Is there variation in utilization of preoperative tests among patients undergoing total hip and knee replacement in the US, and does it affect outcomes? A population-based analysis

Abstract

Study objective To describe recent practice patterns of preoperative tests and to examine their association with 90-day all-cause readmissions and length of stay. Design Retrospective cohort study using the New York Statewide Planning and Research Cooperative System (SPARCS). Setting SPARCS from March 1, 2016, to July 1, 2017. Participants Adults undergoing Total Hip Replacement (THR) or Total Knee Replacement (TKR) had a preoperative screening outpatient visit within two months before their surgery. Interventions Electrocardiogram (EKG), chest X-ray, and seven preoperative laboratory tests (RBCs antibody screen, Prothrombin time (PT) and Thromboplastin time, Metabolic Panel, Complete Blood Count (CBC), Methicillin Resistance Staphylococcus Aureus (MRSA) Nasal DNA probe, Urinalysis, Urine culture) were identified. Primary and secondary outcome measures Regression analyses were utilized to determine the association between each preoperative test and two postoperative outcomes (90-day all-cause readmission and length of stay). Regression models adjusted for hospital-level random effects, patient demographics, insurance, hospital TKR, THR surgical volume, and comorbidities. Sensitivity analysis was conducted using the subset of patients with no comorbidities. Results Fifty-five thousand ninety-nine patients (60% Female, mean age 66.1+/− 9.8 SD) were included. The most common tests were metabolic panel (74.5%), CBC (66.8%), and RBC antibody screen (58.8%). The least common tests were MRSA Nasal DNA probe (13.0%), EKG (11.7%), urine culture (10.7%), and chest X-ray (7.9%). Carrying out MRSA testing, urine culture, and EKG was associated with a lower likelihood of 90-day all-cause readmissions. The length of hospital stay was not associated with carrying out any preoperative tests. Results were similar in the subset with no comorbidities. Conclusions Wide variation exists in preoperative tests before THR and TKR. We identified three preoperative tests that may play a role in reducing readmissions. Further investigation is needed to evaluate these findings using more granular clinical data.

Patient, provider, and system‐level factors associated with preoperative cardiac testing: A systematic review

“…Fourteen of the 28 studies had a moderate risk of bias 22,24,25,27–29,31–34,36,41,43,49 . Nine had a low risk of bias, 30,38–40,42,44–47 and five had a high risk of bias (Table 1; Supporting Information: Tables ) 23,26,35,37,48 …”

“…24,30,34,38,40,43,49 Most of the included studies were observational using either cross-sectional 22,[24][25][26]28,34,40,41,45,46,48 or retrospective cohort designs (Table 1). 27,30,32,33,37,38,[42][43][44]47,49 Of those with an intervention, 23,29,31,35,36,39 none was a randomized controlled trials although one was a nonrandomized controlled trial (Supporting Information: Table 8). 36 Five other studies used pre-post designs 23,29,31,35,39 of which all were prospective, interventional studies except one.…”

“…22,28,29,37,48 Fewer than half used a nationally representative patient sample. 25,27,28,30,33,42,[44][45][46][47]49 The preoperative study of interest varied: nine studies focused on stress tests, 24,31,33,40,41,43,45,46,49 seven on EKGs 23,30,32,35,36,39,42 (two of which 30,42 also studied cascades of care after an EKG), three on echocardiograms, 22,29,37 one on cardiac catheterization, 38 and the remaining eight on a combination of tests. [25][26][27][28]34,44,47,48…”

“…Fourteen of the 28 studies had a moderate risk of bias. 22,24,25,[27][28][29][31][32][33][34]36,41,43,49 Nine had a low risk of bias, 30,[38][39][40]42,[44][45][46][47] and five had a high risk of bias (Table 1; Supporting Information: Tables 1-4). 23,26,35,37,48 Of the studies deemed to have a low-risk of bias, most 30,42,44,46,47 were cohort studies that used a nationally representative data set with well-designed adjustment for patients' clinical characteristics.…”

Construction and validation of a predictive model for lower extremity deep vein thrombosis after total knee arthroplasty

“…In recent years, total knee arthroplasty has been widely used in clinical practice, and the factors associated with postoperative lower extremity DVT and its impact on clinical prognosis have received increasing attention. [13,14] Several previous studies have explored the risk factors for postoperative DVT after total knee arthroplasty, but the conclusions varied between studies, and data on relevant predictive models are still lacking. [15][16][17] In the present study, we used predictors that are common and easily recognized in clinical practice, and developed this nomogram prediction model based on the 5 independent risk factors identified, and after model validation, it was determined that the model established in this study had good predictive ability.…”