Influences of the Organization of Primary Care Practice on Colorectal Cancer Screening Rates

1
Influences of the Organization of Primary Care
Practice on Colorectal Cancer Screening Rates

• Elizabeth Yano, PhD1,2
• Lynn Soban, PhD, RN1,2
• David Etzioni, MD, MSHS1,3
• Patricia Parkerton, PhD1,2
• 1 VA Greater Los Angeles HSRD Center of
  Excellence 2 UCLA School of Public Health 3
  David Geffen-UCLA School of Medicine
• 7th Annual Health Care Organizations Conference
• Richmond, Virginia ? June 2005

2
Acknowledgments

• Colorectal Cancer QUERI (Project CRS 02-163)
• Michelle vanRyn, PhD, Laura Kochevar, PhD,
  Melissa Partin, PhD, Center for Chronic Disease
  Outcomes Research, Minneapolis VA and University
  of Minnesota
• UCLA RWJ Clinical Scholars Program
• VA Office of Quality Performance
• Steven Wright, PhD, Mike Kerr, Dean Bross
• VA Greater Los Angeles HSRD Center of Excellence
• Andy Lanto, MA, MingMing Wang, MPH, Ismelda
  Canelo, Vera Snyder-Schwartz, MA, Lisa
  Rubenstein, MD
• VA Office of Academic Affiliations
• Linda Johnson, PhD, Director, Associated Health
  Postdoctoral Fellowship Program

3
The Old VA?
4
U.S. Veterans

• Veterans are Americans who have been discharged
  from military service
• U.S. citizens, typically English-speaking
• Men and women 18 years and older
• Over 26 million in U.S. (2000) (13 population)
• Nearly 4 decline since 1990 (1 million deaths)
• Represents from 10-17 of civilian
  population/state
• Mostly white non-Hispanic males (average age 58),
  mostly employed high school graduates

Sources U.S. Census (2000), National Survey of
Veterans (2001)
5
U.S. Veterans

• 94 of veterans are men
• Compared to 49 of US adults
• 6 of vets are women
• Women veterans tend to be
• Younger
• Have a higher education level
• Include a greater of minorities
• Fastest growing segment of new VA users
• Projected to be 10 by 2010
• Nearly 20 of OEF/OIF military

Sources 2001 National Survey of Veterans,
Center for Women Veterans.
6
Placing VA Users in ContextSociodemographics
Source Wilson Kizer, Health Affairs, 1997.
7
Placing VA Users in ContextInsurance and Health
Status
Veterans who use the VA have fewer options
and heavier health needs...
Source Wilson Kizer, Health Affairs, 1997.
8
Health Status of U.S. Veterans

• Eye or vision problem (56) ? 63-76 of VA users
• High blood pressure (34) ? 40-48
• Arthritis or rheumatism (26.8) ? 39-55
• Heart trouble (16) ? 20-28
• Severe chronic pain (15) ? 27-54
• Stomach or digestive disorder (13) ? 19-37
• Other ear, nose or throat condition (13) ?
  19-28
• Diabetes requiring insulin or dietary tx (11) ?
  12-18
• Kidney or bladder trouble (10) ? 14-24

Substantially higher chronic disease burden,
lower SF-36 scores, yet rates of physical
decline 2.5 times lower than civilian outcomes
Source 2001 National Survey of Veterans Among
top 3 priority groups
9
Major Reorganization of VA Health Care System
(1996)

• Historically, VA was
• Regionally-supported (4 geographic regional
  offices)
• Individually-managed hospitals focused on
  specialty care
• Funded through prior-year cost reimbursement
• Extremely poor quality of care reputation
• Congressionally approved VA reorganization
• Veterans integrated service networks (VISNs)
  (n22)
• Network-level leadership with budget control and
  executive performance agreements (incentivized
  audit-and-feedback)
• Funded by population served (capitated)
• Simultaneous eligibility reform changed VA to
  health plan/payor
• Computerized patient record system (CPRS)
  w/decision support
• Primary care as platform for restructuring care
  delivery

10
VA Health Care System as the Organizational
Context

• VA health care system largest in U.S.
• 163 hospitals, 750 freestanding community-based
  outpatient clinics, 130 nursing homes, 200 vet
  ctrs
• Served over 5 million patients in past year
• Affiliated with over 1,200 academic institutions
• Including 107 medical schools, 55 dental schools
• 110,000 students and trainees in 45
  disciplines/year
• ½ of all U.S. MDs received part of their
  training in VA
• VA market penetration growing rapidly
• 20 of those 65
• VA health care budget 25.9 billion (2003)

11
Colorectal Cancer Screening

• Colorectal cancer (CRC) represents one of highest
  clinically preventable disease burdens
• Efficacious screening tests available (e.g., FOBT
  w/follow-up, sigmoidoscopy, colonoscopy)
• Widespread recognition of importance of early
  detection
• Yet variation in CRC screening remains high
• Patient-level determinants (e.g., insurance,
  knowledge deficits, patient preferences)
• Contextual factors (e.g., area characteristics)
• Current knowledge of organizational factors
  limited

12
VA Colorectal Cancer Screening

• VA has achieved higher CRC screening rates than
  other systems or settings
• among the lowest performing prevention measures
• among most variable despite nearly 10 years
  effort

13
Study Objective

• To evaluate the contribution of primary care
  practice-level determinants of colorectal cancer
  (CRC) screening variations
• Accounting for contextual factors
• Controlling for patient-level factors

14
Conceptual Model Study Design
Primary Care Practice Organization (99-00)
Patient Outcomes (01)
CENTRALIZATION

Patient Receipt of Colorectal Cancer Screening
Test
INTEGRATION

CONTROL VARIABLES Regional variations Patient
covariates
-
DIFFERENTIATION
Adapted from Bazzoli et al., 1999 Dubbs et
al., 2004 Zapka et al., 2002 Zapka Lemon 2004
15
Organizational Measures

• Centralization
• Primary care authority over decision-making
• Practice structure and operations (k4, a0.89)
• Staffing and human resource management (k5,
  a0.84)
• Organizational influence outside of practice
  (k4, a0.78)
• Financial independence/resource control
• Separate budgetary control over primary care
• Ability to recapture cost savings and re-invest
  in practice
• Resource alignment to PC mission through PC
  service lines

Source Yano et al. VA Primary Care Practices
Survey (1999-2000)
16
Organizational Measures

• Integration
• Sufficiency of resources supporting service
  coordination
• Non-physician staffing (k3, a0.69)
• Admin/clinical space supporting PC functions
  (k3, a0.85)
• Clinical support arrangements (k5, a0.81)
• Differentiation
• Organizational complexity of PC practice
  environment
• Practice size
• Academic affiliation
• Local area managed care penetration

Sources From Yano et al. VA Primary Care
Practices Survey (1999-2000) Except for
practice size (Austin), area managed care
penetration (InterStudy)
17
Patient-Level Measures

• Screened for colorectal cancer
• At least 3 returned fecal occult blood test
• (FOBT) cards (past 12 months) OR
• Flexible sigmoidoscopy (past 5 years) OR
• Colonoscopy (past 10 years)
• Sociodemographics
• Age, gender, race-ethnicity, income,
• insurance status
• Utilization
• Primary care visits (past 12 months)

Chart reviews (EPRP)
Austin admin data (OPC)
18
Sample Determination
Organizations
Patients
Org Survey Data (n219)
Patient Data (n52,323)

• history of CRC
• history of IBD
• history of polyps
• ages 52-85 years
• PC visit patterns

VAMCs Only (n159)
Complete Data (n155)
Patient Data (n38,818)

VA/CMS merge
19
Statistical Analysis

• Correlation coefficients and ANOVA
• To examine bivariate relationships between
  organizational predictors and facility-level CRC
  screening rates
• Generalized estimating equations (GEE)
• To evaluate influence of organizational
  characteristics on a patients probability of
  receiving CRC screening
• Correcting for patient clustering within
  facilities
• Adjusting for patient-level covariates and region

20
Organizational Characteristics of VA Health Care
Facilities (n155)
21
Organizational Characteristics of VA Health Care
Facilities (n155)
22
Characteristics of Patients Eligible for CRC
Screening (n38,818)
23
Organizational Characteristics and Colorectal
Cancer Screening
p 24
Organizational Characteristics and Colorectal
Cancer Screening
p 25
Practice-Level Predictors of Patients Receipt of
CRC Screening
Generalized estimating equations (GEE) model,
adjusted for clustering of patients within
practices and for patient-level age, gender,
race-ethnicity, income, insurance, PC visit rates.
26
Conclusions

• Crossing the quality chasm will require major
  changes in how care is organized
• VA as case study
• Restructured delivery to teams, service lines
• Implemented electronic medical records
• Changed incentives w/audit-and-feedback
• Increased managerial/clinical decision support
• and public sector turnaround
• Substantial changes in quality of care over time
• Demonstrably better performance than Medicare
• Demonstrably better diabetes care than Kaiser

Sources IOM, 2001 Jha, et al., 2003 Asch, et
al., 2004 Kerr, et al., 2005.
27
Conclusions

• Patients were much more likely to receive CRC
  screening if they were attending a PC practice
  with
• Greater autonomy over how care is organized
• Consistent with office-systems approaches
• Necessary but not sufficient substrate for QI and
  org change
• Higher investment in clinical support
  arrangements
• May offset impact of competing demands in busy PC
  practice
• Strong mediating role of VA electronic medical
  records

28
Conclusions

• Patients were much more likely to receive CRC
  screening if they were attending a PC practice
• In smaller care environments
• PC in large, tertiary care settings is less
  organized
• As practices grow, hierarchical levels ? (?
  bureaucracy)
• Practices also in competition for centralized
  resources
• But evidence on influence of practice size mixed
  (e.g., practice size predictor of cervical
  cancer screening)
• Single-encounter screening (eg, paps, flu shots)
  vs.
• multiple-encounter screening (eg, CRC
  screening)
• Multiple-encounter screening requires practice
  autonomy and integration of sufficient resources
  to support work of PCPs

Sources Goldzweig et al. 2004 Battista et al
1990.
29
Implications

• Placing VA in organizational context
• VA as staff-model HMO (akin to Kaiser)
• VA as integrated delivery system (IDS)
• All practices linked through electronic records,
  common policies and procedures
• Hierarchical reporting authority and performance
  accountability
• Common focus on health care needs of specific
  community veterans
• VA organizational traits not unlike other medical
  group practices
• Survival threat and leadership overcame barriers
  to change

30
They always say time changes things, but you
actually have to change them yourself.
Andy Warhol, The Philosophy of Andy Warhol
31
Thank you for your attention
and input