Pyridoxine-dependent epilepsy, or PDE, is caused by loss-of-function mutations in a gene called ALDH7A1. People with the disorder experience severe seizures that, remarkably, can be ameliorated by high doses of pyridoxine, better known as vitamin B6. That response has been the cornerstone of treatment for decades. Yet the vitamin does not solve everything. Psychiatric and cognitive symptoms can persist even after the seizures are controlled, leaving patients and families grappling with problems that the standard therapy does not touch. That clinical observation raised a fundamental question for the Johns Hopkins team: do those lingering symptoms arise through the same mechanism as the seizures, or through something else entirely?

To find out, the researchers engineered mice in which ALDH7A1 could be selectively removed from either the liver or from astrocytes, the star-shaped brain cells that help maintain the chemical environment surrounding neurons. This precision approach allowed the team to disentangle contributions from different organ systems in a way that a conventional whole-body knockout could not. The results were unambiguous. Mice lacking ALDH7A1 in their liver cells became more susceptible to seizures but did not develop the changes in mood and behavior seen in mice lacking the gene throughout the body. Conversely, mice lacking the gene specifically in astrocytes developed behavioral deficits without becoming any more prone to seizures.

The behavioral changes observed in the astrocyte-specific mice were detailed and telling. The animals showed depressive-like behavior, reduced motivation, passive coping and decreased self-care — a constellation of deficits that mirrors the psychiatric burden reported by people with PDE whose seizures are well managed. The researchers traced these changes to a disruption of the normal antioxidant defenses that help cells manage reactive molecules and maintain normal function. In the astrocytes lacking ALDH7A1, that protective machinery faltered, and the disruption was associated with reduced activity among neurons in the prelimbic cortex, a brain region known to be involved in regulating emotional behavior.

That mechanistic chain — gene loss in astrocytes, weakened antioxidant defenses, disturbed redox balance, altered neuronal activity in an emotion-regulating circuit — gave the team a concrete therapeutic target. Because antioxidant defenses can help maintain the balance of reactive molecules in cells, the researchers turned to sulforaphane, an antioxidant compound found in broccoli sprouts that is known to activate NRF2, a cellular pathway that helps protect cells from oxidative stress. When the researchers incorporated sulforaphane into the mice’s diet, it increased NRF2 levels in the ALDH7A1-deficient astrocytes, suggesting increased antioxidant protection and an improved redox balance within the affected brain cells.

Crucially, the compound’s effects were selective. Sulforaphane did not correct the increased seizure susceptibility associated with loss of ALDH7A1 throughout the body. That distinction, the researchers say, provided additional evidence that the psychiatric and seizure symptoms are driven by separate mechanisms, with the psychiatric changes linked to astrocytes in the brain and seizure susceptibility linked to dysfunction involving the liver. For a field that has often treated a genetic mutation as a single point of failure, the finding is a powerful demonstration that one gene can produce distinct disease processes in different tissues.

“An important implication of this work is that the symptoms we see in a rare neurological disorder do not necessarily have to come from the same place or through the same mechanism,” says Akira Sawa, M.D., director of the Johns Hopkins Schizophrenia Center and professor of psychiatry and behavioral sciences at Johns Hopkins Medicine. “In this case, we found that the liver and brain each make distinct contributions to the disease, which gives us a much clearer picture of how these symptoms arise.”

The separation of mechanisms also reframes how the psychiatric burden of PDE might be treated. Vitamin B6 remains the standard for controlling seizures, but sulforaphane could potentially be added to treatment to address the psychiatric symptoms associated with the disorder. “The fact that sulforaphane could improve the psychiatric phenotype while not preventing seizures was especially informative,” says Sawa. “It suggests that targeting the biology of the brain directly may be able to address psychiatric symptoms that persist even when the seizures themselves are controlled.”

The study adds to a growing body of evidence that neurological and psychiatric symptoms arising from the same genetic mutation can stem from distinct biological and metabolic processes. That principle may extend well beyond PDE. If a single inherited mutation can produce seizures through liver-centered metabolic dysfunction and mood and motivational deficits through astrocyte redox imbalance in the brain, then other neurodevelopmental and neuropsychiatric conditions may similarly harbor organ-specific mechanisms that have been blurred together under a single diagnosis. Dissecting those mechanisms, as this study did with tissue-selective genetic engineering, could open the door to therapies aimed at each pathway individually rather than at the gene in the abstract.

The researchers caution that much work remains before sulforaphane reaches the clinic. The findings provide a foundation for further study of the compound as a potential supplemental treatment for psychiatric symptoms associated with pyridoxine-dependent epilepsy, but future research will be needed to determine whether targeting astrocyte redox imbalance can improve these symptoms in people with the disorder. Mechanism-driven clinical trials of sulforaphane represent a potential next step, and any such trials would need to establish whether the NRF2 activation seen in mice translates safely and effectively to human patients. The federally funded study was supported by the National Institute of Mental Health and the National Institute on Drug Abuse. Researchers from the Johns Hopkins University School of Medicine departments of neuroscience, psychiatry and behavioral sciences, physiology, pharmacology and therapeutics, biomedical engineering and genetic medicine led the work, with collaborators from the National Institute on Drug Abuse Intramural Research Program, Heidelberg University and the University of Maine. For now, the study stands as a vivid example of how a humble broccoli-sprout compound, guided by rigorous mechanistic science, can illuminate the hidden architecture of a rare disease — and perhaps point the way toward treating the symptoms that medicine has long left behind.

Subject of Research: Distinct liver and brain mechanisms underlying seizures and psychiatric symptoms in pyridoxine-dependent epilepsy

Article Title: Study reveals both liver and brain pathways in differing symptoms of inherited metabolic disorder

Article References: Study reveals both liver and brain pathways in differing symptoms of inherited metabolic disorder. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: pyridoxine-dependent epilepsy, ALDH7A1, astrocytes, liver, sulforaphane, NRF2, oxidative stress, vitamin B6, seizures, psychiatric symptoms, Johns Hopkins, Science Advances

The controversy began when Drs. Goldman and Martin published a critique arguing that the Johns Hopkins study, led by Dr. Alba Azola along with Dr. Rebecca T. Veenhuis and Dr. Leah H. Rubin, had been framed in a way that overstated what its results could show. Their central claim was that the study, by recruiting patients from specialty clinics rather than from the general population, could not speak to how the RECOVER PASC score would perform as a population-based screening instrument. In their response, the Johns Hopkins team does not dispute the mathematical logic behind that concern. Instead, they argue that Goldman and Martin have misidentified the question the study was designed to answer in the first place.

That question, the authors explain, was deliberately clinical rather than epidemiological. Their investigation asked how well the RECOVER PASC score classifies individuals who received a clinical diagnosis of Long COVID after comprehensive, multidisciplinary evaluation in specialty clinics, compared with individuals who had documented SARS-CoV-2 infection but recovered without persistent symptoms. This is a fundamentally different exercise from estimating how the score would behave if applied to an entire community, where the mix of patients, symptom burdens, and competing diagnoses would look very different. The team emphasizes that their goal was to contribute to the ongoing independent evaluation and iterative refinement of the index, not to certify it as a definitive diagnostic test.

The technical heart of the debate concerns what statisticians call the reference standard, the benchmark against which a new test is measured. For many diseases, a laboratory assay or imaging finding can serve as an objective gold standard. Long COVID has no such benchmark. In the absence of a biological marker, the Johns Hopkins team turned to expert clinical diagnosis based on the 2024 consensus definition issued by the National Academies of Sciences, Engineering, and Medicine, which they describe as the most appropriate clinical reference standard currently available. They acknowledge candidly that this is a pragmatic comparator rather than a true gold standard, but argue that emerging research indices must be evaluated against something, and expert diagnosis grounded in a consensus definition is the best available option.

The authors also point to the evolution of the RECOVER index itself as evidence that such evaluation is expected and welcome. In 2024, the RECOVER-Adult Long COVID research index was updated to incorporate additional participant data, expanded symptom ascertainment informed by input from the patient community, and a revised symptom-weighting model and threshold. An index designed to be revised as new evidence accumulates, they argue, naturally invites the kind of external scrutiny their study provided. Testing a tool in settings that differ from those in which it was developed is a cornerstone of clinical measurement science, and the specialty referral cohort, with its rigorous phenotyping, offers exactly the kind of demanding test case that can reveal where an index succeeds and where it falls short.

One of the sharpest points of contention involved enrollment criteria. Goldman and Martin suggested that requiring participants to have at least one neuropsychiatric symptom, such as brain fog, biased the study toward higher sensitivity, inflating the apparent ability of the score to detect true cases. The Johns Hopkins team agrees that the criterion defines a specific clinical spectrum of Long COVID and must be weighed when interpreting the results, but they reject the suggestion that it contaminated the comparison. The requirement, they explain, reflected the design of the parent study funded by the National Institute of Mental Health and the clinical focus of their Brain Health Program, and it was explicitly described in the original manuscript. Crucially, participants were not selected based on their RECOVER PASC score or on meeting any component of the score’s threshold, and brain fog itself was not required for enrollment. The neuropsychiatric criterion, in other words, shaped the referral population under study rather than smuggling the index into the reference classification.

The choice of comparator group drew similar scrutiny. The Johns Hopkins study compared clinically diagnosed Long COVID patients against people who had documented SARS-CoV-2 infection and recovered without lingering symptoms, a design intended to test whether the score can discriminate between persistent illness and uncomplicated recovery. The authors concede that future studies comparing Long COVID with symptom-overlapping conditions, including other infection-associated chronic illnesses, myalgic encephalomyelitis/chronic fatigue syndrome, fibromyalgia, dysautonomia, and mood disorders, would provide important complementary information about the score’s differential diagnostic performance. Far from invalidating their findings, they argue, such studies would extend them, mapping the score’s behavior across a wider landscape of conditions that mimic or overlap with Long COVID.

On the question of predictive values, the two sides find firmer common ground. Positive and negative predictive values depend heavily on disease prevalence: the same score can yield very different predictive values in a high-prevalence specialty clinic and a low-prevalence community sample. The Johns Hopkins authors agree entirely that these measures should not be generalized beyond the sampled population, and they clarify that the predictive values in their study were presented as descriptive characteristics of the cohort rather than as estimates applicable to broader clinical or community settings. What survives this clarification, they insist, is the study’s principal observation: the RECOVER PASC score demonstrated high specificity against recovered SARS-CoV-2 controls, meaning it rarely mislabeled recovered individuals as having Long COVID, but showed limited sensitivity in a clinically characterized Long COVID cohort, meaning it missed a substantial share of expert-diagnosed cases.

That combination of high specificity and limited sensitivity carries real clinical weight. A score that rarely produces false positives but frequently produces false negatives could, if used as a gatekeeping tool, steer genuinely ill patients away from evaluation and care. The Johns Hopkins team’s willingness to highlight the score’s sensitivity limitation, even while defending their methodology, underscores that their aim is refinement rather than advocacy. They frame the exchange with Goldman and Martin as a dialogue between complementary rather than competing questions: their study characterizes performance in the specialty referral settings where patients with persistent post-COVID symptoms are actually evaluated, while population-based studies, which they call essential, would characterize performance across the full spectrum of SARS-CoV-2 recovery.

The broader lesson may outlast the dispute itself. No single study, the authors conclude, can fully characterize the performance of an emerging research index across all clinical settings; confidence is built instead through complementary studies conducted in community populations, primary care settings, specialty referral clinics, and symptom-overlapping comparator populations. For a condition as heterogeneous and contested as Long COVID, that incremental, multi-setting approach may be the only scientifically defensible path toward standardized classification. The work was supported by the National Institutes of Health and the National Institute of Mental Health, and the authors report no conflicts of interest. As research indices like the RECOVER PASC score continue to evolve, this exchange stands as a reminder that in diagnostic science, what a test is for often matters as much as how well it performs.

Subject of Research: External clinical validation of the RECOVER PASC research index for Long COVID diagnosis

Article Title: Letter to Editor External Clinical Validation of the RECOVER Research Index: A Response to Goldman and Martin

Article References: Azola, A., Veenhuis, R. T., & Rubin, L. H. (2026). Letter to Editor External Clinical Validation of the RECOVER Research Index: A Response to Goldman and Martin. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10851-3

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10851-3

Keywords: Long COVID, RECOVER PASC score, external validation, diagnostic index, SARS-CoV-2, sensitivity, specificity, NASEM consensus definition, specialty referral cohort, predictive values, clinical research, Johns Hopkins

BPH is not cancer, but it is hardly a minor inconvenience. The noncancerous overgrowth of prostate cells becomes increasingly common as men age, and as the gland swells it can compress the urethra, the narrow tube that carries urine out of the body. The consequences are familiar to millions of older men: frequent urination, repeated nighttime trips to the bathroom, a weak urinary stream, and the frustrating sensation that the bladder never fully empties. Despite how widespread the condition is, the biological events that set it in motion — and in particular the question of why it arises almost exclusively in one region of the prostate — have remained stubbornly unresolved. Senior author Vasan Yegnasubramanian, a professor of oncology, pathology, and radiation oncology and molecular radiation sciences at the Johns Hopkins Kimmel Cancer Center and director of inHealth Precision Medicine at Johns Hopkins Medicine, noted that although BPH affects millions of men worldwide, scientists still do not fully understand why it develops primarily in the transition zone while other regions of the gland remain relatively unaffected.

That regional puzzle is rooted in the prostate’s internal architecture. The gland is not a uniform organ but is divided into several anatomically and biologically distinct compartments, most notably the peripheral zone, the central zone, and the transition zone. These regions differ in their cellular makeup and in their vulnerability to disease, and the split is dramatic: BPH arises almost exclusively in the transition zone, the area surrounding the urethra, while the majority of prostate cancers emerge in the peripheral zone. Understanding why the same organ hosts two such different disease patterns has long been a goal of prostate biology, and it suggests that the answer lies in fundamental differences in the cells that populate each zone.

To pursue that question at single-cell resolution, the research team, led by Yegnasubramanian together with Rulin Wang, a research associate and the study’s first author, and Angelo De Marzo, professor of pathology at the Johns Hopkins University School of Medicine and associate director of cancer research pathology at the Kimmel Cancer Center, turned to single-cell RNA sequencing. This technology allows investigators to determine which genes are switched on or off in individual cells, rather than averaging signals across a bulk sample in which rare cell types can vanish into the noise. The team studied noncancerous prostate tissue collected from 10 men undergoing surgery for localized prostate cancer, obtaining samples from all three major zones of the gland and profiling gene expression in nearly 130,000 individual cells.

The first task was cartographic: identifying the major cell types present throughout the prostate and confirming how they were distributed across zones. With that atlas in hand, the researchers zoomed in on basal epithelial cells, a population that helps form the lining of the prostate and is thought to include progenitor-like cells capable of contributing to tissue maintenance and repair. Rather than finding a single homogeneous basal population, the analysis revealed that these cells split into four distinct molecular subtypes — a level of diversity that conventional histology had not exposed.

One of those four subtypes immediately stood out. Its cells expressed high levels of a gene called WIF1, along with several other genes, and their distribution across the gland was anything but random. Additional laboratory tests confirmed that these WIF1-positive basal cells were abundant in the transition zone but nearly absent from the peripheral and central zones. In other words, the researchers had found a cell population whose geographic footprint within the prostate matched, almost exactly, the territory where BPH takes hold.

The molecular profile of these cells offered further clues about what they might be doing. According to Wang, the analysis revealed that the cells possess molecular features associated with tissue remodeling that could contribute to prostate enlargement, as well as an ability to communicate with neighboring cells, suggesting they may help establish the unique biological environment of the transition zone. Tissue remodeling — the orchestrated breakdown and rebuilding of structural components — is a normal part of organ maintenance, but when it runs unchecked it can drive the kind of nodular overgrowth seen in BPH. A cell type equipped with remodeling programs and embedded in the exact region where that overgrowth occurs is a compelling candidate for a local driver of the disease process.

Perhaps most tellingly, the team found that WIF1-positive basal cells are expanded in BPH tissues. That expansion raises the possibility, the researchers say, that these cells contribute to the regional processes that make the transition zone particularly susceptible to BPH. The gene whose name the cells carry, WIF1, encodes a Wnt inhibitory factor, a secreted molecule that modulates Wnt signaling — a pathway with well-established roles in development, stem cell behavior, and tissue growth. The presence of a Wnt-pathway regulator in a remodeling-competent basal population enriched at the BPH epicenter suggests a plausible mechanism by which local signaling environments could shape where and how the gland enlarges, though the study’s authors frame this as a foundation for future work rather than a settled causal story.

The significance of the finding lies less in any single gene than in the strategy it validates. Bulk analyses of prostate tissue have long averaged together the signals of many cell types, obscuring the rare, regionally restricted populations that may hold the key to zonal disease patterns. By dissecting the gland zone by zone at single-cell resolution, the Hopkins team showed that the transition zone carries its own distinctive cellular identity, embodied in a basal cell subtype found almost nowhere else in the organ. That identity, the researchers suggest, may be what renders the transition zone uniquely prone to the proliferative and remodeling processes of BPH — and, by extension, may help explain why the peripheral zone instead becomes the favored soil for prostate cancer.

By uncovering this distinct cell population associated with the region where BPH begins, the research provides new insight into the cellular mechanisms underlying prostate growth and remodeling, and it offers a foundation for future studies aimed at understanding how the condition develops. The long-term hope, the researchers say, is that a clearer picture of the transition zone’s biology will ultimately support more precise ways to prevent, diagnose, and treat BPH — potentially allowing therapies that target the specific cell programs driving enlargement rather than broadly suppressing prostate growth. The study also involved contributions from Qizhi Zheng, Mindy Graham, Ajay Vaghasia, Jianyong Liu, Jordan Gregg, Tracy Jones, Anuj Gupta, Nicole Castagna, Yan Zhang, Kornel Schuebel, Jennifer Meyers, Alyza Skaist, Dixie Hoyle, Jasmine Kung, Jessica Hicks, Alok Mishra, Yuhan Yang, and William Nelson. The work was supported by the National Institutes of Health and the National Cancer Institute, the Prostate Cancer Foundation, and several philanthropic and institutional funds, including the Patrick G. Walsh Fund and the Maryland Cigarette Restitution Fund.

Subject of Research: Identification of a WIF1-positive basal epithelial cell type in the prostate transition zone associated with benign prostatic hyperplasia

Article Title: Newly identified prostate cell type may help explain why enlarged prostate develops in one region of gland

Article References: Newly identified prostate cell type may help explain why enlarged prostate develops in one region of gland. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: benign prostatic hyperplasia, prostate, transition zone, basal epithelial cells, WIF1, single-cell RNA sequencing, Johns Hopkins, tissue remodeling, prostate zones, urology, aging, cell biology

The research team, led by Shiv Ayappa together with Divya Manikandan, Katharine Press Callahan, Travis N. Rieder, Michael Fingerhood, and Margaret S. Chisolm, conducted semi-structured interviews with 20 participants who were in recovery from opioid use disorder. The interviews focused specifically on whether participants had experienced a “rock bottom,” what those experiences were like, and whether they had actually propelled the participants toward recovery. Two members of the study team independently coded each interview transcript using an inductive thematic analysis approach, resolving discrepancies through discussion and consensus before consolidating their findings into a final codebook — a method designed to let themes emerge from the data rather than imposing preconceived categories on it.

The headline finding is striking in its simplicity. Seventy percent of participants — 14 of the 20 interviewed — reported having experienced something they would describe as rock bottom. Yet only 10 percent, just two participants, said that a rock bottom experience actually motivated them toward recovery. In other words, the overwhelming majority of people who reached the kind of devastating low point that popular wisdom treats as a catalyst did not find in it any push toward change. The floor they hit was not a springboard; it was simply more suffering.

The researchers did not stop at counting. Their analysis identified four recurring themes that characterized rock bottom experiences: loss of social connection, character erosion, loss of resources, and a loss of the will to live. Participants described becoming isolated from family and friends, watching their sense of moral identity disintegrate, exhausting their financial and material supports, and in the most severe cases arriving at a state in which life no longer felt worth living. These are not the ingredients of motivation. They are the ingredients of despair — and despair, the study suggests, tends to entrench addiction rather than dissolve it.

Where, then, does the motivation to recover actually come from? For 85 percent of participants — 17 of the 20 — the answer lay in what the researchers call “turning points”: discrete events that redirected the trajectory of a life. Three themes dominated these turning points: parenthood, near-death experiences, and incarceration. Becoming a parent, or facing the prospect of losing a child, gave many participants a future worth protecting. Surviving an overdose or another brush with death confronted them with the fragility of the life they wanted back. And incarceration, whatever its harms, imposed a forced pause that some participants used to reorient. Crucially, these turning points differ from rock bottom in a fundamental way: they point toward something, rather than merely documenting everything a person has lost.

That distinction has profound clinical implications. If rock bottom is not a motivator, then the widespread advice to “let them hit bottom” — the instinct of families, and sometimes of clinicians, to withhold help until a person’s life collapses completely — has no empirical footing. The authors note that the idea of rock bottom may be stigmatizing, framing people with addiction as people who must be broken before they can be fixed, and that it may hinder timely support for addiction. In a period when opioid use disorder remains a major public health crisis and physician workforce shortages already hamper the treatment response, waiting for an imaginary floor is a luxury that individuals, families, and health systems cannot afford.

The study’s most conceptually ambitious move is its connection to the idea of human flourishing. The four themes of rock bottom — lost connection, eroded character, lost resources, and lost will to live — map with striking precision onto the domains that flourishing researchers, including the Harvard-based Global Flourishing Study, use to describe a life going well: close social relationships, character and virtue, material and financial stability, and a sense of meaning and purpose. Rock bottom, in this framing, is not a mysterious psychological event but the systematic collapse of the very domains that make human life flourish. The researchers suggest that the promotion of human flourishing may therefore lead to sustained recovery — a reframing that shifts the clinical question from “how bad must it get?” to “what does this person need to build a life worth staying sober for?”

The findings on what sustains recovery reinforce this reframing. When the researchers examined the factors critical to maintaining recovery over time, two themes emerged: rebuilding social connection and character growth. Participants who stayed in recovery described reweaving relationships, repairing their sense of who they were, and growing into people they could respect. This aligns with a substantial body of prior research on recovery capital — the internal and external resources a person can draw on to initiate and maintain recovery — which has repeatedly linked social networks and identity reconstruction to long-term outcomes. It also echoes earlier qualitative work showing that people who recover, whether through treatment or on their own, typically describe a shift in identity rather than a single moment of surrender.

None of this means that moments of crisis are irrelevant. Near-death experiences, after all, were among the most common turning points, and a growing literature on “hitting bottom” in alcohol use disorder has attempted to operationalize the construct precisely because people do describe such moments. What the Johns Hopkins study clarifies is the direction of the causal arrow. Crisis alone does not generate change; change is generated when crisis is joined to hope, connection, and a plausible future. A near-death experience motivates when there is something on the other side of survival — a child, a relationship, a self worth becoming. Pure degradation, unaccompanied by any of these, tends only to deepen the spiral.

The study has limitations inherent to its design. Twenty participants, all currently in recovery, cannot represent the full population of people with opioid use disorder, including those who died before reaching recovery or who never entered it. Qualitative thematic analysis, however rigorous, is interpretive by nature. Yet the consistency of the themes, the careful dual-coding process, and the convergence with decades of recovery research give the findings considerable weight. And the practical message is urgent: rather than waiting for people with opioid addiction to lose everything, clinicians, families, and policymakers should be working to strengthen the very things addiction destroys — relationships, character, resources, and hope. Recovery, this research suggests, begins not at the bottom, but at the first glimpse of a flourishing life.

Subject of Research: Rock bottom experiences and turning points in recovery from opioid use disorder and their relationship to human flourishing

Article Title: Rock Bottom Is Not a Motivator: From Opioid Addiction to Human Flourishing

Article References: Rock Bottom Is Not a Motivator: From Opioid Addiction to Human Flourishing. (n.d.). https://doi.org/10.1007/s11606-026-10848-y

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10848-y

Keywords: opioid use disorder, rock bottom, addiction recovery, turning points, human flourishing, qualitative research, motivation, recovery capital, stigma, social connection, Johns Hopkins, Journal of General Internal Medicine

The research, led by Sujal Manohar and Eden Noah Gelgoot with colleagues including Alex Walker, Nora McCarthy and Margaret S. Chisolm, set out to answer a question that has lingered around arts and humanities initiatives in medicine: how exactly do these activities enable learning and reflection? While the value of the arts and humanities in medical education has been increasingly recognized in recent years, with the potential to cultivate important clinical attributes, much of the literature has focused on outcomes rather than process. Less is known about how a processfolio actually functions as a vehicle for learning, and even less about the individual, lived experiences of the learners who create one. The Johns Hopkins team designed their study to close that gap, treating the students not as subjects to be measured but as narrators of their own educational journey.

Methodologically, the study took a deliberately qualitative approach. The researchers employed a cross-sectional design anchored in Interpretative Phenomenological Analysis, or IPA, a methodological framework that aims to uncover how individuals make sense of their personal and social world. IPA is well suited to questions of lived experience because it asks researchers to engage in a double hermeneutic: participants are interpreting their own experiences, and the researchers are then interpreting those interpretations. In this case, the data came from three sources. First, the researchers analyzed the completed processfolios themselves, the summative course artefacts in which students had assembled their work. Second, they examined the students’ final written reflections. Third, they conducted individual semi-structured interviews with each participant, allowing them to elaborate on what the processfolio had meant to them.

The sample was small but remarkably complete. All five fourth-year medical students who finished the museum-based course chose to participate in the research study. That total participation is significant in qualitative research of this kind, because it means the findings capture the full cohort rather than a self-selected fragment of it. It also reflects the voluntary, trust-based nature of the study, which was reviewed by the Johns Hopkins University Institutional Review Board and deemed exempt research under reference number IRB00365270. Because the study was classified as exempt, the institutional review board determined that it did not include informed consent procedures or fall under specific guidelines for research with human subjects. The exemption also signals the educational, non-interventional character of the work: the researchers were studying artefacts and reflections that already existed as part of the course.

From their analysis, the researchers distilled four Group Experiential Themes, the shared patterns of meaning that emerged across the five participants’ accounts. The first theme, contemplating the journey of reflection, growth and insight, speaks to the processfolio’s core function: it makes learning visible over time. Rather than submitting a single polished product at the end of a course, students accumulate artefacts that trace how their thinking evolved week by week. Looking back across that accumulation, participants could see their own intellectual and emotional trajectory in a way that conventional assessments rarely allow. The processfolio became, in effect, a mirror held up to the learning process itself, letting students contemplate not just what they had learned but how they had learned it.

The second theme, exploring personal and emotional growth, points to a dimension of medical training that is often squeezed out by the demands of scientific coursework. Medical students are routinely taught to compartmentalize emotion, yet the capacity to recognize and process one’s own emotional responses is increasingly understood as central to physician wellbeing and patient care. In the museum-based course, engagement with art opened a space where emotional responses were not only permitted but expected to be examined. The processfolio gave students a private, structured repository for that examination, and the act of assembling it encouraged them to connect their experiences in the gallery to their emerging identities as physicians. Participants reflected on their development as medical students and future physicians, a conclusion the authors highlight as central to the study’s purpose.

The third theme is perhaps the most clinically resonant: communicating in clinical contexts, with attention to the importance and complexity of both verbal and non-verbal communication. Art museums are, in a sense, laboratories of observation. Standing before a painting or sculpture, students must describe what they see, tolerate ambiguity, and recognize that others may perceive the same object differently. These are precisely the skills that translate to the bedside, where a physician’s reading of a patient’s posture, expression and silence can matter as much as any laboratory value. The processfolio documented this translation, capturing how students moved from describing artworks to thinking about how they describe patients, and how much of clinical communication happens without words at all.

The fourth theme, acknowledging the complex history of medicine and advocating for better practices, reveals a critical, even activist dimension of the experience. Museums and medicine share entangled histories, and engagement with artistic and historical material prompted students to confront uncomfortable truths about how their profession has treated patients, particularly those from marginalized communities. Rather than treating the arts and humanities as a decorative supplement to clinical training, this theme shows them functioning as a lens for professional critique. Students did not simply absorb the history of medicine; they interrogated it, and the processfolio preserved that interrogation as a record of their evolving ethical commitments. The authors frame this as learners advocating for better practices, a striking outcome for a course that takes place far from any hospital ward.

The study is candid about its limitations and about the fact that the processfolio is not a frictionless tool. Learners shared challenges related to its use, an acknowledgment that matters for educators considering adoption. A processfolio demands sustained reflective labor across a course, and students accustomed to the metrics of biomedical assessment may find the open-ended, self-documenting format unfamiliar or burdensome. The authors also emphasize that the research was conducted at a single institution with a small sample size, which constrains how far the findings can be generalized. Five students at one medical school cannot speak for all learners everywhere, and the authors explicitly call for future research in this area to test and extend what they observed.

Even with those caveats, the study offers a rare, granular look inside an arts and humanities intervention that is often praised in the abstract but seldom examined from the learner’s point of view. By combining analysis of the processfolios and written reflections with in-depth interviews, and by interpreting that material through the lens of Interpretative Phenomenological Analysis, the researchers have shown how a single pedagogical artefact can carry a student through contemplation of their own growth, exploration of their emotional life, refinement of their communicative instincts and confrontation with their profession’s history. The work was published open access on 1 October 2026, received on 11 April and accepted on 10 September of that year, with Manohar and Gelgoot serving as co-first authors. Chisolm’s contribution was supported through her direction of the Paul McHugh Program for Human Flourishing at Johns Hopkins. As medical schools worldwide continue to weave the arts and humanities into their curricula, this study suggests that the humble processfolio, a folder of artefacts and reflections assembled over four weeks in a museum, may be one of the most quietly powerful instruments for shaping the physicians of tomorrow.

Subject of Research: Use of processfolios in arts and humanities-based museum education for reflective learning in medical students

Article Title: Exploring the use of processfolios in a museum-based course for medical students

Article References: Manohar, S., Gelgoot, E. N., Walker, A., McCarthy, N., & Chisolm, M. S. (2026). Exploring the use of processfolios in a museum-based course for medical students. BMC Medical Education. https://doi.org/10.1186/s12909-026-10394-5

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10394-5

Keywords: medical education, arts and humanities, processfolio, museum-based education, reflection, Interpretative Phenomenological Analysis, medical humanities, qualitative research, Johns Hopkins, experiential education, pedagogy, communication skills

The system, called CLASS—short for Clinical LLM Abstraction & Structuring System—is described in a new feasibility report published in the Journal of Medical Systems. Led by anesthesiologist and informatics researcher Frederick H. Kuo, the team built CLASS as a modular, Python-based pipeline that runs entirely within a secure institutional computing environment, keeping protected health information behind the hospital’s own walls rather than sending it to external services. That design choice reflects a growing consensus in medical informatics: the power of large language models can be harnessed for clinical data without compromising patient privacy, provided the infrastructure is configured correctly.

Technically, CLASS rests on three pillars. The first is a set of concept lists curated by subject matter experts—in this case, surgeons and informatics specialists who defined exactly which procedures and clinical details the system should look for. The second is a task-specific prompt suite, a collection of carefully worded instructions that steers the large language model toward consistent, clinically grounded interpretations of each note. The third is a schema-constrained output format, which forces the model to return its findings in a structured, predictable structure rather than free-flowing prose. The results are then exported to an interactive dashboard built for expert review, allowing clinicians to verify outputs, spot errors, and analyze the extracted data at scale.

One of the most innovative features of CLASS is its handling of the unknown. Rather than simply classifying notes against a fixed list of predefined concepts, the pipeline actively flags potential variants or entirely novel concepts that do not fit the existing schema, surfacing them for expert consideration. In specialized fields where terminology evolves quickly and procedures are often described in non-standard ways, this ability to propose expansions to the concept vocabulary could fundamentally change how clinical registries and research databases are built and maintained.

To test the system, the researchers applied CLASS to a retrospective corpus of pediatric esophageal airway treatment surgery, or EATS, operative notes at their single center. EATS is a demanding test case: these operative notes describe complex, highly individualized procedures in children with airway and esophageal abnormalities, and much of the procedural detail is not captured in standard billing codes. The team compared CLASS outputs against adjudication by an experienced surgeon on the twenty longest notes in the corpus, yielding 3,960 individual note-procedure pairs for evaluation.

The results were striking. Across those thousands of judgments, the observed concordance between the automated pipeline and the surgeon’s adjudication reached an F1 score of 0.9967—a near-perfect measure of precision and recall combined. In practical terms, the model almost never missed a procedure the surgeon identified, and almost never invented one that was not there. For a task as subtle as parsing operative prose about pediatric airway surgery, that level of agreement suggests that large language models, when properly constrained and prompted, can match expert-level abstraction performance in at least some specialized clinical domains.

The novelty-detection results were more nuanced, and arguably more interesting. CLASS proposed 28 candidate procedure variants or additions to the curated concept list, and the clinical team judged 18 of them—64.3 percent—to be genuinely useful. That is a meaningful yield: nearly two out of every three suggestions from the machine were worth a clinician’s time. At the same time, the surgeon identified 12 additional procedures that CLASS failed to surface, a reminder that the system works best as a collaborator rather than a replacement. The human expert still caught things the machine missed, and the machine still surfaced things the human might not have thought to codify.

The authors are careful to frame the study appropriately. This was an exploratory implementation at a single center, focused on a single surgical service, with evaluation limited to a small set of long notes. Generalizability to other institutions, other note types, and other clinical tasks remains unproven, and the team emphasizes that broader validation is needed before such pipelines could be trusted for high-stakes applications. The full production code and clinical data cannot be released publicly because they involve protected health information and institution-specific infrastructure, but the researchers have shared technical implementation details, template code, pseudocode, and de-identified prompt examples in the online supplementary materials, giving other informatics teams a practical roadmap for building similar systems.

Even with those caveats, the implications are considerable. Clinical research has long been throttled by the bottleneck of manual abstraction: cohort studies that could enroll thousands of patients are often limited to hundreds simply because there are only so many hours in a research coordinator’s day. Quality-improvement programs face the same constraint, unable to measure surgical outcomes comprehensively when the relevant data must be pulled by hand from narrative notes. If pipelines like CLASS can reliably convert unstructured text into analyzable data within standard institutional infrastructure, the effective sample sizes of clinical research could expand dramatically, and hospitals could monitor the quality of specialized care in near real time.

The study also highlights a shift in how medical informatics teams may work in the coming years. Instead of writing brittle rule-based extraction algorithms or training bespoke machine-learning models on small labeled datasets, teams can now curate expert concept lists, design prompts, and review machine-generated suggestions—a workflow in which clinicians define what matters and the model handles the linguistic heavy lifting. The CLASS experience suggests this human-machine partnership can work: the model performs the extraction with near-perfect fidelity, proposes useful vocabulary expansions, and leaves final judgment to the experts who bear clinical responsibility. As large language models continue to demonstrate their grasp of medical language, studies like this one offer a concrete, privacy-conscious template for turning the narrative richness of the medical record into structured knowledge—without a single chart being pulled by hand.

Subject of Research: A large language model pipeline for extracting structured data from unstructured clinical notes

Article Title: Exploratory Implementation and Feasibility Report of CLASS (Clinical LLM Abstraction & Structuring System), A Large Language Model Pipeline for Extracting Unstructured Data From Clinical Notes

Article References: Exploratory Implementation and Feasibility Report of CLASS (Clinical LLM Abstraction & Structuring System), A Large Language Model Pipeline for Extracting Unstructured Data From Clinical Notes. (n.d.). https://doi.org/10.1007/s10916-026-02462-6

Image Credits: AI Generated

DOI: 10.1007/s10916-026-02462-6

Keywords: large language models, clinical notes, natural language processing, electronic health records, medical informatics, data extraction, pediatric surgery, artificial intelligence, CLASS pipeline, clinical research, esophageal airway treatment surgery, Johns Hopkins