Medicare beneficiaries’ demand for total hip arthroplasty (THA) and total knee arthroplasty (TKA) has increased significantly over the past several years, with recent studies reporting 209,945 primary THAs and 243,802 primary TKAs performed annually.1,2 With this demand has come an increase in the percentage of patients discharged to an extended-care facility (ECF) for skilled nursing care or acute rehabilitation—an estimated 49.3% for THA and 41.5% for TKA.1,2 To qualify for discharge to an ECF, Medicare beneficiaries are required to have an inpatient stay of at least 3 consecutive days.3 Although the basis of this rule is unclear, it is thought to prevent hasty discharge of unstable patients.
We conducted a study to explore the effect of this policy on length of stay (LOS) in a population of patients who underwent primary total joint arthroplasty (TJA). Based on a pilot study by our group, we hypothesized that such a statuary requirement would be associated with increased LOS and would not prevent discharge of potentially unstable patients. Specifically, we explored whether patients who could have been discharged earlier experienced any later inpatient complications or 30-day readmission to justify staying past their discharge readiness.
Materials and Methods
Institutional review board approval was obtained for this study. Between 2011 and 2012, the senior authors (Dr. Wellman, Dr. Attarian, Dr. Bolognesi) treated 985 patients with Current Procedural Terminology (CPT) codes 27130 (THA) and 27447 (TKA). Of the 985 patients, 287 (29.13%) were discharged to an ECF and were included in the study. Three of the 287 were excluded: 2 for requiring preadmission for medical optimization and 1 for having another procedure with plastic surgery. All patients were admitted from home on day of surgery and had a standardized clinical pathway with respect to pain control, mobilization, and anticoagulation. Physical therapy and occupational therapy (PT/OT) were initiated on day of surgery and were continued daily until discharge.
The primary outcome was discharge readiness, defined as meeting the criteria of stable blood pressure, pulse, and breathing; no fever over 101.5°F for 24 hours before discharge; wound healing with no concerns; pain controlled with oral medications; and ambulation or the potential for rehabilitation at the receiving facility. Secondary outcomes were changes in PT/OT progress, medical interventions, and 30-day readmission rate. PT/OT progress was categorized as either slow or steady by the therapist assigned to each patient at time of hospitalization. Steady progress indicated overall improvement on several measures, including transfers, ambulation distance, and ability to adhere to postoperative precautions; slow progress indicated no improvement on these measures.
Results for continuous variables were summarized with means, standard deviations, and ranges, and results for categorical variables were summarized with counts and percentages. Student t test was used to evaluate increase in LOS, and the McNemar test for paired data was used to analyze rehabilitation gains from readiness-for-discharge day to the next postoperative day (POD). SAS Version 9.2 software (SAS Institute) was used for all analyses.
Results
Of the 284 patients included in the study, 203 were female (71.5%), 81 male (28.5%). Mean (SD) age was 68 (11) years (range, 21-92 years). One hundred seventy-nine patients (63.0%) underwent TKA, and 105 (37.0%) underwent THA. Two hundred twenty-seven patients (80.0%) were discharged to skilled nursing care, and 57 (20.1%) to inpatient rehabilitation. Mean (SD) LOS was 3.44 (0.92) days (range, 3-9 days). One hundred eighty-three patients (64.4%) were ready for discharge on POD 2, 76 (26.8%) on POD 3, and 25 (8.8%) after POD 3. Delaying discharge until POD 3 increased LOS by 1.08 days (P < .001). Two hundred nine patients (73.6%) were discharged on POD 3, and 75 (26.4%) after POD 3. Reasons for being discharged after POD 3 were lack of ECF bed availability (48 patients, 64.0%) and postoperative complications (27 patients, 36.0%). Patients ready for discharge on POD 2 had fewer complications than patients ready after POD 2 (P < .001).
Analysis of the 183 patients who were ready for discharge on POD 2 demonstrated a statistically significant (P = .038) change in rehabilitation progress by staying an additional hospital day. However, this difference was not clinically significant: Only 17.5% of patients improved, while 82.5% remained unchanged or declined in progress. Most important, among patients who demonstrated rehabilitation gains, the improvement was not sufficient to change the decision regarding discharge destination. Three patients (1.6%) ready for discharge on POD 2 were readmitted within 30 days of discharge (2 for wound infection, 1 for syncope). Risk for 30-day readmission or development of an inpatient complication in patients ready for discharge on POD 2 was not significant (P = .073). Table 1 summarizes the statistical results.
