1 INTRODUCTION

Informal care is considered a cornerstone of all long-term care provision systems in Europe (EU, 2018). It is also gaining increasing recognition in international policy circles as a key issue for future welfare policy. Estimates suggest that as much as 80% of all long-term care in Europe is provided by informal carers (Frontiera et al., 2019). In the United States, approximately 43.5 million people are caregivers to a family member with a disability or illness (Gottschalk et al., 2020).

Caregiving can affect a caregiver's emotional well-being and their social activities as they often feel that they are not prepared for their role due to a wide variety of personal factors and lack of skills (Woodford et al., 2018).

A first step to developing therapeutic interventions provided by clinicians targeting caregiver preparedness is to better understand factors which influence such burden levels. This might enable the development of effective interventions to prepare and support caregivers in their role (Lieshout et al., 2020). A recent thematic analysis reporting on informal caregivers of community dwelling people with traumatic brain injury (TBI) highlighted the health systems’ reliance on informal care and the importance of supporting them: Healthcare professionals need to consider and respond to the impact that changing circumstances have on the capacity of informal caregivers to manage their workload (McIntyre et al., 2018).

The coronavirus disease 2019 (COVID-19) began spreading in China at the end of 2019 and, to date (October 2021) represents an international health emergency without precedents in terms of its health, economic and organizational effects on people's lives (WHO, 2021).

Spain has been one of the most affected countries in the world in terms of relative and absolute number of diagnosed cases. On March 13, 2020 (legally effective on March 15), the government declared a national state of alarm, with regulations targeted to facilitate diagnosis, ensure appropriate treatment of cases and reduce the spread of COVID-19, including measures of national lockdown, confinement of the population and restricted mobility (Laxe et al., 2020). While these experiences have been felt globally, the COVID-19 pandemic has introduced additional vulnerability and marginalization to those with some type of functional impairment—people with disabilities, chronic illness or frailty due to aging (Kang et al., 2020).

Evidence suggests, even in nonpandemic times, that persons with disabilities experience emotional and behavioral disturbances which are distressing for both individuals with disabilities and their close persons (Kim, 2007). There are limited data on the impact or added burden of a pandemic (or other calamities) on informal caregivers of persons with TBI or stroke.

TBI is a growing public health concern and represents the greatest contributor to disability globally among all trauma-related injuries (Dewan et al., 2018). There is consensus over the fact that the disability associated with TBI stems mainly from the cognitive, emotional, and behavioral alterations that complicate its course (Jorege et al., 2015). Furthermore, it has been recognized that these disturbances tend to be chronic, difficult to treat and, occasionally, progressive (Bavisetty et al., 2008).

Stroke is a major cause of long-term disability worldwide with increasing number of young people being affected by stroke in low- and middle-income countries (Katan et al., 2018). Behavioral disturbances have been reported as frequent complications in stroke survivors (Ferro et al., 2009). As reported in previous research, survivors of TBI very often experience symptoms ranging from irritability to aggressive outbursts (Dyer et al., 2006).

Previous research remarked that increased severity of acquired injury is related to higher levels of caregiver stress. For example, Doser and Norub (2016) demonstrated over a long period (from 3 to 6 years) that the main factor influencing the burden of spouse-caregivers was the severity of the injury of the patients. Considering a shorter period after injury (2 years) Laratta et al. (2020) recently identified injury severity being associated with a higher level of spouse-caregiver burden. Similarly, the CONOCES study concluded that both the burden borne by informal caregivers and the likelihood of them being at a high risk of burnout (at 3 and 12 months post event) were associated with the severity of the stroke (Oliva Moreno et al., 2018). Meanwhile complaints on behavioral issues (such as low self-activation, lack of initiative or apathy) have been previously reported in survivors of stroke (Caeiro et al., 2013).

Consequently, we hypothesized that during COVID-19, specific emotional (e.g., anger, irritability and impatience) or behavioral (e.g., lack of initiative and overly dependency) aspects could be identified as sources of increased stress to caregivers when recipients of care are stratified by cause or by injury severity. Such aspects may be taken as an initial step for developing personalized therapeutic interventions provided by clinical professionals (e.g., as online remote services) to informal caregivers.

Therefore, this study aims at comparing the levels of stress in a set of 10 emotional and 10 behavioral standardized items, as reported by informal family caregivers of people with chronic TBI or stroke living in the community before and during the outbreak of COVID-19. We stratified the participants in (i) informal caregivers of people with TBI (IC-TBI) and informal caregivers of people with stroke (IC-STROKE) and (ii) informal caregivers of mild, moderate or severe recipients of care.

2 MATERIALS AND METHODS 2.1 Study design

We conducted a prospective observational study analyzing informal caregivers of individuals with acquired brain injury (TBI or stroke) at chronic phase, with more than 3 years since injury, who were living in the community. The individuals with stroke or TBI were selected from the electronical health records of Institut Guttmann- Neurorehabilitation hospital in Catalonia, Spain.

Before COVID-19, it was usual practice for patients and their informal caregivers to periodically (approximately every 18 months) visit the hospital. Such in-person follow-up visits aimed to assess their medical and psychosocial status. During COVID-19, in-person follow-up visits were suspended and online assessments were implemented as part of the hospital's remote services.

Therefore, only informal caregivers of individuals with stroke or TBI registered in the hospital's electronical health records with at least one in-person visit to the hospital (before the COVID-19 outbreak) and who responded to the online assessment (during COVID- 19) were included in this study. Recruitment period for the online assessment was between June 2020 and April 2021. This study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Guidelines (STROBE, 2021).

2.2 Participants

Eligible participants were informal caregivers of individuals with the diagnosis of stroke or TBI (at the moment of injury were aged ≥18), living in the community with electronical health records including complete data.

Individuals with stroke or TBI were excluded for the following reasons: diagnosis of concomitant comorbidity (e.g., spinal cord injury, brain tumors and anoxia) and a previous history of another disabling condition or not living in the community.

To satisfy the informal caregiver criteria, the participant had to be a first-degree relative (parent, sibling or spouse) who had spent at least 12 months looking after the person with TBI or stroke and who had the following characteristics: (i) is responsible for and makes decisions concerning the care of the person with TBI or stroke, regardless of whether he/she lives with the latter and/or (ii) spends most of the time, daily or weekly, accompanying/looking after the person with TBI or stroke. The informal caregiver might look after no more than one dependent person (De Arroyabe et al., 2013).

Every eligible participant was contacted as part of the routine clinical follow-up. Consequently, participants completed the online measures analyzed in this study as part of a virtual visit involving other assessments, not analyzed in this study. Such assessments include, for example the Patient Competency Rating Scale (PCRS) to assess self-awareness following brain injury.

Informal caregivers answered their follow-up online assessments within 10 days since contacted; therefore, all assessments were completed between June 2020 and April 2021.

2.3 Remote follow-up service and the Head Injury Behavior Scale

The online assessments were implemented during the COVID-19 lockdown in order to provide a remote follow-up service. Each participant received the online assessments by an SMS message sent to the participant's mobile phone. This SMS was sent by the professional from the hospital's Psychosocial Unit in charge of the online follow-up. The online assessments include the same questionnaires and measurements as when participants were assessed (before the COVID-19 outbreak) in-person during their follow-up visits.

In this study, we focused on one of such measurements: the Head Injury Behavior Scale (HIBS) (Godfrey et al., 2003). The Psychosocial Unit of the hospital routinely performs HIBS follow-up every 3 years.

The HIBS was implemented to assess distress in caregivers of persons with acquired brain injury. It is a 20-item scale describing common psychological problems that occur following brain injury. For each of the items in the scale, the caregiver is asked, “Is the behavior a problem?” (yes/no) and “How much distress does this problem cause?.” Distress ratings are recorded for the identified items using a four-point scale: 1 (no distress); 2 (mild distress); 3 (moderate distress) and 4 (severe distress). The HIBS includes two subscales: Emotional Regulation and Behavioral Regulation. The Emotional Regulation subscale assesses behaviors reflecting impaired emotional control and occurring during interactions with caregivers (e.g., sudden/rapid mood changes, depression, anger and aggression). The Behavioral Regulation subscale assesses problems that carry less emotional valence for caregivers (e.g., lack of control over behavior, inappropriate behavior for social situations, lack of motivation, or lack of interest in doing things) (Godfrey et al., 2003).

The HIBS has been previously used in assessing emotional and behavioral problems in individuals with chronic TBI (Villalobos et al., 2020) (Marsh et al., 2006) or stroke (Orfei et al., 2009). In this study, we used the version of the HIBS adapted to the Spanish language (De Arroyabe et al., 2013). A description of the 20 HIBS items is presented in Table S1.

2.4 Clinical and demographic variables

Demographics (age, sex, years of education and marital status), as well as time since onset of the injury, were collected. Stroke severity of the injury was assessed within 24 h before acute phase discharge by the National Institutes of Health Stroke Scale (NIHSS) and stratified as follows: mild (1–4), moderate (NIHSS 5–14) or severe (NIHSS ≥15) (Williams et al., 2000).

The Glasgow Coma Scale (GCS) (Teasdale et al., 1974) was used to assess TBI severity at admission and was stratified as mild (14–15), moderate (9–13) or severe (3-8) (Mena et al., 2011). All data were collected from the hospital’ electronical health records of the individuals with TBI or stroke.

2.5 Statistical analyses

All statistical analyses were performed in R-v3.6.3 (64 bits) (The R project, 2021), and level of significance was set at p = .05. Descriptive statistics were used for demographic and clinical characteristics of individuals with stroke or TBI. Responses to HIBS were compared before and during the COVID-19 outbreak and during it using the McNemar's test for nominal data (2 × 2 contingency tables with a dichotomous trait, caused stress [yes, no]), the Wilcoxon ranked test or paired t-test when appropriate (HIBS total, emotional and behavior subtotals). The Shapiro Wilk test was used to assess normality, Levene test for homogeneity of variances and Cohen's d to assess effects sizes (small effect size [d = 0.1], medium [d = 0.3] and large effect size [d = 0.5]).

2.6 Ethical considerations

The study follows the Declaration of Helsinki and was approved by the Ethics Committee of Clinical Research of Institut Guttmann- Neurorehabilitation hospital. The participants were anonymized and nonidentifiable.

3 RESULTS

The initial number of eligible participants, considering the criteria described in Section 2.2. was n = 191, four of them had difficulties in understanding or expressing in the Spanish language as detected by members of the research team, which prevented him/her from being able to collaborate. One hundred and eighty-seven participants received the SMS in their mobile phones, 24 (12.8%) were not a first-degree relative (parent, sibling or spouse) who had spent at least 12 months looking after the person with TBI or stroke, 3 (1.6%) were looking after more than one dependent person, 11 (5.8%) did not have an available Functional Independence Measure (FIM) in person assessment and 27 (14.4%) did not complete the online HIBS assessment.

Consequently, a total of 122 informal caregivers participated in the study, 76 (62.3%) were IC-STROKE and 46 (37.7%) were IC-TBI.

Table 1 presents the demographics and clinical characteristics of the individuals with TBI and stroke. The mean age at the moment of online assessment was 55 (11) years, and 70% of the individuals were men.

TABLE 1. Demographics and clinical characteristics of all individuals with stroke (n = 76) or traumatic brain injury (n = 46) included in the study Variables Total participants (N = 122) Sex (%) Male 71.3 Female 28.7 Age at the moment of the online assessment, mean (SD) 55.49 (11.82) Age <65 at the moment of the online assessment (%) 78.7 Age ranges at the moment of the online assessment (%) 18–30 1.6 31–45 24.6 46–60 38.5 61–75 32.0 76+ 3.3 Time (in days) since lockdown (March 14) to the online assessment, mean (SD) 191 (73) Time (in years) since closest in-person assessment to lockdown (March 14), mean (SD) 1.74 (0.88) Time (in years) since injury to online assessment, mean (SD) 8.47 (7.48) Age at the moment of injury in years, mean (SD) 44.74 (14.47) Injury origin, n (%) Traumatic brain injury 46 (37.7%) Stroke 76 (62.3%) Type of stroke, n (%) Hemorrhagic stroke 36 (29.5%) Ischemic stroke 40 (32.8%) Severity of stroke (NIHSS) (%) Mild 13.8 Moderate 43.0 Severe 43.2 Severity of traumatic brain injury (GCS) (%) Mild 15.6 Moderate 11.1 Severe 73.3 FIM in-person assessment, mean (SD) Cognitive FIM 26.48 (8.82) Motor FIM 61.32 (26.16) Total FIM 87.81 (33.30) Time in years since FIM in-person assessment to the online assessment, mean (SD) 0.81 (1.31) Years of education at the moment of the online assessment (%) Read and write (<2 years) 6.6 Primary (2–5 years) 38.5 Secondary (6–12 years) 29.5 Higher (>13 years) 25.4 Marital status. Married (%) 64.8 Location where respondents were living at the moment of answering the online assessment (%) Barcelona 72.8 Girona 10.7 Tarragona 9.9 Lérida 6.6 Note: All characteristics are presented as percentages (%), unless otherwise indicated. Abbreviations: FIM, Functional Independence Measure; NIHSS, National institute of Health Stroke Scale; GCS, Glasgow Comma Scale.

In relation to the injury severity, 70% of the individuals with TBI were severe, and 86% of the individuals with stroke were moderate or severe. The mean time since lockdown (March 14) to the online assessment was 191 (73) days. The mean time since the closest in-person assessment to lockdown (March 14) was 1.74 (0.88) years. The mean time since injury to online assessment was 8.47(7.48) years.

3.1 Stratification by IC-TBI or IC-STROKE

We then compared the emotional regulation subtotal, the behavioral regulation subtotal and the total HIBS score for IC-TBI (N = 46, 37.7%) and IC-STROKE (N = 76, 62.3%), as presented in Table 2.

TABLE 2. Paired comparisons for precoronavirus decease 2019 (COVID-19) assessments and during COVID-19 assessments for Head Injury Behavior Scale (HIBS) emotional, behavioral and total Stratification Subscale COVID-19 Median Mean (SD) SE Mean CI mean 0.95% Shapiro-Wilk normality test (p) Wilcoxon Signed rank test (W) Wilcoxon Signed rank test (p) Effect size (d) All (N = 122) Emotional Before 5.50 8.09 (8.47) 0.77 1.52 <.0001 1801 <.001 −0.317 During 9.00 11.25 (10.71) 0.97 1.92 <.0001 Behavioral Before 6.00 8.47 (8.37) 0.76 1.50 <.0001 2060 .004 −0.256 During 9.00 10.82 (9.44) 0.85 1.69 <.0001 Total Before 13.50 16.56 (15.42) 1.40 2.76 <.0001 1853.5 <.001 −0.323 During 18.00 22.07 (18.63) 1.69 3.34 <.0001 IC-TBI (N = 46) Emotional Before 9.00 10.00 (9.30) 1.37 2.76 <.0001 340 .109 −0.236 During 9.50 12.02 (10.36) 1.53 3.08 <.0001 Behavioral Before 8.00 9.91 (8.20) 1.21 2.44 <.0001 305.5 .027 −0.325 During 11.50 12.85 (9.40) 1.39 2.79 <.0001 Total Before 17.00 19.91 (15.79) 2.33 4.69 <.0001 305.5 .027 −0.325 During 21.50 24.87 (17.65) 2.60 5.24 <.0001 IC-STROKE (N = 76) Emotional Before 4.00 6.93 (7.76) 0.89 1.77 <.0001 584.5 .001 −0.365 During 7.00 10.79 (10.97) 1.26 2.51 <.0001 Behavioral Before 5.00 7.59 (8.40) 0.96 1.92 <.0001 797 .072 −0.206 During 8.00 9.59 (9.32) 1.07 2.13 <.0001 Total Before 9.00 14.53 (14.92) 1.71 3.41 <.0001 663 .004 −0.323 During 15.00 20.38 (19.11) 2.19 4.37 <.0001 Abbreviations: IC-TBI, informal caregivers of people with TBI; IC-STROKE, informal caregivers of people with stroke.

When considering emotional regulation subtotals, no significant differences were found in IC-TBI (p = .109), but differences were significant for IC-STROKE (p = .001) and with the highest effect size d = −0.365

Meanwhile, the opposed is observed when considering behavioral regulation subtotals, no significant differences were found in IC-STROKE (p = .072), but differences were significant for IC-TBI (p = .027).

We then compared individuals with stroke with individuals with TBI considering demographics and clinical variables, as presented in Table 3.

TABLE 3. Comparison of demographics and clinical characteristics of individuals with stroke (n = 76) and traumatic brain injury (n = 46) Variables Stroke (N = 76) TBI (N = 46) p Sex (%) .935 Male 71.1 71.7 Female 28.9 28.3 Age at the moment of the online assessment, mean (SD) 59.73 (9.03) 48.49 (12.62) <.001 Age >65 at the moment of the online assessment (%) 26.3 13.0 .083 Age ranges at the moment of the online assessment (%) <.001 18–30 0 4.3 31–45 9.2 50.0 46–60 47.4 23.9 61–75 39.5 19.6 76+ 3.9 2.2 Time (in days) since lockdown (March 14) to the online assessment, mean (SD) 200 (71) 178 (73) .120 Time (in years) since closest in-person assessment to lockdown (March 14), mean (SD) 1.73 (0.92) 1.76 (0.83) .334 Time (in years) since injury to online assessment, mean (SD) 6.74 (7.30) 11.34 (6.93) <.001 Age at the moment of injury in years, mean (SD) 50.71 (9.25) 34.89 (16.16) <.001 Severity (%) .642 Mild 13.8 15.6 Moderately severe and severe 86.2 84.4 FIM in-person assessment, mean (SD) Cognitive FIM 28.13 (8.04) 23.76 (9.45) .008 Motor FIM 63.50 (23.01) 57.73 (30.61) .719 Total FIM 91.63 (29.09) 81.50 (38.80) .378 Time in years since FIM in-person assessment to the online assessment, mean (SD) 0.80 (1.20) 0.82 (1.50) .395 Years of education at the moment of online assessment (%) .130 Read and write (<2 years) 9.2 2.2 Primary (2–5 years) 35.5 43.5 Secondary (6–12 years) 34.2 21.7 Higher (>13 years) 21.1 32.6 Marital status. Married (%) 81.6 37.0 <.001 Support in ADLs