The West African Ebola virus disease (EVD) outbreak struck the vulnerable health systems of Liberia, Sierra Leone, and Guinea in 2014. Liberia, specifically, has one of the lowest physician densities in the world with just 1 physician for every 10 000 people and similarly low numbers of other clinical health care workers with only an estimated 193 midwives.1 Prior to the outbreak, the Liberian health care system had limited infrastructure, as well as limited infection prevention and control (IPC) policies, practices, and supplies. All of these factors contributed to the vulnerability of the health care system during the Ebola epidemic and resulted in the temporary closures at all levels of health care facilities throughout the country during the peak of the outbreak. This not only decreased the chances of surviving Ebola due to late diagnosis and treatment, but also presumably increased morbidity and mortality of other common, treatable diseases due to decreased access to care.
The EVD outbreak contributed significantly to the critical shortage of skilled health care professionals in an overall weakened health system. One hundred ninety-three health care workers died from confirmed Ebola with several additional deaths due to suspect and probable cases.Reference Thomas2,3 Of the total number of health care workers who were infected with Ebola, 61% of those hospitalized and 74% of those not hospitalized for treatment died.3 Health care workers had a disproportionately higher risk of infection caring for patients in a health care system with limited IPC practices compared with the general population. This elevated risk exacerbated the pre-existing shortage of health care workers due to high rates of attrition, uneven distribution, lack of hazard pay, and limited occupational health standards.
It is known among the general population that living under the constant threat of EVD causes physical, social, and psychological distress.Reference Mutua and Wal4 In Liberia, in particular, 90% of community residents during the peak of the case volume of EVD reported fear of contracting the disease.Reference Kobayashi, Beer and Bjork5 How this affects health care workers who must continue to work in a high-risk environment, as well as how health care workers perceive EVD in their workplace, is unknown in the context of West Africa. Studies from preparedness planning and implementation in high-income countries have identified fear of transmission, lack of personal protective equipment (PPE), not being treated professionally, and poor communication as part of nurses’ concerns in caring for patients with EVD.Reference Speroni, Seibert and Mallinson6 Health care workers in Liberia sustained significant trauma during the EVD outbreak related to seeing colleagues infected and die from EVD; fear of being infected themselves and going through repeated, long periods of self-monitoring; as well as the risk of bringing infection home affecting their families and communities.
It was identified early in the outbreak that the implementation of IPC in health care settings was key to stopping the spread of EVD.Reference Kalra, Kelka and Galwankar7 A large portion of response efforts by the Liberian Ministry of Health and Social Welfare (MOHSW), World Health Organization, Centers for Disease Control and Prevention, and implementing partners (including international non-governmental organizations [NGOs] and community organizations) focused on training health care workers in infection control. The MOHSW, with technical assistance from international agencies and partners, created the Keep Safe, Keep Serving (KSKS) training package for IPC at health care facilities in the setting of EVD. KSKS was meant to motivate health care workers to return to work during the EVD outbreak and promote IPC practices within the hospital in order to keep staff safe. KSKS trainings were conducted frequently throughout health care facilities by different partners, often with lack of coordination, resulting in many health care workers receiving more than 1 KSKS training. It is unknown of what impact that several trainings had on affecting health care workers’ attitudes on understanding how to protect themselves and feeling safe from Ebola, as well as their motivation to return to work with the risk of Ebola. The objective of this evaluation was to determine what factors, particularly the number of previous IPC trainings, impacted health care workers’ understanding of how to protect themselves from Ebola and contributed to feeling safe and motivated in returning to work during the West African Ebola epidemic.
METHODS
The Liberian health care system consists of 39 hospitals: 23 public and 16 private. A quantitative survey was conducted at public hospitals to support the Ministry of Health and the hospital facilities that fall under their mandate. Twenty-one Liberian Government hospitals took part in a survey conduction over a 6-month period, January to June 2015. Two public hospitals were not included in this study because they did not conduct KSKS trainings for their health care workers. The survey was developed based on gaps identified within health care facilities, as well as general health care worker complaints voiced when returning to work during the Ebola epidemic. All hospital health care workers who took part in the KSKS trainings conducted by the university-based IPC program implementer, Academic Consortium Combating Ebola in Liberia (ACCEL), were surveyed. The ACCEL IPC KSKS trainings were conducted by Liberian clinical health care workers who served as master trainers and were well equipped in teaching the curriculum. All hospital workers, including non-clinical staff, were trained in KSKS. Health care workers who participated in the training included physicians, physician assistants, nurses, mid-wives, medical technicians, medical aides, janitorial and environmental health staff, pharmacy and lab staff, administration, kitchen staff, and security staff. The 11-question survey was administered by ACCEL staff in paper format prior to any KSKS training delivered by ACCEL. The survey was written at a sixth grade reading level to accommodate as many literacy levels in Liberia as possible. Additionally, to mitigate literacy limitations, questions had yes or no answers, and those with a sliding scale answer were designed using modified Likert-scale faces format. The survey was read out loud to participants as a group for those who needed assistance with reading.
Prior to the survey being handed out, participants in the training were verbally consented for participation in the survey. Those who did not want to or were unable to complete the survey still participated in training. Surveys were conducted prior to the beginning of ACCEL KSKS trainings in an effort to capture the status of health care workers’ attitudes and feelings prior to the start of what was often not their first IPC/Ebola training. This study was institutional review board (IRB)-exempt in Liberia because no personal health information was collected and the Boston Children’s Hospital IRB provided reciprocal IRB exemption. Participant mobile phone numbers were recorded on the survey to ensure that no health care workers participated more than once. All paper surveys were entered into Magpi® open source digital data collection software by ACCEL staff using password-protected tablets.
Upon completion of all ACCEL KSKS trainings over 6 months, the data were exported from Magpi to STATA 13.0 (StataCorp, College Station, TX). Basic demographics and response distributions were analyzed. The primary outcome evaluated was whether health care workers understood how to protect themselves from Ebola while at work. Previous health care worker training in IPC was evaluated as the primary predictor of understanding how to protect oneself.
Logistic regression was used for data analysis. Univariate screening for each individual variable was conducted to determine statistical significance. After a list of significant variables was determined, multi-variable intermediate model building was conducted with omission of variables based on loss of significance in the model. This process was done in order to determine which predictors should be included in the final regression model with attention to including confounders and omitting collinear variables.Reference Bursac, Gauss, Williams and Hosmer8 The number of previous IPC/Ebola-related trainings that a health care worker had received prior to the ACCEL comprehensive IPC package was evaluated as the primary predictor for whether health care workers understood how to protect themselves from Ebola. The 7 additional co-variables included in the model were (1) health care worker’s qualification, (2) how scared a health care worker felt going/returning to work, (3) how motivated a health care worker was to go/return to work, (4) health care worker’s confidence in the hospital’s ability to maintain IPC, (5) health care worker’s hand hygiene before and after each patient, (6) availability of IPC supplies at the workplace, and (7) presence of an IPC focal person at each hospital to answer questions.
Interaction terms were tested and effect modification evaluated between how scared a health care worker felt returning to work, health care worker’s confidence in the hospital’s ability to maintain IPC, and health care worker’s motivation to go to work. None of these interaction terms were included in the final regression model because there was no significant cumulative effect on the primary outcome found between these risk factors.
RESULTS
A total of 2748 out of 2756 health care workers agreed to participate in the survey and were included in the data analysis. The 8 health care workers who chose not to participate in the study did not disclose why. Eighty-six percent (2353) of participants indicated that they understood how to protect themselves from Ebola. Out of all participants, 2096 (76.3%) had received some sort of Ebola training prior to the implementation of the ACCEL comprehensive IPC intervention with the majority receiving 1 or 2 previous trainings, 36% and 33%, respectively (Table 1).
TABLE 1 Training Response Demographics and Response Sizes
Health care workers’ understanding of how to protect themselves significantly increased as the number of previous IPC/Ebola trainings increased up to a total of 3 trainings, OR 2.0, 4.9, and 8.0, respectively (Table 2). Beyond 3 trainings, however, the likelihood of understanding how to protect themselves from Ebola decreased. These findings suggest a saturation point of 3 previous trainings to maximize the likelihood of health care worker understanding on how to protect themselves from Ebola, with no additional benefit from more than 3 trainings.
TABLE 2 Odds Ratios for Predictors of Feeling Protected From Ebola at Work in the Hospital Environment
On evaluation of other contributing predictors of the outcome, almost all showed significance in either positively or negatively affecting health care workers’ understanding of how to protect themselves from Ebola. As educational level increased, the odds of understanding how to protect one’s self also increased, with all qualification odds ratios being statistically significant except a bachelor’s in science. Health care workers who reported washing their hands before and after patient contact were found to be 3.4 times more likely to understand how to protect themselves from Ebola while at work, with all other variables held constant. The availability of IPC supplies at work resulted in health care workers being 3.4 times more likely to understand how to protect themselves at work. Having confidence in the ability to go to someone with IPC questions and get the correct information resulted in health care workers being 2.1 times more likely to understand how to protect themselves from Ebola.
DISCUSSION
This study shows that several factors were found to contribute to hospital-based health care workers’ understanding of how to protect themselves from Ebola while at work and feeling safe practicing in the setting of an Ebola outbreak. Health care workers’ understanding of how to protect themselves from Ebola increased with the number of trainings received, however, hit its maximum effect at 3 trainings. After 3 trainings, the understanding of protection and feeling of safety went down, thus suggesting a saturation point and training fatigue.
Much of the Ebola response efforts in the technical area of IPC focusing on providing trainings as a systemic approach to IPC was new to the Liberian health care system. The fact that training improved the feeling of protection among Liberian hospital health care workers is important because it supports efforts to build capacity in the existing health care system as part of the humanitarian response efforts. It also supports the continual education process and hospital staff empowerment. However, these findings also identify an important threshold effect where the quantity of trainings conducted during humanitarian response efforts can have a negative impact on health care workers’ attitudes toward safety and understanding of the health care system. The reasons for this decline in safety were not explored in this study but some can hypothesize that trainings done by multiple agencies could convey conflicting facts and messages, resulting in confusion and loss of confidence in the ability to protect themselves by health care workers. Humanitarian response efforts need to be better coordinated in the future to prevent duplication of efforts, conflicting messages, and trainee fatigue.
These data imply that training during an Ebola outbreak and possibly other disease outbreaks is an important component of the outbreak response. It identifies a threshold effect with the positive impact on feeling protected based on the number of trainings that health care workers receive, peaking at 3 and then diminishing. This is important as coordination and avoidance of duplication become part of the increasingly complex humanitarian response. Organizations ranging from international and national NGOs, local governments, international agencies, and faith-based organizations often provide aid with a focus on sustainability and capacity building with training serving as a key component of their programming. These data note the importance of not conducting excessive trainings to reach maximum effect and skill/knowledge transfer.
The variables of reported health care worker hand hygiene practices, hospital IPC supply availability, and hospital IPC implementation through an IPC focal person were found to be significant predictors of feeling protected from Ebola at work. The significance of feeling protected through the practice of handwashing is important because handwashing is an inexpensive, time efficient, readily available with running or bucket water and soap and low resource strategy for IPC. All of these benefits of handwashing are low cost with a high yield in results when used. Human behavior change for hand hygiene is often the most difficult thing to change and is seen in low-, middle-, and high-income countries.Reference Srigley, Corace and Hargadon9 This supports a call to action for donors to not only teach proper hand hygiene, but also ensure that the necessary resources are available through water and sanitation renovations and supply chain support for the materials. The need for supply chain support is further supported by finding that the availability of IPC supplies impacts the likelihood of a health care worker to understand how to protect one’s self. The provision of basic IPC supplies and PPE on a health systems level is necessary for quality health care provision and health practice as these items are not commonly included in national drug/supply distribution plans. Unless the presence of IPC materials and PPE is sustainable, the confidence that health care workers feel in practice will not persist and the feeling of safety will diminish. These data support that mentoring and capacity building of the IPC focal person role within hospitals and the development of IPC policies and procedures are key. The ongoing mentorship and leadership from the IPC focal persons to health care workers along with constant IPC resource availability will help maintain a positive feeling of protection from Ebola.
The health care worker’s qualification also played a role in feeling protected from Ebola. As educational level increased, so did understanding of how to protect one’s self and feel safe. Not surprisingly, skilled health care workers (nurses, physician assistants, and physicians) felt more protected with training as they were more likely to be able to absorb this knowledge given their existing baseline medical knowledge. This also aligns with the distribution of health care workers who were infected during the EVD outbreak. One-third of those infected had overall lower educational levels and qualifications (ie, cleaners/janitors), yet higher risk jobs given the nature of exposure during job duties.3 This suggests that special attention should be focused on those health care workers with lower educational levels and higher risk jobs to ensure safety and protection while at work. It is also noted that, during this 6-month period, no health care workers at hospitals contracted Ebola, providing objective validation of these feelings of protection and education level correlation.
Prior data on protection of health care workers practicing in the existing health setting (specifically, West Africa) are limited given the limited size of the prior outbreaks, with the largest being less than 500 patients and local health care workers only minimally affected. Data from the knowledge and attitudes of Western region health care workers showed fear of transmission, lack of PPE, not being treated professionally, and poor communication as part of nurses’ concerns in caring for patients with EVD.Reference Speroni, Seibert and Mallinson6 In this study, having supplies present increased health care workers’ feeling of protection similar to what Western health care workers reported as a barrier to Ebola care.
This study was limited by several factors, including health care worker literacy, understanding of questions, social desirability effect, Hawthorne effect, and missing data. Many participants could not read or write, and ensuring that they understood the questions asked and answered appropriately was challenging given that the survey was conducted in a group setting rather than individually. Language used in the survey was also not Liberian English and thus resulted in some participants’ misunderstanding of questions or leaving them blank/missing, which represented variable sample sizes for each question’s response. This information suggests that respondents may have also answered yes to the question on previous trainings received given they “felt” this to be the socially and/or intellectually “correct” answer. There is also no definition of what “previous training” entailed, and interpretation of this question likely varied greatly from small in-services to day or multi-day long trainings. A contextual and logistical limitation is that this survey is cross sectional in time at each facility with facilities surveyed at different points over the 6-month period, with no pre- and post-comparison. However, the presence of an ongoing, active outbreak regardless of the size should capture the presence of predictors for health care workers feeling protected from Ebola at work.
Additionally, the final logistic regression model accounts for 25% of the variability in outcome. This suggests that the model is likely missing several variables that affect health care workers’ understanding of how to protect themselves from Ebola while at work. Despite these unknown variables, the current model does explain several factors that affect health care workers’ understanding on how to protect themselves from Ebola and sheds light on the Ebola epidemic and how health care workers felt when returning to work to Keep Safe, Keep Serving.
CONCLUSION
This study provides the first look at what factors make West African health care workers understand, feel protected, and be ready to return to the provision of routine medical care, which is critically important during and immediately after an Ebola outbreak. The number of previous trainings; the baseline health care worker qualifications; and the components of IPC, including handwashing, availability of supplies, and the resource of an IPC focal person, showed to be statistically significant in correlation with how safe health care workers felt returning to work in the setting of an Ebola epidemic. The identification of these factors played a critical role in ending the largest ever EVD outbreak, as well as for a future response to recurrent Ebola outbreaks and other infectious pathogens.
Future studies evaluating other variables accounting for the 25% of variability in outcome would be helpful in understanding all aspects of the health care worker and hospital environment that contribute to the feeling of safety during and after an infectious disease outbreak. In addition, the coordination of educational efforts aimed at health care workers and hospital staff is key for response agencies during infectious disease outbreaks given that learners reach a saturation point. Without coordination of efforts, the risk of learning fatigue and the inability to absorb information rises and thus has the potential for interventions to be unsuccessful, resulting in wasted time, money, and resources.
Institutional Review Board (IRB) Statement
This study was IRB-exempt by the Boston Children’s Hospital IRB, as well as the Liberian IRB.
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
