Introduction

The COVID-19 pandemic highlighted the relevance of the nursing profession in our society. During this time, the physical and psychological work challenges of nurses in different settings (hospital, outpatient, and geriatric) increased drastically, including reduced contact time between nurses and geriatric patients (Kramer et al., 2021; Rutten, Backhaus, Hamers, & Verbeek, 2022). Furthermore, the risk of infection was significantly higher in nurses than in other populations (Alshamrani, El-Saed, Al Zunitan, Almulhem, & Almohrij, 2021). Increasing work-related demands are linked to discontent with working conditions as well as increased risks for musculoskeletal disorders, high stress levels, low job satisfaction, and the desire to leave the profession (van der Heijden, Brown Mahoney, & Xu, 2019; Wollesen et al., 2019).

Nurse-specific stressors are manifold, including high workload, emotion suppression, time pressure, work disruptions, multiple responsibilities, shift work, heavy lifting and carrying, and unfavorable postures for prolonged periods (Rohwer, Mojtahedzadeh, Harth, & Mache, 2021). These stressors are typically linked to physical and psychological health problems (McVicar, 2003; Ruotsalainen, Verbeek, Mariné, & Serra, 2016). For instance, the risk of back pain is drastically higher in nurses than in other healthcare professions such as physiotherapists (Çınar-Medeni, Elbasan, & Duzgun, 2017; Dawson et al., 2007; Vinstrup, Jakobsen, & Andersen, 2020). Moreover, musculoskeletal diseases, sleep and concentration disorders, and dysfunctional behaviors (lack of exercise, drug and alcohol consumption) are frequently observed in nurses (Rohwer et al., 2021). To minimize perceived demands and job turnover, workplace health promotion programs (WHPPs) including physical exercise are necessary and should be implemented in nursing settings as an integral part of nursing facilities (Mojtahedzadeh et al., 2021).

To improve the work ability of nurses, resources such as health knowledge, stress management skills, health-promoting lifting and carrying behavior, as well as physical fitness should be increased (Converso et al., 2017). One method for resource enhancement is the implementation of WHPPs with a focus on exercising and/or ergonomics that initiate work behavior without inducing musculoskeletal discomfort and that reduce stress perceptions (Kreis, 2003; Wollesen, Menzel, Lex, & Mattes, 2016).

The goals of WHPPs are to reduce back pain via ergonomics-based interventions (Carta et al., 2010; Karahan & Bayraktar, 2013; Kindblom-Rising, Wahlström, Nilsson-Wikmar, & Buer, 2011; Ratzon, Bar-Niv, & Froom, 2016), “back school” programs (Jaromi et al., 2018), strength training (Cooper, Tate, & Yassi, 1998), or multicomponent interventions (Otto & Wollesen, 2022). The underlying mechanisms of these interventions are to reduce physical stress at the workplace (e.g., transfer situations) by improving movement quality (Carta et al., 2010; Ratzon et al., 2016), lifting techniques, and strength (Jaromi et al., 2018). A few studies demonstrated a reduction in back pain (Carta et al., 2010; Kindblom-Rising et al., 2011), while other studies found no effect (Karahan & Bayraktar, 2013; Ratzon et al., 2016). In a more recent study, Otto and Wollesen (2022) investigated the effects of a multicomponent program in a geriatric nursing setting, combining ergonomics training including exercises for body perception and learning of different lifting techniques and strength training.

In addition, WHPPs aim to improve work ability psychologically via reduced stress (Bischoff, Otto, Hold, & Wollesen, 2019; Fang & Li, 2015; Mohebbi, Dehkordi, Sharif, & Banitalebi, 2019) or enhanced quality of life (QoL; Becker, Angerer, & Muller, 2017). For example, yoga reduces self-reported sleep disturbances and work stress experienced by hospital nurses (Fang & Li, 2015). Similarly, a controlled clinical study investigated the effect of an aerobic exercise program and found reduced stress levels (Mohebbi et al., 2019). Generally, exercising may decrease perceived stress by shifting the psychological stress threshold, by reinforcing personal (e.g., self-efficacy) and social resources (e.g., social support), and by decreasing psychological and physical stress reactivity (stress buffer effect; Gerber & Pühse, 2009; Bischoff et al., 2019).

Beside exercise and training interventions, the World Health Organization (WHO) as well as national guidelines promote the combination of physical activity during daily live (e.g., to go for a walk or to work in the garden) with physical training to overcome sedentary behavior (Wackerhage et al., 2021; Rütten & Pfeifer, 2017). Therefore, WHPPs also include interventions to increase physical activities (Malik, Blake, & Suggs, 2014).

However, oftentimes WHPPs lack the application of training principles to induce discrete training effects. For instance, FITT principles (frequency, intensity, time, type of exercise; Swain, Brawner, & American College of Sports Medicine, 2014) should be integrated to ensure high-quality, effective interventions. Moreover, it has been shown that the training specificity, individualization, and overload are evidence-based training aspects that have to be integrated into the training methodology (Hecksteden, Faude, Meyer, & Donath, 2018). Within the actual practice of WHPPs, it remains unclear whether a more exercise scientific approach might increase the effectiveness of exercise interventions regarding the work ability of nurses. In particular, the intensity and time components of the FITT principle tend to be missing in exercise-based study reports (Billinger, Boyne, Coughenour, Dunning, & Mattlage, 2015). Moreover, aspects of individualization, progression, and control of the training load might be beneficial for gaining positive effects (Holzgreve, Schulte, Germann, & Wanke, 2022).

In regard to individualization, health promotion should be conducted in a setting-specific vs. one-size-fits-all manner (Schwarzer, 2008; Wollesen et al., 2016). As work demands differ between settings, the interventions should be adapted to the requirements of the specific setting (elderly care, outpatient care, or hospital). As demonstrated in the study by Otto and colleagues (2019), the highest strain on the musculoskeletal system was reported for nurses in elderly care compared to hospital care and outpatient care. Also, stressors in an outpatient setting include long driving times and high emotional involvement with patients, while this is less problematic in hospital settings (Vander Elst et al., 2016).

Accordingly, the focus of this systematic review is to address the literature gaps concerning the identification of effective exercise-related WHPP intervention types that enhance work ability in the nursing setting. Moreover, a specific aim is to give an overview of the implementation of FITT and training control principles (individualization aspects, progression, and variety) within the identified interventions in different nursing settings in order to gain more insights of the effect of training mechanisms on health-related outcomes. This will help practitioners and occupational health promotion experts to identify and conduct effective, setting-specific measures. Accordingly, the research questions of the current study include the following:

  • Which exercising and ergonomic intervention types in different nursing settings (hospital, geriatric, outpatient) can be identified?

  • What are the effects regarding physical and psychological health on nurses?

  • Which specific training principles and tailoring aspects can be identified to enhance the effectiveness of the interventions?

Methods

This systematic review was registered at the international prospective register of systematic reviews (PROSPERO; registration ID: CRD42021239665). Further, the review adhered to the guidelines of the Preferred Reporting Items for Systematic Reviews (PRISMA; Page et al., 2021).

Eligibility criteria

Following the PICOS scheme, interventions that tested nurses and nurse aides in different settings (hospital, outpatient, and geriatric) were eligible (P). Studies that included other health personnel (e.g., physicians, janitors, technical or office staff) were excluded due to differing work demands. Interventions qualified for synthesis if at least one component of the intervention was based on exercise or ergonomic work behavior (I). As this review mainly aimed to analyze the composition of the interventions, all potential control conditions were analyzed and reported in the results tables (C). Furthermore, outcomes that may impact nurses’ work ability were synthesized, including individual, social–cognitive, and organizational characteristics such as stress perception, pain, (QoL), physical activity, self-efficacy, ergonomic abilities, work demands, work satisfaction, sickness absence (O). The review included peer-reviewed intervention studies published in English or German since 1990. These included randomized controlled trials (RCTs), controlled quasi-experimental interventions, single-group pre–post studies, and pilot studies (S).

Search strategy and study retrieval

In November 2022 the final database search was conducted in the databases Medline, PsycInfo, and CINAHL with search words comprising MeSH terms “hospital nurses,” “outpatient nurses,” “elderly care nurses,” “physical activity,” “ergonomics,” “stress,” “pain,” “absenteeism,” and respective synonyms thereof (cf. Appendix B for complete search strategy). AKO imported the retrieved studies to the software Citavi 6. After deletion of duplicates, authors LH and AKO independently screened titles and abstracts of remaining studies with regard to the inclusion criteria. In a second screening cycle, LH and AKO reviewed the remaining full texts for eligibility. Disagreements for inclusion were resolved by discussion and, if necessary, via consultation with BW (supervising researcher).

Data extraction and risk-of-bias assessment

Extraction of the study data was performed by LH and AKO. Relevant criteria pertained to (1) participant description, (2) study design, (3) study aims, (4) intervention type, (5) intervention duration, (6) studied outcomes, (7) training principles, and (8) results.

Authors LH and AKO independently assessed the methodological quality of the included studies. The assessment was based on the “Standard Quality Assessment Criteria for Evaluating Primary Research Papers from a Variety of Fields” (Kmet, Cook, & Lee, 2004). Assessed studies are rated with respect to 14 criteria and each criterion is awarded an individual score of 0 (not fulfilled), 1 (partially fulfilled), or 2 (fulfilled) for a potential total score of 28. Accordingly, the assessed studies receive a percentage-based quality score. Quality scoring was adapted from recent studies (Curry, Patterson, Greenley, Pearson, & Forbes, 2021) such that a score of < 70% was rated as “poor,” 70–84% was considered “moderate,” and ≥ 85% was “high” quality. Through discussion, LH and AKO resolved disagreements in quality scoring for perfect agreement. For one study to avoid bias, the quality assessment was performed by consulting a colleague in the research team (acknowledged in the Disclosures and declarations section).

Data synthesis

The data synthesis includes a description of general characteristics (quality assessment, participants, study design, intervention type, intervention duration, and outcome parameters), as well as an outcome-specific synthesis. Thus, we analyzed separately interventions with physical outcomes, psychological outcomes, and self-managed interventions aiming at increasing physical activity. Assignment of intervention type was made through analysis of the study report and, if necessary, via discussion.

A meta-analytic approach was not feasible due to the high heterogeneity in the measured outcomes and intervention types.

Results

A total number of 37 studies were eligible for this review (cf. Fig. 1 for search flow).

Fig. 1
figure 1

Prisma diagram

Full size image

Reasons for exclusion included wrong study design (n = 29), no separate analysis of exercising components (n = 4), and inclusion of non-nursing subjects (n = 21).

Assessment of methodological quality

The overall quality of included studies was moderate (M = 0.76, min = 0.43, max = 0.92). The assessment identified 11 studies with low quality, 19 studies with moderate quality, and seven studies with high quality (cf. Table 1).

Table 1 Quality scores of included studies in alphabetical order
Full size table

Description of included studies

Sample sizes of studies ranged from 14 (Martins & Marziale, 2012) to 316 (Hartvigsen et al., 2005) nurses (overall N = 3487). The general intervention types were aerobic exercising (n = 1), back school (n = 3), ergonomics (n = 6), increasing physical activity (n = 3), multicomponent ergonomics (n = 6), multicomponent exercising (n = 12), multicomponent ergonomics and exercising (n = 1), physiotherapy (n = 2), and yoga (n = 3).

Concerning settings, 29 studies were based in hospitals, one study intervened with outpatient nurses, three studies were conducted with geriatric-care nurses, and four studies included nurses from different settings. Most studies involved female nurses (79–100%) or no sex distribution was found in the study report. Further, the mean age of the treated nurses was 38.82 years (min = 25.3, max = 51.28). Three studies (Karahan & Bayraktar, 2013; Martins & Marziale, 2012; C.-J. Zou et al., 2021) did not report mean sample age (cf. Table 2).

Table 2 Study table
Full size table

Synthesis of evidence for different intervention outcomes

Table 3 shows synthesized results of the included studies according to primary intervention outcomes. Within outcome categories, intervention types are differentiated. Various studies tested more than one primary outcome. Moreover, the frequency and duration of interventions are presented in Table 3. Additionally, Table 4 shows significant effects. In the case of no significant effects, “no” is reported in the table.

Table 3 Results by main intervention type in alphabetical order
Full size table
Table 4 Results by main intervention type in alphabetical order
Full size table

Intervention effects of studies with primarily physical outcomes

A total number of 28 studies aimed to improve physical outcomes (aerobic exercising n = 1, back school n = 3, ergonomics n = 6, increasing physical activity n = 3, multicomponent ergonomics n = 5, multicomponent exercising n = 7, multicomponent exercising and ergonomics n = 1, physiotherapy n = 2; cf. Table 3).

Aerobic exercising.

In a quasi-experimental study, Yuan et al. (2009) used treadmill exercise training for hospital nurses. The intervention group reduced their BMI and increased their strength, flexibility, and cardiopulmonary functioning.

Back school.

One combined intervention (Alexandre et al., 2001) consisting of education about work-related back pain accompanied by strength and flexibility exercises for hospital nurses that was compared to a single 45-minute (min) exercise class led to reduced cervical back pain.

The studies by Jaromi et al. (2012, 2018) used the same back school training curriculum (Spine Care for Nurses ergonomic training, muscle strengthening and stretching) compared to a passive control group for hospital nurses. In both studies, participants were strongly encouraged to carry out the learned exercises at home for five times per week. The interventions reduced back pain and improved ergonomic lifting behavior.

Ergonomics.

Best (1997) intervened with a specific ergonomic method for 32 hours (h) (manutention). The intervention improved the lifting technique of the participating geriatric nurses.

Carta et al. (2010) tested a single-group multimodal ergonomic, combining a 2‑h educative lesson with 3 h of practical training. At the 6‑month follow-up, back pain and patient handling knowledge as well as patient handling techniques improved significantly.

Feldstein et al. (1993) examined multicomponent ergonomic intervention (education: body mechanics, transfer techniques, identification of environmental hazards, and the importance of daily stretching) combined with practical training for hospital nurses. The intervention group improved in various patient transfer variables and reduced back pain while the inactive control group did not.

One intervention with outpatient nurses (Hartvigsen et al., 2005) included education and advice from a physiotherapist regarding ergonomic work behavior (lifting techniques, body mechanics). Further, nurses participated in four group-counselling sessions to deepen the learned transfer techniques and body mechanics. They found no effects on lower back pain (LBP).

Karahan and Bayraktar (2013) conducted a single-group multimodal ergonomic intervention with a 2‑h educational part and a 2‑h practical part (body mechanics, causes and reinforcing factors for back pain, back anatomy, lifting devices, and back protecting exercises). At the 3‑month follow-up, knowledge about work-related back pain increased significantly. Moreover, patient handling tasks improved significantly. The change was maintained at the 3‑month follow-up.

The multimodal ergonomic intervention by Kindblom-Rising et al. (2011) for hospital nurses examined two half-day patient transfer courses with a 2-week interval involving role playing and transfer exercises. At the 12-month follow-up, items on a self-developed questionnaire regarding physical complaints, movement awareness, and patient instructions during transfers improved in the intervention group, whereas in the inactive control group the number of sick leave days decreased.

Multicomponent ergonomics.

An RCT including hospital nurses with musculoskeletal complaints investigated a program including individual-specific job analysis (Ratzon et al., 2016). In addition, nurses performed segmental stabilization exercises. The intervention group exhibited less detrimental body postures at the end of the intervention compared to a passive control group.

Cooper et al. (1998) used a controlled individualized physiotherapy ergonomic intervention for hospital nurses with diagnosed back injuries. Moreover, a work-hardening program was instituted if sickness absence exceeded four working days after participation in the physiotherapy program. They found a significant reduction in back pain in the intervention compared with the control group.

In the RCT by Ewert et al. (2009) nurses from different settings with back pain received either a multicomponent exercise intervention combined with ergonomic and stabilization education or an active control intervention. Both intervention programs exhibited significant improvements in pain and QoL, but effects were not maintained at the 12-month follow-up.

In the study by Warming et al. (2008), hospital nurses either received a multicomponent ergonomic intervention (primary intervention), including patient transfer education as well as exercising components, or patient transfer education (secondary intervention), or no intervention. The patient transfer education was based on using transfer aids, the support of the patient, and ergonomic movement. The exercise regimen included aerobic and strength training. At the 12-month follow-up, the primary intervention group improved in one back pain dimension (disability) compared with the secondary intervention group.

The study by Yang et al. (2021) evaluated the effects of an 8‑week multidimensional ergonomic intervention to reduce work-related musculoskeletal disorders (WRMD) in hospital nurses. The participants received education, health behavior training, and work environment adjustments, whereas the control group only received two lectures on WRMDs and safe work environments. No group effects were found, merely a within-group improvement in WRMDs, risk perception, and health behaviors.

In a cross-over RCT by Otto and Wollesen (2022), geriatric nurses received a multimodal ergonomic and strength training. The intervention group received ergonomics training once per week for 20–30 min over the course of 10 weeks. Subsequently, the intervention group exercised for 12 weeks once per week for 45–60 min. The need-tailored approach yielded improvements in lifting performance and in subjective back pain.

Multicomponent exercising.

In one RCT, hospital nurses with back pain received an intervention incorporating coordination exercises and preventive education (Muller et al., 2001). Coordinative exercises were executed inside a space curl (36 sessions, 30 min each, one to two times per week). Preventive education was based on back-friendly patient transfers. The intervention group was compared with one active control group (received only coordination training), one passive control group (received only preventive education), and one inactive control group. Positive intervention effects (trunk muscle coordination, LBP frequency, and QoL) diminished after the 12-month follow-up.

Martins and Marziale (2012) compared a program for hospital nurses with rotator cuff disorder comprising strengthening and stretching exercises, shoulder proprioception, and cryotherapy with an active control group. At postintervention, pain and rotator cuff disorder-related QoL improved in both groups.

Bauer et al. (2019) conducted an RCT with hospital nurses suffering from LBP. Participants in the intervention group performed specific neuromuscular exercises to restore functioning of lumbar back muscles. The authors reported improvements in the intervention group compared to the control groups lumbar movement variability immediately after the intervention but did not provide data on statistical significance.

In one parallel-group RCT, the research group tested the effects of back exercises (strengthening and stretching) vs. transcutaneous electrical nerve stimulation (TENS) for hospital nurses with nonspecific LBP (Jalalvandi et al., 2022). At postintervention, both groups exhibited reduced pain and disability; however, the TENS group outperformed the exercise group.

Zou et al. (2021) investigated the effects of core stability training on back pain in hospital nurses. The self-selected nurses exercised five times per week for 45 min, for a total of 4 weeks. Subjective pain decreased in intervention group compared to the control group.

In the RCT of Brox and Froystein (2005), geriatric nurses received aerobic exercises, muscle strengthening, and stretching, in weekly meetings for 6 months. Only self-reported physical activity increased compared to an inactive control group.

The research group of Matsugaki et al. (2017) performed an RCT in different nursing settings where female nurses were randomized to either a multicomponent exercise (resistance and aerobic training) group with or without physiotherapist advice. Both groups exercised two times per week for 12 weeks. Compared to the active control group, knee muscle strength, oxidative stress, and low-density lipoprotein cholesterol significantly improved in the intervention group.

Physiotherapy.

In the study by Becker et al. (2017), injured hospital nurses received individualized physiotherapy to restore general physical capacity and psychological group sessions to improve well-being. An active control group only received the physiotherapeutic component of the intervention. At the 3‑month follow-up, the intervention group had reduced pain levels, physical demands, and burnout symptoms, while work-related psychological well-being improved compared to the control group.

Moreira et al. (2021) examined hospital nurses receiving therapeutic exercises for spine stabilization. The intervention group showed significant improvements in strength in trunk flexor muscles compared to the inactive control condition.

Intervention effects in studies aiming to increase physical activity

One single-group study with hospital nurses used a pedometer challenge (Web-based step tracker; Lavoie-Tremblay et al., 2014). The steps per day were recorded for a duration of 8 weeks, with the goal of 10,000 daily steps. At the 6‑month follow-up, participating nurses reduced their sitting time and had less insomnia.

In a non-randomized controlled pilot study, hospital nurses who were mothers participated in a 10-week intervention that aimed at facilitating physical activity during working hours (Tucker et al., 2011). Facilitation was achieved by providing different options in the working environment for being physically active. At the end of the study, intervention group participants had reduced body fat, healthier BMI, and less percent fat mass; however, steps per day did not increase significantly.

Similarly, Tucker et al. (2016) researched a program in different nursing settings with comparable content as an earlier study by Tucker et al. (2011). During work time, the workplace of the nurses was adjusted in a way that physical activity was facilitated. Additionally, participants received individualized mobile coaching messages to increase intervention adherence and success. The single-group study lasted for 6 months, and significantly increased moderate physical activity with increased steps per day while sedentary behavior decreased.

Intervention effects of studies with primarily psychological outcomes

A total of nine interventions aiming to improve psychological outcomes were identified (multicomponent ergonomics n = 1, multicomponent exercising n = 5, yoga n = 3).

Multicomponent ergonomics.

Akyurek and colleagues (2020) investigated the effects of a WHPP for hospital nurses including progressive muscle relaxation, posture exercises, and ergonomic education. Progressive muscle relaxation and posture exercise were trained for 30 min per session, whereas no specific duration for the ergonomic suggestions was reported. At the 12-month follow-up, the intervention group had lower subjective stress and burnout as well as better coping strategies and compassion satisfaction than the inactive control group.

Multicomponent exercising.

Two RCTs applied a similar stretching training (including back pain, core muscle, and relaxation exercises) for hospital nurses and investigated the effects on exercise self-efficacy, pain (Chen et al., 2014), and anxiety (Chen et al., 2016). The interventions were successful in improving exercise self-efficacy, pain, and anxiety, compared to an inactive control group.

The RCT by Mohebbi et al. (2019) investigated effects of an aerobic training regimen on hospital nurses’ stress perceptions compared to an inactive control group. Positive effects on occupational stress levels were not maintained at the 2‑month follow-up.

In the RCT by Ehegartner et al. (2021), nurses from different settings received a 5-day need-tailored practical and educational prevention program with contents being back school (2 × 60 min), aerobic exercises (2 × 60 min), active (4 × 60 min) and passive (2 × 50 min) relaxation, and cognitive-behavioral lectures on stress and work demands (5 × 120 min). At the 9‑month follow-up, stress and health-related well-being improved subjectively between groups.

Das Gecim and Esin (2021) studied the effects of a self-management program including general health education, yoga, and individualized nutrition counselling in an RCT for hospital nurses. The nurses were also encouraged to continue exercises at home. At the 3‑month follow-up, the intervention group showed improved work ability compared with the control group.

Yoga.

In Alexander and colleagues’ (2015) pilot study, hospital nurses either received an 8‑week yoga intervention or no intervention (total duration and frequency not reported). Participants in the yoga group significantly improved in terms of self-care and burnout symptoms.

In another study, yoga compared to an inactive control group was found to improve sleep quality and work stress in hospital nurses (Fang & Li, 2015).

Moreover, in an RCT including hospital nurses (Mandal et al., 2021), the nurses practiced yoga for a total of 12 weeks. The analysis identified intervention effects for perceived stress in the intervention vs. control group.

Applied training principles

Frequency and duration.

Concerning training frequencies of training sessions and total duration, in the aerobic intervention, nurses exercised for 20–30 min three times per week for a total of 12 h.

The duration of three back school studies ranged from 5 to 25.5 h and was delivered twice per week for 45–60 min (Alexandre et al., 2001; Jaromi et al., 2018) or in weekly 50-min sessions (Jaromi et al., 2012).

There was a wide range of duration in the six ergonomic studies, lasting between 4 and 112 h. Session frequencies and intervals were typically reported ambiguously. Two studies (Best, 1997; Kindblom-Rising et al., 2011) reported insufficient information.

Multicomponent ergonomic interventions addressing primarily physical outcomes (n = 6) were conducted between 2.25 and 30.5 h, typically in weekly (Otto & Wollesen, 2022; Ratzon et al., 2016) or twice-weekly (Warming et al., 2008) sessions lasting 45–60 min. However, three studies lacked important information regarding frequency and duration (Cooper et al., 1998; Ewert et al., 2009; Yang et al., 2021). One multicomponent ergonomic intervention that addressed psychological outcomes was conducted twice per week for about 40 min over 5 weeks; however, detailed information regarding ergonomic training is missing (Akyurek et al., 2020).

Multicomponent exercise studies for improved physical outcomes (n = 7) had a wide range of total duration with 4.5–48 h. Jalalvandi et al. (2022) applied a regimen with three sessions per week with a short duration of only 15 min and participants in the intervention by C.-J. Zou et al. (2021) exercised for 45 min five times per week. Other studies typically had weekly training duration of 1 (Brox & Froystein, 2005; Muller et al., 2001) to 2 h (Bauer et al., 2019). Two studies (Martins & Marziale, 2012; Matsugaki et al., 2017) failed to report session durations. In exercise studies aiming to improve psychological outcomes, total duration was between 19.75 and 65 h. While three studies (Chen et al., 2014; Chen et al., 2016; Mohebbi et al., 2019) involved exercise three times per week for 50- to 60-min sessions, Ehegartner et al. (2021) delivered a 5-day intensive prevention program for a total of 19.75 h. Lastly, Das Gecim and Esin (2021) did not provide information regarding session length and session content for a 4-week intervention.

Physiotherapeutic interventions had a total duration of 12–18 h. While nurses in the study by Moreira et al. (2021) exercised twice weekly for 30 min over 12 weeks, those in the study by Becker et al. (2017) trained weekly for 45 min over 10 weeks.

In yoga studies, session duration was typically 50 min in twice-weekly sessions over 3 (Mandal et al., 2021; 20 h) to 6 months (Fang & Li, 2015; 43 h). Alexander et al. (2015) did not provide sufficient information regarding frequency and duration.

Control mechanisms.

In total, 11 studies explicitly stated and applied training control mechanisms. Intensity control was used in seven studies (Ewert et al., 2009; Jalalvandi et al., 2022; Matsugaki et al., 2017; Mohebbi et al., 2019; Otto & Wollesen, 2022; Warming et al., 2008; Yuan et al., 2009), four studies pertained to progressive overload (Bauer et al., 2019; Becker et al., 2017; Martins & Marziale, 2012; Otto & Wollesen, 2022), and one study offered training variety (Alexander et al., 2015).

Further, eight studies reported the intensity of the trainings, of which three were light (Brox & Froystein, 2005; Tucker et al., 2016; Yuan et al., 2009) two were moderate (Mohebbi et al., 2019; Otto & Wollesen, 2022) and three had individualized intensity (Becker et al., 2017; Matsugaki et al., 2017; Ratzon et al., 2016).

Lastly, 19 studies were tailored to the needs of the target group or were individualized. Of those, seven were specific to back pain or LBP (Alexandre et al., 2001; Bauer et al., 2019; Jalalvandi et al., 2022; Jaromi et al., 2018; Jaromi et al., 2012; Muller et al., 2001; C.-J. Zou et al., 2021), one study was specific to shoulder pain (Martins & Marziale, 2012), and 11 studies applied a need-tailored or individualized training regimen (Becker et al., 2017; Cooper et al., 1998; Ehegartner et al., 2021; Ewert et al., 2009; Matsugaki et al., 2017; Otto & Wollesen, 2022; Ratzon et al., 2016; Tucker et al., 2016; Tucker et al., 2011; Warming et al., 2008; Yang et al., 2021).

Discussion

The overall aim of this systematic review was to illuminate the evidence regarding the efficacy of different exercise and ergonomic intervention types on the work ability of nurses. Furthermore, we included a qualitative analysis of training methods and content with respect to the FITT principles (Swain et al., 2014) and parameters of training control (e.g., RPE, progression, or variety).

In this review, nine types of exercise-based interventions in nursing settings were identified. These intervention types included aerobic exercising (n = 1), back school (n = 3), ergonomics (n = 6), increasing physical activity (n = 3), multicomponent ergonomics (n = 6), multicomponent exercising (n = 12), combination of exercising and ergonomics (n = 1), physiotherapy (n = 2), and yoga (n = 3).

Effectiveness of intervention types addressing physical outcomes

Light aerobic exercise on a stair stepper (Yuan et al., 2009) improved BMI, grip strength, flexibility in a sit-and-reach rest, durability of abdominal and back muscles, and cardiopulmonary function compared to an inactive control group. This exercise was done with a volume of 720–1080 min in 3 months. The improvements in the BMI as well as in cardiopulmonary functioning could have been expected due to the type of exercise (Christle & Arena, 2020). The improvements in grip strength as well as abdominal and back muscles are unexpected with respect to aerobic exercise. However, if participants are exercising on a stair-stepper they might have balanced their body movement on the stepper with holding the handrails and might have grabbed the handle bars and squeezed them. This may induce a higher grip strength. Moreover, this leads to an upright body posture with muscle activity in the abdominal and back muscles (which are also involved in the stair-stepping movements). Interestingly, these improvements were reached with light intensities. It may be possible that the participants were untrained and therefore even the light intensities were able to offer positive effects on physical fitness (Bann et al., 2015). In addition, there is some evidence that grip strength could be a marker of overall body composition and fitness (Bohannon, 2019). However, these explanations are only speculative and are not supported by the study description. Moreover, the study quality was low and therefore these results should not be overestimated, especially as this was the only study using stair-stepping exercises identified in this review.

The back school programs were able to improve body posture and the quality of the lifting performance (Jaromi et al., 2012; Jaromi et al., 2018) and to reduce cervical back pain (Alexandre et al., 2001) and non-specific back pain (Jaromi et al., 2012; Jaromi et al., 2018). In the study by Jaromi et al. (2012), the intervention was not more effective than passive physiotherapy in reducing back pain directly after the intervention. However, the positive changes in the intervention group were maintained in the 6‑ and 12-month follow-ups. Together with the positive changes of the lifting techniques and accompanying improvements in body posture, the nurses might have changed their lifting behavior, which led to persistent reduction in back pain (Al Johani & Pascual Pascua, 2019). To gain these effects, a total duration of 5 in 6 weeks was reported.

Included ergonomic studies (n = 6) exhibited improvements in movement quality and pain perceptions. In contrast to the results of the back schools including education, only one ergonomic program succeeded in reducing physical disorders and the number of sick leave days (Kindblom-Rising et al., 2011). None of the programs showed positive effects on reducing back pain compared to the control groups. One might argue that the intervention duration of classic ergonomic workshops is insufficient to gain positive effects on lifting behaviors. However, the length of the identified programs had a duration between 4 h and 112 h. Moreover, reflecting the duration of the back schools, positive results were found for a duration of 5 h in total (Jaromi et al., 2012). Out of three studies that measured follow-ups, two studies (Carta et al., 2010; Karahan & Bayraktar, 2013) reported sustainable intervention effects for ergonomic working. However, low study quality may confound the study results (M = 0.61). Thus, more high-quality interventions are necessary to establish the effects of short-term ergonomic trainings.

In addition, classic ergonomic interventions favor a straight upper body for correct lifting and carrying. A straight upper body with a strong lordotic characteristic is often presented here (the lower back tends to hollow in order to keep the back straight, especially when bending forward) and the assumption of a “rounded back” (kyphotic characteristic) is used as a potentially harmful negative example. However, there is often the contradiction that participants describe the negative example as being more pleasant, and the frequency of observations of these negative examples in practice raises the question of whether the techniques known from classic back schools are actually effective. In a combination of in vivo data and simulation models, Khoddam-Khorasani et al. (2020) compared the compression forces between lordotic, kyphotic, and medium postures and came to the conclusion that the extreme “back-friendly” lordotic upper body posture generates significantly higher compression forces and higher muscular activities than the other two variants. On the basis of the results of the study, the authors advocate adopting a middle position that avoids extreme upper body postures in one direction or the other. Therefore, one might argue that the content and the composition of the ergonomic programs should be changed, e.g., with the expertise of sports science to gain sustainable effects.

One solution might be multicomponent ergonomic programs incorporating aspects of education, ergonomic practical training, as well as strength or resistance training. In multicomponent ergonomic studies (interventions that combined ergonomic intervention and another component, e.g., exercising, meditation), mixed results were obtained regarding work ability enhancement. This may be due to the fact that intervention components trigger different physiological and psychological mechanisms. These mechanisms are linked to specific adaptation processes that require different training types, frequencies, and durations to occur (e.g., Rivera-Brown & Frontera, 2012).

Overall, five out of seven interventions led to positive results on body postures and lifting performance (Otto & Wollesen, 2022, combining ergonomics, education and strength training; Ratzon et al., 2016 combining education and physiotherapeutic exercises). Moreover, low back pain and accompanying disabilities were reduced in three of the programs (Cooper et al., 1998; Otto & Wollesen, 2022; Warming et al., 2008). The three studies combined education and strength exercises, while Warning et al. (2008) also included additional aerobic exercises. Therefore, a combination of the three components of education, ergonomic behavior, and (strength) exercises seems to be most promising. The potential benefits of classic training interventions might be one reason why the extensive ergonomic and exercise intervention by Ewert et al. (2009) did not find superior improvements in the intervention group that received ergonomic training, segmental stabilization, and cognitive-behavioral education in addition to multicomponent exercises compared to physical exercises only. Potential reasons pertain to training content selection, operation on the same psychological mechanisms as the exercise components, and inclusion of motivated participants, who may already have a good basic understanding of ergonomic work behavior.

An intervention combining progressive muscle relaxation, posture exercises, and ergonomics (Akyurek et al., 2020) was not able to gain positive effects on pain and body posture. The program had only 10 min per session for body posture exercises and the amount of ergonomic training was not specified. Therefore, one might conclude that the composition of the training was not specific enough to trigger mechanism to reduce pain and to improve body posture.

Nine multicomponent exercise studies examined physical outcomes, with eight studies reporting positive effects. Decreased pain was found for five studies that used different combinations of strength and stretching training (cf. Table 3; Chen et al., 2016; Chen et al., 2014; Jalalvandi et al., 2022; Martins & Marziale, 2012; Zou et al., 2021). Interestingly, the combination of the physical training with additional therapy used in the intervention by Martins and Marziale (2012) did not add further benefits.

Also, improvements of strength and trunk muscle coordination were reported in three studies (Matsugaki et al., 2017; Muller et al., 2001; Zou et al., 2021). These effects might be explained by the exercise specificity (cf. Table 3). However, exercise and test specificity might be a potential problem with respect to the reported benefits of the interventions. For instance, the study by Muller et al. (2001) examined coordination training for nurses, and while they found an improvement of trunk muscle coordination, there were no posturography-related improvements. With no balance-specific training components, it cannot be expected to find improvements in posturographic outcomes.

The aspect of specificity is also addressed in physiotherapy. In the integrated studies of this review, physiotherapy seems to improve work ability by changing various physical and psychological parameters (mobility, strength, psychological well-being, pain perception; Becker et al., 2017; Moreira et al., 2021). One major advantage of physiotherapy is that it should be designed to tackle the specific complaints of clients. Nevertheless, the extensive requirements regarding financial, personnel, and time-related resources might hinder the feasibility of physiotherapeutic approaches (Kreis, 2003).

Effectiveness of intervention types addressing psychological outcomes

A total number of 10 studies addressed psychological outcomes such as stress and QoL (cf. Tables 2 and 3). In the eight programs with positive benefits on stress outcomes, three used a yoga intervention (Alexander et al., 2015; Fang & Li, 2015; Mandal et al., 2021). In line with previous literature (e.g., Bischoff et al., 2019; Otto & Wollesen, 2022), significant effects on stress perceptions in nursing personnel were observed after a total intervention time of 20 h (50-min sessions two times/week). Unfortunately, no yoga study included follow-up measurements, nor was the yoga intervention compared with other intervention types. Moreover, the study quality was only moderate.

One might argue that yoga has the potential to improve biochemical markers of stress (Riley & Park, 2015); however, Mandal et al. (2021) did not find improvements in biochemical outcomes after a 20‑h yoga intervention.

One study examining multicomponent ergonomics including progressive muscle relaxation and breathing exercises assessed primarily psychological outcomes (Akyurek et al., 2020). Applying the curriculum twice per week for 5 weeks was linked to improved coping strategies as well as less perceived subjective stress and burnout symptoms. However, the lack of information concerning session duration is unfortunate and does not allow further conclusions to be drawn on adaptative mechanism.

Two studies evaluated the effect of a multicomponent exercising regimen on stress outcomes (Ehegartner et al., 2021; Mohebbi et al., 2021). The 5‑day prevention program by Ehegartner et al. (2021; total duration of 19.75 h) led to short-term effects on stress reduction, but desirable effects diminished at the 9‑month follow-up. Mohebbi et al. (2019) assessed a multicomponent aerobic training (24 h, three times/week for 1 h). Participants improved on stress-related outcomes: Unfortunately, no training-appropriate outcomes such as VO2max were assessed to offer conclusions, e.g., with respect to the stress-buffering hypotheses (for an overview, cf. Bischoff et al., 2019).

It should be noted that one study that aimed to increase physical activity (Lavoie-Tremblay et al., 2014) was able to reduce sitting times but led to significant higher stress values in the intervention group. Whether this result was due to a higher effort that had to be managed next to the daily working tasks or whether this aspect refers to, e.g., the physical activity paradox (Janssen & Voelcker-Rehage, 2023) cannot be answered in this review. Moreover, a short-term intervention with the goal to increase general physical activity may not be sufficient to induce stress-buffering effects (Gerber, Kellmann, Hartmann, & Pühse, 2010).

Quality-of-life parameters were examined in three interventions: multicomponent training including several exercises, ergonomics, and cognitive-behavioral aspects (Ewert et al., 2009); self-management training including health behavior, yoga, and nutrition (Das Gecim and Esin, 2021); coordination exercises combined with preventive education (Muller et al., 2001). None of these interventions exhibited an effect on QoL from pre- to post-testing. Previous studies on health-related QoL showed that, e.g., compassion fatigue, burnout, and compassion satisfaction influence the physical and mental components of QoL in nurses (Ruiz-Fernández et al., 2020). These aspects were not specifically addressed in these interventions. This might be one explanation why the measurement of QoL as surrogate marker did not show positive effects. Moreover, it needs to be discussed whether a self-management intervention should be combined with supervised training, as supervision guarantees appropriate training and exercise execution (Ninot et al., 2011).

Effectiveness of intervention types addressing increasing physical activity

Finally, three studies examined a self-managed physical activity challenge in hospital nurses (Lavoie-Tremblay et al., 2014; Tucker et al., 2011; Tucker et al., 2016). The trials had effects on reduced sedentary and inactivity times as well as weight loss and BMI. However, the examined stress outcomes did not change in the desired direction. The interventions were based on a motivational, self-management and therefore non-supervised model. Also, the studies used a gamification approach where participants could compare themselves with their participating colleagues. It has been found that there are moderating effects for gamification elements for participants who do not meet the WHO criteria on health-promoting physical activity (Bischoff, Baumann, Meixner, Nixon, & Wollesen, 2021). Specifically, different personality types may have differential preferences for gamified health promotion (Ghaban & Hendley, 2019). Thus, need-tailored gamification aspects should be considered in future self-managed WHPPs.

Benefits of identified training principles and control of training load

Frequency and duration.

A total of 15 studies failed to include relevant information pertaining to frequency, duration, or training session intervals (cf. Table 3). In particular ergonomic studies omitted details, leading to ambiguity regarding frequency, session duration, and session intervals. Studies with more exercise-related content were more often formulated clear specifications toward frequency and duration of trainings. For intervention studies that do not rely on a self-management model (e.g., mhealth interventions; Krebs, Prochaska, & Rossi, 2010), providing information on training principles should be the default to aid rigorous efficacy analyses.

Further, the high variability in training duration in the intervention types renders analysis difficult. However, it is noticeable that two sessions per week resulted predominantly in a reduction of back pain. This is in line with the findings of Wackerhage et al. (2021), who recommend a frequency of two training sessions per week.

Ergonomic studies with total duration of between 4 and 112 h exhibit poor reporting quality and highly heterogenous effectiveness. For instance, the most extensive intervention time (112 h; Hartvigsen et al., 2005) had no effects on pain and general health. However, the results showed improved lifting techniques with at least 4 h of intervention (Best, 1997; Carta et al., 2010; Feldstein et al., 1993; Karahan & Bayraktar, 2013; Kindblom-Rising et al., 2011). Thereby, the frequency of intervention units is often not specified. Further studies with high quality are needed to verify the effects (Otto et al., 2021).

Multicomponent ergonomic interventions were also highly variable concerning intervention duration, but there is a clear indication that intervention duration of 22 weeks (once a week) is associated with reduced back pain and improved working technique (Otto & Wollesen, 2022). This is linked to the finding that increased intervention effectiveness is associated with higher intervention duration (Greaves et al., 2011). Furthermore, long-term studies showed an average of 66 days for behavioral changes and 12–16 weeks for improvements in strength, reinforcing the study’s findings (Prieske et al., 2019; Scherenberg et al., 2018).

Therapeutic approaches can have subjective psychological effects for nurses at a total duration of 18 h in ten weekly sessions; however, these changes were not associated with objective strength measures (Becker et al., 2017). However, the therapeutic exercises in the study by Moreira et al. (2021) were linked to strength increases at a total time of 12 h, delivered twice weekly within 12 weeks. This difference may be due to the healthy participants in their study (2021) who did not suffer from pain during the intervention.

Lastly, yoga can have psychological effects on sleep quality and stress perception with two 50-min training sessions per week within 12 weeks. This was also found in the review by Zou et al. (2018), who recommend at least 60 min of yoga practice weekly over 12–16 weeks for psychological health effects.

Training intensity.

Concerning intensity, only eight studies explicitly mentioned training intensity (cf. Table 3). For light intensity (n = 3), results were mixed. It might not be sufficient to induce cardiovascular changes with light-intensity exercises, even with relatively high training load (Brox & Froystein, 2005). When moderate training intensities were applied (n = 2), effects on work ability improved. However, Mohebbi et al. (2019) did not analyze training specific outcomes but merely self-reported stress. This is unfortunate in training- and exercise-based studies that are likely to induce training-specific adaptations (e.g., Gibala, Little, MacDonald, & Hawley, 2012). Individual intensity studies (n = 3) were effective in reducing pain and improving body posture, strength, and vital parameters such as VO2max. However, results were heterogeneous between studies, which may be due to very different total intervention volumes. Also, intensity is merely one of many aspects for training individualization and should be combined with further control mechanisms (Simpson et al., 2021).

Individualization and need-tailoring.

Overall, 19 studies intervened with an individualized or need-tailored approach (cf. Table 3). Studies that aimed to improve region-specific pain levels (back or shoulder; n = 9) for injured nurses typically exhibited the desired results. However, maintaining work ability over long periods of time is a multifaceted endeavor and should consider both injured and non-injured nurses. Thus, need-tailoring and individualization procedures should incorporate multiple aspects of individuals, such as physical, psychological, and social circumstances (Bull et al., 2020). Additionally, the other studies with an individualized or need-tailored approach demonstrated predominantly positive effects with regard to their investigated outcomes (n = 8). This highlights the importance for individualized need-tailored interventions that consider nurses’ needs, wishes, and barriers (Otto et al., 2021; van Hoof et al., 2018). This approach has been shown to promote sustainable and long-term behavioral modifications and increase the employees’ motivation (Wollesen et al., 2016).

Training control.

Training control mechanisms were mentioned by 11 studies and included intensity control (n = 7), progressive overload (n = 4), and variety (n = 1; cf. Table 3). Studies with intensity control typically exhibited positive results at postintervention for outcomes that were associated with training aspects that were controlled for intensity. Similarly, progressive overload was associated with outcome improvement for various pain-, strength-, and movement-related parameters at postintervention in strength-based studies. The results point to the added value of sport science-based training control procedures (Borresen & Lambert, 2009).

Training principles and controls should be described as a default in exercise-based studies so as to allow for more thorough and detailed assessment and increase intervention efficacy. Where detailed description of training principles is missing, heterogenous effects of similar exercise interventions ensue (Herold, Müller, Gronwald, & Müller, 2019).

Strengths and limitations

This systematic review is the first in-depth analysis of exercise-based WHPP intervention types in different nursing settings. Additionally, the study shed light on the usage of relevant training principles in WHPPs. To our knowledge, the analysis of FITT and training control principles is unique in a WHPP setting and should become a standard procedure for WHPP and other health promotion reviews in order to establish common ground for the evaluation of intervention quality. Furthermore, we identified a large gap in the health promotion literature in the geriatric and outpatient setting. Additionally, this review could identify specific shortcomings of WHPPs in nurses. For instance, a lack of follow-up measurements and missing information pertaining to relevant training principles prevent the analysis of long-term adaptations to exercise regimens.

One limitation of the study pertains to the fact that our search strategy may not have detected all appropriate studies within the scope of our research aims. After checking the literature lists of related reviews (Bischoff et al., 2019; Otto et al., 2021), we believe that we identified the majority of relevant studies. However, we included all identified study types and did not focus on RCTs specifically. This might influence the evidence gained from the results due to missing control groups. Furthermore, with respect to theories of work ability, work performance, or work capacity, the results of this review might be limited. Future reviews should include additional outcomes regarding psychosomatic symptoms, work ability, and occupational health models. Also, the heterogeneity of outcomes was rather large. Comparisons between studies and a meta-analytical approach are therefore not feasible. Future studies may aim for the assessment of more homogeneous outcomes and rely on validated, often-used questionnaires rather than self-developed scales to allow for meta-analytic analyses in the future. Finally, while we contacted researchers for full texts of retrieved studies, we did not contact authors who failed to report full result details such as missing significance values.

Conclusion and recommendations

The review showed that intervention types such as back school, multicomponent ergonomics and training interventions, as well as yoga positively impact work ability via psychological and physical mechanisms.

Regarding evidence-based training principles, this review showed for the nursing setting that training specificity leads to benefits in the evaluated outcomes. In line with the previous literature, this was shown for improving ergonomic behavior in back schools and some of the ergonomic interventions, reducing back pain in combined interventions using in particular ergonomic and strength training and in reducing stress by yoga interventions. Moreover, regarding multicomponent training interventions including strength, endurance, and flexibility training, these interventions were as effective as physiotherapy. While the review clearly showed that different intervention types can lead to improvements in work ability at postintervention, only one study that included follow-ups found an improvement after 12 months post-intervention, which relied heavily on the autonomy and self-responsibility of the participants. Accordingly, future studies should attempt to innovate workplace health promotion programs (WHPPs) in a way that allows for sustainable long-term work ability improvements. Also, integration of sport scientific expertise concerning training control mechanisms for enhanced health promotion program success is essential to achieve sustainable effects on the work ability of nurses and other target groups.

Moreover, this systematic review showed that WHPPs in nursing settings currently may not adhere to the setting approach sufficiently, as suggested by the lack of studies in two of the analyzed settings. To ensure a high work ability of the entire nursing work force, it is very important to conduct WHPPs in all nursing settings so as to avoid nursing crises in the future. To accomplish this, there is a need for more high-quality intervention studies such as randomized controlled trials that identify feasible exercise-based WHPPs in nursing settings conducted according to evidence-based training principles.