A Person in Crisis Deserves More Than a Case Number
Updated October 29, 2026
Irma Perez was 43 when she died after being rescued from a drainage ditch in Lemon Grove, San Diego County.
Her family wanted answers. A neighbor said he had repeatedly contacted authorities before she was taken to a hospital. Local reporting documented the family’s concerns and the Sheriff’s Department’s announcement of an administrative investigation. This was a specific complaint about a response to someone in distress, not simply an abstract debate about homelessness. [NBC 7 reporting](https://www.nbcsandiego.com/news/local/woman-found-in-ditch-lemon-grove/3896228/), [ABC 10News reporting](https://www.10news.com/news/local-news/family-of-woman-found-in-lemon-grove-ditch-demands-answers-from-county-sheriffs-office)
There has since been a significant development.
On June 4, 2026, San Diego County’s Citizens’ Law Enforcement Review Board sustained two allegations in case 25-098, identifying her as Irma Perez Espinoza: failure to take appropriate action during a welfare-check response, and failure to activate a body-worn camera as required.
The findings describe a July 29, 2025, call, a deputy’s departure without locating her, and a subsequent failure to return after further information. Responders rescued her on August 1; she died in hospital on August 3.
The record also corrects an important detail: the witness did not observe her between July 29 and August 1. It therefore does not establish that she remained continuously in the ditch throughout that interval.
The board’s report records the medical examiner’s cause of death as complications of chronic alcohol abuse, including cirrhosis and acute pancreatitis, with environmental exposure contributing. The manner was classified as accidental. These medical findings and the sustained misconduct findings must both remain part of the account. [CLERB final findings, June 4, 2026](https://www.sandiegocounty.gov/content/dam/sdc/clerb/meetings/2026/06-04-2026/06-04-26%20Final%20Findings%20Report%20AMENDED.pdf)
A civilian oversight finding is not a criminal conviction. It also does not, by itself, establish that a different response would certainly have prevented her death. This review has not located a verified public record establishing the deputy’s final employment discipline, a criminal judgment, or a civil settlement.
Those limits matter. They do not erase the failure that the oversight process identified.
A person’s housing situation should never substitute for an assessment of whether that person needs help. Addiction, illness, poverty, and distress do not make someone less entitled to a competent response.
The public does not need an invented motive to ask why a welfare check was not properly completed.
The numbers must be updated, too
Multnomah County, Oregon, recorded 456 deaths among people experiencing homelessness in 2023. Their average age at death was 46. Those figures were accurately reported, but they are no longer the latest complete annual figures located for this review. [Associated Press reporting on the 2023 report](https://apnews.com/article/fe74f9747830a848aa28f82d93add582)
The county’s report released on December 19, 2025, recorded at least 372 deaths in 2024, an 18 percent decrease. Average age at death was 48. Overdose deaths fell from 282 to 214.
These are deaths among people experiencing homelessness, not a count of people who all froze outdoors or died on sidewalks. The 2024 total includes deaths in hospitals and under medical care. The county also warns that reporting expanded beginning in 2022, complicating comparisons with earlier years. [Multnomah County’s 2024 mortality report announcement](https://multco.us/news/multnomah-county-releases-domicile-unknown-report-homeless-deaths-occurring-2024)
The decline must be acknowledged. Continuing to say that deaths are rising every year would be inaccurate.
But a lower number does not make the remaining deaths acceptable. It should prompt a more useful question: which changes helped, which risks remain, and how can effective protection reach more people?
A decline alone cannot establish that one particular policy caused it. Equally, continued deaths do not prove that every service failed or that no one was helped.
Accountability requires better evidence than either conclusion.
Where did the money go, and what did it achieve?
Public spending deserves scrutiny. That scrutiny becomes stronger when it follows records rather than assumptions.
California’s State Auditor reported in April 2024 that the state had not consistently tracked and evaluated the costs and outcomes of its homelessness programs. Of five programs examined, auditors considered Homekey and the CalWORKs Housing Support Program likely cost-effective. They could not fully evaluate the remaining three because sufficient outcome information was unavailable. [California State Auditor, Report 2023-102.1](https://information.auditor.ca.gov/reports/2023-102.1/index.html)
That finding supports a serious criticism: public authorities lacked information needed to judge parts of their response.
It does not establish that all the money was stolen, that every program was useless, or that service providers deliberately kept people homeless to preserve their jobs. It was also a finding about conditions examined in that audit, not proof that every identified weakness remained unchanged in October 2026.
A separate 2024 audit found that San José and San Diego needed to improve planning and evaluation of their homelessness responses. These are specific institutional findings that can be tracked against later corrective action. [California State Auditor, Report 2023-102.2](https://information.auditor.ca.gov/reports/2023-102.2/index.html)
The demand should be concrete: show the contracts, explain the costs, publish meaningful outcomes, identify failures, and document what changed.
A program should be able to explain how many people received assistance, how long they waited, how many obtained stable housing, and how many returned to homelessness. Those measures also need context, including the needs of the people served, so that providers are not rewarded merely for selecting easier cases.
Why the original accusation needed correction
An earlier version of this article argued that institutions were living off the crisis rather than trying to resolve it.
That expressed anger, but it asserted an intention the evidence reviewed here does not establish.
An organization receiving funding to address a persistent problem is not, by that fact alone, deliberately preserving the problem. Salaries, administration, training, and coordination can be necessary parts of delivering services. The relevant questions are whether spending is reasonable, work is performed, records are accurate, and people receive useful help.
Where fraud, conflicts of interest, falsified outcomes, or deliberate neglect are documented, they should be named and investigated. Where evidence is missing, suspicion should remain a question rather than become a verdict.
Removing an unsupported accusation does not weaken the case for accountability. It directs attention toward conduct that can actually be examined.
Emergency response and long-term care both matter
A welfare check, emergency medical treatment, shelter access, addiction treatment, and permanent housing serve different purposes. One cannot simply replace the others.
An immediate concern requires an appropriate assessment. Long-term support requires somewhere safe to go, continuity of care, and a workable route toward stability.
Respecting a person’s choices does not mean assuming that every apparent refusal is informed or that every person in distress can navigate a complicated referral process. At the same time, concern does not automatically justify coercion. Decisions should rest on the person’s condition, applicable law, professional judgment, and a properly documented assessment.
“Someone contacted her” is not the same as “her needs were assessed.”
“A referral was made” is not the same as “care was received.”
“A shelter placement was offered” does not answer whether the placement was available, suitable, and accessible.
These distinctions are where public promises either become practical help or fail to do so.
People should not disappear inside administrative language
Terms used in public records can help agencies collect consistent data. The problem arises when a category replaces attention to the individual.
Housing status should describe a circumstance, not determine a person’s worth. A history of substance use should inform care, not become a reason to dismiss distress. A difficult encounter should be documented accurately, not turned into a permanent assumption that someone cannot be helped.
Families also deserve explanations that distinguish what is known, what remains uncertain, and what action followed a finding of failure.
Privacy may limit what can be released. It should not become a blanket substitute for explaining institutional decisions.
What remains to be followed
For Irma Perez Espinoza’s case, the next questions concern the Sheriff’s Office’s response to the sustained findings, any disciplinary outcome that can lawfully be disclosed, and any verified subsequent legal proceedings.
For Multnomah County, later mortality reports should be assessed with attention to reporting methods, causes of death, and whether the decline continues. Improvements should be recorded rather than ignored because they complicate an argument.
For homelessness programs, audit recommendations should be followed through to implementation. Announcing a new reporting system is different from producing reliable information. Publishing a budget is different from demonstrating results.
Missing public information should be identified precisely. It should not automatically be described as concealment, nor should an unanswered question be treated as resolved.
Why this record remains open
A person does not have to be blameless, sober, employed, housed, or easy to help before their safety matters.
The responsibility to respond competently does not disappear because a situation is difficult. Neither does the responsibility to report accurately disappear because a story is upsetting.
This record will preserve established failures, acknowledge verified improvements, and correct claims that the evidence does not support. It will distinguish allegations from findings, findings from discipline, and discipline from criminal or civil judgments.
The purpose is to keep attention on what happened to people and whether the institutions responsible learned anything that will protect the next person.
Public concern should not end when a report is published. A report should make the next stage of accountability possible.
Record status: Updated through October 29, 2026. Civilian oversight findings in Irma Perez Espinoza’s case are confirmed. Final disciplinary and legal outcomes remain unverified in this review. The latest complete Multnomah County annual mortality report located covers 2024. Further verified developments will be added.