Readmissions and mortality after outpatient vs inpatient unicompartmental knee arthroplasty in Denmark A propensity score matched study of 5,384 procedures

Background: Limited nationwide data on the development of outpatient unicompartmental knee arthroplasty (UKA) practice and patient safety exist. The primary objective of this study on patients receiving a medial or lateral UKA was to investigate 7, 30-and 90-day readmission risk and 90-day mortality in outpatient vs inpatient surgeries. Secondary to investigate the nationwide development of outpatient UKA surgery in 2014–2018. Methods: Included patients received a medial or lateral UKA in the period January 1, 2014 to December 31, 2018 in any Danish hospital. Data were collected from the Danish National Patient Register. The cohort consisted of 1,059 outpatient and 4,325 inpatient surgeries, hereof 5,182 medial and 202 lateral UKA. After propensity score matching (1:1) 1,057 patients were included in each group. Results: We found a 7-day readmission risk of 1.5 % vs 1.4 % (p = 0.8), 30-day readmission risk of 2.6 % vs 3.2 % (p = 0.3), and 90-day readmission risk of 4.2 % vs 4.8 % (p = 0.4) after outpatient vs inpatient UKA. Similar results were found after matching. We found no signiﬁcant differences in 90-day mortality for the unmatched or matched cohorts. The amount of outpatient UKA surgeries in Denmark increased from 86 in 2014 to 214 in 2018. Conclusion: Outpatient medial or lateral UKA seem to be as safe as inpatient UKA on a nationwide basis. (cid:1) 2022 The Authors. Published by Elsevier B


Contents lists available at ScienceDirect
The Knee j o u r n a l h o m e p a g e : and studies report no increased morbidity or mortality [4,5].However, fast-track programs require a multidisciplinary effort and strict discharge criteria to be followed [6].
Outpatient unicompartmental knee arthroplasty (UKA) surgery protocols have been implemented at several Danish hospitals as a part of well-established fast-track programs [3].Concerns about the safety of outpatient surgery has been raised, after studies based on the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database have shown higher odds of complications after outpatient total hip arthroplasty (THA) and total knee arthroplasty (TKA) compared to inpatient THA and TKA [7][8][9][10].Similar readmission and complication risks after outpatient vs inpatient THA and TKA have been found in a Danish study [11].The safety of outpatient UKA remains unknown.Therefore, the primary objective of this study on patients receiving a medial or lateral UKA was to investigate 7-, 30-and 90-day readmission risk and 90-day mortality in outpatient vs inpatient surgeries.Secondary to investigate the nationwide development of outpatient UKA surgery in 2014-2018.

Patients and methods
This was a retrospective cohort study of nationwide collected data.The RECORD guidelines for the reporting of routinely collected, observational data were followed [12].

Data sources
We retrieved information on surgical procedure codes, somatic diagnosis codes, age and sex from the Danish National Patient Register (DNPR).The DNPR is a national administrative register collecting information on all public and private hospital contacts in Denmark since 1977.Danish hospitals receive reimbursement from the health authorities when reporting to the DNPR ensuring a completeness of > 99 % [13].We used the Charlson Comorbidity Index (CCI) to estimate the burden of comorbidities for each patient [14,15].The CCI was calculated from ICD-10 codes for the 10-year period up to the date of surgery [14,15].The burden of comorbidities was classified into three levels: CCI of 0 (low), CCI of 1-2 (medium), and CCI of 3 or more (high).We collected information on cohabitation and death from the Danish Civil Registration System (CPR), which is a national register containing basic personal information on all who have a civil registration number.Cohabitation was categorized binary as patients being married or cohabiting vs unmarried, widowed or divorced.

Patients
All included patients received a medial or lateral UKA in the period January 1, 2014 to December 31, 2018.We identified 6,162 patients in the DNPR from the procedure codes KNGB01 (uncemented medial UKA), KNGB02 (uncemented lateral UKA), KNGB11 (cemented medial UKA) and KNGB12 (cemented lateral UKA) registered according to the Nordic Medico-Statistical Committee (NOMESCO) classification of surgical procedures (Figure 1) [16].Outpatient surgery did not include an overnight stay and inpatient surgery included length of stay (LOS) of 1-3 days postoperatively.Patients were excluded if LOS was more than 3 days, because prolonged stay indicated unusual circumstances making the patients unsuited for inclusion.The later surgery was excluded for patients receiving bilateral UKA.

Outcome measures
We obtained ICD-10 diagnosis codes for readmissions from the DNPR within 7, 30 and 90 days after discharge.A readmission was defined as a new admission with overnight stay in hospital.The diagnosis codes listed in Appendix A were considered potentially related to the index surgery.Obviously unrelated readmissions (e.g.G44.8:Other specified headache syndrome; D12.3: Benign neoplasm of transverse colon) listed in Appendix B were excluded.Only first-time readmissions after surgery were included.

Statistics
Data were presented as counts and risk estimated as proportions.Continuous data were assessed using Students t-test and nominal variables were assessed using Pearson chi-squared test.Continuous data were inspected for normal distribution with Q-Q plots.We used propensity score matching to control for potential confounders.We estimated the propensity score with a multivariable logistic regression model for the 5,384 patients eligible for this study.The groups were matched according to patients undergoing outpatient or inpatient surgery.We used 1:1 nearest neighbor matching with a caliper of 0.2, discarding units out of common support and sampling without replacement [17].Age, sex, cohabitation, CCI and type of UKA were set as matching variables.We investigated the matched and unmatched groups for imbalance of the matching variables with standardized mean differences (SMD) (Table 1) [18].A SMD > 0.1 indicated imbalance.The outcome of the propensity score model was average treatment effect of the treated (ATT) and statistical significance was reported with p-values.Statistical significance was set at the 5 % level.For all analyses, we used Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC.

Ethics
The study was conducted in accordance with the Declaration of Helsinki and was approved by the Danish Data Protection Agency (Journal no.19/14416).Ethical approval was not needed as the study was non-interventional.Declarations of interest: none.

Results
We identified 5,384 patients receiving a medial or lateral UKA, 5,182 and 202 respectively; 1,059 outpatient UKA and 4,325 inpatient UKA surgeries with LOS of 1-3 days, mean LOS 1.4 (CI 1.4-1.5)(Figure 1).After propensity score matching 1,057 patients were included in each group and no indications of imbalance were present between groups (Table 1).
Risk of readmission were similar for in-and outpatient UKA after 7, 30 or 90 days for both the unmatched and the matched cohorts (Table 2).Mortality risk within 90 days were similar between groups as well (Table 2).
The nationwide cohort included 27 centers, 19 of which performed both in-and outpatient UKA.Overall 19 % of UKA's performed in 2014-2018 in Denmark were outpatient surgeries, with an increase during the study period.86 outpatient UKA surgeries were performed in 2014 and 214 in 2018 (Figures 2 and 3).Outpatient UKA peaked in 2016 with 299 surgeries.

Discussion
Risk of readmission within 7, 30 and 90 days of surgery and 90-day mortality were similar between in-and outpatient UKA performed from 2014 to 2018 for both the matched and unmatched cohorts.An increase in outpatient UKA surgery was observed in the study period.

Readmission risk and mortality
Our findings support those of previous studies considering readmission risks for both in-and outpatient UKA.They found 30-day readmission risks of 0-4 % and 90-day readmission risks of 2-7 % with no differences between in-and outpatient UKA [19][20][21].Our findings of low readmission risks are probably the result of increased use of well-established fast-track programs nationwide.90-readmission risks of 7-9 % after fast-track TKA has been reported [22,23].TKA is associated with higher risk of readmission than UKA, but a recent study has found similar risks for surgeries performed at fast-track centers in Denmark [23].A study considering the safety of outpatient surgery will also be relevant for THA and TKA, for which outpatient surgery can be beneficial as well as for UKA.
Our results were also consistent with previously reported 90-day mortality risks of 0-0.08 % after UKA surgery [20,24].The Propensity Score Model using ATT (Average treatment effect of the treated) in the matched and unmatched cohort estimates P-values.

Development of outpatient UKA
The amount of outpatient UKA surgeries doubled from 2014 to 2018, with a peak in 2016.Few large volume centers contributed with the majority of outpatient UKA.Most centers increased their proportion of outpatient vs inpatient surgeries in the study period.The variation may be explained by the fact that outpatient programs from each center varies, as outpatient surgery requires a well-established protocol to be feasible [20,25].However, we did not investigate this aspect in this study as the DNPR does not supply information on the perioperative setup for the individual surgical center.

Strengths and limitations
A strength of this study is the nationwide coverage with > 99 % completeness of data ensuring a largest study cohort possible for the included procedure codes and complete follow-up [13].Propensity score matching has limited the bias of important confounders.The selection process of patients eligible for outpatient surgery in the different centers may vary, but no data on this was available in the DNPR.Information on diagnosis codes were collected from the DNPR and were not verified from hospital charts and this is a limitation of our study.However, no obvious differences in registrations of readmissions between in-and outpatient procedures should bias the results of this study.

Conclusion
Outpatient medial or lateral UKA seem to be as safe as inpatient UKA on a nationwide basis.

Funding
The study was supported by the Research Council at Lillebaelt Hospital.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Figure 2 .
Figure 2. Outpatient surgeries in Denmark in 2014-2018.The figure presents the number of outpatient unicompartmental knee arthroplasty (UKA) surgeries from the eight hospitals in Denmark performing >10 outpatient surgeries in 2014-2018 and the total number of outpatient UKA surgeries performed in Denmark in the same period.

Figure 3 .
Figure 3. Frequency of outpatient surgeries in Denmark in 2014-2018.The figure presents the frequency of outpatient unicompartmental knee arthroplasty (UKA) surgeries of the total number of UKA surgeries from the eight hospitals in Denmark performing >10 outpatient surgeries in 2014-2018 and the total frequency of outpatient UKA surgeries of the total number of UKA surgeries performed in Denmark.

Table 1
Demographic matching characteristics.
SMD, standardized mean difference, displayed before and after propensity score matching.SMD > 0.1 indicates imbalance between matching variable.UKA = unicompartmental knee arthroplasty.

Table 2
Outcomes in unmatched and propensity score matched cohorts.