The psychometric profile of chiropractic patients in Norway and England: using and comparing the generic versions of the STarT Back 5-item screening tool and the Bournemouth Questionnaire

Translation of the English version of the Bournemouth Questionnaire

A generic version of the BQ for use in all musculoskeletal pain conditions has been developed in English but has to our knowledge not undergone translation or validation. The BQ back pain questionnaire has been translated to Norwegian as part of a large multi-center predictor study [6, 12, 22] but the validation process has not been published. For the predictor study the BQ was first translated into Norwegian and then back translated to English by two different professional translators. The BQ developer compared the original and back-translated English versions for content and wording. The Norwegian version was then tested for patient usability on six subsequent treatment visits in 12 low back pain (LBP) patients. The change in BQ score was compared to the clinician’s record and found to reflect clinical change (unpublished appendix in the predictor study).

Our study utilized the Norwegian translated LBP BQ from the predictor study [12] as a template for the generic condition version we wanted to use. A group of clinicians who are members of the Norwegian clinical research network (KIP) compared the Norwegian LBP BQ to the English version of the generic BQ downloaded from the developer’s website [10]. All clinicians were experienced chiropractors and knowledgeable in British, Australian or American English. The translated versions were discussed, and adjusted to obtain consensus and close equivalence with the original version. Specifically, “painful complaint” was translated to “smerteproblem”, “complaint” to “plage”, and “anxious” to “urolig”. Back-translation was performed by two bilingual individuals with knowledge of Norwegian and English at a professional academic level, and with English as a native language. The original and the back-translated English versions of the generic BQ were later compared by three clinicians. If discrepancies were found, the Norwegian version was adjusted to optimize conceptual overlap. The translated generic BQ version was tested on Norwegian speaking patients presenting to a chiropractor, and no particular problems in answering the questions were reported.

STarT back tool

We used the 5-item STarT Back generic condition tool which is the 9-item psychosocial subscale modified to screen and identify distress in other conditions [5]. The questions addresses; fear (1 item from the Tampa Scale of Kinesiophobia), anxiety (1 item from the Hospital Anxiety and Depression Scale), pessimistic patient expectations (1 item from the Pain Catastrophizing Scale), low mood, (1 item from the Hospital Anxiety and Depression Scale) and how much the patient is bothered by their pain [7]. All items use a response format of ‘agree’ or ‘disagree’, with exception to the bothersomeness item, which uses a Likert scale. Scores range from 0 to 5 and patients scoring 4 or 5 are being classified as high psychosocial risk. We used the five psychometric questions from the 9-item Norwegian SBT version available from the Keele University website April 2012 [10].

Procedure Norwegian population

Objectives and practical procedures were explained to the participating chiropractors at a meeting for a network of chiropractic clinics volunteering in data collection for research where valid data collection and use of questionnaires in clinical decision-making was the topic. All patients who presented to a chiropractor with a painful complaint and were above the age of 18 and who were able to read and write Norwegian were invited to participate. Patients were excluded if they had attended a chiropractic consultation or received treatment less than 3 months prior to the initial visit. The SBT and BQ questionnaires were completed in the waiting room prior to first visit. After the consultation the chiropractor chose a diagnosis among a list of 26 commonly used diagnoses used in Norwegian chiropractic clinics based on the ICPC (International Classification of Primary Care) codes reported to the National Insurance Scheme of the Norwegian Labour and Welfare Service. The de-identified questionnaires were then transferred into an online survey tool (Web survey creator, Dipolar Pty Limited, Sydney, Australia) either by a secretary or the chiropractor. The paper forms were sent to the coordinator to check accuracy of data entry. The 26 diagnoses were further compiled into four areas of complaint, back, neck, extremity and other for convenience of analysis and comparison with other studies.

Procedure UK

The STB and BQ are routinely collected in the participating UK chiropractic clinics through a free online system (Care Response, London) accessed via any internet browser to help health care providers gather and report clinical outcome and patient satisfaction information and which has been previously described [23]. The data is stored in a database with secure encryption, which gives clinicians access only to information that relates to only their clinic. Data were collected with explicit consent to be used anonymously for research purposes.

The initial assessment completed by patients included a brief background medical history, and the 9-item SBT and BQ. A summary was provided for their practitioner. This included calculation of the SBT ranking of the patient’s risk of having disabling pain in three months. Outcome forms are automatically sent to patients by email 14, 30 and 90 days after starting care but for this study we only used the baseline data pertaining to the five psychosocial items. Data from LBP patients with the same inclusion criteria as the Norwegian participants with initial consultation during the month of May 2012 were extracted for analysis.

Ethics

The study was conducted in accordance with the Declaration of Helsinki, registered with the Norwegian Data Protection Authority (registration no 47214) and was deemed exempt from ethics committee approval. No variation was made to patient care as a consequence of this study and all participants consented to anonymous data being used for research purposes.

Data analysis

Demographics (age, sex, duration and area of pain) are presented as a simple distribution. The SBT was scored according to the methods specified by the instrument developers [10] and classified as “distressed by condition” if SBT score >4. BQ was scored on a 0–10 point scale for each question with a maximum total score of 70. Associations between the SBT and BQ were calculated using Spearman’s rank correlations in JMP 10, SAS institute Inc., Cary, NC, USA and in IBM SPSS 17 Statistics, IBM Corporation, Somers, NY, USA. The magnitude of the correlation coefficient was evaluated as small if 0.1-0.3, moderate if 0.3-0.5, and high if greater than 0.5 [24]. The Chi-squared test was used for categorical data and results given in medians with interquartile range (IQR) unless otherwise specified. We used an alpha level of .05 for all statistical tests.

Close association between BQ and SBT for depression scores

This study has provided a comparative analysis of the measuring properties of the SBT and the BQ for patients presenting to a chiropractic clinic. The BQ is only validated as an outcome measure. It is, as opposed to SBT, also validated to measure outcome of treatment for neck pain [11]. It has been developed to be a simple yet thorough tool to measure outcome of treatment. The psychosocial variables measuring anxiety and depression have shown poor responsiveness to change using BQ. Most likely because populations of patients presenting to chiropractic care have a low BQ baseline score (4 or below) for these items and consequently have too little change from baseline to follow-up when used as an outcome measure, i.e. a floor effect [16]. Here we have shown that these items may be of interest for the clinician if the patient present with a high score, especially for depression. This is apparent for all area of complaints included in this study. There is a low association between question 2 of the SBT and question 4 of the BQ addressing anxiety. The reason for this is unclear, but it is questionable if this SBT question actually represents the original anxiety component of the Hospital Anxiety and Depression Scale (HADS-Anx).

The study population are patients with a painful complaint and all groups had a fairly high pain score that was only slightly higher for patients with SBT score >4. There was a moderate to high association between the BQ pain question and the SBT bothersomeness question indicating that these items are related. Pain was less associated with the other SBT items and particularly not associated with the fear and anxiety items.

Patient profile

A Danish study indicate a tenfold risk of psychosocial risk factors such as depression, catastrophizing and fear avoidance being present in the patients SBT identified as “high risk” [25]. Eleven percent was found to be in the high-risk group in this study of Danish LBP patients. This is comparable to what we found in our Norwegian population indicating a similar profile of chiropractic patients in Scandinavia. In contrast, the high risk group has been reported as high as 28% in an UK LBP population [5]. Our study shows a significant difference in the patient profile at baseline between chiropractic patients in Norway and England. Norwegian patients are somewhat younger, are less distressed by condition in general, have significantly lower catastrophization and depression scores, but are mildly more anxious than English patients. It therefore appears to be important that validation of diagnostic tools is restricted to the country it is being utilized within, as significant variations between countries are present.

In both countries there were more women than men seeking care. This concurs with findings from a Canadian chiropractic teaching clinic [26]. In the Canadian study they also reported that a majority of patients seeking care had spinal pain (81.4%) similar to the Norwegian population in our study where 66% presented with back pain and 21% with neck pain. One in five in the Canadian study had an extremity complaint and 10% cervical pain while in our study one in five had a cervical complaint and 10% extremity complaint. In a Belgian study of chiropractic clinics 91.5% of patients presented with spinal pain complaints [27]. In contrast, the prevalence of the different musculoskeletal areas of complaint to those presenting to general practitioners (UK) is quite different to that presenting to chiropractors, the most common area of complaint is low back (16%), and neck pain is fifth (9%) after knee, chest and shoulder complaint [28]. In a Norwegian epidemiological study [29], 34% of the population reported low back pain and 36% neck pain within the last week.

In our study we have shown that the neck pain, LBP and extremity pain patients score differently on BQ total, SBT total, “distressed by condition” and on the different sub-questions for both the BQ and the SBT. The group categorized as “other” in our study however, is small and therefore any statistical significance should be treated with caution. However, these patients show a tendency towards a more distressed psychosocial profile that should be explored in a larger study. There is a need to develop good stratification tools and outcome measures for generic use since as many as one in three seeking care has a complaint other than LBP as well as more than one area of complaint.

Using questionnaires on different populations

The purpose of the SBT is to serve as a tool in deciding a treatment plan for patients presenting to primary care with low back pain, providing a more individualized treatment not based purely on the clinicians’ experience and expertise. Our study has found that the SBT psychological profile for LBP patients is similar to that found by other authors. The treatment package provided for the patient by different health care providers may differ in regards to addressing psychosocial issues depending on the clinician’s interest and formal training in treating musculoskeletal pain patients. This will influence the need to use stratification tools for referral but may give the qualified therapist a way to objectively target treatment for high risk patients.

As the BQ has never been used to stratify patients to “targeted treatment”, it is not possible to determine from our study whether the association between a high BQ score and high SBT score is clinically relevant. In a study assessing the SBT on chiropractic LBP patients showed that the high risk group had greater improvement and were equal to the median and low risk groups by the 30 day assessment [23]. Despite the high-risk group having greater pain and disability at baseline, they experienced greater improvement with regular chiropractic care. The chiropractic treatment paradigm may have aspects that address the psychosocial issues of the high-risk patient as measured by SBT, or the chiropractic patient population may differ from that of the general practitioners’. It is likely that the patients in both Field and Newell [23] and our studies were largely self-selecting and sought chiropractic care directly. In the UK population 41% of the patients had previously seen the chiropractor, and although this was not recorded for the Norwegian population or tested for, there is the possibility that an established therapeutic alliance may affect the patient’s psychological profile at different testing points.

The developers of SBT emphasize that an advantage of the tool is that it is validated for use in primary care, whilst other tools like the Oswestry Disability Index (ODI) and the Roland-Morris Disability Questionnaire are developed for use in secondary care. In the study by Hill and co-workers [5] the patients were referred to a physiotherapist, who then stratified the patient to “targeted treatment” through the use of SBT. Already at this stage in the stratification process the patient is once removed from the first point of contact in primary care. Evaluation of the implementation of diagnostic stratification tools in different areas of health care service may be of importance. This must be considered when using diagnostic and outcome instruments developed for use in a different health care settings than where it has been tested through RCTs.

Patients in primary care and secondary care may have different expectation to treatment success and thereby choice of treatment and this may also affect the baseline psychosocial profile of a patient population. In a study looking at minimal clinical important change in the ODI in a Danish population of LBP patients, one group treated by chiropractors in primary care and another group treated in the secondary sector of the Danish health care system, the investigators found a large difference in the reduction in ODI score needed for the patient to acknowledge a change. The percentage change scores from baseline were substantially higher in chiropractic patients in primary care (71%) compared to patients in secondary Care (27%) for a similar change in minimal important clinical change [30]. Although patients in primary care tend to be more acute and a rapid change in clinical presentation is expected, there may be an expectation of treatment outcome that are different for patients seeking chiropractic care.

BQ has never been validated as a stratifying tool, nor validated to identify predictors for chronicity. A study regarding BQ’s predictive and monitoring properties questioned its accuracy [12]. However, they did find that certain individual items were useful in predicting specific outcomes, making it still a useful predictive tool. They found that patients with a low score for BQ Work had a lower risk of being on sick-leave the following year. Patients scoring high on BQ Control had a higher risk of having LBP during the following year while a low score on this item or on BQ ADL predicted absence of debilitating LBP. They also found that a high score on BQ ADL, BQ Social and BQ Control predicted persistent LBP the following year. Although the use of BQ as a predictive instrument has been criticized [31] it is in use and has been found useful as part of a diagnostic package in clinical practice [32]. The generic use of BQ and SBT in general chiropractic practice is promising but should be used with caution until validated.

Limitations

A validated Norwegian translation of the SBT was not available at the time of our study and may differ from the latest validated version. The SBT has received much attention and is of interest for health care decision makers. The informal translations available on the Keele University website are in clinical use. If the tool is applied on a population where it has not been properly validated it may have unforeseen consequences. We have here addressed the necessity of proper validation for both the generic versions of the SBT and the BQ. The public may assume that the translation process is of the same scientifically quality as the rest of the published work of the research group since the developers have approved translations through written permissions and posted them on their website.

The diagnostic grouping in the Norwegian population is crude, with the “other” group being small and heterogenic. There may also be a bias in which patient the participating chiropractor entered in the database as there was a large difference in number of patients provided from the Norwegian chiropractors. In future studies the distribution of diagnosis should be compared with the reimbursement registry for chiropractors to ensure that the data is representative for regular practice. In this study we did not analyze the influence of age and duration on the psychosocial profile in the different diagnostic groups.

For stratification tools and outcome tools to be of value in routine practice, they need to be easy to collect and evaluate for the clinician. The effort of manually collecting and logging the patient forms from paper may have been the reason why not all Norwegian clinics were able to submit data to the study, this in spite of the fact that it took less than one minute to enter a patient’ data set into the online survey database. Using an online data collection tool with automated data collection from the patient at baseline and follow up, such as the one used for data collection in the UK study population, could improve the utilization of questionnaires for clinical use as well as for research.

Clinical relevance and further research

Clinicians using BQ routinely should be aware that patients with a high total score or with a high score on the depression question may have a psychosocial profile that should be taken into consideration when developing a treatment plan.

This study is the first to give a diagnostic and psychosocial profile of chiropractic patients in Norway. The results show that there is a significant difference in demographics and psychological profile between the Norwegian and UK chiropractic patients, showing that study results cannot necessarily be extrapolated between countries, prompting the need for research across borders, and caution when pooling data from different countries.

This study shows that both SBT and BQ generic versions for use in all musculoskeletal complaints may give valuable insight into the patient psychosocial profile and that there is a need for validation of the generic versions of these questionnaires.