Methods

The survey was the second part of a two-phase study that combined qualitative and quantitative methods.19 It was conducted across three National Health Service (NHS) trusts in the West Midlands, UK: one acute teaching, one primary care trust, and one rural district general hospital. Trusts were purposively recruited to ensure participation from HCWs across a wide demographic range, working in diverse healthcare settings. The target population comprised all categories of HCWs (hospital doctors, nursing staff, healthcare managers, professions allied to medicine (eg, radiographers, pharmacists), ancillary staff (eg, porters, mortuary workers), general practitioners (GPs) and community HCWs. Sampling was stratified by trust, and survey recipients were randomly selected from databases of current staff provided by human resources contacts in each trust.

Three thousand anonymised self-completion surveys were sent to HCWs within the target employment groups across participating trusts, from July to September 2008. Surveys were either mailed directly to individuals at their work address or distributed to staff by clinical managers. Those wishing to participate were able to either complete a paper version of the survey, returned via Freepost (postage-paid) envelope directly to the research team, or submit their responses online. Non-respondents were sent one reminder, and return of the survey was taken as consent to participate. No incentive was offered for completion.

Survey

Survey content was developed from the findings of qualitative research conducted in the first phase of the study.15 The survey included closed questions on demographic characteristics, employment status and home circumstances, including caring responsibilities. To investigate the likelihood that an individual would work, one question outlined a series of circumstances (n = 12) that may arise during a pandemic (eg, infection to self or family), to which respondents indicated their likelihood of continuing to work under each circumstance (“likely”, “unlikely”, “don’t know” or “not applicable”). Agreement or disagreement with various statements regarding ethics was also assessed. These ethics statements were derived from both the qualitative phase of the study and a review of the literature on duty and healthcare provision, encompassing the broad areas of solidarity or loyalty, punishment and reward, risks and benefits, general obligations and reciprocity.

Data analysis

Analysis focused on the sociodemographic characteristics of respondents giving positive or negative ratings to their perceived likelihood of working under different circumstances, and the association between this and agreement or disagreement with each of the ethics principles outlined in the survey. A score for “likelihood” of working (from 0 to 100) was calculated for each individual on the basis of the percentage of the 12 statements detailing potential adverse circumstances that might arise during a pandemic in which respondents said they would be “likely” to work. The survey instrument and a detailed explanation of the method for calculating the likelihood score are provided elsewhere.11

Responses were dichotomised between those with a likelihood score of 100 (the highest potential likelihood of working) and those scoring <100 (lower potential likelihood). Binary logistic regression was used to calculate bivariate and multivariate odds ratios (ORs) to evaluate the association between likelihood of working and agreement or disagreement with ethics principles, first for all respondents, then by selected sociodemographic group—age, gender, job category, caring responsibilities and work type (part-time or full-time employment). Differences between groups were further investigated using parametric and non-parametric statistical tests. Non-response bias on the basis of gender, age and job type was assessed using χ² analysis. All data were analysed using SPSS (version 15.0).

Ethics approval

National Research Ethics Service (NRES) approval for this project was granted by Nottingham Research Ethics Committee 2 (Ref: 07/H0408/120), and the approval of the research and development unit was obtained from each participating trust.

Results

Of 3000 surveys distributed, 1032 (34.4%) were returned complete. A further 70 were returned blank, indicating a desire not to receive a reminder. No statistically significant differences were found between respondents and non-respondents on the basis of gender, age or job type. This suggests that the responses received reflected the demographic composition of the survey population.

Respondent characteristics

The study population (table 1) included more females than males. Those aged 41 to 50 years comprised the largest age group, although numbers were fairly equally distributed across the four age groups surveyed. The largest employment category was ancillary workers (17.3% of respondents) and the smallest was hospital doctors (11.8%). Part-time workers were outnumbered by full-time workers, and HCWs who indicated having some form of caring responsibility for family members (for children aged under 16 or elderly dependents) comprised just under half of all survey responders.

Ethics statements and mean likelihood of working

Table 2 shows the frequency of respondent agreement or disagreement with each of the ethics statements outlined in the survey, and the association between these responses and mean likelihood of working scores. Responses demonstrated that all HCWs have a strong sense that there is a duty to work during a pandemic, even in the face of personal risk. The majority of respondents (76.8%) agreed that doctors and nurses have such a duty and so do HCWs in general (72.7%). Almost all respondents agreed that “everyone should pull together during a pandemic”, and only about a third of respondents agreed that HCWs should have the option to refuse to work with infected patients. In all of these cases, agreement with the statement in question was associated with a significantly higher likelihood of working during a pandemic.

In spite of the general acceptance that HCWs have a duty to work during a pandemic, punishment or sanctions for those who do not fulfil this duty were not widely endorsed. Only 14.2% of respondents agreed that those who refuse to work should be punished; similarly low proportions agreed that those who do not work should be disciplined (18.3%), or lose wages (29.8%). In contrast to the aversion to punishment, many respondents felt that HCWs who continue to work during a pandemic should receive some form of reward for their efforts, either in the form of special priority for treatment if they became ill with the virus as a result of their work (80.5%) or in the form of other (financial) remuneration (76.6%).

Nearly all respondents felt that in addition to their own obligations during a pandemic, their employers had reciprocal obligations to put measures in place to protect HCWs and their families, such as the provision of personal protective equipment and of vaccination for themselves or family members. Agreement or disagreement with these statements did not correspond to a statistically significant difference in the mean likelihood of working scores.

Even though the majority of respondents agreed that HCWs have a duty to work, there was a clear conflict between the duty to treat the sick and the duty to self and family. Three-quarters of respondents stated that their primary responsibility was to themselves and their family. Those agreeing that family responsibilities were important had a significantly lower mean likelihood of working than those who disagreed (54.4 versus 72.7, respectively).

Strength of associations between agreement with statements and likelihood of working

Table 3 shows bivariate and multivariate ORs to demonstrate the strength of the observed associations between agreement with ethics statements and the potential likelihood of working. The perception that there was a duty to work emerged as a strong predictor of potential attendance; those agreeing with the statement that “all HCWS have a duty to work” were significantly more likely to report that they would work than those who disagreed (bivariate OR 3.07). This association remained in the multivariate model, although less strongly (multivariate OR 2.01). Similarly, those who agreed that “doctors and nurses have a duty to the sick” were over four times more likely to work than those who disagreed.

Despite the low acceptability of punitive measures for HCWs who refuse to work during a pandemic, those who endorsed such measures were significantly more likely to report that they would work than those who did not. Agreement with the statement that “people who refuse to work should be punished” showed a bivariate OR of 2.71. Similarly, respondents who agreed that HCWs “should face disciplinary action if they refuse to work” were also significantly more likely to report that they would work than those who disagreed (bivariate OR 1.69). The potential conflict between the duty to patients and the duty to family members was again apparent. Those agreeing that their main responsibility was to themselves and their family were significantly less likely to work than those who disagreed (bivariate OR 0.25). This association remained in the multivariate model.

Sociodemographic predictors of agreement with ethics statements

Table 4 shows the demographic and employment characteristics associated with agreement with each of the ethics statements in the survey, and whether or not there was a significant difference between demographic subgroups (age group was not a statistically significant variable and is therefore not illustrated).

Significant differences in the response to several of the ethics statements were found on the basis of HCW job type. Doctors and nurses agreed with the statement “doctors and nurses have a duty to the sick even when there are high risks to themselves” to a lesser extent than those in other HCW roles (doctors, 76.2%, n = 93; nurses, 70.9%, n = 95; versus ancillary workers, 78.8%, n = 141; managers, 79.5%, n = 120; GPs, 82.3%, n = 116). Very few hospital doctors and GPs agreed that HCWs should not receive special priority during a pandemic (7.4%, n = 9 and 3.5%, n = 5, respectively). Similarly, respondents were differentiated according to job type regarding the option of refusing to work with infected patients. Managers were significantly more likely than those in other HCW groups to agree that HCWs should be allowed to refuse to work with infected patients (bivariate OR 2.22, CI 1.35 to 3.70).

For all demographic subgroups, there was little differentiation on the basis of age, gender, caring responsibility or full-time and part-time working status regarding agreement with ethics statements relating to either the duty to work or the reciprocal obligations of employers towards their employees. This suggests that the importance (or otherwise) of these factors was perceived in a similar way irrespective of the demographic or employment characteristics of survey respondents. Conversely, agreement with the statement that “my main responsibility is to myself and my family” was highly variable across all demographic subgroups (except for full-time versus part-time workers). Female respondents were significantly more likely than males to agree that family comes first (bivariate OR 1.41, CI 1.23 to 1.59), as were respondents with caring responsibilities for others (bivariate OR 1.41, CI 1.18 to 1.67). Nurses and ancillary workers were also significantly more likely than other categories of HCW to agree that they perceived an over-riding duty to themselves and their families (bivariate OR 2.50, CI 1.36 to 4.63 and 1.77, CI 1.03 to 3.02, respectively).

The need to continue earning their wages had a significant impact for many HCWs on their stated likelihood of working during a pandemic. This was particularly the case for full-time workers in comparison with part-time employees (bivariate OR 1.14, CI 1.06 to 1.22), and for nurses and ancillary workers (bivariate OR 4.82, CI 2.82 to 8.24, and 3.92, CI 2.38 to 6.48, respectively) in comparison with other HCWs.

Discussion

The strain that an influenza pandemic will place on health services and the significance of the role of HCWs in the pandemic response is well recognised.6 8 9 10 16 In investigating HCWs’ perceptions of their duty to continue working during an influenza pandemic and the extent to which their stated likelihood of working is guided by factors associated with duty, this survey raises a number of issues.

The duty to work

The sense that all HCWs felt that they had a duty to work despite personal risk emerged strongly. This duty to work is typically assumed to apply only to professional HCWs who are guided by the professional norms embedded within their training and the codes of conduct that govern their practice.2 17 However, this survey suggests that similar norms regarding the duty of care to the sick extend to all HCWs regardless of professional standing. Not only was there a strong sense that “everyone should pull together during a pandemic” (which may imply a duty for everyone and not just HCWs), but agreement with the ethics statements recognising a duty to work even in difficult circumstances correlated strongly with the stated likelihood of working. This suggests that HCWs’ decisions about whether or not they are prepared to work during a pandemic are closely linked to their sense of duty, whether motivated by a professional responsibility, a general duty to help those in need or loyalty to their colleagues.15

Our previous qualitative work has demonstrated that non-professional HCWs articulate various reasons for feeling they have a duty to work during a pandemic. Some of these reasons are associated with the nature of their role(s), and others relate to the perceived obligation to respond to the national crisis or emergency that a pandemic may create.5 15 It is interesting, however, that non-professional HCWs have adopted a sense of duty to patients analogous to that of the professional duty. This raises the question of whether non-professional HCWs are owed some of the respect accorded to professional HCWs, which is in part based on the belief that they will treat the sick regardless of personal risk. Given that salary is the most prominent marker of esteem, and assuming that it is right that salary is more indicative of educational qualification than it is of dedication or hard work, alternative means need to be found of recognising the risks that non-professional HCWs are apparently prepared to take to tend the sick.

Reciprocity

Nearly all respondents believed that their employer has a responsibility both to take special measures to mitigate the risks they face in performing their role (eg, provision of personal protective equipment and vaccination) and to offer rewards in recognition of HCWs’ dedication in a time of crisis. Many HCWs seem to be willing to take necessary risks, but not unnecessary risks, and expectations about whether these risks would be mitigated by their employers may influence HCWs’ perception of whether the risks are perceived as worth taking as part of their duty to patients.

However, the interplay between reciprocity, the notion of unnecessary risk and the duty to work warrants further consideration. Clearly, employers have an obligation to mitigate occupational risk, and this has contributed to HCWs in developed countries routinely facing fewer risks at work than they did in the past.20 21 Because professional HCWs in particular are rarely called upon to face significant risks, risk may no longer be regarded as part of what they take on when they join the profession. Furthermore, since medical and nursing students are protected from risk during training, risk may not be experienced or explicitly considered until events such as SARS (severe acute respiratory syndrome) or an influenza pandemic occur. It is noteworthy that nearly 30% of nurses, 25% of hospital doctors and 18% of GPs did not think that they had a duty to work where doing so would pose a risk to themselves or their families. It was also clear from our results that HCWs do not regard unwillingness to work—ie, abandonment of the duty to work—as something that should attract punishment. Nevertheless, professional HCWs continue to enjoy the benefits of the esteem in which they are held by the public they serve. Some commentators have gone so far as to regard these benefits as a kind of social contract, in which there is an expectation that professional HCWs will fulfil their duty to work in times of crisis.2 3 Our results suggest that for some combinations of risks and other circumstances they may not be willing to do so. This raises the question of the extent to which reciprocity may be lacking on both sides.

Competing duties and responsibilities

Others have argued that a duty to work during a pandemic may be tested most by competing duties to family members.18 Family responsibility emerged from the survey as a potentially important factor in HCWs’ decision-making about whether to work during a pandemic, irrespective of any wider sense of duty to work. This was evident in the survey results, with 72.7% of HCWs agreeing that “all HCWs have a duty to work, even if there are high risks involved”, alongside 74.1% who agreed that “my main responsibility is to myself and my family”. It may be the case that employer-led action in reducing the personal risks faced by HCWs could be enough to tip the balance for many in favour of working. In some sense this is obvious: the more that risks are reduced (eg, through provision of vaccination), the more that working during a pandemic becomes like working in normal times. What is more difficult to address is the relationship between how HCWs feel about their work in the context of their employment circumstances and the extent to which they may be willing to take personal risks. Remuneration, loyalty, respect and responsibility may combine with the sense of duty to patients to outweigh the sense of duty to family.

On the other hand, it would be imprudent to underestimate the force of family responsibilities—particularly where parents and young children are concerned. Being motivated out of parental love to put the interests of one’s children ahead of all other interests is arguably the defining characteristic of being a parent. If so, the responsibility towards young children—particularly if they are themselves ill—may always outweigh responsibilities towards patients, even where those patients are vulnerable and dependent.

Competing duties and responsibilities do, however, highlight the extent to which any HCW, professional or otherwise, can be expected, in the words of the UK’s General Medical Council, to “Make the care of your patient your first concern”.22 This expectation holds even in the guidance that pertains during a pandemic.23 It is questionable whether such an entreaty can apply to the modern professions of nursing and medicine to the extent that it did when doctors were predominantly men whose family obligations extended largely to providing income, discipline and stability, and nurses were expected to be single women who were obliged to leave the profession upon marrying. The changing demography of HCWs inevitably introduces competing duties, which may leave patients vulnerable in times of crisis.

We have already explored the implications of the distinction between willingness and ability to work during a pandemic, and the way in which this distinction may inform pandemic planning.11 15 The results presented here highlight the relationship between willingness and ethical motivation. The extent to which notions and perceptions of duty might be usefully incorporated into discussions about pandemic planning and management requires further consideration, particularly with regard to communication strategies.24 For example, plans are under way to offer vaccination against swine flu to HCWs. The uptake of (seasonal) influenza vaccination by HCWs is typically low, and NHS executives are keen to ensure maximum uptake of the newly licensed swine influenza vaccine.25 Our findings suggest that individuals’ perceptions of their duty to work during a pandemic might positively influence vaccination uptake if such ethical considerations were made an explicit component of regional and national debates on the acceptability of vaccination. Perhaps information sheets currently being prepared for HCWs about influenza vaccination should outline the ethics debate about the duty to work and the duty to be vaccinated, as well as simply providing “factual” information about the vaccine itself and the arrangements for administering it to those prepared to receive it.

Limitations

In this study, participants were asked to consider their potential response to a hypothetical situation. Regardless of their stated likelihood of working and perceptions of whether or not there is a duty to work during an influenza pandemic, those who responded might behave differently during an actual pandemic than their responses to our survey suggest. Equally, if the current pandemic influenza virus remains mild, the willingness to work might be correspondingly greater, as potential risks to self and family will be reduced. Responder bias is also an inherent feature of population-based survey research. In this case, willingness to participate in the survey may be positively correlated with HCWs’ potential likelihood of working during a pandemic and, consequently, individuals’ perceptions of their duty to do so. However, analysis of the sociodemographic characteristics (age and gender) of survey responders versus non-responders did not demonstrate any statistically significant differences between those who returned the survey and those who did not. This suggests that the responses received are representative of the wider population of HCWs.

Conclusions

It is often assumed that professional HCWs have an ethical duty to work, even when faced with personal risk. The duty to put the needs of the patient first contributes to the esteem in which professional HCWs are held. Yet we found that many non-professional workers also felt a similar sense of duty and may, therefore, deserve similar respect. At the same time, this may not be justified if neither professional or non-professional HCWs are willing to put this duty into practice during times of risk. Our respondents’ sense of the duty to work may conflict (unsurprisingly) with their sense of duty to family, as well as other factors such as a perceived lack of reciprocity on the part of their employers.

A willingness to work during the current pandemic should not, however, be taken as evidence that the sense of duty to work in the face of risk remains intact. Significant proportions of those responding may be unwilling to prioritise their sense of duty to patients ahead of that to self and family, and a large proportion of doctors and nurses did not think that they had such a duty at all.