Preoperative anxiety correlates with various outcomes such as postoperative analgesic requirements, postanesthesia care unit and hospital stay, and delayed negative psychological effects [1,2]. Considering the high incidence and the associated adverse outcomes [2-4], treatment may be indicated, and both pharmacological (i.e., premedication) and psychological interventions are used to treat this problem [5,6]. However, premedicant usage patterns are not well documented and have been described only in the British literature . From a public health policy point of view, it may be important to survey practice patterns in order to determine how often a premedication intervention is used to prevent an adverse outcome (e.g., preoperative anxiety). Indeed, survey research is widely used in the medical and surgical literature to report practice patterns [8-11]. A recent editorial in the Journal of the American Medical Association has similarly advocated the concept of the "practical standard of care" . That is, the level of care actually being delivered to patients given the limitations of a particular system (e.g., health maintenance organization [HMO]) or geographical location . We therefore designed a repeated mailing survey to assess the current practice of preoperative sedation among anesthesiologists in the United States (US).
A database containing the names of 5,396 (of 34,079 total) randomly selected anesthesiologists was provided to the investigators by the data processing unit of the American Society of Anesthesiology (ASA, Park Ridge, IL). Retired physicians, trainees, and physicians practicing anesthesia outside the US were excluded from the study. The study protocol was approved by our institutional review board.
The initial version of the survey consisted of 82 questions regarding the characteristics of the respondents and their routine use of sedative premedication for both adults and children (see Figure 1). During the pilot phase of the study, the questionnaire was pretested by 250 anesthesiologists and subsequently revised based on analysis of their responses. The final version of the survey instrument (1) was limited to 73 questions in three sections:
1 The questionnaire is available in its entirety from the investigators.
1. Frequency and type of premedication drugs (sedative/hypnotic/analgesic and anticholinergics) used.
2. Opinions and general practices regarding premedication and parental presence during induction of anesthesia.
3. Demographics of the respondent.
The survey questions reported in this article pertain only to the sedative/hypnotic/analgesic groups of drugs. Several types of scales were used in the study, all of which are commonly used in survey research . The frequency of premedication practice was assessed by presenting a statement to which the respondents were asked to estimate the percentage of patients for whom they used premedicant sedatives. Attitudes toward premedication practice were assessed using a series of statements to which the respondents were asked to answer in a five-point Likert scale ranging from most important to least important . Routes used for administration of premedication (e.g., orally, intravenously) were assessed on a rank ordinal scale.
We mailed questionnaires to the participants in the sample during October 1995. Although the survey was anonymous, return envelopes were coded to permit the identification of nonrespondents, to whom we sent additional surveys in December 1995. To measure potential nonresponse bias, we randomly selected a 10% (n = 300) subsample of the physicians who had not responded by December 1995 and mailed them an anonymous questionnaire containing 25 items from the initial survey.
Statistic and Analytic Approaches
Upon data analysis, frequency of premedication was classified by six US geographic regions:
1. Northeast: Connecticut, New Hampshire, Pennsylvania, Delaware, Vermont, Rhode Island, Maine, New Jersey, Washington, DC, Maryland, Massachusetts, and New York.
2. Southeast: Alabama, South Carolina, Tennessee, Florida, Virginia, Georgia, West Virginia, North Carolina, and Kentucky.
3. South Central: Arkansas, Mississippi, Texas, Kansas, Missouri, Louisiana, and Oklahoma.
4. North Central: Illinois, Minnesota, North Dakota, Indiana, Ohio, South Dakota, Iowa, Wisconsin, Michigan, and Nebraska.
5. Southwest: Arizona, Nevada, California, New Mexico, Colorado, Hawaii, and Utah.
6. Northwest: Alaska, Washington, Idaho, Wyoming, Montana, and Oregon.
Demographic data regarding HMO enrollment was obtained from InterStudy (Saint Paul, MN)2. The data were stratified based on the US geographic regions described above. The association between the frequency of premedication and HMO penetration (i.e., HMO enrollment by total population) was examined using Pearson's correlation coefficient (r). Multivariable linear regression models were used to determine whether this association was independent of other demographic variables.
2 Data were obtained from InterStudy Competative Edge: Industry Report 6.1, Table 16"Combined enrollment as a percentage of state population as of July 1, 1995."
Data were analyzed with the use of SPSS version 6.1.1 (SPSS Inc., Chicago, IL). Demographic data are summarized as the mean and standard deviation for continuous data and frequency for categorical data. For each item in the questionnaire, we computed the frequency or mean response with standard errors and 95% confidence intervals. We compared means between subgroups using the unpaired samples t-test and one-way analyses of variance. Categorical items were analyzed by frequency distribution and chi squared analysis. In the case of skewed data, medians and ranges are reported, and nonparametric tests such as the Mann-Whitney U-test and the Kruskal-Wallis H-test were used for analysis. Stepwise multivariable linear regression models were used to assess the independent effects of demographics and practice characteristics on the frequency of premedication. The final models were limited to the significant predictors of the variable of interest and were performed for outcomes of frequency. Significance level was determined at P < 0.05.
Of the 5396 anesthesiologists surveyed, 121 were found to be ineligible for the survey: 7 anesthesiologists had retired, 111 had left no forwarding address, and 3 had died. Of the remaining physicians, 2421 (46%) returned the questionnaire after two mailings. The demographic and professional characteristics of the respondents are shown in Table 1. Their mean age was 44 years, and most (77%) were men. Respondents had practiced anesthesia from 1 to 51 yr (mean 13 +/- 9), and 52% of the respondents had completed anesthesia subspecialty training (e.g., pediatrics).
The reported prevalence of premedication in the preoperative holding area varies widely among the different age groups and geographical locations in the US (Figure 2). Premedicants were used the least often for children younger than age 3 yr and most often for adults less than 65 yr of age (25% vs 75%, P = 0.001). Analysis by geographical regions revealed that premedicants were used least often in the Southwest and Northeast regions and most often in the Southeast region (P = 0.001) (Figure 2). When the frequency of premedication was examined against HMO penetration in the various geographical regions, correlation coefficients (r) ranged from -0.96 to -0.54 (Table 2). Interestingly, the strength of the correlation decreased with the age of patients, with the highest correlation (-0.96) observed with children aged 4-7 yr and the lowest with adults more than 65 yr (Table 2). A multivariable analysis in which premedication in children was the outcome and predictors included HMO penetration, percentage of time the respondent spends providing pediatric anesthesia, number of years in practice, type and number of beds in the hospital of the respondent, and geographical region revealed that HMO penetration, percentage of time the respondent spends providing pediatric anesthesia, and geographical region are independent predictors for the usage of premedication in children (Table 3). Similarly, for adults, this association between premedication and HMO penetration was independent of geographical region, number of years in practice, specialty training, type of hospital and number of beds in the respondent's hospital (Table 3).
Among adults, the most commonly used sedative premedicant was midazolam, followed by diazepam (7%) and lorazepam (2%) (Figure 3). Similarly, for children, midazolam was used by more than 80% of the respondents; the remaining respondents used mostly ketamine (4%) and transmucosal fentanyl (3%). The majority of respondents premedicate adult patients in the preoperative holding area using an intravenous route (>70%) followed by the intramuscular route (18%) and the oral route (10%) (P = 0.001). In contrast, the majority of respondents (80%) premedicate pediatric patients in the preoperative holding area using the oral route followed by the intranasal route (8%), the intramuscular route (6%), and the rectal route (3%) (P = 0.001).
The respondents were also asked to rate 23 factors influencing their premedication practice (Table 3). For example, when asked about the reasons for using sedative premedication, decreased anxiety and increased cooperation were rated as most important; modulation of sympathetic response and analgesia were the least important (P = 0.001).
Comparative analysis of the questionnaires that were received after the first mailing versus those received after the second mailing revealed no significant differences between the two groups. Similarly, the "nonresponse survey" indicated that the nonresponders did not differ significantly from the responders either in demographic characteristics or in their premedication patterns.
The marked variation among geographical areas in premedicant usage patterns underscores the lack of consensus among anesthesiologists about the need for premedication. Geographic variation in various aspects of medical care is well documented . For example, Welch et al.  described geographic variation in expenditure for physicians' services in the US. Similarly, our survey examined geographic variation in premedicant usage patterns. Interestingly, we found that when these data were correlated with HMO penetration, the geographic variations in usage of premedication coincided with penetration by HMOs. It is important to remember, however, that we have measured for association and not for causation. One possible explanation for this association is that hospitals located in areas that are highly penetrated with HMOs are driven to greater efficiency and lower costs. Since premedication may be associated with additional nursing staff, increased pharmacy costs, and possible delays in the operating room schedule and recovery after short surgical procedures, some hospitals may not encourage its use. The higher correlation observed in the younger age groups may further support this explanation. First, unlike adults who are premedicated intramuscularly, most younger children are premedicated orally in a process that requires at least 15-20 min. Second, short surgical procedures (e.g. placement of pressure-equalizing tubes) are far more common in the pediatric population. Therefore, premedicating children may theoretically result in longer operating room delays than those caused by premedicating adults.
Overall, more anesthesiologists in the US use sedative premedicants for adults undergoing outpatient surgery than their colleagues in Great Britain (89% vs 53%, P = 0.001) . Similarly, more British anesthetists indicated that they never premedicate children under-going outpatient surgery with sedative premedicants (37% vs 20%, P = 0.001) . In contrast, preoperative psychological preparation programs and parental presence during induction of anesthesia are more common in Great Britain [15,16]. The reasons for these practice differences may include a stronger demand and less concern about the legal implications of parental presence in Great Britain [7,17,18]. Different anesthesia induction techniques for children may also affect the reported attitudes; while in the US, mask induction is still the most common practice, the introduction of a eutectic mixture of local anesthetics in Great Britain caused many anesthesiologists to use intravenous induction techniques.
Our study has two methodological limitations. First, this study targeted only anesthesiologists who are members of the ASA. It is possible that this group is not representative of anesthesiologists as a whole, although it is estimated that 90% of US anesthesiologists are ASA members . Second, this survey was limited to anesthesiologists, and because nurse anesthetists occasionally provide anesthetic care without the supervision of an anesthesiologist, it is possible that the data presented in this study do not represent their practice.
In conclusion, the marked variation among geographical areas in premedicant usage patterns under-scores the lack of consensus among anesthesiologists regarding the need for premedication. Interestingly, we found that premedicant usage patterns coincided with HMO penetration. Finally, more anesthesiologists in the US use sedative premedicants for both children and adults undergoing surgery than their colleagues in Great Britain.
The authors would like to thank Paul G. Barash, MD, for his critical review of this manuscript.
1. Johnston M. Pre-operative emotional states and post-operative recovery. Adv Psychosom Med 1986;15:1-22.
2. Kain Z, Mayes L, O'Connor T, Cicchetti D. Preoperative anxiety in children: predictors and outcomes. Arch Pediatr Adolesc Med 1996;150:1238-45.
3. Badner N, Nielson W, Munk S, et al. Preoperative anxiety: detection and contributing factors. Can J Anaesth 1990;37:444-7.
4. Moerman N, van Dam F, Muller M, Oosting H. The Amsterdam Preoperative Anxiety and Information Scale (APAIS). Anesth Analg 1996;82:445-51.
5. Kain Z, Mayes L. Anxiety in children during the perioperative period. In: Borestein M, Genevro J, eds. Child development and behavioral pediatrics. NJ: Lawrence Erlbaum Associates, 1996:85-103.
6. Lichtor J, Zacny J. Psychological preparation and preoperative medication. In: Miller RD, ed. Anesthesia. 4th ed. New York: Churchill Livingstone, 1994:1015-44.
7. Mirakhur RK. Preanaesthetic medication: a survey of current usage. J R Soc Med 1991;84:481-3.
8. Putnam F, Loewenstein R. Treatment of multiple personality disorder: a survey of current practices. Am J Psychiatry 1993;150:1048-52.
9. Campbell J, Condon V. Catheter removal of blunt esophageal foreign bodies in children: survey of the Society for Pediatric Radiology. Radiology 1989;19:361-5.
10. Wellington N, Rieder M. Attitudes and practices regarding analgesia in newborn circumcision. Pediatrics 1993;92:541-3.
11. Tait R, Tuttle D. Prevention of occupational transmission of human immunodeficiency virus and hepatitis B virus among anesthesiologists: a survey of anesthesiology practice. Anesth Analg 1994;79:623-8.
12. Argy O. Standards of care. JAMA 1996;275:1296.
13. Alreck P, Settle R. The survey research handbook. 2nd ed. Chicago: Irwin, 1995:113-42.
14. Welch P, Miller M, Welch G, et al. Geographic variation in expenditures for physicians' services in the United States. N Engl J Med 1993;328:621-7.
15. Mathews A, Ridgeway V. Psychological preparation for surgery. In: Steptoe A, Mathews A, eds. Health care and human behavior. London: Academic Press, 1984:231-59.
16. Kain Z, Ferris C, Mayes L, Rimar S. Parental presence during induction of anesthesia: practice differences between the US and Great Britain. Paediatr Anaesth 1996;6:187-93.
17. Holt S. Paediatric anaesthesia. Br J Anaesth 1985;291:673.
18. Roberts W. Paediatric anaesthesia. Br J Anaesth 1985;291:543.
19. Gaba M, Howard S, Jump B. Production pressure in the work environment. Anesthesiology 1994;81:488-500.