The aim of this survey was to determine if management intended by chiropractors is in accordance with two key recommendations in the Australian evidence-based acute LBP guideline  and more generally how chiropractors intend to support or manage people with acute LBP. Ours is the first study undertaken in Australia, that we are aware of, that has investigated these two key recommendations for people with acute LBP in chiropractic practice.
Where an x-ray was not recommended according to the guideline 68% of chiropractors responded that they would order or take an x-ray, indicating an overall x-ray guideline adherence of 32%. It has been reported previously that chiropractors use plain x-rays at a high frequency for people with acute non-specific LBP, with rates varying from 12% to 63% [15–22]. Our findings are consistent with these studies and suggest that a high proportion of chiropractors who responded to this survey are not complying with the guideline on the use of plain x-ray for acute LBP. Some chiropractors order or take plain x-rays to rule out pathology, screen for contraindications to manipulation, assist in the selection and delivery of treatment, and monitor patient progress . A post-hoc review of the literature found no substantive evidence of harm from chiropractic manipulation because plain x-ray had not been used nor enhanced efficacy.
Across all five vignettes approximately 25% (95% CI: 20%, 29%) of chiropractors would take, or order, a full spine plain x-ray. This was despite the lack of information in the vignettes about any symptoms or examination findings at the other regions of the spine. There is no evidence that full spine x-rays are warranted in these vignettes and this finding is troubling considering the radiation dose that the patient would receive [24, 25].
Around one third of chiropractors surveyed (35%; 95% CI: 29%, 40%) would perform spinal manipulation of the spine in a patient with a likely fracture. We did not investigate the reasons for this, however it may be that a) the potential for fracture diagnosis in the vignette was not considered, or b) the chiropractor intended to apply manipulation to regions above or below the fracture site.
The findings above have implications for patients' safety and health resource use. First, a full spine x-ray will expose the patient to needless ionising radiation, and manipulation of a spinal fracture may worsen the fracture with potentially serious consequences. Second, the findings have implications for third party payments for these extensive and expensive diagnostic imaging.
Giving advice to stay active
Over half the chiropractors (51%; 95% CI: 47%, 56%) indicated that in the first four vignettes they would advise the patient to stay active, while 60% (95% CI: 57%, 64%) responded they would give exercises (back specific or general) to people seen who were similar to the first four vignettes. In this sample of chiropractors, there appears to be a need to encourage the giving of advice to patients with acute LBP to stay active, regardless of any exercises they would recommend to the patient in order to bring chiropractic management in line with evidence-based practice.
The response rate was low (37%) and therefore the results have an increased potential for selection bias where those who choose to participate differ in important characteristics with those who do not choose to participate . It may be speculated that those who responded were more likely to comply with the guideline, in which case we have calculated underestimates of guideline compliance. If this is the case the problem may be greater than our results demonstrate. It also may be speculated that those who responded were less likely to comply with the guideline, in which case we have calculated overestimates of guideline compliance. Regardless, we can only conclude for the respondents to this survey and are not able to generalise to the broader population of chiropractors in these three States of Australia. No data were available for the broader chiropractic population at the time we undertook this study, so we were unable to compare the characteristics of the respondents to the characteristics of the broader Australian chiropractic population to investigate non-response bias. The response rate is lower than those seen in other surveys of health professionals where a mean of 58% has been reported, however it has also been demonstrated that response rates of health professionals are decreasing over time . Recently published surveys of chiropractors have reported response rates ranging from 38% to 88% [27–31].
Patient vignettes are limited by a lack of detail that may be gleaned during a patient consultation, including visual cues relating to pain during an examination. Further, patient vignettes may not be sensitive enough to pick up the fine nuances of practice that would better guide the practitioner . Therefore, our conclusions are based on a measure of proxy behaviour rather than real behaviour. More research is needed to establish the extent to which proxy measures of behaviour can predict actual behaviour.