The Effect of Interventions on Preventing Musculoskeletal Injuries Related to Nurses Work: Systematic Review

Background: The 12-month prevalence of musculoskeletal disorders related to work (MDRW) in nurses rests between 71.8% to 84%, so it is urgent to develop preventive intervention programs with the purpose of avoiding negative physical, psychological, socioeconomic, and working aspects. There are several intervention programs aimed at preventing musculoskeletal disorders related to work for nurses, but few have successfully proven results. Despite the evidence pointing to the benefits of multidimensional intervention programs, it is essential to determine which interventions have positive effects on the prevention of this kind of disorder to create an effective intervention plan. Aim: This review intends to identify the different interventions adopted in the prevention of musculoskeletal disorders related to work in nurses and to compare the effectiveness of these interventions, providing the appropriate and scientific basis for building an intervention to prevent musculoskeletal disorders in nurses. Method: This Systematic Review was guided by the research question, “What are the effects of musculoskeletal disorders preventive interventions on nursing practice?” and carried out in different databases (MEDLINE, CINAHL, and Cochrane Central Register of Controlled Trials, SCOPUS, and Science Direct). Later, the results were submitted to the eligibility criteria, the appraisal quality of the papers, and the data synthesis was performed. Results: 13 articles were identified for analysis. The interventions implemented to control the risk were: training patient-handling devices; ergonomics education; involving the management chain; handling protocol/algorithms; acquiring ergonomics equipment; and no-manual lifting. Conclusions: The studies associated two or more interventions, the majority of which (11 studies) were training-handling devices and ergonomics education, therefore emerging as the most effective instruments in the prevention of MDRW. The studies did not associate interventions that cover all risk factors (individual, associated with the nature of the work, organizational, and psychological aspects). This systematic review can help with making recommendations for other studies that should associate organizational measures and prevention policies with physical exercise and other measures aimed at individual and psychosocial risk factors.


Introduction
Musculoskeletal disorders are a complex health problem transversal to all sectors of activity worldwide. The European institutions with responsibility for health and work have expressed their concern and provided guidelines for their control due to the risk of these injuries becoming pandemic, with repercussions on the economy of the different countries, including the increasing costs in the health systems [1].
In view of the above, the aim of this review was to identify the different interventions adopted in the prevention of musculoskeletal disorders related to work in nurses and to compare the effectiveness of these interventions, providing the appropriate and scientific basis for building an intervention to prevent musculoskeletal disorders in nurses praxis.

Study Design
Given the purpose of the study and the state of the art of the phenomenon under study, a systematic review (SR) was chosen [21,22]. The option for a review with a scientific and systematic methodology is justified by the need to have reliable results from which conclusions can be drawn and decisions made, minimizing the risk of bias and guiding clinics and health policies based on research results [21,23]. The protocol to guide the SR [22] was prepared and agreed upon in December 2021, and it was registered on Prospero with ID No. CRD42022331581 in May 2022.
The research question that guided the definition of eligibility criteria and the research strategy was: What are the effects of MDRW preventive interventions on nursing practice?

Eligibility Criteria
Inclusion criteria were primary experimental and epidemiological study designs (RCT, non-RCT, quasi-experimental, cohort studies, case-control studies, and analytical crosssectional studies), which measure interventions for the prevention and/or reduction of musculoskeletal disorders related to nurses' work. There were no restrictions on country or year of research, but they were limited to studies published in Portuguese, English, and Spanish. Reports, such as unpublished manuscripts and conference abstracts, are not eligible for inclusion.
Exclusion criteria were systematic review, studies with qualitative design or protocols, studies in which the target population is only nursing assistants or home nurses, studies in which the intervention is for treatment or rehabilitation of injuries or illnesses of MDRW, and studies with intervention in the multidisciplinary team that does not identify nurses' results.
The main outcomes appraised were incidence and prevalence of musculoskeletal disorders, absenteeism rate, related pain, back, upper limbs, shoulders, and neck loading, adherence to safe behaviors from a biomechanical point of view, and acceptance and adhesion to the program by nurses.

Research Strategy
The first stage of the search began in August 2021, carried out with natural language terms in association with the medical subject headings-MeSH and SCoR guidelines, conducted on the EBSCO host platform in the MEDLINE, and CINAHL databases. The terms were associated with the Boolean operators OR and AND at the junction of the descriptors identified in PICO.
This first research allowed us to identify the keywords and descriptors used by the indexing of articles and to raise awareness of current scientific knowledge, such as helping the elaboration of the study protocol.
The second stage of the research started in January 2022 and lasted until May of the same year; the survey was conducted on the EBSCO host platform (MEDLINE, CINAHL, and Cochrane Central Register of Controlled Trials), SCOPUS, and Science Direct.
The last research was conducted in May of 2022. An inspection of the bibliographic references of the articles was carried out to identify systematic reviews that report and guide the future research of systematic reviews for the important conclusions about the topic.

Data Extraction, Quality Appraisal, and Data Synthesis
After identifying all the articles in the different databases, they were transferred to EndNote, Clarivate Analytics, Philadelphia, United States, to recognize duplicate articles and eliminate them. For the calculation of the relevance of the article, we transfer all articles in the "RIS" format to "Rayyan".
The process of data extraction started by analyzing the title and synopsis of all articles based on the selection criteria initially defined. This process was carried out by the two reviewers independently, in case of doubts, and they were clarified through a third reviewer.
An excel file was built to extract the results, which was carried out by the same researchers. Each article was summarized and organized according to the following items: study identification (author, year of publication, and country), objective, type of study, sample, and results.
The risk of being biased, the instruments RoB2, quality assessment of before-after studies with no control group, and the quality assessment of papers describing observational and quasi-experimental studies that were used [21,24] were assessed by the team of researchers. Given the heterogeneous nature of the study designs, a narrative synthesis was chosen to answer the research question. For the quality of the evidence, we considered a confidence interval of 95%.

Results
A total of 48 duplicate articles (of the 165 articles submitted) were identified. After analyzing the title and synopsis, we excluded a total of 103 records for not complying with the inclusion criteria and were left with 13 articles to identify their eligibility. 6 of 26 The reasons for the exclusions of the 103 articles were: 28 articles were excluded for analyzing the wrong population (nursing assistants, home nurses, bus drivers, patients, orthodontists, industrial workers), 43 articles for wrong outcomes (treatment of chronic musculoskeletal disorders, prevalence, effects of the workplace, others occupational disorders: stress, anxiety, and dermatology), 31 articles for study design (systematic reviews, literature review, and protocols), and one for foreign language (Arabic writing).
Finally, we had 13 articles for analysis. The research diagram and the study selection process can be seen in Figure 1. literature review, and protocols), and one for foreign language (Arabic writing Finally, we had 13 articles for analysis. The research diagram and the stud process can be seen in Figure 1.
One article was eliminated after the eligibility assessment for low quality of Of the selected articles, six are from the USA and three from Canada, the rema cles are from China, Iran, Germany, and Vietnam.
Regarding the year of publication, we can see that the oldest article was pu 2001 and that 38.5% of the articles were published in the last five years (one artic one article in 2020, two in 2021, and one in 2022).
This diffusion of countries and the growing number of publications on demonstrates the interest and importance that this topic has for the scientific c worldwide.  Table 2 presents the extraction of the results of the 13 articles, identifying tives, the type of study, the time of evaluation of the intervention, the evaluati ment, and its results, and finally, the conclusions of the articles. One article was eliminated after the eligibility assessment for low quality of evidence. Of the selected articles, six are from the USA and three from Canada, the remaining articles are from China, Iran, Germany, and Vietnam.
Regarding the year of publication, we can see that the oldest article was published in 2001 and that 38.5% of the articles were published in the last five years (one article in 2017, one article in 2020, two in 2021, and one in 2022).
This diffusion of countries and the growing number of publications on this topic demonstrates the interest and importance that this topic has for the scientific community worldwide. Table 2 presents the extraction of the results of the 13 articles, identifying the objectives, the type of study, the time of evaluation of the intervention, the evaluation instrument, and its results, and finally, the conclusions of the articles. The baseline was on 1 July 1998, and 6-months, and again at 12-months. The participants 346 nurses Outcome: The frequency of manual patient-handling tasks, i.e., those tasks during which neither mechanical nor nonmechanical assistive equipment is used, was significantly and markedly decreased in the Arm C "no strenuous lifting" wards by 6 months, by an average of 9 tasks per shift. This decrease was sustained at 1 year. In contrast, there was no significant change over 6 months in frequency of manual handling tasks on Arm A or Arm B. By the time of the 6-month follow-up, there was a significant increase in use of assistive devices such as transfer belts and sliding devices on Arm B; however, this increase was not sustained at 1 year. The use of these manual assistive devices increased significantly (P 5 0.021) by 6 months on Arm C, but by 1 year use declined significantly. At 6 months nurses on Arm C reported using the sit-stand lift an average of 4.9 times per shift. This declined significantly to 3.2 times per shift at 1 year. The use of total body lifting equipment over time varied significantly by service type. Instruments: Visual analogic scales (VAS); SF36; Oswestry Low Back Pain Disability Questionnaire; Disability of Arms, Shoulder, and Hands (DASH).
After 6 months increased the using of lifting devices, but there was a decrease in one year. The "no strenuous lifting" program, which combined training with assured availability of mechanical and other assistive patient handling equipment, most effectively improved comfort with patient handling, decreased staff fatigue, and decreased physical demands. The fact that injury rates were not statistically significantly reduced may reflect the less sensitive nature of this indicator compared with the subjective indicators. This multi-faceted program resulted in positive outcomes associated with injury rates, modified duty days, job satisfaction, self-reported safety in performing patient handling Outcomes: The number of injuries by occupation showed that in the control group, the distribution remained unchanged with the exception of therapists (physical therapists, occupational therapists, respiratory therapists) where a significant increase was seen. The most significant change was seen in the decrease in injuries in attendants (from 25.4% to 0%) and increase in injuries in nurse aides (from 1.1% to 11.3%) in the intervention group. Instruments: Analysis of all injuries and time-loss rates.
Significantly reduce both time-loss and no-time-loss injuries and disability related to patient handling. The reductions of claim costs/injury represented a substantial benefit to the intervention hospitals. The program intervention seemed to be more effective in the small hospitals than in the medium or large ones. . No significant difference in loss to follow-up was found between the two groups (χ 2 = 0.074, p = 0.862). GEE showed that the multidimensional intervention program improved the risk perception of WRMDs (OR = 0.517, p < 0.001) and health behavior application (OR = 0.025, p < 0.001), relative to that of the routine specialist training. Interactions between the measurement time and group were observed (p < 0.001).
Age and the length of ICU employment affected the perception of a safe working environment (p = 0.047 and p = 0.011 respectively). The GEE, including age and ICU employment, indicated that the measures of the intervention group and the control group were statistically significant. The perception of an unsafe working environment in the control group was 1.637 times that in the intervention group (OR = 1.637, p = 0.024).
A meticulous planning is essential to make interventions compatible with the daily work routine. The multidimensional intervention program seems applicable from time, financial, and organizational perspectives, he helped to reduce the short-term reported incidence rate of WRMDs, improve the nursing risk perception and health behavior application, and promote a safe working environment. Outcomes: 46 were in the intervention group (53% survey response rate) and 29 (39% survey response rate) were in the control group. There were no significant demographic differences between participants in the intervention and control groups. The combined sample (n = 75) was 95% female and ranged in age from 21 to 59 years (mean 33.7). Nursing staff members from the multifaceted minimal-lift environment experienced a reduction in patient-handling injuries and costs compared to nursing staff working in a non-minimal-lift environment. The intervention unit injury incidence rate was 3.26/100 full-time equivalents (FTEs) whereas the control unit injury incidence rate was 3.43/100 FTEs. Injury costs for the intervention unit were $6566 compared with $11,145 for the control unit (a $4579 difference). After subtracting the cost of peer coach education ($1680), the intervention unit experienced a $2899 return on investment Instruments: Demographic and equipment use data for the intervention and control units were collected through self-report via pen-and-paper survey.
Implementing a successful multifaceted minimal-lift environment for nursing staff can be time consuming and complex. Intervention strategies must match the innovation, target group, and workplace context, and could expand program elements to include ergonomic assessment protocols, after action reviews. Prior to the intervention, there was no significant difference between the two intervention groups and one control group. The repeated measure analysis test confirmed that was important and significant difference 6 and 12 months after the Intervention. There was a significant interaction between the factors "group" and "test time" (p < 0.05, p < 0.001).
Increasing the mean score of attitude, knowledge, perceived self-efficacy, enabling factors, reinforcing factors, quality of life, public health, and preventive behaviors of LBP in intervention group (p < 0.05, <0.001), but no significant change in mean score of knowledge, attitude, Self-efficacy, quality of life, general health, reinforcing Health behaviors require context and access to education through the best and easiest channels, which seems to be appropriate for social media. Different educational approaches can be effective in reducing low back pain, disability and improving the health care workers life. The social media approach has been more successful than long-term face-to-face intervention and may be a better way to deliver training programs because of its ease of access and reduced operating costs factors, enabling factors and preventive behaviors of LBP in the control group (p > 0.05). Instruments: Visual Analogue Scale (VAS) for measuring LBP, for measuring pain-related disability, the Quebec Back Pain Disability Scale (QBPDS) was used. The social media approach to maintaining behavior for a long time (6 months) was more successful than the face-to-face approach However, in 44.2% of situations, the bed was partially raised, and in 11.1%, the bed was not raised at all. In total, 52 care situations in the bathroom were observed. A stool was used in 67.3% of these situations to perform basic care in the sitting position; in 32.7% of the situations, the stool was not used by the nurses. Instruments: The CUELA measurement system and video analyses were used to evaluate this intervention.
This study showed a significant improvement in body postures after implementation of a training concept consisting of instruction on frequent body postures in nursing, physical exercises, instructions in practical ergonomic work at the bedside and in the bathroom, and reorganization of work environment.   Instruments: At the training sessions, participants were given baseline ergonomic surveys (levels of musculoskeletal comfort of different body parts, the presence and severity of pain, and the levels of physical and mental exhaustion experienced "at the end of a typical workday"). Other questions are information on several aspects of their work, such as the degree of support received from their supervisors, the amount of time available to complete tasks, and their level of job satisfaction.
Many nursing staff are reluctant to use mechanical lifts for patient handling tasks, the main reason reported was the lack of perceived need, followed by the lack of time and the lack of maneuvering space. They suggest that must be encouraged by management, and a policy of no manual lifting must be adopted.  The intervention measures are probably effective in reducing the prevalence of MSDs at these neck, shoulder/upper arm, wrists/hand, and lower back. One of the limitations of the educational intervention in this study is that it only provides of theoretical information and knowledge to nurses but does not monitor the application of these measurements by nurses in their actual work.

Quality of the Evidence
Three instruments were used to assess the bias of the 13 studies. Two RCTs have been evaluated by RoB2.0 (Table 3) and five studies for the quality assessment of before-after studies with no control group (Table 4) and the quality assessment of papers describing observational and quasi-experimental studies were the instrument use for the other six studies (Table 5).

Quality of the Evidence
Three instruments were used to assess the bias of the 13 studies. Two RCTs have been evaluated by RoB2.0 (Table 3) and five studies for the quality assessment of before-after studies with no control group (Table 4) and the quality assessment of papers describing observational and quasi-experimental studies were the instrument use for the other six studies (Table 5).  [25] some concerns some concerns Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomization process; D2-Risk of bias due to deviations from the intended interventions (effect of assignment to intervention); D3-Missing outcome data; D4-4: Risk of bias in measurement of the outcome and D5-Risk of bias in selection of the reported result.
The study by Pourhaji et al. [30], presented after its judgment through the ROB2: low risk of bias, and the study Yassi et al. [25], presents some concerns, since it presented in D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bias of the 13 studies. Two RCTs have evaluated by RoB2.0 (Table 3) and five studies for the quality assessment of beforestudies with no control group (Table 4) and the quality assessment of papers descr observational and quasi-experimental studies were the instrument use for the othe studies (Table 5).  [25] some concerns some c cern Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomization process; D2-Risk of bias due to d tions from the intended interventions (effect of assignment to intervention); D3-Missing ou data; D4-4: Risk of bias in measurement of the outcome and D5-Risk of bias in selection reported result.
The study by Pourhaji et al. [30], presented after its judgment through the ROB2 risk of bias, and the study Yassi et al. [25], presents some concerns, since it present D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bias of the 13 evaluated by RoB2.0 (Table 3) and five studies for the quali studies with no control group (Table 4) and the quality asse observational and quasi-experimental studies were the inst studies (Table 5).  [25] some concerns Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomization process tions from the intended interventions (effect of assignment to inter data; D4-4: Risk of bias in measurement of the outcome and D5reported result.
The study by Pourhaji et al. [30], presented after its judg risk of bias, and the study Yassi et al. [25], presents some co D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bia evaluated by RoB2.0 (Table 3) and five studies fo studies with no control group (Table 4) and the observational and quasi-experimental studies w studies (Table 5).  [25] some co cerns Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomiza tions from the intended interventions (effect of assignm data; D4-4: Risk of bias in measurement of the outco reported result.
The study by Pourhaji et al. [30], presented a risk of bias, and the study Yassi et al. [25], prese D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bias of the 13 studies. Two RCTs have been evaluated by RoB2.0 (Table 3) and five studies for the quality assessment of before-after studies with no control group (Table 4) and the quality assessment of papers describing observational and quasi-experimental studies were the instrument use for the other six studies (Table 5).  [25] some concerns some concerns Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomization process; D2-Risk of bias due to deviations from the intended interventions (effect of assignment to intervention); D3-Missing outcome data; D4-4: Risk of bias in measurement of the outcome and D5-Risk of bias in selection of the reported result.
The study by Pourhaji et al. [30], presented after its judgment through the ROB2: low risk of bias, and the study Yassi et al. [25], presents some concerns, since it presented in D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bias of the 13 studies. Two RCTs have evaluated by RoB2.0 (Table 3) and five studies for the quality assessment of beforestudies with no control group (Table 4) and the quality assessment of papers descr observational and quasi-experimental studies were the instrument use for the othe studies (Table 5).  [25] some concerns some c cern Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomization process; D2-Risk of bias due to d tions from the intended interventions (effect of assignment to intervention); D3-Missing ou data; D4-4: Risk of bias in measurement of the outcome and D5-Risk of bias in selection reported result.
The study by Pourhaji et al. [30], presented after its judgment through the ROB2 risk of bias, and the study Yassi et al. [25], presents some concerns, since it present D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bias of the 13 studies. Two evaluated by RoB2.0 (Table 3) and five studies for the quality assessme studies with no control group (Table 4) and the quality assessment of p observational and quasi-experimental studies were the instrument use studies (Table 5).  [25] some concerns Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomization process; D2-Risk of tions from the intended interventions (effect of assignment to intervention); D3data; D4-4: Risk of bias in measurement of the outcome and D5-Risk of bias reported result.
The study by Pourhaji et al. [30], presented after its judgment throu risk of bias, and the study Yassi et al. [25], presents some concerns, sinc D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bias of the 13 evaluated by RoB2.0 (Table 3) and five studies for the quali studies with no control group (Table 4) and the quality asse observational and quasi-experimental studies were the inst studies (Table 5).  [25] some concerns Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomization process tions from the intended interventions (effect of assignment to inter data; D4-4: Risk of bias in measurement of the outcome and D5reported result.
The study by Pourhaji et al. [30], presented after its judg risk of bias, and the study Yassi et al. [25], presents some co D3 some concerns.

Quality of the Evidence
Three instruments were used to assess the bia evaluated by RoB2.0 (Table 3) and five studies fo studies with no control group (Table 4) and the observational and quasi-experimental studies w studies (Table 5).  [25] some co cerns Pourhaji et al.
(2020) [30] Legend: D1-Risk of bias arising from the randomiza tions from the intended interventions (effect of assignm data; D4-4: Risk of bias in measurement of the outco reported result.
The study by Pourhaji et al. [30], presented a risk of bias, and the study Yassi et al. [25], prese D3 some concerns.

Quality of the Evidence
Three instruments were used to a evaluated by RoB2.0 (Table 3) and fiv studies with no control group (Table  observational and quasi-experimenta studies (Table 5).  [25] Pourhaji et al.
The study by Pourhaji et al. [30], risk of bias, and the study Yassi et al D3 some concerns. The study by Pourhaji et al. [30], presented after its judgment through the ROB2: low risk of bias, and the study Yassi et al. [25], presents some concerns, since it presented in D3 some concerns.
Eleven studies associate training-handling devices with ergonomics education [19,[25][26][27][28][29][31][32][33]36] and information on risk factors, use of mechanical devices, and use of the principles of body mechanics during the use of mechanical means, but also in carrying out other activities.

Discussion
The 12 studies in this review are heterogeneous from the point of view of study design, sample size, implemented intervention, assessment instruments, measured outcomes, and contexts where the study was carried out, which does not allow for meta-analysis. Nevertheless, the results allow the evaluation of the methodological quality of the studies and the evidence of the interventions that each study used to control the risk of musculoskeletal injury.
This SR made it possible to identify which interventions prevent MDRW in nurses and synthesize the evidence on which interventions were implemented, their feasibility, and their impact on different outcomes, with special relevance to the prevalence of MDRW. This type of injury is a global and transversal problem in almost all professions, but it assumes a worrying incidence and prevalence among health professionals, especially in nurses [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38] due to the very nature of the professional activity, with the need to mobilize and transfer patients with a high degree of dependence, performing activities in positions that imply dorsiflexion and torsion of the spine, lifting weights above the recommended for the anthropometric characteristics of the professional [37][38][39][40], few rest periods between activities that demand of high physical effort, or the maintenance of painful postures for a long time [37][38][39].
The studies included reinforce the importance of programs aimed at learning the correct handling of patients and/or using the principles of biomechanics in carrying out this activity [19,[25][26][27][28][29][30][31][32][33]36], the use of mechanical means that reduce overload, and the associated risk [19,[25][26][27][28][29][31][32][33][34][35][36], including the policy of no manual lifting [27]. These data are consistent with the recommendations of the Occupational Safety and Health Administration (OSHA) which recommends the increased availability of assistive devices and the use of even the most basic assistive devices as an integral part of safe patient handling [40].
Only two programs involved the management chain [27,32], and the results show an impact of the intervention in reducing the incidence of injuries and absenteeism, which reinforces the authors' recommendations for the development of a safety culture in institutions providing care for health, ensuring knowledge, skills, and competencies for the prevention of injuries in its professionals [36,38,40]. Although organizational cultural changes take time, this is equally applicable to safety culture [40], and it's urgent that healthcare provider settings are increasingly dynamic work environments that benefit from strong organizational programs, policies, and practices around risk identification and reduction [39].
It should be noted that the results of some programs, in relation to the impact on the reduction of injuries and the adoption of safe behaviors, do not observe gains, for example, in the use of equipment or changes in practices [25,28], leaving the question of whether the involvement of the organization with clear policies for the clinic of professionals, guaranteeing the safety of the patient and the professionals, and adhesion to the programs would increase since the beneficial effects of safe patient-handling programs improved over time, which highlights the importance of a long-term and continued effort to make the necessary cultural changes [41].
We corroborate that interventions should take into account not only the ergonomics but also the improvement of the organizational aspects of the work environment [37], but other aspects should be explored, such as communication with the patient to actively involve them in the procedures and promote their rehabilitation, or even involving the caregiver [42], previously planning the activity by unblocking the space around the patient's bed, ensuring an optimization of the interaction between the health professional and the patient [37] or his caregiver [40], and the health professional and the environment [36].
The studies evaluated two or more interventions simultaneously, which goes against the key idea that risk control is carried out through the implementation of systemic and multifactorial programs [37]. Multiple approaches are needed to put changes in practice and to promote a safety culture, including workflow processes, ongoing training, skills, supervising, and communication between professionals about risk [37][38][39][40]. It is necessary to encourage in the units the choice of facilitators to teach, change behaviors, and monitor the appropriate use of mobility aids [40] and also to promote adequate training to improve the knowledge and skills of the nursing staff in the handling of dependent patients [43].
Outcomes focused on incidence, prevalence, self-perceived frequency and intensity of physical discomfort, musculoskeletal pain, time-loss rates, risk perception, and perception of a safe working environment [19,[25][26][27][28][29][30][31][32][33][34][35][36]. It is suggested that future research explores nurses' adhesion to the programs. A study that aimed to provide a systematic review of the international literature, synthesize knowledge, and explore factors that influence nurses' adhesion to patient-safety principles concluded that patients' participation, healthcare providers' knowledge and attitudes, a collaboration by nurses, appropriate equipment and electronic systems, education and regular feedback, and standardization of the care process influenced nurses' adherence to patient-safety principles [44].
Interventions essentially focused on preventive measures aimed at risk and not at promoting the health of professionals. Given the multifactorial nature of risk factors for falls, individual, psychosocial, organizational, and socio-economic, we share the opinion of other studies that recommend preventive measures including physical exercise for muscle strengthening, food education to maintain weight, cognitive-behavioral strategies to control anxiety, and investing in a good work environment to control psychosocial risk factors [41,[45][46][47][48][49][50].
Some studies identify the ineffectiveness of training only one factor [22], and others showed that multifactor training (transfer, lifting, and repositioning), and the multiple interventions (education and training, zero lift policy, provision of assistive devices for patient support and care, individual measure, etc.) are emerging as the most effective instruments in the prevention of MDRW [16,24,25,27,30].
These multidimensional intervention programs reduce the self-reported performance of "unsafe" working environments [16,24], decrease time-loss/injury days, modify duty days, increase job satisfaction, and decrease workers compensation costs [24,25,30,33]. A peer leader program is much more effective than traditional educational approaches and facilitates the implementation of the program, as well as being sustainable over time [24], especially in small hospitals [25].
The procede-proceed model has a significant effect on behaviors as a factor that increases the quality of lifestyles of low back pain (LBP) [27]. Theoretical education was effective in improving knowledge, attitude, and self-efficacy, reinforcing and enabling factors, and behavior immediately after 6-12 months of intervention [27]. Other studies conclude that the effective goal of reducing MDRW is the combining of theoretical education with ergonomics practice [33]. The social media approach to maintaining behavior for a long period of time (6 months) was more successful than the face-to-face approach [27].

Study Limitations
The heterogeneous design of the quantitative studies, the different instruments used to evaluate the intervention, and their differences did not allow for meta-analysis and limits the evidence of this SR. In addition to this factor, the search was conducted only in four databases and the inclusion of studies in Portuguese, Spanish, and English may have excluded studies published in other languages that would have answered the research question.

Conclusions
The interventions implemented to control the risk of MDRW were training patienthandling devices, ergonomics education, involving the management chain, handling protocol/algorithms, acquiring ergonomics equipment, and no-manual lifting. The use of two or more interventions in association allowed a reduction in the prevalence of MDRW and associated symptomatology, increasing risk perception, decreasing frequency and intensity of physical discomfort, musculoskeletal pain, time-loss rates, and risk perception.
Combining theoretical education with ergonomics will be more effective in the goal of reducing MDRW, and the organization must implement appropriate policies to apply the intervention more effectively.
Future studies should associate organizational measures and prevention policies with physical exercise and other measures aimed at individual and psychosocial risk factors because the multifactorial nature of risk can only be controlled with multifactorial interventions with a combination of individual, psycho-organizational, and task-related measures.
In nursing education, both graduate and postgraduate, it s important to introduce curricula content on risk factors and preventive measures, enabling the student to adopt these measures in clinical practice. The simulated practice in the laboratory can be a good pedagogical strategy for the development of these competencies.
For the professionals who are in the clinic, the simulation and video recording of the posture and movements performed in the provision of care can allow the awareness of the individual risk associated with the nature of the professional activity.
The methodological quality of the studies is acceptable and makes it possible to make recommendations for the clinic, for training, and for research. It should be noted that the heterogeneity of the program's conditions, the robustness of the evidence, and the synthetic description of the interventions make it difficult to understand the whole of the program and, above all, how its implementation was carried out in terms of time, involvement of human resources, strategies for adherence to the program, and collaborative work within the contexts (or lack thereof). Funding: The present study was funded by the Center for Research, Innovation, and Development in Nursing, in Portugal, by means of grants provided to some of the authors.

Institutional Review Board Statement:
The present study was carried out according to the guidelines of the Declaration of Helsinki and approved by an Ethics Committee.