Intention of nurses to receiving influenza vaccination before the 2013–14 season

Abstract

A study was conducted to determine the influenza vaccination uptake rate of nurses in Hong Kong after the pH1N1 epidemic, and examine their intention for vaccination before the next influenza season. Questionnaires in Chinese with multiple choice responses were delivered by post through 4 nurses organizations. The following were explored: intentions and reasons for vaccination in the forthcoming season; perceptions of influenza outbreak risk; attitudes toward professional obligation and vaccination policies. Cramer's V and Eta values were calculated to analyze association and effect size. Between March and May 2013, analysis was made on 1,934 (6.8%) valid questionnaires, with 620 (30.7%) having received influenza vaccination before the season of 2012–13. Some 30.7% and 42% intended to accept and decline vaccination respectively in the forthcoming season (2013–14) while 27.3% remained indecisive. Over 80% of indecisive nurses had refused vaccination in the last season. More nurses refusing (18.1%) or indecisive (9.8%) of vaccination perceived being “significantly unwell” after past vaccinations, compared to only 1% in those who intended to accept. Expert opinions and guidelines were important for making decision on vaccination in 40% of nurses. The ranking of professional responsibilities of vaccination in indecisive nurses was between those declining and accepting future vaccination. Overall, past vaccination experience, professional responsibilities and reference to guidelines were major factors affecting nurses' decision on future vaccination against influenza in Hong Kong. The results might however not be generalizable to all nurses as the analyses were restricted to those responding to the appeal from selected professional organisations.

Keywords:

• influenza
• influenza vaccine
• nurses
• survey
• vaccination intention

Abbreviation:

• flu
• influenza

Introduction

Over the years, emergence of new strains of influenza (flu) virus has occurred, which is of concern because of the lack of herd immunity in the community. Vaccination against seasonal flu may reduce the chances of serious outbreaks and minimize the possibility of evolvement of new flu strains, as hybridization of genetic materials between animal and human influenza could occur in the human host.¹ In this connection, vaccination of health care workers is an important prevention strategy, as health care facility related flu tends to affect older and immunocompromised patients, causing high mortality.² The uptake rate of flu vaccination in health care workers has, however, been highly variable, ranging from 9% to 92% across geographic regions,³ which undermines efforts for controlling flu outbreaks.

In a former study conducted in 2009 in Hong Kong, it was shown that acceptance of seasonal flu vaccination in nurses was much higher than the monovalent vaccine against Influenza A(H1N1)pdm09 virus (pH1N1).⁴ With the incorporation of pH1N1 in trivalent seasonal flu vaccine in the ensuing years and following the global dissemination of pH1N1 as a dominant flu strain, it was anticipated that the impacts of monovalent pH1N1 vaccine would wane. A follow up survey was, therefore, conceptualized after the pH1N1 pandemic to determine the latest uptake rate in nurses, and to explore their intention for vaccination in the next flu season in Hong Kong between late 2013 and early 2014. In this study, we attempted to answer the following questions: 1) what is the proportion of nurses who is indecisive on future vaccination? 2) Are there discrepancies on attitudes toward vaccination and outbreak risk between nurses who would definitely receive, decline and those indecisive on future vaccination? 3) How did past experience of flu affect the attitudes, perceptions and acceptance of future vaccination?

Results

Respondents' characteristics and rate of vaccine acceptance

Between March and May 2013, a total of 28300 questionnaires were sent out after which 2,050 completed questionnaires were returned, giving a response rate of 7.24%. Thirty-one were duplicates and excluded. There were 85 completely blank returns. A total of 1,934 valid questionnaires were available for analysis. The demographics of the respondents are shown in Table 1.

Table 1. Demographics of respondents (total percentage in each group do not necessarily add up to 100 because of missing entries)

	Number	Percentage	Cramers'V
Gender: male	168	8.7	0.02
female	1763	91.3
Age: < 25	69	3.6	0.11
26–35	358	18.6
36–45	595	30.8
46–55	668	34.6
>56	239	12.4
Live alone: yes	161	8.5	0.04
no	1743	91.5
Working experience			0.11
< 1 yr	71	4.5
2–5 yr	152	9.7
6–10 yr	140	9.0
11–20 yr	293	18.8
> 21 yr	907	58.0
Qualification			0.02
registered nurse	1522	79.7
enrolled nurse	386	20.3
Work need to contact patients			0.07
none	170	9.0
rarely	151	8.0
sometimes	208	11.0
often	1353	72.0
Public or private sector: public clinics/hospitals	1598	83.9	0.04
private sector	307	16.1

† registered nurses have received 3 years' professional training in nursing prior to their confirmation as 'registered nurses'; enrolled nurses have received 2-years training before registration.

* p value for the Cramers'V was < 0.05.

A majority (91%) of the respondents was female and 65% were between 36–55 y old. About 76% had worked for over 10 y Most (90%) were living with their family and the majority (78%) was registered nurses who had completed formal training of 3 y Among the 1,934 nurses providing valid returns, 619 (30.7%) had received flu vaccination during the season of 2012–13.

Intention to vaccinate before the upcoming season of 2013–14

In anticipation of the forthcoming season of 2013–14, 592 (30.7%) of the 1,928 responding nurses intended to receive their vaccination; 809 (42%) did not intend to be vaccinated; and 527 (27.3%) remained undecided at the time of filling in the questionnaires. Nurses in older age group or with a longer working experience had relatively higher Cramer's V values than the other demographics, in relation to the acceptance of vaccination in the next season (Cramer's V 0.1 and 0.11 respectively, Table 1). The acceptance of vaccination in the next season was strongly associated with any history of vaccination in 2011–12/ 2012–13 (Cramer's V = 0.75 and 0.83 respectively) or when the nursing career started (Cramer's V=0.59, Table 2).

Table 2. Possible factors associated with respondents' plan of receiving flu vaccine in the forthcoming season 2013–14

	Plan for vaccination in season of 2013–14
	Yes	No	Undecided	Cramer's V	Eta
Accepting vaccination in season of 2011–12	487/589(82.7%)	21/808(2.6%)	95/523(18.2%)	0.75	0.55
Accepting vaccination in season of 2012–13	528/591(89.3%)	8/809(1%)	83/444(15.7%)	0.83	0.62
Accepting H1N1pdm09 monovalent vaccination	253/586(43.2%)	54/808(6.7%)	104/525(19.8%)	0.34	0.23
Frequency of vaccination since working as nurse	—	—	—	0.59	0.58
History of vaccination as student nurse	84/586(14.3%)	77/791(9.7%)	82/522(15.7%)	0.09	0.09
Any influenza since January 2013
No	288(49)	459(58.4)	269(51.8)	0.07	0.06
Yes	299(50.9)	323(41.1)	243(46.8)
Lab confirmed flu	1(0.2)	4(0.5)	7(1.3)
Family members in same household with influenza since January 2013
No	236(42.4)	371(49.7)	216(43)	0.03	0.03
Yes	315(56.7)	368(49.3)	277(55.2)
Definite	5(0.9)	8(1.1)	9(1.8)
Unwell after past vaccination
No	187(32.4)	98(26.9)	100(24.4)	0.39	0.37
Mild	385(66.6)	200(54.9)	270(65.9)
Significant	6(1)	66(18.1)	40(9.8)

* p value for Cramers'V > 0.05 implying the Cramers'V value is insignificantly different from 0

^† Ranked (5 classes) responses are not shown individually, which were: never vaccinated, vaccinated less than every 2 years, almost every 2 years, almost yearly and yearly respectively. The Chi Square statistics were all <0.05.

Among the nurses who were indecisive on accepting vaccination before the upcoming season of 2013–14, only 15.7% (83/444) had been vaccinated in the preceding season (Table 2). About two-thirds (> 60%) agreed that the “global and local situation of flu at the time of vaccination,” “efficacy of vaccine,” “possible side effects of vaccine and severity of flu illness” were determining factors before making a decision. Most (35–40%) of the nurses only “slightly agreed” that “expert opinion” and “hospital/government guidelines” were their considerations before vaccination. Details of data are shown in supplementary Figure 1.

Among the 1,368 nurses responding to the question on previous experiences of side effects after past vaccination, 865 (63%) reported mild symptoms and 115 felt significantly unwell. A higher proportion of nurses who refused or were undecided about vaccination for the upcoming season of 2013–14, had experiences of being “significantly unwell” after their past vaccination (18.1% and 9.8% vs 1% in those planning for vaccination, Cramer's V=0.39, Eta value=0.37, Table 2). One's experience of flu symptoms and the definitive diagnosis of pH1N1 gave a very low Cramer's V value (0.07, Eta values 0.06) in relation to vaccine acceptance for the coming season. Similarly, the presence of influenza symptoms or diagnosis in cohabitating family members gave a low Cramer's V value in relation to one's decision of vaccination for the coming season (Cramer's V and Eta = 0.03 respectively).

Association of professional attitudes and risk perceptions with intention to vaccinate before the next season 2013–14

Six statements addressing professional attitudes were listed in the questionnaires: (A) Staff vaccination is an important form of infection control; (B) staff vaccination should be compulsory; (C) vaccination lowers hospital outbreak risk; (D) healthcare staff should set an example for vaccination; (E) the proportion of healthcare staff who are vaccinated is low; and (F) vaccination of healthcare staff should be delivered on a voluntary basis. Participants were asked to grade their agreement on a Likert scale with 6 levels ranging from strongly agree to strongly disagree. The proportion in agreement with the questions for nurses indecisive on vaccination in the coming season was in-between those who intended to decline definitely and those who intended to accept. This was especially prominent for statements (A), (B), (C) and (D) above. Agreeing to these attitudes was moderately associated with vaccination with Cramer's V that ranged between 0.3–0.4 and Eta values between 0.2–0.3. The magnitude of association of factors (E) and (F) were much lower with both Cramer's V and Eta values <0.2 (Table 3).

Table 3. The relative strength of association between attitudes and risk perception with I) vaccination acceptance and II) history of influenza

	I) Plan for vaccination 2013–14		II) Symptoms of influenza since 2013
	Cramer's V	Eta	Cramer's V	Eta
Perceived risk of contracting influenza
a) I have high risk of influenza	0.2	0.2	0.15	0.2
b) Hong Kong has high chance of outbreak	0.1	0.1	0.1	0.03
c) Hong Kong has good preventive measures against influenza	0.06	0.08	0.06	0.02
Professional Attitudes
a) Hospital staff vaccination is an important form of infection control	0.34	0.23	0.08	0.06
b) Staff vaccination should be compulsory	0.37	0.3	0.06	0.04
c) Vaccination lower hospital outbreak risk	0.38	0.26	0.06	0.02
d) Hospital staff should set an example for vaccination	0.44	0.3	0.04	0.01
e) Vaccination rate among staff is relatively low	0.12	0.08	0.08	0.06
f) Hospital staff vaccination should be voluntary	0.19	0.18	0.05	0.02

The higher the Cramers'V the larger the relative strength of associations of the particular factors with (a) and (b).

* P value for Cramers'V > 0.05, implying the Cramers'V value was insignificantly different from 0.

Three statements addressed one's perceived risk of contracting influenza, risk of local outbreak in the coming year and the adequacy of current preventive measures against flu. Agreements to these questions were graded similarly. Intention for vaccination in the coming season of 2013–14 was only weakly associated with one's perceived risk of infection and outbreak risk (Cramer's V and Eta all < 0.2). There was very weak association between intention to vaccinate and the perceived adequacy of local preventive measures against influenza, as reflected by the small Cramer's V and Eta (both < 0.1) (Table 3).

Less than 30% of nurses without flu symptoms in the current year perceived that they had a high risk of contracting flu in the coming season. About 40% who had experienced flu symptoms believed they had high risk. Some 54% of nurses without symptoms of flu believed that there was a high chance of outbreak in Hong Kong, compared to 65.8% in those who had symptoms of flu in the past year. About two-thirds of the 12 nurses with laboratory diagnosed influenza believed themselves to be at high risk of contracting flu and in a high chance of outbreak.

Less than 20% of nurses, regardless of their experiences with flu symptoms in the past year, generally agreed that vaccination should be compulsory. For those agreeing to compulsory vaccination, > 80% believed that vaccination could prevent hospital outbreaks and current health policy could not protect healthcare staff against flu, while about 60% felt that compulsory vaccination was not difficult to administer and the voluntary vaccination rate was too low. For those who disagreed with compulsory vaccination, > 80% believed that vaccination should be a personal choice and compulsory vaccination was difficult to administer.

Discussion

In our study, acceptance of seasonal flu vaccination among nurses in Hong Kong in 2012–13 was around 30%, which was slightly higher than the 24.1% in the preceding year.^5,6 The flu vaccine acceptance rate has been declining gradually since 2005, and the perceived severity of flu was an important factor influencing vaccine uptake. Kraut reported that healthcare workers who perceived pH1N1 to be more severe than seasonal flu had a higher rate of pH1N1 vaccine acceptance than the seasonal flu vaccine.⁷ This finding was consistently reported in other studies.^3,8,9 In a German study targeting healthcare workers, only 16.4% expressed that pH1N1 had influenced their decision of vaccination, while 11% of those declining vaccination did not perceive it as a severe disease.¹⁰ In our study, only 30–40% of nurses perceived they had high risk of contracting flu. According to the WHO, pH1N1 has become the dominant strain causing influenza globally.¹¹ The milder nature of pH1N1 compared to avian influenza likely explained why this was no longer treated as a life-threatening infection. The perceived occurrence of severe adverse reactions after vaccine against a mild form of influenza was another probable deterrent in healthcare workers against subsequent vaccinations. Publicized adverse reactions to the monovalent pH1N1 vaccine had caused public concern when pH1N1 strain was incorporated into the seasonal flu vaccine.⁵ This may, however, change with time, as adverse reactions had not been reported to be excessive following such incorporation in subsequent years. In this study, older age or a longer work experience had a relatively stronger association with acceptance of vaccination compared to other demographics. These findings were similar to 2 recent studies in Spain and Italy in which age was noticed to be associated with vaccine uptake.^12,13 However, in these studies, the presence of chronic illness was also found to be associated with vaccine uptake, which could confound with the age factor. Similarly, working experience could be confounded by age. The extent of making contact with patients at work did not have a strong association with vaccine acceptance. This probably reflected that clinical work, involving patient contact, was not considered as a high risk factor for contracting flu.

Jose et al reported that 58% or more healthcare workers experienced side effects after vaccination, especially in those vaccinated with pH1N1.¹⁴ Our study gave a similar overall rate of over 60%, and these “significant” side effects were associated with refusal of vaccination in the forthcoming season of 2013–14. Side effects have been cited as a major reason for refusal of vaccination in studies and reviews.^3,7,15 Doubtfulness on vaccine efficacy was the second most common reason to decline vaccination. This perceived ineffectiveness was evident in those who intended to decline pH1N1 vaccination. Alkuwari et al reported that healthcare workers who doubted vaccine effectiveness were 0.2 times less likely to vaccinate,⁸ while Savas reported a likelihood of up to 4.89 times higher to vaccinate for those believing in vaccine efficacy.¹⁶ In our study, neither definitive laboratory diagnosed flu, nor experience of flu symptoms was associated with the acceptance of vaccination (Table 2, with a low Cramer's V and Eta value <0.1). As the number of definite diagnosed flu was very small (only 12 with serotypes), generalization of the findings is not possible. The presence of flu in cohabitating family members was also unimportant in the decision of vaccination (Cramer's V =0.03)

In our study, about 27% of nurses were indecisive in vaccination for the coming 2013–14 flu season. The rate was higher compared to 21.6% in the previous survey in 2009.⁴ The ultimate decision of these nurses, and possible means of influencing their attitudes, would affect significantly the overall vaccination rate. Neither perceiving a high risk of contracting flu nor a high chance of outbreak in Hong Kong were important factors for deciding on vaccination, as over 60% of them considered that good preventative measures against flu were already in place. Expert opinion and local government/hospital authority guidelines were relatively less important, a consistent observation for respondents in the season of 2012–13 (data not shown). Elsewhere, Sevencan et al reported that negative attitude toward authority advice reduced the likelihood of vaccination and vice versa.¹⁷ From a study conducted in Qatar, healthcare workers perceiving government advice on the impacts of pH1N1 to be over-exaggerated or over-estimated were less likely to vaccinate.⁸ In our study, the levels of agreement with the infection control purpose of flu vaccinations in indecisive nurses were intermediate between those who definitively agreed and others declining to vaccinate in the coming season of 2013–14. The findings were similar to a local study in which healthcare professionals with a positive attitude in professional responsibility and protecting patients were more likely to accept having the flu vaccine.¹⁸ It has been shown that structural in-service education, campaigns or conferences were associated with improved vaccination rates.^19,20 Indecisive nurses thus constituted an important group to target for enhancing the overall vaccine uptake rate in the healthcare settings.

In designing future vaccination promotion strategy, the main reasons for accepting and rejecting vaccination should both be addressed. Nurses accepting vaccination in the last season 2012–13, were more positive toward their professional responsibility of vaccination in preventing hospital and local outbreak, compared to those declining vaccination. Strengthening the values of professional responsibility is, therefore, a meaningful strategy in promoting vaccination acceptance. Hakim et al reported a very high vaccination acceptance rate of over 80% through a comprehensive campaign in which one component was to strengthen professional responsibility.⁹ Role models of senior staff could be another important strategy to encourage front-line healthcare workers to participate in vaccination programs and have been recommended by the Centers for Disease Control and Prevention (CDC) and National Vaccine Advisory Committee (NVAC) in the US^19,20 Successful examples with role modeling were reported and summarized by Lam et al in a Canadian review.²¹ In comparison, a negative impact on vaccine uptake was evident when the Turkish Prime Minister rejected pH1N1 vaccination.¹⁶ Locally, the Hospital Authority of Hong Kong, the organization running all public hospitals, has adopted various measures with similar principles to enhance vaccination acceptance among healthcare workers (HCW). Campaigns are run within hospitals with posters encouraging HCW to receive vaccination. The vaccination process has been made more convenient by “vaccination on wheel” program, through which a team of mobile nursing staff approach different work places within hospitals proactively to encourage vaccination. As stipulated in the Hospital Authority's policy detailed in the influenza preparedness plan, every hospital cluster “would drive annual influenza vaccination among HCW, in the form of campaign, role models, programs etc., in order to achieve a high coverage rate, for protection of the HCW and the patients they serve.(clause 6.9.4)”²² Finally, while mandatory vaccination is an important policy for ensuring high flu vaccination uptake in places like Singapore,¹⁵ this policy is not supported by our respondents here in Hong Kong. Although mandatory vaccination could be an effective measure to increase vaccination rate and most of the nurses were aware of the benefits, they generally disagreed on the compulsory policy. Mandatory vaccination programs have also received lots of criticism, although it has been shown to reduce hospital related influenza like illness.²³

This study has a number of limitations. Foremost, we used large numbers of positive ended questions which most nurses tended to agree. Other limitations included recall bias and selection bias. Nurses who responded to the questionnaires could have been more motivated in expressing their opinion and choices compared to non-responders. Only about 7% of the nurses responded by sending back the surveys. The responding nurses might differ in demographic characteristics compared to the non-respondents and, thus, potentially the strength of the findings could be biased and, therefore, could not be generalized. Although this study was a follow-up survey to the one administered in 2009 that addressed pH1N1, the differences in demographics of the respondents between the 2 studies implied the results were not directly comparable. There were more nurses in the age group > 56 y old in this study than the previous one (12% vs 8.3 % ) and less in those between 36–55 (65.4% compared to ˜80% in previous studies).⁴ Although this could have been the result of aging of the same population in the previous studies, it could not be verified because of the anonymous nature of the survey. There were also more enrolled nurses with shorter period of training in this study. As the surveys were administered through similar organizations with a similar method, the limitations were comparable. Nevertheless, we tried to sample as representative as possible of the population. On the other hand, it was also suboptimal to identify duplicated entries through self-reporting by the respondents, as accuracy could be heavily affected by willingness to report. Thus the magnitude of duplicates was unknown. Screening the lists of recipients with removal of duplicates prior to posting of the survey would have been ideal, but that implied intrusion of privacy without consent. Lastly, nurses on the mailing list of the professional organizations might not necessarily represent the entire local nursing population. Extrapolation of results should therefore be cautioned.

In conclusion, flu vaccine acceptance in healthcare workers could be heavily influenced by one's past experience of side effects arising from previous vaccinations, attitudes toward implications of adverse reactions, and perceived vaccine effectiveness. From our study, it can be seen that nurses' regular vaccination behavior since the start of their nursing career and their beliefs in professional obligation did positively affect the flu uptake rate. It can be envisaged that education and campaign activities strengthening these professional values, interventions clarifying the myths of vaccine side effects/ineffectiveness, as well as role modeling by senior healthcare colleagues, are potentially effective ways to improve vaccine uptake rate in nurses. Given the high percentage of nurses remaining undecided on vaccination before the next flu season, they are particularly important to be targeted to enhance preparedness of the healthcare profession. To our knowledge, the perception of this particular group of nurses has hardly been specifically investigated in other studies. Thus, our findings could serve as reference for targeting indecisive nurses to increase their vaccine acceptance rate.

Methods

A survey was conducted by using self-administered questionnaires, which were distributed by post through 4 professional bodies of nursing: the Hong Kong Nurses General Union; both the Nurses Branch and Enrolled Nurses Branch of the Hong Kong Chinese Civil Servant's Association; and the Association of Hong Kong Nursing Staff. A total of 28,300 questionnaires were sent out, representing about 60% of nurses in Hong Kong. Respondents were asked to return the completed questionnaires by freepost the delivery of which implied their consent to participate. To avoid duplications, participants were excluded if they indicated they had answered the same survey through the other organizations. Approval was obtained from the Survey and Behavioral Research Ethics Committee of The Chinese University of Hong Kong.

The questionnaire consisted of multiple choice questions in Chinese language divided in 4 sections. Section 1 covered one's reasons of accepting or rejecting vaccination and past experiences in flu vaccination. Section 2 assessed the history and experience of contracting flu. Section 3 examined perceptions toward flu vaccination, the responses to which were ranked. The last section covered demographic data. All returns were anonymous. Statistical analysis was performed using IBM SPSS Statistics 20.0 (IBM, U.S). The association of nominal or ordinal data with more than 2 categories of unequal columns and rows in the data matrix was explored using Cramer's V, which follows the principles of Chi Square statistics with a similar interpretation of value between 0 and 1. Cramer's V allows comparison of relative importance between the factors, a higher value of which indicates the more important the factor would be. Eta values were calculated to describe the effect size where appropriate, as the conventional p value only report either statistical significance or non-significance which could be misleading in interpreting the magnitude of relationship between factors.^24,25 Association and effect size were considered modest when Cramer's V and Eta value reached 0.4 or above. The conventional p value of < 0.05 was taken as the cut off point for describing if the Cramer's V value was significantly different from 0.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website

Acknowledgments

We would like to thank the Hong Kong Nurses General Union, both the Nurses Branch and Enrolled Nurses Branch of Hong Kong Chinese Civil Servant's Association, and the Association of Hong Kong Nursing Staff in supporting our study and distributing the questionnaires. Thanks also go to Miss Mandy Li for creating and managing the data system, and Li Ka Shing Institute of Health Sciences for the technical support provided. Special thanks to Dr. David William Sorrell (EdD) for reviewing the English writing.

Funding

This study was funded by the Hong Kong Mood Disorders Center of The Chinese University of Hong Kong.