The Effects of Pay for Performance System on Tuberculosis Control Treatment in Taiwan

The Effects of Pay-for-Performance System on Tuberculosis Control and Treatment in Taiwan Ya-Hsin Li Doctoral Candidate Department of Health Systems Management Tulane University (Jun 3rd , 2007) Agenda • • • • • • • Backgrounds Objectives Literature Review Research Methods Results Discussions Implications Background • In 2004, about five million new Tuberculosis (TB) cases were found in the whole world (WHO, 2006) • In 2004, 1.7 million died from TB (WHO, 2006) • TB is expected to rank as one of the top ten causes of illness in the world by the year 2020 (WHO, 2006; Kochi et al., 1997). Taiwan Incidence Rate Death Rate China Japan US UK Canada 70 4.23 101 17 • • 30 4 5 0 12 1 5 1 Incidence rate/ death rate: per 100,000 population per year Data Sources: CDC in Taiwan, 2006; WHO, 2006 Background • Bureau of National Health Insurance (BNHI) in Taiwan started in – October, 2001: pay-for-performance payment demonstration project for TB – January, 2004: pay-for-performance on Tuberculosis (P4P on TB) program • Only few research studies have examined how the P4P program worked, or how the P4P affected TB treatment outcomes. Objectives • To compare patients’ cure rate, death rate and length of treatment before and after P4P on TB program approach. • To compare patients’ cure rate, death rate and length of treatment between P4P hospitals and Non-P4P hospitals. • To compare patients’ cure rate, death rate and length of treatment among P4P hospitals with or without case managers. Literature Review - P4P on TB in Taiwan • Goals of P4P on TB – – – – Decreasing the failure rates in tuberculosis treatment Increasing the cure rate of tuberculosis Improving the case management model Improving the payment system to encourage health care providers to provide high quality treatment, and give more responsibilities to physicians or other health care providers. Literature Review - Systems of treating TB in Taiwan New TB Cases TB patients can choose treatment hospital by their own choices. Non-P4P Hospital P4P Hospital Hospital must have case Manager If TB cases are more than 100 Non-P4P Patients Without case manager With case manager Non-P4P Patients P4P Patients Non-P4P Patients P4P Patients Physicians can decide whether TB patients can/can’t join the P4P program based on the criteria set by the Bureau of National Health Insurance (This study organized, 2007) Literature Review -Medical Payment for P4P on TB P4P on TB payment (bonus points) (Hospitals and physicians can get original payment plus this P4P payment) Stage First (1-3 months) Second (4-6 months) Third (7-9 months) Fourth (9-12 months) Total Case Management 1,500 1,500 Physicians 500 (case notification) 1000 if case is cured 1000 if case is cured 1000 if case is cured 1,500 when patient cured 500 per month 500 per month 3,000 and more Study Structure National Dataset from CDC of Taiwan Before P4P implemented After P4P implemented Year 2002 Year 2003 Year 2004 Year 2005 P4P Hospitals Non P4P Hospitals With Case Managers Without Case Managers Compare Patients’ cure rate, death rate and length of treatment Source of Data • National dataset from years 2002 to 2005 has been obtained from the CDC of Taiwan. • Supplementary data – – – – Patients’ gender/age, Treatment hospital/ case manager, Date of registration and date of completion of treatment, Reasons for closing case (cured, dead, or others) Research Methods • Statistics – Descriptive statistics: • Cure rate • Death rate • Length of treatment (LOT) – Bivariate statistics: • Chi-Square/ T-test – to determine any difference of cure rate, death rate, length of treatment under: » Before and After P4P program implemented » P4P hospitals and Non-P4P hospitals » P4P hospitals with and without case managers – Logistic Regression • Factors associated with cure treatment for TB patients – Dependent variable: patient cured or not – Independent variables: patients’ gender, patients’ age, hospital type, hospital ownership, hospital regions, P4P or Non-P4P hospital, with case manager or without. Results Non P4P P4P Total Numbers of hospital joined P4P on TB Medical Center Regional Hospital Local Hospital Clinics 6 18 364 19973 18 60 117 66 (75.00) (76.92) (24.32) (0.33) 24 78 481 20039 Without Case manger Medical Center Regional Hospital 1 14 48 55 With Case manager 17 46 69 11 (94.44) (76.67) (58.97) (16.67) Total 18 60 117 66 • Numbers of hospital with Case Managers Local Hospital Clinics Descriptive Analysis – Before P4P V.S After P4P Before P4P After P4P Implementation N N Sexual Male Female Age <45 yrs old 45-64 yrs old >=65 yrs old Cured cases Death cases Average treatment days (only count for Cured cases) 7057 7785 18355 6754 171 213.12 21.30 23.50 55.30 20.35 0.52 22428 10769 67.60 32.40 33197 % 42.20 Implementation N 45398 30484 14898 9710 10587 25039 12831 875 212.35 % 57.80 67.20 32.80 P 0.252 0.851 21.40 23.40 55.20 28.26 1.93 <0.01 <0.01 0.133 * Before implementation of P4P: year 2002-2003; After P4P: year 2004-2005 * Cases were followed for 9 months since the registration of the cases Descriptive Analysis – P4P Hospital V.S Non P4P Hospital Non-P4P N (%) N Sexual Male Female Age <45 yrs old 1938 (19.90) 7772 (21.80) 6640 (68.20) 3097 (31.80) 23844 (66.90) 11801 (33.10) <0.01 9741 (21.50) P4P N (%) 35657 (78.50) 0.02 P 45-64 yrs old >=65 yrs old Cured cases Death cases Average treatment days (only count for Cured cases) 2008 (20.60) 5781 (59.40) 1913 (19.64) 238 (2.44) 214.31 8579 (24.10) 19258 (54.10) 10918 (30.62) 637 (1.79) 212.01 <0.01 <0.01 <0.01 * Cases were followed for 9 months since registration of the cases Descriptive Analysis – Hospital with Case Manager V.S without Case Manager Without N (%) N Sexual Male 11415 (67.74) 19069 (66.80) 16851 (37.12) with N(%) 28547 (62.88) 0.04 P Female Age <45 yrs old 45-64 yrs old 5429 (32.22) 3683 (21.86) 3728 (22.12) 9469 (33.17) 6027 (21.11) 6859 (24.03) <0.01 >=65 yrs old Cured cases Death cases Average treatment days 9422 (55.91) 4143 (24.59) 336 (1.99) 213.35 15617 (54.71) 8688 (30.43) 539 (1.89) 211.88 <0.01 0.75 0.02 (only count for Cured cases) * Under P4P hospital, if the hospital have more than 100 new TB cases in a year, the hospital should have 1 case manager in the hospital. Factors associated with cure treatment for TB patients Criteria Intercept Sex male (reference) female Age <45 yrs old (reference) 45-64 yrs old >= 65 yrs old Estimate 0.45 -0.07 -0.33 -0.87 Standard Error 0.04 0.02 0.03 0.03 OR P 0.94 0.72 0.42 <0.01 <0.01 <0.01 Hospital Type Medical Center (reference) Regional Hospitals Local Hospitals Clinics P4P Hospital Non-p4p(reference) P4P hospital Case Manager With no case manager (reference) With case manager 0.00 -0.01 -0.61 0.03 0.03 0.07 1.00 0.99 0.54 0.99 0.86 <0.01 0.51 0.03 1.67 <0.01 -0.13 0.03 0.88 <0.01 Owner Public Hospital (reference) Private Hospital Non-profit proprietary hospital Military/Vetaran hospital Public clinic Private clinic -0.17 0.04 -0.11 -0.38 0.54 0.03 0.03 0.04 0.07 0.09 0.84 1.04 0.90 0.68 1.72 <0.01 0.19 0.01 <0.01 <0.01 Discussion • When we compare before/after P4P implemented, the death rate was higher in the group of “after P4P implemented”. – Patients interrupted from treatment, and when returning to the system, patient maybe have multidrug resistance problem. Thus the study should control the interrupted and returning cases. – The TB data may be more completely due to the implementation of P4P. Discussion • When we compare hospitals with case managers and without case managers, there was no significant difference in terms of death rate. – This may be due to the fact that the system of case management started in 2004 and impact on deaths will be observable after some time. – Most case managers were in medical centers or regional hospitals. The severity of illness of patients may be higher in these tertiary level facilities. Severity of patients may have increased the death. Thus, the comparison should control for disease severity. Conclusions and Policy Implication • Comparison of outcomes before and after the implementation of P4P indicate that the cure rate improved significantly after the introduction of P4P. • Comparisons of “P4P hospitals” and “Non-P4P hospitals” show that the cure rate, death rate and average length of treatment all improved significantly. These results confirmed that P4P is a good payment policy for TB treatment. • Hospitals with case manager had better treatment cure rate and length of treatment than hospitals without case manager. These results indicate the importance of case managers in TB treatment. Limitations • The dataset has a number of limitations. • The study could not control the level of severity of the illness of the patients. Ya-Hsin Li yli@tulane.edu