Implementation and Impacts of Lean Redesigns in Primary Care
October 28, 2016
Contents
Slide 1. Implementation and Impacts of Lean Redesigns in Primary Care
Slide 2. Project Team
Slide 3. Palo Alto Medical Foundation (PAMF)
Slide 4. What is "Lean"?
Slide 5. Implementation of Lean in Primary Care
Slide 6. Conceptual Framework
Slide 7. Qualitative Data Sources
Slide 8. Qualitative Analytic Methods
Slide 9. Measures of Implementation
Slide 10. Outer Setting
Slide 11. Intervention Characteristics
Slide 12. Intervention Characteristics
Slide 13. Process of Implementation
Slide 14. Inner Setting
Slide 15. Characteristics of Individuals and Teams
Slide 16. Characteristics of Individuals and Teams
Slide 17. Summary of Qualitative Findings on Implementation
Slide 18. Impact of Lean on System Performance
Slide 19. Quantitative Methods
Slide 20. Phased Implementation of Lean Redesigns
Slide 21. Workflow Efficiency (Office Visit Charts Closed < 2 hours)
Slide 22. Workflow Efficiency
Slide 23. Physician Productivity
Slide 24. Clinical Quality Metrics
Slide 25. Clinical Quality Metrics
Slide 26. Patient Satisfaction
Slide 27. Patient Satisfaction
Slide 28. Physician Satisfaction % Differences (2011 vs. 2014)
Slide 29. Summary of Lean Impacts
Slide 30. Conclusions
Slide 1
Implementation and Impacts of Lean Redesigns in Primary Care
October 28, 2016
Presenter: Dorothy Hung, Ph.D., Associate Scientist
Palo Alto Medical Foundation Research Institute
Moderator: Michael Harrison, Ph.D., Senior Social Scientist
Center for Delivery, Organization, and Markets, AHRQ
Discussant: Arlene Bierman, M.D., M.S., Director
Center for Evidence and Practice Improvement, AHRQ
Slide 2
Project Team:
Carrie Gray, Ph.D., Su-Ying Liang, Ph.D.,
Meghan Martinez, M.P.H., Hal Luft, Ph.D.,
Deanne Wiley, C.R.A., Julie A. Schmittdiel, Ph.D.
This study was supported by a task order under AHRQ's ACTION II contract. Its findings do not reflect the view of AHRQ or any federal agency.
">Slide 3
Palo Alto Medical Foundation (PAMF)
- Multispecialty, not-for-profit ambulatory care delivery system.
- Serves nearly 1 million patients.
- Spans 4 counties in San Francisco Bay Area.
- 1500 physicians, 6700 non-MD staff.
- Majority fee-for-service:
- 70% commercial FFS.
- 12% commercial HMO.
- 13% Medicare/Medicaid.
- 5% Self-pay or Other.
Slide 4
What is "Lean"?
Set of principles, practices, and problem-solving tools that aim to improve efficiency and quality.
Founded on 2 basic principles:
- Respect for people.
- Continuous improvement (PDSA).
Key practices and tools:
- Strategic Deployment
- Translating goals to targets.
- Cross-Functional Management
- Value stream mapping.
- Daily Management
- Standard frontline work.
Standard management work.
Continuous improvement.
- Standard frontline work.
Slide 5
Implementation of Lean in Primary Care
Sequence of spread across primary care clinics:
- 1 "pilot" clinic in 1 region.
- 3 "beta" test clinics in 3 regions.
- 13 remaining clinics across all 4 regions.
Lean-based improvements:
- Value Stream Mapping.
- Call Management.
- 5S of Workspace.
- Workflow ("Flow") Redesign.
"Flow" Redesigns:
- Co-location of MD/MA dyads.
- Daily huddles.
- Agenda setting.
- In-basket management.
Slide 6
Conceptual Framework
- Identify contextual factors impacting frontline acceptance and reported adoption of Lean redesigns in primary care.
Consolidated Framework for Implementation Research modified for studying Process Redesign (CFIR-PR)
Diagram of modified framework for studying process redesign. A circle is surrounded by Outer Setting. The circle has four quadrants: Intervention Characteristics, Individual/Team Characteristics, Inner Setting, and Process of Implementation. The outer setting leads to measures of implementation and then to outcomes.
Examples of CFIR-PR domains
- Outer Setting: external pressures and policy changes.
- Intervention Characteristics: co-location, standardizations of tasks.
- Individual/Team Characteristics: physician and staff roles/scope of work.
- Inner Setting: networks and communications, culture.
- Process of Implementation: implementation style, execution, employee engagement.
- Measures of Implementation: acceptance, adoption, appropriateness, fidelity, adaptability, cost.
- Outcomes: efficiency, clinical quality, physician productivity, workforce and patient satisfaction.
Source: Rojas Smith L, Ashok M, Dy SM, Wines RC, Teixeira-Poit S. Contextual Frameworks for Research on the Implementation of Complex System Interventions. Methods Research Report. (Prepared by the RTI International—University of North Carolina at Chapel Hill Evidence-based Practice Center under Contract No. 290-2007-10056-I). AHRQ Publication No. 14-EHC014-EF. Rockville, MD: Agency for Healthcare Research and Quality; March 2014. www.effectivehealthcare.ahrq.gov/reports/final.cfm.
Slide 7
Qualitative Data Sources
In-depth interviews (N=113):
- Physicians.
- Clinic leaders.
Focus Groups (N=11 groups, 3-6 members each):
- Medical Assistants.
Observations (N=20):
- Events.
- Workflows.
In-depth Interviews by Professional Role
| Contacted | Interviewed | |
|---|---|---|
| Frontline Physicians | ||
| Family Practitioners | 72 | 26 |
| Internists | 58 | 19 |
| Pediatricians | 47 | 24 |
| Organizational Leaders | ||
| Physician Leaders | 21 | 21 |
| Operations Leaders | 24 | 23 |
| Total interviews | 222 | 113 |
Slide 8
Qualitative Analytic Methods
- Coding of fieldnotes and transcripts:
- Codes emerged from data.
- Refined codes to align with CFIR-PR domains.
- Focused analysis on CFIR-PR Measures of Implementation:
Acceptance, reported Adoption of Lean redesigns.- Categorized interviews according to implementation
- Compared differences between context, acceptance, adoption.
- Atlas.ti software used to manage data and facilitate analysis.
Slide 9
Measures of Implementation
Study results focused on two implementation measures:
- Acceptance – Degree to which those impacted by the Lean change effort viewed the changes as acceptable in principle.
- Adoption – The reported adoption, attempt to adopt, or conversely, abandonment of Lean redesigns in practice.
Slide 10
Outer Setting
- External pressures and policy changes influenced the acceptance of Lean as a solution for primary care.
"Burning platform," "Hamster health care"
"Just grinding out patients as a primary care doc…it feels like emptying the ocean with a teaspoon. The psyche of being a primary care doctor these days has got to get better…" "It's hard to be on a treadmill…" – Physician
"The burning platform was really our affordability targets and how are we going to weather [the year] when we come upon it." – Clinic Leader
Slide 11
Intervention Characteristics
- Co-location of care teams had a positive impact on frontline experiences with Lean redesigns.
- Physician-Medical Assistant (MA) dyad now sit side-by-side to facilitate communication, patient care workflows.
"It's really a teachable moment too…we're finding that the physicians are saying, 'Oh, you know that patient that had X, Y, and Z…this is what the diagnosis is and this is what it means,' or 'Here are some symptoms to look out for.' So, it's a really good opportunity for that dyad to have teaching." - MA Supervisor
- Physician-Medical Assistant (MA) dyad now sit side-by-side to facilitate communication, patient care workflows.
Slide 12
Intervention Characteristics
- There were challenges to accepting Lean standardization of some workflows and care processes.
"You have to say please trust me because if we all do it the same way and we all follow the same rules…then the whole team can perform at an optimum level from the patient service representative, to the doctor and everyone in between, and you not only get back more time, you build a better care, you can see more patients, and you feel better about coming to work." – Clinical Director
Slide 13
Process of Implementation
- Engaging frontline employees in developing Lean redesigns is a critical step.
"[I think for Lean to be successful] …make sure that the doctors and the staff continue to have a say in what happens. That's always a big concern is that… people are worried things just happen from above and we're losing control." – Internist
- When participants were involved in developing new work processes or had greater flexibility in adapting redesigns, they reported being more willing to at least "try out" Lean redesigns.
Slide 14
Inner Setting
- Matching implementation style to existing organizational cultures:
- Top-down style not as effective in clinics with "democratic, non-hierarchical" cultures.
- Micro-cultures, largely fostered by local leadership, can powerfully affect implementation efforts.
- Even when highly skeptical of Lean, faith in the department's leadership, and leaders' willingness to be flexible and open, provided an environment where members felt that they should at least "give Lean a chance."
Slide 15
Characteristics of Individuals and Teams
- Work roles and relationships between care team members were dramatically changed.
- Required skillsets, competence of MA as newly designated "Flow Manager."
- MD compliance with Lean redesigns affected team's ability to adopt the new workflows.
Slide 16
Characteristics of Individuals and Teams
- Physician autonomy and adherence:
- Those most resistant to Lean believed they were already highly efficient.
- Some were concerned that Lean threatens their autonomy; others acknowledged they still had authority where it matters most (in exam room):
"I don't feel like my work has changed so much that I'm not in control. I still decide what I'm doing with my patients. It's just that Lean presents my patients to me in a nicer way so that I can do my work better." – Physician
Slide 17
Summary of Qualitative Findings on Implementation
- Outer setting had most impact on frontline acceptance of Lean in principle:
- Market pressures, Overwhelming demand in primary care
- Other contextual factors played critical roles in adoption of Lean in practice:
Intervention characteristics:
- Co-location.
- Standardization.
Process of Implementation:
- Top-down vs. Bottom-up.
- Employee engagement.
Inner setting:
- Organizational culture.
- Local leadership.
Individuals and Teams:
- Work roles & relationships.
- Physician autonomy.
- Hung DY, Gray CP, Martinez MC, Schmittdiel J, Harrison MI. Acceptance of Lean Redesigns in Primary Care: A Contextual Analysis. Health Care Management Review. 2016 Mar 2. [Epub ahead of print]
Slide 18
Impact of Lean on System Performance
- Quantitative analysis of operational metrics indicating Lean's impact in primary care.
- Performance areas examined:
- Workflow Efficiency ("Flow" metrics).
- Physician Productivity.
- Operating Expenses.
- Clinical Quality.
- Patient Satisfaction.
- Physician and Staff Satisfaction.
Slide 19
Quantitative Methods
- Data sourced from dashboards, billing, quality reports, Experience of Work, AMGA, and Press-Ganey surveys.
- Generalized linear mixed models with physician-month as unit of observation (N=328 MDs employed from 2011-2014).
- Estimated overall impacts over time using interrupted time series analysis, non-randomized stepped wedge design.
- Phased implementation of Lean:
Projected values ("counterfactual" in the absence of Lean) vs.
Observed values after Lean redesigns were implemented in all clinics across the system.
Slide 20
Phased Implementation of Lean Redesigns
Bar chart showing schedule of implementation from pilot in September/October 2011 through 3 beta tests and 7 clinics running through May/June 2014.
Slide 21
Workflow Efficiency (Office Visit Charts Closed < 2 hours)
Two charts showing percentage of office visit charts closed in under 2 hours for pilot, 3 beta tests, and 7 clinics from January 2011 through May 2014.
Slide 22
Workflow Efficiency
| Flow Metric | Projected Value (absent Lean) | Observed Value (with Lean) | Mean Difference (95% bootstrap) |
|---|---|---|---|
| Office Visit < 2h | 51.2% | 56.2% | 5.0%* |
| E-messaging < 4h | 79.5% | 77.7% | -1.9% |
| Rx Renewal < 4h | 63.4% | 71.4% | 8.0%* |
| Telephone < 4h | 57.3% | 62.4% | 5.1%* |
Flow Metrics
- Office visit charts closed within 2 hours.
- Electronic patient messages responded within 4 hours.
- Prescription refills renewed within 4 hours.
- Telephone encounters closed within 4 hours.
Slide 23
Physician Productivity
| RVU Metric | Projected Value (absent Lean) | Observed Value (with Lean) | Mean Difference (95% bootstrap) |
|---|---|---|---|
| wRVU/cFTE | 252.3 | 265.0 | 13.9* |
| wRVU/visit | 1.5 | 1.5 | 0.0 |
wRVU: work Relative Value Unit
cFTE: clinical Full-Time Equivalent
- RVUs restated to CMS 2014 v2 valuation.
- wRVU/cFTE: Production per clinical FTE.
- wRVU/visit: Service intensity per office visit.
Slide 24
Clinical Quality Metrics
- Pay for performance clinical quality metrics for each physician over time.
- Interrupted time series analysis on metrics that had an initial statistical difference pre- vs. post-Lean:
- Coordinated Diabetes Care: A1c < 8.0%.
- Coordinated Diabetes Care: A1c < 7.0%.
- Coordinated Diabetes Care: LDL-c < 100 mg/dL.
- Coordinated Diabetes Care: Nephropathy Screening.
- Cervical Cancer Screening, Asymptomatic Women.
- Chlamydia Screening in Women (16-20 yo).
- Adolescent Immunizations: Meningococcal.
Slide 25
Clinical Quality Metrics
| Quality Metric | Projected Value (absent Lean) | Observed Value (with Lean) | Mean Difference (95% bootstrap) |
|---|---|---|---|
| Diabetes: A1c Control < 7.0% | 64.5% | 67.9% | 3.4%* |
| Diabetes: A1c Control < 8.0% | 35.5% | 39.4% | 3.9%* |
| Diabetes: LDL < 100 mg/dL | 48.1% | 53.1% | 5.0%* |
| Diabetic Nephropathy Monitoring | 75.7% | 79.9% | 4.2%* |
| Cervical Cancer Screening | 71.9% | 71.1% | -0.8% |
| Chlamydia Screening 16-20 | 61.7% | 60.7% | -1.0% |
| Immunizations - Meningococcal | 77.9% | 69.0% | -8.9%* |
Slide 26
Patient Satisfaction
- For each physician, examined proportion of satisfaction scores > 90% for each domain and for composite overall score.
- Patient satisfaction domains:
- Access.
- Care Provider.
- Moving Through the Visit.
- Nurse/Medical Assistant.
- Handling of Personal Issues.
- Composite Overall Score.
Slide 27
Patient Satisfaction
| Domain (proportion of 90% satisfied or higher) |
Project Value (absent Lean) |
Observed Value (with Lean) |
Mean difference (95% bootstrap) |
|---|---|---|---|
| Composite Score | 49.1% | 63.2% | 14.1%* |
| Access | 37.4% | 55.4% | 18.1%* |
| Care Provider | 79.0% | 69.8% | -9.2%* |
| Moving through Visit | 50.9% | 49.3% | -1.6% |
| Nurse/MA | 66.2% | 68.0% | 1.7% |
| Personal Issues | 69.0% | 74.5% | 5.5%* |
Slide 28
Physician Satisfaction % Differences (2011 vs. 2014)
By phase of implementation
Bar chart showing percentage difference between 2011 and 2014 in physician satisfaction by phase of implementation (time since completion of Lean redesigns):
Pilot (25 months), Beta (13-15 months), All remaining (4-11 months)
Leadership and communication:
pilot, 9, beta, minus 3, all remaining minus 2; time spent working: pilot, 3, beta, 4, all remaining, minus 2; relationship and staff: pilot, 1, beta, 6; overall satisfaction, pilot and beta, 2, all remaining, minus 4
Staff Satisfaction % Differences (2011 vs. 2014)
All primary care clinics system-wide
Bar chart showing percentage difference between 2011 and 2014 in staff satisfaction for all primary clinics systemwide Composite score, 3; credible leadership, 5; employee engagement, 4; connection to purpose, 4; empowerment and autonomy, 4; respect and recognition, 3; work, structure, and process, 3
Slide 29
Summary of Lean Impacts
| Topic | Conclusions |
|---|---|
| Workflow Efficiency | Increase in timeliness of completing 3 of 4 workflow measures: office visit chart closures, medication renewals, telephone responses. |
| Physician Productivity | Higher wRVUs generated per physician per month. No change in wRVUs per office visit (service intensity) |
| Operating Expenses | Lower total operating expenses (including staff compensation, and drugs and supply costs) standardized per tRVU. Not significant at p<0.05. |
| Clinical Quality | Improvements in coordinated diabetes care metrics, no change in preventive screening metrics, and decreased meningococcal immunization among adolescents. |
| Patient Satisfaction | Higher satisfaction overall and in specific domains, including access to care and handling of personal issues. Lower satisfaction with interactions with care providers. |
| Physician Satisfaction | In pilot and beta clinics: Higher satisfaction overall and in specific domains, including time spent working and relationships with staff. Lower satisfaction overall in last phase of gamma clinics to implement Lean. |
| Staff Satisfaction | Higher satisfaction overall and in specific domains, including credible leadership, employee engagement, growth / development, connection to purpose, healthy partnerships, empowerment and autonomy. |
Hung DY, Harrison MI, Martinez MC, Luft HS. Scaling Lean in Primary Care: Impacts on System Performance. American Journal of Managed Care, Forthcoming March 2017.
Slide 30
Conclusions
- Implementation context matters:
- Process of scaling across multiple sites.
- Frontline engagement.
- Alignment with internal clinic environments.
- Overall, beneficial effects of Lean redesigns on performance without harm to clinical quality.
- Using Lean techniques to redesign care delivery:
- Attention to "flow."
- Change management: involve providers in design and also results.
