Surveillance Systems, Registries, Data Synthesis, Visualization, and Informatics

Quick Answer: Surveillance systems differ in completeness: active surveillance seeks cases proactively, passive surveillance depends on voluntary reporting, and sentinel surveillance monitors selected sites. Registries provide longitudinal case-level data but face coverage gaps and reporting delays. Synthesis methods (systematic review, meta-analysis) combine evidence across studies, while informatics provides the standards, interoperability, and infrastructure that make timely public health action possible.

Surveillance System Types and Trade-offs

Active surveillance involves regular, planned outreach to reporting sites — clinics, labs, hospitals — to identify cases. The CDC's FoodNet and Emerging Infections Program use active surveillance for foodborne and emerging pathogens. Active surveillance yields the most complete case ascertainment but is expensive and labor-intensive. Passive surveillance relies on providers and labs to voluntarily submit reports (e.g., notifiable disease reporting to state health departments). It is inexpensive and covers broad geographies but under-ascertains cases because reporting depends on clinician awareness, testing practices, and compliance. Sentinel surveillance uses selected reporting sites — such as CDC influenza sentinel providers — to monitor trends, detect outbreaks early, and characterize strains; it cannot measure total case counts but is efficient for trend monitoring.

Syndromic surveillance monitors pre-diagnostic symptoms (e.g., emergency department chief complaints, over-the-counter drug sales) to detect outbreaks earlier than lab-confirmed reporting, but signals are non-specific and require validation. Each system has a sensitivity (probability it detects true cases) and timeliness trade-off; increasing sensitivity often increases false positives and follow-up burden.

Registries and Their Limitations

A registry is a systematic collection of case-level data about a defined condition. Population-based cancer registries (SEER, NPCR) collect incidence, stage, treatment, and survival data. Immunization registries (IIS) track vaccination histories. Birth defect registries monitor congenital anomalies. Disease-specific registries (e.g., cystic fibrosis) follow patients longitudinally.

Registries offer longitudinal, case-level detail but face important limitations. Selection bias arises when only diagnosed or treated cases enter the registry, missing undiagnosed or untreated cases. Reporting completeness varies by facility, region, and demographic group. Lag times between diagnosis and registry submission delay trend detection. Data quality depends on abstraction accuracy and variable definitions. A registry may also lack denominator data for rates unless linked to a population file. When interpreting registry-based rates, always ask whether the catchment area covers the target population and whether under-reporting is likely.

Synthesizing Information From Multiple Sources

Public health decisions often rest on multiple studies or data streams. Systematic reviews use pre-specified methods to identify, appraise, and summarize all relevant studies on a question. Meta-analysis statistically combines effect estimates across studies, weighting each by inverse variance; results are displayed in a forest plot with each study as a square sized by weight and a diamond showing the pooled estimate. Heterogeneity (I² statistic, Q test) must be assessed before pooling — substantial heterogeneity suggests the studies measure different effects and a single pooled estimate is misleading. Pooled analysis combines raw individual-level data from multiple studies, a stronger design than meta-analysis of published summaries.

When synthesizing surveillance data with research studies, weigh each source by its timeliness, completeness, and risk of bias. A real-time syndromic signal may justify immediate action even while peer-reviewed studies lag; conversely, a single observational study with residual confounding should not override consistent surveillance trends. Triangulation across data sources — surveillance, registry, survey, and EHR — strengthens causal inference when each source has different biases.

Data Visualization Best Practices

Choosing the correct chart type is the foundation of honest visualization. Bar charts compare categorical counts; line charts show trends over time; histograms display distributions of a continuous variable; scatter plots show relationships between two continuous variables; maps display geographic variation. Misleading practices to avoid include truncated y-axes that exaggerate differences, dual y-axes that imply false correlations, 3D pie charts that distort proportions, and inconsistent color scales. Accessibility matters: use colorblind-friendly palettes (e.g., viridis, ColorBrewer), avoid red-green pairs as the sole encoding, and label axes and legends clearly. A good visualization lets the reader extract the message in under five seconds and supports public health decisions with accurate magnitude perception.

Public Health Informatics

Public health informatics is the systematic application of information and computer science to public health practice, research, and learning. It integrates surveillance, EHRs, immunization systems, lab reporting, and decision support. Core components include interoperability standards (HL7 v2 and FHIR for health data exchange; LOINC and SNOMED CT for coded clinical concepts; ICD-10-CM for diagnoses), data governance (who can access which data elements and under what authority), security and privacy (HIPAA, FERPA, 42 CFR Part 2), and analytics infrastructure (cloud platforms, disease detection algorithms, NLP for free-text reports). Informatics matters because modern outbreaks move faster than manual reporting: electronic lab reporting, syndromic surveillance feeds, and interoperable EHR connections shorten the gap between event and action. The CPH exam emphasizes that informatics is not just IT support — it is a discipline that shapes data quality, timeliness, decision support, and ultimately the effectiveness of public health response.

An emerging consideration is data equity: informatics systems can reinforce disparities when communities lacking EHR access, broadband, or digital health tools are invisible to surveillance. Designing inclusive data systems — capturing underserved populations, validating algorithms across demographic groups, and partnering with community data sources — is part of the informatics responsibility, not a side concern.