The global burden of musculoskeletal injury in low and lower-middle income countries: A systematic literature review.

BACKGROUND: While the global burden of musculoskeletal injury is increasingly recognized, few epidemiologic studies have specifically recorded its incidence or prevalence, particularly in low- and middle-income countries. Understanding the burden of musculoskeletal injury relative to other health conditions is critical to effective allocation of resources to mitigate the disability that results from trauma. The current study aims to systematically review the existing primary literature on the incidence and prevalence of pelvic and appendicular fractures, a major component of musculoskeletal injury, in low- and lower-middle income countries (LMICs).
METHODS: This study conforms to the systematic review and traditional meta-analysis guidelines outlined in the PRISMA-P statement. Incidence rates were calculated as the occurrence of new fracture cases per 100,000 person-years, and prevalence as total fracture cases per population sample, reported as percentages.
RESULTS: The literature search yielded 3497 total citations. There were 21 full-text articles, representing 14 different countries, selected for data extraction. Included studies reported a wide range of incidence and prevalence rates, with an overall mean fracture incidence ranging from 779 (95% CI: 483.0-1188.7) to 1574 (95% CI: 1285.1-1915.1) per 100,000 person-years.
CONCLUSION: Better understanding the unmet burden of musculoskeletal injury in LMICs is critical to effectively allocating resources and advocating for underserved populations. To address existing gaps and heterogeneity within the literature, future research should incorporate population-based sampling with broader geographic representation in LMICs to more accurately capture the burden of disease.

Introduction

The Global Burden of Disease and Risk Factors study (GBD), first published in 1990, sought to quantify the health effects of more than 100 diseases and injuries across 8 regions of the world. This was a landmark study that impacted global health research by providing estimates of mortality and morbidity by age, sex and region. This report found that injuries accounted for more than 15% of all ill-health in the world in 1990 and forecast this to increase to 20% by 2020.[ The GBD update in 2010 corroborated this trend, with notable increases in both road injuries (6.2% increase from 1990 to 2010) and musculoskeletal disorders (84% increase from 1990 to 2010).[ These findings suggest a significant, but nonspecific, global burden of both musculoskeletal disease and injury, resulting in repeated calls for action within the field of global orthopaedics.[

While the global burden of musculoskeletal disease and injury are documented by these studies, they are broadly defined. Musculoskeletal disease encompasses a range of conditions—including traumatic injuries, congenital anomalies, chronic back pain, arthritis, rheumatologic conditions, and others—each of which has varying incidence and cost of intervention.[ Injury descriptions, on the other hand, often focus on mechanisms of injury (e.g., road traffic accidents, falls, interpersonal violence) rather than anatomical characteristics or outcomes. While this focus may assist in identifying areas for primary prevention, it does not adequately describe regional or country-income level differences in incidence or outcome, which would allow for more effective interventional treatment strategies. With optimal medical and surgical management in high-resource settings, many fractures, such as those of the femur, have limited to no long-term sequelae.[ In low-resource settings, however, these same fractures are more likely to receive no treatment or substandard treatment due to numerous barriers to care, posing a risk for lifelong disability.[ For such patients, a loss of physical ability can severely limit productivity or completely prevent their participation in work, leading to a host of problems for individuals, their communities, and their broader societies.[

The stark differences in outcomes following musculoskeletal injuries in countries with different resources highlight an important opportunity for improvement. Effective allocation of resources, including trauma systems, surgical training and infrastructure, and rehabilitation, requires an accurate quantification of the burden of disease in different economic conditions.

In order to determine the current understanding of musculoskeletal burden, the current study aims to systematically review the existing primary literature on the incidence and prevalence of pelvic and appendicular fractures, a major component of musculoskeletal injury in LMICs.

Material and methods

Search strategy

The systematic review utilized in this study conforms to the systematic review and traditional meta-analysis guidelines outlined in the PRISMA-P statement.[ The study was deemed exempt from institutional review board approval. Inclusion criteria were established a priori to minimize potential selection bias. The objective was to identify clinical studies with pelvic or appendicular fractures reported in low- and lower-middle income countries. The following criteria were implemented to find eligible clinical studies for the systematic review: a publication in a peer-reviewed journal and a primary source of original data; description of a pelvic or appendicular fracture, or a broader musculoskeletal injury that could include a pelvic or appendicular fracture; inclusion of a relevant quantitative measure of frequency or burden (incidence, prevalence, DALY, or QALY); and report of patients from countries that are classified as low- or lower-middle income by the 2018 World Bank List of Economies.[

Data sources

A comprehensive search was conducted in 6 academic databases in July 2018: EMBASE (www.elsevier.com/online-tools/embase); PubMed (https://www.ncbi.nlm.nih.gov/pubmed/); African Index Medicus (http://indexmedicus.afro.who.int/); Literatura Latino Americana em Ciencias da Saude (LILACS—http://lilacs.bvsalud.org/en/); Índice Bibliográfico Español en Ciencias de la Salud (IBECS—http://ibecs.iscii.es/); and Pan American Health Organization (PAHO—http://www.paho.org/). With the aid of a research librarian, a standardized search algorithm was developed using search terms in English and Spanish, modified for each database (Appendix A).

Study selection

Article selection and data extraction were conducted through DistillerSR, a systematic review software program (https://www.evidencepartners.com/). After running a duplicate detection scan in DistillerSR, 2 rounds of screening were performed to determine article eligibility: a title and abstract screening, followed by a full-text screening. Two researchers (D.M.C. and T.A.M.) independently evaluated the titles and abstracts of articles identified for possible eligibility (N = 3243). Remaining articles from this pool (N = 62) underwent a full-text screening, which were reviewed in duplicate to determine eligibility for inclusion. The reference list of each study was reviewed to find additional potential citations for full-text screening. Conflicts regarding article inclusion were resolved by consensus of reviewers that included senior authors (D.M.C., T.A.M., S.M., D.W.S., and T.M.). No publications were excluded based on language, with articles in English, French, Spanish, and Ukrainian (N = 21) included in the final review.

Data extraction

A data extraction form was created prior to selection and screening of publications. Data were extracted in duplicate (D.M.C. and T.A.M.) in order to verify accuracy. Extracted information included: country in which study was conducted; study population—number of participants, age, sex; study design—hospital or clinic-based convenience sampling, population-based sampling, or school-based survey; period of data collection; and injury characteristics—type of fracture (osteoporotic/fragility fracture, traumatic fracture, or other), fracture region (lower extremity, upper extremity, or pelvic), diagnosis, and type of quantitative measurement to describe the fracture (incidence, prevalence, DALY, or other).

Incidence rates were reported as the occurrence of new fracture cases per 100,000 person- years, and prevalence as total fracture cases per population sample, reported as percentages. Clopper-Pearson exact 95% confidence intervals were calculated for incidence and prevalence rates.

Results

The literature search yielded a total of 3497 citations (1271 from PubMed, 725 from EMBASE, 674 from African Index Medicus, 628 from LILACS, 199 from IBECS, and 0 from PAHO) (Fig. 1). There were 34 additional citations identified through manual review of included articles’ references. After duplicate detection, 3243 unique citations remained. After title and abstract review, 62 articles were included in a full-text screen, and of those, 21 articles were selected for final data extraction.[

Figure 1

PRISMA flow diagram outlining the study selection process.

Fourteen different countries were represented among the 21 included studies (Table 1). Eight studies were performed in low-income countries, while 13 were conducted in lower-middle income countries. Nine studies used population-based sampling, 10 used hospital- or clinic-based convenience sampling, and 2 used school-based surveys. Ten articles reported incidence rates, while the other 11 reported prevalence. One study reported Disability-Adjusted Life Years (DALYs), in addition to prevalence.[ Five studies included data specific to pediatric and young adult patients, and 7 studies included data specific to elderly patients. The remaining studies did not specify an age group. The manuscripts also differed in the type of fracture-related injury studied: 8 examined fragility fractures, 5 focused on acute fractures, 3 evaluated fracture malunion, and 5 did not specify the type of fracture included.

Table 1

Description of included studies

Table 2 demonstrates the incidence rate of fracture-related injury in each study population, all reported per 100,000 person-years. Among the 12 studies reporting fracture incidence, only 2 studies conducted in Nepal and Sierra Leone reported the overall incidence of all fracture types. Using community-based surveys, these studies reported estimates of incidence of 779 (95% CI: 483.0–1188.7) and 1574 (95% CI: 1285.1–1915.1), respectively. The majority of other studies focused specifically on fragility fractures in older adults. The incidence of hip fracture ranged from 0.67 (95% CI: 0.22–1.56) in Nigeria to 132 (95% CI: 125.5–138.9) in India. Only 1 study from Nigeria reported the incidence of upper extremity fractures, estimating an incidence of distal radial fracture of 2.14 (95% CI: 1.22–3.48).

Table 2

Incidence of musculoskeletal injury in low- and lower-middle income countries

Prevalence data were reported primarily for sequelae of fracture such as malunion (Table 3), which ranged from 0.3 to 0.4% in adults compared to 0.1% in children. In contrast, the prevalence of acute traumatic injuries was much higher, ranging from 13.4 to 18.2% in adults in Nepal and Sierra Leone and 14.7% among children in Zambia. Only 1 study, in addition to incidence, reported Disability-Adjusted Life Years (DALYs) incurred due to musculoskeletal injury in Sierra Leone: 49.2 DALYs per 1000 persons, and Nepal: 11.6 per 1000 persons.[

Table 3

Prevalence of musculoskeletal injury in low- and lower-middle income countries

Discussion

We conducted a systematic review of studies describing the epidemiology of musculoskeletal injury in low- and lower-middle income countries. We found a relatively small number of studies that specifically reported incidence or prevalence of fractures in the pelvic and appendicular skeleton from these settings. Among studies that were identified, many relied on hospital-based convenience samples that are unlikely to capture the true incidence of injury. In addition, most studies focused on low-energy fractures related to osteoporosis rather than high-energy trauma from road-traffic injury. This is despite the increasing recognition of the epidemic of road traffic injuries that disproportionately affects LMICs.[ Only 3 studies reported the prevalence of malunion, and none reported the incidence or prevalence of nonunion.

Despite the dearth of epidemiologic data from LMICs, there have been recent reviews published regarding musculoskeletal injuries in high-income countries (HICs).[ A 2018 study evaluated the burden of acute trauma, using road traffic injuries (RTIs) as a surrogate marker.[ They reported that there were 255 million DALYs in 2016 in HICs due to road traffic injuries. These data have been used to advocate both for injury prevention and trauma care systems. Using a similar methodology, Agarwal-Harding et al[ used RTI data to estimate the global incidence of femur fractures at 15.7–45.5 per 100,000 person-years. Notably, they estimated the incidence of femur fracture was twice as high in LMICs compared to HICs. No studies that we identified can serve as a direct comparison because they were either too broad or focused specifically on low-energy fractures in older adults. Nonetheless, using the incidence of RTIs as a surrogate for injury rates likely remains the best available method in the absence of empiric data from these settings.

Osteoporotic fracture incidence has also been heavily studied in HICs, and an increasing number of resources have been dedicated as a result, including programs aimed at both prevention and treatment of fragility fractures.[ Many of the studies identified in this review were similarly aimed at the increasing incidence of fracture among aging populations, which is increasingly affecting both HICs and LMICs. There did appear to be a lower incidence of hip fracture in Sub-Saharan African countries such as Nigeria and Cameroon compared to countries in other regions including India, Ukraine, and Morocco. Data from Sierra Leone demonstrating high overall incidence of musculoskeletal injury suggests that high-energy trauma predominates in this region, though existing data are not conclusive in this regard. We hypothesize that younger populations and higher incidence of road traffic injury lead to a greater proportion of high-energy musculoskeletal injury in low-income countries compared to middle-income countries. However, the existing data from LMICs identified in this review are insufficient to make this statement conclusively.

The present review also has multiple limitations. First, due to existing gaps in the literature, this review included 21 studies in the final report, representing only 14 of the 82 LMICs (Fig. 2). Moreover, 1 inevitable limitation in a systematic literature review is selection bias, including the potential omission of relevant studies due to limitations of the search strategy. In the present study, selection bias was minimized by conducting a comprehensive search using 6 different electronic databases with broad geographical representation, and by including both English and Spanish search terms. No studies were excluded based on language, with articles in English, French, Spanish, and Ukrainian included in the final review. However, it remains possible that articles archived in smaller databases, or those indexed in languages other than English or Spanish, were excluded from the present study. Another limitation is that many of these studies utilized hospital- or clinic-based study designs, in which incident fractures were only recorded if the patient sought medical care, significantly underestimating the number of individuals suffering from musculoskeletal injury. Further, the heterogeneity of the available data limits our ability to statistically compare the fracture burden between studies and to draw more broadly generalizable conclusions concerning the global burden of fractures in LMICs. There are multiple potential reasons for this variability, including differences in sampling methods, eligibility criteria, and statistical methods. We did not include studies reporting the incidence of spinal injuries or injuries that did not involve fracture (e.g., ligamentous injuries), and therefore cannot comment on their burden in LMICs. Finally, prevalence is difficult to interpret in these studies because many acute fractures heal uneventfully, while sequelae, such as nonunion or malunion may cause long-term disability.

Figure 2

World map representing all LMICs that are (represented in dark gray) and are not (represented in light gray) included in this review.

Conclusions

There are few studies reporting the incidence of musculoskeletal injury in LMICs, and available studies focus largely on fragility fractures in elderly populations despite the known high incidence of road traffic injury. Best available estimates of the overall incidence of musculoskeletal injury in LMICs range from 779 to 1574 per 100,000 person-years. Future research should incorporate population or community-based sampling with broader geographic representation in LMICs to more accurately capture incidence burden. Better understanding the unmet burden of musculoskeletal injury in LMICs is critical to better allocating resources and advocating for this underserved population.