The Association Between Partner Reactions to Chronic Pain and Pain-Related Outcomes: A Systematic Review.

G.F.D. completed all searches, screened abstracts, analysed all results, created tables and figures and compiled the manuscript. N.E.A. and L.P. also completed article screening. N.E.A., J.S. and P.J.M. discussed the results and extensively commented and edited the manuscript. G.F.D., N.E.A., J.S. and P.J.M. made substantial inputs to the study design.

The authors have nothing to report.

The protocol for this systematic review was registered with Prospero (ID number CRD42019120880). The review was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) (Liberati et al.  2009 ). To be included in the review, studies had to: (1) be written in English; (2) report on adults (> 18 years) experiencing persistent pain (pain for over 3 months) who are in a relationship; (3) include a measure of partner response or perceived partner response; (4) include a measure of pain‐related outcomes (i.e., disability, pain severity, quality‐of‐life, functioning); and (5) report on the relationship between partner response and pain‐related outcomes. For the purposes of this review, the term partner refers to individuals in an intimate relationship with the person experiencing chronic pain, regardless of marital status. Although some of the papers included use of the term spouse or caregiver , this review will use the term partner consistently to describe an individual in an intimate relationship with a person experiencing persistent pain. Articles where the participant group consisted solely of cancer‐related pain were excluded, as were studies of children under 18 years. No restriction was placed on the year of publication. Five databases (PubMed, PsycINFO, MEDLINE, CINAHL and EMBASE) were searched up to the date of review (July 2024) using a combination of key words of ‘spousal reaction’ OR ‘solicitous’, OR ‘operant’, OR ‘critical’, OR ‘empathy’, OR ‘validation’, OR ‘invalidation’, AND ‘chronic pain’, OR ‘persistent pain’, AND ‘marital’, OR ‘couple’, OR ‘ICP’ (individuals in chronic pain), AND ‘pain severity’, OR ‘pain adjustment’, OR ‘disability’, OR ‘activity level’. Further, MeSH terms were used such as ‘interpersonal relations,’ ‘marriage,’ ‘spouses,’ ‘disclosure,’ ‘hostility,’ ‘reinforcement, psychology,’ or ‘social support’. References and bibliographic lists of retrieved articles were also searched manually. Two of three potential reviewers (G.F.D., L.P. and N.E.A.) independently screened each title, abstract and full‐text article for eligibility. Screening and extraction disagreements were first resolved by discussion between the two reviewers and, if unresolved, were mediated by another reviewer (P.J.M.). Data was extracted and tabulated by one author (G.F.D.) using a custom‐designed form which was checked by the other authors. The following items were extracted by one author (G.F.D.): (1) author(s) names; (2) publication date; (3) study design; (4) age range of participants; (5) gender of participants; (6) country of study; (7) type of pain; (8) measure of partner reaction or perceived reaction; (9) measure of pain‐related outcome; and (10) main findings of the study. Due to the differences in methodology of included studies and the heterogeneity of the data, a meta‐analysis was not performed. Instead, a narrative synthesis of the data was conducted highlighting the magnitude of relationships between variables and the similarities and differences across included studies. Cohen's criteria were used to interpret effect sizes for correlations (small = 0.10–0.29; medium = 0.30–0.49; large ≥ 0.50) and with comparable thresholds also used for standardised regression coefficients (Cohen  1988 ). Methodological quality of each study was assessed independently by two of three potential reviewers (G.F.D., L.P. and N.E.A.) using the Downs and Black checklist (Downs and Black  1998 ) that was modified to reflect studies that were not randomised control studies. This modified version of the checklist has been used in previous studies (Andrews et al.  2012 ). A copy of the modified version is available in Appendix  S1 . The Downs and Black Checklist has adequate test–retest reliability, inter‐rater reliability and criterion validity, reported elsewhere (Downs and Black  1998 ). Disagreements were resolved by discussion between the two reviewers and, if still unresolved, with another reviewer (P.J.M.). A narrative synthesis of results was then performed.

The process of study selection is outlined in Figure  1 . Of the 5691 records originally identified through searches, 1835 duplicates were removed. Of the 3806 titles and abstracts screened, 3690 were excluded, primarily on the basis that they did not report on a relationship between partner reactions to chronic pain and pain‐related outcomes. Of the 120 full text articles reviewed, 57 were excluded. The main reasons for exclusion of full text articles were: (1) the study population focus was not a couple relationship ( n  = 18); (2) there was no partner reaction measure included ( n  = 8); and (3) the study did not report on pain‐related outcomes ( n  = 15). PRISMA flowchart of screening and selection. All extracted data from included studies are presented in Table  1 and summarised in the text below. To facilitate interpretation and appraisal of findings, studies have been grouped based on similarities in the conceptual underpinnings of the studied partner response constructs. Four broad categories were used for grouping partner responses: Operant‐based responses—behaviours understood primarily through reinforcement or punishment frameworks. Intimacy and emotion regulation‐based responses—behaviours grounded in emotional attunement and interpersonal process models. Hybrid responses—behaviours that combine instrumental support with emotional attunement. Cognitive and coping responses—internal partner processes, such as catastrophising and mindfulness, that influence the interpretation and expression of interpersonal responses to pain. Operant‐based responses—behaviours understood primarily through reinforcement or punishment frameworks. Intimacy and emotion regulation‐based responses—behaviours grounded in emotional attunement and interpersonal process models. Hybrid responses—behaviours that combine instrumental support with emotional attunement. Cognitive and coping responses—internal partner processes, such as catastrophising and mindfulness, that influence the interpretation and expression of interpersonal responses to pain. Study characteristics for all included articles. Punishing responses‐depression: Large effect Punishing responses‐fear of movement: Large effect Nil significant correlations between ambivalence over emotional expression from partner and patient depression or pain ANOVA with repeated measures demonstrated that partners where both people were low on emotional ambivalence had the lowest scores on depression Correlations: Solicitousness‐pain severity: Medium effect Solicitousness‐pain interference: Medium effect Punishing‐catastrophising: Medium effect Punishing‐affective distress: Medium effect Depression (BDI), Catastrophising (CSQ) Pain interference and severity (WHYMPI) Solicitous responses‐pain severity: small effect Solicitous responses‐ pain interference: small effect Punishing responses‐ depression: small effect Punishing responses‐pain interference: small effect Punishing responses‐catastrophising: Small effect Punishing responses‐depression: moderate effect Patient perceived criticism was assessed with a variation from Hooley and Teasedale. As well as a measure to see how critical their partner was in the last 3 h. Spouses were also asked these questions in the daily diary Hostility (C‐MHS) Also coded discussion for hostility and criticism. Criticism was coded by Smith et al. These were coded on a four‐point scale Perceived criticism‐pain intensity: Medium effect size Perceived hostility‐pain behaviours: Small effect Perceived hostility‐pain intensity: Small effect Depression (BDI) (BDI), Hostility (C‐MHS), Pain Severity and Interference (WHYMPI) Punishing‐depression: Medium effect. Punishing‐pain interference: Medium effect Punishing‐ pain severity: Small effect Hostility‐depression: Medium effect Correlations: Patient perceived criticism and catastrophising ( r  = 0.23, p  < 0.05), pain catastrophising (0.19, p  < 0.05), pain intensity (0.20, p  < 0.05), spouse reported criticism and catastrophising ( r  = 0.20, p  < 0.05). Patient perceived hostility was correlated with negative affect ( r  = 0.30, p  < 0.05), pain catastrophising ( r  = 39, p  < 0.05), pain intensity (0.20, p  < 0.05) Spouse reported hostility and catastrophising ( R  = 0.20, p  < 0.05) Concurrent associations between pain catastrophising were significantly related to spouse reported hostility In lagged analyses (items at one time predicting items 3 h later) pain catastrophising at time one was significantly associated with decreases in spouse reported criticism and hostility 3 h later Patient perceived criticism–catastrophising: Small effect Patient perceived criticism–pain catastrophising: Small effect Patient perceived criticism–pain intensity: Small effect Spouse‐reported criticism– CSQ‐CAT: Small effect Patient perceived hostility–CSQ‐CAT: Medium effect Patient perceived hostility–pain catastrophising: Medium effect Patient perceived hostility–pain intensity: small effect Spouse‐reported hostility: Small effect Negative responses‐disability: small effect Negative responses‐depression: small effect Spouse invalidation‐helplessness: Small effect Spouse invalidation‐affective distress about pain: Small effect Men M =  77 Women M =  88 For women patients: Negative partner responses were positively associated with depression ( r  = 0.21, p  < 0.01) and pain severity ( r  = 0.34, p  < 0.001) For male patients: Negative partner responses were positively associated with depression ( r  = 0.45, p  < 0.0001) and pain severity ( r  = 0.39, p  < 0.001). Distracting responses were positively associated with pain severity in males ( r  = 0.24, p  < 0.05) For women patients: Negative partner responses‐depression: Small effect Negative partner responses and pain severity: Medium effect For male patients: Negative partner responses‐depression: Medium effect Negative partner responses‐pain severity: Medium effect Distracting responses‐pain severity: small effect Nil significant results See moderation and mediation table Spouse catastrophising‐pain severity: Small effect Spouse catastrophising–pain interference: Medium effect Spouse catastrophising–depressive symptoms: Medium effect Correlations: solicitous spouse responses and pain catastrophising (0.26, p  < 0.05). Distracting spouse responses and pain catastrophising ( r  = 0.22, p  < 0.05) Negative (punishing responses) and psychological distress ( r  = 0.40, p  < 0.001) and pain severity ( r  = 0.34, p  < 0.001). In hierarchical regression, pain catastrophizing was positively and significantly related to perceived solicitous spouse responses to pain among participants with shorter pain durations Solicitous spouse responses–pain catastrophising: Small effect Distracting spouse responses–pain catastrophising: Small effect Punishing spouse responses–psychological distress: Medium effect Punishing spouse responses–pain severity: Medium effect Correlations: Solicitous and pain ( r  = 0.21, p  < 0.01), Negative responses and depression ( r  = 0.17, p  < 0.01), negative responses and catastrophising ( r  = 0.24, p  < 0.01), Solicitous and catastrophising ( r  = 0.29, p  < 0.01) Partner catastrophizing and pain ( r  = 0.17, p  < 0.01). Partner catastrophizing and depression ( r  = 0.12, p  < 0.01) Solicitous responses–pain: Small effect Negative responses–depression: Small effect Negative responses–catastrophising: Small effect Solicitous responses–catastrophising: Small effect Partner catastrophising‐pain: Small effect Partner catastrophizing‐depression: Small effect Partner validation–pain intensity: Large effect Partner invalidation–pain intensity: Medium effect size Partner validation–pain intensity: β  = −0.65 (large effect) Correlations: Solicitous and pain behaviour ( r  = 0.33, p  < 0.001); solicitous and pain intensity (0.19, p  < 0.01), solicitous and disability ( r  = 0.17, p  < 0.05) Negative responses to pain and pain behaviours ( r  = 0.28, p  < 0.001). Negative responses to pain and pain intensity ( r  = 0.19, 0.001) Negative responses and disability ( r  = 0.27, p  < 0.001), negative responses to pain and depressive symptoms ( r  = 0.41, p  < 0.001) Solicitous pain behaviour–pain behaviour: Medium effect Solicitous pain behaviour‐pain severity: Small effect Solicitous pain behaviour–disability: Small effect Negative responses to pain–pain behaviour: Small effect size Negative responses to pain–pain intensity small effect size Negative responses to pain–disability: Small effect size Negative responses to pain–depressive symptoms: Medium effect size 131 men in the sample, 131 women in the sample Within the chronic pain sample, there were 66 men and 66 women Solicitous responses–observed pain behaviour: Medium effect Punishing responses–observed pain behaviour: Small effect Correlations for patient reported MPI distraction and BDI‐II ( r  = 0.44; p  < 0.01) Spouse reported MPI—Distraction and BDI‐II ( r  = 0.36, p   0.01); Spouse reported Solicitous and BDI‐II ( r  = 0.27, p  > 0.01); Spouse reported Solicitous and social activities ( r  = 0.45 p  < 0.01), Spouse reported Solicitous and general activities (0.31, p  < 0.05) Patient‐reported MPI distraction–depression: Medium effect Spouse‐reported MPI distraction–depression medium effect Spouse‐reported punitive responses–depression: Medium effect Spouse‐reported solicitous responses–depression: Small effect Spouse‐reported solicitous responses–social activities: Medium effect Spouse‐reported solicitous responses–general activities: Medium effect Bivariate correlations: Solicitous responses with physical quality of life ( r  = 0–0.24, p  < 0.01) and PDI ( r  = 0.25, p  < 0.01). Distracting responses with physical quality of life ( r  = −0.17, p  < 0.05), PDI ( r  = 0.15, p  < 0.05) Negative (or punishing responses) and physical quality of life ( r  = −0.26, p  < 0.01), negative responses and mental quality of life ( r  = 0.26, p  < 0.001), PDI ( r  = 0.28, p  < 0.01), CES‐D ( r  = 0.28, p  < 0.01) Solicitous responses–physical quality of life: Small effect Solicitous responses–pain disability (PDI): Small effect Distracting responses–physical quality of life: Small effect Distracting responses–pain disability (PDI): Small effect Punishing (negative) responses–physical quality of life: Small effect Punishing (negative) responses–mental quality of life: Small effect Punishing (negative) responses–pain disability (PDI): Small effect Punishing (negative) responses–depressive symptoms (CES‐D): Small effect Negative responses–mental health–related quality of life: Medium effect Negative responses–disability: small effect Negative responses–depressive symptoms: Medium effect Correlations: T1 empathic responses and T1 physical limitations ( r  = 0.22; p  < 0.05); T1 solicitous responses and T1 physical limitations ( r  = 0.29, p  < 0.05) T2 Solicitous response and T2 physical limitations ( r  = 0.20; p  < 0.05) T2 Punishing responses and T2 physical limitations ( r  = 0.25, p   0.05) Empathic responses (T1)–physical limitations (T1): Small effect Solicitous responses (T1)–physical limitations (T1): Small effect Solicitous responses (T2)–physical limitations (T2): Small effect Punishing responses (T2)–physical limitations (T2): Small effect Punishing responses (T3)–physical limitations (T3): Small effect Solicitous responses–pain severity: Medium effect Solicitous responses–pain interference: Medium effect Mood and anxiety (MASQ) Pain severity (WHYMPI) Solicitous responses–pain severity: Small effect Distracting responses–pain severity: Small effect Punishing responses–depressive symptoms: Small effect No significant associations were observed between base rates of spouse validation or invalidation and pain or mood outcomes in the overall sample However, in gender analyses, greater spouse invalidation was associated with higher patient pain severity in male patient couples ( r  = 0.48, p  < 0.01). In addition, higher base rates of wives' validation were associated with greater pain severity in male patients ( r  = 0.44, p  < 0.01) Spouse invalidation–patient pain severity (male patient couples only): Medium effect Wives' validation–male patient pain severity: Medium effect Correlations: Solicitous responses and activity engagement ( r  = −0.22, p  > 0.01), pain willingness ( r  = −0.31, p   0.001). Punishing responses and activity engagement ( r  = −0.27, p  < 0.001), pain willingness ( r  = −0.29, p  < 0.001), total acceptance ( r  = −0.32, p  < 0.001). Distracting responses and pain willingness ( r  = −0.20, p  < 0.01), total acceptance ( r  = −0.19, p  < 0.01). In hierarchical regression, all responses (solicitous, distracting and punishing) negatively predicted 8.9% of variance in predicting activity engagement and 15% when predicting pain willingness Solicitous responses–activity engagement: Small effect Solicitous responses–pain willingness: Medium effect Solicitous responses–total acceptance: Medium effect Punishing responses–activity engagement: Small effect Punishing responses–pain willingness: Small effect Punishing responses–total acceptance: Medium effect Distracting responses–pain willingness: Small effect Distracting responses–total acceptance: Small effect Correlations: Encouragement received and disability ( r  = −0.19, p  < 0.01) In hierarchical regression analyses, encouragement received independently predicted lower disability ( B  = −0.27, p  < 0.05) Distracting responses–pain behaviours: Small effect Punishing responses–pain behaviours: Medium effect Correlations: SRI facilitative responses to well behaviours and pain behaviour ( r  = 0.35, p  < 0.01), SRI negative responses to well behaviours‐ and pain behaviour (r‐0.60, p  < 0.01), SRI negative responses to well behaviours and pain intensity ( r  = 0.28, p  < 0.05); SRI negative responses to well behaviours and depression CES‐D ( r  = 0.30, p  < 0.05) SRI negative responses to pain behaviours and depression( r  = 0.32, p  < 0.01); SRI solicitous responses to pain behaviours and pain behaviour ( r  = 0.46, p  < 0.05), SRI solicitous responses to pain behaviours and pain intensity ( r  = 0.38, p  < 0.01), SRI solicitous responses to pain behaviours and pain interference ( r  = 0.47, p  < 0.01) Regression models: Negative responses to well behaviours and solicitous responses to pain behaviours predicted pain behaviour Negative responses to well behaviours predicted pain intensity Solicitous responses to pain behaviours predicted pain interference Solicitous responses to pain behaviours predicted depressive symptoms Facilitative responses to well behaviours–pain behaviour: Medium effect SRI negative responses to well behaviours–pain behaviour: Large effect SRI negative responses to well behaviours‐ pain intensity: Small effect SRI negative responses to well behaviours–depression: Medium effect SRI negative responses to pain behaviours–depression: Medium effect SRI solicitous responses to pain behaviours–pain behaviour: Medium effect SRI solicitous responses to pain behaviours–pain intensity: Medium effect SRI solicitous responses to pain behaviours–pain interference: Medium effect Pain intensity: Every 3 h the person with persistent pain responded to the question ‘During the past 3 h, how intense was your pain’ using 9‐point scale Pain behaviour = Spouses rated how many pain behaviours they saw and heard every 3 h Correlations: SRI solicitous responses to pain behaviour and SIP Physical scale (0.24 p  < 0.05) Solicitous responses to pain behaviour and Pain intensity (0.25, p  < 0.05); Negative responses to pain behaviour and CES Depression (0.52 < 0.01); Negative responses to pain behaviour and SIP (0.31, p  < 0.05); negative responses to pain behaviour and Patient reported PBCL (0.43, p  < 0.01), negative responses to pain behaviour and spouse reported PBCL ( r  = 0.39, p  < 0.01) negative responses to pain behaviour and, pain intensity ( r  = 0.24, p  < 0.05); negative responses to pain behaviour and CES‐D depression ( r  = 0.52, p  < 0.52), Negative responses to pain behaviour and SIP physical ( r  = 0.31, p  < 0.05). Facilitative responses to well behaviour correlated with CES‐D Depression ( r  = −0.29, p  < 0.01) Negative responses to well behaviour correlated with CES‐D Depression ( r  = 0.28, p  < 0.01); Negative responses to well behaviour and SIP ( r  = 0.44; p  < 0.01), negative responses to well behaviour and Patient reported PBC ( r  = 0.37, p  < 0.01); negative responses to pain behaviour and Spouse reported PBC ( r  = 0.42; p  < 0.01); negative responses to pain behaviour and Pain Intensity ( r  = 0.30, p  < 0.01) Solicitous responses to pain behaviour–physical functioning: Small effect Solicitous responses to pain behaviour–pain intensity: Small effect Negative responses to pain behaviour–depression: Medium effect Negative responses to pain behaviour–SIP physical functioning: Medium effect Negative responses to pain behaviour–patient‐reported pain behaviours: Medium effect Negative responses to pain behaviour–spouse‐reported pain behaviours: Medium effect Negative responses to pain behaviour–pain intensity: Small effect Facilitative responses to well behaviour–depression: Small effect Negative responses to well behaviour–depression: Small effect Negative responses to well behaviour–physical functioning: Medium effect Negative responses to well behaviour–patient‐reported pain behaviours: Medium effect Negative responses to pain behaviour–spouse‐reported pain behaviours: Medium effect Negative responses to pain behaviour–pain intensity: Medium effect Living in Family Environments (LIFE) coding system was used to code behaviours of patients and partners from videos of tasks Partner behaviours (WHYMPI) Partner solicitous behaviour–verbal pain behaviour: Medium effect Partner solicitous behaviour–non‐verbal pain behaviour: Medium effect Partner negative behaviour–non‐verbal pain behaviour: Small effect Correlations: Solicitous and pain intensity ( r  = 0.20, p  < 0.05) In hierarchical regression, solicitous responses were positively associated with higher vulvovaginal pain intensity ( B  = 0.20, p  = 0.03) Higher facilitative responses were associated with lower pain intensity ( B  = −0.20, p  = 0.04) Correlational analyses indicated that women's perceived solicitous ( r  = 0.28, p  < 0.001), men's reported solicitous ( r  = 0.30, p  < 0.001) and women's perceived negative partner responses ( r  = 0.23, p  < 0.001) were positively associated with women's pain intensity. Women perceived negative partner responses were also associated with greater anxiety ( r  = 0.25, p  < 0.001). Daily diary multilevel analyses showed that women reported higher pain on days when they perceived greater solicitous and negative partner responses and when their male partners reported greater solicitous and lower facilitative responses Within person effects of male partner responses on women's pain: Women's pain increased on days where she perceived greater solicitous and punishing responses from her partner. Women's pain increased on days when their male partner reported greater solicitous responses and decreased on days of greater facilitative responses Women perceived solicitous responses–pain: Small effect Men reported solicitous responses–pain: Medium effect Women perceived negative responses–pain: Small effect Women perceived negative responses–anxiety: Small effect Pain intensity (NRS) Quality of life (Skindex‐29) Women perceived empathy–quality of life: Medium effect Women perceived disclosure–quality of life: Medium effect Pain intensity–partner‐reported solicitous responses: Small effect Women‐reported solicitous responses–catastrophising: Small effect Correlations: Women's pain was correlated with male report of solicitousness ( r  = 0.30, p  < 0.001) and women's report of solicitousness ( r  = 0.28, p  < 0.001) and women's perception of negative spousal response ( r  = 0.23, p  < 0.001) Within‐person (daily) analyses indicated that women reported higher pain intensity on days when they perceived greater solicitous and punishing partner responses and lower pain intensity on days characterised by greater facilitative responses. Women's anxiety symptoms were higher on days when greater solicitous responses were perceived Women's pain–male‐reported solicitousness: Medium effect Women's pain–women‐reported solicitousness: Small effect Women's pain–women's perceived negative spousal responses: Small effect Solicitous responses—pain behaviour: Medium effect Negative responses to pain behaviour—pain behaviour: Medium effect Negative responses to well behaviour—pain behaviour: Small effect Negative responses to pain behaviour—depressive symptoms: Small effect Negative responses to pain behaviour—pain intensity: Small effect Correlations: Daily negative mood correlated with punishing response (0.10, p  < 0.001), Daily negative mood correlated with solicitous response (0.05, p  < 0.05). Solicitous response with sleep quality (0.05, p  < 0.05), daily pain was correlated with solicitous responses (0.06, p  < 0.05). Daily positive mood correlated with partner‐reported empathic responses ( r  = 0.05, p  < 0.05). Daily pain correlated with partner‐reported empathic responses ( r  = 0.04, p  < 0.05) Daily negative mood–punishing responses: Small effect Daily negative mood–solicitous responses: Small effect Solicitous responses–sleep quality: Small effect Daily pain–solicitous responses: Small effect Empathic responses–positive mood: Small effect Empathic responses–daily pain: Small effect Correlational analyses indicated that reinforcement of pain behaviour ( r  = 0.28, p  < 0.001), reinforcement of well behaviour ( r  = 0.17, p  < 0.001), punishment of pain behaviour ( r  = 0.15, p  < 0.01) and punishment of well behaviour ( r  = 0.29, p  < 0.001) were all positively associated with catastrophising In linear regression analyses, reinforcement of pain behaviour predicted higher pain intensity ( β  = 0.19, p  < 0.05), reinforcement of well behaviour predicted lower pain intensity ( β  = −0.16, p  < 0.05) and punishment of well behaviour predicted higher pain severity ( β  = 0.14, p  < 0.05) Reinforcement of pain behaviour–catastrophising: Small effect Reinforcement of well behaviour–catastrophising: Small effect Punishment of pain behaviour–catastrophising: Small effect Punishment of well behaviour–catastrophising: Small effect Multilevel regression analyses demonstrated that reinforcement of pain behaviour was associated with poorer physical capacity ( β  = −0.075, p  < 0.05); Reinforcement of well behaviour was associated with increased physical capacity ( β  = 0.054, p  < 0.05) Punishment of well behaviour was associated with poorer physical capacity: β  = −0.037, p  < 0.05. Punishment of well behaviour was associated with poorer mental capacity ( β  = −0.093, p  < 0.05). Reinforcement of well behaviour was associated with greater mental capacity ( β  = 0.049, p  < 0.05). Punishment of pain behaviour was associated with reduced mental capacity ( β  = −0.040, p  < 0.05) Reinforcement of pain behaviour was also associated with greater immobility ( β  = 0.108, p  < 0.05). Punishment of well behaviour was associated with immobility ( β  = 0.149, p  < 0.05) Reinforcement of well behaviour was associated with lower immobility ( β  = −0.039, p  < 0.05) Reinforcement of pain behaviour—physical capacity: Small effect Reinforcement of well behaviour—physical capacity: Small effect Punishment of well behaviour—physical capacity: Small effect Punishment of well behaviour—mental capacity: Small effect Reinforcement of well behaviour—mental capacity: Small effect Punishment of pain behaviour—mental capacity: Small effect Reinforcement of pain behaviour—immobility due to pain: Small effect Punishment of well behaviour—immobility due to pain: Small effect Reinforcement of well behaviour—immobility due to pain: Small effect Patient‐reported empathic responsiveness–depression (Time 1): Medium effect Patient‐reported empathic responsiveness–depression (Time 2): Medium effect Partner‐reported empathic responsiveness–patient depression (Time 1): small effect MPI/SRI negative responses—pain catastrophising: Medium effect MPI/SRI negative responses—pain intensity: Medium effect Quality and intensity of pain (MPQ) Psychological symptoms (BSI) Autonomy support–pain intensity: small effect Autonomy support–depression: small effect Punishing spouse responses–psychological flexibility: Medium effect Perceived partner responsiveness–non‐reactivity: Medium effect Patient nonverbal pain expression was measured by spouse at end of the day on a Likert scale (1–3) Verbal pain expression was measured by spouse at the end of the day on Likert scale (1–3) Empathic responses–verbal expression: Medium effect Empathic responses–nonverbal expression of pain: small effect Solicitous responses–verbal expression: small effect Solicitous responses–nonverbal expression of pain: small effect Punishing responses–nonverbal expression of pain: small effect Note: Effect size magnitude refers to Cohen's criteria for correlations (small = 0.10–0.29; medium = 0.30–0.49; large ≥ 0.50) and comparable thresholds for standardised regression coefficients. ‘Data not available’ indicates that a standardised effect size could not be obtained from the data provided in the manuscript. Abbreviations: AEQ, The Ambivalence Over Emotional Expressiveness Questionnaire; AIMS, Arthritis Impact Measurement Scale 2; ASES, Arthritis Self‐Efficacy Scale (ASES); BDI, Beck Depression Inventory; BDI‐II, Beck Depression Inventory‐Second Edition; BSI, Brief Symptoms Inventory; CES‐D, Center for Epidemiologic Studies Depression; C‐MHS, Cook‐Medley Hostility Scale; CPAQ, Chronic Pain Acceptance Questionnaire; CSQ, Coping Strategies Questionnaire; DASS, Depression Anxiety Stress Scale; DCI, Dyadic Coping Inventory; FABQ, Fear Avoidance Belief Questionnaire; FFMQ, Five Facet mindfulness Questionnaire; HADS, Hospital Anxiety and Depression Scale; HCCQ, Healthcare Climate Questionnaire; MASQ, Mood and Anxiety Symptom Questionnaire; MPQ, McGill Pain Questionnaire; NRS, Numerical Rating Scale; ODQ, Oswestry Disability Questionnaire; PBCL, Pain Behaviour Checklist; PCS, Pain Catastrophizing Scale; PCS‐S, Pain Catastrophising Scale‐Spouse; PDI, Pain Disability Index; POMS, Profile of Mood States; PRQ, Pain Response Questionnaire; PSQI, Pittsburgh Sleep Quality Index; RMDQ, Roland Morris Disability Questionnaire; SF‐12, Medical Outcomes Study Short Form 12; SF‐16, Short Form‐16; SIP, Sickness Impact Profile; SPPB, Short Physical Performance Batter; SRI, Spouse Response Inventory; STAI, State–Trait Anxiety Inventory; TSK, Tampa Scale of Kinesphobia; WHYMPI, West Haven‐Yale Multidimensional Pain Inventory; WOMAC, Western Ontario McMaster Universities Index; YPAS, Yale Physical Activity Survey. A small number of studies reported on findings from mediation and moderation analyses. The results of these studies have been reported separately in Table  2 . Table showing mediation and moderation analyses within all included studies. Mediators: Perceived partner punishing responses (WHYMPI) Moderators: Pain duration and perceived social support The results of the methodological appraisals are available in Table  1 . Studies were generally of high to moderate quality, with a mean score of 77.5% (range 57.14% to 100%). Only one study (Stephens et al. 2006 ) scored < 64%: it did not provide inclusion and exclusion criteria and failed to meet the external validity criteria of the Downs and Black Checklist. Items from the Downs and Black Checklist that were consistently rated poorly across studies were item 8 (being representative of the entire population) (Burns et al.  1996 , 2013 ; Campbell et al.  2012 ; Cano  2004 ; Cunningham et al.  2012 ; Ekholm et al.  2023 ; Gauthier et al.  2011 ; Stephenson et al.  2014 ; Stragapede et al.  2024 ), item 9 (being representative from the entire population from which they were recruited) (Burns et al.  2015 , 1996 , 2013 ; Campbell et al.  2012 ; Ekholm et al.  2023 ; Gauthier et al.  2011 ; Mittinty et al.  2024 ; Stragapede et al.  2024 ) and item 10 (the staff, people and facilities where patients were treated representative of the treatment the majority of patients receive) (Burns et al.  2013 ; Campbell et al.  2012 ; Cunningham et al.  2012 ; Ekholm et al.  2023 ; Gauthier et al.  2011 ; Stragapede et al.  2024 ). Responses in this category included partner solicitousness, punishing responses, distracting partner responses and critical and hostile responses. Of the 63 studies retained, 41 (65.08%) used a measure of partner solicitousness. Most of these 41 studies used the West Haven Yale Multidimensional Pain Inventory (WHYMPI) subscale of solicitousness. Five studies used the Spouse‐Response Inventory (SRI), and one used a facial coding system to measure solicitousness. Solicitous behaviour was consistently positively correlated with pain severity and intensity in women experiencing provoked vestibulodynia (Rosen et al.  2012 , 2015 , 2014 ) and in other persistent pain conditions (Bergeron et al.  2021 ; Boothby et al.  2004 ; Burns et al.  1996 ; Fillingim et al.  2003 ; Forsythe et al.  2012 ; Issner et al.  2012 ; Lousberg et al.  1992 ; Weiss and Kerns  1995 ). Reported effect sizes in these studies were moderate to large. Additionally, solicitous partner responses were linked to more pain behaviours (small‐large effects) (Buenaver et al.  2007 ; Forsythe et al.  2012 ; Pence et al.  2008 ; Raichle et al.  2011 ; Romano et al.  2000 , 1995 ; Rosen et al.  2013 ; Schwartz et al.  2005 ; Weiss and Kerns  1995 ; Wilson et al.  2013 ) more disability (small‐medium effects) (Forsythe et al.  2012 ; Ginting, Tripp, and Nickel  2011 ; Romano et al.  1995 ; Schwartz et al.  2005 ), more pain interference (small‐medium effects) (Boothby et al.  2004 ; Buenaver et al.  2007 ; Issner et al.  2012 ) and less activity engagement (small effect) (McCracken  2005 ) and greater physical dysfunction (Romano et al.  1995 ). Small to moderate effect sizes were also found between solicitous partner responses and higher levels of depression (Gauthier et al.  2008 ) and with lower levels of pain willingness (i.e., the ability to experience pain without using avoidant or control strategies) (McCracken  2005 ). Solicitous responses to pain behaviours (as opposed to well behaviours) using the SRI were related to increased pain behaviours, pain intensity and pain interference (medium effects) (Pence et al.  2008 ). Within a male cohort (Fillingim et al.  2003 ), high levels of partner solicitousness were related to higher levels of pain ratings and disability when compared to people with chronic pain who rated their partners as low in solicitousness. Women in this study (Fillingim et al.  2003 ) who perceived higher levels of partner solicitousness demonstrated lower pain tolerance and higher levels of pain interference as well as greater use of opioid medications. One study (Cunningham et al.  2012 ) also found that greater levels of solicitous responses were associated with increased opioid use, such that higher solicitousness significantly predicted higher baseline morphine equivalent daily dosage. Several studies found significant positive associations between solicitous responses and catastrophizing cognitions by the person experiencing pain in a group with provoked vestibulodynia (small–moderate effects) (Davis et al.  2015 ; Rosen et al.  2013 ) and in a chronic pain population (small effect) (Buenaver et al.  2007 ). Reinforcement of both pain and well behaviours was associated with catastrophizing, both with small effect sizes (Sorbi et al.  2006a ). Of the 63 retained studies, 35 (55.5%) examined punishing (or negative) responses to partners with chronic pain. Most of these 35 studies ( n  = 29) used the punishing subscale of the WHYMPI. Partner punishing responses were associated with catastrophizing the next day (Martire et al.  2019 ) and with catastrophizing more generally with small to moderate effect sizes reported (Boothby et al.  2004 ; Buenaver et al.  2007 ; Davis et al.  2015 ; Stragapede et al.  2024 ). Catastrophizing from the person in pain was associated with punishment of both well and pain behaviours with small effect sizes reported (Sorbi et al.  2006a ). Punishing responses were also positively related to depression (small to large effect sizes) (Alschuler et al.  2011 ; Cano et al.  2000 ; Ginting, Tripp, and Nickel  2011 ; Ginting, Tripp, Nickel, et al.  2011 ; Kerns et al.  1990 ), affective distress (moderate effect size), (Boothby et al.  2004 ) and psychological distress (moderate effect size) (Cano  2004 ). Punishing or negative partner responses were associated with greater disability generally (Campbell et al.  2012 ; Forsythe et al.  2012 ; Ginting, Tripp, and Nickel  2011 ; Ginting, Tripp, Nickel, et al.  2011 ; Raichle et al.  2011 ). In some studies, associations between punishing or negative partner responses were examined in relation to whether responses occurred in the context of well or pain behaviour by the person in pain using the SRI. Within this literature, punishing or negative responses in relation to well behaviour by the person in pain were associated increased disability (Schwartz et al.  2005 ), immobility (Sorbi et al.  2006b ), depression (Raichle et al.  2011 ) physical dysfunction (Raichle et al.  2011 ) and pain intensity (Pence et al.  2008 ; Raichle et al.  2011 ). All these relationships were small to moderate effects. Negative partner responses to both well and pain behaviour predicted more pain behaviours (small to moderate effect sizes) (Schwartz et al.  2005 ). Punishing responses to pain behaviours were associated with more depression, physical dysfunction and pain behaviours (Raichle et al.  2011 ). These effect sizes were moderate to large. In studies that assessed punishing or negative responses more globally using the WHYMPI, such responses were associated with high pain intensity in the population of people experiencing pain (Boothby et al.  2004 ; Buenaver et al.  2007 ; Cano  2004 ; Cano et al.  2000 ; Grant et al.  2002 ; Sorbi et al.  2006b ) and in a sample experiencing vulvodynia (Rosen et al.  2015 , 2014 ) at small to moderate effects. Grant et al. ( 2002 ), found that the perception of one's partner as being punishing was associated with fluctuations in daily pain. In addition to pain severity, studies demonstrated small to moderate associations between more punishing or negative responses to a partner's pain, as per the WHYMPI and higher levels of pain interference (Buenaver et al.  2007 ; Burns et al.  1996 ). In another study, punishing behaviour was associated with daily negative mood; however, this relationship was only a small effect (Song et al.  2015 ). There was less literature examining distracting partner responses and the impact of these responses on a person's pain experience (14 studies of the 63 studies; 22.2%). Distracting responses were associated with more pain severity (small effects) (Boothby et al.  2004 ; Cano et al.  2000 ; Kerns et al.  1990 ; Mohammadi et al.  2018 ), depression (moderate effect) (Gauthier et al.  2008 ), increased pain behaviour (small effects) (Mohammadi et al.  2018 ) and catastrophizing (Cano  2004 ). In a study by Ginting, Tripp, and Nickel ( 2011 ) distracting responses were associated negatively with physical quality of life and positively with disability as measured by the PDI in men with chronic pelvic pain syndrome. Partner‐ and patient‐perceived partner criticism was examined in five studies (Burns et al.  2015 , 2013 , 2018 , 2019 ; Post et al.  2022 ), primarily using Likert‐scale self‐report measures, with one study employing observational coding (Burns et al.  2019 ). Across studies, higher levels of perceived partner criticism were associated with poorer psychological and pain‐related outcomes. Specifically, partner‐ and patient‐reported criticism was associated with greater catastrophizing and increased pain behaviours (Burns et al.  2015 ) and higher perceived partner criticism was associated with greater pain intensity (Burns et al.  2013 ). In ecological momentary assessment studies, higher momentary partner criticism was associated with greater concurrent depressed affect in the person with chronic pain, with small effect sizes reported (Post et al.  2022 ). However, partner criticism did not predict subsequent increases in depressed affect over time. Hostility was examined in a smaller number of studies and was more consistently associated with pain intensity than with psychological outcomes. Perceived partner hostility was associated with higher concurrent pain intensity, with small effect sizes reported (Burns et al.  2013 , 2018 ). In contrast, one observational study found no significant associations between coded hostility or criticism and patient depression or anxiety (Burns et al.  2019 ). Responses included in this category include empathic partner responses, validation, emotional expression and disclosure and perceived partner responsiveness. Five studies considered the impact of empathic partner responses to pain and their related outcomes. In Song et al. ( 2015 ), partner‐reported empathic responses were very weakly and positively associated with the person with pain's daily positive mood and pain. Likewise, a person with pain's perceptions of their partners' empathic responses showed a small positive association with mood and pain. In one study (Wilson et al.  2017 ), higher daily empathic spousal responsiveness predicted better 18‐month gains in patients' physical function, with the strength of this association indicating a small‐to‐moderate positive effect that reached statistical significance. People whose partners were one standard deviation above the mean in empathic responsiveness improved by 0.26 points on the 0–12 physical‐function scale, compared with only 0.05 for that one standard deviation below, representing a small‐to‐moderate longitudinal effect. Empathic accuracy, defined as the level of concordance between a person in pain's own pain rating and their partner's perception of their pain, was moderately associated with increased levels of pain interference and affective distress (Gauthier et al.  2008 ). These associations likely reflect that pain which is more intense or disruptive is easier for partners to detect accurately, rather than empathic accuracy itself exacerbating distress and increasing pain interference. Higher levels of empathic accuracy were associated with fewer distracting and punishing partner responses with small effect sizes reported (Gauthier et al.  2008 ). In a study related to provoked vestibulodynia (Rosen et al.  2014 ), there was no relationship between empathic responses and pain intensity. Three studies considered the impact of validation and invalidation on chronic pain variables (Cano et al.  2012 ; Ekholm et al.  2023 ; Leong et al.  2011 ). In couples where there was a male who was experiencing chronic pain, their female partner's validation was moderately associated with greater pain intensity and moderately associated with lower marital satisfaction, as measured by the person in pain (Leong et al.  2011 ). Additionally, reciprocal invalidation sequences (partner invalidation followed by a person in pain's invalidation) were positively associated with greater pain, whereas neutral responses to invalidation were related to lower pain. Moderate effect sizes were reported for these associations, and the same effects were not observed in female‐patient couples. In contrast, for women in pain (who experienced vulvodynia), partners' validating communication was significantly and moderately associated with the women's lower level of pain intensity (Ekholm et al.  2023 ). In further study (Cano et al.  2012 ), the presence of partner invalidation was positively associated with pain‐related helplessness and affective distress in the person with pain, reflecting small to moderate effects. One study explored the relationship between perceived partner responsiveness (PPR) and persistent pain in women who experience genito‐pelvic pain (Bergeron et al.  2021 ). In this study, perceived partner responsiveness was defined as verbal and non‐verbal responses that are seen to be validating, understanding and caring. However, no relationship between PPR and pain intensity was found in the study. One study investigated the degree of ambivalence over emotional expression from partners and its impact on pain‐related outcomes in a population of couples where the woman experienced provoked vestibulodynia (Awada et al.  2014 ). In couples where both members were low on ambivalence about emotional expression (i.e., comfortable expressing their emotions), women were significantly less depressed than other configurations of couples (e.g., where one was high on ambivalence and the other was low). Two studies investigated the relationship between disclosure and pain‐related variables. One study in with women experiencing provoked vestibulodynia, revealed no significant relationships between pain intensity and disclosure by partners (Rosen et al.  2016 ). In a study of clients with osteoarthritis in the knee (Zhaoyang et al.  2018 ), partner disclosure or ‘holding back’ (not discussing concerns with one's partner), were not associated with depression in the person experiencing persistent pain. Responses included in this category were facilitative responses, dyadic coping and partner autonomy support. Five studies investigated facilitative responses and pain‐related outcomes. Schwartz et al. ( 2005 ) developed an assessment (Spouse Response Inventory) of partner responses which included responses to both the pain and well behaviours of a partner experiencing chronic pain. In this measure, partner responses to well behaviours are noted to be facilitative in nature and were described as providing words of encouragement and support in response to the person experiencing pain participating in activities or health‐providing behaviours. In Schwartz et al. ( 2005 ), facilitative responses were moderately associated with less disability. In another study, encouragement received independently predicted lower disability, albeit at small effect sizes (McWilliams et al.  2017 ). Meanwhile, another study found that facilitative responses to well behaviours were positively related to solicitous response to pain behaviours (large effect) and negatively related to negative responses to pain behaviour as measured in the SRI in correlational analyses (moderate effect) (Pence et al.  2008 ). Further, in these same analyses, facilitative responses to well behaviour were positively associated with pain behaviour (moderate effect); however, these results did not reach significance in an analysis of variance (Pence et al.  2008 ). In Raichle et al. ( 2011 ), facilitative responses to pain behaviour were negatively related to depression, with a small effect size observed. In a sample of women experiencing provoked vestibulodynia, facilitative responses were correlated with solicitous responses as perceived by women (the person in pain) with small effect sizes (Rosen et al.  2014 ). In another study with women experiencing provoked vestibulodynia (Rosen et al.  2015 ), greater male partner facilitative responses were associated with less pain, with a small effect size. Two studies investigated the impact of dyadic coping (both negative and supportive) on psychological distress (Mittinty et al.  2020 , 2024 ). In the first study (Mittinty et al.  2020 ), the Dyadic Coping Inventory was adapted to suit a chronic pain sample. Supportive pain statements included: ‘I talk to my partner about their pain and help them change their perspective’/‘My partner helps me look at my pain differently’; or ‘When my partner is in pain, I offer help’/‘My partner helps me do things when I am in pain.’ Examples of the negative dyadic coping statements were: ‘I often ignore my partner when he/she is in pain’/‘When I am in pain, my partner withdraws.’ In this study, supportive dyadic coping, from the perspective of the person with pain, was not related to anxiety, stress or depression at any time point. In contrast, perceived negative dyadic coping was associated with higher depressive symptoms observed at baseline and a moderate effect observed at a 6‐month follow‐up. Partner‐reported supportive dyadic coping was related to lower depression in the person with pain at 3 months, although this association was small and no longer evident by 6 months. In another study (Mittinty et al.  2024 ) with people experiencing rheumatoid arthritis and their partners, supportive dyadic coping showed moderate negative associations with depression, anxiety and stress at both six‐ and 12‐month follow‐up, whereas perceived negative dyadic coping showed moderate positive associations with each of these outcomes across the same time periods. Uysal et al. ( 2017 ) found that partner autonomy support, characterised by providing choices and options when assisting a person in pain, minimising pressure and taking an empathic stance towards one's partner, was negatively correlated with depression and pain intensity, at small effect sizes. Responses included in this category were partner catastrophizing and mindfulness. Cano et al. ( 2005 ) reported that catastrophising about one's partner's pain was associated with more pain severity, pain interference and depressive symptoms in the person experiencing pain. Small to moderate effect sizes were found for these associations. These results were supported by another study where partner catastrophising was positively correlated with pain intensity and depressive symptoms in a population of women with provoked vestibuloydnia (Davis et al.  2015 ). However, both of the effect sizes reported for these associations were small. Another sample investigating endometriosis did not find a relationship between partners catastrophizing and the depressive symptoms of the women experiencing endometriosis (Stragapede et al.  2024 ). One study investigated the association between partner mindfulness variables (e.g., partners' non‐judging, non‐reactivity, acting with awareness) and pain‐related variables as well as punishing partner responses (Williams and Cano  2014 ). There were no significant relationships found between pain intensity and the partner mindfulness variables.

Use of Artificial Intelligence: The authors declare that they did not use any artificial intelligence in the development of this review article.

This review is the first to systematically examine, appraise and integrate evidence regarding studies investigating the relationship between real or perceived partner responses pain and pain‐related outcomes. Findings suggest that the type of support a partner provides may help explain differences in pain outcomes, highlighting the couple dynamic as an important focus of clinical interventions. Operant‐based responses, including solicitous, punishing, critical/hostile and distracting responses were consistently linked to poorer pain outcomes. Some of these results contradict operant theory which posits that punishing, critical and distracting responses should decrease pain‐related behaviour and disability. There are a number of possible explanations for these unexpected findings. Prior research suggests that solicitous, punishing and distracting responses may reflect different expressions of a broader partner responsiveness constructs, with both positively and negatively toned attentiveness associated with greater distress and increased pain behaviour (Papas et al.  2001 ). It is possible that drawing any attention to pain through partner responsiveness may inadvertently lead to increased pain perception and distress, helping explain why both solicitous and punishing responses are sometimes associated with higher levels of pain behaviour. Moderation and mediation analyses also indicate that the impact of partner responses may vary according to relational context, with factors such as marital satisfaction, perceived entitlement to support, depressive symptoms and daily relationship satisfaction impacting associations between operant variables and pain outcomes (Cano et al.  2009 ; Kerns et al.  1990 ; Rosen et al.  2014 ; Weiss and Kerns  1995 ). Mediation analyses further indicated that partner responses helped explain the association between cognitive or relational factors (e.g., catastrophizing, partner confidence) and pain‐related outcomes (e.g., Davis et al.  2015 ; Hemphill et al.  2016 ). Overall, these findings may indicate that partner responses do not function uniformly and their influence on pain behaviours appears contingent on the broader emotional and relational context. Further, it is possible that unexpected associations between operant‐based constructs and pain outcomes may reflect measurement and design limitations rather than weaknesses in operant theory itself. As Newton‐John ( 2002 ) noted, the reinforcing or punishing effect of a response can only be demonstrated through direct observation of changes in behaviour over time, as in functional analysis. Most studies instead relied on self‐report questionnaires that label responses by form (e.g., ‘solicitous’ or ‘punishing’) without testing whether they alter pain behaviour. Consequently, when expected associations do not appear, this may reflect limitations in operationalisation rather than failure of the theoretical model. Observational work may be useful for understanding the impact of partner responses, as defined by operant theory and pain behaviour. Using behavioural observation Romano et al. ( 1995 , 2000 ) found that spouse solicitous responses were associated with greater verbal and non‐verbal pain behaviour, independent of pain severity. Future research should therefore integrate questionnaires with direct observation or experimental manipulation to clarify whether operant constructs function as reinforcers or punishers in context. Emotion‐based responses (empathy, validation/invalidation, perceived partner responsiveness and patterns of emotional expression or withholding) were associated with both adaptive and maladaptive outcomes, suggesting that their impact may be explained by other factors. For example, partner empathy was linked to better mood (Song et al.  2015 ), physical function (Wilson et al.  2017 ) and lower levels of depression (Stephenson et al.  2014 ) in some studies however, higher daily pain and increased interference in others (Gauthier et al.  2008 ; Song et al.  2015 ). While evidence was limited, validating spousal responses also seemed to have a different effect depending on if the person in pain is male or female (Leong et al.  2011 ). Attachment theory (Bowlby  1951 ) may be a useful model for explaining differences in how individuals respond to emotion‐based responses such as empathy and validation. Individuals high in attachment anxiety typically require high levels of reassurance and closeness (Mikulincer  2003 ) and therefore may respond favourably to partner validation and empathy. Conversely, those with high attachment avoidance may prioritise autonomy and often avoid close emotional connection in relationships due to a fear of losing their independence (Mikulincer and Shaver  2010 ). Thus, they may respond less favourably to empathy or validation in the absence of other partner responses aimed at facilitating independence. Healthy participants with high attachment avoidance have been shown to respond differently to their partner's empathy in experimentally induced pain experiments compared to those with high levels of attachment anxiety (Hurter et al.  2014 ). Eight studies that investigated partner responses that encompassed both emotional and problem‐focused support that promoted wellness behaviours or autonomy were associated with better pain outcomes. Supportive dyadic coping, facilitative responses and autonomy‐supportive behaviours were associated with better psychological functioning (Mittinty et al.  2020 , 2024 ; Raichle et al.  2011 ; Uysal et al.  2017 ). Facilitative responses were also associated with lower levels of disability (Schwartz et al.  2005 ) and less pain (Rosen et al.  2015 ). Despite the lack of literature examining responses in this category, these results highlight which partner responses may be able augment pain‐related outcomes in helpful ways. Results also point to a potentially important interaction between the content and affective valence of partner responses for promoting function. Cognitive and coping‐based responses refer to internal partner processes that shape how they perceive, interpret and emotionally regulate in response to their partner's pain. Partner catastrophising generally predicted worse outcomes, including greater pain severity, pain interference and depressive symptoms (Cano et al.  2005 ; Davis et al.  2015 ; Gauthier et al.  2011 ). Associations between partner catastrophising, solicitous responses and poorer outcomes in provoked vestibulodynia (Davis et al.  2015 ) suggest that catastrophising may influence pain through a partner's overprotective behaviour where they discourage the person in pain from engaging in activity. Evidence for partner mindfulness was limited, with no significant links to pain intensity (Williams and Cano  2014 ). However, its theorised role in reducing emotional reactivity (Britton et al.  2012 ) and fostering empathic engagement (Block‐Lerner et al.  2007 ) suggests potential to support more adaptive couple coping in some relationships. While partner‐inclusive interventions can improve communication, reduce distress and enhance self‐efficacy in pain management (Keefe et al.  1996 ; Saarijärvi  1991 ; Saarijärvi et al.  1992 ), these findings demonstrate the importance of considering the relational context when designing interventions that involve partners. Interventions must move beyond simple coaching or reinforcement models, to focus on developing couples' skills to promote adaptive coping and wellness behaviours. Interventions such as Mindful Living and Relating (Cano et al.  2018 ) treat both partners as active participants with their own emotional and relational needs, rather than positioning the partner solely as a coach or supporter. The findings of this review, suggest that helping partners distinguish between unhelpful solicitousness and constructive empathy may also be an important focus in addition to behaviour that promote problem‐solving, wellness behaviours, autonomy and activity engagement. Targeting the quality, timing and function of these responses may improve the effectiveness of couple‐based interventions. Attachment theory may also be a helpful framework for the development of partner‐inclusive interventions for chronic pain. For example, Emotionally Focused Therapy for Couples (EFT) is a brief attachment‐based couple therapy that has been shown to improve martial satisfaction with distressed couples coping with a range of difficulties such as depression, trauma and physical illnesses (Beasley and Ager  2019 ; Wiebe and Johnson  2016 ). EFT helps couples identify negative interaction cycles, explore individual needs and develop more emotionally attuned and supportive patterns of relating, enabling couples to draw on a more secure attachment to manage everyday problems. (Johnson  2012 ). Through its experiential approach, EFT supports the development of new emotional experiences and interaction patterns while recognising individual differences and needs. This review has several limitations. The search was limited to English‐language articles, raising the potential for missed studies. The relatively small number of eligible studies limits the robustness of conclusions and highlights the need for larger and more methodologically rigorous designs. Conclusions are limited by the predominance of cross‐sectional designs, which cannot capture the reciprocal nature of partner responses and pain outcomes. Reliance on self‐report questionnaires also limits confidence, as operant effects require direct observation to establish reinforcement or punishment (Newton‐John  2002 ). Future research should employ longitudinal and observational methods (e.g., Romano et al.  1995 ) and incorporate attachment measures to clarify which forms of support are most adaptive and how these effects vary according to relationship quality, attachment patterns and emotional climate.

This was the first systematic review to examine associations between partner reactions and pain‐related outcomes in patients living with persistent pain. Evidence partially aligned with operant theory, with solicitous responses reinforcing poorer outcomes, but punishing and distracting responses often predicted the opposite of theoretical expectations. Findings also highlighted the potential value of alternative frameworks, such as attachment models, in explaining the variability of effects. Future research should investigate partner behaviours linked to positive pain outcomes and develop partner‐inclusive interventions that balance empathy, validation and autonomy support. Longitudinal and observational designs are needed to clarify causal mechanisms and guide the development of clinically relevant interventions.

Over the past decade there has been growing awareness of the importance of the social environment in relation to outcomes for people in pain (Bernardes et al.  2024 , 2017 ; Che et al.  2018 ; Goubert and Bernardes  2025 ). Theoretical perspectives have been formulated to explain the link between social support, particularly that provided by partners and patient pain‐related outcomes, such as disability and pain severity, in the context of persistent pain. Despite this, there has not yet been a comprehensive review that considers the relationships between partner reactions to chronic pain and pain outcomes. This review seeks to fill this gap. Fordyce ( 1976 ) was the first to consider the relevance of the social environment and the role of partner reactions for people with chronic pain. According to his operant theory of chronic pain, romantic partner responses to pain could either reinforce or discourage pain behaviours. Solicitous responses, conceptualised as statements and behaviours that provide attention or express concern to the person experiencing persistent pain, were seen as reinforcing pain behaviours, ultimately discouraging the person from engaging in activity. In contrast, punishing and distracting responses by partners were theorised to discourage pain behaviours of people experiencing persistent pain, thereby increasing function. Punishing responses were described as negative and critical in nature, while distracting responses were related to partners' efforts to divert attention and pursuits away from pain and pain‐related cognitions. Another explanatory model that has emphasised the importance of the social environment for pain outcomes is the Communal Model of Pain Catastrophizing (CCM) (Sullivan et al.  2001 ). According to this model, people living with pain may use catastrophizing as a means to gain support to deal with pain, thereby decreasing engagement in daily activities and increasing disability (Sullivan et al.  2001 ). More recently, there has been increasing empirical interest in the affective quality of interactions between couples where one person experiences persistent pain (Clark‐Sienkiewicz et al.  2018 ). That is, theories regarding intimacy and empathic responses have begun gaining traction as a way of better understanding how people manage their persistent pain. According to these theories, validation of a partner's experience of pain and their emotions increases trust, emotional regulation and closeness, whereas invalidation (i.e., ignoring a partner's pain, being hostile) may create distance between the couple and disrupt emotional regulation (Fruzzetti and Worrall  2010 ). Within the broader health literature, the relationship between social support and health outcomes has been well documented, highlighting the role of social support as a determinant of health (Wang et al.  2003 ). In the absence of any known systematic evaluation regarding the links between partner reactions and pain outcomes for people living with chronic pain, the aim of this systematic review is to provide an in‐depth analysis, appraisal and integration of the available evidence regarding the relationship between partner reactions to people with persistent pain and pain‐related outcomes. Understanding the links between partner responses and pain‐related outcomes may inform ways to optimise partner support within their relationship context, potentially improving outcomes for both.

The authors declare no conflicts of interest.

Appendix S1: ejp70236‐sup‐0001‐Appendix1.docx.