Research brief
Proposed first evaluation: one serious-illness conversation workflow
Version: build cad5eff, built 2026-11-14, dataset v1.24.2. Sharing this brief with an institution is not endorsement by it. Values in Care is an independent educational project. No employer, health system, payer, religious body, or other organization is affiliated with, endorses, sponsors, or reviews it unless that relationship is documented here.
Problem
What matters to a person can be lost, misattributed, or left disconnected from the conversations where care is decided. This is the premise the product is built on. It is a hypothesis for collaborators to test, not a finding this project has measured.
What exists todayBuilt, not evaluated
Software that exists now. None of it has been evaluated with users or in care. Use non-identifiable notes only; a shared export is not anonymized.
- Decision Guide: a person names a decision, marks priorities without ranking them, and prepares questions. Selection order never implies importance.
- Conversation workspace: records whose words priorities are (patient, surrogate, professional, or unknown), whether the wording was confirmed with the person and when, unresolved disagreement, responsible role, and a revisit trigger.
- Confirmation tied to wording: editing confirmed priorities, or whose words they are, clears the confirmation until the person confirms again.
- Sharing review: the person marks each section, and each selected context (its label, questions and source link), Share, Keep private or Decide later. Only Share reaches the preview, print, download or handoff. Keeping a section private does not remove the same information if it also appears in a shared section. Shared outputs do not name or count what was left out.
- Browser storage on the device, or session storage for this tab only on shared devices. Preparation notes are not automatically sent to project servers or analytics. Copying, downloading, printing or sharing can create copies outside the platform.
- Twenty-eight care contexts offer optional prompts, sourced international examples and fictional cases. These support future adaptation; the initial proposed evaluation focuses on serious illness. A caregiver's own needs have their own field.
- Sourced evidence register with a public integrity report, corrections log and instrument-rights register.
- Spanish and Arabic interface text only. Content is English. Not reviewed or culturally validated.
Where the evidence stands
- Software checks:
- CI is configured to run unit tests and core browser checks on code changes, in pull requests and on the main branch, and the full browser suite, including accessibility and print checks, weekly or on demand. Docs-only changes skip them. Only completed, passing runs for the demonstrated commit count. They cover the tested behaviors and are not evidence that it helps anyone.
- Usability research:
- User-testing sessions run: 0. Participants: 0. Findings: 0.
- Expert review:
- Completed: 0. None is claimed beyond what the public review log shows.
- Clinical effectiveness:
- Not studied. No outcome or clinical-validity data.
- Approvals:
- No institutional endorsement, partnership, ethics review or data-use agreement.
Research on chaplaincy, shared decision-making and services in other countries describes those practices and settings. It is not evidence that this software helps, and listing many places does not show the workflow is valid or transferable across cultures or health systems.
Proposed first collaborationProposed
We seek a clinical or chaplaincy collaborator and a research-methods partner to evaluate one serious-illness conversation workflow using fictional cases. The first deliverable is a revised workflow, a record of comprehension and disclosure failures, and a feasibility protocol. No site or collaborator is committed.
The proposed added value is a brief with confirmed attribution, section-level sharing, preserved disagreement, and a clear next step. Compare it with a paper worksheet or the existing documentation workflow at similar facilitator time; superiority is not established.
Initial setting. Proposed, not selected: adults with serious illness who are offered spiritual care in one palliative care or spiritual care service, starting with English-speaking participants. The first stage uses fictional cases only. Collaborators choose the real setting, population and eligibility.
Proposed research questionProposed
“Does a chaplain-supported ValuesInCare workflow improve the accuracy, patient ownership and clinical usefulness of values communication, compared with the existing workflow, at an acceptable burden?”
To estimate what the platform adds, compare chaplain care as usual with chaplain care plus the platform. Extra time or attention from a chaplain, peer worker or other professional is not evidence that the software works.
Proposed chaplain-supported workflow
- Invite: ask whether the person wants to talk about meaning, faith, or what matters. They can decline or postpone.
- Listen: record it in their words.
- Confirm: read it back; record whether they confirmed it, and whose words these are.
- Choose what to share: the person chooses which sections the team sees.
- Prepare the handoff: a neutral summary with decisions, uncertainty and disagreement visible.
- Identify next steps: the support requested and the responsible role.
- Revisit: what should bring the group back.
Proposed evaluation stages A to D
- Stage A: Early co-design and comprehension, with fictional cases. Patients and caregivers, alongside chaplains and clinicians, work through the fictional cases and help shape the wording. Patients and caregivers take part through a route the research partner approves, such as a patient and family advisory group, not through a clinical encounter. Objective comprehension checks are kept apart from open-ended preferences. No real health information. Privacy, attribution and unwanted-disclosure failures count as substantive findings. Proposed; no session has taken place. Needs: A facilitator, an approved recruitment route, 5 to 8 participants per group, the protocol in docs/user-testing-protocol.md.
- Stage B: Bounded feasibility after institutional arrangements. One service, one workflow, a small number of consenting participants. The research partner decides the approved workflow, data collection method and governance for that study, including whether this tool, a paper version or another system fits. Measures: completion, burden, missingness, unintended effects. Needs: Institutional review, privacy and security review, and a site lead. None exists.
- Stage C: Comparative evaluation. Usual chaplain care compared with chaplain care plus the platform, so that added attention from a chaplain is not credited to the software. Design, outcomes and sample size are for collaborators to set. No sample size or effect is proposed here. Needs: A statistician, an ethics-approved protocol, agreed outcome measures.
- Stage D: Cross-setting and cross-language adaptation. Test what transfers across services, countries and languages, and what changes with staffing, financing, law and context. Translation requires qualified reviewers; it does not replace a medical interpreter. Needs: Local partners, reviewed translations, local ethics oversight.
Candidate measures (proposed, not validated)
Proposed measures, not validated scales. None is adopted, scored or embedded.
- Patient recognition and correction of the brief (perceived fidelity to their priorities)
- Usability and comprehension of the brief and the sharing choices
- Time a participant and facilitator need to complete a brief
- Sharing accuracy: each output holds what the person chose to share, and nothing else
- Control over what is shared
- Receiving clinician's understanding
- Whether the concern was addressed
- Requested support offered, declined, unavailable or completed
- Burden, missingness and unintended effects
- Preservation of disagreement
- Continuity: the next professional has the priorities without asking again
- Acceptability, including who declines or stops, and why
A person's values are never scored as right or wrong. Success is not defined as lower costs, greater hospice uptake, less treatment, greater religiosity.
For later adaptation to other care contexts: “Does the workflow help people recognize their own priorities in a brief, control disclosure, and communicate those priorities accurately to a care team, at acceptable burden?”
Data and governance
- Use non-identifiable notes only. Browser storage is not a place for patient records, and deleting cannot recall anything downloaded, printed or copied.
- Analytics records feature use only (event, feature, method), never typed text.
- Existing instruments are not embedded. FICA and FACIT-Sp need written permission or a license first; none is on file.
- A study with real participants needs a workflow, data collection method and governance its research partner approves. No single system is assumed for every study.
- A shared export is not anonymized.
- Chaplaincy credentials and models differ by country; U.S. certification is not assumed elsewhere.