Scaling Remote Work: What 22,000 Employees Taught Me About Policy at Scale

Lessons from chairing the remote work committee at the largest municipal health system in the United States

On March 13, 2020, I walked into what I thought would be a routine meeting about telework policy updates at LA County Department of Health Services. By the time I walked out, I was chairing an emergency committee tasked with sending thousands of employees home—immediately—while keeping a $4 billion health system operational.

The largest municipal health system in the United States had approximately 22,000 employees. About 60% of them had jobs that could theoretically be done remotely. None of our systems, policies, or management practices had been designed for that reality. We had six days to figure it out.

The First 72 Hours

Our initial instinct was to create a comprehensive remote work policy. Define eligibility criteria. Establish approval workflows. Document equipment requirements. Build tracking systems.

We abandoned that approach within 48 hours.

The problem wasn’t that comprehensive policy was wrong—it’s that comprehensive policy takes months to develop properly, and we had days. Every hour spent debating eligibility criteria was an hour that employees who could work from home were instead commuting to facilities where they might get sick or get others sick.

We made a decision that felt reckless at the time: any employee whose manager believed they could work remotely should go home immediately. No forms. No approval chains. No equipment provisioning. Just go.

This created chaos. Employees working from personal laptops that couldn’t access critical systems. Managers who had no idea how to supervise people they couldn’t see. Teams that had never used video conferencing trying to coordinate patient care remotely. IT support tickets increased 400% in the first week.

But it also meant that by March 20, we had reduced on-site headcount by roughly 8,000 employees. Given what we learned later about transmission dynamics in the early pandemic, that decision probably saved lives.

What We Built in the Following Months

Once the immediate crisis passed, we had to build sustainable systems. The committee I chaired met weekly for eighteen months. Here’s what we learned:

Job Classification Is Harder Than It Looks

Our first major project was classifying every position by remote work eligibility. This sounds straightforward. It wasn’t.

Consider a medical records technician. The job is primarily computer-based—scanning documents, entering data, responding to requests. Clearly remote-eligible, right?

Except: some records are physical and require on-site access. Some requests are time-sensitive and require proximity to clinical staff. Some systems can only be accessed from specific terminals for security reasons. And the employee’s home environment might not meet HIPAA requirements for handling protected health information.

We ended up with four categories:

Fully Remote-Eligible: Position can be performed entirely from a remote location with no regular on-site requirements.

Hybrid-Eligible: Position benefits from some in-person presence but core duties can be performed remotely. These positions might require 1-3 days per week on-site.

On-Site Required: Position cannot be performed remotely due to physical requirements, security constraints, or direct patient care responsibilities.

Situationally Remote: Position is normally on-site but can work remotely during specific circumstances (personal illness, family emergency, facility closure).

The challenge was that reasonable people disagreed about where specific positions fell. A supervisor who valued in-person interaction might classify a position as on-site required. A different supervisor might classify the identical position as hybrid-eligible. We needed consistency without micromanaging every classification decision.

Our solution was example-based guidance. Rather than abstract criteria, we published detailed examples: “A position like X is typically classified as Y because of Z.” Supervisors could then reason by analogy. It wasn’t perfect, but it created enough consistency to be defensible.

Equity Issues We Didn’t Anticipate

Within weeks of our initial remote deployment, we started hearing complaints that felt like they were about fairness but were actually about something deeper.

“Why does accounting get to work from home when I have to come in?”

The surface answer was obvious: accountants can do their jobs from anywhere with a laptop, while facilities staff need to be physically present to maintain buildings. But the complaint wasn’t really about logic. It was about the fact that remote-eligible jobs and non-remote-eligible jobs weren’t evenly distributed across the organization.

Remote eligibility correlated strongly with education level and salary. The people who could work from home were disproportionately white-collar professionals. The people who couldn’t were disproportionately frontline workers—often lower-paid, often workers of color, often in roles with less organizational power.

This wasn’t something a telework policy could fix. But acknowledging it mattered. We made sure that any discussion of remote work benefits also addressed what we were doing for employees who couldn’t work remotely: hazard pay, enhanced safety protocols, priority vaccination, additional leave provisions.

The employees who had to come in didn’t resent the employees who could stay home. They resented being invisible in conversations that only talked about remote work as if it were the only thing that mattered.

Management Capability Was the Binding Constraint

We initially focused on technology: laptops, VPN capacity, collaboration tools, security. These were real problems, but they were solvable problems. IT threw resources at them and they got fixed.

The harder problem was management.

About 40% of our supervisors had never managed remote employees before. Their management approach assumed physical presence: walking the floor, observing work in progress, being available for impromptu questions. Remote work broke all of these patterns.

Some supervisors adapted quickly. They established regular check-ins, learned to evaluate results instead of activity, built new rhythms for their teams.

Others struggled. A subset responded by attempting surveillance—requiring employees to be on video all day, checking online status constantly, demanding detailed activity logs. This destroyed trust and drove away good employees. Another subset essentially abdicated management, becoming unreachable and leaving their teams without direction.

We ran management training sessions. We shared best practices. We coached struggling supervisors individually. It helped, but the honest truth is that some supervisors never adapted. When we eventually implemented a hybrid model, the departments that struggled most weren’t the ones with the most complex work. They were the ones with supervisors who had never developed the skills to manage for outcomes rather than presence.

The Return-to-Office Trap

By late 2021, with vaccines widely available, we faced pressure to “return to normal.” Leadership wanted employees back in the office. The question was: which employees, how often, and why?

We made a mistake here that I’ve since seen other organizations repeat. We framed the conversation as “return to office” instead of “what work model makes sense going forward.”

“Return to office” implies that pre-pandemic was correct and remote work was a temporary exception. This framing made it hard to have honest conversations about what we’d learned. Some teams were demonstrably more productive working remotely. Some functions had been overbuilt for in-person work that didn’t actually require presence. Some employees had relocated, changed their childcare arrangements, or reorganized their lives around remote work.

The organizations that handled this transition best were the ones that started fresh: What does each role actually require? What work model optimizes for outcomes? They treated the pandemic as an opportunity to rethink assumptions rather than an interruption to restore.

We eventually landed on a hybrid model with genuine flexibility—not “everyone comes in Tuesday through Thursday” but “your work determines your presence requirements, and your manager has discretion within guidelines.” It’s messier than a uniform policy, but it better reflects the actual diversity of our work.

What I’d Do Differently

If I were advising an organization implementing remote work at scale today, I’d emphasize:

Start with job analysis, not policy. Understand what each role actually requires before writing rules about where it can be done. Most organizations haven’t rigorously examined their jobs’ actual location requirements in decades.

Invest in management capability first. Technology problems get solved. Management problems persist. The organizations that do remote work well are the ones with supervisors who can manage for outcomes, not presence.

Address equity proactively. Remote work creates winners and losers. If you only talk about the winners, you’ll lose the trust of everyone else. Be explicit about what you’re doing for employees who can’t work remotely.

Don’t optimize for surveillance. The instinct to monitor remote employees more closely than in-office employees is understandable and counterproductive. If you can’t trust someone to work without constant observation, you have a hiring or management problem, not a remote work problem.

Plan for the hybrid middle ground. Fully remote and fully in-office are both easier to manage than hybrid. But hybrid is where most knowledge work will land. Build systems for the messy middle, not the clean extremes.

The Lasting Change

Three years later, the health system still operates with substantial remote and hybrid work. Not as much as at the pandemic peak, but far more than before. Patient services remain primarily in-person, as they should. But administrative functions, data analysis, planning, and support work happen from wherever makes sense for the work and the employee.

The employees who gained flexibility value it enormously. The organization gained access to talent that wouldn’t have considered a daily commute to downtown Los Angeles. The facilities footprint has shrunk, reducing real estate costs.

The pandemic forced us to run an experiment we never would have chosen. The results were clear enough that we didn’t go back to the old model. That’s probably the strongest evidence that the change was real.

Remote work at scale isn’t about technology or policy. It’s about management capability, equity, and honest assessment of what work actually requires. The organizations that learned this during the pandemic will retain the benefits. The ones that treated it as a temporary exception are still figuring out why their employees are unhappy.