Comparison of patient-demographics, causes, and costs of 90-day readmissions following primary total shoulder arthroplasty for glenohumeral osteoarthritis

Abstract

Introduction

Readmissions following orthopaedic surgery are associated with worse outcomes and increased healthcare costs. Studies investigating trends, causes, and costs of readmissions following primary total shoulder arthroplasty (TSA) for the indication of glenohumeral osteoarthritis (OA) are limited. The objective was to compare: 1) patient-demographics of those readmitted and not readmitted within 90-days following primary TSA for OA; 2) causes of readmissions and 3) associated costs.

Methods

A retrospective query from 2005 to 2014 was performed using a nationwide administrative claims database. The study group consisted of patients readmitted within 90-days following primary TSA for glenohumeral OA, whereas patients not readmitted served as controls. Causes of readmission were stratified into the following groups: cardiovascular, hematological, endocrine, gastrointestinal (GI), musculoskeletal (MSK), neoplastic, neurological, pulmonary, infectious, renal, and miscellaneous causes. Patient demographics were compared, in addition to the frequency of the causes of readmissions, and their associated costs. Chi-square analyses compared demographics between groups. Analysis of variance was utilized to determine differences in 90-day costs for the causes of readmission. A p-value less than 0.001 was significant.

Results

The overall 90-day readmission rate was 2.4% (3432/143,878). Patients readmitted following primary TSA were more likely to be over the age of 75, female, and higher prevalence of comorbid conditions, including psychiatric and medical conditions. Readmitted patients had a higher overall comorbidity burden per mean Elixhauser-Comorbidity Index (ECI) scores (10 vs. 7,p < .0001). The leading cause of readmissions were due to MSK (17.34%), cardiac (16.28%), infectious (16.26%), and gastrointestinal (11.64%) etiologies. There were differences in the mean 90-day costs of care for the various causes of readmissions, with the leading cost of readmissions being cardiac causes ($10,913.70) and MSK ($10,590.50) etiologies.

Conclusion

Patients with greater comorbidities experienced increased incidence of readmission following TSA for glenohumeral OA. Cardiac and MSK etiologies were the primary cause of readmissions.

Level of evidence

III.

1. Introduction

Hospital readmissions are an undesired occurrence that not only present issues for the patient, but also hold a costly burden for the hospital. Studies have demonstrated that the estimated annual cost of hospital readmissions is close to $20 billion, therefore, making efforts in reducing these costs essential to healthcare systems.¹ With this concept in mind, it is important to look at the most common procedures with limited research and shoulder surgery is one of these as it has been on the rise over the last decade, becoming the third most common joint procedure behind knee and hip surgery.² Between 2011 and 2017, the number of primary shoulder arthroplasties more than doubled, from 51,329 to 104,575, and continues to increase each year.³ Osteoarthritis (OA) of the glenohumeral (GH) joint is one of the most common causes of shoulder pain and although there are several non-surgical options for treating OA of the GH joint, the evidence has shown that total shoulder arthroplasty (TSA) is a highly viable option, providing significant improvements in pain, global health assessment, function, and quality of life.4, 5, 6, 7, 8, 9, 10, 11, 12

As the number of TSA's for the treatment of glenohumeral OA continues to rise each year, more research is necessary to look into who makes up this population and the reasons for readmission. OA of the glenohumeral joint is a common disease among older adults, as the incidence increases over the age of 50 and is more prevalent in women.^13,14 In one study, the average age of patients undergoing TSA for OA was over 65, which presented with a higher number of comorbidities and risk factors, such as hypertension, diabetes, obesity, heart disease, and lung disease.¹⁵ Although the aforementioned study did take inpatient complications into account, they did not look into readmission rates and costs associated with them. Other studies have also found links to poorer TSA post-operative outcomes with diabetes and psychiatric comorbidities,^16,17 while others have found no association between BMI and shoulder dislocation 30 days post-op.¹⁸ However, these studies did not consider the indication for the surgery into account, as fractures and trauma present different complications and costs post-operatively, than OA. Fractures are associated with a longer length of hospital stay, increased medical complications, bleeding transfusions, and 30 day readmission.¹⁹ Furthermore, the center of Medicare services does not distinguish between the type of arthroplasty performed (TSA vs reverse total shoulder arthroplasty) or the reason for the surgery (OA vs fracture), making research that differentiates both procedures and causes instrumental.²⁰

The purpose of this study was to closely investigate the ninety-day readmission of patients following a total shoulder arthroplasty for glenohumeral osteoarthritis. This was done by analyzing a nationwide comprehensive database and determining: 1) Patient-demographics of patients readmitted and not readmitted within 90-days following TSA for OA; 2) Causes of readmissions; 3) the associated costs.

2. Methods

2.1. Database

A retrospective query from January 5, 2005 to March 31st, 2014 was performed using a nationwide administrative claims database using the PearlDiver (PearlDiver Technologies, Fort Wayne, Indiana, United States) program. The PearlDiver database is a subscription based database which houses information from either the Parts A and B Medicare claims or from a private insurance database known as Mariner. For this study, the Medicare Claims data was extracted by using International Classification of Disease, Ninth Revision (ICD-9), ICD-10, and Current Procedural Terminology (CPT) codes. As the downloaded data is devoid from any patient identifying information, the current study was deemed exempt from our institution's Institutional Review Board (IRB) process.

2.2. Cohorts

The database was initially queried for all patients undergoing primary TSA using CPT code 23472 (total arthroplasty of glenohumeral joint with glenoid and proximal humeral replacement). Patients who had glenohumeral OA were subsequently identified using ICD-9 diagnoses code: 715.11 (osteoarthrosis, localized, primary, shoulder region). The inclusion criteria for the study group consisted of all patients in the database undergoing primary TSA exclusively for glenohumeral OA and readmitted within 90-days following the index procedure; whereas patients not readmitted served as the control cohort. Causes of readmission were manually stratified into the following primary etiologies of readmissions: cardiovascular, hematological, endocrine, gastrointestinal (GI), musculoskeletal (MSK), neoplastic, neurological, pulmonary, infectious, renal, and miscellaneous causes.

2.3. Variables of interest

Primary endpoints of the study were to compare demographics of patients who were and were not readmitted within 90 days following the index procedure, determine the frequency of the causes of readmissions, and the associated costs. Demographics analyzed included: age, sex, and presence of comorbidities. Overall comorbidity burden between the cohorts was also compared using mean Elixhauser-Comorbidity Index (ECI) scores.

2.4. Data analyses

Chi-square analyses were used to compare patient demographics between readmitted patients and controls. Analysis of variance (ANOVA) was utilized to determine differences in mean 90-day costs of care between the aforementioned causes of readmission during the study interval. Due to the ease of reporting statistical differences with large administrative databases, a p-value less than 0.001 was considered statistically significant.

3. Results

3.1. Readmission demographics

Of the 3432 patients (2.4%) that were readmitted, patients 75–79 years old had the highest number of readmissions, accounting for 24.04% of readmissions (n = 825), Table 1. This was followed by the 70–74 age cohort (20.86%, n = 716) and 65–69 age range (18.76%, n = 644). The lowest amount of readmissions was among those below the age of 64 (10.23%, n = 351). For sex, female readmissions outnumbered male readmission by a ratio of 1.54 (female readmission n = 2083, male readmission n = 1349, p = .06). Readmitted patients had a higher comorbidity burden per mean Elixhauser-Comorbidity Index (ECI) score (10 vs, 7, p < .0001), Table 1. The most common comorbidities among readmission patients were hypertension (94.87%, p < .0001), fluid/electrolyte imbalance (73.19%, p < .0001), pulmonary disease (57.46%, p < .0001), arrythmias (54.46%, p < .0001), depression (45.72%, p < .0001), diabetes (45.72%, p < .0001), and congestive heart failure (45.63%, p < .0001), Table 1.

Table 1.

Comparison of patients readmitted within 90-days following primary total shoulder arthroplasty for glenohumeral osteoarthritis to controls.

Demographics	Readmitted Patients		Control		p-valuea
Age (Years)	n	%	n	%	<0.0001
<64	351	10.23	12,737	9.07
65 - 69	644	18.76	36,032	25.66
70 - 74	716	20.86	35,810	25.50
75 - 79	825	24.04	30,722	21.87
80 - 84	607	17.69	18,355	13.07
85<	289	8.42	6790	4.83
Sex					0.06
Female	2083	60.69	82,981	59.08
Male	1349	39.31	57,465	40.92
Comorbidities
AIDS	*	N/A	234	0.17	0.255
Alcohol Use Disorder	236	6.88	6591	4.69	<0.0001
Arrhythmias	1869	54.46	53,917	38.39	<0.0001
Blood Loss Anemia	468	13.64	9757	6.95	<0.0001
BMI <19 kg/m	53	1.54	1239	0.88	<0.0001
BMI 19–24 kg/m	76	2.21	2097	1.49	0.001
BMI 25–30 kg/m	121	3.53	3886	2.77	0.008
BMI 30–39 kg/m	419	12.21	15,720	11.19	0.066
BMI 40–70 kg/m	257	7.49	8152	5.80	<0.0001
CHF	1566	45.63	33,377	23.77	<0.0001
Coagulopathies	691	20.13	18,249	12.99	<0.0001
Depression	1569	45.72	46,810	33.33	<0.0001
Diabetes Mellitus	1569	45.72	49,740	35.42	<0.0001
Fluid/Electrolyte Imbalance	2512	73.19	67,405	47.99	<0.0001
Hypertension	3256	94.87	124,604	88.72	<0.0001
Hypothyroidism	1276	37.18	46,552	33.15	<0.0001
Liver Failure	320	9.32	9385	6.68	<0.0001
Lymphoma	118	3.44	3501	2.49	<0.0001
Metastatic Cancer	242	7.05	6071	4.32	<0.0001
Neurological Disorders	689	20.08	17,374	12.37	<0.0001
Paralysis	209	6.09	4740	3.37	<0.0001
Peptic Ulcer Disease	511	14.89	13,811	9.83	<0.0001
Pul. Circ. Disorders	617	17.98	12,913	9.19	<0.0001
Pulmonary Disease	1972	57.46	58,269	41.49	<0.0001
PVD	1325	38.61	37,487	26.69	<0.0001
Renal Failure	387	11.28	7091	5.05	<0.0001
Rheumatoid Arthritis	782	22.79	26,397	18.80	<0.0001
Valvular Disorders	1219	35.52	35,280	25.12	<0.0001
Weight Loss	568	16.55	9998	7.12	<0.0001

AIDS = Acquired Immunodeficiency Syndrome; BMI = Body Mass Index; CHF = Congestive Heart Failure; PVD = Peripheral Vascular Disease; * = <11 Patients; N/A = Not Applicable.

^a= Assessed by Pearson's Chi-Square Analyses.

3.2. Cause of readmission

The leading causes of ninety-day readmission following primary total shoulder arthroplasty for glenohumeral osteoarthritis were musculoskeletal (17.34%), cardiac (16.28%), infections (16.26%), gastrointestinal (11.64%), and neurological (10.40%), Fig. 1.

Fig. 1.

Leading causes of ninety-day readmissions following primary total shoulder arthroplasty for glenohumeral osteoarthritis.

3.3. Associated costs

The highest costs of ninety-day readmissions following TSA for glenohumeral osteoarthritis included cardiac ($10,913.70), musculoskeletal ($10,590.50), infections ($10,075.92), neoplasm ($9816.75), and pulmonary ($9202.17), Table 2.

Table 2.

Costs of ninety-day readmissions stratified by the primary diagnostic cause in patients undergoing primary total shoulder arthroplasty for glenohumeral osteoarthritis.

Year	Cardiac	Pulmn	Neuro	Heme	MSK	GI	Infections	Misc	Neoplasm	Endocrine	Renal
2005	$12,461.93	$6189.91	$9601.40	$4646.50	$8787.74	$6358.81	$9677.00	$6295.57	$8274.56	$6316.80	$5827.96
2006	$14,814.70	$6224.05	$5628.13	$3431.00	$7954.25	$8976.88	$7860.47	$5487.13	$7058.70	$6651.00	$8716.21
2007	$9177.29	$11,630.94	$9388.75	$4793.57	$11,242.53	$10,045.17	$9074.25	$8567.26	$10,986.58	$4909.00	$7224.32
2008	$12,079.86	$14,306.00	$6734.18	$5576.30	$10,252.92	$9509.93	$10,526.38	$5987.81	$14,534.25	$31,475.00	$5538.03
2009	$10,966.37	$10,102.79	$7951.71	$5024.18	$10,882.88	$8149.37	$11,474.95	$6314.85	$10,570.15	$6046.86	$5758.81
2010	$9152.14	$7924.71	$5886.50	$5489.87	$9543.94	$8357.98	$7908.62	$5459.39	$13,142.00	$3747.00	$5376.06
2011	$8629.68	$6362.67	$7811.00	$6251.63	$10,749.45	$7390.13	$11,827.85	$7141.20	$6977.50	$6265.60	$6639.29
2012	$9946.84	$12,120.07	$7766.64	$7596.50	$12,960.88	$11,043.09	$13,078.20	$18,032.10	$8982.75	$8135.00	$5393.68
2013	$11,682.00	$6855.72	$9201.75	$8035.29	$12,519.03	$9119.77	$9219.22	$7067.33	$7680.00	$4800.00	$6889.86
2014	$10,226.15	$10,304.80	$6195.00	$5092.44	$11,011.38	$7082.95	$10,112.28	$7665.46	$9961.00	$16.589.35	$5152.42
Mean	$10,913.70	$9202.17	$7616.51	$5593.73	$10,590.50	$8603.41	$10,075.92	$7801.81	$9816.75	$8705.14	$6251.66
SD	$1894.74	$2898.90	$1472.42	$1382.70	$1544.16	$1427.78	$1691.36	$3724.61	$2551.18	$8631.17	$1116.56

Pulmn = Pulmonary; Neuro = Neurological; Heme = Hematological; MSK = Musculoskeletal; GI = Gastrointestinal; Misc = Miscellaneous; SD = Standard Deviation.

4. Discussion

With the growing demand for TSA for glenohumeral osteoarthritis, it has become imperative to decipher the demographics and risk factors that place patients at an increased risk for readmission. Furthermore, as hospitals face readmission penalties from Medicare,²¹ it's important to understand the most common readmission causes and their associated costs, in order to reduce the burden on healthcare systems. The results from this study yielded valuable information related to these topics. Patients with higher incidence of readmission included females over the age of 75, and higher prevalence of comorbidities, including hypertension, fluid/electrolyte imbalance, pulmonary disease, arrythmias, depression, diabetes, and congestive heart failure. The most frequent causes for ninety-day readmission included musculoskeletal (average cost of $10,590.50), cardiac (average cost of $10,913.70), infections (average cost of $10,075.92), gastrointestinal (average cost of $8603.41), and neurological (average cost of $7616.51).

The findings of this study coincided with the results from prior investigations. We showed that 90-day readmissions following TSA were more likely in patients over 75 years of age and with a higher comorbidity burden (Elixhauser-Comorbidity Index (ECI) score (10 vs 7)). In a study conducted by Cvetanovich et al., evaluating The American College of Surgeons National Surgical Quality Improvement Program for primary total shoulder arthroplasty readmissions, they found increased 30 day readmissions for patients above 80 years of age (RR = 3.1), as well as increased readmission rates for musculoskeletal and infectious etiologies.²² Another study performed by Matsen et al., which studied the factors affecting readmission after shoulder arthroplasty, uncovered similar results with increased readmission rates for increasing patient age and more associated medical comorbidities.²³ A study done by Xu et al. investigated risk factors for readmission following total shoulder arthroplasty and found that increased age over 75 and having multiple comorbidities were significantly associated with readmissions.²⁴ Of these comorbidities, the most common were hypertension, diabetes, chronic obstructive pulmonary disease, and cardiac issues. Because these prior studies are limited to short term (30-day) follow-up, smaller sample size, or restricted to a single statewide database, our study addresses important gaps on evaluating readmissions within the 90-day bundled payment window following TSA strictly for glenohumeral osteoarthritis.

In the present study, the leading etiologies for readmission within 90 days of index procedure were musculoskeletal (17.3%), cardiac (16.3%), infections (16.3%), and gastrointestinal (11.6%). When assessing readmission causes following total shoulder arthroplasty, prior studies have shown similar results with musculoskeletal and infections most responsible.22, 23, 24 Schairer et al. examined 14,602 patients who underwent total shoulder arthroplasty for osteoarthritis, of which 6% were readmitted within 90 days.¹ Of these readmissions, 82% were due to medical issues, such as osteoarthritis and deep vein thrombosis, while 18% accounted for surgical complications, such as infections and musculoskeletal problems. A retrospective review at a single institution from 2005 to 2011 by Mahoney et al. also yielded similar findings, with the most common reasons for readmission related to infection and musculoskeletal causes.²⁵ In a 2014 retrospective cohort study by Chung et al., the top reasons for patient readmission following total shoulder arthroplasty included musculoskeletal complications (dislocations, mechanical loosening, joint fracture) and infections.²⁶ With the exponential rise of shoulder arthroplasty, a better understanding of the medical complications unrelated to orthopaedic surgery or the musculoskeletal system may be useful in the risk stratification process. Further delineation of the specific cardiac and gastrointestinal readmission etiologies is warranted.

By 2030, the demand for shoulder arthroplasties is expected to grow annually with a projected increase of 333.3% for patients younger than 55 and an increase of 755.4% for patients over the age of 55.²⁷ With this foreseeable increase in demand as well as the Medicare penalties for readmissions, it falls on medical professionals and institutions to make financially responsible decisions based on their patient demographics, as we face an aging population that will have a higher prevalence of glenohumeral osteoarthritis. We showed that cardiac ($10,913.70), musculoskeletal ($10,590.50), and infections ($10,075.92) were most responsible for hospital expenditures as a result of readmission within 90 days. In a study by Chung et al., patients who were readmitted within 90 days after undergoing total shoulder arthroplasty cost the hospital 4.95 times as much as those patients who were not readmitted.²⁶ Another study completed by Kennon et al. found that the median readmission cost per total shoulder arthroplasty case was $11,031.²⁸ However, most studies, including these two, do not delineate the indication for undergoing a primary total shoulder arthroplasty. As postulated by Malik et al., there is a higher resource utilization in shoulder arthroplasty patients that have undergone a fracture, as opposed to those who have OA of the glenohumeral joint.¹⁹ Fractures have been associated with longer lengths of stay, 30-day complications, revisions, and readmissions, making them ultimately more costly. There is limited literature that focuses on the readmission of glenohumeral OA patients who have underdone primary total shoulder arthroplasty, making the findings of this study an important contribution to the literature.

This study is not without its limitations, as using an administrative database comes with its own inherited limitations, such as coding or clerical errors.²⁹ One such limitation includes the possibility that patients might not be clinically diagnosed for a risk factor, such as depression or alcohol use disorder. Another limitation of this study is that readmissions were limited to 90 days post-op, leaving any readmission that occurred after that time period to not be accounted for. The granularity of the readmissions were not accounted for and are opportunities for future study. Despite these limitations, the study has a large sample size that ensures a high power and this study is the first to our knowledge to investigate the patient demographics, causes, and costs of 90-day readmissions following primary total shoulder arthroplasty for glenohumeral osteoarthritis.

5. Conclusion

After completing a retrospective analysis of 143,878 patients who had undergone total shoulder arthroplasty for glenohumeral osteoarthritis, 3432 (2.4%) were readmitted within 90 days. Older patients (over the age of 75), female, and higher overall comorbidity burden were risk factors for readmission. This study also determined the top causes and associated costs for readmissions in this patient population with musculoskeletal and cardiac etiologies most responsible. The results from this study are a valuable addition to the literature, as this information can be used by healthcare providers and facilities to not only educate patients on potential surgical outcomes, but to help make data driven decisions to potentially reduce readmission rates and their associated costs. Future studies should aim at comparing other indications for total shoulder arthroplasty and their demographics, as well as determining more granular causes and costs of readmissions.

IRB/ethical approval

The study was exempt from our institution's Institutional Review Board (IRB) approval process.

Funding

None.

CRediT authorship contribution statement

Marcos Vargas: Data curation, Writing – original draft. Giovanni Sanchez: Data curation, Writing – original draft. Adam M. Gordon: Conceptualization, Methodology, Data curation, Writing – original draft. Andrew R. Horn: Conceptualization, Methodology, Data curation, Writing – original draft. Charles A. Conway: Conceptualization, Methodology. Afshin E. Razi: Visualization, Investigation, Supervision, Writing – review & editing. Ramin Sadeghpour: Visualization, Investigation, Supervision, Writing – review & editing.

Declaration of competing interest

All Authors have nothing to disclose.

Appendix A. Supplementary data

The following is the Supplementary data to this article: