BCHHC Intermediate — PDGM Deep Dive

PDGM — Patient-Driven Groupings Model

Effective January 2020, replaced the therapy-based PPS. Payment is per 30-day period (not 60-day episode). Each period classified by five factors:

Factor 1: Timing

Factor 2: Admission Source

Factor 3: Clinical Grouping (12 Groups)

|-------|--------------------------|

Factor 4: Functional Level

From OASIS items: Low / Medium / High

Based on: ambulation, transferring, bathing, dressing, eating, toilet hygiene

Factor 5: Comorbidity Adjustment

Based on secondary diagnoses: None / Low / High

Higher comorbidity complexity = higher payment

LUPA — Low Utilization Payment Adjustment

Form 485 — Plan of Care

Intermediate Topics

The Gap Between Learning and Production

Most tutorials show you the happy path. Intermediate knowledge is about the unhappy paths: what breaks, how to debug it, how to prevent it, and how to recover when it does break anyway.

Real-World Patterns

Configuration Management

Use environment variables for values that differ per environment. Use a secrets manager for sensitive values. Validate all configuration at startup — a service that starts with invalid config and then crashes an hour later is harder to debug than one that refuses to start. Version your configuration alongside your code.

Connection Management

Connections are expensive to create. Use connection pools. Set appropriate pool sizes — too small creates a bottleneck, too large exhausts server resources. Set timeouts: connect timeout (how long to wait for a connection to be established), read/write timeout (how long to wait for a response).

Retry and Backoff

Transient failures are normal in distributed systems. Retry with exponential backoff: wait 1s, then 2s, then 4s, then 8s. Add jitter (random variation) to prevent thundering herd when many services retry simultaneously. Set a maximum number of retries. Use a circuit breaker to stop retrying a service that is clearly down.

Structured Logging

JSON logs are machine-readable. Include: timestamp (ISO 8601), log level, service name, trace ID (for correlation across services), user ID or request ID (for debugging specific requests), and the actual message. Never log sensitive data (passwords, tokens, PII).

Performance Fundamentals

Measure first, optimise second. Use profiling tools — optimise what the profiler shows is slow, not what you think is slow. Common wins: add the right indexes, reduce N+1 queries, add a cache layer, reduce serialisation overhead, batch small operations.

Intermediate Debugging

When basics fail: 1. Enable debug logging temporarily. 2. Add timing instrumentation around suspected slow operations. 3. Check network traffic (tcpdump, Wireshark, or a proxy). 4. Check system resources (top, iostat, netstat). 5. Reproduce with minimal config to eliminate variables.

Testing Strategy

|-----------|--------------|-------------|