Table 3.

The usefulness of each patient characteristic in predicting whether a clinical scenario containing that characteristic was rated inappropriate or appropriate for each therapy (i.e., the percent of scenarios each characteristic helps classify)

Patient characteristics*	Predicting a rating of inappropriate**				Predicting a rating of appropriate**
	Mobilization		Manipulation		Mobilization		Manipulation
Presence of red flags	100.0%	↑	100.0%	↑	65.3%	↓	35.0%	↓
Prior unfavorable experience with spinal manual therapy	86.3%	↑	86.3%	↑	51.6%	↓	61.0%	↓
Major neurologic findings	18.7%	↑	86.3%	↑	41.5%	↓	27.1%	↓
Previous non-manual conservative care failed	13.0%	↓	24.5%	↓	---		---
Additional (e.g., advanced imaging) testing show cervical disc herniation, stenosis, or foraminal osteophytosis	6.5%	↑	2.9%***	↓	100.0%	↓	100.0%	↓
Clinically substantial traumatic etiology	---		21.9%	↑	30.7%	↓	23.5%	↓
No signs of painful/limited active range of motion	---		4.7%	↑	---		---
Radiographs showing advanced spinal degeneration	---		---		26.0%	↓	65.3%	↓
No cervical nerve root radiculopathy	---		---		7.6%	↓	---
Joint dysfunction in upper cervical spine	---		---		---		14.8%	↓
No local pathology	---		---		---		11.9%	↑
Continued psychosocial stress	---		---		---		---

^*The patient characteristics are all fully defined in the Appendix.

^**Predictions of a rating of inappropriate were versus ratings of equivocal or appropriate. Predictions of appropriate were versus ratings of equivocal or inappropriate.

^***Tests negative for serious pathology included here with other test results.

^↑= This patient characteristic predicts this rating.

^↓= This patient characteristic predicts against this rating.