Clinical Decision Support A Population Health Approach Farzad

Clinical Decision Support A Population Health Approach Farzad Mostashari, MD, MSc NYC DOHMH fmostash@health. nyc. gov

Quality of Care With and Without EHR Use No Benzo in Depression Tobacco cessation Statin Use

Government Role? • Will HIT address priority public health issues? • Will HIT adoption reach disadvantaged populations?

Primary Care Information Project • Mission – Improve population health in medically underserved areas through health information technology (HIT) • Resources – NYS- $250 Million over 5 years for HIE – NYC- $27 million Community EHR Project – CDC- Center of Excellence PH Informatics

PCIP Roadmap 1. Citywide EHR Network • • 2. Procure “best of breed” Electronic Health Record Add Public Health/Quality Improvement functionality Prepare Network and Hardware Infrastructure Extend to Correctional Health and community providers Citywide Quality Improvement Network • • • 3. Citywide automated quality measurement and reporting Decision supports and other quality improvement Extend to other ambulatory EHRs Citywide Health Information Exchange Network • • • Interfaces to other systems (e. g. , HHC) Syndromic and Notifiable Disease Reporting Citywide Immunization, Lead Registries, and School Health Linkages to RHIOs Medicaid medication history

HEAL-NY 1 • Impact – All NYC Community Health Centers will have EHRs by 2009 • 648 providers • 500, 000 patients • 50% Medicaid, 20% uninsured – Safety net providers “RHIO”

Other Providers • Small doctors offices – Deliver 85% of primary care – Lowest EHR adoption rates – Greatest challenges in quality and financing • Convening and support – – Medical Societies QIO (IPRO) Hospitals Health Plans?

Eligibility and Public Purpose • Care for underserved / vulnerable populations – Medicaid and uninsured – S Bronx, Harlem, Central Brooklyn • Participate in public health goals – Mandatory indicator reporting (automated, confidential) – Quality improvement (inc. decision support tools) – Public health interfaces (school health, CIR) • Financial Commitment – $4 k per provider contribution to QI fund – Assume all ongoing costs after 2 yr testing phase

What do they get? • • Licenses to “NYC Build” e. Clinical. Works Common interfaces included On-site training Quality improvement technical assistance • Predictable, low ongoing (M&S) costs – Less than $1, 500 per FTE provider/ yr

1. 10 Take New York Indicators Have a Regular Doctor or Other Health Care Provider 2. Be Tobacco-Free 3. Keep Your Heart Healthy 4. Know Your HIV Status 5. Get Help for Depression 6. Live Free of Dependence on Alcohol and Drugs 7. Get Checked for Cancer 8. Get the Immunizations You Need 9. Make Your Home Safe and Healthy 10. Have a Healthy Baby

Smart Web Form • Facilitates n-level structured data collection • Built-in intelligence to make calculations based on data entered – – – – Initial Visit form Tobacco Quit Readiness Assessment form Fax-to-Quit form Asthma Severity Assessment form PHQ 2 and PHQ 9 for Depression Screening AUDIT-C for Alcohol Misuse Screening School Health New Admission Examination form Sexual History form

Registry

Measure Reporting • User Interface to run measure report • Can view with/ without exclusions • Cross tab: facility, provider, insurance & race • Drill-down capability

Decision Support Tools • Based on TCNY measures • Passive alerts and reminders – – Wary of “alert fatigue” Minimal set Actionable (Order, Historical Order or Suppress) Consonant with workflows, not disruptive • Not just alerts – Order sets, templates, clinical knowledge, data presentation, process reengineering

Desired Attributes • Acceptable to small independent providers • Priority health issues (premature deaths) comprehensively addressed • Could be implemented in multiple EHR systems • Mechanism for updating as evidence base changes

Clinical Decision Support – Tobacco Best Practice Alert

Actionable, nonintrusive alert will show on the right -pane.

Measure Defn as CDSS Logic? • • Potential solution to portability of CDSS Consistent message of “what’s good care” Providers “clean” measures as they go But can’t afford lots of “false positives” – Need CDSS Numerator inclusions • “snooze” on order • Incorporates epidemiologic info (never smoker > 26 yo) – Denominator exclusions • Patient refusal, medical contraindication, system reasons

Order Set • Triggered by TCNY quality indicator • Full order set vs. Quick order – Quick - lab • A 1 C testing in patients with diabetes (6 months) • LDL testing in high risk patients (IVD, DM) – Full - medication, referral, education (provider/patient) • A 1 C control in patients with diabetes (good control) • BP control in high risk (130/80) patients (IVD, DM) • LDL control in high risk patients (IVD, DM)

Other Decision Supports • • • Order Sets Templates Adverse drug event alerts Data display and visualization Pricing and formulary information

e. Med. NY Allows treating physicians to access to their respective Medicaid patients’ Medication history at the point of care, at the time of treatment Eliminates steps from the practice workflow by linking Medicaid Eligibility checking directly into e. Clinical. Works’ practice management

“Not Just Alerts” • Practice workflow reorganization – Structured data collection – Registries and panel management – Alternative visit types – Team-based care – Case management – Patient education and self-management

Key Features • Framework for comprehensive, but limited, set of evidence-based interventions • Measure definition= CDSS trigger logic • Incorporation of epidemiologic information • Underpins a distributed query architecture

Future Directions • Human-Computer interface optimization • Extension of concepts to other EHRs • Establish architecture for distributing new measures and order sets • Extension to Public Health alerts and reporting • Evaluation – AHRQ, NORC

DOH receives signal of outbreak of respiratory illness in young adults Cough Practice Alert in EHR for age 18 -45, relevant Sx/Dx, requests provider to do nasopharyngeal wash and call DOH for immediate pick-up and viral ID by DOH lab