Practical Suggestions for Facilitating Quality Improvement Within Health Care Organizations

Organizational Context & Penetration of QI Interventions: Case Studies from Implementing Depression Collaborative Care Elizabeth Yano PhD1, 2; JoAnn Kirchner MD3, 4; Jacqueline Fickel PhD1; Louise Parker PhD3; Mona Ritchie MSW3; Chuan-Fen Liu PhD5,6; Edmund Chaney PhD5,6; Lisa Rubenstein MD1,7,8 1VA Greater Los Angeles HSR&D Center of Excellence; 2UCLA School of Public Health; 3Center for Mental Health Outcomes Research, Little Rock AR; 4University of Arkansas Medical Sciences; 5Northwest Center for Outcomes Research, Seattle WA; 6University of Washington, Seattle; 7UCLA School of Medicine; 8RAND Health Background “It’s not your father’s Army any more…” – It’s not your father’s VA any more either VA’s quality transformation (1990s to current) – Reorganization towards primary care – Adoption of electronic medical records – Incentivized performance audit-and-feedback – Capitated budgets/resource allocation Parallel with substantial HSR investment Quality Enhancement Research Initiative (QUERI) National disease targetsQUERI Centers Research-clinical partnerships designed to implement research into practice Mental Health QUERI – Depression particularly common and disabling – Implementation of depression collaborative care as national strategic priority for primary care Depression Collaborative Care Forges shared care between PC and MH PC provider education Informatics-based decision support Leadership support Depression care manager – Telephone assessment of + screens – Telephone management and follow-up – Based in PC but supervised by MH specialist Substantial Evidence Base Demonstrates Effectiveness of Collaborative Care Feasible, cost-effective care models show – Improved quality of life for up to five years – Reduced job loss – Improved financial status – Higher satisfaction and participation in care – Reduced disparities in care and outcomes – Improved chronic disease status (HbA1C) More than 10 randomized controlled trials Models Increase Efficiency… Reduce primary care visits Maintain current rate of MHS visits Use MHS resources more effectively Cost-saving (due to reduced medical care costs) after first year – One randomized trial, included VA Research Objective Routine-care implementation of depression collaborative care in VA primary care practices – Little known about factors underlying intervention penetration – Objective: To evaluate influences of organizational characteristics on degree of penetration during implementation Factors Associated with Adoption and Diffusion of Collaborative Care as an Organizational Innovation INDIVIDUAL (LEADER) CHARACTERISTICS ORGANIZATIONAL INNOVATION Collaborative Care for Depression in VA INTERNAL CHARACTERISTICS OF ORGANIZATIONAL STRUCTURE Centralization (-) Complexity (+) Formalization (-) Interconnectedness (+) Organizational slack (+) Size (+) EXTERNAL CHARACTERISTICS OF THE ORGANIZATION System openness Source: Adapted from Rogers EM. Diffusion of innovations. New York: The Free Press, 1995. Study Design & Sample Part of larger group RCT of collab care Implementation thru evidence-based QI – Expert-panel consensus development among PC and MH leaders Implementation priorities Care model specifications Seven 1st-generation primary care practices – Across 3 VA networks spanning 5 states Data Sources & Measures VA administrative data (“Austin”) (caseload) Organizational site surveys – Measures of internal organizational structure (e.g., centralization, complexity) – Measures of external organizational context (e.g., urban/rural location) Intervention penetration reports – % PC providers referring patients, # consults/FTE Validated by qualitative data from semistructured stakeholder interviews – Senior/mid-level health care managers, PC/MH providers, depression care managers Principal Findings Practices ranged from 4,600-14,000 patients among 4-11 PCPs Depression diagnosis ranged from 1-10% of population of PC patients Reported level of implementation high (7-9 out of 9-point scale) Sense of PC-MH collaboration variable – Difficulty deciding if PC or MH responsible Penetration highly variable Limited regional consistency – One VISN high penetration but different approaches PC Provider Penetration % PCPs Started 1st 6 Months 100 90 80 70 60 50 40 30 20 10 0 A1 A2 B1 B2 B3 C1 C2 Network #1 Network #2 Network #3 PC Provider Penetration % PCPs Started 1st 6 Months 100 90 80 70 60 50 40 30 20 10 0 A1 A2 B1 B2 B3 C1 C2 0 10 5 15 % PCPs Started Consults/FTE Referrals/PCP FTEs 30 25 20 Network #1 Network #2 Network #3 Organizational Context & Penetration Referrals/PCP FTE 30 25 20 15 10 5 0 A1 A2 B2 C1 C2 B3 MED MED Levels of early PCP penetration MED HIGH HIGH HIGH LOW B1 # Months: 16 Small city 20 Small city 18 Rural 2 Small city 6 Small city 9 Semirural 21 Rural Organizational Context & Penetration High Penetration Low Penetration Low practice authority Variable resources QI activity variable PC education ~low No PC-MH case confs Med-to-high authority Variable resources QI activity variable PC education med-hi No PC-MH case confs Organizational Context & Penetration Speed or extent of penetration not influenced by: – PC and MH provider relationships – Area characteristics (eg, urban/rural location) – Practice size Except for largest practice (>14,000 patients) Initiating early collaborative care referral did not predict future referral behavior Highest referral rates typically among practices with lowest perceived MH staffing Implications VA an exceptional laboratory in which to translate research into practice – Common electronic medical records – Identifiable management structures – Common policies and procedures Effective penetration may have less to do with these enablers than local clinic characteristics, needs and approach – Moderate penetration  time for PDSA – Time to adopt/adapt  as opposed to “high burn”