1 Introduction

The number of primary total joint arthroplasty (TJA) procedures, including hip, knee, shoulder, elbow, ankle or wrist, have increased from 2009 to 2019 and continue to rise.^1–3 As the number of these procedures continues to grow, it is imperative to identify, and thus ameliorate, factors contributing to suboptimal patient outcomes.

Various studies conducted utilizing national databases have demonstrated Medicaid status to be an independent risk factor for increased morbidity, reoperations, and resource utilization.^4–7 For instance, a National Inpatient Sample (NIS) database study by Browne et al. reported patients from 2002 to 2011 with Medicaid insurance undergoing TJA were more likely to experience postoperative complications than TJA patients with non-Medicaid primary insurance. However, this study was limited to hip or knee arthroplasties.^4,8 In addition, this study did not examine Medicare patients or patients who self-paid. To the authors’ knowledge, no study has investigated the impact of primary payer status on postoperative outcomes for all major joint arthroplasties.

Our study expands on the existing literature^5,9–11 by investigating postoperative complications and resource utilization among patients undergoing inpatient, major primary joint arthroplasties (hip, knee, shoulder, shoulder, ankle, or wrist) from 2012 to 2020 stratified by primary payer status. We hypothesized that Medicare or Medicaid insurance would be independently associated with higher rates of postoperative complications and report higher healthcare resource utilization when compared to patients with private insurance, even when controlling for confounding variables.

2 Methods

The Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) database was used to identify patients who underwent total hip, knee, shoulder, elbow, ankle, or wrist joint arthroplasty surgery from 2012 to 2020. The International Classification of Disease-9 (ICD-9) and ICD-10 codes used are detailed in Table 1.

3 Results

3.1

3.1 General characteristics

The number of entries identified from the HCUP NIS database from 2012 to 2020 was 1,940,147, the methodology which is described in Fig. 1. After elimination of entries with incomplete information, 1,875,189 entries remained. After only including entries with payer status as Medicaid, Medicare, private, or no insurance status, 1,813,611 entries remained. These patients were then split into the following cohorts: Medicare ≥65 years old (n = 964,494), Medicare <65 years old (n = 105,159), Medicaid (n = 78,290), no insurance (n = 10,387), and private insurance (n = 679,168). By examining Table 2, matching is clearly necessary; patients in the Medicare ≥65 years cohort tended to be older within their stratification, as 60.3 % of entries were in the 65–74-year age group, while only 14.9 % of private entries were in the 65–74-year age group. The statistics in Table 2 were multiplied by NIS-provided weights to provide national estimates.

Close

Fig. 1 Flowchart showing the matching and analysis of the population. Filtering of the data after acquisition from the NIS and subsequent matching process. Cohorts used for multivariable analysis.

Export to PPT

4 Discussion

The increased incidence of complications and healthcare utilization occurring in Medicaid patients undergoing orthopaedic procedures has been well documented in the literature.^4,7,17,18 With the ever-expanding prevalence of Medicaid being adopted by uninsured patients in the United States, especially since the COVID-19 pandemic,^19–21 it is important to continue to understand why such patterns exist. Our study adds to the existing literature by exploring the effect, not only of Medicaid, but also of Medicare and uninsured patients, on postoperative complications and healthcare utilization after TJA.

Medicaid payer status has been demonstrated to act as an independent risk factor for worse outcome measures with TJA, as enrollees have higher rates of postoperative mortality, complications, resource utilization, longer lengths of stay, and 90-day ED visits and readmissions than patients with Medicare or private insurance.^9,10,22,23 In total hip arthroplasty, both Medicaid and Medicare patients have shown higher hazard ratios for hospital charges and increased complications, such as periprosthetic infections, when compared to privately insured patients.^11,24 In our study, hospital charges were $1170 more for Medicaid patients than they were for patients with private insurance. Given the demographics that constitute the Medicaid patient population—differences in social determinants of health, such as a lack of regular doctor appointments; the presence of more complicated, chronic conditions along with a myriad of social factors—the costly nature of these patients’ admissions should not be entirely surprising, though some studies suggest that Medicaid patients are less costly to beneficiaries, so more research is needed to elucidate these differences.^25–27

Our data further revealed that Medicaid patients tended to stay in the hospital for 0.21 days longer than patients with private insurance. These results were in line with studies suggesting longer lengths of stay for non-privately insured patients, particularly Medicaid patients.^28–30 Previously, increased LOS has also been associated with more short-term complications, which further explains the increased risk of complications for both Medicaid and Medicare <65 patients in this study.^27,31 These increases in LOS are likely secondary to modifiable factors acting as barriers to discharge. For example, the increased incidence of smoking and illicit drug use, social concerns resulting in inadequate pain control, socioeconomic factors, and delay in post-operative mobilization have all been cited as barriers to discharge in this population.^32–34

Medicare patients <65 years old were more likely than Medicare patients ≥65 years old to experience a central nervous system complication and genitourinary complications relative to privately insured patients. Despite our matching and controlling for comorbidities, these differences in estimated relative risk could potentially be explained by the severity of comorbidities impacting Medicare <65 patients who typically qualify for Medicare.³⁵ This includes patients who suffered severe, traumatic accidents, patients with amyotrophic lateral sclerosis, or patients with permanent kidney failure.^36–38 Studies have demonstrated that Medicare patients <65 undergoing lumbar fusion have worse pre- and post-operative patient-reported outcome measures.^11,39 Moreover, it has been shown that Medicare patients with disabilities less than 65 years of age are more likely than elderly beneficiaries to be impoverished minorities, factors which both can contribute to worse outcomes.⁴⁰ Our results showed that hospital charges were $3849 more for Medicare patients <65 and $2243 more for patients ≥65 when compared to patients with private insurance. In recent studies, individuals with private insurance have reported higher costs of care, suggesting that those with public insurance such as Medicare have more cost-effective care; however, this was not focused on surgical procedures.⁴¹ We hypothesize the severity of comorbidities and quality of private insurance coverage in patients who can afford TJA in the United States to be an explanation for our findings.¹¹

A decrease in the number of private-enrolled patients in our sample of patients >65 could be due to several factors. Adults >65 years older are the age group most likely to be covered by more than one plan of insurance secondary to a desire to supplement Medicare coverage with privately purchased plans or coverage through a current or previous employer,⁴² with 29.5 million people aged 65 or older having multiple coverage in 2021. Private insurance is the primary payer when an individual does not sign up for Medicare, or they work for a large company.³⁵ Medicare is the primary payer, typically among retirees and with 86 % of adults 75 and older being retired; this justified the decrease in private-insurance primary payers in our sample above the age of 65.⁴³ The most common type of multiple coverage among this patient population was a combination of Medicare and private insurance.⁴² This puts into context the cohorts being compared when looking at primary payers of Medicare ≥65 years old.

At 65, the age of Medicare eligibility, patients are likely to undergo more frequent screening for malignancies, such as prostate cancer, especially if they are no longer employed and have slightly less strenuous work schedules.^44,45 As such, this cohort of patients may ultimately find malignancies at a lower stage at diagnosis relative to the rest of the population. Patients aged >65 with primary coverage via private insurance are likely still in the workforce⁴⁶ and, as such, are less likely to undergo such consistent screening or consistent visits to the doctor's office, potentially contributing to the increased risk of genitourinary complications we found. In addition, we found patients with primary Medicare coverage to be more likely to experience a cardiac complication than privately insured patients; this is consistent with the literature, which suggests that continued work beyond the age of 65 is likely to result in better health outcomes relative to those who do not work beyond that point.⁴⁷ This may be due to higher activity rates of individuals who remain in the workforce at a later age.

Our multivariable analysis was a strength of this study and allowed us to analyze a large number of patient records while also controlling for several patient variables through matching. Despite this, our matching did not consider race and other, differing hospital characteristics such as hospital location (rural, urban teaching, urban nonteaching). We attempted to account for this by including these characteristics as covariates in our analysis. In addition, dual-eligible patients were unable to be identified because of database limitations. Most dual-eligible patients are classified as Medicaid payers in the NIS. In addition, the analysis of no-insurance-based patient data did not allow us to make any significant conclusions regarding complications in that population. It is important to consider that while our paper demonstrates associations according to payer status and overall postoperative outcomes when it comes to TJAs, the retrospective methodology precludes the establishment of causality. These differences could be accounted for with nonquantifiable social determinants of health, and, as such, a retrospective multivariable analysis cannot aptly capture the full picture.

5 Conclusion

Ultimately, after exact matching for comorbidities, our study found that Medicaid and Medicare patients appeared to have higher risks for postoperative complications after joint arthroplasty. In addition, hospital charges were greater for both Medicare and Medicaid patients, regardless of age. Our findings build upon the available literature investigating insurance status on post-operative outcomes and are important to analyze as the utilization of Medicaid and Medicare continues to expand among US patients, especially after the expansion of the Affordable Care Act. However, we also found that patients who are older than 65 and insured with Medicare are not as affected by payer status, and that the more vulnerable populations on Medicare and Medicaid need to be targeted for healthcare access and better care. Further research will allow a better understanding of which precise factors lead to suboptimal outcomes in these patient populations. Such research could examine how hospital location and accessibility to care, which may be associated with payer status, could affect these results.

CRediT authorship contribution statement

William ElNemer: contributed, Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, and, Funding acquisition. Sribava Sharma: contributed, Conceptualization, Methodology, Software, Writing – original draft, and, Writing – review & editing. John P. Avendano: contributed, Conceptualization, Writing – original draft, and, Writing – review & editing. Myung-Jin Cha: contributed contributed, Conceptualization, Writing – original draft, and, Writing – review & editing. Majd Marrache: contributed, Writing – original draft, Writing – review & editing. Andrew B. Harris: contributed, Writing – original draft, Writing – review & editing, Supervision, Project administration, and, Funding acquisition. Umasuthan Srikumaran: contributed, Writing – original draft, Writing – review & editing, Supervision, Project administration, and, Funding acquisition. Matthew J. Best: contributed, Writing – original draft, Writing – review & editing, Supervision, Project administration, and, Funding acquisition.

Guardian/patient's consent

No patient consent was needed or obtained.

Ethical statement

This study is IRB exempt because of the use of a national deidentified sample database.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.