8.2 Step 6: Planning & Collecting Baseline Performance Measures

Step 6: Planning & Collecting Baseline Performance Measures

Core Principle: In Step 6 of the Center for Health Design's 8-Step EBD Process, the project team executes the empirical measurement plan established by the hypotheses in Step 5. Baseline performance measures represent the quantitative and qualitative reference data collected in the existing physical environment prior to construction, renovation, or occupancy. Without a standardized, pre-intervention baseline, it is very difficult to judge whether post-occupancy outcomes reflect the design interventions or external confounding factors.

A frequent and costly error in healthcare capital projects is the failure to plan for baseline data collection before vacating an old facility. Once a legacy unit is demolished, renovated, or abandoned, the pre-occupancy operating conditions are permanently erased. Retrospective attempts to reconstruct baseline data from incomplete archives or staff recollections introduce massive recall bias and data fragmentation. Step 6 demands proactive, systematic data governance embedded early in project scheduling.

---

The Strategic Imperative of Step 6 in the EBD Process

Within the 8-Step EBD sequence, Step 6 serves as the benchmark against which future design success is evaluated. Its roots are in predesign: EDAC study materials describe the strategic facilities plan as evaluating current practices and establishing an operational baseline, including assessing current processes at a macro level and collecting data on the current level of performance.

┌────────────────────────────────────────────────────────────────────────┐
│         THE DATA COLLECTION TIMELINE IN HEALTHCARE CAPITAL PROJECTS    │
├────────────────────────────────────────────────────────────────────────┤
│  Phase 1: Predesign & Design  ──> STEP 5: Formulate Hypotheses        │
│  Phase 2: Pre-Construction    ──> STEP 6: COLLECT BASELINE MEASURES   │
│  Phase 3: Construction        ──> STEP 7: Monitor Implementation       │
│  Phase 4: Post-Occupancy (POE)──> STEP 8: Measure Performance Results  │
└────────────────────────────────────────────────────────────────────────┘

CHD's current exam content outline places this work in the Design domain: the team should establish a set of baseline metrics and data to be gathered before occupancy for evaluating hypotheses once the building is occupied, including measurable outcomes, system-based criteria (such as a Lean project approach that concurrently evaluates safety, efficiency, satisfaction, and cost), and the estimated budget for measurement.

Collecting baseline measures serves three distinct institutional functions:

1. Establishes Internal Validity: In quasi-experimental field research (the dominant research design in healthcare architecture), researchers cannot randomly assign patients to different hospital buildings. Robust baseline data, paired with a non-equivalent control group or an interrupted time-series design, allows researchers to rule out confounding secular trends.
2. Calibrates Realistic Targets: A project team cannot project a 30% reduction in medication errors without knowing the current baseline error frequency, standard deviation, and underlying clinical distribution.
3. Identifies Latent Operational Dysfunctions: The process of capturing baseline workflow and clinical data often reveals existing operational bottlenecks—such as broken supply chains or charting workarounds—that must be solved through organizational policies in tandem with physical architecture.

---

Selecting Valid Metrics Across the Five Healthcare Domains

An effective baseline data collection plan does not attempt to measure every conceivable operational variable. Instead, it targets the specific dependent variables identified in the Step 5 hypotheses across five primary data domains:

┌────────────────────────────────────────────────────────────────────────┐
│                 FIVE CORE BASELINE MEASUREMENT DOMAINS                 │
├───────────────────────────────────┬────────────────────────────────────┤
│ 1. Clinical Quality & Safety (EHR)│ 2. National Quality Registries     │
│    • Medication error incident logs│    • NDNQI nurse-sensitive falls   │
│    • Hospital-Acquired Conditions  │    • NHSN device-associated HAIs   │
│    • Average Length of Stay (ALOS) │    • Pressure injury prevalence   │
├───────────────────────────────────┼────────────────────────────────────┤
│ 3. Patient & Family Experience     │ 4. Staff Occupational Well-Being   │
│    • HCAHPS domain percentiles     │    • OSHA Form 300 injury logs    │
│    • Press Ganey satisfaction logs │    • Annual voluntary RN turnover  │
│    • Post-discharge noise feedback │    • Overtime hours & absenteeism  │
├───────────────────────────────────┴────────────────────────────────────┤
│ 5. Operational Workflow & Spatial Telemetry                            │
│    • Real-Time Location Systems (RTLS) nurse travel distance (km/shift)│
│    • Call-light response latencies (seconds)                           │
│    • Medication dispensing and pharmacy turnaround intervals           │
└────────────────────────────────────────────────────────────────────────┘

1. Clinical Quality and Electronic Health Record (EHR) Data

Modern healthcare enterprises capture massive clinical data volumes within Electronic Health Record systems (e.g., Epic, Cerner). For EBD baseline tracking, data analysts extract longitudinal clinical metrics, including:

• Medication Administration Errors (MAEs): Categorized by timing errors, dosage errors, omission errors, and wrong-patient errors, standardized per 1,000 doses administered or per 100 patient days.
• Hospital-Acquired Delirium: Measured via standardized nursing assessments such as the Confusion Assessment Method (CAM) or CAM-ICU.
• Risk-Adjusted Length of Stay (ALOS vs. Geometric Mean LOS): Evaluated in conjunction with the hospital's Case Mix Index (CMI) to ensure that changes in stay duration reflect operational and clinical efficiency rather than shifts in patient clinical complexity.

2. National Benchmarking Registries

Standardized national registries provide risk-adjusted, peer-reviewed benchmarks that permit comparison between the study hospital and national percentiles:

• NDNQI (National Database of Nursing Quality Indicators): Managed by Press Ganey (originally developed by the American Nurses Association), NDNQI tracks nurse-sensitive quality indicators, notably:
  • Patient fall rates per 1,000 patient days.
  • Patient falls resulting in injury (minor, moderate, severe, death).
  • Hospital-acquired pressure injury (HAPI) stage 2 and above prevalence rates.
• NHSN (National Healthcare Safety Network): Managed by the Centers for Disease Control and Prevention (CDC), NHSN tracks standardized infection ratios (SIRs) for healthcare-associated infections (HAIs):
  • Central Line-Associated Bloodstream Infections (CLABSI).
  • Catheter-Associated Urinary Tract Infections (CAUTI).
  • Surgical Site Infections (SSI).
  • Methicillin-Resistant Staphylococcus aureus (MRSA) bacteremia.
  • Clostridioides difficile (C. diff) laboratory-identified events.

3. Patient and Family Experience (HCAHPS)

The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) is a standardized, publicly reported survey that CMS requires of hospitals paid under the inpatient prospective payment system to receive their full annual payment update. It measures patient perceptions across key domains directly influenced by the physical environment:

• Quietness of the Hospital Environment at Night: Frequently the lowest-scoring domain nationally; heavily influenced by acoustic materials, alarms, and corridor layout.
• Cleanliness of the Hospital Environment: Influenced by visual clutter, finish materials, lighting, and bathroom layout.
• Nurse Communication and Responsiveness: Influenced by decentralized vs. centralized station sightlines and call-light notification architecture.

4. Staff Occupational Health, Safety, and Retention

The human capital of healthcare delivery represents the largest operational expense of any health system. Physical work environments heavily impact clinical staff safety and retention:

• OSHA Form 300 Logs: Mandated by federal law, these logs document work-related recordable injuries, including acute musculoskeletal back and shoulder strains resulting from patient repositioning, slips, trips, and falls.
• Human Resources (HR) Metrics: Annual voluntary registered nurse turnover rates, first-year nurse turnover rates, average days to fill open clinical vacancies, and overtime expenditure hours.
• Validated Psychometric Surveys: Standardized instruments assessing staff burnout (e.g., Maslach Burnout Inventory [MBI]), physical exhaustion, and perceived environmental support.

5. Operational Workflow and Spatial Telemetry

Objective spatial behavior provides the missing link between architectural layouts and clinical outcomes:

• Real-Time Location Systems (RTLS) & RFID Telemetry: Infrared or ultra-wideband (UWB) tracking tags worn on nurse ID badges to log precise travel trajectories, steps per shift, total kilometers walked per 12-hour shift, and percentage of shift spent at the bedside versus at centralized stations or in supply closets.
• Call-Light Telemetry: Automated electronic call-bell servers that log the exact timestamp of patient call button activation, time elapsed until staff room entry (cancellation of call light), and call frequency by clinical shift.

---

Determining the Baseline Time Horizon: Neutralizing Seasonality

A critical methodological question in Step 6 is: How long must the baseline data collection period last?

[!IMPORTANT]

A Practical Baseline Standard

Where possible, collect about 12 continuous months of baseline data. A full annual cycle helps account for seasonality and cyclical changes in census, acuity, and staffing. This is research good practice rather than a specific CHD-mandated number.

┌────────────────────────────────────────────────────────────────────────┐
│               ANNUAL HEALTHCARE CYCLICALITY & SEASONALITY              │
├────────────────────────────────────────────────────────────────────────┤
│  Q1 (Jan - Mar): Winter respiratory surges (Flu, RSV, COVID-19)        │
│                  • Maximum census, high patient acuity, high nurse OT   │
│  Q2 (Apr - Jun): Elective surgeries ramp up; post-winter decompression │
│  Q3 (Jul - Sep): Summer trauma spikes, academic hospital July turnover, │
│                  elective surgery vacation slowdowns                   │
│  Q4 (Oct - Dec): Year-end insurance deductible rush; holiday staffing  │
│                  dips; early winter viral onset                        │
└────────────────────────────────────────────────────────────────────────┘

Confounding Risks of Abbreviated Baselines

If a project team collects baseline measures for only 30 or 60 days during the summer (e.g., July and August) and compares them to post-occupancy measures collected during the winter (e.g., January and February), the study is completely compromised:

• Census and Acuity Fluctuation: Inpatient census and acuity surge dramatically during winter viral illness seasons. A newly designed emergency department might appear to have worsened patient wait times in post-occupancy simply because it moved during an influenza epidemic.
• Staffing and Turnover Cyclicality: Teaching hospitals experience predictable spikes in clinical errors, extended procedure times, and onboarding delays in July when new resident physicians and newly graduated nurses enter service (the "July Effect").
• When a Full Year Isn't Possible: Use the longest feasible baseline window, compare with the same months from prior years where data exist, and use statistical adjustment (for example, seasonal and case-mix adjustment) to reduce bias.

---

The Pre-to-Post Measurement Invariance Rule

To support internal validity, the project team should keep measurement invariance between Step 6 (pre-occupancy baseline) and Step 8 (post-occupancy evaluation). "Pre-to-post measurement invariance rule" is this guide's teaching label for that principle, not an official CHD term.

┌────────────────────────────────────────────────────────────────────────┐
│                THE PRE-TO-POST INVARIANCE MANDATE                      │
├────────────────────────────────────────────────────────────────────────┤
│  PRE-OCCUPANCY BASELINE (Step 6)   POST-OCCUPANCY EVALUATION (Step 8) │
│  • Same Survey Instrument     ───► • Same Survey Instrument            │
│  • Same Exact Phrasing / Scale ───► • Same Exact Phrasing / Scale      │
│  • Same Sampling Protocol      ───► • Same Sampling Protocol           │
│  • Same EHR Data Extraction    ───► • Same EHR Data Extraction         │
│  • Same Environmental Sensors  ───► • Same Environmental Sensors       │
└────────────────────────────────────────────────────────────────────────┘

Common Violations of Measurement Invariance

1. Altering Survey Questions or Scales: If the pre-occupancy survey asks staff to rate acoustic noise on a 5-point Likert scale, and the post-occupancy survey switches to a 7-point scale or rewords the question from "How disruptive is corridor noise?" to "Does corridor noise interfere with sleep?", the datasets cannot be statistically compared.
2. Shifting Sensor Calibration or Placement: If baseline acoustic sound pressure levels (dBA Leq) were recorded with sound level meters positioned at the patient headwall 1.2 meters above the floor, post-occupancy measurements must be conducted using the exact same meter type, weighting scale (A-weighting, slow response), and physical height/room coordinates.
3. Modifying EHR Extraction Query Logic: If the baseline query extracted fall events occurring between 19:00 and 07:00, the post-occupancy query must not extract 24-hour falls without maintaining the discrete night-shift stratification.

---

Institutional Research Governance, IRB Oversight & HIPAA Compliance

Step 6 involves gathering sensitive human, clinical, and institutional data, requiring rigorous research governance.

1. Research vs. Quality Improvement (QI)

A primary institutional consideration is determining whether the EBD data collection initiative constitutes Human Subject Research or standard Quality Improvement (QI):

• Quality Improvement (QI): Systematic, data-guided activities designed to immediately improve healthcare delivery processes within a specific institution, with no external scientific intent. QI initiatives typically do not require formal Institutional Review Board (IRB) review.
• Human Subject Research (The Common Rule): Systematic investigations designed to develop or contribute to generalizable scientific knowledge (including publishing findings in peer-reviewed journals like HERD or presenting at national conferences). If a project is designed to contribute to generalizable knowledge (for example, findings intended for publication), it is generally treated as research, and the team should obtain an IRB determination before collecting baseline data. The IRB, not the project team, decides whether an activity is research or quality improvement.

2. Levels of IRB Review in Healthcare Design Studies

3. HIPAA Privacy Rule and Protected Health Information (PHI)

Under the Health Insurance Portability and Accountability Act (HIPAA), researchers must safeguard Protected Health Information (PHI). When extracting baseline clinical datasets:

• Safe Harbor Method: Requires removing 18 specified identifiers (including names, most date elements other than year, medical record numbers, contact details, and other unique identifying numbers or codes).
• Limited Data Sets & Data Use Agreements (DUAs): If date stamps or departmental location codes are essential to correlate clinical events with environmental conditions, the team must execute an institutional Data Use Agreement and maintain strict encryption standards.