Abstract
Background:
Pet ownership is common. Growing evidence suggests children form deep emotional attachments to their pets. Yet, little is known about children’s emotional reactions to a pet’s death.
Aims:
To describe the relationship between experiences of pet death and risk of childhood psychopathology and determine if it is “better to have loved and lost than never to have loved at all”.
Method:
Data came from the Avon Longitudinal Study of Parents and Children, a UK-based prospective birth cohort (n=6260). Children were characterized based on their exposure to pet ownership and pet death from birth to age 7 (never loved; loved without loss; loved with loss). Psychopathology symptoms at age 8 were compared across groups using multivariable linear regression.
Results:
Psychopathology symptoms were higher among children who had loved with loss compared to those who had loved without loss (β=0.35, p=0.013; 95% CI=0.07, 0.63), even after adjustment for other adversities. This group effect was more pronounced in males than in females. There was no difference in psychopathology symptoms between children who had loved with loss and those who had never loved (β=0.20, p=0.31, 95% CI =−0.18, 0.58). The developmental timing, recency, or accumulation of pet death was unassociated with psychopathology symptoms.
Conclusions:
Pet death may be traumatic for children and associated with subsequent mental health difficulties. Where childhood pet ownership and pet bereavement is concerned, Tennyson’s pronouncement may not apply to children’s grief responses: it may not be “better to have loved and lost than never to have loved at all”.
Introduction
Pet ownership is common. Roughly half of households in developed countries own at least one pet [1, 2]. For example, 31% of United Kingdom households report owning a dog and 26% report owning a cat, with smaller but substantial percentages reporting ownership of other household animal types [3, 4]. Since the 1980’s, an accumulating body of research into human animal interaction (HAI) and human animal bonding (HAB) suggests that people can form complex bonds to animals [5]. This research has often focused on children, given the particularly high prevalence of pet ownership during childhood [4, 6] as well as the development of child-oriented interventions that capitalize on the developmental benefits of HAI and HAB. From this literature, there is increasing evidence that children often form deep emotional attachments to their pets. These attachments can resemble secure human attachment relationships [6–8] in providing several key resources, such as affection, protection, and reassurance [6, 9]. Previous studies have shown children often turn to pets for comfort and to discuss emotional experiences [10, 11]. Childhood pet ownership and attachment has, in turn, been linked to a number of positive developmental consequences associated with healthy attachment, such as increased empathy [12, 13], self-esteem [14, 15], and greater social competence [16, 17].
Unfortunately, one consequence of the high prevalence of childhood pet ownership is that many children are exposed to the death of a pet. The two most common pet types – dogs and cats – live an average of 12 and 15 years, respectively [18]. Thus, many youth living in households with a pet will experience the death of that pet sometime during childhood. Although relatively little research has been done to empirically study children’s emotional reactions to a pet’s death, children’s grief in response to the loss of other important attachment relationships has been well-documented [19–21]. Though children’s grief responses may be distinct from those of adults––with bereaved children displaying infantile behaviors, fearfulness [22], and somatic reactions, including headaches and stomach aches [23]––their grief may be no less intense [20, 24]. In general, the death of a family member has been associated with an increased risk of childhood psychopathology symptoms [25], including anxiety [26], post-traumatic stress symptoms [27], and depressive symptoms [27]. It has also been shown that although grief reactions for most children abate over time following the death of a loved one, some children can exhibit a high, prolonged grief response known as complicated grief. Complicated grief is a particularly potent predictor of depression in children and adolescents as far as three years after the loss [19].
Despite the prevalence of pet death as a potentially traumatic loss during childhood, very little research has examined the mental health consequences of children’s exposure to the death of a pet. The few cross-sectional and retrospective studies that have explored this topic have primarily studied psychopathology symptoms in adults [28], among whom pet death has been associated with increased risk for neurotic [29] and depressive symptoms [30], though risk for major psychopathology following pet death is low [31]. Prior case reports and empirical studies have found that compared to adults, children’s grief responses to a pet’s death can be profound [32, 33], and can have greater intensity and duration [34].
To our knowledge, no previous studies have explored childhood mental health problems following the death of a pet. Thus, it remains unclear whether pet death is associated with psychopathology symptoms, and if the known positive effects of owning a pet outweigh any negative consequences associated with pet bereavement. In the words of British poet Alfred Lord Tennyson, the question remains: is it “better to have loved and lost than never to have loved at all”? [35]. The current study aimed to answer this question by using data from a deeply characterized prospective longitudinal population-based birth-cohort study, containing serial measures of household pet ownership and child exposure to pet death. With these data, we explored the association between pet death and subsequent psychopathology symptoms during childhood, focusing on differences between non-pet owners (never loved), pet owners who never experienced the death of a pet (love without loss), and pet owners who experienced a pet death (love with loss).
Methods
Sample and Procedures
Data came from the Avon Longitudinal Study of Parents and Children (ALSPAC), a prospective, longitudinal birth cohort of children born to pregnant mothers who were living in the county of Avon England (120 miles west of London) with estimated delivery dates between April 1991 and December 1992 [36, 37]. Approximately 85% of eligible pregnant women agreed to participate (N=14,541), and 76% of eligible live births (N=14,062) who were alive at 12 months of age (N=13,988 children) were enrolled. Response rates to data collection have been good (75% have completed at least one follow-up), with 56% (N=7912) of the original sample participating in the age 8 assessment. Ethical approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committee. More details are available on the ALSPAC website, including a fully searchable data dictionary: http://www.bristol.ac.uk/alspac/researchers/our-data/.
Measures
Pet Ownership and Exposure to Pet Death
Pet ownership and exposure to pet death were determined through mailed questionnaires completed by the mothers.
Pet ownership was assessed in a questionnaire about living arrangements, where the mother indicated if she owned a pet and if so, how many. This questionnaire was completed at five time periods, when the child was 8 months, 21 months (1.75 years), 33 months (2.75 years), 47 months (3.9 years), and 84 months (7 years) of age.
Children’s exposure to pet death was determined through an item in a stressful life events inventory, asking the mother to indicate whether or not the child had been exposed to pet death since the last questionnaire. This questionnaire was completed at six time periods, when the child was 18 months (1.5 years), 30 months (2.5 years), 42 months (3.5 years), 60 months (5 years), 72 months (6 years), and 84 months (7 years) of age. Age of exposure was defined as the age of the child at the time the mother completed the questionnaire indicating her child had experienced pet death. For example, if the mother indicated at the age 30 month assessment that the death of a pet had occurred at some time since the previous assessment (at 18 months), the age of exposure was coded as 30 months.
We used these data to categorize children into one of three mutually exclusive groups: never loved, meaning children who were non-pet owners throughout the entire time period; love with loss, meaning children who were pet owners and experienced the death of at least one pet (in a time period subsequent to the report of pet ownership); and love without loss, meaning children who were pet owners who did not experience the death of a pet.
Given that the focus of ALSPAC is on children and their development rather than pet ownership specifically, these survey measures did not allow us to identify certain relevant details, such as the type of pet that died or the strength of the child’s attachment to that pet. These child-centric measures were, however, unparalleled in their attention to the timing of exposure and measurement of co-occurring adversities. The limitations of these measures are addressed in further detail in the Discussion section.
Child Psychopathology
Child psychopathology symptoms were assessed using the Strengths and Difficulties Questionnaire (SDQ) [38, 39], which mothers completed by mail when the child was 8 years old. The SDQ is one of the most commonly used dimensional rating scales of child psychopathology in epidemiology studies and has excellent psychometric properties [40, 41]. The SDQ contains 25 items, rated on a three-point scale (0=not true, 1=somewhat true, or 2=certainly true), capturing the child’s behavior and feelings within the past six months. We calculated a total SDQ score by summing across items on the first four subscales (conduct problems; emotional symptoms; hyperactivity; peer problems; range 0–40), with higher scores indicating more emotional and behavioral difficulties (α=0.82). This total score has been shown in studies from across the globe to correlate highly with questionnaire and interview measures of psychopathology, including the Child Behavior Checklist as well as clinician-rated diagnoses of child mental disorder [42, 43].
Covariates
We controlled for the following baseline covariates, measured at the time of the child’s birth: child sex; child race/ethnicity; number of previous pregnancies; maternal marital status; highest level of maternal education; maternal age; homeownership; parent social class; singleton or multiple birth; and maternal depression, as assessed by the Edinburgh Postnatal Depression Scale (EPDS) [44]. Covariates were selected for inclusion because they were found to be potential confounders in our sample, or because they have been included routinely in longitudinal birth cohorts when studying child mental health outcomes [45–47]. For example, prior studies have found higher levels of pet ownership among families with lower education levels [4, 48] and lower parent social class (as defined by occupation) [4, 49]. Adjustment for maternal depression allowed us to reduce potential impacts of common rater bias [50], as mothers reported about both their child’s exposure to pet death as well as their child’s emotional and behavioral problems, and maternal mood or other factors may influence reports of adversity exposure [51] and psychopathology [52, 53].
Recognizing that childhood adversities often co-occur, and that the effects of pet death on psychopathology could be confounded by experiences of other adversities, we additionally adjusted for exposure to three major types of childhood adversity: financial hardship, caregiver physical or emotional abuse, and physical or sexual abuse by anyone (see Supplemental Materials for details).
Primary Analyses
To reduce potential bias and minimize loss of power due to attrition [54, 55], we conducted all analyses using multiply imputed datasets, where missing exposure (i.e., pet ownership and pet death) and covariate information were imputed using the MICE package in R [55] (see Supplemental Materials).
Our analysis was based on an analytic sample of 6260 children out of a possible 7912 (79%) who completed the age 8 assessment, which was the last time point of data examined in the current analysis. The analytic sample met two inclusion criteria. First, given that methods for imputation of missing outcomes may induce additional noise [56], we restricted our analyses to children who had a completed outcome measure. This criterion omitted 436 children from the sample who participated in the age 8 assessments. Second, in the interest of deriving exposure groups that were as homogenous as possible, we omitted children from our primary analysis whose mothers reported that the child had experienced the death of a pet although no pet had been indicated to reside in the household in prior assessments (n=1216; 16%) Supplemental Figure 1). The experience of pet loss in the absence of pet ownership was likely due to the child experiencing a pet loss outside of the home (e.g., at a grandparent’s home or in a school classroom, where children often encounter pets with whom they may bond [57, 58]). Further details can be found in Supplemental Materials.
We began the analysis by running univariate and bivariate analyses to examine the distribution of baseline covariates in the total analytic sample and by our three exposure groups. We then used multivariable linear regression to compare child psychopathology symptom scores across the three exposure groups (never loved, love without loss, and love with loss), after adjustment for baseline covariates (Model 1). To ensure these results were not explained by exposure to other types of adversities, we ran a set of models – building from Model 1 – to additionally adjust for the role of exposure to financial hardship (Model 2), caregiver physical or emotional abuse (Model 3), physical or sexual abuse by anyone (Model 4), and all three adversities considered simultaneously (Model 5).
Secondary Analyses
We conducted three sets of secondary analyses. First, given documented differences between girls’ and boys’ grief responses to pet death [59], as well sex differences in psychopathology symptoms [60, 61], we reran the primary analyses stratified by sex.
Second, based on evidence from life course theory that the effects of childhood adversity on risk for childhood psychopathology may vary depending on the characteristics of the exposure, including when it occurs in development, how many times it occurs, and how recently it occurred [62, 63], we capitalized on the availability of the repeated measures of pet death and pet ownership to examine the potential time-dependent effects of pet death on childhood psychopathology symptoms. Specifically, we used a structured life course modeling approach grounded in least angle regression [64, 65] to evaluate which of the three life course theoretical models explained the most variability in child psychopathology symptoms, as determined by r2 values [66]. The life course models tested were: (1) a sensitive period model [66]; (2) an accumulation model [67]; and (3) a recency model [68] (see Supplemental Materials).
Third, recognizing that the experience of pet death may still be impactful for children who lost non-household animals, we examined the effects of being ever exposed to pet death without differentiating between explicit and ambiguous pet ownership. Thus, we reran all models to include the 1216 children who likely experienced pet loss outside of home and were excluded from our primary analysis. These results are reported as Models 6–10.
Results
Sample Characteristics and Distribution of Exposure to Pet Death
The analytic sample was sex-balanced (50.7% male) and comprised of predominately White (97.0%) children from families whose parents were married and owned their home (Table 1). Pet death was common in this sample, with most children experiencing the death of a pet at some point in their lives (52.7%; N=3296). A large percentage of children had pets that were still living (love without loss group N=1682; 26.9%), with only 808 children (12.9%) belonging to the never loved group. These three subgroups differed on some demographic characteristics. Specifically, children in the love with loss group were more likely to be female (p=0.001), non-White (p<0.001), from families with less parental education (p<0.001) and lower parental social class (p<0.001), and were exposed to other forms of childhood adversity (Table 1). Among children in this love with loss group, the most frequent age at first exposure to the death of a pet was 4.75 years (24%) (Figure 1).
Table 1.
Distribution of covariates in the total ALSPAC analytic sample and by the three subgroups defined by pet ownership and pet death
| Total sample (N=6260) | Love with Loss (N=3296) | Love without Loss (N=1682) | Never Loved (N=808) | P-value | |
|---|---|---|---|---|---|
| N (%) | N (%) | N (%) | N (%) | ||
| 0.001 | |||||
| Males | 3175 (50.7) | 1609 (48.8) | 877 (52.1) | 451 (55.8) | |
| Females | 3085 (49.3) | 1687 (51.2) | 805 (47.9) | 357 (44.2) | |
| Race | <0.001 | ||||
| Non-White | 184 (3.0) | 67 (2.1) | 58 (3.5) | 42 (5.3) | |
| White | 5884 (97.0) | 3134 (97.9) | 1582 (96.5) | 751 (94.7) | |
| Maternal education | <0.001 | ||||
| less than O-level | 1290 (21.0) | 697 (21.5) | 340 (20.5) | 109 (13.6) | |
| O-level | 2174 (35.3) | 1234 (38.1) | 533 (32.1) | 252 (31.5) | |
| A-level | 1668 (27.1) | 880 (27.1) | 473 (28.5) | 217 (27.1) | |
| Degree or above | 1023 (16.6) | 431 (13.3) | 315 (19.0) | 222 (27.8) | |
| Maternal marital status | 0.326 | ||||
| Never Married | 764 (12.4) | 384 (11.8) | 212 (12.7) | 86 (10.7) | |
| Widowed/Divorced/Separated | 304 (4.9) | 171 (5.3) | 72 (4.3) | 35 (4.4) | |
| Married | 5115 (82.7) | 2700 (82.9) | 1380 (82.9) | 680 (84.9) | |
| Home ownership | <0.001 | ||||
| Mortgage/own home | 5149 (83.8) | 2691 (83.1) | 1400 (84.7) | 697 (88.3) | |
| Rent home | 828 (13.5) | 446 (13.8) | 215 (13.0) | 87 (11.0) | |
| Other | 164 (2.7) | 100 (3.1) | 37 (2.2) | 5 (0.6) | |
| Age of mother at child birth | 0.029 | ||||
| Ages 15–19 | 99 (1.6) | 47 (1.4) | 29 (1.7) | 3 (0.4) | |
| Ages 20–35 | 5616 (89.7) | 2971 (90.1) | 1503 (89.4) | 720 (89.1) | |
| Age >35 | 545 (8.7) | 278 (8.4) | 150 (8.9) | 85 (10.5) | |
| Parental social class (occupation) | <0.001 | ||||
| Professional | 909 (14.5) | 399 (12.1) | 277 (16.5) | 185 (22.9) | |
| Managerial and technical | 2424 (38.7) | 1261 (38.3) | 663 (39.4) | 335 (41.5) | |
| Skilled, non-manual | 1354 (21.6) | 760 (23.1) | 347 (20.6) | 144 (17.8) | |
| Skilled, manual | 348 (5.6) | 209 (6.3) | 93 (5.5) | 22 (2.7) | |
| Semi-skilled, manual | 103 (1.6) | 63 (1.9) | 25 (1.5) | 5 (0.6) | |
| Unskilled, manual or other | 1122 (17.9) | 604 (18.3) | 277 (16.5) | 117 (14.5) | |
| Number of previous pregnancies | <0.001 | ||||
| 0 | 2782 (45.5) | 1314 (40.9) | 786 (47.8) | 464 (58.7) | |
| 1 | 2222 (36.4) | 1195 (37.2) | 624 (37.9) | 253 (32.0) | |
| 2 | 837 (13.7) | 533 (16.6) | 181 (11.0) | 54 (6.8) | |
| 3+ | 267 (4.4) | 174 (5.4) | 54 (3.3) | 20 (2.5) | |
| Singleton vs. multiple birth | 0.156 | ||||
| Singleton | 6128 (97.9) | 3239 (98.3) | 1645 (97.8) | 786 (97.3) | |
| Multiple birth | 132 (2.1) | 57 (1.7) | 37 (2.2) | 22 (2.7) | |
| Financial hardship | <0.001 | ||||
| Never exposed | 4092 (69.4) | 2073 (66.7) | 1237 (75.7) | 634 (78.7) | |
| Exposed | 1802 (30.6) | 1033 (33.3) | 397 (24.3) | 172 (21.3) | |
| Caregiver physical or emotional abuse | 0.029 | ||||
| Never exposed | 4444 (83.0) | 2295 (81.6) | 1329 (84.4) | 656 (84.3) | |
| Exposed | 908 (17.0) | 519 (18.4) | 246 (15.6) | 122 (15.7) | |
| Physical or sexual abuse by anyone | 0.002 | ||||
| Never exposed | 4533 (87.5) | 2370 (86.0) | 1500 (89.2) | 659 (89.4) | |
| Exposed | 645 (12.5) | 385 (14.0) | 182 (10.8) | 78 (10.6) | |
| Mean (SD) | Mean (SD) | 0.012 | |||
| Maternal depression | 5.16 (4.54) | 5.27 (4.55) | 5.00 (4.50) | 4.81 (4.39) | 0.041 |
Note. The groups reported here were determined before imputation using complete-case data, meaning any child who had complete pet ownership and pet death exposure data (n=5786). The actual group proportions varied slightly across the 20 imputed datasets. The p-values corresponded to chi-squared tests when the covariate was a categorical variable (testing the null hypothesis that the covariates were equally distributed among the three exposure subgroups). For maternal depression (continuous), ANOVA was performed and the corresponding p-value was reported. Since most covariates and the pet death exposure variables had missingness, the cell counts do not sum to the total sample size.
Figure 1.
Child age at first exposure to pet death and number of occasions exposed among the Love with Loss group, meaning children who were pet owners and experienced pet death
Primary Analyses: Association between Pet Death and Child Psychopathology Symptoms
As shown in Table 2 and Figure 2 for Model 1, there were no differences observed in psychopathology symptoms between children in the love without loss group and the children who never loved (p=0.45) after adjustment for baseline covariates. Similarly, there were also no differences in psychopathology symptoms observed between the love with loss group and the never loved group (p=0.31).
Table 2.
Results of linear regression models examining difference in child psychopathology symptom scores between groups in the ALSPAC analytic sample (N=6260), after adjustment for covariates and exposure to other childhood adversities.
| Beta | SE | P-value | 95% CI | |
|---|---|---|---|---|
| Model 1: Baseline covariates only | ||||
| Exposure | ||||
| Never loved vs. Love with Loss | 0.20 | 0.19 | 0.311 | (−0.18, 0.58) |
| Never loved vs. Love without loss | −0.15 | 0.20 | 0.452 | (−0.56, 0.25) |
| Love without loss vs. Love with Loss | 0.35 | 0.14 | 0.013* | (0.07, 0.63) |
| Model 2: Model 1 + Financial hardship | ||||
| Exposure | ||||
| Never loved vs. Love with Loss | 0.13 | 0.19 | 0.507 | (−0.25, 0.51) |
| Never loved vs. Love without loss | −0.19 | 0.20 | 0.358 | (−0.59, 0.21) |
| Love without loss vs. Love with Loss | 0.32 | 0.14 | 0.025* | (0.04, 0.59) |
| Covariate | ||||
| Never vs. ever exposed to financial stress | 0.72 | 0.14 | <.001** | (0.45, 1) |
| Model 3: Model 1 + Caregiver physical or emotional abuse | ||||
| Exposure | ||||
| Never loved vs. Love with Loss | 0.17 | 0.19 | 0.369 | (−0.21, 0.55) |
| Never loved vs. Love without loss | −0.15 | 0.20 | 0.476 | (−0.54, 0.25) |
| Love without loss vs. Love with Loss | 0.32 | 0.14 | 0.023* | (0.04, 0.59) |
| Covariate | ||||
| Never vs. ever exposed to phys/emo abuse | 1.39 | 0.17 | <.001** | (1.05, 1.72) |
| Model 4: Model 1+ Physical or sexual abuse by anyone | ||||
| Exposure | ||||
| Never loved vs. Love with Loss | 0.14 | 0.19 | 0.485 | (−0.24, 0.51) |
| Never loved vs. Love without loss | −0.17 | 0.20 | 0.401 | (−0.57, 0.23) |
| Love without loss vs. Love with Loss | 0.31 | 0.14 | 0.029* | (0.03, 0.58) |
| Covariate | ||||
| Never vs. ever exposed to phys/sex abuse | 1.56 | 0.19 | <.001** | (1.19, 1.94) |
| Model 5: Model 1+ All three childhood adversities | ||||
| Exposure | ||||
| Never loved vs. Love with Loss | 0.07 | 0.19 | 0.722 | (−0.31, 0.45) |
| Never loved vs. Love without loss | −0.19 | 0.20 | 0.351 | (−0.59, 0.21) |
| Love without loss vs. Love with Loss | 0.26 | 0.14 | 0.066 | (−0.02, 0.53) |
| Covariate | ||||
| Never vs. ever exposed to financ. hardship | 1.35 | 0.19 | <.001** | (0.97, 1.73) |
| Never vs. ever exposed to phys/emo abuse | 0.59 | 0.14 | <.001** | (0.31, 0.86) |
| Never vs. ever exposed to phys/sex abuse | 1.17 | 0.17 | <.001** | (0.83, 1.51) |
Note. In these analyses, the first group listed, meaning before the vs., was the referent group. The names of the models indicate what variables were adjusted for when estimating the effects of the pet ownership and exposure status in the regression analyses. The covariate and exposure to other adversity variables are described in the Methods section.
The corresponding beta estimate was significantly different from 0 at p<.05.
The corresponding beta estimate was significantly different from 0 at p<.0001.
Figure 2.
Results of linear regression models examining difference in child psychopathology symptom scores between groups in the full ALSPAC analytic sample, adjusting for covariates and exposure to other adversity.
Note. Each vertical line represents point estimates and the corresponding confidence interval. The psychopathology symptoms in the love with loss group, compared to the love without loss group were significantly higher in Models 1–4, although the magnitude of effect was not as large as the effects of other major types of childhood adversity.
However, psychopathology symptom scores were higher among children who experienced pet death (love with loss), compared those who had pets that were still living (love without loss) (β=0.35, p=0.013; 95% CI=0.07,0.63). This relative increase in psychopathology symptoms persisted, though was slightly attenuated, after adjustment for financial hardship (Model 2), caregiver physical or emotional abuse (Model 3), and physical or sexual abuse by anyone (Model 4). When all three types of adversity were included simultaneously as covariates (Model 5), the difference in psychopathology symptoms associated with pet loss was marginally statistically significant (β=0.26; p=0.06). Notably, in visually examining the magnitude of the difference in psychopathology symptoms between the love with loss group compared to the love without loss group, we can see across Models 2–5 that this effect was at least one third as large as the magnitude of having ever been exposed to each of the adversity covariates (Table 2).
Secondary Analyses: Association between Pet Death and Child Psychopathology Symptoms
Figure 3 shows that the increase in psychopathology symptoms in the love with loss group compared to the love without loss group was more pronounced in males than in females (Model 1: βmale=0.45, pmale=0.035; βfemale=0.28, pfemale=0.14). The patterns of between-group differences in males were similar to the results from the primary analysis; however, we did not observe any group effect in females.
Figure 3.
Results of linear regression models examining difference in child psychopathology symptom scores between groups stratified by sex, adjusting for covariates and exposure to other adversity.
Note. Each vertical line represents a point estimate and the corresponding confidence interval. After stratifying by sex, the effects of the love with loss group relative to the love without loss group were no longer significant in girls, but they were still observed in Models 1–3 in boys.
There were no meaningful differences in risk for psychopathology symptoms based on the developmental timing, recency, and accumulation of exposure to pet death. That is, all life course theoretical models were weak and inconclusive predictors of child psychopathology in both the full sample and among the sample of pet owners (p>0.05; Supplemental Table 1).
As shown in Table 3, children exposed to the death of a pet, whether that pet resided in their household or not, had psychopathology symptoms scores that were slightly higher than their peers who did not experience a pet death (β=0.26; 95% CI=0.03, 0.50; p=0.03), after adjustment for covariates (Model 6). This effect was still observed after accounting for exposure to financial hardship (Model 7), but no longer statistically significant after adjustment for the other two abuse-related adversities (Model 8–10). Compared to the primary analyses, where subgroups were defined based on pet loss and pet ownership status, the effect sizes in this model associated with the ever versus never exposed analyses were smaller, suggesting that defining the pet loss experience with more precision allowed us to see more meaningful patterns.
Table 3.
Results of linear regression models examining difference in child psychopathology symptom scores between those ever versus never exposed to pet death regardless of pet ownership (N=7476), after adjustment for covariates and exposure to other major childhood adversities.
| Beta | SE | P-value | 95% CI | |
|---|---|---|---|---|
| Model 6: Baseline covariates only | ||||
| Never vs. ever exposed to pet death | 0.26 | 0.12 | 0.029* | (0.03,0.5) |
| Model 7: Model 6 + Ever/never exposed to financial hardship | ||||
| Never vs. ever exposed to pet death | 0.24 | 0.12 | 0.047* | (0,0.48) |
| Never vs. ever exposed to financial stress | 0.53 | 0.13 | <.001** | (0.27,0.79) |
| Model 8: Model 6 + Ever/never exposed to caregiver physical or emotional abuse | ||||
| Never vs. ever exposed to pet death | 0.23 | 0.12 | 0.052 | (0,0.47) |
| Never vs. ever exposed to caregiver physical or emotional abuse | 1.32 | 0.18 | <.001** | (0.97,1.66) |
| Model 9: Model 6 + Ever/never exposed to physical or sexual abuse by anyone | ||||
| Never vs. ever exposed to pet death | 0.21 | 0.12 | 0.076 | (−0.02,0.45) |
| Never vs. ever exposed to physical or sexual abuse by anyone | 1.56 | 0.18 | <.001** | (1.19,1.92) |
| Model 10: Model 6 + Ever/never exposed to all three major childhood adversities | ||||
| Never vs. ever exposed to pet death | 0.18 | 0.12 | 0.138 | (−0.06,0.41) |
| Never vs. ever exposed to financial hardship | 1.36 | 0.19 | <.001** | (0.99,1.73) |
| Never vs. ever exposed to caregiver physical or emotional abuse | 0.36 | 0.13 | 0.005 | (0.11,0.62) |
| Never vs. ever exposed to physical or sexual abuse by anyone | 1.11 | 0.18 | <.001** | (0.76,1.46) |
Note. In these analyses, the never exposed group was the referent. The names of the models indicate what variables were adjusted for when estimating the effects of exposure to pet death in the regression analyses. The covariate and exposure to other adversity variables are described in the Methods section.
The corresponding beta estimate was significantly different from 0 at p<.05.
The corresponding beta estimate was significantly different from 0 at p<.0001.
Discussion
To our knowledge, the current study is the first to test the association between exposure to a pet’s death and psychopathology symptoms in childhood. Three main findings emerged from this prospective study. First, we found that pet ownership was common, with most children (88%) in our sample having owned a pet at some point in childhood. Second, pet death was also a common childhood experience, with a substantial proportion (63%) of children having lost a pet during the first seven years of life. Third, we found that these experiences of pet death were associated with elevated psychopathology symptoms. This association was observed even after accounting for other adverse factors known to increase child risk for poor mental health, such as low socioeconomic status, maternal history of depression, and exposure to child abuse. These findings align with previous work in adult grief documenting increased neurotic and depressive symptoms following the death of a pet [28–30]. Our findings also align with the few case reports and empirical studies exploring the psychological sequelae of pet bereavement in childhood [33, 34], which have found that children’s grief responses to a pet’s death can surpass adults’ responses in intensity and duration [34]. Most previous studies of pet bereavement in children and adults have not accounted for the potential psychological benefits of pet ownership. From what we can determine, this is the first study to compare groupings of pet ownership in this manner and thus our findings regarding the differences between love with loss and love without loss are novel.
Three additional findings were observed as well. First, the association between pet death and elevated psychopathology symptoms was stronger in male children than in female children, which was somewhat unexpected given previous research in adolescents suggesting that females reported a more intense grief response to a pet’s death than did males [59]. Additionally, this association was stronger for household pets versus non-household pets; however, even in the case of the death of a non-household pet, children still showed an increase in psychopathology symptoms. Finally, the strength of this association did not vary as a function of when the pet’s death occurred during childhood, how many times it occurred, or how recently it occurred. This finding was somewhat surprising in light of emerging work suggesting that exposure to adversity in the first five years of life may be especially important in shaping risk for psychopathology symptoms in childhood [62] and beyond [69, 70]. We did not, however, find evidence to suggest similar timing effects here.
This study had three major strengths. First, despite the ubiquity of pet ownership [1, 2] and the fact that a pet’s death is likely the first major loss a child will encounter [59], few studies have systematically explored the effect of pet death on children’s risk for experiencing psychopathology symptoms. Our study therefore addresses an important, but understudied issue. Second, we addressed this issue by analyzing data from a large, longitudinal, and population-based sample of children, who were followed from birth and whose mothers had provided repeated measures that allowed us to track experiences of pet ownership and pet loss across time. These serial measurements enabled us to capture events during childhood without relying on retrospective reporting, which is commonplace among studies examining the consequences of childhood adversities. The depth of measurement in ALSPAC also allowed us to adjust for other important potential confounders, notably experiences of co-occurring adversity. Third, we could characterize experiences of pet death in ways that moved beyond the simple classification of children as ever versus never exposed.
Several limitations are noted. Although ALSPAC contains rich data collected from parents and children, the study was not designed to investigate pet ownership and pet death experiences, thus these measures of these constructs lacked some granularity. For example, while there was information available about the type of pet the child had, there was no data available to identify which of the pets had died. Moreover, we were unable to examine the effects of pet death for specific types of pets, including cats or dogs. This was a limitation because prior studies have shown that children tend to form stronger bonds with dogs and cats, and less strong attachments with pet birds or fish [6, 71]. Future studies could extend these findings by examining the role of the type of pet death to elucidate differences that may emerge from different types of animal bonding. Additionally, while earlier child psychopathology may be linked to pet ownership and later psychopathology symptoms, we did not adjust for psychopathology symptoms before age 8, as this would prove difficult for maintaining temporality in the exposure-disease association. In brief, our first indicator of exposure to pet death at age 18 months occurred before the first assessment of psychopathology symptoms in ALSPAC. Thus, inclusion of psychopathology measured after this time point would create temporal ambiguity with respect to our exposure-outcome association. That is, while psychopathology symptoms were assessed at 48 months, adding this measure as a covariate would be problematic as it would likely mediate the relationship between exposure to pet ownership and pet death that occurred before 48 months and psychopathology symptoms at age 8. We hope future studies will be able to more carefully account for time-varying covariates so that the prospective and longitudinal association between pet death and child psychopathology can be studied. Finally, the high prevalence of pet death (above 50%) in the analytic sample indicated that the classification likely covered a wide range of experiences spanning in severity. In future studies, the experience of pet death could be further characterized to capture more subtle distinctions within the love with loss group, which likely reflect not only different pet types but different durations of pet ownership and the strength of attachments between children and their pets.
In conclusion, Tennyson’s pronouncement may not, in fact, apply to children’s grief responses to pet bereavement: where childhood pet ownership is concerned, it may not be “better to have loved and lost than never to have loved at all”. Our study results suggest that pet death may be traumatic for children and that children who have pets may show signs of mental health difficulties if their pet dies. Especially when pets feel like members of the family and children are attached to their pets, parents and other caregivers may find it beneficial to recognize children’s short- and long-term psychological reactions, which may mimic responses to the loss of other important human attachments. The death of a pet should be treated as the loss of other strong emotional attachments, and parents and physicians should be prepared to treat it as such.
Supplementary Material
Funding Statement and Acknowledgments:
We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The UK Medical Research Council and the Wellcome Trust (Grant ref: 102215/2/13/2) and the University of Bristol provide core support for ALSPAC. A comprehensive list of grants funding is available on the ALSPAC website (http://www.bristol.ac.uk/alspac/external/documents/grant-acknowledgements.pdf).
This publication is the work of the authors, each of whom serve as guarantors for the contents of this paper. This work was conducted with support from the National Institutes of Health Award K01MH102403 and R01MH113930. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors thank Janine Cerutti for her assistance in preparing this manuscript for publication.
Footnotes
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
Conflict of Interest Statement: On behalf of all authors, the corresponding author, states that there is no conflict of interest. All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.
License for Publication: The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, a worldwide licence to the Publishers and its licensees in perpetuity, in all forms, formats and media (whether known now or created in the future), to i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include within collections and create summaries, extracts and/or, abstracts of the Contribution, iii) create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) license any third party to do any or all of the above.
Ethical approval: Ethical approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committee. Informed consent for the use of data collected via questionnaires and clinics was obtained from participants following the recommendations of the ALSPAC Ethics and Law Committee at the time. More details are available on the ALSPAC website (www.bristol.ac.uk/alspac), including a fully searchable data dictionary.
Transparency statement: Dr. Erin C. Dunn, the manuscript’s senior author and grantor, affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.
Contributor Information
Kristen Nishimi, Center for Genomic Medicine, Massachusetts General Hospital, Boston, MA; Department of Social and Behavioral Sciences, Harvard T.H. Chan School of Public Health, MA.
Andrew D.A.C. Smith, Applied Statistics Group, University of the West of England, UK.
Erin C. Dunn, Center for Genomic Medicine, Massachusetts General Hospital, Boston, MA; Department of Psychiatry, Harvard Medical School, Boston, MA; Center on the Developing Child at Harvard University, Cambridge, MA.
References
- 1.Voith VL, The Impact of Companion Animal Problems on Society and the Role of Veterinarians. Veterinary Clinics of North America: Small Animal Practice, 2009. 39(2): p. 327–345. [DOI] [PubMed] [Google Scholar]
- 2.Utz R, Walking the Dog: The Effect of Pet Ownership on Human Health and Health Behaviors. Social Indicators Research, 2014. 116(2): p. 327–339. [Google Scholar]
- 3.Murray JK, et al. , Number and ownership profiles of cats and dogs in the UK. Veterinary Record, 2010. 166(6): p. 163. [DOI] [PubMed] [Google Scholar]
- 4.Westgarth C, et al. , Family Pet Ownership during Childhood: Findings from a UK Birth Cohort and Implications for Public Health Research. International Journal of Environmental Research and Public Health, 2010. 7(10): p. 3704–3729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hosey G and Melfi V, Human-animal interactions, relationships and bonds: a review and analysis of the literature. 2014.
- 6.Hawkins RD, Williams JM, and A. Scottish Society for the Prevention of Cruelty to, Childhood Attachment to Pets: Associations between Pet Attachment, Attitudes to Animals, Compassion, and Humane Behaviour. International Journal of Environmental Research and Public Health, 2017. 14(5): p. 490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Beck L and Madresh EA, Romantic Partners and Four-Legged Friends: An Extension of Attachment Theory to Relationships with Pets. Anthrozoös, 2008. 21(1): p. 43–56. [Google Scholar]
- 8.Carr S and Rockett B, Fostering secure attachment: experiences of animal companions in the foster home. Attachment & Human Development, 2017. 19(3): p. 259–277. [DOI] [PubMed] [Google Scholar]
- 9.Zilcha-Mano S, Mikulincer M, and Shaver PR, Pets as safe havens and secure bases: The moderating role of pet attachment orientations. Journal of Research in Personality, 2012. 46(5): p. 571–580. [Google Scholar]
- 10.Bryant BK, The neighborhood walk: Sources of support in middle childhood. Monographs of the Society for Research in Child Development, 1985. 50(3): p. 122–122. [PubMed] [Google Scholar]
- 11.Covert AM, et al. , Pets, Early Adolescents, and Families. Marriage & Family Review, 1985. 8(3–4): p. 95–108. [Google Scholar]
- 12.Poresky RH and Hendrix C, Differential Effects of Pet Presence and Pet-Bonding on Young Children. Psychological Reports, 1990. 67(1): p. 51–54. [DOI] [PubMed] [Google Scholar]
- 13.Vidović VV, Štetić VV, and Bratko D, Pet Ownership, Type of Pet and Socio-Emotional Development of School Children. Anthrozoös, 1999. 12(4): p. 211–217. [Google Scholar]
- 14.Poresky RH, et al. , Young Children’s Companion Animal Bonding and Adults’ Pet Attitudes: A Retrospective Study. Psychological Reports, 1988. 62(2): p. 419–425. [DOI] [PubMed] [Google Scholar]
- 15.Poresky RH, et al. , Children’s pets and adults’ self-concepts. (Tennessee Self Concept Scale, Companion Animal Bonding Scale). The Journal of Psychology, 1988. 122(5): p. 463. [Google Scholar]
- 16.Guttman G, Predovic M, and Zemanek M, Influence of pet ownership on non-verbal communication and social competence in children, in Proceedings of the International Symposium on the Human–Pet Relationship. 1983: Vienna. p. 58–63. [Google Scholar]
- 17.Endenburg N and Baarda DB, The roles of pets in enhancing human well-being: effects on child development, in The Waltham Book of Human–Animal Interaction: Benefits and Responsibilities of Pet Ownership., Robinson I, Editor. 1995, Pergamon: Exeter. p. 1–17. [Google Scholar]
- 18.Grimm D, Why we outlive our pets. Science, 2015. 350(6265): p. 1182–1185. [DOI] [PubMed] [Google Scholar]
- 19.Melhem NM, et al. , Grief in Children and Adolescents Bereaved by Sudden Parental Death. Archives of General Psychiatry, 2011. 68(9): p. 911–919. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wolraich M, et al. , The pediatrician and childhood bereavement. American Academy of Pediatrics. Committee on Psychosocial Aspects of Child and Family Health. Pediatrics, 2000. 105(2): p. 445. [PubMed] [Google Scholar]
- 21.Bugge K, et al. , Young Children’s Grief: Parents’ Understanding and Coping. Death Studies, 2014. 38(1): p. 36. [DOI] [PubMed] [Google Scholar]
- 22.Dowdney L, Childhood bereavement following parental death: annotation. Journal of Child Psychology and Psychiatry and Allied Disciplines, 2000. 41(7): p. 819–830. [PubMed] [Google Scholar]
- 23.Sood A, et al. , Children’s reactions to parental and sibling death. Current Psychiatry Reports, 2006. 8(2): p. 115–120. [DOI] [PubMed] [Google Scholar]
- 24.Schultz K, Bereaved children. Canadian Family Physician, 1999. 45: p. 2914–2921. [PMC free article] [PubMed] [Google Scholar]
- 25.Brent D, et al. , The Incidence and Course of Depression in Bereaved Youth 21 Months After the Loss of a Parent to Suicide, Accident, or Sudden Natural Death. American Journal of Psychiatry, 2009. 166(7): p. 786–794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kranzler EM, et al. , Early Childhood Bereavement. Journal of the American Academy of Child & Adolescent Psychiatry, 1990. 29(4): p. 513–520. [DOI] [PubMed] [Google Scholar]
- 27.Kaplow JB, Howell KH, and Layne CM, Do Circumstances of the Death Matter? Identifying Socioenvironmental Risks for Grief-Related Psychopathology in Bereaved Youth. Journal of Traumatic Stress, 2014. 27(1): p. 42–49. [DOI] [PubMed] [Google Scholar]
- 28.Brown OK and Symons DK, “My Pet Has Passed”: Relations of Adult Attachment Styles and Current Feelings of Grief and Trauma after the Event. Death Studies, 2015. 40(4). [DOI] [PubMed] [Google Scholar]
- 29.Kimura Y, Kawabata H, and Maezawatype M: Full Paper, Frequency of Neurotic Symptoms Shortly after the Death of a Pet. The Journal of Veterinary Medical Science, 2014. 76(4): p. 499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Hunt M, Al-Awadi H, and Johnson M, Psychological Sequelae of Pet Loss Following Hurricane Katrina. Anthrozoös, 2008. 21(2): p. 109–121. [Google Scholar]
- 31.Adrian J, Deliramich A, and Frueh B, Complicated grief and posttraumatic stress disorder in humans’ response to the death of pets/animals. Bulletin of the Menninger Clinic, 2009. 73(3): p. 176–187. [DOI] [PubMed] [Google Scholar]
- 32.Sharkin BS and Knox D, Pet Loss: Issues and Implications for the Psychologist. Professional Psychology: Research and Practice, 2003. 34(4): p. 414–421. [Google Scholar]
- 33.Kaufman K and Kaufman N, And then the dog died. Death Studies, 2006. 30(1): p. 61–76. [DOI] [PubMed] [Google Scholar]
- 34.Jarolmen J, A Comparison of the Grief Reaction of Children and Adults: Focusing on Pet Loss and Bereavement. OMEGA — Journal of Death and Dying, 1998. 37(2): p. 133–150. [Google Scholar]
- 35.Tennyson AT, In memoriam. 1852, Boston. [Google Scholar]
- 36.Boyd A, et al. , Cohort profile: The ‘Children of the 90’s’- the index offspring of the Avon Longitudinal Study of Parents and Children. International Journal of Epidemiology, 2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Fraser A, et al. , Cohort profile: The Avon Longitudinal Study of Parents and Children: ALSPAC mothers cohort. International Journal of Epidemiology, 2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Goodman R, The strenghts and difficulties questionnaire: A research note. Journal of Child Psychology and Psychiatry, 1997. 38: p. 581–586. [DOI] [PubMed] [Google Scholar]
- 39.Goodman R, Psychometric properties of the strenghts and difficulties questionnaire. Journal of the American Academy of Child and Adolescent Psychiatry, 2001. 40: p. 1337–1345. [DOI] [PubMed] [Google Scholar]
- 40.Ezpeleta L, et al. , Psychometric properties of the strengths and Difficulties Questionnaire3–4 in 3-year-old preschoolers. Comprehensive Psychiatry, 2013. 54(3): p. 282–291. [DOI] [PubMed] [Google Scholar]
- 41.Muris P, Meesters C, and van den Berg F, The Strengths and Difficulties Questionnaire (SDQ)--further evidence for its reliability and validity in a community sample of Dutch children and adolescents. European Child & Adolescent Psychiatry, 2003. 12(1): p. 1–8. [DOI] [PubMed] [Google Scholar]
- 42.Goodman A, Lamping DL, and Ploubidis GB, When to use broader internalizing and externalizing subscales instead of the hypothesized five subscales on the strengths and difficulties questionnaire (SDQ); Data from British parents, teachers, and children. Journal of Abnormal Child Psychology, 2010. 38: p. 1179–1191. [DOI] [PubMed] [Google Scholar]
- 43.Goodman A and Goodman R, Population mean scores predict child mental disorder rates: validating SDQ prevalence estimators in Britain. Journal of Child Psychology and Psychiatry, 2011. 52(1): p. 100–8. [DOI] [PubMed] [Google Scholar]
- 44.Cox JL, Holden JM, and Sagovsky R, Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. The British Journal of Psychiatry, 1987. 150: p. 782–6. [DOI] [PubMed] [Google Scholar]
- 45.Hibbeln JR, et al. , Maternal seafood consumption in pregnancy and neurodevelopmental outcomes in childhood (ALSPAC study): An observational cohort study. The Lancet, 2007. 369(9561): p. 578–585. [DOI] [PubMed] [Google Scholar]
- 46.Suren P, et al. , Parental obesity and risk of autism spectrum disorder. Pediatrics, 2014. 133(5): p. e1128–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Oliver BR, et al. , Social Cognition and Conduct Problems: A Developmental Approach. Journal of the American Academy of Child & Adolescent Psychiatry, 2011. 50(4): p. 385–394. [DOI] [PubMed] [Google Scholar]
- 48.Parslow RA, et al. , Pet Ownership and Health in Older Adults: Findings from a Survey of 2,551 Community-Based Australians Aged 60–64. Gerontology, 2005. 51(1): p. 40–47. [DOI] [PubMed] [Google Scholar]
- 49.Mullersdorf M, et al. , Aspects of health, physical/leisure activities, work and socio-demographics associated with pet ownership in Sweden. Scandinavian Journal Of Public Health, 2010. 38(1): p. 53–63. [DOI] [PubMed] [Google Scholar]
- 50.Podsakoff PM, et al. , Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of Applied Psychology, 2003. 88(5): p. 879–903. [DOI] [PubMed] [Google Scholar]
- 51.Holt S, Buckley H, and Whelan S, The impact of exposure to domestic violence on children and young people: a review of the literature. Child Abuse & Neglect, 2008. 32(8): p. 797–810. [DOI] [PubMed] [Google Scholar]
- 52.Chilcoat HD and Breslau N, Does psychiatric history bias mothers’ reports? An application of a new analytic approach. Journal of the American Academy of Child and Adolescent Psychiatry, 1997. 36(7): p. 971–9. [DOI] [PubMed] [Google Scholar]
- 53.Ringoot AP, et al. , Parental depression and child well-being: young children’s self-reports helped addressing biases in parent reports. Journal of Clinical Epidemiology, 2015. 68(8): p. 928–38. [DOI] [PubMed] [Google Scholar]
- 54.van Buuren S, Boshuizen HC, and Knook DL, Multiple imputation of missing blood pressure covariates in survival analysis. Statistics in Medicine, 1999. 18(6): p. 681–94. [DOI] [PubMed] [Google Scholar]
- 55.van Buuren S and Groothuis-Oudshoorn K, mice: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software, 2011. 45: p. urn:issn:1548–7660. [Google Scholar]
- 56.White IR, Royston P, and Wood AM, Multiple imputation using chained equations: issues and guidance for practice. Statistics in Medicine, 2011. 30(4): p. 377–399. [DOI] [PubMed] [Google Scholar]
- 57.Rud AG and Beck AM, Companion animals in Indiana elementary schools. Anthrozoös, 2003. 16(3): p. 241–251. [Google Scholar]
- 58.Uttley C, Animal Attraction: Including Animals in Early Childhood Classrooms. YC Young Children, 2013. 68(4): p. 16–21. [Google Scholar]
- 59.Brown BH, Richards HC, and Wilson CA, Pet Bonding and Pet Bereavement Among Adolescents. Journal of Counseling & Development, 1996. 74(5): p. 505–509. [Google Scholar]
- 60.Eme RF, Sex differences in childhood psychopathology: A review. Psychological Bulletin, 1979. 86(3): p. 574. [PubMed] [Google Scholar]
- 61.Nolen-Hoeksema S and Girgus JS, The emergence of gender differences in depression during adolescence. Psychol Bull, 1994. 115(3): p. 424–443. [DOI] [PubMed] [Google Scholar]
- 62.Keiley MK, et al. , The timing of child physical maltreatment: a cross-domain growth analysis of impact on adolescent externalizing and internalizing problems. Dev Psychopathol, 2001. 13(4): p. 891–912. [PMC free article] [PubMed] [Google Scholar]
- 63.Dunn EC, et al. , What life course theoretical models best explain the relationship between exposure to childhood adversity and psychopathology symptoms: recency, accumulation, or sensitive periods? Psychological Medicine, 2018. 48(15): p. 2562–2572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Simpkin AJ, et al. , Longitudinal analysis of DNA methylation associated with birth weight and gestational age. Hum Mol Genet, 2015. 24(13): p. 3752–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Smith AD, et al. , A structured approach to hypotheses involving continuous exposures over the life course. Int J Epidemiol, 2016. 45(4): p. 1271–1279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 66.Ben-Shlomo Y and Kuh D, A life course approach to chronic disease epidemiology: conceptual models, empirical challenges, and interdisciplinary perspectives. International Journal of Epidemiology, 2002. 31: p. 285–293. [PubMed] [Google Scholar]
- 67.Evans GW, Li D, and Whipple SS, Cumulative risk and child development. Psychological Bulletin, 2013. 139(1): p. 342–396. [DOI] [PubMed] [Google Scholar]
- 68.Shanahan L, et al. , Child-, adolescent- and young adult-onset depressions: differential risk factors in development? Psychological Medicine, 2011. 41(11): p. 2265–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Dunn EC, et al. , Developmental timing of child maltreatment and symptoms of depression and suicidal ideation in young adulthood: results from the National Longitudinal Study of Adolescent Health. Depression and Anxiety, 2013. 30(10): p. 955–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 70.Teicher MH, Anderson CM, and Polcari A, Childhood maltreatment is associated with reduced volume in the hippocampal subfields CA3, dentate gyrus, and subiculum. Proceedings of the National Academy of Science, 2012. 109(9): p. E563–E572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 71.Westgarth C, et al. , Pet ownership, dog types and attachment to pets in 9–10 year old children in Liverpool, UK. BMC Veterinary Research, 2013. 9: p. 102. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.