| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Department of Anesthesia and Critical Care Medicine, The Children’s Hospital of Philadelphia; and Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia
Address correspondence to Giovanni Cucchiaro, MD, Department of Anesthesia and Critical Care Medicine, The Children’s Hospital of Philadelphia, 34th St. and Civic Center Blvd., Philadelphia, PA 19104. Address e-mail to Cucchiaro{at}email.chop.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionAppendix 1References |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix 1References |
|---|
Medical interventions for younger patients, particularly in postoperative care settings, are typically focused around achieving the best possible pain control. Effective management of postoperative pain relies on the administration of opioids. The most frequent side effects caused by the administration of opioids are nausea and vomiting. Several studies have been conducted in adult populations using different methods to measure the risk-benefit relationship and to examine patients’ preferences for outcomes in the postoperative period, including a willingness to pay scenarios or responding to virtual scenarios (3,4). The results have clearly shown that nausea and vomiting are adverse outcomes of major concern for the adults.
Health economists have long been confronted with similar challenges of quantifying the tradeoffs patients make (5) with respect to possible treatment outcomes. One of the most successful assessment methods is the standard gamble (SG) technique, which traditionally assigns numeric values to health outcomes with respect to death (utility score of 0) and perfect health (score of 1.0) (6).
The purpose of this study was to determine whether adolescents can contribute to the decision-making processes that affect their care. We selected the two most frequent postoperative complications, pain and vomiting, and examined adolescents’ preferences with respect to the risk of developing these complications. Specifically, we examined adolescents’ willingness to tolerate some pain during the postoperative period to avoid vomiting. We used the SG technique to assess patients’ perceptions of a hypothetical treatment that would provide them with perfect pain control, with respect to different rates of vomiting during the postoperative period.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix 1References |
|---|
The IRB at The Children’s Hospital of Philadelphia approved this study. The study was fully explained to both parents and children. A written consent was signed by the parents, and patients who agreed to participate to the study gave their assent by signing the same form, as requested by the local IRB.
All patients underwent the standard surgical procedure and postoperative analgesia used at The Children’s Hospital of Philadelphia. Analgesia in the postoperative period was provided using a patient-controlled analgesia device in patients who underwent ACL reconstruction and a patient-controlled epidural analgesia device in patients who underwent the Nuss procedure. In this study, patients connected to a patient-controlled analgesia received IV morphine, whereas patients connected to a patient-controlled epidural analgesia received a combination of bupivacaine 0.1% and fentanyl 5 µg/mL infused in an epidural catheter.
The ACL and Nuss groups were observed for 24 and 48 h, respectively. Pain level and presence of side effects secondary to the administration of opioids were recorded every 4 h. Pain scores were measured using a Visual Analog Scale (VAS), graduated from 0 (no pain) to 10 (unbearable pain). With respect to the side effects caused by the administration of opioids, we considered in the final analysis only the presence and number of episodes of vomiting recorded during the study period.
We interviewed patients between 24 and 48 h after surgery to minimize the possible effect of general anesthesia on their performance when responding to the study questionnaire. Patients who experienced pain (defined as VAS score more than 3, this being the cutoff value for administration of IV opioids at this institution) at any time during the study period (either in the recovery room or postoperative period) were asked to participate in the study. The study questionnaire consisted of an initial comprehension task followed by the 2 outcome assessments of hypothetical vignettes and a SG test.
Study comprehension task. Only patients with an adequate knowledge of probabilities could be expected to reliably participate in the study. We used a three-item numeric literacy measure, similar to the tool developed by Woloshin et al. (8) to screen the respondents according to their cognitive sophistication with numbers and ratios. Patients were asked the following three questions by the interviewer:
Patients were excluded if they answered none of the three questions correctly.
Hypothetical vignettes. The second part of the questionnaire consisted of two vignettes about hypothetical patients, Mary and John (Appendix 1). In summary, Mary and John had both undergone an operation and were both connected to a device for postoperative analgesia. John had excellent analgesia. However, he was experiencing moderate nausea and vomiting, such that these symptoms affected his ability to drink and eat. Mary was not experiencing any nausea or vomiting. However, she was feeling moderate pain, which affected her ability to sleep and get out of bed. Patients were then asked to rate Mary’s and John’s experience on a VAS from 0 to 100, where 0 represented a "very unhappy patient" and 100 represented a "very happy patient" (Appendix 1). The ratings assigned to Mary and John were used to verify the patient’s consistency across assessment methods. If a respondent rated Mary’s outcome (pain but no vomiting) better than John’s (no pain but vomiting) on the VAS, then during the SG assessment (to be described below), the respondent was expected to accept a lower risk of vomiting.
The third part of the questionnaire consisted of a SG scenario. To assess how patients evaluate the tradeoff between perfect pain control and side effects (vomiting), we used a procedure similar to the Ping-Pong method developed by Lenert et al. (9). We chose the one-item method primarily because of its ease of administration, especially when interacting with patients in the postoperative period, when they are often unable to use a more reliable, computer-interactive SG assessment.
The patient was first asked to imagine the pain level they experienced since the operation and to choose a pain management method for the selected outcome. They were told "You have two options. You can choose to do nothing. If you choose to do nothing, you will have the same amount of pain for another day before it goes away. Your other option is to get a pain medicine now. Your pain will immediately go away. However, the pain medicine causes some people to feel sick and nauseous. If that happens, you will feel sick and nauseous, and you will vomit for at least a few times a day."
The interviewer repeated the options if required to ensure that the child understood the choice alternatives. The interviewer explained the hypothetical gamble in more detail, "There is no way to know for sure whether or not the pain medicine will make you feel sick. What if there is a 1% chance that it will make you sick?" Typically, the child says that a 1% chance would be acceptable. Then the interviewer followed the Ping-Pong procedure and changed the risk to "99% that it will make you sick," and so on using 10% increments, as described later (10% chance that it will make you sick followed by 90% chance that it will make you sick, etc.), until the patient felt that the status quo was about the same as the hypothetical gamble.
The interviewer used a visual aid to help the child think probabilistically while asking the question. The visual aid was composed of a series of 11 pie charts, the first one being entirely white, and the others having different proportions of white and black sections. The black portions went from 0% (all white pie chart) to 100% (all black pie chart) that increased by 10% increments in between. As the options were presented verbally, patients were simultaneously shown different pie charts and told that the different black portions of the pie charts represented the different risks of vomiting (from the 0% all white pie chart to the 100% all black pie chart) that he or she would take if he or she accepts the new analgesic. The all white pie chart (0% risk of vomiting) was shown first. It is assumed that children who understood the concept of SG would choose to take the new medication. Children were next shown the all black pie chart (100% risk of vomiting). The point when the child would refuse to take the new medication represented the child’s cutoff point for vomiting. To limit external human factors that could have affected the way the SG test was conducted and influenced the results, children were interviewed by only one of the investigators.
Statistical analysis included a Student’s t-test for continuous independent samples and a 
2 test for categorical data. Nonparametric tests (Mann-Whitney) were used to compare non-normally distributed variables. Univariate and multivariate linear regression models were fitted to relate the willingness to take a risk of vomiting to possible predictors. The following variables were analyzed in the regression: risk of vomiting (dependent variable), sex, age, maximum VAS score, VAS score in the recovery room (pain score during the first 60 min after the operation), number of episodes of vomiting during the study period, technique of analgesia, and type of surgery serving as independent variables. Because of the small sample size, the appropriateness of the model was verified by obtaining a robust estimate of variances. A stepwise logistical regression was performed to identify independent variables that could predict the risk of vomiting children were willing to take. The Shapiro-Wilk test of normality was conducted to verify the normal distribution of the SG scores.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix 1References |
|---|
There was no statistical difference between boys and girls with respect to age, VAS score in the recovery room, maximum or mean postoperative VAS score, and incidence of vomiting (Table 1). There was a significant difference in sex distribution between the two procedures, with a significantly larger percentage of boys in the Nuss compared with the ACL group.
|
The overall patients’ rating of Mary’s (experienced pain and no vomiting) and John’s (experienced vomiting and no pain) postoperative course was similar (52 ± 23 and 54 ± 22, respectively). When data were analyzed by sex, boys rated Mary’s experience more positively than did girls (62 ± 20 and 42 ± 19, respectively) (P = 0.008). Girls and boys rated John’s experience similarly (51 ± 14 and 52 ± 26, respectively; P = 0.5). Episodes of vomiting during hospitalization significantly affected children’s rating of John’s experience. Children who vomited rated John’s experience higher than those who did not (59 ± 23 versus 47 ± 23; P = 0.04). We found a positive correlation between girls’ ratings of John’s experience and the risk of vomiting they were willing to take (r = 0.23; P = 0.04). Similarly, we found a negative correlation between boys’ rating of Mary’s experience and the risk of vomiting they were willing to take (r = –0.29; P = 0.035).
The distribution of the SG score was normal (P value for the Shapiro-Wilk test of normality was 0.1; normality not rejected). The mean maximum risk of vomiting that the overall study population was willing to accept to decrease the pain level to zero was 32% ± 24% (median, 30; range, 1–100). The distribution of the risks patients were willing to take is shown in Figure 1: 67% of the children were willing to take <35% risk of vomiting to reduce their pain level to zero. The mean risk of vomiting that girls were willing to take (41% ± 24%) (median, 35; range, 1–100) to reduce their pain level to zero was significantly higher when compared with boys (25% ± 22%; median, 20; range, 1–100; P = 0.02). Also, children who had already experienced vomiting during the postoperative period were more willing to take a higher risk of vomiting (46% ± 26%) compared with those who did not have any episode of vomiting during their hospital stay (23% ± 17%; P = 0.035). There was no difference in the willingness to take a risk of vomiting between ACL reconstruction and Nuss procedure (28% ± 22% and 35% ± 24%, respectively; P = 0.08).
|
Willingness to take the risk of vomiting was significantly correlated with sex (r = 0.15; P = 0.04) and previous episodes of vomiting during the hospitalization (r = 0.25; P = 0.001). No significant correlations were found between the willingness to take a risk of vomiting and maximum VAS score, VAS score in the recovery room, patients’ ages, technique of analgesia used in the postoperative period, or surgical procedure.
The multivariate regression analysis confirmed sex and previous experience of vomiting as significant predictors of children’s willingness to take a risk of vomiting (Table 2). Previous experience of vomiting seemed the strongest predictor of the patients’ decisions (P = 0.002), followed by sex (P = 0.04). The robust variance estimates test indicated that the point estimates are the same with similar standard errors. Henceforth, the model seemed appropriate. The stepwise logistical regression indicated that previous experience of vomiting, VAS in the recovery room, and sex are independent variables for risk of vomiting (R2 = 0.44; P = 0.007).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix 1References |
|---|
The difference between boys and girls in willingness to take a higher risk of vomiting to eliminate pain seems to confirm previous data, which indicate a more frequent prevalence of acute and chronic pain among girls with lower pain thresholds and tolerance in girls compared with boys (10,11).
Although using different methodologies, our study seems to indicate a difference between children and adults, for whom vomiting seemed to always be the major concern (4,12). Another difference between adolescents and adults (3) shown in this study is that adolescents’ willingness to risk vomiting was significantly influenced by the child’s own experience with previous episodes of vomiting. Patients who had already experienced vomiting in the postoperative period were willing to take a higher risk of vomiting to achieve better analgesia, as compared with children who had not previously experienced vomiting. The choice made by adolescents in this study may be explained from the perspective of cognitive behavioral models, which have been used to effectively treat depressive and anxiety disorders (13). Exposure and subsequent desensitization to a noxious or feared stimulus (either imaginary or in vivo) is considered to be a fundamental and active ingredient in the treatment of anxiety for children (14) and has also been found to be effective in the treatment of chronic pain (15). From this perspective, children who have vomited during the postoperative period have in essence been exposed or desensitized to this feared or unwanted outcome. Thus, these patients seem to view their pain as the primary symptom of concern and, in turn, are more likely to pursue pain treatment despite side effects.
Postoperative vomiting is one of the most common complications after general anesthesia with an incidence that can be as frequent as 80% (16,17). Whereas adults consider postoperative nausea and vomiting as one of the most undesirable postoperative symptoms, there are no data on children’s feelings about postoperative vomiting. One of the top priorities of caregivers and parents in the postoperative period is to assure that children do not experience needless suffering. When setting targets for achieving ideal postoperative analgesia in children, physicians base their practice on personal experience, which is supported by data from a few published studies, indicating that parents expect perfect control of postoperative pain and prompt administration of analgesia when children experience minimal pain (18,19).
There are very few studies in the literature addressing the problem of children’s and adolescents’ preferences with respect to different medical outcomes. A challenge when measuring health-related issues in pediatric patients is that parent proxy reports are often used. The use of parents’ proxy reports instead of patient self-reports fails to consider that parents and children may differ in their perception of the child’s health (20). Although patient self-reporting of health-related quality-of-life has gained increasing acceptance in pediatrics when managing children with chronic diseases, such as cystic fibrosis and rheumatoid arthritis, parents often remain the principle medical decision-makers. However, several studies have shown inconsistencies between proxy-perceived value of health-related quality-of-life in children and adults (21,22). This is also true when managing pain. Health care workers assessment of pain experienced by children has been shown to be unreliable compared with a child’s self-report (23,24). When considering previous research, it is also doubtful that parents’ estimates of their children’s pain are reliable as a clinical tool.
SG techniques are used to value global health states and health-related quality-of-life. Although, the goal of SG is typically to find the maximum risk of immediate death the subject would accept in order to avoid the health condition in question, it has also been used to test for other adverse health conditions other than death (9). We used this technique to address only two aspects of postoperative health state: vomiting and pain. Research has shown that the measurements obtained in quantitative preference elicitation are strongly influenced by the elicitation process (25). To make the process clearer to children and minimize possible differences in the test presentation, we used graphic tools (black and white pie charts) (9) to capture and hold the subject’s attention and limited the number of interviewers to one.
SG is a methodologically complicated technique that may seem inappropriate in children (26). However, a few studies have shown that children as young as eight years old can use the SG method to assign preferences for outcomes in childhood chronic diseases (27,28). A limitation in the study design is that we did not ask the parents and medical care providers their opinion so as to compare them with the children’s preferences. Also, the SG technique incorporates the attitudes toward risk in the decision-making process, and children’s risk aversion may have overshadowed other considerations when making the final decision.
In conclusion, this initial study shows that adolescents are capable of assigning a value to different health status, with girls and boys giving different ratings to the same status. These preliminary data seem to suggest that adolescents’ preferences may be different from those of adults. Future studies will benefit from examining a larger sample of patients to test the validity and reliability of the SG technique in children and adolescents. Finally, these studies should directly examine and compare the preferences of parents, health care providers, and pediatric patients to assess potential differences among these groups and help to increase our understanding of what postoperative outcomes matter the most and to whom.
|
Appendix 1 |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix 1References |
|---|
After you are done with that, think of yourself, and give a number using the same scale to rate your experience in the past couple of days with the pain medications.
Story 1: John
After his surgery, John gets medicine for his pain. John is connected to a machine that controls how much medicine John is getting. But John can also push a button to get extra pain medicine if he requires it. For the past couple of days, the medicine has been helpful. John feels no pain, and he feels no need to push the button for extra medicine, but he feels queasy most of the time. He throws up once or twice a day and does not have too much of an appetite because he feels queasy and nauseous. He feels tired, uncomfortable, and wants to go home.
Story 2: Mary
Mary gets medicine for her pain after a procedure. Mary is connected to a machine that controls how much medicine Mary is getting. Mary can push a button to get extra pain medicine if she needs it. For the past couple of days the medicine helps a bit, but Mary sometimes feels pain. Pushing the button for extra medicine helps, and she needs to get that extra pain medicine 2 to 3 times a day, but the pain comes back when the medicine wears out. Mary cannot sleep through the night because of pain. However, she can eat and drink because her appetite is fine. She does not feel queasy. However, she feels uncomfortable and wants to go home.
Story 3: Yourself
Think about yourself and how you have been feeling for the past couple of days. Pay attention to the pain you may have had and the side effects if you have experienced them. Compare your experience with John’s and Mary’s and mark that on the scale, too.
|
Footnotes |
|---|
3Response options included "1%," "10%," "20%," "100%," and "don’t know." One percent was the correct answer.
Accepted for publication December 15, 2005.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix 1References |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. E. Kelly Jr, T. F. Cash, R. C. Shamberger, K. K. Mitchell, R. B. Mellins, M. L. Lawson, K. Oldham, R. G. Azizkhan, A. V. Hebra, D. Nuss, et al. Surgical Repair of Pectus Excavatum Markedly Improves Body Image and Perceived Ability for Physical Activity: Multicenter Study Pediatrics, December 1, 2008; 122(6): 1218 - 1222. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. C. Vreeman, S. M. E. Finnell, E. R. Cernkovich, and A. E. Carroll The Effects of Antiemetics for Children With Vomiting Due to Acute, Moderate Gastroenteritis Arch Pediatr Adolesc Med, September 1, 2008; 162(9): 866 - 869. [Full Text] [PDF]
|
|
|
|
 |
|
 T. Weber, J. Matzl, A. Rokitansky, W. Klimscha, K. Neumann, E. Deusch, and Medical Research Society Superior postoperative pain relief with thoracic epidural analgesia versus intravenous patient-controlled analgesia after minimally invasive pectus excavatum repair. J. Thorac. Cardiovasc. Surg., October 1, 2007; 134(4): 865 - 870. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Cucchiaro, S. N. Adzick, J. B. Rose, L. Maxwell, and M. Watcha A comparison of epidural bupivacaine-fentanyl and bupivacaine-clonidine in children undergoing the Nuss procedure. Anesth. Analg., August 1, 2006; 103(2): 322 - 7, table of contents. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
