- Email: [email protected]
Impact of the Hawthorne Effect in a Longitudinal Clinical Study
Impact of the Hawthorne Effect in a Longitudinal Clinical Study: The Case of Anesthesia Donatella De Amici, MD, Catherine Klersy, MD, Felice Ramajoli, MD, Loretta Brustia, MD, and Pierluigi Politi, MD Department of Epidemiology and Clinical Biometry-Scientific Direction (D.D.A., C.K.); Department of Anesthesiology and Intensive Care (F.R., L.B.); Service of Psychiatry (P.P.); IRCCS Policlinico San Matteo, Pavia, Italy
ABSTRACT: Clinical research can be influenced by many factors that are capable of invalidating results, and one of these factors is known as the Hawthorne effect: the mere awareness of being under observation can alter the way in which a person behaves. In experimental research this effect can be the undesired effect of the experiments themselves, and the stronger its presence, the greater it can influence the results. In anesthesia practice, owing to the particular emotional condition of a patient facing a surgical operation, the Hawthorne effect could be especially strong. The aim of our study was to show the impact that the knowledge of being included in a study has (Hawthorne effect), by comparing the postoperative changes in psychological well-being in two groups of patients undergoing knee arthroscopy and receiving different information about the study from the anesthetist during the preoperative interview. Other signs and symptoms such as postoperative knee pain, nausea, vomiting (the most feared occurrences), headache, return of spontaneous diuresis, analgesic request, anesthesia complications, as well as the intensity of anxiety were also assessed as secondary endpoints. Our results show that subjects who were aware that they were part of a study scored significantly better on postoperative measures of psychological well-being and postoperative knee pain, compared to subjects who were unaware. The size of the effect, as measured by the odds ratio, remains unchanged when controlling for potential confounding factors. The study has enabled us to demonstrate the presence of the Hawthorne effect in clinical research. Therefore, the Hawthorne effect should be acknowledged and accounted for in the design of a study and in the interpretation of results. Control Clin Trials 2000;21:103–114 Elsevier Science Inc. 2000 KEY WORDS: Hawthorne effect, anesthesiologic research
Address reprint requests to: Dott. De Amici Donatella, M.D., Department of Epidemiology and Clinical Biometry, Scientific Direction, IRCCS Policlinico, San Matteo, Italy. E-mail: [email protected] Received June 1, 1999; revised October 29, 1999; accepted November 11, 1999. Controlled Clinical Trials 21:103–114 (2000) Elsevier Science Inc. 2000 655 Avenue of the Americas, New York, NY 10010
0197-2456/00/$–see front matter PII S0197-2456(99)00054-9
104
D. De Amici, et al.
INTRODUCTION Clinical research can be influenced by many factors that are capable of invalidating results. One of these factors is known as the Hawthorne effect, a name derived from experimental studies carried out from 1924 through 1932 in the Hawthorne Works Plant of the Western Electric Company in Chicago [1]. These studies were conducted with a view to improving the productivity of workers by changing their environmental conditions. For example, more light was accompanied by increased productivity but less light also increased productivity. Thus the company’s efforts produced conflicting results, because any change in working conditions, however insignificant, led to greater productivity. The subjects increased their productivity in response to the increased attention and to the subtle pressure of being observed. It was recognized that what later came to be called the Hawthorne effect was responsible for this. Man has a natural propensity to be influenced by being observed, and therefore the mere awareness of being under observation can alter the way in which a person behaves. In experimental research this can be the undesired effect of the experiments themselves, and the stronger its presence, the greater it can influence the results. Therefore risk exists that what is simply the consequence of a patient knowing that he is observed can be mistaken for the outcome of a specific intervention. This should be borne in mind both when preparing the study design and when analyzing the results [2–5]. In the clinical setting an emotional involvement related to the patient-doctor relationship is inevitable, and could induce the Hawthorne effect. Actually, the latter could be especially strong in the practice of anesthesia, owing to the particular emotional condition of a patient facing a surgical operation, and it might therefore influence the results of clinical research in anesthesiology. More than 30 years ago it was discovered that the meeting between patient and anesthetist and the type of information given to the patient can have a positive or negative influence on his state of psychological well-being and on postoperative morbidity [6–11]. The aim of our study was to show the impact that the knowledge of being included in a study has on a trial (Hawthorne effect), by comparing the postoperative changes in psychological well-being in two groups of patients undergoing knee arthroscopy and receiving different information about the study from the anesthetist at the preoperative interview. Other signs and symptoms such as postoperative knee pain, nausea, vomiting (the most feared occurrences), headache, return of spontaneous diuresis, analgesic request, anesthesia complications, as well as the intensity of anxiety were also assessed as secondary endpoints. METHODS Study Design The study was designed as a randomized controlled clinical trial. Two groups of patients were compared: an “additional information group” (exposed group), patients who were informed that they were taking part in a trial; and a “routine information group” (control group), patients who were not. Patients were assigned to one of the two groups according to a simple randomization scheme.
The Hawthorne Effect in Anesthesia
105
It was calculated that a sample size of 58 patients per group would guarantee a power for the study of 80% with a type I error of 5% (two-tailed). In the control group, we hypothesized a decrease of psychological well-being after surgery in 50% of patients. In the exposed group we estimated the postoperative decrease in psychological well-being to be present in only 25% of patients. During the preoperative interview an informed consent form (Appendix A) relating to loco-regional anesthesia was submitted to both groups. The control group received the standard form. The exposed group received a similar form, the sole difference being that the following statement was inserted at the beginning: “We are conducting research to evaluate the acceptability of loco-regional anesthesia. Therefore you are part of a study and will be followed with particular attention and interest to record which side effects of the anesthesia are least acceptable to you.” Both groups were treated identically in all other respects. Approval of the local Ethics Committee was obtained. Setting and Patients Potential participants in the study were consecutive patients, of both sexes and of age ⭓18 years, admitted to the IRCCS Policlinico San Matteo in Pavia, from June 1, 1997 to June 30, 1998, to undergo knee arthroscopy under spinal anesthesia. Unlike general anesthesia, spinal anesthesia makes it possible to interview the patients in the postoperative period with the certainty that their replies will be credible. Furthermore, arthroscopy is an operation of brief duration, performed on young, healthy subjects. The exclusion criteria were the use of tranquilizers, history of psychiatric disease, and the abuse of alcohol or drugs. Patients requesting general anesthesia were also excluded from the study. One hundred twenty-one patients were eligible according to the former criteria. Five of them preferred general anesthesia, so 116 patients were entered into the study. Assessments All the assessments were carried out by two anesthesiologists from the research staff. The first performed the preoperative examination and the subsequent anesthesia. He learned of the patient’s group assignment only at the end of the first examination. The second performed the postoperative examination. He was blinded to the patient’s group assignment. Preoperative Assessments The first anesthetist, in charge of the anesthetic procedure, conducted the enrollment interview privately with each patient on the afternoon before the operation. The examination lasted about 30 minutes. The information collected covered the patient’s personal history and points of anesthesiological importance. The patients underwent a physical examination as well. Subsequently the patients were administered the Italian version of Goldberg’s General Health Questionnaire (GHQ) [12] to determine his preoperative state of psychological well-being. The GHQ is a self-rating questionnaire extensively used to detect non-psychotic mental disorders. The original 60-item
106
D. De Amici, et al.
version and shorter 30-, 28-, 20-, and 12-item versions have been used in a variety of settings and among different social and cultural groups. The GHQ30 was chosen for this study because of its well-validated ability to identify individuals with psychological problems in a nonpsychiatric setting [13], and to measure psychological changes and changes in well-being over time in such situations [14]. According to Goldberg and Williams [12], a score of 5 or 6 was the cut-off used to identify patients with psychiatric pathology. The GHQ-30 was scored by assigning a score of 0 for each item in the first two response categories and of 1 in the last two [14]. Each patient’s scores, obtained for each single item, were then summed; they will be defined as “GHQ score” from here on. Higher scores correspond to a worse quality of life. The Zung SelfRating Anxiety Scale (SAS) [15] was also administered to measure state of anxiety: higher scores represent higher anxiety levels. Finally, the patients were handed the informed consent form according to group allocation. The patients were given about 20 minutes to read and sign it. On the evening before the operation no tranquilizers were administered and no pre-anesthesia was performed on the day of the operation. No analgesics or antiemetics were administered on discharge from the operating theatre. Details on the anesthesiological technique are reported in Appendix B. Postoperative Assessments The surgeon examined each patient shortly after the operation. About 7 hours after the operation the second anesthetist performed the postoperative examination to assess complete recovery from spinal anesthesia. This time is sufficient to allow complete recovery of sensitivity and motricity from the administered dosage of anesthetic. Any request for analgesics and the return of spontaneous urination or the positioning of a urinary catheter were recorded. Then the postoperative GHQ and Zung questionnaires were administered. Finally the presence of a series of symptoms was assessed and/or quantified, such as intensity of postoperative knee pain and headache, and occurrence of nausea and/or vomiting. Postoperative knee pain intensity was measured by a Visual Analog Scale (VAS) [16]. The VAS rating system consists of a 100-mm line, which represents two behavioral extremes at either end of the continuum, i.e., “no pain” at one end and “worst possible pain” at the other. Patients were instructed in the use of the VAS; none of the patients asked for the instruction to be repeated. The intensity of headache was also recorded by VAS. Nausea and vomiting were recorded as occurring or not occurring. Patients were also requested to rank from 0 to 100 the subjective importance that they attributed to each of the symptoms measured, and to give an overall judgement on the anesthesiological technique in terms of discomfort. An overall score for the subjective rating of importance of symptoms was computed as the sum of the single scores. Patients also indicated the presence, if any, of other problems (dizziness, pain at the injection site, pain at any other site, any unusual sensation, dysesthesia or hyperesthesia). On the following morning the request for analgesics during the night as well as the occurrence of any complications were recorded. The patients were then discharged.
The Hawthorne Effect in Anesthesia
107
Outcomes The primary outcome is the change in the state of psychological well-being after the operation, as measured by the GHQ questionnaire. For the primary analysis of the effectiveness of the intervention, change in psychological wellbeing was considered a binary outcome (patients who remain stable or do worse vs. patients whose GHQ score shows a reduction, and so do better). The secondary endpoints were the intensity of postoperative knee pain and/ or headache and the presence of nausea and/or vomiting. Finally, differences in the return of spontaneous diuresis, the possible occurrence of anesthesia-related problems, and the need for analgesic were also assessed, as well as the reduction in anxiety. Statistical Analysis Characteristics at baseline were assessed both for the control and the exposed group. Mean and standard deviation (SD) are reported for continuous variables and absolute and relative frequencies for categorical variables. Student’s t test and a 2 test were used to compared means and proportions respectively. To assess the presence of the Hawthorne effect, a logistic model was fitted to the primary outcome by including at first the group indicator alone and then by adding a series of potential confounders (sex, age, education, marital status, symptoms and subjective rating of importance of symptoms). Thus both the crude and the adjusted odds ratios (OR) together with their 95% confidence interval were computed to quantify the association between the awareness of being observed and the outcome. A similar approach was used to assess this association with the secondary endpoints. A p value ⬍ 0.05 was considered statistically significant. RESULTS A total of 116 patients were enrolled in the study: 56 patients were included in the control group and 60 in the exposed group. Baseline characteristics of the patients are shown in Table 1. Mean GHQ score was 2.3 (SD 3.1) in the control group and 2.7 (SD 2.6) in the exposed group. Fourteen patients in both groups (25.0 and 23.3%, respectively) had a score ⭓ 6 at the baseline assessment. Anesthesia Characteristics All patients were of American Society of Anesthesiologists (ASA) status 1 and no patient had previously undergone spinal anesthesia for surgical operations of any type. Adequate surgical anesthesia was achieved in all patients. No problem arose related to anesthetic technique. The average duration of the operation was 45 minutes (SD 6). Postoperative Assessments Data on primary and secondary endpoints are summarized in Table 2 and demonstrate a lower prevalence of signs of discomfort in the exposed group. The proportion of patients with increased well-being after surgery was significantly higher (p ⫽ 0.03) in the exposed group. A similar trend was observed
108
D. De Amici, et al.
Table 1 Patients’ Baseline Characteristics by Study Group
Characteristic Sex (M/F) Age (years) Height (cm) Weight (kg) Education Low grade (⭐8 years) High grade (⭓13 years) Family Married At least 1 child Occupation None Employee Independent work Manager/freelance
Control Group N ⫽ 56 Mean (SD) or N (%)
Exposed Group N ⫽ 60 Mean (SD) or N (%)
35/21 30.0 (7) 169.2 (5) 64.5 (9)
36/24 32.1 (8) 168.1 (6) 66.4 (8)
26 (46.4%) 30 (53.6%)
34 (56.7%) 26 (43.3%)
19 (33.9%) 13 (23.2%)
27 (54.0%) 18 (30.0%)
4 33 17 2
15 33 9 3
(7.2%) (58.8%) (30.4%) (3.6%)
(25.0%) (55.0%) (15.0%) (5.0%)
for mean percent changes from baseline (see confidence interval), although the difference did not reach statistical significance. Furthermore, the interval between the performance of spinal anesthesia and the postoperative anesthesiologic examination was 7 hours 16 minutes (SD 50 minutes) in the control group and 7 hours 55 minutes (SD 1 hour 2 minutes) in the exposed group (p ⫽ 0.08). Complete recovery of sensory and motor function was observed in all patients during the post-anesthesiological examination the evening after the operation. No patients reported neurologic symptoms such as dysesthesia or hyperesthesia. Seven patients from the control group and 5 patients from the exposed group required catheterization due to difficulty in urinating (p ⫽ 0.63). During the afternoon examination, 21 patients in the control group and 6 in the exposed group had already requested an analgesic owing to postoperative knee pain at the operation site (p ⫽ 0.002). Twenty-one patients in the control group and four in the exposed group asked for it during the night (p ⫽ 0.0001). The Zung SAS reduction was ⫺1.8 (SD 0.4) in the control group and ⫺3.3 (SD 0.4) in the exposed group (p ⫽ 0.011). Association Between the Knowledge of Being Part of a Study and Changes in Psychological Well-being To uncover the presence of the Hawthorne effect, logistic models were fitted to assess the association between the knowledge of being part of a study and a better postoperative psychological outcome. Crude odds ratios for group effect were computed as well as odds ratios controlled for sex, age, education, marital status, symptoms, and subjective rating of importance of symptoms, for the primary outcome, and for postoperative knee pain and headache as secondary outcomes. The symptoms considered for controlling group effect
GHQ ⫽ General Health Questionnaire. a Normal-based 95% confidence interval.
Patients who remain stable or do worse on GHQ Primary outcome Preoperative GHQ score Postoperative GHQ score GHQ score percent change Secondary outcomes Postoperative knee pain rated ⭓ 50 Presence of headache Presence of nausea Presence of vomiting Overall discomfort Overall rating of importance of symptoms 30.2 (2.9)
(39.3%) (30.4%) (10.7%) (1.8%) (16.1%)
(10.0%) (15.0%) (5.0%) (0.0%) (3.3%) 31.8 (2.6)
6 9 3 0 2
2.7 (2.6) 1.1 (1.7) ⫺61.5 (48.8)
2.3 (3.1) 1.5 (2.4) ⫺30.8 (116.7) 22 17 6 1 9
22 (36.7%)
Exposed Group N ⫽ 60 Mean (SD) or N (%)
32 (57.1%)
Control Group N ⫽ 56 Mean (SD) or N (%)
Postoperative Assessment by Study Group
Characteristic
Table 2
(⫺44.2; ⫺14.4) (⫺30.6; ⫺1.4) (⫺15.6; 4.2) (⫺5.2; 1.6) (⫺23.4; ⫺2.1) 1.60 (0.6; 2.6)
⫺29.3 ⫺15.4 ⫺5.7 ⫺1.8 ⫺12.8
0.4 (⫺0.7; 1.4) ⫺0.3 (⫺1.1; 0.4) ⫺30.7 (⫺69.8; 8.4)
⫺20.4 (⫺38.6; ⫺0.023)
Exposed-Control Mean Difference (95% CI)a
The Hawthorne Effect in Anesthesia
109
110
D. De Amici, et al.
Table 3 Measure of the Hawthorne Effect: Logistic Models Fitted to Test the Association of the Awareness of Being Part of a Study and Changes in Psychological Well-being Modelsa Primary endpoint Change in GHQ (raw model) Change in GHQ (stratified model) Secondary endpoints Postoperative knee pain rated ⭓ 50 (raw model) Postoperative knee pain rated ⭓ 50 (stratified model) Headache (raw model) Headache (stratified model)
ORb
95% Confidence Interval
Parameter p Value
Model p Value
2.3 3.1
1.1–4.9 1.1–9.1
— 0.034
0.0266 0.0001
5.8
2.1–15.8
—
0.0002
5.6 2.5 1.9
1.7–18.4 1.0–6.1 0.6–5.7
0.004 — 0.280
0.0001 0.0464 0.0160
OR ⫽ odds ratio; GHQ ⫽ General Health Questionnaire. a
Raw model: OR for group effect is computed (no other covariates in the model); Stratified model: OR for group effect is computed while controlling for sex, age, education, symptoms and total importance of symptoms. b OR quantifies the relative risk for a patient unaware of being part of a trial of a postoperative decrease in well-being, as compared to a patient aware of this fact.
were postoperative knee pain, headache, spontaneous recovery of diuresis and nausea. Data on model fitting are reported in Table 3, showing a clear association between unawareness of being observed and a negative outcome in terms of psychological well-being, postoperative knee pain, and to a lesser degree headache. The size of the effect, as measured by the odds ratio, remains unchanged when controlling for the potential confounding factors. No model was fitted to nausea and vomiting, because of the extremely low prevalence of those symptoms. DISCUSSION Results Our study was aimed at uncovering the Hawthorne effect, namely the tendency of people to change their behavior because of their being observed, in a clinical longitudinal study. Most articles published in the literature were designed to acknowledge the Hawthorne effect in health care research (and sometimes suggest solutions to it), with particular concern for the behavior of physicians or other health care professionals [3–5, 17, 18]. As far as we are aware, a single study was published in the last few years whose purpose was to assess the presence of aspecific effects in clinical longitudinal studies, the Hawthorne effect among others; the outcome measure was quality of life, as in the present study. We feel that a subjective phenomenon such as the Hawthorne effect can be best measured by evaluating changes in the subjective perception of quality of life. It appears from our data that the group of patients who knew they were included in a study reported a better quality of life after the procedure, with a twofold decrease in their GHQ score, compared to unaware
The Hawthorne Effect in Anesthesia
111
patients. The exposed group also had a lower proportion of patients with a stable or increased score (see Table 2). The size of the study has been calculated for changes in proportion of patients with well-being variation; this can explain the lack of significant difference between mean percent changes of GHQ. Both groups of patients were comparable in baseline GHQ score, time of postoperative evaluation and their rating of symptoms importance, as well as for their demographic characteristics. This effect appears to be stronger than in the Bouchet et al. work [2] probably due to the design of the study, which is longitudinal in our case, and to the particular condition of a patient scheduled for an invasive procedure. Results on secondary outcomes further support the evidence for a Hawthorne effect in a clinical setting: fewer patients reported symptoms such as strong pain, headache and nausea in the exposed group; the anxiety level decreased to a greater extent and the request for analgesic was reduced in the hours following surgery. With logistic regression, the Hawthorne effect was maintained for both the primary and the secondary endpoints, after controlling for a series of factors that could influence the response of the patients such as age, sex, marital status, education and symptoms. Clinical Research in Anesthesiology A peculiarity of clinical research in anesthesiology is that the patients’ agreement to take part in a study has to be sought in the preoperative phase. This period is characterized by tension, apprehension, and nervousness. The patients feel particularly vulnerable and frightened and feel that there is nothing they can do to control the contingent situation. The exact percentage of patients who are anxious preoperatively is not known but the literature suggests between 60 and 80% [19, 20]. Therefore, devoting special attention to patients in this particular emotional moment is of paramount importance. This emotional setting could explain the large influence that “being under observation” has in this study. Other clinical settings might uncover Hawthorne effects of a different amount, as could study populations with different age distributions. While the debate on the type of information provided to patients before surgery or the form in which it is presented is on-going [21–26], our study has shown that the subjects’ physiological and psychological responses to a stressful situation like a surgical operation can be made more benign by the awareness of being under observation. The Hawthorne Effect in Clinical Research The very definition of the Hawthorne effect makes it impossible to elicit the effect when subjects provide their informed consent to be included in a study. Actually, the studies reported in the literature, both in health care research and in the clinical setting, are all masked to the subject [2–5], and so is ours. This has enabled us to demonstrate the presence of the Hawthorne effect in clinical research, where some kind of intervention will be applied after randomization of patients. However, the informed consent is part of a classical clinical trial, so the underlying Hawthorne effect should be acknowledged and accounted for in the design of the study and in the interpretation of results. When the Hawthorne
112
D. De Amici, et al.
effect is unequally present in the comparison group, bias of the trial results may occur; also, the Hawthorne effect in a clinical trial may produce results not applicable to everyday events. Whereas this study answers the question concerning the importance of the Hawthorne effect in a field where subjective perception is predominant, the impact of this phenomenon on more “objective” parameters remains open.
REFERENCES 1. Parson HM. What happened at Hawthorne? Science 1974;193:922–932. 2. Bouchet C, Guillemin F, Brianc¸on S. Nonspecific effects in longitudinal studies: Impact on quality of life measures. J Clin Epidemiol 1996;49:15–20. 3. Cebul RD. Randomized, controlled trials using the Metro Firm System. Med Care 1991;29:JS9–JS18. 4. Winkens RAG, Knottnerus JA, Kester AD, Grol RPTM, Pop P. Fitting a routine health care activity into a randomized trial: an experiment possible without informed consent? J Clin Epidemiol 1997;50:435–439. 5. Campbell JP, Maxey VA, Watson WA. Hawthorne effect. Implications for prehospital research. Ann Emerg Med 1995;26:590–594. 6. Egbert LD, Battit GE, Turndorf H, Beecher HK. The value of the preoperative visit by an anesthetist. A study of doctor-patient rapport. JAMA 1963;185:553–555. 7. Egbert LD, Battit GE, Welch CE, Bartlett NK. Reduction of postoperative pain by encouragement and instruction of patients. A study of doctor-patient rapport. N Engl J Med 1964;270:825–827. 8. Miller SM, Mangan CE. Interacting effects of information and coping style in adapting to gynecologic stress: should the doctor tell all? J Personality Soc Psych 1983; 45:223–236. 9. Lonsdale M, Hutchinson GL. Patients’ desire for information about anaesthesia. Scottish and Canadian attitudes. Anesthesia 1991;46:410–412. 10. Perry F, Parker RK, White PF, Clifford A. Role of psychological factors in postoperative pain control and recovery with patient-controlled analgesia. Clin J Pain 1994; 10:57–63. 11. Elsass P, Duedahl H, Friis B, Møller IW, Sørensen MB. The psychological effects of having a contact-person from the anesthetic staff. Acta Anaesthesiol Scand 1987; 31:584–586. 12. Goldberg DP, Williams P. A User’s Guide to the General Health Questionnaire. Windsor: NFER/Nelson; 1988. 13. Malt UF. The validity of the General Health Questionnaire in a sample of accidentally injured adults. Acta Psychiatr Scan 1989;80 Suppl 355:103–112. 14. Beck P, Malt UF, Dencker SJ, et al. Scale for Assessment of diagnosis and severity of mental disorders. General Health Questionnaire (GHQ). Acta Psychiatr Scand 1993;87 Suppl 372:34. 15. Zung WW, Magruder-Habib K, Velez R, Alling W. The comorbidity of anxiety and depression in general medical patients: a longitudinal study. J Clin Psychiatry 1990; 51:77–80. 16. Cella DF, Perry SW. Reliability and concurrent validity of three visual analogue mood scales. Psychol Rep 1986;59:827–833. 17. Zucker TP, Flesche C, Germing U, et al. Patient-controlled versus staff-controlled analgesia with pethidine after allogenic bone marrow transplantation. Pain 1998;75:305–312.
The Hawthorne Effect in Anesthesia
113
18. Kanter MH. The transfusion audit as a tool to improve transfusion practice: a critical appraisal. Transfus Sci 1998;19:69–81. 19. Shevde K, Panagopoulos G. A survey of 800 patients’ knowledge, attitudes and concerns regarding anesthesia. Anesth Analg 1991;73:190–198. 20. Moerman N, Van Dam FSM, Muller MJ, Oosting H. The Amsterdam preoperative anxiety and information scale (APAIS). Anesth Analg 1996;82:445–451. 21. Suls J, Wan CK. Effects of sensory and procedural information on coping with stressful medical procedures and pain: a meta-analysis. J Consult Clin Psychol 1989; 57:372–379. 22. Webber GC. Patient education: a review of the issues. Med Care 1990;28:1089–1103. 23. Devine EC. Effects of psychoeducational care for adult surgical patients: a metaanalysis of 191 studies. Patient Educ Couns 1992;19:129–142. 24. Klafta JM, Roizen MF. Current understanding of patients’ attitudes toward and preparation for anesthesia: A review. Anesth Analg 1996;83:1314–1321. 25. Panda N, Bajaj A, Pershad D, Yaddanapudi LN. Chiari P. Pre-operative anxiety. Effect of early or late position on the operating list. Anaesthesia 1996;51:344–346. 26. Inglis S, Farnill D. The effects of providing preoperative statistical anaesthetic-risk information. Anaesth Intens Care 1993;21:799–805.
APPENDIX A Informed Consent Form The control group received the form with the exclusion of the first paragraph; the exposed group received the full text including the italicized section. We are conducting research to evaluate the acceptability of loco-regional anesthesia. Therefore you are part of a study and will be followed with particular attention and interest to record which side effects of the anesthesia are least acceptable to you. You are at liberty to abstain from participating in this study and you are free to withdraw your consent to participation at any time. The spinal anesthesia to which you will be subjected is a type of anesthesia with which it is possible, with only one drug, to obtain a perfect anesthesia and an excellent postoperative analgesia. To obtain these results, the anesthetic is injected into the back and so the patient becomes unable to move or to feel pain from approximately the waist down. These effects last for some hours after the operation, until the effects of the anesthetic have worn off. You may be unable to urinate: in this case you will be temporarily catheterized. One side effect, which can occur with this type of anesthesia, is the onset of headache. It is important to remain lying down and not to attempt to sit up. Any unpleasant effect vanishes in a few hours.
APPENDIX B Anesthetic Technique On the morning of the operation, spinal anesthesia is performed in lateral decubitus with a 27-gauge needle and 0.5–0.8 mL of hyperbaric bupivacaine at 0.5% is injected into the subarachnoid space. The dosage is calculated on
114
D. De Amici, et al.
the basis of the patient’s height so that the level of anesthesia reached does not go beyond D12. Acetated Ringer solution (1500 mL) is rapidly infused during the operation followed by a further 1000 mL in the postoperative phase. Problems related to the anesthetic technique, e.g., multiple attempts to puncture the dura mater, paresthesias or bleeding are recorded.
ABSTRACT: Clinical research can be influenced by many factors that are capable of invalidating results, and one of these factors is known as the Hawthorne effect: the mere awareness of being under observation can alter the way in which a person behaves. In experimental research this effect can be the undesired effect of the experiments themselves, and the stronger its presence, the greater it can influence the results. In anesthesia practice, owing to the particular emotional condition of a patient facing a surgical operation, the Hawthorne effect could be especially strong. The aim of our study was to show the impact that the knowledge of being included in a study has (Hawthorne effect), by comparing the postoperative changes in psychological well-being in two groups of patients undergoing knee arthroscopy and receiving different information about the study from the anesthetist during the preoperative interview. Other signs and symptoms such as postoperative knee pain, nausea, vomiting (the most feared occurrences), headache, return of spontaneous diuresis, analgesic request, anesthesia complications, as well as the intensity of anxiety were also assessed as secondary endpoints. Our results show that subjects who were aware that they were part of a study scored significantly better on postoperative measures of psychological well-being and postoperative knee pain, compared to subjects who were unaware. The size of the effect, as measured by the odds ratio, remains unchanged when controlling for potential confounding factors. The study has enabled us to demonstrate the presence of the Hawthorne effect in clinical research. Therefore, the Hawthorne effect should be acknowledged and accounted for in the design of a study and in the interpretation of results. Control Clin Trials 2000;21:103–114 Elsevier Science Inc. 2000 KEY WORDS: Hawthorne effect, anesthesiologic research
Address reprint requests to: Dott. De Amici Donatella, M.D., Department of Epidemiology and Clinical Biometry, Scientific Direction, IRCCS Policlinico, San Matteo, Italy. E-mail: [email protected] Received June 1, 1999; revised October 29, 1999; accepted November 11, 1999. Controlled Clinical Trials 21:103–114 (2000) Elsevier Science Inc. 2000 655 Avenue of the Americas, New York, NY 10010
0197-2456/00/$–see front matter PII S0197-2456(99)00054-9
104
D. De Amici, et al.
INTRODUCTION Clinical research can be influenced by many factors that are capable of invalidating results. One of these factors is known as the Hawthorne effect, a name derived from experimental studies carried out from 1924 through 1932 in the Hawthorne Works Plant of the Western Electric Company in Chicago [1]. These studies were conducted with a view to improving the productivity of workers by changing their environmental conditions. For example, more light was accompanied by increased productivity but less light also increased productivity. Thus the company’s efforts produced conflicting results, because any change in working conditions, however insignificant, led to greater productivity. The subjects increased their productivity in response to the increased attention and to the subtle pressure of being observed. It was recognized that what later came to be called the Hawthorne effect was responsible for this. Man has a natural propensity to be influenced by being observed, and therefore the mere awareness of being under observation can alter the way in which a person behaves. In experimental research this can be the undesired effect of the experiments themselves, and the stronger its presence, the greater it can influence the results. Therefore risk exists that what is simply the consequence of a patient knowing that he is observed can be mistaken for the outcome of a specific intervention. This should be borne in mind both when preparing the study design and when analyzing the results [2–5]. In the clinical setting an emotional involvement related to the patient-doctor relationship is inevitable, and could induce the Hawthorne effect. Actually, the latter could be especially strong in the practice of anesthesia, owing to the particular emotional condition of a patient facing a surgical operation, and it might therefore influence the results of clinical research in anesthesiology. More than 30 years ago it was discovered that the meeting between patient and anesthetist and the type of information given to the patient can have a positive or negative influence on his state of psychological well-being and on postoperative morbidity [6–11]. The aim of our study was to show the impact that the knowledge of being included in a study has on a trial (Hawthorne effect), by comparing the postoperative changes in psychological well-being in two groups of patients undergoing knee arthroscopy and receiving different information about the study from the anesthetist at the preoperative interview. Other signs and symptoms such as postoperative knee pain, nausea, vomiting (the most feared occurrences), headache, return of spontaneous diuresis, analgesic request, anesthesia complications, as well as the intensity of anxiety were also assessed as secondary endpoints. METHODS Study Design The study was designed as a randomized controlled clinical trial. Two groups of patients were compared: an “additional information group” (exposed group), patients who were informed that they were taking part in a trial; and a “routine information group” (control group), patients who were not. Patients were assigned to one of the two groups according to a simple randomization scheme.
The Hawthorne Effect in Anesthesia
105
It was calculated that a sample size of 58 patients per group would guarantee a power for the study of 80% with a type I error of 5% (two-tailed). In the control group, we hypothesized a decrease of psychological well-being after surgery in 50% of patients. In the exposed group we estimated the postoperative decrease in psychological well-being to be present in only 25% of patients. During the preoperative interview an informed consent form (Appendix A) relating to loco-regional anesthesia was submitted to both groups. The control group received the standard form. The exposed group received a similar form, the sole difference being that the following statement was inserted at the beginning: “We are conducting research to evaluate the acceptability of loco-regional anesthesia. Therefore you are part of a study and will be followed with particular attention and interest to record which side effects of the anesthesia are least acceptable to you.” Both groups were treated identically in all other respects. Approval of the local Ethics Committee was obtained. Setting and Patients Potential participants in the study were consecutive patients, of both sexes and of age ⭓18 years, admitted to the IRCCS Policlinico San Matteo in Pavia, from June 1, 1997 to June 30, 1998, to undergo knee arthroscopy under spinal anesthesia. Unlike general anesthesia, spinal anesthesia makes it possible to interview the patients in the postoperative period with the certainty that their replies will be credible. Furthermore, arthroscopy is an operation of brief duration, performed on young, healthy subjects. The exclusion criteria were the use of tranquilizers, history of psychiatric disease, and the abuse of alcohol or drugs. Patients requesting general anesthesia were also excluded from the study. One hundred twenty-one patients were eligible according to the former criteria. Five of them preferred general anesthesia, so 116 patients were entered into the study. Assessments All the assessments were carried out by two anesthesiologists from the research staff. The first performed the preoperative examination and the subsequent anesthesia. He learned of the patient’s group assignment only at the end of the first examination. The second performed the postoperative examination. He was blinded to the patient’s group assignment. Preoperative Assessments The first anesthetist, in charge of the anesthetic procedure, conducted the enrollment interview privately with each patient on the afternoon before the operation. The examination lasted about 30 minutes. The information collected covered the patient’s personal history and points of anesthesiological importance. The patients underwent a physical examination as well. Subsequently the patients were administered the Italian version of Goldberg’s General Health Questionnaire (GHQ) [12] to determine his preoperative state of psychological well-being. The GHQ is a self-rating questionnaire extensively used to detect non-psychotic mental disorders. The original 60-item
106
D. De Amici, et al.
version and shorter 30-, 28-, 20-, and 12-item versions have been used in a variety of settings and among different social and cultural groups. The GHQ30 was chosen for this study because of its well-validated ability to identify individuals with psychological problems in a nonpsychiatric setting [13], and to measure psychological changes and changes in well-being over time in such situations [14]. According to Goldberg and Williams [12], a score of 5 or 6 was the cut-off used to identify patients with psychiatric pathology. The GHQ-30 was scored by assigning a score of 0 for each item in the first two response categories and of 1 in the last two [14]. Each patient’s scores, obtained for each single item, were then summed; they will be defined as “GHQ score” from here on. Higher scores correspond to a worse quality of life. The Zung SelfRating Anxiety Scale (SAS) [15] was also administered to measure state of anxiety: higher scores represent higher anxiety levels. Finally, the patients were handed the informed consent form according to group allocation. The patients were given about 20 minutes to read and sign it. On the evening before the operation no tranquilizers were administered and no pre-anesthesia was performed on the day of the operation. No analgesics or antiemetics were administered on discharge from the operating theatre. Details on the anesthesiological technique are reported in Appendix B. Postoperative Assessments The surgeon examined each patient shortly after the operation. About 7 hours after the operation the second anesthetist performed the postoperative examination to assess complete recovery from spinal anesthesia. This time is sufficient to allow complete recovery of sensitivity and motricity from the administered dosage of anesthetic. Any request for analgesics and the return of spontaneous urination or the positioning of a urinary catheter were recorded. Then the postoperative GHQ and Zung questionnaires were administered. Finally the presence of a series of symptoms was assessed and/or quantified, such as intensity of postoperative knee pain and headache, and occurrence of nausea and/or vomiting. Postoperative knee pain intensity was measured by a Visual Analog Scale (VAS) [16]. The VAS rating system consists of a 100-mm line, which represents two behavioral extremes at either end of the continuum, i.e., “no pain” at one end and “worst possible pain” at the other. Patients were instructed in the use of the VAS; none of the patients asked for the instruction to be repeated. The intensity of headache was also recorded by VAS. Nausea and vomiting were recorded as occurring or not occurring. Patients were also requested to rank from 0 to 100 the subjective importance that they attributed to each of the symptoms measured, and to give an overall judgement on the anesthesiological technique in terms of discomfort. An overall score for the subjective rating of importance of symptoms was computed as the sum of the single scores. Patients also indicated the presence, if any, of other problems (dizziness, pain at the injection site, pain at any other site, any unusual sensation, dysesthesia or hyperesthesia). On the following morning the request for analgesics during the night as well as the occurrence of any complications were recorded. The patients were then discharged.
The Hawthorne Effect in Anesthesia
107
Outcomes The primary outcome is the change in the state of psychological well-being after the operation, as measured by the GHQ questionnaire. For the primary analysis of the effectiveness of the intervention, change in psychological wellbeing was considered a binary outcome (patients who remain stable or do worse vs. patients whose GHQ score shows a reduction, and so do better). The secondary endpoints were the intensity of postoperative knee pain and/ or headache and the presence of nausea and/or vomiting. Finally, differences in the return of spontaneous diuresis, the possible occurrence of anesthesia-related problems, and the need for analgesic were also assessed, as well as the reduction in anxiety. Statistical Analysis Characteristics at baseline were assessed both for the control and the exposed group. Mean and standard deviation (SD) are reported for continuous variables and absolute and relative frequencies for categorical variables. Student’s t test and a 2 test were used to compared means and proportions respectively. To assess the presence of the Hawthorne effect, a logistic model was fitted to the primary outcome by including at first the group indicator alone and then by adding a series of potential confounders (sex, age, education, marital status, symptoms and subjective rating of importance of symptoms). Thus both the crude and the adjusted odds ratios (OR) together with their 95% confidence interval were computed to quantify the association between the awareness of being observed and the outcome. A similar approach was used to assess this association with the secondary endpoints. A p value ⬍ 0.05 was considered statistically significant. RESULTS A total of 116 patients were enrolled in the study: 56 patients were included in the control group and 60 in the exposed group. Baseline characteristics of the patients are shown in Table 1. Mean GHQ score was 2.3 (SD 3.1) in the control group and 2.7 (SD 2.6) in the exposed group. Fourteen patients in both groups (25.0 and 23.3%, respectively) had a score ⭓ 6 at the baseline assessment. Anesthesia Characteristics All patients were of American Society of Anesthesiologists (ASA) status 1 and no patient had previously undergone spinal anesthesia for surgical operations of any type. Adequate surgical anesthesia was achieved in all patients. No problem arose related to anesthetic technique. The average duration of the operation was 45 minutes (SD 6). Postoperative Assessments Data on primary and secondary endpoints are summarized in Table 2 and demonstrate a lower prevalence of signs of discomfort in the exposed group. The proportion of patients with increased well-being after surgery was significantly higher (p ⫽ 0.03) in the exposed group. A similar trend was observed
108
D. De Amici, et al.
Table 1 Patients’ Baseline Characteristics by Study Group
Characteristic Sex (M/F) Age (years) Height (cm) Weight (kg) Education Low grade (⭐8 years) High grade (⭓13 years) Family Married At least 1 child Occupation None Employee Independent work Manager/freelance
Control Group N ⫽ 56 Mean (SD) or N (%)
Exposed Group N ⫽ 60 Mean (SD) or N (%)
35/21 30.0 (7) 169.2 (5) 64.5 (9)
36/24 32.1 (8) 168.1 (6) 66.4 (8)
26 (46.4%) 30 (53.6%)
34 (56.7%) 26 (43.3%)
19 (33.9%) 13 (23.2%)
27 (54.0%) 18 (30.0%)
4 33 17 2
15 33 9 3
(7.2%) (58.8%) (30.4%) (3.6%)
(25.0%) (55.0%) (15.0%) (5.0%)
for mean percent changes from baseline (see confidence interval), although the difference did not reach statistical significance. Furthermore, the interval between the performance of spinal anesthesia and the postoperative anesthesiologic examination was 7 hours 16 minutes (SD 50 minutes) in the control group and 7 hours 55 minutes (SD 1 hour 2 minutes) in the exposed group (p ⫽ 0.08). Complete recovery of sensory and motor function was observed in all patients during the post-anesthesiological examination the evening after the operation. No patients reported neurologic symptoms such as dysesthesia or hyperesthesia. Seven patients from the control group and 5 patients from the exposed group required catheterization due to difficulty in urinating (p ⫽ 0.63). During the afternoon examination, 21 patients in the control group and 6 in the exposed group had already requested an analgesic owing to postoperative knee pain at the operation site (p ⫽ 0.002). Twenty-one patients in the control group and four in the exposed group asked for it during the night (p ⫽ 0.0001). The Zung SAS reduction was ⫺1.8 (SD 0.4) in the control group and ⫺3.3 (SD 0.4) in the exposed group (p ⫽ 0.011). Association Between the Knowledge of Being Part of a Study and Changes in Psychological Well-being To uncover the presence of the Hawthorne effect, logistic models were fitted to assess the association between the knowledge of being part of a study and a better postoperative psychological outcome. Crude odds ratios for group effect were computed as well as odds ratios controlled for sex, age, education, marital status, symptoms, and subjective rating of importance of symptoms, for the primary outcome, and for postoperative knee pain and headache as secondary outcomes. The symptoms considered for controlling group effect
GHQ ⫽ General Health Questionnaire. a Normal-based 95% confidence interval.
Patients who remain stable or do worse on GHQ Primary outcome Preoperative GHQ score Postoperative GHQ score GHQ score percent change Secondary outcomes Postoperative knee pain rated ⭓ 50 Presence of headache Presence of nausea Presence of vomiting Overall discomfort Overall rating of importance of symptoms 30.2 (2.9)
(39.3%) (30.4%) (10.7%) (1.8%) (16.1%)
(10.0%) (15.0%) (5.0%) (0.0%) (3.3%) 31.8 (2.6)
6 9 3 0 2
2.7 (2.6) 1.1 (1.7) ⫺61.5 (48.8)
2.3 (3.1) 1.5 (2.4) ⫺30.8 (116.7) 22 17 6 1 9
22 (36.7%)
Exposed Group N ⫽ 60 Mean (SD) or N (%)
32 (57.1%)
Control Group N ⫽ 56 Mean (SD) or N (%)
Postoperative Assessment by Study Group
Characteristic
Table 2
(⫺44.2; ⫺14.4) (⫺30.6; ⫺1.4) (⫺15.6; 4.2) (⫺5.2; 1.6) (⫺23.4; ⫺2.1) 1.60 (0.6; 2.6)
⫺29.3 ⫺15.4 ⫺5.7 ⫺1.8 ⫺12.8
0.4 (⫺0.7; 1.4) ⫺0.3 (⫺1.1; 0.4) ⫺30.7 (⫺69.8; 8.4)
⫺20.4 (⫺38.6; ⫺0.023)
Exposed-Control Mean Difference (95% CI)a
The Hawthorne Effect in Anesthesia
109
110
D. De Amici, et al.
Table 3 Measure of the Hawthorne Effect: Logistic Models Fitted to Test the Association of the Awareness of Being Part of a Study and Changes in Psychological Well-being Modelsa Primary endpoint Change in GHQ (raw model) Change in GHQ (stratified model) Secondary endpoints Postoperative knee pain rated ⭓ 50 (raw model) Postoperative knee pain rated ⭓ 50 (stratified model) Headache (raw model) Headache (stratified model)
ORb
95% Confidence Interval
Parameter p Value
Model p Value
2.3 3.1
1.1–4.9 1.1–9.1
— 0.034
0.0266 0.0001
5.8
2.1–15.8
—
0.0002
5.6 2.5 1.9
1.7–18.4 1.0–6.1 0.6–5.7
0.004 — 0.280
0.0001 0.0464 0.0160
OR ⫽ odds ratio; GHQ ⫽ General Health Questionnaire. a
Raw model: OR for group effect is computed (no other covariates in the model); Stratified model: OR for group effect is computed while controlling for sex, age, education, symptoms and total importance of symptoms. b OR quantifies the relative risk for a patient unaware of being part of a trial of a postoperative decrease in well-being, as compared to a patient aware of this fact.
were postoperative knee pain, headache, spontaneous recovery of diuresis and nausea. Data on model fitting are reported in Table 3, showing a clear association between unawareness of being observed and a negative outcome in terms of psychological well-being, postoperative knee pain, and to a lesser degree headache. The size of the effect, as measured by the odds ratio, remains unchanged when controlling for the potential confounding factors. No model was fitted to nausea and vomiting, because of the extremely low prevalence of those symptoms. DISCUSSION Results Our study was aimed at uncovering the Hawthorne effect, namely the tendency of people to change their behavior because of their being observed, in a clinical longitudinal study. Most articles published in the literature were designed to acknowledge the Hawthorne effect in health care research (and sometimes suggest solutions to it), with particular concern for the behavior of physicians or other health care professionals [3–5, 17, 18]. As far as we are aware, a single study was published in the last few years whose purpose was to assess the presence of aspecific effects in clinical longitudinal studies, the Hawthorne effect among others; the outcome measure was quality of life, as in the present study. We feel that a subjective phenomenon such as the Hawthorne effect can be best measured by evaluating changes in the subjective perception of quality of life. It appears from our data that the group of patients who knew they were included in a study reported a better quality of life after the procedure, with a twofold decrease in their GHQ score, compared to unaware
The Hawthorne Effect in Anesthesia
111
patients. The exposed group also had a lower proportion of patients with a stable or increased score (see Table 2). The size of the study has been calculated for changes in proportion of patients with well-being variation; this can explain the lack of significant difference between mean percent changes of GHQ. Both groups of patients were comparable in baseline GHQ score, time of postoperative evaluation and their rating of symptoms importance, as well as for their demographic characteristics. This effect appears to be stronger than in the Bouchet et al. work [2] probably due to the design of the study, which is longitudinal in our case, and to the particular condition of a patient scheduled for an invasive procedure. Results on secondary outcomes further support the evidence for a Hawthorne effect in a clinical setting: fewer patients reported symptoms such as strong pain, headache and nausea in the exposed group; the anxiety level decreased to a greater extent and the request for analgesic was reduced in the hours following surgery. With logistic regression, the Hawthorne effect was maintained for both the primary and the secondary endpoints, after controlling for a series of factors that could influence the response of the patients such as age, sex, marital status, education and symptoms. Clinical Research in Anesthesiology A peculiarity of clinical research in anesthesiology is that the patients’ agreement to take part in a study has to be sought in the preoperative phase. This period is characterized by tension, apprehension, and nervousness. The patients feel particularly vulnerable and frightened and feel that there is nothing they can do to control the contingent situation. The exact percentage of patients who are anxious preoperatively is not known but the literature suggests between 60 and 80% [19, 20]. Therefore, devoting special attention to patients in this particular emotional moment is of paramount importance. This emotional setting could explain the large influence that “being under observation” has in this study. Other clinical settings might uncover Hawthorne effects of a different amount, as could study populations with different age distributions. While the debate on the type of information provided to patients before surgery or the form in which it is presented is on-going [21–26], our study has shown that the subjects’ physiological and psychological responses to a stressful situation like a surgical operation can be made more benign by the awareness of being under observation. The Hawthorne Effect in Clinical Research The very definition of the Hawthorne effect makes it impossible to elicit the effect when subjects provide their informed consent to be included in a study. Actually, the studies reported in the literature, both in health care research and in the clinical setting, are all masked to the subject [2–5], and so is ours. This has enabled us to demonstrate the presence of the Hawthorne effect in clinical research, where some kind of intervention will be applied after randomization of patients. However, the informed consent is part of a classical clinical trial, so the underlying Hawthorne effect should be acknowledged and accounted for in the design of the study and in the interpretation of results. When the Hawthorne
112
D. De Amici, et al.
effect is unequally present in the comparison group, bias of the trial results may occur; also, the Hawthorne effect in a clinical trial may produce results not applicable to everyday events. Whereas this study answers the question concerning the importance of the Hawthorne effect in a field where subjective perception is predominant, the impact of this phenomenon on more “objective” parameters remains open.
REFERENCES 1. Parson HM. What happened at Hawthorne? Science 1974;193:922–932. 2. Bouchet C, Guillemin F, Brianc¸on S. Nonspecific effects in longitudinal studies: Impact on quality of life measures. J Clin Epidemiol 1996;49:15–20. 3. Cebul RD. Randomized, controlled trials using the Metro Firm System. Med Care 1991;29:JS9–JS18. 4. Winkens RAG, Knottnerus JA, Kester AD, Grol RPTM, Pop P. Fitting a routine health care activity into a randomized trial: an experiment possible without informed consent? J Clin Epidemiol 1997;50:435–439. 5. Campbell JP, Maxey VA, Watson WA. Hawthorne effect. Implications for prehospital research. Ann Emerg Med 1995;26:590–594. 6. Egbert LD, Battit GE, Turndorf H, Beecher HK. The value of the preoperative visit by an anesthetist. A study of doctor-patient rapport. JAMA 1963;185:553–555. 7. Egbert LD, Battit GE, Welch CE, Bartlett NK. Reduction of postoperative pain by encouragement and instruction of patients. A study of doctor-patient rapport. N Engl J Med 1964;270:825–827. 8. Miller SM, Mangan CE. Interacting effects of information and coping style in adapting to gynecologic stress: should the doctor tell all? J Personality Soc Psych 1983; 45:223–236. 9. Lonsdale M, Hutchinson GL. Patients’ desire for information about anaesthesia. Scottish and Canadian attitudes. Anesthesia 1991;46:410–412. 10. Perry F, Parker RK, White PF, Clifford A. Role of psychological factors in postoperative pain control and recovery with patient-controlled analgesia. Clin J Pain 1994; 10:57–63. 11. Elsass P, Duedahl H, Friis B, Møller IW, Sørensen MB. The psychological effects of having a contact-person from the anesthetic staff. Acta Anaesthesiol Scand 1987; 31:584–586. 12. Goldberg DP, Williams P. A User’s Guide to the General Health Questionnaire. Windsor: NFER/Nelson; 1988. 13. Malt UF. The validity of the General Health Questionnaire in a sample of accidentally injured adults. Acta Psychiatr Scan 1989;80 Suppl 355:103–112. 14. Beck P, Malt UF, Dencker SJ, et al. Scale for Assessment of diagnosis and severity of mental disorders. General Health Questionnaire (GHQ). Acta Psychiatr Scand 1993;87 Suppl 372:34. 15. Zung WW, Magruder-Habib K, Velez R, Alling W. The comorbidity of anxiety and depression in general medical patients: a longitudinal study. J Clin Psychiatry 1990; 51:77–80. 16. Cella DF, Perry SW. Reliability and concurrent validity of three visual analogue mood scales. Psychol Rep 1986;59:827–833. 17. Zucker TP, Flesche C, Germing U, et al. Patient-controlled versus staff-controlled analgesia with pethidine after allogenic bone marrow transplantation. Pain 1998;75:305–312.
The Hawthorne Effect in Anesthesia
113
18. Kanter MH. The transfusion audit as a tool to improve transfusion practice: a critical appraisal. Transfus Sci 1998;19:69–81. 19. Shevde K, Panagopoulos G. A survey of 800 patients’ knowledge, attitudes and concerns regarding anesthesia. Anesth Analg 1991;73:190–198. 20. Moerman N, Van Dam FSM, Muller MJ, Oosting H. The Amsterdam preoperative anxiety and information scale (APAIS). Anesth Analg 1996;82:445–451. 21. Suls J, Wan CK. Effects of sensory and procedural information on coping with stressful medical procedures and pain: a meta-analysis. J Consult Clin Psychol 1989; 57:372–379. 22. Webber GC. Patient education: a review of the issues. Med Care 1990;28:1089–1103. 23. Devine EC. Effects of psychoeducational care for adult surgical patients: a metaanalysis of 191 studies. Patient Educ Couns 1992;19:129–142. 24. Klafta JM, Roizen MF. Current understanding of patients’ attitudes toward and preparation for anesthesia: A review. Anesth Analg 1996;83:1314–1321. 25. Panda N, Bajaj A, Pershad D, Yaddanapudi LN. Chiari P. Pre-operative anxiety. Effect of early or late position on the operating list. Anaesthesia 1996;51:344–346. 26. Inglis S, Farnill D. The effects of providing preoperative statistical anaesthetic-risk information. Anaesth Intens Care 1993;21:799–805.
APPENDIX A Informed Consent Form The control group received the form with the exclusion of the first paragraph; the exposed group received the full text including the italicized section. We are conducting research to evaluate the acceptability of loco-regional anesthesia. Therefore you are part of a study and will be followed with particular attention and interest to record which side effects of the anesthesia are least acceptable to you. You are at liberty to abstain from participating in this study and you are free to withdraw your consent to participation at any time. The spinal anesthesia to which you will be subjected is a type of anesthesia with which it is possible, with only one drug, to obtain a perfect anesthesia and an excellent postoperative analgesia. To obtain these results, the anesthetic is injected into the back and so the patient becomes unable to move or to feel pain from approximately the waist down. These effects last for some hours after the operation, until the effects of the anesthetic have worn off. You may be unable to urinate: in this case you will be temporarily catheterized. One side effect, which can occur with this type of anesthesia, is the onset of headache. It is important to remain lying down and not to attempt to sit up. Any unpleasant effect vanishes in a few hours.
APPENDIX B Anesthetic Technique On the morning of the operation, spinal anesthesia is performed in lateral decubitus with a 27-gauge needle and 0.5–0.8 mL of hyperbaric bupivacaine at 0.5% is injected into the subarachnoid space. The dosage is calculated on
114
D. De Amici, et al.
the basis of the patient’s height so that the level of anesthesia reached does not go beyond D12. Acetated Ringer solution (1500 mL) is rapidly infused during the operation followed by a further 1000 mL in the postoperative phase. Problems related to the anesthetic technique, e.g., multiple attempts to puncture the dura mater, paresthesias or bleeding are recorded.