Predictors of Length of Stay and Discharge Disposition After Shoulder Arthroplasty

A Systematic Review

Jacob E. Berman, MBA; Ana Mata-Fink, MD; Hafiz F. Kassam, MD; Theodore A. Blaine, MD; David Kovacevic, MD

Disclosures

J Am Acad Orthop Surg. 2019;27(15):e696-e701. 

In This Article

Results

The combined MEDLINE, Scopus, Embase, and Cochrane Library database queries with the use of the search terms yielded 12,611 unique references through our search methods (Figure 1). After screening titles and abstracts, we retrieved 42 references for full-text review. Twenty-nine studies fulfilled our inclusion criteria after full-text review.[3,10–29,33–40] Seven of these studies were then excluded because of significant overlaps in their patient populations. Table 1 (Supplemental Digital Content 1, http://links.lww.com/JAAOS/A287) outlines the characteristics of the 22 studies that were used for this review.

Figure 1.

Flowchart describing the literature search.

Of the included studies, 19 were level III evidence,[3,11–14,16,17,19–21,34,35,38,40] and 3 were level IV evidence.[27,28,33] Fourteen of the studies used large national databases including the National Inpatient Sample,[12–14,17,19,22,23,25,26] the National Surgical Quality Improvement Program,[3,11] the Statewide Planning and Research Cooperative System,[16,35] and Premier.[27] Two studies used data from internal databases,[20,21] and six studies were retrospective studies,[24,28,33,34,38,40] one of which had prospectively collected data.

The average age at the time of surgery was 69.3 years, with 59% of surgeries performed on women. Twenty studies examined hospital length of stay after shoulder arthroplasty.[3,11–13,16,17,19,20,22–25,27,28,33–35,38,40] Twelve studies examined factors associated with discharge to a facility after shoulder arthroplasty.[13,14,19,21–26,33–35] Patient factors and procedure type responsible for extended length of stay and increased probability of discharge to a facility are summarized in Table 2 (Supplemental Digital Content 2, http://links.lww.com/JAAOS/A288) and Table 3 (Supplemental Digital Content 3, http://links.lww.com/JAAOS/A289), respectively.

Age

The effect of age on length of stay was reported in five studies.[3,12,16,17,24] All four studies showed that older age was significantly associated with extended length of stay. In two studies, ORs of 1.38 and 3.69 were found for patients older than 85 years.[16,17] Two additional studies reported ORs of 2.82 and 3.23 for age greater than 80.[3,24]

The effect of age on discharge disposition was reported in two studies.[24,26] Older age was significantly correlated with probability of being discharged to a facility. Compared with patients younger than 65 years, patients older than 65 years were more likely to be discharged to a facility after shoulder arthroplasty (OR, 2.1 for ages 65 to 74 years; OR, 5.8 for ages 75 to 84 years; OR, 17.9 for ages ≥85 years).[26]

Sex

Sex as a factor of length of stay was reported in three studies.[3,16,17] All three studies showed that female sex was associated with an extended length of stay after shoulder arthroplasty. In these studies, ORs ranged from 1.16 to 2.63.[3,16,17] In one study, which was excluded because of overlapping populations, length of stay for female patients was shown to be greater by 0.3 days for anatomic and reverse total shoulder arthroplasties (TSAs) and by 0.1 days for hemiarthroplasties.[39]

The effect of sex on discharge disposition was reported in one study. Female sex was correlated with an increased likelihood of discharge to a facility (OR, 2.8).[26]

Medical Comorbidities

The effect of obesity was reported in three studies.[3,17,21] Although the definition of obesity differed between the studies, body mass index ≥40 kg/m[2] was associated with an extended length of stay (OR, 1.97)[3] and discharge to a facility (OR, 8).[21]

The effect of diabetes on length of stay was assessed in two studies.[3,17] These studies showed that diabetes had a statistically significant effect on the probability of an extended length of stay (OR range, 0.96 to 1.91).[3,17] One study, excluded for population similarities, showed that insulin-dependent diabetes mellitus increased the probability of an extended length of stay compared with non–insulin-dependent diabetes mellitus.[36] One study reported that diabetes increased the likelihood of discharge to a facility.[26]

Chronic obstructive pulmonary disease (COPD) was investigated in two studies.[3,17] Both studies showed a significant correlation of COPD with the probability of extended length of stay (OR, 1.54 to 2.41). No data were collected regarding the effect of COPD on discharge disposition.

The effect of heart disease was examined in two studies.[3,17] Both studies reported a significant correlation between heart disease and probability of extended length of stay (OR, 1.89 to 8.22). No data were collected regarding the effect of heart disease on discharge disposition.

Kidney disease was studied in two studies.[3,17] Both studies demonstrated an association between renal insufficiency and extended length of stay (OR, 1.65 to 13.71). No data were collected regarding the effect of kidney disease on discharge disposition.

Individual studies examined the effect of malnutrition,[11] metabolic syndrome,[19] opioid misuse,[18] and alcohol disorder[22] on postoperative length of stay. These factors were associated with extended length of stay. Charlson Comorbidity Index score greater than one was also associated with an extended length of stay.[16] Two studies reported that opioid misuse and alcohol disorder were correlated with an increased likelihood of discharge to a facility.[18,22] One study reported that opioid use had no significant effect on length of stay or discharge disposition.[34]

Hospital Volume

The effect of hospital volume on length of stay was examined in two studies.[14,25] Increasing volume of shoulder arthroplasties performed was found to reduce the probability of extended length of stay. Hospitals performing more than five shoulder arthroplasties per year were less likely to have extended length of stay postoperatively (OR, 0.79 for hospital volume 5 to 9, OR, 0.67 for hospital volume 10 to 14, OR, 0.60 for hospital volume 15 to 24, OR, 0.45 for hospital volume ≥25).[25]

The effect of hospital volume on discharge disposition was evaluated in two studies.[14,25] Increased hospital volume (>5 shoulder arthroplasties per year) correlated with a reduced probability of discharge to a facility postoperatively. In one study, the ORs were nearly equivalent regardless of whether the patient underwent shoulder arthroplasty at a low-volume (5 to 14 procedures per year) or a high-volume (>15 procedures per year) hospital (5 to 9 procedures, OR, 0.84; 10 to 14 procedures, OR, 0.86; 15 to 24 procedures, OR, 0.84; ≥25 procedures, OR, 0.75).[25]

One study examined the effect of surgeon volume on length of stay and discharge disposition.[14] Increased surgeon volume was associated with a decreased length of stay and a decreased likelihood of discharge to a facility. Patients undergoing surgery from physicians performing between two and five shoulder arthroplasty procedures per year had lengths of stays on average 0.3 days longer for TSA and 0.5 days longer for hemiarthroplasty and nonroutine discharge disposition OR of 1.3 compared with physicians who performed five or more procedures per year.

Surgical Procedure and Anesthesia

The effect of surgical procedure (ie, reverse shoulder arthroplasty versus anatomic TSA) on length of stay was examined in two studies.[23,28] In one study, patients undergoing reverse shoulder arthroplasty was associated with an increased length of stay (OR, 1.9).[23] In the other study, no significant difference was found in hospital length of stay.[28] Two studies investigated the effect of the surgical procedure on discharge disposition.[23,26] Both studies reported an increased likelihood of discharge to a facility for patients undergoing reverse TSA (OR, 1.3 to 1.5).

The effect of peripheral nerve block on length of stay was examined in one study.[27] Peripheral nerve block was associated with a decreased length of stay (OR, 0.89). No data were collected regarding the effect of peripheral nerve block on discharge disposition.

The effect of general versus regional anesthesia on length of stay and discharge disposition was examined in one study.[35] General anesthesia was correlated with a shorter length of stay compared with regional anesthesia with averages of 2.0 and 2.3 days, respectively (P < 0.001). In this study, the likelihood of discharge to a facility with the use of regional anesthesia versus general anesthesia decreased 6.7% and 10.2%, respectively (P = 0.008).

Continuous interscalene brachial plexus blockade was evaluated in two studies.[33,40] Findings were disparate because in one study, this was significantly correlated with a decreased length of stay (0.3 days fewer),[33] whereas in another study,[40] this was associated with increased length of stay (0.09 days greater). Both studies were statistically significant with P values of <0.001 and 0.002, respectively.

Local injection of liposomal bupivacaine combined with intravenous dexamethasone was associated with reduced length of stay in one study.[38] The authors found a decrease from two days to one day in patients treated with the cocktail.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....