Use of Routine Home Health Care and Deviations from an Uncomplicated Recovery Pathway after Radical Prostatectomy

Abstract

Objective

To evaluate the statistical association between routine home health use after prostatectomy, short-term surgical outcomes, and payments.

Methods

We identified all men who underwent a robotic radical prostatectomy from 04/01/2014 through 10/31/2015 in the Michigan Urological Surgery Improvement Collaborative (MUSIC) with insurance from Medicare or a large commercial payer. We calculated rates of “routine” home care use after prostatectomy by urology practice. We defined “routine” home care as home care initiated within four days of discharge among patients discharged without a pelvic drain. We then compared ED visits, readmissions, prolonged catheter use, catheter reinsertion rates, and 90- day episode payments, unadjusted and in a propensity-adjusted analysis, for those who did and did not receive home care.

Results

We identified 647 patients, of whom 13% received “routine” home health care. At the practice level, the use of routine home care after prostatectomy varied from 0 to 53% (p=0.05) (mean: 3.6%, median: 0%). Unadjusted, patients with routine home care had increased ED visits within 16 days (15.5% vs 6.9%, p<0.01), similar rates of catheter duration > 16 days (3.6% vs 3.0%, p=0.79) and need for catheter replacement (1.2% vs 2.5%, p=0.46), and a trend toward decreased readmissions (0 vs 4.1%, p=0.06). Only the increased ED visits remained significant in adjusted analyses (p<0.01). Home health had an average payment of $1,000 per episode.

Conclusion

Thirteen percent of patients received “routine” home health care after prostatectomy, without improved outcomes. Patients don’t require “routine” home health care to improve short-term outcomes after radical prostatectomy.

Medicare and commercial payers are transitioning from paying for quantity of services to paying for the value of care, which includes costs and quality. Variation exists in the costs and quality of many conditions, including total episode costs for radical prostatectomy^1,2. The predominant drivers of the variation in 90-day episode costs after radical prostatectomy are readmissions (related to complications), professional payments and post-acute care, which all contribute almost equally to the variation².

Although we know the broad categories where variation exists, a better understanding of how specific physician practice patterns impact both short-term outcomes and payments is critical to improving outcomes and decreasing costs. After a radical prostatectomy, some patients are discharged with home health care for routine post-operative care. This “routine” use of home health care is in contrast to patients who receive catheter teaching in the hospital and are only discharged with home health care for the treatment of specific patient-related concern or complication that occurred during the hospital admission (e.g. a wound infection requiring packing).

“Routine” home health care is frequently prescribed to decrease patient and caregiver anxiety related to the urinary catheter and prevent catheter-related issues that would otherwise result in emergency department (ED) visits or readmission. However, clinical and cost outcomes following the use of “routine” home health care have not been evaluated previously and the “routine” use of home health care is a potential source of practice pattern variation. Home health care could prevent catheter-related complications, resulting in lower emergency department visits and readmissions. One would expect to see decreased ED visits and readmissions secondary to routine home health use during the time the catheter was in place. In contrast, routine home health may be unnecessary and associated with equivocal, or even worse outcomes. An evaluation of the use of home health care after pancreatectomy demonstrated increased 30-day readmissions when patients were discharged home with home health compared to no home care, even after adjusting for the propensity to receive home health³. In another analysis of all post-acute care discharges for patients >= 18 years of age from a health system, using propensity score matching, the authors found that patients discharged with home health had an increased risk of readmission compared to those discharged home⁴.

In this context, we aim to evaluate routine home health care after radical prostatectomy. Specifically, we assess variation in its use and whether its use is associated with catheter-related complications, frequency of ED visits, and readmissions. We evaluate ED visits and readmissions during the post-operative period when routine home care visits are occurring and the urinary catheter is likely in place. The availability of such data would clarify the value of routine home health care after prostatectomy and guide clinicians about its use.

METHODS

Data Sources

Data for this analysis was collected through the Michigan Urological Surgery Improvement Collaborative (MUSIC) and the Michigan Value Collaborative (MVC).

Michigan Urological Surgery Improvement Collaborative (MUSIC)

MUSIC is a physician-led collaborative established in 2012 with the goal of improving the quality and decreasing the costs of care associated with prostate cancer management. Currently about 90% of urologists in Michigan at 43 sites participate in the program and share knowledge and experience in managing and treating prostate cancer. A cornerstone of the collaborative is that each site has a trained data abstractor that collects detailed and standardized clinical information on each prostatectomy, including patient characteristics, pathology, and 30-day post-operative outcomes, including duration of use of pelvic drains. Each site’s data is collectively available for analysis and quality improvement activities.

As part of MUSIC’s efforts to improve short-term recovery after radical prostatectomy, the Notable Outcomes and Trackable Events after Surgery (NOTES) criteria were previously established and validated⁵. The criteria define an uncomplicated early recovery pathway following radical prostatectomy. We use two NOTES benchmarks for our outcome analyses (catheter placement 16 days or less and no 30-day catheter replacement), two benchmarks for our inclusion criteria (drain placement less than or equal to two days and no 30-day mortality), and one benchmark for a sensitivity analysis (length of stay two days or less).

Michigan Value Collaborative (MVC)

MVC is a quality improvement collaborative funded by the Blue Cross Blue Shield of Michigan (BCBSM) that aims to improve health care quality and decrease costs. Part of the collaborative involves maintaining a claims-based registry that provides detailed information on episodes of care for over 25 medical and surgical conditions⁶. Services are identified and categorized in four component categories: the index admission, readmissions, professional services, and post-acute care services. Post-acute care services include home health care, ED visits, skilled nursing faculty (SNF), rehabilitation services, and other outpatient post-discharge care. The 4 component categories are summed to create 90-day total episode payments, and all payments are price-standardized using the Michigan Medicare fee schedule to allow for comparisons across institutions. Both Blue Cross Blue Shield of Michigan preferred provider organization and Medicare fee-for-service claims are included in the database. Using claims allows identification of all services utilized during an episode, including ED visits and readmissions, regardless of the site of these services.

Study population

Our study population included all men who underwent a robotic radical prostatectomy (RP) in the Michigan Urological Surgical Improvement Collaborative who also had claims in MVC. We matched cases in the two datasets using birth date, surgery date, and insurance type. In claims data, RP index hospitalizations were identified by an International Classification of Diseases, Ninth Revision (ICD)-9 diagnosis code of 185 and either a professional claim with one of the following Current Procedural Terminology codes: 55810, 55812, 55815, 55840, 55842, 55845, or 55866 or a facility claim with an ICD-9 procedure code of 60.5 or 60.6x. We excluded claims with ICD-9 procedure codes of suprapubic prostatectomy, benign retro-pubic prostatectomy, cystectomy, or pelvic exenteration. These methods to identify claims in prostatectomy claims in MVC data have been used previously². We additionally excluded patients who died within 30 days of surgery, patients who had an index hospitalization longer than 30 days, and patients who had drain in place for more than two days. We excluded patients who had a drain in place for more than two days because we wanted to evaluate patients with routine peri-operative courses and prolonged use of pelvic drains likely signified the presence of a urinary leak. All related claims within 90 days after discharge are included in the episode. We focused on the use of home health and its associated outcomes within the first 16 days after discharge since over 95% of catheters in MUSIC are removed by post-operative day 16⁵.

Exposure variable

Routine home care was defined by initiation of home care within four days of discharge, identified by the first date of service on the home health claim, among patients discharged without a pelvic drain.

Outcomes

We measured four short-term postoperative outcomes and price standardized 90-day episode and component payments. The short-term postoperative clinical outcomes included: emergency department (ED) visits within 16 days of discharge (MVC claims), readmissions within 16 days of discharge (MVC claims), catheter in place for more than 16 days (MUSIC NOTES registry measure), and requirement for catheter replacement (MUSIC NOTES registry measure).

Statistical Analysis

We first analyzed variation in routine home health use after prostatectomy. To do this, we compared the proportion of patients receiving routine home health across MUSIC practices. Next, we performed univariate statistical analyses to compare patient and cancer characteristics of patients who received routine home health care compared to those that did not. We then examined the association between routine home care and short-term post-operative outcomes. Since patients that received routine home health may have had other clinical needs driving the decision, we also created a propensity score for the receipt of home health based on age and number of Charlson co-morbidities. We then fit a logistic regression model using the propensity score while also controlling for other patient (age, body mass index, race, and Charlson co-morbidity score) and cancer (t stage) characteristics. Fourth, we assessed the reasons for ED visits and readmission using claims data. Fifth, we performed univariate statistical analyses to compare 90-day episode payments, index admission, professional services, readmissions, and home health payments and other post-acute care services.

Finally, we performed several sensitivity analyses to examine the robustness of our findings. In order to examine the impact of our definition of “routine” home care, we repeated our analyses after modifying the definition of routine home care to home health services provided within 2 days of discharge or within 7 days of discharge (instead of the 4 days used in the primary analysis). To evaluate whether or not we were capturing patients who were discharged with home care for directed reasons (rather than routine reasons), we further limited our analyses to patient length of stay less than or equal to 2 days.

All analyses were performed using SAS v9.4 (NC), at the 5% significance level. The University of Michigan institutional review board deemed this study exempt from review.

RESULTS

We identified 705 men from 33 MUSIC practices who underwent a prostatectomy and could be matched between claims and clinical registry data. Among this group, 647 (92%) met our inclusion criteria and 84 patients (13%) received routine home care. Patient and cancer characteristics were similar for patients that received or did not receive routine home care (Table 1).

Table 1.

Patient Demographics

	Home Health	No Home Health	p-value
No. pts (%)	84 (13.0)	563 (87.0)
Median age	63.2	62.0	0.14
Race (%)			0.64
White	85.2	82.7
Black	7.4	10.8
Other	7.4	6.6
BMI			0.63
<25	13.4	15.9
25–30	51.2	44.5
30–35	29.3	30.6
≥35	6.1	9.1
No Charlson comorbidity (%)			0.95
0	0.0	0.2
1	4.8	5.5
2	32.1	30.2
3+	63.1	64.1
Gleason score			0.45
≤6	23.8	26.3
7	57.1	59.8
8 through 10	19.1	13.9
Pathologic T stage (%)			0.64
T1	0.0	0.4
T2	66.7	70.5
T3	33.3	29.1

Figure 1 demonstrates the use of routine home health care after prostatectomy, by institution. The use of routine home health care after prostatectomy varied from 0–53% across MUSIC practices (p=0.05). Unadjusted, patients with routine home care had increased rates of ED visits (15.5% vs. 6.9%, p<0.01) and a trend toward decreased readmissions within 16 days (0 vs. 4.1%, p=0.06). Catheter use longer than 16 days (3.6 vs. 3.0%, p=0.79) and need for catheter replacement (1.2% vs. 2.5%, p=0.46) were similar between those receiving and not receiving home health (Figure 2). Only the differences in ED visits (p<0.001) remained after controlling for the propensity to receive home health and other patient and cancer characteristics in multivariable logistic regression analyses (Supplementary Table 1).

Figure 1.

Proportion of patients receiving routine home health, by practice

Figure 2.

Unadjusted short-term clinical outcomes after radical prostatectomy

* p>0.05

The most common reason for ED visits and readmissions are presented in Supplementary Table 2. Catheter-related complications were the second most frequent reason for ED visits amongst patients with home health (2.4%) compared to it being third for patients without routine home health (1.1%). The most prevalent reasons for readmissions amongst patients without routine home health were gastrointestinal distress (0.89%) and pulmonary embolism (0.71%).

Although index hospitalizations payments were similar between patients who received home health care and those who didn’t ($9,403 vs $9,193, p=0.172), average total episode payments were significantly different ($15,556 vs $13,788, p<0.001). The difference was mostly driven by post-acute care spending ($1,862 for patients with routine home health compared to $641 without it, p<0.001). Home health cost an average of approximately $1,000 per episode and was the predominant driver of the cost-differential between the groups (Figure 3).

Figure 3.

90-day total episode and component payments following radical prostatectomy

* p>0.05

Our sensitivity analyses identified no substantive changes to our principal findings.

DISCUSSION

In this study, there was considerable variation in the receipt of routine home health care after prostatectomy. After controlling for patient and cancer characteristics, patients with home health care had increased rate of visits to the emergency department and no differences in readmissions, prolonged catheter duration or requirement for a catheter replacement. Home health care cost $1,000 per episode and was the predominant driver in post-acute care spending. Collectively, these findings suggest that routine home care adds cost without improving short-term postoperative quality, suggesting that careful patient selection for home health care is necessary to improve health care value.

Few have evaluated the impact of routine home health care on outcomes following radical prostatectomy. An evaluation of ED visits after prostatectomy at a tertiary care hospital in Canada, demonstrated a 24% ED visit rate after prostatectomy, with infrequent readmission (4.7%). In the Canadian analysis, 18% of patients who returned to the ED presented for “genitourinary device complications” and 16% for urinary tract infections. The evaluators therefore developed a pilot program where a nurse met with the patient prior to discharge and served as a resource for questions via telephone in the early post-operative period. The nurse was also available for catheter removal. This demonstration program resulted in a relative reduction of 19% for ED visits and 2.1% for readmissions⁷. Separately, home health care is frequently used after total joint arthroplasty. A recent study demonstrated that when home care was not used, patients had similar 30-day complication rates, and patient satisfaction remained high⁸.

Our analysis is unique because it evaluates home health care use in multiple practice types across Michigan, using both claims data and detailed clinical registry data. Using robust clinical registry data combined with administrative claims, we are able to precisely describe variation in home health use and its effects on short-term post-surgical outcomes and costs. We previously reported on our novel NOTES clinical registry that reliably tracks clinically relevant preoperative complication rates⁵. Claims data can supplement registry data by identifying claims for emergency department visits or readmissions at remote facilities, unbeknownst to the primary surgeon.

There are several potential reasons that the use of “routine” home care after prostatectomy was not associated with lower ED visits, readmissions or catheter-related complications. First, the most common complications following laparoscopic radical prostatectomy are anastomotic leaks requiring prolonged catheterization, post-operative hemorrhage, bladder neck contracture, and urinary retention^9,10. Home care can do little to prevent these complications from occurring. There is also little evidence that home care can prevent the most common reasons for readmission following radical prostatectomy: thromboembolic (13.6%), wound (12.2%), renal/genitourinary (11.8%) and gastrointestinal (11.8%) complications¹¹. If the most common reasons for ED visits following prostatectomy were related to catheter care, including clogged catheters, need for replacement catheter, inability to empty a urine drainage bag, or even catheter-related urinary tract infections, then it would be conceivable that home care may improve outcomes. However, in our cohort the most common reason for ED visits among those who did not receive home health were gastrointestinal (GI) distress with comparable rates of urinary retention/obstruction/dysuria to those who received home health. In fact, those who received home health had a higher share of catheter-related visits (2.4% vs. 1.1%) (Supplementary Table 2).

Our results demonstrated an increase in ED visits associated with “routine” home care. It is unclear why this occurred. Home health nurses may identify problems early, lending to early referrals to the ED, where a complication may be easily managed as an outpatient and the patient discharged home. In contrast, perhaps, home care leads to overly sensitive identification of issues that may not have needed an ED visit. An evaluation of home health after pancreatectomy revealed an increased risk of readmissions with home health for non-severe conditions, but not severe conditions³. The authors speculated that this increased risk could be secondary to reflex admissions by surgeons when a home health provider notifies the surgeon reflecting either the surgeon’s unfamiliarity with home health staff or concern for medicolegal implications. Regardless of the exact mechanism, our study illuminates the relationship between routine home health care and drivers of variation and 90-day payments in prostatectomy episodes in a diverse set of Michigan private and academic practices.

Our study has several limitations. First, we only analyzed hospitals in Michigan, thus the results may not be generalizable to the entire United States. However, our cohort includes differing practice types (academic/non-academic) in an array of geographic locations (urban/rural) within Michigan and thus captures real-world experience using home health care. Second, we defined routine home health care by home care starting within four days of discharge in patients without a drain. However, the exact reason that home care was initiated is not available and these patients may have received home care for non-routine reasons that we were unable to capture (i.e., lack of social support, patient request, etc). While this definition of home care may over-estimate “routine” health care and add selection bias to our cohort definition, patient and cancer characteristics were similar between the cohorts and our sensitivity analyses with varying definitions of routine home care did not affect our results. Third, it is unclear whether the patients with home health returned to the ED in response to advice from the home care provider or if they would have returned to the ED even without home care. We are also unable to identify whether a complication was identified and managed definitively in the ED among home care patients and thus prevented a readmission. Fourth, we were unable to measure factors such as patient satisfaction and decreased anxiety that could be a benefit of routine home health care. Fifth, we have identified if any home care visit took place and are not able to evaluate the number and/or quality of visits. Higher quality home care providers could potentially provide better outcomes than lower-quality home care providers. Sixth, we have only identified if any home care visit took place and are not able to evaluate the number and/or quality of home care visits. It is conceivable that higher quality home care providers would provide better outcomes than lower-quality home care providers. However, we are evaluating the use of home care in a community setting, not the use of specific home care practices or agencies. Last, we were not able to measure physician practice characteristics, such as the availability of an in-office nurse or multiple physicians that are available to field patient questions, that may also lead to differences in short-term outcomes after radical prostatectomy.

These limitations notwithstanding, our findings have important implications for physicians, patients, and hospital administrators. For physicians, our results demonstrate that the need for home health care should be determined on a case-by-case basis and its routine use does not improve outcomes for all patients. Our findings can also provide further reassurance to patients. Although patients may initially be concerned about providing their own routine catheter maintenance, our results demonstrate that routine home care does not provide measurable benefits in terms of decreased readmissions, ED visits, or catheter-related complications. Our analyses further suggest that hospital administrators should evaluate alternate programs to decrease readmissions, ED visits and catheter-related complications after a prostatectomy and formalize catheter care teaching during the hospital stay rather than relying on home health care to provide that teaching.

Collectively, our findings suggest that there is great variation in the use of routine home care after radical prostatectomy, which does not improve outcomes, but increases costs. Thus, home health care should not be routinely prescribed for all patients after radical prostatectomy, but should instead be used selectively to increase value. Moving forward, further research should explore which patients may benefit most from home health care services, if there are any physician or practice characteristics (i.e., those in more rural areas, smaller practices, higher complication or ED visit rates) whose patients may benefit from home health care, and if there are certain attributes of the home care visit or agency that improves outcomes. Additionally, it will be important to understand patients’ perceptions of routine home care and how it impacts their perceived post-operative.