DMADV: designing new healthcare processes with Six Sigma

DMADV (Define, Measure, Analyze, Design, Verify) is the best-known road map for Design for Six Sigma (DFSS). DMAIC improves a process that already exists. DMADV designs a new process, service or product so that it meets its requirements from the start. ASQ's Six Sigma Green Belt Body of Knowledge lists DMADV as a DFSS road map that lines up with DMAIC. Some sources call the last phase Validate rather than Verify; the idea is the same.

In a hospital or clinic, the five phases look like this:

• Define: state what the new process or service must achieve, its scope, the sponsor and the team, and who it serves.
• Measure: find out what patients, families and staff need from it, and turn those needs into measurable requirements with targets.
• Analyze: develop a few alternative designs and compare them against the requirements, including how each one could fail.
• Design: work out the chosen design in detail, including the steps, roles, space, equipment, forms and system changes, and test it on a small scale.
• Verify: run a pilot, check performance against the targets, and hand the process to its owner with a plan for monitoring it.

The choice depends on whether there is a process worth improving. Use DMAIC when a process exists, can be measured and has causes the data can reveal. Use DMADV when there is no process yet, such as a new clinic, service line, unit or screening program, or when the current process is so far from what is needed that fixing it step by step won't reach the target.

DMADV asks for more work up front, because the team has to understand requirements before there is a process to measure. The payoff is less rework later: problems that would surface after go-live are found while the design is still on paper or in tape on the floor.

The Measure phase is what sets DMADV apart. Teams gather the voice of the customer through interviews, focus groups, observation and patient and family advisors, then translate it into measurable requirements. A common tool is quality function deployment (QFD), whose matrix, the "house of quality," links each requirement to the design features that will meet it.

In a Swedish hospital's DFSS project to improve the medication process, QFD made the team aware that the process served several different groups of customers, and at the cardiology clinic the time doctors spent prescribing medication fell by more than 20%. At a children's hospital in Illinois, a pediatric ICU team used focus groups, affinity diagrams and Kano analysis to identify 15 requirements for daily morning surgical rounds, then built a new rounding model from a house of quality.

Published DMADV projects in healthcare are usually single-site reports, but they show the kinds of problems the method suits:

• Depression screening in trauma care: a level I trauma center in Ohio used DMADV to design a PHQ-2 depression screening process for admitted trauma patients. Completion rose from 60.7% to 80.6%, and weekend completion from 42.9% to 82.8%. Where the screening form was located turned out to be critical.
• Operation notes: a private hospital in Dublin, where only 20% of operation notes were completed accurately, used DMADV to co-design a digital operation note with mandatory fields. Afterward, all notes were completed digitally and legibly, and completing one took 30% less time.
• Staff flu vaccination: a New England hospital used DFSS to design its employee influenza vaccination process. The authors note that a deliberately designed process, with well-defined and documented steps, is easier to teach, monitor, control and transfer to other units.

Design projects tend to fail in two places. The first is requirements. Teams skip the patient's voice and design around staff convenience or a favorite solution, or they collect requirements but never turn them into targets, which leaves Verify with nothing to check against.

The second is the handoff. Go-live is treated as the finish line, the pilot data is never compared with the targets, and the new process has no owner once the project team disbands. The result is a new process that drifts like an old one.

• Set a measurable target for every critical requirement in the Measure phase.
• Keep a patient or family advisor on the team through Design, not just in the interviews.
• Plan the verification measures, and who will collect them, before the pilot starts.
• Name the process owner and the review cadence before the project closes.

• Improvement Projects supports DMADV alongside DMAIC, A3s and PDSA cycles. It ships a ready-made DMADV template, and project templates can be customized, so a DMADV project can follow the phases and prompts your coaches teach.
• Plan steps carry an owner, a team and a due date, so each phase's work is assigned and visible.
• Linked metrics show starting, current and goal values on the project, which gives the Verify phase its targets.
• Analysis tools such as process mapping and FMEA are available for design and risk work, and pilot tests can be recorded as PDSA cycles with their predicted and actual effects.
• Follow-up reviews after implementation record whether the new process held, and lessons learned carry over to the next design project.