(Originally publish in Spring 2019)
Five of these were prospective studies, which followed a group of students over a period of time to determine whether the weight of backpacks was associated with back pain. Two of these studies were rated as having a “high risk of bias” and three were “moderate risk of bias,” due to high attrition rate, mixed testing methods, confounding factors, and poor prognosis.
Two of these studies—one longitudinal (2) and one randomized-controlled trial (RCT) (3)—examined the association between schoolbag weight and the risk of getting low back pain, yet neither reported such association. Another two longitudinal studies evaluated the students’ perception of schoolbags’ weight and the risk of low back pain. (4,5) Although both reported an association between both factors, Yamato et al. considered them to be of low-quality evidence because of high attrition rate, moderate to high confounding factors, and moderate to high risk of bias in reporting and measurements.
The last study (longitudinal) did not examine bag weight or its perception. However, none of the students reported any back pain from carrying a schoolbag. (6) Overall, none of these studies reported how long the students were carrying the bag, how many times they sought back pain care, or how many times they were absent from school.
The rest of the studies, which were a mix of cross-sectional studies (following a population and examine data from one point in time) and case-controls (comparing an individual with back pain and another similar individual with no back pain). Most of them “did not find an association between schoolbag characteristics and [low back pain],” despite having mixed sample sizes, different measurements in schoolbag weight, and definitions of back pain. Even so, Yamato et al. suggested that the review should be “interpreted with caution.” (1)
“There are few longitudinal cohort studies in children and adolescents in the literature, and in most of these cases identifying risk factors for back pain was not the primary aim,” they reported. “Consequently, the choice of exposure variables, measurement instruments and timing of data collection are not optimal to reveal causal relationships. These problems may explain why systematic reviews in this field do not have strong conclusions.” (1)
Zooming Out the Literature
Previous reviews also found similar results as Yamato et al. had cited. One major systematic review from 2007, published in Physiotherapy Theory and Practice, reviewed 10 qualified studies with more than 17,000 children and teens on upper body pain (neck, upper back, shoulders). The highest predictors of who is more likely to get upper body pain are static posture, depression, stress, psychosomatic symptoms, gender, and age, with girls having higher reports of pain than boys, particularly those with depression. (7)
In one study, somehow almost all the girls who participated reported weekly of having pain in association with sitting in a static posture while hardly any boys reported any pain. (8,9) Other studies included found that those who hunched while working or playing on a laptop for a long time have higher reports of shoulder pain as well as those who sit at school or at home for one to more than four hours. (10,11) Although having a kyphotic posture can lead to such pain, it is likely less about the specific posture, and more about being in one position for too long.
Because of the different methods of measurements used in the selected studies, it is hard to determine how much influence any one posture has upon pain development. “The results of the selected studies should therefore be interpreted with caution,” the researchers said, “especially the pain measurements because none of the studies defined the frequency, duration, or intensity of pain in the same manner. This could create either an under- or an overestimation of the associations made between risk factors and pain.” (7)
They also cited that self-reported pain measurements can be influenced by psychosocial and cultural factors that could vary among different populations of youths. Also, the recall period in the studies vary from a week to 12 months, which tends to diminish accuracy of reporting the longer the time and increase with age. (7)
Another review, published in Journal in Physiotherapy in 2010, examined five qualified papers on the first risk factors that are likely to contribute to chronic low back pain among children. One of their findings is that spinal asymmetry increased the risk of back pain a year later among girls compared to those with less or no spinal asymmetry, but when the population was re-examined eight years later, there was no difference between the groups. Further studies also found that lumbar mobility was not associated with back pain risk. (12)
“In summary, as many variables were examined in the five studies, many of the significant associations with future low back pain may have been chance findings. Although 13 risk factors were identified, none was confirmed as significant in an independent study. Four failed to be validated as predictive in a subsequent study, which amplifies the need for validation studies.
“The remaining nine that await validation are spinal symmetry, lumbar spine extension endurance, the ratio of lumbar flexion mobility to extension endurance, the ratio of lumbar extension mobility to extension endurance, the ratio of lumbar flexion and extension mobility to extension endurance, high levels of physical activity, part-time work, abdominal pain, and psychosocial difficulties. Future research should use a standard definition of low back pain, use short recall periods, and report raw data to enable results to be meaningfully pooled across studies.” (12)
Finally, a joint review from Japan and Australia cited some early evidence that heavy backpacks may increase the risk of back pain. However, that alone may not reduce back pain because more data shows that 89 percent of backpack injuries did not affect a sample of 247 American children who were admitted to the emergency room for such injuries. (13) “Eliminating heavy backpacks from children, or even reducing the load they carry therefore, may not be a panacea in reducing overall [back pain] rates,” the researchers wrote. (14)
Even among adult research, it’s almost the same story. One 2008 Danish systematic review published in Journal of Manipulative and Physiological Therapeutics pooled 54 qualified studies with a total of more than 20,700 participants, which was a mix of cross-sectional, cohort, and case-control studies. (15) The researchers rated each study based on a 4-point criteria for quality, including:
1. Description of the selection process, participation rate, and response rate (for cross-sectional and cohorts) or having a minimum of one age and sex-matched setup for the case-controls;
2. Exposure: accuracy and reproducibility of the method;
3. Outcome: validation of the measurements and statistics;
4. Odds ratio or risk ratio provided. (What's the difference?)
RELATIVE RISK vs. ODDS RATIO, a Brief Glimpse
Some researchers report relative risk (RR) and odds ratio (OR) when comparing two groups’ outcome after they were exposed to a risk factor. While these concepts can help us make better sense of risk assessment, some clinicians can get confused between both terms.
First, risk is the probability of someone getting sick or hurt, while odds is the ratio between those who got sick or hurt and those who didn’t. For example, a study that finds 6 out of 100 runners got a knee injury over a six-month period would say that this population of runners has a 6% chance of getting a knee injury during this period of time.
The odds of getting a knee injury would be 6/94, which is 0.064. Compared to the risk, the odds is very insignificant.
RR takes the existing data and compares the risk of those who were exposed to the risk factor and those who didn’t. In one study Yamato et al. cited (see infographic on the right), the RR measured is 2.1, which means those who had difficulty carrying a heavy schoolbag have twice the risk of getting back pain than those who don’t—with the baseline of 1.0.
OR is similar to RR but it compares ratios instead of probability. Another study that compared weight perception and back pain found an OR of 2.2, where those who perceived a heavy bag were twice as likely to get back pain than those who didn’t.
However, RR and OR are meaningless if we don’t know how many people were involved in the study. This is where we need the absolute risk to determine the baseline risk. For example, if we say heavy backpacks may increase the likelihood of getting back pain by 50%, that number seems huge. However, if we use absolute risk and reveal that heavy backpacks increase risk of back pain from 1 per 100 students to 2 per 100 students, then we have a different perspective. (That perspective changes further if we increase (1,000) or decrease (10) the sample size, which changes the RR.)
We should also consider that having an increased risk does not guarantee that an illness or injury will happen, and neither does having a decreased risk mean that it will not happen. Given the current pool of data, we can say that carrying a heavy backpack may contribute to back pain, but the risk is quite small compared to other factors, such as mental health, family and social support, parental communication about pain, and personal beliefs and perceptions about pain.
RelativelyRisky on Twitter offer some examples of the differences by using mainstream news reports.
1. Noordzij M, van Diepen M, Caskey FC, Jager KJ. Relative risk versus absolute risk: one cannot be interpreted without the other. Nephrol Dial Transplant. 2017 Apr 1;32(suppl_2):ii13-ii18. doi: 10.1093/ndt/gfw465.
“We found insufficient evidence for an association between sagittal spinal curves and all other health related outcomes (low back pain, symptomatic degenerative lumbar disk disease, disk herniation, cervical pain, thoracic pain, tension, type headache, bone mass loss, ventilatory disorders, fracture, risk of injurious falls.”
The only things they found that have a “moderate” association between sagittal spine posture and health issues are temporomandibular disorder, pelvic organ prolapse, daily function, and death. (15) Regarding “death,” the researchers mentioned that it is more likely that the participants with kyphosis were at high risk for cardiovascular disease.
Further studies in the last 20 years also found various results:
1. Weak association between kyphosis and non-traumatic shoulder pain at the acromion based on 10 qualified studies. (16)
2. No association of low back pain with the size of the lumbar lordosis, pelvic tilt, leg length discrepancy, and the length of abdominal, hamstring, and iliopsoas muscles are not associated with the occurrence of LBP among 600 young to older adults in Tehran, Iran. (17)
3. A correlation was found between subjects having lower lordotic curve and low back pain vs. those with higher lordotic curve and no pain in a South Korean systematic review of 13 studies where nine of them found such correlation with “statistical significance” while the other four did not. This only demonstrate correlation, not causation, since they do not know which came first: the pain or the lack of lordotic curve? (18)
4. A systematic review cited six high-quality studies that do not find an association between “awkward occupational postures” and low back pain. If there is any, it would be “weak,” the researchers stated. While there are some reports that such work postures could cause low back pain, such as prolonged kneeling or squatting, the evidence is conflicting because of the lack of reporting of duration, frequency, and different methods of measurements of posture and back pain. (19)
In fact, one research found that cricket players with abdominal muscle symmetry had higher reportings of back pain than those with higher muscle development on one side of their body (20), and there is hardly any difference between abdominal muscle development and back pain among a sample of ballet dancers. (21)
While it may be intuitive to connect the dots between heavy backpacks and back pain when we see or imagine a child hunched forward while walking home, research with larger populations suggest otherwise. However, it did not start as so. One of such earliest research was done in India with six boys between the ages of 9 to 15 that was published in Ergonomics in 1965. (22)
Researchers M.S. Malhotra and J. Sen Gupta from the Defence Institute of Physiology and Allied Sciences gave the students a 6-pound schoolbag to carry, kind of like the one that Paddington Bear carries but with a long strap. Four of the boys carried the bag and walked with a respiratory in their mouth over a certain distance, and during that four-day experiment, each boy carried the bag in a different way—between the shoulder blades (rucksack), hanging toward the lower back, one strap across the body, and carrying with one hand. The two other boys carried a Douglas bag that measured oxygen consumption and basal metabolic rate as the other boys walked.
Long story short, they found that the rucksack position is the most energy efficient way to carry and walk and does not “lead to deformity of the body posture...no impediment in movement.” They also mentioned that the worse position to carry is with one hand.
While this preliminary study has plenty of methodological errors (not to mention that the biopsychosocial model of pain had not been developed yet and the gate-control theory of pain was as novel as satellites and space travel), it may have fueled further beliefs among clinicians and the public at the time that back pain is biomechanical in nature. In 2002, British physiotherapist Jenny Wigram wrote in Education and Health that many students blamed heavy schoolbags for their low back pain, in addition to sitting for long hours in a classroom. (23)
Around that time, some clinicians suggested that students should carry between 10 to 15 percent of their body weight, but no more than 20 percent. (24) As we can see in the last 20 years, the evidence is suggesting that back pain among youths is much more than just the back.
While these studies that were reviewed are not perfect, they challenge the common belief that carrying a heavy backpack is a primary causation of back pain. Even if backpacks do not contribute to the majority of people’s back pain, we should take individual considerations in case we meet clients or patients who believe their pain stems mostly from hauling a heavy bag.
Remember that pain is a subjective experience, a mixture of many different cast members that contribute different amounts of each individual. For some, it may be mostly physical, and for some, it is more psychological or social. What we can do as massage therapists for this population is provide quality touch in an environment that “calms down” the nervous system, and promote analgesic effects in the body and the brain. (25) Our young clients or patients deserve such treatment, and most of the time, that may be enough for them.
1. Yamato TP et al. Do schoolbags cause back pain in children and adolescents? A systematic review. Br J Sports Med. 2018; 52: 1241–1245. doi: 10.1136/bjsports-2017-098927
2. Jones GT, Watson KD, Silman AJ, et al. Predictors of low back pain in British schoolchildren: a population-based prospective cohort study. Pediatrics. 2003;111:822–8.
3. Negrini S, Politano E, Carabalona R, et al. The backpack load in schoolchildren: clinical and social importance, and efficacy of a community-based educational intervention. A prospective controlled cohort study. Eura Medicophys. 2004;40:185–90.
4. Jones GT, Macfarlane GJ. Predicting persistent low back pain in schoolchildren: a prospective cohort study. Arthritis Rheum. 2009;61:1359–66.
5. Szpalski M, Gunzburg R, Balagué F, et al. A 2-year prospective longitudinal study on low back pain in primary school children. Eur Spine J. 2002;11:459–64.
6. Sjolie AN. Persistence and change in nonspecific low back pain among adolescents: a 3-year prospective study. Spine 2004;29:2452–7.
7. Prins Y, Crous L, Louw QA. A systematic review of posture and psychosocial factors as contributors to upper quadrant musculoskeletal pain in children and adolescents. Physiother Theory Pract. 2008;24:221–242. doi: 10.1080/09593980701704089.
8. Niemi SM, Levoska S, Kemila J, Rekola K, Keinanen-Kiukaanniemi SM 1996 Neck and shoulder symptoms and leisure time activities in high school students. Journal of Orthopaedic Sports and Physical Therapy 24: 25–29.
9. Niemi SM, Levoska S, Rekola KE, Keinanen-Kiukaanniemi SM 1997 Neck and shoulder symptoms of high school students and associated psychosocial factors. Journal of Adolescent Health 20: 238–242.
10. Harris C, Straker L 2000 Survey of physical ergonomics issues associated with school children’s use of laptop computers. Industrial Ergonomics 26: 337–346.
11. Ramos EMA, James CA, Bear-Lehman J 2005 Children’s computer usage: Are they at risk of developing repetitive strain injury? Work 25: 143–154.
12. Hill JJ, Keating JL. Risk factors for the first episode of low back pain in children are infrequently validated across samples and conditions: a systematic review. J Physiother. 2010;56(4):237-44.
13. Wiersema BM, Wall EJ, Foad SL. Acute backpack injuries in children. Pediatrics 2003; 111: 163-6.
14. Smith DR, Leggat PA. Back Pain in the Young: A Review of Studies Conducted Among School Children and University Students. Current Pediatric Reviews. 2007;3(1): 69-77(9).
15. Christensen ST, Hartvigsen J. Spinal curves and health: a systematic critical review of the epidemiological literature dealing with associations between sagittal spinal curves and health. J Manipulative Physiol Ther. 2008 Nov-Dec;31(9):690-714. doi: 10.1016/j.jmpt.2008.10.004.
16. Barrett E, O'Keeffe M, O'Sullivan K, Lewis J, McCreesh K. Is thoracic spine posture associated with shoulder pain, range of motion and function? A systematic review. Man Ther. 2016 Jul 21;26:38-46. doi: 10.1016/j.math.2016.07.008.
17. Nourbakhsh MR, Arab AM. Relationship between mechanical factors and incidence of low back pain. J Orthop Sports Phys Ther. 2002 Sep;32(9):447-60.
18. Chun SW, Lim CY, Kim K, Hwang J, Chung SG. The relationships between low back pain and lumbar lordosis: a systematic review and meta-analysis. Spine J. 2017 Aug;17(8):1180-1191. doi: 10.1016/j.spinee.2017.04.034. Epub 2017 May 2.
19. Roffey DM, Wai EK, Bishop P, Kwon BK, Dagenais S. Causal assessment of awkward occupational postures and low back pain: results of a systematic review. Spine J. 2010 Jan;10(1):89-99. doi: 10.1016/j.spinee.2009.09.003.
20. Gray J, Aginsky KD, Derman W, Vaughan CL, Hodges PW. Symmetry, not asymmetry, of abdominal muscle morphology is associated with low back pain in cricket fast bowlers. Journal of Science and Medicine in Sport. 2015 Apr 23. pii: S1440-2440(15)00091-2. doi: 10.1016/j.jsams.2015.04.009.
21. Gildea JE, Hides JA, Hodges PW. Morphology of the abdominal muscles in ballet dancers with and without low back pain: a magnetic resonance imaging study. J Sci Med Sport. 2014 Sep;17(5):452-6. doi: 10.1016/j.jsams.2013.09.002.
22. Malhotra MS, Gupta JS. Carrying of school bags by children. Ergonomics. 1965;8:55–60.
23. Wigram J. Why is low back pain common in adolescence? Education and Health. 2002;20:36–7.
24. Dockrell S, Simms C, Blake C. Schoolbag weight limit: can it be defined? J Sch Health 2013;83:368–77.
25. Vigotsky A. D., Bruhns R. P. The role of descending modulation in manual therapy and its analgesic implications: a narrative review. Pain Research & Management. 2015;2015:11. doi: 10.1155/2015/292805.292805.