Biostatistics & Population Health

Surveillance systems and public health data sources

Clinical Overview and When to Suspect Surveillance Gaps

— Passive surveillance: providers/labs report to health departments (e.g., notifiable diseases) — cheap, broad, but underreports

— Active surveillance: health department contacts providers/labs (e.g., during outbreaks, measles, polio) — more complete, resource intensive

— Sentinel surveillance: selected reporting sites monitor specific conditions (e.g., ILINet for influenza-like illness)

— Syndromic surveillance: real-time symptom/ED chief-complaint data (BioSense) — early outbreak detection before lab confirmation

— Stem mentions "state health department," "CDC notification," "reportable disease," cluster of cases, or asks "best data source to estimate prevalence/incidence"

— Outbreak investigation steps (Snow-style)

— Vaccine safety signal → think VAERS

— Cancer incidence trends → SEER

— National prevalence of chronic disease/risk factors → BRFSS or NHANES

Public health surveillance = ongoing, systematic collection, analysis, and interpretation of health data essential to planning, implementation, and evaluation of public health practice (CDC definition)

Purpose on Step 3: recognize which data source answers a given clinical/epidemiologic question and when a clinician must report a case

Core categories

When to suspect a surveillance question on the exam:

Board pearl: Surveillance answers population-level questions; clinical trials answer efficacy questions. If the stem asks "how would you monitor adverse events after FDA approval of a vaccine," the answer is VAERS / VSD, not an RCT.

Step 3 management: Suspected case of a notifiable disease (e.g., measles, TB, syphilis, pertussis, N. meningitidis) → call the local health department immediately, do not wait for lab confirmation. Reporting is mandatory and provider-initiated; failure can carry licensure consequences.

Presentation Patterns and Key History — How Surveillance Data Arrive

— Clinician diagnoses or suspects a nationally notifiable condition (NNDSS list maintained by CDC/CSTE)

— Examples: measles, mumps, rubella, pertussis, Neisseria meningitidis, hepatitis A/B/C, HIV (in all states), TB, syphilis, gonorrhea, chlamydia, Lyme, Zika, rabies exposure, foodborne (Salmonella, Shigella, E. coli O157, Listeria), botulism, anthrax, plague, smallpox, viral hemorrhagic fevers

— Mechanism: phone/fax/electronic to local or state health department → state → CDC

— Most reportable infections also trigger lab reporting (positive culture, PCR, serology) — redundancy improves capture

— BRFSS (telephone, adult risk factors), YRBSS (high schoolers), NHANES (exam + labs), NHIS (household interview), NSDUH (substance use), PRAMS (postpartum)

— SEER (cancer, ~48% US coverage), USRDS (ESRD), National Trauma Data Bank

Surveillance data flow into clinicians' workflow in predictable patterns

Provider-initiated reporting

Laboratory-initiated reporting

Death certificate data → National Vital Statistics System (NVSS); feeds leading-cause-of-death statistics

Birth certificate data → natality, congenital anomaly surveillance

Hospital discharge data → HCUP/NIS for inpatient utilization, NEDS for ED visits

Survey-based (no clinical encounter required)

Registry-based

Key distinction: Incidence questions → use case-based surveillance (NNDSS, SEER). Prevalence of behaviors/conditions → use cross-sectional surveys (BRFSS, NHANES). Mortality → NVSS. Adverse drug/vaccine events → FAERS / VAERS.

Board pearl: If a stem asks "what proportion of US adults have measured HbA1c ≥6.5%?" — the answer source is NHANES (it includes physical exam + labs), not BRFSS (self-report only).

Physical Exam Findings — Anatomy of a Surveillance System

— Proportion of true cases detected by the system

— Low sensitivity = underreporting (typical of passive systems)

— Proportion of reported cases that are true cases

— Drives resource use; low PPV wastes investigations

— Categories: suspected, probable, confirmed (e.g., measles confirmed = lab-positive IgM or PCR or epi-link to confirmed case)

— Tightening definition ↑ specificity/PPV, ↓ sensitivity

— Loosening definition ↑ sensitivity, ↓ PPV

"Exam" of a surveillance system = the attributes CDC uses to evaluate it (MMWR 2001 framework). Tested frequently.

Simplicity — structure and ease of operation

Flexibility — adapts to new diseases or changing case definitions (e.g., COVID-19 added rapidly)

Data quality — completeness and validity of recorded data

Acceptability — willingness of persons/organizations to participate

Sensitivity

Positive predictive value (PPV)

Representativeness — accurately describes occurrence over time and distribution in the population

Timeliness — speed between steps (onset → report → action)

Stability — reliability and availability

Case definition is the diagnostic "physical exam" of surveillance

Key distinction: Sensitivity of a surveillance system ≠ sensitivity of a diagnostic test. Surveillance sensitivity captures system-level case capture (provider reporting + lab + investigation), not assay performance.

Board pearl: If an outbreak investigation expands the case definition mid-investigation, expect incidence to appear to rise artifactually — this is a classic distractor. Always ask whether the case definition changed before concluding a real epidemiologic trend.

Step 3 management: Evaluating a new electronic lab reporting feed → assess timeliness and completeness first; these are the attributes most often improved by automation.

Diagnostic Workup — Choosing the Right Data Source (Initial)

— Weekly MMWR tables; state-level granularity

— FoodNet (active surveillance, 10 sites, ~15% US population) — best for trends in lab-confirmed enteric infections

— PulseNet — molecular subtyping (WGS) to link cases across states → outbreak detection

— NORS — outbreak-level reporting

— SEER (incidence, survival; high-quality, ~48% population)

— NPCR (national coverage, all states) — combined as US Cancer Statistics

Match the question to the system — the single most testable concept in this topic

Notifiable infectious disease incidence (US) → NNDSS (CDC)

Vaccine-preventable disease, real-time outbreak → NNDSS + state health dept active case finding

Influenza activity → FluView = ILINet (outpatient ILI), NREVSS (lab), pediatric mortality, hospitalization (FluSurv-NET)

COVID-era respiratory → RESP-NET (hospitalization), wastewater surveillance (NWSS)

Foodborne illness

HIV → eHARS (enhanced HIV/AIDS reporting); CD4/viral load reporting in most states

STIs → NNDSS + STD Surveillance Network

Healthcare-associated infections → NHSN (mandatory for CMS reimbursement — CLABSI, CAUTI, SSI, C. difficile, MRSA bacteremia)

Antimicrobial resistance → NHSN AR Option, NARMS (enteric)

Cancer

Birth defects → state birth defects registries; MACDP (metropolitan Atlanta)

Board pearl: SEER vs. NPCR — SEER for survival/research depth, NPCR for national geographic completeness. USCS combines both.

CCS pearl: When the stem describes a multi-state cluster of Salmonella, the orders that move the case are call public health and send isolate for WGS (PulseNet), not repeat stool cultures.

Diagnostic Workup — Advanced/Specialized Data Sources

— VAERS (Vaccine Adverse Event Reporting System) — co-run by CDC/FDA; passive, anyone can report, used for signal detection only; cannot establish causality

— VSD (Vaccine Safety Datalink) — large integrated health system database (~12M lives); active, used to test hypotheses generated by VAERS

— CISA — clinical consultation for complex vaccine adverse events

— FAERS — FDA Adverse Event Reporting System for drugs/biologics (non-vaccine); MedWatch is the reporting portal

— Sentinel Initiative — FDA's active drug surveillance using claims/EHR (>100M lives); equivalent of VSD for drugs

— PRAMS — postpartum survey on behaviors/experiences

— Pregnancy Mortality Surveillance System (PMSS) — maternal deaths

— MMRCs — state Maternal Mortality Review Committees (case review, preventability)

— NVDRS — National Violent Death Reporting System (homicide, suicide details)

— WISQARS — query tool for injury data

— BRFSS — adults, state-level, telephone, self-report

— YRBSS — high school students

— NSDUH — substance use, mental health

— MTF (Monitoring the Future) — adolescent drug use trends

— NHANES — interview + physical + labs; only source for objectively measured HTN, lipids, HbA1c, lead, BMI

— NHIS — interview only, household-based

Adverse event and safety surveillance (high yield)

Pregnancy/maternal

Injury/violence

Behavioral/risk factor (population estimates)

Health/exam (gold standard for prevalence)

Key distinction: VAERS signal → VSD confirmation. A causal claim ("vaccine X causes Y") cannot rest on VAERS alone — classic distractor on Step 3 ethics/epi crossover.

Board pearl: Need measured (not self-reported) national prevalence of an obesity/diabetes/lead exposure metric → NHANES every time.

Risk Stratification — Outbreak Investigation Logic (CDC 10 Steps)

— 1. Prepare for fieldwork (supplies, team, contacts)

— 2. Establish existence of outbreak — compare observed to expected (endemic baseline)

— 3. Verify the diagnosis — review labs, exam findings

— 4. Construct a working case definition (person, place, time, clinical/lab criteria)

— 5. Find cases systematically and record information (line list)

— 6. Perform descriptive epidemiology — epidemic curve, person/place/time

— 7. Develop hypotheses

— 8. Evaluate hypotheses — typically retrospective cohort (defined population, e.g., wedding) or case-control (no defined population)

— 9. Refine hypotheses and conduct additional studies (environmental, lab)

— 10. Implement control and prevention measures — can be initiated at any step once sufficient evidence exists

— 11. Communicate findings

— Point source — sharp peak, all cases within one incubation period (potluck salmonella)

— Continuous common source — plateau (contaminated water supply)

— Propagated — successive peaks ~1 incubation apart (measles, person-to-person)

— Defined cohort (wedding guests, cruise ship) → retrospective cohort, calculate attack rates and RR

— Undefined source population (community-wide) → case-control, calculate OR

When surveillance signals a potential outbreak, follow the CDC outbreak investigation steps (frequently tested in order):

Epidemic curve shapes

Choosing the analytic study

Board pearl: Attack rate = ill / (ill + well) exposed; compare exposed vs. unexposed to find the vehicle (highest RR with biologic plausibility).

Step 3 management: Do not wait for hypothesis confirmation before implementing control measures — close the implicated facility, remove the product, or vaccinate contacts as soon as evidence is suggestive. Stepwise: report → line list → epi curve → cohort/case-control → control.

Pharmacotherapy — Mandatory Reporting "Drug Regimen"

— Infectious: measles, mumps, rubella, pertussis, diphtheria, polio, smallpox, anthrax, plague, tularemia, botulism, rabies (animal bites), viral hemorrhagic fevers, novel influenza, SARS-CoV-2, TB, N. meningitidis, H. influenzae invasive, hepatitis A/B/C, HIV, syphilis, gonorrhea, chlamydia, Lyme, RMSF, Zika, dengue, malaria, typhoid, cholera, Salmonella, Shigella, STEC, Listeria, C. botulinum

— Non-infectious: elevated blood lead, certain cancers (to state cancer registry), birth defects (state-dependent), occupational lung disease (some states)

— Child abuse/neglect → CPS (mandated, all states, no privilege exceptions)

— Elder abuse / vulnerable adult abuse → APS (most states)

— Intimate partner violence → varies; injury from weapon/burns generally reportable

— Gunshot/stab wounds → law enforcement (most states)

— Suspected impaired driver / specific conditions (e.g., seizures) → state DMV in some states

— Vaccine adverse events meeting VAERS table → VAERS (mandated by NCVIA 1986)

— Serious drug adverse events → FDA MedWatch (mandatory for manufacturers, voluntary but encouraged for providers)

— Class A (urgent: measles, meningococcal, foodborne with public risk) → immediate phone report, often within 24 h

— Most others → within 1–7 days, electronic acceptable

Treat reporting as a prescribed regimen with indications, timing, and route

Always reportable to public health (representative US list — varies slightly by state)

Reportable outside infectious disease channels (clinician duty)

Timing

Board pearl: Confidentiality is overridden by mandatory reporting statutes — you may report without patient consent, and HIPAA explicitly permits disclosures to public health authorities.

CCS pearl: A suspected case of meningococcemia → dexamethasone + ceftriaxone now, droplet isolation, and notify public health so household/close contacts get rifampin, cipro, or ceftriaxone chemoprophylaxis.

Procedures — Linking Surveillance to Public Health Action

— Single case of high-consequence pathogen (measles, meningococcal, novel influenza, anthrax, VHF, polio) → immediate isolation + health department notification + contact tracing and post-exposure prophylaxis

— Foodborne cluster (PulseNet match) → traceback to common food/facility, recall via FDA/USDA, public advisory

— Healthcare-associated infection signal (NHSN) → infection control bundle review, root cause analysis, public reporting via CMS Hospital Compare

— VAERS signal → VSD or Sentinel hypothesis testing → ACIP review → potential label change/pause (historical: rotavirus tetravalent intussusception 1999; J&J COVID vaccine TTS 2021)

— Cancer incidence rise (SEER) → etiologic investigation, screening guideline review (USPSTF)

— Antibiotic resistance trend (NARMS, NHSN) → antibiotic stewardship, formulary changes

— TB → identify close contacts, IGRA or TST + symptom screen + CXR; treat LTBI

— Pertussis → azithromycin prophylaxis for household and high-risk contacts regardless of vaccination

— Meningococcal → rifampin, ciprofloxacin, or ceftriaxone within 24 h ideally

— HIV → partner notification services; PEP within 72 h

— STIs → expedited partner therapy (EPT) legal in most states for chlamydia/gonorrhea

Surveillance is only valuable when it triggers action. Step 3 stems often pair a data source with the correct next intervention.

Action by signal type

Contact tracing principles

Board pearl: EPT allows the index patient's clinician to provide medication/prescription for partners without examining them — high-yield ethical/legal crossover.

CCS pearl: TB case identified → place in negative-pressure airborne isolation, start RIPE, send sputum AFB ×3, notify health department for directly observed therapy (DOT) and contact investigation.

Special Populations — Renal/Hepatic… and Data Quality Limitations

— Worst for diseases with mild/asymptomatic presentations (chlamydia, hepatitis C)

— True incidence may be 5–50× reported; CDC publishes adjusted estimates (e.g., STIs)

— BRFSS: landline/cell telephone — misses unhoused, institutionalized, non-English speakers (partially addressed by weighting)

— NHANES: oversamples Black, Hispanic, older adults, low-income to improve subgroup estimates — must use sampling weights in analysis

— Increased screening → apparent ↑ incidence without true change (e.g., thyroid cancer with neck ultrasound)

— Improved test sensitivity (PCR vs. culture) → ↑ apparent incidence

— Numerator (cases) ≠ population (denominator) source → biased rates

— Use census or ACS for denominators

— Cancer registry data typically reported with 2–4-year lag

— NVSS mortality has provisional and final files; cause-of-death miscoding common

Surveillance "comorbidities" = the biases and limitations that distort interpretation

Underreporting (low system sensitivity)

Selection bias in surveys

Recall and social desirability bias — self-reported behaviors (alcohol, sexual activity, exercise) systematically biased

Differential ascertainment

Denominator problems

Lag and provisional data

Privacy thresholds — small cell suppression (rates based on <20 events suppressed or flagged unstable) to protect identity and statistical reliability

Key distinction: Crude rate uses total population; age-adjusted rate uses a standard population (US 2000) — required when comparing populations with different age structures (e.g., Florida vs. Utah cancer rates).

Board pearl: A rising age-adjusted incidence with stable mortality often signals overdiagnosis from screening, not a true epidemic — classic for thyroid cancer and prostate cancer (pre-2012 USPSTF change).

Special Populations — Vulnerable Groups and Specialized Registries

— PRAMS — postpartum behaviors, contraception, breastfeeding, depression (states + NYC)

— MMRCs — review every pregnancy-associated death for preventability; data → CDC ERASE MM

— NVSS Linked Birth/Infant Death — neonatal/infant mortality with maternal risk factors

— National Immunization Survey (NIS) — vaccination coverage in children 19–35 mo and adolescents

— YRBSS — high schoolers; risk behaviors, mental health, sexual activity, substance use

— Newborn screening programs — state-mandated; data feed back to public health for case management of PKU, CAH, SCD, CF, etc.

— Indian Health Service surveillance; CDC's Tribal Epidemiology Centers

— Historical undercount in NNDSS for AI/AN — now improved race/ethnicity completeness

— NIOSH systems — fatal injuries (CFOI), pneumoconiosis (NIOSH Coal Workers' Health Surveillance), needlestick (EPINet)

— SENSOR — state-based occupational disease (asthma, pesticide poisoning)

Specific systems target groups underrepresented in general surveillance

Maternal/perinatal

Pediatric

Indigenous populations

Occupational

Refugee/immigrant — CDC Electronic Disease Notification (EDN) for arrivals

LGBTQ+ — increasingly captured in NHIS, BRFSS optional modules, YRBSS; historic data gap

Incarcerated — generally excluded from household surveys (NHANES, NHIS, BRFSS); separate Bureau of Justice Statistics data

Board pearl: Several Step 3-relevant populations (homeless, incarcerated, institutionalized, military active duty) are systematically excluded from household surveys — answer choices that hinge on these groups should not cite BRFSS/NHANES estimates.

Step 3 management: Adolescent flagged on YRBSS-mirroring screen for suicidality → immediate safety assessment, lethal means counseling, mental health referral, document plan; YRBSS itself is anonymous and not actionable individually.

Complications — Misuse and Misinterpretation of Surveillance Data

— A new case definition, lab test, or active case-finding push will transiently spike incidence — not a true epidemic

— Example: PCR replacing culture for pertussis ↑ apparent incidence

— HIV name-based reporting initially controversial; balanced by improved care linkage

— Small-cell suppression protects identity in rare disease/small-area data

— VAERS reports used to claim vaccine-autism link — invalid because VAERS lacks denominator, controls, causality assessment

— Missing race/ethnicity in 20–40% of some COVID early reports → inability to detect disparities

Ecological fallacy — inferring individual-level causation from group-level data (e.g., state-level smoking rate correlates with state cancer rate ≠ smokers in that state have cancer)

Simpson's paradox — aggregated data reverse the direction of effect seen in subgroups; classic when comparing crude vs. stratified mortality

Berkson bias — hospital-based surveillance overrepresents sicker populations

Lead-time and length-time bias — screening-driven surveillance can falsely suggest improved survival without mortality benefit

Surveillance artifact

Reporting delays — recent weeks/months in NNDSS appear artificially low; always interpret with lag-adjusted curves

Numerator/denominator mismatch — rates calculated with non-matching populations (e.g., COVID deaths from NVSS over census denominator missing recent migration)

Confidentiality breaches

Inappropriate causal claims from passive systems

Inequitable data

Key distinction: Association ≠ causation even when surveillance trend lines move together (Bradford Hill criteria still required: strength, consistency, specificity, temporality, biologic gradient, plausibility, coherence, experiment, analogy).

Board pearl: A question showing rising case counts after a new screening program → answer is usually surveillance artifact / overdiagnosis, not a true incidence rise.

When to Escalate — Outbreak and Public Health Emergency Triggers

— Any case of a Category A bioterrorism agent: anthrax, botulism, plague, smallpox, tularemia, viral hemorrhagic fevers

— Single case of measles, polio, diphtheria, rabies (human), novel influenza A, SARS, MERS, Ebola

— Cluster of unexplained severe illness or death

— Multistate enteric cluster matched by PulseNet WGS

— Healthcare-associated outbreak (NICU, hemodialysis, surgical)

— Clinician → local health department (first call) → state health department → CDC → WHO under International Health Regulations (IHR 2005) for events of international concern

— Health Alert (highest), Health Advisory, Health Update, Info Service

— Clinicians should subscribe; HAN messages frequently appear as question stem framing

— Suspected high-consequence pathogen → single-patient airborne isolation room, PPE per pathogen, notify hospital epidemiology and infection prevention immediately, do not transport through general areas

— Specimen handling per BSL level; alert lab before sending

— Governors and HHS Secretary can declare emergencies enabling EUAs, quarantine orders, crisis standards of care

— Quarantine (asymptomatic exposed) vs. isolation (symptomatic/infectious) — terminology tested

Triggers for state/CDC escalation

Escalation pathway

CDC Emergency Operations Center (EOC) activates for major events (H1N1 2009, Ebola 2014, Zika 2016, COVID-19 2020, mpox 2022)

Health Alert Network (HAN) — CDC tiered messaging

Hospital-level escalation

Legal authorities during emergencies

CCS pearl: Returning traveler from Uganda with fever and bleeding → immediate single-patient negative-pressure room, full PPE including PAPR, notify hospital infection control and state health department, do not draw routine labs in open lab — call CDC EOC for Ebola/Marburg guidance and specimen routing.

Key Differentials — Distinguishing Among Surveillance System Types

— NNDSS: passive, national, all notifiable conditions, less complete

— FoodNet: active, 10 sites, enteric pathogens only, gold-standard trends

— VAERS: passive, hypothesis-generating, no denominator

— VSD: active cohort, hypothesis-testing, has denominator

— CISA: clinical consultation for individual complex cases

— FAERS: passive drug AE reports (MedWatch portal)

— Sentinel: active distributed database, drug safety hypothesis testing

— SEER: depth (survival, treatment), ~48% population

— NPCR: breadth (all 50 states), incidence only historically

— USCS: combined SEER+NPCR for national incidence/mortality

— BRFSS: state-level, telephone, adult risk behaviors

— NHIS: national, in-person interview, health conditions/access

— NHANES: national, interview + physical exam + labs, only source for measured biomarkers

— NVSS: aggregate vital statistics

— NDI (National Death Index): record-level matching for research follow-up

— NHSN is the HAI system (CLABSI, CAUTI, SSI, C. diff, MRSA bacteremia) tied to CMS payment

— NNDSS captures community-acquired notifiable pathogens

Common Step 3 confusions, side by side:

NNDSS vs. FoodNet

VAERS vs. VSD vs. CISA

FAERS vs. Sentinel Initiative

SEER vs. NPCR vs. USCS

BRFSS vs. NHIS vs. NHANES

NVSS vs. NDI

NHSN vs. NNDSS for HAIs

Key distinction: When stem asks "best source to estimate national prevalence of uncontrolled hypertension" → NHANES (measured BP). When stem asks "state-level adult smoking prevalence" → BRFSS. When stem asks "trends in lab-confirmed Salmonella" → FoodNet.

Board pearl: Self-reported vs. measured is the single most common discriminator between BRFSS/NHIS and NHANES on the exam.

Key Differentials — Surveillance vs. Other Epidemiologic Tools

— Surveillance: population-level ongoing data collection

— Screening: individual-level testing of asymptomatic persons (USPSTF domain)

— Registry: organized collection of data on persons with a specific condition (often clinical-quality focus, e.g., STS cardiac surgery, NCDR cardiology)

— Some registries (SEER) function as surveillance systems

— Surveillance is exempt from IRB review when used for routine public health practice

— When surveillance data are analyzed for generalizable knowledge beyond program evaluation, it becomes research requiring IRB

— Active case finding: seek undetected cases in a population (TB in shelters)

— Contact tracing: identify exposed individuals from a known case

— Cluster: aggregation of cases that may or may not exceed expected — investigate to determine

— Outbreak: occurrence above expected baseline

— Epidemic = outbreak, often larger geographic scope

— Pandemic = epidemic across multiple countries/continents

— Endemic: usual baseline level

— Hyperendemic: persistently high

— Holoendemic: nearly universal in early life (some malaria settings)

Distinguish surveillance from related but distinct concepts

Surveillance vs. screening

Surveillance vs. registry

Surveillance vs. research study

Active case finding vs. contact tracing

Cluster investigation vs. outbreak investigation

Endemic vs. hyperendemic vs. holoendemic

Sporadic — irregular, infrequent cases without epidemiologic linkage

Key distinction: Quality improvement (QI) data feed clinical operations and are typically not human-subjects research; surveillance data inform public health practice — both are distinct from research requiring IRB and informed consent.

Board pearl: "Investigators want to publish national surveillance data analysis comparing two states' opioid trends" — if descriptive for public health practice, exempt; if inferential for generalizable knowledge, IRB required.

Secondary Prevention — Using Surveillance to Set Policy

— VAERS/VSD signals on rotavirus (RotaShield, 1999) → withdrawal

— VSD/CISA on RZV, mpox, COVID vaccines → schedule changes

— NIS coverage data → identify undervaccinated areas, drive outreach

— BRFSS/NHIS smoking trends → MPOWER, taxation, cessation coverage mandates

— Youth e-cigarette surge (NYTS) → flavor restrictions, Tobacco 21

— SEER incidence/mortality trends → USPSTF updates (e.g., 2021 lung cancer screening expansion to 50–80 y, 20 pack-year; 2021 colon cancer to start at 45 y)

— NHANES BP/lipid trends → JNC/ACC-AHA targets; statin eligibility

— NHANES HbA1c data, USRDS ESRD trends → diabetes prevention program coverage, SGLT2 use in CKD

— NVSS overdose mortality, prescription drug monitoring programs (PDMPs), DAWN ED data → CDC 2016/2022 prescribing guidelines, naloxone distribution

— MMRCs → AIM bundles (hemorrhage, hypertension, sepsis); Medicaid extension to 12 mo postpartum

— NHSN/NARMS data → Joint Commission stewardship requirement (2017)

Surveillance closes the loop by driving guideline and policy change — high-yield linkages:

Vaccine policy (ACIP)

Tobacco control

Cancer screening

Cardiovascular

Diabetes

Opioid epidemic

Maternal mortality

Antimicrobial stewardship

Step 3 management: For each patient discharge, link individual secondary prevention (e.g., post-MI: statin, beta-blocker, ACEi/ARB, aspirin, P2Y12, cardiac rehab, smoking cessation, BP/lipid follow-up) to the population data that justified the guideline — exam stems reward this systems framing.

Board pearl: USPSTF grades A/B = cover without cost-sharing under ACA; surveillance-driven evidence directly translates to insurance benefit design.

Follow-Up, Monitoring Parameters, and Counseling

— Annual evaluation against CDC's 9 attributes

— Completeness audits (compare to lab data, hospital discharge data)

— Capture-recapture methods estimate true case counts from overlapping sources

— Subscribe to CDC HAN, state health department alerts, MMWR

— Use EHR-integrated electronic case reporting (eCR) — increasingly mandatory; reduces clinician burden, improves timeliness

— Explain that diagnosis triggers public health notification and possible contact outreach

— Reassure about confidentiality protections — only public health personnel receive identifiers

— STIs: discuss partner notification options (provider-referral, patient-referral, EPT)

— TB: explain DOT and household contact testing

— Date of report, person notified, control measures initiated

— In CCS, write "report to public health department" as an explicit order

— Health department typically contacts clinician within 24–72 h for class A diseases

— Provide additional clinical data (exposures, travel, vaccination history) as requested

— Lead poisoning (BLL ≥3.5 µg/dL CDC reference value, 2021): recheck per CDC schedule, environmental investigation if ≥5

— Newborn screening positive: confirmatory testing within days; subspecialty referral

— Active TB: monthly clinical visits during therapy, sputum monthly until conversion, DOT

Continuous quality monitoring of surveillance

Clinician participation

Counseling patients about reportable conditions

Documentation

Follow-up cadence after reporting

Patient follow-up specific to condition

CCS pearl: After ordering "notify state health department" for syphilis, also order HIV test, hepatitis B/C serology, treat with benzathine penicillin G 2.4 million units IM, and counsel on partner notification / EPT not applicable for syphilis — partners need clinical evaluation and serology, not EPT.

Board pearl: Lead reference value changed from 5 to 3.5 µg/dL in 2021 — newer questions reflect this lower threshold.

Ethical, Legal, and Patient Safety Considerations

— HIPAA Privacy Rule §164.512 permits disclosure to public health authorities without patient authorization

— Patients cannot opt out of communicable disease reporting

— Clinicians have a duty to report; failure can incur fines and licensure action

— Provider-referral, patient-referral, or contract referral

— EPT legal in most states for chlamydia and gonorrhea; never for syphilis or HIV

— HIV partner notification is voluntary in most states; partners contacted without revealing index identity

— Newborn screening: typically opt-out (varies by state); ethical tension between parental autonomy and state interest in child welfare

— Public health emergency: EUAs require disclosure of investigational status, alternatives, right to refuse

— Rare disease in small geography → cell suppression in published data

— Genomic data in PulseNet: case-level identifiers protected

— Patient with newly diagnosed TB discharged home: ensure DOT handoff to health department, written follow-up date, medication in hand, household contacts list given to public health

— Failure to complete public health handoff = high-risk transition error

— Surveillance can stigmatize (HIV reporting, immigration concerns) — counter with assurances and culturally appropriate outreach

— Missing demographic data (race, ethnicity, sex/gender) hides disparities; ASTHO and CDC push for complete data

Mandatory reporting overrides confidentiality

Tarasoff duty — separate from public health reporting; duty to warn/protect identifiable third parties from credible threats by a patient (varies by state but tested as a general principle)

Partner notification and STIs

Informed consent edge cases

Confidentiality in small populations

Transition-of-care safety

Equity considerations

Board pearl: A patient with syphilis refuses to allow partner notification → you still must report to public health, but partner notification is conducted by public health staff who do not disclose the index patient's identity. Patient confidentiality and mandatory reporting coexist.

High-Yield Associations and Rapid-Fire Clinical Facts

NNDSS → CDC, weekly MMWR, passive, all states

MMWR → CDC's flagship surveillance publication, published Fridays

NHANES → measured biomarkers, oversamples minorities, mobile exam center

BRFSS → telephone, state-level, adult risk behaviors

YRBSS → high school students, biennial, anonymous

NHIS → in-person interview, household, national health conditions

NSDUH → substance use, mental health, ages 12+

NIS → child/adolescent vaccination coverage

PRAMS → postpartum mothers, state-level

VAERS → passive, vaccine AE, signal generation only

VSD → active, vaccine safety, hypothesis testing

FAERS → drug AE, MedWatch portal

Sentinel Initiative → FDA active drug surveillance, >100M lives

SEER → cancer depth, ~48% population, survival data

NPCR → cancer breadth, all 50 states

USCS → SEER + NPCR combined

FoodNet → active, 10 sites, enteric infections

PulseNet → molecular subtyping (WGS), multistate cluster detection

NORS → outbreak-level enteric reporting

NHSN → HAIs, tied to CMS payment

NARMS → enteric antimicrobial resistance

WONDER → CDC online query tool (NVSS, cancer, environmental)

WISQARS → injury data query

NVDRS → violent deaths

eHARS → enhanced HIV/AIDS reporting

NWSS → wastewater (SARS-CoV-2, polio, mpox)

ILINet/FluView → influenza outpatient + lab + hospitalization

HAN → CDC tiered emergency clinician alerts

IHR (2005) → WHO international reporting framework; PHEIC declarations

Board pearl: Wastewater surveillance gained prominence with SARS-CoV-2 and detected the 2022 New York polio case — modern stems may reference NWSS for early outbreak signals.

Step 3 management: When uncertain which agency receives a report, default answer is the local (county/city) health department, which then routes upward — never the FBI, never the WHO directly, never the CDC as first contact.

Board Question Stem Patterns

— Stem: researcher wants measured prevalence of obesity, HbA1c, lead → NHANES

— State-level adult smoking → BRFSS

— National cancer survival → SEER

— Vaccine adverse event after EUA approval → VAERS → VSD

— Multistate Salmonella outbreak detection → PulseNet

— Suspected measles in clinic → call local health department immediately, isolate (airborne), order IgM and PCR

— Single case meningococcemia → droplet isolation, ceftriaxone, notify public health, contact chemoprophylaxis

— TB suspect on chest x-ray → airborne isolation, AFB ×3, notify public health, IGRA on contacts

— Gunshot wound presenting to ED → stabilize and notify law enforcement

— Suspected child abuse → CPS report, hospital admission for safety if needed

— Step ordering: establish outbreak → verify diagnosis → case definition → line list → epi curve → hypothesis → analytic study → control → communicate

— Defined cohort (wedding) → retrospective cohort with attack rates and RR

— Undefined population → case-control with OR

— Rising thyroid cancer incidence with stable mortality → overdiagnosis / surveillance artifact

— VAERS used to claim causation → no denominator, no causality — invalid

— State-level correlation extended to individuals → ecological fallacy

— Refusal of consent for STI reporting → report anyway; HIPAA permits

— Newborn screening parental refusal → varies by state but generally mandatory with limited exemptions

— EPT for chlamydia → legal in most states, not for syphilis or HIV

Pattern 1: "Best data source"

Pattern 2: "Most appropriate next step"

Pattern 3: "Outbreak investigation"

Pattern 4: "Bias identification"

Pattern 5: "Ethics/legal"

Board pearl: Whenever the stem mentions "first step" alongside a notifiable disease, the answer is almost always isolate (if needed) + report to public health, even before confirmatory testing returns.

One-Line Recap

Public health surveillance is the ongoing, systematic collection and analysis of population health data — chosen by question type, evaluated by CDC's 9 attributes, mandated by reporting laws that override HIPAA, and acted upon through outbreak investigation and policy — and the Step 3 task is to match the right source to the right question and trigger the right next step.

— Measured biomarker prevalence → NHANES; state behavioral → BRFSS; notifiable disease incidence → NNDSS; vaccine AE → VAERS → VSD; drug AE → FAERS/Sentinel; cancer → SEER/NPCR/USCS; foodborne → FoodNet/PulseNet; HAIs → NHSN

— Passive = cheap, undersensitive; active = complete, costly; sentinel = focused providers; syndromic = real-time symptoms

— HIPAA explicitly permits public health disclosures; clinicians must report Class A diseases immediately, others within days; failure carries legal consequences

— Establish → verify → define → line list → epi curve → hypothesize → test (cohort if defined population, case-control if not) → control → communicate

— Case definition changes, new tests, screening programs all inflate apparent incidence — always rule out before declaring a true trend

— For every reportable diagnosis, write three orders: treat the patient, isolate appropriately, notify public health — and document the handoff so transition-of-care is safe

Match the source to the question

Passive vs. active vs. sentinel vs. syndromic

Mandatory reporting trumps confidentiality

Outbreak investigation is a stepwise CCS algorithm

Surveillance artifacts mimic real epidemics

Step 3 voice

Board pearl: When in doubt on a surveillance stem, the safest answer pairs local health department notification with the least invasive but most timely clinical action — that combination is rarely wrong on Step 3.