INTRODUCTION: Ageing populations and the increasing prevalence of multimorbidity are a challenge for healthcare delivery and health system design. Integrated care has been discussed as a solution to address these challenges. In Portugal, Local Health Units (LHU) promote vertical integration of healthcare, with one of the expected effects being a decrease of readmission rates in individuals with chronic conditions. Readmissions are frequently studied for its negative impacts on individuals, carers, and providers, with excessive unplanned readmission rates among hospitals being a sign of frail integrated care. Thus, we assume as the main aim of this study to assess the impact of vertical integration on the readmission of individuals with chronic conditions.
METHODS: A database including administrative data from 1 679 634 inpatient episodes from years 2002-14 was considered. We identified readmissions with the hospital-wide all-cause unplanned readmission measure methodology of Centers for Medicare and Medicaid Services. The considered outcome was 30-day hospital-wide all-cause unplanned readmissions (1: readmitted), and risk-standardized readmission ratio. Chronic conditions were identified from all diagnoses coded with International Classification of Diseases – 9th version – Clinical Modification codes (1: chronic). In order to assess the impact of LHU on the readmission of individuals with chronic conditions, we compared 30-day readmissions before and after the creation of each LHU. We used difference-in- differences technique to address our main aim. In addition, to understand the associations between individuals’ risk factors and time to readmission, we developed a Cox regression model for LHU and control group.
RESULTS: Difference-in-differences results suggest that vertical integration promoted a decrease on risk-standardized readmission ratio in four LHU, but significant only in LHU 1. In addition, when analysed the individual risk of readmission we observed that it was reduced for four LHU, but only significantly for LHU 3 and LHU 5. A sensitivity analysis was performed for annual evolution of odds ratio of risk of readmission, and initial results were considered stable for most years. Cox regression results suggest that for LHU and control hospitals, female individuals were less at risk of readmission than men, the risk increased with increasing age and number of comorbidities. At LHU, we observed a decreased risk of readmission with increasing number of chronic conditions.
Impact of vertical integration on the readmission of individuals with chronic conditions
1. Impact of vertical integration on the
readmission of individuals with
chronic conditions
Óscar Brito Fernandes
Master in Health Management
10th Edition
2014-2016
Supervisors
Rui Santana, PhD
Sílvia Lopes, PhD
2. • Avaliação do impacto da criação das Unidades Locais de Saúde em Portugal, study carried out by
Escola Nacional de Saúde Pública, Universidade NOVA de Lisboa, and funded by Fundação Calouste
Gulbenkian (2014-2016).
• Research team:
Ana Patrícia Marques
Bruno Moita
João Sarmento
Óscar Brito Fernandes
Rui Santana (Coordinator)
Sílvia Lopes
DISCLOSURE
3. BACKGROUND
• Integrated care
• Readmissions
• Chronic conditions
#1
RESEARCH AIMS
• Main aim
• Specific objectives#2
METHODOLOGY
#3
RESULTS
• Characteristics of the sample
• Individuals’ risk factors and
readmission
• Impact of vertical integration
#4
DISCUSSION
• Discussion of results
• Study limitations#5
FINAL REMARKS
#6• Study design
• Data
• Variables
• Statistical analysis
5. Integrated care is an organizational principle for
care delivery[1] as a managerial response to
differentiation and fragmentation[2].
INTEGRATED CARE
Many integrated care approaches aim to provide a
more independent life to individuals with chronic
conditions[3-4], highlighting improvements to the
patients’ care experience and health outcomes.
#1 BACKGROUND
6. PORTO
VISEU
GUARDA
COIMBRA
CASTELO BRANCO
LEIRIA
SANTARÉM
PORTALEGRE
ÉVORA
VIANA DO
CASTELO
BRAGA
VILA REAL
BRAGANÇA
AVEIRO
BEJA
SETÚBAL
LISBOA
FARO
Matosinhos
1999
Alto Minho
2008
2008
2009
2007
2008
Litoral
Alentejano
2012
Norte
Alentejano
Baixo
Alentejo
Guarda
Castelo
Branco
Nordeste
2011 12% Population
PORTUGAL MAINLAND
Local Health Units
15% Budget
NHS HOSPITALS[5]
#1 BACKGROUND
Resident population by county in
LHU’s catchment area was retrieved
from National Statistics Institute on
May 2016. Last data update by June
16, 2015.
7. Readmission is a subsequent inpatient admission
to any acute care facility which occurs within 30
days of the discharge date of an eligible index
admission[6].
READMISSIONS
Excessive unplanned readmission rates among
hospitals could be a sign of frail integrated
care[7].
#1 BACKGROUND
8. Chronic conditions[8] include health conditions
that persist across time and require healthcare,
including non-communicable diseases, mental
disorders, some communicable conditions and on-
going physical impairments.
CHRONIC CONDITIONS
Individuals with chronic conditions are more likely
to experience hospital readmission since they
are more vulnerable to non-effective home
transitions after hospital discharge[9].
#1 BACKGROUND
9. • Describe 30-day readmission frequency in individuals with chronic
conditions, from 2002 to 2014.
• Analyze the association between individuals’ risk factors and
readmission.
• Analyze the impact of vertical integration on the readmission rates
and risk of readmission of individuals with chronic conditions.
Assess the impact of vertical integration
on the readmission of individuals with
chronic conditions
#2 RESEARCH AIMS
10. • Datasets provided by ACSS,
Portuguese Central
Administration for
Healthcare system;
• Data refers to Portugal
mainland hospital morbidity
from 2002 to 2014.
0201
• Outcome research;
• Observational,
analytical,
longitudinal, and
retrospective cohort
study.
Study
Design
Data
Sources
#3 METHODOLOGY
METHODOLOGY
11. • Selected 9 523 432 index
admissions;
• Treatment and Control
group accounted for
1 679 634 index
admissions;
• Time frame: 8 years, 5
years pre-integration, 3
post-integration.
03
Data
Analyzed
Variables Statistical
Analysis
#3 METHODOLOGY
METHODOLOGY
Control group
6
Public hospitals
Treatment
7
Local Health Units
Selection criteria
• Be part of the same
ACSS hospital
benchmark group as
LHU;
• Excluded hospitals with
different contexts
• Data available from pre-
and post-integration
periods for each LHU.
12. • Selected 9 523 432 index
admissions;
• Treatment and Control
group accounted for
1 679 634 index
admissions;
• Time frame: 8 years, 5
years pre-integration, 3
post-integration.
03
Data
Analyzed
Variables Statistical
Analysis
#3 METHODOLOGY
METHODOLOGY
18%
Treatment group 845 275
Control group 834 359
Analysed sample
13. Generalized linear mixed model
at the specialty cohort (AHRQ)
• Readmissions identified using
CMS hospital-wide all-cause
unplanned readmission
measure;
• AHRQ Condition Classification
System for principal diagnosis;
• CMS Condition Category
groups for comorbid diseases;
• Hierarchical logistic regression
models at the specialty cohort.
Generalized linear mixed models
SAS University Edition
Independent variables
Age
Principal diagnosis
Selected comorbidities
Outcome
Individual risk of readmission
Dependent variable
30-day readmission
#3 METHODOLOGY
METHODOLOGY
14. Cox regression
IBM SPSS (v.23)
Covariates
Gender
Age group
# Chronic conditions
# Elixhauser comorbidities
Outcome
Association between individuals’ risk factors and
time to readmission
Time variable
Days until readmission
Status variable
1: Readmitted
#3 METHODOLOGY
METHODOLOGY
Cox regression
• Elixhauser comorbidity index;
• Chronic condition indicator by
AHRQ;
• Initial assessment of covariates
by univariate Cox regression;
• Kaplan-Meier plots visual
inspection;
• Analyses conducted separately
for LHU and control group.
15. Difference-in-differences
STATA (v.13)
Outcome
Risk of readmission (odds ratio) for LHU compared
to the control group
Dependent variable
30-day readmission
#3 METHODOLOGY
METHODOLOGY
Difference-in-differences
• Unconditional logit model with
fixed effects using dummy
variables;
• Parallel trend assumption
tested by a non-linear
restriction:
18. INDIVIDUALS’ RISK FACTORS AND TIME TO READMISSION
#4 RESULTS
LOCAL HEALTH UNITS CONTROL GROUP
Odds Ratio=1 Odds Ratio=1
0.906
0.928
0.839
GENDER
(male)
FEMALE
AGE
(0-19)
20-44
45-64
65-84
85+ 1.716
1.281
0.861
0.683
0.713
1.197
1.755
19. #4 RESULTS
LOCAL HEALTH UNITS CONTROL GROUP
Odds Ratio=1 Odds Ratio=1
1.298
1.280
1.398
CHRONIC CONDITIONS
(0)
1
2 1.287
3
4
5+
1.266
1.233
1.201
ELIXHAUSER COMORBIDITY INDEX
(0)
1
2
3
4
5+
1.604
1.896
2.296
2.509
1.456
1.472
1.396
1.362
1.285
1.583
1.935
2.192
2.403
INDIVIDUALS’ RISK FACTORS AND TIME TO READMISSION
20. RISK OF READMISSION: LHU VERSUS CONTROL GROUP
#4 RESULTS
Odds Ratio=1
1.017
LHU 1
LHU 2
LHU 3
LHU 4
LHU 5
LHU 6
LHU 7
0.991
0.911
1.240
0.860
1.076
0.937
Parallel trend
assumption not verified
21. Vertical integration faces different barriers within
each organization.
Different interventions addressed to reduce hospital
readmissions have different potential of
effectiveness.[10-11]
The risk of
readmission does
not follow a clear
pattern among
LHU.
#5 DISCUSSION
22. In LHU, the risk of readmission decreases with
increasing # chronic conditions, after adjusting for
gender, age group and comorbidities.
Possible evidence of better coordinated care for
these patients?
Groups with higher
#chronic
conditions
presented
decreased risk of
readmission.
#5 DISCUSSION
23. Readmission rates reflect not solely the quality of
hospital care[12-14]
, but also factors in one’s home
and communities[15-17]
.
Lack of national studies to compare results,
specifically regarding readmissions and chronic
conditions.
One cannot
measure vertical
integration impact
solely considering
readmission
indicator.
#5 DISCUSSION
24. Track the hospitals’
organizational evolution
Analytical and selection biasReliability on administrative
data
LIMITATIONS OF THE STUDY
#5 DISCUSSION
Limitation due to the model selected to identify
readmissions, chronic conditions: Also, the
criteria to compose the control group might
have incurred in selection bias.
Study limited in its ability to prove causation.
Difficult to account for the area of residence of
individuals treated at LHU, as well as the
intense hospital horizontal integration
phenomena.
25. FINAL REMARKS
Mixed evidence over 30-day
readmission of individuals with
chronic conditions
More research needed to better
evaluate
It’s a long road to reach integrated
care
#6 FINAL REMARKS
26. REFERENCES
#7 REFERENCES
[1] Shaw S, Rosen R, Rumbold B. What is integrated care? [Internet]. 2011. Available from:
http://www.nuffieldtrust.org.uk/sites/files/nuffield/publication/what_is_integrated_care_research_report_june11.pdf
[2] Lillrank P. Integration and coordination in healthcare: an operations management view. J Integr Care [Internet]. Emerald Group Publishing Limited; 2012 Feb 10 [cited 2016 Apr 18];20(1):6–
12. Available from: http://www.emeraldinsight.com/doi/abs/10.1108/14769011211202247
[3] Dorling G, Fountaine T, McKenna S, Suresh B. The Evidence for Integrated Care [Internet]. 2015. Available from:
http://www.mckinsey.com/~/media/McKinsey/dotcom/client_service/Healthcare Systems and Services/PDFs/The evidence for integrated care.ashx
[4] OECD. Health Reform: Meeting the Challenge of Ageing and Multiple Morbidities [Internet]. Meeting the Challenge of Ageing and Multiple Morbidities. 2011. Available from:
http://www.oecd-ilibrary.org/social-issues-migration-health/health-reform_9789264122314-en
[5] Portugal. Ministério da Saúde. Administração Central do Sistema de Saúde. Termos de referecia para contratualização hospitalar no SNS: Contrato-Programa 2016 [Terms of reference for
hospital contractualization in the NHS. Contract-program 2016] [Internet]. Lisboa; 2016. Available from: http://tinyurl.com/hfumhjr
[6] Horwitz L, Grady J, Zhang W, DeBuhr J, Deacon S, Krumholz H, et al. 2015 Measure Updates and Specifications Report: Hospital-Wide All-Cause Unplanned Readmission Measure - Version
4.0. 2015.
[7] Bisognano, M, Boutwell A. Improving transitions to reduce readmissions. Front Health Serv Manage. 2009;25(3):3–10.
[8] WHO. Innovative care for chronic conditions: building blocks for action: global report. Noncommunicable Diseases and Mental Health. 2002. p. 1–99.
[9] Jackson CT, Trygstad TK, DeWalt DA, DuBard CA. Transitional care cut hospital readmissions for North Carolina medicaid patients with complex chronic conditions. Health Aff.
2013;32(8):1407–15.
[10] Kansagara D, Chiovaro JC, Kagen D, Jencks S, Rhyne K, O’Neil M, et al. Transitions of care from hospital to home: an overview of systematic reviews and recommendations for improving
transitional care in the Veterans Health Administration [Internet]. 2015. Available from: http://tinyurl.com/h52xjlj
[11] Hansen LO, Young RS, Hinami K, Leung A, Williams M V. Interventions to reduce 30-day rehospitalization: A systematic review. Ann Intern Med [Internet]. 2011;155(8):520–8. Available from:
http://tinyurl.com/h4eh3n5
[12] Bianco A, Molè A, Nobile CGA, Di Giuseppe G, Pileggi C, Angelillo IF. Hospital Readmission Prevalence and Analysis of Those Potentially Avoidable in Southern Italy. PLoS One. 2012;7(11).
[13] Fischer C, Lingsma HF, Marang-van De Mheen PJ, Kringos DS, Klazinga NS, Steyerberg EW. Is the readmission rate a valid quality indicator? A review of the evidence. PLoS One.
2014;9(11):1–10.
[14] Horwitz LI, Partovian C, Lin Z, Grady JN, Herrin J, Conover M, et al. Development and use of an administrative claims measure for profiling hospital-wide performance on 30-day unplanned
readmission. Ann Intern Med. 2014;161:S66–75.
[15] Kangovi S, Grande D, Meehan P, Mitra N, Shannon R, Long JA. Perceptions of readmitted patients on the transition from hospital to home. J Hosp Med. 2012;7(9):709–12.
[16] Hu J, Gonsahn MD, Nerenz DR. Socioeconomic status and readmissions: evidence from an urban teaching hospital. Health Aff (Millwood) [Internet]. 2014 May;33(5):778–85. Available from:
http://www.ncbi.nlm.nih.gov/pubmed/24799574
[17] Joynt KE, Jha AK. A path forward on Medicare readmissions. N Engl J Med [Internet]. 2013 Mar 28;368(13):1175–7. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23465069
27. Impact of vertical integration on the readmission of
individuals with chronic conditions
ØMixed evidence over 30-day readmission of individuals with chronic conditions
within LHU
ØIt’s a long road to reach integrated care
ØMore research needed to better evaluate, and better serve
Óscar Brito Fernandes
oscar.fernandes@chlc.min-saude.pt