Public health emergencies, such as influenza pandemics and major earthquakes, can place an enormous burden on a health care workforce. During an influenza pandemic, the health care workforce is essential to provide care to ill persons and to help prevent further spread of infection.1 Additionally, large-scale disasters, such as earthquakes, can impact the availability of health care services in the affected community. Responses to such emergencies depend on close collaboration among health departments, community partners, health care facilities, and an accessible and willing health care workforce. Responsibilities of the health care workforce could include the direct provision of preventative and therapeutic services as well as supportive nonclinical activities that could mitigate the health impact of a public health emergency. This collaboration is particularly important in urban areas where the high population density can increase the spread of an epidemic dramatically or, in the case of a severe earthquake, impact critical infrastructure.Reference Smith, Morgans and Qureshi2 Yet, uncertainty exists about whether those in the health care workforce are both able and willing to respond to public health emergencies. Given that pandemics and other disasters, natural and human-induced, will continue to occur, we need to understand the reasons associated with the decision to report or not report to work during public health emergencies.Reference Iserson, Heine and Larkin3
The attitudes and capabilities of health care workers to report to work during disasters is a topic of growing interest to those involved in surge planning. Health care workers' ability and/or willingness to report to work during influenza pandemic has been reported in several studies, suggesting between 25% and 82% of workers will report during this type of disaster. The sudden acute respiratory distress syndrome (SARS) experience in Toronto resulted in 25% of nurses staying out of work to avoid exposure.Reference Imai, Takahashi, Hasegawa, Lim and Koh4 Qureshi et al conducted a survey of health care workers in New York City and reported that most indicated they would be able to report to work during most catastrophic disasters, but fewer were willing to report to work for certain disasters, such as a SARS outbreak.Reference Qureshi, Gershon and Sherman5 Similarly, among employees of three local health departments surveyed in Maryland, 53.8% were likely to report to work during an influenza pandemic.Reference Balicer, Omer and Barnett6 The most influential factor associated with willingness to report to work in this survey was the perception of the importance of one's role in the agency's overall response. Barnett et al reported that 84% of the local public health department workers were willing to respond to a pandemic flu emergency regardless of its severity.Reference Barnett, Balicer and Blodgett7 Basta et al conducted a survey of Florida county health department employees to assess willingness to respond and reported that nearly half of public health department employees are unwilling to report to work during the peak of an influenza pandemic when the public health response will be a vital component of pandemic containment and mitigation.Reference Basta, Edwards and Schulte8 A recent report indicated that most Florida providers were willing to respond to bioterrorism events within their local community, with 82.7% of providers willing to respond in their local community, and 53.6% within the state.Reference Crane, McCluskey, Johnson and Harbison9 Irvin et al reported that 50% of hospital health care personnel respondents would and 42% might report to work as usual in the event of an avian pandemic.Reference Irvin, Cindrich, Patterson and Southall10 Additional studies on the attitudes of various health care workers toward reporting to work during pandemic influenza in different settings have been reported.Reference Barnett, Balicer and Thompson11Reference Balicer, Barnett and Thompson12Reference Barnett, Levine and Thompson13
The ability and willingness of staff to report to work during a public health emergency was identified by the King County Healthcare Coalition as a significant issue for King County, Washington, a region of over 1.89 million persons and over 60 000 health care workers. The King County Healthcare Coalition was established in 2005 to facilitate strategic and collaborative emergency planning among health care agencies and providers and to coordinate emergency response activities in the region. King County is the most populous county in Washington State and is the 14th most populous county in the nation. King County has a total of 22 hospitals, 5 major medical groups with over 42 sites throughout the county, a number of specialty medical providers, over 1000 licensed residential long-term care facilities, 2 tribal clinics, 33 safety net clinics, 387 pharmacies, and a number of hospital, residential, and outpatient mental health and chemical dependency treatment providers and home health and home care providers. The goal of the regional health care worker survey was to provide an assessment of the factors that may impact the ability and willingness of health care workers in King County, Washington, to report to work during influenza pandemic or a severe earthquake.
METHODS
Survey Sample
The sample was drawn from two groups: the State of Washington Licensure Database and King County health care agencies. This approach was used to help ensure a broad representation of different types of health care workers in the sample. The licensure database group consisted of physicians, nurses, pharmacists, and mental health providers who, according to the licensure records, lived or worked in King County. The King County health care agency group included health care workers from seven of the health care sectors represented by the King County Healthcare Coalition, specifically: hospitals, home health and palliative care agencies, long-term care providers, mental health providers, safety net clinics, specialty providers, and tribal clinics. Survey participants from the licensure database group were selected at random and stratified based on license type (physician, nurse, pharmacist, and mental health professional, ie, licensed mental health counselors, licensed social workers, licensed marriage and family therapists, and registered counselors). All King County hospitals were included in the health care agency group. Other participating health care agencies were selected at random and stratified based on health care sector and the three King County emergency management geographic zones. Health care workers associated with the individual agencies were selected at random and stratified based on job title (physician, nurse, clinical support staff, administration, and non-clinical support staff). If an agency was unable or unwilling to provide job titles, employees were selected at random from a list of agency personnel. A total of 2400 persons were selected from the licensure database group and 6811 persons were selected from the agency group.
Survey Administration
The survey was administered by the University of Washington Survey Research Division. Survey participants were sent an invitation letter with instructions for completing the survey electronically via a secure Web site. A reminder letter or e-mail was sent to participants within two weeks of receiving the invitation letter. A final reminder letter and a paper survey were sent to those participants who had not responded to the web survey within six weeks of receiving the invitation letter. In an attempt to increase the response rate, a lottery component was included as an incentive. If they chose to do so, respondents could enter a lottery for one of 70 $100 checks by providing their e-mail address. With permission, e-mail addresses provided for the purpose of entering the lottery were used for the lottery only and were in no way connected to a respondent's survey responses. Data collection lasted for approximately six months. Data analysis was completed at the University of Washington Northwest Center for Public Health Practice. The project was reviewed and granted a certificate of exemption by the University of Washington Human Subjects Division.
Survey Content
The 70-item survey contained questions about demographics, ability and willingness to report to work during two different disaster scenarios, barriers that would prevent workers from reporting to work during the disaster scenarios, strategies that would help workers report to work during the disaster scenarios, and job responsibilities and training. The first disaster scenario described an influenza pandemic in which patients had been admitted to local area hospitals, fellow workers were out sick with severe influenza, and some patient deaths had occurred. The second scenario described a severe earthquake during which there had been structural damage to buildings and facilities, disruption of public transportation and major roadways, and reports of injuries. Respondents were asked if they would be able and willing to report to work for each scenario assuming that they and their families were okay. The response options were “yes,” “no,” and “not sure.” Items dealing with barriers and strategies were presented as statements with which respondents could agree or disagree. Response options were “strongly agree,” “somewhat agree,” “somewhat disagree,” “strongly disagree,” and “not sure.” Only those respondents who reported having children under the age of 18 years at home were asked about child care obligations as a barrier and the availability of child care as a strategy.
Analysis
With the exception of respondent demographics, data for the two disaster scenarios were analyzed separately. Response categories for the items dealing with barriers and strategies were collapsed into two categories “agree” and “disagree.” All analyses were performed using SAS®, 9.2 (SAS Institute, Cary, NC). Odds ratios with 95% confidence intervals from logistic regression were used to examine the association between demographic variables and employer-related factors, and self-reported ability and willingness to report to work. Multivariate logistic regression models were used to examine the association between job category and self-reported ability and willingness to report to work while controlling for age, gender, children and adult dependents at home, employment status, and flu vaccination within the last 12 months. Multivariate models were also used to examine the association between employer-related factors and self-reported ability and willingness to report to work while controlling for job category, age, gender, children and adult dependents at home, employment status, and flu vaccination within the last 12 months. All logistic regression models were analyzed with the “proc surveylogistic” command. Sampling weights were used in the regression analysis to account for the unequal probability of selection as well as non-response. A P -value of .05 was used to determine statistical significance.
RESULTS
During six months of data collection, 4484 of the 9211 surveys were returned. A total of 138 surveys were returned as undeliverable, 15 respondents refused to complete the survey, 215 were ineligible due to employment status (retired or unemployed) or because of lack of an agency/licensure identification code, and 40 were excluded from the analysis due to insufficient data. The final overall response rate for those reached was 50% with 4306 respondents, ie, response rate = (completes + ineligibles + partials + refusals) / (total − return to sender). Within the licensure database group, pharmacists (48%), mental health providers (46%), and nurses (45%) had the highest response rates, while physicians (33%) had the lowest response rate. Among health care facilities, safety net clinics (73%), mental health providers (58%), specialty providers (50%), and hospitals (49%) had the highest sector response rates within the agency group. Home health and palliative care (37%), long term care providers (30%), and tribal clinics (28%) had the lowest sector response rates.
Characteristics of Respondents
Seventy-five percent of the respondents were female, 73% were employed full time, and 69% were married or living with a partner (Table 1). Just over half of the respondents (56%) were between 40 and 59 years old. Thirty-five percent of respondents reported being the primary caregiver for children under the age of 18 years and 7% indicated that they had adult dependents at home. In describing their primary job category, 24% of respondents indicated nurse and 20% clinical support services. Seventy-two percent of respondents reported that they had worked in their current profession for six or more years, and 57% reported that they had been with their current employer for six or more years.
TABLE 1 Respondent Demographics
Ability and Willingness to Report to Work
When asked about an influenza pandemic scenario, 95% of respondents reported that they would be able to report to work and 89% reported that they would be willing to report to work. When asked about a severe earthquake scenario, 74% of respondents reported that they would be able to report to work and 88% reported that they would be willing to report to work. Levels of self-reported ability and willingness for both scenarios were high among the different job categories (Table 2).
TABLE 2 Self-Reported Ability and Willingness to Report to Work by Job Category*
Barriers and Strategies
Among all respondents fear or concern for their family (51%), fear or concern for themselves (39%), and personal health problems (34%) were the barriers most often reported for the influenza scenario (Table 3). In addition, over half of the respondents (57%) with children under 18 living at home reported that child care obligations would be a barrier when reporting to work. For the earthquake scenario, more respondents reported that transportation problems (71%), fear or concern for their family (63%), and fear or concern for themselves (45%) would be barriers when reporting to work. For those respondents with children at home, child care obligations (64%) were identified as a barrier when reporting to work following a severe earthquake.
TABLE 3 Self-Reported Barriers and Strategies by Scenario
When asked about strategies that would help respondents report to work during the influenza scenario, the most frequent responses included anti-viral treatment for themselves (92%), anti-viral treatment for their families (86%), and the ability to work from home (70%). Sixty-four percent of respondents with children at home indicated that the availability of child care would be a helpful strategy during an influenza pandemic. Strategies reported most often for the earthquake scenario included a recorded phone message providing the status of the workplace (91%), a shuttle service within walking distance of home (84%), a pre-arranged carpool of employees (76%), and the ability to report to a facility closer to home (72%). As in the influenza scenario, respondents with children at home reported that the availability of child care (66%) would be a helpful strategy following a severe earthquake.
Sixty-three percent of respondents reported that their employer had provided adequate information and programs to help them report to work during a disaster. Just over half (55%) of the respondents reported that their employer had communicated what the respondents job responsibilities would be during a disaster. Sixty-two percent of respondents indicated that they would like to receive more training about how to fulfill their job responsibilities during a disaster. Of these respondents who indicated the desire for more training, 62% would be interested in training on the incident command structure, 61% would be interested in training on specific job duties, 58% would be interested in training on personal preparedness, 56% would be interested in training on family preparedness, and 35% indicated interest in training on infection control.
Ability and Willingness to Report to Work–Multivariate Models
After controlling for age, gender, children and adult dependents at home, employment status, and flu vaccine in the last 12 months, only one job category (administration) had increased odds of self-reported ability as compared to mental health provider reference group, but all job categories had increased odds of self-reported willingness to report to work during an influenza pandemic as compared to mental health provider (Table 4). When controlling for job category and demographic characteristics, only perceived importance of role in an emergency response was associated with increased odds of both self-reported ability [OR 3.50; 95% CI 2.38-5.16] and willingness [OR 4.80; 95% CI 3.72-6.19] to report to work during an influenza pandemic. Having adequate information from an employer about getting to work during a disaster was associated with increased odds of self-reported willingness [OR 1.72; 95% CI 1.33-2.24] to report to work.
TABLE 4 Factors Associated With Self-reported Ability and Willingness to Report to Work During an Influenza Pandemic
For the severe earthquake scenario, after controlling for the demographic characteristics mentioned above, four job categories (nurse, clinical support services, environmental services, and administration) had increased odds of self-reported ability as compared to mental health provider reference group, and all but two job categories (administrative support and environmental services) had increased odds of self-reported willingness to report to work following a severe earthquake. When controlling for job category and demographic characteristics, perceived importance of role in an emergency response was associated with higher odds of self-reported ability [OR 2.19; 95% CI 1.85-2.58] and willingness [OR 3.25; 95% CI 2.57- 4.12] to report to work following a severe earthquake. Having adequate information from the workplace about getting to work was also associated with both self-reported ability and willingness to report to work following a severe earthquake (Table 5).
TABLE 5 Factors Associated With Self-reported Ability and Willingness to Report to Work Following a Severe Earthquake
Willingness to Participate in Other Responsibilities
Respondents were asked to indicate in which activities they would be willing to participate, in lieu of their normal job responsibilities, during a disaster. Of the respondents who provided a response, 73% (n = 3038) indicated that they would be willing to report to work at a facility closer to home and perform similar job duties, 67% (n = 2808) would be willing to participate in a community neighborhood response team, 57% (n = 2371) would be willing to staff a call center, and 52% (2158) would be willing to staff an alternate care facility.
COMMENT
Our findings suggest that, if health care workers and their families are okay, respondents would be willing and able to report to work in each of the two disaster scenarios presented. More respondents reported the ability to get to work during an influenza pandemic rather than following a severe earthquake, while willingness to report to work differed only slightly by scenario. The difference in ability to report to work could be due to the unique nature of the two scenarios, each of which would present respondents with different types of problems that would need to be addressed. The most frequently cited barriers for the influenza scenario were fear or concern for family and fear or concern for self, while fewer respondents indicated that transportation problems would be a barrier. Fear or concern for themselves and their families were frequently cited barriers for the earthquake scenario as well, but the most frequently cited barrier was transportation problems. For those respondents with children at home, the availability of child care was also an important barrier reported for both scenarios. Concern for family has been identified by a number of researchers as having an impact on willingness to report to work in different disaster scenarios.Reference Imai, Takahashi, Hasegawa, Lim and Koh4Reference Qureshi, Gershon and Sherman5Reference Basta, Edwards and Schulte8Reference Barnett, Levine and Thompson13Reference Qureshi, Merrill, Gershon and Calero-Breckheimer14Reference Damery, Wilson and Draper15Reference Dimaggio, Markenson, T Loo and Redlener16Reference Martinese, Keijzers, Grant and Lind17
The most frequently cited strategies that would help respondents report to work during an influenza pandemic were anti-viral influenza treatment for themselves, anti-viral treatment for their household, and the ability to work from home. Martinese, et alReference Martinese, Keijzers, Grant and Lind17 found that the availability of immunizations and/or anti-viral medications impacted whether or not hospital staff reported to work during an influenza pandemic. Though a very small sample, Mackler, et alReference Mackler, Wilkerson and Cinti18 reported that the majority of the first responders surveyed would not remain on duty during a smallpox outbreak without access to vaccine. Unlike an influenza pandemic, a severe earthquake could impact the type of transportation available to get to work and the safety of that trip. Given these concerns, the most frequently cited strategies were a recorded phone message telling employees the status of their workplace, a shuttle service with a pick-up site close to home, and a prearranged carpool of employees.
Health care workers in all job categories reported a high degree of ability and willingness to report to work in both scenarios. Martinese, et alReference Martinese, Keijzers, Grant and Lind17 reported similar results in that job description did not have a significant effect on work absenteeism. Balicer et alReference Balicer, Omer and Barnett6 however, reported that clinical staff was more likely than technical or support staff to report to work. Controlling for job category and demographic characteristics a respondent's perceived importance in the emergency response to a disaster was associated with increased odds of reported ability and willingness. According to Balicer et al,Reference Balicer, Omer and Barnett6 the perception of one's role in the agency's overall response was the single most influential factor associated with willingness to report to work. Regarding employer-related factors, it is apparent that some health care workers were unsure about what their job responsibilities are during a disaster and that not all employers have provided enough support and guidance about how to get to work during a disaster. However, a substantial number of respondents would like more training about how to fulfill their job responsibilities during a disaster. Desired training included individual and family preparedness, specific job responsibilities, and the incident command structure.
While a large number of respondents reported that they would be willing and able to report to work, the findings indicate that there are things that health care agencies could do to help their employees report to work during a disaster. Willingness to report to work for many health care workers is impacted tremendously by the need to provide care and reassurance to their families.Reference Smith19 Administrators could work within their agencies or with other agencies to provide employees with child care options or anti-viral treatments for family members during a disaster. It is also apparent from the findings that respondents were interested in additional training about how to fulfill their job responsibilities during a disaster. Health care agencies need to work with their employees before a disaster occurs so as to avoid the problems seen in past disasters such as the SARS outbreak.Reference Imai, Takahashi, Hasegawa, Lim and Koh4Reference Nickell, Crighton and Tracy20Reference Koh, Lim and Chia21 Bernstein and HawryluckReference Bernstein and Hawryluck22 state that the best defense for a disaster such as SARS is to have a contingency plan in place, to have well conceived and developed plans well known in advance and rehearsed so as to limit the potential damage of such a disaster. Garrett et al reported that preferential access to either antiviral therapy, protective equipment, or both for the employee as well as his or her immediate family will have the greatest impact on mitigating absenteeism during a pandemic.Reference Garrett, Park and Redlener23
This study does have some limitations. While we did reach a 50% response rate, responder bias is still an issue. Those who responded to the survey may be more motivated to report to work than those who did not respond. In addition, certain health care sectors such as long-term care and home health/palliative care had response rates lower than 50%. Given the way in which the four questions about ability and willingness to report to work were phrased (with the assumption that one's family is ok) the positive responses are probably inflated. In a real emergency, health care workers may not know the condition of their family, which could impact the decision to report to work. There are many factors that will influence a person's decision to report to work during an actual emergency that could not be accounted for in this survey. As Ives, et alReference Ives, Greenfield and Parry24 put forth, a decision to work (or not to work) is the result of a combination of motivations and beliefs, which interact with both genuine and constructed barriers to ability. For some health care workers this may result in a genuine barrier to ability or a barrier to willingness that is perceived as a barrier to ability, and for others it may simply be a barrier to willingness. While our questionnaire was guided by previous assessments of health care workers' willingness to respond, it is not based on the application of a behavioral model, such as Witte's Extended Parallel Process Model, that has been recently applied by Barnett et al and Balicer et al.Reference Barnett, Balicer and Thompson11Reference Balicer, Barnett and Thompson12 Finally, what people report on surveys about their behavior during a disaster is not necessarily predictive of how they might respond in an actual event.
The implications of our findings for planning by public health and health care organizations are multi-fold. At the individual health care facility level, organizations should assess how well they are communicating with their staff about roles during disasters and the training needed to perform those roles. Gershon et al reported that a large proportion of New York City home health care workers reported that they would be either unable or unwilling to provide care to their patients during a pandemic and that the vast majority of respondents had not received any work-based, pandemic-related training.Reference Gershon, Magda and Canton25 Perceived importance in an emergency response was independently associated with increased odds of both self-reported ability and willingness to report to work during a pandemic or a severe earthquake. Other employer-related factors, such as addressing strategies that could help people get to work in an earthquake, could help organizations better prepare for disasters. If a goal is to keep the regional health system functional and appropriately staffed during a disaster, our findings suggest strategies that also need to be considered at a regional level (eg, county or multicounty level). For example, the earthquake scenario highlights vulnerability of staff getting to and from their workplace. This reinforces the need to develop regional transportation strategies that support the movement of health care workers to their places of employment throughout the area, which may be a more efficient use of resources, than planning by individual facilities. Additionally, given the potential barriers for staff getting to and from their normal workplace, it highlights the need for more collaborative planning across the health care system about sharing staff. This includes credentialing, privileging and talking to staff about expectations and strategies for where they could work and what they would be doing. The high percentage of respondents who said they would participate in a neighborhood response team suggests opportunities to develop strategies for supplementing neighborhood level medical response if persons cannot get to their facilities in an effort to try and augment triage and treatment at that level reduce influx of folks going to hospitals unnecessarily. This step, however, points to the importance of doing good planning about what that level of medical response would look like, and what resources would be needed. The data on the percentage of health care workers willing to work from home suggest potential opportunities for regional planning on for call centers with medical triage and clinical support capabilities. Additionally, the responses pertaining to pandemic influenza preparedness reinforces the need for regional discussions about strategies for providing antiviral medicines and immunization to staff and families and seeking standardization as much as possible across a given geographical region.
CONCLUSION
Health care staff is the most important asset to health care organizations. Based on our findings, we identified opportunities to better inform and educate the health care workforce about their roles in a pandemic influenza and an earthquake response. Given the limitations of self-reported data, further understanding of variables that impact staff willingness and ability to report to work under different scenarios is needed before we can build effective surge capacity strategies that are reliant on staffing.
