From Shared Meals to Home Visits: How Taitung’s Vineyard Home Supports Indigenous Elders Living Across the City

On October 19, 2026, Taiwan's Council of Indigenous Peoples reported the opening of the Vineyard Home Cultural Health Station in Feng-tian Village, Taitung City. The council said Indigenous elders there are dispersed across urban and general residential communities rather than concentrated in one tribal settlement. The station will therefore combine on-site and outreach services according to each elder's functional condition and service preference.

This is more than another activity room. The release records 534 cultural health stations nationwide, 114 in Taitung County, and eighteen in Taitung City after the new opening. The operational challenge is to turn a fixed site into a care node that can reach elders across dispersed neighborhoods while maintaining dependable daily service.

The 2027 Cultural Health Station Implementation Plan, approved on October 7, supplies a fuller institutional baseline. It traces today's network to daytime care stations created in 2006. The network grew from 121 stations serving 4,259 people in 2016 to 534 stations serving 19,038 people in 2026. Its goals include culturally safe care, stable professional staffing, and responsiveness to urban needs.

Services are not a single class. The plan covers basic health support, culturally grounded health and disability-prevention activities, nutritious meals, home visits, telephone check-ins, needs meetings, and locally appropriate help such as accompaniment outside the home or to medical care. Stations generally open five weekdays for at least eight hours. A station serving twenty to twenty-nine registered elders receives two care workers, with staffing increasing by service tier.

These rules clarify the distinction between opening and operating. A sign and a room establish an entrance, but public value emerges through repeated contact, judgment, service, and follow-up. In an urban setting, having a station does not mean every elder can find it, and making contact does not mean every elder can or wants to attend regularly.

From an Indigenous local-economy perspective, a care station is social infrastructure. It can affect elder health, unpaid family-care time, local care employment, meals, and transport, and can reveal risks earlier. The release does not publish Vineyard Home's registered population, service radius, on-site and outreach shares, transport arrangements, referrals, or initial outcomes. Those benefits must not be treated as already proven.

Five care-continuity cards can turn the institutional rules into a practical observation framework. They are not five new forms for elders to complete repeatedly. They connect who remains unreached, how an elder chooses services, what happens each day, how risks are handed over, and whether care continues.

The first is a service-area card. It maps the real area that staff can reach rather than merely listing an administrative district. It records neighborhood distribution, walking and transport conditions, language, accessibility, living arrangements, permitted contact channels, and the person responsible for periodic updates. Public versions remain aggregated and never display home addresses.

The card asks who may be missing. Regular attendees are visible, while elders who seldom leave home, live farther away, lack digital access, or have hearing and mobility barriers may not appear on ordinary activity lists. Community organizations, care networks, and families can help locate needs, but each contact must explain its purpose and respect refusal.

The second is a service-choice card. The elder, with a trusted supporter if desired, records preferred attendance days, outreach methods, language, meals, activities, health concerns, contact times, and activities they do not want. Each item is optional and changeable. A blank field never becomes presumed consent.

On-site and outreach support are not permanent labels. An elder may attend routinely but need home contact during severe weather, recovery, or a transport interruption. Another may begin with telephone contact or a short visit and later join meals and activities. The card keeps reasons for changes so service follows real life.

The third is a daily-care card. Each contact records only necessary facts: method, service, meals, any vital signs taken under procedure, functional or emotional observations, the elder's own words, follow-up tasks, owners, and deadlines. Cultural activities also record language, facilitator, context, and elder feedback so culture does not become a decorative class title.

Attendance is not a grade assigned to an elder. Lower participation may reflect transport, family schedules, illness, unsuitable activities, or physical barriers. Staff workload must also capture outreach travel, meal preparation, family contact, accompaniment, documentation, and handover. Otherwise, expanding service simply shifts hidden work to frontline staff.

The fourth is a risk-and-referral card. When workers observe fall risk, abnormal vital signs, nutrition, medication, emotional distress, possible cognitive decline, unsafe housing, family-care strain, or possible abuse, the card records only necessary observations, urgency, the elder's preference, the receiving contact, time, acceptance status, and reply deadline. Details do not circulate in general chat groups.

Sending a message does not complete a referral. The station must know whether the receiving service replied, whether the elder can reach it, whether language and culture are understood, whether family members know the next step, and who reconnects if service breaks. Emergencies follow established professional procedures. AI cannot assign urgency or reject a case.

The fifth is a continuity-and-outcomes card. At seven, thirty, and ninety days it asks whether a newly contacted elder completed a needs conversation, whether service choices remain suitable, whether a referral connected, whether missed attendance was understood, whether family strain changed, and whether the elder wishes to continue. Annual review adds coverage, waiting time, staffing stability, and cultural-safety feedback.

Total visits are not enough. Repeated visits by the most mobile group can rise while others remain undiscovered. Outreach can rise while uncounted travel and documentation time erode quality. Public indicators need denominators, time periods, definitions, missing-data shares, and limitations so activity is not mistaken for better daily life.

The five cards can share an elder-selected identifier with layered access. Elders and authorized family supporters see their own choices and tasks. Care workers see what each contact requires. A station manager sees staffing, service gaps, and referral status. County and central agencies receive only necessary aggregates. Medical detail, addresses, family conflict, and non-public cultural knowledge stay out of general dashboards.

Retention must also be purpose-specific. Contact lists, health observations, referral data, and activity images serve different purposes. One consent cannot authorize unlimited reuse. Elders need to know why information is collected, who can see it, how long it remains, and how to correct or stop future use. Paper and digital systems both need staff-exit access revocation.

AI can assist with low-risk tasks: converting confirmed schedules into reminders, checking whether a record lacks an owner or deadline, aggregating anonymous service gaps, drafting multilingual notices, flagging overdue referral replies, and comparing staffing with actual workload. Staff verify every output.

AI must not infer Indigenous identity or health from names, images, accents, or addresses; predict that low-attendance elders are not worth contacting; select consent on an elder's behalf; invent home-visit records; rank service priority with one score; or place case data into external models. Resource decisions require transparent rules, professional judgment, and elder participation.

Frontline conditions matter as much as data. The opening release mentions nonslip surfaces, safe entrances, handrails, emergency support, professional training, and defined roles. The cards should connect facility checks, completed training, relief staffing, and incident response. A system must not hide unsafe space or staff shortages behind high completion rates.

A minimum version needs no large platform. Vineyard Home can begin with an aggregated living-area map, four access-separated forms, a referral-status board, a weekly handover meeting, and the three follow-up intervals. After four to eight weeks, elders, families, care workers, and the operating organization can remove unused fields and add real service gaps.

Becoming Taitung City's eighteenth cultural health station is a new service entrance, not the end of evaluation. Official evidence establishes the dispersed urban context and supplies a formal service baseline. The next evidence must show whether people in need are found, can choose suitable service, have risks accepted by a responsible party, and receive continuous support.

The five cards keep coverage from missing people, prevent service from overriding choice, separate daily care from raw attendance, require referrals to reach an outcome, and measure continuity through elder experience. When cultural safety, feasible frontline work, and data minimization operate together, a station becomes genuine infrastructure for aging in place.

Sources

• Council of Indigenous Peoples, Taiwan: Opening of the Vineyard Home Cultural Health Station
• Council of Indigenous Peoples, Taiwan: 2027 Cultural Health Station Implementation Plan