Cross-sectional Study of Psychosocial and Pain-related Variables Among Patients With Chronic Pain During a Time of Social Distancing Imposed by the Coronavirus Disease 2019 Pandemic

Valerie Hruschaka; K. Mikayla Flowers; Desiree R. Azizoddin; Robert N. Jamison; Robert R. Edwards; Kristin L. Schreiber


Pain. 2021;162(2):619-629. 

In This Article


Study Design and Participants

This observational cross-sectional cohort study was conducted at an academic medical center in Boston, Massachusetts, between April 28, 2020, and May 22, 2020. The study was approved by the Partners Human Research Committee (PHRC)/Institutional Review Board (IRB). Participants were recruited through Partners Healthcare Rally website, an online platform that provides the public with information about current research studies they may be eligible to participate in. In addition, we contacted patients from our previous studies that belonged to cohorts that had postsurgical pain (mastectomy, TKA, C-section, thoracotomy, and spine surgery), fibromyalgia, and/or low back pain and invited them to complete the survey. Individuals who expressed interest in participating in the study were emailed a link to a secure data entry system to complete the eligibility survey.

Individuals were eligible to participate if they self-reported persistent pain for 3 months or longer, were English speaking, were ≥18 years of age, and were Massachusetts residents under the statewide social distancing rules at the time of survey. All participants (N = 150) who completed the survey provided consent and were compensated with a $20 Amazon electronic gift code. The study survey took approximately 30 to 45 minutes to complete.


Participants completed a series of questionnaires assessing sociodemographic, clinical, psychosocial characteristics, as well as degree of social and physical distancing, pain severity, and pain interference, which included both free form questions and validated questionnaires.

Sociodemographic and Social Distancing Questions. In addition to basic demographics, participants completed the Brief Basic Needs Questionnaire. This questionnaire contains items relevant to basic needs security.[27] This measure included 1 question with a 5-point Likert scale (very hard, hard, somewhat hard, not very hard, or not hard at all) on financial resource strain in which participants were asked how challenging it is for them to afford basic necessities such as food, housing, medical care, and heating. In addition, there were 2 items assessing food insecurity, 2 items examining transportation needs, and 3 items specific to housing stability. The questionnaire has a range of 1 to 14 in which higher score suggests increased challenges in obtaining basic needs. Participants also reported number of household residents, changes in employment, and whether they received a stimulus check. Patients were also asked if they got tested for COVID-19 and also if they had a positive test result. Participants were queried regarding perceived changes in physical and social isolation during COVID-19, access to health care, use of pain self-management techniques, and methods of social communication and connection since social distancing, including social media usage. Specifically, participants were also asked to indicate a percentage (operated a slider initially set at 50% to indicate %) of both social and physically isolation, with anchors at 0% being "not isolated at all" and 100% being "completely isolated." Participants also reported utilization of various self-management techniques, physical and behavioral pain management resources, from a checklist. For those selected, they further indicated one of 3 choices regarding their utilization of the modality (it has decreased, it has stayed the same, or it has increased). Social contact was also examined by asking patients to indicate one of 6 options regarding video chat, phone calls or texting, connecting with social media, online classes, and in-person events (N/A—I have never performed this, I stopped doing this completely since I started social distancing, I do this less now than before social distancing, I do this the same amount now as before social distancing, I do this more often now than before social distancing, or I only started doing this since I started social distancing).

Pain Outcomes. The 9-item Brief Pain Inventory (BPI) questionnaire was used to assess the primary outcomes of pain severity and pain interference with general functioning. The BPI has high test–retest reliability, construct validity, and criterion validity in chronic pain populations.[21] Patients filled out this questionnaire twice consecutively, once with the text of each item referring to the time "before social distancing," and once with the text referring to "Now, since social distancing." Participants were asked to provide a 5-point Likert-scale (0 = "strongly disagree" to 4 = "strongly agree") rating to indicate how much they attributed changes in pain to different elements of social distancing, including loneliness, mood, exercise, and access to pain treatments or self-management practices.

Psychosocial Characteristics. The Pain Catastrophizing Scale (PCS) was used to examine catastrophic thinking associated with pain, including the domains of magnification, helplessness, and rumination. The PCS consists of 13 items scored from 0 to 4, resulting in a total possible score of 52, with higher scores indicating greater pain-related catastrophic thinking. The PCS has been widely validated in populations within both pain and controls.[64,65] The validated 3-item UCLA Loneliness Scale (Version 3) was used to assess loneliness and has been found to be highly reliable, both in terms of internal consistency and test–retest reliability. Each question is rated on a 3-point scale, and all items are summed to give a total score of up to 9, with higher scores indicating greater loneliness.[32,57] The 10-item extroversion and introversion subscale from the Myers–Briggs Type Indicator (MBTI) was used to assess introversion. The MBTI has been found to have construct validity. The introversion subscale has a range of 0 to 10, with higher scores indicating a preference towards higher levels of introversion.[27] The Distress Intolerance Index is a 10-item tool which was used to examine participants' ability to tolerate distress, with scores ranging from 10 to 40 with higher scores representing higher distress intolerance, and has demonstrated good internal consistency and reliability in pain patients and has been used to assess cognitive/affective responses and avoidance behaviors in response to distress.[44] The Perceived Stress Scale (PSS) is a 10-item validated questionnaire used to assess the extent patients find their lives unpredictable, uncontrollable, and overloaded,[17] with higher scores indicating greater perceived stress. The Primary Care PTSD Screen for DSM-5 (PC-PTSD-5) was used to capture the prevalence of PTSD in our sample. The PC-PTSD-5 is a 5-item measure that reflects the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5) PTSD diagnostic criteria and has demonstrated validity, with a score of 3 or higher considered as positive for PTSD.[10] In addition, validated short-form instruments from the Patient-Reported Outcome Measurement Information System (PROMIS) were used to assess anxiety, depressive symptoms, and sleep disturbance.[12,16]

Statistical Approach

Descriptive statistics were used to examine frequencies, percentages, and measures of central tendency for demographics and behavioral health characteristics. Paired t tests were used to compare participants' current vs recalled ratings of outcomes (social isolation, physical isolation, pain severity, and pain interference), and effect sizes are reported as differences in the means with 95% confidence intervals (CI). Our power analysis was based on our previous chronic pain study where patients reported a worst pain severity with mean of 6.5 with SD of 2.6.[58] Using a power of 80% and 2-sided level of significance of 5%, we calculated that to detect a small effect size[16] in a paired group comparison, 135 subjects would be needed, so we planned to recruit 150 patients in case of missing questionnaire items. To correct for multiple testing of changes related to the social distancing mandate across 4 outcomes (social isolation, physical isolation, pain severity, and pain interference), we used a Bonferroni correction and considered P < 0.0125 significant for these outcomes.

We used linear regression to assess associations between patient characteristics and both pain outcomes (pain severity and pain interference). Categorical variables were dichotomized or made continuous when possible to decrease degrees of freedom. First, we performed univariable (simple) linear regression analyses for each individual baseline characteristic and each pain outcome. Next, we conducted a multivariable (multiple) linear regression analysis for each pain outcome, including all variables that were associated with the outcome in the univariable analyses at the P < 0.1 level. Multicollinearity between candidate predictors in the multivariable linear regression models was assessed using variable inflation factor (VIF) values, with a score of VIF > 5 considered problematic for model stability.[63] VIF score ranges for predictor variables included in each model indicated minimal instability from multicollinearity amongst predictors (VIF range 1.11–3.52). All statistical analyses were performed using SPSS version 26.