TY - JOUR T1 - Value of small sample sizes in rapid-cycle quality improvement projects JF - BMJ Quality & Safety JO - BMJ Qual Saf SP - 202 LP - 206 DO - 10.1136/bmjqs-2015-005094 VL - 25 IS - 3 AU - E Etchells AU - M Ho AU - K G Shojania Y1 - 2016/03/01 UR - http://qualitysafety.bmj.com/content/25/3/202.abstract N2 - Quality improvement initiatives can become bogged down by excessive data collection. Sometimes the question arises—are we doing an adequate job with respect to a recommended practice? Are we complying with some guideline in at least X% of our patients? The perception that one must audit large numbers of charts may present a barrier to initiating local improvement activities. The model for improvement and its Plan–Do–Study–Act (PDSA) cycles typically require frequent data collection to test ideas and refine the planned change strategy. The perception that data collection must involve many patients can lead to insufficiently frequent PDSA cycles.1 In this review, we demonstrate the important contributions that small samples can make to improvement projects, including local audits, PDSA cycles and during broader implementation and evaluation.Suppose you are a hospital-based clinician who has joined a medication reconciliation working group. Medication reconciliation refers to efforts to avoid unintentional changes to medication regimens at transition points such as hospital admission and discharge.2 You notice that medication reconciliation did not occur for several patients on your service this week. Your institution sets a target medication reconciliation rate of at least 80%, based on external standards and internal commitments to patient safety. You decide to audit 20 consecutive admissions, and find that only 10 charts (50%) have completed medication reconciliation. You present your findings at the weekly team meeting. Your colleagues tactfully point out that your sample is far too small to draw any meaningful conclusions.Surprisingly, your sample of 20 consecutive admissions actually provides strong evidence that local performance falls short of your performance target. If your service were actually performing medication reconciliation 80% of the time, a sample of 20 charts would produce an observed reconciliation rate of only 50% (or worse) about three times out of every 1000 similar audits. … ER -