Decreased costs with maintained patient satisfaction after total joint arthroplasty in a physician-owned hospital

Abstract

Objective

Comparing total joint arthroplasty (TJA) costs and patient-reported outcomes between a physician-owned hospital (POH) and a non-POH.

Methods

Costs for each 90-day TJA episode at both facilities were determined, and patients were asked to complete a patient satisfaction questionnaire.

Results

Average TJA episode cost was $19,039 at the POH, compared to $21,302 at the non-POH, a difference of $2,263 (p = 0.03), largely driven by decreased skilled nursing facility utilization in the POH group. There were no differences between groups for patient satisfaction.

Conclusion

TJA can be performed at reduced cost with comparable patient satisfaction at POHs, compared to non-POH facilities.

1. Introduction

Section 6001 of the 2010 Affordable Care Act (ACA) placed restrictions on the expansion of physician-owned hospitals (POH) in the United States.^1,2 This policy continues to be debated in the healthcare literature, and multiple studies have been published on both sides of the debate. Proponents of the ACA restrictions argue that POHs may distort financial incentives, which may negatively affect practice patterns. For example, some authors have raised concerns that POHs could preferentially target healthier and more profitable patients, thus hindering large full-service hospitals (FSH) from supplementing their less profitable services.3, 4, 5 Others have argued that POHs do not necessarily produce more efficient care.⁶

However, there is now a growing body of literature suggesting POHs are more efficient, have lower costs, and provide better patient outcomes, particularly regarding total joint arthroplasty (TJA). Several recent investigations have demonstrated lower costs, less complications, and higher patient satisfaction in POHs compared with non-POHs.7, 8, 9 Most of the previous studies on this topic have utilized large patient databases, including data from a multitude of surgeons and hospitals at either a state or national level.7, 8, 9 While compelling, these data could potentially introduce a significant degree of heterogeneity despite attempting to control for confounding variables. Thus, there exists a need in the literature for studies which directly compare episode-of-care costs and patient satisfaction between POH and non-POH facilities, while keeping important variables such as geographical location, patients, care pathways, surgeons, implants, and surgical technique comparable.

Thus, the purpose of the present study was to directly compare TJA costs and patient-reported outcomes (PRO) between a single POH with a single non-POH in which the geographical location, surgeons, and implants, and care pathways were essentially identical. We asked the following research questions for this study: 1) does a specialty POH provide total joint arthroplasty (TJA) at reduced cost compared to a non-POH tertiary care facility? and 2) Is patient satisfaction higher at a specialty POH compared to a non-POH tertiary care facility?

2. Methods

2.1. Study design

Approval by the Institutional Review Board at our institution was obtained prior to undertaking the present study. Using deidentified CMS data that was aggregated by our bundled payments for care improvement (BPCI) convener, Medicare TJA patients aged >65 years who are managed under our physician-owned contract bundle with a diagnosis-related group (DRG) of either 469 (Major Joint Replacement or Reattachment of Lower Extremity with Major Complications or Comorbidities) or 470 (Major Joint Replacement or Reattachment of Lower Extremity without Major Complications or Comorbidities) were identified and reviewed. We included for study all patients who underwent unilateral total joint arthroplasty performed between October 2018, when the bundle was initiated, through March of 2020. All procedures were performed by two fellowship-trained arthroplasty specialists at one of two locations: 1) a 30-bed specialty orthopaedic hospital that is physician-owned (POH); and 2) an 875-bed tertiary care hospital that is not physician-owned (non-POH). These hospitals are located approximately one mile from each other in a large metropolitan area, and serve the same geographical patient population. Identical implant systems, surgical technique, perioperative protocols, care pathway, anesthesia team, and physician extenders were used at both institutions. We excluded five patients: four who underwent non-elective hip hemiarthroplasty for a femoral neck fracture, and one patient who upon further review was found to have undergone a revision rather than primary total knee arthroplasty (TKA).

After application of these inclusion and exclusion criteria, a total of 209 consecutive TJAs were identified for inclusion in the study, 144 TJAs at the POH and 65 TJAs at the non-POH. Patient demographic data was collected from the electronic medical record (EMR) at our institution and is displayed in Table 1. Using the CMS data provided by our convener, we determined the Total Cost (TC) associated with each 90-day TJA episode. We further determined breakdown of costs for each patient including the Anchor Cost (AC) (a summation of costs associated with the surgery itself including DRG payment to hospital, surgeon fee, anesthesia costs, etc), as well as durable medical equipment (DME) cost, Part “B” costs (PB), outpatient costs (OP) including physical therapy or doctors’ visits, home health (HH) costs, skilled nursing facility (SNF) costs, and costs associated with re-admission. We further recorded the number of home health visits, SNF days, and hospital re-admission incidence for each patient group.

Table 1.

Perioperative demographic variables.

	POH (n = 144)	Non-POH (n = 65)	p
Age (years)	73.3 (range 65.5–89.3)	74.5 (range 65.4–92.1)	0.11
Gender
Male (n, %)	56 (39%)	32 (49%)	0.23
Female (n, %)	88 (61%)	33 (51%)
Body Mass Index (kg/m2)	28.6 (range 19.5–38.9)	30.0 (range 16.1–40.3)	0.07

Patient satisfaction was determined via questionnaire administered post-operatively at the most recent follow-up visit, asking the following questions: 1) How would you rate your overall satisfaction with the surgery outcome? And 2) How would you rate your overall satisfaction with the hospital experience? For each question, the response was scored as: Very Satisfied, Satisfied, Neutral, Unsatisfied, or Very Unsatisfied.

The primary outcome of our study was the average Total Cost per TJA patient, calculated for both the POH and non-POH hospitals. Secondary outcomes included the average cost for each of the cost sub-categories in each group, including Anchor, DME, PB, OP, HH, SNF, and re-admission costs. As a final secondary outcome, we compared patient satisfaction in each group as determined by the patient questionnaire.

2.2. Statistical analysis

All data were entered, stored, and analyzed using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) and GraphPad QuickCalcs (GraphPad Software, San Diego, CA, USA). Continuous variables were analyzed using the Student's t-test, while categorical variables were analyzed using Fisher's exact test. In all cases, statistical significance was set at p = 0.05.

3. Results

When assessing baseline demographic variables between groups, no significant differences were noted between POH and non-POH groups for gender, age, or BMI (all p > 0.05). The average total cost (TC) of a total joint replacement episode was $19,039 at the POH, compared to $21,302 at the non-POH, a difference of $2,263 (p = 0.03). This difference appeared to be largely driven by decreased SNF utilization in the POH group. Including zero-dollar costs for those who did not get admitted to SNFs, the average SNF cost per TJA episode was $209 in the POH group, compared to $2,219 in the non-POH group, a difference of $2,010 (p = 0.003). There were a total of six SNF admissions in the POH group (4.2%), compared to twelve SNF admissions (18.5%) in the non-POH group (p = 0.002). The average Anchor Cost (AC) per episode was $14,006 for the POH group, compared to $14,630 in the non-POH group, a difference of $624 (p = 0.0006). There were no other significant differences between groups for any of the additional cost sub-categories analyzed (Table 2).

Table 2.

Cost analysis for total joint arthroplasty.

	POH (n = 144)	Non-POH (n = 65)	p
Total Cost	$19,039 (range $14,028—$48,305)	$21,302 (range $14,547–82,544)	0.03
Anchor Cost	$14,006 (range $12,720—$15,137)	$14,630 (range $10,469—$22,573)	0.0006
DME Cost	$54 (range $0—$1,026)	$109 (range $0–1269)	0.08
Part “B” Cost	$1,143 (range $27—$4518)	$1,175 (range $0—$5,305)	0.83
Outpatient Cost	$849 (range $0—$21,691)	$485 (range $0—$3,923)	0.17
HH Visits	6 (range 0–23)	7 (range 0–23)	0.47
HH Cost	$2,192 (range $0—$3,783)	$2,042 (range $0—$5,266)	0.41
SNF Admissions	6 (4.2%)	12 (18.5%)	0.002
Mean SNF Cost per patient*	$209 (range $0—$11,733)	$2,219 (range $0—$54,976)	0.003
Mean SNF Days per admission**	10.2 days (range 6–23 days)	22.2 days (range 6–100 days)	0.29
Cost per actual SNF admission**	$5,010 ($1,861—$11,733)	$12,021 ($2,344—$54,976)	0.27
Readmissions	3 (2.1%)	4 (6.2%)	0.21
Readmission Cost	$17,262 (range $8,488—$22,546)	$11,787 (range $6,793—$17,394))	0.37

All costs listed as mean cost per TJA. DME = Durable Medical Equipment; HH = Home Health; SNF = Skilled Nursing Facility; * = calculated for all study patients, including those not admitted to SNF; ** = calculated only for patients who had a SNF admission.

A total of 141 patients (141/209, 67.5%) completed the patient satisfaction questionnaire. There was no difference between POH patients (mean 1.5 years, range 0.8–2.2 years) and non-POH patients (mean 1.3 years, range 0.8–2.2 years) for duration of follow-up (p = 0.064). When assessing patient satisfaction responses, patients in the POH group demonstrated similar overall satisfaction with the surgery outcome (97.8% very satisfied or satisfied) compared with patients in the non-POH group (96.2% very satisfied or satisfied) (p = 0.63). Additionally, patients in the POH group demonstrated similar satisfaction with the hospital experience (96.6% very satisfied or satisfied) compared with patients in the non-POH group (92.3% very satisfied or satisfied) (p = 0.42) (Table 3).

Table 3.

Patient satisfaction responses.

Overall Satisfaction						P*
	Very Satisfied	Satisfied	Neutral	Unsatisfied	Very Unsatisfied
POH (n)	76	11	0	1	1	0.63
Non-POH (n)	42	8	1	1	0

Hospital Satisfaction						P*
	Very Satisfied	Satisfied	Neutral	Unsatisfied	Very Unsatisfied
POH (n)	75	11	0	3	0	0.42
Non-POH (n)	42	6	3	0	1

POH = physician-owned hospital; * = p values calculated to compare response rates of “very satisfied or satisfied” versus “not very satisfied or satisfied”.

4. Discussion

In the setting of the restrictions placed on physician-owned hospitals (POH) by the Affordable Care Act, debate continues as to whether POHs can improve quality and reduce costs associated with total joint replacement. While attempting to conserve such variables as surgeon technique, implant, perioperative protocol, care pathways, and patient population, the present study demonstrates a cost savings of more than $2,000 per case, driven largely by savings from reduced SNF utilization, when TJA is performed in a POH. Patients were equally satisfied with their surgery outcome and hospital experience whether the joint replacement was performed in the POH or the non-POH.

Our study has several limitations. The most obvious limitation is the potential for selection bias within the two study groups. It could be argued that less healthy patients who are more likely to utilize healthcare resources may have been preferentially diverted to the tertiary care non-POH facility, while healthier patients were more likely to receive care at the specialty orthopaedic POH facility. We attempted to mitigate this potential bias by only including BPCI patients aged >65 in our study population. By virtue of inclusion in our bundled payment program, all patients had undergone an identical pre-optimization protocol, which uses established criteria to screen high risk patients with increased likelihood of complications. For example, these criteria require patients with excessive BMI, uncontrolled diabetes, malnutrition, and other medical issues to be optimized by a medical doctor prior to undergoing surgery. Additionally, when assessing perioperative demographic variables between the two groups, there were no differences in age, gender, or BMI between POH and non-POH patients. Thus, despite the possibility of selection bias in this study, we feel our study design of including only pre-optimized BPCI patients creates a relatively homogenous study population which helps to mitigate this potential bias.

A second potential limitation is that our study involves a direct comparison between two separate hospitals, and we limited our study population to only include patients from two arthroplasty specialists in the same orthopaedic practice. While this was done purposefully to reduce variability between the study groups (as both surgeons use the same implants, techniques, surgical staff, anesthesia team, and BPCI pre-optimization pathway in both hospitals), we cannot rule out that our results may not be applicable to other patient populations and hospitals outside of our geographic region. Regional differences in cost structure, SNF utilization, or hospital type may potentially affect these reported outcomes.

A third limitation is that only 67.5% of patients in our study completed the patient satisfaction questionnaire. Thus, while patient satisfaction appears high in both hospital settings, we cannot rule out that this component of our study may have lacked the necessary power to detect a difference in patient satisfaction between groups.

Debate continues in the current arthroplasty literature regarding the role of POHs. Courtney et al. compared 45 POHs to 2,657 non-POHs using a large CMS database, and found that POHs had lower mean Medicare costs, improved risk-adjusted complication scores, and higher scores in all patient-satisfaction categories.⁷ Similarly, Cram et al. have compared TJAs at specialty and general hospitals concluded that patients had a lower risk of adverse events at specialty hospitals, even when accounting for patient characteristics and procedure volume.¹⁰ Bae et al. evaluated POHs versus full-service hospitals (FSH) across multiple procedures and found that procedures done at FSH facilities had a significantly higher risk of complications.⁹

Conversely, Lundgren and colleagues analyzed CMS data from 145 hospitals (eight POH, sixteen proprietary hospitals, and 121 general hospitals) and found that POHs had higher average total episode spending, though noting higher patient satisfaction scores at the POH facilities.⁸ Carey et al. reviewed hospital inpatient discharge data from three states in which POHs are common (Texas, Arizona, and California) and concluded that specialty hospitals have higher levels of cost inefficiency when compared to full-service hospitals.⁶ Additionally, other studies have suggested that POHs can indirectly increase non-POH healthcare costs by selectively targeting healthier patients and more lucrative procedures, thus taking these cases away from the non-POH facilities.^4,5 However, in a recent retrospective study comprising 14 full-service hospitals (FSH) and two surgical hospitals (SH) in which the two SH were joint ventures between physicians and their hospital partners, the authors found after controlling for medical comorbidities and demographic variables that TJA at a SH facility was associated with significant overall episode cost savings for both Medicare ($3266) and private insurance ($13,132) payers.¹¹ The present study is in agreement with the above findings of Courtney et al., in that Medicare costs were significantly decreased in the POH setting. Interestingly, the majority of savings associated with the POH in our study was driven by reduced SNF utilization, which may underscore the increased level of control that surgeons have on post-operative disposition when surgery is performed at a POH compared to a large tertiary-care facility.

5. Conclusion

In conclusion, the moratorium placed on the expansion of physician-owned hospitals by the Affordable Care Act continues to be questioned. The present study demonstrates that total joint replacement in Medicare beneficiaries can be performed at reduced cost and with comparable patient satisfaction at physician-owned hospitals, when compared to non-POH facilities.

Ethical approval

The study was approved by the Institutional Review Board at our medical center.

Funding

None declared.

Declaration of competing interest

The authors have declared no conflict of interest.