Diagnosed Before I Arrived
When the Chart Tells the Physician What He Thinks
The latest update to my hospital’s version of Epic has made something explicit that had previously been implicit. My notes now automatically begin with diagnoses I did not make and cannot remove. Homelessness. Food insecurity. Substance Use Disorder. BMI greater than 30. These appear beside pneumonia, heart failure, and sepsis, as though they occupy the same category of thought. They do not. They are not diseases. They are descriptions of how a person lives.
The justification is, as always, the dollar, and more specifically, what the government requires one to do to earn it. Programs such as Merit-based Incentive Payment System reward the documentation of “complexity,” and social conditions have been folded into that accounting. The chart expands. The metrics improve. Higher payment follows. None of this improves the care delivered in the room. It does, however, reveal the conceptual framework now guiding much of modern medicine.
We are told these are “social determinants of health.” That language bothers me a great deal. The terminology was taught to me in medical school to refer to epidemiological variables that were associated with poor health outcomes, but the word “determinant” carries a misleading connotation, perhaps intentionally, although perhaps that is overly cynical. The ordinary meaning of “determinant” implies that the factor in question exerts a decisive influence, that it fixes the outcome in an inexorable way. That implication is never justified or even admitted, yet it is precisely how the term is now used in policy, documentation, and everyday clinical language.
I do not think this is a trivial semantic complaint. When a risk factor is described as a determinant, probability is equivocated with certainty, correlation masquerades as causation, and context is recast as mechanism. Most importantly, these so-called “social determinants” all reflect discrete, ongoing lifestyle choices made by human beings with agency, and yet the very term “determinant” strips away the idea of choice and replaces it with inexorable fate.
Consider what these categories actually represent. Homelessness does not, by itself, produce disease in the way that a pathogen does. It is associated with illness, often strongly, but there is no causal mechanism by which not having a home causes a disease. Camping is not considered a social determinant of health for this reason. Famed rock climber Alex Honnold spent a decade “homeless” and living out of a car at various crags and climbing locales, yet no one would consider him to be high risk for diabetic foot amputations. Why is that?
The association is entirely because “homelessness” as a social determinant functions as a proxy word, a euphemism to describe a pattern of extremely poor and unhealthy lifestyle choices. Poor nutrition, violence, criminality, mental illness, drug use, irregular care, and nonadherence all contribute, and those who we describe as “homeless” tend to behave this way. These choices are so deeply intwined in the connotation of homelessness that even though Alex Honnold was by definition homeless, no one ever described him that way, because he was at the same time athletic, healthy, famous, and rich, living at what essentially served as his gymnasium, instead of drug-addled, mentally ill and violent, harassing pedestrians on the streets of L.A.
The same is true of substance use and obesity. There are direct physiologic effects. There is also a broader constellation of choices that accompany these conditions, choices that more immediately drive outcomes. It is therefore more accurate to say that many so-called determinants function as proxies. They mark a way of living. They signal a pattern. They do not, in themselves, explain the outcome. When we attribute causation to the label, we obscure the mechanisms that matter.
This matters because people are not interchangeable units acted upon by external forces. They are agents. They act within constraints, but they act. Two individuals facing similar circumstances do not reliably arrive at the same place. Some adhere to treatment, seek care, and modify behavior despite disadvantage. Others do not. That variation is not incidental. It is central. It reflects differences in decision-making, priorities, and response to adversity. Put differently, it reflects differences in character.
The current framework does not account for this. By describing these factors as determinants, it suggests inevitability. The patient becomes the product of his environment. Outcomes are framed as the natural consequence of circumstance rather than as the result of an interaction between circumstance and choice. This is often presented as compassion. It has the opposite effect. It strips the patient of agency. It reduces a human life to a set of inputs and outputs.
When the electronic record automatically assigns these conditions as diagnoses, it encodes this worldview into the practice of medicine. The chart becomes a document not only of illness but of ideology. It asserts, before any clinical reasoning has taken place, why the patient is unwell. It does so in a way that is convenient for measurement and reimbursement, not necessarily for understanding or care.
It also directs attention away from the proximate drivers of health. If the problem is homelessness, the solution becomes housing. If the problem is food insecurity, the solution becomes access. These may be worthwhile aims. They are not sufficient explanations. They do not address the behaviors that mediate risk: adherence, follow-up, diet, substance use, and the countless small decisions that accumulate over time into health or illness. Without engaging those, the label remains a description, not a solution. Simply giving the undomiciled a house would not make an abused body healthy, for it would not change any part of the lifestyle of homelessness. Rather, I predict, it would nearly immediately destroy what was previously a perfectly nice house.
The defenders of this system will object, and their objection deserves a direct answer. The argument goes as follows: even if these conditions are not diseases in any strict sense, documenting them serves a clinical purpose. It alerts the care team. It triggers social work referrals. It connects patients to community resources. The physician who sees only the sore throat, the argument continues, misses the forest for the trees.
This is not an unreasonable position. It would be convenient if it were supported by evidence. It is not, at least not in any robust way. The published literature on SDOH documentation in electronic records shows consistent improvement in one category of outcome: process measures. Screening rates go up. Referrals are generated. Boxes are checked. What the evidence does not reliably show is improvement in actual health outcomes or reductions in healthcare utilization and cost. A systematic review examining 28 studies, half of which were randomized controlled trials, found that while patients demonstrated improved access to resources after SDOH screening interventions, findings on the effects on health outcomes and cost were mixed.1 Generating a referral and improving a patient’s health are not the same thing, and the literature treats them as interchangeable far too often.
This matters because the entire justification for mandatory, non-removable SDOH documentation rests on the assumption that documentation leads to action and that action leads to better outcomes. The first link in that chain is weak. Even proponents of EHR-based SDOH integration acknowledge that the evidence base remains limited, and that barriers to actually capturing codes and connecting patients to resources persist despite the presence of documentation infrastructure.2 The second link is weaker still. There is, as yet, no compelling demonstration that routine SDOH documentation in emergency or acute care settings produces meaningful improvements in the chronic social conditions it purports to address.
The financial architecture underlying this effort makes the situation worse. CMS has now elevated housing instability codes to “complication or comorbidity” status under the Medicare Severity Diagnosis Related Groups payment system, directly tying SDOH documentation to higher inpatient reimbursement.3 The incentive is therefore not to improve the patient’s housing situation. The incentive is to document that the housing situation exists, collect the additional payment, and move on. One should not be surprised when a system optimized for documentation produces excellent documentation.
The broader MIPS framework, of which SDOH reporting is now a component, has already been shown to operate this way. A 2022 study in JAMA examining more than 80,000 primary care physicians found that MIPS scores were inconsistently associated with performance on clinical process and outcome measures, and that physicians caring for more medically complex and socially vulnerable patients were more likely to receive low scores despite providing high-quality care.4 A metric that systematically disadvantages physicians who treat the most difficult patients while rewarding documentation compliance is not a quality metric. It is an administrative exercise. Appending social determinants to that exercise does not redeem it.
There is also a more immediate objection, one that will be familiar to any emergency physician. The argument for SDOH documentation assumes a continuity of care in which a flag placed in the record propagates meaningfully through a system. In the emergency department, that assumption fails almost entirely. The patient presenting with chest pain at two in the morning will be seen, treated, and discharged or admitted on the basis of the clinical picture in front of me. The fact that my note begins with his BMI, placed there automatically and without my input, does not trigger a nutritional intervention. It does not prompt a behavioral health referral that would not otherwise have occurred. It does not change his diet on the drive home. It sits in the chart, contributes to a complexity score, and supports a billing calculation. The patient is not meaningfully better for it.
None of this is an argument against social workers, community health programs, or genuine efforts to connect vulnerable patients with resources they need. Those efforts have value, and physicians who work in continuity practices have reasonable tools to pursue them. The argument is against the specific practice of encoding social circumstances as automatic diagnoses in the medical record, presented as equivalent to clinical findings, non-removable by the treating physician, and justified primarily by a reimbursement incentive rather than demonstrated patient benefit.
Social conditions influence health. They do not determine it. To speak as though they do is to misunderstand both the evidence and the patient. It replaces a complex human being with a simplified model, and it builds a system around that model. When that system is then reinforced by financial incentives that reward documentation rather than outcomes, the distortion compounds. The chart becomes more complete on paper and less faithful to reality.
What is lost in this arrangement is the physician’s most basic function: to meet the patient in front of him with clear eyes and accurate language. The automatic diagnosis forecloses that encounter before it begins. It tells the physician what he thinks before he has examined anyone. It tells the patient, implicitly, that his circumstances explain him, that the system has already accounted for what ails him before he has said a word. This is not medicine practiced in service of the patient, but rather, in service of the ledger.
The correction does not require abandoning concern for patients who face genuine hardship. It requires only honesty about what medicine can and cannot do, and about what the record is and is not. A chart is a clinical document. It should reflect what the physician observed, assessed, and decided. When it is converted into an instrument of social accounting, it serves neither function. The physician loses authorship of his own reasoning. The patient loses the physician who might otherwise have attended to him without a predetermined conclusion already entered at the top of the page.
Medicine’s defenders of neutral language mistake the absence of judgment for the presence of compassion. But judgment is precisely what the physician does. He distinguishes, draws distinctions, decides. A clinical vocabulary deliberately emptied of evaluative content is not a kinder vocabulary; it is an emptier one, and ultimately a less honest one. When we describe a circumstance as a diagnosis, we commit ourselves to a set of assumptions about cause, mechanism, and responsibility. Those assumptions have consequences, for how patients are understood, for how resources are allocated, and for what kind of care is actually delivered. The terminology of social determinism, embedded now in the electronic record by regulatory mandate, carries assumptions that do not survive scrutiny. Medicine should insist on better.
Yan AF, Chen Z, Wang Y, et al. Effectiveness of social needs screening and interventions in clinical settings on utilization, cost, and clinical outcomes: a systematic review. Health Equity. 2022;6(1):454-475. doi:10.1089/heq.2022.0010
Buitron de la Vega P, Losi S, Sprague Martinez L, et al. Implementing an EHR-based screening and referral system to address social determinants of health in primary care. Med Care. 2019;57(suppl 6):S133-S139. doi:10.1097/MLR.0000000000001029
Centers for Medicare and Medicaid Services. FY 2025 Inpatient Prospective Payment System Final Rule. August 1, 2024. Accessed May 16, 2026. https://www.cms.gov/medicare/payment/prospective-payment-systems/acute-inpatient-pps/fy-2025-ipps-final-rule-home-page
Bond AM, Schpero WL, Casalino LP, Zhang M, Khullar D. Association between individual primary care physician Merit-based Incentive Payment System score and measures of process and patient outcomes. JAMA. 2022;328(21):2136-2146. doi:10.1001/jama.2022.20619