A research-informed
medical Spanish tool.
Consulta is built around peer-reviewed research in medical-Spanish education and language-concordance outcomes in US primary care. The practice activities draw on that literature. The AI rubric and numerical score are our experimental design, not a validated instrument.
Consulta is in beta testing. It is practice software, not a clinical tool, and is not intended to provide clinical guidance, diagnostic support, or decisions about real patients. The simulated patient is a language-learning persona; nothing it says should inform treatment of any real person. The coach’s score assesses your Spanish in a practice encounter and is not a clinical competency credential.
Practice communication, then reflect.
Medical-Spanish education research supports structured practice and assessment. Ortega and colleagues found that fluency ratings and medical-interview ratings differed in the same elective. That motivates our decision to keep Spanish feedback separate from clinical coverage; it does not establish that those skills are statistically independent or validate this app.
Possible misunderstandings, with speech-recognition uncertainty acknowledged.
Specific suggestions grounded in what you said.
Consistency of address and appropriate wording, without requiring imitation of an accent.
Language behaviors that may help a patient share. A withheld simulated fact is not evidence of poor Spanish.
Clarification, paraphrasing, and acknowledging when you do not understand.
An experimental score, not a credential.
The 0–100 score summarizes one practice transcript. We have not established its repeatability, agreement with bilingual expert raters, or relationship to ILR proficiency levels. It cannot assess pronunciation from text, establish clinical readiness, or authorize care without an interpreter. Focus on the concrete feedback and your next practice goal.
The simulated patient may respond inconsistently. Language-concordance studies inform this design, but do not prove that a particular phrase causes disclosure. The Hsueh review found mixed associations across studies; the Lopez Vera case study reported greater comfort in its setting and called for larger samples.
Use fictional information only.
Your recording is sent to Mistral for transcription. Transcript text goes to OpenAI for the patient reply and coach feedback. Patient reply text goes to the configured voice provider: ElevenLabs, OpenAI, or Mistral. Vercel hosts the application. These services process the content outside your browser.
Completed encounters and cached reviews are stored in this browser tab, without an application database. Preferences stay in browser storage. Delete this encounter on the review page clears its local transcript and cached reports. It does not delete provider-side records. Provider retention depends on the service and account configuration; we do not promise zero retention. Never enter real patient identifiers or confidential information.
Sources and scope.
These sources inform the educational design. None evaluates Consulta.
- Evaluation of a Medical Spanish Elective for Senior Medical Students: Improving Outcomes through OSCE Assessments
Ortega P, Park YS, Girotti JA. Medical Science Educator. 2017;27(2):329–337.
- Medical Spanish Standardization in U.S. Medical Schools: Consensus Statement From a Multidisciplinary Expert Panel
Ortega P, Diamond L, Alemán MA, et al. Academic Medicine. 2020;95(1):22–31. Published online 2019.
- Patient-Provider Language Concordance and Health Outcomes: A Systematic Review, Evidence Map, and Research Agenda
Hsueh L, Hirsh AT, Maupomé G, Stewart JC. Medical Care Research and Review. 2021;78(1):3–23. Published online 2019.
- Physician-Patient Communication and Satisfaction in Spanish-Language Primary Care Visits
Haskard-Zolnierek K, Martin LR, Bueno EH, Kruglikova-Sanchez Y. Health Communication. 2023;38(4):714–720. Published online 2021.
- A Case Study of the Impact of Language Concordance on Patient Care, Satisfaction, and Comfort with Sharing Sensitive Information During Medical Care
Lopez Vera A, Thomas K, Trinh C, Nausheen F. Journal of Immigrant and Minority Health. 2023;25:1261–1269.
Who built this.
- Johnattan García RuizMPH, MBACo-creator · Founder & Managing Director, DecilionLinkedIn →
- Adam BeckmanMD, MBACo-creator · Clinical advisorLinkedIn →
Consulta is developed by Decilion, a global health consultancy working across Latin America and the Caribbean. Decilion delivers health policy research, strategic planning, data visualization, and applied AI for global-health organizations, universities, and NGOs; in English, Spanish, and Portuguese.
Clients include Harvard T.H. Chan School of Public Health, Boston University SPH, Georgetown O’Neill Institute, U4 Anti-Corruption Resource Centre, and Movement Health Foundation.
decilion.com →