Table III.

Risk of bias assessment.

Authors	Study design	Level of evidence	Selection				Comparability	Outcome		Quality	Comment
			Representativeness of the sample (+)	Sample size (+)	Nonrespondents (+)	Ascertainment of exposure (++)	Based on design and analysis (++)	Assessment of outcome (++)	Statistical test (+)
Rogers et al44	Cross-sectional	Level 4	+		+	+		++	+	5/10 Satisfactory	Selection Bias: metropolitan, Medicaid expansion state
Li et al30	Cross-sectional	Level 4	+	+	+	++		++	+	8/10 Good	Large sample size
Patterson et al39	Cross-sectional	Level 4	+		+	++		++	+	6/10 Satisfactory	Selection bias: North Carolina
Curry et al12	Cross sectional	Level 4	+	+		+		++	+	6/10 Satisfactory	Selection bias: urban populations

Authors	Study design	Level of evidence	Selection				Comparability	Outcome			Quality	Comment
			Representativeness of the exposed cohort (+)	Selection of the nonexposed cohort (+)	Ascertainment of exposure (+)	Outcome not present before study (+)	Based on design and analysis (++)	Assessment of outcome (+)	Follow-up length (+)	Follow-up cohort (+)
Sabesan et al48	Retrospective cohort	Level 3	+	+	+	+	+		+	+	7/9 Good	Small cohort
Jain et al24	Cohort	Level 2	+	+		+	+		∗		5/9 Satisfactory	30% dropout rate
Chapman et al8	Retrospective cohort	Level 3	+	+	+	+	++	+	+	+	9/9 Excellent	Selection bias: narrow definition for RCT, Medicare only

RCT, rotator cuff tear.

Top: Newcastle Ottawa Scale adjusted for cross-sectional studies²³; excellent = 9-10, good = 7-8, satisfactory = 5-6, unacceptable = 1-4. Bottom: Newcastle Ottawa Scale for cohort studies⁶¹; excellent = 8-9, good = 6-7, satisfactory = 4-5, unacceptable = 1-3.