Perioperative anxiety is a complex combination of fear, apprehension, and worry often accompanied by physical sensations.1 Even minor surgery can be a frightening experience for children.2–4 Anxiety influences patients’ subjective perceptions,1 and preoperative anxiety is associated with higher level of postoperative pain.5–7
Interest in children’s anxiety has increased in recent years, and trials have investigated the best way to measure children’s anxiety. The standard Spielberger State-trait Anxiety Inventory (STAI)8 is used most frequently in the literature. Because, this gold standard tool is not appropriate for evaluating anxiety in a busy operating room setting,9 the modified Yale Preoperative Anxiety Scale (m-YPAS) has been developed to assess anxiety for children aged 2–12 yr undergoing surgery. In the postoperative period, the Post-Hospital Behavioral Questionnaire10,11 has been proven to be a valid and robust instrument for the assessment of behavioral change in children, but it is not recognized as a measure of anxiety.
Anxiety evaluation is essential for the clinical perioperative follow-up of children, but the multiplicity of instruments depending on the age of the child or on the period of the evaluation is limited in clinical practice. A useful and unique anxiety instrument may have multiple benefits in following anxiety throughout the hospital stay.
Furthermore, in the early postoperative period, pain can be a confusion factor for a scale based on behavioral aspects. Self-report measurements could be a more effective solution. Crandall et al.12 have demonstrated the validity of numeric 0–10 anxiety self-report scale to measure preoperative anxiety in children 7–13 yr. Because anxiety evaluation is also important postoperatively to optimize children’s pain management, the aims of this study were to evaluate the perioperative validity of the visual analog scale (VAS)-anxiety in children and evaluate its clinical relevance for pain management.
This was a longitudinal observational study. Approval was obtained from the Montpellier Institutional Review Board. Children were selected from the Children’s Hospital Surgery Center, Montpellier University Hospital (France). Children of ASA status I, II, or III, aged 7–16 yr who were undergoing elective surgery and general anesthesia were enrolled from September 2005 to June 2006. Patients were excluded if they had a diagnosis of mental retardation or chronic pain. Outpatient surgery was not included in the study.
Study information was given the day before surgery on arrival at the hospital. Parents’ written and children’s oral consents were obtained, and demographic information was collected. Anxiety was self-reported using a VAS at four timepoints (Table 1): the evening before surgery (D-1), the day of surgery before anesthesia (D0), postoperative day one (POD1), and the day of discharge (DD). Parents blinded to their child’s responses evaluated the child’s anxiety and their own anxiety using VAS. The VAS-anxiety consists of a 100-mm horizontal line with the two end points labeled “no anxiety or fear” and “worst possible anxiety or fear.” The patient is required to show the point that corresponds to their level of anxiety at that moment. VAS has been validated for adults in a previous study.13 It is a pertinent tool to assess parents’ anxiety.14,15
Pain intensity was also self-reported on a 100-mm VAS (VAS-pain).16 Children also completed the state version of the STAI8,17 adapted for age. This instrument is a self-administered questionnaire with two separate 20-question rating scales to evaluate “trait” and “state” anxiety. Only the state items were used. The pediatric version (State-Trait Anxiety Inventory for Children [STAIC]) was administered to children younger than 12 yr. The youth version (State-Trait Anxiety Inventory for Youth [STAIY]) was administered to the older children. The scores for all items were summed to create the total score. Research assistants were told to read the STAI questionnaire to young children and explain items that the children found difficult to understand. It was not administrated in the operating room before induction (D0) because it takes more than 5 min to answer. At this time, observational state anxiety was assessed using the modified m-YPAS, which is a behavioral scale developed for assessing state preoperative anxiety in children aged 7–12 yr.9 It consists of 27 items in five dimensions (activity, emotional, expressivity, state of arousal, vocalization, and use of parents). The item “Use of parents” was not included in the total score because the parents do not enter the operating room. The score is based on the sum of partial weights for each category. The psychologist who completed the m-YPAS was trained in a previous study using videotapes of children before anesthesia. The reliability levels were excellent. This research did not modify the management of the patient in any way and did not impose a protocol of sedative premedication and/or anesthesia.
Children aged 7–11 yr were in the youngest group, and children aged 12 yr and older were in the oldest group. The sample size was based on the correlation between children’s anxiety assessed by the m-YPAS before induction in the youngest group. Given a correlation r = 0.50 between VAS-anxiety and m-YPAS, a power of 90% and a two-sided α level of 0.05, 37 subjects were needed. This study also included children aged 12 years and older, so we decided to increase the total number of subjects: 100 children were needed to complete this study.
Continuous data are expressed as mean ± sd or median for non-Gaussian variables. Categorical data are expressed as frequencies (%). Continuous variables were compared with Student’s t-test or the Mann–Whitney U-test for the non-Gaussian variables. Categorical variables were compared with the χ2 test.
Changes over time in the two age groups were analyzed by time × group interaction in a repeated-measures analysis of variance. The convergent validity was evaluated by the analysis of the association of VAS-anxiety with other validated measures of the same construct (m-YPAS, STAIC, and STAIY). The discriminant validity measures that association with measures of different constructs (VAS-pain). Convergent and discriminant validity were evaluated by Pearson correlations.
Because m-YPAS is a valid tool to assess preoperative anxiety in children, sensitivity, specificity, and positive and negative predictive values were examined for different cutoff points of VAS-anxiety. A receiver operating characteristic (ROC) determined VAS-anxiety cutoff using m-YPAS with 30 as the reference point of high-level anxiety.9 This cutoff has been used to differentiate groups of high anxiety (≥30/100) of children and parents.
Pearson correlation was used to investigate the relationship among parents’ VAS-anxiety and both self and proxy reports of children’s anxiety. The median differences were tested by a paired comparison.
To study the influence of both children’s and parents’ anxiety on children’s postoperative pain, groups of anxiety were determined in each period using the cutoff resulting from the ROC analysis. Postoperative pain differences between groups of anxiety were tested using Student’s t-test.
Data were analyzed using the SAS package Version 9 software (Carry, CA).
One hundred children were enrolled in this study. Children’s characteristics are reported in Table 2. No statistically significant differences were reported between the age groups. All children were able to use VAS-anxiety. In the preoperative period, VAS-anxiety did not differ significantly among history of previous surgery, gender, age, ASA status, and type of surgery. Postoperatively, VAS-anxiety did not differ with the type of surgery (POD1: P = 0.42; DD: P60).
The VAS-anxiety scale changed over time from the first day of hospital stay to the DD (P < 0.001) (Fig. 1). The highest anxiety level was on the day of surgery before induction when 50% of the children scored higher than 30 for their anxiety. Children’s anxiety decreased in the postoperative period but was still present. The difference between VAS before induction and the day after surgery was significant (P = 0.003) (Fig. 1). The analysis by time × group interaction reported that VAS-anxiety did not differ significantly between the two age groups (P = 0.24) from the first day of hospitalization to the DD (Fig. 2). VAS-anxiety changed over time in the two age groups (P = 0.04) (Fig. 2)
Table 3 describes the concurrent validity of VAS-anxiety. Before induction, VAS-anxiety correlated significantly with m-YPAS (r = 0.67, P < 0.001) in children younger than 12 yr. We found significant correlations between VAS and STAIY on the first day of hospitalization, on the POD1, and on the DD (r = 0.67, P < 0.001; r = 0.63, P < 0.001; r = 0.62, P = 0.002). The correlation between VAS and STAIC was only significant on the DD (r = 0.66, P < 0.001). For each period, the intercept between VAS-anxiety and two versions of STAI differed significantly from 0. These significant intercepts seem to be due from the range of STAI. Moreover, changes over time were not significant in the youngest age group using STAIC (P > 0.05) (Fig. 3a), whereas it was significant in the oldest group using STAIY (P < 0.001) (Fig. 3b).
Table 4 shows the sensitivity, specificity, and positive and negative predictive values for different cutoff points of VAS-anxiety. A ROC curve was generated to determine the cutoff of VAS-anxiety (Fig. 4). The area under the curve was 0.80 (CI: 0.69–0.91, P < 0.001). A score of 30 maximized the sum of the sensitivity and specificity and was considered clinically relevant. A VAS-anxiety score of 30 or more was used to identify the high-anxiety groups of children and parents. The study of parental assessment showed a significant correlation between the children’s self-report and the parents’ proxy report (r = 0.72, P = <0.001; r = 0.67, P = <0.001; r = 0.84, P = <0.001). Intercepts when parents versus children assess the children’s anxiety differed significantly from 0 on the day before surgery (P = 0.0005) and the day after surgery (P = 0.0284). The difference (d = 1.04; P = 0.002) between the self-reports and the proxy reports is significant (Fig. 5) on the day before surgery. There was no significant correlation between parents’ preoperative anxiety and child’s preinduction anxiety. Nevertheless, the proxy report was significantly higher when the parents were anxious (VAS ≥30) in the three periods (P = 0.008; P = 0.001; and P = 0.0005) (Fig. 6). The coefficient of correlation between self-reports of children’s anxiety and self-reports of parents’ anxiety is not significant on the day before surgery (r = 0.18, P = 0.19). In the postoperative period, the coefficient of correlation is significant but low (POD1: r = 0.44, P = 0.002; DD: r = 0.53, P = 0.001). The mean differences are significant in the three periods (P = 0.00001; P = 0.04; and P = 0.03) (Fig. 5). Moreover, the children’s anxiety levels were significantly higher when the parents were anxious in the postoperative period (POD1: P = 0.002; DD: P = 0.002) (Fig. 6).
A significant but moderate correlation coefficient was found between pain and anxiety levels in the POD1 (r = 0.37, P = 0.0015) and the DD (r = 0.52, P < 0.001). A significant difference was found between VAS-pain and VAS-anxiety (P < 0.05).
Postoperative pain did not differ significantly in the preoperative anxiety groups (POD1: P = 0.47; DD: P = 0.61). Nevertheless, simultaneous assessment of both pain and anxiety showed that the group of anxious children had a significantly higher level of pain in the postoperative period (POD1: P = 0.04; DD: P = 0.01) (Fig. 7a). The children’s pain was significantly higher in the group with anxious parents on the day after surgery (POD1: P = 0.02) (Fig. 7b).
This study clearly demonstrates that VAS-anxiety is a useful method and has good psychometric properties to assess children’s anxiety in perioperative periods. Although VAS-pain is the universal tool to assess pain in older children, the choice of children’s anxiety scales was limited by the period of assessment or children’s age. Today, following up children’s anxiety from the first day of hospital stay to the DD has been difficult with a unique scale. In addition, children’s anxiety studies suffered from a reduced age span. This study shows that VAS-anxiety is a valid tool and sensitive to change over time, pain, and parents’ anxiety.
Because anxiety influences pain levels, children’s anxiety must be assessed in clinical practice.5–7 Optimal anxiety management requires a statistically valid perioperative scale. The first step of our validation study was to compare VAS-anxiety with the fully recognized scales. Our results provide the concurrent validity of VAS-anxiety. Most importantly, we have demonstrated that the level of anxiety is dependent on the scale but not on age. In medical literature, STAI is a self-report instrument widely used to measure anxiety. The correlations between VAS-anxiety and STAI are significant in older children but only significant in children younger than 12 yr on the DD. These results might suggest that this difference is due to the children’s age. However, VAS-anxiety is able to detect statistically important changes over time in two age groups, whereas STAIC is not sensitive enough with the youngest children. Finally, the lower correlation between STAIC and VAS-anxiety did not limit the use of VAS-anxiety but demonstrated the difficulty of using STAIC in young children.
This result is consistent with the study of Poma et al.,18 which showed that VAS changed significantly but the Spileberger’questionnaire did not. Kain et al.9 also showed in their validation study that STAIC, which requires at least 5–10 min to complete, is not an easy tool in a busy operating room setting. Schisler et al.19 found that young children’s vocabularies may be too limited to easily understand some items of the STAIC. Furthermore, understanding the items is more difficult with premedication or after general anesthesia. The youngest age at which to use a VAS scale is also controversial. McGrath et al.20 have shown that VAS is reliable for use by children aged 5 yr and older. However, the conceptual complexity of the VAS requires the user to translate a subjective sensory experience into a linear format. Berk21 found that the ability to seriate does not appear until 7-years old. In our study, all children were able to use VAS-anxiety from the first day of hospitalization to the DD.
Although m-YPAS9 cannot be used easily in clinical practice, its use is now frequently described in the medical literature to assess young children’s anxiety before anesthetic induction. This scale has the advantage of being used from the age of 5 yr and assessing multimodal aspects of anxious behavior.
The good correlation between VAS-anxiety and m-YPAS in our study confirms the validity of VAS to detect anxiety levels before surgery. The m-YPAS scale has been chosen to study the sensitivity, specificity, and positive and negative predictive values of VAS-anxiety. The maximal sum of specificity and sensitivity showed a cutoff at 30 with 78% sensitivity and 67% specificity. In an adult study, Kindler et al.13 showed a lower sensitivity at this cutoff and preferred a low-cutoff point of VAS but accepted a higher number of false-positive scores (35.1%). We chose 30 as a threshold to detect high levels of anxiety in both children and their parents.
The last point of the analysis of the concurrent validity of VAS-anxiety is the comparison between self-report and proxy report. The statistical test most frequently used to validate a new method of assessment is Pearson correlation coefficient. The majority of studies, which compare different groups of patients, is limited by this test. However, a high correlation between scores does not systematically indicate high agreement.22,23 When using the VAS scale for both children and their parents, the difference can be tested. Although children’s anxiety and parents’ anxiety are significantly correlated, the paired comparison indicates that parents are more anxious than children in the preoperative period. Furthermore, our results show that parents overestimate the children’s own assessment. In addition, proxy-reporting scores are higher when parents are anxious. These results prove the difficulties in everyday clinical practice of perception of anxiety and demonstrate the importance of self-reporting to assess subjective measurements. Although parents’ proxy reports may be a useful alternative, children’s self-reporting is preferred whenever possible. This study proves that the VAS-anxiety detects differences between self-reporting and proxy reporting and that the self-report scale is also preferred for young children. In addition, because pain may induce bias on behavioral assessment, self-reporting of anxiety should be the primary method in the postoperative period.
An important property to validate in a new scale is also the responsiveness to change.24 Our results show that VAS-anxiety is a sensible instrument to detect changes over time for the two age groups we studied. In addition, in agreement with Caumo et al.,4 we reported that the anxiety level is significantly more important before surgery than after. This result can be explained simply by the definition of anxiety, which is a future-oriented emotion characterized by an apprehensive anticipation threat.25 The surgical act or the anesthesia constitutes a real threat for children.
Postoperative anxiety is reduced, but its assessment is still important because postoperative pain interferes in this experience. The significant correlation between VAS-pain and VAS-anxiety shows the relationship between these anxiety and pain, but the moderate correlation and the significant difference proved that children are able to differently assess anxiety and pain. Because cries and agitation in the postoperative period are common behavior of pain and anxiety, the main difficulty with proxy reporting of postoperative pain is to discriminate between pain and anxiety.26 Using VAS, children were able to differently assess pain and anxiety. This result highlights the importance of self-reporting, especially postoperatively.
Our study did not prove that a high score for VAS-anxiety in the preoperative period constitutes a predictor for postoperative symptoms. However, simultaneous assessment of both pain and anxiety showed that children with a high level of anxiety had a significantly higher level of pain. This result confirms that VAS-anxiety assesses state and not trait anxiety and the relevance of anxiety measures in the postoperative period.
Some authors have studied the presence of the parents to facilitate induction of anesthesia. Because parents could transfer anxiety to the child, their presence rarely seems beneficial.27 Kain et al.28 clearly reported the superiority of midazolam premedication over untrained parents. In the postoperative period, we reported that the children of the anxious parents’ group were significantly more anxious and that those children had the highest pain levels. In agreement with Kotiniemi et al.,11 our study indicates that children’s and their parents’ problems are not over when they leave the hospital. We demonstrate that anxiety is almost always present on the DD and that the optimal management of postoperative symptoms cannot be reduced to pain assessment. Our validation of VAS-anxiety constitutes an opportunity to optimize children’s pain management, and this useful instrument should be recommended for clinical practice.
Several limitations should be noted in our study. First, it would be interesting to evaluate the concurrent validity of VAS-anxiety in postanesthesia care unit, but there is no gold standard to assess anxiety in the postanesthesia care unit, and it was impossible to administer 20 items in this period. The second concern is the lack of use of the test–retest reliability coefficients. This important test shows the temporal stability property of a scale. Because anxiety level can change rapidly,29 this test would not be appropriate in this context. Further, randomized trials are needed to test whether VAS-anxiety is sensitive to postoperative treatment and to analyze its clinical pertinence using confounding variables. Furthermore, further studies should include outpatient surgery.
In conclusion, VAS-anxiety is a useful and valid tool to assess perioperative anxiety in children 7 yr and older. Furthermore, our study demonstrated that VAS-anxiety should also be measured in the postoperative period in routine clinical practice to optimize anxiety and pain management.
1. Spielberger C. Anxiety as an emotional state. Anxiety: current trends in theory and research. Vol. 1. New York: Academic Press, 1972
2. Wollin SR, Plummer JL, Owen H, Hawkins RM, Materazzo F, Morrison V. Anxiety in children having elective surgery. J Pediatr Nurs 2004;19:128–32
3. Ben-Amitay G, Kosov I, Reiss A, Toren P, Yoran-Hegesh R, Kotler M, Mozes T. Is elective surgery traumatic for children and their parents? J Paediatr Child Health 2006;42:618–24
4. Caumo W, Broenstrub JC, Fialho L, Petry SM, Brathwait O, Bandeira D, Loguercio A, Ferreira MB. Risk factors for postoperative anxiety in children. Acta Anaesthesiol Scand 2000;44:782–9
5. Lamontagne LL, Hepworth JT, Salisbury MH. Anxiety and postoperative pain in children who undergo major orthopedic surgery. Appl Nurs Res 2001;14:119–24
6. de Groot KI, Boeke S, van den Berge HJ, Duivenvoorden HJ, Bonke B, Passchier J. The influence of psychological variables on postoperative anxiety and physical complaints in patients undergoing lumbar surgery. Pain 1997;69:19–25
7. Kain ZN, Mayes LC, Caldwell-Andrews AA, Karas DE, McClain BC. Preoperative anxiety, postoperative pain, and behavioral recovery in young children undergoing surgery. Pediatrics 2006;118:651–8
8. Spielberger C. Manual for the state-trait anxiety inventory (form Y). Palo Alto, CA: Consulting Psychologists Press, 1983
9. Kain ZN, Mayes LC, Cicchetti DV, Bagnall AL, Finley JD, Hofstadter MB. The yale preoperative anxiety scale: how does it compare with a “gold standard”? Anesth Analg 1997;85:783–8
10. Vernon DT, Schulman JL, Foley JM. Changes in children’s behavior after hospitalization. Some dimensions of response and their correlates. Am J Dis Child 1966;111:581–93
11. Kotiniemi LH, Ryhanen PT, Moilanen IK. Behavioural changes in children following day-case surgery: a 4-week follow-up of 551 children. Anaesthesia 1997;52:970–6
12. Crandall M, Lammers C, Senders C, Savedra M, Braun JV. Initial validation of a numeric zero to ten scale to measure children’s state anxiety. Anesth Analg 2007;105:1250–3, table of contents
13. Kindler CH, Harms C, Amsler F, Ihde-Scholl T, Scheidegger D. The visual analog scale allows effective measurement of preoperative anxiety and detection of patients’ anesthetic concerns. Anesth Analg 2000;90:706–12
14. Chlan LL. Relationship between two anxiety instruments in patients receiving mechanical ventilatory support. J Adv Nurs 2004;48:493–9
15. Davey HM, Barratt AL, Butow PN, Deeks JJ. A one-item question with a Likert or Visual Analog Scale adequately measured current anxiety. J Clin Epidemiol 2007;60:356–60
16. Huskisson E. Measurement of pain. Lancet 1974;2:1127–31
17. Spielberger C. Manual for the state-trait anxiety inventory for children. Palo Alto, CA: Consulting Psychologists Press, 1973
18. Poma SZ, Milleri S, Squassante L, Nucci G, Bani M, Perini GI, Merlo-Pich E. Characterization of a 7% carbon dioxide (CO2
ation paradigm to evoke anxiety symptoms in healthy subjects. J Psychopharmacol 2005;19:494–503
19. Schisler T, Lander J, Fowler-Kerry S. Assessing children’s state anxiety. J Pain Symptom Manage 1998;16:80–6
20. McGrath P, de Veber L, Hearn M. Multidimensional pain assessment in children. Adv Pain Res Ther 1985;9:387–93
21. Berk L. Child development. 3rd ed. Needham Heights, MA: Allyn and Bacon, 1994
22. Theunissen NC, Vogels TG, Koopman HM, Verrips GH, Zwinderman KA, Verloove-Vanhorick SP, Wit JM. The proxy problem: child report versus parent report in health-related quality of life research. Qual Life Res 1998;7:387–97
23. Cremeens J, Eiser C, Blades M. Factors influencing agreement between child self-report and parent proxy-reports on the Pediatric Quality of Life Inventory 4.0 (PedsQL) generic core scales. Health Qual Life Outcomes 2006;4:58
24. Hays RD, Hadorn D. Responsiveness to change: an aspect of validity, not a separate dimension. Qual Life Res 1992;1:73–5
25. Rhudy JL, Meagher MW. Negative affect: effects on an evaluative measure of human pain. Pain 2003;104:617–26
26. von Baeyer CL, Spagrud LJ. Systematic review of observational (behavioral) measures of pain for children and adolescents aged 3 to 18 years. Pain 2007;127:140–50
27. Lerman J. Anxiolysis–by the parent or for the parent? Anesthesiology 2000;92:925–7
28. Kain ZN, Mayes LC, Wang SM, Caramico LA, Hofstadter MB. Parental presence during induction of anesthesia versus sedative premedication: which intervention is more effective? Anesthesiology 1998;89:1147–56
29. Kain ZN, Caldwell-Andrews AA, Mayes LC, Weinberg ME, Wang SM, MacLaren JE, Blount RL. Family-centered preparation for surgery improves perioperative outcomes in children: a randomized controlled trial. Anesthesiology 2007;106:65–74