Table 2 Summary of results (n = 5)

Kongsted et al. 2011 [17]. Feasibility of the STarT back screening tool in chiropractic clinics: a corss-sectional study of patients with low back pain

-Test Danish patients’ ability to fill out the Danish version of the SBT ^a

-to see if the SBT was able to identify 3 SBT subgroups in the study population

-to examine differences between the SBT 3 subgroups (in age, gender, symptoms, depression, fear avoidance beliefs and catastrophic coping strategies)

Cross-sectional

n = 475

19 chiropractic clinics in Denmark

Low: 59%

Medium: 29%

High: 11%

The high risk group had the greatest number of days with LBP ^e in previous year and previous 2 weeks as well as the longest current episode duration.

-Study population had a higher percentage of low risk participants than the population studied in SBT validation study (8) (59%) versus 47% in medical care setting).

-Primary outcomes: SBT, MDI^b, FABQ^c, CSQ^d

-Positive dose-response relationship between SBT and MDI scores (5% of low risk SBT group had signs of depression compared to 31% in high risk group).

-FABQ positively correlated with SBT risk groups (1% in low risk group, 31% in high risk group).

-CSQ positively correlated to SBT risk group (7% in low risk group, 55% in high risk group).

Only a small percentage of participants had high scores for primary outcomes percentages and pain and disability weren’t measured.

Field and Newell 2012 [18]. Relationship between STarT Back Screening Tool and prognosis for low back pain patients receiving spinal manipulative therapy

Compare outcomes for participants in the low, medium, and high risk SBT group after a course of usual chiropractic care.

Prospective cohort

n = 404

6 chiropractic clinics in England

Low: 42%

Medium: 32%

High: 27%

Over half (56.2%) of the sample had pain for <1 month at the onset of care.

-Primary outcome: PGIC^g

-Secondary outcomes: BQ pain subcategory and total BQ score

-BQ pain and total score was strongly associated with SBT high risk group at baseline, but by the 30 day follow-up, there was no difference between SBT groups for these 2 outcomes.

-SBT groupings were not statistically significantly associated with PGIC scores at any follow-up point.

-When stratified by gender, males in the SBT high risk group had 3 times the odds of poor outcome compared to low risk males at 90 days.

-SBT high risk group not statistically significantly associated with low pain improvement (defined as </=2 2 points on BQ pain scale). The high risk group improved just as much as the other risk groups at each follow up.

-Duration of current pain episode and reoccurrence of the pain/problem for >30 days in the last year provided some prognostic ability, but variance and predictive accuracy was low.

-Patients received usual care; it is unknown whether the chiropractors were tailoring treatment for high risk patients, according to their clinical expertise.

−36.5% of the sample was lost to follow-up by the 90 day endpoint.

Irgens et al. 2013 [19]. 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

Examine the correlation between the SBT and BQ ^fscores for low back pain patients presenting for chiropractic care in Norway and England.

Cross-sectional

n = 214

18 chiropractic clinics in Norway

n = 186

12 chiropractic clinics in England

Episode duration for both UK and Norwegian populations:

<3 weeks: 45%

>12 weeks: 37%

Norwegian patients were younger, less distressed by their condition, and had lower catastrophization and depression rates, but higher anxiety rates than their English counterparts.

-Positive association between BQ total score and SBT score for all areas of musculoskeletal complaint.

-Each BQ question was positively associated with overall SBT score.

-Strong association between the SBT depression sub-score and BQ low mood sub-score for neck pain, moderate for back pain, and low for extremity pain.

-Strong association between the BQ pain sub-score and the SBT bothersomeness sub-score for back pain in Norwegian participants (this association was moderate for UK participants).

-Moderate association between the BQ pain and SBT bothersomeness sub-score for neck pain in both countries.

-The association between BQ and SBT anxiety sub-scores was low to moderate.

-A validated Norwegian version of the SBT was not available at the time of the study.

Newell et al. 2015 [20]. Using the STarT Back Tool: does timing of stratification matter?

Determine if categorizing participants to SBT group at care onset led to differences in prognostic accuracy compared to those categorized 2 days after their first chiropractic visit.

Prospective cohort

n = 749

11 chiropractic clinics in the UK

Low: 39%

Medium: 37%

High: 24%

Patients over 16 years of age with nonspecific LBP who completed a BQ ^f at the onset of care.

Duration of Pain at onset:

<1 mo: 43%

1–3 mo: 10%

>3 mo: 47%

Those in high risk SBT category were older with a more acute presentation and higher condition severity.

-Primary Outcomes: PGIC^g and BQ^f

-No difference in SBT prognostic ability in participants categorized before versus after initial visit.

-Medium and high risk groups had a greater change in pain and total BQ^f score than low risk group.

-Improvement at 14 and 30 day follow-up predicted improvement at 90 day follow-up.

- > 1/3 of participants changed SBT category in the time just before initial visit and 2 days post initial visit.

-SBT category post initial visit was predictive of 30 day follow-up outcomes in medium risk group, when adjusting for baseline variables.

-medium risk group improved more than high risk group in spite of fewer treatment visits

−58% drop out rate by 90 day follow-up

-Patients were self referring for chiropractic care; results may not be generalizable to entire nonspecific LBP population.

Kongsted et al. 2016 [9]. Prediction of outcome in patients with low back pain—A prospective cohort study comparing clinicians’ prediction with those of the Start Back Tool

Determine how clinicians’ expectations performed compared to that of the SBT and to what extent combining clinicians’ expectations with the SBT increased the amount of variation explained in outcome.

Cross-sectional (secondary analysis of a 2014 study by Eirikstoft and Kongsted) and a Longitudinal component.

n = 859

(710 responded at 2-weeks, 676 at 3 months, 636 at 12 months)

17 chiropractic offices in Denmark

Overall group mean scores not provided in this analysis

-Primary outcomes: NRS^h, RMDQ ⁱ

-The ability of both clinicians and SBT to predict future outcomes was low.

-Clinicians’ expectations combined with the SBT were slightly better at predicting activity limitation outcomes (RMDQ) than either of the two by themselves, but the proportions of those accurately predicted was still low.

-The measure of clinicians’ expectations was not validated and it is uncertain how clinicians’ define terms like “short/uncomplicated” versus “prolonged” and “long-lasting.”

Field and Newell 2016 [22]. Clinical outcomes in a large cohort of musculoskeletal patients undergoing chiropractic care in the United Kingdom: a comparison of self- and national health service-referred routes

Compare outcomes of self-referred and National Health Service-referred patients presenting for chiropractic care.

Prospective cohort

n = 8,222

Chiropractic clinics in the south of the UK.

National Health Service funded patients:

Low: 26%

Medium: 35%

High: 39%

Privately funded patients:

Low: 46%

Medium: 32%

High: 22%

-National Health Service funded patients had a statistically significantly lower percentage of patients in the low risk group compared to the self-referred group.

-National Health Service referred patients were more chronic, in more distress, displayed more co-morbidity, received more treatment visits, but were less likely to continue care past 30 days than self-referred patients.

-Primary Outcomes: BQ^fand PGIC ^g

-Group * Time interactions existed for medium and high risk SBT categories.

-Possible between-group differences in patient expectations.

-Results may not be generalizable to all patients presenting for spinal care, or patients outside of the south of the UK.

Morso et al. 2016 [21]. The prognostic ability of the StarT Back Tool was affected by episode duration

To determine whether the SBT’s prognostic ability was affected by care setting, baseline episode duration, and outcome time point being predicted.

Secondary analysis of prospective cohort data

n = 416

17 chiropractic clinics in Denmark

n = 265 (General medical practice)

Taken from a Danish audit project

n = 974 (secondary care outpatient medical setting)

Spine Centre in Denmark

n = 200 (physiotherapy)

27 physiotherapy offices in Denmark

Low: 52%

Medium: 39%

High: 10%

-There were statistically significant differences across care settings for most baseline characteristics (age, gender, current episode duration, previous LBP episodes, pain intensity, activity limitation, and SBT score).

-Most chiropractic patients were younger, male, presenting with very short episode duration (across all SBT risk groups), and most likely to score in low risk SBT group. In contrast, those at the Spine Centre had the longest episode duration.

-Primary outcome: RMDQ

-The prognostic ability of the high risk versus low risk SBT group was stronger for those whose episode duration was 2 weeks or longer. The prognostic ability of the medium risk group versus the low risk group was different for people whose episode duration was greater than 12 weeks.

-low risk SBT groups had the lowest activity limitation scores.

-Across all care settings, the SBT’s prognostic ability is weakest for those whose episode duration is < 2 weeks. -Chiropractic settings had the greatest percentage of patients with this short episode duration.

-Secondary data analysis; there may have been other variables that had an influence on outcomes.

-Measurements weren’t made on all cohorts at all time points, which may have weakened the capacity to separate the effects of setting and outcome time points.