Patient Safety & Systems-Based Practice

Population health management and risk stratification

Clinical Overview and When to Suspect Population Health Gaps

— Risk stratification: classify panel into low/rising/high/very-high risk tiers

— Care gap closure: identify overdue preventive services, uncontrolled chronic disease, missed transitions

— Targeted intervention: match intensity of resources (care manager, pharmacist, social worker) to risk tier

— Practice/clinic with low screening rates (mammography, colonoscopy, HbA1c, BP control)

— Rising 30-day readmission rate or ED utilization in a defined panel

— ACO/value-based contract where quality metrics (HEDIS, MIPS, STAR) are below benchmark

— Disparities in outcomes by race, language, or zip code within a panel

— Triple Aim (better care, better health, lower cost) → expanded to Quadruple Aim adding clinician well-being, and Quintuple Aim adding health equity

— Chronic Care Model: self-management support, delivery system design, decision support, clinical info systems, community resources, health system

— Kaiser/PHM pyramid: 70-80% low-risk (wellness/prevention) → 15-20% rising-risk (disease management) → 3-5% high-risk (complex case management) → <1% very-high (palliative/intensive)

Board pearl: When a stem describes a clinic with poor diabetes control across the panel, the answer is rarely "see each patient sooner" — it's build a registry, stratify by HbA1c, and deploy a pharmacist-led care team to the high-risk tier. Step 3 rewards system-level solutions over individual heroics.

Definition: Population health management (PHM) is the proactive, data-driven care of a defined patient panel to improve outcomes, close care gaps, and reduce cost — distinct from episodic, reactive office visits.

Core triad:

When to "suspect" a population health problem in a Step 3 stem:

Frameworks you must recognize:

Data sources: EHR registries, claims data, HIE feeds, ADT (admit-discharge-transfer) alerts, social determinants screens, patient-reported outcomes.

Presentation Patterns and Key History

— "You are the medical director of a primary care clinic with 8,000 patients..." → systems-level intervention question

— "An ACO contract requires you to improve colorectal cancer screening from 52% to 70%..." → care-gap closure logic

— "Compared with neighboring clinics, your panel has higher 30-day readmissions for CHF..." → transitions-of-care intervention

— "A 58-year-old has missed three appointments and his HbA1c is 11.2%..." → rising-risk patient needing outreach, not just rescheduling

— Utilization markers: ≥2 ED visits or ≥1 hospitalization in past 6 months, polypharmacy (≥5 meds), ≥3 chronic conditions

— Adherence markers: missed appointments ("no-shows"), medication possession ratio <80%, lapsed refills

— Social determinants (SDOH): food insecurity, housing instability, transportation barriers, low health literacy, limited English proficiency, social isolation

— Behavioral health overlay: depression PHQ-9 ≥10, substance use, untreated SMI — doubles risk of poor chronic disease control

— Demographics, payer mix, prevalence of HTN/DM/CHF/COPD/CKD

— Quality metric performance vs. benchmarks

— Attribution lists from payers (who "counts" toward your metrics)

— PRAPARE and AHC-HRSN — standardized SDOH screening

— LACE index, HOSPITAL score — predict 30-day readmission risk at discharge

— CMS-HCC, Charlson, Elixhauser — comorbidity-based risk adjustment

Step 3 management: When a stem lists a patient with multiple no-shows, polypharmacy, and food insecurity, the right next step is enrollment in a care management program with SDOH navigation, not simply "schedule a follow-up appointment in 2 weeks." Recognize the rising-risk phenotype.

Typical Step 3 stem framings:

History elements that flag population health risk at the individual level:

Panel-level history (clinic data):

Screening tools:

Physical Exam Findings — Panel-Level "Vital Signs"

— Preventive: breast (50-74, q2y mammogram), cervical (21-65 per USPSTF), colorectal (45-75), lung (50-80 with 20 pack-yr LDCT), AAA (men 65-75 who smoked)

— Chronic disease control: HbA1c <8 in diabetics, BP <140/90, statin in ASCVD/DM, ACEi/ARB in CHF/CKD with proteinuria

— Utilization: 30-day readmission rate, ED visits per 1,000, ambulatory-care-sensitive admissions (CHF, COPD, DM)

— Patient experience: CAHPS surveys

— Run chart: plot metric over time; ≥6 consecutive points above/below median = non-random shift

— Statistical process control (SPC) chart: points outside 3σ = special cause variation requiring investigation

Key distinction: Common cause variation = inherent system noise (fix the system); special cause variation = identifiable event (fix the cause). A sudden drop in mammography rates after a radiologist retires = special cause. Persistently low rates across 3 years = common cause requiring redesign, not blame.

Board pearl: If a Step 3 quality-improvement vignette shows a run chart with a shift below the median for 8 consecutive months, the answer is investigate for special cause and implement a targeted PDSA, not "continue current workflow."

Population health has no stethoscope exam; instead, the "physical" is the dashboard of panel-level metrics. Step 3 expects fluency with these:

HEDIS / MIPS / STAR measure categories:

Equity stratification: all measures should be disaggregated by race, ethnicity, language, payer, zip code — unstratified data hides disparities (Quintuple Aim).

Run charts and control charts:

PDSA cycle: Plan-Do-Study-Act — small, rapid tests of change before full rollout

Benchmarking: compare to national HEDIS percentiles (75th = good, 90th = excellent), CMS STAR ratings (5-star = top tier).

Diagnostic Workup — Risk Stratification Tools and Registries

— CMS-HCC (Hierarchical Condition Categories): Medicare risk score; predicts cost. RAF (risk adjustment factor) score >2.0 = high complexity.

— Charlson Comorbidity Index: predicts 10-year mortality; score ≥5 = high risk

— Elixhauser: 30 comorbidity categories; better for inpatient prediction

— LACE index (Length of stay, Acuity, Comorbidities, ED visits): predicts 30-day readmission; score ≥10 = high risk

— PRISM, ACG, Milliman: proprietary commercial risk scores

— Recent utilization (ED visits, admits in last 6-12 months)

— Medication complexity, adherence (MPR/PDC)

— SDOH screen results, behavioral health diagnoses

— Frailty indicators (gait speed, falls, FRAIL scale)

— Diabetes: HbA1c tiers (<7, 7-9, >9), presence of CKD, CVD, neuropathy → escalate intensity

— CHF: NYHA class, recent hospitalization, BNP trajectory

— COPD: GOLD A/B/E groups by symptoms and exacerbations

— CKD: eGFR + albuminuria → KDIGO heat map (green/yellow/orange/red)

Step 3 management: A 72-year-old with DM, CKD stage 3, CHF (EF 30%), 2 ED visits, and PHQ-9 of 14 belongs in complex care management with a multidisciplinary team (pharmacist, RN care manager, behavioral health, social work), not standard quarterly visits. Recognize the very-high-risk phenotype on a stem.

Step 1: Build the registry. A registry is a structured list of patients meeting criteria (e.g., all adults with DM, all post-MI patients). Source: EHR problem list + diagnosis codes + lab/med data. Without a registry, PHM is impossible.

Step 2: Apply a risk model. Common validated tools:

Step 3: Layer on dynamic and social data:

Disease-specific stratification examples:

Output: every patient assigned to a tier (low/rising/high/very-high) with a corresponding care plan.

Diagnostic Workup — Predictive Analytics and Care Gap Identification

— Descriptive: what happened (last HbA1c, last mammogram date)

— Predictive: what will happen (probability of readmission, ED visit, decompensation in next 6 months)

— Prescriptive: what to do about it (suggested intervention)

— Preventive screening (per USPSTF A/B recommendations)

— Immunizations (influenza annually, COVID per CDC, pneumococcal age/risk-based, Tdap q10y, zoster ≥50, RSV ≥75)

— Chronic disease labs (HbA1c q3-6 mo, lipids, microalbumin, eGFR)

— Medication gaps (statin in ASCVD, ACEi in CHF, anticoagulation in AFib with CHA₂DS₂-VASc ≥2)

— Patient portal message → text/IVR reminder → letter → MA phone call → RN call → home visit

— Match intensity to risk tier; don't burn high-touch resources on low-risk gaps

Board pearl: When a stem shows a clinic with 45% colorectal screening despite physician reminders, the highest-yield next step is standing orders empowering nursing staff to send FIT kits to all eligible patients, not "physician education." System redesign > exhortation.

CCS pearl: On CCS-style cases, ordering "diabetic registry review" or "care gap report" is reasonable when managing a complex outpatient with multiple comorbidities.

Predictive vs. descriptive analytics:

Machine learning risk models increasingly used (e.g., Epic's readmission model, Optum Impact Pro). Step 3 won't ask the math but will ask how to act on the output — usually outreach by a care manager.

Care gap reports: automated EHR queries identifying patients overdue for:

Closing the gap — outreach hierarchy:

Pre-visit planning: review care gaps the day before each visit; embed in huddle. Closes 2-3× more gaps than reactive care.

Standing orders and protocols: allow MA/RN to give flu shot, order screening colonoscopy, repeat HbA1c without physician click-through.

Risk Stratification — The Population Pyramid and Tier-Based Intervention

— Tier 1 — Healthy (50-60%): focus = primary prevention, screening, immunization, lifestyle counseling. Tools: portal nudges, group visits, community programs.

— Tier 2 — At-risk (20-30%): 1-2 risk factors (HTN, prediabetes, smoking, obesity). Focus = risk factor modification. Tools: health coaching, MyPlate/DPP (Diabetes Prevention Program — covered by Medicare for prediabetes), tobacco cessation.

— Tier 3 — Rising-risk (10-15%): established chronic disease, controlled or mildly uncontrolled. Focus = disease management. Tools: pharmacist-led titration, RN education, remote monitoring (BP cuffs, CGM).

— Tier 4 — High-risk (3-5%): multiple chronic conditions, recent utilization. Focus = care coordination. Tools: longitudinal RN care manager, monthly touchpoints, transitional care management (TCM) after each discharge.

— Tier 5 — Very-high/complex (1-2%): frail elderly, end-stage disease, high utilizers. Focus = intensive case management or palliative care. Tools: home-based primary care, ACP, hospice referral when appropriate.

— Don't put a healthy 30-year-old in case management; don't leave an end-stage CHF patient with portal reminders alone.

— CCM (Chronic Care Management, 99490): ≥2 chronic conditions, 20 min/mo non-face-to-face

— TCM (99495/99496): post-discharge, contact within 2 business days, visit within 7-14 days

— AWV (Annual Wellness Visit, G0438/G0439): required HRA + care plan

— Principal Care Management, BHI, CoCM for single complex condition or behavioral integration

Step 3 management: Post-CHF discharge → schedule TCM visit within 7 days, RN phone contact within 48 hours, med reconciliation, weight log, diuretic titration plan. This bundle reduces 30-day readmission by ~25%.

The PHM pyramid (memorize for Step 3):

Matching resources to tier ("right care, right place, right time"):

Billable Medicare codes supporting PHM:

Pharmacotherapy — Population-Level Medication Optimization

— ASCVD: high-intensity statin (atorvastatin 40-80, rosuvastatin 20-40); add ezetimibe → PCSK9i if LDL >70 on max statin

— HFrEF (EF ≤40%): quadruple therapy — ARNI (or ACEi/ARB) + β-blocker (carvedilol, metoprolol succinate, bisoprolol) + MRA (spironolactone/eplerenone) + SGLT2i (dapa/empa)

— Type 2 DM with ASCVD/CKD/HF: SGLT2i and/or GLP-1 RA regardless of HbA1c

— CKD with albuminuria: ACEi/ARB + SGLT2i + finerenone (if T2DM)

— AFib with CHA₂DS₂-VASc ≥2 (men) or ≥3 (women): DOAC unless mechanical valve or moderate-severe mitral stenosis

— Post-MI: DAPT 12 mo, statin, β-blocker, ACEi if EF<40 or HTN/DM

— 90-day fills, mail order, synchronized refills, blister packs

— Combination pills (polypill) for HTN/lipids

— Pharmacist collaborative practice agreements for titration

— Generic substitution; $4 list awareness

— Beers criteria (AGS) — avoid in ≥65: long-acting benzos, anticholinergics (diphenhydramine, oxybutynin), sulfonylureas (glyburide), NSAIDs chronic

— STOPP/START — European equivalent; START flags omissions (e.g., missing statin, anticoagulant)

Board pearl: When a Step 3 stem lists a HFrEF patient on lisinopril + metoprolol only, the gap is add MRA and SGLT2i — quadruple therapy is the new standard. Closing GDMT gaps is a quality metric in most ACO contracts.

PHM "pharmacotherapy" = ensuring the right drug reaches the right patient across the panel, not individual prescribing.

Guideline-directed medical therapy (GDMT) gaps to close:

Medication adherence interventions:

Deprescribing in high-risk elderly:

Medication reconciliation at every transition (admit, transfer, discharge, AWV) — TJC National Patient Safety Goal.

Procedures / System Interventions — Care Models and Team-Based Care

— Continuous, comprehensive, coordinated, team-based, accessible, quality-focused

— Empanelment (each patient assigned to a PCP/team), 24/7 access, registry use

— Group of providers jointly accountable for cost and quality of an attributed population

— MSSP (Medicare Shared Savings), REACH ACO — upside/downside risk

— Shared savings if total cost of care < benchmark AND quality targets met

— Contact within 2 business days of discharge (phone/portal OK)

— Face-to-face visit within 7 days (high complexity) or 14 days (moderate)

— Med reconciliation, follow-up on pending labs, education

— Reduces 30-day readmits ~20-25%

— Collaborative Care Model (CoCM, billable 99492-99494): PCP + behavioral care manager + psychiatric consultant — strongest evidence for depression in primary care, ~2× remission rates

CCS pearl: For a recently discharged CHF patient, ordering "schedule TCM visit within 7 days" + "RN telephone call within 48 hours" + "home BP/weight monitoring" demonstrates Step 3-level transitions-of-care competence.

PHM's "procedures" are care delivery redesigns. Know these models:

Patient-Centered Medical Home (PCMH, NCQA-recognized):

Accountable Care Organization (ACO):

Bundled payments: single payment for episode (e.g., CJR for hip/knee, BPCI-A) — incentivizes coordination across hospital + post-acute

Transitional Care Management (TCM):

Hotspotting / high-utilizer programs: intensive outpatient interventions for top 1-5% utilizers (Camden model — note: 2020 RCT showed mixed results, but concept tested on boards)

Hospital-at-home, SNF-at-home, remote patient monitoring (RPM): CMS-reimbursed RPM requires ≥16 days of data in 30-day period

Behavioral health integration:

Group visits / shared medical appointments: diabetes, prenatal (CenteringPregnancy) — improve outcomes and access.

Special Populations — Elderly and Renal/Hepatic Impairment

— Mind: cognition (Mini-Cog, MoCA), mood (PHQ-9, GDS), delirium prevention

— Mobility: gait/balance, falls (ask all ≥65 annually; Timed Up & Go >12 sec = high risk)

— Medications: Beers, STOPP/START, deprescribing, anticholinergic burden

— Multicomplexity: multimorbidity, frailty, social context

— Matters most: patient goals, advance care planning

— HRA, cognitive screen, depression screen, fall risk, functional/ADL/IADL assessment, ACP, personalized prevention plan

— NOT a physical exam; distinct from "Welcome to Medicare" IPPE (within first 12 mo of Part B)

— Fried phenotype: weight loss, exhaustion, weakness, slow gait, low activity (≥3 = frail)

— Clinical Frailty Scale (Rockwood) 1-9: ≥5 = mildly frail, often triggers shift toward comfort-oriented goals

— KDIGO heat map drives nephrology referral (eGFR <30 or rapid decline or A3 albuminuria)

— Drug adjustments: avoid NSAIDs, dose-adjust DOACs (apixaban preferred in CKD 4-5), avoid metformin if eGFR <30, prefer SGLT2i down to eGFR 20

— Contrast/gadolinium counseling, vaccinate (Hep B, pneumococcal, influenza, COVID, RSV)

Key distinction: Treating a frail 88-year-old's HbA1c to <7% causes hypoglycemia and falls; target HbA1c 7.5-8.5% in frail elderly per ADA. Aggressive control in the wrong tier is harm, not quality.

Step 3 management: Identify frailty before applying disease-specific targets — goals must be individualized to life expectancy and patient priorities.

Geriatric PHM priorities — the 5 M's (AGS framework):

Annual Wellness Visit (Medicare, G0438 initial / G0439 subsequent):

Frailty stratification:

Renal impairment in PHM:

Hepatic impairment: avoid acetaminophen >2 g/d in cirrhosis, dose-adjust statins (rosuvastatin caution), avoid NSAIDs (variceal bleed risk), vaccinate Hep A/B.

Special Populations — Pregnancy, Pediatrics, and Equity-Focused Subgroups

— Universal early prenatal: dating US, CBC, type/screen, HIV, syphilis, HBsAg, HCV, rubella, varicella, urine culture, GC/CT

— Risk-based: TSH, diabetes (early if BMI ≥25 + risk factor)

— Universal screening 24-28 wk: 50-g GCT, repeat CBC, RhoGAM if Rh-negative

— 35-37 wk: GBS culture

— Postpartum: depression screen (EPDS) at 1, 2, 4, 6 mo; comprehensive postpartum visit by 12 weeks (ACOG "fourth trimester"); contraception, BP if HDP, glucose tolerance if GDM (6-12 wk OGTT)

— Maternal mortality disparity: Black women 3× higher mortality — equity stratification mandatory

— Well-child visit schedule: birth, 3-5d, 1, 2, 4, 6, 9, 12, 15, 18, 24, 30 mo, then annually

— Developmental screen at 9, 18, 30 mo; autism screen at 18 and 24 mo (M-CHAT-R)

— Lead screen at 12 and 24 mo (Medicaid-required), dyslipidemia 9-11 and 17-21

— Adolescent: HEADSS, depression screen ≥12, HIV once 15-21, HPV vaccine starting 9-12

— Disaggregate every measure by REaL (race, ethnicity, language) + SOGI (sexual orientation/gender identity) + disability

— Targeted universalism: universal goal, tailored strategies by subgroup

— Community Health Workers (CHWs), promotores, doulas — proven to reduce disparities

Board pearl: Postpartum follow-up shifted from "single 6-week visit" to a comprehensive postpartum care plan with contact within 3 weeks — Step 3 favorite. Especially critical for women with HDP, GDM, or postpartum depression.

Maternal population health:

Pediatric PHM (Bright Futures / AAP):

Health equity stratification (Quintuple Aim):

LGBTQ+ panel: PrEP eligibility (HIV-negative with risk factors), gender-affirming care continuity, cervical/breast screening based on organ inventory not gender marker.

Complications and Adverse Outcomes of Poor PHM

— Preventable hospitalizations (ambulatory-care-sensitive conditions: CHF, COPD, DM, asthma, HTN, dehydration)

— Avoidable ED utilization

— 30-day readmissions (CMS HRRP penalizes hospitals for excess CHF, AMI, PNA, COPD, CABG, THA/TKA readmits)

— Polypharmacy adverse events, falls, delirium

— Late-stage cancer diagnosis due to screening gaps

— Disparities widening when interventions aren't equity-stratified

— Quality bonuses lost (MIPS penalties up to ±9%, STAR rating downgrades)

— ACO shared savings forfeited; downside risk losses

— HRRP penalties up to 3% of base Medicare DRG payments

— HAC (Hospital-Acquired Condition) penalty: bottom quartile loses 1%

— Reputational and accreditation risk (TJC, NCQA)

— Cherry-picking / lemon-dropping: dismissing high-risk patients to improve scores

— Overscreening / overtreatment: PSA in 85-year-old to "close" a gap inappropriately; tight glycemic control in frail elderly causing hypoglycemia

— Documentation gaming: upcoding HCC for risk score inflation without true clinical change → False Claims Act exposure

— Workforce burnout from metric pressure (counter-Quadruple-Aim)

— Algorithmic bias (e.g., 2019 Obermeyer study — race-blind algorithm systematically under-referred Black patients because it used cost as a proxy for need)

— Must audit algorithms for equity before deployment

Key distinction: Closing a care gap appropriately ≠ blindly applying screening to every patient. USPSTF age cutoffs and shared decision-making for individuals with limited life expectancy prevent the harm of overscreening — and Step 3 will test this.

Patient-level consequences of failed risk stratification:

System-level consequences:

Unintended consequences of metric-driven care ("Goodhart's Law" — when a measure becomes a target, it ceases to be a good measure):

Health equity harms:

When to Escalate — Care Management, Specialty Referral, and Inpatient Triage

— Healthy → at-risk: add health coach, DPP, tobacco quitline

— At-risk → rising-risk: pharmacist titration, RN disease management calls

— Rising-risk → high-risk: assign longitudinal RN care manager, CCM enrollment, monthly touchpoints

— High-risk → very-high: complex care management team, home-based care, palliative consult

— Nephrology: eGFR <30, rapid decline >5/yr, A3 albuminuria, refractory HTN

— Cardiology: new HFrEF, refractory HTN despite 3 agents incl. diuretic, advanced HF for transplant/LVAD

— Endocrine: uncontrolled DM despite multiagent + insulin, type 1, MODY suspicion

— Palliative care: life-limiting illness with symptom burden or unclear goals — early palliative referral improves survival in metastatic NSCLC (Temel 2010)

— Behavioral health: PHQ-9 >20, suicidality, treatment-resistant depression after 2 SSRIs

— Chest pain with abnormal ECG/troponin

— CHF with hypoxia, hemodynamic compromise, or unresponsive to outpatient diuresis

— COPD exacerbation with hypoxia, hypercapnia, accessory muscle use

— DKA, severe hyperglycemia with ketones, hypoglycemia with altered MS

— Acute stroke symptoms (call 911, not "come to clinic")

CCS pearl: On Step 3 CCS, after discharging a CHF patient, advancing the clock to 48 hours for nurse phone follow-up and 7 days for TCM clinic visit mirrors real-world quality bundles and is the expected sequence.

Within the PHM pyramid, "escalation" means moving a patient to higher-intensity resources:

Specialty referral criteria (high-yield examples):

When to send to ED / admit (ambulatory triage):

ADT alerts: EHR notification when a paneled patient is admitted/discharged anywhere → triggers care manager outreach within 24-48h.

Key Differentials — Same-Category (Quality/Systems) Concepts to Distinguish

— Public health: entire community/society; governmental; vaccination campaigns, sanitation, surveillance (e.g., CDC, state health dept)

— Population health: defined panel (clinic, ACO, payer); clinical/healthcare-driven outcomes

— Overlap but Step 3 stems usually mean the panel-level clinical definition

— QI: local, iterative, PDSA, no IRB usually required; intent = improve local care

— Research: generalizable knowledge, requires IRB and consent (Common Rule)

— If results are intended for publication and generalization, treat as research — IRB review

— Risk stratification: clinical use — tier patients for care intensity

— Risk adjustment: financial/statistical — normalize outcomes by case-mix for fair comparison (CMS-HCC, APR-DRG)

— Disease management: single condition focus (DM, CHF), protocol-driven

— Case management: patient-centered, multi-condition, social complexity

— Care coordination: linking across settings/providers; transitions

— Screening: asymptomatic, population-based (USPSTF A/B)

— Case-finding: opportunistic during visit for other reason

— Diagnostic: symptomatic patient; different pretest probability

Board pearl: A clinic implementing a PDSA cycle to reduce no-shows does not need IRB approval; the same intervention designed to produce a generalizable peer-reviewed RCT does. The intent and dissemination plan determine the category.

Population health vs. public health:

Quality Improvement (QI) vs. Research:

Risk stratification vs. risk adjustment:

Disease management vs. case management vs. care coordination:

Screening vs. case-finding vs. diagnostic testing:

Triple/Quadruple/Quintuple Aim distinctions — know all three and what each added (well-being, then equity).

Key Differentials — Other-Category Concepts (Biostatistics & Payment Models)

— NNS (number needed to screen) for mortality benefit — colonoscopy ~1,250, mammography ~1,300 to prevent one death

— Lead-time bias, length-time bias, overdiagnosis — confound screening apparent benefit

— Sensitivity vs. specificity vs. PPV — PPV drops when prevalence drops (apply when screening low-risk groups)

— Absolute risk reduction vs. relative risk reduction — boards punish RRR-only thinking; ARR and NNT drive shared decisions

— FFS (fee-for-service): volume rewarded — root of many PHM problems

— P4P (pay-for-performance): quality bonuses on top of FFS

— Capitation: PMPM payment regardless of utilization — incentive to keep patients healthy AND risk of stinting

— Shared savings (upside only) vs. two-sided risk (upside + downside)

— Global budget / full risk: total cost accountability

— MA: capitated, risk-adjusted by HCC, plans incentivized to capture diagnoses; STAR ratings drive bonus payments

— Traditional: FFS, MIPS/MSSP for clinicians

— CMS increasingly adjusts measures for dual-eligible status, LIS, area deprivation index (ADI) — avoids penalizing safety-net providers

— Health outcomes ≈ 20% clinical care, 30% health behaviors, 40% social/economic, 10% environment — most PHM ROI is upstream of the exam room.

Key distinction: Recognize when a stem is asking about statistical interpretation of a screening program vs. systems design of how to deliver it — both fall under PHM but have different right answers.

Biostatistics for PHM measures:

Payment models (alternative payment models, APMs):

Medicare Advantage (MA) vs. Traditional Medicare:

HEDIS, CAHPS, HOS: quality, experience, and health-outcome surveys feeding STAR ratings

Social risk vs. clinical risk:

Determinants framework:

Secondary Prevention and Long-Term Panel Plan

— Antiplatelet (ASA 81 mg ± P2Y12 per indication)

— High-intensity statin → LDL <70 (consider <55 in very high risk)

— BP <130/80 (ACC/AHA), ACEi/ARB if HFrEF, CKD, DM

— Cardiac rehab referral (Class I; underused — major care gap)

— Smoking cessation, Mediterranean/DASH diet, 150 min/wk moderate activity

— SGLT2i / GLP-1 RA if DM or HFrEF

— HbA1c q3-6 mo, BP each visit, lipids annually, eGFR + UACR annually, dilated eye exam annually (q2y if low risk), comprehensive foot exam annually, dental q6 mo

— Vaccines: influenza, COVID, pneumococcal (PCV20 or PCV15+PPSV23), HepB if <60, RSV ≥75, Tdap, zoster ≥50

— CoCM enrollment, measurement-based care (PHQ-9 q month until remission, then q3 mo)

— Bill ACP code 99497 (first 30 min) at AWV or any visit

— POLST/MOLST for seriously ill with limited prognosis

— Healthcare proxy documented for every adult panel patient

Step 3 management: Cardiac rehab referral after MI/PCI/CABG is a Class I, often-missed quality measure — when a stem says "discharged stable after STEMI, what's next?" the answer frequently includes referral to cardiac rehab along with the GDMT bundle.

Cardiovascular secondary prevention bundle (post-ASCVD event):

Stroke secondary prevention: antiplatelet or DOAC (if AFib), statin, BP control, carotid eval if applicable

Cancer survivorship: survivorship care plan, surveillance schedule, cardio-oncology if anthracycline/trastuzumab, second-malignancy screening, psychosocial support

Diabetes long-term plan:

Behavioral health continuity:

Advance care planning:

Follow-Up, Monitoring Parameters, and Counseling Cadence

— HTN controlled: q3-6 mo; uncontrolled: q2-4 wk until at goal

— DM at goal: q6 mo with HbA1c; off-target: q3 mo

— HFrEF stable: q3-6 mo; recent decompensation: q1-2 wk during titration

— CKD stage 3: q6-12 mo; stage 4: q3 mo; stage 5: monthly with nephrology

— Depression: PHQ-9 monthly until remission, then q3 mo for 6-12 mo

— BP: annually if normal; more often if elevated

— Lipids: q4-6 yr if low risk, more often if on therapy

— Colorectal: colonoscopy q10y, FIT annually, Cologuard q3y (start 45)

— Mammography q2y 50-74 (USPSTF 2024 lowered to start at 40 q2y)

— Cervical: cytology q3y 21-29; co-test q5y or cytology q3y 30-65

— Lung LDCT annually 50-80 if 20 pack-yr and quit <15 yr

— Bone density q2y in women ≥65 (earlier if risk)

— AAA US once in men 65-75 ever smoked

— 5 A's (Ask, Advise, Assess, Assist, Arrange) for tobacco

— SBIRT for alcohol/substance use

— Motivational interviewing for behavior change

— Physical activity prescription (FITT: frequency, intensity, time, type)

— Diet: Mediterranean and DASH have strongest CV outcome data

Board pearl: When a stem asks "next best step" in a stable HFrEF patient discharged 5 days ago, the answer is the TCM visit at 7 days plus PHQ-9 (depression is common and treatable in CHF, doubles mortality).

Standard follow-up intervals for chronic disease (well-controlled → uncontrolled):

Preventive cadence (USPSTF, ACS, CDC):

Counseling — high-yield:

Patient-reported outcomes (PROMs): PHQ-9, GAD-7, PROMIS, KCCQ in CHF — track over time as "vital signs."

Ethical, Legal, and Patient Safety Considerations

— Up to 50% of patients have a medication discrepancy at hospital discharge

— Mandatory med reconciliation at every transition (TJC NPSG.03.06.01)

— Discharge summary to PCP within 48 h; failure is both quality and malpractice risk

— Pre-diagnosis genetic testing/screening: discuss insurance implications (GINA protects health insurance/employment but NOT life, disability, long-term care insurance)

— Mass-outreach interventions (text-message campaigns): require opt-in for HIPAA-compliant PHI sharing; opt-out for general health education

— Algorithm-driven risk scores: patients deserve to know if AI/risk score influenced care decisions

— Suspected child/elder/dependent-adult abuse — report to APS/CPS without consent; immunity if good-faith

— Reportable communicable diseases (TB, syphilis, HIV new dx, measles, pertussis) to local health dept

— Impaired driving risk (varies by state; CA, OR, PA mandate physician reporting of certain conditions)

— Gunshot/stab wounds, suspicious burns

— 42 CFR Part 2: SUD treatment records — stricter than HIPAA; specific consent required

— Adolescent confidentiality: contraception, STI, mental health, SUD — protected in most states; billing/EOB can inadvertently breach

— Risk-stratification algorithms must be audited for bias (Obermeyer 2019)

— Section 1557 ACA — language access, disability accommodation

— Don't perform PSA in an 85-year-old with dementia "to close a gap"

— Don't push HbA1c <7 in a frail elder — Beers/Choosing Wisely violation

Step 3 management: When a clinic's CHF readmission rate spikes, the response is a root-cause analysis (RCA) with a Just Culture lens, not disciplinary action against the discharging hospitalist.

Transitions of care — the highest-risk PHM moment:

Informed consent edge cases in PHM:

Mandatory reporting (Step 3 staples):

Confidentiality:

Equity / non-discrimination:

No-harm in metric chasing:

Just Culture: distinguish human error, at-risk behavior, reckless behavior — supports reporting and learning, not blame.

High-Yield Associations and Rapid-Fire Clinical Facts

Board pearl: Step 3 loves the TCM bundle, AWV components, Beers list, and Quintuple Aim — memorize all four.

PHM pyramid percentages: ~70% healthy/at-risk, ~20% rising/high-risk, ~5% complex, ~1% catastrophic

LACE ≥10 = high readmission risk

TCM visit: contact ≤2 business days, visit ≤7 (high) or ≤14 (moderate) days

CCM: ≥2 chronic conditions, ≥20 min/mo non-face-to-face, billable 99490

Triple → Quadruple → Quintuple Aim: + clinician well-being → + equity

80% of health outcomes driven by SDOH + behaviors, not clinical care

DPP (Diabetes Prevention Program): Medicare-covered for prediabetes; ~58% reduction in progression to DM

HRRP-penalized conditions: AMI, CHF, PNA, COPD, CABG, THA/TKA

CMS STAR ratings: 5-star = top tier; drives MA enrollment and bonus

HEDIS Comprehensive Diabetes Care: HbA1c control, eye exam, nephropathy screen, BP <140/90

USPSTF Grade A/B: must be covered without cost-sharing (ACA preventive services)

Beers criteria: updated 2023; high-yield: anticholinergics, benzos, sulfonylureas (glyburide), NSAIDs chronic, PPIs >8 wk without indication

Obermeyer 2019 bias study: cost-as-proxy-for-need under-referred Black patients — algorithmic equity audit required

CoCM (Collaborative Care): 99492/93/94 — strongest evidence for depression in PC; ~2× remission

Goodhart's Law: measure becomes target → ceases to be a good measure; metric gaming

Annual Wellness Visit ≠ physical exam — HRA + care plan + cognitive/fall/depression screen + ACP

GINA protects health insurance & employment; NOT life/disability/LTC

Cardiac rehab Class I after MI/PCI/CABG/HFrEF — major underused metric

Postpartum: comprehensive visit by 12 wk, with first contact by 3 wk (ACOG)

Vaccines ≥65: annual flu, COVID, PCV20 or PCV15+PPSV23, RSV ≥75, Tdap q10y, zoster (Shingrix ≥50)

Board Question Stem Patterns

— "Your clinic's colorectal screening is 45% vs. benchmark 70%." → answer: system intervention (standing orders for FIT, mailed outreach, navigator), not "remind physicians."

— Patient with multiple chronic conditions + recent admissions + SDOH + behavioral health. → enroll in complex care management / CCM.

— Patient discharged 3 days ago. → TCM visit within 7 days + medication reconciliation + RN call within 48h.

— Run chart shows 8 consecutive points below median. → special cause; investigate and PDSA.

— 82-year-old with dementia, low function — "should we order PSA/mammogram/colonoscopy?" → no; life expectancy < benefit horizon; shared decision and discontinue.

— Frail elder on glyburide with HbA1c 6.2% and fall. → stop glyburide (Beers); relax HbA1c target to 7.5-8%.

— Diabetes control disparity by race in your panel. → stratify data by REaL, deploy CHW/language-concordant care, audit algorithms.

— Local intervention intended only for internal improvement. → QI, no IRB. If publishing for generalization → IRB review.

— ACO context, downside risk, quality not met. → must improve both cost and HEDIS measures; shared savings forfeited otherwise.

— Risk model under-flags minority patients. → audit for bias, retrain on outcome-based labels not cost (Obermeyer).

— Missed visits and screens. → schedule per ACOG / Bright Futures cadence, send navigator if SDOH barriers.

Step 3 management: When two answer choices both seem reasonable, pick the system-level intervention over the individual-level reminder — that's the PHM voice the test rewards.

Pattern 1 — Care gap closure:

Pattern 2 — Risk stratification:

Pattern 3 — Transitions of care:

Pattern 4 — Quality measurement:

Pattern 5 — Overscreening harm:

Pattern 6 — Deprescribing:

Pattern 7 — Equity:

Pattern 8 — QI vs. research:

Pattern 9 — Payment models:

Pattern 10 — Algorithmic bias:

Pattern 11 — Postpartum / pediatric well-care gap:

One-Line Recap

Rapid recap bullets:

Board pearl: On Step 3, when a vignette describes a clinic, panel, ACO, or health system rather than a single patient encounter, default your thinking to registry → risk tier → team-based intervention → measurement → equity audit — that sequence will land you on the right answer more often than any drug or dose choice ever will.

Population health management is the disciplined practice of stratifying a defined panel by clinical and social risk, closing care gaps with the lowest-intensity intervention that works, and reserving high-touch resources for the small fraction of high-risk patients where they change outcomes.

Stratify, then match resources: the PHM pyramid (healthy → at-risk → rising → high → very-high) drives every Step 3 systems answer; don't put a healthy 30-year-old in case management or leave an end-stage CHF patient on portal reminders.

Close care gaps with systems, not exhortation: standing orders, registries, pre-visit planning, team-based care (PCMH, CoCM, pharmacist titration), and TCM bundles outperform "remind the doctor" every time — and align with HEDIS, MIPS, STAR, and ACO quality contracts.

Equity and avoidance-of-harm are co-equal with metrics: disaggregate data by REaL/SOGI/SDOH (Quintuple Aim), audit algorithms for bias, and refuse to overscreen the frail or deprescribe-resistant elder just to "close a gap" — Beers, USPSTF age cutoffs, and shared decision-making protect against metric-driven harm.

Transitions of care are the highest-leverage moment: post-discharge contact within 2 business days, TCM visit within 7-14 days, full medication reconciliation, and ADT-triggered care manager outreach reduce 30-day readmissions by ~20-25% and protect patients from the discharge cliff that drives most ambulatory-care-sensitive harm.