Advertisement

An Ecological Momentary Assessment Study of Pain Intensity Variability: Ascertaining Extent, Predictors, and Associations With Quality of Life, Interference and Health Care Utilization Among Individuals Living With Chronic Low Back Pain

  • M. Gabrielle Pagé
    Correspondence
    Address reprint requets to M Gabrielle Pagé, PhD, Clinical Scientist, Centre de recherche du Centre hospitalier de l'Université de Montréal, 850 rue St-Denis, office S01-122, Montreal, QC Canada H2X 0A9
    Affiliations
    Centre de recherche du Centre hospitalier de l'Université de Montréal & Department of Anesthesiology and Pain Medicine, Faculty of Medicine, Université de Montréal, Montreal, Quebec, Canada
    Search for articles by this author
  • Lise Gauvin
    Affiliations
    Centre de recherche du Centre hospitalier de l'Université de Montréal & Department of Social and Preventive Medicine, École de santé publique de l'Université de Montréal, Montreal, Quebec, Canada
    Search for articles by this author
  • Marie-Pierre Sylvestre
    Affiliations
    Centre de recherche du Centre hospitalier de l'Université de Montréal & Department of Social and Preventive Medicine, École de santé publique de l'Université de Montréal, Montreal, Quebec, Canada
    Search for articles by this author
  • Roy Nitulescu
    Affiliations
    Centre de recherche du Centre hospitalier de l'Université de Montréal & Centre d'intégration et d'analyse en données médicales du Centre hospitalier de l'Université de Montréal, Montreal, Quebec, Canada
    Search for articles by this author
  • Alina Dyachenko
    Affiliations
    Centre de recherche du Centre hospitalier de l'Université de Montréal & Centre d'intégration et d'analyse en données médicales du Centre hospitalier de l'Université de Montréal, Montreal, Quebec, Canada
    Search for articles by this author
  • Manon Choinière
    Affiliations
    Centre de recherche du Centre hospitalier de l'Université de Montréal & Department of Anesthesiology and Pain Medicine, Faculty of Medicine, Université de Montréal, Montreal, Quebec, Canada
    Search for articles by this author
Open AccessPublished:January 20, 2022DOI:https://doi.org/10.1016/j.jpain.2022.01.001

      Highlights

      • There is substantial variability in momentary reports of pain intensity.
      • It remains difficult to identify risk and protective factors of pain intensity variability.
      • Pain intensity variability does not seem associated with healthcare utilization.

      Abstract

      This ecological momentary assessment (EMA) study examined the extent of pain intensity variability among 140 individuals with chronic low back pain and explored predictors of such variability and psychosocial and health care utilization outcomes. Individuals completed momentary pain intensity reports (0–10 numeric rating scale) several times daily for two periods of seven consecutive days, one month apart. Participants also completed online questionnaires at baseline which tapped into pain characteristics, pain-related catastrophization, kinesiophobia, activity patterns, and depression and anxiety symptoms. Questionnaires assessing quality of life and health care utilization were administered online one month after completion of the last EMA report. Data were analyzed using linear hierarchical location-scale models. Results showed that pain intensity fluctuated over the course of a week as shown by an average standard deviation of 1.2. The extent of variability in pain intensity scores was heterogeneous across participants but stable over assessment periods. Patients’ baseline characteristics along with psychosocial and health care utilization outcomes were not significantly associated with pain intensity variability. We conclude that pain intensity variability differs across patients yet correlates remain elusive. There is an important gap in our knowledge of what affects this variability. Future EMA studies should replicate and extend current findings.

      Perspective

      This study provides evidence indicating that there is substantial variability in momentary reports of pain intensity among individuals living with chronic low back pain. However, risk and protective factors for greater lability of pain are elusive as is evidence that greater pain intensity variability results in differential health care utilization.

      Key words

      Introduction

      Chronic low back pain (CLBP) develops in an alarming proportion of patients.
      • Beneciuk JM
      • Robinson ME
      • George SZ.
      Low back pain subgroups using fear-avoidance model measures: Results of a cluster analysis.
      ,
      • Leeuw M
      • Goossens ME
      • Linton SJ
      • Crombez G
      • Boersma K
      • Vlaeyen JW.
      The fear-avoidance model of musculoskeletal pain: Current state of scientific evidence.
      ,
      • Wertli MM
      • Rasmussen-Barr E
      • Weiser S
      • Bachmann LM
      • Brunner F.
      The role of fear avoidance beliefs as a prognostic factor for outcome in patients with nonspecific low back pain: A systematic review.
      The lifetime prevalence of low back pain (LBP) is estimated at 80%,
      • Gross DP
      • Ferrari R
      • Russell AS
      • Battie MC
      • Schopflocher D
      • Hu RW
      • Waddell G
      • Buchbinder R.
      A population-based survey of back pain beliefs in Canada.
      and incidence of CLBP among individuals experiencing an acute LBP episode could be as high as 59%.
      • Heneweer H
      • Picavet HS
      • Staes F
      • Kiers H
      • Vanhees L.
      Physical fitness, rather than self-reported physical activities, is more strongly associated with low back pain: Evidence from a working population.
      ,
      • Henschke N
      • Maher CG
      • Refshauge KM
      • Herbert RD
      • Cumming RG
      • Bleasel J
      • York J
      • Das A
      • McAuley JH.
      Prognosis in patients with recent onset low back pain in Australian primary care: Inception cohort study.
      ,
      • Lee HI
      • Song J
      • Lee HS
      • Kang JY
      • Kim M
      • Ryu JS.
      Association between cross-sectional areas of lumbar muscles on magnetic resonance imaging and chronicity of low back pain.
      Direct and indirect costs of LBP in the United States have been estimated between US$84.1 billion to US$624.8 billion; similar costs have been reported in other countries.
      • Dagenais S
      • Caro J
      • Haldeman S.
      A systematic review of low back pain cost of illness studies in the United States and internationally.
      Research also shows that among chronic medical conditions, CLBP is associated with the highest level of medical resource use.
      • Rapoport J
      • Jacobs P
      • Bell NR
      • Klarenbach S.
      Refining the measurement of the economic burden of chronic diseases in Canada.
      Interestingly, a study of health care costs among individuals with LBP seen in primary care showed that 21% of patients had high total health care costs (≥US$2,700), which represented 68% of total costs for all patients in the study.
      • Engel CC
      • von Korff M
      • Katon WJ.
      Back pain in primary care: predictors of high health-care costs.
      Substantial fluctuations in pain intensity across time and/or individuals have been shown in multiple studies (see
      • Kongsted A
      • Kent P
      • Axen I
      • Downie AS
      • Dunn KM.
      What have we learned from ten years of trajectory research in low back pain?.
      for a review). Pain intensity variability is often operationalized either as a composite of within-day variations in pain intensity scores
      • Affleck G
      • Tennen H
      • Keefe FJ
      • Lefebvre JC
      • Kashikar-Zuck S
      • Wright K
      • Starr K
      • Caldwell DS.
      Everyday life with osteoarthritis or rheumatoid arthritis: Independent effects of disease and gender on daily pain, mood, and coping.
      ,
      • Allen KD
      • Golightly YM
      • Olsen MK.
      Pilot study of pain and coping among patients with osteoarthritis: A daily diary analysis.
      ,
      • Bartley EJ
      • Robinson ME
      • Staud R.
      Pain and fatigue variability patterns distinguish subgroups of fibromyalgia patients.
      ,
      • Bellamy N
      • Sothern RB
      • Campbell J.
      Rhythmic variations in pain perception in osteoarthritis of the knee.
      ,
      • Bellamy N
      • Sothern RB
      • Campbell J
      • Buchanan WW.
      Rhythmic variations in pain, stiffness, and manual dexterity in hand osteoarthritis.
      ,
      • Burstrom L
      • Jonsson H
      • Bjor B
      • Hjalmarsson U
      • Nilsson T
      • Reuterwall C
      • Wahlstrom J.
      Daily text messages used as a method for assessing low back pain among workers.
      ,
      • Ghaly MS
      • Handouka NM
      • Khashana AK
      • Shalaby SA
      • Fahim AE
      • Ramadan AE.
      WIthin-day variability of pain and its relation to quality of life in children with juvenile primary fibromyalgia syndrome.
      ,
      • Harris RE
      • Williams DA
      • McLean SA
      • Sen A
      • Hufford M
      • Gendreau RM
      • Gracely RH
      • Clauw DJ.
      Characterization and consequences of pain variability in individuals with fibromyalgia.
      ,
      • Keefe FJ
      • Affleck G
      • France CR
      • Emery CF
      • Waters S
      • Caldwell DS
      • Stainbrook D
      • Hackshaw KV
      • Fox LC
      • Wilson K.
      Gender differences in pain, coping, and mood in individuals having osteoarthritic knee pain: A within-day analysis.
      ,
      • Kratz AL
      • Murphy SL
      • Braley TJ.
      Ecological momentary assessment of pain, fatigue, depressive, and cognitive symptoms reveals significant daily variability in multiple sclerosis.
      ,
      • Nguyen RH
      • Turner RM
      • Sieling J
      • Williams DA
      • Hodges JS
      • Harlow BL.
      Feasibility of collecting vulvar pain variability and its correlates using prospective collection with smartphones.
      ,
      • O'Brien EM
      • Waxenberg LB
      • Atchison JW
      • Gremillion HA
      • Staud RM
      • McCrae CS
      • Robinson ME.
      Intraindividual variability in daily sleep and pain ratings among chronic pain patients: Bidirectional association and the role of negative mood.
      ,
      • Schneider S
      • Junghaenel DU
      • Keefe FJ
      • Schwartz JE
      • Stone AA
      • Broderick JE.
      Individual differences in the day-to-day variability of pain, fatigue, and well-being in patients with rheumatic disease: Associations with psychological variables.
      ,
      • Stone AA
      • Schwartz JE
      • Broderick JE
      • Shiffman SS.
      Variability of momentary pain predicts recall of weekly pain: A consequence of the peak (or salience) memory heuristic.
      ,
      • Timmermans EJ
      • de Koning EJ
      • van Schoor NM
      • van der Pas S
      • Denkinger MD
      • Dennison EM
      • Maggi S
      • Pedersen NL
      • Otero A
      • Peter R
      • Cooper C
      • Siviero P
      • Castell MV
      • Herbolsheimer F
      • Edwards M
      • Limongi F
      • Deeg DJH
      • Schaap LA.
      Within-person pain variability and physical activity in older adults with osteoarthritis from six European countries.
      ,
      • Toussaint L
      • Vincent A
      • McAllister SJ
      • Whipple M.
      Intra- and inter-patient symptom variability in fibromyalgia: Results of a 90-day assessment.
      ,
      • Williams DA
      • Gendreau M
      • Hufford MR
      • Groner K
      • Gracely RH
      • Clauw DJ.
      Pain assessment in patients with fibromyalgia syndrome: A consideration of methods for clinical trials.
      ,
      • Zakoscielna KM
      • Parmelee PA.
      Pain variability and its predictors in older adults: Depression, cognition, functional status, health, and pain.
      but sometimes also as a function of recall of average pain experienced over months, weeks, single weeks or days.
      • Bendayan R
      • Ramirez-Maestre C
      • Ferrer E
      • Lopez A
      • Esteve R.
      From acute to chronic back pain: Using linear mixed models to explore changes in pain intensity, disability, and depression.
      ,
      • Dobscha SK
      • Morasco BJ
      • Kovas AE
      • Peters DM
      • Hart K
      • McFarland BH.
      Short-term variability in outpatient pain intensity scores in a national sample of older veterans with chronic pain.
      ,
      • Koukkari WL
      • Sothern RB
      Introducing Biological Rhythms.
      ,
      • Lagersted-Olsen J
      • Bay H
      • Jorgensen MB
      • Holtermann A
      • Sogaard K.
      Low back pain patterns over one year among 842 workers in the DPhacto study and predictors for chronicity based on repetitive measurements.
      ,
      • Parry E
      • Ogollah R
      • Peat G.
      Significant pain variability in persons with, or at high risk of, knee osteoarthritis: Preliminary investigation based on secondary analysis of cohort data.
      ,
      • Tupper SM
      • Rosenberg AM
      • Pahwa P
      • Stinson JN.
      Pain intensity variability and its relationship with quality of life in youths with juvenile idiopathic arthritis.
      Variability can be estimated using various indicators, such as intra-individual standard deviation in pain scores, mean squared successive differences, and autocorrelation.
      • Mun CJ
      • Suk HW
      • Davis MC
      • Karoly P
      • Finan P
      • Tennen H
      • Jensen MP.
      Investigating intraindividual pain variability: Methods, applications, issues, and directions.
      In recent years, intra-individual variability has been identified as an important aspect to study and understand, as opposed to conceptualizing it as noise.
      • Lester HF
      • Cullen-Lester KL
      • Walters RW.
      From nuisance to novel research questions: Using multilevel models to predict heterogeneous variances.
      There is a growing body of literature focusing on pain intensity variability in acute LBP,
      • Bendayan R
      • Ramirez-Maestre C
      • Ferrer E
      • Lopez A
      • Esteve R.
      From acute to chronic back pain: Using linear mixed models to explore changes in pain intensity, disability, and depression.
      ,
      • Suri P
      • Rainville J
      • Fitzmaurice GM
      • Katz JN
      • Jamison RN
      • Martha J
      • Hartigan C
      • Limke J
      • Jouve C
      • Hunter DJ.
      Acute low back pain is marked by variability: An internet-based pilot study.
      CLBP,
      • Burstrom L
      • Jonsson H
      • Bjor B
      • Hjalmarsson U
      • Nilsson T
      • Reuterwall C
      • Wahlstrom J.
      Daily text messages used as a method for assessing low back pain among workers.
      ,
      • Kongsted A
      • Kent P
      • Axen I
      • Downie AS
      • Dunn KM.
      What have we learned from ten years of trajectory research in low back pain?.
      ,
      • Lagersted-Olsen J
      • Bay H
      • Jorgensen MB
      • Holtermann A
      • Sogaard K.
      Low back pain patterns over one year among 842 workers in the DPhacto study and predictors for chronicity based on repetitive measurements.
      and other adult
      • Affleck G
      • Tennen H
      • Keefe FJ
      • Lefebvre JC
      • Kashikar-Zuck S
      • Wright K
      • Starr K
      • Caldwell DS.
      Everyday life with osteoarthritis or rheumatoid arthritis: Independent effects of disease and gender on daily pain, mood, and coping.
      ,
      • Allen KD
      • Golightly YM
      • Olsen MK.
      Pilot study of pain and coping among patients with osteoarthritis: A daily diary analysis.
      ,
      • Bartley EJ
      • Robinson ME
      • Staud R.
      Pain and fatigue variability patterns distinguish subgroups of fibromyalgia patients.
      ,
      • Bellamy N
      • Sothern RB
      • Campbell J.
      Rhythmic variations in pain perception in osteoarthritis of the knee.
      ,
      • Bellamy N
      • Sothern RB
      • Campbell J
      • Buchanan WW.
      Rhythmic variations in pain, stiffness, and manual dexterity in hand osteoarthritis.
      ,
      • Clauw DJ
      • Crofford LJ.
      Chronic widespread pain and fibromyalgia: What we know, and what we need to know.
      ,
      • Dobscha SK
      • Morasco BJ
      • Kovas AE
      • Peters DM
      • Hart K
      • McFarland BH.
      Short-term variability in outpatient pain intensity scores in a national sample of older veterans with chronic pain.
      ,
      • Harris RE
      • Williams DA
      • McLean SA
      • Sen A
      • Hufford M
      • Gendreau RM
      • Gracely RH
      • Clauw DJ.
      Characterization and consequences of pain variability in individuals with fibromyalgia.
      ,
      • Keefe FJ
      • Affleck G
      • France CR
      • Emery CF
      • Waters S
      • Caldwell DS
      • Stainbrook D
      • Hackshaw KV
      • Fox LC
      • Wilson K.
      Gender differences in pain, coping, and mood in individuals having osteoarthritic knee pain: A within-day analysis.
      ,
      • Kratz AL
      • Murphy SL
      • Braley TJ.
      Ecological momentary assessment of pain, fatigue, depressive, and cognitive symptoms reveals significant daily variability in multiple sclerosis.
      ,
      • Nguyen RH
      • Turner RM
      • Sieling J
      • Williams DA
      • Hodges JS
      • Harlow BL.
      Feasibility of collecting vulvar pain variability and its correlates using prospective collection with smartphones.
      ,
      • O'Brien EM
      • Waxenberg LB
      • Atchison JW
      • Gremillion HA
      • Staud RM
      • McCrae CS
      • Robinson ME.
      Intraindividual variability in daily sleep and pain ratings among chronic pain patients: Bidirectional association and the role of negative mood.
      ,
      • Parry E
      • Ogollah R
      • Peat G.
      Significant pain variability in persons with, or at high risk of, knee osteoarthritis: Preliminary investigation based on secondary analysis of cohort data.
      ,
      • Schneider S
      • Junghaenel DU
      • Keefe FJ
      • Schwartz JE
      • Stone AA
      • Broderick JE.
      Individual differences in the day-to-day variability of pain, fatigue, and well-being in patients with rheumatic disease: Associations with psychological variables.
      ,
      • Stone AA
      • Schwartz JE
      • Broderick JE
      • Shiffman SS.
      Variability of momentary pain predicts recall of weekly pain: A consequence of the peak (or salience) memory heuristic.
      ,
      • Timmermans EJ
      • de Koning EJ
      • van Schoor NM
      • van der Pas S
      • Denkinger MD
      • Dennison EM
      • Maggi S
      • Pedersen NL
      • Otero A
      • Peter R
      • Cooper C
      • Siviero P
      • Castell MV
      • Herbolsheimer F
      • Edwards M
      • Limongi F
      • Deeg DJH
      • Schaap LA.
      Within-person pain variability and physical activity in older adults with osteoarthritis from six European countries.
      • Toussaint L
      • Vincent A
      • McAllister SJ
      • Whipple M.
      Intra- and inter-patient symptom variability in fibromyalgia: Results of a 90-day assessment.
      • Treister R
      • Honigman L
      • Lawal OD
      • Lanier RK
      • Katz NP.
      A deeper look at pain variability and its relationship with the placebo response: Results from a randomized, double-blind, placebo-controlled clinical trial of naproxen in osteoarthritis of the knee.
      ,
      • Williams DA
      • Gendreau M
      • Hufford MR
      • Groner K
      • Gracely RH
      • Clauw DJ.
      Pain assessment in patients with fibromyalgia syndrome: A consideration of methods for clinical trials.
      ,
      • Zakoscielna KM
      • Parmelee PA.
      Pain variability and its predictors in older adults: Depression, cognition, functional status, health, and pain.
      and pediatric
      • Ghaly MS
      • Handouka NM
      • Khashana AK
      • Shalaby SA
      • Fahim AE
      • Ramadan AE.
      WIthin-day variability of pain and its relation to quality of life in children with juvenile primary fibromyalgia syndrome.
      ,
      • Tupper SM
      • Rosenberg AM
      • Pahwa P
      • Stinson JN.
      Pain intensity variability and its relationship with quality of life in youths with juvenile idiopathic arthritis.
      populations experiencing chronic pain (see
      • Schneider S
      • Junghaenel DU
      • Broderick JE
      • Ono M
      • May M
      • Stone II, AA.
      Indices of pain intensity derived from ecological momentary assessments and their relationships with patient functioning: An individual patient data meta-analysis.
      ,
      • Schneider S
      • Junghaenel DU
      • Ono M
      • Broderick JE
      • Stone III, AA.
      Detecting treatment effects in clinical trials with different indices of pain intensity derived from ecological momentary assessment.
      ,
      • Stone AA
      • Broderick JE
      • Goldman RE
      • Junghaenel DU
      • Bolton A
      • May M
      • Schneider S.I.
      Indices of pain intensity derived from ecological momentary assessments: Rationale and stakeholder preferences.
      for an overview). Increased pain intensity variability has been shown to be positively associated with pain recall bias.
      • Stone AA
      • Schwartz JE
      • Broderick JE
      • Shiffman SS.
      Variability of momentary pain predicts recall of weekly pain: A consequence of the peak (or salience) memory heuristic.
      Given the well-known peak-and-end effect
      • Ariely D.
      Combining experiences over time: The effects of duration, intensity changes and on-line measurements on retrospective pain evaluations.
      ,
      • Kahneman D
      • Fredrickson BL
      • Schreiber CA
      • Redelmeier DA.
      When more pain is preferred to less: Addiont a better end.
      ,
      • Varey C
      • Kahneman D.
      Experiences extended across time: Evaluations of moments and episodes.
      that influences recall bias, measuring pain as a dynamic process in real-time using methodologies such as ecological momentary assessment (EMA) offers significant advantages.
      • Stone AA
      • Broderick JE
      • Goldman RE
      • Junghaenel DU
      • Bolton A
      • May M
      • Schneider S.I.
      Indices of pain intensity derived from ecological momentary assessments: Rationale and stakeholder preferences.
      ,
      • Stone AA
      • Obbarius A
      • Junghaenel DU
      • Wen CKF
      • Schneider S.
      High-resolution, field approaches for assessing pain: Ecological momentary assessment.
      Pain intensity variability is not simply a byproduct of different measurement approaches and is influenced by a multitude of factors. For example, demographic (eg, younger age, higher body mass index),
      • Parry E
      • Ogollah R
      • Peat G.
      Significant pain variability in persons with, or at high risk of, knee osteoarthritis: Preliminary investigation based on secondary analysis of cohort data.
      cognitive and psychological factors (eg, using distraction strategies as opposed to focusing on pain, higher levels of depressive symptoms) and neurophysiological factors (eg, greater amplitude in gamma oscillations)
      • Hu L
      • Iannetti GD.
      Neural indicators of perceptual variability of pain across species.
      have been associated with higher pain intensity variability.
      • Nouwen A
      • Cloutier C
      • Kappas A
      • Warbrick T
      • Sheffield D.
      Effects of focusing and distraction on cold pressor-induced pain in chronic back pain patients and control subjects.
      ,
      • Peyron R
      • Laurent B
      • Garcia-Larrea L.
      Functional imaging of brain responses to pain. A review and meta-analysis (2000).
      ,
      • Ploghaus A
      • Tracey I
      • Gati JS
      • Clare S
      • Menon RS
      • Matthews PM
      • Rawlins JN
      Dissociating pain from its anticipation in the human brain.
      ,
      • Zakoscielna KM
      • Parmelee PA.
      Pain variability and its predictors in older adults: Depression, cognition, functional status, health, and pain.
      It is also important to consider the broader context in which pain intensity variability takes place. Several studies have shown that pain intensity variability correlates with variations in the presence/absence and intensity of other symptoms, such as pain severity, pain unpleasantness, fatigue, anxiety, and low mood.
      • Bartley EJ
      • Robinson ME
      • Staud R.
      Pain and fatigue variability patterns distinguish subgroups of fibromyalgia patients.
      ,
      • Schneider S
      • Junghaenel DU
      • Broderick JE
      • Ono M
      • May M
      • Stone II, AA.
      Indices of pain intensity derived from ecological momentary assessments and their relationships with patient functioning: An individual patient data meta-analysis.
      ,
      • Zakoscielna KM
      • Parmelee PA.
      Pain variability and its predictors in older adults: Depression, cognition, functional status, health, and pain.
      The clinical impact and correlates of pain intensity variability are poorly understood. Some studies associated greater pain intensity variability with greater disability levels, lower functioning and poorer quality of life,
      • Schneider S
      • Junghaenel DU
      • Keefe FJ
      • Schwartz JE
      • Stone AA
      • Broderick JE.
      Individual differences in the day-to-day variability of pain, fatigue, and well-being in patients with rheumatic disease: Associations with psychological variables.
      ,
      • Suri P
      • Rainville J
      • Fitzmaurice GM
      • Katz JN
      • Jamison RN
      • Martha J
      • Hartigan C
      • Limke J
      • Jouve C
      • Hunter DJ.
      Acute low back pain is marked by variability: An internet-based pilot study.
      whereas other studies suggest that pain intensity variability is associated with levels of physical activities in men but not women.
      • Timmermans EJ
      • de Koning EJ
      • van Schoor NM
      • van der Pas S
      • Denkinger MD
      • Dennison EM
      • Maggi S
      • Pedersen NL
      • Otero A
      • Peter R
      • Cooper C
      • Siviero P
      • Castell MV
      • Herbolsheimer F
      • Edwards M
      • Limongi F
      • Deeg DJH
      • Schaap LA.
      Within-person pain variability and physical activity in older adults with osteoarthritis from six European countries.
      Little is known about the importance of pain intensity variability and patterns of health care utilization. To our knowledge, only one study examined weekly variations in pain intensity trajectories and health care utilization.
      • Hutchings A
      • Calloway M
      • Choy E
      • Hooper M
      • Hunter DJ
      • Jordan JM
      • Zhang Y
      • Baser O
      • Long S
      • Palmer L.
      The longitudinal examination of arthritis pain (LEAP) study: Relationships between weekly fluctuations in patient-rated joint pain and other health outcomes.
      Results showed that higher weekly pain intensity variability in adults with osteoarthritis is linked with greater health care utilization.
      • Hutchings A
      • Calloway M
      • Choy E
      • Hooper M
      • Hunter DJ
      • Jordan JM
      • Zhang Y
      • Baser O
      • Long S
      • Palmer L.
      The longitudinal examination of arthritis pain (LEAP) study: Relationships between weekly fluctuations in patient-rated joint pain and other health outcomes.
      ,
      • Schneider S
      • Junghaenel DU
      • Broderick JE
      • Ono M
      • May M
      • Stone II, AA.
      Indices of pain intensity derived from ecological momentary assessments and their relationships with patient functioning: An individual patient data meta-analysis.
      In an effort to better understand the extent and predictors of pain intensity variability, we report on an EMA study designed to (1) describe the extent and stability of variability in pain intensity among participants living with CLBP; (2) identify demographic and psychosocial predictors of variability in pain intensity; and (3) examine the associations between variability in pain intensity and psychosocial outcomes (ie, quality of life, disability) and health care utilization (ie, number of health care providers consulted, polypharmacy).

      Methods

      Subjects

      Participants were adults (≥18 years of age) suffering from CLBP (≥3 months duration) for which they have consulted a health care provider about their pain in the last two months. Patients also had to be fluent in written and spoken French and have access to Internet. Patients who had significant cognitive impairment, a diagnosis of cancer in the lower back, or cancer-related pain were excluded.

      Recruitment

      Recruitment took place in various medical and paramedical settings (ie, physiatry, physiotherapy and chiropractic clinics, hospital billboards, emergency rooms, primary care clinics [including family medicine groups and local community services centers]), through patient associations, and via conventional/social media (including hospital intranet) between January 2015 and September 2017.
      Study pamphlets and posters were distributed in health care settings and information about the study was disseminated in the media. Potential participants contacted the research team, and their eligibility was assessed during a telephone conversation with one of the members of the research team. If eligible, verbal and subsequently written consent were sought. Individuals who provided written consent received an email containing a link to the baseline questionnaires, which tapped into pain characteristics, psychosocial characteristics, and demographic information. This study was approved by the Research Ethics Committee of the Centre hospitalier de l'Université de Montréal.

      Ecological Momentary Assessment Design

      Participants were instructed to complete a momentary pain intensity report which was programmed on a mobile app (iSurveySoft [https://www.harvestyourdata.com/]) using an interval-based prompting design at each of the following times: 1) when they got up in the morning; 2) twice randomly during the day when they heard the auditory signal programmed to beep between 9:00 AM and 12:00 PM and again between 1:00 PM and 6:00 PM); 3) just before taking pro re nata (P.R.N.) analgesics; and 4) before going to bed in the evening. The momentary reports allowed for estimation of two models of pain intensity variability, namely a momentary model to assess current pain intensity levels, and a coverage model to assess average, worst, and least pain intensity levels since the previous prompt.
      • May M
      • Junghaenel DU
      • Ono M
      • Stone AA
      • Schneider S.
      Ecological momentary assessment methodology in chronic pain research: A systematic review.
      The momentary reports used the Numeric Rating Scale (0–10) to assess current pain intensity, as well as average, worst, and least pain intensity levels since the previous prompt. Scale anchors remained the same across pain intensity questions (0 = no pain; 10 = worst possible pain), but the time of reference differed (right now for current pain intensity versus since the previous diary entry for average, worst, and least pain intensity). Additional questions tapped into types of activities participants had engaged in since the previous momentary report that might have influenced their pain intensity ratings (ie, analgesic medication intake, exercise or strenuous physical activity of moderate to severe intensity, prolonged posture, rest or relaxation). Patients were instructed to complete the momentary reports for two periods of seven consecutive days, one month apart.

      Questionnaires

      The Numerical Rating Scale (NRS) for Pain Intensity (baseline (T0), completion of electronic diary (T1), and follow-up one month after completion of electronic diary (T2)) is a self-administered scale that measures pain intensity and can be adapted to different contexts based on its frame of reference (currently vs over the past X hours/days/weeks, etc) or pain characteristics (current pain, worst pain, least pain, averaged pain). The endpoints represent the extremes. For pain intensity, they range from 0 = “no pain at all” to 10 = “worst possible pain”. The NRS has been shown to have good reliability and validity as a measure of pain intensity in patients with chronic musculoskeletal pain or low back pain.
      • Childs JD
      • Piva SR
      • Fritz JM.
      Responsiveness of the numeric pain rating scale in patients with low back pain.
      ,
      • Salaffi F
      • Stancati A
      • Silvestri CA
      • Ciapetti A
      • Grassi W.
      Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale.
      Although its psychometric properties in EMA studies are unknown, it is one of the most commonly used measures for pain intensity in these types of studies.
      • May M
      • Junghaenel DU
      • Ono M
      • Stone AA
      • Schneider S.
      Ecological momentary assessment methodology in chronic pain research: A systematic review.
      It was the main instrument used in the EMA reports.
      Pain characteristics (T0). Participants were asked a series of questions about their past and current pain history in order to assess pain onset, type of pain (presence of radicular pain or not), circumstances surrounding pain onset, number of previous LBP episodes, perceived pain predictability, average length of episode, and frequency of LBP pain episodes.
      ThePain catastrophizing scale (PCS)
      • Sullivan MJL
      • Bishop SR
      • Pivik J.
      The pain catastrophizing scale: Development and validation.
      (T0) is a valid and reliable 13-item scale measuring pain-related rumination, magnification, and helplessness. The scale has been validated both in English and French. Participants rate each item on a 5-point scale ranging from 0 (not at all) to 4 (all the time), for a total score of up to 52. Cronbach alpha of .87 for the total scale is satisfactory. The scale also has good convergent validity with measures of anxiety (r = .32) and negative affect (r = .32). Ten-week test-retest reliability held up well (r = .70).
      • Sullivan MJL
      • Bishop SR
      • Pivik J.
      The pain catastrophizing scale: Development and validation.
      • Sullivan MJL
      • Lynch ME
      • Clark MR.
      Dimensions of catastrophic thinking associated with pain experience and disability in patients with neuropathic pain conditions.
      • Sullivan MJL
      • Thorn B
      • Haythornthwaite JA
      • Keefe F
      • Martin M
      • Bradley LA
      • Lefebvre JC.
      Theoretical perspectives on the relation between catastrophizing and pain.
      ThePattern of activity measure – pain (POAM-P)
      • Cane D
      • Nielson WR
      • McCarthy M
      • Mazmanian D.
      Pain-related activity patterns: Measurement, interrelationships, and associations with psychosocial functioning.
      (T0) is a 30-item scale assessing the patterns of engagement in daily activities when experiencing pain. The three subscales of the POAM-P (avoidant, overdoing, and pacing) have good internal consistency (α = .86-.94).
      • Cane D
      • Nielson WR
      • McCarthy M
      • Mazmanian D.
      Pain-related activity patterns: Measurement, interrelationships, and associations with psychosocial functioning.
      The POAM-P correlates low to moderately with measures of depression, anxiety, pain disability, and pain attitudes.
      • Cane D
      • Nielson WR
      • McCarthy M
      • Mazmanian D.
      Pain-related activity patterns: Measurement, interrelationships, and associations with psychosocial functioning.
      This questionnaire was translated into French using a forward-backward translation method by native speakers of French and English.
      TheHospital anxiety and depression scale (HADS)
      • Bjelland I
      • Dahl AA
      • Haug TT
      • Neckelmann D.
      The validity of the hospital anxiety and depression scale. An updated literature review.
      ,
      • Zigmond AS
      • Snaith RP.
      The hospital anxiety and depression scale.
      (T0, T2) is a 14-item scale that assesses levels of anxiety and depression in non-psychiatric patients. For each item, patients are asked to check the answer that most closely describes how they have been feeling in the past week on a scale from 0 to 3. Total scores for each subscale range from 0 to 21; higher scores indicate higher levels of anxiety or depression. The internal consistency of the depression (α = .67–.90) and anxiety (α = .68–.93) subscales of the HADS is adequate.
      • Bjelland I
      • Dahl AA
      • Haug TT
      • Neckelmann D.
      The validity of the hospital anxiety and depression scale. An updated literature review.
      The HADS also has good discriminant and convergent validity as well as specificity and sensitivity in detecting clinically significant levels of depression and anxiety.
      • Bjelland I
      • Dahl AA
      • Haug TT
      • Neckelmann D.
      The validity of the hospital anxiety and depression scale. An updated literature review.
      The psychometric properties of the French version of the HADS have been examined in a large primary-care population; results showed that the scale has good reliability (α = .79–.89) and a similar factor structure as the English version of the HADS (depression and anxiety subscales).
      • Roberge P
      • Dore I
      • Menear M
      • Chartrand E
      • Ciampi A
      • Duhoux A
      • Fournier L.
      A psychometric evaluation of the French Canadian version of the Hospital Anxiety and Depression Scale in a large primary care population.
      The Brief pain inventory-10 (BPI-10)
      • Tyler EJ
      • Jensen MP
      • Engel JM
      • Schwartz L.
      The reliability and validity of pain interference measures in persons with cerebral palsy.
      (T2) is a modified version of the 7-item BPI
      • Cleeland CS
      Measurement of pain by subjective report.
      • 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.
      • Cleeland CS
      • Ryan KM.
      Pain assessment: Global use of the brief pain inventory.
      that contains a total of 10 items assessing various aspects of pain interference. Participants are asked to rate on a scale from 0 to 10 the extent to which pain has interfered with general activity, mood, mobility, normal work, relationships, sleep, enjoyment of life, self-care, recreational activities, and social activities. The BPI has been shown to have good reliability and validity in a sample of patients with LBP.
      • 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.
      The BPI has been translated into French using a forward-backward translation method.
      • Larue F
      • Carlier AM
      • Brasseur L
      • Colleau SM
      • Cleeland CS
      Assessing the prevalence and severity of cancer pain in France: The French Brief Pain Inventory.
      The Short-form-12 health survey version 2 (SF-12v2)
      • Ware Jr., J
      • Kosinski M
      • Keller SD.
      A 12-item short-form health survey: Construction of scales and preliminary tests of reliability and validity.
      (T2) is a valid and reliable 12-item scale that assesses quality of life using two subscales: physical health and mental health. For each item patients are asked to check the box that best describes their answers; answer options vary from one question to the next. The SF-12 has good test-retest reliability (r = .76–.89) and internal consistency (α = .81–.84).
      Health Care Utilization Form (T2). A simplified and modified version of the Ambulatory and Home Care Record, focusing on utilization (and not healthcare costs), was used to assess healthcare utilization.
      • Guerriere DN
      • Choiniere M
      • Dion D
      • Peng P
      • Stafford-Coyte E
      • Zagorski B
      • Banner R
      • Barton PM
      • Boulanger A
      • Clark AJ
      • Gordon AS
      • Guertin MC
      • Intrater HM
      • Lefort SM
      • Lynch ME
      • Moulin DE
      • Ong-Lam M
      • Racine M
      • Rashiq S
      • Shir Y
      • Taenzer P
      • Ware M.
      The Canadian STOP-PAIN project - Part 2: What is the cost of pain for patients on waitlists of multidisciplinary pain treatment facilities?.
      ,
      • Guerriere DN
      • Ungar WJ
      • Corey M
      • Croxford R
      • Tranmer JE
      • Tullis E
      • Coyte PC.
      Evaluation of the ambulatory and home care record: Agreement between self-reports and administrative data.
      Participants were asked a series of ad-hoc questions to rate the extent to which they utilized health-care resources because of their LBP. Questions asked about the categories of health care providers consulted (using a pre-established list and allowing participants to select ‘others’ and specify additional types of providers consulted) and the types of analgesics and over-the-counter pain medication consumed to manage pain. These questions were used to compute two variables that were used in the statistical models. Data on consultations with health care providers was used to compute a variable that represented the number of different health care providers consulted during the reference period. Furthermore, participants’ medication intake was classified as polypharmacy, if they reported taking two or more drugs from different pain medication classes.

      Study Procedures

      Following completion of baseline questionnaires online (T0), participants installed two mobile applications (iSurveySoft [https://www.harvestyourdata.com/] and Random Alarms [https://itunes.apple.com/us/app/random-alarms/id660069126?mt=8]) on their smart device (phone or tablet). For those who did not own such a device, one was mailed to them for use during the diary study period. They then completed the EMA portion of the study (T1). One month following the completion of the second EMA recording period, participants received an email with a link to the follow-up questionnaires measuring pain-related outcomes and health care utilization (T2). For both baseline and follow-up time points, participants had a period of 7 days to complete the questionnaires. A reminder was sent after 4 days, if questionnaires had not been completed yet.
      Questionnaires assessing pain characteristics, pain-related catastrophization, kinesiophobia, activity patterns, and depression and anxiety symptoms were administered at baseline (T0). Questionnaires measuring depression and anxiety symptoms, pain interference, work absenteeism/presenteeism, quality of life, health care utilization and satisfaction with treatments were administered at the follow-up (T2). Based on the fear avoidance model of chronic pain,
      • Asmundson GJ
      • Norton PJ
      • Vlaeyen JWS
      Fear-avoidance models of chronic pain: An overview.
      ,
      • Vlaeyen JW
      • Linton SJ.
      Fear-avoidance model of chronic musculoskeletal pain: 12 years on.
      ,
      • Vlaeyen JWS
      • Linton S.
      Fear-avoidance and its consequences in chronic musculoskeletal pain: A state of the art.
      the following variables were identified as potential predictors of pain intensity variability: pain catastrophization, anxiety, depressive symptoms, and activity pattern. Age, sex, and pain characteristics were also included. Outcomes of pain intensity variability that were examined included physical and mental health-related quality of life, pain interference, number of health care providers consulted, and polypharmacy.

      Data Analyses

      Data analysis was based on a sample of convenience. Continuous variables were centered to their means. Categorical variables were handled using deviation contrast coding. Details of data manipulation can be found in Supplementary Material.
      A standard measure of intraclass correlation was measured using a basic linear mixed-effects model specified as follows. The dependent variable was the momentary pain intensity rating, and the independent variables were a global intercept and a random intercept by subject. A 95% confidence interval was computed using a parametric bootstrap method. This was implemented using the bootMer function in the lme4 library in R.
      R Core Team
      R: A Language and Environment for Statistical Computing.
      Obj. 1: To describe the extent and stability of pain intensity variability. Using a linear hierarchical location-scale model
      • Gerhart JI
      • Burns JW
      • Bruehl S
      • Smith DA
      • Post KM
      • Porter LS
      • Schuster E
      • Buvanendran A
      • Fras AM
      • Keefe FJ.
      Variability in negative emotions among individuals with chronic low back pain: Relationships with pain and function.
      ,
      • Hedeker D
      • Mermelstein RJ
      • Demirtas H.
      Modeling between-subject and within-subject variances in ecological momentary assessment data using mixed-effects location scale models.
      with a normal likelihood function, we aimed to quantify the intra-individual variability in momentary pain intensity ratings across the two 7-day recording periods. Pain intensity variability was operationalized using intra-individual standard deviations in current pain intensity scores because it is the most commonly used indicator of pain intensity variability,
      • Mun CJ
      • Suk HW
      • Davis MC
      • Karoly P
      • Finan P
      • Tennen H
      • Jensen MP.
      Investigating intraindividual pain variability: Methods, applications, issues, and directions.
      can capture the degree of variations in pain scores regardless of their temporality, can be computed with pain intensity measurements not equally spaced, and requires few observations to be captured reliably.
      • Schneider S
      • Junghaenel DU
      • Ono M
      • Broderick JE
      • Stone III, AA.
      Detecting treatment effects in clinical trials with different indices of pain intensity derived from ecological momentary assessment.
      Location-scale models can be used when one is interested in modeling the mean and variance structure of intensive longitudinal data measured multiple times across multiple waves in the same participants. This model included both a location component and a scale component (with a log link function) to model the mean and the variance in the data, respectively. The location component consisted of the following variables, each with both fixed and random effects: intercept, time trend (time/prompt for completing an EMA report), indicator for period (1st week vs 2nd week), and interaction between time trend and period. These variables are sufficient to describe the individual-level trends in reported current pain across the two periods. The scale component consisted of the following variables, each with both fixed and random effects: intercept and indicator for period. These variables are sufficient to allow us to measure the subject-specific variability around the time trends in each period. Possible pain scores (0–10) were the only knowledge used to inform priors for parameters. All posterior distributions were manually compared to their respective prior distributions to make sure the priors did not have any non-negligible influence on posterior estimates. Analyses were conducted using Stan and R statistical software.
      • Carpenter B
      • Gelman A
      • Hoffman MD
      • Lee D
      • Goodrich B
      • Betancourt M
      • Brubaker M
      • Guo J
      • Li P
      • Riddell A.
      Stan: A probabilistic programming language.
      ,
      R Core Team
      R: A Language and Environment for Statistical Computing.
      The statistical approach was chosen over more standard statistical models such as hierarchical mixed-effects models as it allows to explore within-subject variability as a dependent variable, which was the central objective of this study.
      Intra-individual standard deviation derived from location-scale modeling was chosen over other variability measures (such as autoregressive correlation, mean square of successive differences) because of its ease of interpretation and because it allows for the use of a single parametric model that is fit to the entire unprocessed data.
      Obj 2. Identify demographic and psychosocial predictors of variability in pain intensity. In order to address the second objective, the model used for the first objective was extended in the following manner. Participant-level predictors (age, sex, pain characteristics (overall pain intensity, duration of pain, pain type, pain predictability), pain-related catastrophization, anxiety, depression and activity patterns [overdoing, avoidance, pacing]) were added to the scale component of the model to predict intra-individual variability in pain scores. Categorical variables with more than two categories (ie, pain type, pain predictability) were coded using deviation contrast coding to compare each category's mean to the overall mean.
      Obj 3. Examine the associations between variability in pain intensity and psychosocial and health care utilization outcomes. Since the predictor of interest (ie, pain intensity variability) is an inferred quantity, rather than an observed one, a unique modeling approach was employed to measure its associations with the five outcomes of interest (physical health-related quality of life, mental health-related quality of life, pain interference, number of health care providers consulted, and polypharmacy). Specifically, the predictor (ie, pain intensity variability) was entered into the model as a latent variable having a subject-specific probability distribution as inferred in objective 1 for the random intercept in the scale component of the model (ie, normal with known mean and variance). Missing outcomes were treated as interval censored over the possible range of the data. Physical health-related quality of life and pain interference outcomes were modeled using a linear model with a normal distribution whereas mental health-related quality of life was modeled with a linear model and a Student's t distribution, due to a lack of fit in the tails. The number of health care providers consulted was modeled as count data using a log link and a mixture between a Bernoulli distribution and a Poisson distribution truncated below 1 (ie, hurdle model). Patients with missing outcomes were excluded from this model. Finally, the polypharmacy outcome was modeled as a dichotomous variable using a logistic regression model. Patients with missing outcomes were excluded from this model as well. For all these models, priors were informed by the known measurement scale of data where possible, and otherwise diffuse priors were used.

      Results

      The study flow chart appears in Fig 1. A total of 183 individuals participated in the study. They were primarily recruited from physiotherapy clinics (n = 42), intranet of the Centre hospitalier de l'Université de Montréal (n = 33), patient associations (n = 33), Centre hospitalier de l'Université de Montréal physiatry department (n = 17), conventional/social media (n = 16), primary care clinics (n = 15), other health care clinics (n = 12), hospital billboards (n = 12), and emergency rooms (n = 3) of the Greater Montreal area. A final number of 140 participants were selected for analyses on the criteria that a minimum of 10 observations (out of a potential 28, not taking into account those completed before taking p.r.n. analgesics, according to the EMA design) had to be recorded per weekly period (sensitivity analyses were carried out on the sample of all participants who completed at least one diary entry (N = 160) and point and interval estimates remained stable compared to the main results). Fig 2 shows the number of EMA reports completed per participant for each of the two periods (1st week and 2nd week) as well as overall for both periods. Independent Student's t-tests and chi-square tests were used to determine whether or not those who were included in the analyses for Objectives 1-3 were significantly different at baseline from those who completed the baseline questionnaire but were excluded from the analyses due to low compliance with the EMA procedure or due to dropping outs in terms of sociodemographic and pain characteristics. There were no significant differences between those included and those excluded in terms of age, sex, education level, average pain intensity, pain duration, and scores on the psychological measures (anxiety, depression, catastrophizing, kinesiophobia, and patterns of activity) at baseline (all P > .05). Follow-up questionnaires were completed on average 68.0 (standard deviation: 14.1) days after completion of baseline questionnaires.
      Figure 2
      Figure 2Ecological Momentary Assessment (EMA) study compliance across participants (N = 160) and study period. Graphs A and B represent the number of EMAs completed per participant within the first and second weeks, respectively. Graph C represents the number of EMA reports completed per participant during the entire study.

      Participants’ Characteristics

      Sociodemographic, pain, and psychological characteristics are presented in Table 1. Most participants had been living with CLBP for many years (median: 5.2 years, interquartile range: 1.7–15.1) and reported an average pain intensity of 5.0 (±2.0) on a 0–10 scale. At follow-up, more than half of participants had consulted 2 or more different types of health care providers over the past 2 months, and two thirds were taking 2 or more different types of pain medication. Levels of psychological distress, such as depressive symptoms, anxiety, and pain catastrophizing were globally below the clinically significant levels. The intraclass correlation for subjects’ momentary pain intensity was substantial at .62 (95% confidence interval: .55–.67).
      Table 1Sociodemographic and Clinical Characteristics of the Study Participants (N = 140)
      Sociodemographic Characteristics
      Sex [N(%)]
       Female98 (70.0)
       Male42 (30.0)
      Age (mean ± SD)46.5 ± 13.3
       Ethnicity [N(%)]
       Caucasian124 (88.6)
       Black6 (4.3)
       Others10 (7.1)
      Education [N(%)]
       High school16 (11.4)
       Collegial43 (30.7)
       University81 (57.9)
      Living status [N(%)]
       Partners and/or children99 (70.7)
       Alone28 (20.0)
       Others13 (9.3)
      Work status [N(%)]
       Full-time70 (50.0)
       Part-time14 (10.0)
       Retired17 (12.1)
       Disability20 (14.3)
       Others19 (13.6)
      Baseline Pain Characteristics
      Pain type [N(%)]
       LBP with radiculopathy69 (49.3)
       LBP without radiculopathy24 (17.1)
       Diffuse low back pain20 (14.3)
       Radicular pain only3 (2.1)
       Other24 (17.1)
       Pain duration (mean ± SD)

      (median [interquartile range])
      8.9 ± 9.4

      5.2[1.7; 15.1]
      Pain frequency [N(%)]
       Continuous112 (80.0)
       Intermittent28 (20.0)
      Perceived pain predictability [N(%)]
       My pain is unpredictable39 (27.9)
       My pain is mostly worse in the morning24 (17.1)
       My pain is mostly worse in the afternoon6 (4.3)
       My pain is mostly worse in the evening/night22 (15.7)
       My pain is mostly worse when I do specific physical activities49 (35.0)
      Average pain intensity in past 7 d (NRS-11) (mean±SD)5.1 ± 2.0
      Worst pain intensity in past 7 d (NRS-11) (mean±SD)6.8±2.0
      Least pain intensity in past 7 d (NRS-11) (mean±SD)2.6±2.1
      ER visits in previous 6 mo due to LBP [N(%)]
       Yes33 (23.6)
       No107 (76.4)
      Hospitalization in previous 6 mo due to LBP [N(%)]
       Yes6 (4.3)
       No134 (95.7)
      Health Care Utilization at Follow-Up
      Number of health care providers consulted [N(%)]
       036 (25.9)
       130 (21.6)
       232 (23.0)
       321 (15.1)
       ≥420 (14.4)
      Polypharmacy [N(%)] (≥2 different pain medication classes)
       Yes93 (66.9)
       No46 (33.1)
      Current medication use [N(%)]
       Acetaminophen84 (60.4)
       Muscle relaxants50 (36.0)
       Non-selective NSAIDS50 (36.0)
       Antidepressants32 (23.0)
       Anticonvulsants24 (17.3)
       Short-acting opioids24 (17.3)
       Selective NSAIDS21 (15.1)
       Long-acting opioids11 (7.9)
       Tramadol10 (7.2)
       Topical agents8 (5.8)
       NMDA antagonists1 (0.7)
      Psychosocial characteristics (mean ± SD)
      BaselineFollow-up
      HADS-Anxiety8.5 ± 3.67.9 ± 3.4
      HADS-Depression6.3 ± 4.65.8 ± 4.3
      PCS21.5 ± 10.5
      POAMP-Avoidance19.8 ± 8.8
      POAMP-Overdoing23.2 ± 7.1
      POAMP-Pacing21.0 ± 10.2
      SF12.V2-Physical37.4 ± 7.3
      SF12.v2-Mental43.2 ± 7.8
      Abbreviations: NSAID, non-steroidal anti-inflammatory drugs; NMDA, N-methyl-D-aspartate; HADS-A, hospital anxiety and depression scale- anxiety subscale; HADS-D, hospital anxiety and depression scale- depression subscale; PCS, pain catastrophization scale; TSK, tampa scale kinesophobia; POAMP, patterns of activity measure- pain; SF12.V2-physical, physical health-related norm-based score of the short-form 12 version 2; SF12.V2-Mental, mental health-related norm-based score of the short-form 12 version 2.
      Obj. 1: To describe the extent and stability of pain intensity variability. The EMA portion of the study resulted in 6,730 momentary pain reports (3,397 during the 1st week and 3,333 during the 2nd week) nested within the 140 participants. There were 452 entries (253 and 199 in weeks 1 and 2, respectively) related to medication intake, while the 6,278 remaining entries represented the fixed (n = 3,273) and random (n = 2,932) entries or additional entries that participants made (n = 73). As such, on average, participants responded to 79.1% of fixed and random EMA momentary reports (79.6% in 1st week and 78.7% in 2nd week). As a reference, completion rates were of 74.9% during the 1st week and 72.6% during the 2nd week for the sample of participants who completed at least one EMA (N = 160). Among the EMA reports completed by the 140 participants included in the analyses, completion rates for the getting up, going to bed, and random prompts were 87.6%, 81.4% and 74.7% during the 1st week and 85.8%, 79.2% and 74.8% during the 2nd week, respectively.
      Fig 3 shows two examples of participants’ pain intensity variability, represented as the model's point and interval predictions for the outcome, given the observed covariates, plotted against the observed outcomes. As shown in this figure, individuals differ in the extent to which their EMA-reported pain intensity scores vary across time. For some participants, there is little variation in pain within the week whereas for others pain scores fluctuate substantially. The point and interval estimate of the parameters from the hierarchical location-scale model are shown in Table 2. Findings reveal interesting patterns in the data. Firstly, the average patient's mean momentary pain intensity score was 3.75 (95% highest density interval [HDI]: 3.43–4.07), and this pain intensity increased slightly over time (posterior mode: .16/wk (95% HDI: .00 to .32), though it did not differ between the two EMA periods (posterior mode: -.10 (95% HDI: -.34 to .15)), nor did its rate of change (posterior mode: .04 per week (95% HDI: -.29 to .37)). Secondly, after accounting for EMA period (1st week vs 2nd week) and time trend (time variable within each weak), the average patient's pain intensity fluctuations over time had a standard deviation of 1.2. This standard deviation differed slightly between weekly periods—approximately 1.3 for the first one and 1.1 for the second (coefficient: -.30, 95% HDI: -.42 to -.18). There was substantial heterogeneity between study participants in terms of both mean pain intensity and pain intensity variability over time. Specifically, the standard deviation of the random intercept of the location component of the model was 1.94, indicating that 95% of patient-specific mean pain intensities were between scores of 0 and 8. Furthermore, the standard deviation of the random intercept of the variance component of the model was .75, indicating that 95% of patient-specific standard deviations were between .57 and 2.56.
      Figure 3
      Figure 3Momentary pain intensity scores for 2 participants (top graphs depict data from a participant with greater pain intensity variability whereas the bottom graphs illustrate momentary reports from a participant expressing lower pain intensity variability) across the two periods. The raw observations of the momentary pain scores are plotted over time in a solid black line for period 1 (on the left) and period 2 (in the center). Overlaid in red are the model's posterior predictions for current pain given the observed covariates: median (central notch), first and third quartiles (vertical line segment), and 2.5 and 97.5 percentiles (bottom and top crosses, respectively). Finally, to the far right are displayed the point and interval estimates of the participant-specific standard deviation, in blue: mode (central notch) and 95% highest density interval (vertical line segment). Period refers to the week that the ecological momentary assessment took place (1st vs 2nd wk) whereas time refers to the specific moment at which EMAs were completed within each period.
      Table 2Point and Interval Estimates of the Parameters From the Hierarchical Location-Scale Model Examining Predictors of 6,730 Ecological Momentary Assessment Pain Intensity Reports Among 140 Participants Living With Chronic Low Back Pain
      ParameterPosterior modes (95% highest density interval)
      Location component of model-
       Intercept, overall3.75 (3.43; 4.07)
       Time, overall (per week).16 (.00; .32)
       Intercept, period 2 vs period 1-.10 (-.34; .15)
       Time, period 2 vs period 1.04 (-.29; .37)
      Variance component of model, within subject (log-scale)-
       Intercept, overall.38 (.25; .51)
       Intercept, period 2 vs period 1-.30 (-.42; -.18)
      Random effect standard deviations, between subjects-
       (1) Location, intercept, overall1.94 (1.72; 2.20)
       (2) Location, time, overall.67 (.47; .85)
       (3) Location, intercept, period 1 vs period 11.19 (1.01; 1.43)
       (4) Location, time, period 2 vs period 11.34 (1.03; 1.71)
       (5) Variance, intercept, overall.75 (.65; .85)
       (6) Variance, intercept, period 2 vs period 1.58 (.46; .70)
      NOTE: Period refers to the week the ecological momentary assessment took place (1st vs 2nd wk) while time refers to the time at which EMAs were completed within each period. The overall intercept in the variance component of the model can be converted to a standard deviation by taking the square-root of its exponent (ie, sqrt (exp (0.38)) = 1.2). Similarly, the period effect in the variance component of the model can be converted to a multiplicative effect to the standard deviation using the same transformation.
      Obj 2. Identify demographic and psychosocial predictors of variability in pain intensity. In this expanded model, age, sex, average pain at baseline, pain duration, type of CLBP, pain predictability, anxiety, depression, pain catastrophizing, and activity patterns (ie, avoidance, overdoing, pacing) were used to explain between-participant variability in momentary pain intensity scores. Coefficient estimates for these explanatory variables are shown in Table 3. The marginal effects implied by these coefficients can be converted to multiplicative effects to the intra-subject standard deviation by taking the square-root of the exponent of the coefficient. For example, a marginal effect of ±.2 is equivalent to a 10.5% higher standard deviation or a 9.5% lower one, respectively. These are reported in the last column of Table 3 for ease of interpretation.
      Table 3Coefficient Estimates of the Multivariable Linear Model That Examines the Association Between Patient-Level Factors and Intra-Subject Variability in EMA Momentary Pain Intensity Scores Over Time
      ParameterPosterior mode95% highest density intervalEquivalent % change in SD
      Age at baseline, per 10 y.00-.11; .11.0
      Sex (female vs male)-.05-.34; .26-2.5
      Average pain in past 7 d at baseline.04-.05; .122.0
      Duration of pain at baseline, per 5 y.04-.04; .112.0
      Pain type at baseline (compared to the average patient)
      Back pain only-.20-.52; .14-9.5
      Back pain with radiculopathy-.01-.26; .28-.5
      Radiculopathy only.05-.65; .852.5
      Diffuse back pain.03-.33; .371.5
      Other
      Deviation contrast coding used here (ie, each category mean is compared to the overall mean). As in all contrast coding types, one category is always left out. All continuous variables are centered around their means.
      Pain predictability at baseline (compared to the average patient)
      Unpredictable-.18-.44; .09-8.6
      Worse in the morning.16-.15; .458.3
      Worse in the afternoon-.13-.64; .41-6.3
      Worse in the evening.25-.06; .5613.3
      Worse based on specific activities
      Deviation contrast coding used here (ie, each category mean is compared to the overall mean). As in all contrast coding types, one category is always left out. All continuous variables are centered around their means.
      Anxiety score at baseline-.03-.08; .02-1.5
      Depression score at baseline.00-.04; .04.0
      Pain catastrophizing at baseline, per 5 points-.07-.15; .01-3.4
      Avoidance pattern at baseline, per 5 points.03-.06; .131.5
      Overdoing pattern at baseline, per 5 points.02-.09; .111.0
      Pacing pattern at baseline, per 5 points.00-.07; .09.0
      low asterisk Deviation contrast coding used here (ie, each category mean is compared to the overall mean). As in all contrast coding types, one category is always left out. All continuous variables are centered around their means.
      Results showed an equivocal pattern in interval estimates regarding the role of pain type and pain predictability in participants’ pain intensity variability. Specifically, all interval estimates included the null value, zero. As such, these patient-level predictors do not play a statistically significant role in explaining intra-individual variability in pain. Furthermore, once the marginal effects are converted into their implied percentage impact on pain intensity variability, effects do not appear to be clinically meaningful either, varying between -9.5% and 13.3%. These effects are negligible in comparison to the between-subject heterogeneity (which had an effect of 45% per standard deviation). Another way to gain more insight into this question is by comparing the estimate for the standard deviation of the random intercept of the variance component of the model between the original model used in objective 1 and this expanded model. Essentially, if these predictors were useful, then the variance of this random intercept would decrease noticeably, since between-subject differences would be explainable by those subject-level predictors and thus already accounted for in the model. However, the standard deviation of that random effect was .74 (95% HDI: .65–.87) in this expanded model, compared to .75 (95% HDI: .65–.85) in the original one—ie, virtually identical. It is, therefore, likely that these predictors indeed have limited explanatory value.
      Obj 3. Examine the association between momentary pain intensity variability and psychosocial and health care utilization outcomes. This objective aimed to determine whether or not intra-individual variability in pain scores during the weekly periods was associated with pain-related outcomes measured one month later. No additional covariates were included in these models since results from the analyses conducted for Objective 2 suggested no material associations between patient characteristics and pain intensity variability.
      Results showed no statistically significant associations between intra-individual variability in pain scores and physical and mental health-related quality of life, pain interference, number of health care providers consulted, and polypharmacy. Parameter estimates are shown in Table 4.
      Table 4Parameter Estimates of the Associations Between Ecological Momentary Assessment (EMA) Pain Intensity Variability and Pain-Related Outcomes
      Outcome of interestInterceptSlope
      Coefficient95% highest density intervalCoefficient95% highest density interval
      Physical health-related quality of life (SF12v2-P)38.537.2; 39.7-.17-1.9; 1.6
      Mental health-related quality of life (SF12v2-M)43.942.6; 45.2-.18-2.0; 1.7
      Pain interference (BPI)4.54.2; 4.8.2-.2; .6
      Number of health care providers consulted
      Hurdle component (zero outcome, logit scale)

      Poisson component (log scale)
      -1.1

      .8
      -1.5; -.7

      .7; 1.0
      -.1

      -.1
      -.6; .5

      -.3; .1
      Polypharmacy (yes/no, logit scale).7.4; 1.1.4-.1; .9
      Abbreviations: SF12v2-P, physical health-related norm-based summary score from the short-form-12 version 2; SF12v2-M, mental health-related norm-based summary score from the short-form-12 version 2. BPI, brief pain inventory.
      Though there is no evidence of any statistical association between pain intensity variability and pain-related outcomes, we might still consider the clinical significance of these estimates. Indeed, a consensus statement from the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials underscores the importance of examining statistical significance but also clinical meaningfulness of results, since statistical significance is dependent upon sample size and not only variability and magnitude of treatment effects.
      • Dworkin RH
      • Turk D
      • Wyrwich KW
      • Beaton C
      • 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.
      So for example, we can consider physical health-related quality of life. The mean of this outcome was 38.5 (95% highest density intervals: 37.2; 39.7), and the association between momentary pain intensity variability and this outcome was -.17 (95% highest density intervals: -1.9; 1.6). Since the standard deviation of the random intercept from the variance component of the model was .75, then the impact of one standard deviation of our predictor on the outcome is -.13 points on the physical quality of life scale, which is not clinically meaningful. The story is very similar for mental health-related quality of life and pain interference. Looking at the number of providers’ outcome, the intercept from the hurdle component of the model, -1.1, can be transformed into the probability of having had zero providers by using the inverse logit transformation—this yields a probability of approximately .25. The slope coefficient can be transformed into an odds ratio (OR) by taking the exponent of -.1, which yields an OR of .9, meaning that for each unit increase in pain intensity variability the odds of having zero providers is reduced by 10%. Furthermore, the intercept from the Poisson component of the model, .8, can be converted to the mean number of providers during follow-up by taking its exponent, yielding a rate of 2.2 providers per 1 month of follow-up. The slope coefficient can be transformed into a rate ratio (RR) by taking the exponent of -.1, which yields a RR of .9, meaning that for each unit increase in pain intensity variability the rate of provider visits is reduced by 10%. Similar calculations for the polypharmacy outcome yield that 67% of subjects used two or more different types of pain medication, and that the OR for this outcome was 1.5—ie, for each unit increase in pain intensity variability the odds of using two or more different types of pain medication was 50% greater.

      Discussion

      The study aimed to examine intra-individual variability in current pain intensity scores among individuals with CLBP, identify baseline predictors of variability, and explore its impact on pain outcomes and health care utilization. Results showed that the average patient's momentary pain intensity fluctuated with a standard deviation of 1.2 (12% relative to the scale of the data) over a week of observation, with some individuals experiencing more limited fluctuations in pain intensity levels (<5% of measurement scale) and others evincing a standard deviation greater than 25% of measurement scale. The study did not identify predictors of this intra-individual pain intensity variability. Finally, this variability was not conclusively associated with psychosocial characteristics and health care utilization either.

      Intra- and Inter-Individual Pain Intensity Variability

      Pain intensity variability has been shown in different populations. For example, among individuals with fibromyalgia, pain intensity variability had a mean ± SD of 1.61 ± .66.
      • Harris RE
      • Williams DA
      • McLean SA
      • Sen A
      • Hufford M
      • Gendreau RM
      • Gracely RH
      • Clauw DJ.
      Characterization and consequences of pain variability in individuals with fibromyalgia.
      Pain intensity variability in the present study was broadly similar over the course of a week. A recent study showed that various pain indices are perceived as having most impact on one's experience; individuals living with chronic pain consider their worst pain and more time in high intensity pain as most meaningful whereas clinicians give more weight only to worst pain intensity scores.
      • Stone AA
      • Broderick JE
      • Goldman RE
      • Junghaenel DU
      • Bolton A
      • May M
      • Schneider S.I.
      Indices of pain intensity derived from ecological momentary assessments: Rationale and stakeholder preferences.
      This has important implications not only for clinical care but also for research. Ignoring pain intensity variability could lead to studies being underpowered due to measurement error, and could be associated with false positive findings in terms of directionality and magnitude of effects.
      • Button KS
      • Ioannidis JP
      • Mokrysz C
      • Nosek BA
      • Flint J
      • Robinson ES
      • Munafo MR.
      Power failure: Why small sample size undermines the reliability of neuroscience.
      ,
      • Gelman A
      • Carlin J.
      Beyond power calculations: assessing type S (Sign) and type M (Magnitude) errors.
      Determination of sample sizes in multilevel studies need to consider many aspects of data, such as the intraclass correlation coefficients (a measure of homogeneity), in addition to power, nested structure, level of significance, and effect sizes.
      • Cohen MP.
      Sample size considerations for multilevel surveys.
      ,
      • Maas CJM
      • Hox JJ.
      Robustness issues in multilevel regression analysis.
      ,
      • Maas CJM
      • Hox JJ.
      Sufficient sample sizes for multilevel modeling.
      ,
      • Mathieu JE
      • Aguinis H
      • Culpepper SA
      • Chen G.
      Understanding and estimating the power to detect cross-level interaction effects in multilevel modeling.
      ,
      • Snijders TAB
      Power and sample size in multilevel modeling.

      Predictors of Pain Intensity Variability

      Surprisingly, none of the sociodemographic, pain-related or psychological characteristics at baseline were associated with momentary pain intensity variability. Multiple factors can influence one's pain perception, including for example nociception and other biological factors, cognitions, contexts, and mood; these factors can also vary from moment to moment.
      • Madden VJ
      • Kamerman PR
      • Catley MJ
      • Bellan V
      • RUssek LN
      • Camfferman D
      • Moseley L.
      Variability in experimental pain studies: Nuisance or opportunity?.
      As such, predictors of pain intensity variability might not be individual difference characteristics, but rather time-varying characteristics. Indeed, a recent study examined daily measures of pain intensity and psychosocial characteristics (ie, pain catastrophizing, pain self-efficacy, and negative affect) among 54 individuals with LBP and found that daily increases in pain intensity were associated with daily increases in catastrophizing, negative affect and lower levels of pain self-efficacy.
      • Wesolowicz DM
      • Bishop MD
      • Robinson ME.
      An examination of day-to-day and intraindividual pain variability in low back pain.
      It is also possible that low levels of psychological distress, pain catastrophizing, and pain avoidance behaviors in this community sample contributed to the inconclusive results obtained.
      Our results are in contrast with some other studies that showed depressive ratings, self-efficacy, and emotional and physical functioning to be associated with intra-individual variability in pain intensity scores.
      • Schneider S
      • Junghaenel DU
      • Keefe FJ
      • Schwartz JE
      • Stone AA
      • Broderick JE.
      Individual differences in the day-to-day variability of pain, fatigue, and well-being in patients with rheumatic disease: Associations with psychological variables.
      ,
      • Schneider S
      • Junghaenel DU
      • Ono M
      • Stone AA.
      Temporal dynamics of pain: An application of regime-switching models to ecological momentary assessments in patients with rheumatic diseases.
      ,
      • Zakoscielna KM
      • Parmelee PA.
      Pain variability and its predictors in older adults: Depression, cognition, functional status, health, and pain.
      Some of these studies, however, either looked at variability across daily pain averages,
      • Schneider S
      • Junghaenel DU
      • Keefe FJ
      • Schwartz JE
      • Stone AA
      • Broderick JE.
      Individual differences in the day-to-day variability of pain, fatigue, and well-being in patients with rheumatic disease: Associations with psychological variables.
      ,
      • Zakoscielna KM
      • Parmelee PA.
      Pain variability and its predictors in older adults: Depression, cognition, functional status, health, and pain.
      or used different metrics of pain intensity variability such as average duration of pain states.
      • Schneider S
      • Junghaenel DU
      • Ono M
      • Stone AA.
      Temporal dynamics of pain: An application of regime-switching models to ecological momentary assessments in patients with rheumatic diseases.
      In the present study, pain intensity variability was examined using momentary pain scores measured at five different times: two fixed times (getting up and going to bed), two random times (once in the morning and once in the afternoon), and prior to taking medication. It is possible that such methodological choices across studies impact on the magnitude of pain intensity variability and its predictors.
      • Kwasnicka D
      • Kale D
      • Schneider V
      • Keller J
      • Yeboah-Asiamah Asare B
      • Powell D
      • Naughton F
      • Ten Hoor GA
      • Verboon P
      • Perski O
      Systematic review of ecological momentary assessment (EMA) studies of five public health-related behaviours: Review protocol.
      Harmonization across studies would facilitate comparability of results across samples.

      Impact of Pain Intensity Variability

      Results overall showed no conclusive association of pain intensity variability and patient-reported outcomes. These findings are consistent with those from a patient data meta-analysis that indicated that pain intensity variability is not associated with various functioning domains.
      • Schneider S
      • Junghaenel DU
      • Broderick JE
      • Ono M
      • May M
      • Stone II, AA.
      Indices of pain intensity derived from ecological momentary assessments and their relationships with patient functioning: An individual patient data meta-analysis.
      In contrast, other measures of pain derived from EMAs, such as maximum pain, pain intensity variability or amount of time in high pain, were associated with some aspects of physical functioning, depressive symptoms, and social functioning.
      • Schneider S
      • Junghaenel DU
      • Broderick JE
      • Ono M
      • May M
      • Stone II, AA.
      Indices of pain intensity derived from ecological momentary assessments and their relationships with patient functioning: An individual patient data meta-analysis.
      These findings show that pain intensity variability on its own might not capture the full impact of fluctuating pain levels on patient`s functioning, psychosocial outcomes and health care utilization. For example, 10 to 20% variations in pain intensity scores might not have the same impact on individuals when pain is mild compared to when it is moderate to severe. It might also be that what is driving health care utilization is high pain intensity levels or high time spent in pain more so than variability in pain intensity scores. Measures of pain intensity variability in this study are dependent on the frequency of pain measurement, in this case at least four times per day. This type of measurement does not provide metrics about the amount of time spent in high versus low pain intensity nor the number of low-high pain cycles one has encountered during the day since repeated measurements would have to occur twice as frequently as the shortest cycle to allow for proper detection. Multiple dimensions associated with pain intensity variability could be examined, such as the speed of pain cycles or the frequency of changes in pain scores, the predictability of pain intensity variability, and perceived control over pain variations. Such dimensions of the pain experience might influence differentially pain appraisal, pain behaviors, and pain-related outcomes. Indeed, a study of individuals with musculoskeletal pain including LBP showed that pain catastrophizing partially mediates the association between pain intensity and number of health care visits and fully mediated the relationship between pain intensity and condition-specific healthcare costs.
      • Lentz TA
      • Rhon DI
      • George SZ.
      Predicting opioid use, increased health care utilization and high costs for musculoskeletal pain: What factors mediate pain intensity and disability?.
      This suggests that one's appraisal of pain experiences influences health-care seeking behavior. It would be interesting to explore the extent to which various metrics of the pain experience, in addition to momentary pain intensity variability, influence one's overall pain appraisal.

      Strengths and Limitations

      The study used an EMA design involving multiple within-day ratings, over two distinct time periods. In addition, compliance with momentary pain intensity reports was high, at 80% on average for study participants. It also relied on a community sample of individuals with CLBP recruited through multiple strategies and channels. The study has some limitations, however. First, more than one third of eligible individuals declined participation in the study due to lack of time or interest or other reasons. It is thus possible that some selection bias is present – ie, selection of individuals through application of inclusion/exclusion criteria, refusals, and attrition can lead to a sample that may not be representative of all individuals with CLBP. Second, the structure and timing of administration of the momentary pain intensity reports did not allow for examination of all pain indices, such as amount of time spent in high pain. Future work could explore various components of momentary pain intensity variability in relation to pain outcomes and health care utilization. Furthermore, since outcome measures were not taken at baseline—to minimize participant burden—it was not possible to adjust for baseline levels of outcomes. Third, the use of random alarms as prompts for some of the EMAs did not allow for monitoring of time between the alarm and completion of the EMA report. However, EMA reports completed outside of the possible timeframe (9 A.M.–12 P. M.–1 P. M.–6 P. M.) were discarded. Fourth, study design required all individuals to have some level of comfort with digital technology. Efforts were made to minimize this impact, however, by providing individuals with pre-programmed iPod Touch devices, if they did not own a smartphone or other items alike. Finally, healthcare utilization was self-reported by participants and might not reflect actual use. However, the risks of under- or over-reporting were minimized by choosing a limited period of reference, namely one month. This method, however, has the advantage of capturing all types of health care utilization, including those not typically included in medico-administrative datasets.

      Conclusions and Future Directions

      Data generated from this study revealed that individuals living with CLBP experience within-week and between-week variability in pain intensity scores and the magnitude of this variability differs between individuals. Momentary pain intensity variability appears to be a stable facet of the pain experience across time. No significant predictors of pain intensity variability could be identified, and this construct was not conclusively associated with patterns of pain-related outcomes and healthcare utilization.
      These findings are important because they contribute to a growing body of literature highlighting the need to consider multiple dimensions of the pain experience beyond average pain scores.
      • Stone AA
      • Obbarius A
      • Junghaenel DU
      • Wen CKF
      • Schneider S.
      High-resolution, field approaches for assessing pain: Ecological momentary assessment.
      EMAs allow for the exploration of various dimensions of the pain experience unfolding within-day in the real world, but heterogeneity in methodologies across studies to implement them makes it difficult to compare results across studies. Increased harmonization in this methodology would improve inter-study comparability as would advancement of meta-analytic methods for synthesizing results from multilevel studies.

      Data Availability Statement

      The data on which results are based are available from the corresponding author MG Pagé. Data access requests will be granted pending proper ethics approval for a secondary data analysis. The dataset cannot be made available since participants did not provide consent for this.

      Author Contributions

      MG Pagé, L Gauvin, M-P Sylvestre, and M Choinière were involved in study design. MG Pagé recruited participants and collected data. Data analyses were conducted by R Nitulescu and A Dyachenko. MG Pagé and R. Nitulescu drafted the manuscript. All authors revised it critically and approved the final version.

      Appendix. Supplementary data

      References

        • Affleck G
        • Tennen H
        • Keefe FJ
        • Lefebvre JC
        • Kashikar-Zuck S
        • Wright K
        • Starr K
        • Caldwell DS.
        Everyday life with osteoarthritis or rheumatoid arthritis: Independent effects of disease and gender on daily pain, mood, and coping.
        Pain. 1999; 83: 601-609
        • Allen KD
        • Golightly YM
        • Olsen MK.
        Pilot study of pain and coping among patients with osteoarthritis: A daily diary analysis.
        J Clin Rheumatol. 2006; 12: 118-123
        • Ariely D.
        Combining experiences over time: The effects of duration, intensity changes and on-line measurements on retrospective pain evaluations.
        J Behav Decis Mak. 1998; 11: 19-45
        • Asmundson GJ
        • Norton PJ
        • Vlaeyen JWS
        Fear-avoidance models of chronic pain: An overview.
        in: Asmundson G.J.G. Vlaeyen J.W.S. Crombez G. Understanding and Treating Fear of Pain. Oxford University Press, New York2004: 3-24
        • Bartley EJ
        • Robinson ME
        • Staud R.
        Pain and fatigue variability patterns distinguish subgroups of fibromyalgia patients.
        J Pain. 2018; 19: 372-381
        • Bellamy N
        • Sothern RB
        • Campbell J
        • Buchanan WW.
        Rhythmic variations in pain, stiffness, and manual dexterity in hand osteoarthritis.
        Ann Rheum Dis. 2002; 61: 1075-1080
        • Bellamy N
        • Sothern RB
        • Campbell J.
        Rhythmic variations in pain perception in osteoarthritis of the knee.
        J Rheumatol. 1990; 17: 364-372
        • Bendayan R
        • Ramirez-Maestre C
        • Ferrer E
        • Lopez A
        • Esteve R.
        From acute to chronic back pain: Using linear mixed models to explore changes in pain intensity, disability, and depression.
        Scand J Pain. 2017; 16: 45-51
        • Beneciuk JM
        • Robinson ME
        • George SZ.
        Low back pain subgroups using fear-avoidance model measures: Results of a cluster analysis.
        Clin J Pain. 2012; 28: 658-666
        • Bjelland I
        • Dahl AA
        • Haug TT
        • Neckelmann D.
        The validity of the hospital anxiety and depression scale. An updated literature review.
        J Psychosom Res. 2002; 52: 69-77
        • Burstrom L
        • Jonsson H
        • Bjor B
        • Hjalmarsson U
        • Nilsson T
        • Reuterwall C
        • Wahlstrom J.
        Daily text messages used as a method for assessing low back pain among workers.
        J Clin Epidemiol. 2016; 70: 45-51
        • Button KS
        • Ioannidis JP
        • Mokrysz C
        • Nosek BA
        • Flint J
        • Robinson ES
        • Munafo MR.
        Power failure: Why small sample size undermines the reliability of neuroscience.
        Nat Rev Neurosci. 2013; 14: 365-376
        • Cane D
        • Nielson WR
        • McCarthy M
        • Mazmanian D.
        Pain-related activity patterns: Measurement, interrelationships, and associations with psychosocial functioning.
        Clin J Pain. 2013; 29: 435-442
        • Carpenter B
        • Gelman A
        • Hoffman MD
        • Lee D
        • Goodrich B
        • Betancourt M
        • Brubaker M
        • Guo J
        • Li P
        • Riddell A.
        Stan: A probabilistic programming language.
        J Stat Software. 2017; 76: 1-32
        • Childs JD
        • Piva SR
        • Fritz JM.
        Responsiveness of the numeric pain rating scale in patients with low back pain.
        Spine. 2005; 30: 1331-1334
        • Clauw DJ
        • Crofford LJ.
        Chronic widespread pain and fibromyalgia: What we know, and what we need to know.
        Best Pract Res Clin Rheumatol. 2003; 17: 685-701
        • Cleeland CS
        Measurement of pain by subjective report.
        in: Chapman C.R. Loeser J.D. Advances in Pain Research and Therapy. Raven Pr., New York, NY1989: 391-403
        • 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: 267-273
        • Cleeland CS
        • Ryan KM.
        Pain assessment: Global use of the brief pain inventory.
        Ann Acad Med Singapore. 1994; 23: 129-138
        • Cohen MP.
        Sample size considerations for multilevel surveys.
        Inte Stat Rev. 2005; 73: 279-287
        • Dagenais S
        • Caro J
        • Haldeman S.
        A systematic review of low back pain cost of illness studies in the United States and internationally.
        Spine J. 2008; 8: 8-20
        • Dobscha SK
        • Morasco BJ
        • Kovas AE
        • Peters DM
        • Hart K
        • McFarland BH.
        Short-term variability in outpatient pain intensity scores in a national sample of older veterans with chronic pain.
        Pain Med. 2015; 16: 855-865
        • Dworkin RH
        • Turk D
        • Wyrwich KW
        • Beaton C
        • 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: 105-121
        • Engel CC
        • von Korff M
        • Katon WJ.
        Back pain in primary care: predictors of high health-care costs.
        Pain. 1996; 65: 197-204
        • Gelman A
        • Carlin J.
        Beyond power calculations: assessing type S (Sign) and type M (Magnitude) errors.
        Perspect Psychol Sci. 2014; 9: 641-651
        • Gerhart JI
        • Burns JW
        • Bruehl S
        • Smith DA
        • Post KM
        • Porter LS
        • Schuster E
        • Buvanendran A
        • Fras AM
        • Keefe FJ.
        Variability in negative emotions among individuals with chronic low back pain: Relationships with pain and function.
        Pain. 2018; 159: 342-350
        • Ghaly MS
        • Handouka NM
        • Khashana AK
        • Shalaby SA
        • Fahim AE
        • Ramadan AE.
        WIthin-day variability of pain and its relation to quality of life in children with juvenile primary fibromyalgia syndrome.
        MYOPAIN. 2017; 23: 165-171
        • Gross DP
        • Ferrari R
        • Russell AS
        • Battie MC
        • Schopflocher D
        • Hu RW
        • Waddell G
        • Buchbinder R.
        A population-based survey of back pain beliefs in Canada.
        Spine (Phila Pa 1976). 2006; 31: 2142-2145
        • Guerriere DN
        • Choiniere M
        • Dion D
        • Peng P
        • Stafford-Coyte E
        • Zagorski B
        • Banner R
        • Barton PM
        • Boulanger A
        • Clark AJ
        • Gordon AS
        • Guertin MC
        • Intrater HM
        • Lefort SM
        • Lynch ME
        • Moulin DE
        • Ong-Lam M
        • Racine M
        • Rashiq S
        • Shir Y
        • Taenzer P
        • Ware M.
        The Canadian STOP-PAIN project - Part 2: What is the cost of pain for patients on waitlists of multidisciplinary pain treatment facilities?.
        Can J Anaesth. 2010; 57: 549-558
        • Guerriere DN
        • Ungar WJ
        • Corey M
        • Croxford R
        • Tranmer JE
        • Tullis E
        • Coyte PC.
        Evaluation of the ambulatory and home care record: Agreement between self-reports and administrative data.
        Int J Technol Assess Health Care. 2006; 22: 203-210
        • Harris RE
        • Williams DA
        • McLean SA
        • Sen A
        • Hufford M
        • Gendreau RM
        • Gracely RH
        • Clauw DJ.
        Characterization and consequences of pain variability in individuals with fibromyalgia.
        Arthritis Rheum. 2005; 52: 3670-3674
        • Hedeker D
        • Mermelstein RJ
        • Demirtas H.
        Modeling between-subject and within-subject variances in ecological momentary assessment data using mixed-effects location scale models.
        Stat Med. 2012; 31: 3328-3336
        • Heneweer H
        • Picavet HS
        • Staes F
        • Kiers H
        • Vanhees L.
        Physical fitness, rather than self-reported physical activities, is more strongly associated with low back pain: Evidence from a working population.
        Eur Spine J. 2012; 21: 1265-1272
        • Henschke N
        • Maher CG
        • Refshauge KM
        • Herbert RD
        • Cumming RG
        • Bleasel J
        • York J
        • Das A
        • McAuley JH.
        Prognosis in patients with recent onset low back pain in Australian primary care: Inception cohort study.
        BMJ. 2008; 337: a171
        • Hu L
        • Iannetti GD.
        Neural indicators of perceptual variability of pain across species.
        Proc Natl Acad Sci U S A. 2019; 116: 1782-1791
        • Hutchings A
        • Calloway M
        • Choy E
        • Hooper M
        • Hunter DJ
        • Jordan JM
        • Zhang Y
        • Baser O
        • Long S
        • Palmer L.
        The longitudinal examination of arthritis pain (LEAP) study: Relationships between weekly fluctuations in patient-rated joint pain and other health outcomes.
        J Rheumatol. 2007; 34: 2291-2300
        • Kahneman D
        • Fredrickson BL
        • Schreiber CA
        • Redelmeier DA.
        When more pain is preferred to less: Addiont a better end.
        Psychol Sci. 1993; 4: 401-405
        • Keefe FJ
        • Affleck G
        • France CR
        • Emery CF
        • Waters S
        • Caldwell DS
        • Stainbrook D
        • Hackshaw KV
        • Fox LC
        • Wilson K.
        Gender differences in pain, coping, and mood in individuals having osteoarthritic knee pain: A within-day analysis.
        Pain. 2004; 110: 571-577
        • 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: 309-318
        • Kongsted A
        • Kent P
        • Axen I
        • Downie AS
        • Dunn KM.
        What have we learned from ten years of trajectory research in low back pain?.
        BMC Musculoskelet Disord. 2016; 17: 220
        • Koukkari WL
        • Sothern RB
        Introducing Biological Rhythms.
        Springer, New York, NY2006: 655 (p)
        • Kratz AL
        • Murphy SL
        • Braley TJ.
        Ecological momentary assessment of pain, fatigue, depressive, and cognitive symptoms reveals significant daily variability in multiple sclerosis.
        Arch Phys Med Rehabil. 2017; 98: 2142-2150
        • Kwasnicka D
        • Kale D
        • Schneider V
        • Keller J
        • Yeboah-Asiamah Asare B
        • Powell D
        • Naughton F
        • Ten Hoor GA
        • Verboon P
        • Perski O
        Systematic review of ecological momentary assessment (EMA) studies of five public health-related behaviours: Review protocol.
        BMJ Open. 2021; 11e046435
        • Lagersted-Olsen J
        • Bay H
        • Jorgensen MB
        • Holtermann A
        • Sogaard K.
        Low back pain patterns over one year among 842 workers in the DPhacto study and predictors for chronicity based on repetitive measurements.
        BMC Musculoskelet Disord. 2016; 17: 453
        • Larue F
        • Carlier AM
        • Brasseur L
        • Colleau SM
        • Cleeland CS
        Assessing the prevalence and severity of cancer pain in France: The French Brief Pain Inventory.
        in: American Pain Society 10th Annual Scientific Meeting, New Orleans, LA1991 (abstract)
        • Lee HI
        • Song J
        • Lee HS
        • Kang JY
        • Kim M
        • Ryu JS.
        Association between cross-sectional areas of lumbar muscles on magnetic resonance imaging and chronicity of low back pain.
        Ann Rehabil Med. 2011; 35: 852-859
        • Leeuw M
        • Goossens ME
        • Linton SJ
        • Crombez G
        • Boersma K
        • Vlaeyen JW.
        The fear-avoidance model of musculoskeletal pain: Current state of scientific evidence.
        J Behav Med. 2007; 30: 77-94
        • Lentz TA
        • Rhon DI
        • George SZ.
        Predicting opioid use, increased health care utilization and high costs for musculoskeletal pain: What factors mediate pain intensity and disability?.
        J Pain. 2020; 21: 135-145
        • Lester HF
        • Cullen-Lester KL
        • Walters RW.
        From nuisance to novel research questions: Using multilevel models to predict heterogeneous variances.
        Orga ResMethods. 2019; : 1-47
        • Maas CJM
        • Hox JJ.
        Robustness issues in multilevel regression analysis.
        Stat Neerl. 2004; 58: 127-137
        • Maas CJM
        • Hox JJ.
        Sufficient sample sizes for multilevel modeling.
        Methodology. 2005; 1: 86-92
        • Madden VJ
        • Kamerman PR
        • Catley MJ
        • Bellan V
        • RUssek LN
        • Camfferman D
        • Moseley L.
        Variability in experimental pain studies: Nuisance or opportunity?.
        Br J Anaesth. 2021; 126: e61-e64
        • Mathieu JE
        • Aguinis H
        • Culpepper SA
        • Chen G.
        Understanding and estimating the power to detect cross-level interaction effects in multilevel modeling.
        J Appl Psychol. 2012; 97: 951-966
        • May M
        • Junghaenel DU
        • Ono M
        • Stone AA
        • Schneider S.
        Ecological momentary assessment methodology in chronic pain research: A systematic review.
        J Pain. 2018; 19: 699-716
        • Mun CJ
        • Suk HW
        • Davis MC
        • Karoly P
        • Finan P
        • Tennen H
        • Jensen MP.
        Investigating intraindividual pain variability: Methods, applications, issues, and directions.
        Pain. 2019; 160: 2415-2429
        • Nguyen RH
        • Turner RM
        • Sieling J
        • Williams DA
        • Hodges JS
        • Harlow BL.
        Feasibility of collecting vulvar pain variability and its correlates using prospective collection with smartphones.
        Pain Res Treat. 2014; 2014659863
        • Nouwen A
        • Cloutier C
        • Kappas A
        • Warbrick T
        • Sheffield D.
        Effects of focusing and distraction on cold pressor-induced pain in chronic back pain patients and control subjects.
        J Pain. 2006; 7: 62-71
        • O'Brien EM
        • Waxenberg LB
        • Atchison JW
        • Gremillion HA
        • Staud RM
        • McCrae CS
        • Robinson ME.
        Intraindividual variability in daily sleep and pain ratings among chronic pain patients: Bidirectional association and the role of negative mood.
        Clin J Pain. 2011; 27: 425-433
        • Parry E
        • Ogollah R
        • Peat G.
        Significant pain variability in persons with, or at high risk of, knee osteoarthritis: Preliminary investigation based on secondary analysis of cohort data.
        BMC Musculoskelet Disord. 2017; 18: 80
        • Peyron R
        • Laurent B
        • Garcia-Larrea L.
        Functional imaging of brain responses to pain. A review and meta-analysis (2000).
        Neurophysiol Clin. 2000; 30: 263-288
        • Ploghaus A
        • Tracey I
        • Gati JS
        • Clare S
        • Menon RS
        • Matthews PM
        • Rawlins JN
        Dissociating pain from its anticipation in the human brain.
        Science. 1999; 284: 1979-1981
        • R Core Team
        R: A Language and Environment for Statistical Computing.
        R Foundation for Statistical Computing, Vienna, Austria2018 (Available online at)
        • Rapoport J
        • Jacobs P
        • Bell NR
        • Klarenbach S.
        Refining the measurement of the economic burden of chronic diseases in Canada.
        Chronic Dis Can. 2004; 25: 13-21
        • Roberge P
        • Dore I
        • Menear M
        • Chartrand E
        • Ciampi A
        • Duhoux A
        • Fournier L.
        A psychometric evaluation of the French Canadian version of the Hospital Anxiety and Depression Scale in a large primary care population.
        J Affect Disord. 2013; 147: 171-179
        • Salaffi F
        • Stancati A
        • Silvestri CA
        • Ciapetti A
        • Grassi W.
        Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale.
        Eur J Pain. 2004; 8: 283-291
        • Schneider S
        • Junghaenel DU
        • Broderick JE
        • Ono M
        • May M
        • Stone II, AA.
        Indices of pain intensity derived from ecological momentary assessments and their relationships with patient functioning: An individual patient data meta-analysis.
        J Pain. 2021; 22: 371-385
        • Schneider S
        • Junghaenel DU
        • Keefe FJ
        • Schwartz JE
        • Stone AA
        • Broderick JE.
        Individual differences in the day-to-day variability of pain, fatigue, and well-being in patients with rheumatic disease: Associations with psychological variables.
        Pain. 2012; 153: 813-822
        • Schneider S
        • Junghaenel DU
        • Ono M
        • Broderick JE
        • Stone III, AA.
        Detecting treatment effects in clinical trials with different indices of pain intensity derived from ecological momentary assessment.
        J Pain. 2021; 22: 386-399
        • Schneider S
        • Junghaenel DU
        • Ono M
        • Stone AA.
        Temporal dynamics of pain: An application of regime-switching models to ecological momentary assessments in patients with rheumatic diseases.
        Pain. 2018; 159: 1346-1358
        • Snijders TAB
        Power and sample size in multilevel modeling.
        in: Everitt B.S. Howell D.C. Encyclopedia of Statistics in Behavioral Science. Wiley, Chicester2005: 1570-1573
        • Stone AA
        • Broderick JE
        • Goldman RE
        • Junghaenel DU
        • Bolton A
        • May M
        • Schneider S.I.
        Indices of pain intensity derived from ecological momentary assessments: Rationale and stakeholder preferences.
        J Pain. 2021; 22: 359-370
        • Stone AA
        • Obbarius A
        • Junghaenel DU
        • Wen CKF
        • Schneider S.
        High-resolution, field approaches for assessing pain: Ecological momentary assessment.
        Pain. 2021; 162: 4-9
        • Stone AA
        • Schwartz JE
        • Broderick JE
        • Shiffman SS.
        Variability of momentary pain predicts recall of weekly pain: A consequence of the peak (or salience) memory heuristic.
        Pers Soc Psychol Bull. 2005; 31: 1340-1346
        • Sullivan MJL
        • Bishop SR
        • Pivik J.
        The pain catastrophizing scale: Development and validation.
        Psychol Assess. 1995; 7: 524-532
        • Sullivan MJL
        • Lynch ME
        • Clark MR.
        Dimensions of catastrophic thinking associated with pain experience and disability in patients with neuropathic pain conditions.
        Pain. 2005; 113: 310-315
        • Sullivan MJL
        • Thorn B
        • Haythornthwaite JA
        • Keefe F
        • Martin M
        • Bradley LA
        • Lefebvre JC.
        Theoretical perspectives on the relation between catastrophizing and pain.
        Clin J Pain. 2001; 17: 52-64
        • Suri P
        • Rainville J
        • Fitzmaurice GM
        • Katz JN
        • Jamison RN
        • Martha J
        • Hartigan C
        • Limke J
        • Jouve C
        • Hunter DJ.
        Acute low back pain is marked by variability: An internet-based pilot study.
        BMC Musculoskelet Disord. 2011; 12: 1-9
        • Timmermans EJ
        • de Koning EJ
        • van Schoor NM
        • van der Pas S
        • Denkinger MD
        • Dennison EM
        • Maggi S
        • Pedersen NL
        • Otero A
        • Peter R
        • Cooper C
        • Siviero P
        • Castell MV
        • Herbolsheimer F
        • Edwards M
        • Limongi F
        • Deeg DJH
        • Schaap LA.
        Within-person pain variability and physical activity in older adults with osteoarthritis from six European countries.
        BMC Musculoskelet Disord. 2019; 20: 12
        • Toussaint L
        • Vincent A
        • McAllister SJ
        • Whipple M.
        Intra- and inter-patient symptom variability in fibromyalgia: Results of a 90-day assessment.
        Musculoskelet Care. 2015; 13: 93-100
        • Treister R
        • Honigman L
        • Lawal OD
        • Lanier RK
        • Katz NP.
        A deeper look at pain variability and its relationship with the placebo response: Results from a randomized, double-blind, placebo-controlled clinical trial of naproxen in osteoarthritis of the knee.
        Pain. 2019; 160: 1522-1528
        • Tupper SM
        • Rosenberg AM
        • Pahwa P
        • Stinson JN.
        Pain intensity variability and its relationship with quality of life in youths with juvenile idiopathic arthritis.
        Arthritis Care Res. 2013; 65: 563-570
        • Tyler EJ
        • Jensen MP
        • Engel JM
        • Schwartz L.
        The reliability and validity of pain interference measures in persons with cerebral palsy.
        Arch Phys Med Rehabil. 2002; 83: 236-239
        • Varey C
        • Kahneman D.
        Experiences extended across time: Evaluations of moments and episodes.
        J Behav Decis Mak. 1992; 5: 169-185
        • Vlaeyen JW
        • Linton SJ.
        Fear-avoidance model of chronic musculoskeletal pain: 12 years on.
        Pain. 2012; 153: 1144-1147
        • Vlaeyen JWS
        • Linton S.
        Fear-avoidance and its consequences in chronic musculoskeletal pain: A state of the art.
        Pain. 2000; 85: 317-332
        • Ware Jr., J
        • 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: 220-233
        • Wertli MM
        • Rasmussen-Barr E
        • Weiser S
        • Bachmann LM
        • Brunner F.
        The role of fear avoidance beliefs as a prognostic factor for outcome in patients with nonspecific low back pain: A systematic review.
        Spine J. 2014; 14: 816-836
        • Wesolowicz DM
        • Bishop MD
        • Robinson ME.
        An examination of day-to-day and intraindividual pain variability in low back pain.
        Pain Med. 2021; 22: 2263-2275
        • Williams DA
        • Gendreau M
        • Hufford MR
        • Groner K
        • Gracely RH
        • Clauw DJ.
        Pain assessment in patients with fibromyalgia syndrome: A consideration of methods for clinical trials.
        Clin J Pain. 2004; 20: 348-356
        • Zakoscielna KM
        • Parmelee PA.
        Pain variability and its predictors in older adults: Depression, cognition, functional status, health, and pain.
        J Aging Health. 2013; 25: 1329-1339
        • Zigmond AS
        • Snaith RP.
        The hospital anxiety and depression scale.
        Acta Psychiatr Scand. 1983; 67: 361-370