Model-Informed Precision Dosing Using Machine Learning for Levothyroxine in General Practice: Development, Validation and Clinical Simulation Trial.

Levothyroxine is one of the most prescribed drugs in the western world. Dosing is challenging due to high-interindividual differences in effective dosage and the narrow therapeutic window. Model-informed precision dosing (MIPD) using machine learning could assist general practitioners (GPs), but no such models exist for primary care. Furthermore, introduction of decision-support algorithms in healthcare is limited due to the substantial gap between developers and clinicians' perspectives. We report the development, validation, and a clinical simulation trial of the first MIPD application for primary care. Stable maintenance dosage of levothyroxine was the model target. The multiclass model generates predictions for individual patients, for different dosing classes. Random forest was trained and tested on a national primary care database (n = 19,004) with a final weighted AUC across dosing options of 0.71, even in subclinical hypothyroidism. TSH, fT4, weight, and age were most predictive. To assess the safety, feasibility, and clinical impact of MIPD for levothyroxine, we performed clinical simulation studies in GPs and compared MIPD to traditional prescription. Fifty-one GPs selected starting dosages for 20 primary hypothyroidism cases without and then with MIPD 2 weeks later. Overdosage and underdosage were defined as higher and lower than 12.5 µg relative to stable maintenance dosage. MIPD decreased overdosage in number (30.5 to 23.9%, P < 0.01) and magnitude (median 50 to 37.5 µg, P < 0.01) and increased optimal starting dosages (18.3 to 30.2%, P < 0.01). GPs considered lab results more often with MIPD and most would use the model frequently. This study demonstrates the clinical relevance, safety, and effectiveness of MIPD for levothyroxine in primary care.

Trends in low-value GP care during the COVID-19 pandemic: a retrospective cohort study.

BACKGROUND: Several studies showed that during the pandemic patients have refrained from visiting their general practitioner (GP). This resulted in medical care being delayed, postponed or completely forgone. The provision of low-value care, i.e. care which offers no net benefit for the patient, also could have been affected. We therefore assessed the impact of the COVID-19 restrictions on three types of low-value GP care: 1) imaging for back or knee problems, 2) antibiotics for otitis media acuta (OMA), and 3) repeated opioid prescriptions, without a prior GP visit. METHODS: We performed a retrospective cohort study using registration data from GPs part of an academic GP network over the period 2017-2022. The COVID-19 period was defined as the period between April 2020 to December 2021. The periods before (January 2017 to April 2020) and after the COVID-19 period (January 2022 to December 2022) are the pre- and post-restrictions periods. The three clinical practices examined were selected by two practicing GPs from a top 30 of recommendations originating from the Dutch GP guidelines, based on their perceived prevalence and relevance in practice (van Dulmen et al., BMC Primary Care 23:141, 2022). Multilevel Poisson regression models were built to examine changes in the incidence rates (IR) of both registered episodes and episodes receiving low-value treatment. RESULTS: During the COVID-19 restrictions period, the IRs of episodes of all three types of GP care decreased significantly. The IR of episodes of back or knee pain decreased by 12%, OMA episodes by 54% and opioid prescription rate by 13%. Only the IR of OMA episodes remained significantly lower (22%) during the post-restrictions period. The provision of low-value care also changed. The IR of imaging for back or knee pain and low-value prescription of antibiotics for OMA both decreased significantly during the COVID-restrictions period (by 21% and 78%), but only the low-value prescription rate of antibiotics for OMA remained significantly lower (by 63%) during the post-restrictions period. The IR of inappropriately repeated opioid prescriptions remained unchanged over all three periods. CONCLUSIONS: This study shows that both the rate of episodes as well as the rate at which low-value care was provided have generally been affected by the COVID-19 restrictions. Furthermore, it shows that the magnitude of the impact of the restrictions varies depending on the type of low-value care. This indicates that deimplementation of low-value care requires tailored (multiple) interventions and may not be achieved through a single disruption or intervention alone.

Are specialist-provided end-of-life scenarios key to initiation of advance care planning in primary care? A mixed-methods study.

OBJECTIVES: Specialist-provided end-of-life scenarios (SP-EOLS) may improve advance care planning (ACP) implementation in primary care by helping overcome barriers such as uncertain prognosis and poor interprofessional collaboration. We aimed to explore the current use and potential impact of SP-EOLS on ACP in Dutch primary care. METHODS: We performed a mixed-methods study. From patients discussed in a hospital-based academic palliative care multidisciplinary team meeting between 2016 and 2019 and died, we collected primary care electronic medical records data on SP-EOLS, actual EOLS, and ACP initiation and applied descriptive and comparative analyses. Subsequently, we interviewed general practitioners (GPs) and thematically analyzed the transcripts. RESULTS: In 69.7% of 66 reviewed patient files, SP-EOLS were found. In patients whose GP had received SP-EOLS, ACP conversations were more often reported (92.0 vs. 61.0%, p = 0.006). From 11 GP interviews, we identified 4 themes: (1) SP-EOLS guide GPs, patients, and relatives when dealing with an uncertain future perspective; (2) SP-EOLS provide continuity of care between primary and secondary/tertiary care; (3) SP-EOLS should be tailored to the individual patient; and (4) SP-EOLS need to be personalized and uniformly transferred to GPs. SIGNIFICANCE OF RESULTS: SP-EOLS may facilitate ACP conversations by GPs. They have the potential to help overcome existing barriers to ACP implementation by providing guidance and supporting interprofessional collaboration. Future research should focus on improving SP-EOLS and tailor them to the needs of all end users, focusing on improving their effect on ACP conversations.