Objective: To investigate whether false-positive outcomes on neonatal hearing screening cause long-lasting parental concerns.
Methods: A general population of parents whose children had participated in the universal neonatal hearing screening (UNHS) programme were examined. Parents filled out a questionnaire 6 months after UNHS. Outcomes were compared for all parents whose child tested positive or inconclusive in at least one of three tests but afterwards proved not to have hearing impairment (cases, n = 154) and a random sample of parents whose child passed the first test (controls, n = 288). Parental anxiety as measured with the State-Trait Anxiety Inventory (STAI), attitude towards the child (child health rating and experienced problems) and sensitivity to hearing problems were measured.
Results: Median STAI score was equal for cases and controls. Parental attitudes toward the child also did not differ. The difference in the proportion of parents who worried about their child’s hearing was statistically significant between cases and controls (p = 0.001) and varied with the number of screens; 4% of controls were worried about the child’s hearing, as compared to 10% of cases whose children were tested twice, and 15% of cases whose children were tested three times.
Conclusions: False-positive UNHS test results do not cause long-term general parental anxiety. However, 6 months after screening, a considerable proportion of parents continued to experience hearing-specific worries regarding their child.
Statistics from Altmetric.com
As early intervention can improve the cognitive and developmental outcomes of young infants with hearing loss,1–3 implementation of universal neonatal hearing screening (UNHS) is recommended in Europe4 and the USA.5 6 However, a major problem with UNHS is that false-positive results occur quite frequently7 and there are concerns that false-positive screen results at any stage may cause long-lasting parental concern, even when normal hearing has been confirmed.8 9 False-positive results in neonatal screening, for example for congenital hypothyroidism,10 phenylketonuria,11 Down syndrome12 or cystic fibrosis,13 are known to produce increased levels of parental concern for up to 12–24 months after the test. False-positive results on UNHS were also found to result in concerns immediately after the screen, with the concern increasing if the child failed two subsequent tests.14–17 It is not known how long parental concerns continue after false-positive UNHS and, if they do exist for a longer period, whether they influence parents’ attitudes towards their child. We investigated these issues in the present study.
This study is part of a larger project, the primary goal of which was to compare different methods of implementing UNHS as regards screen results, acceptance and costs.18 UNHS was performed in three pilot regions by trained nurses either at the infant’s home or at a well-baby clinic (WBC). Portable devices were used to record transient evoked otoacoustic emissions (TEOAE).18 The target age for the first screen was between day 4 and week 4 of life. Children who failed a screen (unilateral/bilateral refer or invalid result) were referred for a rescreen. After three failed screens, referral to audiology was indicated for further diagnostic evaluation. Children born between 1 November 1999 and 1 November 2000 were eligible for participation. Neonates placed in a neonatal intensive care unit were excluded as they have a separate screening programme.19 A total of 3137 children were screened. The study was approved by the local medical ethics committee.
Cases and controls
Cases were defined as all parents of children who failed the first screen but, according to the following screens or further diagnostic evaluation, later proved to have normal hearing (false-positive result). A sample of parents of children who passed the first screen bilaterally served as controls. To increase the power, we included 1.5 controls per case instead of 1 control per case. We therefore counted the number of cases per region each month and randomly sampled 1.5 times as many controls from the same region.
A questionnaire developed for this study was mailed to cases and controls 6 months after the first screen. To increase the participation rate, we included a stamped addressed envelope and sent a reminder to non-respondents. Respondents could also indicate whether they would like to receive a summary of the study results and could make comments or suggestions on hearing screening which were taken into account when implementing UNHS nationally.
The questionnaire asked parents about: (1) their anxiety state (ie, how does the respondent feel at that moment?), (2) their attitude towards their child and (3) their sensitivity to hearing problems.
The six-item version of the state scale of the Spielberger State-Trait Anxiety Inventory (STAI) was used to examine anxiety state.20 Each item has four possible answers (“not at all”, “somewhat”, “moderately” and “very much”). Answers to the items “I am tense”, “I feel upset” and “I am worried” were scored as 0 (“not at all”) to 3 (“very much”), and for the items “I feel calm”, “I am relaxed” and “I feel content” as 3 (“not at all”) to 0 (“very much”). For each parent, scores were combined to form a sum-STAI score. It therefore follows that sum-STAI scores can range from 0 (no anxiety) to 18 (high anxiety).
We hypothesised that concerns due to UNHS might negatively affect parents’ attitude towards their child and might cause them to consider them less healthy than other children. In the questionnaire, we therefore assessed parents’ perception of their child’s general health by asking them to rate it on a scale from 1 (not healthy) to 10 (very healthy). Parents might also experience more problems with feeding, sleeping, making contact and/or crying. Hence we also asked parents whether they experienced such problems (yes/no). The presence of other significant health problems, such as asthma or Down syndrome, was also recorded, as these may independently affect parental anxiety. The encountered health problems were independently rated as mild or severe by two paediatricians.
To investigate whether a false-positive screen resulted in a higher sensitivity to hearing problems, we asked how often the parents worried about their child’s hearing, and how often they tried to test the child’s hearing, for example by whispering or making small noises. We also asked how often they wondered if their baby could hear well. Possible answers these questions were “never”, “sometimes’, “often” or “very often”. The latter two categories were combined in the analyses due to small numbers. We also recorded whether, since the last screen, a general practitioner had been visited to check the child’s hearing (yes/no).
Finally, the questionnaire addressed the background variables family size, level of formal education of both parents, and the relationship of the respondents to the child screened.
Differences between groups were tested by non-parametric tests (χ2, Fisher’s exact, Mann-Whitney, Kruskal-Wallis). We used multivariate linear, ordinal and logistic regression to assess differences between cases and controls, and whether the number of screens affected the sensitivity of the parents to the child’s hearing and the STAI score. Potentially confounding variables were age at first screen (in days), the presence of a health problem, the location of screening (at home or a WBC), and the aforementioned background variables. In a backward selection procedure, these variables were eliminated if a p value of >0.20 was reached. p Values <0.05 were considered significant.
We sent questionnaires to 217 cases and 335 controls. A total of 173 (80%) questionnaires from cases and 288 (86%) from controls were eligible for analysis. Nineteen (11%) cases were excluded as their children appeared to have a conductive or sensorineural hearing loss (diagnosed at audiology, n = 14) or might have a hearing loss (referred to audiology but not diagnosed because the parents did not comply with the referral, n = 5). Analyses were thus based on 154 cases and 288 controls. Cases and controls did not differ with respect to region, education of the mother, number of children in the family or respondent. Age at first screen was higher for the children of cases than of controls (median (range) 16 (0–55) days for controls and 19 (1–103) days for cases, p = 0.048). Mothers filled out 84% of the questionnaires.
Sum-STAI scores ranged from 0 to 15 for controls, from 0 to 14 for cases whose children underwent two screens, and from 0 to 13 for cases whose children underwent three screens (table 1). For all three groups, the median score was 3. Differences between groups were not statistically significant (p = 0.94); multivariate linear regression analysis also showed no relationship.
Of cases, 79% reported no health problem or illness, 14% a mild health problem and 7% a severe health problem. For controls, these percentages were 82%, 15% and 3%, respectively (no significant difference, p = 0.26). For both cases and controls, the median health score was 9 (range 1–10, p = 0.27; not significant).
Amongst cases, 7% experienced feeding problems, 6% indicated sleeping problems, 1% had problems making contact and 3% had problems with crying behaviour. For controls these percentages were 3%, 4%, 1% and 2%, respectively. Differences were not significant (p = 0.06, 0.23, 1.00 and 0.75, respectively) and remained so after adjustment for confounders.
Sensitivity to possible hearing problems
Table 2 shows the percentages of parents who often/very often tested their child’s hearing. This differed significantly for cases and controls (p = 0.003) and with the number of screens (p = 0.01). However, significance could not be reached in a multivariate ordinal regression analysis to correct for confounding variables. We also found no statistically significant differences between cases and controls with respect to frequency of visits to a general practitioner or the frequency of wondering about the child’s hearing.
The proportion of parents who worried about their child’s hearing differed significantly for cases and controls (p = 0.001) and varied significantly with the number of screens (p = 0.001); 4% of controls worried sometimes, often or very often about their child’s hearing, while 10% of cases whose children had two screens did so, and 15% of cases whose children had three screens. Adjustment for confounders did not change the results; odds ratios (95% CI) were 2.6 (1.0 to 6.6) for two screens and 4.1 (1.6 to 10.5) for three screens, where one screen served as a reference. The final regression model, in addition to number of screens, only contained age at first screen. Other variables did not contribute significantly.
A considerable proportion of parents continued to feel worried about their child’s hearing 6 months after a false-positive UNHS test result. There was no relationship between the parents’ attitude towards their child and false-positive UNHS test results: parents of cases did not experience their child as less healthy than other children and did not experience more problems with the child.
Cases did not have a higher level of anxiety as measured with the STAI than controls 6 months after the first screen. However, the STAI might not be sensitive enough to identify the hypothesised raised levels of anxiety. To examine this, we investigated whether there is a relationship between the seriousness of a child’s illness and parental anxiety as measured with the STAI. A statistically significant, positive relationship was found, with a median STAI score of 3 for children without illness, 3.5 for children with mild illness and 4.5 for children with severe illness (p = 0.006). We concluded that the STAI is sensitive enough, and that a false-positive screen result does not have long-lasting effect on parental anxiety.
Our findings are confirmed by other smaller studies. Watkin et al21 used the STAI to compare the anxiety state of a retest group of 57 patients with false-positive results with a control group. They showed a significant difference 6 weeks after the screen. Poulakis and co-workers22 also found that 18% of parents whose children were referred for further diagnostic evaluation continued to feel “a little” worried 6 months after being informed that there was no permanent hearing loss. Clemens et al23 found that 5 months after UNHS 14% of parents whose child had passed a rescreen, said that this had caused them some or much anxiety. For comparison, we found that 10–15% of cases worried about their child’s hearing 6 months after the screen, depending on the number of tests.
Our questionnaire was sent out 6 months after screening, by which time parents knew definitely that their child did not have a hearing problem. Therefore, anxiety caused by the screen may have faded. Such an effect was found by Magnuson and Hergils,24 who performed in-depth interviews 8–12 months after screening with 49 parents of 26 children, half of whom had passed the first hearing screen and the other half of whom had undergone a retest before passing. They found that anxiety was evoked when the child did not pass the first retest; this was a critical point. When parents finally were given a definite answer as to whether or not the child had a hearing impairment, the anxiety seemed to fade.
As the response to our questionnaire was high (80% for cases and 86% for controls), selection bias is unlikely to have had a large impact on the results. We cannot exclude that the questionnaire might have re-ignited concerns in cases or controls, and that this might happen more often with parents whose children had a false-positive screen than with controls. The finding that parents whose children had at least one false-positive screen were more likely to worry about their child’s hearing over the next 6 months, when compared to controls, could be partly explained by this effect. The other outcome measures showed no differences between cases and controls. Thus, the impact of false-positive screen results on parents might be smaller due to this effect.
From our results it follows that the number of false-positives in UNHS should be kept to a minimum. In our study we found that age at first screen was higher for cases than for controls. In another study we showed that referral rate is lowest between days 5 and 13 after birth.25 Before and after this period the false-positive rate is higher. Therefore, children tested at a later age are more likely to have a false-positive result. Possibly, at a later age children are more likely to be awake and more likely to have middle-ear problems. In the Netherlands, we therefore encourage initial screening on days 4–7 after birth. Also, it might be considered whether the A-ABR method should be included in the programme as it has higher specificity as compared to the OAE method. In the Netherlands, we have introduced the A-ABR for the third screen. First results are promising as the percentage of referrals now is less than 0.5% compared to 2.2% before introduction of the A-ABR.
Nevertheless, false-positive test results cannot be ruled out completely. To minimise hearing-specific worries in parents, personnel should be aware of the possibility of parental worries when informing them about the screening procedure. Also, in the Netherlands, we found that timing between screens might increase concerns in parents; rescreening within 1–2 days led to suspicion among parents that something serious was wrong with their child, whereas a long delay between screens extends the period during which the parents feel uncertain about the child’s hearing. Therefore, we encouraged rescreening between 4 and 7 days after the previous test.
In conclusion, false-positive UNHS test results do not cause long-term general parental anxiety. However, 10–15% of parents of children who have a false-positive test continue to feel worried about their child’s hearing 6 months after screening. UNHS should be implemented in such a way that the number of false-positives is kept to a minimum.
What is already known on this topic
Early intervention can improve the cognitive and developmental outcomes of young infants with hearing loss.
To prevent delays in detecting hearing impairment in newborns, implementation of universal neonatal hearing screening (UNHS) has been recommended in Europe and the USA.
There are concerns that false-positive screen results at any stage may cause long-lasting parental concern, even when normal hearing has been confirmed.
Large long-term follow-up studies on parental concern following UNHS are lacking.
What this study adds
In a relatively large study of 154 cases and 288 controls, we found that 6 months after UNHS a false-positive screen result does not have a lasting effect on parental general anxiety or perceived general child health.
Worries regarding the child’s hearing remained among 10–15% of parents of children who had false-positive tests.
UNHS should be implemented in such a way that the number of false-positive screen results is kept to a minimum to prevent unnecessary long-lasting concerns.
We thank Jeanet Bruil for her advice concerning the development of the questionnaire.
Funding: This study was funded by the Netherlands Organisation for Health Research and Development (ZonMW, grant 28-2549-1).
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.