Pain scores are often used to assess the adequacy of analgesia in research studies. Based on changes in the pain score or final pain score, an investigator may decide that one analgesic medication or treatment is superior to another. The numeric rating scale (NRS) is an example of a pain scale used in research studies. The NRS is administered by asking the patient to verbally estimate pain on a scale of 0 to 10, with 0 representing no pain and 10 representing the worst imaginable pain. In clinical trials of patients with rheumatoid arthritis or cancer, the NRS has been found reliable, easy for the patient to understand, and easy for the investigator to score (1–3). Jensen et al. (4), in a study evaluating six different pain scales, found the NRS to have many advantages over other pain scales, including ease of administration, and recommended using the NRS for patients with chronic pain. We find the same advantages of the NRS in the parturient and have been using the scale in our labor analgesia research studies.
During labor and delivery, one typically administers epidural pain medicine until the woman is satisfied with the degree of pain relief and does not desire more pain medication. The NRS score associated with desire for additional analgesic medication in the parturient has not been determined, although it has been established with the visual analog scale (VAS) (5). The VAS is used by asking the patient to place a mark on a 100-mm horizontal line that is labeled at one end “no pain at all” and at the other end with “worst imaginable pain.” Plummer and Brownridge (5) found a median VAS score of 19 mm (interquartile range, 4–40 mm) among patients reporting satisfactory pain relief after the first dosing of the epidural catheter. In this study, we performed a post hoc analysis of data from three previous studies that we had conducted regarding labor epidural analgesia to define the NRS scores that are associated with the desire for additional analgesic medication in the parturient (6–8).
We previously conducted three studies regarding labor analgesia in the parturient that were approved by the IRB of Mount Sinai School of Medicine, New York, NY. In these three studies, written informed consent was obtained from each patient before her enrollment in the study (6–8). The studies all involved women in active labor (uterus contracting at least once every 5 min) who requested a continuous lumbar epidural for labor analgesia. In the first study, women were randomized to receive different doses of ropivacaine (ropivacaine study) (6). In the second study, women were randomized to receive 13 mL of bupivacaine 0.25% either while they were positioned supine with a 30 degrees left lateral tilt or while in the left lateral decubitus position (position study) (7). In the third study, women were randomly assigned to one of two groups depending on the medium used for the loss-of-resistance technique (air or 0.9% saline) to identify the epidural space, followed by 13 mL of bupivacaine 0.25% (air versus saline study) (8). In all of these studies, women were asked to quantify their pain level on a 0–10 pain scale (NRS) before and after the intervention. The women were told that a score of 0 represents no pain and a score of 10 represents the worst imaginable pain. After the intervention, in addition to the pain scale, the woman was asked if she wanted additional analgesic medication (clinical outcome).
All data were entered into an Excel database and converted to a SAS file (SAS/STAT User’s Guide, Version 6, SAS Institute Inc, Cary, NC) for statistical analysis (1). A post hoc analysis of the data was performed to look for NRS scores or changes in NRS scores that are associated with patient desire for more analgesic medication. Data for treated and control patients were combined and described for each of the three studies before combining the data across studies. Although sample sizes differed among the three studies, the cases were pooled without weighting by sample sizes because all patients were volunteers from the same service in the same hospital in similar calendar years. Ordinal and categorical data were compared using the χ2 test or the χ2 test for trend. Changes in pain score were analyzed using the Fisher’s exact test. Differences were considered significant at P < 0.05.
Data were reviewed from 311 patients: 69 from the ropivacaine study, 96 from the position study, and 146 from the air versus saline study. The initial median pain scores were 6, 7, and 7 in the ropivacaine, position, and air versus saline study, respectively (Table 1).
In the individual studies, hardly any of the patients with a pain score of 0 or 1 requested additional medication. The results were also homogeneous across the three studies when the pain score was >3, with almost all of the patients wanting more medication. The proportions were intermediate between these two extremes when the pain scores were 2 or 3 but were substantially larger among the ropivacaine patients than the two other groups (Table 2).
Combining the results from all three studies, 2% of the patients with a NRS score of 0 or 1, 51% of the patients with a score of 2 or 3, and 93% of the patients with a score >3 wanted more medication. The consistency of the responses when the final pain score was 0 or 1 and when the pain score was >3, respectively, suggests that no information about wanting additional medication would be lost by grouping outcomes into three categories. Category I = a pain score of 0 or 1, Category II = a pain score of 2 or 3, and Category III = a pain score of >3, corresponding with patient desire for more medication of 2%, 51%, and 93%, respectively (Table 3).
We also sought an association between a change in pain score and desire for additional pain medication. Because very few patients wanted additional medication when the final pain score was <2 and essentially all patients wanted more pain medication when the final pain score was >3, we only analyzed the data from the patients with a final pain score of 2 or 3 to seek an association between a change in pain score and desire for additional pain medication. We combined the data from the three studies and divided the patients with a final pain score of 2 or 3 into two groups: those with a change in pain score of <5 and those with a change in pain score of ≥5. The choice of a change of 5 was somewhat arbitrary because review of the data did not identify an obvious cutoff. However, a cutoff of 5 apportioned a reasonable number of patients to both groups. We were not able to find a statistical difference in desire for additional pain medication between the two groups (P = 0.16; Table 4).
We found that the NRS pain score associated with patient desire for additional medication in the parturient is very small. More than 90% of the patients with a pain score >3 wanted more analgesic medication, approximately 50% of the patients wanted additional analgesic medication with a pain score of 2 or 3, and very few patients (2%) with a pain score of 0 or 1 wanted more medication (Table 2). We also found that the change in pain score from baseline was not as important as the final pain score.
The appropriate clinical outcome in obstetric analgesia studies varies depending on the purpose of the study. The outcome used may be the amount of pain, desire for more analgesic medication, patient satisfaction, and presence of lower-extremity motor block. A pain score allows for comparison among studies and is occasionally used as a surrogate for clinical outcomes.
A number of different scales have been developed to assess pain. Besides the NRS as we used, the verbal rating scale (VRS) and the VAS are often used. The VRS typically consists of a series of verbal pain descriptors ordered from the least to most intense. The patient chooses the descriptor that best matches his or her level of pain. The VAS is probably the most frequently used scale in research studies, including anesthesia studies. It is relatively easy to use, minimally intrusive, conceptually simple, and its ratio scale properties allow meaningful interpretation of percentage differences in VAS measurements (9). Collins et al. (10) found that the VAS correlates with a standard four-point categorical scale (none, mild, moderate, or severe pain), where a VAS score of >30 mm indicates moderate pain. We prefer not to use the VAS because the parturient experiencing labor pain does not want to be inconvenienced with marking a line or does not always understand the scale. These drawbacks of the VAS have been noted by other investigators (1,2).
Jensen et al. (4) compared six different pain scales including the NRS, VAS, and VRS in patients with chronic pain. The criteria used to compare scales were ease of administration and scoring, rates of incorrect responses, sensitivity as defined by statistical power and the number of available responses, and magnitude of the relationship between each scale and a linear combination of pain intensity indices. They found the NRS has many advantages as compared with the other scales and seems to be the most practical index. They recommended using the NRS for patients with chronic pain. We find the NRS useful for the parturient because it does not require additional paperwork and it is easy for the patient to understand and respond.
The question of how to arrive at a clinically relevant interpretation of pain scale scores has been addressed in a variety of settings, including in the parturient. Carlsson (11) evaluated patients under treatment for chronic pain and found, as we did, that posttreatment pain scores reflected treatment success better than any change in pain scores. In a study of postoperative pain measured by the VAS (12), we found that when the VAS score was ≤3, most patients did not want more analgesic medication, and when the pain score was >7, most patients wanted more medication. Only when the score was between these two extremes was the desire for more medication influenced by the amount of change in pain score from baseline. In the parturient, Plummer and Brownridge (5) found that patients who reported satisfactory pain relief had a median VAS score of 19 mm (interquartile range, 4–40).
The difference in results between this study and the one we performed in the postoperative setting, (12) in which patients were satisfied with their quality of analgesia at greater pain scores, may be related to a difference in pain scales or to the patient setting. In the study regarding postoperative pain, the VAS scale was used, and in the studies regarding labor analgesia, the NRS was used. However, it should be noted that other investigators have found an excellent correlation between NRS and VAS scores when measured concurrently in the same patients (1,4,13).
The patient setting may be a more important reason for the difference in results between this study and the one performed in the postoperative setting (12). Postoperative patients may not have the expectation of complete pain relief, nor may it be safe or easy to achieve. In the context of labor analgesia, parturients who request epidural analgesia expect almost complete pain relief, which is both safe and relatively easy to achieve in regard to labor analgesia.
Overall, pain scores that we found associated with adequate analgesia were lower than those found by Plummer and Brownridge (5). The different pain scales used might explain this difference, VAS in their study (5) versus the NRS in ours, or by the different study population, Australian (5) versus American women.
The results of the ropivacaine study were not consistent with those of the other two studies (position and air versus saline). In the ropivacaine study, unless the final pain score was 0 or 1, almost all the patients wanted additional medication, whereas in the other two studies, many patients did not want additional medication until the pain score was >3. The explanation for this difference in results is not obvious. The design of the three studies was similar. Also, all women were in active labor in the same hospital during a similar time period with similar initial pain scores, and all of them had requested epidural analgesia (Table 1). The women were all assessed 15 minutes after the study intervention in the same manner; they were asked if they wanted additional analgesic medication, and they were asked to quantify their pain on a 0–10 pain scale.
We reviewed the data to see if the larger proportion of multiparous women in the ropivacaine study could explain the difference in responses. Although it is true that in all three studies, the multiparous women were generally more likely than the nulliparous to desire additional analgesic medication, both the nulliparous and the multiparous women in the ropivacaine study with pain scores ≤3 exceeded their counterparts in the other studies. A second possible explanation is that women who volunteer to participate in research studies specifically devoted to analgesia, as in the ropivacaine study, may expect better quality of analgesia than women in studies with other end points. This hypothesis deserves further consideration. As pointed out by Feinstein (14), in discussions about features of clinical trials, the process by which patients are accrued into a trial may distinguish them from the general population. Finally, we should note that there were fewer patients in the ropivacaine study than in the other two studies (22% of the total), and we cannot exclude the possibility that the observed difference is a random event.
Even if the results of the ropivacaine study reflect a true difference from the other two studies, that does not diminish the main point of this paper that when the NRS score is the outcome measure in comparative studies, grouping the scores into categories is more useful to the clinician than using individual scores. However, it is possible that the same cutoffs may not be appropriate for all circumstances.
The results of our study suggest that when desire for additional analgesic medication is the outcome measure in pain studies about the parturient, grouping the NRS scores for analysis into 3 categories (0 or 1, 2 or 3, and >3) is more useful to the clinician than using the full spectrum of NRS scores. Grouping patients by pain scores rather than using the full spectrum of pain scores has been recommended by Serlin et al. (3) with the NRS in cancer patients and by the authors of this study (12) with the VAS in postoperative patients.
In conclusion, we found that unless the NRS score is low (0 or 1), the parturient is likely to want additional medication for labor epidural analgesia. We also found that grouping the NRS values into categories for analysis is more useful to the clinician than using the full spectrum of NRS scores. We anticipate that this study will stimulate further research into appropriate use of the NRS and other pain scores in the parturient.
We would like to express appreciation to James B. Eisenkraft, MD, Andrew B. Leibowitz, MD, and Meg A. Rosenblatt, MD, for their critical review of this manuscript.
1. Ferraz MB, Quaresma MR, Aquino LR, et al. Reliability of pain scales in the assessment of literate and illiterate patients with rheumatoid arthritis. J Rheumatol 1990; 17: 1022–4.
2. Joos E, Peretz A, Beguin S, Famaey JP. Reliability and reproducibility of visual analogue scale and numeric rating scale for therapeutic evaluation of pain in rheumatic patients. J Rheumatol 1991; 18: 1269–70.
3. Serlin RC, Mendoza TR, Nakamura Y, et al. When is cancer pain mild, moderate or severe: grading pain severity by its interference with function. Pain 1995; 61: 277–84.
4. Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain 1986; 27: 117–26.
5. Plummer JL, Brownridge P. Epidural analgesia in labour using intermittent doses determined by midwives. Int J Obstet Anesth 1998; 7: 88–97.
6. Beilin Y, Galea M, Zahn J, Bodian CA. Epidural ropivacaine for the initiation of labor epidural analgesia: a dose finding study. Anesth Analg 1999; 88: 1340–5.
7. Beilin Y, Abramovitz SE, Zahn J, et al. Improved epidural analgesia in the parturient in the 300 tilt position. Can J Anaesth 2000; 47: 1176–81.
8. Beilin Y, Arnold I, Telfeyan C, et al. The quality of analgesia when air vs. saline is used for identification of the epidural space in the parturient. Reg Anesth 2000; 25: 596–9.
9. Katz J, Melzack R. Measurement of pain. Surg Clin North Am 1999; 79: 231–52.
10. Collins SL, Moore RA, McQuay HJ. The visual analogue pain intensity scale: what is moderate pain in millimetres? Pain 1997; 72: 95–7.
11. Carlsson AM. Assessment of chronic pain. I. Aspects of the reliability and validity of the visual analogue scale. Pain 1983; 16: 87–101.
12. Bodian CA, Freedman G, Hossain S, et al. The visual analog scale for pain: clinical significance in postoperative patients. Anesthesiology 2001; 95: 1356–61.
13. DeLoach LJ, Higgins MS, Caplan AB, Stiff JL. The visual analog scale in the immediate postoperative period: intrasubject variability and correlation with a numeric scale. Anesth Analg 1998; 86: 102–6.
14. Feinstein A. The role of observational studies in the evaluation of therapy. Stat Med 1984; 3: 341–5.