County briefing details shelter-referral practices, pilot data; officials cite HMIS gaps and propose open-referral conversation with Portland

Multnomah County's Homeless Services Department on Feb. 20 presented a briefing on shelter-referral practices, pilot data from eight shelters and next steps to improve data and referral transparency after a budget note from District 3 Commissioner Julia Brim Edwards.

Commissioner Julia Brim Edwards asked for the briefing last year after hearing community concerns that people living in neighborhoods adjacent to shelters sometimes could not access shelter services nearby. "Instead of coming to a conclusion about what was happening, that ask for data was the most important place to sort of just start the discussion," she told the board.

Deputy Director Anna Plum and Research Evaluation Analyst Senior Doc Ramblings described the county's current data and referral landscape. Plum said the county's January 2024 by-name list showed 5,394 people experiencing unsheltered homelessness but that staff are working to deduplicate records and prepare a monthly dashboard to show inflow, outflow and demographics. "We do not systematically collect data on where someone was prior to entering a shelter or other services," Plum said, explaining that HMIS is designed for federal reporting and that collecting prior-location information can create barriers for people seeking shelter.

To help fill that gap, Doc Ramblings led a two-month pilot (Oct.-Nov.) in which eight shelters from three providers recorded prior-living situations for people entering shelter. Ramblings said 465 individuals entered the pilot shelters during the period; about 70% were coming from unsheltered situations and 123 individuals (about 26%) were coming from previously sheltered or housed situations. Of the 465 entries, staff could map 335 responses (72%) with sufficient geographic detail.

Pilot mapping and shelter-level findings

The pilot mapped entrances by shelter and found both local and countywide patterns: some people entered shelters close to where they had been unsheltered, while many others traveled from other parts of the county. Examples provided during the briefing included:

- Willamette Center (congregate shelter, District 1): 222 entries during the pilot; 168 responses were mappable; several entrants reported living unsheltered close to the shelter before entry.

- Laurelwood Center (congregate, District 3): 98 entries; 72 mappable; at least seven entrants reported living unsheltered nearby.

- Walnut Park (congregate, District 2): 51 entries; 35 mappable; relatively few participants came from East County.

- Motel/alternative shelters (example: Stark Street motel, Menlo Park SRV): lower entry counts and most placements to motels were internal referrals from congregate shelters.

Ramblings said average travel distances varied by site; for example, downtown shelters tended to have shorter average distances because of the concentration of unsheltered people there.

Referral systems and policy issues

Plum described four major shelter systems the county manages or funds: family, youth, survivors of domestic violence, and adults, plus a smaller veterans system. Referral pathways differ by system: family placements are handled through coordinated access and consider geographic preference where possible; youth shelters can be first-come-first-served or case-manager referred; domestic-violence shelters keep locations confidential and use multiple referral paths.

The adult system, Plum said, is the most complex. She outlined three common shelter types (congregate, alternative/pod-style, motel) and two referral categories (open and closed). Open-referral shelters accept self-referrals or referrals from community partners. Closed-referral shelters accept referrals only from specified agencies, navigation teams or the Street Services Coordination Center (SSCC). Plum said many motels receive internal referrals from congregate shelters and that a set-aside program (about 100 beds across 11 shelters) takes referrals tied to the city's camping impact-reduction work.

Plum told the board that shelters sometimes prioritize neighborhood referrals at opening under "good neighbor" agreements, but that ongoing operations do not generally prioritize geographic preference; instead placement is typically based on service needs, identity and equity. She said the county is "beginning" a mapping of referral policies and hopes to convene a community-wide discussion on whether to standardize referral policies across jurisdictions and providers.

Data, HMIS and next steps

Plum and Ramblings said the county is procuring a new HMIS contractor and seeking geolocation and outreach data features in the new system. Plum said staff piloted an outreach tool, Survey 123, as part of the Housing Multnomah Now effort to capture geolocation in outreach data.

Next steps identified by staff included: completing a deduplicated population dashboard that will present monthly inflow/outflow and demographics; convening partners to discuss referral-policy standardization for the adult system; and incorporating geolocation features in the new HMIS contract. Plum and Ramblings emphasized that collecting prior-location data in shelter intake is not currently standard because it does not affect eligibility, adds administrative burden and may discourage people from seeking shelter.

Board reaction and policy proposals

Commissioners pressed staff for more detail. Commissioner Brim Edwards said she had sent a letter to Portland Mayor Ted Wheeler (referred to in the meeting as Mayor Wilson) requesting a shift to an open-referral approach across the shelter system and asked county staff to work with city staff to develop how an open-referral model would operate. Plum said the city's temporary alternative shelters and Safe Rest Villages currently use closed referrals tied to navigation teams and that the county has delayed planned transfers of some city-operated sites while the parties work through the details.

Commissioner Moyer raised concerns about exclusion policies at shelters and legal obligations under disability law, saying providers must ensure behavior-related exclusions comply with federal nondiscrimination obligations. Plum said the county has a small number of shelter programs designed for people with behavioral-health needs and that a lived-experience advisory committee will examine shelter exclusion practices.

Plum and Ramblings said staff will return with a memo and dashboards when deduplication and QA are finished, and that a community process on referral practices will begin in coming months. Plum said Oak Street Village had a soft launch and that the shelter-availability tool piloted for the set-aside beds would expand to show open-referral shelter availability by April and then be rolled out to outreach workers.

What was not resolved

Staff said the population dashboard will not include a consistent, person-level geographic field until the HMIS procurement and outreach-data solutions are in place. Commissioners requested more visibility on city-run closed-referral sites and how camp-clearance referrals are routed into shelters. Plum said the county does not systematically capture which specific camp a referred person came from via existing HMIS intake and that those data gaps are among the reasons for the pilot and the vendor procurement.

The briefing closed with an acknowledgement from county staff that the discussion on referral policy, equity and data collection is at an early stage and will continue with city partners and community stakeholders.