Economic Evaluation of a Pre-ESRD Pay-for- Performance Program in - - PowerPoint PPT Presentation

Economic Evaluation of a Pre-ESRD Pay-for- Performance Program in Advanced Chronic Kidney Disease Patients

Hsieh, Hui-Min, Ph.D. Associate Professor at Department of Public Health, Kaohsiung Medical University, Taiwan

Co-authors: Lin, M. Y., Chiu, Y. W., Wu, P. H., Cheng, L.J, Jian, F. S., Hsu, C. C., and Hwang, S. J.

Acknowledgement: This project was funded by National Health Insurance Administration, Ministry of Health and Welfare (M0HW103-NH-1001), Kaohsiung Medical University Hospital (KMUH103-3R10) and Ministry of Science and Technology (MOST-103-2314-B-037-001).

1

Introduction

End-stage renal disease (ESRD) is a major cause of morbidity and mortality worldwide.  The number of patients with ESRD receiving renal replacement therapy (RRT) is projected to grow from 2,618 million in 2010 to 5,439 million worldwide by 2030. One of the key concerns is that these patients consume a substantial share of national health care resources. Therefore, it became a key priority to identify strategies to provide optimization

• f pre-ESRD care for chronic kidney disease (CKD) patients in order to delay

progression to ESRD, improve patient survival and reduce economic costs.

2

Pre-ESRD P4P program

Taiwan has high incidence and prevalence rate of ESRD and undertakes a substantial economic burden from this disease.  A nationwide pre-ESRD Pay-for-performance (P4P) program was launched under the National Health Insurance Administration (NHIA) coverage in late 2006. The primary goals for this program are:

• To provide adequate care for CKD stage 3b, 4 and 5 patients
• To slow down the deterioration of renal function
• To reduce incidence of ESRD
• To provide early preparations for dialysis
• To potentially reduce substantial health expenditure

3

Key Features of the Pre-ESRD P4P Program

Multidisciplinary care team：at least one nephrologist as well as well-trained renal nurses and dieticians). Require to follow guidelines of the National Kidney Foundation Dialysis Outcomes Quality Initiative (NKF-K/DOQI) for the standard care of patients. Regular follow-up care every 3 months Financial Incentives：an extra NT$1,200 (US$40) per initial enrollment visit, NT$600 (US$20) per follow-up visit, and NT$600 (US$20) per annual evaluation visit.

4

Objectives

To investigate： Whether the nationwide pre-ESRD P4P program reduced the risks of all-cause mortality and dialysis initiation for advanced CKD paitents. Whether the program allowed for a cost-effective use of resources among advanced CKD patients from a single payer perspective.

5

Data Sources

Nationwide Pre-ESRD P4P registry：precisely identifying whether patients were enrolled in the P4P program NHIA claims database：obtaining information on patient comorbidities and health provider characteristics. A registry of catastrophic illnesses：providing information for a long-term dialysis ESRD patients and the date on which dialysis began. Death registry data：providing accurate date of death information

6

Study Population

To identify pre-ESRD patients if he or she had the first three primary diagnosis codes [ICD-9-CM] codes 585 or 581.9 given the Pre-ESRD P4P program policies in at least one outpatient visit or inpatient admission and was followed by nephrologists between January 1, 2007 and December 31, 2009. Defined advanced CKD patients for those who had received ESA treatment within two years, one year prior and post, of the index date given the NHIA reimbursement policy regulated that advanced CKD patients who have serum creatinine >6 mg/dL and a hematocrit <28% could receive an erythropoiesis-stimulating agent (ESA). P4P/ Non-P4P group: newly enrolled P4P patients in the P4P registry; Non-P4P group: who were not found to be enrolled in the P4P program.

7

Outcomes of Interest

Effectiveness:

• Presence of ESRD warranting renal replacement therapy (RRT) initiation;
• Mortality due to any cause were the primary end-points.
• Life years gained

Direct medical costs:

• From a single payer perspective. 3% discount rate.
• CKD related medical costs
• Other-cause medical costs
• Costs were measured separately for patients who progressed to ESRD and

underwent dialysis or not during the follow-up period. Time Horizon: 3 years

8

Economic and Statistical Analytical Approach

Propensity score matching (PSM) approach to determine adequate comparison groups. The covariates included age, gender, income status, urbanization, pre-ESRD related comorbidities, charlson comorbidity index, health care institution characteristics (accreditation level, teaching status, and geographic location). Competing-risk regression models based on cause-specific and subdistribution hazards were used to analyze the association between the effect of pre-ESRD P4P program and two primary effectiveness end-points.  Cost-effectiveness analysis Incremental cost effectiveness ratio (ICER) as the ratio of the difference in costs and divided by the difference in effectiveness between groups. Bootstrapping with 500 replications.

9

Incremental cost effectiveness ratio (ICER)

10

Descriptive results for P4P and non-P4P pre-ESRD patient characteristics (before and after matching)

Before PSM matching After PSM matching

Variables P4P Non-P4P p-value P4P Non-P4P p-value

N

2,093 2,161 1,473 1,473

Patient Demographic Characteristics Gender (N, %) Female

1,082 (51.70%) 1,123 (51.97%) 0.859 743 (50.44%) 750 (50.92%)

0.796 Male

1,011 (48.30%) 1,038 (48.03%) 730 (49.56%) 723 (49.08%)

Age in years (mean±SD)

63.79 (±13.52) 64.87 (±13.70) 0.010

64.37 (±13.62) 64.57 (±13.78) 0.693 Income Status(N, %) 1-19999

423 (20.21%) 505 (23.37%) 0.092 346 (23.49%) 326 (22.13%)

0.687 Dependents

714 (34.11%) 707 (32.72%) 489 (33.20%) 478 (32.45%)

20000-39999

812 (38.80%) 813 (37.62%) 537 (36.46%) 561 (38.09%)

40000+

144 (6.88%) 136 (6.29%) 101 (6.86%) 108 (7.33%)

Urbanization (N, %) Rural

587 (28.05%) 624 (28.88%) 0.548 403 (27.36%) 412 (27.97%)

0.711 Urban

1,506 (71.95%) 1,537 (71.12%) 1,070 (72.64%) 1,061 (72.03%)

Patient Clnical Characteristics Pre-ESRD related comobidities (N, %) Diabetes

1,051 (50.22%) 1,048 (48.50%) 0.262

727 (49.36%) 747 (50.71%) 0.461 Hyptertension

1,637 (78.21%) 1,648 (76.26%) 0.129

1,148 (77.94%) 1,152 (78.21%) 0.859 MI

28 (1.34%) 22 (1.02%) 0.333 17 (1.15%) 18 (1.22%)

0.865 CHF

238 (11.37%) 306 (14.16%) 0.006

194 (13.17%) 201 (13.65%) 0.705 Stroke

138 (6.59%) 182 (8.42%) 0.023

106 (7.20%) 121 (8.21%) 0.300 Gout

369 (17.63%) 387 (17.91%) 0.812

257 (17.45%) 254 (17.24%) 0.884 PVD

80 (3.82%) 73 (3.38%) 0.436 47 (3.19%) 53 (3.60%)

0.542 Charlson Comobidity Index (Mean±SD)

4.16 (±2.04) 4.00 (±2.16) 0.014

4.09 (±2.08) 4.11 (±2.11) 0.785

11

Matched Sample P4P Non-P4P P4P Non-P4P P4P Non- P4P Incidence rate/ mortality rate comparisons

• No. of total person-years

Incidence of dialysis and mortality Incidence rate/ mortality rate per 1000 person-years Incidence rate ratio N (%) N (%) p- value (95%CI) p-value N 1,473 1,473 Incidence of dialysis 2,188 1,853 1,163 1,173 0.649 531.52 632.98 0.84

<0.001

(78.95%) (79.63%) (0.77, 0.91) All-cause mortality 4,102 3,955 283 341 0.009 68.99 86.23 0.80

0.003

(19.21%) (23.15%) (0.68, 0.93) Time-to-event comparisons P4P Non-P4P p-value Time to dialysis (in days) 430.95 (±293.83) 347.26 (±297.45) <0.001 Time to death (in days) 682.22 (±280.79) 595.17 (±285.58) <0.001 Percent of receiving PD (N,%)* 183 (15.74%) 142 (12.11%) 0.011 Percent of receiving HD (N,%)* 980 (84.26%) 1,031 (87.89%) 0.011

Descriptive results for incidence of dialysis and all-cause mortality between matched P4P and non-P4P pre-ESRD cohorts within 3-year follow-ups

Note: HD=hemodialysis; PD=peritoneal dialysis * Only for patients who underwent dialysis after enrollment into the P4P program

12

Incidence of dialysis All-cause mortality Variables SHR 95%CI p-value SHR 95%CI p-value Policy effect P4P 0.845 (0.779,0.916) <0.001 0.792 (0.673,0.932) 0.005

Results from Competing-risk regression analysis of death and dialysis

13

* Models were adjusted for age, gender, income status, urbanization, patient comorbidities listed in Table 1.

SHR=0.845 (P4P v.s. Non-P4P) (95%CI=0.779, 0.916) p-value<0.001

Initiation of Dialysis

.5 1 1.5 2 1 2 3 Follow-up time (in years) Non-P4P P4P

All-cause Mortality

SHR=0.792 (P4P vs Non-P4P) 95%CI=0.673, 0.932 p-value=0.005 .05 .1 .15 .2 Cumulative Incidence 1 2 3 Follow-up time (in years) Non-P4P P4P

• Fig1. Adjusted cumulative incidence curve for risks of death and dialysis initiation between P4P and non-

P4P advanced CKD cohorts

14

Results for Incremental Effectiveness and Costs Between matched P4P and Non-P4P

Patients with Dialysis Patients without Dialysis

Measurements

P4P Non-P4P

Adjusted Difference

P4P Non-P4P

Adjusted Difference Mean±SD Mean±SD Bootstrapped

• Coef. (95% CI)

Mean±SD Mean±SD Bootstrapped

• Coef. (95% CI)

Incremental Effectiveness

Life Years

2.83 ± 0.48 2.74 ± 0.61

0.08

2.63 ± 0.78 2.46 ± 0.90

0.16

(0.08,0.08) (0.15,0.17)

Incremental Direct Costs

Total CKD related direct costs

1,155,401 ± 593,059 1,274,343 ± 698,444

• 114,704

186,802 ± 246,183 192,087 ± 315,821

• 3,434

(-117,061,-112,348) (-5,498,-1,370)

Total Other-Cause medical costs

323,098 ± 386,182 353,801 ± 485,608

• 32,420

178,699 ± 227,382 232,469 ± 395,375

• 45,836

(-33,955,-30,884) (-48,131,-43,542)

15

Patients with Dialysis Patients without Dialysis

Measurements Adjusted Difference Adjusted Difference

Bootstrapped

• Coef. (95% CI)

Bootstrapped

• Coef. (95% CI)

Incremental Cost-Effectiveness Ratio (ICER)

Total CKD related costs per life year gained

• 1,547,067
• 16,231

(-1,617,977,-1,476,158) (-35,192,2,730)

Total other-cause costs per life year gained

• 423,122
• 404,979

(-448,020,-398,224) (-501,570,-308,388)

Results for ICERs Between matched P4P and Non-P4P

16

Discussion

We observed advanced CKD patients enrolled in a pre-ESRD P4P program had a lower risk of progressing to ESRD and mortality, and the time to dialysis therapy was significantly longer (430 versus 347 days) (p<0.001). The superior outcomes might be explained by several potential reasons in terms of specific aspects of the P4P programs:

• Guidelines (NKF-K/DOQI) adherence
• Multidisciplinary team approach: Previous studies have supported the improvement of

multidisciplinary team approaches on quality outcomes for CKD patients (Chen et al. 2015a;

Chen et al. 2013; Chen et al. 2014; Cho et al. 2012; Fenton et al. 2010; Goldstein et al. 2004; Wu et al. 2009).

• Frequent nephrology visits: A previous study suggests that CKD patients who received

consistency of care from required nephrology visits at least once every 3 months in the 6 months prior to RRT starting had a superior prognosis(Singhal et al. 2014)

17

Discussion

P4P significantly increased adjusted LYs and made possible the cost-effective use of resources, particularly for patients who began dialysis. Consistent results were found in

• Hopkins et al. (2011): examined the cost-effectiveness of a multifaceted pre-

ESRD care model for CKD stage 3 and 4 patients in a randomized clinical trial in Canada and suggested cutoff levels for eGFR below 40 or 45 ml/min were more cost-effective than above the cutoff(Hopkins et al. 2011).

• Because of the better preparation for RRT, previous studies also found MDC

patients had less unplanned urgent dialysis, shorter hospital days and a lower incidence of cardiovascular events at dialysis or in the postdialysis period(Chen et al. 2015a; Chen et al. 2014; Wei et al. 2010; Yu et al. 2014).

18

Conclusions

In conclusion, compared with advanced CKD patients who did not enroll in the P4P program, P4P patients had lower risks of both incidence of dialysis initiation and death. In addition, our empirical findings suggest the pre-ESRD P4P program in Taiwan provided long-term cost-effective use of resources and cost-savings for advanced CKD patients.

19

A N Y Q U E S T I O N S ?

20