Determining the magnitude of surveillance bias in the assessment of lower extremity deep venous thrombosis

Abstract

Prognostic/epidemiologic study, level III; therapeutic study, level III.

Does Surveillance Bias Influence the Validity of Measures of Inpatient Complications? A Systematic Review

“…Ten of the articles dealt with VTE; 7 dealt with a trauma population, 6 of which used either local or national trauma registry data; the other used chart review data. 9,11,14-18 Three studies specifically examined postoperative VTE, one of which used the actual PSI definition, based on hospital claims data 8 ; the other 2 used chart-abstracted data, one from the Veterans Affairs Surgical Quality Improvement Program (VASQIP), 19 the other from the American College of Surgeons National Surgical Quality Improvement Project (ACS-NSQIP). 20 The only non-VTE study dealt with CLABSI rates and used survey data.…”

Early Detection of Deep Venous Thrombosis in Trauma Patients

“…Prior investigations identified major head injury, spinal cord injury, long-bone and pelvic fractures, hypotension, advanced age, the transfusion of blood products, and an injury severity score greater than 15 all as risk factors [ 2 , 11 , 12 , 18 ]. Supporters of surveillance VDU in high-risk trauma patients argue earlier diagnosis of DVT results in earlier treatment, possibly preventing propagation of the DVT and development of a pulmonary embolism [ 10 , 13 , 16 ]. However, other studies show screening for DVT in trauma patients to have little effect in preventing pulmonary emboli and suggest that surveillance studies are not cost-effective [ 19 , 20 ].…”

“…Factors found to place trauma patients at moderate to high-risk for VTE have historically included the following: blunt mechanism of injury, spinal cord injury, long-bone or pelvic fractures, major head injury, severe vascular endothelial damage, the need for transfusion of blood products, advanced age, an injury severity score (ISS) greater than 15 and hypotension [ 2 , 8 , 11 , 12 ]. While screening VDU imaging of high-risk trauma patients remains a controversial practice, employing this modality naturally leads to an increase in the number of DVTs diagnosed [ 13 - 15 ].…”

Patient Co-Morbidity and Functional Status Influence the Occurrence of Hospital Acquired Conditions More Strongly than Hospital Factors

“…19 In addition, several studies documented inherent bias in the CMS Hospital-Acquired Condition Reduction Program (HACRP), reporting different HACRP scores based solely on hospital size or use of surveillance (e.g.,DVT surveillance), in situations with identical complication rates. 20,21 Finally, and ironically, Rajaram et al reported that the CMS non-payment policy, as currently designed, actually penalized the highest-performing hospitals; those that were accredited by the Joint Commission, offered advanced services, were major teaching institutions, and had better performance on process and outcome measures. 22 These authors recommended reevaluation and reform of the CMS policy.…”

“…Shackford et al compared VTE incidence between trauma medical centers that did or did not use routine duplex DVT screening and noted that while the incidence of PE was identical (0.4%) at both centers, the odds of a DVT diagnosis were 5.3 times higher in the center using routine DVT surveillance. 21 Several risk factors for HAC were particularly important. The risk of HAC in patients with major or extreme loss of function before surgery was greater than six times compared to patients with minor or moderate loss of function before surgery.…”