Factors influencing long-term care facility performance during the COVID-19 pandemic: a scoping review

Abstract

Background

The COVID-19 pandemic wreaked havoc on long-term care facilities (LTCFs). Some LTCFs performed better than others at slowing COVID-19 transmission. Emerging literature has mostly described infection prevention and control strategies implemented by LTCFs during the pandemic. However, there is a need for a comprehensive review of factors that influenced the performance of LTCFs in containing COVID-19 spread to inform public health policy.

Objective

To build on the existing literature, we conducted a scoping review of factors that influenced LTCF performance during the COVID-19 pandemic using a multidimensional conceptual framework of performance.

Methods

We followed the Joanna Briggs Institute’s methodology for scoping reviews. We queried CINAHL, MEDLINE (Ovid), CAIRN, Science Direct, Scopus, and Web of Science for peer-reviewed literature in English or French published between January 1^st, 2020 and December 31^st, 2021. Retrieved records were screened for context (COVID-19 pandemic), population (LTCFs), interest (internal and external factors that influenced LTCF performance), and outcomes (dimensions of performance: equity, accessibility, reactivity, safety, continuity, efficacy, viability, efficiency). Descriptive characteristics of included articles were summarized. Dimensions of performance as well as internal (e.g., facility characteristics) and external (e.g., visitors) factors identified to have influenced LTCF performance were presented.

Results

We retained 140 articles of which 68% were classified as research articles, 47% originated in North America, and most covered a period between March and July 2020. The most frequent dimensions of performance were “efficacy” (75.7%) and “safety” (75.7%). The most common internal factors were “organizational context” (72.9%) and “human resources” (62.1%), and the most common external factors were “visitors” (27.1%) and “public health guidelines” (25.7%).

Conclusions

Our review contributes to a global interest in understanding the impact of the COVID-19 pandemic on vulnerable populations residing and working in LTCFs. Though a myriad of factors were reported, a lack of randomized controlled trials makes it impossible to establish causality between the identified factors and LTCF performance. The use of a multidimensional framework can be recommended to evaluate healthcare system performance not merely in terms of efficacy and safety, but alongside other critical dimensions such as efficiency and equity.

Trial registration

Research Registry ID: researchregistry7026

Background

On March 11^th 2020 the World Health Organization declared COVID-19 a pandemic in response to the alarming propagation of the highly transmissible and virulent virus [1, 2]. Residents of long-term care facilities (LTCFs) around the world bore the brunt of the pandemic, experiencing high morbidity and mortality [3]. As of July 2020, LTCF residents represented 47% of COVID-19-related deaths across 12 countries within the Organization for Economic Co-operation and Development (OECD), while in Canada this number approached 80% [4]. In the Canadian province of Québec, between March and July of 2020, more than 90% of all deaths due to COVID-19 occurred among adults over the age of 70, of whom 64% resided in LTCFs [5]. As of February 2021, an average of 41% of COVID-19-related deaths across 25 OECD countries occurred among LTCF residents, with the highest percentage of 75% recorded in Australia, compared to 39% in the United States of America (USA) and 59% in Canada [6]. As of July 2022, 43% of Canada’s overall deaths due to COVID-19 were among LTCF residents and staff [7].

Amidst the crisis, some LTCFs performed better than others, slowing COVID-19 transmission more effectively and, as a consequence, experiencing fewer cases and lower mortality [8]. It is critical to understand the reasons behind this differential LTCF performance to inform the prevention of and response to outbreaks in the future. Literature describing COVID-19 in LTCFs has primarily focused on reporting disease outcomes [3], identifying effective infection prevention and control (IPC) strategies [8, 9], and formulating IPC guidance [10]. A growing number of publications have described factors that influenced the effectiveness or failure of implemented IPC strategies to prevent or mitigate COVID-19 outbreaks. Factors associated with more successful control of COVID-19 spread in LTCFs include: strong partnerships between LTCFs, laboratory services, hospitals, and public health officials; greater funding with more care hours per resident; proactive leadership that enforced IPC measures; access to a multidisciplinary IPC team; monitoring of human and material resources; universal testing of residents and staff; and visitor restrictions [11,12,13,14,15,16,17,18,19,20,21,22,23,24]. Less effective responses have been linked to: inadequate national long-term care policies; a lack of integration between LTCFs, hospitals, and the public health sector; chronic underfunding; shortages of material resources and IPC trained staff; and the failure to detect and manage asymptomatic COVID-19 cases [19, 24,25,26,27,28,29,30,31,32,33]. In addition, the magnitude of COVID-19 outbreaks in LTCFs has been positively correlated with larger sized facilities, urban location, lower quality ratings, higher occupancy rooms, a greater proportion of racial/ethnic minority residents, and fewer care hours per resident [34,35,36,37,38,39,40,41,42].

Given the breadth and heterogeneity of the existing literature, there is a need for a comprehensive review of factors that influenced LTCF performance during the COVID-19 pandemic. Several reviews have been conducted to elucidate these factors. Gmehlin and Munoz-Price [43] offered an overview of the epidemiology, clinical manifestations, and interventions implemented to combat COVID-19 by LTCFs, underscoring the effectiveness of testing and subsequent cohorting of both residents and staff. Similarly, Dykgraaf et al.’s [8] rapid narrative review of strategies that helped limit COVID-19 spread in LTCFs stressed the importance of serial universal screening of residents and staff. A narrative review by Calcaterra et al. [44] synthesized IPC measures used by LTCFs to manage COVID-19 in several Asian countries, noting similarity of measures used elsewhere in the world, such as visitor restrictions. The authors speculated that the effective response demonstrated by LTCFs in these Asian countries may have been due to better integration into the healthcare system and greater preparedness informed by previous experience. In their scoping review, Palacios-Ceña et al. [45] synthesized the qualitative evidence relating to the experiences of LTCF staff, residents, and residents’ families during the COVID-19 pandemic. Themes that emerged from the data underscored LTCFs’ challenges with managing material and human resources, the emotional toll exerted by the pandemic on residents and staff, as well as the innovative solutions and adaptive strategies implemented by LTCF staff in response to the hardship. Furthermore, Frazer et al.’s [46] rapid systematic review linked the likelihood of experiencing a COVID-19 outbreak in a LTCF to a facility’s bigger size, for-profit status, higher crowding index, greater number of employees, and staff working at multiple facilities. In Konetzka et al.’s [47] systematic review of facility characteristics associated with COVID-19-related outcomes, a larger facility size and higher disease prevalence in the community emerged as the strongest predictors of COVID-19 cases and deaths. A more comprehensive review of factors that drove COVID-19 transmission in LTCFs was performed by Giri et al. [48]. Here, identified factors included: asymptomatic COVID-19 transmission, factors linked to the facilities (e.g., for-profit status), residents (e.g., malnutrition), or staff (e.g., high staff turnover), as well as external factors (e.g., underfunding).

The aforementioned studies offered insight into a range of factors that influenced the performance of LTCFs during the COVID-19 pandemic. However, no studies have applied a multidimensional approach to assessing LTCF performance during this challenging period. Accordingly, to examine LTCF performance during the first 2 years of the COVID-19 pandemic and more broadly, we performed a scoping review using a multidimensional conceptual framework of performance within healthcare systems. For the purposes of this scoping review, the performance of a healthcare system was defined as the system’s ability to achieve its objectives in relation to the population’s health as a function of the following dimensions of performance: equity, accessibility, reactivity, safety, continuity, efficacy, viability, and efficiency [49]. This framework helped to determine factors that could be viewed as either facilitators or barriers to the management of COVID-19 within LTCFs.

Methods

Conceptual framework

Despite being commonly expressed in terms of case numbers and death rates [50, 51], the performance of a healthcare system in the context of the COVID-19 pandemic can be conceptualized in multiple ways. For example, the OECD defines a healthcare system’s performance as the attainment of goals at the lowest possible cost [52], whereas the framework used by the Canadian Institute for Health Information evaluates healthcare system performance by answering the questions “How healthy are Canadians?” and “How healthy is the health system?” [53]. In Québec, the Ministère de la Santé et des Services Sociaux (MSSS) adopted a conceptual framework entitled the Cadre de référence ministériel d’évaluation de la performance du système public de santé et de services sociaux à des fins de gestion, where a healthcare system’s performance is conceptualized as the system’s ability to reach its objectives in relation to the population’s health, taking into consideration the optimization of resources and the quality and accessibility of services [54].

We chose the MSSS framework as the foundation of our scoping review’s conceptual framework because of its multifaceted approach to assessing healthcare system performance. Thus, the conceptual framework that guided our review encompassed eight dimensions of performance that fell under three fundamental elements of performance: 1) equity and accessibility within the accessibility of services; 2) reactivity, safety, continuity, and efficacy within the quality of services; and 3) viability and efficiency within the optimization of resources (see Fig. 1).

Fig. 1

The conceptual framework of healthcare system performance adapted from the Cadre de référence ministériel d’évaluation de la performance du système public de santé et de services sociaux à des fins de gestion developed by the Québec’s Ministère de la Santé et des Services Sociaux (19)

The definitions of the dimensions of performance included in the original MSSS framework are presented in Table 1. To optimize database queries, we slightly modified the original MSSS framework’s dimensions in our search strategies. Specifically, we substituted the dimension of reactivity with the terms adaptability and satisfaction, whose definitions within the databases better aligned with the original MSSS definition of reactivity. By the same token, the terms resource management and resource mobilization were used in our searches in lieu of the dimension of viability. In addition, the CINAHL database terms effectiveness and security were identified as synonyms of the original MSSS dimensions of efficacy and safety, respectively, and thus were added to the search strategies to ensure all relevant literature was captured. Throughout the rest of the manuscript, only the original MSSS definitions are used (Table 1).

Table 1 Population, Interest, Comparison, Outcome, and Time (PICOT) model

In addition to the dimensions of performance, the conceptual framework incorporated factors that could influence LTCF performance, as informed by the previous literature [48, 55]. Factors were categorized into eight internal and four external factors, as presented in Table 1.

Methodological framework

We followed the methodology for scoping reviews developed by the Joanna Briggs Institute, which expands on the work done by Arksey and O’Malley as well as Levac et al. [56, 57]. Accordingly, we followed these nine steps: 1) formulating and aligning the review’s objective(s) and question(s); 2) developing eligibility criteria in keeping with the established objective(s) and question(s); 3) describing the approach to database queries, article selection, data extraction, and presentation of findings; 4) searching for the evidence; 5) selecting the evidence; 6) extracting the evidence; 7) analyzing the evidence; 8) presenting the evidence; and 9) summarizing the evidence with respect to the review’s objective(s) and question(s), drawing conclusions, and noting potential implications.

To ensure our review included all necessary elements, we completed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) Checklist (see Supplementary File 1) [58]. We also consulted the updated guideline for reporting scoping reviews that was outlined in the PRISMA 2020 Statement [59, 60].

Eligibility criteria

Table 1 presents the eligibility criteria using the Population, Interest, Comparison, Outcome, and Time (PICOT) model.

Data sources and search strategy

The protocol for this review was registered with the Research Registry [researchregistry7026] and published [61]. We queried the databases CINAHL, MEDLINE (Ovid), CAIRN, Science Direct, Scopus, and Web of Science for records that met the eligibility criteria, including the context (COVID-19 pandemic), population (LTCFs), interest (factors that influenced LTCF performance), and outcomes (dimensions of performance: equity, accessibility, reactivity, safety, continuity, efficacy, viability, efficiency). We restricted our searches to peer-reviewed literature written in English or French and published during the first 2 years of the COVID-19 pandemic, between January 1^st 2020 and December 31^st 2021. We excluded records that focused on infections other than COVID-19 and healthcare settings other than LTCFs. We also excluded records that focused on pharmaceutical treatments or COVID-19 vaccination rates.

Two co-authors (JL, EB) performed database searches independently using strategies developed by our team (see Supplementary Files 2–7). All searches were performed using descriptors with the Boolean operators “AND” and “OR”. Retrieved records were imported to the EndNote software, and duplicates were removed.

Article selection

From Endnote, retrieved records were exported to the Rayyan web platform [62], duplicates were removed, and records were screened in accordance with an algorithm developed by our research team [63]. Two co-authors (JL, EB) pilot tested the algorithm for reliability by independently screening the titles and abstracts of the first 10% of articles and then comparing results. The screening algorithm was then elaborated upon, for example, by adding the definitions of the dimensions of performance.

Once the algorithm had been tested and refined, the first complete screening round took place, where the titles and abstracts of all records were screened by two co-authors (JL, EB) as well as divided between the other co-authors, such that each record was screened by three co-authors. All three co-authors had to deem a record eligible for it to pass to the second screening round, where each record was read in its entirety by one co-author who decided whether to include or exclude it (see Fig. 2). The decision of whether to retain a record was based on three questions, answered in order: 1) Does the record focus on management activities during the COVID-19 pandemic?; 2) Does the population include residents and/or staff of a single or multiple LTCF(s) during the pandemic?, and 3) Does the record discuss at least one factor (e.g., resident characteristics, financial resources, etc.) that influenced LTCF performance during the pandemic? At this stage, it was not required to identify all factors discussed in each record; once any one factor was identified in a record, the record was retained for data extraction.

Fig. 2

Screening algorithm. Adapted from previous work by Tchouaket et al. [63]. Legend: ^Q1 MANAGEMENT OF COVID-19: Is the primary focus of the record on management activities (e.g., organisation, planning, IPC) within a healthcare facility during the COVID-19 pandemic? ^Q2 POPULATION: Does the record specify its population as LTCFs housing elderly residents or hospice/palliative care services occurring within the context of a single or multiple LTCF(s)? (Synonyms: nursing homes, assisted-living facilities, homes for the aged, aged care homes, retirement homes, long term care). ^Q3 INTERESTS: Does the record discuss at least one factor related to the management of COVID-19 and outcomes of the pandemic within a single or multiple LTCF(s)? Internal factors: 1) resident characteristics; 2) facility characteristics; 3) staff characteristics; 4) human resources; 5) material resources; 6) technological resources; 7) financial resources; and 8) organizational context. External factors: 1) admissions; 2) visitors; 3) virus circulation in the community; and 4) public health policies/guidelines. Note: All keywords were translated into French for the CAIRN database

Data extraction

Relevant data from retained articles were extracted in three stages. First, descriptive characteristics were extracted using a data charting form developed by two co-authors (JL, EB) and reviewed by the entire research team. These same two co-authors (JL, EB) pilot tested the form, after which data extraction by five co-authors (KK, FEM, SR, JL, EB) occurred using Google Forms. The following data were extracted from each article: citation, year of publication, country of origin, study objectives, design, setting, participants, and framework (if one was used).

Following the first stage, dimensions of performance were extracted using a coding frame, where each dimension was defined as per the adopted conceptual framework. To ensure reliability, two co-authors (KK, SR) were first assigned an identical set of 20% (28/140) of included articles. The co-authors read the articles independently to extract identified dimensions of performance along with verbatim or paraphrased examples (1–3 sentences). For each record, dimensions were coded as “1”, if extracted, or “0”, if not extracted. Inter-coder reliability was calculated by summing the number of dimensions upon which both co-authors had agreed (e.g., viability coded as “1” or “0” for both coders), dividing it by eight (total number of dimensions), and multiplying the result by 100% [64, 65]. Mean inter-coder reliability across the 28 records was 81.3%. The two co-authors (KK, SR) met to share feedback about the coding process, and, in the case of conflicts, a third co-author (ENT) arbitrated. The coding frame was then fine-tuned by specifying what did and did not qualify as supporting evidence for each dimension (see Supplementary File 8). Lastly, the remaining records were randomly allocated to the same co-authors (KK, SR), who completed the extraction of dimensions.

All relevant factors reported to have influenced LTCF performance were extracted from each retained article. To ensure reliability, two co-authors (KK, SR) were first assigned an identical set of 10% (14/140) of records. The co-authors read the articles independently to extract identified factors into an Excel spreadsheet along with verbatim or paraphrased examples (1–3 sentences). In the case of identifying a factor that was not specified in the screening algorithm, the factor was extracted under the category “Other”. The co-authors then met to discuss the coding process and resolve any conflicts. Following this discussion, the remaining articles were randomly allocated to the same co-authors (KK, SR), who completed the extraction of factors.

Data synthesis and presentation

Three co-authors (KK, SR, ENT) analyzed extracted data. Countries of origin were collapsed into geographic regions. Descriptive characteristics were summarized, and the data collection period was extracted to situate the evidence within the temporal progression of the COVID-19 pandemic. Dimensions were extracted and their frequencies displayed in tabular form as were the factors identified to have influenced at least one dimension of LTCF performance. The completed data charting form is provided in Supplementary File 9.

Research team

The principal investigator (ENT) was supported by a registered nurse (JL), who is an expert in the assessment of factors that contribute to outbreaks, IPC, and performance evaluation in nursing. The research team also included two professors (DS, IB) with extensive experience with scoping review methodology in the field of nursing sciences, as well as research professionals (KK, SR, EB, FEM, SS) and a doctoral student in nursing (MJ), who are all proficient in database searches, article selection, and manuscript preparation.

Results

Article selection

Results of the selection process are summarized in a PRISMA flow chart (see Fig. 3). After removing duplicates and screening titles and/or abstracts, 448 of the 9,895 retrieved records underwent full-text assessment. After eliminating 15 records for which full text was not accessible, 433 records were read in their entirety, of which 139 were retained. Reasons for exclusion included: addressing a wrong subject (n = 159), having a wrong interest (n = 108), focusing on a wrong population (n = 16), being in a language other than English or French (n = 9), and being a duplicate (n = 2). In addition, one record was added using a snowballing approach for the total of 140 articles.

Fig. 3

PRISMA flow chart, outlining the identification and selection stages of this review. Adapted from The PRISMA 2020 statement: An updated guideline for reporting systematic reviews [59]. The term “report” signifies “a document (paper or electronic) supplying information about a particular study”, such as a journal article or government report, while the term “record” signifies “The title or abstract (or both) of a report indexed in a databased or website” [59]

Characteristics of retained articles

Table 2 summarizes descriptive characteristics of the 140 retained articles, including publication type, study design, setting, participants, data collection/analysis period, month/year of publication, country and region of origin, and framework (if one was used).

Table 2 Article characteristics (N = 140)

Countries and regions of origin

North America was the most represented region with 66 out of 140 (66/140, 47.1%) articles, of which 53 (80.3%) originated in the USA [11,12,13,14,15, 18, 19, 25, 29, 30, 34,35,36,37,38,39,40,41, 47, 50, 70, 73, 74, 76, 81, 89, 90, 92, 100,101,102,103,104, 108, 110, 111, 114, 117, 121,122,123, 126, 127, 129, 131, 137, 138, 141, 147, 148, 150, 159, 162], 12 (18.2%) came from Canada [16, 22, 24, 27, 107, 109, 113, 118, 124, 132, 139, 164], and one (1.5%) article covered both the USA and Canada [125]. North America was followed by Europe with 34 out of 140 (24.3%) articles, of which nine (26.5%) originated in France [21, 23, 28, 75, 78, 80, 87, 96, 136], six (17.6%) in England [72, 88, 119, 120, 143, 151], five (14.7%) in Italy [26, 67, 69, 77, 82], three (8.8%) in Germany [17, 94, 105], two (5.9%) in Spain [97, 140], and three (8.8%) covered several European countries [79, 95, 128], while the rest represented one country each, including Austria [33], Belgium [31], Cyprus [158], Norway [116], the Netherlands [91], and Turkey [71]. Nineteen of the 140 articles (13.6%) came from Asia, of which six (31.6%) originated in China [112, 115, 145, 160, 161, 163], five (26.3%) in Japan [66, 98, 133,134,135], two (10.5%) in Hong Kong [85, 144], two (10.5%) in Singapore [154, 155], two (10.5%) in South Korea [20, 149], and two (10.5%) in Taiwan [84, 146]. Australia was represented by six (4.3%) articles [99, 106, 152, 156, 157, 165], South America by two (1.4%) articles both coming from Brazil [93, 130], and the Middle East by one (0.7%) article originating in Saudi Arabia [68]. In addition, one (0.7%) article covered both Europe and Australia [83], and 11 (7.9%) articles—including seven reviews—adopted a global perspective [8, 32, 42, 43, 45, 46, 48, 55, 86, 142, 153].

Publication types

Of the 140 retained articles, 95 (67.9%) were classified as research articles and included, for example original articles, reviews, and brief reports [8, 11,12,13,14,15,16,17,18,19,20,21,22,23, 28, 33,34,35,36,37,38,39,40,41, 43, 45,46,47,48, 50, 55, 66,67,68,69,70,71,72, 74,75,76,77, 80, 82, 87, 89, 91,92,93, 96, 97, 100,101,102,103, 105,106,107,108,109,110,111,112,113,114, 116,117,118,119,120, 122,123,124, 126,127,128,129, 132, 135,136,137, 139, 142, 143, 145, 147,148,149, 151, 152, 158, 161,162,163,164]. In addition to research articles, our sample included 18 (12.9%) discussion papers [29,30,31,32, 78, 81, 85, 94, 125, 130, 131, 138, 140, 150, 153, 159, 160, 165]; 11 (7.9%) policy analysis papers [24, 26, 27, 83, 86, 88, 90, 95, 98, 115, 134]; seven (5.0%) letters to the editor [73, 84, 133, 144, 146, 154, 155]; three (2.1%) medical news articles [25, 104, 141]; two (1.4%) professional journal articles [99, 157]; one (0.7%) communication [156]; one (0.7%) guideline [79]; one (0.7%) conference proceedings paper [42]; and one (0.7%) morbidity and mortality weekly report [121].

Designs of research articles

The 95 research articles included 39 (41.1%) studies that analyzed secondary quantitative data [13, 18, 20, 34,35,36,37,38,39, 41, 50, 66, 69, 71, 72, 76, 77, 80, 89, 92, 100,101,102,103, 105,106,107,108, 110, 111, 118, 122, 123, 129, 132, 137, 139, 148, 162], 38 (40.0%) studies that used primary quantitative data [11, 12, 14, 15, 17, 19, 21,22,23, 28, 33, 40, 67, 68, 70, 74, 75, 87, 91, 93, 96, 97, 109, 112,113,114, 117, 119, 126, 127, 135, 136, 143, 145, 147, 149, 158, 161], eight (8.4%) qualitative studies [16, 112, 120, 124, 151, 152, 163, 164], one (1.1%) study that used both primary and secondary quantitative data [82], as well as nine (9.5%) reviews, including three rapid reviews [8, 55, 142], two scoping reviews [45, 48], one systematic review [47], one rapid systematic review [46], one literature review [43], and one narrative review [128].

The 39 studies that analyzed secondary quantitative data encompassed 24 (61.5%) cross-sectional studies [34,35,36,37,38,39, 50, 69, 71, 72, 92, 100,101,102, 105, 106, 108, 110, 111, 122, 123, 137, 139, 148], seven (17.9%) retrospective cohort studies [80, 89, 103, 107, 118, 132, 162], three (7.7%) case-control studies [41, 76, 129], three (7.7%) descriptive epidemiological studies [13, 20, 66], one (2.6%) repeated cross-sectional study [77], and one (2.6%) prospective cohort study [18].

The 38 studies that collected primary quantitative data included 10 (26.3%) cross-sectional studies [11, 15, 23, 87, 91, 112, 119, 136, 143, 145], eight (21.1%) prospective cohort studies [14, 19, 28, 40, 67, 68, 117, 147], four (10.5%) implementation research studies [70, 74, 97, 127], two (5.3%) implementation research/cross-sectional studies [21, 109], two (5.3%) descriptive epidemiological studies [114, 149], two (5.3%) mixed-methods studies [93, 161], two (5.3%) retrospective cohort studies [17, 75], one (2.6%) cross-sectional study based on computational modeling [126], one (2.6%) mixed-methods pre-post evaluation study [113], one (2.6%) quasi-experimental before-and-after study [22], one (2.6%) repeated cross-sectional study [12], one (2.6%) retrospective cohort/cross-sectional study [33], one (2.6%) retrospective pre-post study [96], one (2.6%) complex intervention with a prospective cohort/pre-post study [158], and one (2.6%) interventional pre-post study [135].

Settings and participants of research articles

Of the 38 studies that used primary quantitative data, 27 (71.1%) were based on evidence from multiple LTCFs with the number of facilities ranging from 2 to 5,126 [11, 15, 19, 22, 23, 33, 40, 70, 74, 75, 87, 93, 96, 97, 109, 112,113,114, 117, 119, 127, 136, 143, 145, 149, 158, 161], whereas 11 (28.9%) focused on a single facility [12, 14, 17, 21, 28, 67, 68, 91, 126, 135, 147]. Of the 39 studies that were based on secondary quantitative data, 32 (82.1%) pooled evidence from multiple LTCFs with the number of facilities ranging from 3 to 15,390 [18, 20, 34,35,36,37,38,39, 41, 50, 69, 72, 76, 77, 89, 92, 100,101,102,103, 105, 106, 110, 111, 118, 122, 123, 129, 137, 139, 108, 148], whereas seven (7.7%) centered on a single facility [13, 66, 71, 80, 107, 132, 162]. Of the eight qualitative studies, seven (87.5%) presented evidence from multiple facilities [116, 120, 124, 151, 152, 163, 164] and one (12.5%) described the experiences of a single facility [16]. The one study based on both primary and secondary quantitative data used evidence from multiple facilities [82].

Of the 38 studies that relied on primary quantitative data, 20 (52.6%) included both residents and staff as participants [12, 17, 21, 28, 33, 67, 68, 75, 91, 97, 114, 117, 119, 126, 135, 136, 143, 147, 149, 158], 10 (26.3%) focused on staff [23, 70, 74, 87, 93, 109, 112, 113, 145, 161], six (15.8%) focused on residents [11, 14, 15, 19, 96, 127], and two (5.3%) collected facility-level data [22, 40]. Of the 39 studies that analyzed secondary quantitative data, 27 (69.2%) used facility-level data [18, 34,35,36,37,38,39, 41, 50, 76, 89, 92, 100,101,102,103, 106, 108, 110, 111, 118, 122, 123, 129, 137, 139, 148], nine (23.1%) included both residents and staff as participants [13, 20, 66, 71, 80, 105, 107, 132, 162], one (2.6%) focused on residents [77], one (2.6%) employed municipal- and facility-level data [69], and one (2.6%) used facility- and resident-level data [72]. The one study that combined primary and secondary quantitative data performed a facility-level analysis [82]. All participants in the eight qualitative studies were staff members [16, 116, 120, 124, 151, 152, 163, 164].

COVID-19 pandemic timeframe

To situate the evidence within the temporal progression of the COVID-19 pandemic, we plotted the period of data collection or analysis for the 95 research articles (see Fig. 4). Most research articles (81/95, 85.3%) used one or more consecutive weeks/months during 2020 as their period of data collection or analysis [8, 11,12,13,14,15,16,17,18,19,20,21,22,23, 28, 33,34,35,36,37,38,39,40,41, 43, 46, 50, 66,67,68,69,70,71,72, 75, 82, 91,92,93, 96, 97, 100, 102, 103, 105,106,107,108,109,110,111,112,113,114, 116,117,118,119,120, 122,123,124, 126,127,128, 132, 136, 137, 142, 143, 145, 147,148,149, 151, 152, 158, 161,162,163,164]. In addition, of the 95 research articles, eleven (11.6%) articles used a period beginning in 2020 and extending into 2021 [45, 47, 48, 55, 74, 76, 87, 89, 101, 129, 139], one (1.1%) article focused on a 2-month period in 2021 [80], one (1.1%) article reported results of an interventional study conducted between April 2019 and March 2021 [135], and one (1.1%) article covered two non-consecutive periods, one corresponding to the first wave of the pandemic in 2020 and the other to the second wave in 2021 [77].

Fig. 4

Plot of the period of data collection/analysis for research articles (n = 95); numbers represent the number of articles covering each period. *Data collection began in 2019

Frameworks

Of the 140 articles, 12 (8.6%) applied a conceptual or theoretical framework, including: a checklist of measurable IPC practices adapted from guidelines for managing COVID-19 in LTCFs [16]; an adapted organizational framework analysis focusing on social ties and the interdependency between individuals and organizations [120]; Barker’s behavior setting theory [161]; Ernst and Chrobot-Mason’s framework of boundary spanning leadership [81]; a “fuzzy comprehensive evaluation method” that integrates internal and external factors to evaluate performance in terms of IPC [112]; the organizational crisis management theory [112]; the health and social care framework [45]; the Infection Prevention and Control Assessment and Response Tool [40]; the Plan-Do-Study-Act Cycle [14]; Macrae and Wiig’s resilience framework [116]; the vulnerable population conceptual framework [162]; Behrens and Naylor’s operational framework for a coordinated response to the COVID-19 pandemic [153]; and the framework of the determinants of the risk of severe infection in conjunction with the concept of deep defense within infection management [134].

Dimensions of LTCF performance

Table 3 displays dimension(s) of LTCF performance extracted from all retained articles. Efficacy and safety were the most frequently discussed dimensions; these were extracted from 106 (75.7%) of the 140 articles each, followed by viability (81/140, 57.9%), continuity (43/140, 30.7%), accessibility (38/140, 27.1%), and reactivity (37/140, 26.4%). Equity and efficiency were the least frequently addressed dimensions; these were extracted from 19 (13.6%) and 18 (12.9%) articles, respectively.

Table 3 Dimensions of performance (N = 140)

Factors that influenced COVID-19 management in LTCFs

Factors that were discussed in relation to the management of COVID-19 within LTCFs were extracted from all 140 articles (see Table 4). Among internal factors, “organizational context” and “human resources” were the most frequently reported factors; these were extracted from 102 (72.9%) and 87 (62.1%) of the 140 articles, respectively, followed by “material resources” (63/140, 45.0%), “facility characteristics” (53/140, 37.9%), “staff characteristics” (52/140, 37.1%), “resident characteristics” (37/140, 26.4%), “technological resources” (26/140, 18.6%), and “financial resources” (18/140, 12.9%) (see Table 4). Among external factors, “visitors” and “public health policies/guidelines” were the most frequently reported factors; these were extracted from 38 (27.1%) and 36 (25.7%) of the 140 articles, respectively, followed by “virus circulation in the community” (22/140, 15.7%), and “admissions” (17/140, 12.1%). Sixteen (11.4%) articles reported factors that were extracted under the category “Other”. These included asymptomatic or atypical presentation of COVID-19, the lack of knowledge about COVID-19 symptoms, and vaccination.

Table 4 Internal and external factors influencing the performance of long-term care facilities (N = 140)

Discussion

This scoping review identified and synthesized factors that could influence LTCF performance by analyzing a vast body of international literature from early in the COVID-19 pandemic until the end of 2021. By applying a multidimensional conceptual framework of healthcare system performance, this study brings to light various factors that were reported to have influenced the accessibility and quality of healthcare services during the pandemic, how the available resources were optimized, and how these parameters contributed to overall LTCF performance.

LTCF performance during the COVID-19 pandemic

The majority of included articles considered the performance of a single or multiple LTCF(s) along the dimensions of efficacy (75.7%) and safety (75.7%). As per this review’s conceptual framework, these two dimensions gauge healthcare system performance in terms of the quality of the services provided. This finding reflects the urgency of the pandemic response demanded from LTCF management around the world, which led to the prioritization of the delivery of safe care amidst a public health crisis. Efficacy was predominantly discussed in terms of the degree to which IPC interventions implemented by LTCFs were successful at decreasing the transmission of COVID-19 or managing COVID-19 outbreaks once they occurred. Effective reduction of COVID-19 incidence, morbidity, and mortality was linked to early implementation of and adherence to IPC measures, including: the use of personal protective equipment (PPE), hand hygiene, sanitation, isolation, universal serial testing of residents and staff, cohorting, vaccination, visitor restrictions, and the availability of IPC trained staff (e.g., [11, 12, 15,16,17,18, 43, 71, 73, 76, 79, 80, 85, 105, 115, 121, 129]). Furthermore, close integration between acute and long-term care, communication between LTCFs and public health authorities, and access to a multidisciplinary clinical team emerged as strategies that contributed to a more successful COVID-19 response (e.g., [8, 21, 31, 81, 109, 124]). In terms of safety, IPC measures put in place by LTCFs to limit viral spread were most commonly reported. Multiple publications provided evidence of how LTCF staff acted as vectors of COVID-19 introduction and transmission, for example when working at multiple facilities or when not using PPE properly, if at all, due to severe resource shortages (e.g., [24, 79, 141]). Safety risks associated with LTCF staff working at multiple facilities can be addressed by asking staff to reside with residents under isolation, which was reported to be effective at reducing COVID-19 incidence and mortality [48, 75]. However, the implementation of any IPC protocol must be weighed against its potential impact on the mental health of residents and staff and on overall quality of life. One proposed antidote to negative mental health outcomes as a consequence of strict IPC measures was a person-centered approach to long-term care, where the delivery of care is negotiated between the LTCF resident and the healthcare professional and delivered in a personalized and respectful manner [94].

Approximately 13% of the included articles considered LTCF performance along the dimensions of efficiency and equity, which evaluate performance in terms of the optimization of resources and the accessibility of services, respectively. Staff producing their own PPE, allocation of unused areas for staff self-confinement, storage of material supplies by pharmacies for distribution to multiple LTCFs, and technological innovations, such as portable infection equipment trolleys, were highlighted as strategies employed by some LTCFs to ensure an efficient use of available resources (e.g., [75, 81, 116, 120]). In relation to equity, several articles, most of which originated in the USA, noted how the COVID-19 pandemic put a spotlight on the existing health inequities and, in some cases, exacerbated them among already vulnerable populations, notably low-income individuals and communities of color (e.g., [72, 90, 162]). An association was commonly reported between a LTCF housing a greater proportion of racial/ethnic minority residents and a higher probability of COVID-19 cases (e.g., [35, 39, 43, 48]).

The two most frequently extracted dimensions of performance, efficacy and safety, are strongly linked to the two factors most commonly reported by the included articles—organizational context (72.9%) and human resources (62.1%). This finding underscores the pressing need for increased investment in long-term care infrastructure and workforce globally to ensure satisfactory working conditions, job security, decent pay, and adequate levels of staff who are competent in IPC, geriatrics, and palliative care [166]. Enforcement of safety standards and continuous human resource development are necessary steps towards reductions in staff absenteeism and turnover rates, better continuity of care, and improved care delivery [166].

About one in nine (11.4%) articles included in the review noted unexpected factors, including: the initial lack of knowledge about the virus and the symptoms it causes [42, 78, 132]; the difficulty detecting and managing asymptomatic and atypical COVID-19 cases [13, 43, 48, 68, 94, 114, 134] which was subsequently addressed by universal serial testing of residents and staff [23, 73, 130]; and the ability to curb COVID-19 transmission with early administration of vaccines [32, 129].

Our findings are in line with a scoping review by Giri et al. [48], which showed that the ability of LTCFs to contain COVID-19 outbreaks was influenced by an interplay of various internal (related to residents, staff, and facilities) and external (material resources and public health policies) factors, as well as asymptomatic transmission of COVID-19. Building on Giri et al.’s classification of factors, our review provides a more fine-grained understanding of why some LTCFs performed better than others by extracting a greater number of factors and linking these factors to multiple dimensions of performance using a conceptual framework. Furthermore, our findings emphasized the contribution of the LTCF’s organizational structure and the quality of its leadership to the facility’s capacity to effectively respond to the COVID-19 pandemic. Organizational challenges leading to shortages of staff and material resources during the first wave of the pandemic were also highlighted by Palacios-Ceña et al.’ [45] scoping review of qualitative evidence. Consistent with our findings, this narrative synthesis found evidence of LTCF staff taking initiative in procuring supplies by cultivating connections within the community.

Directions for future research

The descriptive characteristics of the articles included in our scoping review revealed a gap in the existing knowledge pertaining to the performance of LTCFs during the COVID-19 pandemic. First, except for one quasi-experimental pre-post study [22], all research articles included in our review were non-experimental studies, which precludes the establishment of causality between the identified factors and LTCF performance. Therefore, experimental studies should be prioritized in the future to help understand which factors could predict LTCF performance during the ongoing COVID-19 pandemic or future outbreaks. When experimental designs are not feasible, qualitative studies can help gain a deeper, more nuanced understanding of the various factors that facilitated or hindered the management of COVID-19 in LTCFs by gleaning perspectives of diverse stakeholders, such as facility leadership and staff, residents and their families, and public health officials [167]. Second, the majority of research articles in our review were cross-sectional, thus providing only a snapshot of LTCF response to COVID-19, with the bulk of evidence coming from the first wave of the pandemic (between March and July 2020). The skewness towards the early phases of the pandemic might have been behind efficacy and safety being the most commonly reported dimensions of performance, reflecting the urgency to provide care to LTCF residents in a safe and effective manner amidst a public health crisis. Additional population-based longitudinal studies are warranted to explore long-term impacts of the reported factors on LTCF performance [168]. Third, only 9% of the included articles used a framework to guide their inquiry. We recommend future studies adopt a framework to help understand complex healthcare systems, as it would allow for inter-study comparability and a fine-grained analysis of structures and processes that influence the delivery and outcomes of care [169]. Fourth, a few articles included in our review discussed LTCF performance as a function of the dimensions of efficiency and equity. Thus, more research is needed to examine LTCF performance in terms of the optimization of available resources and fair access to services both in the context of the COVID-19 pandemic and more broadly. Fifth, the reviewed qualitative literature mostly examined the practices and perspectives of LTCF staff. Studies exploring LTCF residents’ experiences in the pandemic context, for example with respect to their feelings of loneliness, may provide valuable information that could help maintain and improve residents’ quality of life (e.g., [170, 171]). Lastly, of the 11 Canadian-based studies included in our review, 10 originated in the provinces of Ontario and British Columbia, and only one came from Québec. This represents a substantial knowledge gap given that Québec experienced the highest COVID-19 mortality rate in Canada between March 2020 and October 2021 [172]. Future work should examine the failures of IPC in hard-hit regions of the world to obtain insight into the causes and mechanisms of inadequate healthcare, as well as provide opportunities to innovate and apply higher quality care in the future [173, 174].

Limitations and strengths

Due to the evolving nature of the COVID-19 pandemic, the results of this review may not reflect the current public health profile. Furthermore, because our review focused on the management of COVID-19, the findings may not fully capture indirect effects of the IPC measures implemented by LTCFs that have, nonetheless, had an influence on residents’ mental health. For example, the restriction or prohibition of family visits and group activities in LTCFs resulted in residents experiencing social isolation and loneliness [175,176,177]. Though six databases were used for queries we could have included more. We did not search for articles written in languages other than English or French, and French articles that were retrieved were screened and excluded.

Despite these limitations, our study offers a comprehensive review of various factors reported to have influenced the management of COVID-19 within LTCFs by applying a multidimensional conceptual framework of performance. The adopted framework guided us throughout the entire review process by providing a clear focus on our review’s aim and objectives. In addition to the conceptual framework, adherence to the JBI methodology for scoping reviews helped us maintain the overall rigor of the study. This work may inform the development of more effective IPC interventions to help prevent or mitigate future outbreaks in LTCFs, while being sensitive to the needs, preferences, and values of residents and staff. The lessons learned thus far should be considered on an evolving basis when developing IPC programs specific to long-term care.