pith. sign in

arxiv: 2606.09952 · v1 · pith:NLRPOFVNnew · submitted 2026-06-08 · 🧬 q-bio.QM · physics.med-ph

Adjusted trajectory of medication exposure taking into account the periodicity of dispensations and the number of dispensed packs and comparative analysis on EFEMERIS database

C\'ecile Chouquet , Anna-Belle Beau , Christine Damase-Michel , David Jeauneau , Isabelle Lacroix , Sabine Mercier This is my paper

Pith reviewed 2026-06-27 14:28 UTC · model grok-4.3

classification 🧬 q-bio.QM physics.med-ph
keywords medication exposure trajectoriesdispensed packsdispensation periodicityEFEMERIS databasedefined daily doseclustering analysispsychotropic medicationsneonatal outcomes
0
0 comments X

The pith

Adjusting medication exposure trajectories for dispensed pack counts and dispensation periodicity changes profiles for 35 percent of cases and improves clustering quality on EFEMERIS data.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces an adjustment method for calculating medication exposure trajectories that incorporates the number of dispensed packs and whether dispensations occur occasionally or regularly. On the EFEMERIS database, three calculation scenarios are compared, showing that 65 percent of trajectories remain unchanged while the remaining cases exhibit shifts in defined daily dose totals or exposure profiles. Four percent of trajectories switch clusters under the adjusted method, which also produces higher clustering quality. This adjustment alters the distribution of some maternal characteristics and neonatal outcomes across clusters, including a stronger association between high psychotropic doses and neonatal pathology.

Core claim

The adjustment method, by factoring in the number of dispensed packs and the periodicity of dispensations, produces revised exposure trajectories that differ from simpler calculations in 35 percent of cases, resulting in 4 percent of trajectories moving to different clusters with measurably better clustering quality and observable shifts in associations between clusters and neonatal outcomes such as pathology rates linked to high psychotropic exposure.

What carries the argument

The adjustment method that recalculates exposure trajectories using the number of dispensed packs and dispensation type (occasional or regular).

If this is right

  • Significant changes occur in the number of defined daily doses for some exposed women.
  • Four percent of trajectories are reassigned to different clusters.
  • Clustering quality improves under the adjusted calculation.
  • Cluster distributions shift for certain maternal characteristics and neonatal outcomes.
  • The link between high psychotropic doses and neonatal pathology is reinforced.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method could be tested on other prescription databases to check if similar trajectory shifts appear.
  • If validated, it might change how exposure is estimated in studies linking prenatal medication to child health.
  • Future work could examine whether the adjustment affects conclusions for medication classes beyond psychotropics.

Load-bearing premise

Accounting for the number of dispensed packs and the periodicity of dispensations yields a more accurate representation of actual medication exposure than simpler calculations.

What would settle it

Direct validation of the adjusted trajectories against patient pill counts or pharmacy refill records over time would show whether the observed changes in DDD totals and cluster assignments better match actual intake patterns.

read the original abstract

We presented an adjustment method for the calculation of medication exposure trajectories based on the number of dispensed packs and the type of dispensations (occasional or regular). A comparative study based on the EFEMERIS data was carried out using three different scenarios of trajectory calculation depending on whether or not the number of packs and the periodicity of medication dispensations were taken into account. The impact of the scenario was highlighted using global indicators on the number of Define-Daily Dose (DDD) on all women exposed; the study of changes in individual trajectories from one scenario to another was carried out; we also compared the results of a clustering into four groups. If 65% of the trajectories remained unchanged, we could observe on the rest significant changes in number of DDD and/or on individual exposure profile. We observed 4% of trajectories that were attributed to a different cluster, and the clustering was of better quality with the adjustment method. Depending on the study context, an impact on cluster distribution could be observed for some maternal characteristics and neonatal outcomes. This was the case for a higher occurrence of neonatal pathology for neonates from mothers belonging to the cluster with high doses of psychotropics, thus reinforcing the conclusions of previous studies of a link between high exposure to psychotropic medications and presence of pathology for the newborn.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The manuscript presents an adjustment method for computing medication exposure trajectories that incorporates the number of dispensed packs and dispensation periodicity (occasional versus regular). On the EFEMERIS database, three scenarios are compared; 65% of trajectories are unchanged while the remainder exhibit changes in total DDD or exposure profile, 4% of trajectories switch clusters, clustering quality improves under the adjusted method, and some maternal/neonatal outcome associations shift (e.g., stronger link between the high-psychotropic cluster and neonatal pathology).

Significance. If the adjustment demonstrably reduces measurement error relative to simpler methods, the approach could improve exposure assessment in pharmacoepidemiologic studies of pregnancy medication use and strengthen causal inferences about dose-response relationships with neonatal outcomes. The large database and explicit scenario comparison are positive features, but the absence of ground-truth validation limits the strength of the accuracy claim.

major comments (3)
  1. [Methods] Methods section: the precise algorithmic rules for adjusting DDD counts from pack numbers and periodicity classification are not specified with sufficient detail (e.g., decision thresholds, handling of overlapping dispensations, or edge cases), rendering the central comparative results non-reproducible and preventing assessment of whether the adjustment is unbiased.
  2. [Results] Results (clustering subsection): the claim that improved clustering quality demonstrates superior accuracy is not supported, because no external ground-truth measure of actual consumption exists; any reparameterization that increases between-cluster separation can improve internal metrics without increasing fidelity to real intake. The 4% cluster-switch rate and outcome-association shifts could equally reflect introduced bias.
  3. [Discussion] Discussion: the interpretation that the adjustment 'reinforces' prior findings on psychotropic exposure and neonatal pathology assumes the adjusted trajectories are closer to truth, yet the paper supplies no sensitivity analysis or external validation to rule out systematic bias introduced by the pack/periodicity rules.
minor comments (1)
  1. [Abstract] Abstract: the three scenarios are referenced but never enumerated explicitly, forcing the reader to infer their definitions from later text.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We appreciate the referee's constructive feedback on our manuscript. We have carefully considered each major comment and provide point-by-point responses below. Where appropriate, we will revise the manuscript to address the concerns raised.

read point-by-point responses

  1. Referee: [Methods] Methods section: the precise algorithmic rules for adjusting DDD counts from pack numbers and periodicity classification are not specified with sufficient detail (e.g., decision thresholds, handling of overlapping dispensations, or edge cases), rendering the central comparative results non-reproducible and preventing assessment of whether the adjustment is unbiased.

    Authors: We agree that additional detail is necessary for reproducibility. In the revised manuscript, we will expand the Methods section to include the precise algorithmic rules, including decision thresholds for classifying dispensations as occasional or regular, procedures for handling overlapping dispensations, and examples of edge cases. revision: yes

  2. Referee: [Results] Results (clustering subsection): the claim that improved clustering quality demonstrates superior accuracy is not supported, because no external ground-truth measure of actual consumption exists; any reparameterization that increases between-cluster separation can improve internal metrics without increasing fidelity to real intake. The 4% cluster-switch rate and outcome-association shifts could equally reflect introduced bias.

    Authors: We acknowledge the validity of this point. The manuscript will be revised to clarify that the improved clustering quality is an internal validation metric and does not constitute evidence of superior accuracy or reduced measurement error. We will rephrase the relevant sections to present the changes in cluster assignments and outcome associations as descriptive findings without implying that they demonstrate the adjustment's accuracy. revision: yes

  3. Referee: [Discussion] Discussion: the interpretation that the adjustment 'reinforces' prior findings on psychotropic exposure and neonatal pathology assumes the adjusted trajectories are closer to truth, yet the paper supplies no sensitivity analysis or external validation to rule out systematic bias introduced by the pack/periodicity rules.

    Authors: We agree that the language should be tempered to avoid implying validation of the adjustment. In the revised Discussion, we will modify the interpretation to state that the adjusted method leads to a stronger observed association in this dataset, while noting the lack of external validation and the possibility of introduced bias. We will avoid phrases that suggest the adjusted trajectories are necessarily closer to the true exposure. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical comparison of adjustment scenarios on external database

full rationale

The paper defines an adjustment method for exposure trajectories based on pack counts and dispensation periodicity, then applies three scenarios to the EFEMERIS database. It reports empirical changes (65% trajectories unchanged, 4% cluster switches, improved clustering quality) and shifts in associations with maternal/neonatal outcomes. No derivation chain reduces a claimed result to its own inputs by construction; there are no fitted parameters renamed as predictions, self-definitional equations, or load-bearing self-citations. The clustering-quality comparison is an internal metric evaluated against the same external records, not a self-referential validation loop. This is a standard methodological comparison study with no circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no mathematical details, so no free parameters, axioms, or invented entities can be identified.

pith-pipeline@v0.9.1-grok · 5799 in / 1217 out tokens · 29986 ms · 2026-06-27T14:28:33.213507+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

16 extracted references · 1 canonical work pages

  1. [1]

    Exposed or not exposed? Explor ing exposure classific ation in studies using administra tive data to investigate outcomes following medication use during pregnancy

    Grzeskowiak LE, Gilbert AL, Morrison JL. Exposed or not exposed? Explor ing exposure classific ation in studies using administra tive data to investigate outcomes following medication use during pregnancy. European journal of clinical pharma cology. 2012; 68:459– 467. https://doi.or g/10.100 7/s00228-011- 1154-9 PMID: 220801 82

    2012

  2. [2]

    Methods for time-varyin g exposure related problems in pharmaco epidemiology : an overview

    Pazzagli L, Linder M, Zhang M, Vago E, Stang P, Myers D, et al. Methods for time-varyin g exposure related problems in pharmaco epidemiology : an overview. Pharmac oepidemiology and drug safety. 2018; 27(2):148– 160. https://doi.or g/10.1002/ pds.4372 PMID: 29285840

    2018

  3. [3]

    Longitudi- nal methods for modelin g exposures in pharmaco epidemiologic studies in pregnanc y

    Wood ME, Lupattelli A, Palmsten K, Bandoli G, Hurault- Delarue C, Damase-M ichel C, et al. Longitudi- nal methods for modelin g exposures in pharmaco epidemiologic studies in pregnanc y. Epidem iologic Reviews. 2021; 43(1):130– 146. https://doi.or g/10.1093/ epirev/mxab 002

    2021

  4. [4]

    Group-ba sed tra- jectory models: a new approac h to classifying and predicting long-term medication adherenc e

    Franklin JM, Shrank WH, Pakes J, Sanfe ´ lix-Gimeno G, Matlin OS, Brennan TA, et al. Group-ba sed tra- jectory models: a new approac h to classifying and predicting long-term medication adherenc e. Medical care. 2013; 51(9):789– 796. https://doi.or g/10.109 7/MLR.0b013e 318298 4c1f PMID: 23685406

    2013

  5. [5]

    Modelin g exposure s of medication s used episodica lly during preg- nancy: Triptans as a motivating example

    Harris GME, Wood M, Nordeng H. Modelin g exposure s of medication s used episodica lly during preg- nancy: Triptans as a motivating example. Pharma coepidemi ology and drug safety. 2020; 29(9):1111 –

    2020

  6. [6]

    https://d oi.org/10.100 2/pds.508 9 PMID: 327485 40 PLOS ONE Adjusted trajectory of medicat ion exposure and comparati ve analysis on EFEMERIS database PLOS ONE | https://doi.or g/10.137 1/journal.po ne.03087 67 February 13, 2025 9 / 10

    2025

  7. [7]

    Frank AS, Lupattel li A, Matteson DS, Nordeng H. Maternal use of thyroid hormon e replacem ent therapy before, during, and after pregnancy: agreem ent between self-report and prescrip tion records and group-based trajector y modelin g of prescription pattern s. Clinical Epidemiology . 2018; p. 1801–181 6. https://doi.or g/10.214 7/CLEP.S 175616 PMID: 30584374

    2018

  8. [8]

    How to take into account exposure to drugs over time in pharmaco epidemiology studies of pregnant women? Pharma - coepide miology and drug safety

    Hurault-De larue C, Chouquet C, Savy N, Lacroix I, Beau AB, Montastruc JL, et al. How to take into account exposure to drugs over time in pharmaco epidemiology studies of pregnant women? Pharma - coepide miology and drug safety. 2016; 25(7):770– 777. https://doi. org/10.1002/p ds.4000 PMID: 27018245

    work page doi:10.1002/p 2016

  9. [9]

    Patterns of prenatal antidepress ant exposure and risk of preeclam psia and postpartu m haemorrhage

    Palmsten K, Chambers CD, Wells A, Bandoli G. Patterns of prenatal antidepress ant exposure and risk of preeclam psia and postpartu m haemorrhage . Paediatric and perinat al epidemi ology. 2020; 34 (5):597–60 6. https://do i.org/10.1111 /ppe.126 60 PMID: 32207549

    2020

  10. [10]

    Prescription of drugs during preg- nancy: a study using EFEME RIS, the new French database

    Lacroix I, Hurault C, Sarramon M, Guitard C, Berrebi A, Grau M, et al. Prescription of drugs during preg- nancy: a study using EFEME RIS, the new French database . European journal of clinical pharma cology. 2009; 65:839– 846. https://doi.or g/10.100 7/s00228-009- 0647-2 PMID: 19365629

    2009

  11. [11]

    Definitions and general considera tions; 2018

    WHO CC. Definitions and general considera tions; 2018. Available from: https://atcd dd.fhi.no/ddd/ definition_an d_gene ral_considera

    2018

  12. [12]

    Healthcare data- bases in Europe for studying medicine use and safety during pregnancy

    Charlton RA, Neville AJ, Jordan S, Pierini A, Dama se-Michel C, Klungsøy r K, et al. Healthcare data- bases in Europe for studying medicine use and safety during pregnancy . pharmaco epidemiology and drug safety. 2014; 23(6):586– 594. https://doi.or g/10.100 2/pds.3613 PMID: 24664855

    2014

  13. [13]

    Prescri ption drug use during pregnanc y in France: a study from the national health insurance perman ent sample

    Demailly R, Escolano S, Quantin C, Tubert-Bitter P, Ahmed I. Prescri ption drug use during pregnanc y in France: a study from the national health insurance perman ent sample. Pharmac oepidemiology and drug safety. 2017; 26(9):1126 –1134. https://doi.or g/10.100 2/pds.426 5 PMID: 28758270

    2017

  14. [14]

    Interest of the trajec- tory method for the evaluatio n of outcomes after in utero drug exposure : example of anxiolytics and hyp- notics

    Hurault-De larue C, Chouquet C, Savy N, Lacroix I, Beau AB, Montastruc JL, et al. Interest of the trajec- tory method for the evaluatio n of outcomes after in utero drug exposure : example of anxiolytics and hyp- notics. Pharmac oepidemiology and Drug Safety. 2017; 26(5):561– 569. https://doi.or g/10.1002/ pds. 4199 PMID: 283378 23

    2017

  15. [15]

    KmL: k-means for longitudinal data

    Genolini C, Falissard B. KmL: k-means for longitudinal data. Comp utational Statistics. 2010; 25(2):317–

    2010

  16. [16]

    https://do i.org/10.1007 /s00180- 009-0178- 4 PLOS ONE Adjusted trajectory of medicat ion exposure and comparati ve analysis on EFEMERIS database PLOS ONE | https://doi.or g/10.137 1/journal.po ne.03087 67 February 13, 2025 10 / 10

    2025