The Impact of Nursing Staffs’ Working Conditions on the Quality of Care Received by Older Adults in Long-Term Residential Care Facilities: A Systematic Review of Interventional and Observational Studies

Background: Little documentation exists on relationships between long-term residential care facilities (LTRCFs), staff working conditions and residents’ quality of care (QoC). Supporting evidence is weak because most studies examining this employ cross-sectional designs. Methods: Systematic searches of twelve bibliographic databases sought experimental and longitudinal studies, published up to May 2021, focusing on LTRCF nursing staff’s working conditions and the QoC they provided to older adults. Results: Of the 3577 articles identified, 159 were read entirely, and 11 were retained for inclusion. Higher nursing staff hours worked per resident per day (HPRD) were associated with significant reductions in pressure sores and urinary tract infections. Overall staff qualification levels and numbers of RNs had significant positive influences on QoC. Conclusions: To the best of our knowledge, this systematic review is the first to combine cohort studies with a quasi-experimental study to explore associations between LTRCF nursing staff’s working conditions and older adult residents’ QoC. Human factors (including HPRD, staff turnover, skill mix, staff ratios) and the specific working contribution of RNs had overwhelmingly significant influences on QoC. It seems essential that LTRCF supervisory and decision-making bodies should promote optimal working conditions for nursing staff because these have such a direct impact on residents’ QoC.


Introduction
A challenge facing most modern societies, caused by ageing populations, is the increasing demand on health and care services [1]. Ageing increases the risks of developing multiple chronic conditions, leading to patients with complex long-term care needs [2]. Across European countries, ageing populations are creating a growing demand for health care staff, particularly nursing specialists in geriatric or psychogeriatric care, to care for older adults with complex, age-related, somatic and psychopathological conditions [3,4]. Thus, it seems likely that a major part of the nursing and medical care required for this population will be redirected from hospitals to home health care or assisted living teams, intensifying the need for highly-specialized geriatric care [5]. The lack of highly qualified

Materials and Methods
The present work followed the Joanna Briggs Institute's guidelines for systematic reviews and was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations [37]. Systematic review registration: PROSPERO 2021: CRD42021226656

Types of Studies
This review included publications on primary research if they (1) examined the relationship between nursing staff's working conditions and the QoC received by nursing-home residents, (2) only included LTRCFs, and (3) were original research articles describing observational, longitudinal, or experimental quantitative studies. There were no restrictions on language, country of origin, or publication date.

Types of Participants
The types of staff who had participated in the studies were certified nursing assistants (CNAs), licensed practical nurses (LPNs), and registered nurses (RNs). CNAs are defined as professional health care workers who have had several months of health care training and who provide direct care to residents, under the supervision of LPNs or RNs. LPNs are defined as professional health care workers who have completed a two-to-three-year health care training program and who provide support services to RNs. RNs are defined as professional health care workers who have completed a three-to-four-year health care studies program, obtained a bachelor's degree or equivalent, and who perform many basic and advanced nursing tasks.

Factors of Interest
Factors of interest were any individual or organizational variables associated with staff working conditions. The European Foundation for the Improvement of Living and Working Conditions describes working conditions as "the conditions in and under which work is performed. A working condition is a characteristic or a combination of characteristics of work that can be modified and improved" [38]. According to the International Labour Organization, working conditions cover a broad range of topics and issues, from working time (hours of work, rest periods, and work schedules) to remuneration, as well as the physical conditions and mental demands that exist in the workplace [39].
In the health field, the World Health Organization defines several key components of working conditions for nursing staff, including working hours, shift work, workload, staffing levels, and clinical rotation [40]. Nurse staffing levels are a major factor in nurses' working conditions and are directly related to the QoC [41]. According to the International (6-7 stars), or high (8-9 stars) quality [37]. Researchers discussed any disagreements to reach consensus.
We used the validated Robins-I tool for assessing the risk of bias in non-randomized studies of interventions (NRSIs) [38]. This tool covers two dimensions and seven domains through which bias might be introduced into an NRSI: (i) pre-intervention and at-intervention bias (due to confounding, the selection of study participants, or the classification of the intervention), and (ii) post-intervention bias (due to missing data, deviations from intended interventions, bias in the measurement of outcomes, or bias in the selection of the reported result) [38]. Any disagreements in quality assessments were resolved through discussion.

Search Strategy Results
Our strategy of searching bibliographic databases retrieved a total of 3577 articles after the elimination of duplicates. Based on their titles and abstracts, 159 articles were retained as potentially eligible, and their entire texts were evaluated. In the end, only 11 articles met our selection criteria and were included: 10 cohort studies and one quasi-experimental interventional study ( Figure 1).

Characteristics of Studies, Participants, and Institutions
The eleven included studies were carried out in four countries (Germany, China, South Korea, and the USA), across three continents (Europe n = 2, Asia n = 3, and North America n = 6), and published between 1977 and 2018. Ten were cohort studies [40][41][42][43][44][45][46][47][48][49], and one was a quasi-experimental interventional study [50] (Table 1).  (2006), Germany study n = 29 (4 months) qualified personnel and incidence of pressure ulcers with low (< 50%); 2) medium (50-60%); and 3) high proportions of qualified personnel (≥ 60%)  Incidences of pressure ulcers: number of residents with a pressure ulcer in relation to the total number of residents Popp et al. [55] (2006), Germany Cohort study Nursing homes n = 29 0.33 (4 months) between proportions of qualified personnel and incidence of pressure ulcers equivalent posts occupied  Residents classified into three groups: 1) cared for with low (< 50%); 2) medium (50-60%); and 3) high proportions of qualified personnel (≥ 60%)  Incidences of pressure ulcers: number of residents with a pressure ulcer in relation to the total number of residents Popp et al. [55] (2006), Germany Cohort study Nursing homes n = 29 0.33 (4 months) Examine relationships between proportions of qualified personnel and incidence of pressure ulcers  Data source: Hamburger Qualitätsvergleich in der Dekubitusprophylaxe  Proportions of qualified personnel: full-time equivalent posts occupied  Residents classified into three groups: 1) cared for with low (< 50%); 2) medium (50-60%); and 3) high proportions of qualified personnel (≥ 60%)  Incidences of pressure ulcers: number of residents with a pressure ulcer in relation to the total number of residents  [52] Kwong et al. [50]        Popp et al. [55] (2006), Germany Cohort study Nursing homes n = 29 0.33 (4 months) between proportions of qualified personnel and incidence of pressure ulcers  Residents classified into three groups: 1) cared for with low (< 50%); 2) medium (50-60%); and 3) high proportions of qualified personnel (≥ 60%)  Incidences of pressure ulcers: number of residents with a pressure ulcer in relation to the total number of residents Popp et al. [55] (2006), Germany Cohort study Nursing homes n = 29 0.33 (4 months) between proportions of qualified personnel and incidence of pressure ulcers  Residents classified into three groups: 1) cared for with low (< 50%); 2) medium (50-60%); and 3) high proportions of qualified personnel (≥ 60%)  Incidences of pressure ulcers: number of residents with a pressure ulcer in relation to the total number of residents Popp et al. [55] (2006), Germany Cohort study Nursing homes n = 29 0.33 (4 months) Examine relationships between proportions of qualified personnel and incidence of pressure ulcers Dekubitusprophylaxe  Proportions of qualified personnel: full-time equivalent posts occupied  Residents classified into three groups: 1) cared for with low (< 50%); 2) medium (50-60%); and 3) high proportions of qualified personnel (≥ 60%)  Incidences of pressure ulcers: number of residents with a pressure ulcer in relation to the total number of residents Examine relationships between proportions of qualified personnel and incidence of pressure ulcers  Data source: Hamburger Qualitätsvergleich in der Dekubitusprophylaxe  Proportions of qualified personnel: full-time equivalent posts occupied  Residents classified into three groups: 1) cared for with low (< 50%); 2) medium (50-60%); and 3) high proportions of qualified personnel (≥ 60%)  Incidences of pressure ulcers: number of residents with a pressure ulcer in relation to the total number of residents   [43,46,49,56].
The quasi-experimental interventional study by Burgio et al. [62] compared two models of professional caregiver staffing numbers: a first group of 104 residents cared for by a permanent staff of 91 CNAs and a second group of 192 residents cared for by a rotating staff of 178 CNAs. There were no major significant differences in the characteristics of the residents and CNAs available for comparison except for the permanent or rotating staffing models.
Given the heterogeneity of the data included in our selected studies, it was impossible to carry out a meta-analysis of their groups or subgroups.

Methodological Quality of the Studies
The overall methodological quality of the cohort studies included in this review was poor-to-moderate ( Table 2).
None of our included studies was rated as having a high total methodological quality score (eight to nine stars): five were given a moderate score (six to seven stars) [43,46,50,53,55], and five scored as having poor methodological quality (fewer than six stars) [49,56,[58][59][60]. The different domains of evaluation used on the quasi-experimental interventional study by Burgio et al. [62] were scored as having methodological quality ranging from a weak to moderate risk of bias, with an overall risk of bias classed as moderate.

Description of the Staffing Levels in the Studies
The authors of the selected studies used different means of defining staffing levels, generally referring to national laws and standards. Several authors collected data on HPRD, which is calculated by dividing the total number of hours worked by all the professional caregivers in an LTRCF over 24 h by the number of residents [43,46,49,53,56,58] (Table 3).
Konetzka et al. [49] looked at RNs' HPRD and the staffing skill mix, defined as RNs' proportion of the total hours of care given by RNs, LPNs, and CNAs. RNs' mean HPRD was 0.35 (SD = 0.22), and their mean proportion of the skill mix was 0.12 (SD = 0.06) [49]. Shin et al. [56] looked at the HPRD of RNs, CNAs, and qualified care workers (QCWs), who, according to some authors, correspond to the USA's definition of CNAs [56].
Yoon et al. [59] defined the nurse staffing level as the number of nursing staff (RNs and LPNs) per 100 beds. They also looked at the ratio of RNs, defining that as the ratio of RNs to all nursing staff. The mean nurse staffing level was 15.58 nurses per 100 beds, and the ratio of RNs was 0.56, suggesting that RNs made up about half of the total nursing staff [59].
Kwong et al. [50] examined the number of full-time CNAs per 100 residents. Across the four LTRCFs in their sample, the mean was 14.85 (SD = 9.85) [50]. Note: A study could be awarded a maximum of one star for each numbered item in the Selection and Outcome categories. A maximum of two stars could be given for Comparability. Studies were evaluated on a scale from 0 to 9 stars and classified into groups of low (<6 stars), moderate (6-7 stars), or high (8-9 stars) quality. X* = X star. Table 3. Synthesis of the HPRD. * HPRD data were collected for RNs, LPNs, and CNAs for five distinct groups of residents; ** professionals defined as CNAs in the study; *** professionals defined as QCWs, but equivalent to CNAs.
Finally, German law states that RNs must make up at least 50% of the nursing staff in LTRCFs [55,60]. Zimmerman et al. [60] defined staffing levels as the ratios between the number of full-time equivalents (FTEs) for each type of caregiver and the residents. The proportion of RNs in their sample of LTRCFs ranged from 31.6% to 90.6%, with a mean of 56.7% [60]. Popp et al. [55] considered the proportion of FTE staff who were active, qualified personnel caring for residents. The proportion of qualified personnel in each different establishment ranged from 46% to 75%, with a mean of 58.1% [55].

Clinical Outcomes
Four studies examined associations between working conditions and the development of pressure ulcers [49,50,55,58] (Table 4), with three of them evaluating relationships between nurse staffing levels or their HPRD and the development of pressure ulcers [49,50,58]. Konetzka et al. [49] indicated that an increase in RNs' HPRD significantly reduced the probability of developing pressure ulcers (p = 0.01). The total number of hours worked by care staff, i.e., the combined hours of RNs, LPNs, and CNAs, did not have a statistically significant influence on the development of pressure ulcers in the studies by Konetzka et al. [49] (p > 0.05) or Temkin et al. [58] (p = 0.61). Two studies looked at nursing staff's levels of qualification and the development of pressure ulcers [50,55]. In the study by Kwong et al. [50], residents in nursing homes where there were RNs had a significant 26% lower probability of developing pressure ulcers (p ≤ 0.001). Finally, one study examined the relationship between the prevalence of pressure ulcers and teamwork and several managerial aspects [58]. Temkin et al. [58] revealed a significant reduction in the probability of developing pressure ulcers in LTRCFs displaying better team cohesion (p = 0.03) and greater nursing autonomy (p = 0.03).
Three studies investigated associations between working conditions and residents' urinary function or the development of urinary infections [49,58,59] (Table 4). In the study by Konetzka et al., increases in RNs' HPRD and the combined number of hours worked by RNs, LPNs, and CNAs led to statistically significant reductions in the probability of developing urinary infections (p = 0.01) [49]. According to Yoon et al. [59], an increase in the standard deviation (0.19) of RN staffing levels led to a significant 80% increase in the probability of improved or stable urinary incontinence (p = 0.02). By contrast, the combined number of hours worked by RNs, LPNs, and CNAs in the study by Temkin et al. [58], and the combined staffing levels of RNs and LPNs per 100 beds in the study by Yoon et al. [59], had no statistically significant influence on the probability of urinary incontinence (p = 0. 22) or the probability of improved or stable urinary incontinence (p = 0.37), respectively. A 0.23 increase in the standard deviation of the team cohesion score in the study by Temkin et al. [58], reduced the probability of incontinence by a statistically significant 7.6% (p < 0.001).
Linn et al. [53] evaluated associations between the respective numbers of hours worked by RNs, LPNs, and CNAs and nursing home residents' health status over a period of six months. Only the number of hours worked by RNs had a significant positive influence on residents' health status, as evidenced by the fact that the LTRCFs where RNs worked the most hours had lower mortality rates, less deterioration in residents' health status, and fewer hospital admissions (p < 0.05) ( Table 4).
Zimmermann et al. [60] examined associations between the number of residents per RN, the number of residents per CNA, and residents' weight loss. One extra resident per RN significantly increased the probability of residents losing weight by 2.3 times (p ≤ 0.01), whereas one extra resident per CNA had no statistically significant influence on weight loss (p ≥ 0.05) ( Table 4).    Structural and organizational factors, such as bed occupancy rates, nursing home size, whether the institution was private or public, whether it was in an urban or rural location, and whether nursing staff were assigned to residents on a permanent or rotating basis, had no statistically significant influence on the development of pressure ulcers, urinary infections, changes in urinary function, health status, or weight loss among residents [49,53,[58][59][60] (Table 4).
Finally, the quasi-experimental interventional study by Burgio et al. [62] sought to differentiate between the effects on residents of having permanent or rotating CNA staffing assignments, especially by looking at the spoken interactions between caregivers and residents, as well as at residents' disruptive behaviors, hygiene, appearance, and their self-perceived emotions. This study only found a statistically significant difference between staffing systems for the quality of care [62]. Higher scores were also noted for residents' personal appearance and hygiene under the permanent CNA staffing model (p = 0.04) [62] ( Table 5).

Process-Related Outcomes
Two studies investigated associations between the HPRD of RNs, LPNs, and CNAs and the number of deficiencies linked to care or the quality of care (QoC) declared in each LTRCF [43,46] (Table 4). Kim et al. [46] reported that an increase in the total number of nursing hours worked (RN plus LPN plus CNA hours), significantly reduced the total number of deficiencies (p < 0.001), the number of deficiencies linked to the QoC (p < 0.001), and the number of severe deficiencies altering the safety of care (p < 0.05). By looking at the hours worked by each professional group, a significant reduction in the total number of deficiencies and the number of deficiencies linked to the QoC was observed as more hours were worked by RNs (p < 0.001 and p < 0.01, respectively), by LPNs (p < 0.001), and by CNAs (p < 0.001) [46]. The number of hours worked by RNs and LPNs had no statistically significant effect on the number of severe deficiencies altering the safety of care (p > 0.05), whereas an increase in the number of hours worked by CNAs significantly reduced the number of severe deficiencies (p < 0.05) [46]. However, Hyer et al. [43] considered the joint influence of the hours worked by LPNs and RNs together (and separately from those worked by CNAs) on the total number of declared deficiencies and the number of deficiencies linked to the QoC. It highlighted, on the contrary, that the only variable associated with a statistically significant reduction in deficiencies linked to the QoC was an increase in the number of hours worked by CNAs (p = 0.02).  Finally, Shin et al. [56] looked at the associations between the HPRDs of RNs, LPNs, and CNAs, the skill mix (the ratio of RNs to LPNs, and the ratio of RNs to CNAs), staff turnover, and 15 other indicators of the QoC (the prevalence of falls, pressure sores, aggressive behavior, depression, cognitive decline, incontinence, urinary tract infection, weight loss, dehydration, tube feeding, bed rest, activities of daily living, residents' range of motion, antidepressant or sleeping pill use, and the need for physical restraint). A one-hour increase in the HPRD of RNs was associated with a statistically significant 3.9% lower rate of depression among residents (p = 0.002), a 5.7% lower prevalence of bedridden residents (p = 0.05) and a 1.1% lower use of physical restraints (p = 0.02) [56]. Furthermore, LTRCFs employing more RNs than LPNs observed significantly lower levels of aggressive behavior (p = 0.03), depression (p = 0.02), weight loss (p = 0.03), and being bedridden among their residents (p = 0.04) [56]. A greater ratio of RNs to CNAs was significantly associated with residents suffering less weight loss (p = 0.05) [56]. Finally, a significant positive statistical relationship was observed between the administration of antidepressants and sleeping pills and RN staff turnover (p < 0.001) and LPN staff turnover (p = 0.02), whereas no significant associations were noted between CNA staff turnover and different indicators of QoC [56]. When RN staff turnover rose by 5.9%, the prevalence of residents taking antidepressants or sleeping pills rose by 27.2%, whereas when LPN staff turnover rose by 9.7%, the prevalence of residents taking antidepressants or sleeping pills rose by 18% [56] (Table 4).

Discussion
The present systematic review aimed to identify cohort and experimental studies exploring associations between the working conditions of nursing staff and the quality of care (QoC) received by older-adult residents living in LTRCFs. We identified and incorporated ten cohort studies and one quasi-experimental interventional study into our review, covering a total of 64,139 residents and 406,632 observations. These combined pieces of research helped us to distinguish the influence of nursing staff's working conditions on two types of results: residents' clinical outcomes and results linked to processes and care pathways.
Regarding residents' clinical outcomes, higher overall rates of nursing staff's total HPRD were associated with the significantly better prevention of poor clinical outcomes such as the development of pressure ulcers or urinary tract infections (UTIs). Specifically, the greater the number of hours worked by registered nurses (RNs) or the greater the number of RN staff employed, the greater the real positive impacts on the different clinical outcomes measured among residents, notably in preventing the development of pressure ulcers and UTIs, improving urinary function and general health status, and reducing hospital admissions and the mortality rate. However, this was not true for licensed practical nurses (LPNs) and certified nursing assistants (CNAs). The importance of nursing staff's qualification levels was also observed because RNs' specific skills and knowledge were associated with greater positive influences on preventing the development of pressure ulcers and UTIs and improving urinary function than were those of LPNs and CNAs. The number of residents cared for per member of the nursing staff was also an important factor because an increase of one resident per RN was associated with a significantly higher risk of weight loss among those residents. Certain organizational aspects, such as effective teamwork, good team cohesion, and more nursing autonomy, were associated with positive impacts on residents' clinical outcomes. Other organizational factors, such as permanent or rotating staff assignments to residents or the LTRCF's occupancy rate, had no influence on residents' clinical outcomes, nor did certain structural factors such as the size of LTRCFs, whether they were privately or publicly run, and whether they were situated in urban or rural areas.
With regards to results linked to care processes, the importance of higher total HPRD for all nursing staff was also highlighted because this was favorably associated with lower numbers of deficiencies linked to care or to the QoC, as well as with lower numbers of severe deficiencies declared by LTRCFs. The QoC was ensured by RNs' specific contributions to improving QoC indicators. Finally, faster staff turnover was associated with a significant negative impact on QoC indicators.
The present systematic review had some limitations. Despite a thorough literature search using recognized guidelines and recommendations on methodology, our review may have missed some studies which met all the selection criteria due to study search errors or investigator mistakes. Three of the studies selected used the Online Survey, Certification, and Reporting (OSCAR) database [65] to collect data on nursing staff's HPRD and the structural characteristics of the LTRCFs participating. However, OSCAR's accuracy and validity, in these studies, were somewhat contested. Indeed, nursing professionals' HPRDs were only calculated over a two-week period, which may not have been adequately representative of their true HPRD over a longer timeframe. Eight studies evaluated residents' clinical outcomes using data reported by nursing staff themselves, which creates a risk of bias. In addition, the selected studies predominantly used nursing staff's HPRD as the independent variable of interest, which may have led to an over-representation of this variable compared to other factors influencing the QoC. It is also difficult to draw any conclusions on the influence of the structural characteristics of LTRCFs as most of the studies did not explore the direct impacts of those variables on the QoC; instead, they used them as control variables during statistical analyses. Furthermore, there was a lot of heterogeneity in the follow-up periods chosen by the different cohort studies, varying between four months and nine years. Moreover, none of the cohort studies was given a high score for the quality of its methodology: five were considered moderate and four were of poor methodological quality. The one quasi-experimental interventional study, for its part, had a moderate risk of bias. Finally, any generalization of the present findings should be made with caution as the LTRCFs studied were always representative of a particular region or country.
Overall, the present systematic review included ten cohort studies and one quasiexperimental interventional study examining large samples of LTRCFs, residents, and observations using accurate, valid measurement instruments. Furthermore, we used highly recommended methodological norms and guidelines, making our findings very reliable. To the best of our knowledge, no systematic reviews incorporating cohort and experimental studies have been published to date on how nursing staff's working conditions affect the QoC received by older adults living in LTRCFs. Other systematic reviews on this interesting topic mainly drew together studies of a transversal design, potentially biased by the numerous confounding factors inherent in such designs. The present systematic review thus helps to provide a higher level of proof.
In view of the small number of experimental studies in our field of interest to date, there is a need for further interventional research on the impact of nursing staff's working conditions on the QoC received by older adults living in LTRCFs. Providing safe, highquality care is the primary objective of all health care institutions. With a view to attaining continuous improvements in quality and safety, more research data on the relationship between nursing staff's working conditions and the QoC provided to residents would help to support recommendations to health care managers, supervisors, political decisionmakers, and other stakeholders involved in long-term care. More data would help to establish better working conditions, notably with a view to defining a standard minimum level of nursing staff necessary to ensure optimal care for older adults living in LTRCFs.
Finally, most of the studies identified in this systematic review underlined the tendency for LTRCFs to reduce their numbers of RNs and hire more LPNs and CNAs in order to reduce the overall costs of nursing personnel. However, most of these studies also pointed out the specific contributions of RNs in maintaining and improving the QoC. Thus, particular attention should be given to the presence of enough RNs in an LTRCF to supervise and monitor the care dispensed by their LPN and CNA colleagues. This approach will enable staff to better prevent adverse events, halt residents' worsening health statuses, and avoid the necessity of beginning burdensome treatments to heal pressure ulcers or infections-actions that, in themselves, will save institutions money in the long term.

Conclusions
To the best of our knowledge, the present systematic review is the first to have integrated longitudinal cohort and interventional studies exploring associations between nursing staff's working conditions and the QoC given to older adults living in LTRCFs. The review highlighted the predominant influence of human factors on the QoC. Higher overall nursing staff hours worked per resident per day, a suitable number of residents attributed to each caregiver, a reduction in staff turnover, as well as the specific contribution of enough working hours carried out by RNs, along with their special skills and knowledge, can all have a significant positive influence on residents' clinical outcomes and on results linked to the processes of care. Some organizational elements, such as effective teamwork, more cohesive care teams, and greater levels of nursing autonomy, were all associated with positive impacts on the QoC, whereas other organizational factors, such as assigning permanent or rotating members of staff to residents or the LTRCF's occupancy rate, only had a relatively small influence on the QoC. Structural factors (such as the size of the LTRCF, whether it was privately or publicly owned, and whether it was located in an urban or rural area) were only weakly associated with the QoC. In the end, it is essential that each LTRCF's supervisory board, management committee, or decision-making organ makes sure that it promotes optimal working conditions for its nursing staff because these valuable health care professionals have a direct impact on the QoC provided to residents. Particular attention should be given to ensuring that the overall nursing staff's HPRD is sufficient and that there are enough RNs in the mix of nursing professionals.