Prevalence and associated characteristics of recurrent non-specific low back pain in Zimbabwean adolescents: a cross-sectional study

The study used a cross-sectional design and was conducted in government secondary schools based in Harare, Zimbabwe. The government-administered secondary schools are classified into two categories (S₁ and S₂) based on location. The S₁ schools are located in the low density areas where people of high socio-economic status live [25]. The S₂ schools are located in the high density suburbs where people of lowest socio-economic status reside [25]. At the time of the study, a total of 71 458 adolescents were registered in government secondary schools.

As no study has been conducted in Zimbabwe on adolescent recurrent NSLBP, Epi Info version 7.1.1.0 Statcalc package for population surveys was used to calculate the minimum sample size. The following parameters were considered: target population (71 458), estimated prevalence of recurrent NSLBP in school-children found regionally (13.5%) [26], precision effect 3%, design effect (1%) and 95% confidence interval. The sample size was calculated to be 495. However, the number was adjusted upwards in anticipation of attrition from school absenteeism and possible refusals. The final sample size was 620 students.

A two-stage cluster sampling method was used to select schools and study participants. There are 55 relevant government-administered secondary schools in Harare, Zimbabwe. There are 17 secondary schools in the S₁ and 38 secondary schools in the S₂ category. In the first-stage of sampling, a list of all government secondary schools was constructed in clusters of two categories, S₁ and S₂. Each school in the two clusters was assigned a numerical number. Based on a strategy that considered proportionality between the S₁ and S₂ schools, one secondary school was randomly selected from the S₁ category and two from S₂ cluster box.

In the second stage of sampling, each school selected was asked to provide a list of all classes in each form. Six classes were randomly selected where all the school-children in the relevant classes were then eligible to participate in the study. Inclusion criteria considered adolescents registered full-time in the participating schools. The World Health Organisation (WHO) definition of adolescents was adopted as people ranging between 10 and 19 years [16]. In addition, school-children eligibility entailed obtaining parental/legal guardians’ permission and submission of a completed medical health questionnaire. However, participants were excluded if they had parental reports of spinal pathologies, orthopaedic conditions, a previous history of trauma to the back, any central or peripheral nervous problem affecting muscle quality in the extremities and trunk. In addition, they were excluded if they had overt or covert physical deformities such as leg length discrepancy, scoliosis and lordosis [27].

As there was no Zimbabwean questionnaire on adolescent recurrent NSLBP, questions asked in the questionnaire were derived from previously validated instruments used in literature [10, 16, 27‐29]. The choice of a questionnaire as a survey tool was based on the fact that pain is a highly subjective phenomenon best evaluated by self-report [29‐31].

The first section of the questionnaire gathered information on the primary outcome measures of lifetime prevalence, point prevalence and the prevalence of recurrent NSLBP. For lifetime prevalence, respondents were specifically asked the following question “Have you ever experienced pain or discomfort in the lower part of your back which lasted for one day (24 hours) or longer in your life, not associated with menstruation in females?” To assist the respondents in understanding the anatomical region of the lower back, a mannequin was used with an arrow pointing to a posterior view of the lumber region. Subsequently, respondents were asked to provide the age when they first experienced an episode of low back pain in order to determine the mean age of onset of the condition.

For point prevalence, respondents were specifically asked the following question “Do you have pain in the lower part of your back now?” The operational definition of recurrent NSLBP adopted was “pain which occurred at least two times over the past year with each episode of lasting at least 24 hours, with pain intensity of greater than two on VAS with at least a 30-day pain free period between the episodes” [21]. Recurrent NSLBP was specifically asked regarding the last 12 months. In addition, to understand the nature and the character of the recurrent NSLBP experienced by adolescents, the researcher sought information on the frequency, intensity and the duration of the recurrent episodes. Pain intensity was evaluated based on the VAS from 0 (no pain) to 10 (maximum pain). The health-seeking behaviour for adolescents with recurrent NSLBP was also ascertained by the following question “Have you ever sought treatment for non-specific recurrent low back pain or information from any health-care professional regarding the pain that you frequently feel? This behaviour described seeking formal or informal health care services.

The medical health questionnaire was used to determine the medical history of school-children as reported by parents or legal guardians. The questionnaire was adapted and modified to suit the design of the research study from a study conducted by Fanucchi et al. [27]. The questionnaire provided the criteria for exclusion from the study. Students were excluded if they had spinal pathologies, or deformities such as scoliosis, spinal tumours, spinal trauma, and orthopaedic conditions such as fractures of the pelvis and lower limbs, leg length discrepancy and any neurological conditions which alter muscle tone of the lower limbs and the spine. These conditions fell out of the scope of the study.

Prior to use, the low back pain study questionnaire was subjected to content validity testing and one-week test-retest reliability test. Five lecturers from the Department of Nursing, Physiotherapy and Community Medicine at the University of Zimbabwe were used as content experts. They were chosen on the basis of experience in epidemiological research studies and on preferential interest in musculoskeletal research. Subsequently, the questionnaire was administered to a sample of 40 students twice. During the initial test, students were not told that they would be re-tested again after one week. For the primary outcome measures of NSLBP, moderate to substantial kappa coefficients (0.48-0.72) were observed (Table 1). These results were consistent with findings from other studies [29, 32]

The study was approved by the Human Research Ethics Committee (HREC) from the University of Cape Town [ref: 189/2012], Medical Research Council of Zimbabwe (MRCZ) [ref: MRCZ/B/356], Ministry of Education, Sports, Arts and Culture [ref: C/426/3] and Harare Provincial Education Director [ref: G/42/1]. School-children as well as their parents/legal guardians gave assent and consent respectively to participate in the study. The fieldwork for the first phase of the study was conducted during the second academic school term. The fieldwork was subdivided into three distinct stages: preparatory stage, intermediate stage, and questionnaire administration.

After the schools had agreed to participate, the researcher (MC) visited the schools to establish background information. The researcher utilised this opportunity to address school authorities on the nature and procedural issues of the research project and to agree on specific dates for data collection. In addition, the researcher utilised the opportunity to identify the participating classes and students in each of the participating secondary schools for the main study.

On an agreed date, the researcher attended the selected schools to address the participating students. Verbal information was given explaining the rationale, procedure, risks and benefits of the research project to the students. At that stage, students who verbally agreed to participate were given a letter with standard information regarding the study, and consent forms for parents/legal guardians. In addition, students were given the medical health questionnaire to be completed by parents/legal guardian at home. The students were given seven days to return the parental documents to the class teacher in a sealed envelope. During this period, reminders were sent to the respective class teachers using cell-phone text messages and voice calls to ensure prompt return of the documents by students. The researcher collected returned parental documents from the class teachers intermittently to the end of the seven days.

On agreed dates with the school authorities, the researcher visited the three participating schools consecutively. Because each school had six participating classes, a maximum of three days of data collection were spent at each school. This was done to minimise disruptions of scheduled lessons. Eligible students were given an information sheet to read then sign an assent form. The questionnaires were completed in the classrooms during school hours in the presence of the researcher and the class teacher.

Raw data from the questionnaires was entered into Microsoft Excel before being imported to STATISTICA version 11 for analysis. Questionnaires with missing data were not analysed. Descriptive statistics were used to describe baseline demographic characteristics of the participants, including means with standard deviations for continuous data and frequencies for categorical data. The primary outcome measures of lifetime prevalence, point prevalence and prevalence of recurrent NSLBP were expressed as percentage of the total population. Exact 95% confidence intervals (CI) were given. Chi-square test was used to evaluate the effect of gender and age on NSLBP prevalence. The severity of recurrent NSLBP was evaluated using VAS and the independent t-test was used to assess the difference between males and females. Pain intensity was dichotomised as mild pain (1–4 score on VAS) and severe pain (5–10 score on VAS) [16].

Table 2 depicts the demographic characteristics of the study participants. The mean age of the sample was 16 ± 1.72 years (range 13–19 years). Of the total school-children, 53.8% (n = 286) were female students. As indicated by the independent t-test, there was a significant difference in the mean age of participants between male students (16.2 ± 1.79) and female students (15.8 ± 1.65) [t (530) = 2.34, p = 0.02].

The lifetime prevalence was 42.9% (Table 3). There was no significant difference in the lifetime prevalence between male students (42.7%) and female students (43.0%) as indicated by the Chi-square test [X ²(1) = 0.006, p = 0.94]. There was a significant trend towards increase in the proportion of school-children with NSLBP as age increased despite some inconsistencies at some ages [X ² for linear trend = 73.3, p < 0.001] (Figure 2). The mean age of onset for NSLBP in the school-children was 14.4 ± 1.90 years. There were different peaks of onset between the sexes. Figure 3 shows that NSLBP peaked earlier in female students. The mean age at onset was 13.9 ± 1.91 years for female students compared to 15.0 ± 1.75 years for male students [t(226) = 4.21, p < 0.001].

Point prevalence indicated that 9.2% of the school-children had NSLBP on the day of questionnaire adminstration. Female students were affected significantly compared to male students [X ² (1) = 11.2, p < 0.001]. Overall, there was a significant trend towards increase in the proportion of school-children with point NSLBP as age incresaed [X ² _trend = 29.8, p < 0.001] (Figure 2). The mean intensity of point NSLBP was 3.5 ± 1.64 on the VAS. An independent t- test showed no significant difference between male students (3.08 ± 1.12) and female students (3.68 ± 1.76) in the intenstity [t(51) = -1.15, p = 0.26]. In adittion, as indicated by the one-way analysis of variance (ANOVA) test, there were no significant differences in the mean intensity of point NSLBP across the age categories [F (6,46) = 1.69, p = 0.14].

The prevalence of recurrent NSLBP was 28.8%. Male students were more affected (29.7%) compared to female students (28.0%), although this was not statistically significant [X² (1) = 0.19, p = 0.67]. The prevalence of NSLBP increased with age despite a slight drop for the 14 year olds (Figure 2).

Table 4 presents data on the characteristics of recurrent NSLBP. The majority of school-children (n = 82, 53.6%) experienced more than three episodes in 12 months. However, episodes of recurrent NSLBP were reported to last for less than a week in the majority, with both male and female students equally affected. Of the 153 school-children with recurrent NSLBP, 26.8% (n = 41) reported having sought medical treatment for the symptoms. Health-seeking behaviour was not gender-related. However, it was significantly related to the report of sciatica [χ ² (1) =47.9, p < 0.001]. Sciatica was reported in 20.9% (n = 32) of school-children with recurrent NSLBP. The mean intensity of recurrent NSLBP reported was 4.8 ± 1.94 on the VAS. Female students reported to experience intense episodes of recurrent NSLBP [t (151) = -5.66, p < 0.001].

Health-seeking behaviour was significantly associated with pain intensity [X ² (1) = 6.55, p = 0.01]. Moreover, there was a significant difference in the mean intensity of recurrent NSLBP reported by adolescents who sought treatment compared to those who did not [t (151) = -5.03, p < 0.001].

The lifetime prevalence rate observed in this study falls within the range (7%-80%) reported in systematic reviews of adolescent NSLBP studies [3, 8, 9, 23]. Variations in the prevalence figures have been attributed to methodological or population differences [9]. Using a relatively small sample size, a lifetime prevalence of 57.8% was reported in 400 school-children aged between 10 and 18 years in Kuwait [15]. These findings were consistent with results obtained in South Africa using a relatively larger sample of 1 123 adolescents between the ages of 13 and 18 years [17]. In contrast, a large cross-sectional study conducted among 43 630 school-children aged between 9 and 15 years, found a relatively lower lifetime prevalence of 32.1% in Japan [33]. However, the study could be criticised on the grounds that the authors left the judgment of NSLBP to the respondents. In addition, the sample was relatively younger compared to the present study sample. Nevertheless, the possibility that lifetime prevalence could have been under-reported in the present study cannot be over-ruled. This is because the screening question on lifetime prevalence had a kappa coefficient of 0.72 indicating substantial but not perfect agreement. This finding highlights the possible existence of recall bias secondary to a phenomenon described as memory decay [34]. It is possible that adolescents forget past lifetime episodes of NSLBP resulting in under-estimated prevalence figures.

The prevalence increased with advancing age despite inconsistencies at some ages. This has been reported in other cross-sectional studies [13, 16]. This indicates that NSLBP becomes increasingly common with age. These findings together with a well-known fact that adolescent NSLBP predicts adult NSLBP expose a significant future health-care problem in Zimbabwe [23]. In addition, NSLBP affects males and females equally during adolescence. This is consistent with other studies. A cross-sectional study conducted in Nigeria reported lifetime prevalence figures of 59.7% and 57.5% for male and female students respectively [χ² = 1.49, p = 0.222] among 4 400 school-children aged between 10 and 19 years [16]. There is no explanation that has been postulated in the literature for this [16]. In contrast, the majority of studies indicate to a higher prevalence of NSLBP in females compared to males [11, 30, 35]. Other studies report a higher prevalence in males [17, 36]. Although the actual reasons for the higher prevalence of spinal pain among females are unknown, several authors have attributed this to menses and accelerated growth spurt [27, 37, 38]. On the other hand, Jordaan et al. [17] cited higher concentration of testosterone during pubertal growth in males for the higher prevalence.

A small subset (28.8%) of adolescents reported having experienced recurrent NSLBP in the last 12 months. Both sexes were equally affected and the prevalence increased with chronological age. These findings provide support to recent epidemiological reports that NSLBP is recurrent in nature and confirms the indiscriminate nature of the condition [10, 13]. There is sparse data available regarding the prevalence of recurrent NSLBP in adolescents in the literature. The few studies available have reported different rates because of definitional issues. This makes comparison between studies difficult [13].

In Norway, recurrent NSLBP was reported to be 32% in adolescents with mean age of 14.7 years [18]. Apart from the small sample size used (n = 88), the operational definition of recurrent NSLBP considered pain experienced for more than seven days for over 12 months. There was no specific emphasis on the intensity and frequency of episodes that constituted a recurrence of NSLBP. In contrast, a cross-sectional study conducted in the United Kingdom among 500 school-children aged 10 and 16 years reported a lower prevalence of 13.1% [13]. Recurrent NSLBP was specified as pain experienced “regularly” without clarity on duration, intensity and frequency of episodes [13]. Similarly, a study conducted in Mozambique reported a recurrent prevalence of 13.5% in a study involving 204 school-children aged between 11 and 16 years [26].

The seemingly high prevalence of recurrent NSLBP in Zimbabwean adolescents could be attributed to several factors. Recurrent NSLBP may indeed be a problem of concern among Zimbabwean adolescents. Since the study had no monetary benefits, it is highly unlikely that the school-children exaggerated or reported a non-existent problem. However, due to the relatively small sample size used (n = 153) and the cross-sectional nature of the study, it is possible that the 12-month prevalence of recurrent NSLBP could have been overestimated through a concept of forward telescoping [34]. The reliability of the screening question for recurrent NSLBP showed moderate agreement (k = 0.51) between the test and re-test. In addition, the adopted definition of recurrent NSLBP has not been validated in adolescents.

A number of parameters were evaluated to assess the characteristics of recurrent NSLBP reported in Zimbabwean adolescents. This was done to understand the nature of recurrent NSLBP in adolescents with regard to frequency, intensity, and duration of episodes. However, a limited number of studies have specifically examined these aspects of recurrent NSLBP. In the present study, more than half (n = 82) of the school-children with recurrent NSLBP reported having experienced at least three episodes of NSLBP in 12 months. This strongly indicates that recurrent NSLBP is a common phenomenon and adolescents are vulnerable to many recurring episodes in a period of time; a finding shared by many authors [13, 23, 24]. In the majority of adolescents (85.6%), an episode of NSLBP was reported to last for a short period of time (less than seven days). Rarely was an episode reported to last for more than two weeks. This could reflect a favourable natural history of recurrent NSLBP in adolescents, a finding supported by Burton et al. [22]. This finding is consistent with previous research conducted by Jones et al. [13] who found that NSLBP episodes lasted less than seven days in the majority of school-children.

The school-children were asked to rate, on average, the perceived intensity of the various episodes of NSLBP they experienced over a 12 months period. The mean intensity was reported to be 4.8 on the VAS. These findings support the conclusions of many epidemiological studies that reported on adolescent LBP being benign or mild [16, 29, 39]. Although with no clear details on the method used to evaluate the intensity of back pain, Ayanniyi et al. [16] observed that 56% (n = 3185) of the school-children with back pain reported the pain to be mild. From a primary prevention perspective, this finding is comforting and partially dismisses the need to lend exaggerated importance to symptoms of NSLBP in adolescents. However, recurrent NSLBP was reported to be intense in females than in males. This is not surprising as males and females have been reported to perceive pain differently [18]. Females are known to have a lower pain tolerance and lower pain thresholds [11].

NSLBP has been reported to occur with or without symptoms of sciatica [40]. The present study found that a small proportion of adolescents (20.9%) had recurrent NSLBP associated with sciatic symptoms. The presence of sciatica in NSLBP patients has been regarded as an important indicator for severe and continuous pain [24, 41]. This suggests that adolescents with recurrent NSLBP characterised by sciatic symptoms deserve attention. In the present study, health-seeking behaviour was associated with sciatic symptoms. Few studies have specifically evaluated the presence of sciatica in adolescents with recurrent NSLBP. In a clinical study conducted on 36 hospitalised patients between the ages of 10 and 18 years, Bockowski et al. [41] found that 52% had NSLBP and sciatica. Despite the small sample size used, the study highlighted the common occurrence of sciatica in adolescents with NSLBP. Watson et al. [30] reported a 31% prevalence of sciatica among 1 446 school-children aged between 11–14 years. Harreby et al. [28] considered sciatica as radiating pain down to the leg to below the knee and reported a prevalence of 4.7% among a cohort of 1 389 Danish school-children aged between 13 and 16 years.

The study uniquely investigated lifetime prevalence, point prevalence and prevalence of recurrent NSLBP across all the age categories of adolescents in schools in Zimbabwe. Recurrent NSLBP cases were identified based on a consensus-agreed definition of the condition [21]. However, it is premature to accept and generalise these findings until such a study has been repeated. The response rate from both the school-children and parents was satisfactory. Non-participation bias had no effect on the observed findings. In addition, the study questionnaire had content validity and moderate to substantial reliability.

The main limitation of this study was the cross-sectional and subjective nature of the data collected. A cause-effect relationship cannot be deduced. Reliance on subjective morbidity for retrospective data creates a possibility of recollection bias. This may result in imprecise prevalence figures. However, since pain has been described as a subjective phenomenon, subjective recall has been regarded as the most valid way to assess pain [13]. In addition, two methods were used to assist respondents appreciate the anatomical region of the lower back in an attempt to avoid the influence of recall bias: a direct question and a pre-shaded manikin, with an arrow pointing to a posterior view of the lumber region.