Abstract
The Columbia-Bassett Medical School Program developed the SLIM (Systems, Leadership, Improvement, and Management) curriculum to give future doctors tools to improve health care systems. Medical students trained as Lean Six Sigma Green Belts lead clinical performance improvement projects using the DMAIC process: Define, Measure, Analyze, Improve, and Control.
Our project focused on medicine inpatient pain management at Bassett Hospital, a 180-bed, acute care teaching facility in rural Cooperstown, New York, affiliated with Columbia University College of Physicians and Surgeons. Pain management is one of 10 measures included on the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey. 1 On this survey, an average of 29% of inpatients say their pain is not always well controlled, leaving room for improvement. 2 Bassett’s HCAHPS pain scores were below the national average and lowest for inpatient medicine. Interviews with patients, nurses, physicians, and administrators during the Define phase identified possible drivers in our hospital that fell into 3 categories: timeliness, communication, and therapy choice.
In the Measure and Analyze phases, this qualitative information was used to design a quantitative patient survey. Timeliness of addressing pain (P = .016) and feeling the care team understood their pain (P = .008) significantly correlated with being “Very Satisfied” (Fisher exact test, n = 46) and were the top 2 factors in a Pareto analysis accounting for 71% of defects. This aligns with previous findings of significant drivers as prompt response, asking about pain, and expressing care and concern. 3
We chose to focus on response time for pain-related calls, from patient call to nurse contact. Pain call response time was within the customer specification limit of 3 minutes only 39% of the time (29/75). Process mapping identified that nurses are frequently occupied with other patients or tasks at the time of the pain call. To target this root cause, a procedure was created wherein the unit clerks trigger a backup system if nurses will be unable to answer a call within 3 minutes.
In the Improve phase, the intervention was introduced over 1 week and measured over the next 6 weeks; χ2 analysis or the Wilcoxon rank-sum test were conducted. During the pilot, 14% of pain-related calls used the backup system and nurses appreciated the help. The postintervention median pain call response time was 2.4 minutes (n = 130), a significant decrease from 3.5 (P = .006), and the percentage within 3 minutes increased from 39% to 60% (P = .003). Patient satisfaction had a nonsignificant increase from 63% to 67% reporting they were “Very Satisfied” and from 67% to 73% feeling their pain was always addressed quickly (P = .76 and P = .52; n = 24).
The intervention was successful in meeting the primary aim of decreasing nurse response time to pain calls. Further rounds of DMAIC cycles to address other drivers of patient satisfaction will be necessary to achieve a significant impact on HCAHPS scores.