Screening for Pain in the Ambulatory Cancer Setting: Is 0-10 Enough?

Asking about pain frequency and intensity and reconsidering threshold values on pain intensity scales may be practical strategies to better identify patients with cancer who have relevant pain.

Abstract

Purpose:

The purpose of this study was to explore concordance between patient self-reports of pain on validated questionnaires and discussions of pain in the ambulatory oncology setting.

Methods:

Adult, ambulatory patients (N = 452) with all stages of cancer were included. Three pain measures were evaluated: two items from the Symptom Distress Scale (frequency [SDSF] and intensity [SDSI]) and the Pain Intensity Numeric Scale (PINS). Relevant pain was defined as: (1) scores 3 of 5 on SDSF or SDSI or 5 of 10 on the (PINS); or (2) discussion of existing pain in an audio-recorded clinic visit. For each scale, McNemar's test assessed concordance of patient self-reports of relevant pain with discussions of relevant pain in the audio-recorded clinic visit. Sensitivity, specificity, and accuracy were calculated and a receiver operating characteristic analysis evaluated thresholds on self-report pain questionnaires to best identify relevant pain discussed in clinic.

Results:

Identification of relevant pain by self-report was discordant (P < .001) with discussed pain coded in audio-recorded visits for all three measures. Specificity was higher for intensity (SDSI, 0.94; PINS, 0.97) than frequency (SDSF, 0.87); sensitivity was higher for frequency (SDSF, 0.35) than intensity (SDSI, 0.24; PINS, 0.12). Accuracy was higher for the SDS pain items (SDSF, 0.57; SDSI, 0.54) than for PINS (0.48). Receiver operating characteristic analysis curves suggest that lower threshold scores may improve the identification of relevant pain.

Conclusion:

Self-report pain screening measures favored specificity over sensitivity. Asking about pain frequency (in addition to intensity) and reconsidering threshold scores on pain intensity scales may be practical strategies to more accurately identify patients with cancer who have relevant pain.

Introduction

Despite significant efforts over the past 30 years to improve its management, cancer pain remains a pervasive problem for many patients.^1,2 A recent review estimated that 62% to 86% of patients with advanced cancer experience pain, as do 24% to 60% of patients receiving active cancer treatment and 26% to 35% of cancer survivors.³ Additionally, in a survey of over 800 hospitalized patients with cancer in the United States, the mean patient-reported pain intensity score was 5.7 (on a 0 to 10 scale), and 25% of patients reported being in frequent, or constant, severe pain more than 50% of the time.⁴ Research consistently demonstrates that undertreated pain can have serious negative effects on patient and caregiver quality of life.^2,5 Key stakeholder organizations, such as the American Pain Society, the World Health Organization, and the Institute of Medicine, continue to call upon health care providers and institutions to improve pain management, including promptly recognizing distressing pain.^6–8

Barriers to providing quality cancer pain management are multifactorial and include those related to health care providers, patients, and the health care system.⁹ Optimal pain management relies foundationally on accurate pain assessment, as it is highly unlikely that an unknown or poorly measured symptom will receive proper clinical attention and intervention.^10,11 Recent research has also debated optimal threshold scores (or so-called cut-points) when screening for pain in various patient populations.^12–21 Accordingly, this report contributes to the literature by discussing two critical issues related to assessing pain in outpatients with cancer: how can we best identify patients who are experiencing potentially distressing cancer pain, and what are the most appropriate threshold scores by which to do so?

The purpose of this study was to explore the concordance between patient self-reports of pain on validated, standardized questionnaires and subsequent patient-provider discussions of relevant pain in the ambulatory oncology setting.

Methods

This secondary analysis was completed with data from a randomized trial, Electronic Self-Report Assessment for Cancer (ESRA-C II), that was conducted at two comprehensive cancer centers.⁵ Institutional review boards at the Fred Hutchinson Cancer Research Center and the Dana-Farber Cancer Institute approved the original trial. A total of 752 eligible, adult, ambulatory patients with any type and stage of cancer who started a new therapeutic regimen were enrolled. Patients reported symptoms related to quality of life (SxQOL) before commencing treatment and at specific time points throughout therapy. Approximately 6 weeks after study enrollment and treatment initiation, a regularly scheduled clinic visit between the patient, family caregiver (if present), and clinician was audio-recorded. Study procedures related to the audio-recorded visits are reported elsewhere.¹¹

Operationalization and Coding of Relevant Versus Nonrelevant Pain

Questionnaires.

Patients reported SxQOL within 24 hours before the clinic visit using an online platform either in the clinic or at home. The SxQOL assessment included previously validated self-report tools covering symptom distress,²² depression,²³ peripheral neuropathy,²⁴ spiritual distress,²⁵ and cancer-specific quality of life.²⁶ Pain was assessed using: (1) two items from the Symptom Distress Scale (SDS),¹² frequency (SDSF) and intensity (SDSI), to measure pain during the last week, and (2) the Pain Intensity 0 to 10 Numeric Scale (PINS) to measure current pain. Details of the pain scale items, and clinical meaning of corresponding scores, are described in Appendix Table A1 (online only). In the original trial design, the pain intensity (SDSI) item was asked only if the patient answered with a 2 (“I have pain once in a while”) or higher on the pain frequency (SDSF) question. For this secondary analysis, any question that was not answered by the patient on either of the SDS items or the PINS was assumed to be nonrelevant, and the value was set to the lowest possible value (1 on SDS and 0 on PINS). A sensitivity analysis was conducted that removed patients with missing data for any of the scale items. With consensus of the investigators of the parent study, and congruent with commonly accepted clinical pain management guidelines, moderate-to-severe pain (indicated by a 3 or higher on SDS items, or a 5 or higher on PINS) was considered to be relevant pain at a level that warranted intervention. A graphic display of the patient's self-reported pain scores was provided to the clinician before the patient's clinic visit in both intervention and control groups.

Audio-recorded clinic visits.

Pain was coded as relevant during the audio-recorded clinic visit if the patient or family member reported pain (over approximately the last 2 weeks), including pain that was improving in response to a previous intervention but was still present. Synonyms for pain, such as discomfort or soreness, were also coded. We coded these discussions of pain in the audio-recorded visit as relevant for two key reasons. First, if a patient or family member believed pain, in any form or severity, was important enough to warrant discussion with a provider during the clinic visit, we believed it was indeed relevant. Second, even if a patient's pain was potentially managed with clinician-prescribed medications, standard pain management guidelines recommend reassessment and continued monitoring of pain levels after pharmacologic intervention.^6,27 In other words, if pain was expressed as an ongoing or chronic issue, and the pain had been addressed and managed to some extent by prescriber intervention, pain continued to be relevant for that particular individual and their caregiver and was coded accordingly. If no discussion of pain occurred during the clinic visit, or if the patient denied pain, pain was coded as nonrelevant.

Analytic Methods

The discussion of relevant pain in the audio-recorded visit was considered the reference standard for comparison in this analysis. Patient self-reports of relevant pain on the SDS and PINS pain scales were then evaluated for the ability to identify relevant pain as coded in the audio-recorded clinic visit.

Sensitivity, specificity, and accuracy were calculated for all pain screening measures. Sensitivity was defined as the proportion of self-reports exceeding the predefined threshold for relevant pain of the total number of patients indicating pain as relevant in the audio-recorded visit. Specificity was defined as the proportion of self-reports not exceeding the threshold for relevant pain of the total number of patients not indicating relevant pain in the audio-recorded visit. Accuracy was defined as the total concordance of self-reported and audio-recorded definitions of relevant pain of the total number of patients. McNemar's exact test was used to assess the concordance of self-reported relevant pain with the audio-recorded reports for each scale. Additionally, the effect of adding pain frequency to pain intensity was explored; specifically, if a patient indicated moderate-to-severe pain on the SDSF, but not on intensity scales (SDSI or PINS), the patient was considered to have relevant pain.

A receiver operating characteristic (ROC) analysis was performed to evaluate various patient-report screening threshold scores to identify relevant pain in the audio-recorded clinic visit. The ROC curve plots sensitivity versus 1 minus the specificity. The area under the curve (AUC) was calculated using the trapezoidal rule for SDSF, SDSI, and PINS. Possible new threshold scores were evaluated using sensitivity, specificity, and accuracy; new threshold scores were selected such that sensitivity and specificity were both greater than 0.50.

To explore whether cancer stage affects the ability of pain screening scales to identify discussions of relevant pain in clinic, the analysis was first conducted in patients with advanced stage disease (stage IV), and then in patients with early-stage disease (carcinoma in situ [CIS], stage I-II). Patients with stage III disease were excluded because it was believed that patients with this stage disease were difficult to classify as either early- or late-stage due to the high degree of variability in tumor burden across cancer type. The analysis was also performed with all stages (CIS, I-II, III, IV) combined.

Results

Of 752 eligible participants, 452 patients had complete staging information and an audio-recorded clinic visit. Patients with non-stageable disease were excluded from the analysis. Baseline participant characteristics are presented overall and by stage in Table 1. Of the 452 patients, 10 (2%) were stage CIS, 83 (18%) were stage I, 127 (28%) were stage II, 94 (21%) were stage III, and 138 (31%) were stage IV. Results were similar across stage; therefore, the combined stage (CIS-IV) sample of the 452 patients is the focus of this report. Furthermore, results from the sensitivity analysis excluding patients with missing data (< 1% of the data: one response was missing on SDSI, four responses on SDSF, and six responses on PINS) were similar (data not shown).

Table 1.

Patient Characteristics

Characteristic	All Stages No. (%)	Stage CIS-II No. (%)	Stage III No. (%)	Stage IV No. (%)
Total	452	220 (49)	94 (21)	138 (31)
Age, years
Median (range)	57.6 (21.8-87.2)	58.1 (22.8-83.9)	55.8 (21.8-87.2)	57.6 (25.6-86.4)
Sex
Male	228 (50)	95 (43)	51 (54)	82 (59)
Female	224 (50)	125 (57)	43 (46)	56 (41)
Minority
Yes	42 (9)	23 (10)	10 (11)	9 (7)
No	367 (81)	180 (82)	75 (80)	112 (81)
Unknown	43 (10)	17 (8)	9 (10)	14 (10)
Work status
Not working	132 (29)	51 (23)	29 (31)	50 (36)
Working	277 (61)	147 (67)	60 (64)	70 (51)
Unknown	44 (10)	22 (10)	5 (5)	18 (13)
Service
Medical oncology	284 (63)	105 (48)	62 (66)	117 (85)
Radiation oncology	168 (37)	115 (52)	32 (34)	21 (15)
Cancer type
Bladder	13	3	3	7
Breast	149	109	21	19
Colorectal	49	9	22	18
Esophageal	12	3	7	2
GI, other*	15	4	2	9
Gastric	7	3	1	3
Head and neck	25	3	3	19
Lymphoma, Hodgkin	6	4	1	1
Lymphoma, non-Hodgkin	14	6	3	5
Myeloma	1	1	0	0
Pancreatic	10	3	3	4
Prostate	95	56	21	18
Renal cell	12	0	3	9
Sarcoma	17	2	1	14
Testicular	15	10	2	3
Other†	8	4	1	3
Unknown primary	4	0	0	4

Abbreviation: CIS, carcinoma in situ.

^*Adenocarcinoma (appendix, small bowel, ampullary gland, duodenum, gallbladder), Bismuth IV Klatskin tumor, carcinoid tumor (small bowel), cholangiocarcinoma, GI stromal tumor, mucinous neoplasm (appendix), and pancreatic neuroendocrine.

^†Carcinoid tumor (unknown primary origin, bronchial), Merkel cell carcinoma, retroperitoneal germ cell tumor, and Wilms tumor.

Relevant pain was discussed during 263 (58%) of the audio-recorded visits, compared with self-reported relevant pain on SDS and PINS (26%, SDSF; 17%, SDSI; 8%, PINS). Relevant pain identified on the self-report questionnaires was discordant (P < .001) with relevant pain discussed in the audio-recorded visits for all three scales. For all patients, relevant pain was discussed in clinic but not reported on the self-report questionnaire (38%, SDSF; 44%, SDSI; 51%, PINS) more often than relevant pain was reported on the questionnaires and not discussed in clinic (5%, SDSF; 2%, SDSI; 1%, PINS). All pain scales favored specificity over sensitivity. Specificity was higher for the intensity scales (0.94, SDSI; 0.97, PINS) than for the frequency scale (0.87, SDSF); sensitivity was higher for the frequency scale (0.35, SDSF) than for the intensity scales (0.24, SDSI; 0.12, PINS). Accuracy was higher for the SDS pain items (0.57, SDSF; 0.54, SDSI) than for the PINS (0.48). When pain frequency (SDSF) was added to the pain intensity items (SDSI or PINS), the accuracy, sensitivity, and specificity mirrored that of frequency (SDSF) alone. Identification of relevant pain on SDS and PINS remained statistically different from the audio-recorded coding of relevant pain when pain frequency and intensity were combined. The same trend existed when looking at patients with early-stage (CIS-II) and late-stage (IV) disease (Table 2).

Table 2.

Accuracy, Sensitivity, and Specificity for Threshold Scores Compared With Audio-Recorded Visit–Defined Relevant Pain

Pain Scale	Accuracy			Sensitivity			Specificity
	All	Stage CIS-II	Stage IV	All	Stage CIS-II	Stage IV	All	Stage CIS-II	Stage IV
Moderate-Severe Threshold
SDSF (threshold ≥ 3)	0.57	0.57	0.54	0.35	0.29	0.40	0.87	0.89	0.80
SDSI (threshold ≥ 3)	0.54	0.56	0.54	0.24	0.20	0.31	0.94	0.94	0.94
PINS (threshold ≥ 5)	0.48	0.51	0.44	0.13	0.11	0.16	0.97	0.98	0.94
SDSI plus SDSF (threshold ≥ 3)	0.59	0.57	0.58	0.39	0.31	0.45	0.86	0.88	0.80
PINS plus SDSF (thresholds ≥ 5, 3)	0.58	0.57	0.55	0.37	0.30	0.42	0.87	0.89	0.78
New Threshold Consideration
SDSF (threshold ≥ 2)	0.67	0.67	0.69	0.69	0.64	0.74	0.66	0.71	0.60
SDSI (threshold ≥ 2)	0.63	0.63	0.62	0.52	0.46	0.57	0.79	0.82	0.70
PINS (threshold ≥ 2)	0.57	0.56	0.61	0.45	0.41	0.58	0.72	0.74	0.66
SDSI plus SDSF (threshold ≥ 2)	0.68	0.67	0.68	0.70	0.64	0.74	0.65	0.71	0.58
PINS plus SDSF (threshold ≥ 2)	0.66	0.65	0.68	0.72	0.68	0.77	0.58	0.63	0.52

Abbreviations: CIS, carcinoma in situ; PINS, Pain Intensity Numeric Scale; SDSF, Symptom Distress Scale Frequency; SDSI, Symptom Distress Scale Intensity.

The ROC analysis (Fig 1) indicated comparable AUC values for SDSF, SDSI, and PINS (0.69, 0.67, and 0.65, respectively). Given the shape of the curves and the AUC, SDSI and SDSF seemed to better identify relevant pain compared with the audio-recorded visit than PINS. New threshold scores were selected to achieve sensitivity and specificity greater than 0.50. The comparisons between patient-reported scores on the pain scales and the audio-recorded visit data for the newly selected threshold scores are displayed in Table 2. Using a lower threshold value of 2 increased the sensitivity to 0.69 and 0.52 and decreased specificity to 0.66 and 0.79 for SDSF and SDSI, respectively. When the threshold for SDSF was set at 2 (“I have pain once in a while”), coded audio discussions and patient self-reports of relevant pain were no longer discordant (P = .19).

Figure 1.

Receiver operating characteristic curve for all possible threshold scores for each pain scale item (n = 452). AUC, area under the curve; PINS, Pain Intensity Numeric Scale; SDS, Symptom Distress Scale.

To meet the sensitivity and specificity criteria for PINS, a threshold score of 1 was suggested; however, a value of 1 was thought to be remarkably low (and of limited clinical relevance), so a threshold of 2 was explored. The threshold of 2 for PINS increased the sensitivity from 0.13 to 0.45; it decreased specificity from 0.97 to 0.72. Adding SDSF to SDSI and PINS, all at a threshold of 2, increased the sensitivity and decreased the specificity as compared to the intensity items alone. Adding SDSF to either intensity scale, at a threshold score of 2, produced results similar to that of SDSF alone (Table 2). The same general trend existed when calculated for patients with early- and late-stage disease.

Discussion

Our results suggest that asking a corollary assessment screening question about pain frequency (in addition to pain intensity) and reconsidering threshold scores on pain intensity scales are simple strategies to help busy clinicians more accurately identify ambulatory patients with cancer who may be struggling with pain. Improving assessment is only one facet of enhancing cancer pain management, but it is the crucial first step.

Previous research has demonstrated that ambulatory patients with cancer commonly self-report pain as a distressing symptom, and even patients with advanced stage disease are likely to receive suboptimal pain care.^28–31 Initial screening prompts trigger clinicians to probe more specifically about a potentially troubling symptom; therefore, it is vital that initial screening approaches adequately identify patients who are experiencing relevant pain. For adult patients in the United States, the most commonly used scale to screen for pain is the PINS (0 to 10; 0 indicates no pain, whereas 10 indicates worst possible pain). In fact, since the American Pain Society advocated documenting pain scores along with routine vital signs³² in the mid-1990s, use of the PINS scale has dramatically increased to satisfy Joint Commission regulatory requirements.³³ In busy ambulatory settings, often a patient care technician first screens for the presence of pain using the PINS scale. A score of 1 to 3, 4 to 6, and 7 to 10 corresponds to mild, moderate, and severe pain, respectively. Institutional protocols vary, but scores of 4 or higher (or sometimes 7 or higher) on PINS traditionally mandate further clinician attention.

Much research to date has focused on the performance of brief (1 to 2 item) screening tools, the benefit of composite distress indicators, and optimal threshold scores (cut points) to accurately identify troubling cancer symptoms.^{12,17,18,20,34–38} A recent systematic review of the cancer literature acknowledges gaps in evidence to support optimal cut points and recommends a threshold score of ≥ 4 to identify symptoms that require additional follow-up.¹⁸ More specifically, a cross-sectional study of 210 oncology outpatients determined an optimal cut point of 4 to identify clinically significant pain,¹² as did a study analyzing pain severity in almost 200 patients with symptomatic bone metastases who were referred to a palliative radiotherapy clinic.²⁰ Additionally, Butt et al³⁵ surveyed almost 600 ambulatory patients with solid tumors to evaluate the usefulness of a single-item screening question for fatigue, pain, distress, and anorexia. Optimal threshold scores were determined by calculating the sensitivity and specificity using ROC analysis and the proportion of patients at each screening score who reported that some relief of the target symptom would significantly improve their quality of life. The authors found that optimal cutoff scores ranged from 4 to 6 depending on the target symptom and concluded that single-item intensity screening instruments for fatigue, pain, distress, and anorexia are helpful screening tools to identify distressing symptoms in the outpatient oncology setting. In a related study, Jacobsen et al³⁷ explored whether the single-item Distress Thermometer compared favorably with longer, multi-item measures to screen for psychological distress in 380 ambulatory patients with cancer; they concluded that a Distress Thermometer cutoff score of 4 yielded optimal sensitivity and specificity. Our work extends this previous body of research by demonstrating that even lower threshold scores may more robustly identify relevant cancer pain.

Importantly, we found that patients reported relevant pain to providers during clinic visits more frequently than relevant pain was detected on the self-report screening pain questionnaires. This result raises interesting questions about the use of commonly accepted pain scales and our reliance on them to identify patients experiencing distressing pain. First, the SDS and PINS may favor specificity over sensitivity, detecting patients who do not have pain better than identifying those who do. Clinically, this translates into potential missed opportunities to help patients who are experiencing relevant pain. Second, we found that patients who rated their pain at levels not commonly considered to be threshold trigger scores (ie, 2 or 3 of 10) still discussed relevant pain during their clinic visit.

An interesting finding of our study was that concordance between patient self-reports of relevant pain on questionnaires and discussions of relevant pain in clinic significantly improved when we considered pain frequency along with pain intensity. Because frequency improved concordance when added to intensity, but adding intensity to frequency made no difference, one may conclude that pain frequency is the key metric and could be the sole measure used to assess cancer pain. Although this is an intriguing finding, it is important to keep in mind that the audio-recorded discussions of pain were not coded for severity, and this limits our ability to make conclusions about the role of pain intensity. Additionally, much regulatory and institutional effort has been focused on raising awareness of pain and incorporating pain intensity screening into routine patient assessment. Given these realities, it seems an unwise, and unrealistic, recommendation to abandon the clinical practice of inquiring about pain intensity in screening for cancer pain. Instead, we suggest consideration of routinely inquiring about pain frequency, in addition to pain intensity.

It is also important to note that our findings held across stage of disease. We did not find significant differences when we compared patients with early-stage and later stage disease. This reflects the reality that pain warranting discussion in clinic is a relevant concern for all patients with cancer across the illness trajectory, regardless of disease stage or etiology of pain.

Clinical Implications

Our results suggest a reconsideration of threshold scores that are typically considered triggers on pain intensity scales. Generally, a score of 2 of 10 on the PINS screening tool is viewed as nonrelevant and may not be addressed by a clinician. Our ROC analysis challenges this clinical practice and underscores the importance of clinician willingness to explore the distress that may be concomitant with a low pain score. Such discussions take time, and time is a high-premium commodity in a busy ambulatory setting. Being able to gauge and prioritize the importance of exploring a patient PINS pain score of 2, for example, may be assisted by adding a screening question related to frequency of pain after the standard inquiry about pain intensity. This could be as simple as following the standard intensity-related question, “What is your pain level now?” with a question related to frequency, such as: “During the last week, how frequently has pain been a problem for you?” Using this combination of simple screening questions may be an extremely effective, efficient, and practical method of accurately identifying more patients with cancer whose quality of life is being negatively affected by distressing pain.

Our research also demonstrates the crucial importance of optimally assessing for pain across the entire illness trajectory for all patients with cancer, regardless of stage of disease. Certainly, patients with stage IV cancer are at high risk for pain, but our findings provide an important reminder that clinicians must remain attuned to relevant pain in patients with earlier stage disease, which may be a result of the effects of treatment or other etiologies. Future research should explore pain outcomes when lower PINS threshold scores and frequency screening questions are used in the ambulatory oncology setting. Additionally, many commonly used screening tools for cancer symptoms other than pain (eg, fatigue) also rely on a 0 to 10 intensity rating scale; future research should explore if similar considerations related to lower threshold and frequency exist for other cancer symptom screening intensity scales.

Limitations

As with any secondary analysis, data have been bound by the original parameters of the parent study, in which a different set of research questions were used for study planning and design. An important limitation of our study is that the variable of relevant pain in the audio-recorded clinic visits could not be coded practically for severity; subsequently, any persistent or present pain, or pain requiring medication, regardless of severity, that was discussed was considered relevant. This conservative and broad approach in defining relevant pain is important to consider in interpreting our findings, given that pain discussions may have occurred for minimal pain, or for moderate-to-severe pain that had been successfully lowered before the clinic visit. Despite this limitation, we argue that any ongoing pain, regardless of whether it has been previously addressed or not, remains relevant for patients with cancer and their caregivers and is a key factor in overall quality of life.

Audio-recorded clinic visits also occurred at one fixed point in time. Of course, it is possible that relevant pain was discussed at earlier, or later, clinic visits that were not audio-recorded. Patients who did not answer the SDS or PINS items were considered to have nonrelevant pain for this analysis; this assumption may be inaccurate. However, given the small percentage of unanswered questions and our sensitivity analysis, this assumption did not affect the final results. It is also important to emphasize that this study primarily captures cancer-related pain; the findings may not apply to patients with other types of pain. Last, our sample consisted primarily of white patients from two comprehensive cancer centers in metropolitan areas; results may not generalize to more diverse patient samples receiving care at other types of institutions in differently populated areas.

In conclusion, the traditional clinical practice of relying on PINS (0 to 10 intensity scale) to identify cancer pain that warrants further discussion between patients and clinicians is potentially inadequate. Reconsidering lower threshold scores for pain intensity, and also inquiring about pain frequency, are strategies that should be considered to accurately identify more patients who are experiencing relevant cancer pain.

Appendix

Table A1.

Description of the Pain Scale Items Used in Analysis

Scale	Score and Description
SDS*
SDSF score of pain during the last week	1: I almost never have pain	2: I have pain once in a while	3: I have pain several times per week	4: I am usually in some degree of pain	5: I am in some degree of pain almost constantly
SDSI score of pain during the last week	1: When I do have pain, it is very mild	2: When I do have pain, it is mildly distressing	3: When I do have pain, it is usually fairly intense	4: The pain I have is very intense	5: The pain I have is almost unbearable
PINS†
Intensity of pain currently	0: No pain	1-3: Mild pain	4-6: Moderate pain	7-9: Severe pain	10: Worst pain imaginable

Abbreviations: PINS, Pain Intensity Numeric Scale; SDS, Symptom Distress Scale; SDSF, SDS Frequency; SDSI, SDS Intensity.

^*SDS internal consistency in parent trial⁵ was 0.86.

^†For a detailed discussion of psychometric properties related to pain scales, see Jensen M: J Pain 4:2-21, 2003.

Authors' Disclosures of Potential Conflicts of Interest

Disclosures provided by the authors are available with this article at jop.ascopubs.org.

Author Contributions

Conception and design: All authors

Financial support: Donna L. Berry

Administrative support: Donna L. Berry

Provision of study materials or patients: Donna L. Berry

Collection and assembly of data: Traci M. Blonquist, Barbara Halpenny, Donna L. Berry

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Screening for Pain in the Ambulatory Cancer Setting: Is 0-10 Enough?

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or jop.ascopubs.org/site/misc/ifc.xhtml.

Virginia T. LeBaron

No relationship to disclose

Traci M. Blonquist

No relationship to disclose

Fangxin Hong

No relationship to disclose

Barbara Halpenny

No relationship to disclose

Donna L. Berry

Research Funding: Medocity (Inst)