JKJohn Kim

John Kim

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


John Kim has a total of 86 co-authors. Here are the top co-authors by number of shared publications:

  • Elizabeth W. Tucker

    Center for Infection and Inflammation Imaging Research, Johns Hopkins University School of Medicine, Baltimore, MD 21287, USA

    5 shared publications
  • Sanjay K. Jain

    Center for Infection and Inflammation Imaging Research, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA

    5 shared publications
  • Alvaro Andres Ordoñez

    Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, USA

    4 shared publications
  • Kelly Flavahan

    Center for Infection and Inflammation Imaging Research, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA; Center for Tuberculosis Research, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA; Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA

    4 shared publications
  • Marc D. Succi

    Harvard Medical School, Boston, USA

    4 shared publications
  • Melissa Bahr

    Center for Infection and Inflammation Imaging Research, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA; Center for Tuberculosis Research, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA; Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA

    4 shared publications

Latest Publications

  • Point‐of‐Care Ultrasound in Hospital Medicine

    by John Kim, David Lee, Michael Schwartz, Tanping Wong, Kevin Piro, James C. Hudspeth, Lao‐Tzu Allan‐Blitz

    Point‐of‐care ultrasound (POCUS) is increasingly used in hospital medicine to support bedside diagnosis, management, and procedural safety. This narrative review aims to synthesize current evidence on POCUS applications in pneumonia, heart failure, deep vein thrombosis, and bedside procedures. While POCUS may be clinically useful for such use cases, its broader adoption is limited by training gaps, infrastructure constraints, and credentialing challenges. Standardizing curricula and institutional support are critical next steps to optimize POCUS adoption and impact. Broader implementation of POCUS may enhance diagnostic accuracy and care efficiency in inpatient settings.


  • Open Access

    Medicare spending and use of subcutaneous biologic formulations with hyaluronidase

    by John Kim, Aaron S Kesselheim, Edward R Scheffer Cliff, Benjamin N Rome

    Several intravenously administered biologic drugs have been reformulated with hyaluronidase to enable subcutaneous delivery. This offers greater convenience and fewer infusion reactions while also increasing overall spending if biosimilar competition for the subcutaneous versions begins later than for the original intravenous versions. As of December 2024, at least 9 biologics had investigational or approved hyaluronidase versions. Medicare spending on these drugs totaled $10.3B in 2022. For 4 of these drugs, hyaluronidase versions accounted for 5%-83% of Medicare spending in 2022, with hyaluronidase versions accounting for the highest share of spending for pertuzumab-trastuzumab and daratumumab and a lower share of spending for 2 drugs that had biosimilar competition for the original versions: rituximab and trastuzumab. The benefits of subcutaneous hyaluronidase versions must be balanced against the challenges that come with higher prices if these versions are introduced before biosimilar competition begins for the original versions. Policymakers should ensure manufacturers cannot use “hyaluronidase hopping” to delay biosimilar competition or eligibility for Medicare price negotiation.


  • Open Access

    Feasibility of a Non‐pharmacological Intervention for Impaired Financial Capacity in Mild Cognitive Impairment

    by Milap A. Nowrangi, John D. Outen, Paul B. Rosenberg, John Kim

    Background: A pilot trial of financial literacy training (FL) and computerized cognitive training (CCT) was conducted to assess tolerability, feasibility, and the effect on financial capacity (FC) in participants with normal cognition and MCI.Methods: Cognitively normal and MCI participants were randomized to receive either targeted FL content (N = 4) (Savvy Saving Seniors (“SSS”) workshop) narrated and presented online or the active control CCT (N = 3) (BrainHQ, ©Posit Science) over 6‐weeks. “SSS” participants were instructed to complete 25‐minute sessions twice per week and BrainHQ participants were instructed to complete 60‐minute sessions three times per week.Results: No significant differences were found between baseline and follow‐up in either group. Table 1. Two out of 3 BrainHQ participants reported that their processing speed had improved and they felt challenged and experienced positive cognitive benefits. Three out of 4 “SSS” participants reported that they felt stimulated and challenged and that the activities were a good use of their time. All “SSS” participants agreed their financial literacy had improved. The most common reported benefits were accomplishment, enjoyment, and increased alertness. The most common reported discomforts were frustration, boredom and one participant complained of eye strain.Discussion/Conclusions: We demonstrated good efficacy, tolerability, and adherence in delivering FL education and CCT interventions in cognitively normal and MCI participants.

  • Open Access

    Linezolid does not improve bactericidal activity of rifampin-containing first-line regimens in animal models of TB meningitis

    by Elizabeth W. Tucker, Camilo A. Ruiz-Bedoya, Filipa Mota, Clara Erice, John Kim, Patricia de Jesus, Ravindra Jahdav, Melissa Bahr, Kelly Flavahan, Xueyi Chen, et al.


  • Open Access

    2663. Comprehensive at-home Sexually Transmitted and Blood Borne Infection (STBBI) Testing Program: A Pilot Initiative

    by Brandon L Christensen, Caley B Shukalek, Grace (Juyoung) Kwon, Nolan E Hill, Deirdre L Church, Christopher T Naugler, John Kim, Ranjani Somayaji

    Sexually transmitted and blood borne infection (STBBIs) remain a growing concern in North America with rates increasing dramatically since 2000. Although self-testing has been demonstrated to be valid and acceptable, no fully remote STBBI testing programs are available in Canada. To evaluate feasibility and acceptability of at-home STBBI testing, we implemented a fully remote, web-based platform of risk-responsive, non-invasive STBBI diagnostics in a large Canadian health zone. A web-based risk assessment and intake form was developed using RedCap software through a community-facing website (athomestitesting.ca) in partnership with a local sexual health centre. Demographic, sexual practices, and medical history were collected from interested participants through the platform. Diagnostic kits inclusive of testing for gonorrhea/chlamydia (urine and self-swabs of vagina/pharyngeal/rectal) and syphilis/HIV/Hepatitis C (dried blood spot) were mailed to participants based on their survey responses, along with instructions and supplies for collection and return. Outcomes to be assessed include feasibility and acceptability of the web-based platform, testing and positivity rates for STIs, and time to testing and treatment (for positive cases). In the first month to date with limited launch, eleven participants completed the intake survey and were mailed testing kits through the platform. Of the completed testing, no new STIs were identified. The median age of the cohort was 38 years old (IQR 32, 41) and six identified as men, four as women and one non-binary. Regarding gender identity, seven identified as gay/queer/bisexual and three as straight/heterosexual. In a follow-up survey, all participants to date stated reported satisfaction or extremely satisfaction with the online testing. The platform model demonstrates feasibility in the early pilot phase with acceptability based on participant responses. Next steps include expanded launch of the platform and assessment of testing rates and feasibility of a web-based model for STI diagnostics. Caley B. Shukalek, MD, MSc, MPH, Gilead Sciences: Advisor/Consultant|Gilead Sciences: Grant/Research Support|Gilead Sciences: Honoraria|Merck Canada: Grant/Research Support|Merck Canada: Honoraria|PurposeMed Inc: Advisor/Consultant|PurposeMed Inc: Ownership Interest Ranjani Somayaji, M.D., CF Foundation: Grant/Research Support|CIHR: Grant/Research Support|Oncovir: Data and Safety Monitoring Board Fees|Vertex Pharmaceuticals: Advisor/Consultant|Vertex Pharmaceuticals: Grant/Research Support|Vertex Pharmaceuticals: Honoraria

  • Top Cited Papers
    Open Access

    Assessing the Utility of ChatGPT Throughout the Entire Clinical Workflow: Development and Usability Study

    by Arya Rao, Michael Pang, John Kim, Meghana Kamineni, Winston Lie, Anoop K Prasad, Adam Landman, Keith Dreyer, Marc D Succi

    Background: Large language model (LLM)–based artificial intelligence chatbots direct the power of large training data sets toward successive, related tasks as opposed to single-ask tasks, for which artificial intelligence already achieves impressive performance. The capacity of LLMs to assist in the full scope of iterative clinical reasoning via successive prompting, in effect acting as artificial physicians, has not yet been evaluated. Objective: This study aimed to evaluate ChatGPT’s capacity for ongoing clinical decision support via its performance on standardized clinical vignettes. Methods: We inputted all 36 published clinical vignettes from the Merck Sharpe & Dohme (MSD) Clinical Manual into ChatGPT and compared its accuracy on differential diagnoses, diagnostic testing, final diagnosis, and management based on patient age, gender, and case acuity. Accuracy was measured by the proportion of correct responses to the questions posed within the clinical vignettes tested, as calculated by human scorers. We further conducted linear regression to assess the contributing factors toward ChatGPT’s performance on clinical tasks. Results: ChatGPT achieved an overall accuracy of 71.7% (95% CI 69.3%-74.1%) across all 36 clinical vignettes. The LLM demonstrated the highest performance in making a final diagnosis with an accuracy of 76.9% (95% CI 67.8%-86.1%) and the lowest performance in generating an initial differential diagnosis with an accuracy of 60.3% (95% CI 54.2%-66.6%). Compared to answering questions about general medical knowledge, ChatGPT demonstrated inferior performance on differential diagnosis (β=–15.8%; P<.001) and clinical management (β=–7.4%; P=.02) question types. Conclusions: ChatGPT achieves impressive accuracy in clinical decision-making, with increasing strength as it gains more clinical information at its disposal. In particular, ChatGPT demonstrates the greatest accuracy in tasks of final diagnosis as compared to initial diagnosis. Limitations include possible model hallucinations and the unclear composition of ChatGPT’s training data set.


  • Top Cited Papers

    Evaluating GPT as an Adjunct for Radiologic Decision Making: GPT-4 Versus GPT-3.5 in a Breast Imaging Pilot

    by Arya Rao, John Kim, Meghana Kamineni, Michael Pang, Winston Lie, Keith J. Dreyer, Marc D. Succi


  • Open Access

    Assessing the Utility of ChatGPT Throughout the Entire Clinical Workflow: Development and Usability Study (Preprint)

    by Arya Rao, Michael Pang, John Kim, Meghana Kamineni, Winston Lie, Anoop K Prasad, Adam Landman, Keith Dreyer, Marc D Succi

    BACKGROUND Large language model (LLM)–based artificial intelligence chatbots direct the power of large training data sets toward successive, related tasks as opposed to single-ask tasks, for which artificial intelligence already achieves impressive performance. The capacity of LLMs to assist in the full scope of iterative clinical reasoning via successive prompting, in effect acting as artificial physicians, has not yet been evaluated. OBJECTIVE This study aimed to evaluate ChatGPT’s capacity for ongoing clinical decision support via its performance on standardized clinical vignettes. METHODS We inputted all 36 published clinical vignettes from the Merck Sharpe & Dohme (MSD) Clinical Manual into ChatGPT and compared its accuracy on differential diagnoses, diagnostic testing, final diagnosis, and management based on patient age, gender, and case acuity. Accuracy was measured by the proportion of correct responses to the questions posed within the clinical vignettes tested, as calculated by human scorers. We further conducted linear regression to assess the contributing factors toward ChatGPT’s performance on clinical tasks. RESULTS ChatGPT achieved an overall accuracy of 71.7% (95% CI 69.3%-74.1%) across all 36 clinical vignettes. The LLM demonstrated the highest performance in making a final diagnosis with an accuracy of 76.9% (95% CI 67.8%-86.1%) and the lowest performance in generating an initial differential diagnosis with an accuracy of 60.3% (95% CI 54.2%-66.6%). Compared to answering questions about general medical knowledge, ChatGPT demonstrated inferior performance on differential diagnosis (β=–15.8%; P<.001) and clinical management (β=–7.4%; P=.02) question types. CONCLUSIONS ChatGPT achieves impressive accuracy in clinical decision-making, with increasing strength as it gains more clinical information at its disposal. In particular, ChatGPT demonstrates the greatest accuracy in tasks of final diagnosis as compared to initial diagnosis. Limitations include possible model hallucinations and the unclear composition of ChatGPT’s training data set.


  • Open Access

    Dynamic 18F-Pretomanid PET imaging in animal models of TB meningitis and human studies

    by Filipa Mota, Camilo A. Ruiz-Bedoya, Elizabeth W. Tucker, Daniel P. Holt, Patricia De Jesus, Martin A. Lodge, Clara Erice, Xueyi Chen, Melissa Bahr, Kelly Flavahan, et al.

    Pretomanid is a nitroimidazole antimicrobial active against drug-resistant Mycobacterium tuberculosis and approved in combination with bedaquiline and linezolid (BPaL) to treat multidrug-resistant (MDR) pulmonary tuberculosis (TB). However, the penetration of these antibiotics into the central nervous system (CNS), and the efficacy of the BPaL regimen for TB meningitis, are not well established. Importantly, there is a lack of efficacious treatments for TB meningitis due to MDR strains, resulting in high mortality. We have developed new methods to synthesize 18F-pretomanid (chemically identical to the antibiotic) and performed cross-species positron emission tomography (PET) imaging to noninvasively measure pretomanid concentration-time profiles. Dynamic PET in mouse and rabbit models of TB meningitis demonstrates excellent CNS penetration of pretomanid but cerebrospinal fluid (CSF) levels does not correlate with those in the brain parenchyma. The bactericidal activity of the BPaL regimen in the mouse model of TB meningitis is substantially inferior to the standard TB regimen, likely due to restricted penetration of bedaquiline and linezolid into the brain parenchyma. Finally, first-in-human dynamic 18F-pretomanid PET in six healthy volunteers demonstrates excellent CNS penetration of pretomanid, with significantly higher levels in the brain parenchyma than in CSF. These data have important implications for developing new antibiotic treatments for TB meningitis.