Research Paper
National Breakthrough Pain Study: prevalence, characteristics, and associations with health outcomes Arvind Narayanaa, Nathaniel Katzb,c, Alicia C. Shillingtond, Judith J. Stephensone, Qing Harshawd, Carla B. Fryef, Russell K. Portenoyg,*
Abstract The National Breakthrough Pain Study is a large observational study that assessed breakthrough pain (BTP) in a population of commercially insured community-dwelling patients with opioid-treated chronic pain. Eligible patients were identified from an administrative claims database, and consenting patients were asked to complete a structured telephone interview and several validated questionnaires. Questionnaires assessed pain interference with function (Brief Pain Inventory-Short Form), health status (Short Form 12 [SF-12] Health Survey), disability (Sheehan Disability Scale), work performance (World Health Organization Health and Work Performance Questionnaire), and mood (Generalized Anxiety Disorder-7 Screener [GAD-7] and Patient Health Questionnaire-2 [PHQ-2]). Of 2198 patients interviewed, 1278 patients had persistent pain controlled with opioid therapy; 1023 (80%) of these patients reported BTP. Patients had a median of 2.0 episodes of BTP per day (range, 1-50) and a median duration of BTP of 45 minutes (range, 1-720). Compared with patients without BTP, patients with BTP had more pain-related interference in function (Brief Pain Inventory, mean 6 SD: 34.2 6 15.6 vs 25.0 6 15.7 [P , 0.001]), worse physical health (SF-12 physical component score: 29.9 6 9.6 vs 35.1 6 10.4 [P , 0.001]) and mental health (SF-12 mental component score: 47.4 6 11.3 vs 49.3 6 10.4 [P , 0.001]), more disability (Sheehan Disability Scale global impairment score: 15.1 6 9.1 vs 10.6 6 8.5; World Health Organization Health and Work Performance Questionnaire absolute absenteeism: 12.4 6 59.9 vs 7.7 6 44.9 hours [both P , 0.001]), and worse mood (GAD-7 score: 7.4 6 5.9 vs 5.9 6 5.4; PHQ-2 anhedonia score: 1.2 6 1.1 vs 0.9 6 1.0 [both P , 0.001]). In this population of community-dwelling patients with opioid-treated chronic pain, BTP was highly prevalent and associated with negative outcomes. This burden of illness suggests the need for specific treatment plans. Keywords: Epidemiology, Breakthrough pain, Quality of life, Patient-reported outcomes, National Breakthrough Pain Study
1. Introduction Breakthrough pain (BTP) may be defined as a transitory episode of pain that occurs on a background of persistent pain that is otherwise controlled on a stable opioid regimen.26 Prevalence of BTP varies with diagnosis, severity of illness, and treatment setting. Among patients with controlled persistent cancer pain, 33% to 89% experience BTP.24,26,27,31 Studies of varied populations with opioid-treated pain unrelated to cancer have reported BTP in 48% to 74% of patients.23,24,32 Although controversy surrounds the clinical implications of BTP in populations without cancer,5,19,20 these studies suggest at minimum that the phenomenon is prevalent in all opioid-treated chronic pain populations.
Several surveys of patients with cancer indicate that BTP is associated with more severe pain, less effective analgesic treatment, impaired ability to function, negative psychosocial outcomes, and relatively poorer quality of life.4,25,27 More limited data in patients with chronic pain unrelated to cancer suggest similar associations.25 All of these surveys have underscored the need for additional research to assess the clinical impact of BTP in larger samples of patients. The National Breakthrough Pain Study (NBTPS) was designed as a large observational study of community-dwelling patients with chronic pain identified from a national database of commercially insured patients. The aim was to assess the prevalence and characteristics of BTP and to explore its associations with diverse health-related outcomes.
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article. a
2. Methods
*Corresponding author. Address: MJHS Institute for Innovation in Palliative Care, 39 Broadway, 12th Floor, New York NY 10006, USA. Tel.: (212) 649-5560; fax: (212) 649-5544. E-mail address: [email protected] (R. K. Portenoy).
The NBTPS used telephone survey methodology to conduct a cross-sectional observational study. Patient interviews were conducted between April and December 2010. An institutional review board approved the survey, and all patients provided verbal consent before responding to questions.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.painjournalonline.com).
2.1. Patient selection
Teva Pharmaceuticals, Frazer, PA, USA, b Tufts University School of Medicine, Boston, MA, USA, c Analgesic Solutions, Natick, MA, USA, d EPI-Q, Inc, Oak Brook, IL, USA, e HealthCore, Inc, Wilmington, DE, USA, f Thayer County Health Services, Hebron, NE, USA, g MJHS Institute for Innovation in Palliative Care, New York, NY, USA
PAIN 156 (2015) 252–259 © 2015 International Association for the Study of Pain http://dx.doi.org/10.1097/01.j.pain.0000460305.41078.7d
252
·
A. Narayana et al. 156 (2015) 252–259
Potentially eligible patients were identified from the HealthCore Integrated Research Database (HIRDSM), an administrative database containing claims data from one of the largest health PAIN®
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
benefits companies in the United States. The claims originated from patients enrolled in 1 of 14 US health plans. For this study, the HIRD sampling frame consisted of 6.4 million patients who were $18 years of age and were currently active health plan members enrolled for $12 months before the patient sample list was extracted. From this sampling frame, 50,000 health plan members were randomly selected and stratified to approximate the US general population (based on the 2000 US National Census) according to the following characteristics: age (18-34 years, 35-49 years, 50-64 years, 65-74 years, and $75 years), sex, and health plan census region (Northeast, South, Midwest, and West). Members who were likely to have chronic pain based on diagnosis codes (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM]; see Supplemental Digital Content, http://links.lww.com/PAIN/A15) and opioid use were oversampled to ensure adequate representation in the applicable analysis cohorts. To exclude patients receiving methadone for treatment of addiction, patients with an ICD-9-CM diagnosis code (medical claims) or Healthcare Common Procedure Coding System code (ambulatory services) for drug abuse or dependence concurrent with a pharmacy claim for methadone were rejected from the sample (see Supplemental Digital Content, http://links.lww.com/PAIN/A16). Patients with an invalid or missing telephone number and those who were on the HealthCore do-not-call list also were excluded. From this representative sample of 50,000 health plan members, 2 potential analysis cohorts were selected (Fig. 1). To create these analysis cohorts, a series of smaller samples (replicates) was selected, with each replicate being representative of the larger stratified sampling frame but also characterized by the criteria used to define each of the analysis cohorts. These replicates were released into a telephone
www.painjournalonline.com
253
survey field, which provided the information necessary to call each patient. One of the analysis cohorts—a “pain” cohort—included patients with controlled persistent pain, with or without BTP. This cohort was initially identified from the HIRD by a medical claim with an ICD-9-CM diagnosis code consistent with chronic pain and pharmacy claims consistent with long-term opioid use. Specifically, patients were eligible for this cohort if they had (1) at least 2 medical claims separated by $3 months that included an ICD-9-CM code consistent with back pain, neuropathic pain, traumatic injury, osteoarthritis, fibromyalgia, cancer pain, or other chronic pain conditions, and (2) at least 3 opioid prescription claims within 3 months with a medication refill adherence of $90%. Medication refill adherence assesses daily use by dividing the total days’ supply of medication by the number of days in the study period and multiplying by 100 to provide a percent adherence rate.11,12 At the time of the subsequent telephone interview, the characterization of the patient as appropriate for the pain cohort was confirmed. The second analysis cohort—a “control” cohort—was identified by lack of evidence of a pain diagnosis or chronic opioid use. This status also was confirmed during the subsequent telephone contact. 2.2. Telephone survey Up to 6 calls were made in an effort to identify and speak with the patient identified through the sampling process. The interviewers were trained by investigators (A.C.S. and C.B.F.) by completing a question and answer session on the survey objectives, inclusion and exclusion criteria, survey script, and procedures for the callback protocol. The interviewers were counseled to strictly
Figure 1. National Breakthrough Pain–Survey Design. The original study protocol consisted of 5 analysis groups (control cohort, patients with cancer with controlled persistent pain without breakthrough pain (BTP), patients with cancer with controlled persistent pain with BTP, patients without cancer with controlled persistent pain without BTP, and patients without cancer with controlled persistent pain with BTP). For this analysis, data for patients with and without cancer were combined.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
254
·
PAIN®
A. Narayana et al. 156 (2015) 252–259
adhere to the fully structured interview script and were not allowed to provide improvised assistance or instruction in an attempt to interpret any portions of the script in response to questions. They were permitted only to reread script portions in response to questions and were randomly monitored by floor supervisors to ensure strict adherence to the script (interview script available on request). Patients were interviewed if they were fluent in English, could communicate by telephone, and consented to the study. Those who were capable of participating were informed that the survey was designed “to learn about the health care status and needs in the population.” They were told that patients were randomly selected, participation was voluntary, responses were confidential, and participation would not affect health care benefits. Verbal informed consent was elicited before proceeding with the interview, and each patient was paid $25 as compensation for their time. The telephone survey required less than 30 minutes on average (range, 5-50 minutes) to complete. The interview began with questions about demographics not captured in claims data (eg, marital status, household income, ethnicity, educational level) and questions designed to confirm eligibility. Patients selected for the pain cohort were first asked whether they had persistent pain and could respond affirmatively to a question designed to identify clinically significant pain: “Some people have pain that causes significant problems with daily functioning, mood, sleep, or their overall quality of life. Have you had pain that has caused these types of significant problems over the past 3 months?” Patients with acute or intermittent pain and those who did not respond affirmatively to the statement did not continue with the survey. Patients with clinically significant persistent pain were then asked to confirm that they took opioid medication daily; those who did not also discontinued the interview. Patients with clinically significant opioid-treated pain were then queried about the extent to which the pain was controlled by the medication. Persistent pain was considered controlled when the patient reported baseline pain as not more than moderate on a verbal rating scale of pain intensity (absent, mild, moderate, or severe or excruciating). Only patients who described their persistent pain as absent, mild, or moderate were permitted to continue with the rest of the telephone survey. Through this initial series of questions, patients who were not eligible for the pain cohort were discontinued, and all patients remaining in the pain cohort were characterized by clinically significant persistent pain controlled by an opioid regimen. These patients were then asked whether they experienced BTP (Fig. 1), which was defined as temporary flares of moderate-to-severe pain that were more intense than the persistent pain. Eligibility for the control cohort was confirmed through a series of similar questions. Patients who were allowed to continue with the rest of the survey verified that they had no pain on the day of the survey; no pain that caused significant problems with daily function, mood, sleep, or quality of life (thus, some patients in the control group may have had mild pain); and no history of chronic opioid use. Patients in both the pain cohort and the control cohort provided verbal responses to several validated questionnaires intended to assess quality of life, functioning, and productivity. Verbal responses to the Brief Pain Inventory (BPI)-Short Form6,7,13 were elicited from patients with pain, and patients in the pain cohort with BTP responded to a structured interview designed to characterize their BTP.
2.3. Questionnaires and structured interview
2.3.1. Short Form 12 Health Survey (version 2) The Short Form 12 (SF-12) includes a subset of questions from the widely used Short Form 36 Health Survey. Short Form 12 is a valid measure of health status and includes a physical component summary score and the mental component summary score to assess physical and mental health, respectively.30 2.3.2. Sheehan Disability Scale The Sheehan Disability Scale (SDS) is a widely used measure of disability that assesses the impact of health conditions on productivity, specifically days out of role.1,10,22,28 The SDS has a global interference score, which is the sum of the 3 separate interference subscales. 2.3.3. World Health Organization Health and Work Performance Questionnaire-Short Form The World Health Organization Health and Work Performance Questionnaire-Short Form evaluates workplace absenteeism and presenteeism and thereby assesses 1 element associated with disability.14,15 2.3.4. Generalized Anxiety Disorder-7 Screener The Generalized Anxiety Disorder-7 (GAD-7) is a brief and validated screen for anxiety disorders.18 2.3.5. Patient Health Questionnaire-2 The validated Patient Health Questionnaire-2 (PHQ-2) screens for depressed mood and includes separate items for dysphoria and anhedonia.16 2.3.6. Brief Pain Inventory-Short Form The BPI-Short Form is a widely used instrument that includes pain intensity measures assessed on a numeric scale of 0 to 10 and a 7-item subscale that evaluates the extent to which pain interferes with varied domains of function, such as sleep, walking, social relations, and enjoyment of life.6,7,13 2.3.7. Breakthrough Pain Questionnaire The Breakthrough Pain Questionnaire (BPQ) is a structured interview adapted from a format used in previous studies.23,26,27 Responses to the BPQ characterized persistent pain and BTP, if experienced, in terms of temporal characteristics, location, duration, quality, and other features. 2.4. Statistical analysis Patients’ administrative claims data from the HIRD were merged with their survey data acquired during the telephone interview to conduct the analysis. The HIRD provided demographic data for the 12-month period before the date of the survey (eg, age group, gender, region), information about comorbidities (documented by ICD-9-CM codes on medical claims), medical resource utilization (eg, inpatient hospitalizations, emergency department visits, outpatient services), direct health care costs, and pharmacy claims and costs. Comorbidity burden was assessed through the
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
Deyo-modified Charlson Comorbidity Index8 and was reported as mean 6 SD. Demographic characteristics, prevalence of BTP, and pain phenomenology (eg, location, duration, severity) were analyzed descriptively using frequencies, means, and SDs, as appropriate. Mean 6 SD 24-hour pain intensity and interference subscores from the BPI-Short Form were analyzed between patients with BTP and those without BTP using a Student t test. Persistent pain severity (ordinal scale; recorded as absent, mild, or moderate) was compared between those with BTP and those without BTP using frequency distributions and x2 tests with a gamma statistic. Continuously distributed characteristics of BTP (eg, number of BTP flares per day, duration until peak intensity, duration from start of flare to end) were examined descriptively (means, SDs, and medians) only in patients with BTP. The ability to predict BTP flares was reported as proportions, rated on a Likert scale (never, sometimes, often, almost always, or always). The relationship between start of flare and last medication dosing was reported as a proportion of those responding affirmatively (pain temporally related to the last dose) vs negatively. Location of the worst BTP was reported as a proportion of those responding back, legs, neck, shoulder, arm, or head. Data from the questionnaire responses were summarized for the control cohort, the pain cohort with BTP, and the pain cohort without BTP. The SF-12 physical and mental health component summary scores were calculated as normed scores.30 The SDS scales were coded into global and interference subscales. Days out of role were recorded as past 30 days and past 365 days. Each scale was reported as mean 6 SD and compared across the 3 cohorts using an analysis of variance with a post hoc test (Bonferroni) for pairwise comparisons if the overall test result among groups was statistically significant. A P value of ,0.05 was considered statistically significant for all analyses. Univariate associations with pain interference (BPI-Short Form subscore) were assessed by Pearson r and Cohen d. Parameters with effect sizes $0.3 and/or with P values ,0.1 were generally considered for multivariate model inclusion after examining for potential multicollinearity. Impact of BTP on pain interference (BPI-Short Form subscore) was modeled using an ordinary least squares regression analysis, controlling for relevant demographics, persistent pain severity, mood, and comorbidity.
3. Results Interviews were attempted with 8703 patients. Of these, 3725 refused to participate, 2780 were deemed ineligible after a partial interview, and 1 was excluded because of missing medical claims data. Of the 2197 eligible patients who completed the interview, 1278 were in the pain cohort and 919 were in the control cohort. Of the 1278 in the pain cohort, 1023 (80.0%) reported BTP (Fig. 1). Patient demographics and other baseline clinical characteristics are summarized in Table 1. Across the control and pain cohorts, the mean age was 47.8 to 52.0 years, and the proportion reporting white race was 87.9% to 95.2%. Only 145 of the 1278 patients (11.3%) with controlled persistent pain had cancer. Work-related disability was reported by 0.7% of patients in the control cohort, 16.9% in the pain cohort without BTP, and 31.3% in the pain cohort with BTP. The majority of patients in the pain cohort, all of whom had controlled persistent pain, reported moderate (vs absent or mild) persistent pain intensity, regardless of the presence of BTP (Table 2). However, the proportion of patients reporting moderate vs mild persistent pain was greater among those with BTP than it was among those without BTP (P 5 0.027). The multivariate regression model evaluating the impact
www.painjournalonline.com
255
Table 1
Patient demographics and other baseline characteristics. Variable Age, years Mean (SD) Sex, n (%) Male Female Race, n (%)* White Black Hispanic Native American Asian Other† Employment status* Employed Self-employed Disabled Retired Other‡ Education ,High school High school Junior college College Postgraduate Annual income ,$20,000 $20,000-$40,000 $40,000-$80,000 .$80,000 Pain due to cancer, n (%) Yes Deyo-Charlson comorbidity index Mean (SD) Visit to pain specialist, n (%) Yes
Control cohort
Pain cohort without BTP
Pain cohort with BTP
n 5 915 47.8 (17.6) n 5 919 426 (46.4) 493 (53.6) n 5 911 801 (87.9) 37 (4.1) 42 (4.6) 12 (1.3) 41 (4.5) 26 (2.8) n 5 918 535 (58.3) 151 (16.4) 6 (0.7) 121 (13.2) 215 (23.4) n 5 133 3 (2.3) 48 (36.1) 19 (14.3) 57 (42.9) 6 (4.5) n 5 110 10 (9.1) 16 (14.5) 38 (34.5) 46 (41.8) NA NA NA
n 5 255 52.0 (11.4) n 5 255 106 (41.6) 149 (58.4) n 5 254 236 (92.9) 6 (2.4) 4 (1.6) 6 (2.4) 1 (0.4) 4 (1.6) n 5 254 110 (43.3) 43 (16.9) 43 (16.9) 30 (11.8) 50 (19.7) n 5 253 8 (3.2) 106 (41.9) 43 (17.0) 87 (34.4) 9 (3.6) n 5 237 30 (12.7) 64 (27.0) 76 (32.1) 67 (28.3) n 5 255 33 (12.9) n 5 255
n 5 1023 49.8 (9.8) n 5 1023 418 (40.9) 605 (59.1) n 5 1023 974 (95.2) 22 (2.2) 24 (2.3) 26 (2.5) 6 (0.6) 12 (1.2) n 5 1021 367 (35.9) 142 (13.9) 320 (31.3) 68 (6.7) 273 (26.7) n 5 1021 43 (4.2) 430 (42.1) 181 (17.7) 345 (33.8) 22 (2.2) n 5 916 132 (14.4) 236 (25.8) 318 (34.7) 230 (25.1) n 5 1023 112 (10.9) n 5 1023
NA NA
1.1 (2.0) n 5 255
1.0 (1.8) n 5 1023
NA
99 (38.8)
436 (42.6)
* Respondents could select more than 1 race or employment status or could choose not to answer. † “Other” includes mixed race, Arabic, Armenian, Creole, Cuban-Dominican, Hebrew, Indian, Mediterranean, or Middle Eastern. ‡ “Other” includes unemployed and looking for work, unemployed and not looking for work, student, homemaker, something else, temporarily laid off, maternity leave, and illness. BTP, breakthrough pain; NA, not applicable or not available.
of BTP on pain interference controlled for the difference in the proportion of patients with moderate pain. The BPQ provided descriptive information about the BTP experienced by the 1023 patients in the pain cohort (Table 3). Overall, 688 (67.3%) patients with BTP reported that they treated their BTP; of these, 358 (52.0%) patients used medication. The most common ($5%) of these medications were hydrocodone with acetaminophen (181/358 [50.6%]), oxycodone (51/358 [14.2%]), and oxycodone with acetaminophen (44/358 [12.3%]); only 2 patients reported using transmucosal immediate-release fentanyl formulations for BTP. The perceived effects of medication or other treatments for BTP were not evaluated. Patients with BTP reported a median of 2.0 episodes per day (range, 1-50), with a median duration of 45 minutes (range, 1-720). The back was the most common location for BTP flares (55.2%). Approximately two-thirds of patients could not reliably predict the timing of BTP, but approximately one-third noted that BTP usually occurred at some predictable number of hours after taking a regularly scheduled dose of an analgesic.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
256
·
PAIN®
A. Narayana et al. 156 (2015) 252–259
Table 2
Responses to the Brief Pain Inventory and persistent pain severity. Variable BPI Pain intensity over 24 h, mean (SD) Worst Least Average Interference in 24 h, mean (SD) Activity Mood Ability to walk Ability to work Social relations Sleep Enjoyment of life Total score in past 24 h, mean (SD) Persistent pain severity Baseline pain over past week, n (%) Mild Moderate Absent
P†
Control cohort (n 5 133)*
Pain cohort without BTP (n 5 255)
Pain cohort with BTP (n 5 1023)
2.0 (2.0) 0.6 (1.2) 1.5 (1.7)
5.0 (2.1) 2.6 (1.8) 4.0 (1.8)
5.9 (1.9) 3.2 (1.8) 4.6 (1.7)
,0.001 ,0.001 ,0.001
0.7 (1.5) 0.7 (1.6) 0.7 (1.7) 0.9 (1.8) 0.5 (1.4) 0.8 (1.7) 0.9 (1.7) 5.0 (9.4)
3.6 (2.7) 3.6 (2.8) 3.0 (2.9) 4.0 (2.9) 2.6 (2.6) 4.2 (3.0) 4.0 (2.6) 25.0 (15.7)
5.3 (2.8) 4.7 (2.8) 4.1 (3.0) 5.5 (2.9) 3.7 (3.0) 5.6 (2.9) 5.2 (2.8) 34.2 (15.6)
,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001
— — —
63 (24.7) 187 (73.3) 5 (2.0)
191 (18.7) 823 (80.4) 9 (0.9)
0.027
* Respondents in the control group who had mild pain that did not cause significant problems with daily function, mood, sleep, or their quality of life and who did not have pain on the day of the survey received the BPI. † t-tests were performed between patients with BTP and patients without BTP. P , 0.05 was considered statistically significant. BPI, Brief Pain Inventory; BTP, breakthrough pain.
Across the control cohort, the pain cohort with BTP, and the pain cohort without BTP, the SF-12 physical and mental component summary scores, SDS, World Health Organization Health and Work Performance Questionnaire, GAD-7, and PHQ-2 all showed statistically significant group differences (Table 4). In all cases,
Table 3
Responses to the breakthrough pain (BTP) Questionnaire (pain cohort with BTP only)*. Variable Number of BTP flares per day Median Mean Duration until peak pain, min Median Mean Duration from flare start to end, min Median Mean Ability to predict BTP flares, n (%) Never Sometimes Often Almost always Always Do not know When pain flares occur, they start at the same time after the last scheduled dose of medication, n (%)† Worst BTP pain location, n (%)‡ Back Leg(s) Neck Shoulder(s) Arm(s) Head * 1023 patients had BTP, which was 5% of overall sampled population in the study. † Observed n 5 961, not all patients answered all of the questions. ‡ Respondents could select more than 1 location.
N 5 1023 2.0 2.9 10.0 32.0 45.0 101.4 324 (31.6) 363 (35.4) 161 (15.7) 107 (10.4) 55 (5.4) 14 (1.4) 323 (33.6)
n 5 1022 564 (55.2) 218 (21.3) 166 (16.2) 141 (13.8) 66 (6.5) 58 (5.7)
patients with persistent pain and BTP had the most impairment. Patients in this subgroup experienced the worst health status, more depressed and anxious mood, and more disability than patients in the pain cohort with persistent pain who did not report BTP and, to a larger extent, patients in the control group. The BPI provided additional information about the extent to which pain interfered with varied domains of function. Compared with those who did not experience BTP, patients with BTP had significantly greater mean pain interference in all functional domains (Table 2). In a multivariate model controlling for relevant covariates (including persistent pain intensity, anxiety, depressed mood, and patient demographics), the major factors associated with pain interference (P , 0.001) included the presence of BTP, level of baseline constant pain, depressed mood (as indicated by GAD-7 scores), and anxiety (as indicated by PHQ-2 scores) (Table 5). These variables together explained 36% of the variation of pain interference (adjusted R2 5 0.360 [P , 0.001]). For patients with BTP, the predicted pain interference (0-10 scale) was 0.907 points higher than for patients without BTP, holding all other variables constant. Further subgroup analyses revealed that between patients with mild baseline pain or moderate baseline pain, those with BTP had significantly greater mean pain interference than those without BTP (mild baseline pain [3.6 vs 2.6, respectively; P 5 0.002] and moderate baseline pain [5.2 vs 3.9, respectively; P , 0.001]).
4. Discussion The NBTPS is an unprecedented attempt to gather data about BTP from a nationally representative population of commercially insured community-dwelling patients with clinically significant chronic pain managed with opioid therapy. More patients were surveyed than in any similar study to date, and the study was the first to identify patients using administrative claims data from 1 of the largest health benefits companies in the United States. The data substantially augment the results of earlier surveys of BTP, which have been conducted in smaller samples of ambulatory patients,23,24 samples of cancer inpatients or cancer patients
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
www.painjournalonline.com
257
Table 4
Responses to health status, functioning, mood, and productivity survey instruments. Variable, mean (SD) SF-12 physical component summary score SF-12 mental component summary score SDS global Interference with work or school Interference with social or leisure activities Interference with family life or home responsibilities SDS days out of role Past 30 days Past 365 days HPQ-SF Absolute absenteeism, h‡ Absolute presenteeism§ GAD-7‖ PHQ-2{ Anhedonia Dysphoria
Control cohort
Pain cohort without BTP
Pain cohort with BTP
53.4 (6.6) 54.7 (5.9) 1.7 (4.0) 0.6 (1.5) 0.6 (1.6) 0.5 (1.4)
35.1 (10.4) 49.3 (10.4) 10.6 (8.5) 3.7 (3.3) 3.6 (3.0) 3.3 (2.9)
29.9 (9.6)*† 47.4 (11.3)*† 15.1 (9.1)*† 5.3 (3.6)*† 5.1 (3.2)*† 4.8 (3.1)*†
0.3 (1.8) 3.2 (14.0)
4.5 (7.4) 59.1 (99.6)
8.9 (10.3)*† 112.2 (134.3)*†
1.4 (45.7) 87.0 (12.0) 1.9 (2.6)
7.7 (44.9) 83.5 (14.9) 5.9 (5.4)
12.4 (59.9)*† 79.8 (18.7)*† 7.4 (5.9)*†
0.1 (0.3) 0.8 (0.3)
0.9 (1.0) 0.6 (0.9)
1.2 (1.1)*† 0.9 (1.0)*†
ANOVA test was used for numeric variables with post hoc testing (Bonferroni) for pairwise comparison if overall test among groups was statistically significant. x2 test was used for overall and pairwise comparisons for categorical variables. P , 0.05 was considered statistically significant. * P , 0.001 for BTP group vs control group. † P , 0.001 for BTP group vs no BTP group. ‡ Scored in terms of raw hours lost per 4 wk (a higher score indicates a higher amount of absenteeism). § Lower bound of 0 (total lack of performance during time on the job) and upper bound of 100. ‖ GAD-7 score .8 suggests anxiety or panic disorder. { PHQ-2 scores $3 (sum of anhedonia and dysphoria scores) suggest major depression. BTP, breakthrough pain; GAD-7, Generalized Anxiety Disorder-7; HPQ-SF, Health and Work Performance Questionnaire; PHQ-2, Patient Health Questionnaire-2; SDS, Sheehan Disability Scale; SF-12, 12-Item Short Form Health Survey.
with advanced illness,3,4,26,27,31 and a sample of patients with advanced illnesses other than cancer.23,32 Most patients in this study had pain unrelated to cancer. Breakthrough pain was highly prevalent in this population (80%), exceeding the prevalence of 48% to 74% reported in earlier surveys of patients with chronic noncancer pain.23,24,32 To some extent, this higher prevalence may have been related to the definitions applied to BTP in the various surveys. The current survey acquired information about pains that were experienced as more intense than the baseline persistent pain, whether the intensity of the temporary flare was moderate or severe; other surveys limited assessment to pain perceived to be severe or excruciating. Compared with the largest survey of ambulatory patients to date (N 5 228),23 this study showed a similar number of episodes of BTP per day (median, 2.0 for both studies), time until peak pain (median, 10 minutes for both studies), and duration of BTP episodes (median, 45 minutes and 60 minutes, respectively). Breakthrough pain likely related to the end of the dosing interval
Table 5
Multiple regression model with pain interference (Brief Pain Inventory subscore) as the dependent variable. Model
Unstandardized coefficients Parameter estimate
BTP (present vs absent) Education Male gender GAD-7 score Income Level of constant pain PHQ-2 Constant
0.907 20.065 20.291 0.102 20.169 1.105 0.410 0.585
t
P
6.383 21.129 22.477 7.498 22.671 8.302 6.944 1.261
0.000 0.259 0.013 0.000 0.008 0.000 0.000 0.207
SE 0.142 0.058 0.117 0.014 0.063 0.133 0.059 0.464
P , 0.05 was considered statistically significant. BTP, breakthrough pain; GAD-7, Generalized Anxiety Disorder-7; PHQ-2, Patient Health Questionnaire-2.
(reported as BTP flares that occurred at the same time after the last scheduled dose of medication) was more common in this study (33% vs 19%) and suggests that there may be an opportunity to address BTP in a substantial proportion of patients by adjusting the dose or interval of the around-the-clock opioid regimen or using medications with a longer duration of action. However, it is important to note that the interpretation of this finding is limited in that patients were not specifically asked whether their BTP episode occurred before the next scheduled dose of their medication was due. In this study, patients with controlled persistent pain and BTP reported substantially worse health status, mood, pain interference with functioning, and disability than patients with controlled persistent pain and no BTP. Not surprisingly, the differences were greatest when comparing those with BTP and those in the control cohort; patients with persistent pain and no BTP had better scores than patients with BTP and worse scores than patients in the control cohort. Similarly, previous surveys of patients with cancer found that BTP is associated with functional impairment and relatively poorer quality of life,4,27 and a more recent survey of community-dwelling patients with and without cancer reported comparable findings.25 These previous surveys did not assess measures such as the SDS and GAD-7. Clinical significance of the group differences in adverse outcomes may be illuminated by comparison with the limited data on minimum clinically important differences (MCID) available in the literature. Dworkin et al.9 suggested that the MCID for BPI interference is $1 point in patients with chronic pain, and Bennett et al.2 reported that $1.26- and $2.28-point differences in the SF-12 physical component and mental component summary scores, respectively, were clinically significant for heart failure patients. The present survey found that these MCIDs were exceeded for all BPI interference assessments and the SF-12 physical component summary score when comparing differences in average scores between patients in the pain cohort with
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
258
·
PAIN®
A. Narayana et al. 156 (2015) 252–259
and without BTP. To the best of our knowledge, MCIDs for the SDS have not been previously reported; however, in previous studies, scores of approximately $5 on the 3 subscales of the SDS (ie, work/school, social/leisure life, and family/home life) were associated with significant functional impairment.17,28 In the present survey, SDS subscale scores were 4.8 to 5.3 among pain cohort patients with BTP and 3.3 to 3.7 among pain cohort patients without BTP. The importance of BTP as a predictor of adverse health outcomes was further supported by multivariate analysis, which revealed that BTP contributed significantly to pain-related interference in functioning, after controlling for persistent pain intensity, patient demographics, and other baseline characteristics. This finding was consistent with both preliminary multivariate tests performed on the same data set (with 1 additional patient) using ordinary least squares regression21 and a previous survey of patients with cancer.27 The earlier analyses on the same data set found that patients with BTP compared with those without BTP had both worse physical functioning on the SF-12 and higher disability on the SDS after controlling for factors such as demographics, level of persistent pain, anxiety, and depressed symptoms (both P , 0.0001). In analyses including only the pain cohort patients with BTP, greater BTP intensity significantly predicted pain interference (P , 0.0001) and worse physical functioning (P 5 0.003), and greater days per week with BTP predicted higher disability (P 5 0.034).21 In the previous cancer survey, multivariate analysis controlled for demographics, intensity of background pain, supplemental analgesic drugs, and performance status and confirmed that BTP contributed significantly to pain-related interference in function (P , 0.001), accounting for 13% of the shared variance. 27 Together, these analyses support the conclusion that BTP is an identifiable and common phenomenon in a commercially insured community-dwelling population and is associated with negative outcomes in multiple health domains. Although these associations do not establish causality, they are consistent with previous studies in their suggestion that BTP should be independently assessed as a potential target for treatment in populations with chronic pain. The interpretation of these data should be tempered by several important limitations in the study design. The assessment of BTP is difficult, and there is potential for errors in measurement. As noted previously, the BPQ used in this study differed from previous versions23,26,27 in defining a flare of moderate-intensity pain as a BTP. Previous surveys limited BTP to pain reported as severe or excruciating and, as a result, the prevalence of BTP in this study was higher. The method for patient identification yielded a relatively small number of participants with cancer pain (n 5 145), largely owing to difficulty in identifying patients with cancer through health care claims data alone, and the generalizability of the data to cancer populations is uncertain. Moreover, the information obtained about opioid administration was based on pharmacy claims data, lacked detail, and could not establish the actual use of these drugs. Finally, the study sample included only US citizens with health insurance and had a greater representation of white patients (88%-95%) than the overall US population (78%)29; extrapolation to other populations must be done cautiously.
5. Conclusions Breakthrough pain is highly prevalent and is strongly associated with adverse outcomes in a population of commercially insured community-dwelling patients with opioid-treated
chronic pain. Further studies are needed to determine whether these adverse outcomes related to health status, mood, function, and disability are causally related to the occurrence of BTP and whether treatment of BTP reduces these adverse health consequences. Given the burden of illness associated with BTP, empirical treatment of BTP should be considered in selected patients.
Conflict of interest statement At the time of this study, A. Narayana was an employee of Cephalon, Inc, now a wholly owned subsidiary of Teva Pharmaceuticals, and C. B. Frye was an employee of EPI-Q, Inc, a health outcomes research company contracted by Teva Pharmaceuticals to develop and manage this research study. Q. Harshaw and A. C. Shillington are employees of EPI-Q, Inc. J. J. Stephenson is an employee of HealthCore, Inc, a research company contracted by Teva Pharmaceuticals to develop and manage this research study. N. Katz consults, advises, and conducts research for Teva Pharmaceuticals. R. K. Portenoy has presented at a conference sponsored by Grupo Ferrer; his department has received educational or research grants from Boston Scientific, Covidien Mallinckrodt Inc, Medtronic, Otsuka Pharma, Pfizer, Purdue Pharma, and St Jude Medical, Allergan, Endo, and Prostrakan; and he has received royalties from Oxford University Press and UpToDate. This study was sponsored by Cephalon, Inc, now a wholly owned subsidiary of Teva Pharmaceuticals. Writing support was provided by Bina J. Patel, PharmD, CMPP, of Peloton Advantage, LLC, funded by Teva Pharmaceuticals.
Acknowledgements The authors thank Susan Larijani for her help with the operational aspects of the management of the study, and DataStat, Inc, for their survey data collection services. The authors also thank the participants who completed the telephone surveys.
Appendix A. Supplemental Digital Content Supplemental Digital Content associated with this article can be found online at http://links.lww.com/PAIN/A15 and http://links. lww.com/PAIN/A16. Article history: Received 18 April 2014 Received in revised form 24 September 2014 Accepted 16 October 2014
REFERENCES [1] Arnold LM, Clauw DJ, Wohlreich MM, Wang F, Ahl J, Gaynor PJ, Chappell AS. Efficacy of duloxetine in patients with fibromyalgia: pooled analysis of 4 placebo-controlled clinical trials. Prim Care Companion J Clin Psychiatry 2009;11:237–44. Available at: http://www.ncbi.nlm.nih. gov/pmc/articles/PMC2781036/pdf/pcc11237.pdf. Accessed: Dec. 10, 2014. [2] Bennett SJ, Oldridge NB, Eckert GJ, Embree JL, Browning S, Hou N, Chui M, Deer M, Murray MD. Comparison of quality of life measures in heart failure. Nurs Res 2003;52:207–16. [3] Caraceni A, Bertetto O, Labianca R, Maltoni M, Mercadante S, Varrassi G, Zaninetta G, Zucco F, Bagnasco M, Lanata L, De Conno F. Episodic (breakthrough) pain prevalence in a population of cancer pain patients. Comparison of clinical diagnoses with the QUDEI—Italian questionnaire for intense episodic pain. J Pain Symptom Manage 2012;43:833–41. Available at: http://download.journals.elsevierhealth.com/pdfs/journals/ 0885-3924/PIIS0885392412000139.pdf. Accessed: Dec. 10, 2014.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
[4] Caraceni A, Martini C, Zecca E, Portenoy RK, Ashby MA, Hawson G, Jackson KA, Lickiss N, Muirden N, Pisasale M, Moulin D, Schulz VN, Rico Pazo MA, Serrano JA, Andersen H, Henriksen HT, Mejholm I, Sjogren P, Heiskanen T, Kalso E, Pere P, Poyhia R, Vuorinen E, Tigerstedt I, Ruismaki P, Bertolino M, Larue F, Ranchere JY, Hege-Scheuing G, Bowdler I, Helbing F, Kostner E, Radbruch L, Kastrinaki K, Shah S, Vijayaram S, Sharma KS, Devi PS, Jain PN, Ramamani PV, Beny A, Brunelli C, Maltoni M, Mercadante S, Plancarte R, Schug S, Engstrand P, Ovalle AF, Wang X, Alves MF, Abrunhosa MR, Sun WZ, Zhang L, Gazizov A, Vaisman M, Rudoy S, Gomez Sancho M, Vila P, Trelis J, Chaudakshetrin P, Koh ML, Van Dongen RT, Vielvoye-Kerkmeer A, Boswell MV, Elliott T, Hargus E, Lutz L. Breakthrough pain characteristics and syndromes in patients with cancer pain. An international survey. Palliat Med 2004;18:177–83. [5] Chou R, Fanciullo GJ, Fine PG, Adler JA, Ballantyne JC, Davies P, Donovan MI, Fishbain DA, Foley KM, Fudin J, Gilson AM, Kelter A, Mauskop A, O’Connor PG, Passik SD, Pasternak GW, Portenoy RK, Rich BA, Roberts RG, Todd KH, Miaskowski C; for the American Pain Society–American Academy of Pain Medicine Opioids Guidelines Panel. Clinical guidelines for the use of chronic opioid therapy in chronic noncancer pain. J Pain 2009;10:113–30. Available at: http://download.journals.elsevierhealth.com/pdfs/journals/1526-5900/ PIIS1526590008008316.pdf. Accessed: Dec. 10, 2014. [6] Cleeland CS, Nakamura Y, Mendoza TR, Edwards KR, Douglas J, Serlin RC. Dimensions of the impact of cancer pain in a four country sample: new information from multidimensional scaling. PAIN 1996;67:267–73. [7] Cleeland CS, Ryan KM. Pain assessment: global use of the Brief Pain Inventory. Ann Acad Med Singapore 1994;23:129–38. [8] Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–19. [9] Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, Haythornthwaite JA, Jensen MP, Kerns RD, Ader DN, Brandenburg N, Burke LB, Cella D, Chandler J, Cowan P, Dimitrova R, Dionne R, Hertz S, Jadad AR, Katz NP, Kehlet H, Kramer LD, Manning DC, McCormick C, McDermott MP, McQuay HJ, Patel S, Porter L, Quessy S, Rappaport BA, Rauschkolb C, Revicki DA, Rothman M, Schmader KE, Stacey BR, Stauffer JW, von Stein T, White RE, Witter J, Zavisic S. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain 2008;9:105–21. [10] Galvez R, Marsal C, Vidal J, Ruiz M, Rejas J. Cross-sectional evaluation of patient functioning and health-related quality of life in patients with neuropathic pain under standard care conditions. Eur J Pain 2007;11:244–55. [11] Hamilton RA, Briceland LL. Use of prescription-refill records to assess patient compliance. Am J Hosp Pharm 1992;49:1691–6. [12] Hess LM, Raebel MA, Conner DA, Malone DC. Measurement of adherence in pharmacy administrative databases: a proposal for standard definitions and preferred measures. Ann Pharmacother 2006;40:1280–8. [13] Keller S, Bann CM, Dodd SL, Schein J, Mendoza TR, Cleeland CS. Validity of the brief pain inventory for use in documenting the outcomes of patients with noncancer pain. Clin J Pain 2004;20:309–18. [14] Kessler RC, Ames M, Hymel PA, Loeppke R, McKenas DK, Richling DE, Stang PE, Ustun TB. Using the World Health Organization Health and Work Performance Questionnaire (HPQ) to evaluate the indirect workplace costs of illness. J Occup Environ Med 2004;46(6 suppl):S23–37. [15] Kessler RC, Barber C, Beck A, Berglund P, Cleary PD, McKenas D, Pronk N, Simon G, Stang P, Ustun TB, Wang P. The World Health Organization
www.painjournalonline.com
[16] [17]
[18]
[19]
[20] [21]
[22]
[23]
[24]
[25]
[26] [27] [28] [29]
[30]
[31]
[32]
259
Health and Work Performance Questionnaire (HPQ). J Occup Environ Med 2003;45:156–74. Kroenke K, Spitzer RL, Williams JB. The Patient Health Questionnaire-2: validity of a two-item depression screener. Med Care 2003;41:1284–92. Leon AC, Olfson M, Portera L, Farber L, Sheehan DV. Assessing psychiatric impairment in primary care with the Sheehan Disability Scale. Int J Psychiatry Med 1997;27:93–105. Lowe B, Decker O, Muller S, Brahler E, Schellberg D, Herzog W, Herzberg PY. Validation and standardization of the Generalized Anxiety Disorder Screener (GAD-7) in the general population. Med Care 2008;46:266–74. Manchikanti L, Singh V, Caraway DL, Benyamin RM. Breakthrough pain in chronic non-cancer pain: fact, fiction, or abuse. Pain Physician 2011;14: E103–17. Available at: http://www.painphysicianjournal.com/linkout_vw.php? issn51533-3159&vol514&page5E103. Accessed: Dec. 10, 2014. Markman JD. Not so fast: the reformulation of fentanyl and breakthrough chronic non-cancer pain. PAIN 2008;136:227–9. Narayana A, Katz N, Shillington AC, Stephenson JJ, Portenoy RK. The National Breakthrough Pain Survey: mulitvariate predictors of functional status, pain interference, and disability [poster 71]. Presented at PAINWeek 2012; September 5-8, 2012; Las Vegas, NV. Perez C, Navarro A, Saldana MT, Martinez S, Rejas J. Patient-reported outcomes in subjects with painful trigeminal neuralgia receiving pregabalin: evidence from medical practice in primary care settings. Cephalalgia 2009;29:781–90. Portenoy RK, Bennett DS, Rauck R, Simon S, Taylor D, Brennan M, Shoemaker S. Prevalence and characteristics of breakthrough pain in opioid-treated patients with chronic noncancer pain. J Pain 2006;7: 583–91. Portenoy RK, Bruns D, Shoemaker B, Shoemaker SA. Breakthrough pain in community-dwelling patients with cancer pain and noncancer pain, part 1: prevalence and characteristics. J Opioid Manag 2010;6:97–108. Portenoy RK, Bruns D, Shoemaker B, Shoemaker SA. Breakthrough pain in community-dwelling patients with cancer pain and noncancer pain, part 2: impact on function, mood, and quality of life. J Opioid Manag 2010;6:109–16. Portenoy RK, Hagen NA. Breakthrough pain: definition, prevalence and characteristics. PAIN 1990;41:273–81. Portenoy RK, Payne D, Jacobsen P. Breakthrough pain: characteristics and impact in patients with cancer pain. PAIN 1999;81:129–134. Sheehan DV, Harnett-Sheehan K, Raj BA. The measurement of disability. Int Clin Psychopharmacol 1996;11(suppl 3):89–95. US Census Bureau. State and Country QuickFacts. US Department of Commerce. Available at: http://quickfacts.census.gov/qfd/states/ 00000.html. Accessed April 16, 2014. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care 1996;34:220–33. Zeppetella G, O’Doherty CA, Collins S. Prevalence and characteristics of breakthrough pain in cancer patients admitted to a hospice. J Pain Symptom Manage 2000;20:87–92. Available at: http://download.journals. elsevierhealth.com/pdfs/journals/0885-3924/PIIS0885392400001615. pdf. Accessed: Dec. 10, 2014. Zeppetella G, O’Doherty CA, Collins S. Prevalence and characteristics of breakthrough pain in patients with non-malignant terminal disease admitted to a hospice. Palliat Med 2001;15:243–6.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
National Breakthrough Pain Study: prevalence, characteristics, and associations with health outcomes Arvind Narayanaa, Nathaniel Katzb,c, Alicia C. Shillingtond, Judith J. Stephensone, Qing Harshawd, Carla B. Fryef, Russell K. Portenoyg,*
Abstract The National Breakthrough Pain Study is a large observational study that assessed breakthrough pain (BTP) in a population of commercially insured community-dwelling patients with opioid-treated chronic pain. Eligible patients were identified from an administrative claims database, and consenting patients were asked to complete a structured telephone interview and several validated questionnaires. Questionnaires assessed pain interference with function (Brief Pain Inventory-Short Form), health status (Short Form 12 [SF-12] Health Survey), disability (Sheehan Disability Scale), work performance (World Health Organization Health and Work Performance Questionnaire), and mood (Generalized Anxiety Disorder-7 Screener [GAD-7] and Patient Health Questionnaire-2 [PHQ-2]). Of 2198 patients interviewed, 1278 patients had persistent pain controlled with opioid therapy; 1023 (80%) of these patients reported BTP. Patients had a median of 2.0 episodes of BTP per day (range, 1-50) and a median duration of BTP of 45 minutes (range, 1-720). Compared with patients without BTP, patients with BTP had more pain-related interference in function (Brief Pain Inventory, mean 6 SD: 34.2 6 15.6 vs 25.0 6 15.7 [P , 0.001]), worse physical health (SF-12 physical component score: 29.9 6 9.6 vs 35.1 6 10.4 [P , 0.001]) and mental health (SF-12 mental component score: 47.4 6 11.3 vs 49.3 6 10.4 [P , 0.001]), more disability (Sheehan Disability Scale global impairment score: 15.1 6 9.1 vs 10.6 6 8.5; World Health Organization Health and Work Performance Questionnaire absolute absenteeism: 12.4 6 59.9 vs 7.7 6 44.9 hours [both P , 0.001]), and worse mood (GAD-7 score: 7.4 6 5.9 vs 5.9 6 5.4; PHQ-2 anhedonia score: 1.2 6 1.1 vs 0.9 6 1.0 [both P , 0.001]). In this population of community-dwelling patients with opioid-treated chronic pain, BTP was highly prevalent and associated with negative outcomes. This burden of illness suggests the need for specific treatment plans. Keywords: Epidemiology, Breakthrough pain, Quality of life, Patient-reported outcomes, National Breakthrough Pain Study
1. Introduction Breakthrough pain (BTP) may be defined as a transitory episode of pain that occurs on a background of persistent pain that is otherwise controlled on a stable opioid regimen.26 Prevalence of BTP varies with diagnosis, severity of illness, and treatment setting. Among patients with controlled persistent cancer pain, 33% to 89% experience BTP.24,26,27,31 Studies of varied populations with opioid-treated pain unrelated to cancer have reported BTP in 48% to 74% of patients.23,24,32 Although controversy surrounds the clinical implications of BTP in populations without cancer,5,19,20 these studies suggest at minimum that the phenomenon is prevalent in all opioid-treated chronic pain populations.
Several surveys of patients with cancer indicate that BTP is associated with more severe pain, less effective analgesic treatment, impaired ability to function, negative psychosocial outcomes, and relatively poorer quality of life.4,25,27 More limited data in patients with chronic pain unrelated to cancer suggest similar associations.25 All of these surveys have underscored the need for additional research to assess the clinical impact of BTP in larger samples of patients. The National Breakthrough Pain Study (NBTPS) was designed as a large observational study of community-dwelling patients with chronic pain identified from a national database of commercially insured patients. The aim was to assess the prevalence and characteristics of BTP and to explore its associations with diverse health-related outcomes.
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article. a
2. Methods
*Corresponding author. Address: MJHS Institute for Innovation in Palliative Care, 39 Broadway, 12th Floor, New York NY 10006, USA. Tel.: (212) 649-5560; fax: (212) 649-5544. E-mail address: [email protected] (R. K. Portenoy).
The NBTPS used telephone survey methodology to conduct a cross-sectional observational study. Patient interviews were conducted between April and December 2010. An institutional review board approved the survey, and all patients provided verbal consent before responding to questions.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.painjournalonline.com).
2.1. Patient selection
Teva Pharmaceuticals, Frazer, PA, USA, b Tufts University School of Medicine, Boston, MA, USA, c Analgesic Solutions, Natick, MA, USA, d EPI-Q, Inc, Oak Brook, IL, USA, e HealthCore, Inc, Wilmington, DE, USA, f Thayer County Health Services, Hebron, NE, USA, g MJHS Institute for Innovation in Palliative Care, New York, NY, USA
PAIN 156 (2015) 252–259 © 2015 International Association for the Study of Pain http://dx.doi.org/10.1097/01.j.pain.0000460305.41078.7d
252
·
A. Narayana et al. 156 (2015) 252–259
Potentially eligible patients were identified from the HealthCore Integrated Research Database (HIRDSM), an administrative database containing claims data from one of the largest health PAIN®
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
benefits companies in the United States. The claims originated from patients enrolled in 1 of 14 US health plans. For this study, the HIRD sampling frame consisted of 6.4 million patients who were $18 years of age and were currently active health plan members enrolled for $12 months before the patient sample list was extracted. From this sampling frame, 50,000 health plan members were randomly selected and stratified to approximate the US general population (based on the 2000 US National Census) according to the following characteristics: age (18-34 years, 35-49 years, 50-64 years, 65-74 years, and $75 years), sex, and health plan census region (Northeast, South, Midwest, and West). Members who were likely to have chronic pain based on diagnosis codes (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM]; see Supplemental Digital Content, http://links.lww.com/PAIN/A15) and opioid use were oversampled to ensure adequate representation in the applicable analysis cohorts. To exclude patients receiving methadone for treatment of addiction, patients with an ICD-9-CM diagnosis code (medical claims) or Healthcare Common Procedure Coding System code (ambulatory services) for drug abuse or dependence concurrent with a pharmacy claim for methadone were rejected from the sample (see Supplemental Digital Content, http://links.lww.com/PAIN/A16). Patients with an invalid or missing telephone number and those who were on the HealthCore do-not-call list also were excluded. From this representative sample of 50,000 health plan members, 2 potential analysis cohorts were selected (Fig. 1). To create these analysis cohorts, a series of smaller samples (replicates) was selected, with each replicate being representative of the larger stratified sampling frame but also characterized by the criteria used to define each of the analysis cohorts. These replicates were released into a telephone
www.painjournalonline.com
253
survey field, which provided the information necessary to call each patient. One of the analysis cohorts—a “pain” cohort—included patients with controlled persistent pain, with or without BTP. This cohort was initially identified from the HIRD by a medical claim with an ICD-9-CM diagnosis code consistent with chronic pain and pharmacy claims consistent with long-term opioid use. Specifically, patients were eligible for this cohort if they had (1) at least 2 medical claims separated by $3 months that included an ICD-9-CM code consistent with back pain, neuropathic pain, traumatic injury, osteoarthritis, fibromyalgia, cancer pain, or other chronic pain conditions, and (2) at least 3 opioid prescription claims within 3 months with a medication refill adherence of $90%. Medication refill adherence assesses daily use by dividing the total days’ supply of medication by the number of days in the study period and multiplying by 100 to provide a percent adherence rate.11,12 At the time of the subsequent telephone interview, the characterization of the patient as appropriate for the pain cohort was confirmed. The second analysis cohort—a “control” cohort—was identified by lack of evidence of a pain diagnosis or chronic opioid use. This status also was confirmed during the subsequent telephone contact. 2.2. Telephone survey Up to 6 calls were made in an effort to identify and speak with the patient identified through the sampling process. The interviewers were trained by investigators (A.C.S. and C.B.F.) by completing a question and answer session on the survey objectives, inclusion and exclusion criteria, survey script, and procedures for the callback protocol. The interviewers were counseled to strictly
Figure 1. National Breakthrough Pain–Survey Design. The original study protocol consisted of 5 analysis groups (control cohort, patients with cancer with controlled persistent pain without breakthrough pain (BTP), patients with cancer with controlled persistent pain with BTP, patients without cancer with controlled persistent pain without BTP, and patients without cancer with controlled persistent pain with BTP). For this analysis, data for patients with and without cancer were combined.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
254
·
PAIN®
A. Narayana et al. 156 (2015) 252–259
adhere to the fully structured interview script and were not allowed to provide improvised assistance or instruction in an attempt to interpret any portions of the script in response to questions. They were permitted only to reread script portions in response to questions and were randomly monitored by floor supervisors to ensure strict adherence to the script (interview script available on request). Patients were interviewed if they were fluent in English, could communicate by telephone, and consented to the study. Those who were capable of participating were informed that the survey was designed “to learn about the health care status and needs in the population.” They were told that patients were randomly selected, participation was voluntary, responses were confidential, and participation would not affect health care benefits. Verbal informed consent was elicited before proceeding with the interview, and each patient was paid $25 as compensation for their time. The telephone survey required less than 30 minutes on average (range, 5-50 minutes) to complete. The interview began with questions about demographics not captured in claims data (eg, marital status, household income, ethnicity, educational level) and questions designed to confirm eligibility. Patients selected for the pain cohort were first asked whether they had persistent pain and could respond affirmatively to a question designed to identify clinically significant pain: “Some people have pain that causes significant problems with daily functioning, mood, sleep, or their overall quality of life. Have you had pain that has caused these types of significant problems over the past 3 months?” Patients with acute or intermittent pain and those who did not respond affirmatively to the statement did not continue with the survey. Patients with clinically significant persistent pain were then asked to confirm that they took opioid medication daily; those who did not also discontinued the interview. Patients with clinically significant opioid-treated pain were then queried about the extent to which the pain was controlled by the medication. Persistent pain was considered controlled when the patient reported baseline pain as not more than moderate on a verbal rating scale of pain intensity (absent, mild, moderate, or severe or excruciating). Only patients who described their persistent pain as absent, mild, or moderate were permitted to continue with the rest of the telephone survey. Through this initial series of questions, patients who were not eligible for the pain cohort were discontinued, and all patients remaining in the pain cohort were characterized by clinically significant persistent pain controlled by an opioid regimen. These patients were then asked whether they experienced BTP (Fig. 1), which was defined as temporary flares of moderate-to-severe pain that were more intense than the persistent pain. Eligibility for the control cohort was confirmed through a series of similar questions. Patients who were allowed to continue with the rest of the survey verified that they had no pain on the day of the survey; no pain that caused significant problems with daily function, mood, sleep, or quality of life (thus, some patients in the control group may have had mild pain); and no history of chronic opioid use. Patients in both the pain cohort and the control cohort provided verbal responses to several validated questionnaires intended to assess quality of life, functioning, and productivity. Verbal responses to the Brief Pain Inventory (BPI)-Short Form6,7,13 were elicited from patients with pain, and patients in the pain cohort with BTP responded to a structured interview designed to characterize their BTP.
2.3. Questionnaires and structured interview
2.3.1. Short Form 12 Health Survey (version 2) The Short Form 12 (SF-12) includes a subset of questions from the widely used Short Form 36 Health Survey. Short Form 12 is a valid measure of health status and includes a physical component summary score and the mental component summary score to assess physical and mental health, respectively.30 2.3.2. Sheehan Disability Scale The Sheehan Disability Scale (SDS) is a widely used measure of disability that assesses the impact of health conditions on productivity, specifically days out of role.1,10,22,28 The SDS has a global interference score, which is the sum of the 3 separate interference subscales. 2.3.3. World Health Organization Health and Work Performance Questionnaire-Short Form The World Health Organization Health and Work Performance Questionnaire-Short Form evaluates workplace absenteeism and presenteeism and thereby assesses 1 element associated with disability.14,15 2.3.4. Generalized Anxiety Disorder-7 Screener The Generalized Anxiety Disorder-7 (GAD-7) is a brief and validated screen for anxiety disorders.18 2.3.5. Patient Health Questionnaire-2 The validated Patient Health Questionnaire-2 (PHQ-2) screens for depressed mood and includes separate items for dysphoria and anhedonia.16 2.3.6. Brief Pain Inventory-Short Form The BPI-Short Form is a widely used instrument that includes pain intensity measures assessed on a numeric scale of 0 to 10 and a 7-item subscale that evaluates the extent to which pain interferes with varied domains of function, such as sleep, walking, social relations, and enjoyment of life.6,7,13 2.3.7. Breakthrough Pain Questionnaire The Breakthrough Pain Questionnaire (BPQ) is a structured interview adapted from a format used in previous studies.23,26,27 Responses to the BPQ characterized persistent pain and BTP, if experienced, in terms of temporal characteristics, location, duration, quality, and other features. 2.4. Statistical analysis Patients’ administrative claims data from the HIRD were merged with their survey data acquired during the telephone interview to conduct the analysis. The HIRD provided demographic data for the 12-month period before the date of the survey (eg, age group, gender, region), information about comorbidities (documented by ICD-9-CM codes on medical claims), medical resource utilization (eg, inpatient hospitalizations, emergency department visits, outpatient services), direct health care costs, and pharmacy claims and costs. Comorbidity burden was assessed through the
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
Deyo-modified Charlson Comorbidity Index8 and was reported as mean 6 SD. Demographic characteristics, prevalence of BTP, and pain phenomenology (eg, location, duration, severity) were analyzed descriptively using frequencies, means, and SDs, as appropriate. Mean 6 SD 24-hour pain intensity and interference subscores from the BPI-Short Form were analyzed between patients with BTP and those without BTP using a Student t test. Persistent pain severity (ordinal scale; recorded as absent, mild, or moderate) was compared between those with BTP and those without BTP using frequency distributions and x2 tests with a gamma statistic. Continuously distributed characteristics of BTP (eg, number of BTP flares per day, duration until peak intensity, duration from start of flare to end) were examined descriptively (means, SDs, and medians) only in patients with BTP. The ability to predict BTP flares was reported as proportions, rated on a Likert scale (never, sometimes, often, almost always, or always). The relationship between start of flare and last medication dosing was reported as a proportion of those responding affirmatively (pain temporally related to the last dose) vs negatively. Location of the worst BTP was reported as a proportion of those responding back, legs, neck, shoulder, arm, or head. Data from the questionnaire responses were summarized for the control cohort, the pain cohort with BTP, and the pain cohort without BTP. The SF-12 physical and mental health component summary scores were calculated as normed scores.30 The SDS scales were coded into global and interference subscales. Days out of role were recorded as past 30 days and past 365 days. Each scale was reported as mean 6 SD and compared across the 3 cohorts using an analysis of variance with a post hoc test (Bonferroni) for pairwise comparisons if the overall test result among groups was statistically significant. A P value of ,0.05 was considered statistically significant for all analyses. Univariate associations with pain interference (BPI-Short Form subscore) were assessed by Pearson r and Cohen d. Parameters with effect sizes $0.3 and/or with P values ,0.1 were generally considered for multivariate model inclusion after examining for potential multicollinearity. Impact of BTP on pain interference (BPI-Short Form subscore) was modeled using an ordinary least squares regression analysis, controlling for relevant demographics, persistent pain severity, mood, and comorbidity.
3. Results Interviews were attempted with 8703 patients. Of these, 3725 refused to participate, 2780 were deemed ineligible after a partial interview, and 1 was excluded because of missing medical claims data. Of the 2197 eligible patients who completed the interview, 1278 were in the pain cohort and 919 were in the control cohort. Of the 1278 in the pain cohort, 1023 (80.0%) reported BTP (Fig. 1). Patient demographics and other baseline clinical characteristics are summarized in Table 1. Across the control and pain cohorts, the mean age was 47.8 to 52.0 years, and the proportion reporting white race was 87.9% to 95.2%. Only 145 of the 1278 patients (11.3%) with controlled persistent pain had cancer. Work-related disability was reported by 0.7% of patients in the control cohort, 16.9% in the pain cohort without BTP, and 31.3% in the pain cohort with BTP. The majority of patients in the pain cohort, all of whom had controlled persistent pain, reported moderate (vs absent or mild) persistent pain intensity, regardless of the presence of BTP (Table 2). However, the proportion of patients reporting moderate vs mild persistent pain was greater among those with BTP than it was among those without BTP (P 5 0.027). The multivariate regression model evaluating the impact
www.painjournalonline.com
255
Table 1
Patient demographics and other baseline characteristics. Variable Age, years Mean (SD) Sex, n (%) Male Female Race, n (%)* White Black Hispanic Native American Asian Other† Employment status* Employed Self-employed Disabled Retired Other‡ Education ,High school High school Junior college College Postgraduate Annual income ,$20,000 $20,000-$40,000 $40,000-$80,000 .$80,000 Pain due to cancer, n (%) Yes Deyo-Charlson comorbidity index Mean (SD) Visit to pain specialist, n (%) Yes
Control cohort
Pain cohort without BTP
Pain cohort with BTP
n 5 915 47.8 (17.6) n 5 919 426 (46.4) 493 (53.6) n 5 911 801 (87.9) 37 (4.1) 42 (4.6) 12 (1.3) 41 (4.5) 26 (2.8) n 5 918 535 (58.3) 151 (16.4) 6 (0.7) 121 (13.2) 215 (23.4) n 5 133 3 (2.3) 48 (36.1) 19 (14.3) 57 (42.9) 6 (4.5) n 5 110 10 (9.1) 16 (14.5) 38 (34.5) 46 (41.8) NA NA NA
n 5 255 52.0 (11.4) n 5 255 106 (41.6) 149 (58.4) n 5 254 236 (92.9) 6 (2.4) 4 (1.6) 6 (2.4) 1 (0.4) 4 (1.6) n 5 254 110 (43.3) 43 (16.9) 43 (16.9) 30 (11.8) 50 (19.7) n 5 253 8 (3.2) 106 (41.9) 43 (17.0) 87 (34.4) 9 (3.6) n 5 237 30 (12.7) 64 (27.0) 76 (32.1) 67 (28.3) n 5 255 33 (12.9) n 5 255
n 5 1023 49.8 (9.8) n 5 1023 418 (40.9) 605 (59.1) n 5 1023 974 (95.2) 22 (2.2) 24 (2.3) 26 (2.5) 6 (0.6) 12 (1.2) n 5 1021 367 (35.9) 142 (13.9) 320 (31.3) 68 (6.7) 273 (26.7) n 5 1021 43 (4.2) 430 (42.1) 181 (17.7) 345 (33.8) 22 (2.2) n 5 916 132 (14.4) 236 (25.8) 318 (34.7) 230 (25.1) n 5 1023 112 (10.9) n 5 1023
NA NA
1.1 (2.0) n 5 255
1.0 (1.8) n 5 1023
NA
99 (38.8)
436 (42.6)
* Respondents could select more than 1 race or employment status or could choose not to answer. † “Other” includes mixed race, Arabic, Armenian, Creole, Cuban-Dominican, Hebrew, Indian, Mediterranean, or Middle Eastern. ‡ “Other” includes unemployed and looking for work, unemployed and not looking for work, student, homemaker, something else, temporarily laid off, maternity leave, and illness. BTP, breakthrough pain; NA, not applicable or not available.
of BTP on pain interference controlled for the difference in the proportion of patients with moderate pain. The BPQ provided descriptive information about the BTP experienced by the 1023 patients in the pain cohort (Table 3). Overall, 688 (67.3%) patients with BTP reported that they treated their BTP; of these, 358 (52.0%) patients used medication. The most common ($5%) of these medications were hydrocodone with acetaminophen (181/358 [50.6%]), oxycodone (51/358 [14.2%]), and oxycodone with acetaminophen (44/358 [12.3%]); only 2 patients reported using transmucosal immediate-release fentanyl formulations for BTP. The perceived effects of medication or other treatments for BTP were not evaluated. Patients with BTP reported a median of 2.0 episodes per day (range, 1-50), with a median duration of 45 minutes (range, 1-720). The back was the most common location for BTP flares (55.2%). Approximately two-thirds of patients could not reliably predict the timing of BTP, but approximately one-third noted that BTP usually occurred at some predictable number of hours after taking a regularly scheduled dose of an analgesic.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
256
·
PAIN®
A. Narayana et al. 156 (2015) 252–259
Table 2
Responses to the Brief Pain Inventory and persistent pain severity. Variable BPI Pain intensity over 24 h, mean (SD) Worst Least Average Interference in 24 h, mean (SD) Activity Mood Ability to walk Ability to work Social relations Sleep Enjoyment of life Total score in past 24 h, mean (SD) Persistent pain severity Baseline pain over past week, n (%) Mild Moderate Absent
P†
Control cohort (n 5 133)*
Pain cohort without BTP (n 5 255)
Pain cohort with BTP (n 5 1023)
2.0 (2.0) 0.6 (1.2) 1.5 (1.7)
5.0 (2.1) 2.6 (1.8) 4.0 (1.8)
5.9 (1.9) 3.2 (1.8) 4.6 (1.7)
,0.001 ,0.001 ,0.001
0.7 (1.5) 0.7 (1.6) 0.7 (1.7) 0.9 (1.8) 0.5 (1.4) 0.8 (1.7) 0.9 (1.7) 5.0 (9.4)
3.6 (2.7) 3.6 (2.8) 3.0 (2.9) 4.0 (2.9) 2.6 (2.6) 4.2 (3.0) 4.0 (2.6) 25.0 (15.7)
5.3 (2.8) 4.7 (2.8) 4.1 (3.0) 5.5 (2.9) 3.7 (3.0) 5.6 (2.9) 5.2 (2.8) 34.2 (15.6)
,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001
— — —
63 (24.7) 187 (73.3) 5 (2.0)
191 (18.7) 823 (80.4) 9 (0.9)
0.027
* Respondents in the control group who had mild pain that did not cause significant problems with daily function, mood, sleep, or their quality of life and who did not have pain on the day of the survey received the BPI. † t-tests were performed between patients with BTP and patients without BTP. P , 0.05 was considered statistically significant. BPI, Brief Pain Inventory; BTP, breakthrough pain.
Across the control cohort, the pain cohort with BTP, and the pain cohort without BTP, the SF-12 physical and mental component summary scores, SDS, World Health Organization Health and Work Performance Questionnaire, GAD-7, and PHQ-2 all showed statistically significant group differences (Table 4). In all cases,
Table 3
Responses to the breakthrough pain (BTP) Questionnaire (pain cohort with BTP only)*. Variable Number of BTP flares per day Median Mean Duration until peak pain, min Median Mean Duration from flare start to end, min Median Mean Ability to predict BTP flares, n (%) Never Sometimes Often Almost always Always Do not know When pain flares occur, they start at the same time after the last scheduled dose of medication, n (%)† Worst BTP pain location, n (%)‡ Back Leg(s) Neck Shoulder(s) Arm(s) Head * 1023 patients had BTP, which was 5% of overall sampled population in the study. † Observed n 5 961, not all patients answered all of the questions. ‡ Respondents could select more than 1 location.
N 5 1023 2.0 2.9 10.0 32.0 45.0 101.4 324 (31.6) 363 (35.4) 161 (15.7) 107 (10.4) 55 (5.4) 14 (1.4) 323 (33.6)
n 5 1022 564 (55.2) 218 (21.3) 166 (16.2) 141 (13.8) 66 (6.5) 58 (5.7)
patients with persistent pain and BTP had the most impairment. Patients in this subgroup experienced the worst health status, more depressed and anxious mood, and more disability than patients in the pain cohort with persistent pain who did not report BTP and, to a larger extent, patients in the control group. The BPI provided additional information about the extent to which pain interfered with varied domains of function. Compared with those who did not experience BTP, patients with BTP had significantly greater mean pain interference in all functional domains (Table 2). In a multivariate model controlling for relevant covariates (including persistent pain intensity, anxiety, depressed mood, and patient demographics), the major factors associated with pain interference (P , 0.001) included the presence of BTP, level of baseline constant pain, depressed mood (as indicated by GAD-7 scores), and anxiety (as indicated by PHQ-2 scores) (Table 5). These variables together explained 36% of the variation of pain interference (adjusted R2 5 0.360 [P , 0.001]). For patients with BTP, the predicted pain interference (0-10 scale) was 0.907 points higher than for patients without BTP, holding all other variables constant. Further subgroup analyses revealed that between patients with mild baseline pain or moderate baseline pain, those with BTP had significantly greater mean pain interference than those without BTP (mild baseline pain [3.6 vs 2.6, respectively; P 5 0.002] and moderate baseline pain [5.2 vs 3.9, respectively; P , 0.001]).
4. Discussion The NBTPS is an unprecedented attempt to gather data about BTP from a nationally representative population of commercially insured community-dwelling patients with clinically significant chronic pain managed with opioid therapy. More patients were surveyed than in any similar study to date, and the study was the first to identify patients using administrative claims data from 1 of the largest health benefits companies in the United States. The data substantially augment the results of earlier surveys of BTP, which have been conducted in smaller samples of ambulatory patients,23,24 samples of cancer inpatients or cancer patients
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
www.painjournalonline.com
257
Table 4
Responses to health status, functioning, mood, and productivity survey instruments. Variable, mean (SD) SF-12 physical component summary score SF-12 mental component summary score SDS global Interference with work or school Interference with social or leisure activities Interference with family life or home responsibilities SDS days out of role Past 30 days Past 365 days HPQ-SF Absolute absenteeism, h‡ Absolute presenteeism§ GAD-7‖ PHQ-2{ Anhedonia Dysphoria
Control cohort
Pain cohort without BTP
Pain cohort with BTP
53.4 (6.6) 54.7 (5.9) 1.7 (4.0) 0.6 (1.5) 0.6 (1.6) 0.5 (1.4)
35.1 (10.4) 49.3 (10.4) 10.6 (8.5) 3.7 (3.3) 3.6 (3.0) 3.3 (2.9)
29.9 (9.6)*† 47.4 (11.3)*† 15.1 (9.1)*† 5.3 (3.6)*† 5.1 (3.2)*† 4.8 (3.1)*†
0.3 (1.8) 3.2 (14.0)
4.5 (7.4) 59.1 (99.6)
8.9 (10.3)*† 112.2 (134.3)*†
1.4 (45.7) 87.0 (12.0) 1.9 (2.6)
7.7 (44.9) 83.5 (14.9) 5.9 (5.4)
12.4 (59.9)*† 79.8 (18.7)*† 7.4 (5.9)*†
0.1 (0.3) 0.8 (0.3)
0.9 (1.0) 0.6 (0.9)
1.2 (1.1)*† 0.9 (1.0)*†
ANOVA test was used for numeric variables with post hoc testing (Bonferroni) for pairwise comparison if overall test among groups was statistically significant. x2 test was used for overall and pairwise comparisons for categorical variables. P , 0.05 was considered statistically significant. * P , 0.001 for BTP group vs control group. † P , 0.001 for BTP group vs no BTP group. ‡ Scored in terms of raw hours lost per 4 wk (a higher score indicates a higher amount of absenteeism). § Lower bound of 0 (total lack of performance during time on the job) and upper bound of 100. ‖ GAD-7 score .8 suggests anxiety or panic disorder. { PHQ-2 scores $3 (sum of anhedonia and dysphoria scores) suggest major depression. BTP, breakthrough pain; GAD-7, Generalized Anxiety Disorder-7; HPQ-SF, Health and Work Performance Questionnaire; PHQ-2, Patient Health Questionnaire-2; SDS, Sheehan Disability Scale; SF-12, 12-Item Short Form Health Survey.
with advanced illness,3,4,26,27,31 and a sample of patients with advanced illnesses other than cancer.23,32 Most patients in this study had pain unrelated to cancer. Breakthrough pain was highly prevalent in this population (80%), exceeding the prevalence of 48% to 74% reported in earlier surveys of patients with chronic noncancer pain.23,24,32 To some extent, this higher prevalence may have been related to the definitions applied to BTP in the various surveys. The current survey acquired information about pains that were experienced as more intense than the baseline persistent pain, whether the intensity of the temporary flare was moderate or severe; other surveys limited assessment to pain perceived to be severe or excruciating. Compared with the largest survey of ambulatory patients to date (N 5 228),23 this study showed a similar number of episodes of BTP per day (median, 2.0 for both studies), time until peak pain (median, 10 minutes for both studies), and duration of BTP episodes (median, 45 minutes and 60 minutes, respectively). Breakthrough pain likely related to the end of the dosing interval
Table 5
Multiple regression model with pain interference (Brief Pain Inventory subscore) as the dependent variable. Model
Unstandardized coefficients Parameter estimate
BTP (present vs absent) Education Male gender GAD-7 score Income Level of constant pain PHQ-2 Constant
0.907 20.065 20.291 0.102 20.169 1.105 0.410 0.585
t
P
6.383 21.129 22.477 7.498 22.671 8.302 6.944 1.261
0.000 0.259 0.013 0.000 0.008 0.000 0.000 0.207
SE 0.142 0.058 0.117 0.014 0.063 0.133 0.059 0.464
P , 0.05 was considered statistically significant. BTP, breakthrough pain; GAD-7, Generalized Anxiety Disorder-7; PHQ-2, Patient Health Questionnaire-2.
(reported as BTP flares that occurred at the same time after the last scheduled dose of medication) was more common in this study (33% vs 19%) and suggests that there may be an opportunity to address BTP in a substantial proportion of patients by adjusting the dose or interval of the around-the-clock opioid regimen or using medications with a longer duration of action. However, it is important to note that the interpretation of this finding is limited in that patients were not specifically asked whether their BTP episode occurred before the next scheduled dose of their medication was due. In this study, patients with controlled persistent pain and BTP reported substantially worse health status, mood, pain interference with functioning, and disability than patients with controlled persistent pain and no BTP. Not surprisingly, the differences were greatest when comparing those with BTP and those in the control cohort; patients with persistent pain and no BTP had better scores than patients with BTP and worse scores than patients in the control cohort. Similarly, previous surveys of patients with cancer found that BTP is associated with functional impairment and relatively poorer quality of life,4,27 and a more recent survey of community-dwelling patients with and without cancer reported comparable findings.25 These previous surveys did not assess measures such as the SDS and GAD-7. Clinical significance of the group differences in adverse outcomes may be illuminated by comparison with the limited data on minimum clinically important differences (MCID) available in the literature. Dworkin et al.9 suggested that the MCID for BPI interference is $1 point in patients with chronic pain, and Bennett et al.2 reported that $1.26- and $2.28-point differences in the SF-12 physical component and mental component summary scores, respectively, were clinically significant for heart failure patients. The present survey found that these MCIDs were exceeded for all BPI interference assessments and the SF-12 physical component summary score when comparing differences in average scores between patients in the pain cohort with
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
258
·
PAIN®
A. Narayana et al. 156 (2015) 252–259
and without BTP. To the best of our knowledge, MCIDs for the SDS have not been previously reported; however, in previous studies, scores of approximately $5 on the 3 subscales of the SDS (ie, work/school, social/leisure life, and family/home life) were associated with significant functional impairment.17,28 In the present survey, SDS subscale scores were 4.8 to 5.3 among pain cohort patients with BTP and 3.3 to 3.7 among pain cohort patients without BTP. The importance of BTP as a predictor of adverse health outcomes was further supported by multivariate analysis, which revealed that BTP contributed significantly to pain-related interference in functioning, after controlling for persistent pain intensity, patient demographics, and other baseline characteristics. This finding was consistent with both preliminary multivariate tests performed on the same data set (with 1 additional patient) using ordinary least squares regression21 and a previous survey of patients with cancer.27 The earlier analyses on the same data set found that patients with BTP compared with those without BTP had both worse physical functioning on the SF-12 and higher disability on the SDS after controlling for factors such as demographics, level of persistent pain, anxiety, and depressed symptoms (both P , 0.0001). In analyses including only the pain cohort patients with BTP, greater BTP intensity significantly predicted pain interference (P , 0.0001) and worse physical functioning (P 5 0.003), and greater days per week with BTP predicted higher disability (P 5 0.034).21 In the previous cancer survey, multivariate analysis controlled for demographics, intensity of background pain, supplemental analgesic drugs, and performance status and confirmed that BTP contributed significantly to pain-related interference in function (P , 0.001), accounting for 13% of the shared variance. 27 Together, these analyses support the conclusion that BTP is an identifiable and common phenomenon in a commercially insured community-dwelling population and is associated with negative outcomes in multiple health domains. Although these associations do not establish causality, they are consistent with previous studies in their suggestion that BTP should be independently assessed as a potential target for treatment in populations with chronic pain. The interpretation of these data should be tempered by several important limitations in the study design. The assessment of BTP is difficult, and there is potential for errors in measurement. As noted previously, the BPQ used in this study differed from previous versions23,26,27 in defining a flare of moderate-intensity pain as a BTP. Previous surveys limited BTP to pain reported as severe or excruciating and, as a result, the prevalence of BTP in this study was higher. The method for patient identification yielded a relatively small number of participants with cancer pain (n 5 145), largely owing to difficulty in identifying patients with cancer through health care claims data alone, and the generalizability of the data to cancer populations is uncertain. Moreover, the information obtained about opioid administration was based on pharmacy claims data, lacked detail, and could not establish the actual use of these drugs. Finally, the study sample included only US citizens with health insurance and had a greater representation of white patients (88%-95%) than the overall US population (78%)29; extrapolation to other populations must be done cautiously.
5. Conclusions Breakthrough pain is highly prevalent and is strongly associated with adverse outcomes in a population of commercially insured community-dwelling patients with opioid-treated
chronic pain. Further studies are needed to determine whether these adverse outcomes related to health status, mood, function, and disability are causally related to the occurrence of BTP and whether treatment of BTP reduces these adverse health consequences. Given the burden of illness associated with BTP, empirical treatment of BTP should be considered in selected patients.
Conflict of interest statement At the time of this study, A. Narayana was an employee of Cephalon, Inc, now a wholly owned subsidiary of Teva Pharmaceuticals, and C. B. Frye was an employee of EPI-Q, Inc, a health outcomes research company contracted by Teva Pharmaceuticals to develop and manage this research study. Q. Harshaw and A. C. Shillington are employees of EPI-Q, Inc. J. J. Stephenson is an employee of HealthCore, Inc, a research company contracted by Teva Pharmaceuticals to develop and manage this research study. N. Katz consults, advises, and conducts research for Teva Pharmaceuticals. R. K. Portenoy has presented at a conference sponsored by Grupo Ferrer; his department has received educational or research grants from Boston Scientific, Covidien Mallinckrodt Inc, Medtronic, Otsuka Pharma, Pfizer, Purdue Pharma, and St Jude Medical, Allergan, Endo, and Prostrakan; and he has received royalties from Oxford University Press and UpToDate. This study was sponsored by Cephalon, Inc, now a wholly owned subsidiary of Teva Pharmaceuticals. Writing support was provided by Bina J. Patel, PharmD, CMPP, of Peloton Advantage, LLC, funded by Teva Pharmaceuticals.
Acknowledgements The authors thank Susan Larijani for her help with the operational aspects of the management of the study, and DataStat, Inc, for their survey data collection services. The authors also thank the participants who completed the telephone surveys.
Appendix A. Supplemental Digital Content Supplemental Digital Content associated with this article can be found online at http://links.lww.com/PAIN/A15 and http://links. lww.com/PAIN/A16. Article history: Received 18 April 2014 Received in revised form 24 September 2014 Accepted 16 October 2014
REFERENCES [1] Arnold LM, Clauw DJ, Wohlreich MM, Wang F, Ahl J, Gaynor PJ, Chappell AS. Efficacy of duloxetine in patients with fibromyalgia: pooled analysis of 4 placebo-controlled clinical trials. Prim Care Companion J Clin Psychiatry 2009;11:237–44. Available at: http://www.ncbi.nlm.nih. gov/pmc/articles/PMC2781036/pdf/pcc11237.pdf. Accessed: Dec. 10, 2014. [2] Bennett SJ, Oldridge NB, Eckert GJ, Embree JL, Browning S, Hou N, Chui M, Deer M, Murray MD. Comparison of quality of life measures in heart failure. Nurs Res 2003;52:207–16. [3] Caraceni A, Bertetto O, Labianca R, Maltoni M, Mercadante S, Varrassi G, Zaninetta G, Zucco F, Bagnasco M, Lanata L, De Conno F. Episodic (breakthrough) pain prevalence in a population of cancer pain patients. Comparison of clinical diagnoses with the QUDEI—Italian questionnaire for intense episodic pain. J Pain Symptom Manage 2012;43:833–41. Available at: http://download.journals.elsevierhealth.com/pdfs/journals/ 0885-3924/PIIS0885392412000139.pdf. Accessed: Dec. 10, 2014.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
February 2015
·
Volume 156
·
Number 2
[4] Caraceni A, Martini C, Zecca E, Portenoy RK, Ashby MA, Hawson G, Jackson KA, Lickiss N, Muirden N, Pisasale M, Moulin D, Schulz VN, Rico Pazo MA, Serrano JA, Andersen H, Henriksen HT, Mejholm I, Sjogren P, Heiskanen T, Kalso E, Pere P, Poyhia R, Vuorinen E, Tigerstedt I, Ruismaki P, Bertolino M, Larue F, Ranchere JY, Hege-Scheuing G, Bowdler I, Helbing F, Kostner E, Radbruch L, Kastrinaki K, Shah S, Vijayaram S, Sharma KS, Devi PS, Jain PN, Ramamani PV, Beny A, Brunelli C, Maltoni M, Mercadante S, Plancarte R, Schug S, Engstrand P, Ovalle AF, Wang X, Alves MF, Abrunhosa MR, Sun WZ, Zhang L, Gazizov A, Vaisman M, Rudoy S, Gomez Sancho M, Vila P, Trelis J, Chaudakshetrin P, Koh ML, Van Dongen RT, Vielvoye-Kerkmeer A, Boswell MV, Elliott T, Hargus E, Lutz L. Breakthrough pain characteristics and syndromes in patients with cancer pain. An international survey. Palliat Med 2004;18:177–83. [5] Chou R, Fanciullo GJ, Fine PG, Adler JA, Ballantyne JC, Davies P, Donovan MI, Fishbain DA, Foley KM, Fudin J, Gilson AM, Kelter A, Mauskop A, O’Connor PG, Passik SD, Pasternak GW, Portenoy RK, Rich BA, Roberts RG, Todd KH, Miaskowski C; for the American Pain Society–American Academy of Pain Medicine Opioids Guidelines Panel. Clinical guidelines for the use of chronic opioid therapy in chronic noncancer pain. J Pain 2009;10:113–30. Available at: http://download.journals.elsevierhealth.com/pdfs/journals/1526-5900/ PIIS1526590008008316.pdf. Accessed: Dec. 10, 2014. [6] Cleeland CS, Nakamura Y, Mendoza TR, Edwards KR, Douglas J, Serlin RC. Dimensions of the impact of cancer pain in a four country sample: new information from multidimensional scaling. PAIN 1996;67:267–73. [7] Cleeland CS, Ryan KM. Pain assessment: global use of the Brief Pain Inventory. Ann Acad Med Singapore 1994;23:129–38. [8] Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–19. [9] Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, Haythornthwaite JA, Jensen MP, Kerns RD, Ader DN, Brandenburg N, Burke LB, Cella D, Chandler J, Cowan P, Dimitrova R, Dionne R, Hertz S, Jadad AR, Katz NP, Kehlet H, Kramer LD, Manning DC, McCormick C, McDermott MP, McQuay HJ, Patel S, Porter L, Quessy S, Rappaport BA, Rauschkolb C, Revicki DA, Rothman M, Schmader KE, Stacey BR, Stauffer JW, von Stein T, White RE, Witter J, Zavisic S. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain 2008;9:105–21. [10] Galvez R, Marsal C, Vidal J, Ruiz M, Rejas J. Cross-sectional evaluation of patient functioning and health-related quality of life in patients with neuropathic pain under standard care conditions. Eur J Pain 2007;11:244–55. [11] Hamilton RA, Briceland LL. Use of prescription-refill records to assess patient compliance. Am J Hosp Pharm 1992;49:1691–6. [12] Hess LM, Raebel MA, Conner DA, Malone DC. Measurement of adherence in pharmacy administrative databases: a proposal for standard definitions and preferred measures. Ann Pharmacother 2006;40:1280–8. [13] Keller S, Bann CM, Dodd SL, Schein J, Mendoza TR, Cleeland CS. Validity of the brief pain inventory for use in documenting the outcomes of patients with noncancer pain. Clin J Pain 2004;20:309–18. [14] Kessler RC, Ames M, Hymel PA, Loeppke R, McKenas DK, Richling DE, Stang PE, Ustun TB. Using the World Health Organization Health and Work Performance Questionnaire (HPQ) to evaluate the indirect workplace costs of illness. J Occup Environ Med 2004;46(6 suppl):S23–37. [15] Kessler RC, Barber C, Beck A, Berglund P, Cleary PD, McKenas D, Pronk N, Simon G, Stang P, Ustun TB, Wang P. The World Health Organization
www.painjournalonline.com
[16] [17]
[18]
[19]
[20] [21]
[22]
[23]
[24]
[25]
[26] [27] [28] [29]
[30]
[31]
[32]
259
Health and Work Performance Questionnaire (HPQ). J Occup Environ Med 2003;45:156–74. Kroenke K, Spitzer RL, Williams JB. The Patient Health Questionnaire-2: validity of a two-item depression screener. Med Care 2003;41:1284–92. Leon AC, Olfson M, Portera L, Farber L, Sheehan DV. Assessing psychiatric impairment in primary care with the Sheehan Disability Scale. Int J Psychiatry Med 1997;27:93–105. Lowe B, Decker O, Muller S, Brahler E, Schellberg D, Herzog W, Herzberg PY. Validation and standardization of the Generalized Anxiety Disorder Screener (GAD-7) in the general population. Med Care 2008;46:266–74. Manchikanti L, Singh V, Caraway DL, Benyamin RM. Breakthrough pain in chronic non-cancer pain: fact, fiction, or abuse. Pain Physician 2011;14: E103–17. Available at: http://www.painphysicianjournal.com/linkout_vw.php? issn51533-3159&vol514&page5E103. Accessed: Dec. 10, 2014. Markman JD. Not so fast: the reformulation of fentanyl and breakthrough chronic non-cancer pain. PAIN 2008;136:227–9. Narayana A, Katz N, Shillington AC, Stephenson JJ, Portenoy RK. The National Breakthrough Pain Survey: mulitvariate predictors of functional status, pain interference, and disability [poster 71]. Presented at PAINWeek 2012; September 5-8, 2012; Las Vegas, NV. Perez C, Navarro A, Saldana MT, Martinez S, Rejas J. Patient-reported outcomes in subjects with painful trigeminal neuralgia receiving pregabalin: evidence from medical practice in primary care settings. Cephalalgia 2009;29:781–90. Portenoy RK, Bennett DS, Rauck R, Simon S, Taylor D, Brennan M, Shoemaker S. Prevalence and characteristics of breakthrough pain in opioid-treated patients with chronic noncancer pain. J Pain 2006;7: 583–91. Portenoy RK, Bruns D, Shoemaker B, Shoemaker SA. Breakthrough pain in community-dwelling patients with cancer pain and noncancer pain, part 1: prevalence and characteristics. J Opioid Manag 2010;6:97–108. Portenoy RK, Bruns D, Shoemaker B, Shoemaker SA. Breakthrough pain in community-dwelling patients with cancer pain and noncancer pain, part 2: impact on function, mood, and quality of life. J Opioid Manag 2010;6:109–16. Portenoy RK, Hagen NA. Breakthrough pain: definition, prevalence and characteristics. PAIN 1990;41:273–81. Portenoy RK, Payne D, Jacobsen P. Breakthrough pain: characteristics and impact in patients with cancer pain. PAIN 1999;81:129–134. Sheehan DV, Harnett-Sheehan K, Raj BA. The measurement of disability. Int Clin Psychopharmacol 1996;11(suppl 3):89–95. US Census Bureau. State and Country QuickFacts. US Department of Commerce. Available at: http://quickfacts.census.gov/qfd/states/ 00000.html. Accessed April 16, 2014. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care 1996;34:220–33. Zeppetella G, O’Doherty CA, Collins S. Prevalence and characteristics of breakthrough pain in cancer patients admitted to a hospice. J Pain Symptom Manage 2000;20:87–92. Available at: http://download.journals. elsevierhealth.com/pdfs/journals/0885-3924/PIIS0885392400001615. pdf. Accessed: Dec. 10, 2014. Zeppetella G, O’Doherty CA, Collins S. Prevalence and characteristics of breakthrough pain in patients with non-malignant terminal disease admitted to a hospice. Palliat Med 2001;15:243–6.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
