Quality assurance
Quality assurance items test whether you can measure care, interpret the data, and change the system - and whether you can tell QA from QI. QA asks whether standards were met; QI asks how to make them met more reliably.
The machinery
- Define the indicator and the denominator. "Medication errors" is not an indicator; "wrong-time errors per 100 medication passes audited monthly" is.
- Set thresholds and trend them. A single month's rate is noise; three months of upward movement is a signal.
- Common data sources in disability services: medication administration record audits, incident reports, falls, weight and nutrition trends, pressure injury incidence, restraint and seclusion use, emergency department visits, hospitalization, infection rates, training compliance, ISPs reviewed on time, and person/family satisfaction.
- Analysis, not just counting. Use root cause analysis or a similar structured method for sentinel events; use plan-do-study-act cycles for smaller process problems.
- Plan of correction. Every identified deficiency needs an action, an owner, a due date, and a re-measurement.
- External oversight includes state licensing surveys, ICF/IID certification, HCBS quality requirements, and quality-of-life measurement systems such as National Core Indicators.
Worked example
A residential program reports six medication errors in one quarter. Counting is not enough. The review shows five of six were wrong-time errors occurring on the 7 a.m. pass when one nurse covers two houses. The process change: stagger the pass, add a second trained staff member at 7 a.m., and simplify the MAR layout. The team sets a threshold of zero wrong-time errors per quarter, audits 30 passes per month, and re-measures in 90 days. That is QA identifying the gap and QI closing it.
Traps
- Reporting rates with no denominator.
- Blaming an individual when the data show a system problem.
- Writing a corrective action with no owner or re-measurement.
- Auditing only what is easy to count rather than what matters to the person.
- Treating a good survey result as proof that care is good.
- Excluding people receiving services and their families from quality measurement.