Hospital Spending Intensity and Readmission

1
Hospital Spending Intensity and Readmission

• 20130806
• Speaker Chih-Yuan Huang
• Corresponding Author Ching-Chih Lee

2
Introduction(1)

• It is believed that good hospital quality lead to
  good outcomes.
• And hospital quality depends on effective
  measures, more measures require more cost or
  spending.
• Thus we expected that hospital spending intensity
  is proportional to hospital quality and then, the
  most important, good outcomes.
• However, the results are conflicting. For
  example, those favor high spending is associated
  with better outcomes
• Silber JH, Health Serv Res. 201045(6, pt
  2)1872-1892
• Barnato AE, Med Care. 201048(2)125-132
• And those do not
• Fisher ES,  Ann Intern Med. 2003138(4)273-287
• Fisher ES  Ann Intern Med. 2003138(4)288-298
• Kaplan RM,J Palliat Med. 2011 Feb14(2)215-20.
• Yasaitis L, Health Aff (Millwood). 2009
  Jul-Aug28(4)w566-72.

3
Introduction(2)

• Reverse causality(????) Sicker patients use
  more services, higher-spending hospitals may
  appear to have worse outcomes, in part because
  patients are more severely ill. To avoid this
• 1. Selected acute conditions who were likely to
  present with similar mean illness severity,
  rather than all admissions.
• 2. The exposure of interest was spending at the
  hospital level rather than the patient level.
• 3. Estimates of a hospitals spending intensity,
  were based on individuals in their last year of
  life, to further remove potential reverse
  causality between study cohort illness and
  spending.

4
EOL spending intensity-hospital attribute

• Hospitals' end-of-life intensity varies in the
  use of specific life-sustaining treatments that
  are somewhat emblematic of aggressive end-of-life
  care. End-of-life intensity is a relatively
  stable hospital attribute that is robust to
  multiple measurement approaches.
• Barnato AE, Med Care. 2009 Oct47(10)1098-105.

5
Hospital Spending Index

• The primary exposure measure was the hospital
  end-of-life expenditure index (EOL-EI),
  calculated as the mean adjusted spending per
  capita on hospital, emergency department (ED),
  and physician services provided to decedents in
  their last 6 months in their life in our study.

6
Method(1)

• We use NHIRD 2002-2004 and choose patients who
  were admitted due to major diseases such as AMI,
  Stroke, CHF, Pneumonia, DM, and Liver cirrhosis .
  
• Patients were enrolled according to their
  earliest admission(index admission) and underwent
  follow-up for 1 year after the index admission
  date.

7
Methods(2)

• We divided hospitals into three quatiles of
  different end-of-life spending intensity by low,
  moderate and high spending index, as exposure, to
  explore the outcomes of the selected diseases
  under different spending intensities.

8
Method(3)

• The primary outcomes
• 30-day and 1-year mortality
• 30-day and 1-year readmissions
• Cox proportional hazards models were used to
  compare rates of mortality and readmissions
  across hospital expenditure categories.

9
Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431) Table 1. Baseline characteristics, According to Hospital Expenditure Index . (n890431)
Characteristics Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index P value
Characteristics High (n299109) High (n299109)   Moderate (n288456) Moderate (n288456)   Low (n302866) Low (n302866) P value
Characteristics N   N   N P value
Disease                 lt0.001
AMI 21935 7.3   22036 7.6   20234 6.7  
Stroke 57146 19.1   58330 20.2   59356 19.6  
Congestive heart failure 21485 7.2   20252 7.0   22157 7.3  
Pneumonia 101276 33.9   94017 32.6   100150 33.1  
DM 77611 25.9   73871 25.6   80652 26.6  
Liver cirrhosis 19656 6.6   19950 6.9   20317 6.7  
Age, years (mean SD) 5826 5826   5726 5726   5825 5825 lt0.001
Age group                 lt0.001
lt44 66488 22.2   63750 22.1   67473 22.3  
45-54 32671 10.9   32389 11.2   31205 10.3  
55-64 44534 14.9   44467 15.4   42262 14.0  
65-74 72065 24.1   70005 24.3   72047 23.8  
?75 83351 27.9   77845 27.0   89879 29.7  
Gender                 lt0.001
Male 174256 58.3   165249 57.3   514328 57.8  
Female 124853 41.7   123207 42.7   376103 42.2  
CCIS (mean SD)                 lt0.001
0 107235 35.9   101255 35.1   107629 35.5 lt0.001
?1 191874 64.1   187201 64.9   195237 64.5  
Socioeconomic status                 lt0.001
High 100817 33.7   88559 30.7   83829 27.7  
Moderate 113414 37.9   125486 43.5   127544 42.1  
Low 84878 28.4   74411 25.8   91493 30.2  
Urbanization                 lt0.001
Urban 92587 31.0   61338 21.3   50651 16.7  
Suburban 125149 41.8   12474 43.2   135664 44.8  
Rural 81373 27.2   102371 35.5   116551 38.5  
Geographic region                 lt0.001
Northern 193956 64.8   129681 45.0   137386 45.4  
Central 36229 12.1   53928 18.7   61961 20.5  
Southern 65100 21.8   88311 30.6   88224 29.1  
Eastern 3824 1.3   16536 5.7   15295 5.1  
Caseload (mean SD) 7971040 7971040   6981178 6981178   170291 170291 lt0.001
Hospital characteristics                  
Hospital ownership                 lt0.001
Public 95516 31.9   62048 21.5   97516 32.2  
Nonprofit 147764 49.4   153794 53.3   94921 31.3  
Profit 55829 18.7   72614 25.2   110429 36.5  
Hospital accreditation level                 lt0.001
Medical center 164808 55.1   143023 49.6   41467 13.7  
Regional hospital 116033 38.8   107520 37.3   144651 47.8  
District hospital 18268 6.1   37913 13.1   116748 38.5  
                   
10
Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index Table 2. Outcomes for the disease patients, According to Hospital expenditure index
Cohort Outcomes Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index Hospital Expenditure Index P value
Cohort Outcomes High (n299109) High (n299109)   Moderate (n288456) Moderate (n288456)   Low (n302866) Low (n302866) P value
Cohort Outcomes n   n   n P value
AMI 21935 34.2   22036 34.3   20234 31.5  
Death                  
Within 30 days of admission 545 2.5   447 2.0   449 2.2 lt0.001
Within 1year of admission 1457 6.6   1210 5.5   1325 6.5 lt0.001
Major cardiac event                  
Within 30 days of readmission 1670 7.6   1711 7.8   2034 10.1 lt0.001
Within 1year of readmission 6102 27.8   5956 27.0   5894 29.1 lt0.001
Stroke 57146 32.7   58330 33.4   59356 34.0  
Death                  
Within 30 days of admission 1211 2.1   962 1.6   965 1.6 lt0.001
Within 1year of admission 3478 6.1   3089 5.3   3410 5.7 lt0.001
Stroke event                  
Within 30 days of readmission 5611 9.8   5653 9.7   7494 12.6 lt0.001
Within 1year of readmission 19794 34.6   20042 34.4   22664 38.2 lt0.001
Congestive heart failure 21485 33.6   20252 31.7   22157 34.7  
Death                  
Within 30 days of admission 427 2.0   982 1.9   441 2.0 0.684
Within 1year of admission 1988 9.3   1701 8.4   2017 9.1 0.005
Major cardiac event                  
Within 30 days of readmission 2128 9.9   2203 10.9   2542 11.5 lt0.001
Within 1year of readmission 9244 43.0   8729 43.1   9370 42.3 0.169
Pneumonia 101276 34.3   94017 31.8   100150 33.9  
Death                  
Within 30 days of admission 3011 3.0   2066 2.2   2521 2.5 lt0.001
Within 1year of admission 8958 8.8   6194 6.6   7611 7.6 lt0.001
Pneumonia event                  
Within 30 days of readmission 4315 4.3   3999 4.3   5417 5.4 lt0.001
Within 1year of readmission 19470 19.2   18125 19.3   22860 22.8 lt0.001
DM 77611 33.4   73871 31.8   80652 34.7  
Death                  
Within 30 days of admission 578 0.7   507 0.7   700 0.9 lt0.001
Within 1year of admission 3303 4.3   2812 3.8   3633 4.5 lt0.001
DM event                  
Within 30 days of readmission 4968 6.4   5291 7.2   6903 8.6 lt0.001
Within 1year of readmission 26205 33.8   26058 35.3   30315 37.6 lt0.001
Liver cirrhosis 19656 32.8   19950 33.3   20317 33.9  
Death                  
Within 30 days of admission 643 3.3   536 2.7   603 2.7 0.003
Within 1year of admission 2651 13.5   1984 9.9   2172 10.7 lt0.001
Liver cirrhosis event                  
Within 30 days of readmission 2234 11.4   2367 11.9   2720 13.4 lt0.001
Within 1year of readmission 10168 51.7   10321 51.7   10620 52.3 0.457
                   
11
(No Transcript)
12
Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index Table 3. Odds ratios for the mortality, According to Hospital expenditure index
Variable 2002 2002 2002   2003 2003 2003   2004 2004 2004
Variable OR 95 CI P value   OR 95 CI P value   OR 95 CI P value
AMI (n64205)                      
Within 30 days of admission                      
High 1       1       1    
Moderate 0.77 0.61-0.99 0.038   0.99 0.78-1.26 0.968   1.24 0.99-1.55 0.067
Low 0.79 0.60-1.04 0.095   1.20 0.91-1.59 0.202   1.19 0.90-1.59 0.230
Within 1year of admission                      
High 1       1       1    
Moderate 0.81 0.70-0.94 0.006   1.09 0.94-1.27 0.251   1.19 1.03-1.38 0.020
Low 0.92 0.78-1.09 0.346   1.16 0.97-1.38 0.105   1.21 1.02-1.45 0.032
Stroke (n174832)                      
Within 30 days of admission                      
High 1       1       1    
Moderate 0.78 0.68-0.90 0.001   0.83 0.71-0.98 0.026   0.96 0.82-1.12 0.577
Low 0.73 0.62-0.85 lt0.001   0.94 0.79-1.12 0.493   0.94 0.78-1.13 0.499
Within 1year of admission                      
High 1       1       1    
Moderate 0.85 0.78-0.93 lt0.001   0.97 0.88-1.07 0.572   0.97 0.88-1.06 0.499
Low 0.81 0.74-0.89 lt0.001   0.93 0.84-1.03 0.179   0.97 0.88-1.08 0.974
Congestive heart failure (n63894)                      
Within 30 days of admission                      
High 1       1       1    
Moderate 0.71 0.56-0.90 0.005   0.83 0.63-1.10 0.201   1.42 1.06-1.89 0.018
Low 0.71 0.56-0.91 0.006   0.75 0.56-1.02 0.062   1.18 0.87-1.61 0.284
Within 1year of admission                      
High 1       1       1    
Moderate 0.81 0.71-0.91 lt0.001   0.90 0.78-1.04 0.157   1.08 0.94-1.24 0.277
Low 0.80 0.70-0.90 0.001   0.96 0.83-1.11 0.578   1.01 0.87-1.17 0.889
Pneumonia (n295443)                      
Within 30 days of admission                      
High 1       1       1    
Moderate 0.80 0.73-0.89 lt0.001   0.91 0.81-1.02 0.090   0.85 0.77-0.95 0.003
Low 0.85 0.76-0.95 0.003   1.00 0.89-1.13 0.999   0.97 0.86-1.09 0.553
Within 1year of admission                      
High 1       1       1    
Moderate 0.79 0.4-0.84 lt0.001   0.82 0.77-0.88 lt0.001   0.86 0.81-0.92 lt0.001
Low 0.89 0.83-0.96 0.002   0.86 0.80-0.93 lt0.001   0.89 0.82-0.95 0.001
DM (n232134)                      
Within 30 days of admission                      
High 1       1       1    
Moderate 0.82 0.68-0.99 0.034   0.98 0.78-1.24 0.857   0.81 0.62-1.04 0.098
Low 0.77 0.64-0.94 0.008   1.08 0.85-1.35 0.570   0.93 0.73-1.20 0.592
Within 1year of admission                      
High 1       1       1    
Moderate 0.88 0.81-0.96 0.002   0.92 0.83-1.02 0.106   1.01 0.91-1.13 0.807
Low 0.83 0.76-0.90 lt0.001   0.97 0.87-1.07 0.517   1.03 0.93-1.16 0.560
Liver cirrhosis (n59923)                      
Within 30 days of admission                      
High 1       1       1    
Moderate 0.79 0.65-0.95 0.015   0.95 0.74-1.22 0.708   1.10 0.88-1.39 0.404
Low 0.76 0.62-0.83 0.007   0.94 0.73-1.21 0.607   1.13 0.89-1.44 0.317
Within 1year of admission                      
High 1       1       1    
Moderate 0.71 0.64-0.79 lt0.001   0.88 0.77-1.01 0.062   0.99 0.87-1.12 0.848
Low 0.70 0.63-0.78 lt0.001   0.85 0.74-0.97 0.018   0.96 0.84-1.10 0.575
Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload. Adjusted for patient age, gender, Charlson Comorbidity Index Score, socioeconomic status, urbanization level, geographic region, hospital characteristics and caseload.
13
Results

• The high expenditure hospital is associated with
  higher caseload and medical center.
• The expenditure index is not associated regularly
  with survival and readmission rates of those
  diseases we choose to look. This is against some
  of the papers published.
• Romley JA, Annals of Internal Medicine
  154160-167.

14
Results

• If we inspect the data year by year, the lower
  expenditure index hospital is associated with
  lower mortality in 2002, but seldom in 2003 and
  2004. And Moderate expenditure index is
  associated with less readmission rate in 2004.
• Conclusion In contrast with previous study,
  higher spending intensity is not always
  associated with lower readmission and mortality
  rates in selected major diseases in Taiwan.

15
Discussion

• In Taiwan our data may suggest higher spending
  doesnt necessarily mean better outcomes.
• Some suggest that better outcomes is related to
  effective process measures. But in 2006 two
  attempts to link process measures with short term
  mortality failed to explain most of their data.
• Bradley EH, JAMA 296 72-78
• Werner RM, JAMA 296 2694-2702

16
Discussion

• High-spending regions are more likely than other
  regions to use recommended care but are also more
  likely to use discretionary and nonrecommended
  care, the latter of which has adverse outcomes
  for patients.
• Landrum MB, Health Aff (Millwood). 2008
  Jan-Feb27(1)159-68

17
Weakness

• We use DD files(in hospital mortality) which may
  not include death before arriving hospital.
• The chosen disease may be to heterogenous in
  severity.

18
Thank you for your attention!
19
Karen E. Joynt, JAMA. 2013309(24)2572-2578
20
????????? ?????HSI???
Karen E. Joynt, JAMA. 2013309(24)2572-2578
21

• Physician patient-sharing networks and the cost
  and intensity of care in US hospitals.
• Hospital-based physician network structure has a
  significant relationship with an institution's
  care patterns for their patients. Hospitals with
  doctors who have higher numbers of connections
  have higher costs and more intensive care, and
  hospitals with primary care-centered networks
  have lower costs and care intensity
• Barnett ML, Med Care. 2012 Feb50(2)152-60.

22
Reference

• JAMA. 2012 Mar 14307(10)1037-45. doi
  10.1001/jama.2012.265. Association of hospital
  spending intensity with mortality and readmission
  rates in Ontario hospitals. Stukel TA, Fisher ES,
  Alter DA, Guttmann A, Ko DT, Fung K, Wodchis WP,
  Baxter NN, Earle CC, Lee DS.

23
Restriction of admission condition to purify the
disease

• To capture incident admissions, we excluded
  patients with AMI and hip fracture admitted for
  these conditions during the previous year and
  patients with CHF having a CHF admission in the
  previous 3 years. We included patients with a
  first diagnosis of colon cancer undergoing
  potentially curative resection within 6 months,
  excluding those who presented with metastatic
  cancer or who were diagnosed with any other
  cancer within the previous 5 years. We excluded
  patients with AMI having a stay of less than 3
  days. Patients were assigned to the cohort
  corresponding to their earliest admission and
  underwent follow-up for 1 year after the index
  admission date. We created an index episode of
  care beginning at initial admission and ending at
  the final discharge, incorporating transfers. To
  ensure stability of the hospital-specific
  measures, we restricted to 129 hospitals with
  more than 10 study condition admissions per year,
  resulting in exclusion of 27 of hospitals but
  only 3 of patients.

Therese A. Stukel, PhD JAMA. 2012 March 14
307(10) 10371045
24
AMI

• ICD9 code 410.x,
• AMI we excluded patients with AMI admitted for
  these conditions during the previous year(index
  admission ?????AMI??????, ?fresh case)
• We excluded patients with AMI having a stay of
  less than 3 days(??????case, ???????????)

25
CHF

• Coding 428.x
• Excluding having a CHF admission in the
  previous 3 years

26
Stroke

• Coding Cerebrovascular disease 430.x438.x
• ?7-21?????????????????.

Shorter Length of Stay Is Associated With Worse
Functional Outcomes for Medicare Beneficiaries
With Stroke.
O'Brien SR, Phys Ther. 2013 Jul 25. Epub ahead
of print
The United States had the shortest LOS (6 days)
in contrast to Canada with the longest LOS (34-47
days). average LOS was 13.9 14.1 days (range
1-129).
Huang YC, J Stroke Cerebrovasc Dis. 2012 Dec 14.
pii S1052-3057(12)00358-8.
27
Liver cirrhosis

• Mild liver disease 571.2, 571.4571.6 gtfilter
  out
• Moderate or severe liver disease 456.0456.21,
  572.2572.8 gt including(???????case ?)

28
DM

• Focus???????debridement?case
• ????????????case ?
• Or with CKD/HD diagnosis

29
Pneumonia

• ?ventilator support,
• ???shock???case

30
Thinking

• Turn to do HSI index admission and subsequent
  emergent department(ED) visits(?????????)
• ??????, Paper ???HSI???