the death alerts per expected deaths during the three months before the flare-up in Monrovia, to 63% of the alerts per expected deaths during the flare-up (P < 0.0001). Significant improvement was verified in seven of the 10 health districts in the county. On the other hand, the number of live alerts investigated and the whole blood tested for EVD decreased from 0.5 alerts per 100 population to 0.4 alerts per 100 population (P = 0.0003). The reduction was significant in Greenville (P < 0.0001) and Butaw (P = 0.05) districts (Table 3). All the specimens from both live and death alerts came negative for EVD. The total number of health facilities investigating death alerts increased from 19 (Mean = 54% per health district, SD = 34%) to 28 (Mean = 83% per health district, SD = 22%) health facilities (P = 0.006) in all 10 health districts in the county (Fig. 4).

Contact tracing

All the reporting and monitoring tools used during the original outbreak were available and would be reproduced according to number of eventual contacts listed. All the county supervisors and partners, including World Health Organization (WHO) had experiences from the original outbreak. In case of eventual importation of cases, the county surveillance officer, the county monitoring and evaluation officer and the epidemiologists from WHO present in County would be in charge of contact listing as well as to provide one day orientations to the experienced community volunteers responsible for home visits to the contacts. All the 10 district health officers (DHOs) present in county also had experiences from the previous outbreak and would support the supervision of contact tracers together with county supervisors and partners. There was no formal commitment from the local partners to provide incentives for contact tracers and supervisors as well as food and water for contacts. The IPC materials and thermometers in stock in county would be used also for contact tracing in case of need.

Table 3: Live and death alerts investigated in Sinoe County before and during the flare-up in Monrovia, Liberia, in 2016 *P-value was calculated using the chi-squared test for comparing two proportions; it was considered statistically significant when ≤ 0.05

Table 3: Live and death alerts investigated in Sinoe County before and during the flare-up in Monrovia, Liberia, in 2016. *P-value was calculated using the chi-squared test for comparing two proportions; it was considered statistically significant when ≤ 0.05

OBK-17-0018 Fig 4

Fig. 4: Percentage of healthcare facilities investigating Ebola virus disease alerts in Sinoe County three months before and during the flare-up in Monrovia, Liberia, in 2016.

The error bars show the standard deviations (SD) for the percentage of health facilities per health district (total = 10 health districts) investigating at least one alert from January to March and from April to June, 2016. P = 0.3 for live alerts and P = 0.006 for death alerts in Sinoe County, Liberia, calculated using the chi-squared test for comparing two proportions; P was considered statistically significant when ≤ 0.05. N = 34 health facilities in Sinoe County.

Discussion

The interventions implemented in Sinoe County in preparation to respond to the reintroduction of EVD cases from the last flare-up in Monrovia had a significant impact on EVD surveillance, leading to an increase in percentage of death alerts from 26% alerts per expected deaths during the three months before to 63% alerts per expected deaths during the flare-up (P < 0.0001). This finding may have been affected by community engagement activities, awareness and mentoring of the health workers, and improved availability of clinicians in the facilities during the flare-up. Similar findings were reported in Lofa County, Liberia, during the original outbreak in 2014, when more community deaths were reported to the health authorities to be investigated for EVD after community sensitization and acceptance28. The reasons for reduction of live alerts were not clear from our study, and the same may have occurred owing to the increased attention focused on the death alerts during the flare-up, considering that the reporting and investigation of live alerts was already high in the county before the flare-up.

The improvements on the indicators of triage (from 60% at the first assessment to 77% on reassessment; P = 0.002) suggest that the mentoring of the staff during the first assessment led to behavioral change among health workers, leading to better triaging for EVD when people visited the facility. The on-site training of the staff also improved the indicators pertaining to personnel/staff training from 78% to 89% (P = 0.04), as these indicators do not require any interventions other than on-site mentoring. The mentoring of the staff also contributed to an improvement in the system for checking and reporting staff health issues in the facilities, and led to the permanent presence of trained clinicians whenever the facility was open. However, no significant improvement was verified in other indicators like equipment and infrastructure for isolation units, which require interventions from higher administration levels, as the procurement of supplies and construction of infrastructure may be slow because the county is not currently reporting any EVD-positive cases. A similar assessment conducted in Sierra Leone in 2014 led to a quick intervention from stakeholders and partners evidenced by the immediate provision of equipment and supplies, considering that, unlike Sinoe County in Liberia, these districts were facing an active EVD outbreak29.

The presence of trained staff in 90% of the health facilities; the small attrition of RRT members, contact tracers, and trained burial team members trained during the original outbreak; the availability of stock of IPC supplies for at least one month in 27 (90%) facilities; and other logistics, including the availability of three vehicles, rendered the reactivation of the multi-sectoral EVD control strategy fast and relatively affordable. In addition, despite the lack of supplies to attain full functionality, the presence of an isolation space for receiving patients in 43% of the health facilities, and the presence of an ETU with a capacity of 60 beds and a CCC with a capacity of 12 beds suggest that few additional interventions would be required to respond to any eventual importation of cases from Monrovia or elsewhere.

Our study has several potential limitations. First, the assessments were conducted as emergency interventions, and not enough time was available to train the personnel who conducted these assessments. However, the personnel received orientations during a two-hour meeting conducted before the assessment. In addition, a new improved tool for reassessment was introduced before the end of the flare-up. Therefore, two different tools were used for assessment and reassessment, although some indicators did not change. Thus, only 17 (49%) facilities were assessed and re-assessed using the same tool. To minimize this limitation, our analysis only included the indicators present in both tools. The time interval from assessment to reassessment was not the same for all health facilities, varying from one to five months. Thus, some facilities may have had more time to improve than others did.

Additionally, there may have been an information bias due to fatigue of the interviewed staff, as the facilities were receiving more visitors than usual during the flare-up. This may have led them to provide answers that would not require follow-up questions. Further, the interventions were implemented when the county did not report any confirmed EVD cases. Thus, some key areas of EVD response, like access to food and other supplies by contacts during contact tracing and case management, were not assessed. Our study did not determine if the supplies provided and the staff trained would remain available in the county if a flare-up occurred two or more years after the end of the original outbreak, nor did it assess the implications of this. Despite these potential limitations, our findings may be considered for assessing the preparedness for EVD and other future outbreaks, leading to improved surveillance, early detection, and control, as well as prevention of infection among health workers.

In conclusion, as part of outbreak preparedness, community engagement may contribute to the reporting of community-based events, specifically community deaths for EVD surveillance. The mentoring of the staff at health facilities, combined with the assessment of IPC, would lead to behavioral change among the health workers, thereby increasing IPC compliance and improving outbreak surveillance. The low attrition among the personnel trained in outbreak response, and presence of supplies at health facilities made easier, faster, and affordable to achieve the reactivation of the response structures.

We recommend a periodic reassessment of IPC supplies and equipment in health facilities, combined with mentoring of health workers, early advocacy for partners and stakeholders to provide the required equipment and to facilitate the construction of isolation units, and the implementation of reinforcement measures to reduce attrition among the trained health workers, especially within the first year after the end of any outbreak.

Data Availability

All data supporting this study are openly available from figshare, https://doi.org/10.6084/m9.figshare.4902929.

Competing Interests

The authors have declared that no competing interests exist.

Corresponding Author

Jeremias Naiene: naienej@who.int

Appendix

S1 File

]]>
http://currents.plos.org/outbreaks/article/enhancing-ebola-virus-disease-surveillance-and-prevention-in-counties-without-confirmed-cases-in-rural-liberia-experiences-from-sinoe-county-during-the-flare-up-in-monrovia-april-to-june-2016/feed/ 0
Vaccine Hesitancy: In Search of the Risk Communication Comfort Zone http://currents.plos.org/outbreaks/article/vaccine-hesitancy-in-search-of-the-risk-communication-comfort-zone/ http://currents.plos.org/outbreaks/article/vaccine-hesitancy-in-search-of-the-risk-communication-comfort-zone/#respond Fri, 03 Mar 2017 11:30:10 +0000 http://currents.plos.org/outbreaks/?post_type=article&p=70808 Introduction: This paper reports the findings of a national online survey to parents of children aged 5 and younger. The objectives of the study were to assess parental understanding of childhood immunizations, identify sources of information that they trust for vaccine-related content, assess where parents with young children stand on the key issues in the public debate about vaccination, and identify which risk communication messages are most effective for influencing the behaviours of vaccine hesitant parents.

Methods: A total of 1,000 surveys (closed and open-ended questions) were administered in November 2015 using the Angus Reid Forum Panel, a key consumer panel consisting of approximately 150,000 Canadian adults aged 18 and older, spread across all geographic regions of Canada.

Results: Approximately 92% of the Canadian parents surveyed consider vaccines safe and effective, and trust doctors and public health officials to provide timely and credible vaccine-related information. However, a concerning number of them either believe or are uncertain whether there is a link between vaccines and autism (28%), worry that vaccines might seriously harm their children (27%), or believe the pharmaceutical industry is behind the push for mandatory immunization (33%). Moreover, despite the common assumption that social media are becoming the go-to source of health news and information, most parents still rely on traditional media and official government websites for timely and credible information about vaccines and vaccine preventable diseases, particularly during community-based disease outbreaks. Finally, parents reported high levels of support for pro-vaccine messaging that has been demonstrated in previous research to have little to no positive impact on behaviour change, and may even be counterproductive.

Discussion: The study’s results are highly relevant in a context where public health officials are expending significant resources to increase rates of childhood immunization and combat vaccine hesitancy. The data offer insight into where parents stand on the political and public debate about mandatory vaccination, what aspects of vaccine science remain uncertain to them, which media and institutional sources they use and trust to navigate the health information environment, how they look for information and whom they trust during periods of health emergency or crisis, and which communication strategies are considered most effective in persuading vaccine hesitant parents to immunize their children.

]]>

Introduction

In March 2014, Toronto Public Health launched a media campaign to pressure the popular ABC daytime talk show, The View, to reverse its decision to hire outspoken celebrity, Jenny McCarthy, as one of its standing hosts. The concerns, outlined in a series of tweets, Facebook posts and news interviews, were that McCarthy’s well-known anti-vaccine activism would continue to undermine efforts by medical and public health authorities to increase vaccination uptake at a time when measles, whooping cough, chicken pox, mumps and other vaccine preventable diseases (VPD) have been making a comeback. McCarthy co-hosted The View for a year, and was eventually, and suddenly, released from the show in August 2015, reportedly due in part to her unpopular views about vaccines.

Anti-vaccine sentiment has emerged as a recurring topic of media attention alongside each new outbreak of vaccine-preventable disease. As the Toronto Public Health campaign illustrates, it has also become a focus of professional advocacy. While the number of Canadian parents who hold strident anti-vaccine beliefs and refuse to vaccinate their children is generally low (fewer than 3%), increasing numbers of parents (up to 35% in Canada) fall somewhere along a spectrum of beliefs and behaviours that we today call ‘vaccine hesitant’ (VH). Vaccine hesitant parents are a heterogeneous group who express worries about the risks and side effects of vaccines, including when and how frequently they should be administered. While some parents avoid vaccination altogether, others will agree to some vaccines but not all. Vaccine hesitant parents may also ultimately consent to have their kids vaccinated, but do so reluctantly because they remain uncertain about whether this is the correct course of action to take for their child’s health 1.

We argue that the uncertainty associated with vaccine hesitancy represents a ‘wicked’ risk communication problem for public health officials. Wicked problems are by their nature difficult to define, and the solutions for addressing them are often uncertain 2. Although there is broad consensus among health scientists, medical professionals, policymakers and a majority of Canadians that childhood vaccination is a vital public health intervention, there is less agreement about how how public health officials can effectively communicate with parents to address their worries, beliefs and concerns about vaccines. Parents live within and have to navigate an increasingly complex, noisy mediascape in which conflicting claims about vaccines and VPDs compete for attention and thus shape their affective environment. As illustrated by the Toronto Public Health campaign, health officials are challenged to cut through this noise to make the most persuasive case possible that vaccines are both necessary for community protection and essential to saving lives.

In large part, this is a framing problem—how do we talk about vaccine hesitancy, to what do we attribute this phenomenon, and what solutions do we consider? As Scheufele3 argues, frames are both cognitive schemata (i.e., the stories and scripts we carry around in our heads) and elements of public discourse (i.e., manifest or latent patterns of media texts and public communication). Vaccine hesitancy is commonly framed in both media and public discourse as a knowledge-deficit problem—if only parents had access to accurate, scientifically verifiable information about the safety and efficacy of vaccines then surely they would embrace the miracle of immunization. It is also often framed as an example of irrational or misinformed thinking—if only people were not so easily tricked by junk science or the power of Big Pharma and celebrity culture then they would surely be less afraid of the risks and line their kids up for their shots.

Compelling as these accounts may be, neither is empirically tenable. Rather, as we know from recent research, vaccine hesitancy is driven by a complex configuration of issues, including, but not limited to, poor public health literacy, public perceptions of risk that are disproportionate to expert assessments, cultural values, access to large amounts of conflicting information, declining trust in experts, and a host of socio-demographic factors 4,5,6,7. How media, health advocates and others frame the problem of vaccine hesitancy is of course influenced by and informs these issues. And today, public health officials struggle with how best to respond to parental worries, concerns and vaccine refusal behaviours 8.

Our objectives for this study were to know more about perceptions of Canadian parents regarding vaccines and vaccine-preventable disease, to determine what information sources they use, and which institutions they trust. We also explored the responses of parents to common risk messages intended to persuade those who are vaccine hesitant to have their kids vaccinated.

Methodology

To understand and inform strategies for addressing vaccine hesitancy we administered an online survey to 1,000 Canadians parents with young children (5 and younger) regarding immunization-related decision-making, and focusing specifically on the MMR vaccine. The survey instrument included 25 questions organized into 4 major categories: perceptions about vaccines and vaccination; views on the public debate about vaccines and vaccine-preventable disease; information seeking needs and practices, including media usage and trust in institutional sources; and communication strategies.

Participants were provided a 5-point Likert Scale (Strongly Agree, Agree, Disagree, Strongly Disagree, No Opinion/Don’t Know) to respond to the majority of questions, particularly those relating to the ‘Perceptions of Vaccination’ and ‘Public Debate’ items. We also posed three open-ended questions to provide respondents with an opportunity to: (a) explain their vaccination choices; (b) describe where they get their health information; and (c) offer suggestions of other communication strategies that might persuade vaccine hesitant parents. We had two screening questions and a battery of brief questions designed to provide demographic profiles of our respondents. The survey took approximately 15-18 minutes to complete.

Research participants were recruited in December 2015 using the Angus Reid Online Forum panel, a key consumer panel consisting of approximately 150,000 Canadian adults aged 18 and older, spread across all geographic regions of Canada. The panel is benchmarked to known Census targets, such as age, region, income, and education, to ensure a representative sample of the Canadian population. Angus Reid’s online panels are constituted by Canadians who register themselves as participants in surveys for compensation, such as cash, gift cards or vouchers. To be eligible, participants must complete a demographic profile of themselves so that a sample can be created which matches the targeted population of each study. In this case, our interest was to survey only parents with vaccine-aged children, thus the survey would have been distributed only to registered participants matching this profile. Participants are under no obligation to complete surveys once they have registered.

The participation rate for the study was 89 percent. We reached our target participation number of 1,000 respondents after distributing the survey to 1,121 participants who met our qualifying criteria.

Results

How many parents vaccinate, and what are their perceptions about vaccines?

A majority of parents we surveyed, 92 percent, have had their children immunized with the MMR vaccine. Nearly half of the 6 percent of parents whose children had not been vaccinated reported that the age of their child was the primary factor (i.e., they were too young to receive the vaccine). In Canada, the vaccine schedule varies depending on where one lives in the country. In Ontario, children receive two doses, the first at aged 12 to 15 months, and a second dose (MMRV) at 18 months or anytime thereafter and before beginning school. Among the 3.9 percent of survey respondents who indicated that they have intentionally chosen to not vaccinate children, the major concerns given were fears about the possibility of serious adverse reactions (25 percent) and skepticism about vaccine effectiveness (11 percent).

There is a strong consensus within the international public health community (WHO, European Commission, NHS, CDC, Public Health Agency of Canada, American Academy of Paediatrics, etc.) that vaccines are safe, provide tremendous benefits to children’s health, and produce relatively minor side effects. Yet, despite the strong record and scientific consensus about vaccine safety and efficacy, some parents remain skeptical, nervous and uncertain about whether vaccination is appropriate for their children. We were interested, first, to know whether parents have the same levels of understanding and confidence about vaccine safety and efficacy as medical professionals.

A majority of parents correctly answered a series of basic questions about vaccines and vaccination, although some questions elicited higher rates of incorrect and uncertain responses. More than 90 percent of parents agreed with the statement, “vaccines are safe and effective at preventing childhood illnesses” and 82 percent agreed that there is “clear consensus among medical experts that vaccines are safe.” Similarly, 83 percent also disagreed with the statement, “vaccination is less important than it used to be because we have eradicated most childhood diseases.”

Yet, our survey results also revealed some worrisome findings. First, that nearly 17 percent of parents consider vaccination “less important today than in the past” should be a source of concern given the resurgence of VPDs such as measles, mumps and whooping cough. It also reflects findings of other research showing rising rates of vaccine hesitancy attributed to beliefs about vaccine necessity 9. Most troubling, 14 percent of parents agreed with the statement, “vaccines can cause autism,” and an additional 14 percent expressed uncertainty about this statement. Second, although serious vaccine related injuries are rare, more than one-quarter of respondents agreed or were unsure about the statement, “there is a strong likelihood that the MMR vaccine will produce serious adverse reactions.” This indicates that even among parents who vaccinate, some of the most dangerous claims raised by the anti-vaccination movement—vaccines will injure your child and can cause autism—have achieved a concerning degree of resonance.

Where do parents stand on the public debate about vaccines?

Vaccine hesitancy has become a high profile news topic, as illustrated by the Toronto Public Health campaign described in the introduction to this paper. We wanted to understand where parents position themselves on the myriad issues that constitute this increasingly polarized, public debate.

Parent views on vaccines as a ‘public issue’ were more varied than they were on the knowledge of vaccine and vaccination questions. For example, the issue of mandatory vaccination was particularly contentious. Survey respondents were almost evenly split in agreement when asked, “should parents be able to choose whether their children are vaccinated?” Even though more than 90 percent of parents have had their kids vaccinated, 44 percent agreed that vaccination should still be a matter of parental choice (49 percent disagreed, and 7 percent were unsure). We also asked whether schools and daycare facilities should refuse children who are not vaccinated, except for those with medical exemptions. Approximately 65 percent of our respondents answered this question affirmatively, and 66 percent agreed with the statement, “parents who do not have their children immunized (except in cases involving medical exemptions) are irresponsible.” Furthermore, while a strong majority of parents have had their kids vaccinated, and many of them agree that there is a scientific consensus about vaccine safety and effectiveness, 33 percent agree with the statement, “drug companies are behind the government’s push for mandatory vaccination.” Here again, as with the responses to the autism link statement, a key piece of the anti-vaccine movement’s rhetorical artillery has struck a chord with parents.

Where do parents get their health information?

We asked parents an open-ended question: which media sources do you most use for health news and information? For the majority of respondents (57 percent), online news and information sources (e.g., Google, social media, websites, etc.) are their preferred source for health news and information, followed by television or radio (29 percent). A very small number of parents (5 percent) indicated that they most often use scientific sources, such as medical journals, to navigate the ever-changing health information landscape.

We also posed a series of situational questions designed to understand where parents would go first and most often for information and guidance if an outbreak of VPD were to pose a risk to their community. We expected social media sites such as Twitter or Facebook to be obvious top information sources given how many parents indicated that they use the Internet to keep abreast of health news and events, and because of the ways in which these platforms have become key sites of breaking news and collaborative news curation 10. However, this was not the case at all—only 14 percent of our respondents indicated that they would consult social media for breaking news and information if an outbreak of VPD were to threaten their communities. Instead, the website of major news organizations was the overwhelming top choice (33 percent), followed by official government websites (24 percent) and television stations (19 percent). These patterns also roughly translated to the preferred information sources that parents would consult most often.

Which institutions and sources do parents trust?

Past research is clear that the most trusted source of vaccine related information are family physicians and other medical professionals 11. The growth of the anti-vaccination movement, and anxiety among media and public health professionals about the influence of high-profile vaccine skeptics like Jenny McCarthy suggests that popular celebrities and alternative healthcare providers (e.g. naturopaths, chiropractors) may be the source of rising concern among parents about vaccine safety.

Our results show that the three most trusted sources for health information overall are physicians (89 percent trust, 7 percent do not trust), public health officials (83 percent trust, 12 percent do not trust), and academics (77 percent trust, 15 percent do not trust). By contrast, the least trusted sources are popular celebrities such as Gwyneth Paltrow and Oprah Winfrey (8 percent trust, 82 percent do not trust). Celebrity physicians such as Dr. Oz have higher levels of trust than other celebrities do (25 percent trust, 59 percent do not trust), but overall remain strongly distrusted in comparison with most other sources. This should also be a point of concern given the tendency of these sources to pedal junk science and the fears that these types of sources will influence medical decision-making 12,13. Finally, roughly equal numbers of parents expressed trust and distrust in the news media (46 trust, 47 do not trust). This finding is intriguing given how many of our respondents stated they would rely on media outlets as their main source of information during an outbreak event 14.

Trust in institutional sources is polarized when we compare parents who vaccinate their children and those who do not. Not surprisingly, levels of trust in physicians, public health authorities and academics were much higher among vaccinators than non-vaccinators: approximately 90 percent of parents whose children have been vaccinated agreed that these sources can be “trusted to do what is right” compared to only 6 percent who disagreed. Among non-vaccinators, a much smaller group overall, 55 percent expressed trust in these sources compared to 37 percent who did not. Vaccination status was also significant when considering levels of trust in the pharmaceutical industry. While drug industry distrust was generally high with both groups, it was significantly higher among non-vaccinators (80 percent compared to 51 percent). And in the case of celebrities, despite the concerns of public health authorities that parents will blindly follow the advice of celebrity anti-vaxxers like Jenny McCarthy, there was no significant difference between vaccinators and non-vaccinators: both groups agree that popular celebrities and medical celebrities are not trustworthy sources of vaccine information.

Vaccine Hesitancy: What’s to be done?

Health officials have used numerous approaches to persuade parents who do not vaccinate their children to change their views and behaviours and to reinforce positive vaccine behaviour among those who already do. Attempts to frighten parents about the risks of disease or correct false claims about vaccines have been largely ineffective, and may be counterproductive 15. Sandman and Lanard16 argue that there is a fine line between warning the public that a given risk may be potentially worrisome without actually scaring people. Yet, officials often resort to scare tactics and have occasion to use dramatic and vivid imagery to frighten or shame parents into having their children vaccinated. Images of sick children may effectively provoke fear, worry, and other emotions that can be persuasive. Yet, the use of emotionally evocative images may also strengthen beliefs in a vaccine/autism link among a core group of parents, while dramatic narratives that describe the risks to infants of under-immunization can increase self-reported beliefs about serious side effects of vaccines 15.

While our study did not involve experimental testing of risk communication interventions, we did ask participating parents to reflect on messages that public health officials often use to persuade those who are vaccine hesitant, and to indicate which, if any, they feel work best at increasing vaccine uptake among this group (Table 1).

The majority of respondents believed that all of the suggested messages, with the exception of shaming, are likely to be effective in persuading parents to have their children vaccinated. Interestingly, while almost two-thirds of our parents agreed with the statement, “parents who do not have their children immunized (except in cases involving medical exemptions) are irresponsible,” nearly as many (64 percent) believed that this message would be unlikely to change the immunization behaviours of other parents. The relationship between these two questions was moderate and statistically significant: parents who strongly disagree that those who do not vaccinate their children are irresponsible are more likely to dispute the efficacy of shame-based messaging to change vaccine hesitant behaviour. However, parents who do vaccinate their children also believe that this approach is unlikely to be effective in changing behaviour.

The messages that respondents overall felt would most likely work with vaccine hesitant parents were those that emphasized the scientific evidence showing that vaccines are safe and effective (47 percent), followed by messages about the likelihood of catching a serious childhood illness without vaccine protection (40 percent) and those which vividly detail the effects of childhood diseases (37 percent). Among the small number of parents who self-identified as holding anti-vaccine beliefs, the only message that showed any hope of effectively persuading parents like them was, “provide positive encouragement and emphasize that vaccines are strongly recommended, but ultimately the decision is theirs to make” (77 percent). All other messages generated very strong negative reactions for non-vaccinating parents, indicating they would all be unlikely to ameliorate their hesitant beliefs and behaviours.

Finally, research participants were invited to suggest other possible risk communication approaches for persuading vaccine hesitant parents to change their beliefs and behaviours. We coded the 857 discrete responses to this question into numerous other categories, of which the most commonly cited recommendation (28 percent of respondents) was, “use of research to debunk vaccine myths.” Among parents with strongly vaccine hesitant views, messages based on “showing compassion” and “communicating honestly about risk” were most common, although this represents a very small baseline of responses.

The results of our research illustrate a potential disconnect between what parents of young children believe will be effective in persuading parents who are vaccine hesitant with what the available experimental research already tells us: more evidence, statistics and debunking strategies are the least likely to work 15,17,18. Of concern, all of these approaches have been shown in other research to have little to no positive impact on vaccine uptake, and may be counterproductive. If it is assumed that parents who do not have their children vaccinated do so because they lack appropriate knowledge and information, or because they have been duped by anti-vaccine celebrities and activists, then perhaps it is not surprising that parents (and health professionals) would try to address that problem with more science, data and evidence.

The responses about vaccine risk messaging from parents who hold more strident vaccine-hesitant views are worth considering to the extent they reflect the value and importance of expressing empathy and compassion, and providing support and positive encouragement as a means for building trust with parents over the longer term, even if in the short term that does not lead to changes in immunization behaviour. Openness, dialogue, empathy, and respect are foundational values to ethical and effective risk communication 19,20,21. And while they might not yield an immediate shift in vaccination behaviours, they may, over the longer term, be our most effective protection against the wicked problem of vaccine hesitancy.

Table 1

Risk Messaging to increase vaccine uptake

Very Likely Somewhat Likely Not Very Likely Not At All Likely N=
a. Messages should emphasize scientific evidence showing that vaccines have a strong record of safety and effectiveness at reducing serious childhood illnesses. 469 (47.0%) 411 (41.2%) 94 (9.4%) 24 (2.4%) 998
b. Messages should emphasize the statistical likelihood of catching a serious childhood illness like measles or whooping cough without being vaccinated. 401 (40.1%) 449 (44.9%) 117 (11.7%) 33 (3.3%) 1000
c. Messages should use shaming techniques to persuade parents they have a moral duty to vaccinate their children and protect their community. 112 (11.3%) 248 (24.9%) 313 (31.5%) 322 (32.4%) 995
d. Messages should vividly detail the negative effects of childhood diseases, for example photographs of seriously ill children and the consequences for their parents. 365 (36.5%) 440 (44.0%) 145 (14.5%) 50 (5.0%) 1000
e. Messages should provide positive encouragement to parents and emphasize that vaccines are strongly recommended, but ultimately the decision to vaccinate their children is their choice to make. 294 (29.4%) 418 (41.8%) 212 (21.2%) 76 (7.6%) 1000

Limitations and Conclusions

There are some limitations to this study worth noting. First, although the online panel used for our survey is constructed to be representative of the Canadian population in terms of age, region of residence, income and education, selection bias and non-response bias cannot be ruled out. However, the sociodemographic characteristics of our respondents are not significantly different from those of the Canadian population of parents with children aged 5 and younger. Second, the MMR vaccination decision for the child was self-reported by parents which could lead to recall bias, and there was no other measure within the study to assess parental vaccine hesitancy attitudes along a broader spectrum. Hence, as most respondents reported that their child was vaccinated, their reflections on the standard communication messages used by public health to persuade parents about the benefits of vaccination, as well as those suggestions provided by parents that could be persuasive in encouraging parents to vaccinate their children, cannot be expected as being effective specifically for vaccine hesitant parents. Relatedly, messaging deemed to be more acceptable by anti-vaccination parents – namely, public health messaging that both strongly recommends childhood vaccinations while equally expressing empathy and compassion for parental choice – needs further empirical testing, either in an experimental design or through intensive qualitative research.

Despite these limitations, our findings make important contributions to our understanding of vaccine hesitancy and the communication challenges it presents. The data illustrate a combination of positive and troubling news for health communicators who are struggling to find the risk communication comfort zone that builds trust with hesitant parents while increasing vaccination rates. First, our survey indicates that a strong majority of Canadian parents consider vaccines safe and effective, and trust doctors and public health officials to provide timely and credible vaccine-related information, while expressing much lower trust in politicians and industry sources. At the same time, indicators of vaccine hesitancy are prevalent: a significant number of parents either believe or are uncertain about the scientifically unproven link between vaccines and autism, and a concerning number of them worry that vaccines could seriously harm their children.

Despite the widely held belief that social media are replacing legacy news organizations as a major source of health news and information, and thus present a potential threat to public health communication as a site of ‘fake news’ and disinformation, most parents of young children still rely on traditional media and official government websites for timely and credible information about vaccines. This is particularly true during outbreaks of vaccine preventable disease—there are clearly important opportunities to use these extraordinary moments to reinforce positive messaging about the benefits of vaccination. Our data also show that parents place a high level of trust in government websites and expect or assume those sites to be regularly updated with reliable and timely information. Medical and public health officials working in government departments and agencies should pay special attention to this finding.

Finally, parents report high levels of support for pro-vaccine messaging that emphasizes the science of vaccine safety and effectiveness, vividly depicts the consequences of disease, and debunks the myths and misinformation about vaccination. Future research should build on these results to test risk communication interventions with parents who occupy different standpoints along the vaccine hesitancy spectrum.

Taken together, these results contribute to a growing body of research on vaccine hesitancy and health risk communication, and are highly relevant in a context where health officials continue to struggle to strengthen rates of community protection and inspire confidence about the benefits of vaccines for protecting public health.

Competing Interests

The authors have declared that no competing interests exist.

Data Availability

Data are available from the figshare repository at the following URL: https://figshare.com/s/055c898abb39690b0871.

Corresponding Author

Josh Greenberg: joshuagreenberg@cunet.carleton.ca

]]>
http://currents.plos.org/outbreaks/article/vaccine-hesitancy-in-search-of-the-risk-communication-comfort-zone/feed/ 0