EE

Epidemiology of High-Impact Chronic Pain – Dr. Mark Pitcher

Introduction to the Session

  • Speaker: Dr. Mark Pitcher, PhD
    • Epidemiologist & Director of Health Sciences Interprofessional Research, University of Bridgeport
    • Background: Neuroscience training at McGill; epidemiology work at NIH (National Center for Complementary & Integrative Health)
    • Expertise: Integrative pain management, chronic pain epidemiology
  • Context: Part of a lecture series; audience encouraged to interrupt with questions via chat or voice

Fundamentals of Epidemiology

  • CDC definition: “Scientific, systematic, data-driven study of the distribution (frequency & pattern) and determinants (causes & risk factors) of health-related states or events in populations, and the application of this study to control health problems.”
  • Two archetypes of epidemiologists
    • Field (e.g., investigating COVID-19 outbreak in Wuhan)
    • Desk / Data (Dr. Pitcher: large-dataset analyst)
  • Key term — Prevalence
    \text{Prevalence} = \frac{\text{# people with condition at a specific time}}{\text{Total population at same time}}

Historical Estimates & the “100 Million” Debate

  • 2008 estimate: ≈ 100 million U.S. adults (~40 %) reported “chronic pain”
  • Critique: Term was ill-defined; conflated mild, intermittent, & disabling pain
  • Federal response: HHS convened CDC, AHRQ, FDA, NIH, VA, DoD → National Pain Strategy (NPS)

National Pain Strategy – Operational Definitions

  • Chronic Pain (CP): Pain on ≥ ½ the days for ≥ 6 months
  • High-Impact Chronic Pain (HICP): CP plus “substantial restriction” in work, social, or self-care activities for ≥ 6 months
  • Gap: No prevalence data for HICP when definition proposed

Initial HICP Study (Pitcher et al., 2019)

Data Source & Methodology

  • Data: 2011 National Health Interview Survey (NHIS)
    • 120 000–150 000 respondents; complex sampling (oversamples e.g., Native Americans)
    • Publicly downloadable; large files require statistical software know-how
  • Operationalization
    1. CP → Answered “most days” or “every day” to: “During the last 3 months how often did you have pain?”
    2. Activity-limitation items (yes/no): inability to work, attend school, perform household chores, participate socially, etc.
    3. HICP = CP + ≥ 1 activity limitation
    4. CP-w/o-Limitations = CP but no limitations
    • Note: Conservative—no gradation of “partial” limitation

Core Findings (2011)

  • U.S. adult CP prevalence: 18.4 % (~ 40 M)
  • CP-w/o-Limitation: 13.6 % (~ 30 M)
  • HICP: 4.8 % (~ 10 M) (“floor” estimate)

Demographic Patterns in HICP

  • Sex: ~ 55–60 % female
  • Age: prevalence climbs sharply ≥ 65 yrs
  • BMI: higher BMI → higher HICP
  • Race/Ethnicity:
    • More Whites numerically (reflects population)
    • Higher probability in Blacks & Native Americans
  • Education: Lower education ↑ prevalence
  • Marital Status: Higher in divorced / separated

Psychosocial & Health Burden

  • Pain descriptors (vs CP-w/o-Limitation & general population)
    • Greater intensity (“a lot”), all-day duration, daily occurrence
  • Mental health
    • Higher daily depression & anxiety frequency/intensity
    • ↑ medication use for mood disorders
  • General health
    • > 2× more likely to say health worse than 12 mo ago
    • > 11 bed-days/yr; need help with ADLs & self-care
  • Healthcare utilization
    • More specialist visits, ≥ 10 office visits/yr, ≥ 1 surgery/yr
    • Socio-economic paradox: greatest need, least resources/insurance
  • Caveat: NHIS surgery question covers any surgery ➔ interpret cautiously (wisdom teeth to spine fusion)

Activity-Limitation Odds Ratios

  • Compared to “no / occasional pain”
    \text{OR}{\text{Demog only}} = 5.88 \text{OR}{\text{Demog + chronic conditions}} = 4.19
  • CP outranks kidney failure, stroke, emphysema, cancer for predicting functional limitation

Section Summary

  • HICP = severe CP subtype; ≥ 10 M adults (likely closer to 20 M)
  • Limitations due primarily to pain, not comorbidities
  • Disproportionate mental-health burden & healthcare costs
  • Non-pharmacologic & integrative therapies critical amid opioid crisis

Subsequent Literature & Updated Numbers

  • Kaiser Permanente study (Von Korff et al., 2019)
    • 14 % of clinic cohort had HICP; 50 % of high service users had HICP
  • 2016 NHIS (new HICP-specific questions)
    • CP = 20.4 %
    • HICP = 8.0 % (~ 19.6 M)
    • Highest among adults in poverty, < HS education, public insurance
  • Elderly (Health & Retirement Study)
    • ≥ 50 yrs: HICP = 8 %; lowest wealth quartile → 17 +
  • Chronic spinal pain (NHIS 2019)
    • CP-spine = 6 %; HICP-spine = 2.2 %
    • HICP-spine users ↑ opioid use (dose & prevalence)

From Low-Impact to High-Impact: What Turns the Dial?

  • Possible drivers
    • Pain location, intensity, chronicity
    • Intrinsic traits: resilience, grit
    • Adverse childhood experiences
    • Socioeconomic constraints
    • Genetic / physiological differences
    • Occupation-related biomechanical stress

Pilot Data – University of Bridgeport (UB) Clinics

Clinic Context

  • Community teaching clinic; ≈ 18 000 visits/yr
  • Low-SES, uninsured / under-insured patients; services: chiropractic, acupuncture, naturopathy, dental hygiene

Sample & Instruments

  • n = 99; blood, diet, ICD-10, & NPS pain survey
  • Pain survey yields:
    • Frequency categories (most days, every day, every day all day)
    • Pain Severity Score (PSS) out of 30 (sum of intensity 0-10 + interference with ADL 0-10 + interference with enjoyment 0-10)

Key Observations

  • 66 % met CP criteria
  • Sex: Females report higher PSS within CP group
  • Age vs PSS: No clear trend (sample under-powered)
  • Increasing frequency → visually higher severity, but small n ⇒ non-significant
  • Inter-item correlations (CP subgroup)
    • Intensity ↔ Enjoyment r=0.347 (modest)
    • Intensity ↔ Activities r=0.347
    • Activities ↔ Enjoyment r=0.773 (strong) → essentially same construct
  • Individual-level heterogeneity: similar frequency/intensity yet disparate functional impact—suggests multifactorial modifiers

Machine-Learning Exploration (NHIS 2017)

  • 95 candidate variables ➔ XGBoost classifier
  • Top predictors of HICP vs CP-w/o-Limitation
    1. # Bed days past yr
    2. BMI
    3. Age
    4. Pain frequency
    5. Moderate-exercise frequency
    6. Sleep hours
    7. “Effort” (self-reported difficulty)
    8. Alcohol drinks/week
    9. Marital status
  1. Feelings of restlessness
  • “Alternative” modalities (guided imagery, Tai Chi, progressive relaxation) not differentiating—used across pain severities
  • ROC AUC ≈ 0.75 → model substantially better than chance but room for refinement

Socioeconomic Status (SES) & Occupation Hypothesis

  • Lower SES → less higher-ed access → higher likelihood of physically demanding jobs → ↑ injury risk → ↑ CP & HICP
  • NHIS occupation codes (Standard Occupational Classification)
    • Examples (prevalence of HICP):
    • Healthcare Support = 11.6 %
    • Construction & Extraction = 9-10 %
    • Computer & Mathematical = 3.3 %
    • Largest absolute # of HICP cases in Office & Administrative Support (prevalence moderate but occupation common)
  • Education & income gradients mirror occupation risk profiles

Gender-Related Considerations

  • U.S. adult CP prevalence: Female 56 % vs Male 44 %
  • Biological factors: sex hormones, autoimmune diseases
  • Social factors: workplace stress in male-dominated sectors, communication norms around reporting pain

Ethical, Philosophical & Practical Implications

  • HICP represents severe suffering + socioeconomic vulnerability
  • Integral to opioid-overdose narrative—most prescriptions concentrated in HICP
  • Equity issue: populations in poverty & physically demanding jobs shoulder disproportionate burden
  • Public-health strategy must integrate: prevention (workplace safety), early-intervention, non-pharma modalities, mental-health support

Methodological Caveats

  • NHIS = cross-sectional: cannot infer temporality or causation
  • Survey questions set by CDC → limited granularity (e.g., surgery reason)
  • Conservative definitions likely underestimate HICP prevalence
  • Small pilot datasets (e.g., UB study) under-powered; hypothesis-generating

Connections to Previous & Future Work

  • Builds on national calls (NPS) to refine pain definitions & metrics
  • Informs integrative pain-management research agendas (non-opioid, complementary therapies)
  • Necessitates longitudinal cohorts to identify predictors & causal pathways
  • COVID-19 era may shift occupation patterns, mental health, tele-work prevalence → future analyses warranted

Clinical & Research Take-Aways

  • Screening toolkit should include:
    1. Frequency (≥ ½ days × 6 mo)
    2. Intensity (0-10)
    3. Interference with work, social, self-care
    4. Mental-health check (depression, anxiety)
  • CP ≠ HICP: stratify patients; tailor interventions & resource allocation
  • Recognize occupation & SES context; advocate workplace modifications & policy reform
  • Integrative modalities (acupuncture, mindfulness, Tai Chi) widely adopted—need robust efficacy & implementation studies

Concluding Remarks from Dr. Pitcher

  • HICP affects ≥ 10–20 million U.S. adults and is more disabling than stroke or kidney failure in terms of daily function
  • Pain itself—not just comorbid disease—drives inability to work or care for self
  • Epidemiologic evidence underscores urgency for multi-disciplinary, non-opioid strategies
  • Future work: leverage machine learning & longitudinal datasets to unravel causal pathways and optimize intervention targeting