Interventions to manage back pain in agriculture,horticulture,and seafood workers – A scoping review

Protocol and registration

As recommended by the Joanna Briggs’ Institute (JBI), a protocol was developed and registered through the Open Science Framework on March 26, 2024. https://doi.org/10.17605/OSF.IO/FB3ST. The review process was guided by the JBI methodology for scoping reviews (JBI Manual for Evidence Synthesis – JBI Global Wiki) using PRISMA-ScR as the reporting guideline. ²³

Eligibility criteria

The eligibility criteria for this review were determined using the Participants, Concept, Context (PCC) framework (https://jbi-global-wiki.refined.site/space/MANUAL/355862707/10.2.4±Inclusion ± criteria). Papers were included if: 1) the target population involved workers in physically demanding outdoor jobs in agriculture, horticulture (nursery, landscape) or seafood (aquaculture, fishing), sectors with back pain or who were at increased risk for back pain, 2) ergonomic interventions involving training or work modifications (movement, equipment, processes, breaks) or interventions involved non-pharmacological self-management interventions for back pain (exercise, positive thinking, coping strategies, pacing, relaxation, breathing, heat or ice, herbal remedies) were studied, 3) outcomes focused on prevention and/or rehabilitation, that may or may not have provided detailed intervention effectiveness; and may have included but were not limited to; pain, function, and return to work, and 4) completed papers were written in English or Spanish. We also included studies where the aim was to investigate tasks involved in the relevant sectors where participants without experience were recruited for simulations as this methodology is often reported in ergonomic literature. Papers were excluded if: 1) the population involved workers in healthcare, manufacturing, construction, schools or daycare settings, 2) interventions involved surgical or pharmacological interventions, 3) study design and methods were cross sectional, correlational, or descriptive, 4) studies reported prevalence of low back pain without linkage to self-management or ergonomic strategies, or 5) opinion, perspective, commentary, or study protocol papers. No publication date limits were imposed.

Searching and information sources

A health sciences librarian developed and conducted the literature search, with research team input. The preliminary search ensured that no reviews or protocols focused on interventions for back pain in these sectors had been published using the terms ((back) AND (ergonomic OR self-management) AND (agriculture OR agricultural OR farm OR aquaculture OR seafood) was conducted using the health-related repositories (BioMed Central Systematic Reviews, Campbell Collaboration, Cochrane Database of Systematic Reviews, Cochrane Public Health Review Group, JBI Evidence Synthesis, JBI Systematic Review Register, and PROSPERO: International Prospective Register of Systematic Reviews.

The second search was undertaken on September 22, 2022 to identify terms for the final search strategy yielding 1419 articles from the following databases selected for coverage of agriculture, aquaculture, business, and health topics: PubMed (National Library of Medicine), Biological and Agricultural Index Plus (EBSCOhost), Business Source Premier (EBSCOhost), CINAHL (EBSCOhost), Fish Fisheries and Aquatic Biodiversity Worldwide (EBSCOhost), SPORTDiscus (EBSCOhost), CAB Abstracts (Clarivate), and Web of Science (Clarivate).

The text words contained in the titles and abstracts and index terms from articles identified in the test search were used for the final search strategy. The search strategy was peer reviewed by a second librarian and adapted for the databases above, in addition to AGRICOLA (ProQuest), Agriculture Science Database (ProQuest), Embase (Elsevier), Scopus (Elsevier), PsycINFO (EBSCOhost), REHABDATA (National Rehabilitation Information Center), and Institute for Work and Health. (Website for search protocol to be added after deidentified paper is available) The searches for all databases occurred on April 23–25, 2024, and repeated on June 25–27 for additional papers published up to mid-2025. Search strategies are available at https://doi.org/10.1079/searchRxiv.2026.01255. An additional search was conducted once the screening was complete (described below) using the reference lists of all included sources of evidence.

Screening and extraction

References were imported into Covidence software for systematic reviews (https://www.covidence.org), screened, and duplicates removed using the software and manually.(Figure 1) We piloted the search criteria in two rounds using titles and abstracts from 20 papers chosen randomly to refine the inclusion and exclusion criteria and enhance reviewer consistency. Studies in schools or daycare settings, prevalence only, opinion or perspective, editorial, narrative review, feasibility, or protocol papers were excluded, while retaining proceedings and dissertations.

OPEN IN VIEWER

Figure 1.

Consort diagram template source: PRISMA 2020 flow diagram.

Once the inclusion and exclusion criteria were confirmed, all titles and abstracts were screened by two authors independently, with a third resolving discrepancies. Papers written by two authors were reviewed by other members of the team. The full texts were then assessed in detail for eligibility, again by two blinded pairs with a third reviewer resolving disagreements if needed. Reasons for exclusion of papers were recorded. (Figure 1)

Data were extracted from papers and recorded by one author and checked by a second using a rubric created in the Covidence software. General reference data (author (s), DOI, year of publication, study title, journal, research design, region, work sector (agriculture, horticulture, seafood), location of study activities (field, lab simulation, clinical studies), inclusion and exclusion criteria, population (description, sample size, presence of pain, age), tasks studied, outcome measures, intervention (self-management, ergonomic, clinical), and key findings or conclusions were recorded. After completion, the results were collated and tabulated, with discussion for any queries. As a scoping review, the rigor of research methods for individual studies was not a component for inclusion and analysis.

Analysis

After organization of papers by type of intervention: 1) ergonomic (changes to the workplace or equipment to reduce job or task risk), 2) self-management (interventions aimed at promoting individual's ability to manage and reduce pain), or 3) both ergonomic and self-management. Interventions were further categorized for ergonomic interventions (engineering - equipment or task adjustments to reduce forces, administrative - adjusting processes or procedures to reduce cumulative forces, both engineering and administrative controls), and self-management and clinical interventions (education, exercise, both education and exercise). Ergonomic engineering studies were further stratified by methods for reducing mechanical stress.

Publication date and location

The publications meeting inclusion criteria were relatively recent with only 3 articles published prior to 2005. Since 2006, 17–20 studies have been published per 5-year interval; 17 (2006–2010), 19 (2011–2015), 16 (2016–2020), 19 (2021- mid-2025). Studies were conducted mostly in the USA (28, 38%) and Asia Pacific (28, 38%), with lower numbers of studies from Europe (7, 9%), Australia/New Zealand (4,5%), Canada (3, 4%), South or Central America (2, 3%), and Africa (2,3%).

Frequency of papers by sector

As expected, most studies targeted agricultural settings (61, 82%), with lower numbers in agriculture and horticulture (3, 4%), horticulture only (4, 5%), and seafood (5, 7%) sectors (Table 1). Agricultural studies covered a variety of animal, crop, fruit, tree, or vegetable farming, horticulture included forestry, landscape, and nursery, and seafood involved crab and lobster fishing and clam aquaculture studies. (Appendix 1)

Type of tasks studied

Interventions addressing lifting and repeated or prolonged stooped or flexed positions were common across all sectors. Digging, shoveling, clearing land, and weeding were studied in both agriculture and horticulture settings. Animal studies such as milking, sheep shearing, or poultry slaughtering involved managing unpredictable movement of the animals; for crops, fruit, or vegetables, harvesting demands depend on the height or type of the product (e.g., strawberry harvesting close to the ground, peppers or grapes on trellises, or date harvesting requiring tree climbing) while moving product, feed, soil, and equipment were other tasks addressed in all sectors. The seafood context involves specific challenges from slippery or unstable surfaces on the boats or near the water impacting balance. Three studies were related to vibration and requirements for operating machinery such as tractors.^24–26

Type of interventions

In settings where resources prohibited expensive solutions, low-cost modifications were used to provide appropriate options for the context.^27–32 while higher cost technologies such as exoskeletons,^33–37 or robotic assistance ³⁸ were predominately reported in higher income countries. Ergonomic interventions (53, 72%) were studied the most, with lower numbers of papers encompassing self-management or clinical interventions (12, 16%), or combining ergonomic and self-management interventions (9, 12%). For the subcategories, ergonomic engineering controls were the most frequently studied (48, 65%), with low numbers of studies for administrative controls (2, 3%), engineering and administrative controls (2, 3%), and education, exercise, or both education and exercise (4, 5% for each) (Table 1). Summaries for studies are provided in Table 2.

Ergonomic – engineering controls

Interventions in the engineering control category (Table 2) involved strategies or designs to reduce overall ergonomic stress in 9 subcategories:

lifting lighter loads e.g., adjusting the loads based on body weight, ³⁹ lighter milking systems,^40,41 smaller harvest tubs. ⁴² (Figure 2A)

OPEN IN VIEWER

Figure 2A.

Lifting lighter loads using smaller tubs, baskets, or bags. ⁴²

OPEN IN VIEWER

Figure 2B.

Work-related forces were compared using different types of available equipment e.g., using weedeaters or brushcutters supported with a strap on one shoulder or bilateral shoulder straps. ⁴⁵

OPEN IN VIEWER

Figure 2C.

Modified equipment for land preparation: i) types of shovels, ii) rakes for harvesting blueberries, iii) handles and types of hoes.^47,49,50

OPEN IN VIEWER

Figure 2D.

Modified straps or belts to support baskets for carrying product during harvesting.^29,30,51,56

OPEN IN VIEWER

Figure 2E.

Exoskeletons and harnesses used to distribute forces: i) arm support for harvesting, ³⁵ ii) elastic trunk exosuit,^33,34,36,37 iii) harness distributing forces to the shoulders for carrying loads on the head. ⁶⁵

OPEN IN VIEWER

Figure 2F.

Transplanting plants with sensor-based robotic assistance (cobot). ³⁸

Ergonomic engineering interventions generally improved perceived comfort and exertion, and reduced muscle activity, although benefits were sometimes task- or posture-specific and some redesigned systems introduced usability concerns.^{29,30,51,53–56,58} However, effectiveness varied across tasks, even for the same intervention. Some devices improved lifting, others carrying or dumping, and some designs performed worse than traditional equipment for certain postural outcomes.^47,48,52,57 Across field studies where tools or options for carrying product were modified, modest benefits for productivity, fatigue, awkward postures, or pain were noted, but workers did not always prefer new designs.^{27,28,48–50} Adjustments for mechanized equipment were associated with reductions in pain, strain, vibration, or ergonomic risk, although some outcomes were unchanged.^{24,26,52,60,61,71} Studies using exoskeletons or external load-distribution devices such as harnesses were also mixed. Most reported reduced muscle activation although samples were often small limiting statistical power.^32,34,36,37 Two studies reported reduced spinal flexion range^33,34 while one study using a harness for sheep shearing reported reduced time in twisted postures but increased time in lateral flexion and reduced time in neutral trunk postures. ⁶⁴ Harnesses reduced compressive forces in the spine for workers who were shearing sheep,^64,68 but authors questioned about whether the device would protect the spine from unexpected forces. ⁶⁴ In a study comparing two types of exoskeletons to no support during apple harvesting and found that muscle activity varied working at different heights, with lower activity at higher heights with the exoskeleton, but at other heights, load was shifted to the back and lower extremities. ³⁵ In a field study, energy expenditure, pain, and absenteeism was reduced in a study with a back support with straps attached to a land preparation device, ³² lower heart rate and reduced fatigue was found using a personal assist suit with digging, ³⁷ and with harnesses distributing loads for carrying objects while walking. ⁶⁵ Generally, these studies did distribute loads, however, concentration of forces in other areas, practicality, and comfort of devices need to be considered for future work.^34,37

Ergonomic – administrative controls

Only a few studies evaluated administrative controls alone. Two field study papers focused on rest breaks with reduced pain and fatigue.^73,74

Ergonomic – engineering and administrative controls

Three studies combined engineering and administrative controls,^75–77 e.g., in a participatory ergonomic study, aquaculture workers selected and used 3 relevant strategies from a list of options such as pacing, lifting techniques, teamwork, work rotation, using or adjusting equipment to use for 8 weeks. Workers reported significantly reduced pain with work tasks and disability compared to baseline.^76,77 (Figure 3)

OPEN IN VIEWER

Figure 3.

Pulley systems rigged to an overhead support to reduce forces for pulling clam aquaculture nets from the water.^76,77

Self-management – exercise, education, or both exercise and education

While ergonomic interventions were often specific to work tasks, self-management interventions were applicable across settings. The self-management interventions rely on uptake and regular use of interventions either during work activities or at home and often require behavior change, although some exercise interventions were conducted in groups as part of work requirements. Education frequently covered lifting or body mechanics. with outcomes sometimes limited to reporting knowledge or lifting mechanics.^78,79 In one of the few randomized studies, back pain booklets and face-to-face advice or booklets alone resulted in lower disability in both groups up to 12 months, with lower pain up to 6 months in the combined group. ⁸⁰ Exercise or activity aimed at prevention was introduced in the workplace with general warmup programs such as walking or stretching, ⁸² or more comprehensive core strengthening programs in gym settings with positive results for pain reduction.^83–85 disability, ⁸⁴ trunk endurance and flexibility, ⁸³ and ankle range of motion. ⁸² Combined exercise and education interventions also reported reduced pain,^86–88 prevention behaviors, ⁸⁹ trunk endurance,^87–89 and functional disability.^86–88

Clinical interventions were also included in this category with a greater focus on management of symptoms and return to function. Individualized interventions were delivered by healthcare professionals (Physical Therapists, Occupational Therapists, Psychologists) in clinical settings,^87,88 or online, ⁹⁰ Clinical interventions showed varying degrees of improvement in pain, quality of life, impairments, sleep, mobility, and disability. After improvements in impairments, disability, and pain following multi-disciplinary rehabilitation, there was limited retention on returning to work. ⁸⁷ A psychoeducation study reported small short-term and delayed improvements in stress, anxiety, pain acceptance, and quality of life effects. ⁹⁰

Ergonomic and self-management

A growing number of more recent studies combined ergonomic and self-management studies in different formats and sequences. Decreased frequency of back and neck pain was reported after combined education, rest breaks, and ergonomic modifications, with authors concluding that the amount of time for rest breaks differed for older and younger workers. ⁹¹ In one of the most comprehensive combined programs, ergonomic training, along with 3 months of regular exercise, reducing high-risk tasks and hours, reduced pain which was retained 1 year later and reduced interference in work activities. ⁸¹ A comprehensive 3 week program including ergonomic, physical, and psychological content reduced the frequency of high risk work tasks and posture and resulted in significant improvements in pain which were retained after 1 year, ⁹ Conversely, a study comparing workers who identified high risk ergonomic positions along with small group training and stretching did not show improvements at a 2 month follow up compared to those who just received training. ⁹² Three studies in this category used participatory ergonomics, involving the workers’ involvement in planning, creating, and implementing ergonomic or self-management solutions with reduced ergonomic risk and worker satisfaction but pain was either the same or not reported.^31,93–95

Study design, settings, methods, participants

Most studies were non-randomized experimental trials (54, 73%), often with pre-post within-subject comparison, followed by case series (14, 19%), with few randomized experimental trials (5, 7%), and 1 case study (1%). Inclusion or exclusion criteria for recruitment were often missing. There were several papers reporting multiple phases starting with initial prevalence, risk or needs assessment, followed by intervention phases building from earlier results. The self-management interventions compared different forms of exercise or education more often than the ergonomic interventions.

Just over half of the studies (42, 57%) were conducted in the field, with 24 (32%) using simulations in laboratory settings, 4 (5%) using both field and laboratory simulations, and 4 (5%) conducted in rehabilitation settings or online. Simulation studies typically used small convenience samples (range 1–32, mean 17), often young participants without specific work experience in the field with no pain or did not specify if the participants were experiencing pain.

One format specific to workplace interventions was participatory ergonomics, starting with overall education in the first phase, followed by participant involvement in planning, creating, and implementing solutions, and finally testing ergonomic solutions.^{31,77,93–95} These studies reported positive participant satisfaction although small sample sizes did impact overall results.^76,77,94 One study reported reduced ergonomic risk but no statistically significant differences in frequency of musculoskeletal disorders compared to normal activities in a study of pistachio farm workers. ⁹⁴ There was also reduced risk in a small study using education, exercise, rest breaks, and equipment program, ⁷² and reduced numbers of higher risk tasks in another small case series. ³¹ The clam aquaculture study using participatory ergonomic methods reported high adoption and feasibility, along with reduced work task pain and disability,^76,77 while in a case series for lobster fishing, implementation was limited despite positive feedback due to limited financial support, productivity demands, and resistance to change work practices. ⁹³

Overall, generalizing findings from these studies is difficult due to study demographics such as small sample sizes, and recruiting young participants without pain. Half of the studies recruited less than 20 participants (38, 51%), with varied demographics, ages, and specific work tasks. One of the major observations was the recruitment of individuals who were not experiencing pain; less than a quarter of studies (17, 23%) explicitly recruited individuals with pain. There were 15 studies (20%) where pain, disability, or discomfort outcomes were collected but pain was not a specific inclusion criterion. The trend of studying young participants without work experience also limits the application of findings for working populations.

Outcome measures in the field studies varied from self-reported measures of pain, discomfort, or effort, disability questionnaires, productivity, or observational risk assessment. Most field studies also collected preferences and opinions of interventions. Simulations (whether in laboratory or field settings) used physiological measures reflecting efficiency and effort, electromyography, joint angles, forces and moments using motion analysis equipment, used to predict risk for musculoskeletal injury rather than direct impact on pain or discomfort. Risk assessment also included observation and visual analysis of videos or photographs, using REBA, RULA or OWAS risk assessment instruments. Although a number of studies reported participant preferences and satisfaction with the interventions, only a few studies evaluated adoption of equipment and practices^72,76,93,96 or implementation outcomes.^76,96 Outcomes measures are reported in Table 2 along with the results of the studies to inform future research choices.

Ergonomic interventions

Overall, the lower-cost ergonomic interventions are more applicable across countries, while the higher technology options still need to be evaluated in field settings. Adjusting lifting mechanics, movement techniques, modifying or adding platforms, stools, and heights of tables for sorting or harvesting products, principles related to redistributing loads using different bags or containers, and reducing loads are all applicable across sectors. Interventions for digging, planting, clearing land, and weeding apply both to agriculture and horticulture. Some interventions were very specific to the type of harvesting or setting which would not be transferable to other settings, e.g., testing a tree climbing device to harvest dates,^60,61 or systems to empty crab or lobster pots.^58,59,93 While some studies may not be generalizable to other sectors or task contexts, commonalities in the type of movement, position, or the intervention can potentially spark innovative ideas for interventions that are relevant for other sectors.

The higher-cost and newer technologies may not be applicable across settings and research conducted in laboratory simulations still needs to be tested in the field. Studying participants of all ages who represent the workforce and those who do report back pain is important for future research. While some of the exoskeletons or devices to distribute load did report lower muscle activity and calculated forces only one study reported pain and some reported effects in other areas. While forces and muscle activity were affected positively, discomfort in other areas or difficulty performing work activities are issues that need to be considered for future studies.

While there were only a few studies using administrative control interventions, rest breaks are applicable across multiple sectors and were found to be effective for reducing pain for workers working in stooped positions for harvesting, ⁷⁴ and tree nursery tasks, ⁷³ or combined with education. ⁹¹ In a multimodal study, education was paired with adjusted work-rest schedules, limiting the length of time in flexed milking tasks, and by combining tasks, improved efficiency. ⁸¹ Rest breaks, job rotations, and changes in processes do require organizational buy-in and change management as well as individual worker buy-in, making these studies more difficult to implement, however administrative control interventions should be considered for future research alone or combined with other interventions, especially when there are limited options for sophisticated equipment modifications.

Self-management and combined self-management and ergonomic interventions

Although there were less studies in this area, self-management interventions are applicable across settings and have the potential to help workers manage their own work practices and pain. There were some positive outcomes with these interventions which have not been studied as frequently. If workers continue to work with pain, strategies that can be implemented at home or at work are especially important if there are barriers to accessing and covering costs of clinical care. The studies using education alone typically emphasized prevention related to ergonomic risks with limited focus on symptom management. Measuring the impact of education focused interventions is challenging for a variety of reasons, however only assessing knowledge provides limited information for effectiveness of interventions. However, combined approaches often require education in some form, especially with participatory ergonomic interventions and using current educational theories and best practices may potentially enhance some of the methods for application and retention.

Exercise interventions did result in reduced pain and disability, with limited reduction in overall risk though one study using walking before weeding resulted in greater ankle mobility and thereby reduced the degree of spinal flexion for weeding. ⁸² As pain has a significant impact on psychological well-being, interventions that target pain-reduction are warranted and further study of the intertwined physical and psychological aspects targeting motivation, and self-efficacy critical for adoption of interventions and attitudes towards work participation are warranted. Combined multimodal interventions (e.g., education, exercise, and ergonomics) or multimodal treatment approaches are an important direction for future research. Most recent physical therapy clinical practice guidelines for acute and chronic low back pain recommend using multimodal approaches ⁹⁹ and integrating evidence from clinically-based care could be helpful when developing future study questions and designs.

Research design and generalizability

Consistent with the purpose of a scoping review, we did not specifically evaluate the rigor of the research design, power, or statistical analysis, however, small sample sizes and weak statistical analysis limit generalizability for several studies. There are limited studies comparing types of interventions and the wide variety of outcome measures prohibit a systematic review and make definitive conclusions across studies extremely difficult. Of importance, the intervention research needs to build towards real-world application and implementation in the workplace.

Surprisingly, there were lower numbers of studies explicitly focused on impact of interventions on pain or prevention of pain than expected, and the definitions of pain and discomfort ranged extensively. Using standardized definitions of pain and musculoskeletal questionnaires (rather than discomfort), would be helpful for comparison of studies.^99,100 The trend for laboratory-based ergonomic simulation is limits findings that can be directly linked to reducing musculoskeletal pain. Pain is complex, movement patterns vary, and application in the workplace involves numerous factors. While sophisticated outcome measures including electromyography, motion capture, and physiological monitoring of heart rate and oxygen consumption offer many advantages, most simulation studies included healthy asymptomatic participants with no work experience in settings of interest. Recruiting healthy, college-aged students with no history of back pain is logistically easier than recruiting skilled workers from production settings, but this convenience reduces external validity and generalizability, and these research findings should be interpreted cautiously until these studies are replicated in the workplace.

In contrast, field trials conducted with experienced laborers capture authentic pacing, environmental conditions, and movement patterns but rely heavily on observation, surveys, interviews, or video footage as outcome measures which are not as specific and require qualitative or mixed method analysis. Comprehensive field settings provide impactful data to advance intervention research; however, randomization and more rigorous follow-up may not be feasible due to work productivity expectations and are also more complex when time is limited. While laboratory simulations are valuable for generating precise biomechanical and physiological data, field trials remain essential for understanding real-world adoption and implications, durability, and risk reduction. There is a need to progress from simulations where new interventions can be pilot tested, to field studies with larger samples of workers studying pain-related outcomes, productivity, including follow-up which was seldom part of the methods. Few studies examined implementation, and implementation variables such as adoption, engagement, and context for interventions, collected longer-term outcomes, or built off previous work. Building from other studies in the same or different settings is a critical step to move intervention research forwards to translation and application in the field. Thus, a recommendation for future work to promote a more extensive and practical body of evidence is to examine longer-term outcomes in real-world settings with workers, building from previous work and extending into implementation studies.

Strengths and limitations

Strengths of this scoping review were the comprehensive and reproducible search of multiple global sources over the past 3 decades using 1) established methodological and reporting guidelines, 2) review by a multidisciplinary team, and 3) comprehensive analysis. Our multidisciplinary team provided perspectives from multiple angles which is considered a strength for scoping reviews.^101,102 We included and synthesized available ergonomic, self-management, and clinical intervention studies which is a novel addition to the literature. By gathering evidence from all three sectors, the search provides a collection of sources to assist researchers with future studies for multiple sectors.

Conclusions

Most intervention studies included in this scoping review focused on agriculture and ergonomic engineering controls. The research gaps identified in this study highlight an important need for 1) practical intervention studies for all 3 sectors but especially for horticulture and seafood sectors, 2) field studies with workers who have pain, 3) comprehensive multimodal approaches, and 4) rigorous research designs with larger sample sizes and longer term follow up. Currently understudied interventions including administrative controls and self-management approaches along with multi-modal approaches warrant further study. Future intervention research in agriculture, horticulture, and seafood could build from existing studies and draw from studies in other settings using common principles or tasks.