2.1 Surveillance System Design, Objectives & Case Definitions

2.1 Surveillance System Design, Objectives & Case Definitions

Surveillance of healthcare-associated infections (HAIs) is the foundational pillar of an effective Infection Prevention and Control (IPC) program. The Centers for Disease Control and Prevention (CDC) defines public health surveillance as the ongoing, systematic collection, analysis, interpretation, and dissemination of health data essential to the planning, implementation, and evaluation of public health practice. In the healthcare setting, surveillance provides the empirical evidence required to establish baseline infection rates, detect outbreaks, evaluate prevention bundle efficacy, and satisfy regulatory reporting mandates.

---

1. Surveillance Methodologies & Structural Designs

An Infection Preventionist (IP) must select surveillance methodologies aligned with organizational goals, patient acuity, available technological infrastructure, and regulatory requirements. Surveillance systems are classified along three operational axes:

Active vs. Passive Surveillance

• Active Surveillance: The IP proactively searches for HAIs by systematically reviewing microbiology culture results, electronic health record (EHR) clinical notes, imaging reports, pharmacy antimicrobial orders, and patient census lists. Active surveillance yields high sensitivity and data accuracy but demands substantial personnel time.
• Passive Surveillance: Relies on primary care providers, bedside nurses, or external clinicians to voluntarily or mandatorily report suspected infections to the IPC department. Passive surveillance requires minimal IPC resources but suffers from severe underreporting, inconsistent case ascertainment, and low sensitivity.

Targeted (Focused) vs. Facility-Wide Surveillance

• Targeted (Focused) Surveillance: Concentrates surveillance activities on specific high-risk units (e.g., Neonatal Intensive Care Unit [NICU], Surgical Intensive Care Unit [SICU]), specific high-risk devices (e.g., central lines, indwelling urinary catheters, mechanical ventilators), or high-volume surgical procedures (e.g., total joint arthroplasty, coronary artery bypass grafting). Targeted surveillance optimizes resource allocation and delivers actionable data where infection risk and impact are highest.
• Facility-Wide Surveillance: Monitors all hospitalized patients across all inpatient units for all HAI types. While comprehensive, facility-wide manual surveillance is extremely resource-intensive and often dilutes the IP's ability to execute targeted clinical interventions. Electronic surveillance systems (ESS) leveraging automated algorithm-driven data mining are increasingly rendering facility-wide screening feasible.

Prospective vs. Retrospective Surveillance

• Prospective Surveillance: Monitors patients for infection development concurrently during their inpatient admission. Prospective surveillance allows the IP to observe clinical signs in real time, collaborate with bedside staff, implement immediate transmission-based precautions, and intervene during cluster emergence. It is the gold standard for IPC practice.
• Retrospective Surveillance: Identifies infections post-discharge or after clinical management has concluded, primarily through post-discharge chart audits, administrative billing code analysis (ICD-10-CM), or registry reviews. Retrospective data cannot support real-time clinical intervention and administrative coding frequently misclassifies secondary complications as pre-existing conditions.

---

2. Quantitative Evaluation of Surveillance Criteria

To ensure surveillance tools reliably distinguish true infections from non-infections, IPs evaluate surveillance criteria using performance metrics derived from a $2 \times 2$ epidemiological contingency table. Surveillance criteria act as diagnostic tests evaluated against a gold standard (e.g., rigorous expert clinical panel review).

The $2 \times 2$ Contingency Framework

Key Performance Formulas

1. Sensitivity ($Sens$): The probability that the surveillance criteria will correctly identify an actual HAI when present. $\text{Sensitivity} = \frac{TP}{TP + FN} \times 100\%$ High sensitivity minimizes missed cases (false negatives).
2. Specificity ($Spec$): The probability that the surveillance criteria will correctly exclude a patient who does not have an HAI. $\text{Specificity} = \frac{TN}{TN + FP} \times 100\%$ High specificity prevents false alarms and unnecessary resource expenditure.
3. Positive Predictive Value ($PPV$): The probability that a patient flagged by the surveillance criteria actually has a true HAI. $\text{PPV} = \frac{TP}{TP + FP} \times 100\%$
4. Negative Predictive Value ($NPV$): The probability that a patient not flagged by the criteria is genuinely free of an HAI. $\text{NPV} = \frac{TN}{TN + FN} \times 100\%$

Epidemiological Rule: Sensitivity and Specificity are intrinsic parameters of the surveillance definition. However, PPV and NPV depend directly on disease prevalence in the population. As HAI prevalence increases, PPV increases and NPV decreases; conversely, in low-prevalence settings, PPV drops significantly.

Step-by-Step Worked Mathematical Example

An IP validates an automated electronic surveillance algorithm for central line-associated bloodstream infections (CLABSI) against manual expert chart review in a cohort of $1,000$ ICU patients:

• $TP = 80$ patients correctly flagged by the algorithm as having CLABSI.
• $FP = 10$ patients flagged by the algorithm who did not meet clinical CLABSI criteria.
• $FN = 20$ patients with true CLABSI missed by the algorithm.
• $TN = 890$ patients correctly unflagged by the algorithm.

Calculations:

• Sensitivity: $\frac{80}{80 + 20} = \frac{80}{100} = 0.800 = 80.0%$
• Specificity: $\frac{890}{890 + 10} = \frac{890}{900} = 0.9889 = 98.9%$
• Positive Predictive Value (PPV): $\frac{80}{80 + 10} = \frac{80}{90} = 0.8889 = 88.9%$
• Negative Predictive Value (NPV): $\frac{890}{890 + 20} = \frac{890}{910} = 0.9780 = 97.8%$

---

3. Developing Standardized Case Definitions

A surveillance case definition is a set of standardized, objective criteria used to establish whether a patient has a specific HAI for epidemiologic tracking. Standardized case definitions differ fundamentally from clinical diagnostic definitions:

• Clinical Diagnostic Definitions: Designed to guide individual patient treatment; prioritize high sensitivity to ensure no infected patient goes untreated, often incorporating subjective physician impressions and empiric treatment decisions.
• Surveillance Case Definitions: Designed to monitor population disease trends and permit valid comparisons across institutions; prioritize objectivity, consistency, and specificity to eliminate observer bias.

Key Elements of a Robust Case Definition

1. Clinical Criteria: Objective clinical signs and symptoms (e.g., documented temperature $> 38.0^\circ\text{C}$, purulent sputum, dysuria, suprapubic tenderness).
2. Laboratory Criteria: Diagnostic microbiological, serological, or molecular evidence (e.g., $\ge 10^5 \text{ CFU/mL}$ of a single uropathogen, positive blood culture for a recognized pathogen).
3. Imaging / Diagnostic Criteria: Radiographic or ultrasonic evidence (e.g., new or progressive infiltrate on chest radiograph).
4. Device & Temporal Exposure Criteria: Specific duration thresholds for invasive device presence (e.g., central line in place for $> 2$ calendar days) and defined surveillance observation windows (e.g., 30-day post-operative window).