Introduction
Low back pain (LBP) is a recurrent condition that is frequently diagnosed worldwide and found in high, middle and low-income countries; it can occur at least once in a lifetime, and impacts the person's quality of life and their disability-adjusted life years.1,2 LBP is also one of the main causes of retirement from work, generating economic uncertainty at the personal and governmental level.3 According to the WHO, the prevalence of non-specific LBP in industrialized countries ranges between 60% and 70%.4 In Colombia, Duque-Vera et al.5 reported in 2011 that the prevalence of LBP was 67.8% for a sample of professional nurses and workers in two hospitals in Manizales. Furthermore, according to Londoño et al.,6 who conducted a study in 6 representative cities of the country (Bogotá, Medellín, Cali, Barranquilla, Bucaramanga, and Cúcuta), in 2018, LBP was the third leading cause of rheumatic disease (7.24%; 95%CI: 6.28-8.34%), being slightly more prevalent in males than in females (8.58 vs. 8.19%).
Given the high prevalence of LBP, it is necessary to establish timely and adequate diagnosis and treatment with tools that allow establishing prognostic indicators and guide the treatment to be followed. It is also necessary to implement instruments that allow the early identification of physical, biomechanical, psychological, social, and occupational factors7 in patients with LBP that favor their treatment and follow-up.
LBP is a multicausal condition that strongly impacts the productivity of the patients who suffer from it. For this reason, all the factors that may cause it should be taken into account, given that psychosocial factors are usually overlooked, because a proper assessment of these factors, together with the physical aspects, can help physicians to make a complete prognosis.8
One of the most important instruments for timely diagnosis of LBP and implementation of proper treatment is the Start Back Screening Tool (SBST) described by Hill et al.,9 which is designed to guide the initial treatment plan of patients with LBP in primary care and to help health care staff identify subgroups that are at risk of persistent pain and disability. This instrument comprises 9 items that assess physical and psychosocial factors (previously described as indicators of poor prognosis) in order to classify patients into 3 subgroups: patients at low risk (total score <4), patients at medium risk (total score ≥4 and ≤4 on the psychosocial factors subscale), and patients at high risk (total score ≥4 and ≥4 on the psychosocial subscale).
The SBST has been subjected to various tests, studies and validations for its application in countries such as Turkey, where adequate parameters of reproducibility and reliability were found,8 and South Africa, where it was established that it had adequate internal consistency with a Cronbach's alpha of 0.68 and 0.77 at two moments of measurement. Furthermore, in the latter country, it was established that the instrument was well perceived by the population, since 33% of the participants in the study indicated that the instrument was very easy to understand and 40% stated that it was very easy to complete.10 Similarly, studies in Denmark11 and Sweden12 have shown the importance of using SBST to determine the prognosis of patients with LBP and to guide therapeutic decision-making in postoperative spine and hip surgery (the incidence of which is high in LBP) to contribute to the success of these interventions.
On the other hand, Pilz et al.13 published in 2014 a study that aimed to translate and cross-culturally adapt the SBST to Brazilian Portuguese and found that it obtained a reliability of 0.79 (95%CI:0.63-0.95) and an internal consistency of 0.74 for a total score 0.72 for the psychosocial items according to Cronbach's alpha. There is a Spanish version of the SBST questionnaire, which was validated in the Spanish population in the study conducted by Gusi et al.14 in 2011, where they found that the majority of participants interviewed reported that the questionnaire items were clear and understandable, although there was greater difficulty in understanding the disability and anxiety dimensions, as well as poor comprehension on the part of older adults.
Given the usefulness of this questionnaire, there is a need to calculate its metric properties for the Colombian population in order to use it successfully in the country. Thus, the objective of the present study was to determine the internal consistency, as well as the convergent, discriminant, and criterion validity of the SBST questionnaire14 in a Colombian sample.
Materials and methods
Study type and population
Quantitative and longitudinal scale validation study conducted in 68 patients (47 men and 21 women) aged 18 to 63 years (mean=38.32; SD=13.96) and diagnosed with nonspecific LBP, who had undergone a therapeutic intervention (10 sessions) between 2019 and 2020 at Clínica Universidad de la Sabana in Chía, Colombia.
Nonspecific LBP was defined as the presence of painful symptoms, both at rest and on movement, between the floating ribs and the gluteal region, with no other signs suggesting structural damage to the spine or adjacent structures. Chronic LBP was defined as a condition in which symptoms lasted for more than three months.14 The present study is derived from the doctoral thesis of Alfonso-Mora15, entitled El Método Meziéres para el "uso de sf en personas con dolor lumbar. Estudio Mixto (Meziéres Method for the "use of self" in people with low back pain. Mixed study).
Participants were recruited by purposive sampling and the following inclusion criteria were considered for their selection: a score >4 on the Roland-Morris Disability Questionnaire (RMDQ)16 and a score >2 on the Numerical Pain Rating Scale (NPRS) 17 2 weeks prior to the start of their participation in the study. On the other hand, exclusion criteria were: presenting red flag signs of LBP or a serious disease associated with LBP; having a body mass index >35; suffering LBP associated with compression or radicular lesion; and experiencing any physical modification that affected the lumbar or pelvic regions, such as vertebral fixations, segmental post-surgical changes, and/or tumors.
Instruments
The SBST was used to classify the risk of a poor prognosis of LBP. This is a self-reported scale that, as mentioned above, consists of 9 items that allow classifying the risk as low, medium, or high based on the complexity of the symptoms reported by the patient (including both physical and emotional health dimensions) and, in this way, determine the most effective clinical approach.
Furthermore, scales to assess pain and disability were included to test the concurrent validity of the SBST. Thus, pain intensity was measured using the NPRS,16 which uses a numerical scale from 1 to 10 (where 1 equals very low intensity and 10 equals the worst pain experience) and inquiries about pain intensity over a given period of time. For the purposes of this research, participants were requested to rate pain in the last two weeks.
In turn, the RMDQ questionnaire, which consists of 24 items that measure the level of limitation attributable to LBP in different activities of daily living, was used to assess disability. In order to complete it, patients must check the items they consider to be true regarding their limitation status due to LBP; thus, each checked answer is assigned 1 point, so the score may vary between 0 and 24, the latter being the highest level of disability attributable to LBP.17 The cut-off point for this scale is 4 points.
Procedures
Each patient was asked to complete the three questionnaires simultaneously at three different times: 1) before starting the physical therapy sessions; 2) immediately after completing the physical therapy sessions; and 3) six weeks after completing the physical therapy sessions. Ten one-hour sessions of exercises for LBP management were held twice a week. The exercises included neuromuscular techniques (contraction-relaxation and contraction-inhibition), stretching and strengthening, as well as manual myofascial release therapy, deep tissue massage on the muscular system of the lower back, and sedative therapy with physical means.
Statistical analysis
Data on the scores obtained with the three scales and at the three time points were entered into a Microsoft Excel database, which was exported to Jasp version 14.1 for subsequent analysis. First, the internal consistency of the SBST was calculated using Cronbach's alpha, omega, Guttman's lambda-6, and greatest lower bound (GLB) coefficients, where values >0.7 were considered as high internal consistency.18,19
Then, the construct and criterion validity of the SBST questionnaire was determined by establishing its correlations (using Pearson's correlation coefficient) with the NPRS and RMDQ scales, with the understanding that positive and significant correlations (p<0.05) imply that the instrument evaluates similar aspects: the greater the pain and the higher the level of disability, the greater the risk of poor prognosis.
Finally, the discriminant validity of the SBST questionnaire was estimated by means of a repeated measures analysis of variance (ANOVA) between the mean scores obtained at the three moments when the instrument was completed (pre-intervention and post-intervention, and at 6-week follow-up). This was done on the assumption that if there are significant differences between the mean scores obtained at the three time points, then the capacity of the questionnaire to differentiate between individuals with or without a poor prognostic risk of LBP is demonstrated. Similarly, the Bonferroni post hoc comparison test was used to indicate in which pair of measures significant differences are found. It should be noted that prior to the ANOVA, a Mauchly's sphericity test was performed to guarantee sphericity between the variances of the measurements. All analyses were performed in the Jasp software version 14.1, and a significance level of p<0.05 was considered in all analyses.
Ethical considerations
The study took into account the ethical principles for research involving human subjects established in the Declaration of Helsinki20 and the health research provisions of Resolution 8430 of 1993 issued by the Colombian Ministry of Health.21 In addition, the research was approved by the Academic Research Ethics Committee of the Clínica Universidad de la Sabana, as stated in the letter issued by that body on October 9, 2018. Informed consent was obtained from all participants.
Results
Table 1 shows the SBST internal consistency coefficients obtained, which range from 0.634 to 0.780. These results demonstrate acceptable and high internal consistency. The correlations between the SBST questionnaire and the NPRS scale and the RMDQ questionnaire were significant (Table 2), demonstrating the convergent validity of the SBST.
Table 1 Internal consistency coefficients of the Start Back Screening Tool questionnaire.
Criterion | McDonald | Cronbach | Guttman | GLB |
---|---|---|---|---|
Estimation | 0.648 | 0.634 | 0.664 | 0.780 |
Lower limit (95%) | 0.553 | 0.599 | 0.569 | 0.744 |
Upper limit (95%) | 0.759 | 0.666 | 0.792 | 0.891 |
Source: Own elaboration.
Table 2 Correlations between the Start Back Screening Tool questionnaire and the Numerical Pain Rating Scale and the Roland Morris Disability Questionnaire.
Variable | NPRS | RMDQ | SBST | |
---|---|---|---|---|
1. NPRS | Pearson's r | - | ||
p-value | - | |||
2. RMDQ | Pearson's r | 0.274 | - | |
p-value | 0.024 | - | ||
3. SBST | Pearson's r | 0.257 | 0.475 | - |
p-value | 0.035 | <0.001 | - |
ENCD: Numerical Pain Rating Scale; RMD: Roland-Morris Disability Questionnaire; SBST: Start Back Screening Tool.
Source: Own elaboration.
Regarding Mauchly's test for sphericity, the following values were found: W=0.974, chi-square=0.643, and p=0.725, which demonstrated the sphericity of the measurements and, therefore, no correction factors were proposed. With respect to the ANOVA, an F=33.722 and a significance <0.001 were obtained, finding statistically significant differences in the mean scores between the pre-intervention, post-intervention, and follow-up measurements. Table 3 displays the differences between measurements with Bonferroni post hoc comparisons.
Table 3 Post hoc comparisons of the mean scores obtained with the Start Back Screening Tool questionnaire between the three measurement moments.
Measurement moment | Difference of means | Standard deviation | t | p B onferroni | |
---|---|---|---|---|---|
Pre-test | Post-test | 2.577 | 0.328 | 7.863 | <0.001 |
Follow-up | 1.962 | 0.328 | 5.985 | <0.001 | |
Post-test | Follow-up | -0.615 | 0.328 | -1.878 | 0.199 |
Source: Own elaboration.
The data presented in Table 3 show that there are statistically higher values for SBST scores in the pre-intervention measures compared to the post-intervention and follow-up measures.
Discussion
The objective of this research was to determine the psychometric properties of the SBST in a Colombian sample. In this sense, the study data showed that the scale has acceptable to high internal consistency coefficients (Cronbach's alpha=0.634; omega=0.648; lamb-da-6=0.664; GLB=0.780), which allows us to state that this instrument has adequate internal consistency and reliability. Likewise, positive and significant correlations were found between the mean score obtained in the SBST and the mean scores in the NPRS (r=0.257; p=0.035) and the RMDQ (r=0.475; p<0.0010), demonstrating that the instrument has adequate convergent and criterion validity.
Moreover, in the repeated measures ANOVA, significant differences in mean SBST scores were observed between pre-intervention, post-intervention, and follow-up (6 weeks later) measurements (p<0.001). Likewise, according to the post hoc test, there was no significant association between post-treatment SBST scores and follow-up scores (p=0.199), but there was a significant association between pre-intervention scores and follow-up scores (p<0.001). This demonstrates that the scale can discriminate the level of poor prognostic risk between individuals who have experienced an improvement in their LBP following physical therapy interventions and those who have not undergone treatment; that is, the SBST has discriminatory ability.
As mentioned above, the SBST questionnaire has been validated in different countries, and in those studies, with respect to the internal consistency of the instrument, Cronbach's alpha values similar to the 0.634 reported here have been found: Pilz et al.,12 in a study conducted in 52 patients with LBP to validate the questionnaire in Brazilian Portuguese, found a Cronbach's alpha of 0.74; Schmidt & Naidoo,9 in a study conducted on 30 people to make a cross-cultural adaptation to isiZulu (language spoken in South Africa), found a Cronbach's Alpha of 0.68 and 0.77 for test 1 and test 2, respectively; Yilmaz-Yelvar et al.,7 in a study investigating the psychometric properties of the Turkish version of the SBST in 120 LBP patients, found a Cronbach's alpha of 0.747; Mbada et al.,22 in a validation study in Yoruba (a language spoken in West Africa) carried out on 100 patients, reported Cronbach's alpha of 0.85; Wiangkham et al.,23 in a study with 200 participants in which the SBST was translated and cross-culturally adapted to a Thai version, obtained a Cronbach's Alpha of 0.86; Billis et al.,24 in a cross-cultural adaptation and validation study of the SBST to Greek conducted on 124 patients, found Cronbach's alpha of 0.70; and Giusti et al.,25 in a study conducted on 146 patients evaluating the factorial structure and psychometric properties of the Italian version of the SBST, found a Cronbach's alpha of 0.64.
On the other hand, Fritz et al.,19 developed a prospective case series in which they aimed to describe the use of the SBST questionnaire in 214 patients from the United States receiving physical therapy for LBP. For this purpose, they compared the results obtained with this instrument with those of an NPRS and an adaptation of the Oswestry Low Back Pain Disability questionnaire, finding that the SBST is able to predict the level of disability in which the patient will be at the end of treatment. Along these lines, Flitz et al.19 stated that the SBST is able to discriminate risk by predicting pain intensity and disability at the beginning of physical therapy sessions. This finding is consistent with the correlations identified in the present study, the correlation with the RMDQ being stronger.
For their part, Beneciuk et al.,26 in a prospective observational cohort study conducted in 146 patients treated in Florida (United States), confirmed the predictive validity of the SBST questionnaire by comparing it with other psychological tests [Fear-Avoidance Beliefs Questionnaire (FABQ), Pain Catastrophizing Scale (PCS), 11-Item Version of the Tampa Scale of Kinesiophobia (TSK-11) and 9-Item Patient Health Questionnaire (PHQ-9)] and found significant correlations between the SBST and PHQ9, PCS and TSK-11 (r=0.28 to 0.63, p<0.01). On this point, it is worth noting that in the present study no other psychological tests were used to compare the results, suggesting future lines of study that could confirm our findings and thus provide further evidence of the validity and criterion validity of this questionnaire.
Although the present study found acceptable and high indicators of the reliability and internal consistency of the SBST in a sample of the Colombian population, we recommend conducting research with a larger number of participants and taking into account different types of LBP, which would allow identifying the applicability of the test in different patients with this condition. Furthermore, future research on the topic should measure the participants' perception of their understanding of the SBST. Finally, it should be kept in mind that the SBST is a self-report questionnaire, and therefore its proper administration requires guidance from a professional.