Disparities in Pain Medicine: A Psychological Perspective

1
Disparities in Pain MedicineA Psychological
Perspective

• Raymond C. Tait, PhD
• Saint Louis University

2
Disclosures

• Spouse is on the Speakers Bureau for Lilly
• Center for World Health Medicine (direct
  report) has a project jointly funded by Lilly and
  JJ
• No discussion of unapproved uses

3
The study of error is not only in the highest
degree prophylactic, but it serves as a
stimulating introduction to the study of truth.

• --Walter Lippmann (1922)

4
Organization of Comments

• Review general evidence on disparities in pain
  care
• Primary focus race/ethnicity
• Race/ethnicity and SES
• The clinical encounter
• Pain as subjective phenomenon
• Variability in provider assessments
• Provider as a social judge
• Patient influences (reported pain severity,
  race/ethnicity)
• Is high pain severity a condition for
  context-mediated judgments?
• Situational (medical evidence)
• Provider (specialty)
• Imputed response cost a common pathway?

5
Disparities in Pain Care

• Race/ethnicity
• Todd et al., 1993 Green et al., 2003 IOM, 2003
  Chibnall et al., 2005 Anderson et al., 2009
  IOM, 2003, 2011 Meghani et al., 2012
• Gender
• Martin Lemos, 2002 Taylor et al., 2005
• Age
• Old (Hadjistravropoulos et al., 2007 Weiner et
  al, 2002)
• Young (Howard, 2003 Anthony Schanberg, 2005)
• Socioeconomic status
• Morrison et al., 2000 Mayberry et al., 2000
• Conditions
• Low back pain (Tait et al., 2004)
• Acute pain (Salmon Manyande, 1996)
• Recurrent pain (Elander et al., 2006)
• Cancer pain (Cleeland et al., 1997)

6
Disparities in Pain Care

• Race/ethnicity
• Todd et al., 1993 Green et al., 2003 IOM, 2003
  Chibnall et al., 2005 Anderson et al., 2009
  IOM, 2011 Meghani et al., 2012
• Gender
• Martin Lemos, 2002 Taylor et al., 2005
• Age
• Old (Hadjistravropoulos et al., 2007 Weiner et
  al, 2002)
• Young (Howard, 2003 Anthony Schanberg, 2005)
• Socioeconomic status
• Morrison et al., 2000 Mayberry et al., 2000
• Conditions
• Low back pain (Tait et al., 2004)
• Acute pain (Salmon Manyande, 1996)
• Recurrent pain (Elander et al., 2006)
• Cancer pain (Cleeland et al., 1997)

7
Patient Factors in Racial Disparities
Experimental Studies of Pain Perception

• African Americans demonstrate lower pain
  thresholds and tolerances than nHws
• Zborowski, 1969 Zatzick Dimsdale, 1990
• Negative affect may mediate findings
• Sheffield et al., 2000 Campbell et al., 2008
• Higher vigilance may mediate findings
• Campbell et al., 2005
• Cautionary notes
• Experimental data may not predict response to
  clinical pain (Edwards et al., 2001)
• Experimenter-subject effects (e.g., ? pain
  behavior in ethnically concordant pairs Hsieh et
  al., 2011)

8
Patient Factors in Racial Disparities Clinical
Studies of Pain Perception

• Acute pain
• ? AA relative to nHw (Faucett et al., 1994)
• No diffs (Barak Weisenberg, 1988)
• Chronic pain
• ? AA relative to nHw (McCracken et al., 2001
  Selim et al., 2001 White et al., 1999)
• No diffs (Thomason et al., 1998 Jordan et al.,
  1998 Tait Chibnall, 2001)
• Methodologic challenges
• Control over prior treatment, sampling (most
  patients come from single clinic), social
  reactivity

9
Patient Factors in Racial Disparities
Pain-Related Coping and Adjustment

• ? post-treatment disability for AA vs nHw
• Chronic LBP (Selim et al., 2001 Chibnall et al.,
  2005), osteoarthritis (Allen et al., 2010),
  general chronic pain (Edwards et al., 2001)
• ? distress for AA vs nHw
• Allen et al., 2010 McCracken et al., 2001
• Coping
• ? passive coping (Jordan et al., 1998 Clark et
  al., 1999)
• ? catastrophizing (Fabian et al., 2011)
• Distrust of medical professionals
• Lillie-Blanton et al., 2000
• Expectations of benefit
• LaVeist et al., 2000 Ibrahim et al., 2002

10
Racial Disparities Situational/Public Health
Factors

• Racism
• Long-term implications for health (Clark et al.,
  1999)
• Socioeconomic factors co-vary with minority
  status (Mayberry et al., 2000 Meghani et al.,
  2012)
• Access
• Analgesics (Morrison et al., 2000 Green et al.,
  )
• Medical care (Meghani et al., 2012)
• Insurance (Zuvekas and Taliaferro, 2003)
• Resources (Tait Chibnall, 2012)

11
Relative Contributions ofRace and Socioeconomic
Status
12
Disparities in Occupational Lumbar Injury
Outcomes Research (DOLOR)(Agency for Healthcare
Research and Quality, R01 HS13087-01)

• Missouri cases of LB injuries that were settled
  between 1/01 and 6/02
• St. Louis city, St. Louis county, Jackson county
• 90 of African Americans in the state
• 2,934 cases
• 50.3 completed survey
• 14.7 refused survey
• 35.0 could not be traced
• Data sources
• WC database
• Telephone survey instruments

13
Demographics(N 1,475)

• Age ? 43.6 years
• Education 13.07 years
• Gender 896 males, 533 females
• Race (self-identified) 889 Caucasian, 540
  African Americans, 43 mixed, 3 refused
• Working full-time at time of injury 95.2
• Working full-time now 62.8

14
WC Database Surgical vs. Non-Surgical
Treatment(Chibnall et al., Spine, 2006)
Surgery No Surgery Yes Total
Caucasian N 624 70 w/in race N 268 30.0 w/in race N 892 100.0
African American N 536 92.4 w/in race N 44 7.6 w/in race N 580 100.0
Total N 1160 78.8 w/in race N 312 21.2 w/in race N 1472 100.0
?2(1) 106.1, P lt 0.0001 OR 4.0 95 CI 2.9
5.4
15
WC Management Database by Race
Characteristic Caucasian African-American
Medical costs mean (SD) 11,354 (17,755) 3,778 (6,752)
Settlement award mean (SD) 15,328 (13,890) 7,795 (8,927)
Disability rating mean (SD) 12.9 (10.7) 7.5 (6.4)
Claim duration, months mean (SD) 23.2 (12.9) 18.2 (10.6)
16
Predictors of WC ManagementDemographics, SES,
and Injury(Tait et al., Pain, 2004)
Predictors Treatment Costs Disability Rating Settlement Award Claim Duration
Demographics (race, gender, age) 0.10 0.08 0.14 0.10
SES (comp rate, education, income) 0.04 0.02 0.10 0.01
Injury (disc, legal rep, sprain/strain) 0.11 0.10 0.15 0.05
Model 0.25 0.20 0.39 0.16
Simultaneous entry multiple hierarchical
regression R2 change (all Ps lt 0.001)
17
Clinical Outcomes 2 Years Post-Settlement (Chibn
all et al., Pain, 2005)
African American Caucasian
Mean / SD Mean / SD
Usual Pain 6.4 / 2.0 5.0 / 2.0
Catastrophizing 28.3 / 12.8 22.7 / 12.3
PDI 35.7 / 18.1 28.0 / 17.4
P lt 0.0001
18
6-Year Follow-Up High Levels of Pain,
Catastrophizing, and Disability(Chibnall Tait,
Pain Medicine, 2011)
Criterion Variable Long-term Follow-up Significant Predictors (P lt 0.05) OR (95 CI)
Pain intensity Race (0 Caucasian, 1 African American) 2.6 (1.5-4.3)
Pain intensity Pain (high vs. lt high) at baseline 6.2 (3.6-10.6)
Catastrophizing (PCS) Race (0 Caucasian, 1 African American) 1.9 (1.2-3.1)
Catastrophizing (PCS) SES 0.48 (0.34-0.67)
Catastrophizing (PCS) Catastrophizing (high vs. lt high) at baseline 3.6 (2.2-6.1)
Disability (PDI) SES 0.62 (0.43-0.83)
Disability (PDI) Disability (high vs. lt high) at baseline 7.2 (4.1-12.7)
1 high (pain 7 PCS 30 PDI 45) vs. 0
less than high
19
Path Analysis for Predicting Adjustment
(pain/distress/disability PDD)
Circled values indicate Multiple R at that stage
of the model, P lt 0.001
20
Race Effects on Financial Court Actions 5 Years
Post-Settlement(Tait Chibnall, Spine, 2012)
21
Implications for Race SES

• Race/ethnicity and SES are associated with
  differences in patient/provider approach to
  treatment and intermediate-term outcomes
• Race/ethnicity appears to account for greater
  effect during active clinical management
• Race/ethnicity and SES are associated with
  differences in long-term outcomes
• SES accounts for greater long-term effects
• What accounts for disparate clinical management?

22
Judging Pain in Others A Social Interaction
ENVIRONMENT ENVIRONMENT
P A T I E N T P R O V I D E R
23
(No Transcript)
24
Judging Pain in Others A Projective Test?
One Patient
Two Providers
Opinion 1
Opinion 2
25
Internist Judgments of Chronic Low Back
Pain(Chibnall, Dabney Tait, Pain Medicine,
2000)

• 48 internists from an academic school of medicine
• 2 x 4 mixed between and within-subjects design
• Vignettes describing hypothetical low back pain
  patients varied by pain severity (low vs. high)
• Internists provided 4 waves of clinical
  information (history ? physical exam findings ?
  functional disability ? diagnostic test results)
• Measures MD judgments regarding patient
  medical/psychological/disability status,
  treatment, diagnostic testing, and referral
  options

26
MD Judgments Reliability Across 4 Waves of
Information
Outcome Variables Intra-Class Correlation Intra-Class Correlation
Outcome Variables Within MD Between MD
Vertebral or diskal lesion 0.72 0.07
Soft tissue, musculoskeletal 0.71 0.03
Personality factors 0.75 0.06
Orthopedic surgery referral 0.87 0.04
Psychiatry referral 0.49 0.04
Physical therapy referral 0.91 0.04
Prescribe opioids 0.88 0.05
Order MRI 0.69 0.12
Occupational disability level 0.63 0.11
27
Patient Characteristics (The
Target)
28
Pain Presentation Factors that Influence
Judgments

• Chronicity
• Klein et al., 1982 Teske et al., 1983 Taylor et
  al., 1984 Leclere et al., 1990 Eccleston et
  al., 1997 Hahn, 2001
• Distribution
• Ransford et al., 1976 Von Baeyer et al., 1983
  Margolis et al., 1986 Tait et al., 1990
• Behavior
• Prkachin et al., 1994 Krause et al., 1994
  Solomon et al., 1997 Prkachin et al., 2001
• Severity
• Grossman et al., 1991 Zalon, 1993 Chibnall and
  Tait, 1995 Lieberman et al., 1996 Solomon et
  al., 1997 Tait and Chibnall, 1997 Marquie et
  al., 2003

29
Subject vs Patient Pain RatingsChibnall, Tait
Ross, J Behav Med, 1997
30
RATES OF AGREEMENT IN PATIENT CAREGIVER PAIN
RATINGS(from Grossman et al., Correlation of
patient and caregiver ratings of cancer pain, J.
Pain Symp Manag, 1991 653-57)
31
Observer Perceptions of Low Back Pain
Effects of Pain Report and Other Contextual
Factors(Chibnall Tait, 1995)

• 2 hi/lo pain x (2 medical evidence x 2
  relationship valence x 2 victim status)
• Dependent Variables
• estimated pain, disability, emotional distress
• personality characteristics
• 80 undergraduates
• Context influences evident in main effects,
  2-way, 3-way, and 4-way interactions

32
Study Design 2 x (2 x 2 x 2) Mixed Between and
Within(Chibnall Tait, 1995)
Med Evidence Med Evidence Med Evidence Med Evidence No Med Evidence No Med Evidence No Med Evidence No Med Evidence
Control Control No Control No Control Control Control No Control No Control
- - - -
S1 S80 ? ? ? ? ? ? ? ? ? ? ? ? ? ?
Note 8 vignettes counterbalanced for order and
sequence
Pain High (7/10) vs. Low (3/10) Medical
Evidence Present vs Absent Control Present vs
Absent Relationship Positive vs Negative
33
Example Vignettes and Levels of Independent
Variables
Vignette IV Levels
A Close personal friend of many years ... Waiting at a red light, struck from behind CT scan and myelogram show bulging disc, spinal instability -grade II spondylolisthesis B. Submitted claim you believe is fraudulent Lifted heavy box at work despite warning... MRI study and clinical exams reveal no clear pathology. Positive No control Evidence Negative Control No evidence
34
Reported Pain Severity Medical Evidence
Effects on Pain Estimates(Chibnall Tait, 1995)
IV IV Simple Effects Simple Effects T-tests T-tests T-tests
Pain Med Evidence M (SD) F (1,78) M T(39) eta2
High Yes 6.7 (1.3) 72.07 7 -1.4 .05
High No 4.7 (2.2) 72.07 7 -6.5 .52
Low Yes 3.4 (1.6) 14.67 3 1.6 .06
Low No 2.4 (1.8) 14.67 3 -2.1 .10
P lt 0.05 P lt 0.001
35
High Pain Severity Implications for Clinical
Judgment
Pain Report
Low (1-3)
Moderate (4-6)
High (7-10)
Little likelihood of context effects
Some likelihood of context effects without
objective evidence
High likelihood of context effects with/without
objective evidence
36
Sociodemographic Factors that Influence Judgments

• Gender
• Martin Lemos, 2002 Taylor et al., 2005
• Age
• Old (Hadjistravropoulos et al., 2007 Weiner et
  al, 2002)
• Young (Howard, 2003 Anthony Schanberg, 2005)
• Socioeconomic status
• Morrison et al., 2000 Mayberry et al., 2000
• Race/ethnicity
• Todd et al., 1993 Green et al., 2003 IOM, 2003
  Chibnall et al., 2005 Anderson et al., 2009
  IOM, 2011 Meghani et al., 2012

37
Claimants with HNP Predictors of Surgery
Variables Odds Ratio OR 95 CI P
Race 0.32 0.21-0.49 lt.001
Radicular pain 2.86 1.21-6.74 lt.05
Legal rep (2o to Rx dissatisfaction) 0.57 0.39-0.84 lt.01
Legal representation 1.42 0.87-2.33 .16
SES 1.14 0.96-1.36 .13
Gender 0.91 0.61-1.35 .63
Age 1.05 0.81-1.12 .57
Lumbar degeneration 1.01 0.71-1.42 .98

• No surgery vs surgery ?2(8) 59.6, P lt .001
  R2 0.13 (N 640)

38
Contextual Characteristics(The Situation)
39
Factors that Influence Judgments Situational
Features

• Compensation status
• Hadler, 1994 Kennedy, 1997 Chibnall and Tait,
  1999 Merskey and Teasell, 2000 Kappesser et
  al., 2006
• Medical evidence
• Carey et al., 1988 Birdwell et al., 1993 Tait
  and Chibnall, 1994 Chibnall and Tait, 1995
  Chibnall et al., 1997 Tait et al., 2006

40
Study Design 2 x 2 x 2 Within-Subjects(Tait
Chibnall, 1994)
Med Evidence Med Evidence Med Evidence Med Evidence No Med Evidence No Med Evidence No Med Evidence No Med Evidence
Control Control No Control No Control Control Control No Control No Control
- - - -
S1 S48 ? ? ? ? ? ? ? ? ? ? ? ? ? ?
Note 8 vignettes counterbalanced for order and
sequence
Medical Evidence Present vs Absent Control
Present vs Absent Relationship Positive vs
Negative
41
Main Effects Medical Evidence
Variable Medical Evidence M SD eta2
Pain Severity Present 6.3 1.6 0.63
Pain Severity Absent 4.7 2.0
Disability Present 32.0 11.8 0.63
Disability Absent 22.3 12.4
Negative Character Present 22.3 7.9 0.39
Negative Character Absent 26.9 9.1
P lt 0.05 P lt 0.001
42
Incremental Certainty of Disability Low Back
Pain(Carey et al., J Clin Epidemiol
198841691-697)
43
Provider Characteristics (The Judge)
44
Factors that Influence JudgmentsObserver
Features

• Affect
• Tait Chibnall, 1994 Sharpe et al., 1994
  Chibnall Tait, 1995
• Empathy
• Goubert et al., 2005 Tait et al., 2005 Tait,
  2008
• Experience/Specialty
• Lenburg et al., 1970 Choiniere et al., 1990
  Chibnall Tait, 2000 Prkachin et al., 2001
  Marquie et al., 2003 Tait et al., 2010

45
Judging Pain Physician Specialty(Tait et al.,
2010)
Variable Neurosurgeons Mean (SD) Internists Mean (SD)
Pain severity (0-10) 4.46 (1.56) 5.71 (1.54)
Home disability (0-10) 3.83 (1.77) 4.61 (1.70)
Social disability (0-10) 3.69 (1.97) 4.71 (2.01)
Work disability (0-10) 4.40 (2.12) 5.37 (1.62)
Occupational disability (0-100) 27.6 (19.2) 45.1 (22.8)
P lt 0.05 P lt 0.001
46
Pain Management A Social Transaction

• Pain management is an interactive phenomenon
• Social transaction (Craig et al., 2010)
• Participative decision-making (Frantsve Kerns,
  2006)
• Social contract (Kappesser et al., 2008)
• Contract influenced by assumed relational roles
  (patient and provider)
• Implied contractual demands of patients with
  severe, chronic pain
• Fix me
• At least help meanalgesic medications (opioids?)
• Handle any regulatory implications
• Assume long-term management (not cure)
• Embrace high (ongoing?) time demands
• Recognize the likelihood of associated
  psychological distress
• Tackle disability-related sequelae
• Manage sick role, litigation and other system
  issues
• Treat likely co-morbidities
• Prepare for high costs of care

47
Imputed Response Cost A Common Pathway?

• Definition Response cost or negative punishment
  is a way to make behavior less frequent
• Increasing the cost of a response decreases the
  likelihood that it will occur
• Imputed response cost The expectation that a
  given action will result in costly and/or
  burdensome consequences
• Treating patients with chronic pain occasions
  high imputed costs secondary to social contract
  responsibilities
• Especially with high pain severity
• How to reduce response cost
• Decline (or share) treatment responsibility
• Discount severity (reduce responsibility)

48
Imputed Response Cost Effects on Pain Judgment

• Severe chronic pain presentation
• 43 yo WM w/ severe pain x 12 mos (prior L4-5
  discectomy)
• Grade III spondylolisthesis at L5-S1, DJD at
  L4-5, SLR on left, equivocal EMGnot deemed a
  surgery candidate
• Comorbidities HTN, ? Psych distress, frequent
  work absences
• Oxycodone 30mg bid, ibuprofen 800mg tid, vicodin
  prn
• Personal cost (hi/lo)
• Ongoing treatment responsibility vs evaluation
  only
• Societal cost (hi/lo)
• Evaluation for disability determination vs FYI
  only
• Dependent variables
• Psychological attribution, pain-related
  dysfunction

49
Pain Dysfunction Expected Effects of Imputed
Personal/Societal Costs
Mean T ScorePain Dysfunction
Personal Cost
50
Pain Dysfunction Actual Effects of Imputed
Personal/Societal Costs
Mean T ScorePain Dysfunction
Personal Cost
51
Psychosocial Overlay Expected Effects of
Personal and Societal Costs
Mean T ScorePsychosocial Overlay
Personal Cost
52
Psychosocial Overlay Effects of Personal and
Societal Costs
Mean T ScorePsychosocial Overlay
Personal Cost
53
Conclusions I Pain management as Social Judgment

• Judgments regarding pain assessment and treatment
  fit a (complex) social cognition model
• Social cognitive influences are greatest when
  pain is severe and supporting medical evidence is
  lacking
• Patient factors (e.g., race/ethnicity)
  situational context likely to influence
  anticipated response cost
• High anticipated response cost likely to
  contribute to symptom discounting and disparities
  in treatment

54
Conclusions II Strategies to Reduce Anticipated
Response Cost

• Treatment guidelines/education may minimize
  errors in judgment
• Evidence-based medicine?
• Multidisciplinary approaches
• Maximize sources of input
• Distribute treatment burden
• Lessen empathy influences
• Multiple sources of information to reduce bias
• Pain severity, pain distribution, pain behavior
• Levels of pain-related disability
• Psychological distress as modifying, not
  invalidating pain experience

55
Conclusions III Research Questions

• Is anticipated response cost a common pathway?
• Quantify perceived burden of patient features
• Quantify perceived burden of context features
• Assess impact of anticipated response cost on
  patient interaction and shared decision-making
  model
• Measure links between elements of social
  cognition and actual treatment decisions
• So what?
• Can attention to social cognition improve quality
  and/or costs of care?