Consumer reporting of adverse events following immunization (AEFI)

Abstract

Passive reporting of adverse events following immunization (AEFI) by consumers or healthcare professionals is the primary mechanism for post-marketing surveillance of vaccine safety. Although recent initiatives have promoted consumer reporting, there is a lack of research concerning consumer reporters. Computer assisted telephone interviews (CATI) were conducted in 2011 of a cross-sectional, random, general population sample of 191 South Australian parents who stated that their children had previously experienced an AEFI. We compared awareness of surveillance, vaccine safety opinions, and demographics of parents reporting an AEFI to either healthcare professionals or surveillance authorities with those who did not report their children’s AEFI. Multivariate regression analyses measured: the association between reporting and safety views; and demographic predictors of reporting an AEFI. Reporting an AEFI to a healthcare professional or a surveillance authority was not significantly associated with awareness of a surveillance system. AEFI reporters, when compared with non-reporters, were more likely to be Australian-born (OR = 4.58, [1.64, 12.78], P = 0.004); were associated with the perception that a serious reaction was more likely to occur at their children’s last immunization (OR = 2.54 [95%CI 1.22, 5.30], P = 0.013); and were less accepting of the risk of febrile convulsion, (OR = 3.59 [95%CI 1.50, 8.57], P = 0.004).

Although reporting an AEFI was not associated with awareness of surveillance or most socio-demographics, the results suggest some difference in safety opinions. Further studies are required to ascertain if these differences pre-date the occurrence of an AEFI or are a consequence of the AEFI and how consumers can contribute further to vaccine safety surveillance.

Keywords: :

• adverse event following immunization
• vaccine safety attitudes
• parent
• health survey
• surveillance

Introduction

Post-marketing safety surveillance (PMS) is a crucial component of pharmacovigilance for all medications and vaccines.^1,2 Passive surveillance is considered essential for PMS and is recommended by the World Health Organisation (WHO) in all countries to provide a minimal capacity for vaccine safety monitoring and build national strategies dedicated to vaccine pharmacovigilance.³ This mechanism of surveillance involves consumers (vaccinees or their parents) and/or healthcare professionals recognizing and spontaneously submitting reports of suspected AEFI to health authorities for subsequent investigation and response when required. It is widely acknowledged that this method of safety surveillance is limited by incomplete reporting information, under-reporting, biased reporting, the inability to distinguish coincidental from causal events and delayed notifications.^2,4-7 The reasons for this are multifactorial but include a poor understanding of safety surveillance by healthcare providers, limitations on provider time, and unclear interpretations of what constitutes a reportable AEFI.^8,9 In a recent study we investigated the experience of AEFI reporting and found that healthcare professionals’ definitions of a reportable AEFI varied across professional groups and work settings.¹⁰

Consumer involvement is important for effective pharmacovigilance, with an increasing recognition of the benefits of consumer reporting of adverse events following the administration of medications or vaccines.^11,12 Until recently, most PMS systems worldwide have restricted reporting of adverse events to healthcare professionals.¹³ As under-reporting by healthcare professionals is a known, major limitation of passive reporting systems, it is becoming more common to incorporate consumer reporting into PMS. Recent measures to strengthen pharmacovigilance in Europe include new legislation introduced by the European Medicines Agency (EMA), that as of July 2012, requires all member states to incorporate direct consumer reporting of adverse events to their national medicine authority.¹⁴

In Australia, consumer reporting of adverse drug reactions, (ADRs) for all medications, including vaccines, was introduced at a national level in 2003.¹⁵ Currently, direct consumer reporting of ADRs is promoted through the Adverse Medicine Events Line, a telephone reporting facility.¹⁶ In addition, consumers can report directly to the Australian national regulatory authority, the Therapeutics Goods Administration (TGA), by submitting a web based report.¹⁷ In 2011, there were 2,327 AEFI reports collected by the TGA and of these, only 2% were consumer reports.¹⁸ In addition to reporting at a national level, most jurisdictions in Australia accept consumer reports,¹⁹ with some jurisdictions actively promoting consumer reporting. For example, in Western Australia during the past two years, approximately 20% of all AEFI reports were notified by consumers.²⁰

The need for a rigorous and timely vaccine safety surveillance has received much attention in recent years in Australia, as a result of the suspension of a pediatric seasonal trivalent influenza vaccine, (STIV), Fluvax, CSL in 2010, due to increased rates of febrile convulsion in children aged ≤5 years.¹⁹ One of the key recommendations of a national review of the Australian AEFI surveillance system commissioned as a result of the STIV safety signal, was that consumer knowledge and awareness of the surveillance system for AEFI reporting required improvement.¹⁹ Although consumer reporting is encouraged as a strategy to complement healthcare professional reporting, little is known about consumers who report an AEFI. For example, identifying who is likely to report or not report an AEFI and an understanding of the reasons for doing so could inform strategies aimed at improving consumer reporting. We have previously found that parents reporting their children’s AEFIs was primarily related to media attention within the context of a known vaccine safety signal and that parents frequently report their children’s AEFIs to healthcare professionals, rather than direct to a surveillance authority.²¹ There is limited research on factors such as socio-demographics or vaccine safety opinions that may be associated with parental reporting of their children’s AEFI symptoms to healthcare professionals and/or surveillance authorities. Since the detection of an AEFI may commence with a parent’s concern and report, it is critical to understand these factors.

In this study, the primary aim was to examine the vaccine safety opinions and socio-demographic characteristics of parent AEFI reporters and non-reporters in the general population. Our second aim was to test for predictors of reporting a child’s AEFI to a surveillance authority or healthcare professional.

Results

To complete the 2002 interviews, a total of 4700 telephone listings were selected. Of these, 3600 were eligible numbers: 1100 were ineligible either because they were not residential numbers, were disconnected, were fax/modem numbers or corresponded to households located outside of SA. Of the remaining households, 993 refused to be interviewed, 275 were not contactable after six attempts, 229 were either not available or unwell and 101 spoke no English. a participation rate of all eligible households of 55.6%.

Of the total adult sample of 2002, 469 (23.4%) were parents or legal guardians of one or more children aged <18 years of age (n = 929). Of these 469 parents 191 (41%) reported that one or more of their children (n = 269) had previously experienced an AEFI.²² At the time of interview approximately 60% of all children who had experienced the AEFI were aged ≥6 years.²² The type of symptoms described by parents has been previously described, with the majority being common and expected reactions associated with childhood immunizations.²²

Parents were asked “Were any of the symptoms ever reported to any of the following?” with the response options recorded as: the “Department of Health,” a “GP” (family physician), “Parent Helpline” (a South Australian parent telephone information service funded by the Women’s and Children’s Health Network), “Immunisation nurse,” “other,” or “did not report.” Of the 191 parents, 32% (n = 62) stated they had reported their child’s AEFI. Most (96%) AEFI symptoms were reported to healthcare professionals.

Parent demographics

The mean age of the parents interviewed was 38.8 y (95% CI: 37.7 to 39.9) and a range of 19 to 62 y. When comparing demographics, a greater proportion of AEFI non-reporters were born outside Australia (24%) compared with the AEFI reporters (9%), (P = 0.007). No other statistically significant difference in demographics was observed between the parent AEFI reporters and AEFI non-reporters (Table 1).

Table 1. Household demographics of survey respondents (n = 191)

	AEFI reporters N/62 (%)	AEFI non-reporters N/129 (%)
Sex
Male	15 (23.2)	46 (35.7)
Female	47 (76.8)	83 (64.3)
Age
18–34 y	19 (30.7)	36 (28.0)
35–34yrs	32 (52.6)	66 (50.7)
≥45 y	10 (16.7)	28 (21.3)
Total Children
1 child	16 (26.0)	35 (26.8)
2 children	25 (41.2)	65 (49.8)
3 children	10 (16.7)	21 (16.1)
>3 children	10 (16.1)	9 (7.3)
Education
Secondary School	21 (34.8)	28 (21.4)
Trade/certificate/diploma	23 (37.7)	56 (43.3)
Bachelor degree or higher	17 (27.5)	46 (35.3)
Annual household income ($AU)
≤20 000	2 (4.0)	3 (2.2)
20 001–40 000	3 (5.0)	9 (7.1)
40 001–60 000	8 (13.0)	18 (14.1)
60 001–80 000	12 (18.7)	26 (20.0)
80 001–100 000	12 (19.6)	27 (21.1)
>100 000	18 (28.7)	32 (24.8)
Not stated	7 (11.0)	14 (10.8)
Country of birth
Australian born	57 (91.5)	98 (23.9)
Born elsewhere	5 (8.5)	31 (76.1)
Main language spoken at home
English	61 (98.6)	123 (95.9)
Other	1 (1.4)	6 (4.1)
IRSD
Least disadvantaged (IRSD 1)	10 (15.8)	22 (16.9)
Middle (IRSD 2–4)	39 (63.7)	83 (64.3)
Most disadvantaged (5)	13 (20.5)	24 (18.8)

Totals for each characteristic are weighted totals, rounded to zero decimal place. Proportions for each household characteristic may not add up to 100% due to rounding of figures to one decimal place. AEFI reporters are parents who indicated that they had previously reported any of their children’s AEFI symptoms to either a healthcare professional or surveillance authority. AEFI non-reporter are parents who indicated that they had not previously reported any of their children’s AEFI to either a healthcare professional or surveillance authority. Socio-Economic Indexes for Areas (SEIFA) area-based index of relative socioeconomic disadvantage (IRSD) derived from residential postcode and based on the Australian census data.

Safety views

The majority of parents surveyed (91% of reporters and 96% of non-reporters) believed vaccines were safe (Table 2). Almost double the proportion of AEFI reporters compared with non-reporters believed a serious AEFI was “very” or “somewhat likely” at their youngest child’s last immunization (21% vs 11%, P = 0.004). The AEFI non-reporters stated lower levels of concern for vaccine safety for nearly all survey items, although most of the differences observed in response proportions were not statistically significant (Table 2). Approximately two thirds of respondents in both groups stated they were aware of a surveillance system for vaccine safety.

Table 2. Parental safety opinion by AEFI reporting status

Respondents	N
AEFI reporters	62
AEFI non-reporters	129
	Response, N (%)					X^2(P)
In general, how safe would you say the vaccines given to children are?
	Very unsafe	Unsafe	Undecided	Safe	Very safe
AEFI reporters	2 (2.7)	2 (2.5)	3 (4.3)	27 (43.3)	29 (47.2)
AEFI non-reporters	1 (0.6)	2 (1.5)	3 (2.1)	57 (43.7)	68 (52.3)	3.0 (0.60)
How concerned are you that new vaccines have been adequately tested for safety before they are released to the public in Australia?
	Very concerned	Somewhat concerned	Undecided	Not too concerned	Not at all concerned
AEFI reporters	18 (28.8)	10 (17.0)	1 (1.4)	18 (28.5)	15 (24.4)
AEFI non-reporters	28 (21.5)	44 (33.8)	0 (0)	35 (26.9)	23 (17.8)	8.2 (0.12)
How likely did you think he/she would experience a reaction such as fever, irritability or redness at the injection site?^b
	Very likely	Somewhat likely	Undecided^a	Not too likely	Not at all likely
AEFI reporters^c	23 (39.2)	21 (35.9)	0 (0)	10 (17.5)	4 (7.4)
AEFI non-reporters^d	38 (29.8)	56 (44.5)	0 (0)	23 (18.1)	10 (7.6)	1.9 (0.64)
How likely did you think your child would experience a reaction that would need medical treatment from a hospital or GP?^e
	Very likely	Somewhat likely	Undecided	Not too likely	Not at all likely
AEFI reporters^f	4 (6.6)	8 (13.9)	0 (0)	29 (50.5)	17 (29.0)
AEFI non- reporters^g	3 (2.3)	10 (8.4)	0 (0)	46 (37.3)	63 (51.9)	9.9 (0.04)
Acceptability of febrile convulsion risk^h
	Not acceptable		Undecided		Acceptable
AEFI reporters	18 (29.9)		4 (6.5)		39 (63.7)
AEFI non-reporters	20 (15.7)		8 (6.2)		101 (78.0)	5.6 (0.09)
Acceptability of Anaphylaxis risk^i
	Not acceptable		Undecided		Acceptable
AEFI reporters	10 (16.4)		7 (11.6)		44 (72.0)
AEFI non-reporters	21 (16.3)		7 (5.3)		104 (78.4)	2.7 (0.29)
Are you aware that a system for checking and assessing vaccine safety exists in Australia?
	No		Undecided		Yes
AEFI reporters	19 (31.1)		5 (8.0)		38 (60.9)
AEFI non-reporters	47 (36.6)		10 (7.6)		72 (55.8)	0.6 (0.79)

Totals and proportions may not add up to sample totals and 100% due to rounding of figures. ^aCategory “undecided” was omitted from question 3 and question 4 analyses. ^bThe question: “Thinking back to when your youngest child was last immunized, how likely did you think it would be that (s)he would experience a reaction, such as fever, irritability or redness at the injection site? Would you say this type of reaction was: very likely, somewhat likely, not too likely, or not at all likely?” The response categories, “Did not consider it at the time,” and “Don’t know” were recorded if stated by parent respondent. ^cExcludes 3 missing. ^dExcludes 1 “did not consider,” 1 missing . ^eThe question: “Thinking back to your youngest child’s last immunization, how likely did you think it would be that (s)he would experience a reaction that would need medical treatment from a hospital or GP? Would you say: very likely somewhat likely, not too likely or not at all likely?” ^f3 missing. ^gexcludes 6 “did not consider,” 1 missing. ^hThe question:“Young children who develop a fever after immunization may sometimes go on to have a fit or seizure, known as a febrile convulsion. The risk of a febrile convulsion is approximately one in every 12,000 children immunized. When considering whether to vaccinate your child would you say this risk of febrile convulsion is: acceptable ornot acceptable.” A response of “undecided” was recorded if parents stated they did not know. ⁱThe question: “Anaphylaxis is a severe, allergic reaction that can occur after immunization. It requires immediate, medical treatment. The risk of experiencing an anaphylactic reaction to any vaccine ranges from approximately 1 to 10 for every 1 million doses of vaccine. When considering whether to vaccinate your child would you say this risk of anaphylaxis is: acceptable or not acceptable. A response of “undecided” was recorded if parents stated they did not know.

In both univariate and multivariate regression analyses that tested for demographic predictors of reporting, of all the variables, Australian-born parents were significantly associated with reporting an AEFI, unadjusted OR = 3.38 (1.34, 8.53), P = 0.01, adjusted OR = 4.58, (95%CI 1.64, 12.78), P = 0.004, compared with parents born elsewhere (Table 3).

Table 3. Socio-demographic predictors of parental reporting of children’s AEFI, n = 191

Predictor Variable	% (95%CI)	Unadjusted Odds Ratio, (95%CI)	P	Adjusted Odds Ratio, (95% CI)	P
Sex
Male	31.7 (24.4, 39.9)	referent		referent
Female	68.3 (60.1, 75.6)	1.83 (0.81, 4.19)	0.147	1.93 (0.78, 4.75)	0.153
Country of Birth
Overseas Born	18.9 (13.8, 25.3)	referent		referent
Australian Born	81.1 (74.7, 846.2)	3.38 (1.34, 8.53)	0.01	4.58 (1.64, 12.78)	0.004
Parent Age
18–24	28.9 (21.9, 37.0)	referent		referent
35–44	51.3 (43.6, 58.9)	0.95 (0.42, 2.13)	0.898	1.13 (0.46, 2.79)	0.792
45+	19.8 (15.0, 25.7)	0.72 (0.28, 1.82)	0.482	0.78 (0.25, 2.41)	0.662
Total Children
1 child	26.5 (20.4, 33.7)	referent		referent
2 children	47.0 (39.5, 54.7)	0.85 (0.38, 1.90)	0.696	0.88 (0.32, 2.18)	0.711
3 children	16.3 (11.4, 22.7)	1.07 (0.39, 2.88)	0.893	1.06 (0.33, 3.36)	0.927
>3 children	10.1 (6.2, 16.2)	2.27 (0.68, 7.60)	0.182	2.12 (0.51, 8.72)	0.296
Education
Secondary School	25.7 (19.7, 32.8)	referent		referent
Trade/certificate/diploma	41.5 (34.2, 49.2)	0.54 (0.25, 1.17)	0.116	0.54 (0.22, 1.33)	0.178
Bachelor degree/higher	32.8 (25.9, 40.4)	0.48 (0.20, 1.14)	0.095	0.59 (0.22, 1.63)	0.307
Household Income *
≤20 000	3.1 (1.5, 6.4)	referent		referent
20 001–40 000	7.2 (3.9, 12.9)	0.39 (0.05, 2.78)	0.344	0.59 (0.06, 5.62)	0.644
40 000–60 000	15.4(10.4, 22.2)	0.51 (0.10, 2.76)	0.433	0.72(0.08, 6.15)	0.762
60 000–80 000	21.9 (15.7, 29.9)	0.52 (0.10, 2.71)	0.432	0.93 (0.12, 7.14)	0.943
80 000–100 000	23.1 (17.1, 30.4)	0.51 (0.10, 2.55)	0.413	0.95 (0.13, 7.11)	0.957
>100000	29.2 (22.4, 37.2)	0.64 (0.13, 3.13)	0.578	1.52 (0.20, 11.47)	0.683
IRSD
Least disadvantaged **	16.5 (11.8, 22.6)	referent		referent
Middle	64.1 (56.4, 71.1)	1.06 (0.45, 2.51)	0.9	1.08 (0.41, 2.85)	0.883
Most disadvantaged	19.4 (13.7, 26.6)	1.16 (0.39, 3.45)	0.785	0.85 (0.25, 2.91)	0.793

*21 Refused/don’t know. **Socio-Economic Indexes for Areas (SEIFA) area-based index of relative socioeconomic disadvantage (IRSD) derived from residential postcode and based on the Australian census data.

After adjusting for all possible demographic confounders in the multivariate analyses, AEFI reporters when compared with non-reporters, were associated with more concern for vaccine safety on two survey items. They perceived a serious reaction was more likely to occur at their children’s last immunization OR = 2.54 (95%CI 1.22, 5.30), P = 0.013, (Table 4). AEFI reporters were less accepting of the febrile convulsion risk of vaccines when compared with the non-reporters, OR = 3.59 (95%CI 1.50, 8.57), P = 0.004, (Table 4).

Table 4. Association between reporting an AEFI and vaccine safety views, n = 191

	Unadjusted OR (95% CI)	P	Adjusted OR^a (95% CI)	P
Vaccine safety in general	1.34 (0.39, 1.43)	0.380	2.00 (0.89, 4.51)	0.095
Concern about safety testing	0.86 (0.45, 1.64)	0.651	1.05 (0.52, 2.12)	0.896
Likelihood of a mild reaction	1.30 (0.69, 2.43)	0.405	1.35 (0.67, 2.70)	0.401
Likelihood of a serious reaction	2.49 (1.33, 4.67)	0.005	2.54 (1.22, 5.30)	0.013
Acceptability of febrile convulsion risk	2.09 (1.04, 4.20)	0.038	3.59 (1.50, 8.57)	0.004
Acceptability of anaphylaxis risk	1.32 (0.65, 2.68)	0.434	1.36 (0.57, 3.28)	0.483
Awareness of a surveillance system	0.80 (0.42, 1.52)	0.496	0.69 (0.33, 1.45)	0.325

^aadjusted for gender, country of birth, age, total children, education, income, and IRSD.

Discussion

In this study we examined and compared the socio-demographics and vaccine safety views of parents who had previously reported their children’s AEFI symptoms to either health professionals or surveillance authorities with those who did not report their children’s AEFI. The study results demonstrate that reporting an AEFI was not related to most parental socio and other demographics and that having previously reported to a healthcare professional or surveillance authority was not associated with greater awareness of a surveillance system. Our study shows that the reason parents report an AEFI was mostly independent of socio-demographic factors.

The aim of the general population survey analyses presented in this paper was to examine if safety views and/or socio-demographics differed according to whether parents reported their children’s AEFI to a health professional or surveillance authority. Although the reporters’ safety opinion results indicated greater concern (Table 2) on most survey items, only one was statistically significant. This may suggest parent reporters and non-reporters did not differ greatly in safety opinions. Furthermore, the types of AEFI parents described were common and expected effects of immunisation,²² which may also explain why safety opinions did not differ greatly.

With regards to predicting characteristics of parents who will report an AEFI to a surveillance authority or health professionals, of all demographic data collected, we found no significant predictors of reporting other than country of birth, with Australian parents more likely to report, than those born outside Australia. We cannot attribute reporting to parents’ language as nearly all reporters and non-reporters had indicated English was the main language spoken at home. This may also be explained by the exclusion of survey sample respondents who were not fluent in English, as the interviews were conducted in English only. However, reporting an AEFI in this general population study to either surveillance or health professionals was significantly associated with a greater expectation of a serious AEFI occurring and less acceptance of the risks of febrile convulsion (Table 4).

In a previous analysis of all 469 parents in the initial study sample, we found that parents who stated their children had experienced an AEFI (“AEFI parents”) were significantly more concerned for vaccine safety in general and were more likely to expect a mild AEFI, than those whose children who did not experience an AEFI (“no-AEFI parents”).²² The present analyses that compare the opinion results by whether parents reported their children’s AEFI to a health professional or surveillance authority or not demonstrated statistical significance on two different survey items. The AEFI reporters were more likely to perceive a serious AEFI could occur and were more concerned about febrile convulsion risk.

The study results should be considered with potential limitations. The analyses presented are based on cross-sectional data and from a relatively small sample, which may have reduced our statistical power to detect differences. However, the more powerful tests of association, ordinal logistic regressions did detect statistical significance on two of the survey items. Due to the sampling methodology of the Health Monitor it was not possible to attain the proportion of non-responders who were parents of children aged < 18 y from the 3600 total eligible households sampled. Parents were identified after the selected adult had agreed to participate in the interview via the survey questions. The study participants were asked to recall the occurrence and description of their children’s AEFI (self-reports), in addition to whether the symptoms were reported to health professionals or a surveillance authority, the Department of Health. Given that almost two-thirds of the children who had experienced an AEFI in this study were aged ≥ 6 y, it is likely that the children’s previous vaccinations and AEFI may have occurred several years prior to interview and as such, may have influenced respondents’ recall of AEFI and reporting. We did not verify parents’ reports with health providers as it was not within the scope of this study and because of the potentially long period since the event occurred and was reported. Similarly, we did not collect information regarding the timing of children’s onset of AEFI, the associated vaccine(s), the circumstances, and timing of reporting their children’s AEFI were also not known and not elicited due to potential for inaccuracy in details collected retrospectively. However, the nature of the AEFI in the study sample was similar in both reporter and non- reporter groups and therefore likely to be valid. The results regarding the likelihood of experiencing a mild/serious AEFI at their last immunization should be interpreted with caution as these questions were asked for only the youngest child. It may be possible that parents’ perceptions were influenced by the youngest child’s earlier vaccinations and/or vaccination experiences of older children. A further limitation is that responses to the survey questions were not validated independently with specific validation methods/tools, which suggest it is difficult to assess respondents’ understandings of each question. However, it is likely that the use of clearly defined questions regarding the type of AEFI and presenting the febrile convulsion and anaphylaxis risks numerically to respondents (Table 2) would have minimised the potential risk of misunderstanding of the questions’ meanings.

The framing of the safety opinion questions, whereby respondents were asked for a fixed response, did not allow for a more in-depth exploration of respondents’ views. For example, a qualitative research method approach may be useful in understanding parental acceptability of vaccination risks. One approach could be to explore understandings of the risks we presented with perceptions of the potential risks of vaccine preventable diseases. In particular, as parents indicated awareness of surveillance, but chose to report to health professionals, a qualitative approach may provide a deeper understanding of parent motivation to reporting²³ and further examine parent knowledge of reporting channels other than health professionals. We did not ascertain reasons for parents reporting or not reporting their children’s AEFI and this may provide a further distinction in determining factors associated with reporting. Finally, the interpretation of reporting as a predictor of safety opinion in the general population cannot be interpreted as causative, because of the cross-sectional study design. Having reported may have biased their responses to how likely they thought their children would be to experience an AEFI. In order to address the limitations of such bias and that of using retrospective data, a larger, prospective study examining opinions before or at the time of immunization, and then obtaining information on subsequently reported AEFIs to surveillance and medical providers, together with reasons for reporting or not would be required.

Studies in countries with existing patient reporting schemes comparing ADR reporting for medications or vaccines by consumers and healthcare professionals have demonstrated that consumers can contribute significantly to successful pharmacovigilance, identifying new ADRs not previously reported by health professionals, and in providing relevant reports.^24-30

Blenkinsopp (2007) proposes that patients will report an ADR if they believe their doctors will not.¹¹ Not reporting an ADR or an AEFI may relate to consumers accepting an adverse event in a trade-off for the perceived benefits of a medication or vaccine. This would seem a reasonable explanation for parents not reporting an AEFI in this general population study, given the reactions described were known vaccine side-effects.

The results in this study suggest that generally, there was no difference in vaccine safety opinion between reporters and non-reporters. Reporters were however, more concerned about the likelihood of a serious AEFI occurring and less accepting of febrile convulsion risk. Although the majority of parents stated they were aware of the existence of a system for vaccine safety surveillance, this study cannot explain why they chose to report to health professionals on the whole. As the majority of reports in this study were notified to healthcare professionals,²² it is not surprising that knowledge of a system was not a factor associated with reporting. This could relate to parental preference toward medical professionals as sources of vaccine safety.^31-34 It may also suggest that parents in this study had a general sense or trust that there is a safety system, yet were not actually aware that an AEFI could be reported directly to local immunization authorities or of the national reporting scheme operated by the TGA. This should be investigated further in future research.

The strengths of this study were that parents were selected from a random sample of the general South Australian population. Parents were asked to respond to clearly defined questions regarding perception of the likelihood of an AEFI occurring. The study results showed that direct parent reporting to a surveillance authority was low. Preference toward reporting to healthcare professionals was also noted in a recent Australian study of public awareness of consumer ADR reporting, in which 85% of all respondents who had previously experienced an ADR stated they reported to healthcare professionals and only 10% stating awareness of the Australian national reporting scheme.³⁵ The low levels of reporting to relevant surveillance authorities in both studies imply a need to introduce communication strategies to the public about channels of reporting.

Australian federal health authorities are currently investigating strategies for improved communication with consumers about vaccine safety monitoring and raising awareness of reporting systems. Future research and public awareness-raising campaigns should aim to address the provision of information about reporting schemes to people of non-Australian-born background. In particular, ensuring that relevant information was available at the time of immunization in a range of languages would help to meet the needs of the non-English speaking population. Finally, in order to better understand factors associated with direct reporting to a surveillance authority, rather than via a healthcare professional, further research with consumers who report direct to surveillance authorities is required.

Methods

Study population

The study was conducted as part of the Health Monitor program of the Population Research and Outcomes Studies Unit, (PROS) University of Adelaide.³⁶The Health Monitor survey is a commissioned computer aided telephone interview (CATI) survey of a randomly selected household community sample of approximately 2000 adults, conducted three times per year. It is used by government and non-government health professionals and policy makers for research and policy planning, with the main focus being population health surveillance. Each participating organization contributes toward the cost of their survey questions. The study was conducted as part of a multi-study survey (one of four separate studies) between March and May 2011. In addition to the present study’s questions, other immunization topic questions regarding public opinion on data linkage for vaccine safety surveillance and safety attitudes,³⁷ and other health-related topics related to physical activity and smoking were included. Households in both urban and rural areas were randomly selected from the South Australian (population 1.6 million) electronic white pages directory and an invitation letter was sent introducing the survey. Adults in each household who were aged 18 y or over, and who had the most recent birthday were selected for interview. There was no replacement for non-contactable persons. Each respondent was asked if he or she was a parent or legal guardian and, if so, further questions were asked about the vaccination experience of each child in his or her care, enabling classification of the respondent as having a child or children who had received a previous immunization. Further questions were asked to establish whether the children had ever experienced an AEFI. The scripted interview surveys were conducted by trained researchers on each day of the week and at different times of the day. Respondents with poor or no English were excluded from participating. No incentive was offered for participation. A pilot survey of 50 respondents was conducted to check question format and sequence.

The term “AEFI” was defined as “any untoward medical occurrence which follows immunization and which does not necessarily have a causal relationship with the useage of the vaccine.”³⁸ Hence an AEFI does not imply causality, but only a temporal relationship to an immunization that was judged to be vaccine related. We defined a parent “AEFI reporter” as a parent who indicated that they presented their child’s AEFI symptoms to a healthcare professional or reported their child’s AEFI to a surveillance authority. An “AEFI non-reporter” was defined as a parent who stated that they did not report their children’s symptoms to a healthcare professional or surveillance authority.

The survey included seven questions that sought to examine parents’ views of vaccine safety in general; pre-licensure safety testing of vaccines; acceptability of two vaccine safety risks, anaphylaxis, and febrile convulsion; and awareness of a system for vaccine safety surveillance. In order to examine vaccine safety risk perceptions, respondents were asked to recall their opinion of the likelihood of their children experiencing a mild AEFI (described as fever, irritability, and injection site swelling) and a serious AEFI (defined as requiring medical attention), at their youngest child’s last immunization. Ethical approval to conduct the study was granted by the University of Adelaide and South Australian Department of Health ethics committees.

Statistical analysis

We compared differences in parental safety opinion response proportions of parent AEFI reporters with the non-reporters, using the χ² test. Multivariate ordinal logistic regression analyses were conducted to determine: the socio-demographic predictors of reporting an AEFI to a health professional or surveillance authority; and to measure the association between reporting a child’s AEFI and parental views on vaccine safety. The analyses were adjusted for potential confounders among the socio-demographic variables collected. Preliminary checks confirmed the proportional odds assumption was not violated.³⁹ The demographic covariates included: parent age, gender, total children; education (secondary school, certificate/trade, and university), income, and country of birth (Australia or other). We used the Index of Relative Socioeconomic Disadvantage (IRSD) as a measure of socio-economic status.⁴⁰ Statistical tests were two-tailed, with a significance level of 5%. Odds Ratios and 95% confidence intervals were used to examine the strength and precision between sociodemographic predictors of reporting and the association with reporting and vaccine safety opinions. Statistical analyses were conducted using STATA 11.2 (StataCorp). All statistical analyses were performed using data weighted to the age, gender, probability of selection within a household and geographical area profile of the South Australian population. Individual data were weighted by the inverse of the individual's probability of selection and then reweighted by age, gender, and area, derived from the Australian Bureau of Statistics estimated resident population for 30 June, 2009.^36,41 All results presented are weighted for both numbers and proportions.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

H.M. acknowledges support of the National Health and Medical Research Council of Australia: Career Development Fellowship (1016272).

10.4161/hv.27459