01 · The constraint
Clinicians were idle while the front desk processed paper.
The clinic ran roughly 400 patients a month through paper intake. Patients filled forms out at the desk on arrival, which meant the queue stalled before anyone reached a room. Schedule density looked fine on paper; in practice the day was already 30 minutes behind by mid-morning, and missed slots compounded into lost revenue.
The first question was whether the bottleneck was supply (clinicians, rooms) or process. The data answered process: clinicians were idle a non-trivial fraction of the morning while the front desk worked through intake.
02 · The work
From paper at the desk to a digital pre-arrival workflow.
Pulled intake records, scheduling data, and patient-flow timings with SQL. Cleaned the patient panel against the Peruvian double-surname problem: each patient typically carries two last names (paternal and maternal), and the systems were inconsistent about which they stored, in which order, and whether spaces or hyphens separated them. Wrote a deduplication pass that resolved canonical patient IDs across years of Excel exports before any throughput analysis ran.
With clean data, the bottleneck story was unambiguous. Designed a digital pre-arrival workflow so patients completed forms before they walked in. Presented findings to clinic leadership and led the rollout across front-desk and clinical staff so the new flow actually got adopted, not just documented.
Built operational dashboards backed by the same SQL pipeline, tracking throughput, utilization, and intake cycle time so the clinic can see when the constraint shifts somewhere new.
03 · Outcome
Faster days, fewer empty slots, dashboards in place of paper.
Reference
"Katherine redujo nuestros tiempos de espera a la mitad y nuestra tasa de inasistencia un 40% en cuatro meses. Empezó limpiando datos de pacientes que nadie más quería tocar (solo el problema de los dos apellidos tomó semanas), después rediseñó nuestro proceso de admisión para que los pacientes llegaran preparados en vez de esperar en recepción. Los tableros que construyó son cómo opera el equipo ahora."
Translation: "Katherine cut our wait times by half and our no-show rate by 40% in four months. She started by cleaning patient data nobody else wanted to touch (the double-surname problem alone took weeks), then redesigned our intake so patients arrive ready instead of waiting at the desk. The dashboards she built are how the team operates now."
Dr. Mariela Seguil · Policlínico "El Bosque"04 · Behind the work
Notes from the engagement.
The hardest part of this engagement was not the workflow redesign. It was the data. Patient identity in Peru routinely involves two surnames, and the source systems handled them differently from each other. Resolving canonical patient IDs across Excel exports, paper-form data entry, and the scheduling system took longer than the actual operational work.
The lesson generalizes: operational interventions live or die on whether the data underneath them is reconciled. A throughput recommendation downstream of bad joins is worse than no recommendation at all, because the team acts on it. The first deliverable on any operations engagement should be a clean panel and an honest account of where the joins fail, before anything is proposed.
The dashboards were also a deliverable in themselves. Front-desk staff had been keeping their own paper check-in sheets to estimate throughput, which the clinic leadership did not have visibility into. Putting the same numbers in front of both made the change conversation possible.