Abstract
BACKGROUND:
Prospective payment models have incentivized reductions in length of stay after surgery. The benefits of abbreviated postoperative hospitalization could be undermined by increased readmissions or post-acute care use, particularly for older adults or those with comorbid conditions.
OBJECTIVE:
To determine whether hospitals with accelerated postsurgical discharge accrue total episode savings, or incur greater post-discharge payments, among patients stratified by age and comorbidity.
DESIGN:
Retrospective cross-sectional study.
SETTING:
National data from the 100% Medicare Provider Analysis and Review files for July 2012-June 2015.
PATIENTS:
We included Medicare beneficiaries undergoing elective colectomy and stratified the cohort by age (65–69 years, 70–79, ≥80) and Elixhauser comorbidity score (low: ≤0; medium: 1–5; and high: >5). Patients were categorized by the hospital’s mode length of stay, reflecting “usual” care.
MAIN OUTCOMES MEASURES:
In a multilevel model, we compared mean total episode payments and components thereof among age and comorbidity categories, stratified by hospital mode length of stay.
RESULTS:
Among 88,860 patients, mean total episode payments were lower in shortest vs. longest length of stay hospitals across all age and comorbidity strata, and were similar between age groups (65–69: $28,951 vs $30,566, p=0.014; 70–79: $31,157 vs $32,044, p=0.073; ≥ 80 $33,779 vs. $35,771, p=0.005), but greater among higher comorbidity (low: $23,107 vs $24,894, p=0.001; medium $30,809 vs $32,282, p=0.038; high: $44,097 vs $46641, p<0.001). Post-discharge payments were similar among length of stay hospitals by age (65–69 years: Δ$529; 70–79 years: Δ$291; ≥80 years: Δ$872, p=0.25), but greater among high comorbidity (low Δ$477, medium Δ$480, high Δ$1,059, p=0.02).
LIMITATIONS:
Administrative data does not capture patient-level factors that influence post-acute care use (preference, caregiver availability).
CONCLUSIONS:
Hospitals achieving shortest length of stay after surgery accrue lower total episode payments without compensatory increase in post-acute care spending, even among patients at oldest age and with greatest comorbidity.
Keywords: Comorbidities, Payments, Post-acute care, Readmissions
CONSECUENCIAS DE LA EDAD Y LAS COMORBILIDADES ASOCIADAS, EN EL COSTO DE LA ATENCIÓN EN PACIENTES SOMETIDOS A COLECTOMÍA EN PROGRAMAS DE ALTA POSOPERATORIA “ACELERADA”
ANTECEDENTES:
Los modelos de pago prospectivo, han sido un incentivo para reducir la estancia hospitalaria después de la cirugía. Los beneficios de una hospitalización posoperatoria “abreviada” podrían verse afectados por un aumento en los reingresos o en la necesidad de cuidados postoperatorios tempranos luego del periodo agudo, particularmente en los adultos mayores o en aquellos con comorbilidades.
OBJETIVO:
Determinar si los hospitales que han establecido protocolos de alta posoperatoria “acelerada” generan un ahorro en cada episodio de atención o incurren en mayores gastos después del alta, entre los pacientes estratificados por edad y por comorbilidades.
DISEÑO:
Estudio transversal retrospectivo.
AJUSTE:
Revisión a partir de la base de datos nacional del 100% de los archivos del Medicare Provider Analysis and Review desde julio de 2012 hasta junio de 2015.
PACIENTES:
Se incluye a los beneficiarios de Medicare a quienes se les practicó una colectomía electiva. La cohorte se estratificó por edad (65–69 años, 70–79, ≥80) y por la puntuación de comorbilidad de Elixhauser (baja: ≤0; media: 1–5; y alta:> 5). Los pacientes se categorizaron de acuerdo con la modalidad de la duración de la estancia hospitalaria del hospital, lo que representa lo que se considera es una atención usual para dicho centro.
PRINCIPALES MEDIDAS DE RESULTADO:
En un modelo multinivel, comparamos la media de los pagos por episodio y los componentes de los mismos, entre las categorías de edad y comorbilidad, estratificados por la modalidad de la duración de la estancia hospitalaria.
RESULTADOS:
En los 88,860 pacientes, los pagos promedio por episodio fueron menores en los hospitales con una modalidad de estancia más corta frente a los de mayor duración, en todos los estratos de edad y comorbilidad, y fueron similares entre los grupos de edad (65–69: $ 28,951 vs $ 30,566, p = 0.014; 70–79 : $ 31,157 vs $ 32,044, p = 0.073; ≥ 80 $ 33,779 vs $ 35,771, p = 0.005), pero mayor entre los pacientes con comorbilidades más altas (baja: $ 23,107 vs $ 24,894, p = 0.001; media $ 30,809 vs $ 32,282, p = 0.038; alta: $ 44,097 vs $ 46641, p <0,001). Los pagos generados luego del alta hospitalaria fueron similares con relación a la estancia hospitalaria de los diferentes hospitales con respecto a la edad (65–69 años: Δ $ 529; 70–79 años: Δ $ 291; ≥80 años: Δ $ 872, p = 0,25), pero mayores en aquellos con más alta comorbilidad (baja Δ $ 477, medio Δ $ 480, alto Δ $ 1,059, p = 0,02).
LIMITACIONES:
Las bases de datos administrativas no capturan los factores del paciente que influyen en el cuidado luego del estado posoperatorio agudo (preferencia, disponibilidad del proveedor del cuidado).
CONCLUSIONES:
Los hospitales que logran una estancia hospitalaria más corta después de la cirugía, acumulan pagos más bajos por episodio, sin un incremento compensatorio del gasto en la atención pos-aguda, incluso entre pacientes de mayor edad y con mayor comorbilidad.
INTRODUCTION
Prospective payment for surgical hospitalization has incentivized efforts to reduce length of stay after inpatient surgery. For colorectal surgery specifically, enhanced recovery protocols or “fast track” postoperative pathways have accelerated postoperative recovery milestones through laparoscopic surgical approaches, early ambulation, early feeding, and minimization of opioids.1 Evidence has shown that enhanced recovery protocols are not associated with increased complications or readmissions in older adults and those with multimorbidity.2–5 Minimizing harmful hospital-associated exposures including in-hospital immobility, iatrogenic illnesses, and nosocomial infections may also motivate short hospital stay for older adults.6–8 Longer hospital stays may also place older adult patients at risk for hospital-associated disability and delirium.9,10
The benefits of abbreviated postoperative hospitalization could have unintended consequences, however, if they result in more readmissions, or avoidable use of expensive post-acute care. Conversely, lengthening hospitalization has the potential to offset the costs of unplanned readmissions or avert the need for post-acute care.11 New episode-based payment models, such as Medicare’s Bundled Payments for Care Improvement-Advanced (BPCI-A) hold hospitals accountable for the costs of readmissions and post-acute care use within 90 days of discharge. We have previously found that hospitals achieving shortest lengths of stay after inpatient surgery did not incur increased post-discharge spending or readmissions.12 However, older adults and patients with multiple comorbid conditions may have greater hospital and post discharge care needs, including readmissions, than their younger and healthier counterparts, and thus, it is not clear whether these findings may extend to the most vulnerable patients.13 As the age and medical complexity of patients undergoing major surgery continues to increase, an understanding of optimal perioperative care pathways for older patients and those with multiple comorbidities will be essential for hospitals subjected to bundled payments and other episode-based payment incentives.
To evaluate the effect of accelerated discharge pathways after inpatient surgery, we analyzed surgical episode expenditures among older adults and patients with multiple comorbid conditions undergoing elective colectomy, stratified by their hospital’s typical postoperative length of stay. Specifically, we hypothesized that adverse consequences of accelerated discharge, if present, would be represented by no difference in total episode payment by length of stay. We hypothesized that cost savings from a shorter inpatient stay would be offset by a compensatory increase in post-discharge payments outweighing the financial benefit of shorter length of stay. We focused on elective colectomy because it is a common procedure in older adults, a focus of initiatives to hasten recovery and shorten LOS, has high readmission rate and post-discharge care use, and is featured in BPCI-A.
METHODS
This study was based on 100% fee-for-service Medicare claims data for patients ages 65–99 who underwent elective colectomy from July 1, 2012 until June 30, 2015. Patients were identified from the Medicare Provider Analysis and Review, Part B, Outpatient, and Home Health Agency files according to the presence of appropriate colectomy procedure codes from the International Classification of Diseases, version 9 (17.3x, 45.7x, or 45.81–45.83). We selected colectomy because it is a target of enhanced recovery protocols focused on decreasing length of stay, which could have variable impact on patient subpopulations. Colectomy also has the potential for considerable post-operative morbidity. We excluded emergent and synchronous operations (e.g. liver resection for colorectal cancer metastases), claims for services provided to Medicare managed care patients, and those not enrolled in both Medicare parts A and B at time of their index procedure, as in previous work.12,14–16 Additionally, we excluded hospitals that performed fewer than 10 of the included operations per year. Hospital characteristics were obtained from the American Hospital Association Annual Survey.
Patient Categorization
Patients were analyzed both by age and health status. We stratified patients into three age categories: 65–69, 70–79, and ≥80 years. To classify health status, we used Elixhauser comorbidity scores and divided patients into terciles.17 Patients with scores less than or equal to 0 were categorized as “low,” those with scores 1–5 were categorized as “medium,” and those with scores greater than 5 were categorized as “high.”
Hospital Length of Stay
Postoperative length of stay was defined as the number of days from surgery to discharge during the index hospitalization. To quantify institutions’ most typical practice pattern, each hospital was classified according to its integer mode postoperative length of stay for patients following elective colectomy, as in previous work.12 Mode length of stay represents the most common discharge pattern and therefore likely the intended postoperative pathway.18 We stratified hospitals in four approximate quartile groups by mode integer length of stay: ≤3 days, 4 days, 5 days, or ≥6 days.
As a sensitivity analysis, we restricted the patient cohort to those discharged within 1 day of modal LOS, to focus on patients most likely to have experienced the typical care pathways, rather than conflating the effects of complications and other deviations from routine recovery.
Outcomes
The primary outcome measured was the risk adjusted, price-standardized Medicare payments for entire 90-day surgical episodes from the date of discharge, as developed by researchers for the Dartmouth Atlas of Healthcare19 and used in our previous work.12 All payments were calculated according to the 2015 Medicare payment schedule and were inflation adjusted to 2015 US dollars. Payment data were calculated at the hospital level and included total episode payments, index hospitalization, post-acute care, readmissions, and professional fees. Index hospitalization included DRG payment plus outlier payments when present. All readmissions were included when initiated within 90 days of discharge. Post-acute care payments were calculated within the 90-day window based on category: skilled nursing facility payments were calculated using per diem payments and home health care and rehabilitation hospital payments were pro-rated for the period.
In the analysis, we included the following index hospitalization complications: pulmonary failure, pneumonia, myocardial infarction, deep venous thrombosis/pulmonary embolism, acute renal failure, postoperative hemorrhage, surgical site infection, and gastrointestinal bleeding. The aforementioned complications are identified in administrative data with acceptable reliability.20,21
Statistical Analysis
We compared characteristics of patients and hospitals in each of the LOS categories using analysis of variance/F-test for continuous variables and chi-square tests for categorical values.
In all statistical models related to episode payments, adjustments were made for patient demographics and expenditures occurring in six months prior to the index operation as a means to capture the baseline expenditures and care utilization, with the exclusion of age and comorbidity when analyzing age and comorbidity based analyses respectively (eTable 1). The model was also adjusted for procedure complexity classified as segmental colectomy without stoma vs. total colectomy or colectomy with stoma vs. abdominal perineal resection. We used interaction terms to determine the effects of age or comorbidity groups and LOS on payments. The model also adjusted for hospital-level characteristics, including profit status (non-profit vs. for-profit), location (urban vs. rural), teaching status (academic vs. non-academic hospital), bedsize (4 categories: <200, 200–349, 350–499, ≥500 beds), annual procedure volume (continuous variable), and percent Medicaid patient-days (continuous variable).
Using a multilevel (patients clustered in hospitals) mixed-effects linear regression, we compared total episode and individual components of episode expenditures for patients in each of the three age categories (65–69, 70–79, and ≥80 years) and three health status categories (low, medium, high comorbidity), by hospital mode length of stay. These adjusted means were compared utilizing t-tests if comparing two means and analysis of variance/F-test for comparisons of three or more groups. We subsequently repeated this analysis for comparison of post-acute care, readmission, and professional fee payments.
Sensitivity Analysis
As a sensitivity analysis, we restricted the cohort to those discharged within 1 day of hospital mode length of stay, as in our previous work, and repeated multilevel linear regression for all payments.12 This analysis focuses on a homogeneous typical care group, minimizing outsized influence of patients with serious complications and outlier spending.
This study was deemed exempt by the University of Michigan Institutional Review Board.
RESULTS
Patients, Hospitals, and LOS
During the study period, 88,860 patients met inclusion criteria. The 1090 hospitals were stratified into four groups by mode LOS, ≤3 days (281 hospitals, 25.8%), 4 days (419 hospitals, 38.4%), 5 days (257 hospitals, 23.6%) and ≥6 days (133 hospitals, 12.2%). Hospital LOS groups differed by procedure volume, but did not differ by profit status, urban location, teaching status, number of beds, or percent Medicaid patient-days (Table 1).
Table 1:
Hospital Characteristics, by mode length of stay
| Hospital Characteristics | ≤3 days | 4 days | 5 days | ≥6 days | p-value |
|---|---|---|---|---|---|
| Number of hospitals | 281 | 419 | 257 | 133 | |
| Non-profit | 232 (83.2%) | 334 (80.1%) | 192 (74.7%) | 103 (77.4%) | 0.35 |
| Urban | 278 (98.9%) | 417 (99.5%) | 256 (99.6%) | 133 (100%) | 0.52 |
| Teaching hospital | 199 (70.8%) | 314 (74.9%) | 177 (68.9%) | 86 (64.7%) | 0.10 |
| Beds <200 | 70 (25.1%) | 85 (20.4%) | 63 (24.5%) | 39 (29.3%) | 0.07 |
| 200–349 | 100 (35.8%) | 136 (32.6%) | 90 (35.0%) | 53 (39.9%) | |
| 350–499 | 52 (18.6%) | 91 (21.8%) | 50 (19.5%) | 26 (19.6%) | |
| ≥500 | 57 (20.4%) | 105 (25.2) | 54 (21.0%) | 15 (11.3%) | |
| Annual Procedure Volume (mean, SD) | 24.7 (14.5) | 24.3 (14.4) | 22.4 (14.8) | 18.8 (10.4) | <0.001 |
| % Medicaid patient-days (%) | 18.3 | 19.7 | 18.4 | 19.5 | 0.12 |
Patients were then stratified by age: 65–69, 70–79, and ≥80 years, with comparisons between these groups summarized in eTable 2. The oldest group of patients had higher Elixhauser scores, were more likely to be female, more likely to be white, had a higher proportion of postoperative complications, and had higher readmission rates.
Patients were also stratified by Elixhauser score (low, medium, high) summarized in eTable 3. In the entire cohort, the mean Elixhauser score was 4.0 (SD 6.6), and median was 1 (IQR 0–7). The highest comorbidity group was older, more likely to be male, with a higher proportion of Black patients, had a higher proportion of postoperative complications and readmissions.
When stratified by LOS, the patients at highest (≥6 days) LOS hospitals had higher Elixhauser comorbidity scores, higher rates of postoperative complications, and were more likely to be in the high comorbidity group and oldest age group (≥80 years) (Table 2). Readmission rates were similar, but highest among the 5 day LOS group (12.5% vs 13.4%, 15.5%, and 17.9% for ≤3 days, 4 days, and ≥6 days LOS, respectively). A laparoscopic surgical approach was more common in lower LOS hospitals.
Table 2:
Patient characteristics, by hospital mode length of stay.
| Patient and Clinical Characteristics | ≤3 days | 4 days | 5 days | ≥6 days | p-value |
|---|---|---|---|---|---|
| Patients in Total Cohort, N | 24321 | 35610 | 20182 | 8747 | |
| Patients discharged within 1 day of hospital mode LOS (n, % of total cohort) | 12825 (52.7%) | 19358 (54.4%) | 10671 (52.9%) | 4194 (48.0%) | <0.001 |
| Age at Index Date, mean (SD) | 74 | 74.3 | 74.6 | 75.4 | <0.001 |
| Age Group: 65–69 years | 3777 (29.5%) | 5447 (28.1%) | 2872 (26.9%) | 944 (22.5%) | <0.001 |
| 70–79 years | 6456 (50.3%) | 9664 (49.9%) | 5313 (49.8%) | 2105 (50.2%) | |
| ≥80 years | 2592 (20.2%) | 4247 (21.9%) | 2486 (23.3%) | 1145 (27.3%) | |
| Health Group: Healthy | 8038 (62.7%) | 11276 (58.2%) | 5751 (53.9%) | 1997 (47.6%) | <0.001 |
| Medium | 2380 (18.6%) | 3981 (20.6%) | 2177 (20.4%) | 994 (23.7%) | |
| Sick | 2407 (18.8%) | 4101 (21.2%) | 2743 (25.7%) | 1203 (28.7%) | |
| Male, n (%) | 5361 (41.8%) | 8041 (41.5%) | 4408 (41.3%) | 1685 (40.2%) | 0.31 |
| Race: White | 11779 (91.8%) | 17409 (89.9%) | 9671 (90.6%) | 3799 (90.6%) | <0.001 |
| Black | 676 (5.3%) | 1171 (6.0%) | 688 (6.4%) | 295 (7.0%) | |
| Asian | 77 (0.6%) | 210 (1.1%) | 71 (0.7%) | 29 (0.7%) | |
| Hispanic | 75 (0.6%) | 135 (0.7%) | 55 (0.5%) | 17 (0.4%) | |
| Other | 218 (1.7%) | 433 (2.2%) | 186 (1.7%_ | 54 (1.3%) | |
| Number of Comorbidities, mean (SD) | 2.04 (1.53) | 2.12 (1.55) | 2.33 (1.65) | 2.62 (1.74) | <0.001 |
| Elixhauser Comorbidity score, mean (SD) | 3.68 (6.45) | 3.98 (6.51) | 4.26 (6.74) | 4.41 (6.75) | <0.001 |
| Comorbidities: | |||||
| Congestive heart failure | 434 (3.4%) | 817 (4.2%) | 544 (5.1%) | 268 (6.4%) | <0.001 |
| Hypertension | 8360 (65.2%) | 12655 (65.4%) | 7262 (68.1%) | 2972 (70.9%) | <0.001 |
| Chronic pulmonary disease | 1911 (14.9%) | 2984 (15.4%) | 1654 (15.5%) | 772 (18.4%) | <0.001 |
| Diabetes | 2625 (20.5%) | 4407 (22.8%) | 2547 (23.9%) | 1111 (26.5%) | <0.001 |
| Renal failure | 668 (5.2%) | 1221 (6.3%) | 734 (6.9%) | 355 (8.5%) | <0.001 |
| Liver disease | 196 (1.5%) | 283 (1.5%) | 160 (1.5%) | 54 (1.3%) | 0.72 |
| Solid tumor w/out metastasis | 330 (2.6%) | 510 (2.6%) | 318 (3.0%) | 132 (3.1%) | 0.07 |
| Coagulopathy | 175 (1.4%) | 296 (1.5%) | 217 (2.0%) | 93 (2.2%) | <0.001 |
| Fluid and electrolyte disorders | 1054 (8.2%) | 2052 (10.6%) | 1426 (13.4%) | 844 (20.1%) | <0.001 |
| Anemias | 2207 (17.2%) | 3569 (18.4%) | 2273 (21.3%) | 1116 (26.6%) | <0.001 |
| Laparoscopic Surgery | 9192 (71.7%) | 11652 (60.2%) | 4786 (44.9%) | 1316 (31.4%) | <0.001 |
| Postoperative Complications | |||||
| Pulmonary failure | 55 (0.4%) | 119 (0.6%) | 64 (0.6%) | 55 (1.3%) | <0.001 |
| Pneumonia | 12 (0.1%) | 24 (0.1%) | 17 (0.2%) | 20 (0.5%) | <0.001 |
| Postoperative myocardial infarction | 18 (0.1%) | 40 (0.2%) | 34 (0.3%) | 15 (0.4%) | 0.008 |
| Deep vein thrombosis or pulmonary embolism | 13 (0.1%) | 31 (0.2%) | 16 (0.1%) | 14 (0.3%) | 0.012 |
| Acute renal failure | 215 (1.7%) | 460 (2.4%) | 280 (2.6%) | 204 (4.9%) | <0.001 |
| Hemorrhage | 127 (1.0%) | 191 (1.0%) | 133 (1.2%) | 74 (1.8%) | <0.001 |
| Surgical site infection | 34 (0.3%) | 75 (0.4%) | 61 (0.6%) | 51 (1.2%) | <0.001 |
| Gastrointestinal tract hemorrhage | 44 (0.3%) | 88 (0.5%) | 57 (0.5%) | 32 (0.8%) | 0.004 |
| Any postoperative complication | 482 (3.8%) | 938 (4.8%) | 612 (5.7%) | 430 (10.3%) | <0.001 |
| Readmission | 1599 (12.5%) | 2589 (13.4%) | 1652 (15.5%) | 749 (17.9%) | <0.001 |
Note: DVT = deep vein thrombosis; PE = pulmonary embolism
Total Episode Payments
Mean total episode payments were lower in shortest LOS vs. longest LOS hospitals in all age categories, although the difference was not statistically significant among 70–79 year olds (65–69: $28,951 vs $30,566, p=0.014; 70–79: $31,157 vs $32,044, p=0.073; ≥ 80 $33,779 vs. $35,771, p=0.005). When comparing total episode payments for each LOS category, there was no statistically significant difference among the age groups (Figure 1a).
Figure 1:
Variation in payments, by age group (a-e) or comorbidity group (f-j) and hospital mode length of stay. a,f: Total episode payments. b,g: Index payments. c,h: Post-discharge payments. d,i: Readmission payments. e,j: Professional fees. * p<0.001, ** p<0.01, ***p<0.05. †Comparisons of the payment difference (Δ) between mode ≥6 and ≤3 day length of stay hospitals between highest and lowest comorbidity group with p value =0.02.
Similarly, for patients with greatest comorbidity, mean total episode payments were lower in shortest vs longest LOS hospitals in all 3 categories (low comorbidity: $23,107 vs $24,894, p=0.001; medium comorbidity $30,809 vs $32,282, p=0.038; high comorbidity: $44,097 vs $46641, p<0.001) (Figure 1f). When comparing total episode payments for each LOS category, there was no statistically significant difference among the comorbidity groups.
Index Hospitalization Payments
As anticipated, index payments were lower for shorter LOS hospitals. Index payments exhibited stepwise increases by each additional day LOS for all age groups and were similar between the age groups, with the difference between longest LOS vs shortest LOS hospitals similar between age groups (65–69: Δ$1,025; 70–79: Δ$561; ≥80: $729; p=0.52) (Fig 1b). In the comorbidity groups, again there was stepwise increase for each additional day LOS, with similar differences between shortest LOS vs. longest LOS hospitals between all comorbidity groups (low Δ$1,059; medium Δ$891; high Δ$1268; p=0.61). Index payments for all LOS categories, however, increased as comorbidity group complexity increased (Fig 1g).
Post-Discharge Payments
The oldest patients had greater post-discharge care expenditures than youngest patients for all hospital mode LOS groups. Shorter LOS had the lowest post discharge payments for all age groups. By age group, there was little difference between the post-discharge payments for the shortest and longest usual LOS hospitals (payment for ≥6 days minus payment for ≤3 days: 65–69 years: Δ$529; 70–79 years: Δ$291; ≥80 years: Δ$872; p=0.25, Figure 1c).
The highest comorbidity group also had greater post-discharge care payments than the medium comorbidity group, which was higher than the lowest comorbidity group. Again, shorter LOS was not associated with higher post-discharge payments, rather payments increased progressively as LOS increased. The difference between post-discharge payments from LOS ≥6 days minus payment for ≤3 days was greatest in the high comorbidity group (Δ$1,059), and lower in the medium group (Δ$480) and low group (Δ$477), p=0.02 (Figure 1h). The patients with high numbers of comorbid conditions in hospitals with the longest usual LOS had the highest post-discharge payments.
Readmission Payments
Readmission payments were similar by age group and LOS. (Figure 1d). In the comorbidity groups, readmission payments increased with comorbidity, but there were no differences by LOS. (Figure 1i).
Professional Fees
Professional fees were similar by age group and LOS (Fig 1e).The highest comorbidity group had greater professional payments than the low and medium comorbidity groups. These did not vary by LOS (Fig 1j).
Sensitivity Analysis
For the sensitivity analysis, restricting to patients discharged within 1 day of the hospital’s mode LOS, 47,048 (53.0%) patients remained in the cohort. The proportion retained by LOS group ranged from 48.0 to 54.4%. The relationships were maintained, and differences between length of stay groups became more distinct: both by age group and comorbidity group total payments were lowest for LOS ≤3 days, with stepwise increases with each day LOS. Differences in mean total episode payments between hospitals with mode LOS ≤3 days and ≥6 days were statistically significant (p<0.001) for all age groups and all comorbidity groups. Post-discharge payments showed the same pattern, lowest for LOS ≤3 days, with stepwise increases with each day LOS. Differences in mean post-discharge payments between hospitals with mode ≤3 days and ≥6 days were statistically significant (p<0.001) for all age groups and all comorbidity groups. For readmissions, again there was stepwise increase in payments by age and comorbidity groups, with statistically significant differences between LOS groups for all age groups and for low and medium comorbidity groups (although not for the high comorbidity group).
DISCUSSION
We found that shorter length of stay is associated with lower total episode payments without compensatory increase in post-acute care payments across the age and comorbidity spectrum among Medicare beneficiaries. Regardless of age, post-acute care payments were lower in shorter LOS hospitals. Patients with highest comorbidity burden, however, consistently demonstrated higher total episode payments and post-acute care payments at all lengths of stay, compared to patients with lower comorbidity burden. There were no differences in readmission payments by length of stay across the age and comorbidity groups. While we similarly found cost savings after early post-surgical hospital discharge in a previous study of a Medicare population and three surgical procedures, this work adds to those findings by establishing that the efficiencies hold true even for those with most advanced age or comorbid conditions, Medicare patients accrue cost savings with expedient discharge after colectomy.
Our findings support previous studies associating shorter post-operative lengths of stay with lower total episode payments.12,22 We find that hospitals with shorter routine lengths of stay do not achieve accelerated discharge through greater use of post-acute care.23,24 We also find that although the oldest patients and those with higher comorbidity burden have the highest post-acute care payments, there is no evidence of compensatory increase in use among hospitals with shortest LOS. Low LOS hospitals had greater procedure volume and greater use of laparoscopy, both of which may be markers of quality that permit such efficiency of care, while lower utilization of laparoscopy in high LOS hospitals may reflect a patient population at higher operative risk or differences in provider experience.25
Independent of hospital quality, for patients that will have high post-acute care utilization, prolonging LOS in order to decrease post-acute care spending may not achieve that goal. In the oldest age group, ≥80 years, readmission payments were similar across all LOS groups, so keeping an older adult in the hospital longer may not avoid readmission payments, and earlier discharge is not associated with increased readmission payments. Instead, efforts to decrease LOS in skilled nursing facilities may be an alternative mechanism to decrease post-acute care spending.26 Additionally, shorter lengths of stay may minimize some of the deleterious effects of medical care on older adults, particularly hospitalization-associated disability.9 We also found that patients with higher comorbidity burden consistently accounted for the highest payments across payment types: index, post-discharge spending, readmissions, and professional fees. The finding of a consistent relationship between comorbidities and all components of episode spending lends further support to the need for rigorous risk adjustment in episode-based reimbursement programs. Without a detailed accounting for comorbidities, hospitals with higher average case complexity will be more likely to incur reimbursement penalties, even if they are able to achieve efficient hospitalization and post-discharge utilization for their patients. Thus, bundled payment programs will need to be designed with patient-specific metrics of complexity in mind, as has been shown in previous studies of episode-based incentives around both colectomy and joint replacement.27,28 While we show that patients with more comorbid conditions incur higher payments overall, a challenge in stratifying payments moving forward will be differentiating between expected higher payments and those higher payments reflecting lower quality care.
This study has several limitations. First, this study was conducted using administrative data, therefore limiting our ability to capture individual patient complexity that may significantly alter a post-operative length of stay. Using individual patient comorbidity burden, we attempted to account for patient complexity, but are limited by variables captured in the Medicare administrative datasets. Although there were differences in comorbidity and age among the mode length of stay groups, the magnitude of these differences was not sufficient to account for an additional two- or three-day difference in length of stay. We also adjusted for comorbidity and age in our payment models. Diagnosis related group (DRG) status also drives Medicare charges and payments, and we are unable to evaluate whether DRG was assigned to a patient pre-operatively due to comorbid conditions, or post-operatively due to complications. Additionally, administrative data does not capture the patient-level factors associated with post-acute care use, including patient preferences for discharge disposition29 and availability of caregiver and social support, which may influence the lengths of stay and discharge destination. Physician-level factors including preferences and biases also influence discharge disposition, another factor we cannot control for in administrative data. We utilized a multilevel model and the mode length of stay for each hospital included in this study in order to study usual care, which should capture some physician-level factors. While we cannot ascertain use of enhanced recovery pathways, many efforts to accelerate post-operative care for colectomy predate our study period, though adoption of these pathways has been heterogeneous and dynamic throughout the United States.30 Although NSQIP began measuring enhanced recovery metrics for colectomy in 2014 indicating rising uptake, our previous work finds low uptake of these enhanced recovery pathways in Michigan during the 2012–2015 study period.31,32 We limited to hospitals that performed 10 eligible cases, and those hospitals performing fewer cases may have more straightforward operations and healthy patients, which may bias our results to the null. Our dataset was limited to fee-for-service Medicare claims and therefore did not capture patients who are privately insured and/or under the age of 65 undergoing colectomy. We intentionally focused, however, on older adults given their previously established tendency to utilize post-acute care at high levels. Finally, there may be additional unmeasured or unidentified confounding variables that influence payments, length of stay, and post-acute care spending, particularly ownership of post-acute care facilities.
While administrative data does not identify care pathways, such as enhanced recovery protocols, these pathways are often key contributors to accelerated postoperative discharge. Laparoscopic surgery is a key part of enhanced recovery protocols, and we found an association between hospitals’ use of laparoscopic surgery and shorter average LOS. The lack of compensatory increases in post-acute care utilization in the short LOS hospitals, even for older adults and those with multiple comorbidities, suggests that efforts to shorten postoperative hospitalization with accelerated post-operative care pathways are well aligned with episode-based payment initiatives. Patients with a greater comorbidity burden have higher payments overall and may require additional attention in planning of payment reports. Older adults, while demonstrating higher post-acute care payments than their younger counterparts, remain similar in total episode payment and may not require such policy attention. In the setting of BPCI-A including major bowel procedures, risk adjusting for propensity to require post-acute care spending could be a way for hospitals to continue to care for patients that require costlier care: particularly those with multiple comorbid conditions.
Supplementary Material
Acknowledgments
Funding/Support:
This work was supported by the Agency for Healthcare Research and Quality (grant number T32HS000053 (ADR)); the National Clinician Scholars Program (ADR); and the National Institute on Aging (grant number 1K08AG047252-01A1 (SER)).
Footnotes
Financial Disclosures: None reported.
REFERENCES
- 1.Adamina M, Kehlet H, Tomlinson GA, Senagore AJ, Delaney CP. Enhanced recovery pathways optimize health outcomes and resource utilization: a meta-analysis of randomized controlled trials in colorectal surgery. Surgery. 2011;149:830–840. [DOI] [PubMed] [Google Scholar]
- 2.Bagnall NM, Malietzis G, Kennedy RH, Athanasiou T, Faiz O, Darzi A. A systematic review of enhanced recovery care after colorectal surgery in elderly patients. Colorectal Dis 2014;16:947–956. [DOI] [PubMed] [Google Scholar]
- 3.Slieker J, Frauche P, Jurt J, et al. Enhanced recovery ERAS for elderly: a safe and beneficial pathway in colorectal surgery. Int J Colorectal Dis 2017;32:215–221. [DOI] [PubMed] [Google Scholar]
- 4.Lee GC, Hodin RA. Applying enhanced recovery pathways to unique patient populations. Clin Colon Rectal Surg 2019;32:134–137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Braga M, Pecorelli N, Scatizzi M, Borghi F, Missana G, Radrizzani D; PeriOperative Italian Society. Enhanced recovery program in high-risk patients undergoing colorectal surgery: results from the perioperative Italian society registry. World J Surg 2017;41:860–867. [DOI] [PubMed] [Google Scholar]
- 6.Brown CJ, Friedkin RJ, Inouye SK. Prevalence and outcomes of low mobility in hospitalized older patients. J Am Geriatr Soc 2004;52:1263–1270. [DOI] [PubMed] [Google Scholar]
- 7.Zisberg A, Shadmi E, Sinoff G, Gur-Yaish N, Srulovici E, Admi H. Low mobility during hospitalization and functional decline in older adults. J Am Geriatr Soc 2011;59:266–273. [DOI] [PubMed] [Google Scholar]
- 8.Riedinger JL, Robbins LJ. Prevention of iatrogenic illness: adverse drug reactions and nosocomial infections in hospitalized older adults. Clin Geriatr Med 1998;14:681–698. [PubMed] [Google Scholar]
- 9.Covinsky KE, Pierluissi E, Johnston CB. Hospitalization-associated disability: “She was probably able to ambulate, but I’m not sure”. JAMA. 2011;306:1782–1793. [DOI] [PubMed] [Google Scholar]
- 10.Covinsky KE, Palmer RM, Fortinsky RH, et al. Loss of independence in activities of daily living in older adults hospitalized with medical illnesses: increased vulnerability with age. J Am Geriatr Soc 2003;51:451–458. [DOI] [PubMed] [Google Scholar]
- 11.Carey K. Measuring the hospital length of stay/readmission cost trade-off under a bundled payment mechanism. Health Econ 2015;24:790–802. [DOI] [PubMed] [Google Scholar]
- 12.Regenbogen SE, Cain-Nielsen AH, Norton EC, Chen LM, Birkmeyer JD, Skinner JS. Costs and consequences of early hospital discharge after major inpatient surgery in older adults. JAMA Surg 2017;152:e170123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Al-Refaie WB, Parsons HM, Habermann EB, et al. Operative outcomes beyond 30-day mortality: colorectal cancer surgery in oldest old. Ann Surg 2011;253:947–952. [DOI] [PubMed] [Google Scholar]
- 14.Regenbogen SE, Gust C, Birkmeyer JD. Hospital surgical volume and cost of inpatient surgery in the elderly. J Am Coll Surg 2012;215:758–765. [DOI] [PubMed] [Google Scholar]
- 15.Miller DC, Gust C, Dimick JB, Birkmeyer N, Skinner J, Birkmeyer JD. Large variations in Medicare payments for surgery highlight savings potential from bundled payment programs. Health Aff (Millwood). 2011;30:2107–2115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res 2010;45:1783–1795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care. 2009;47:626–633. [DOI] [PubMed] [Google Scholar]
- 18.Galai N, Israeli A, Zitser-Gurevich Y, Simchen E. Is discharge policy a balanced decision between clinical considerations and hospital ownership policy? The CABG example. J Thorac Cardiovasc Surg 2003;126:1018–1025. [DOI] [PubMed] [Google Scholar]
- 19.Gottlieb DJ, Zhou W, Song Y, Andrews KG, Skinner JS, Sutherland JM. Prices don’t drive regional Medicare spending variations. Health Aff (Millwood). 2010;29:537–543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ghaferi AA, Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients. Ann Surg 2009;250:1029–1034. [DOI] [PubMed] [Google Scholar]
- 21.Iezzoni LI, Daley J, Heeren T, et al. Identifying complications of care using administrative data. Med Care. 1994;32:700–715. [DOI] [PubMed] [Google Scholar]
- 22.Schmidt HM, El Lakis MA, Markar SR, Hubka M, Low DE. Accelerated recovery within standardized recovery pathways after esophagectomy: a prospective cohort study assessing the effects of early discharge on outcomes, readmissions, patient satisfaction, and costs. Ann Thorac Surg 2016;102:931–939. [DOI] [PubMed] [Google Scholar]
- 23.Dexter F, Epstein RH. Reductions in average lengths of stays for surgical procedures between the 2008 and 2014 United States national inpatient samples were not associated with greater incidences of use of postacute care facilities. Anesth Analg 2018;126:983–987. [DOI] [PubMed] [Google Scholar]
- 24.Dexter F, Epstein RH, Dexter EU, Lubarsky DA, Sun EC. Hospitals with briefer than average lengths of stays for common surgical procedures do not have greater odds of either re-admission or use of short-term care facilities. Anaesth Intensive Care. 2017;45:210–219. [DOI] [PubMed] [Google Scholar]
- 25.Reames BN, Sheetz KH, Waits SA, Dimick JB, Regenbogen SE. Geographic variation in use of laparoscopic colectomy for colon cancer. J Clin Oncol 2014;32:3667–3672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Chen LM, Norton EC, Banerjee M, Regenbogen SE, Cain-Nielsen AH, Birkmeyer JD. Spending on care after surgery driven by choice of care settings instead of intensity of services. Health Aff (Millwood). 2017;36:83–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ellimoottil C, Ryan AM, Hou H, Dupree J, Hallstrom B, Miller DC. Medicare’s new bundled payment for joint replacement may penalize hospitals that treat medically complex patients. Health Aff (Millwood). 2016;35:1651–1657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sheetz KH, Dimick JB, Regenbogen SE. How patient complexity and surgical approach influence episode-based payment models for colectomy. Dis Colon Rectum 2019;62:739–746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Halawi MJ, Vovos TJ, Green CL, Wellman SS, Attarian DE, Bolognesi MP. Patient expectation is the most important predictor of discharge destination after primary total joint arthroplasty. J Arthroplasty. 2015;30:539–542. [DOI] [PubMed] [Google Scholar]
- 30.Ozhathil DK, Li Y, Smith JK, et al. Colectomy performance improvement within NSQIP 2005–2008. J Surg Res 2011;171:e9–e13. [DOI] [PubMed] [Google Scholar]
- 31.George E, Krapohl GL, Regenbogen SE. Population-based evaluation of implementation of an enhanced recovery protocol in Michigan. Surgery. 2018;163:1189–1190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Berian JR, Ban KA, Liu JB, Ko CY, Feldman LS, Thacker JK. Adherence to enhanced recovery protocols in NSQIP and association with colectomy outcomes. Ann Surg 2019;269:486–493. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.