Table 4.

Significant Prognostic Factors identified in all included studies

ID	First Author	Statistical analysis	Outcomes measured	Statistically significanta predictors of poor outcome	Strength of association	Statistically significanta predictors of good outcome	Strength of association	Comments
1	Balague [12]	Multivariate analysis (stepwise logistic regression)	"Recovery" (composite score including pain, disability & muscle strength) Recovery defined as: ODI Score ≤ 20 VAS pain ≤ 15 Normal muscle strength test (score 5)	Positive neurological examination (Neurotot)	OR 4.3 (95%CI; 1.37, 13.28)			It is unclear whether the odds ratio given is crude or adjusted.
2	Beauvais [13]	Recovery and failure groups compared using Fishers test, Chi squared test or Wilcoxon test	"Recovery" Complete = return to usual work/activities, little or no analgesia Partial = residual pain, frequent analgesic use, complete or partial return to work, limited athletic activities Failure = persistent pain, continuous analgesic use, unable to return to work	Hospital admission because of severity of sciatic pain	Not reported
3	Carragee [14]	Multivariate analysis (multiple logistic regression)	Composite measure of overall outcome comprising sum of scores on 0-10 scale for self-reported pain, medication use, activity restriction and satisfaction, total divided by 4 to give outcome score > 6 = good ≤6 = poor	Larger ratio of disc to remaining canal (in conservatively treated patients)	R = 0.50	Shorter duration of symptoms Absence of litigation Younger age	Not reported	Data from surgically and non-surgically treated patients analysed separately. Only data from conservatively treated patients presented
4	Hasenbring [15]	Multivariate regression analysis	Pain Intensity Self report 8 point scale	Lesser degree of disc displacement Scoliosis High score for non-verbal pain behaviour Low score for direct search for social support Tendency to ignore pain experience Poor ability to imagine coping with the pain Low social status	β = -0.32 β = 0.15 β = 0.31 β = -0.35 β = 0.29 β = -0.20 β = -0.17			Pain intensity was the only outcome studied. 73 (65.8%) underwent surgical treatment but the analysis adjusted for treatment which was not found to be a significant predictor in this study.
5	Jensen [16]	Multivariate analysis adjusted for age, sex and treatment	"Recovery" (composite score including pain on 11 point VRS & disability on RMDQ) Recovery defined as: Pain score < 1 & RMDQ ≤ 3			Broad based disc protrusion Disc extrusion Male gender Absence of canal stenosis (males only)	OR 13.6 (95% CI; 1.9, 95.4) OR 10.6 (95% CI; 1.9, 58.7) OR 2.6 (95% CI; 1.3, 5.0) OR 4.2 (95% CI; 1.2, 14.7)
6	Komori [17]	Non-parametric methods (not further specified)	Outcome defined according to residual self-reported symptoms and disability on 3 point scale (poor, fair, good)	Smaller herniated disc Greater symptom severity at initial assessment	Not reported			The findings of this study should be interpreted with caution due to poor methodological quality
7	Miranda [19]	Multivariate logistic regression	Outcome defined as persistence of pain based on self report of sciatic pain Persistence = sciatica pain on >30 days/year in 2 consecutive years (1994 & 1995) on modified NMQ)	Poor job satisfaction Ex-smoker Jogging	OR 2.8 (95% CI; 1.2,6.7) OR 2.3 (95% CI; 1.3,4.3) OR 3.9 ( 95% CI;1.4,10.7)			Diagnosis of sciatica based on self-reported symptoms only
8	Vroomen [18]	Multivariate logistic regression	Poor outcome defined as absence of any improvement at 3 months based on self-reported change in symptoms	Duration of pain > 30 days Positive SLR	OR 10 (95%CI;2.5,33.3)* OR 2.5 (95%CI;1.25,20)* * see footnote			Patients undergoing eventual surgery excluded from this analysis. Follow up period only 3 months.

* We have recalculated the odds ratios for poor outcome from the original report of the analysis of patients treated conservatively throughout p < 0.05