Outpatient surgical procedures are increasing, as are the number of available anesthetic treatments providing comfortable same-day discharge (1). Evaluating outcomes that can be traced to anesthetic advances, such as continuous nerve block techniques, requires the accurate assessment and monitoring of three variables: pain, physical function, and mental function during the postoperative period. Although a few methods to measure these outcomes have been reported in the anesthesia literature (2–5), they are rarely used, and there is no consensus among anesthesiologists about the best instrument to use in the days after outpatient surgery. In most research studies, anesthesiologists assess pain levels using a 10-point or 100-point visual or verbal analog scale, or use a ranking of verbal descriptors on a nominal, ordinal, or Likert scale (6–8). In clinical practice, the ideal instrument to assess daily outpatient outcomes would logically be administered rapidly and efficiently via the telephone by staff of the ambulatory surgery facility and/or surgeon's office.
The goal of our study was to evaluate the pain, physical function, and mental function components of three different outcome measures to determine which of these instruments was optimal for the postsurgical assessment of patients who received a femoral nerve block during anterior cruciate ligament reconstruction (ACLR) with spinal anesthesia. The three instruments were: (i) the numeric rating scale (NRS, scale 0–10) for pain (9–12); (ii) the 8-item Short-Form (SF-8) Health Survey (13); and (iii) the 40-item Quality of Recovery (QoR) from Anesthesia Survey (14–16). The SF-8 is a psychometrically valid daily instrument which we selected to overcome disadvantages of the SF-36 and SF-12 (17–18). The SF-8 has been used for a number of patient populations (13,19). The QoR has been validated using inpatient populations, and provides a total score, as well as six composite scores (14–16). Our objective was to conduct a robust psychometric assessment of these instruments for surgical outpatients undergoing regional anesthesia without general anesthesia.
Since the brevity of the SF-8 fulfills our requirements for an ideal instrument, we assessed whether it would be an acceptable outcome measure for pain when compared with the NRS pain score with movement (NRSM) and QoR pain questions, and whether the SF-8 would sufficiently capture patient-reported outcomes for physical function and mental function when compared with the QoR. Evaluations of these instruments were based on the concepts of psychometric theory.
Participants and Procedures
Our study was part of a larger randomized clinical trial that was approved by the IRB of the University of Pittsburgh Medical Center, and was described in detail by Williams et al. (20).
Patients were eligible for participation if they were aged 14–65 years, had an ASA physical status classification of I or II, and were scheduled to undergo outpatient ACLR without additional complex knee procedures. Patients were recruited on or before the day of surgery and were prospectively assigned to one of three groups (1): the placebo group, which would receive a 30-mL saline bolus plus a saline infusion of 270 mL given at a rate of 5 mL/h (2); a treatment group that would receive a 0.25% levobupivacaine bolus with a saline infusion or (3) a treatment group that would receive a 0.25% levobupivacaine bolus plus a 0.25% levobupivacaine infusion. All patients received spinal anesthesia, standardized multimodal analgesia and antiemesis, and a perineural femoral catheter. For the current psychometric study, only the placebo and second treatment group were used, and we will refer to these two groups in the current study as the placebo and the intervention group.
During the preoperative interview, we obtained informed consent and demographic data including age, sex, race/ethnicity, height, weight, smoking status, ASA physical status, and a telephone number at which each participant could be reached for postoperative assessment and follow-up. We described the NRS, SF-8, and QoR to study participants and explained that their participation would entail their reporting verbal NRSM scores and completing the SF-8 and QoR in writing on five occasions: immediately before their surgery, and again on postoperative days 1 through 4. We told them that we would telephone them between 3 pm and 5 pm on the four postoperative days to record their NRS and NRSM scores for pain, and to remind them to complete and return the SF-8 and QoR.
For assessments of pain, we evaluated five measures: the NRSM, the response to the SF-8 question about bodily pain (SF-8 Pain), the QoR pain composite score (QoR Pain Comp), the response to the QoR question about moderate pain duration (QoR Moderate Pain), and the response to the QoR question about severe pain duration (QoR Severe Pain). We did not include the NRS at rest score, because it did not distinguish pain as well as did the NRSM.
For assessment of physical function, we evaluated three measures: the SF-8 physical component score (PCS), the QoR physical independence composite score (QoR Phys Indep), and the QoR physical comfort composite score (QoR Phys Comf). For assessment of mental function, we evaluated three measures: the SF-8 mental component score (MCS), the QoR emotional state composite score (QoR Emot State), and the QoR psychological independence composite score (QoR Psy Indep).
The SF-8 composites are linear combinations of the eight items based on the factor loadings for physical functioning and mental functioning. The QoR composites are sums of Likert scale items.
The goal of this study was to determine whether or not the SF-8 can be used to assess short-term pain, physical functioning, and mental functioning without giving up any additional information provided by the NRSM and QoR. To do this, we assessed whether the SF-8 was as reliable, valid, and responsive as the NRSM and QoR composites measuring similar constructs. Reliability assesses how consistently individuals would respond to the items. Validity assesses whether the instrument is measuring what we expect it to, and responsiveness evaluates whether the instrument can capture change in a construct when it occurs in the sample.
We chose to include subjects in the placebo and second intervention group because the placebo group pain levels changed appreciably over the four days, whereas the second intervention group showed very little change in pain over the four days (20). Consequently we expected responsiveness to be different within these two groups, and felt that this anchor-based approach to assessing responsiveness was appropriate (21–23). Assessments of reliability and validity were done with both the placebo and intervention groups combined.
To describe and compare baseline demographics for the placebo and intervention groups, we used means, medians, standard deviations (sd), and percentages, as appropriate, with the corresponding t-tests or χ2 tests. Reliability was assessed using Cronbach's α to evaluate the internal consistency of the questions within each measure. Values of 0.7 or higher are considered acceptable. Typically, validity is assessed using Pearson correlation coefficients. In our case, since we have four time periods, convergent and discriminant validity were assessed using the method proposed by Hamlett et al. (24) and Roy (25), which calculates an overall correlation coefficient using the paired data from all four days. The approach uses mixed linear modeling where both of the paired variables are joined into one variable and an indicator variable is used to distinguish each (varid = 1 if var1 and 2 if var2). This indicator becomes a fixed effect to build the model and a random effect to create the correlation matrix. Time (Day 1, 2, 3, or 4) is a repeated class variable nested within subject identification. The resultant correlation matrix identifies a single correlation of the two variables at the same time point which is used in this assessment and correlations between different time points. We were sufficiently powered to detect significant correlations of 0.22 or higher. Confidence intervals assumed normality, and used the Fisher transformation.
To assess internal responsiveness of the three study instruments, we calculated the standardized response means for postoperative day 1 minus postoperative day 4 in the manner described by Husted et al., (26) and we calculated the probability of change statistic P̂ as described by Zou (27). Standardized Response Means of 0–0.2 indicate no responsiveness to change, whereas values of 0.2–0.5, 0.5–0.8, and >0.8 indicate minimal, moderate, and high responsiveness, respectively. Values for P̂ range from 0.5 (which indicates no ability to detect change) to 1.0 (which indicates perfect ability to detect change). If the P̂ confidence intervals did not contain 0.5 (indicating at least some responsiveness), pairwise comparisons of consecutive days were done to see where significant change occurred. We used the nonparametric Wilcoxon's ranked sum test because the data had a nonnormal distribution, and because the NRSM is measured on an ordinal scale. The study was sufficiently powered to detect the following changes in our measures from one day to the next: NRSM (0.8 U), SF-8 Pain (0.3 U), SF-8 PCS (2.1 U), SF-8 MCS (2.8 U), QoR Pain Composite (0.8 U), QoR physical independence (1.0 U), and QoR Emotional State (1.4 U). We also believe these changes represent clinically meaningful improvements based on how these translate into changes in response patterns to the individual questions.
For all analyses, we used SPSS for Windows version 13.0 (SPSS, Chicago, IL) and Stata version 9.0 (StataCorp, College Station, TX).
Of the 270 patients who were enrolled in the original clinical trial, 233 were retained for follow-up, as described earlier (20), and 154 were included in the current study (placebo and one intervention group).
Participant Characteristics and Baseline Scores
Of the 154 study participants, 44% were women and 91% were Caucasian. The proportions who completed the NRS, SF-8, and QoR on all four postoperative days were 80%, 87%, and 94%, respectively. There were no significant differences between the placebo and intervention group in terms of demographic data (Table 1) or baseline scores on the pain, physical function, and mental function measures.
Although reliability of the SF-8 and QoR composites are reported elsewhere (13,16), we wanted to assess whether the internal consistency of the items within the composites held true for our sample of ACLR outpatients. Cronbach's α for the SF-8 was 0.73 on Day 1, 0.78 on Day 2, 0.81 on Day 3, and 0.84 on Day 4. All values were above 0.7, and the increasing pattern over the four days is consistent with patients in the sample recuperating over time. Across the four days, each QoR composite had the following coefficient α values: Phys Comf (0.77, 0.76, 0.78, 0.80), Emot State (0.79, 0.83, 0.84, 0.84), Phys Indep (0.62, 0.60, 0.58, 0.53), Psy Indep (0.80, 0.86, 0.86, 0.91), and Pain Comp (0.60, 0.50, 0.66, 0.64). The lower values for QoR Phys Indep were primarily driven by the “able to return to work” question, which is a more stringent criterion than the other questions. Questions assessing pain in particular nonsurgical areas (mouth, throat, and head) tended to reduce α more for QoR Pain Comp than did the overall severe or moderate pain questions.
For the five pain measures, the correlations were in the expected directions, and were significantly different from 0, indicating overall convergent validity (Table 2). All measures of pain were highly correlated, regardless of postoperative day. Because the QoR Pain Comp, QoR Moderate Pain, QoR Severe Pain are all part of the QoR composite score, these measures had the highest correlations.
For the three physical function measures, the scores were positively correlated, but only moderately so. The SF-8 PCS and QoR Phys Indep correlation was 0.52 (95% CI: 0.40, 0.63), the correlation of SF-8 PCS and QoR Phys Comf was 0.40 (95% CI: 0.26, 0.53), and the QoR Phys Indep and QoR Phys Comf correlation coefficient was 0.58 (95% CI: 0.46, 0.67). For the three mental function measures, the scores exhibited moderate to small correlation. The correlation of SF-8 MCS with QoR Emot State was 0.60 (95% CI: 0.49, 0.69) and the correlation of QoR Emot State with QoR Psy Indep was 0.43 (95% CI: 0.30, 0.55). The lower correlation of 0.32 (95% CI: 0.17, 0.45) for SF-8 MCS with QoR Psy Indep implies that these instruments are measuring different aspects of mental health.
Valid questions or instrument composites should show low correlation with questions or composites assessing different dimensions. Although the results of discriminant correlation coefficients are too numerous to report, we highlight a few. The correlation of SF-8 PCS and QoR Psy Indep was 0.23 [95% CI: 0.08, 0.37], the NRSM and SF-8 MCS correlation was −0.20 [95% CI: −0.35, −0.04], and the correlations of NRSM with QoR Psy Indep was −0.28 [95% CI: −0.43, −0.14].
Table 3 displays the responsiveness results for the measures in our study. In this evaluation, we compared the score for each measure on postoperative day 1 with that on postoperative day 4. Our premise was that we would see responsiveness on the pain measures for the placebo group (in which pain decreased from moderate to mild) but not for the intervention group (in which pain remained mild throughout the four days). We would expect some responsiveness on the physical and mental function measures for both groups. This is because the longer the patients are at home and having decreased pain or sustained mild pain, the easier it is for them to resume everyday life activities after surgery. Our goal was to make sure the SF-8 measures are as responsive as the other measures.
For pain measures in the placebo group (Table 3), the NRSM and SF-8 Pain were moderately responsive to three-day change; the QoR Moderate Pain question showed minimal responsiveness, and the QoR Pain Comp and QoR Severe Pain question showed no responsiveness. In the placebo group, the NRSM, SF-8, and QoR Moderate Pain showed significant improvement by Day 3 (P < 0.001, P < 0.001, P = 0.002). In the intervention group, there was little variability in measures of pain, as indicated by the lack of responsiveness.
For physical function measures in the placebo group (Table 3), the SF-8 PCS was highly responsive to change; the QoR Phys Indep had slightly lower responsiveness values than did the SF-8 PCS, and the QoR Phys Comf had considerably lower values. The SF-8 PCS and QoR Phys Indep showed significant improvement each day (all P values <0.004), and leveling off by Day 4. The QoR Phys Comf did not show improvement until Day 3 (P = 0.002). In the intervention group, SF-8 PCS and QoR Phys Indep were highly responsive. SF-8 PCS recognized improvement by Day 3 (P < 0.001), whereas QoR Phys Indep saw significant improvements each day (all P values <0.001). Although QoR Phys Comf showed only moderate responsiveness values, it demonstrated significant improvement by Day 3 (P < 0.001).
For mental function measures (Table 3), the SF-8 MCS showed no responsiveness for either the placebo or intervention group. QoR Emot State was moderately responsive in both groups, demonstrating statistical improvement between Days 2 and 3 (P < 0.001). The QoR Psy Indep had low SRMs and P̂ values for both groups, showing only initial improvement between Days 1 and 2 (P = 0.05, placebo; P = 0.004 intervention).
The 154 participants in this study were part of a clinical trial in which patients were randomized to receive placebo saline versus levobupivacaine infusion via a perineural femoral catheter to reduce pain after undergoing ACLR surgery with spinal anesthesia. We evaluated five measures of pain, three measures of physical function, and three measures of mental function. Our measures demonstrated acceptable reliability and validity, but differing levels of responsiveness.
Measures of Pain
The NRSM and the SF-8 question about bodily pain (SF-8 Pain) yielded similar results. In the placebo group, the response patterns of the two measures across the four days were almost the same. In fact, although NRSM has a 10-point scale and SF-8 Pain has a 6-point scale, when we recoded the NRSM scale into six categories (0 = 1; 1–2 = 2; 3–4 = 3; 5–7 = 4; 8–9 = 5; and 10 = 6), we found that the distributions of scores for Days 1 and 4 were nearly identical (Fig. 1). The responsiveness of the two measures was identical. With either measure, the placebo group showed moderate responsiveness and the intervention group showed low responsiveness. These findings were consistent with our expectations.
The QoR pain questions and composite score did not provide useful daily assessments of short-term outpatient pain in our study participants. Questions about moderate and severe pain (QoR Moderate Pain and QoR Severe Pain) did not distinguish participants in the placebo group from those in the intervention group, primarily because the two measures concern the duration of different types of pain during the course of a day, rather than the presence and magnitude of pain. The QoR Pain Comp incorporates items addressing muscle pain, headache, and backache, which we believe would be unlikely to differ in individuals undergoing the same surgical procedure (ACLR) with a standardized anesthesia care protocol. In the placebo group, the pattern of responses to QoR Pain Comp was not similar to that of NRSM and SF-8 Pain, and the responsiveness of QoR was lower than that of NRSM and SF-8 Pain.
Based on the above assessment, NRSM and SF-8 Pain are the preferred measures over the QoR to assess pain on a short-term daily basis, when regional anesthesia in outpatient surgery is being evaluated. The two measures similarly distinguish pain across days and have the same responsiveness. The introduction of a general anesthetic technique with an airway device, superimposed upon a regional technique such as a nerve block, seems likely to introduce enough outcome variability rendering the QoR more valuable, but confirmatory research is required.
Measures of Physical Function
The SF-8 PCS and the QoR Phys Indep behaved similarly for the placebo and intervention groups. Whereas both measures focus on the resumption of daily physical activities, QoR Phys Indep asks about more detailed tasks, such as washing, brushing teeth, and grooming. It makes sense that both groups would show improvements in scores for these measures. The placebo group had progressively less pain over the four days, so patients in this group were probably able to gradually increase their activities of daily living. The intervention group experienced low pain from the start, so patients in this group were probably able to quickly resume more activities of daily living.
The QoR Phys Comf demonstrated a similar pattern of short-term change in the two groups, namely, improvement in the middle of the four-day period. This measure focuses on improvements in sleep patterns and alleviation of some of the short-term effects (e.g., dizziness, nausea, vomiting, retching, appetite loss, and restlessness) that occur after surgery and anesthesia. The fact that the QoR Phys Comf was only moderately responsive to change from Day 1 to Day 4 was primarily because all of the questions were not pertinent to our time period. Previously, we reported that the incidence of postdischarge nausea on postoperative days 1–4 was a low 2%–6% (28).
These results indicate that QoR Phys Indep and SF-8 PCS provide about the same information. However, QoR Phys Comf captures additional information about side effects and may provide valuable insights into when these problems subside in patients undergoing outpatient surgery with anesthesia. If the SF-8 was chosen to be the outcome measure for an anesthesia-related study, a coadministered measure of nausea, vomiting, and other comfort-related outcomes, provided by measures such as the Symptom Distress Scale (29–30), could be an option.
Measures of Mental Function
In both the placebo and intervention groups, QoR Psy Indep showed early change, and demonstrated minimal-to-moderate responsiveness to change from Day 1 to Day 4. This makes sense, because this measure addresses support received in the hospital, and would not appear to apply to patients once they have been sent home unless the items were rephrased to reflect home care.
In both groups, the SF-8 MCS and the QoR Emot State showed little change in actual scores across the four days and responsiveness was lower for the placebo group than for the intervention group. Both SF-8 MCS and QoR Emot State are heavily weighted by how emotional problems affect the activities of daily living, and these problems would most likely diminish when pain is low. SF-8 MCS proved to be irrelevant for patients in our study, but research is needed to indicate if the same is true for patients receiving general anesthesia and an airway device (or in comparisons of general-airway device patients versus regional-only patients). The QoR Emot State composite asks general questions about anxiety, depression, and anger that a patient feels each day, which is expected to decrease as the patient gets better. The difficulty, however, would be in determining whether a decrease in anxiety, depression, and anger (as measured by the QoR) would be due to the dissipating effects of anesthesia, or to general euphoria over the apparent success of surgery.
Two potential limitations of this study deserve mention. First, the study participants all underwent the same surgical procedure at the same medical center. All underwent ACLR with spinal anesthesia and received a perineural femoral catheter, although some received perineural local anesthetic and others received placebo saline. ACLR is commonly performed on an outpatient basis, but a few centers still routinely admit patients to the hospital overnight. The homogeneity of our patient population may limit the generalizability of our results to patients undergoing other surgical procedures. Second, we limited our study to 11 outcome measures, and opted not to study additional measures reported in the anesthesia literature. In 2000, when we proposed the study to the National Institutes of Health, we chose to evaluate the SF-8 because the SF health surveys appeared to have the best psychometric track record for general health status, and the SF-8 was the first-available SF instrument that could be administered every 24 h. We also chose to evaluate the QoR because it appeared to be the most rigorously tested instrument in 2000. Its use has been documented several times since then (14,15). It does not appear that other anesthesia-related outcome instruments have been used in more than one study (4–5), with the exception of the Iowa Satisfaction with Anesthesia Scale (2,31) and the Symptom Distress Scale (29–30).
To summarize, for patients undergoing outpatient knee surgery with regional anesthesia (and without general anesthesia), the use of the SF-8 appears to adequately assess pain, general physical function, and independence. The QoR Phys Comf composite, which includes questions about side effects such as nausea and vomiting, may add additional valuable information to populations other than those we sampled. Although our subjects showed no responsiveness across the four days for the SF-8 MCS, this does not preclude this measure from being used to assess short-term mental functions for other types of outpatients undergoing anesthesia treatments. Future studies are needed to compare findings in patients undergoing different types of knee surgeries (e.g., total knee arthroplasty versus minimally invasive unicondylar knee arthroplasty), different types of outpatient surgeries, and different anesthesia plans (e.g., use of inhaled volatile anesthetic gases versus use of regional anesthetic techniques with propofol sedation). We suspect that the SF-8 may logically be coadministered with the Symptom Distress Scale to yield a lower burden to respondents (and for data analysts) than does the QoR, but proper psychometric evaluation of SF-8 plus Symptom Distress Scale, versus QoR, is needed. Finally, one may consider testing these described instruments measuring other components of postoperative care, particularly in response to follow-up surgical office visits, and physical therapy visits.
We would like to acknowledge the teamwork provided by enrolling anesthesiologists Michael L. Kentor, MD, Raymond B. Schwartz, MD, and Steven L. Orebaugh, MD (University of Pittsburgh Medical Center—South Side Hospital Department of Anesthesiology, Pittsburgh, PA). We also wish to thank the surgeons from the University of Pittsburgh, Department of Orthopedic Surgery (Pittsburgh, PA), Center for Sports Medicine, who allowed us to enroll their patients: Drs. Freddie H. Fu, Christopher D. Harner, Robin V. West, Patrick J. McMahon, and Craig H. Bennett. We also wish to acknowledge previous research coordinators for this study based at the University of Pittsburgh: Chiara M. Figallo, MLIS, and Kimberly A. Francis, MS, MPA; and offer special thanks to the current and former Directors of Orthopaedic Clinical Research (University of Pittsburgh), James J. Irrgang, PhD, PT, ATC, and Molly T. Vogt, PhD, Dr.P.H.
Nerve stimulation needles (Prolong PL-50) were provided by Spinal Specialties, Inc., San Antonio, TX; Life-Tech®, Inc., Stafford, Texas; and I-Flow Corporation, Lake Forest, California. Elastomeric nerve block infusion devices were provided by McKinley Medical, Wheat Ridge, Colorado. Patient samples of rofecoxib were provided by Merck & Co., Inc., Whitehouse Station, New Jersey.
1. Kitz DS, Slusarz-Ladden C. Hospital resources used for inpatient and ambulatory surgery. Anesthesiology 1998;69:383–6
2. Dexter F, Aker J, Wright WA. Development of a measure of patient satisfaction with monitored anesthesia care: the Iowa Satisfaction with Anesthesia Scale. Anesthesiology 1997;87: 865–73
3. Myles PS, Hunt JO, Nightingale CE, Fletcher H, Beh T, Tanil D, Nagy A, Rubinstein A, Ponsford JL. Development and psychometric testing of a quality of recovery score after general anesthesia and surgery in adults. Anesth Analg 1999;88:83–90
4. Hogue SL, Reese PR, Colopy M, Fleisher LA, Tuman KJ, Twersky RS, Warner DS, Jamerson B. Assessing a tool to measure patient functional ability after outpatient surgery. Anesth Analg 2000;91:97–106
5. Swan BA, Maislin G, Traber KB. Symptom distress and functional status changes during the first 7 days after ambulatory surgery. Anesth Analg 1998;86:739–45
6. Inan UU, Sivaci RG, Ermis SS, Ozturk F. Effects of fentanyl on pain and hemodynamic response after retrobulbar block in patients having phacoemulsification. J Cataract Refract Surg 2003;29:1137–42
7. Pastuovic MN, Cohen ME, Burton RG. Propofol: an alternative general anesthetic for outpatient oral surgery. J Oral Maxillofac Surg 1996;54:943–8
8. Hassenbusch SJ, Pillay PK, Magdinec M, Currie K, Bay JW, Covington EC, Tomaszewski MZ. Constant infusion of morphine for intractable cancer pain using an implanted pump. J Neurosurg 1990;73:405–9
9. Ohnhause AR. Methodological problems in the measurement of pain: a comparison between the verbal rating scale and the visual analog scale. Pain 1975;1:379–84
10. Williams A, Hogart B. Pain: a review of three commonly used pain rating scales. Issues Clin Nurs 2005;14:798–804
11. Gagliese L, Weizblit N, Ellis W, Chan VW. The measurement of postoperative pain: a comparison of intensity scales in younger and older surgical patients. Pain 2005;117:412–20
12. Simanski CJ, Maegele LG, Lefering R, Lehnen DM, Kaewl N, Riess P, Yucel N, Tiling T, Bouillon B. Functional treatment and early weightbearing after an ankle fracture: a prospective study. J Orthop Trauma 2006;20:108–14
13. Ware JE, Kosinski M, Dewey JE, Gandek B. How to score and interpret single-item health status measures: a manual for users of the SF-8 health survey. Lincoln, RI: Quality Metric Inc., 2001
14. Myles PS, Hunt JO, Fletcher H, Solly R, Woodward D, Kelly S. Relation between quality of recovery in hospital and quality of life at 3 months after cardiac surgery. Anesthesiology 2001;95:862–7
15. Leslie K, Troedel S, Irwin K, Pearce F, Ugoni A, Gilles R, Pemberton E, Dharmage S. Quality of recovery from anesthesia in neurosurgical patients. Anesthesiology 2003;99:1158–65
16. Myles PS, Weitkamp B, Jones K, Melic J, Hensen S. Validity and reliability of a postoperative quality of recovery score: the QoR-40. Br J Anaesth 2000;84:11–15
17. Wu CL, Naqibuddin M, Rowlingson AJ, Lietman SA, Jermyn RM, Fleisher LA. The effect of pain on health-related quality of life in the immediate postoperative period. Anesth Analg 2003;97:1078–85
18. Wurm WH, Concepcion M, Sternlicht A, Carabuena JM, Robelen G, Goudas LC, Strassels SA. Carr DB. Preoperative interscalene block for elective shoulder surgery: loss of benefit over early postoperative block after patient discharge to home. Anesth Analg 2003;97:1620–6
19. Turner-Bowker DM, Bayliss MS, Ware JE Jr, Kosinski M. Usefulness of the SF-8 Health Survey for comparing the impact of migraine and other conditions. Qual Life Res 2003;12:1003–12
20. Williams BA, Kentor ML, Vogt MT, Irrgang JJ, Bottegal MT, West RV, Harner CD, Fu FH, Williams JP. Reduction of verbal pain scores after anterior cruciate ligament reconstruction with 2-day continuous femoral nerve block. Anesthesiology 2006;104: 315–27
21. Lydick E, Epstein R. Interpretation of quality of life changes. Qual Life Res 1993;2:221–6
22. Lydick EG, Epstein RS. Clinical significance of quality of life data. In: Spilker B, ed. Quality of life and pharmacoeconomics in clinical trials. 2nd ed. Philadelphia: Lippincott-Raven, 1996: 461–5
23. Crosby RD, Kolotkin RL, Williams GR. Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol 2003;56:395–407
24. Hamlett A, Ryan L, Wolfinger R. On the use of PROC MIXED to estimate correlation in the presence of repeated measures. SAS Users Group International, Proceedings of the Statistics and Data Analysis Section. Paper 198–29, 1–7
25. Roy A. Estimating correlation coefficient between two variables with repeated observations using mixed effects model. Biometrical J 2006 48:2:286–301
26. Husted JA, Cook DJ, Farewell VT, Gladman DD. Methods for assessing responsiveness: a critical review and recommendations. J Clin Epidemiol 2000;53:459–68
27. Zou GY. Quantifying responsiveness of quality of life measures without an external criterion. Qual Life Res 2005;14:1545–52
28. Williams BA, Kentor ML, Irrgang JJ, Bottegal MT, Williams JP. Nausea, vomiting, sleep, and restfulness upon discharge home after outpatient anterior cruciate ligament reconstruction with regional anesthesia and multimodal analgesia/antiemesis. Reg Anesth Pain Med 2007;32:193–202
29. Zhao SZ, Chung F, Hanna DB, Raymundo AL, Cheung RY, Chen C. Dose-response relationship between opioid use and adverse effects after ambulatory surgery. J Pain Sympt Manage 2004;28:35–46
30. Apfelbaum JL, Gan TJ, Zhao S, Hanna DB, Chen C. Reliability and validity of the perioperative opioid-related symptom distress scale. Anesth Analg 2004;99:699–709
31. Fung D, Cohen M, Stewart S, Davies A. Can the Iowa Satisfaction with Anesthesia Scale be used to measure patient satisfaction with cataract care under topical local anesthesia and monitored sedation at a community hospital? Anesth Analg 2005;100:1637–43