Table 4.

Instances of non-adherence to specific AHRQ methodological practices in cardiac literature using the national inpatient sample.

Methodological practice	Overall No./Total (%)	Impact Factor < 10 No./Total (%)	Impact factor ≥ 10 No./Total (%)	P -value*
1: Recorded hospital events as patients	170/445 (38.2)	167/421 (39.7)	3/24 (12.5)	0.056
2: Performed state-level analyses	10/445 (2.2)	8/421 (1.9)	2/24 (8.3)	0.672
3: Performed hospital-level analyses after 2011	9/58 (15.5)†	9/51 (17.6)	0/7 (0)	1.0
4: Performed physician-level analyses	12/445 (2.7)	9/421 (2.1)	3/24 (12.5)	0.154
5: Employed non-specific secondary diagnosis codes	159/445 (35.7)	155/421 (36.8)	4/24 (16.7)	0.350
6: Used statistical methods that did not account for the complex survey structure of the NIS	250/445 (56.2)	247/421 (58.7)	3/24 (12.5)	0.0005
7: Did not adjust for major transition periods (1997–1998, 2011–2012) for trend analyses	102/122 (83.6)‡	94/114 (79.8)	8/8 (100)	1.0

^*P value computed via Fisher exact test with Bonferroni correction.

^†58/445 studies performed hospital-level analyses.

^‡122/445 studies performed trend analysis involving major transition periods in the NIS (1997–1998 or 2011–2012).