A Single Set of Numerical Cutpoints to Define Moderate and Severe Symptoms for the Edmonton Symptom Assessment System

Article Outline

• Abstract
• Introduction
• Methods
  • Sample and Settings
  • Instruments and Procedures
  • Data Analysis
• Results
  • Demographics
  • Symptom Prevalence
  • Pain
  • Other Symptoms
• Discussion
• Conclusions
• Appendix. Palliative Performance Scale Version 2 (PPSv2)
• References
• Copyright

Abstract 

Symptom intensity in cancer and palliative care patients is frequently assessed using a 0–10 ranking score. Results are then often grouped into verbal categories (mild, moderate, or severe) to guide therapy. Numerical cutpoints separating these categories are often variable, with previous work suggesting different cutpoints across different symptoms, which is unwieldy for clinical use. The Edmonton Symptom Assessment Symptom (ESAS) assesses nine common symptoms using this 0–10 scale. The primary aim of this study was to examine the relationship between the numerical and verbal scores using the ESAS and to identify a single cutpoint to separate severe from nonsevere symptomatology. A second goal was to similarly identify a cutpoint to separate moderate or severe from none or mild symptom intensity. Consenting patients (n=400) completed both a standard ESAS and an identical form that replaced 0–10 with none, mild, moderate, and severe. Receiver operating characteristic curves were generated to identify the best fit between sensitivity and specificity. For the “severe” ranking, six symptoms had a best fit of 7, with sensitivity for the remaining three symptoms still greater than 80%. For the combined grouping of moderate or severe, results were less uniform. A cutpoint of either 4 or 5 would be supported by our data, with a greater sensitivity using 4 and improved specificity using 5 as the cutpoint. Across all ESAS symptoms, then, 7 or higher represents a severe symptom by patient definition, whereas a cutpoint of either 4 or 5 could reasonably define combined moderate and severe symptoms.

Key Words: ESAS, symptom intensity, palliative care, cutpoints

Back to Article Outline

Introduction 

Palliative care and cancer patients experience a wide array of disease- and treatment-related symptoms throughout the course of their illness, resulting in an ongoing need to improve both identification of these symptoms and communication about them. Symptom assessment tools have been developed to help identify burdensome symptoms and to assess the success of their management. These tools vary in clinical focus from comprehensive symptom and functional assessments to in-depth analyses of single symptoms.¹ One tool devised and validated for rapid symptom identification and monitoring with minimal patient burden is the Edmonton Symptom Assessment System (ESAS).², ³

The ESAS is a patient-rated numerical scale consisting of 10 symptoms (pain, fatigue, nausea, depression, anxiety, drowsiness, appetite, sense of well-being, shortness of breath, and “other” symptom) evaluated on an 11-point scale (0=no symptom and 10=worst possible symptom). It has been used predominately in cancer and palliative care, although it also has been validated in dialysis patients and in intensive care settings.⁴, ⁵, ⁶, ⁷ Patients rate the severity of each symptom at the time of assessment by circling the appropriate number. Interpretation of the number circled can be challenging though, as little is known about what meaning patients attach to any particular numerical rating. Previous published studies have looked at establishing numerical cutpoints to divide the 11-point scale into groupings of mild, moderate, and severe symptoms, both to guide the urgency of therapy and to simplify communication about symptoms.⁸ Specifically, cutpoints have been proposed for pain, fatigue, depression, anxiety, and anorexia,⁹, ¹⁰, ¹¹, ¹², ¹³, ¹⁴, ¹⁵ but there is little uniformity in these recommendations, with definitions of moderate or clinically important symptom intensity, for example, ranging from 2¹¹ to 7.¹⁵

From a clinical perspective, having a series of different critical cutpoints for different symptoms is unwieldy and unlikely to be of clinical utility. To our knowledge, there are no published studies examining cutpoints across all ESAS symptoms, and the current study was designed in an effort to find a single set of patient-defined cutpoints that would reliably identify first, severe level symptoms, and second, the combined grouping of moderate or severe intensity symptoms.

Back to Article Outline

Methods 

Sample and Settings 

This prospective longitudinal design recruited 400 patients from the Palliative Care Clinic and the Rapid Response Radiotherapy Program in the Odette Cancer Centre and inpatient referrals to the Palliative Care Consult Team at Sunnybrook Health Sciences Centre between September 2006 and June 2008. Participants were older than 18 years, were able to speak English, and provided informed consent. This study was approved by the Research Ethics Board at Sunnybrook Health Sciences Centre.

Instruments and Procedures 

After providing informed written consent, participants were asked to rate their symptom distress using the ESAS by circling the appropriate number for each symptom. They then scored each ESAS item again on a new page, using a verbal scale with options of none, mild, moderate, or severe. Physicians or research assistants rated participants' functional status using the second version of the Palliative Performance Scale (PPS) (Appendix) and collected demographic information.

Data Analysis 

Patients' demographic and disease data were assessed using descriptive statistics and frequency distributions. Sensitivity and specificity were calculated for symptoms ranked severe and for the combined group of symptoms ranked moderate or severe for each ESAS symptom. Receiver operating characteristic (ROC) curves were used to identify optimum cutpoints.

Back to Article Outline

Results 

Demographics 

A total of 770 patients were screened, with 166 deemed ineligible because of impaired cognition or decreased level of consciousness and 137 ineligible because of language barriers. Of the 467 patients approached about participation, 67 declined, leaving 400 who completed the study. The median age of participants was 60 years (range 22–95), and 58.5% were female. The median PPS was 70 (range 10–100). The most prevalent malignant diagnoses were breast (n=77), lung (n=63), and gastrointestinal (n=60) cancers. Patient demographics are summarized in Table 1.

Table 1. Patient Characteristics (n=400)

Demographics	n (%)
Gender
Male	166 (41.5%)
Female	234 (58.5%)
Age, years
Median (range)	60.0 (22–95)
PPS
Median (range)	70 (10–100)
Primary cancer site
Breast cancer	77 (19.25%)
Lung cancer	63 (15.8%)
GI cancera	60 (15.0%)
GU/prostate cancer	46 (11.5%)
Hematological cancer	38 (9.5%)
Gynecological cancer	34 (8.5%)
Othersb	77 (19.75%)
Nonmalignant	5 (1.25%)

PPS=Palliative Performance Scale; GI=gastrointestinal.

Symptom Prevalence 

Table 2 summarizes the prevalence of each symptom by word rankings and the median ESAS score within each category. There was a very low prevalence of severe nausea (n=10) and shortness of breath (n=14) in this study.

Table 2. Symptom Prevalence by Word Ranking

	None (n)	Mild (n)	Moderate (n)	Severe (n)
Pain	55	128	150	67
Median ESAS score for each word option	0	3	5	8
Tired	35	140	158	67
Median ESAS score	1	4	7	8
Nausea	284	73	33	10
Median ESAS score	0	2	6	7
Depression	180	126	71	23
Median ESAS score	0	3	6	8
Anxiety	153	120	99	28
Median ESAS score	0	3	6	8
Drowsiness	113	131	114	42
Median ESAS score	0	2	6	8
Loss of appetite	142	99	86	73
Median ESAS score	1	4	6	10
Well-being	84	131	128	56
Median ESAS score	1	4	6	9
Shortness of breath	240	99	47	14
Median ESAS score	0	3	6	9

Pain 

The distribution of results for pain is shown in Fig. 1, with the bar graph showing the breakdown of verbal rankings for each numerical ESAS score. For each number on the ESAS scale, sensitivity and specificity were calculated to determine the best cutpoint for the word “severe.” ROC curves were used to identify the best balance between sensitivity and specificity. An ROC curve (Fig. 2) plots sensitivity vs. 1−specificity, with an ideal curve showing perfect sensitivity and specificity and having an area under the curve (AUC) of 1.0. The data point closest to the upper left corner of an ROC curve will represent the best fit between sensitivity and specificity and can be calculated as the distance from the optimum. Using distance from optimum on the ROC curve (Fig. 2), the best cutpoint to differentiate “severe” pain from “not severe” pain for our study was 7, meaning any score of 7 or higher represented severe pain, with a sensitivity of 89.6% and specificity of 79.6% (Table 3).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1 

  Distribution of pain responses.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2 

  ROC curve for severe pain.

Table 3. Sensitivity, Specificity, and Distance From Optimum: Severe Ranking

CI=confidence interval.

Similarly, to find the cutpoint separating combined moderate or severe pain from mild or no pain, a cutpoint of 5 yielded the optimum cutpoint by ROC criteria, though lowering the cutpoint to 4 improved sensitivity from 84.3% to 90.3% (Table 4).

Table 4. Sensitivity, Specificity, and Distance From Optimum: Moderate Plus Severe

Other Symptoms 

Using a similar approach, sensitivity, specificity, and ROC curves were generated for each of the remaining eight symptoms for both “severe” and “moderate or severe” symptom intensity (Table 3, Table 4). For the “severe” category, similar to “pain,” anxiety, drowsiness, loss of appetite, and well-being all had an optimum cutpoint of 7. For tiredness, the best fit was a cutoff of 8, but this sacrificed sensitivity to 74.6%. Nausea showed a best-fit number of 5, but this was based on only 10 data points. Similarly, shortness of breath, which had only 14 respondents who ranked the symptom “severe,” had a lower best cutpoint at 6.

The area under the ROC curve for eight of the nine symptoms was 0.90 or higher (0.85 for tiredness), showing excellent agreement between the verbal and numerical scales. There was greater variability in the best fit for moderate plus severe cutpoint for the nine symptoms, with nausea, depression, and shortness of breath having a best value of 4, whereas pain, anxiety, drowsiness, and loss of appetite all had a best value of five. Both well-being and tiredness appeared to have an optimum value of 6, though in both cases, this resulted in a significant fall in sensitivity (75.1 and 66.3, respectively). Again, the AUC showed excellent test performance, with six of nine areas showing 0.90 or higher and the remaining three 0.80 or higher.

Back to Article Outline

Discussion 

Although numerical scales are frequently used in assessing symptom severity, both patients and health care professionals typically communicate using word substitutes, and hence, it is useful to know how these words correspond with the numbers generated on symptom scales. Further, guidelines for symptom management generally divide symptom intensity into categories of mild, moderate, and severe to direct therapy.¹⁶, ¹⁷, ¹⁸, ¹⁹, ²⁰ Previous authors have compared 0–10 symptom severity rankings with other validated symptom assessment tools with the goal of identifying a cutpoint that would indicate presence of clinically significant symptomatology. The goal of the current study was to compare a patient's verbal description of symptoms on the ESAS (using none, mild, moderate, and severe) with the standard numerical rankings, to identify a single numerical cutpoint to differentiate symptom intensity across all ESAS symptoms, and further, to see how these cutpoints compared with those from previous studies.

For all symptoms at both severe and moderate plus severe rankings, the AUC for the ROC curves was greater than 0.80 and, for most cases, it was greater than 0.90, suggesting that the words and numbers were indeed strongly related. For the severe ranking, a cutpoint of 7 was ideal for six of nine symptoms. However, it is cumbersome both clinically and for research purposes to have differing cutpoints for multiple symptoms, and ideally, a single cutpoint would be suitable across all nine symptoms. In examining the remaining three symptoms (shortness of breath, nausea, tiredness), sensitivity and specificity remain excellent with the cutpoint set at 7, and our data would, therefore, support a single cutpoint of 7 to differentiate “severe” from “not severe” symptom intensity for each of the nine ESAS symptoms.

For the combined category of moderate plus severe, the results were less consistent, with an identified cutpoint of 4 for nausea, depression, anxiety, and shortness of breath; 5 for pain, drowsiness, and loss of appetite; and 6 for tiredness and well-being. However, accepting a cutpoint of 6 for tiredness and well-being yields low test sensitivity at 75.1% and 66.3%, respectively. Given that one key purpose in using symptom assessment tools is identification of patient suffering, this loss of sensitivity is problematic, and a lower cutpoint with its improved sensitivity is more clinically useful. Similarly, one can argue that if the goal were a single cutpoint that operates well across all symptoms, the lower score with its improved sensitivity would serve the patient population better, as there will be fewer false negatives. Finally, for the symptoms with a low prevalence of moderate or severe rankings (e.g., nausea, shortness of breath), it is possible that lower numbers may, in fact, still represent significant symptom burden, again arguing for choice of the cutpoint at 4 rather than 5 to identify moderate or severe symptom severity.

Looking at previous cutpoint work in cancer patients, there are data available on pain,⁹, ¹², ¹³, ¹⁵ fatigue,⁹, ¹⁴ depression,⁹, ¹⁰, ¹¹ anxiety,⁹, ¹¹ and anorexia⁹ to help further determine ideal cutpoints. Each study had different goals and methods, and there were an array of proposed cutpoints ranging from 2 to 6 (summarized in Table 5) for identifying either moderate level or clinically important symptom levels.

Table 5. Comparison of Cutpoints in Previous Studies of Cancer Patients for Presence of Clinically Important or Moderate Symptom Levels

x = cutpoint proposed by cited study.

Butt et al.⁹ asked patients to rank pain, fatigue, distress, and anorexia on a 0–10 scale, recording the most severe level they had experienced in the past three days, and compared these values with other validated symptom scores (e.g., Functional Assessment of Cancer Therapy-General [FACT-G], FACT-fatigue subscale, Brief Pain Inventory [BPI], and others). Furthermore, patients were asked to record whether an improvement in the target symptom would improve their quality of life (QOL). Interestingly, the derived cutpoints for the five symptoms (distress was subdivided into anxiety and depression) ranged from 4 to 6, similar to our results for moderate plus severe symptoms. For each of the cutpoints established, greater than 85% of patients with that symptom reported that relief would improve their QOL, confirming the clinical importance of the symptom. These data, along with data from our study, suggest that the cutpoint representing moderate plus severe symptom intensity can be interpreted as the presence of clinically important symptom levels and that an ideal cutpoint is between 4 and 6 for symptom detection.

Although Butt et al.⁹ used comparison with validated tools to establish cutpoints, other authors have used interference with function to define mild, moderate, and severe symptom intensity. In a study of pain in cancer patients, Serlin et al.¹² used the interference items of the BPI and established cutpoints of 1–4 for mild pain, 5–6 for moderate, and 7–10 for severe pain based on nonlinear interference with function at each level. Paul et al.,¹⁵ using a similar approach, confirmed the finding of 5 as the lower limit for moderate pain but found a cutpoint of 8 or higher for severe pain. In addition, Chow et al.,¹³ found a lower limit of 5 for defining moderate pain and 8 for severe pain in a group of patients with bone metastases. Similarly, a number of studies on nonmalignant pain have found a cutpoint of 5 or higher to define moderate pain,²¹, ²², ²³, ²⁴ with severe pain results ranging from 7²¹, ²³, ²⁴, ²⁵ to 9.²⁶

Chang et al.¹⁴ looked at fatigue in cancer patients, again in terms of its interference on function, using the Brief Fatigue Inventory (BFI) and established cutpoints for mild, moderate, and severe fatigue. They reported that “usual” fatigue (as opposed to “worst” fatigue) correlated best with the BFI interference items, as well as other tools used, including the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-C30 and the M. D. Anderson Symptom Inventory, and that the optimal values were 4–7 for moderate fatigue and 8–10 for severe fatigue.

Historically, results have varied for cutpoints in depression and anxiety. Vignaroli et al.¹¹ compared ESAS scores with the Hospital Anxiety and Depression Score (HADS) but had very few patients identified with severe symptoms (five of 216 for depression and six of 216 for anxiety by HADS criteria). They found that any score greater than 2 warranted further investigation but did not differentiate symptoms into mild, moderate, or severe categories. Lloyd-Williams et al.,¹⁰ looking specifically at depression, compared the single 0–10 score with the results of an interview using Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria and suggested that a cutpoint of greater than 3 had the best-fit sensitivity (80%) and specificity (42.6%) for the presence of depression but did not quantify its severity. Butt et al.,⁹ as shown earlier, suggested a higher cutpoint of 5 for both anxiety and depression to identify clinically important symptom severity.

In summary, previous authors have assessed numerical scores by comparison with validated assessment tools or by interference on function. Some have focused on identifying numerical levels that warrant further investigation without linking these numbers to symptom severity,⁹, ¹⁰, ¹¹ whereas others have identified cutpoints that correlate with mild, moderate, or severe levels of symptoms.¹², ¹³, ¹⁴, ¹⁵ Results of the latter group of studies yielded numerical cutpoints very similar to ours. Furthermore, our study demonstrated a strong relationship between the verbal and numerical scales (AUC consistently greater than 0.8). However, it is worth noting that although the relationship is indeed strong, there remain a number of outliers, as demonstrated in Fig. 1, with rankings that are well outside the established cutpoints (e.g., for pain, “moderate” rankings ranged from 1 to 10). This highlights the importance of using assessment scales only as a starting point for discussion of symptom experience. Overall, though, from both previous work cited and our results, a cutpoint of either 4 or 5 would be suitable to identify at least moderate symptom severity for most patients, with a cutpoint of 4 improving sensitivity if the goal is case finding and a score of 5 improving specificity. For identification specifically of severe level symptoms, our study supports a cutpoint of 7 to correspond best with the patient descriptor of severe.

Our study has a number of limitations and only begins the process of addressing the meaning patients attribute to different numerical values. Unlike previous work, we used patients' own verbal descriptor as the comparator to define cutpoints as opposed to comparing numerical values with instruments designed to measure interference with function or with tools previously validated for each symptom. However, as shown, our results were similar to the cited studies that used such scales. Our study was limited to English-speaking patients and may, therefore, not be applicable to the non-English population. In addition, the vast majority of our subjects had a malignant diagnosis and hence, the results may not be applicable to patients with nonmalignant illnesses. Furthermore, this was a heterogeneous group in terms of malignant diagnosis and stage of illness; thus, it is possible that specific subgroups of patients may have unique cutpoints different from those found in our study.

Back to Article Outline

Conclusions 

Our data suggest identifying any symptom ranked 7 or higher on the ESAS as severe and any symptom ranked 4 or higher as moderate or severe and warranting further inquiry or investigation.

Back to Article Outline

Appendix. Palliative Performance Scale Version 2 (PPSv2) 

PPS Level (%)	Ambulation	Activity and Evidence of Disease	Self-Care	Intake	Consciousness Level
100	Full	Normal activity and work No evidence of disease	Full	Normal	Full
90	Full	Normal activity and work Some evidence of disease	Full	Normal	Full
80	Full	Normal activity with effort Some evidence of disease	Full	Normal or reduced	Full
70	Reduced	Unable normal job/work Significant disease	Full	Normal or reduced	Full
60	Reduced	Unable hobby/house work Significant disease	Occasional assistance necessary	Normal or reduced	Full or confusion
50	Mainly sit/lie	Unable to do any work Extensive disease	Considerable assistance required	Normal or reduced	Full or confusion
40	Mainly in bed	Unable to do most activity Extensive disease	Mainly assistance	Normal or reduced	Full or drowsy±confusion
30	Totally bed bound	Unable to do any activity Extensive disease	Total care	Normal or reduced	Full or drowsy±confusion
20	Totally bed bound	Unable to do any activity Extensive disease	Total care	Minimal to sips	Full or drowsy±confusion
10	Totally bed bound	Unable to do any activity Extensive disease	Total care	Mouth care only	Drowsy or coma±confusion
0	Death	—	—	—	—

Back to Article Outline

References 

1. Kirkova J, Davis MP, Walsh D, et al. Cancer symptom assessment instruments: a systemic review. J Clin Oncol. 2006;24:1459–1473
   • View In Article
2. Bruera E, Kuehn N, Miller MJ, Selmser P, Macmillan K. The Edmonton Symptom Assessment System (ESAS): a simple method for the assessment of palliative care patients. J Palliat Care. 1991;7:6–9
   • View In Article
   • MEDLINE
3. Nekolaichuk C, Watanabe S, Beaumont C. The Edmonton Symptom Assessment System: a 15-year retrospective review of validation studies (1991–2006). Palliat Med. 2008;22:111–122
   • View In Article
   • CrossRef
4. Chang VT, Hwang SS, Feuerman M. Validation of the Edmonton Symptom Assessment Scale. Cancer. 2000;88:2164–2171
   • View In Article
   • CrossRef
5. Rees E, Hardy J, Ling J, Bradley K, A'Hern R. The use of the Edmonton Symptom Assessment Scale (ESAS) within a palliative care unit in the UK. Palliat Med. 1998;12:75–82
   • View In Article
   • MEDLINE
   • CrossRef
6. Nelson JE, Meier DE, Oei EJ, et al. Self-reported symptom experience of critically ill cancer patients receiving intensive care. Crit Care Med. 2001;29:277–282
   • View In Article
   • MEDLINE
   • CrossRef
7. Davison SN, Jhangri GS, Johnson JA. Cross sectional validity of a modified Edmonton Symptom Assessment System (ESAS) in dialysis patients: a simple assessment of symptom burden. Kidney Int. 2006;69:1621–1625
   • View In Article
   • MEDLINE
   • CrossRef
8. Anderson KO. Role of cutpoints: why grade pain intensity?. Pain. 2005;113:5–6
   • View In Article
   • Full Text
   • Full-Text PDF (48 KB)
   • CrossRef
9. Butt Z, Wagner LI, Beaumont JL, et al. Use of a single-item screening tool to detect clinically significant fatigue, pain, distress, and anorexia in ambulatory cancer practice. J Pain Symptom Manage. 2008;35:20–30
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (252 KB)
   • CrossRef
10. Lloyd-Williams M, Dennis M, Taylor F. A prospective study to compare three depression screening tools in patients who are terminally ill. Gen Hosp Psychiatry. 2004;26:384–389
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (263 KB)
    • CrossRef
11. Vignaroli E, Pace E, Willey J, et al. The Edmonton Symptom Assessment System as a screening tool for depression and anxiety. J Palliat Med. 2006;9:296–303
    • View In Article
    • MEDLINE
    • CrossRef
12. Serlin RC, Mendoza TR, Nakamura Y, Edwards KR, Cleeland CS. When is cancer pain mild, moderate, or severe? Grading pain severity by its interference with function. Pain. 1995;61:277–284
    • View In Article
    • Abstract
    • Full-Text PDF (735 KB)
    • CrossRef
13. Chow E, Doyle M, Li K, et al. Mild, moderate, or severe pain categorized by patients with cancer with bony metastases. J Palliat Med. 2006;9:850–853
    • View In Article
    • MEDLINE
    • CrossRef
14. Chang YJ, Lee JS, Lee CG, et al. Assessment of clinical relevant fatigue level in cancer. Support Care Cancer. 2007;15:891–896
    • View In Article
    • CrossRef
15. Paul SM, Zelman DC, Smith M, Miaskowski C. Categorizing the severity of cancer pain: further exploration of the establishment of cutpoints. Pain. 2005;113:37–44
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (168 KB)
    • CrossRef
16. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology (v.1.2008), cancer-related fatigue. Available from http://www.nccn.org/professionals/physician_gls/PDF/fatigue.pdf. Accessed November 15, 2008.
    • View In Article
17. National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology (v.1.2008), adult cancer pain. Available from http://www.nccn.org/professionals/physician_gls/PDF/pain.pdf. Accessed November 15, 2008.
    • View In Article
18. National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology (v.1.2008), Distress Management. Available from http://www.nccn.org/professionals/physician_gls/PDF/distress.pdf. Accessed November 15, 2008.
    • View In Article
19. Mid Trent Cancer Network-Palliative Care Group. Palliative care cancer pain standards, guidelines and patient information for hospitals July, 2005. Available from http://www.information4u.org.uk/files/Hospital%20Guidelines%20For%20Nottingham%20July%202005.pdf. Accessed November 30, 2008.
    • View In Article
20. World Health Organization. Cancer pain relief. Geneva, IL: World Health Organization; 1986;
    • View In Article
21. Jensen MP, Smith DG, Ehde DM, Robinsin LR. Pain site and the effects of amputation pain: further clarification of the meaning of mild, moderate and severe pain. Pain. 2001;109:317–322
    • View In Article
22. Kapstad H, Hanestad BR, Langeland N, Rustøen T, Stavem K. Cutpoints for mild, moderate and severe pain in patients with osteoarthritis of the hip or knee ready for joint replacement surgery. BMC Musculoskelet Disord. 2008;9:55–63
    • View In Article
    • CrossRef
23. Turner JA, Franklin G, Heagerty PJ, et al. The association between pain and disability. Pain. 2004;112:307–314
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (247 KB)
    • CrossRef
24. Mendoza TR, Chen C, Brugger A, et al. Lessons learned from a multiple-dose post-operative analgesic trial. Pain. 2004;109:103–109
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (116 KB)
    • CrossRef
25. Zelman DC, Dukes E, Brandenburg N, Bostrom A, Gore M. Identification of cut-points for mild, moderate and severe pain due to diabetic peripheral neuropathy. Pain. 2005;115:29–36
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (211 KB)
    • CrossRef
26. Zelman DC, Hoffman DL, Seifeldin R, Dukes EM. Development of a metric for a day of manageable pain control: derivation of pain severity cut-points for low back pain and osteoarthritis. Pain. 2003;106:35–42
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (112 KB)
    • CrossRef