Table 3.

Quality assessment of included cohort studies

Study (author, country, year)	Type of study	Base population	Confounding	Losses to follow-up	Analysis
Sowell, USA, 199924	Cohort	All Medicare population at risk for lower extremity amputation in 1993–1994	Not addressed—only looked at 1 variable—acknowledged as a limitation	No losses to follow-up	Amputation incidence rates with and without exposure to podiatry
Lipscombe, Canada, 200337	Cohort	Patients in Peritoneal Dialysis program at University Health Network, between January 1997 and December 1999	Data on confounding variables collected	No losses to follow-up	Descriptive stats
Lavery, USA, 201021	Cohort	Patients with diabetes attending Scott and White Health Plan, Texas, USA	Data on confounding variables collected	150 consecutive patients with at least 30 months follow-up from the time of diagnosis recruited so no losses to follow-up	Descriptive stats
Sloan, UK, 201038	Cohort	All individuals with a DM-related LEC diagnosis between 1994 and 2001	Data on confounding variables collected	No losses to follow-up	HRs adjusted for Medicare expenditures from care received from non-study health professionals