pith. sign in

arxiv: 1906.09085 · v2 · pith:NIE4KCJ5new · submitted 2019-06-21 · ⚛️ physics.soc-ph · q-bio.PE

Evaluating the impact of PrEP on HIV and gonorrhea on a networked population of female sex workers

Alba Bernini , Elodie Blouzard , Alberto Bracci , Pau Casanova , Iacopo Iacopini , Benjamin Steinegger , Andreia Sofia Teixeira , Alberto Antonioni
show 1 more author
Eugenio Valdano
This is my paper

Pith reviewed 2026-05-25 18:19 UTC · model grok-4.3

classification ⚛️ physics.soc-ph q-bio.PE
keywords PrEPHIVgonorrheafemale sex workerssexual networksrisk compensationepidemic simulationSTI interventions
0
0 comments X

The pith

Simulations on an empirical network show that PrEP reduces HIV prevalence even with risk compensation, but the HIV-gonorrhea interplay requires different strategies at different compensation levels and uniform PrEP adoption outperforms use仅

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

The paper simulates HIV and gonorrhea transmission on a real sexual-contact network linking female sex workers and clients. When only HIV is present, PrEP lowers prevalence regardless of increased condomless sex. When both infections circulate, the paper finds that the level of risk compensation determines which intervention works best. Uniform PrEP coverage across the population reduces overall prevalence more than giving it only to the highest-activity individuals.

Core claim

Using an empirical network of sexual contacts among female sex workers and clients, the authors simulate the joint spread of HIV and gonorrhea. PrEP lowers HIV prevalence even under high risk compensation when HIV circulates alone. With both diseases present, the complex interactions mean that different amounts of risk compensation call for different intervention designs. Providing PrEP only to the most active female sex workers proves less effective than uniform adoption across the group.

What carries the argument

Agent-based simulation on an empirical sexual-contact network that incorporates PrEP adherence, transmission probabilities, and a risk-compensation parameter controlling the fraction of condomless acts.

If this is right

  • PrEP remains effective against HIV even when risk compensation is high if HIV is the only circulating infection.
  • Joint circulation of HIV and gonorrhea means that intervention choice must be matched to the prevailing level of risk compensation.
  • Uniform PrEP distribution across female sex workers reduces prevalence more than targeting only the most active subset.
  • Effects that emerge only from the interaction of multiple diseases and multiple prophylactics must be included when planning real-world strategies.

Where Pith is reading between the lines

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

  • Public-health programs may need to monitor actual condom-use changes after PrEP rollout rather than assume fixed behavior.
  • The same network-based approach could be applied to other pairs of infections to identify when uniform versus targeted distribution is preferable.
  • If risk compensation proves higher than modeled, the paper's results suggest that PrEP alone may not suffice and additional measures would be required.

Load-bearing premise

The chosen empirical network, transmission probabilities, adherence rates, and risk-compensation parameters correctly capture real behavior and biology in the studied population.

What would settle it

Direct observation of changes in condom-use rates and measured HIV and gonorrhea incidence among female sex workers who begin PrEP in a comparable setting.

Figures

Figures reproduced from arXiv: 1906.09085 by Alba Bernini, Alberto Antonioni, Alberto Bracci, Andreia Sofia Teixeira, Benjamin Steinegger, Elodie Blouzard, Eugenio Valdano, Iacopo Iacopini, Pau Casanova.

Figure 1
Figure 1. Figure 1: ) in which individuals are divided into six different compartments, according to their status with respect to the considered diseases: • S: susceptible individuals who can acquire both HIV and gonorrhea by means of a sexual contact; • GE: exposed individuals who have acquired gonor￾rhea but are not yet able to transmit it; • G: gonorrhea-infectious individuals who have gon￾orrhea and can transmit it; • H: … view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
read the original abstract

Sexual contacts are the main spreading route of HIV. This puts sex workers at higher risk of infection even in populations where HIV prevalence is moderate or low. Alongside condom use, Pre-Exposure Prophylaxis (PrEP) is an effective tool for sex workers to reduce their risk of HIV acquisition. However, PrEP provides no direct protection against sexually transmitted infections (STIs) other than HIV, unlike condoms. We use an empirical network of sexual contacts among female sex workers (FSWs) and clients to simulate the spread of HIV and gonorrhea. We then investigate the effect of PrEP adoption and adherence, on both HIV and gonorrhea prevalence. We also study the effect of a potential increase in condomless acts due to lowered risk perception with respect of the no-PrEP scenario (risk compensation). We find that when HIV is the only disease circulating, PrEP is effective in reducing HIV prevalence, even with high risk compensation. Instead, the complex interplay between the two diseases shows that different levels of risk compensation require different intervention strategies. Finally, we find that providing PrEP only to the most active FSWs is less effective than uniform PrEP adoption. Our work shows that the effects emerging from the complex interactions between these diseases and the available prophylactic measures need to be accounted for, to devise effective intervention strategies.

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

2 major / 1 minor

Summary. The paper uses agent-based simulations on an empirical sexual-contact network of female sex workers (FSWs) and clients to model the joint spread of HIV and gonorrhea. It examines how PrEP adoption, adherence, and risk compensation (increased condomless acts) affect prevalence of both infections, and compares uniform versus targeted (most-active FSWs only) PrEP distribution. The central claims are that PrEP remains effective against HIV even with high risk compensation when gonorrhea is absent, that the HIV-gonorrhea interplay implies different risk-compensation levels require different intervention strategies, and that uniform PrEP is more effective than targeting the most active FSWs.

Significance. If the simulation results prove robust, the work would usefully illustrate how co-circulating STIs and behavioral feedback can alter the ranking of PrEP strategies in a high-risk population. The use of an empirical contact network is a positive feature, as is the explicit treatment of risk compensation.

major comments (2)
  1. [Abstract / Methods] Abstract and Methods: the headline ranking (uniform PrEP more effective than targeting most-active FSWs) and the claim that different risk-compensation levels require different strategies rest entirely on simulation outputs. The abstract states that an empirical network is used but supplies no validation, calibration procedure, or robustness checks against alternative networks or parameter ranges; these inputs are load-bearing for the reported interplay and strategy conclusions.
  2. [Results] Results: the four free parameters (HIV transmission probability per act, gonorrhea transmission probability per act, PrEP adherence fraction, risk-compensation multiplier) are not accompanied by reported sensitivity analyses or ranges; without such checks the conclusion that risk-compensation level dictates intervention choice cannot be assessed for stability.
minor comments (1)
  1. [Methods] Notation for the risk-compensation multiplier and the precise functional form by which it increases condomless acts should be stated explicitly in the model description.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments, which help clarify how to strengthen the presentation of our simulation results. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract / Methods] Abstract and Methods: the headline ranking (uniform PrEP more effective than targeting most-active FSWs) and the claim that different risk-compensation levels require different strategies rest entirely on simulation outputs. The abstract states that an empirical network is used but supplies no validation, calibration procedure, or robustness checks against alternative networks or parameter ranges; these inputs are load-bearing for the reported interplay and strategy conclusions.

    Authors: We agree that the abstract and Methods would benefit from explicit statements on network provenance, validation against the original contact data, and basic robustness checks. The network is taken from a prior empirical study of FSW-client contacts; our focus is on relative intervention effects rather than absolute prevalence forecasts. In revision we will add a short subsection describing the network construction, any calibration steps performed, and limited checks (e.g., rewiring or degree-preserving randomization) to confirm that the reported ranking of uniform versus targeted PrEP is not an artifact of the specific network realization. revision: yes

  2. Referee: [Results] Results: the four free parameters (HIV transmission probability per act, gonorrhea transmission probability per act, PrEP adherence fraction, risk-compensation multiplier) are not accompanied by reported sensitivity analyses or ranges; without such checks the conclusion that risk-compensation level dictates intervention choice cannot be assessed for stability.

    Authors: The referee correctly notes the absence of systematic sensitivity reporting for these four parameters. While we performed exploratory runs across plausible ranges to verify that the qualitative ordering of strategies persists, those checks were not documented. We will add a dedicated sensitivity subsection (or supplementary figure) that varies each parameter individually and jointly, showing the stability of the key claims: (i) PrEP efficacy against HIV under risk compensation and (ii) the dependence of optimal strategy on risk-compensation level. revision: yes

Circularity Check

0 steps flagged

No circularity: simulation outputs independent of self-defined quantities.

full rationale

The paper reports agent-based simulations of HIV/gonorrhea spread on an external empirical contact network, using standard transmission rules, adherence rates, and a parameterized risk-compensation term. No derivation step reduces by the paper's own equations to a quantity defined in terms of its outputs; results on uniform vs. targeted PrEP and risk-compensation thresholds are generated by forward simulation rather than forced by construction or self-citation chains. The central claims therefore retain independent content relative to the inputs.

Axiom & Free-Parameter Ledger

4 free parameters · 2 axioms · 0 invented entities

The simulation depends on transmission probabilities, adherence fractions, risk-compensation multipliers, and the empirical network structure; none of these are derived inside the paper.

free parameters (4)
  • HIV transmission probability per act
    Standard parameter fitted or taken from literature; central to prevalence outcomes.
  • Gonorrhea transmission probability per act
    Standard parameter; interacts with HIV via co-infection effects.
  • PrEP adherence fraction
    Directly controls HIV reduction; varied across scenarios.
  • Risk-compensation multiplier
    Increase in condomless acts; the paper varies this to show strategy dependence.
axioms (2)
  • domain assumption The empirical contact network is a faithful representation of the underlying sexual-contact process.
    Invoked when the simulation is initialized from the network data.
  • standard math Transmission events are independent across acts conditional on network edges and PrEP status.
    Standard compartmental or individual-based epidemic modeling assumption.

pith-pipeline@v0.9.0 · 5809 in / 1469 out tokens · 18896 ms · 2026-05-25T18:19:39.675870+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

50 extracted references · 50 canonical work pages

  1. [1]

    W. H. Organization et al., (2012)

    work page 2012

  2. [2]

    Baral, C

    S. Baral, C. Beyrer, K. Muessig, T. Poteat, A. L. Wirtz, M. R. Decker, S. G. Sherman, and D. Kerrigan, The Lancet infectious diseases 12, 538 (2012)

    work page 2012

  3. [3]

    Shannon, S

    K. Shannon, S. A. Strathdee, S. M. Goldenberg, P. Duff, P. Mwangi, M. Rusakova, S. Reza-Paul, J. Lau, K. Deer- ing, M. R. Pickles, and M.-C. Boily, The Lancet 385, 55 (2015)

    work page 2015

  4. [4]

    Ghimire, W

    L. Ghimire, W. C. S. Smith, E. R. van Teijlingen, R. Da- hal, and N. P. Luitel, BMC Women’s Health 11, 42 (2011)

    work page 2011

  5. [5]

    Shannon, S

    K. Shannon, S. A. Strathdee, J. Shoveller, M. Rusch, T. Kerr, and M. W. Tyndall, American journal of public health 99, 659 (2009)

    work page 2009

  6. [6]

    L. A. Urada, D. E. Morisky, N. Pimentel-Simbulan, J. G. Silverman, and S. A. Strathdee, PLoS One 7, e33282 (2012)

    work page 2012

  7. [7]

    Eakle, R

    R. Eakle, R. Bothma, A. Bourne, S. Gumede, K. Motsosi, and H. Rees, PLOS ONE 14, e0212271 (2019)

    work page 2019

  8. [8]

    C. A. Flash, S. K. Dale, and D. S. Krakower, Interna- tional Journal of Women’s Health 9, 391 (2017)

    work page 2017

  9. [9]

    V. A. Fonner, S. L. Dalglish, C. E. Kennedy, R. Baggaley, K. R. Oreilly, F. M. Koechlin, M. Rodolph, I. Hodges- Mameletzis, and R. M. Grant, AIDS (London, England) 30, 1973 (2016)

    work page 1973

  10. [10]

    J. T. Galea, J. J. Kinsler, X. Salazar, S.-J. Lee, M. Giron, J. N. Sayles, C. C´ aceres, and W. E. Cunningham, Inter- national journal of STD & AIDS 22, 256 (2011)

    work page 2011

  11. [11]

    Reza-Paul, L

    S. Reza-Paul, L. Lazarus, M. Doshi, S. H. U. Rah- man, M. Ramaiah, R. Maiya, M. Venugopal, K. Venuku- mar, S. Sundararaman, M. Becker, et al. , PloS one 11, e0166889 (2016)

    work page 2016

  12. [12]

    Eakle, N

    R. Eakle, N. Mutanha, J. Mbogua, M. Sibanyoni, A. Bourne, G. Gomez, F. Venter, and H. Rees, BMJ open 8, e019292 (2018)

    work page 2018

  13. [13]

    K. F. Ortblad, M. M. Chanda, D. K. Musoke, T. Ngabi- 8 rano, M. Mwale, A. Nakitende, S. Chongo, N. Ka- mungoma, C. Kanchele, T. B¨ arnighausen, et al. , BMC infectious diseases 18, 503 (2018)

    work page 2018

  14. [14]

    Mboup, L

    A. Mboup, L. B´ ehanzin, F. A. Gu´ edou, N. Ger- aldo, E. Goma-Mats´ ets´ e, K. Gigu` ere, M. Aza-Gnandji, L. Kessou, M. Diallo, R. K. Kˆ ekˆ e,et al., Journal of the International AIDS Society 21, e25208 (2018)

    work page 2018

  15. [15]

    A. R. Bazzi, K. Yotebieng, S. Otticha, G. Rota, K. Agot, S. Ohaga, and J. L. Syvertsen, Journal of the interna- tional AIDS society 22, e25266 (2019)

    work page 2019

  16. [16]

    Grant, Z

    H. Grant, Z. Mukandavire, R. Eakle, H. Prudden, G. B Gomez, H. Rees, and C. Watts, Journal of the International AIDS Society 20 (2017)

    work page 2017

  17. [17]

    M. M. Cassell, D. T. Halperin, J. D. Shelton, and D. Stanton, Bmj 332, 605 (2006)

    work page 2006

  18. [18]

    Blumenthal and R

    J. Blumenthal and R. Haubrich, The virtual mentor: VM 16, 909 (2014)

    work page 2014

  19. [19]

    Nguyen, Z

    V.-K. Nguyen, Z. R. Greenwald, H. Trottier, M. Cadieux, A. Goyette, M. Beauchemin, L. Charest, D. Longpr´ e, S. Lavoie, H. G. Tossa, et al., AIDS (London, England) 32, 523 (2018)

    work page 2018

  20. [20]

    S. C. Kalichman, J. Pellowski, and C. Turner, Sexually transmitted infections 87, 183 (2011)

    work page 2011

  21. [21]

    K. K. Mugwanya, D. Donnell, C. Celum, K. K. Thomas, P. Ndase, N. Mugo, E. Katabira, K. Ngure, J. M. Baeten, P. P. S. Team, et al., The Lancet infectious diseases 13, 1021 (2013)

    work page 2013

  22. [22]

    Pilgrim, N

    N. Pilgrim, N. Jani, S. Mathur, C. Kahabuka, V. Saria, N. Makyao, L. Apicella, and J. Pulerwitz, PloS one 13, e0196280 (2018)

    work page 2018

  23. [23]

    Unemo and W

    M. Unemo and W. M. Shafer, Clinical microbiology re- views 27, 587 (2014)

    work page 2014

  24. [24]

    S. M. Jenness, K. M. Weiss, S. M. Goodreau, T. Gift, H. Chesson, K. W. Hoover, D. K. Smith, A. Y. Liu, P. S. Sullivan, and E. S. Rosenberg, Clinical Infectious Dis- eases 65, 712 (2017)

    work page 2017

  25. [25]

    Zhong, Q

    L. Zhong, Q. Zhang, and X. Li, Scientific Reports 8, 2432 (2018)

    work page 2018

  26. [26]

    Pretorius, J

    C. Pretorius, J. Stover, L. Bollinger, N. Baca¨ er, and B. Williams, PloS one 5, e13646 (2010)

    work page 2010

  27. [27]

    Mushayabasa, J

    S. Mushayabasa, J. M. Tchuenche, C. P. Bhunu, and E. Ngarakana-Gwasira, BioSystems 103, 27 (2011)

    work page 2011

  28. [28]

    Moreno, R

    Y. Moreno, R. Pastor-Satorras, and A. Vespignani, The European Physical Journal B-Condensed Matter and Complex Systems 26, 521 (2002)

    work page 2002

  29. [29]

    Salath´ e, M

    M. Salath´ e, M. Kazandjieva, J. W. Lee, P. Levis, M. W. Feldman, and J. H. Jones, Proceedings of the National Academy of Sciences 107, 22020 (2010)

    work page 2010

  30. [30]

    Barrat, M

    A. Barrat, M. Barthelemy, and A. Vespignani, Dynami- cal processes on complex networks(Cambridge University Press, 2008) p. 347

    work page 2008

  31. [31]

    Gauvin, A

    L. Gauvin, A. Panisson, C. Cattuto, and A. Barrat, Scientific reports 3, 3099 (2013)

    work page 2013

  32. [32]

    Barrat, M

    A. Barrat, M. Barthelemy, and A. Vespignani, Dynami- cal processes on complex networks (Cambridge university press, 2008)

    work page 2008

  33. [33]

    Holme and N

    P. Holme and N. Litvak, PLoS computational biology 13, e1005696 (2017)

    work page 2017

  34. [34]

    Schiffman, Archives of pathology & laboratory medicine 127, 930 (2003)

    M. Schiffman, Archives of pathology & laboratory medicine 127, 930 (2003)

    work page 2003

  35. [35]

    Cohen, C

    T. Cohen, C. Colijn, and M. Murray, Proceedings of the National Academy of Sciences 105, 16302 (2008)

    work page 2008

  36. [36]

    Poletto, S

    C. Poletto, S. Meloni, A. Van Metre, V. Colizza, Y. Moreno, and A. Vespignani, Scientific reports 5, 7895 (2015)

    work page 2015

  37. [37]

    L. Chen, F. Ghanbarnejad, and D. Brockmann, New Journal of Physics 19, 103041 (2017)

    work page 2017

  38. [38]

    Pinotti, ´E

    F. Pinotti, ´E. Fleury, D. Guillemot, P.-Y. B¨ oelle, and C. Poletto, bioRxiv (2018)

    work page 2018

  39. [39]

    L. E. Rocha, F. Liljeros, and P. Holme, Proceedings of the National Academy of Sciences 107, 5706 (2010)

    work page 2010

  40. [40]

    L. E. Rocha, F. Liljeros, and P. Holme, PLoS computa- tional biology 7, e1001109 (2011)

    work page 2011

  41. [41]

    Holme and J

    P. Holme and J. Saram¨ aki,Temporal networks (Springer, 2013)

    work page 2013

  42. [42]

    M. E. Kretzschmar, Y. T. v. Duynhoven, and A. J. d. Severijnen, American journal of epidemiology 144, 306 (1996)

    work page 1996

  43. [43]

    Bekker, L

    L.-G. Bekker, L. Johnson, F. Cowan, C. Overs, D. Be- sada, S. Hillier, and W. Cates, The Lancet 385, 72 (2015)

    work page 2015

  44. [44]

    H. W. Chesson and S. D. Pinkerton, JAIDS Journal of Acquired Immune Deficiency Syndromes 24, 48 (2000)

    work page 2000

  45. [45]

    Mukandavire, K

    Z. Mukandavire, K. M. Mitchell, and P. Vickerman, Epi- demics 14, 62 (2016)

    work page 2016

  46. [46]

    D. R. MacFadden, D. H. Tan, and S. Mishra, Journal of the International AIDS Society 19 (2016)

    work page 2016

  47. [47]

    Tripathi, R

    A. Tripathi, R. Naresh, and D. Sharma, Applied math- ematics and computation 184, 1053 (2007)

    work page 2007

  48. [48]

    Stehl´ e, N

    J. Stehl´ e, N. Voirin, A. Barrat, C. Cattuto, V. Colizza, L. Isella, C. R´ egis, J.-F. Pinton, N. Khanafer, W. Van den Broeck, and P. Vanhems, BMC Medicine” 9, 87 (2011)

    work page 2011

  49. [49]

    D. T. Dimitrov, B. R. Mˆ asse, and D. Donnell, Journal of acquired immune deficiency syndromes (1999) 72, 444 (2016)

    work page 1999

  50. [50]

    Cohen, S

    R. Cohen, S. Havlin, and D. ben Avraham, Phys. Rev. Lett. 91, 247901 (2003)

    work page 2003