Introduction
Prolonged mechanical ventilation (PMV) (defined as >6 h daily ventilation for >21 days) is an increasingly common and challenging area of long-term care (Zilberberg et al., Reference Zilberberg, de Wit and Shorr2012). While the number of older adult patients treated with PMV in long-term care facilities (LTCFs) continues to rise, increasing numbers of PMV patients and their families are now opting to remain at home. The growing interest and availability of complex homecare models, incorporating increasingly sophisticated medical technology, have opened up the possibility of long-term care for PMV patients at home. Indeed, models of home hospital in general and PMV home hospital specifically have been found to be both a humane and preferred alternative to hospitalization from the patient perspective and an attractive cheaper alternative to healthcare providers (Rose et al., Reference Rose, McKim and Katz2015; Simonds, Reference Simonds2016; Nonoyama et al., Reference Nonoyama, McKim and Road2018; Jacobs et al., Reference Jacobs, Marcus and Stessman2021a). In contrast to research concerning the impact of long-term care upon informal caregivers across a range of patient populations with common chronic diseases (Adelman et al., Reference Adelman, Tmanova and Delgado2014), remarkably little is known concerning the profile and unique challenges which typify the PMV patient–informal caregiver dyad. Furthermore, despite increasing numbers of PMV patients treated at home, it is unknown if the burden upon their informal caregiver is different at home compared with patients treated in the LTCF. An accepted prerequisite for optimal patient long-term care, particularly at home, is the maintenance and safeguarding of caregiver health, competence, and resilience (Adelman et al., Reference Adelman, Tmanova and Delgado2014). Indeed, the informal caregiver may play a pivotal role in the decision concerning whether the PMV patient will be cared for in the LTCF or at home. Therefore, identifying the specific needs of informal caregivers of PMV patients is important for healthcare professionals in order to highlight potential avenues for intervention, with the ultimate goal of improving the wellbeing, health, and overall care of patient and informal caregiver alike, incorporating multidisciplinary palliative and supportive care.
The current study forms part of a recent research program, aimed at gaining a deeper understanding of different aspects of care among a cohort of PMV patients, comparing and contrasting those being cared for either at home or in the LTCF. Recently reported findings have included the patients’ clinical characteristics, complications, and mortality; patients’ mood, distressing symptomatology, overall wellbeing, and attitudes concerning the decision to be ventilated; predictors of site of care, as well as a comparison of costs to the Health Maintenance Organization (HMO) (Jacobs et al., Reference Jacobs, Marcus and Stessman2021a, Reference Jacobs, Marcus and Stessman2021b). The objective of the present study was to describe the profile of informal caregivers of PMV patients, to characterize the specific areas of caregiver strain, and compare informal caregiver strain among those caring for PMV patients at home vs. LTCF. We hypothesized that informal caregiver strain would be greatest for patients treated at home.
Methods
Study design
The current study is a cross-sectional observational study of 106 informal caregivers of 120 PMV patients (ventilation via tracheostomy ≥21 days) either treated at home or in the LTCF, enrolled between May 1st 2016 and April 31st 2018. Study methodology and findings have been recently described in detail (Jacobs et al., Reference Jacobs, Marcus and Stessman2021a, Reference Jacobs, Marcus and Stessman2021b).
Setting
We approached all PMV patients and their primary informal caregivers living within the Jerusalem area, who were insurees of the Clalit Health Services, the largest health maintenance organization in Israel, which provides a mandatory comprehensive health coverage to Israeli citizens. The vast majority of patients lived within the city, while five patients were cared for in a surrounding rural setting. The primary caregiver was identified either by the patient where possible, or by self-definition. PMV care in Jerusalem is provided either as in-patient at the Herzog Medical Center, a single specialized university-affiliated LTCF with capacity for 120 PMV adult patients at the time of the study, or at home by the Jerusalem Home Hospital (Jacobs et al., Reference Jacobs, Cohen and Rozengarten2007, Reference Jacobs, Marcus and Stessman2021a). The decision concerning the site of care depends upon patient's and/or custodian's agreement, sufficient informal/formal homecare, and medical condition. Both settings provide comprehensive multidisciplinary care, with 24/7 on-call medical and respiratory back up at home, and on-site medical, nursing, and respiratory care in the LTCF.
Participants
All adults treated with PMV (aged ≥18 and Clalit Health Service insurees) in Jerusalem, and their informal caregivers were eligible for inclusion. Informal caregivers self-identified as the most involved in caregiving. We identified a total of 123 potential PMV participants (47 at home and 76 in the LTCF), among whom one home participant and two LTCF guardians declined. Among the resulting 120 PMV participants enrolled in the study, we enrolled 106 informal caregivers: 34 caregivers/46 home PMV subjects (28 caregivers/40 communicative PMV subjects and 6 caregivers/6 non-communicative PMV subjects) and 72 caregivers/74 LTCF PMV subjects (21 caregivers/22 communicative PMV subjects and 51 caregivers/52 non-communicative PMV subjects). In brief, PMV subjects at home vs. LTCF were younger (mean age 53.8 ± 21.3 vs. 73.4 ± 16.5 years, p < 0.001), more educated, less likely to be married, had fewer children, were ventilated for a longer duration (median 51.2 months, IQR 28–199, vs. 17.9 months, IQR 9–36.9, p < 0.001). The reason for PMV was more frequently degenerative neurological conditions (amyotrophic lateral sclerosis/neuromuscular diseases) and chronic lung diseases, compared with acute neurological and medical changes (following stroke, head trauma, status post-resuscitation, and sepsis) among LTCF subjects.
Data collection
The study assistant was a nurse who specialized in PMV care and who collected data by direct structured interview of patients and caregivers and reviewed patient records. Data, coded with a unique identifier, were stored in a secured research database. Informed consent was given by the PMV participants or their custodians, and caregivers. The local ethics committee approved the study proposal.
Measures
Patients
Standardized patient questionnaire included sociodemographic, medical, and functional data. For communicative PMV subjects, attitudes toward ventilation were assessed, and symptomatology was assessed using the revised Edmonton Symptom Assessment System (r-ESAS) (a 10-item set of patient reported outcomes assessing the current feeling of tiredness, lack of appetite, pain, drowsiness, nausea, shortness of breath, overall wellbeing, depression, anxiety, and other problems), each symptom graded by patient from 0 to 10 (most severe), with total score ranging from 0 to 100 (100 most severe) (Hui and Bruera, Reference Hui and Bruera2017). The r-ESAS has been validated in 20 languages including Hebrew.
Caregivers
Caregivers were interviewed for sociodemographic data, and caregiver strain was assessed using the 13-item Modified Caregiver Strain index (Mod CSI) (Thronton and Travis, Reference Thronton and Travis2003), validated Hebrew version (Bar et al., Reference Bar, Shapira and Marcus2012). Individual items assess: (1) sleep disturbance; (2) inconvenience; (3) physical strain; (4) confinement; (5) family adjustments; (6) changes to personal plans; (7) demands on time; (8) emotional adjustments; (9) feeling upset by patient behavior; (10) feeling upset that the person being cared for has changed; (11) work adjustments; (12) financial strain; (13) feel completely overwhelmed. Each item scored 0, 1, and 2 for “never”, “sometimes”, and “always”, respectively, with maximum severity scored as 26 points.
Statistical methods
Caregiver's characteristics are presented as mean and SD or percentages, and compared between caregivers of home and LTCF patients using t-test, χ 2, or Fisher exact, as appropriate. Individual strain items among caregivers of home and LTCF patients were compared using the t-test for the continuous score and a Wald test for the binary variable. We used Ward hierarchical clustering for the 13 individual strain items to determine three separate clusters. Strain within each cluster was analyzed as both a continuous variable (calculated as the sum of all elements in that cluster, divided by the maximum score of that cluster) and a binary variable (defined as strain either sometimes or always in >1 element within the cluster). These were compared between home and LTCF patients using the t-test or the Wald test as appropriate. Multivariate linear regression models were performed to identify factors associated with caregiver strain. Caregiver and patient variables with p ≤ 0.3 in the univariate analysis were included in the multivariate model, which accordingly adjusted for caregiver age, gender, relationship to patient, site of care, and patient communicative status. An additional model examined only caregivers of communicative PMV patients for whom r-ESAS symptomatology was available (n = 47), adjusting for caregiver age, gender, caregiver relationship to patient, site of care, and patient r-ESAS score. Both models were repeated for either the total caregiver strain score, or the score in each cluster separately. Statistical analysis was performed using R 3.4.3.
Results
As shown in Table 1, the mean age of caregivers was 58.9 years old; 60.4% of caregivers were females; 82.1% were married; 29.2% were patient's spouses; and 40.6% were patient's children. Differences between the caregivers of PMV patients at home vs. LTC are shown in Table 1. The mean total caregiver strain (Mod CSI score) was 13.58 (SD 6.52), with similar levels between home and LTCF (14.30 SD 7.50 vs. 13.26 SD 6.03, p = 0.50). Similarly, no differences were observed between the 49 caregivers of communicative patients vs. the 57 caregivers of non-communicative patients (13.50 SD 7.12 vs. 13.64 SD 6.04, p = 0.93). As shown in Figure 1 and Table 2, among the 13 different individual items of the Mod CSI, the most frequent complaint was being upset (sometimes or always) that “the person I care for has changed” (93.3%), upsetting patient behavior (81.9%), feeling frequently overwhelmed (81.6%), sleep disturbance (69.5%), and emotional adjustment (67.6%). Caregivers of PMV patients at home vs. LTCF experienced significantly higher levels of physical strain (47.1% vs. 21.1%, p < 0.01), confinement (82.4% vs. 59.2%, p < 0.01), work adjustment (64.7% vs. 43.7%, p = 0.04), and financial strain (64.7% vs. 39.4%, p = 0.01), while caregivers of LTCF patients vs. caregivers of patients at home reported high levels of emotional adjustment (74.6% vs. 52.9%, p = 0.03), upset that the patient has changed (97.2% vs. 85.3%, p = 0.06), that the patient's behavior is upsetting (90.1% vs. 64.7%, p < 0.01), and feelings of being overwhelmed (88.4% vs. 67.6%, p = 0.02). For most items, findings were similar when strain items were examined either as a binary variable as presented above or as a continuous mean. Hierarchical clustering identified three clusters within the 13 items of the Mod CSI, which broadly identified domains of “burden” (physical/financial/time), “mood” (emotional adjustment/feeling overwhelmed/upsetting changes in patient/upsetting patient behaviors), and general “disturbance” (personal plans/sleep/work/family adjustment/feeling confined/inconvenient). The most frequent complaint among the informal caregivers was mood (91%), followed by disturbance (80%), and lowest for the burden cluster (32%). As shown in Figure 2, caregivers for home PMV patients displayed statistically significantly higher burden cluster (48% vs. 25%, p = 0.016), while the mood cluster was consistently higher among caregivers for LTCF PMV patients (97% vs. 79%, p = 0.024). Results were consistent for both binary and continuous measures of strain. As shown in Table 3, after adjusting for caregiver age, gender, relationship to patient, site of care (home or LTCF), and patient's communicative status, the only variable that was statistically significantly associated with caregiver strain was the site of care, with caregiving at home associated with increased burden cluster (β 1.74; CI 0.8, 2.68; p < 0.001), and reduced mood cluster (β −1.7; CI −2.85, −0.54; p = 0.004). Among the 47 caregivers of communicative patients, after adjusting for caregiver age, gender, relationship to patient, site of care, and patient r-ESAS score, caregiving at home remained associated with reduced caregiver mood cluster (β −2.28; CI −3.95, −0.61; p = 0.009). The patient r-ESAS score was statistically significantly associated with the total Mod CSI (β 0.17; CI 0.03,0.31; p = 0.018), burden cluster (β 0.04; CI 0.01,0.08; p = 0.02), mood cluster (β 0.06; CI 0.01,0.11; p = 0.031), and approached near statistical significance for disturbance cluster (β 0.07; CI −0.01,0.15; p = 0.07).
Fig. 1. ICGs of PMV patients: patterns of caregiver strain. Each of the 13 items of the Mod CSI are shown separately, according to the percentage of caregivers who reported strain always, sometimes, or never for each item.
Fig. 2. ICG strain clusters: home vs. LTCF. We compared the differences in caregiver strain clusters at home vs. LTCF. Clusters were examined as both binary (% of caregivers with strain in >1 cluster element) and continuous (mean strain score) variables, and p-values (pv) were determined. *Binary pv = 0.024; continuous pv = <0.001. **Binary pv = 0.67; continuous pv = 0.12. ***Binary pv = 0.016; continuous pv = <0.001.
Table 1. ICG of PMV patients: baseline characteristics at home and in the LTCF
LTCF, long-term care facility.
Table 2. Caregiver strain items among ICGs of PMV patients: home vs. LTCF
Table 3. Predictors of caregiver strain: multivariate models
r-ESAS, Revised Edmonton Symptom Assessment System.
Discussion
In this study of PMV patient–informal caregiver dyads, the mean age of caregivers was approximately 60 years of age, 60.4% were females, and they were more likely to be children of LTCF patients compared with spouses or parents of patients treated at home. Caregivers most frequently reported complaints concerning the negative impact upon emotional domains, followed by general disturbance, and least for overall burden. Despite the finding that the total caregiver strain score was similar irrespective of either site of care (home vs. LTCF), or patient's ability to communicate or not, nonetheless significantly different patterns of individual caregiver strain items, and clusters were observed between home compared with LTCF. Caregiving for a PMV patient at home was significantly associated with a lower level of mood strain, yet greater financial, time, and physical strain, and among communicative patients able to express their symptomatology, increased patient symptom burden was independently associated with an increase in all aspects of caregiver strain.
The small body of research among caregivers of PMV patients has frequently focused upon caregivers’ attitudes toward ventilation and their role as surrogate decision-makers (Sviri et al., Reference Sviri, Garb and Stav2009; Cox et al., Reference Cox, White and Hough2019; Lee et al., Reference Lee, Hsieh and Chang2020). In one of the few existing studies concerning caregiver burden associated with PMV in Canada, 21 informal caregivers of home PMV patients with progressive neuromuscular disease were assessed using the Caregiver Burden Inventory as well as semi-structured interview (Evans et al., Reference Evans, Catapano and Brooks2012). Measured caregiver burden was high, specifically concerning time constraints, disturbance of normal life activities, emotional, physical burden, alongside recurring themes of sense of duty, physical and emotional burden, financial strain, restricted day-to-day life, and a perceived need for improved training and education. The emotional burden of caring for patients treated with PMV was noted by Douglas and Daly (Reference Douglas and Daly2003), who found that about a half (51.2%) of caregivers of PMV patients reported symptoms consistent with some degree of depression at acute hospital discharge with the percentage dropping to 36.4% by six months after discharge. In a qualitative study of 18 caregivers of PMV patients at a specialized weaning center, caregivers reported long-term physical and psychological health changes including alteration to sleep, energy, nutrition, and body weight (Dale et al., Reference Dale, Carbone and Istanboulian2020). In a study of caregivers of chronically critically ill patients in a long-term acute care hospital, 40% of whom were ventilator-dependent, the caregivers reported anxiety, worries about their relatives’ situation, the choices they made and finances and disturbances in everyday life (Lamas et al., Reference Lamas, Owens and Nace2017). A study of 160 PMV patients and their informal caregivers in Taiwan, equally divided between home and LTCF, also described generally high levels of caregiver physical strain, sleep disturbance, and low back pain, with significantly higher levels of physical strain and somatic complaints among caregivers of home PMV patients (Liu et al., Reference Liu, Lu and Fang2017). Unlike our findings, greater financial strain was reported by informal caregivers of patients in LTCF — reflecting the need for families to cover much of the LTCF costs within their local healthcare system.
The Israeli healthcare system provides an obligatory comprehensive healthcare coverage for all citizens, which is delivered by four HMOs (Clarfield et al., Reference Clarfield, Manor and Nun2017). The HMO coverage includes the full cost of care in LTCF for PMV patients, while for home PMV patients and caregivers, the HMO care package includes home visits and telephone coverage by a dedicated team of multidisciplinary specialized health care professionals, all aspects of technical equipment, logistic backup, disposable items, as well as chronic medications. In addition to comprehensive healthcare, it is relevant to note that most older people in Israel still live at home, with only 2.1% residing in any type of long-term care setting (old age homes and skilled nursing facilities), a rate lower than that found in most Organization for Economic Cooperation and Development (OECD) countries (Shnoor and Cohen, Reference Shnoor and Cohen2021; Vinarski-Peretz and Halperin, Reference Vinarski-Peretz and Halperin2021). The low rate of institutionalization is also true among those who are disabled, among whom almost 80% live in the community (Shnoor and Cohen, Reference Shnoor and Cohen2021). It is most likely that this accomplishment is largely attributable to the active role of family members in the provision of care, which in turn reflects the core value of respect and support of the older population in both the Jewish and Muslim traditions. In addition, the Israeli National Insurance Institute funds the Community Long-term Care Insurance Law, which was implemented in 1988, as well as additional benefits for young people with special needs (Dwolatzky et al., Reference Dwolatzky, Brodsky and Azaiza2017; Gal and Bleikh, Reference Gal and Bleikh2019). Among other things, these benefits provide a range of possible service options, all related to the direct care of basic functions normally provided by families or paid part-time caregivers, such as personal care and housekeeping. Benefits might also be used to purchase laundry services and absorbent aids for incontinence. These benefits may be used to partially finance a full-time worker. However, eligibility for benefits is determined according to financial means testing, and while the majority of people are indeed entitled to assistance, even maximum benefits are generally insufficient to completely cover the cost of complex care at home. Consequently, despite significant coverage by both HMO and social security benefits, the extremely complex set of needs surrounding PMV homecare incurs a significant financial burden on both the patient and their family. This would explain the findings among our study population, who unlike those in Taiwan, reported a higher financial strain among caregivers at home rather than at the LTCF.
A large study in the Netherlands comparing aspects of homecare to the LTCF used the Older Persons and Informal Caregiver Survey Minimum DataSet (TOPICS-MDS), which comprises a set of questionnaires for collecting information concerning the physical and mental health of older adults and their primary caregivers. This study found that the subjective burden reported by the caregivers of older adults at home was similar to that of caregivers of older adults at the LTCF (not ventilated), a finding which was also observed in our study, albeit among PMV patients and their informal caregivers (Metzelthin et al., Reference Metzelthin, Verbakel and Veenstra2017). It is interesting to note that in the current study, 61.4% of the caregivers of PMV patients at the LTCF visited their relatives six to seven times a week, while 8.6% visited only once a week or less.
It has been found that caregivers overate the severity of impaired quality of life among PMV patients, compared with the patients themselves or nursing staff, measured using the 5-item European Quality of Life scale (EQ-5D) (Hung et al., Reference Hung, Yan and Fan2010). This might explain the higher levels of mood strain, which we observed among caregivers of patients in LTCF, many of whom were unable to communicate, and more frequently had suffered with acute and catastrophic changes leading to the initiation of PMV. In contrast, home PMV patients and their caregivers often witnessed a more gradual physical deterioration, along with improved levels of patient communication. The findings of our study synthesize several of the above findings, confirming higher levels of physical, time, and financial burden at home, yet lower mood strain. Greater proximity to the PMV patient at home, alongside the fact that more home patients were able to communicate, might explain the lower levels of caregiver mood strain at home. Although the familial relationship of the caregivers to the patient showed differences between home and LTCF; nonetheless, this factor was not significant after multivariable adjustment.
The dynamic relationship between the dyad of caregiver and the patient being cared for has received increasing recognition over recent years. Thus, for example, Lyons et al (Reference Lyons, Zarit and Sayer2002) assessed 63 dyads of caregivers and care recipients. The care recipients were older adults with a wide range of both physical and mental impairment. In that study, there was little disagreement between the caregiver and care recipient over the care recipient's needs, but a significant amount of variation in how much the care recipient and the caregiver agreed on their appraisals of caregiving difficulties. Caregivers appeared to report more caregiver difficulties than care recipients did. Relationship strain perceived by the caregiver, but not the care recipient, was significantly associated with this disagreement. Among cancer patients in particular, the caregiver literature has often focused upon the patient–caregiver dyad, emphasizing the bi-directional influence on health outcomes of both the patient and the caregiver (O'Hara et al., Reference O'Hara, Hull and Lyons2010; Goldzweig et al., Reference Goldzweig, Baider and Andritsch2016; Kelley et al., Reference Kelley, Kent and Litzelman2019). While beyond the scope of the current research, this is an area which may warrant attention, particularly given the extended duration of PMV in certain patient populations, their ability to communicate, and the close and long-term relationship and dependency that exists between patients and informal caregivers.
The finding in the current study that caregiver strain was associated with patients’ level of symptoms emphasizes the need for integration of palliative care in the management of PMV patients and their caregivers (Nelson and Hope, Reference Nelson and Hope2012; Lamas et al., Reference Lamas, Owens and Nace2017; Dale et al., Reference Dale, Carbone and Istanboulian2020). Core elements in such care include alleviation of symptom distress, communication with patients and families about care goals, alignment of treatment with the patients’ values and preferences, transitional planning (to LTCF or home), and family support (Nelson and Hope, Reference Nelson and Hope2012). Palliative care should be implemented during hospitalization in the intensive care unit (ICU) and continued at the LTCF or home, depending on the site of care. In the ICU setting, growing attention has been evident concerning the need for incorporation of palliative care for PMV patients (Chatterjee et al., Reference Chatterjee, Goyal and Kakkera2018; Schorr and Angelo, Reference Schorr and Angelo2018). Palliative and supportive care should also be tailored to the caregivers’ needs (Dale et al., Reference Dale, Carbone and Istanboulian2020). Palliative care interventions have been shown to decrease depression among caregivers of patients with advanced cancer (Alam et al., Reference Alam, Hannon and Zimmermann2020). Palliative care can be delivered by either a specialist palliative care consultant, by the interdisciplinary team caring for the patient in the different setting, or by a dedicated social worker. Working in unison, different members of the team will aim to engage and provide emotional support for both the patient and the caregiver, provide information, address financial concerns, as well as enable the possibility of facilitating support groups for caregivers, and if practical, for patients as well (Peres, Reference Peres2016; Muller-Kluits and Slabbert, Reference Muller-Kluits and Slabbert2020). Chaplains can provide pastoral care and emotional support for caregivers, enhance the positive aspects of caregiving, and aid caregivers in spiritual needs (Steinhauser et al., Reference Steinhauser, Olsen and Johnson2016; Cooper, Reference Cooper2018).
Limitations deserve mentioning. Firstly, our study was limited to patients and caregivers in a single LTCF or home hospital, such that the local standards and model of healthcare, as well as local cultural norms, clearly affect both patient and caregiver measures. It is worthy to note that the overwhelming majority of patients and their caregivers were cared for within the urban setting, with only five patients and caregivers living in nearby villages. Thus, our findings may not necessarily reflect the needs of caregivers in rural areas. Secondly, although the modified CSI instrument is a widely accepted and reliable tool in caregiver research; nonetheless, its use among PMV informal caregivers has not been previously described.
Conclusions and implications
In conclusion, we found that while overall caregiver strain was similar for patients treated at home or LTCF, nonetheless different patterns of strain were observed according to the site of care. Recognizing the different challenges faced by caregivers of PMV patients at home compared with LTCF is an important step toward improving healthcare for both patients and their informal caregivers alike. Within this complex dyad of the patient and the caregiver, an appreciation of their palliative care needs is critical, and a necessary step toward the integration of multidisciplinary palliative care as an essential element within standard care, tailored according to the specific care settings. Prioritizing the wellbeing and resilience of the informal caregiver is crucial, since optimal caregiver status is often the decisive factor in determining whether the PMV patient will be treated in the LTCF or at home.
Acknowledgments
We would like to acknowledge Mrs. Ruth Steiner (RN) for the many hours spent collecting data and interviewing the patients and their caregivers both in the LTCF and at home. We also acknowledge Ms. Aliza Hammerman-Rozenberg (MA) and Mr. Hananel Giat (BA) for their statistical consultation.
Author contributions
JMJ and JS had full access to all study data and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: JS, ELM, and JMJ; acquisition of data: JMJ, ELM, and JS; analysis and interpretation of data: JMJ, ELM, andJS; drafting of the first manuscript and critical revision of the manuscript for important intellectual content: JMJ, ELM, and JS; statistical analyses: JMJ, ELM, and JS.
Funding
This work was supported by the National Institute for Health Policy Research, Israel (Grant 185/2014).
Conflict of interest
None declared.
Role of the sponsors
The sponsors had no role in the study design, data analysis, preparation, or approval of the manuscript.
Transparency statement
JMJ, ELM, and JS affirm the manuscript to be an honest, accurate, and transparent account of the study, and that no important aspects of the study have been omitted.
Open access
