The Study On Risk Factors For Non-Specific Low Back Pain Among School Children And Adolescents

Background and objectivesLow back pain is a subjective feeling of pain in the low-back region, from the lower lumbar, lumbosacral, sacroiliac, to hip or leg. Non-specific low back pain (NLBP) is a general class of low back pain, not related to tumor, inflammation, injury, and lumbar disc protrusion. NLBP in children and adolescents was considered for many years to be rare. In the last two decades, epidemiological studies have shown that NLBP is a very common disease affecting the health of school children and adolescents. The lifetime prevalence varies between 11% and 74%. The cause of NLBP is still unclear until now. However, there are many risk factors closely related to the increase of NLBP.For further research on the risk factors for NLBP among school children and adolescents, prevention and control of NBLP, and health promotion, we carried out the series studies. This study has two parts:The first chapter is a systematic review of the prevalence and the risk factors for non-specific low back pain among school children and adolescents. Whether or not, there are differences in NLBP prevalence by gender or by place? Are there genetic factors in NLBP? Whether or not the prevalence of NLBP can be affected by the factors such as physical activity factors, sedentary factors, boy mass index, smoking, psychological and social factors? For these problems the different studies have yielded different results. With international Cochrane Collaboration methods, this study intends to systematic review the present cross-sectional studies, cohort studies and case-control studies all over the world regarding the prevalence and risk factors for NLBP among school children and adolescents to obtain better evidence for prevention and control of NBLP.The second part is an epidemiological study of non-specific low back pain among school children and adolescents in Guangzhou city. There are no information about the prevalence and risk factors among Chinese school children and adolescents so far. This study intends to carry out a cross-sectional survey with a self-designed questionnaire with high reliability, In order to learn about the prevalence and the characteristics of NLBP among school children and adolescents. We also intend to investigate the possible risk factors for NLBP among adolescents by carrying out a 1:1 matched case-control study.Methods1. Systematic review methods:Guided by the principles of evidence-based medicine and the methods of Cochrane systematic reviews, PUB MED (1966-2009.1), EMBASE (1966-2008.12), OVID, CMAC (1998-2009.12),CBMdisc (1978-2009.12),CNKI (1994-2009.12),VIP (1994-2009.12) and the reference lists of identified studies were searched. We identified all cross-sectional studies, cohort studies and case-control studies of prevalence and risk factors for NLBP in school children and adolescents. The quality of included studies was evaluated by two reviewers. Aggregate value and confidence interval of the prevalence was conducted with MetaDiSc1.4. Meta-analysis with the risk factors was conducted including homogeneous studies and sensitivity analysis with Revman5.0.23.2. Cross-sectional survey methods:Survey sample was school children and adolescents aged between 9 to 18 years of age. We selected a total of 2,100 students from two randomly selected ordinary public schools, one primary and one secondary school in Guangzhou as our subject. We designed a "non-specific low back pain questionnaire for adolescents". With the permission of each school for the students to participate this study, questionnaires were sent to the head teachers of the selected classes during July-August in 2009 by three groups of researchers and collected together by the head teachers two days later. The height and weight of the students were measured by researchers. The data base was set up according to questionnaires with EpiData 3.1, analyzed with SPSS13.0. To evaluate the test-retest reliability of the questionnaire, we calculated Kappa coefficient and Spearman coefficient separately on the count data and grade data. We calculated Cronbach's a coefficient to evaluate internal consistency of the questionnaire. We evaluated the content validity of the questionnaire by semi-structured interviews with students, parents and teachers. The comparison between constituent ratios was made withχ2 test (Crosstabs, Pearson Chi-Square or Fisher's Exact test). The comparison between ranked data was made with Kruskal Wallis Test. The comparison between averages was made with t test, P <0.05 (bilateral) was statistically significant.3. Case-control study methods:According to the results of the cross-sectional survey, we included 607 NLBP patients as the case group, and selected 607 non-NLBP students from the same school as the control group with the matching condition of the same sex and the same class. We measured and compared the exposure proportion to the risk factors including activity condition, school bag factors, learning situation and living conditions between the two groups. The data base was set up according to questionnaires with EpiData 3.1, analyzed with SPSS 13.0. The comparison between constituent ratios was made withχ2 test (Crosstabs, Pearson Chi-Square or Fisher's Exact test). The comparison between ranked data was made with Kruskal Wallis Test. The comparison between averages was made with t test, P<0.05 (bilateral) was statistically significant. Logistic regression analysis was carried out on the statistically significant independent variables after single-factor analysis.Results[Chapter one]1. Twenty-three studies were included in the systematic reviews with prevalence, four studies were graded A, twelve were graded B and the other seven were graded C. Twenty-eight studies were included in the systematic reviews with risk factors. Seven studies were graded A, fifteen were graded B and the other six were graded C.2. Among the twenty-three studies which reported the prevalence of NLBP, twelve reported history prevalence, seventeen reported period prevalence and three reported point prevalence. Heterogeneity was found among the studies by Meta analysis. The summary of the three types of prevalence values were 36% (95%CI:35%-36%),24%(95%CI:23%-24%)and 11%(95%CI:10%-11%)。3. Meta-analysis showed that the prevalence of male students was lower than female students (OR=0.64,95%CI:0.52-0.79)4. The prevalence of students with family history was higher than the students without family history (OR=1.98,95%CI:1.47-2.67)5. The difference of NLBP prevalence between high-amounts of physical activity groups and low-amounts groups had no statistical significance (OR=1.02, 95%CI:0.88-1.17)6. The prevalence of students with high sedentary activity was higher than the students with low sedentary activity (OR=1.11,95%CI:1.01-1.22)7. Smoking is a risk factor for NLBP (OR=2.34,95%CI:1.45-3.76)8. The correlation of boy mass index and NLBP was also statistically significant (OR=1.57,95%CI:1.25-1.97)9. The height of the chair being described as too low was significantly associated with NLBP (OR=2.18,95%CI:1.26-3.77).The perception that the backrest of the chair was too high was significantly associated with NLBP (OR=5.06,95%CI:2.06-12.41). Being not satisfied with school chairs was significantly associated with NLBP (OR=2.60,95%CI:1.96-3.80)10.A certain correlation has found between the time and the carrying manner of school bag and NLBP.11.The prevalence of students walking to school was lower than the students using other transport. 12.The prevalence of students with negative psychosocial experiences was higher than other students.13.Heterogeneity was found among the studies by Meta analysis on the six risk factors. Publication bias may exist by analyzing the funnel plot about the 16 included studies on gender. No significant publication bias was found by analyzing the funnel plot about the 6 included studies on sedentary activity.[Chapter two]1. The Cronbach'sαcoefficient of total items of the questionnaire is 0.912, the Cronbach'sαcoefficient of each part of the questionnaire ranged from 0.418 to 0.945; Test-retest reliability of each item of the questionnaire ranged from 0.338 to 0.885, more than 1/2 coefficients were greater than 0.5. Experts appraised that the questionnaire had good content validity2. After 2235 questionnaires were given out, a total of 2100 surveys were collected. 2100 response made the response rate reach up to 93.96%.17 were eliminated from data analysis because many of their responses were incomplete. Answers from 2083 schoolchildren (977 from males,1106 from females)were considered to be valid (valid response rate 99.19%). The mean age was 14.43(±2.37) years old, and the age span was 10-18 years old.3. The results of cross-sectional survey:①The prevalence of NLBP among school children and adolescents in Guangzhou was 29.1%(95%CI: 27.1%-31.1%).②The difference of the frequency of NLBP between male students and female students was statistically significant (χ2=16.318, P=0.003). The difference of the nature of NLBP between male students and female students was also statistically significant(χ2=22.922, P=0.000). The difference of the duration of NLBP between male students and female students had no statistical significance (χ2=6.068, P=0.194).③4.6% students were absent from school due to NLBP, and 15.3% students visit the doctor due to NLBP. The proportion with respect to the need of medical service was significantly higher in males than in the group of female students (χ2=7.249, P=0.007). The proportion with respect to the impact on daily life was also significantly higher in males than in the group of female students (χ2=8.216, P=0.016).4. The result of case-control study by univariate analysis:①The difference of family history of NLBP between the case group(47.1%) and the control group(27.5%) was statistically significant(χ2=43.740, P=0.000).②The proportion of playing basketball was significantly higher in the case group than in the control group (χ2=5.164, P=0.023).③The proportion of doing gymnastics was significantly higher in the case group than in the control group (χ2=4.125, P=0.042).④The proportion of activity between classes was significantly higher in the control group than in the case group (27.6%vs.22.1%,χ2=9.434, P=0.024).⑤The proportion with respect to the time of carrying school bag was significantly higher in the case group than in the control group (χ2=12.045, P=0.001).⑥The proportion with respect to the feeling of the school bag heavier was significantly higher in the case group than in the control group (χ2=21.963, P=0.000).⑦The proportion of feeling uncomfortable with the school bag was significantly higher in the case group than in the control group (χ2=11.787, P=0.001).⑧The proportion of using bus to school was higher in the case group than in the control group, the proportion of walking to school was lower in the case group than in the control group, the difference was statistically significant (χ2=9.152, P=0.027).⑨The proportion of dissatisfaction with school desks and chairs was significantly higher in the case group than in the control group (χ2=8.401, P=0.004).⑩The proportion of feeling high pressure was significantly higher in the case group than in the control group (χ2=9.745, P=0.002). (11)The proportion of drinking was significantly higher in the case group than in the control group (χ2=8.161, P=0.004)5. The result by multivariate logistic regression analysis:①Family history was a risk factor for NLBP. The prevalence of students with family history was higher than the students without family history (OR=2.303 95%CI=1.757-3.017).②The heavier the students felt the weight of school bags, the greater risk of NLBP was (OR=1.279,95%CI=1.087-1.506)③Playing basketball was a risk factor for NLBP(OR=1.349, 95%CI=1.037-1.756).④Doing gymnastics was a risk factor for NLBP(OR=2.494,95%CI= 1.019-6.104).⑤Satisfaction with school desks and chairs was a protective factor for NLBP(OR=0.750,95%CI=0.575-0.979).⑥Drinking was a risk factor for NLBP(OR= 1.347,95%CI=1.017-1.783).Conclusions[Chapter one]1. The prevalence of NLBP among school children and adolescents had large regional differences.2. Female students were more vulnerable to NLBP than male students. There may be genetic factors in NLBP among school children and adolescents. NLBP patient's weight may be relatively heavier, but still not sure that a high body mass index increased the risk of NLBP.3. Still not sure whether physical activity will affect the occurrence of NLBP or not. Long sit-in may be risk factors for NLBP. Active journeys to school such as walking and bicycling may be healthier to waist, but still not sure the appropriate amount of journey.4. Smoking may increase the risk of NLBP. A certain correlation has found between school chair height, character, comfort level and NLBP. Inappropriate manner of carrying school bag may lead to NLBP. Psychological factors may be an important factor for NLBP.5. After strict quality evaluation, consider the possibility of a higher measurement bias and publication bias. More relevant cohort studies should be carried out to provide solid and reliable evidences. [Chapter two]1. NLBP is a common disease of young people in Guangzhou city.2. The frequency of NLBP, the nature of NLBP, the need of medical service and the impact on daily life are all gender-related.3. There may be genetic factors in NLBP among school children and adolescents in Guangzhou. Basketball and gymnastics may increase the occurrence of NLBP. The perception that the school bag was too heavy was a risk factor for NLBP. The lower level of satisfaction the students felt with their academic may increase the risk of NLBP. Drinking could lead to NLBP.4. We selected a large sample, studied many factors and strictly controlled the process quality, and this study was the first case-control study on risk factors for NLBP among school children and adolescents in Guangzhou. We selected gender and class as matching factors, therefore the results were reliable. However, as the survey was in primary school students, mental development not yet mature, there may be recall bias. As the case-control study of exposure and disease have often difficult to judge the time. Cohort studies need to be conducted to further demonstrate the causal relationship.Innovation and shortage1. Innovations①This is the first systematic review with the prevalence and the risk factors for NLBP among school children and adolescents. Sensitivity analysis showed that the result of the Meta-analysis was stable and reliable, and provided good evidence for the prevention and treatment.②This is the first survey of the prevalence of NLBP among school children and adolescents in Guangzhou city. It can provide reference for the future research in China.③This is the first study of the risk factors for NLBP among school children and adolescents using case-control method.2. Shortages ①Attention should be paid when applying the conclusion as data provider bias and publication bias may exist.②The subject of the cross-sectional survey was only from two schools.