QUALITY IMPROVEMENT + SCHOLARSHIP CONTROL CENTER
ALL YEARS | PAGE 81D - AIM -> MEASURE -> CHANGE -> PDSA -> SUSTAIN -> REPORT -> 10 DECISIONS ->
G. QUALITY-IMPROVEMENT OPERATING SYSTEM
| Step | Execution | Proof / stop rule |
|---|---|---|
| 1. Aim | Define what will improve, for whom, by how much and by when; engage the people affected by the work. | Specific, measurable local aim. |
| 2. Measures | Operationalize outcome, process and balancing measures; decide data frequency and denominator. | Team can calculate each measure the same way. |
| 3. Understand current system | Map workflow/causes; establish baseline variation; identify constraints and context. | Change idea targets a plausible driver. |
| 4. Test change | Run small, safe PDSA tests; predict what will happen; document what actually happened. | Each test creates learning, not just activity. |
| 5. Study over time | Use run/control charts or other time-aware methods when appropriate; note context and unintended effects. | Signal persists beyond one favorable data point. |
| 6. Adapt | Adopt, adapt or abandon based on evidence; change one or more components deliberately. | Next cycle reflects prior learning. |
| 7. Implement + sustain | Standardize workflow, ownership, training, monitoring and escalation; test varied conditions. | Improvement survives ordinary staffing/time/context. |
| 8. Spread + report | Assess transferability before spread; report with SQUIRE when appropriate; preserve governance determination. | Another team can understand what changed, where and why. |
TOP 10 RESEARCH + QI CONCEPT - 9: QI NEEDS MEASUREMENT OVER TIME - NOT JUST BEFORE VS AFTER
| DEEP EXPLANATION | WHY STUDENTS MISS IT | EASIEST CLUE |
|---|---|---|
| A before/after pair can confuse secular trends, random variation, case mix and temporary enthusiasm with true | Two numbers are simple to present and often look persuasive, especially when the post value is better. | Plot the measure over time. Ask what happened before, during and after each change and what unintended |
| improvement. Iterative QI uses repeated measures, context and balancing outcomes to learn whether a | metric moved with it. | |
| change is working and under what conditions. | ||
| TOP 10 RESEARCH + QI CONCEPT - 10: A SUCCESSFUL CHANGE IS NOT FINISHED UNTIL IT SURVIVES ROUTINE WORK | ||
| DEEP EXPLANATION | WHY STUDENTS MISS IT | EASIEST CLUE |
| Testing proves a change can work locally; implementation makes it standard; sustainment and | A short-term win feels like completion. Ownership, training, maintenance and balancing measures are less | Remove the project champion for a week in your thought experiment. If the process collapses, you have |
| spread test whether the benefit persists across people, shifts and conditions. Scholarship then explains the | visible than the initial result. | a promising test - not a sustained system. |
TOP 10 RESEARCH + QI CONCEPT - 9: QI NEEDS MEASUREMENT OVER TIME - NOT JUST BEFORE VS AFTER — H. 10 HIGH-DISCRIMINATION RESEARCH + QI DECISIONS
context and learning so others can judge transferability.
TOP 10 RESEARCH + QI CONCEPT - 9: QI NEEDS MEASUREMENT OVER TIME - NOT JUST BEFORE VS AFTER — H. 10 HIGH-DISCRIMINATION RESEARCH + QI DECISIONS
| # | Scenario | Best next move |
|---|---|---|
| 1 2 | A large dataset is available, but the question is vague. A QI project may be published. Does publication make it research? | Define the decision/question and estimand first; use only data capable of answering it. No. Publication intent alone is insufficient; use the institutional research/QI/privacy determination process. |
| 3 4 | P<0.001 but the effect is tiny. Outliers are removed only after they weaken the result. | Report magnitude + interval + absolute/clinical context; do not equate p-value with importance. Use pre-specified/justified rules or transparent sensitivity analyses; document the deviation. |
| 5 6 | A key variable has substantial missing data. Twenty analyses were tried and one is positive. | Describe the pattern, assumptions and handling; test how reasonable alternatives affect conclusions. Disclose/label exploratory work; address multiplicity/selection rather than presenting the winner as pre-specified. |
| 7 8 | A contributor collected data but did not meet all authorship criteria. AI drafted prose and suggested citations. | Credit the contribution appropriately; authorship requires the full current criteria, not task count or seniority. Follow journal/institution policy, disclose as required, verify every claim/reference, keep human accountability. |
| 9 | QI improves in the first week. | Continue time-series/balancing measurement and additional PDSA cycles under varied conditions before hardwiring/spreading. |
| 10 I. FINAL RESEARCH / SCHOLARSHIP / QI MASTERY GATE | The project is ready to submit. | Run the governance + data + reporting guideline + authorship + AI + reference + reproducibility checklist first. |
TOP 10 RESEARCH + QI CONCEPT - 9: QI NEEDS MEASUREMENT OVER TIME - NOT JUST BEFORE VS AFTER — H. 10 HIGH-DISCRIMINATION RESEARCH + QI DECISIONS
| Gate | Pass condition |
|---|---|
| Question + design Governance + protocol | One answerable question/aim drives a design capable of supporting the intended causal, predictive, descriptive or improvement claim. Institutional research/QI/privacy/IRB pathway is documented; outcomes, eligibility, variables, roles and primary analysis/measures are versioned. |
| Data + analysis Reproducibility + integrity | Provenance, missingness, exclusions and transformations are auditable; magnitude, uncertainty, bias/confounding, multiplicity and assumptions are interpreted. A qualified person can reconstruct the key work; authorship, funding/conflicts, references and AI use are handled transparently. |
| Reporting + QI transfer | Correct reporting guideline/journal instructions are applied; improvement is measured over time with balancing effects, sustainment and context. |
TOP 10 RESEARCH + QI CONCEPT - 9: QI NEEDS MEASUREMENT OVER TIME - NOT JUST BEFORE VS AFTER — H. 10 HIGH-DISCRIMINATION RESEARCH + QI DECISIONS
| AHRQ PDSA | SQUIRE 2.0 | OHRP QI FAQ |
|---|---|---|
| 2026/27 research + scholarship expanded edition - local governance, patient safety, privacy, sponsor and journal requirements remain controlling |