Early mobilization within the first 12 hours (day zero) of total joint arthroplasty (TJA) has been shown to reduce length of stay (LoS) without risking clinical outcomes, patient safety or satisfaction. The purpose of this study was to investigate associations between the degree of mobilization on day zero (i.e., standing at the bedside versus walking in the hallway) and LoS in TJA patients. In addition, we investigated predictors of LoS and day zero mobilization.

A retrospective cohort study was undertaken of the health records of patients in a community hospital setting who had an elective unilateral primary TJA between June 2015 and May 2017 and had mobilized on day zero.

The total sample was 283 patients (184 TKA and 99 THA) across four mobilization categories: Sat on beside (n = 76), Stood by bed/marched in place (n = 83), Walked in the room (n = 79), and Walked in hall (n = 45). Analysis of variance found no significant group differences in age, ASA score, Charlson Comorbidity Index score, anesthesia, surgeon, procedure type, pain medication, and patient reported symptoms recorded by physiotherapists. Significantly more women were in the Sat group and significantly more men were in the Hall group (p < .001). Patient reported symptoms of nausea and drowsiness were significantly greater for the Sat group (p < .001). LoS was also significantly different across the groups. Post hoc Tukey comparisons found the Walked Hall group had significantly shorter LoS (M = 2.7 days) than the Sat group (M = 3.9, p < .001), Stood group (M = 3.4, p = .011), and the Walked Room group (M = 3.5, p = .004).

A hierarchical regression was performed to determine predictors of LoS. Block 1 consisted of demographic, medical status, and patient reported symptoms as variables. Mobilization was entered in Block 2. The first model was significant (p < .001) and explained 24% of variance in LoS. The final model was also significant (p < .001), accounting for a total of 26% of the variance in LoS. Thus, block 2 (i.e., mobilization) accounted for a small but significant 2% incremental variance (p = .008) beyond the block 1 variables in the prediction of LoS. With mobilization added, only male gender (p = .002), lower BMI (p = .026), and lower ASA scores (p = .006) remained significant predictors of shorter LoS, and the predictive ability of several of the block 1 variables were reduced to non-significant levels. A simultaneous regression model was then used to predict degree of mobilization. The model accounted for 24% of the variance in mobilization (p < .001). Variables significantly associated with a greater degree of mobilization included: younger age, male gender, lower BMI, and fewer symptoms, namely nausea, numbness, lightheadedness, and drowsiness.

This study found length of stay was shorter when patients mobilized farther on the day of surgery. Some factors predictive of mobilization may be modifiable. Focusing on symptom management could increase opportunities for farther mobilization on the day of surgery, and thus decrease length of stay.

Introduction: The ten-year survivorship of Oxford Unicompartmental Knee Arthroplasty (OUKA) has ranged from 98% in the hands of the developers to only 82–90% in reports from independent centers and national registries. This study was performed to investigate the effects of surgeon training and correct patient selection on the expected outcome of this procedure.

Methods: We created a computer-simulated joint registry consisting of 20 surgeons who performed OUKA on 1,000 patients. Mathematical models of the patient and surgeon populations and corresponding hazard functions were formulated using data from the Swedish and Australian joint registries. The long-term survivorship of UKA was assumed to average 94% at 10 years and was modeled as the product of hazard functions quantifying risk factors under the surgeon’s control, risk factors presented by the patient, and the inherent revision risk of the procedure. We performed four simulations looking at the effect of surgeon training by pairing surgeons and patients based on surgeon experience and patient risk factors.

Results: When experienced surgeons (> 40 cases) performed OUKA on low risk patients (bottom quintile), the revision rate dropped from 6.0% to 4.5%. The same surgeons had a revision rate of 7.5% when assigned to the highest risk patient group (top quintile). Conversely, when the least experienced surgeons (< 10 cases) selected the least fit patients, the revision rate increased from 6% to 8.25%. However, when these surgeons were assigned to the lowest risk group, only 5.25% of patients were revised. Taken simultaneously, these results indicate that the overall revision rate of this procedure can vary between 4.5% to 8.25%, depending upon the experience of the surgeon and the patients selected.

Conclusions:

Mathematical models of patients and surgeons can be built using joint registry data. These models can then be used in a computer simulation yielding results comparable to what has been reported in the literature.