Psychosocial factors and their predictive value in chiropractic patients with low back pain: a prospective inception cohort study

The nature of pain

Pain is defined as an unpleasant sensory and emotional experience [11]. It is highly subjective being dependent on the individual's personal perceptions and therefore cannot be standardised as people respond in different ways to similar physical pain. The presence of adverse psychological factors, such as anxiety, fear or distress/depression, may have the effect of intensifying the perceived severity of pain and may play an important role in progress towards chronic pain and disability [12].

Acute pain gives rise to anxiety about its aetiology and prognosis [11], whereas chronic pain is distressing and may reinforce fears that the cause is serious and untreatable. In some individuals this may lead to feelings of helplessness and hopelessness and to withdrawal from social interaction. Beliefs also influence our perceptions of events, affect the way we cope and therefore impact significantly on an individual's response to pain and treatment [13].

Research in other populations

Better care strategies for patients at risk of poor outcome require that they first be recognised as being at risk. The literature suggests that the main psychosocial risk factors relate predominantly to depression, distress and role issues, especially with regard to work [1, 14–17]. Waddell, Burton and Main [18] found that the strongest psychosocial and socio-demographic predictors of chronic pain and disability were older age, poor general health/perceptions, abnormal pain behaviour, unemployment and expectations about return to work. However, different stakeholders may prioritise different outcomes. For example, pain relief may be the patient's priority, while return-to-work may be that of the employer and cost-of-care that of the insurer.

Research has identified risk factors with predictive value for chronicity in various public sector settings [19], notably in general practice [20] and physical therapy populations [21]. There is much less evidence in relation to chiropractic patients. Work-based strategies are effective, especially for sub-acute back pain [22] but are expensive and generally unavailable to small companies and the self-employed. Thus many such patients seek the help of chiropractors. These, however, may be different to patients in the public sector and we have no systematic knowledge of which patients who consult chiropractors are at risk of poor outcome in this population. The current study sought to discover which of the currently considered biopsychosocial risk factors for chronic disability are prominent and predict poorer outcome in non-specific back patients seeking help from chiropractors in the independent health sector.

Recruitment

Over a one-year period, 200 consecutive new patients contacting a chiropractic clinic in the market town of Salisbury in the UK for an appointment were asked by the clinic receptionists to confirm if their primary complaint was a new episode of pain in the lower back. If it was, they were informed about the study and that, if eligible and willing, their participation would involve the completion of two questionnaires prior to treatment on first attendance at the clinic, followed by one further, short (6 questions) questionnaire six weeks after initial presentation. This follow-up questionnaire was administered via the telephone. Two years on from initial presentation, all patients who had completed the 6-week follow-up were contacted to see if they would be willing to complete the 6-item questionnaire one more time.

Eligibility and consent

If the patient was willing in principle to participate, the receptionist was required to ask four questions to determine eligibility. To be eligible, patients had to be aged between 18 and 65 years, to have not undergone previous back surgery, not be pregnant and to not have pain below the knee.

If the patient remained eligible and willing to join the study, s/he was asked to attend the clinic 20 minutes prior to their scheduled appointment with the chiropractor. Upon arrival, the patient was given a consent form to read and sign. The receptionist was only allowed to answer clarifying questions if asked. The patient was also given an information sheet that they could take away with them detailing the study. A contact number for the senior researcher was provided on this form in case the patient had any queries or concerns. On completion of the questionnaires, the patient was then examined by the chiropractor who confirmed their final eligibility as having non-specific back pain. Reasons for non-eligibility were recorded.

When the receptionist received back the patient notes and eligibility form, the details of those patients for whom it was appropriate to join the study were entered onto a patient log. This recorded patient contact details, date of recruitment, the granting of consent and the completion date of the first questionnaires. Completed questionnaires, signed consent and eligibility forms and the patient logs were retrieved from the clinic on a weekly basis. A date for a 4 and 6-week follow-up for each participating patient was calculated. The purpose of the 4-week follow-up was to check that the original diagnosis of non-specific (simple) back pain had not changed. This was done by review of the patient's clinical notes at the practice. The results of this re-examination of the patient's notes were also recorded in the log.

Data collection

The baseline questionnaire requested information about age, gender, occupation, work status (part-time or full-time, shift work or not, whether the patient enjoyed their work and how much time, if any, they had taken off work with back pain in the past 3 years). Duration of current episode and chronicity [23] were determined, along with a standardised 'Core Set' of outcome measures that included bothersomeness, interference with work, attitude to persistent pain, days of reduced activities and days off work or school in the past 4 weeks [24]. The presence of aggravating features was also investigated. In addition, patients were asked to complete the Fear-Avoidance Beliefs Questionnaire (FABQ) [25], the inevitability scale of the Back Beliefs Questionnaire (BBQ) [26], the anxiety and coping scales of the Coping Strategies Questionnaire (CSQ) [27] and the 12-item version of the General Health Questionnaire (GHQ-12) [28] as a measure of psychological distress. At 6 week follow-up, pain impact was measured as at baseline, along with satisfaction with care on a 5-point numerical scale.

6 week follow-up

When the date for the 6 week follow-up was known, the patient was sent a reminder card showing the date and time for its completion. The time of day was determined by the patient's stated preference on the consent form as being the most convenient time for them to be contacted by telephone. This was also the case regarding where they were to be contacted, i.e. at home, work or on their mobile telephone. The patient was asked to keep the reminder somewhere prominent and to contact the investigators if the scheduled appointment was not convenient. An alternative appointment could be made for a period of up to 5 days from the original due date. On the agreed date and time, the patient was contacted by telephone and the questionnaire administered. The due date for completion of the 6 week follow-up was noted on the patient log, as was the date it was actually completed.

At 2 years, the number of attendances after the first 6-weeks from presentation was recorded from a note search. In addition, patients were contacted by telephone and asked to complete Deyo's core set [24] again, but without the item about satisfaction with care.

Analysis

Descriptive analysis of the baseline characteristics of participating patients was initially performed, followed by correlation analysis between selected baseline and follow-up variables. Baseline and 6 week outcomes (bothersomeness, cut down days for activity and for work and satisfaction with care) were compared and effect sizes [29] were calculated. Independent 2-sample t-tests were used to compare interval data from population subgroups. The follow-up outcomes of bothersomeness and interference with work were then dichotomised, with all scores of moderate and above taken as higher severity. Scores relating to psychosocial variables were averaged and then dichotomised, with a cut-off from 50% and above. Risk ratios were then calculated between baseline and follow-up variables, including the effect of co-morbidity.

Patients

One factor which may partly explain our inability to identify any psychosocial predictors of poor outcome in LBP patients may be the relatively low number of participants. It is also possible that this inability relates to the practice being in the independent sector. Studies [30] from the public healthcare sector report high scores on psychosocial assessment and distress to be significantly associated with non-recovery at one-year. However, in the current study, participants were recruited from private practice and were not particularly distressed at the time of initial presentation. Nor were the majority overly work-disabled by their condition despite reported high bothersomeness scores. Yet these are two common predictors of poor outcome in other studies [19]. Although a proportion of patients reported moderate to extreme interference with normal work due to LBP at baseline, this was not severe enough to stop the majority from working and certainly not for any protracted length of time. Virtually all who were off work returned to work, which may be due to promptness in seeking care. One large study of chiropractic practice in Europe [5] found that the majority of patients sought care within the first 4 weeks of onset and a further study that apparent high levels of satisfaction bear little relation to the degree of functional improvement achieved [31].

Outcome measures

The Deyo 'Core Set' of outcome measures proposed for low back pain research [24] was used in this study. This is a short, 6-item questionnaire which we chose for its conciseness within an otherwise large questionnaire battery [32]. The main measure of pain symptoms in this questionnaire was how much the pain bothered the patient. In a study seeking to classify primary care patients from general practice with low back pain, Dunn & Croft (2005) [33] found bothersomeness to be a valid measure of severity, being associated with measures of pain, disability, psychological health and work absence. Following its use as a measure of pain severity in a large UK back pain trial [34], Parsons et al (2006) [35] also used bothersomeness (anglicising the term to troublesomeness) in their comparative study which evaluated the troublesomeness of chronic, multi-site pain within individuals. In their validation study of the 'core set', Ferrer at al [32] concluded that it had the potential to be a useful tool in conjunction with other well-established outcome measures in future studies of LBP. However, they also concluded that, as their subjects and their back conditions were not typical of those presenting most commonly in primary care, further validation was needed before it could be widely recommended across LBP populations. While the 'core set' has been proposed as a concise research instrument, both Ferrer et al [32] and Dunn & Croft [33] concluded that further work is needed to verify the usefulness of bothersomeness in clinical practice. Moreover, if it is ultimately to be widely used as a single measure of severity, it will be important to be confident that both patients and clinicians interpret the term in the same way.

Predictors of outcome

Dunn and Croft (2005) [33] found that higher back pain bothersomeness at baseline in a UK general practice population predicted a greater risk of work absence at 6 months. Our chiropractic population did not, however, have appreciable work absence despite the level of bothersomeness. Moreover, although reported high levels of interference with normal work due to LBP at baseline was significantly associated with high levels of interference with work at 6 weeks the 95% confidence interval was very wide, suggesting that this finding should be interpreted with some caution.

The proportion in our sample with work absence, however, may be untypical. Sorensen et al (2006) [36], conducted a large survey of Danish chiropractic patients, (most of whom also consulted early and for low back pain) and found that that nearly a third had been off work, compared to our 16%. Most of their work loss was, as with our sample, of less than a week's duration. Our results also suggest that early intervention may be an important factor in successful care. While distress and depression are generally considered to be among the major predictors of poor outcome [2], these are, in any case, particularly prevalent when back pain has become chronic [37]. Although little is known about predictors of outcome that may be apparent in the very early stages, one small, but carefully controlled inception cohort study in general practice in France [38] found that delayed recovery (by 3 months) was associated with higher baseline disability and low self-related health. In the current study, the first of these was not a predictor. However, co-morbidity was associated with higher levels of interference with normal work at 6 weeks and it has been suggested that greater attention should be given to the existence of co-morbidities in the treatment of non-specific LBP [39].