OBJECTIVE. The purpose of this work was to describe levels of maternal anxiety, depressive symptoms, and perceptions of infant vulnerability associated with newborn genetic screening for susceptibility to type 1 diabetes.
PATIENTS AND METHODS. Mothers of infants tested at birth for genetic susceptibility to type 1 diabetes as part of a prospective study investigating potential environmental triggers of autoimmunity were recruited to this study. Three mother-infant cohorts were studied: 38 infants at increased genetic risk, 73 at low risk, and 76 who had not undergone testing. The Vulnerable Baby Scale, Edinburgh Postnatal Depression Scale, and state subscale of the State Trait Anxiety Inventory were administered at the 9-week, 4-month, and 1-year postnatal ages. Genetic-risk notification occurred at the 10-week postnatal age. Mothers whose infants had undergone genetic testing were also asked to subjectively rate how much they thought and worried about their child's genetic test result. Statistical analyses were conducted to test for differences in questionnaire scores among the 3 groups.
RESULTS. No difference among the groups was detected in Vulnerable Baby Scale or Edinburgh Postnatal Depression Scale scores using linear mixed-effects model analysis. Maternal anxiety was paradoxically slightly lower in the increased-risk group shortly after notification of results, but there were no significant differences among the groups by 1 year. Mothers of infants in the high-risk group reported thinking and worrying about their child's test result significantly more than mothers of low-risk infants at both time points after notification of results.
CONCLUSIONS. Newborn genetic screening to identify infants at risk for type 1 diabetes is not associated with elevated levels of maternal anxiety, depressive symptoms, or heightened perceptions of infant vulnerability. However, responses to subjective assessment questions suggest that it is possible that more subtle effects on mothers do occur, and this requires further investigation.
There is speculation that future newborn screening programs may include testing for genetic variants (polymorphisms) that confer susceptibility to multifactorial diseases, such as type 1 diabetes.1,2 At present, population screening of newborns focuses on the presymptomatic identification of individuals who have highly penetrant Mendelian conditions, such as phenylketonuria.3 Although not entirely uncontroversial,4 screening for such disorders has become a standard component of preventive public health services in the developed world.5 In contrast, screening for the genetic markers associated with multifactorial conditions entails much more uncertainty in that it simply predicts whether an individual is more or less susceptible to developing a disease at some point in the future.6
Most current research programs concentrate on individual diseases, such as type 1 diabetes, but “genetic profiling,” whereby genetic susceptibility to a range of diseases can be determined, is also a possibility.7 It is widely speculated that this type of genetic testing will become commercially viable within the next 20 years.1 The underlying hope is that early identification of people at risk of a specific disorder will lead to reduced morbidity and mortality through targeted surveillance and preventative measures.8 However, though the ultimate goal of preventing the onset of many common diseases is compelling, concerns have been raised regarding the potential for adverse psychological reactions to the testing process.
Predictive genetic testing in childhood remains controversial. In general, most official policies, including those of the American Society of Human Genetics and the American Academy of Pediatrics, strongly advise against testing children for a disease in which preemptive or definitive medical treatment is not available in childhood.9–12 These policies refer explicitly to single-gene disorders but are also applicable to testing for genetic susceptibility to multifactorial diseases, for which no specific guidelines exist. The debate centers on the potential imbalance between risks and benefits for the child. Where there is a lack of evidence of clinical benefit from early identification, greater weight may be given to concerns over the potential psychological impact of genetic-risk predictions on parents and the parent-child relationship.13 For example, some contend that parents may experience prolonged distress on receiving “positive” results, the parent-child bond may be affected, and a “vulnerable-child syndrome” may eventuate.13–15 This latter descriptor is given to situations where illness during infancy (eg, prematurity) seems to be linked to a persistent sense of vulnerability and protectiveness on the part of the parent despite the child's objective good health.16–18 It has been suggested that parents may experience adverse newborn screening results as significant threats to their child's well being, which may contribute to the development of such a syndrome.19
Despite these concerns, there is currently a paucity of data concerning the psychological effects of testing young children for genetic susceptibility to multifactorial disorders. However, type 1 diabetes represents a useful disease model, both for studying the complex genetic and environmental interactions underlying multifactorial disease development and for investigating psychological reaction to newborn testing. It is one of the most common chronic childhood diseases, with a rising incidence (3%–4% per year in most developed countries), particularly in the 0- to 4-year age group.20,21 At present, development of the disease necessitates lifelong adherence to a difficult therapeutic regime that is only partially effective in preventing acute and chronic complications.22 These facts, as well as the existence of a long prediabetic phase, make the possibility of early identification and interruption of the disease process an attractive and potentially achievable goal.23 The condition has a multifactorial etiology with disease development dependent on an undetermined number of genetic and environmental factors.24 The chief genetic determinant of susceptibility to diabetes lies within the class II region of the major histocompatibility complex on chromosome 6.25 More than 90% of patients who develop type 1 diabetes have either the DR3 and/or the DR4 allele of the HLA-DRB1 gene, whereas <40% of healthy control subjects have these alleles. Depending on the population, people homozygous for the high-risk DR4 allele have a 10-to 15-fold increased risk of type 1 diabetes, and people heterozygous for the DR3/DR4 alleles have a 20- to 30-fold increased risk.26 Testing for these high-risk alleles by newborn screening in the general population is currently the object of several longitudinal prospective studies27–30 and, more recently, a larger prospective study involving a consortium of 6 centers in the United States, Scandinavia, and Europe.31 The explicit aim of these studies is to further elucidate the natural history of type 1 diabetes, to identify environmental exposures that may trigger autoimmunity, and ultimately to test interventions that may prevent the disease.
The purpose of the study reported here was to evaluate maternal psychological reaction to receipt of genetic information concerning the newborn infant's risk of type 1 diabetes during a prospective study. We hypothesized that receiving a result stating that the infant was at increased genetic risk for type 1 diabetes would cause adverse psychological reaction compared with receiving a lower genetic-risk result. This hypothesis reflects many of the premises underlying guidelines from professional societies. We report here the prospective psychological evaluation of a cohort of mother and infant pairs subject to newborn genetic screening for susceptibility to type 1 diabetes.
Participants for this project were recruited during the Dunedin (New Zealand)–based Key Environmental Aspects of Type 1 Diabetes (KEA) longitudinal study investigating the natural history of type 1 diabetes.
KEA Longitudinal Study Investigating Potential Environmental Triggers in Children Genetically Susceptible to Type 1 Diabetes
Identification of Newborns at Increased Genetic Risk of Type 1 Diabetes Recruitment
Pregnant women were enrolled at 28 weeks' gestation via obstetricians or midwives at 2 large maternity centers in Dunedin, an urban center of population 120000. Informed consent was obtained from all of the participants, and ethical approval was granted by the Otago Ethics Committee. Written information was provided for all of the prospective participants explaining that the study aimed to “screen children at birth and identify the 10% of children at increased genetic risk for developing diabetes.” Increased risk was defined as “a 1-in-16 chance of developing diabetes before 20 years old.” Eighty-six percent of all parents approached gave consent for a cord blood sample to be taken for genetic analysis of type 1 diabetes susceptibility alleles (phase 1 of the KEA Study). Only infants with severe congenital anomalies or who were stillborn were excluded.
Cord blood was obtained at delivery, and genomic DNA was prepared from 1 to 5 mL of blood using a guanidine hydrochloride-based method. Allele-specific polymerase chain reaction was used to type HLA-DRB1.32 Samples were first genotyped for HLA-DRB1 DR1-4 and DR6-12. Individuals positive for HLA-DRB1*04 were subtyped to identify those having the *0401 and *0404 alleles. Infants were assigned to either an increased-genetic-risk group (HLADR3/DR*0401, DR3/DR*0404, DR*0401/*0401, DR*0401/DR*0404, and DR*0404/DR*0404) or a lower-genetic-risk group (all other HLA-DRB1 alleles).
Communication of Results
The dissemination of results to parents and the support offered to them over this period was intended to simulate how a population-based newborn screening program might operate.
Parents of Infants at Increased Genetic Risk
These parents were informed of their infant's results by letter. The letter opened with both of the parents' first names and then reminded them of the collection of the cord blood sample as part of the KEA Study. It went on to state that their infant had been found to be at “increased risk of developing diabetes by the age of 20 years, with a 1 in 16 risk of developing diabetes compared with the general population risk of 1 in 300.”33 It was also stated that “not all genetically high risk infants go on to develop diabetes because development of diabetes also depends on other poorly understood factors, some of which are found in the environment.” The parents were advised that they would be contacted by telephone a few days later and offered a face-to-face counseling meeting with a pediatrician. All of the parents accepted this offer. At this meeting, the above risk statistics were reiterated, there was general discussion about type 1 diabetes in childhood, and they were invited to continue with phase 2 of the study.
Phase 2 of the KEA Study
This involved 3- to 6-month assessments of autoantibody titers and assessment of potential environmental triggers, such as diet, household cleanliness, and evidence of enteroviral infection.
Parents of Infants in the Low-Genetic-Risk Group
These parents were also informed by letter. The letter opened in the same way as that addressed to parents of infants at increased risk but went on to say that their infant was “not at high genetic risk of diabetes, but this did not imply or guarantee no risk.” This was followed by a brief description of the presenting symptoms of type 1 diabetes and advice to contact their general practitioner if they had any concerns about their child's health in the future. A telephone contact was offered if parents wished to discuss the result further. These children were not invited to participate in phase 2 of the KEA Study.
Parents of children in the low-genetic-risk group who were affected with type 1 diabetes themselves (n = 2 fathers and 1 mother) were sent a modified letter stating that “although the HLA DRB1 gene test was negative their child remains at increased risk [∼1 of 50 quoted if the mother was affected and 1 of 15 if the father was affected] of diabetes because of their family history. This is because of sharing of other genes and unknown environmental factors.”34–38 These children were also not invited to participate in phase 2 of the KEA Study. Although the KEA Study has now been discontinued, children at increased genetic risk continue to be followed with yearly autoantibody titers and will be offered entry into TrialNet39 if positive results occur.
Psychological Effects Arm of the Study
Three mother-infant cohorts were recruited to the psychological effects arm of the study. Infants at increased genetic risk of type 1 diabetes and age-matched infants with low genetic risk were recruited from the KEA Study. Age-matched control infants, who had not undergone any genetic testing, were recruited from the postnatal ward at Dunedin Public Hospital. Informed consent was obtained from all of the participants, and ethical approval was granted by the Otago Ethics Committee.
The study used a longitudinal design with the baseline assessment of maternal psychological reaction at 9 weeks postpartum (baseline, questionnaire 1), notification of newborn genetic screening result for type 1 diabetes susceptibility at 10 weeks, and follow-up assessment of maternal reaction at 16 weeks (questionnaire 2) and 54 weeks (questionnaire 3). Questionnaires were mailed to participants or handed directly to mothers of increased-risk infants if they were attending a KEA Study appointment at an appropriate time. Nonresponse prompted a follow-up telephone call after 2 weeks.
These consisted of 5 sections, as described below.
The Vulnerable Baby Scale
The Vulnerable Baby Scale (VBS) is a 10-item self-report scale used to assess maternal perceptions of their infant's vulnerability.40 It was developed specifically for use in this study to address the important hypothesis that parents may experience adverse newborn screening results as significant threats to their child's well-being, perhaps leading to the vulnerable-child syndrome.19 The VBS was based on an existing Child Vulnerability Scale suitable for use with mothers of 4- to 8-year-old children,41 with modifications to make it developmentally appropriate for use with mothers of young infants. The resulting VBS (Appendix 1) has been validated in such a setting.40 The measure demonstrated good internal consistency (first assessment, α = .72; second assessment, α = .70; third assessment, α =.73). A mean score of 23.1 and SD of 3.1 has been reported for mothers of a group of healthy 10-week-old infants, with a score of ≥27 indicative of increased parental perceptions of vulnerability.40
The Edinburgh Postnatal Depression Scale
The Edinburgh Postnatal Depression Scale (EPDS) is a 10-item self-report scale used to screen for postnatal depression. Its use is well validated both in the postnatal period42,43 and in cohorts of nonpostnatal women with older children.44 The measure demonstrated excellent internal consistency (first assessment, α = .82; second assessment, α = .87; third assessment, α = .85). When the EPDS is used as a postnatal depression screening tool, a cutoff score of >12 is usually used, though other authors use lower cutoff scores to identify those with more minor depressive symptoms, and a difference in score of ≥3 at different time points has been considered clinically significant even below the threshold of 12.45
The State Trait Anxiety Inventory State Subscale
The state component (Form Y-1) of the State Trait Anxiety Inventory (STAI) is a 20-item self-report form that is a reliable and well-validated measure of acute anxiety.46 The standard instructions accompanying the form were used in this study. The measure demonstrated excellent internal consistency (first assessment, α = .90; second assessment, α =.92; third assessment, α = .90). Mean scores on the State Anxiety Scale vary with age and gender. A study of postpartum women reported mean scores of 30.4 at 14 weeks and 31.2 at 30 weeks.47 Permission to use the questionnaire was obtained from Mind Garden Inc.
Subjective Rating of Level of Concern About Type 1 Diabetes Genetic Susceptibility Test Results
This section was administered only to mothers whose infants had undergone genetic testing and incorporated 2 direct questions concerning how much the mother thinks (4.1) and worries (4.2) about her child's genetic test result. Responses were recorded on a 5-point scale (Appendix 2).
Maternal Perception of Infant's Risk of Developing Type 1 Diabetes
This section was administered at 16 weeks and 1 year only to mothers whose infants had undergone genetic testing. Mothers were asked to record how they viewed their infant's risk of developing type 1 diabetes on a simple 5-point rating scale, ranging from “no risk at all” to “certain to develop diabetes” (Appendix 3).
Possible influences on maternal reaction to genetic-risk information that were considered included maternal factors such as age, parity, education status, and time from notification of results to questionnaires. Paternal factors included education status, and infant factors included gestation, birth weight, mode of delivery, NICU admission, and family history of diabetes. Categorical variables were created for several of these covariates as follows: maternal age, using tertiles; parity class categorized as 1, 2, or ≥3; birth weight class categorized as low birth weight if <2500 g, otherwise normal weight; and gestation classed as premature if <37 weeks, otherwise normal.
Sample sizes to permit rejection of the null hypothesis were calculated using Stata 848 and normative data available for each of the 3 questionnaires. The STAI required the largest sample size (33 mothers of infants at increased genetic risk and 66 mothers of control infants and infants at low risk) for 80% power at a .05 level of significance to detect a clinically significant difference in score (defined as >5 points on the STAI, >2 on the EPDS, and >2 on the VBS).
To account for the correlated measurements from each mother, a linear mixed model was constructed for each variable, initially including the possible covariates listed above, and modeling the covariances between the time periods using the most appropriate structure as determined using Akaike and Bayesian information criteria (where autoregressive, Toeplitz, compound-symmetric, and unstructured covariance matrices were considered). The denominator degrees of freedom were calculated using Sattherwaite's approximation.
During the study period, 41 infants at increased genetic risk of diabetes were identified (8.2% of all of the infants tested). Of this group, 38 mothers (93%) agreed to participate in the study evaluating maternal psychological reactions to newborn genetic screening for type 1 diabetes. In addition, 76 mothers of age-matched infants whose cord blood analysis revealed low genetic risk of type 1 diabetes were recruited (97% participation rate). Three of these infants had a parent with type 1 diabetes, putting them at increased risk, despite the absence of specific HLA-DRB1 markers. Initially, these mothers and infants were included in the low-risk group but were excluded from the final analysis, because mothers were aware of their child's elevated risk status in relation to family history. Seventy-six mothers of age-matched control infants who had not undergone any genetic testing were also recruited (92% participation rate). Demographic characteristics of participants are shown in Table 1.
Retention of Participants and Timing of Questionnaires
Participation rates at the 3 assessments during the study period were: 95%, 100%, and 100% for the increased-risk, low-risk, and control groups, respectively, for questionnaire 1; 100%, 96%, and 99% for questionnaire 2; and 100%, 96%, and 86% for questionnaire 3. Questionnaires were administered during similar time periods for each group at each of the 3 assessments. (Means of 10.2, 8.6, and 10.2 weeks for the increased-risk, low-risk, and control groups, respectively, for questionnaire 1, P for difference of .01; means of 16.0, 15.8, and 16.5 weeks for questionnaire 2, P = .7; and means of 53.9, 53.8, and 54.8 weeks for questionnaire 3, P = .3.)
An autoregressive error covariance was used to account for correlation among questionnaire scores for individuals over time. There was no significant difference in scores among the groups (P = .1). A time effect (gradual decrease in VBS score as infant's age increased) was observed (P < .0001) with no significant difference in this time effect among the 3 groups (P = .47; Fig 1).
A compound symmetry structure was used to account for correlation among questionnaire scores for individuals over time. There was no significant difference in scores among the groups (P = .61). A time effect (gradual decrease in EPDS score as infant's age increased) was observed (P < .0001). The increased-risk group seemed to have a larger initial drop in EPDS score and subsequent small rebound effect. This group-time interaction was not statistically significant (P = .30; Fig 2).
Eight percent of mothers had scores greater than the clinical cutoff12 at the first assessment, 7% at the second, and 3% at the third. This is consistent with a previous study in New Zealand,45 though slightly lower than the average prevalence of 13% recorded in a meta-analysis.49
The STAI (State Subscale) Scores
A compound symmetry structure was used to account for correlation among questionnaire scores for individuals over time. There was no significant difference in scores among the groups (P = .6). A time effect was again observed (gradual decrease in STAI score as infant's age increased, P = .0009). Again, the increased risk group displayed a larger initial decrease in score followed by a small rebound. This group-time interaction was found to be weakly significant (P = .04), though posthoc tests were unable to identify a statistically significant difference between the groups at any specific time point (Fig 3). Questionnaire scores (SD and 95% confidence interval [CI]) for the 3 groups at each time period are displayed in Table 2.
The only covariate that was statistically significant was gestation category in the VBS model (P = .0015). (The effect of being in the gestational category “premature” [<37 weeks] was to raise the VBS by 2.5 [95% CI: 1.0–4.0] over scores for mothers with infants in the term category.) Infants in the control group were more likely to be admitted to NICU (Table 1), but this variable was not close to significance in any model (P > .5 in all cases). Including or omitting this variable from the models made no substantial difference to the parameter estimates.
Subjective Rating of Level of Concern About Type 1 Diabetes Genetic Susceptibility Test Results
“How Much Do You Think About Your Child's Test Result?”
A group effect was observed (P < .0001) with the increased-genetic-risk group thinking about their child's test result more than the mothers of infants in the low-risk group. There was a group-time interaction (P < .0001), with the high-risk group thinking about their child's test result more at both time periods after notification of results (Table 3 and Fig 4).
“How Much Do You Worry About Your Child's Test Result?”
A group effect was observed (P < .0001) with the increased-genetic-risk group worrying about their child's test result more than the mothers of infants in the low-risk group. There was a group-time interaction (P < .0001), with the high-risk group worrying about their child's test result more at both time periods after notification of results (Table 4 and Fig 5).
Maternal Perception of Infant's Risk of Developing Type 1 Diabetes
At both 16 weeks and 1 year, 92% of mothers of infants in the low-risk group correctly defined their infant's risk of developing type 1 diabetes as the “same” or “less than most other people.” The remaining 8% were uncertain or incorrectly considered their infant to be at “no risk at all” of developing type 1 diabetes. At 16 weeks, 87% of mothers of infants at increased genetic risk of type 1 diabetes correctly defined their infant to be at “higher risk than most people” for developing type 1 diabetes, and at 1 year this figure rose to 92%. The remaining mothers thought their infant would “definitely develop type 1 diabetes,” was at the “same risk as most people,” or were uncertain (Table 5).
There is increasing speculation that genomic profiling will eventually become part of routine medical practice7 and perhaps of newborn screening.1 Testing newborns for genetic susceptibility to type 1 diabetes already occurs as part of several large prospective studies,27–31 and it is likely that other complex disorders, such as asthma and obesity, will also be studied in this manner. The ultimate goal of preventing, or even delaying, the onset of some of these common diseases is compelling, but concerns have been raised regarding the potential for adverse psychological reactions to such testing. However, there has been little empirical data on which to base further discussion. We undertook the study reported here to provide empirical evidence of maternal psychological reaction to newborn screening for genetic susceptibility to type 1 diabetes.
In this sample, mothers who were informed that their infant was at increased genetic risk for type 1 diabetes did not experience adverse psychological effects compared with control mothers when assessed on 3 different psychological rating scales. (Figs 1⇑–3 and Table 2.) At the final assessment, when infants were 1 year of age, scores on all 3 of the questionnaires were very similar for each of the 3 groups.
Measures of anxiety and depression are commonly used to assess the impact of genetic-risk assessment, screening, or testing in the adult population50–54 and in the newborn period.55–59 The measures used in this study (STAI and EPDS) are well validated, including for women in the postnatal period.42–44,46,60 Results from this study were very similar for both measures (Figs 2 and 3) with no difference between the groups at baseline or 1 year. Three previous reports documenting levels of maternal anxiety, stress, or depression in relation to genetic screening for type 1 diabetes support these findings.61–63 However, shortly after receiving the genetic-risk information, the increased-genetic-risk group exhibited lower questionnaire scores than the other 2 groups (only the STAI results reached a level of borderline statistical significance). This initial dip in anxiety levels exhibited by the increased-genetic-risk group is somewhat counterintuitive, though it could be explained by the reassurance of the face-to-face counseling visit and regular assessments involved in the KEA study that were specific to this group.
An important additional component of our study was to test the hypothesis that knowledge of an increased genetic risk of type 1 diabetes may affect maternal perceptions of their child's vulnerability and, hence, contribute to a vulnerable-child syndrome. This has not previously been empirically tested, though “parental stress” has been measured in newborn screening research63–65 using the Parenting Stress Index,66 which includes questions concerning parental perceptions of their child. However, the appropriateness of the Parenting Stress Index in research with mothers of very young infants has been questioned,56,67 because when used at this time, there are frequently high numbers of missing item responses that potentially invalidate results.63,65
Using the developmentally appropriate VBS, we found that there was no discernible difference in VBS score among the 3 groups at any of the assessments (Fig 1). Mean scores at the second assessment were similar to the mean of a previously reported healthy control group of similar age,40 and, as expected, scores gradually decreased over time. These findings suggest that mothers cope well with the genetic information and do not seem to begin perceiving their child as ill on the basis of the genetic screening test, as has been proposed previously.13,68
Subjective Levels of Concern
Despite this reassuring data, when asked to rate their own degree of concern about their infant's genetic risk of type 1 diabetes, mothers of infants at increased genetic risk reported significantly higher levels than mothers of infants with low genetic risk (Tables 3 and 4 and Figs 4 and 5). In many ways, this is not a surprising result: it seems unrealistic to think that there will be no difference in psychological reaction between mothers of genetically susceptible or low-risk infants. In fact, one of the aims of any newborn screening program involving genetic susceptibility testing must be to create some degree of heightened awareness of their child's health risks among parents so that that they participate in surveillance and/or preventative measures. In the case of type 1 diabetes, though no preventative measure is currently available, early recognition of disease leads to a milder initial clinical course,69 and parents play an important role in this process. Although it would clearly be wrong to overburden parents and create problems akin to the vulnerable-child syndrome, some degree of parental concern is necessary and may be justified. Determining precisely what an appropriate or acceptable parental response constitutes, and how best to achieve this in practice, are challenges for the future. Interestingly, in our study, despite the fact that mothers reported heightened levels of concern regarding their child's genetic test result, this did not seem to translate into an increased frequency of visits to a health professional. This information was derived from question 10 of the VBS that asks specifically about such consultations: mean scores for this question in isolation were very similar for each of the 3 groups (1.3, 1.3, and 1.4 at baseline; 1.3, 1.3, and 1.3 at 16 weeks; and 1.4, 1.5, and 1.4 at 1 year for the increased-risk, low-risk, and control groups, respectively).
The psychological questionnaires used in this study undoubtedly represent a useful way of assessing reactions to genetic-risk information and allow us to state that clinically significant levels of anxiety and depression are unlikely to occur as a result of newborn genetic susceptibility testing for type 1 diabetes. However, the responses to the subjective assessment questions suggest that mothers of children at increased genetic risk of type 1 diabetes do experience some sort of psychosocial reaction to their child's testing. The implications of these subtle effects remain unclear but merit further investigation before population-wide screening is considered.
Perception of Genetic Risk
Although it is becoming increasingly possible to accurately determine people's risk of developing conditions such as type 1 diabetes, considerable uncertainty remains over how this risk information will be evaluated.70 People often have difficulty in processing statistical information, and interpretation of results may depend on a range of factors, including personal and cultural preferences and ethical concerns.70 Our data (Table 5) suggest that the majority of mothers in this study evaluated their infant's risk of developing type 1 diabetes correctly. However, the risk categories that we used to test mothers' risk perception were broad, so that within each category, there would likely be a range of maternal understanding. In addition, despite these broad categories, some mothers in this study did underestimate or overestimate their child's risk: in the increased risk group, 3 mothers thought their infant was at the same risk as most people for developing type 1 diabetes (underestimating), and 1 mother overestimated thinking that her infant would definitely develop diabetes. In the low-risk group, an appropriate risk assessment was considered to be that the infant was at the same or less risk than most people of developing type 1 diabetes, but several mothers (5 at 16 weeks and 6 at 1 year) considered their infant to be at no risk of type 1 diabetes, despite having been informed by letter that their infant was “not at high genetic risk of diabetes, but this did not imply or guarantee no risk.” This “false reassurance” could be harmful if parents ignored symptoms of type 1 diabetes, considering it impossible for their child to develop the condition. Because the numbers of mothers misinterpreting genetic-risk information in this study were very small, no assessment of the relationship between perceived risk and scores on psychological assessment measures was possible, but we consider further exploration of these issues to be important. Consideration should be given not only to how best to ensure accurate understanding of genetic susceptibility test results, but also to how parents may then incorporate this information into their daily lives and use it to make health decisions concerning their children.
The limitations of this study include the relatively small size of the cohorts, though both participation and retention rates were excellent. In addition, our study design precluded us from detecting psychological distress occurring immediately after genetic-risk notification, and it is also possible that despite the scores on the STAI and EPDS being similar at this study end point (1 year), they may continue to rise in the increased-genetic-risk group. Our study concentrated on maternal reaction to genetic-risk information without including the views or responses of fathers. (Fathers' reactions to genetic testing are likely to be important, but in this study mothers were primary caregivers of their young infants, making several aspects of the questionnaires used most relevant to them.) Our study population was predominantly of European origin (5.7% were Maori, which is representative of this region of New Zealand), and, hence, analysis by ethnicity was not undertaken. We do not consider that New Zealanders differ markedly in their attitudes to genetics from other populations, but it is well known that cultural differences may impact on health beliefs, which may, in turn, influence many aspects of the genetic testing process.71
Efforts were made to simulate the type of informative process that would occur if newborn screening routinely included this type of test, in that participants were initially informed of their child's results by letter, followed by a counseling visit with a pediatrician. However, parents were then offered the opportunity to participate in a trial involving monitoring of autoantibodies and aspects of the child's “environment” (but without an intervention arm). Autoantibodies were measured on a 3- to 6-month basis, and no child had a positive test result during the course of this study. It is unlikely that low maternal scores on psychological questionnaires were related to reassurance from negative autoantibody results, because these were not available to parents until several weeks after their 3- to 6-month visits, whereas psychological questionnaires coincided with the visits themselves. However, it is possible that contact with the research team and opportunity to ask questions alleviated anxiety, though it is also possible to conceive that bringing one's young infant for a blood test could have the opposite effect. Immunologic follow-up would be necessary in any type 1 diabetes risk testing program, which may differentiate genetic susceptibility testing for type 1 diabetes from similar testing for other multifactorial disorders. Further research will be necessary to determine how parents cope if follow-up is less intensive, and it may be that even if monitoring is unnecessary, clinically it continues to be so at a psychological level.
Balancing Benefits and Risks
Genetic susceptibility testing in the newborn period has been controversial largely because of the potential imbalance between benefits and risks for the child.13 At present, relatively good predictions of type 1 diabetes can be obtained by genotyping and measuring autoantibodies, but no preventative measure is available. However, children at increased genetic risk who were enrolled in longitudinal studies already show improved health status at onset of type 1 diabetes compared with children who were not involved in these studies.22,69 This is presumably related to medical surveillance and the heightened awareness of parents. Further research in this area aims to reveal more details of the pathogenesis of the disease and eventually to lead to some form of prevention. At this level of benefit, screening entire populations of newborns for genetic susceptibility to type 1 diabetes would become significantly more attractive.
At the same time, evidence, including the results of this study, is gradually accruing to suggest that families generally cope well with type 1 diabetes genetic-risk information concerning their children, if it is conveyed sensitively.61–63 Currently, the research remains fairly limited both in focus and duration, with the longer-term impact of more subtle alterations in parental perception of their child's health status and the potential effects this may have on parenting style still to be evaluated.
There is much activity underway to develop a consensus on what types of newborn screening are appropriate,20 and it is clearly important that those involved in this process research and reflect on all of the likely harms, benefits, and ethical dilemmas associated with expanded newborn screening. Studies such as that described here will aid in such a decision-making process but do not provide all of the information required. Further studies must be sufficiently sophisticated so that they can capture and evaluate the range of psychosocial responses that may arise from conveying relatively complex genetic information to parents of young children and assess how parents manage this information within the context of family life.
A Masonic Child Health Fellowship awarded to Dr Kerruish supported this study. The KEA study was supported by an Otago University Research Grant and grants from the Otago Medical Research Foundation and Novo Nordisk.
We are grateful to Shirley Jones and Debra McNamara for assistance with recruitment and data collection, Marilyn Merriman for HLA-DRB1 genotyping, and Adell Cox, clinical psychologist, for assistance with psychological questionnaires.
- Accepted January 24, 2007.
- Address correspondence to Nicola J. Kerruish, MRCP, FRACP, Department of Women's and Children's Health, Otago Medical School, University of Otago, PO Box 913, Dunedin, New Zealand. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Human Genetics Commission. Profiling the newborn. 2005. Available at: www.hgc.gov.uk/Client/news_item.asp?NewsId=38. Accessed January 31, 2006
- ↵Evans JP, Skrzynia C, Burke W. The complexities of predictive genetic testing. BMJ.2001;322 :1052– 1056
- ↵Clarke A. The genetic testing of children. Working Party of the Clinical Genetics Society (UK). J Med Genet.1994;31 :785– 797
- Human Genetics Society of Australasia. Predictive testing in children and adolescents. 2003. Available at: hgsa.com.au/images/UserFiles/Attachments/Predictivetesting(General)APRIL2005.pdf. Accessed May 26, 2007
- ↵Nelson RM, Botkjin JR, Kodish ED, et al. Ethical issues with genetic testing in pediatrics. Pediatrics.2001;107 :1451– 1455
- ↵Michie S, Marteau T. Predictive genetic testing in children: the need for psychological research. In: Clarke A, ed. The Genetic Testing of Children. Oxford, United Kingdom: BIOS Scientific Publishers Ltd; 1998:331
- ↵Green M, Solnit AJ. Reactions to the threatened loss of a child: a vulnerable child syndrome. Pediatric management of the dying child: part III. Pediatrics.1964;34 :58– 66
- ↵Leslie LK, Boyce WT. The vulnerable child. Pediatr Rev.1996;17 :323– 326
- ↵Devendra D, Liu E, Eisenbarth GS. Type 1 diabetes: recent developments. BMJ.2004;328 :750– 754
- ↵The Environmental Determinants of Diabetes in the Young (TEDDY). Available at: http://teddy.epi.usf.edu. Accessed December 6, 2005
- ↵Bleich D, Polak M, Eisenbarth GS, Jackson RA. Decreased risk of type I diabetes in offspring of mothers who acquire diabetes during adrenarchy. Diabetes.1993;42 :1433– 1439
- Dahlquist G, Blom L, Tuvemo T, Nystrom L, Sandstrom A, Wall S. The Swedish childhood diabetes study: results from a nine year case register and a one year case-referent study indicating that type 1 (insulin-dependent) diabetes mellitus is associated with both type 2 (non-insulin-dependent) diabetes mellitus and autoimmune disorders. Diabetologia.1989;32 :2– 6
- Tillil H, Kobberling J. Age-corrected empirical genetic risk estimates for first-degree relatives of IDDM patients. Diabetes.1987;36 :93– 99
- Tuomilehto J, Lounamaa R, Tuomilehto-Wolf E, et al. Epidemiology of childhood diabetes mellitus in Finland–background of a nationwide study of type 1 (insulin-dependent) diabetes mellitus. The Childhood Diabetes in Finland (DiMe) Study Group. Diabetologia.1992;35 :70– 76
- ↵TrialNet. Available at: www.diabetestrialnet.org. Accessed April 5, 2006
- ↵Forsyth BW, Horwitz SM, Leventhal JM, Burger J, Leaf PJ. The child vulnerability scale: an instrument to measure parental perceptions of child vulnerability. J Pediatr Psychol.1996;21 :89– 101
- ↵Boyce P, Stubbs J, Todd A. The Edinburgh Postnatal Depression Scale: validation for an Australian sample. Aust NZ J Psychiatry.1993;27 :472– 476
- ↵Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry.1987;150 :782– 786
- ↵Webster ML, Thompson JM, Mitchell EA, Werry JS. Postnatal depression in a community cohort. Aust NZ J Psychiatry.1994;28 :42– 49
- ↵Spielberger C, Gorsuch R, Lushene R, Vagg P, Jacobs G. STAI Manual for the State Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press; 1970
- ↵Intercooled Stata for Windows [computer program]. Release 8.0. College Station, TX: Stata Corporation; 2003
- Gritz ER, Peterson SK, Vernon SW, et al. Psychological impact of genetic testing for hereditary nonpolyposis colorectal cancer. J Clin Oncol.2005;23 :1902– 1910
- ↵Boland C, Thompson NL. Effects of newborn screening of cystic fibrosis on reported maternal behaviour. Arch Dis Child.1990;65 :1240– 1244
- Parsons EP, Clarke AJ, Hood K, Lycett E, Bradley DM. Newborn screening for Duchenne muscular dystrophy: a psychosocial study. Arch Dis Child Fetal Neonatal Ed.2002;86 :F91– F95
- Tluczek A, Koscik RL, Farrell PM, Rock MJ. Psychosocial risk associated with newborn screening for cystic fibrosis: parents' experience while awaiting the sweat-test appointment. Pediatrics.2005;115 :1692– 1703
- ↵Bennett Johnson S, Baughcum AE, Carmichael SK, She JX, Schatz DA. Maternal anxiety associated with newborn genetic screening for type 1 diabetes. Diabetes Care.2004;27 :392– 397
- Hood KK, Bennett Johnson S, Carmichael SK, Laffel LM, She JX, Schatz DA. Depressive symptoms in mothers of infants identified as genetically at risk for type 1 diabetes. Diabetes Care.2005;28 :1898– 1903
- ↵Gurian EA, Kinnamon DD, Henry JJ, Waisbren SE. Expanded newborn screening for biochemical disorders: the effect of a false-positive result. Pediatrics.2006;117 :1915– 1921
- ↵Abidin R, ed. Parenting Stress Index: Professional Manual. 3rd ed. Odessa, FL: Psychological Assessment Resources; 1995
- ↵Barker JM, Goehrig SH, Barriga K, et al. Clinical characteristics of children diagnosed with type 1 diabetes through intensive screening and follow-up. Diabetes Care.2004;27 :1399– 1404
- ↵Eiser JR. Communication and interpretation of risk. Br Med Bull.1998;54 :779– 790
- Copyright © 2007 by the American Academy of Pediatrics