Why Your Doctor and Your Aide Cannot Talk to Each Other

Helen Park, 76, is a retired elementary school principal from Sacramento who spent thirty years running institutions and knows exactly how badly institutions can fail. She has a primary care physician she has seen for eleven years. She has a cardiologist she was referred to after a mild event three years ago. She has a home health aide named Rosa who comes three mornings a week through an agency her daughter found. She has a pharmacy two miles from her house where she fills every prescription. She has a daughter, Jennifer, who lives forty minutes away and who coordinates all of this by phone and by memory because there is no other way to coordinate it.

None of these providers talk to each other.

Three weeks ago, Helen’s PCP changed her blood pressure medication. The new prescription was called in to the pharmacy without a note to the cardiologist, because sending a note requires logging into a separate portal, and the PCP’s office uses one system and the cardiologist’s office uses another. The cardiologist does not know about the change.

Rosa noticed last week that Helen is dizzy in the mornings. She mentioned it to the agency nurse, who logged it in the agency’s care management software. The PCP cannot access that software. The observation sits in a system the relevant physician cannot see.

The pharmacy filled the new prescription. It did not flag an interaction with a magnesium supplement Helen’s daughter ordered from Amazon last fall, because the pharmacy does not know about the supplement. Helen did not mention it because she does not think of supplements as medications.

Jennifer took a week off work three weeks ago when this constellation of failures resolved into a fall. Helen is home and recovering. Jennifer is calculating, not for the first time, what she can afford to lose.

This is not a technology failure. The technology to share Helen’s medication change with her cardiologist exists. The technology to surface Rosa’s observation to the PCP exists. The technology to flag a supplement interaction at the point of dispensing exists, provided the pharmacy knows about the supplement. The failure is institutional.

Each of Helen’s providers has optimized its systems for its own workflow, its own billing requirements, its own liability exposure, and its own regulatory obligations. The PCP’s office uses an EHR from one vendor. The cardiologist’s office uses a different one. The home care agency uses a third system designed for home care documentation, not clinical communication. The pharmacy uses a dispensing system built around what it fills, not what the patient takes. These systems were not designed to talk to each other because the institutions that use them were not designed to coordinate.

The person at the center of these separate optimizations was expected to be her own coordination layer. Or to have a daughter who could be.

“Interoperability” is a word that appears frequently in healthcare policy discussions and means little to anyone sitting at a kitchen table. At the kitchen table it means this: your cardiologist can see what your PCP prescribed last week. Your aide’s observation about your dizziness reaches your physician before the fall, not after the emergency department visit. Your pharmacy knows about every medication, supplement, and over-the-counter product you take, not only the prescriptions it filled.

The infrastructure for this exists in fragments. Health information exchanges operate in most states with varying levels of provider participation. The 21st Century Cures Act established interoperability mandates for electronic health records. The Trusted Exchange Framework and Common Agreement, known as TEFCA, is expanding the technical plumbing for data sharing. CMS has billing codes for care coordination — Chronic Care Management, Transitional Care Management — that allow physicians to be paid for the coordination work Helen’s providers are not doing.

None of these frameworks has produced an environment where Rosa’s morning observation reaches Helen’s PCP before the fall. The technical and regulatory foundations are incomplete. But the larger obstacle is institutional, not technical.

Why has this not happened? Four reasons, and the reader deserves to know all of them.

First, data sharing reduces competitive advantage. A hospital that knows which of its patients are seeing competing specialists uses that information. A health system that integrates its data with independent home care agencies may reduce its own leverage over those referrals. Institutional data sharing is an institutional cooperation problem before it is a technical one.

Second, integration costs money. Building and maintaining interoperability infrastructure is expensive. The return on that investment accrues to the patient and to the overall system — fewer hospitalizations, better outcomes, reduced duplication — but not to the individual provider who pays for the integration. The incentives are misaligned at the institutional level in the same way they are misaligned at the agency level in 17.01.

Third, liability is unclear. If the cardiologist acts on information shared by the PCP’s system and something goes wrong, who is responsible? If the aide’s observation is surfaced to the physician and the physician does not act on it, what is the standard of care? The legal framework for shared clinical data is still forming, and the caution of legal counsel has slowed institutional decisions faster than the enthusiasm of clinicians.

Fourth, the person who bears the cost of fragmentation is not the person who bears the cost of integration. Helen and Jennifer bear the cost. Helen’s providers bear the cost of building systems that share data. This is a classic case in which the people with the ability to solve a problem do not face the consequences of leaving it unsolved.

The reader who cannot fix these institutional failures can still reduce their personal risk. These are workarounds, not solutions. They reduce the probability of the failure Helen experienced without eliminating the structural conditions that made it possible.

Maintain a complete medication list and bring it to every appointment. Not just prescriptions. Every supplement, every over-the-counter medication, every vitamin. Ask the pharmacist to review it for interactions every time a new prescription is added. This is a service pharmacists are trained to provide and rarely asked to perform, because most patients do not arrive with a complete list.

Ask your PCP whether the practice participates in the state health information exchange. The answer tells you whether your clinical information is moving between providers automatically or remaining siloed in one system. If the practice does not participate, ask why.

Ask your home care agency whether their documentation is accessible to your physician. Many agencies can establish a communication channel with a patient’s clinical team if asked. Few offer it proactively.

Ask your specialist to send a note to your PCP after every appointment. This was standard practice before electronic health records fragmented communication across incompatible systems. It is still available as a manual process when the automated one does not work.

The institutional changes that would have prevented Helen’s fall are not complicated to describe. They are complicated to implement because they require institutional cooperation that is not currently incentivized.

Mandatory participation in health information exchanges would mean that when Helen’s PCP changes her blood pressure medication, the cardiologist receives an automatic notification. Mandatory home health integration would mean Rosa’s morning observation enters a shared care record that Helen’s PCP can access in real time. Caregiver access to a shared care dashboard would mean Jennifer’s coordination work is supported by information rather than dependent on memory and phone calls.

Every one of these changes is technically feasible today. The obstacles are institutional: financial, legal, competitive, and organizational. Naming the obstacles does not dissolve them. It does allow the reader to evaluate what she hears from providers and policymakers about why integration “hasn’t happened yet.” The answer is not that the technology is not ready.

Helen’s fall was predicted. Rosa observed the dizziness six days before the fall. She logged it in the system she was trained to use. The observation was accurate. The clinical relevance was clear. The physician who could have acted on it never received it.

Jennifer took a week off work when the coordination failed. At her hourly rate as a marketing manager in Sacramento, that week cost approximately $1,400 in lost income, not counting the lost billable time from client relationships she managed by text from her mother’s apartment. The $870 billion in annual informal care labor that Blue Gray Matters documented in its reporting on the hidden economy of caregiving is composed of weeks like this one, multiplied across 53 million Americans who are performing coordination work that institutions should perform but do not.

Helen knows her providers are not talking to each other. She has known it for years. She assumed there was a reason, something technical she did not understand, some system that would eventually catch up. There is a reason. It is not technical. It is the result of institutional decisions about data, competition, liability, and cost, made by organizations that bear none of the consequences Helen bears when those decisions compound into a fall on a Tuesday morning in Sacramento.

She now has the questions to ask.