GA indicates general anesthesia.
eAppendix. MEDLINE Search Strategy
eFigure. Risk of Bias Assessment for Randomized Clinical Trials
eTable. Risk of Bias Assessment for Observational Studies
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Chow CHT, Rizwan A, Xu R, et al. Association of Temperament With Preoperative Anxiety in Pediatric Patients Undergoing Surgery: A Systematic Review and Meta-analysis. JAMA Netw Open. 2019;2(6):e195614. doi:10.1001/jamanetworkopen.2019.5614
Preoperative anxiety is associated with poor behavioral adherence during anesthetic induction and adverse postoperative outcomes. Research suggests that temperament can affect preoperative anxiety and influence its short- and long-term effects, but these associations have not been systematically examined.
To examine the associations of temperament with preoperative anxiety in young patients undergoing surgery.
Studies from MEDLINE, Embase, CINAHL, PsycINFO, Web of Science, and the Cochrane Central Register of Controlled Trials databases were searched from database inception to June 2018.
All prospective studies reporting associations of temperament with preoperative anxiety were included. Overall, 43 of 5451 identified studies met selection criteria.
Data Extraction and Synthesis
Using the PRISMA guidelines, reviewers independently read 43 full-text articles, extracted data on eligible studies, and assessed the quality of each study. Data were pooled using the Lipsey and Wilson random-effects model.
Main Outcomes and Measures
Primary outcome was the association of temperament with preoperative anxiety in patients undergoing surgery.
A total of 23 studies, with 4527 participants aged 1 to 18 years, were included in this review. Meta-analysis of 12 studies including 1064 participants revealed that emotionality (r = 0.11; 95% CI, 0.04 to 0.19), intensity of reaction (r = 0.29; 95% CI, 0.11 to 0.46), and withdrawal (r = 0.40; 95% CI, 0.23 to 0.55) were positively associated with preoperative anxiety, whereas activity level (r = −0.23; 95% CI, −0.31 to −0.16) was negatively correlated with preoperative anxiety. Impulsivity was not significantly associated with preoperative anxiety.
Conclusions and Relevance
This systematic review and meta-analysis provided evidence suggesting that temperament may help identify pediatric patients at risk of preoperative anxiety and guide the design of prevention and intervention strategies. Future studies should continue to explore temperament and other factors influencing preoperative anxiety and their transactional effects to guide the development of precision treatment approaches and to optimize perioperative care.
Surgery can be a fearful event for many younger patients, as they face the threat of parental separation, loss of control, pain and discomfort, a strange environment, and uncertainty about the anesthetic procedure.1,2 The feelings of nervousness, worry, and tension related to an impending surgical experience have been formally recognized as preoperative anxiety,1,3 which can manifest as crying, anger, behavioral unrest, or verbal unrest.1
Nearly 5 million patients 18 years or younger in North America are at risk of developing preoperative anxiety each year.4 Preoperative anxiety is associated with important perioperative outcomes, including lengthened period of anesthetic induction and postoperative recovery.1 Higher levels of preoperative anxiety have also been associated with an increased risk of postoperative delirium, anxiety-related negative behavior changes, postoperative pain, and increased analgesia use.5-8 Given the adverse psychological and clinical implications of preoperative anxiety, identifying patients at greater risk presents a clinically important opportunity to improve their surgical experience and outcomes. Such knowledge can also help to inform a more appropriate allocation of finite hospital resources to patients who would most benefit from perioperative interventions.9
Previous research has found an association of temperament with anxiety in younger patients under stressful situations. Temperament is broadly defined as an individual’s characteristic nature or personality disposition, and it includes susceptibility to emotional stimulation, the strength and speed of response, the quality of the prevailing mood, the fluctuations and intensity of mood, and emotional regulation and reactivity.10-13 According to the diathesis-stress model, the interaction of individual vulnerability and stress leads to the development of psychopathology.14 Certain temperamental traits have been implicated as vulnerability factors for the development of psychological problems, such as anxiety and depression.15,16 Over the past 2 decades, behavioral inhibition, the tendency toward behavioral restraint and withdrawal in novel situations, has been widely studied and is thought to be an important risk factor for anxiety disorders.17-19 Meanwhile, negative affectivity or neuroticism, a temperamental sensitivity to negative stimuli, have also been implicated as a risk factor in the development of internalizing disorders, such as anxiety and depression.15,20 Thus, individual differences in temperament may provide important insights into which patients may fail to cope with or successfully manage challenging or novel situations (eg, surgery) and, as such, may be more likely to experience an elevated stress response.
Given the plethora of adverse outcomes associated with preoperative anxiety, the number of studies on temperament factors of preoperative anxiety has increased over the past 3 decades. However, these studies often appear in journals in disparate disciplines (eg, pediatrics, anesthesia, surgery, psychology) and yield conflicting results. For instance, some evidence has suggested that shy or inhibited patients may be at a greater risk of preoperative anxiety,9 whereas others have suggested that intensity of response, withdrawal, or low activity are risk factors.21-23 The growing extant literature on this topic coupled with a lack of consolidation or consensus among studies highlight the need for a systematic literature synthesis that provides an overview of the current state of knowledge of the association of temperament with preoperative anxiety.
To our knowledge, no systematic review has both qualitatively and quantitatively synthesized the available literature on associations of temperament with preoperative anxiety. Accordingly, we conducted a systematic review and meta-analysis of existing evidence to determine whether temperament is associated with preoperative anxiety in pediatric patients undergoing surgery under general anesthesia. These results could have important implications for the screening and identification of patients most at risk of preoperative anxiety while also helping to inform and guide the design of individualized prevention or intervention strategies.
A protocol for this systematic review was registered on the PROSPERO international prospective register of systematic reviews (CRD42016038028).24 Both narrative and meta-analytic approaches (Preferred Reporting Items for Systematic Reviews and Meta-analyses [PRISMA] reporting guideline) were used to synthesize and analyze the data.25
The research question for this systematic review was generated using the population, intervention (exposure), comparison, outcome, and study design approach. Study eligibility criteria as well as inclusion and exclusion criteria were also established using this framework. Prospective studies (ie, randomized clinical trials [RCTs], nonrandomized clinical trials, and observational study designs) that measured temperament before surgery were eligible for review.
The population of interest was patients aged 1 to 18 years undergoing surgery under general anesthesia at research, community, or university-affiliated hospitals. Only studies that measured temperament using validated scales (eg, Emotionality Activity Sociability Impulsivity [EASI] Temperament Scale) were eligible for inclusion. The outcome of interest was preoperative anxiety in patients undergoing surgery, as measured using validated anxiety scales (eg, the modified Yale Preoperative Anxiety Scale).
A search strategy was developed after consultation with a librarian at McMaster University. Systematic searches were conducted on articles published from database inception to June 2018 using 6 databases: MEDLINE, Embase, CINAHL, PsycINFO, Web of Science, and the Cochrane Central Register of Controlled Trials. No language restriction was applied. The search strategy used medical subject heading terms, which were combined with keywords if necessary (eAppendix in the Supplement). Reference lists were individually searched, and the results were included in this review.
Two of us (C.H.T.C. and A.R.) independently screened titles and abstracts (κ = 75%). After screening, the review authors met and selected studies eligible for full-text screening. A third author (L.A.S.) was consulted to resolve disagreements.
A data extraction form was developed and piloted on 2 randomly selected studies included in the review. The information extracted from each study included study characteristics, population characteristics, details of the exposure, outcomes, summary of results, and risk of bias assessments. Risk of bias for RCTs and observational studies was assessed at the study level using the Cochrane Collaboration risk of bias tool and the Newcastle-Ottawa Scale, respectively.
A minimum of 2 studies was required for meta-analysis. The bivariate correlations (ie, Pearson correlation coefficients [r]) of child temperament with preoperative anxiety reported in available studies were used in meta-analyses. Using random-effects models of Lipsey and Wilson, we first converted all the correlation coefficients, r, for each study to a common metric using Fisher z transformations.26,27 The results were interpreted as significant if r did not cross the 0 line. We calculated the mean of z, inverse-variance weight, standard error of the mean of z, and z test for mean of z. We calculated 95% CIs using these formulas: lower = ES − 1.96(seES) and upper = ES + 1.96(seES), where ES indicates effect size and SE indicates standard error. Mean ESs were converted back to r for interpretation. Sensitivity analyses were performed on experimental studies for each temperament dimension to examine whether interventional effects were influencing the results. Cohen d criteria were used as a guideline for interpreting the size of mean ES: small, r = 0.10 to 0.29; medium, r = 0.30 to 0.49; and large, r = 0.50 or greater.28 Statistical analyses were conducted in Excel (Microsoft Corp).
We identified 23 eligible studies (19 cohort studies and 4 RCTs) (Figure 1). A total of 4527 participants aged 1 to 18 years were included. Most studies were conducted in the United States (13 [57%]), followed by Canada (5 [22%]), Portugal (3 [13%]), Australia (1 [4%]), and the United Kingdom (1 [4%]).
Within and across studies, all 4 RCTs demonstrated moderate to high risk of bias; 2 did not describe masking of participants and outcome in sufficient detail (eFigure in the Supplement). The overall Newcastle-Ottawa Scale scores on the 19 observational studies ranged from 5 to 7 (maximum score, 9). The use of self-report and lack of description in ascertainment of exposure were the most common sources of bias in the studies (eTable in the Supplement).
Data were available on 12 studies for meta-analysis and were pooled for 1064 unique participants.5,21,23,29-37 The included studies reported on the following temperamental traits: activity, emotionality, sociability, shyness, impulsivity, withdrawal, and intensity of reaction.
Activity is defined as the degree of energy expenditure through movement.38 The weighted average correlation from 7 studies5,29-32,34,37 (combined participants, 583) of the negative association of activity with preoperative anxiety was statistically significant with a small ES (r = −0.23; 95% CI, −0.31 to −0.16). Individual ESs ranged from −0.39 to −0.09. The negative correlation suggested that patients who scored as less active exhibited higher preoperative anxiety.
Emotionality is defined as the tendency to become easily and intensely upset.39 The weighted average correlation from 7 studies23,30-33,35,37 (combined participants, 680) of the association of emotionality with preoperative anxiety was statistically significant but had a small ES (r = 0.11; 95% CI, 0.04-0.19). Individual ESs ranged from −0.04 to 0.25. Overall, patients who scored higher on emotionality exhibited higher preoperative anxiety.
Sociability is the tendency to seek social interactions.40 Among the 5 studies30,31,34,36,37 (combined participants, 338) that measured the negative association of sociability with preoperative anxiety, the weighted average correlation was −0.10 (95% CI, −0.21 to 0.01), a small ES. Individual ESs ranged from −0.37 to 0.09. These studies showed that patients who scored as less social exhibited higher preoperative anxiety.
Shyness is defined as the tendency to avoid social interactions or situations.41 Among the 3 studies30-32 (combined participants, 285) that measured the association of shyness with preoperative anxiety, the weighted average correlation was 0.10 (95% CI, −0.02 to 0.22).30-32 Individual ESs ranged from 0.08 to 0.13. These studies indicated that patients who scored as more shy exhibited higher preoperative anxiety.
Impulsivity is defined as the predisposition toward rapid, unplanned reactions to internal or external stimuli with diminished regard to the negative consequences of these reactions to the individual or to others.42 The weighted average correlation from 3 studies23,33,37 (combined participants, 133) of the association of impulsivity with preoperative anxiety was not significant (r = −0.01; 95% CI, −0.19 to 0.17). Individual ESs ranged from −0.26 to 0.44. This result suggested that higher impulsivity may not impart a greater risk of preoperative anxiety.
Withdrawal is defined as the tendency to retreat from novel situations and people.43 The weighted average correlation from 2 studies21,29 (combined participants, 110) of the association of withdrawal with preoperative anxiety was statistically significant, with a medium ES (r = 0.40; 95% CI, 0.23-0.55). Individual ESs ranged from 0.29 to 0.60. These studies suggested that patients who were more withdrawn exhibited higher preoperative anxiety.
Intensity of reaction is defined as the typical strength of an individual’s responsiveness to a situation.44 The weighted average correlation from 2 studies21,29 (combined participants, 110) of the association of intensity of reaction with preoperative anxiety was statistically significant with a small ES (r = 0.29; 95% CI, 0.11-0.46). Individual ESs ranged from 0.27 to 0.33. These studies suggested that patients who had higher intensity of reaction exhibited higher preoperative anxiety.
Sensitivity analyses were performed to assess whether experimental and observational studies generated different findings, and the results showed that the directionality of the weighted average correlation for certain temperament dimensions remained robust: sociability (number of studies, 2; r = −0.13; 95% CI, −0.31 to 0.06); impulsivity (number of studies, 2; r = −0.15; 95% CI, −0.36 to 0.08); and activity (number of studies, 3; r = −0.32; 95% CI, −0.42 to −0.21). Representative forest plots are shown in Figure 2 and Figure 3.
Overall, 3 studies22,45,46 reported the associations of activity with preoperative anxiety. Activity was found to be associated with higher anxiety in the preoperative holding area. A 1996 study22 found that activity also interacted with parental presence, and this interaction was associated with greater preoperative anxiety at anesthetic induction (Tukey test, 2.54; P = .01; R2 = 0.15).
Of the 23 studies, only 1 study45 reported emotionality to be associated with greater preoperative anxiety. It found significant associations of emotionality with preoperative anxiety in the preoperative holding area and on separation at the operating room.
Overall, 1 study reported on low sociability.47 It found low sociability to be associated with greater preoperative anxiety (β = −0.57; SE, 0.21; P = .007).
Of the 23 studies, 2 studies48,49 reported the associations of shyness with preoperative anxiety. In Quinonez et al,49 shyness was significantly associated with anxiety during preseparation (R2 = 0.16; F = 9.23; df = 49; P = .003) and during separation from the parent at the entrance of the operating room (R2 = 0.10; F = 5.12; df = 49; P = .03). In a 2017 study,48 temperamental shyness was found to be associated with lower anxiety during the preoperative clinic visit (β = −10.78; P = .03) and in the holding area on the day of surgery (β = −12.31; P = .03).
In a 2004 study,7 patients at high risk, defined by exhibiting preoperative anxiety and postoperative maladaptive behavioral changes, were found to be more active, more emotional, and less sociable than patients in the low-risk group. A 2011 study50 also reported that greater preoperative anxiety was associated with internalizing behavior (F1,47 = 4.5; P = .04), somatic complaints (F1,49 = 4.0; P = .05), and fear (F1,50 = 5.2; P = .03). A 2006 study51 reported that patients with less anxiety scored lower on activity and impulsivity and that activity was associated with anxiety at anesthetic induction when the parent was present (R2 change = 0.016; P = .007).
In contrast, a 2006 study46 reported no difference in temperament styles between patients in the high-anxiety group vs patients in the low-anxiety group. Two other studies52,53 also reported nonsignificant associations of EASI temperamental dimensions with anxiety as well as easy vs difficult temperament with preoperative anxiety.
Study characteristics are summarized in Table 1. The scales used, the number of assessments and their time points, and outcome summaries for each study appear in Table 2.
To our knowledge, this is the first systematic review and meta-analysis to examine the association of temperament with preoperative anxiety among pediatric patients. This review provided evidence that individual differences in temperament may help identify young patients at risk of preoperative anxiety and guide the design of future prevention and intervention strategies. This review included 23 studies (observational and experimental), involving 4527 participants aged 1 to 18 years undergoing elective same-day surgery. The meta-analytic results of 12 pooled studies revealed that certain temperament styles were significantly associated with preoperative anxiety. Specifically, emotionality, intensity of reaction, and withdrawal were found to be associated with increased preoperative anxiety, whereas activity level was associated with less anxiety. The ESs ranged from small to medium. Impulsivity was not associated with preoperative anxiety.
Our findings are consistent with previous research investigating the association of temperament with psychopathology in children and youth in other stressful, nonclinical contexts54,55 and extends this work by examining these associations in the surgical setting. Importantly, this broadens our understanding of the development of anxiety in a distinct clinical context. Our results suggest that both negative emotionality (small ES) and high intensity of reaction (small ES) were associated with preoperative anxiety. These findings are congruent with previous longitudinal studies, in which negative emotionality in infancy and middle childhood were found to be associated with anxious behaviors 2 years later56 as well as with anxiety symptoms in adulthood.57,58 The results of this meta-analysis are also consistent with a 2007 study,59 which found that negative emotionality was associated with dental fear, as were shyness and activity.
We also noted the association of inhibited temperament (ie, shyness, withdrawal behaviors) with preoperative anxiety, consistent with previous literature examining this association in everyday normative contexts, such as school and home.60-64 Our findings also suggested a negative association of sociability with preoperative anxiety. We found that shyness and sociability exerted small ESs, while withdrawal behaviors exerted medium ESs. The modest ES magnitudes can be understood in the context of research indicating that stronger effects are only seen when inhibited temperament in early childhood is combined with other risk factors, such as parental factors or psychophysiological reactivity.65 Thus, temperamental traits should be examined within a biopsychosocial framework, which includes both biological factors (ie, age, sex, or physiological reactivity) and environmental moderating factors (ie, socioeconomic status, previous surgical experiences, or parental anxiety) to best predict preoperative anxiety and guide future directions for tailored, individualized approaches to managing preoperative anxiety.
In terms of activity level, our results showed that low activity (small ES) was associated with higher preoperative anxiety in patients undergoing surgery. This result is consistent with a longitudinal study58 that showed negative associations of activity levels with anxious behaviors at ages 4 years and 8 to 9 years. Finally, the association of impulsivity with preoperative anxiety in patients undergoing surgery was not significant. This might be explained by the fact that impulsivity is commonly associated with externalizing behaviors, such as aggression, delinquency, and hyperactivity, but not internalizing problems, such as anxiety.66 This is further supported by studies that showed patients with externalizing problems were more impulsive than patients with internalizing problems.67-69
Taken together, patients with negative emotionality and/or inhibited temperaments seem more prone to experiencing preoperative anxiety. Patients with behavioral inhibition or a so-called difficult temperament (eg, negative mood, slow to adapt to new situations) are reported to be at a heightened risk of developing anxiety disorders later in life. Particularly, a difficult temperament was identified as the single most important risk factor for heightened anxiety symptoms.70 This can be further explained by the view that inhibited temperament reflects low temperamental behavioral reactivity.69 Patients who are behaviorally inhibited appear to be more rigid and inflexible in novel or stressful contexts,68 and this inability to adapt may predispose them to greater anxiety in an unfamiliar and stress-inducing environment, like the surgical setting.
The present review has a number of important clinical implications. Its findings contribute to the body of evidence supporting the relevance of temperament in the development and/or maintenance of anxiety. This review provides support that certain temperamental traits (ie, emotionality and withdrawal) might be risk factors for preoperative anxiety and may predict how patients will respond in this unique and stressful setting. This knowledge can be used to help with refinement of screening processes and prevention strategies for preoperative anxiety and to design interventions (eg, improving emotional regulation and coping skills). As temperament represents only a single risk factor, future research should continue to study these individual-level characteristics with other individual-level (eg, psychophysiological reactivity) and family-level (eg, parental behaviors) factors to develop more holistic prognostic models with greater predictive potential.
Several limitations should be noted. First, none of the reviewed studies examined the associations of temperament with postoperative outcomes (eg, pain, emergence delirium, recovery) beyond anxiety. Second, quantitative correlational data for the meta-analysis were only available for a limited number of studies (12), as some of the studies reviewed did not report usable ES statistics for temperament and anxiety. However, the narrative summary from these studies generally showed congruent results. Third, only 2 or 3 studies examined certain temperament dimensions. Fourth, study designs were variable (ie, observational vs experimental). However, sensitivity analyses revealed that most results remained robust when these variations were accounted for, with some demonstrating even stronger effects, whereas others were attenuated owing to a lack of statistical power. Fifth, the included studies analyzed only a subset of the various temperament dimensions that have been implicated in anxiety and/or internalizing disorders. Other temperament traits that are associated with psychopathology, such as surgency, should be considered in future work.67 Sixth, subjective reports on temperament and/or anxiety (eg, by parents, patients, or research staff) are prone to reporting bias and interobserver bias. Seventh, the standards for the collection of temperament and anxiety data (eg, measurement or timing) have not been established and varied across studies. In this review, the EASI Temperament Scale was the primary temperament measure in 18 of 23 studies. Although the EASI Temperament Scale has been widely used as a measure of temperament in the literature, a 2017 systematic review conducted by Walker et al71 suggested that the EASI Temperament Scale may have inconsistent psychometric properties with variable internal consistency and poor factor structure. Thus, future studies should use different, more psychometrically sound measures of temperament and/or include a modified version of the EASI Temperament Scale to improve the reliability and validity of results. Future studies should also take into consideration the timing of temperament measures, as the concurrent assessments of temperament and preoperative anxiety might result in inflated estimates of ESs associating temperament with anxiety. Although the findings of this review are informative, future studies should address the aforementioned limitations in design and data collection to provide more definitive and robust conclusions that can guide future clinical practice.
Our systematic review and meta-analysis suggests that temperament styles are significantly associated with preoperative anxiety for young patients undergoing surgery. The findings showed that patients with negative emotionality and inhibited temperament are more prone to experiencing preoperative anxiety, whereas active and social patients are less likely to experience preoperative anxiety. Future studies should continue delineating the role of temperament with other biological and environmental determinants of preoperative anxiety and their transactional effects by using more standardized measures, such as behavioral observations or noninvasive physiological measures (eg, cortisol or electrocortical activity). Furthermore, future studies should examine the association of temperament with other perioperative outcomes that are of significance to patients, families, and practitioners, such as postoperative pain, emergence delirium, and postoperative maladaptive behaviors, to advance precision medicine approaches in perioperative management. Given the negative impact of preoperative anxiety, identifying etiological factors that may predict its emergence can help to guide the design of future detection, prevention, and individualized management strategies aimed at reducing the adverse effects of preoperative anxiety.
Accepted for Publication: April 26, 2019.
Published: June 7, 2019. doi:10.1001/jamanetworkopen.2019.5614
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Chow CHT et al. JAMA Network Open.
Corresponding Author: Cheryl H. T. Chow, PhD, Department of Psychology, Neuroscience, and Behaviour, McMaster University, 1280 Main St W, PC 135, Hamilton, ON L8S 4K1, Canada (email@example.com).
Author Contributions: Drs Chow and Schmidt had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Chow, Rizwan, Van Lieshout, Buckley, Schmidt.
Acquisition, analysis, or interpretation of data: Chow, Rizwan, Xu, Poulin, Bhardwaj, Schmidt.
Drafting of the manuscript: Chow, Rizwan, Xu, Bhardwaj, Van Lieshout, Schmidt.
Critical revision of the manuscript for important intellectual content: Chow, Rizwan, Xu, Poulin, Van Lieshout, Buckley, Schmidt.
Statistical analysis: Chow, Rizwan, Schmidt.
Obtained funding: Schmidt.
Administrative, technical, or material support: Xu, Bhardwaj, Van Lieshout, Buckley, Schmidt.
Supervision: Chow, Van Lieshout, Buckley, Schmidt.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by an Ontario Mental Health Foundation doctoral scholarship awarded to Dr Chow and an operating grant from the Canadian Institutes of Health Research (CIHR) awarded to Drs Van Lieshout, Buckley, and Schmidt.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Research librarian Stephanie Sanger, MLIS (McMaster University, Hamilton, Ontario, Canada), assisted with the search strategy. Sara Miller, MSc (Scientific Editor, Department of Anesthesia, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada), provided editing services, and Toni Tidy, HBSc (Research Coordinator, Department of Anesthesia, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada), provided unconditional support. None were compensated for their time.
Create a personal account or sign in to: