The study, known by the acronym MAESTRI, was conducted by a team of researchers spanning several German institutions, including the Institute for Digital Medicine at Philipps-Universität Marburg, the Department of Rheumatology and Clinical Immunology at Justus-Liebig University Giessen, and collaborating centers in dermatology and orthopedics. Led by first authors Anna Roemer and Tim Wilhelmi, with Philipp Klemm and Johannes Knitza as senior authors, the trial was designed as a multicenter, parallel-group, randomized pilot study. It recruited adults who were considered to be at increased risk of an underlying inflammatory rheumatic disease and who were presenting to outpatient services in three specialties: rheumatology, dermatology, and orthopedics. Participants were randomly assigned either to complete a pre-visit computer-assisted history-taking questionnaire or to receive usual care.

The technical architecture of the intervention is what distinguishes it from the countless patient questionnaires that have come and gone in digital health. Specialty-specific questionnaires were designed to capture a predefined set of screening-relevant items, the specific pieces of information a clinician would need to judge whether an inflammatory rheumatic disease might be present. Crucially, the software did not simply store the answers in a separate silo. Instead, it generated a structured narrative summary that was formatted to be copied directly into the clinical documentation. In the intervention group, every single clinical note contained this generated summary, meaning the tool was not merely available but was actually integrated into the workflow of writing notes. That integration step, often the weakest link in digital health interventions, appears to have been decisive.

The primary outcome of the trial was documentation completeness, assessed by a blinded reviewer as the percentage of predefined screening-relevant items that made it into the final clinical documentation. The results were striking. The median completeness score in the intervention group was 100.0 percent, with an interquartile range of 100.0 to 100.0, meaning that essentially every patient in that arm had every relevant item documented. In the control group, the median was 46.9 percent, with an interquartile range stretching from 10.0 to 68.8 percent, a spread that illustrates how inconsistent conventional history-taking can be when time is short and demands are high. The difference between the groups was statistically significant, with a p-value below 0.001, and the pattern held consistently across all three participating specialties.

Those numbers deserve a moment of unpacking, because documentation completeness is not an abstract quality metric. When a physician documents the presence or absence of key symptoms, that record becomes the foundation for subsequent clinical decisions, referrals, insurance determinations, and follow-up care. A record that captures fewer than half of the screening-relevant items is a record that forces the next clinician to start from scratch, or worse, one that quietly omits a clue that would have triggered earlier diagnostic workup. In screening contexts, where the entire purpose of the encounter is to determine whether a disease process warrants further investigation, the gap between 47 percent and 100 percent completeness is the gap between a sieve and a net.

Equally important for the future of this approach is the question of acceptance, because even the most elegant software fails if patients refuse to use it or physicians resent the extra step. Here the pilot study delivered encouraging news. Of the 43 randomized and analyzed patients, 41, or 95.3 percent, said they were willing to complete the pre-visit questionnaire. Among the six physicians who responded, all six, a full 100 percent, said they would be willing to offer digital pre-visit questionnaires before future appointments. In a health technology landscape littered with tools that clinicians quietly abandon, a unanimous willingness to continue is a notable signal, even in a small sample.

The qualitative feedback collected in the study added texture to those numbers and revealed why the tool resonated. Patients reported that the pre-visit questionnaires could improve preparation for the consultation, support their recall of symptoms, and help physicians focus the visit on the issues that mattered most. Anyone who has sat in an examination room trying to remember when the morning stiffness started or which joints have been affected will recognize the value of being prompted to think through those details in advance, at home, without the pressure of a ticking clock. The structured questionnaire effectively moves part of the cognitive burden of history-taking out of the consultation and into a moment when the patient has time to reflect.

At the same time, the study captured an important caution from participants. Some emphasized that digital questionnaires should complement, rather than replace, direct physician interaction. That distinction goes to the heart of what computer-assisted history-taking is and is not. The tool in this trial did not diagnose, did not triage patients away from clinicians, and did not substitute an algorithm’s judgment for a physician’s examination. It collected structured information, formatted it for the record, and freed the human encounter to be more focused. The patients in the study seemed to understand that distinction, embracing the efficiency while defending the irreplaceable value of talking with their doctor. Any future deployment would be wise to preserve that boundary.

The trial’s scale demands honest caveats. This was a pilot study, with 45 patients screened and 43 randomized and analyzed, and a mean age of 54.2 years with about two-thirds of participants female. The physicians surveyed numbered only six. Small samples can overestimate effects, and the near-perfect completeness scores in the intervention group reflect a tool specifically engineered to guarantee that the predefined items appear in the note. The authors themselves frame the findings as supporting further evaluation in a larger trial that would assess both workflow impact and clinical impact, meaning whether better documentation actually translates into earlier diagnosis and improved outcomes for patients with inflammatory rheumatic disease. The study was registered prospectively at the German Clinical Trials Register under identifier DRKS00036519, and it received ethics approval from the Ethics Committee of Philipps University Marburg.

Even with those limitations, the MAESTRI pilot offers a compelling glimpse of where outpatient medicine may be heading. The intervention required no exotic technology, no artificial intelligence making diagnostic decisions, and no fundamental restructuring of the clinic. It asked patients to answer structured questions before their visit, converted those answers into a copy-ready narrative summary, and ensured the summary landed in the clinical note. The result was a doubling or more of documentation completeness, near-universal patient acceptance, and unanimous physician buy-in across three different specialties. As health systems worldwide grapple with aging populations, specialist shortages, and the mounting administrative burden that consumes clinician time, the lesson from this German trial is that some of the most impactful digital health tools may be the ones that work quietly in the background, before the consultation even begins, making sure that nothing a patient knows about their own body gets lost between the waiting room and the medical record.

Subject of Research: Pre-visit computer-assisted history-taking for documentation completeness in screening for inflammatory rheumatic diseases

Article Title: Effect of pre-visit computer-assisted history-taking with a structured copy-ready summary on documentation completeness in screening for inflammatory rheumatic diseases: a randomized pilot study

Article References: Roemer, A., Wilhelmi, T., Fink, D., Göbel, M., Hack, J., Bamberger, J., Mühlensiepen, F., Gente, K., Kuhn, S., Klemm, P., & Knitza, J. (2026). Effect of pre-visit computer-assisted history-taking with a structured copy-ready summary on documentation completeness in screening for inflammatory rheumatic diseases: a randomized pilot study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15808-0

Image Credits: AI Generated

DOI: 10.1186/s12913-026-15808-0

Keywords: computer-assisted history-taking, documentation completeness, inflammatory rheumatic diseases, rheumatology, screening, digital health, randomized pilot study, pre-visit questionnaires, clinical documentation, patient acceptance, health services research, outpatient care

The rationale for the trial rests on a well-documented clinical problem. Infants with NOWS tend to lose more weight after birth than other newborns, and earlier work by the same research group mapped early weight-loss percentile curves in opioid-exposed infants, showing that excessive weight loss is common and can trigger additional medical interventions. Because withdrawal disrupts feeding behavior, some clinicians have wondered whether high-calorie formula, which delivers more energy per milliliter than standard formula, could help these babies maintain their weight and possibly recover faster. But the idea has never been firmly established, and concerns linger about whether concentrated feeds might cause gastrointestinal upset, alter stooling patterns, or interfere with breastfeeding, which multiple studies have linked to milder withdrawal courses and shorter treatment.

To test whether such a trial is even possible, the researchers designed a double-blind randomized pilot study conducted in Connecticut and registered as NCT04419857. Between December 2020 and August 2023, they enrolled infants less than 48 hours old who had been prenatally exposed to methadone or buprenorphine, the two medications most commonly used to treat opioid use disorder in pregnancy. The babies were at risk for excessive weight loss and were being cared for under the Eat, Sleep, Console model, a function-based approach to assessing and treating NOWS that has rapidly replaced older scoring systems in many hospitals since a major randomized trial in the New England Journal of Medicine in 2023 showed it shortened hospital stays and reduced the need for medication.

Eligible infants were randomized to receive either standard-calorie or high-calorie feeds, with the feeding approach kept blinded to families and clinical staff wherever possible. The primary outcomes were deliberately pragmatic, reflecting the purpose of a pilot study: the investigators wanted to know whether they could enroll enough families, whether the randomized feeding assignments were actually followed at the bedside, and whether families stayed in the study. Adherence was measured as the proportion of feeds that followed the assigned approach, and retention was tracked both during the birth hospitalization and for two weeks after discharge, a window in which follow-up of this population is notoriously difficult.

The feasibility results were mixed in ways that are informative for anyone planning a definitive trial. Of 65 eligible infants, 35, or 53.8 percent, were randomized, a rate the authors judged comparable to other feasibility trials in similar settings. Once enrolled, families and staff largely stuck to the protocol: 84 percent of all feeds during hospitalization adhered to the randomized feeding approach, suggesting that a high-calorie feeding strategy can be implemented reliably in a real-world nursery environment. In-hospital retention reached 71.4 percent, but the picture changed sharply after discharge, with only 28.6 percent of families retained at the 14-day follow-up point. That attrition, the authors note, mirrors broader literature on post-discharge follow-up among families affected by opioid use disorder and represents the single biggest obstacle to a full-scale efficacy trial.

On the exploratory clinical outcomes, the two groups looked remarkably similar. There were no significant differences between the standard-calorie and high-calorie groups in demographic characteristics, in the maximum percentage of weight loss from birth, in the length of hospital stay, or in the frequency of stooling. The weight-loss finding, with a p-value of 0.12, and the length-of-stay finding, with a p-value of 0.88, do not rule out effects in either direction, since a pilot study of 35 infants is not powered to detect differences; rather, they establish that neither outcome showed an obvious signal of harm or benefit that would demand immediate changes to practice. The stooling comparison, which addressed concerns that concentrated feeds might slow the gut, was similarly unremarkable at p equal to 0.59.

One exploratory result stood out. Infants in the high-calorie group were more frequently receiving breastmilk at the time of discharge, a difference that reached statistical significance at p equal to 0.04. The finding runs counter to a plausible worry: that introducing fortified or high-calorie formula early might displace breastfeeding and reduce a mother’s milk supply or her confidence in nursing. Prior observational studies, including work showing that breastmilk feeding is associated with shorter treatment courses in NOWS, and clinical protocols from the Academy of Breastfeeding Medicine supporting breastfeeding in the setting of stable maternal opioid agonist treatment, have made the preservation of breastfeeding a central goal of NOWS care. The pilot’s data, while exploratory and drawn from a small sample, suggest that early high-calorie supplementation did not appear to be negatively associated with breastfeeding in this cohort.

The study also contributes to a small but growing evidence base on formula composition for opioid-exposed newborns. An earlier randomized feasibility study by Bogen and colleagues tested standard versus high-calorie formula in methadone-exposed infants, and other trials have examined energy-enriched formula in infants recovering from congenital heart surgery, another population at high risk of growth failure. Meanwhile, a systematic review of milk feed osmolality has raised general questions about the tolerability of concentrated feeds in newborns, making controlled data in the NOWS population particularly valuable. The Yale-led trial adds the important element of randomization under the modern Eat, Sleep, Console care model, which changes the context in which feeding decisions influence pharmacologic treatment and length of stay.

The authors are careful about interpretation, and appropriately so. As a pilot study, its purpose was not to prove that high-calorie feeding works or fails, but to determine whether a definitive randomized trial is achievable and to identify the design features it would need. The enrollment and adherence figures suggest the answer is a qualified yes within the hospital, while the 28.6 percent post-discharge retention signals that any future trial should invest heavily in retention strategies, an area with an established methodological literature ranging from participant-centric contact approaches to systematic reviews of retention interventions in pediatric trials. The statistical framework also had to contend with missing data, a challenge the team addressed using established methods for inference under missingness.

For clinicians caring for the rising number of infants born with opioid exposure, the practical message is one of cautious reassurance rather than a change in practice. High-calorie feeds were delivered as intended in more than four out of five feeds, did not visibly worsen weight loss trajectories, stooling, or hospital stay, and were not associated with reduced breastfeeding at discharge. For researchers, the study maps the terrain for the next step: a larger, adequately powered randomized trial with robust post-discharge follow-up to determine whether enriched nutrition can meaningfully reduce excessive weight loss in NOWS without compromising the human-milk feeding that so much evidence suggests benefits these infants. Until such a trial is completed, feeding decisions for newborns in withdrawal will continue to rest on clinical judgment, family preference, and the growing but still incomplete evidence base this pilot study has helped to build.

Subject of Research: Feasibility of high-calorie versus standard feeding in newborns with Neonatal Opioid Withdrawal Syndrome managed with Eat, Sleep, Console

Article Title: Randomized pilot study comparing standard vs high-calorie feeding in infants with Neonatal Opioid Withdrawal Syndrome

Article References: Randomized pilot study comparing standard vs high-calorie feeding in infants with Neonatal Opioid Withdrawal Syndrome. (n.d.). https://doi.org/10.1038/s41372-026-02913-5

Image Credits: AI Generated

DOI: 10.1038/s41372-026-02913-5

Keywords: Neonatal Opioid Withdrawal Syndrome, high-calorie formula, Eat Sleep Console, randomized pilot study, breastfeeding, neonatal weight loss, methadone exposure, buprenorphine exposure, neonatal nutrition, Journal of Perinatology, clinical trial feasibility, opioid use disorder in pregnancy