HIV prevention patterns stayed stable, but first choices shaped later risk pathways
In the French ANRS-PREVENIR cohort, HIV prevention strategies among men who have sex with men followed a strikingly sticky pattern: once a protection state was adopted, participants were most likely to remain in that same state at the next follow-up visit. The analysis separated prevention into three states. Self-dependent protection, or SDP, included condom use and/or PrEP. Partner-dependent protection, or PDP, relied on a partner’s declaration of being HIV-negative or living with an undetectable viral load. No protection, or NP, meant neither strategy was in place.
The highest transition probabilities over the study period favored persistence rather than movement. For SDP, the probability of remaining in the same state at the next visit was 0.920 [0.916, 0.924]. For PDP, it was 0.783 [0.769, 0.797]. For NP, it was 0.584 [0.546, 0.615].
Persistence dominated behavior.
Those numbers matter because they suggest that prevention is not simply a series of isolated decisions before each sexual encounter. Instead, prevention appears to settle into a habitual pattern, and the first strategy adopted may anchor future choices. In practical terms, a person who begins with self-controlled prevention may stay there, while a person who starts with reliance on partner status or no protection may continue along that pathway unless counseling interrupts the trajectory.
PrEP experience and event-driven use were linked to movement away from self-controlled protection
The most clinically relevant signal came from the transitions out of SDP. Two factors increased the probability of moving from self-dependent protection to partner-dependent protection: event-driven PrEP, with a hazard ratio of 3.34 [2.89–3.85], and PrEP naivety, with a hazard ratio of 1.19 [1.04–1.36]. The same two factors were also associated with a higher probability of moving from SDP to no protection. PrEP naivety carried a hazard ratio of 1.44 [1.23–1.69] for the transition from SDP to NP, while event-driven PrEP carried a hazard ratio of 4.15 [3.47–4.96].
The association is biologically and behaviorally plausible. Event-driven PrEP is a non-daily regimen tied to anticipation of risk, which can make protection feel episodic rather than embedded in routine. A person who is not yet experienced with PrEP may also be less likely to maintain a stable self-managed prevention pattern, especially if sexual risk perception changes over time. In that setting, prevention can become reactive instead of structured, and reactive prevention is more vulnerable to interruption.
A shift away from self-controlled prevention is not a trivial statistical movement. It can mean a transition toward strategies that depend on partner disclosure, partner viral suppression, or in some cases no active protection at all, all of which may introduce greater vulnerability when risk is misjudged or disclosures are incomplete.
Perceived and actual risk did not move all transitions in the same direction
Investigators also constructed a composite measure of concordance between actual risk and perceived risk. Higher scores reflected higher actual risk together with either greater discordance or no consideration of risk. A higher concordance score was associated with a reduced probability of transitioning from SDP to PDP, with a hazard ratio of 0.96 [0.92–0.99]. It was not associated with the transition from SDP to NP.
This pattern suggests that risk awareness exerts selective influence rather than a broad behavioral reset. When actual risk is recognized, some people may be less likely to shift from self-controlled protection to reliance on a partner’s status. Yet the same measure did not predict a move into no protection, implying that abandoning all protection may be driven by other forces: relationship dynamics, sexual context, prior experience with PrEP, trust, convenience, or simple fatigue with prevention routines.
The model used here was a continuous-time multi-state Markov approach, which is well suited to describing movement between clinically meaningful states over time. It captures the reality that prevention behavior is dynamic, not fixed. It also allows the next step to be modeled as a function of the current step, which is exactly the pattern seen in this cohort. For sexual health counseling, that is a useful framework: interventions may need to target the moment when a person first adopts a prevention method, because that first choice appears to shape what comes next.
Counseling at the first prevention decision may carry the greatest leverage
The direct bedside implication is straightforward. People initiating PrEP or choosing a prevention strategy for the first time may benefit from closer counseling, not just on how to use the method, but on how that choice can influence future prevention behavior. A person starting event-driven PrEP may need explicit discussion about adherence, risk anticipation, and how quickly prevention can erode when routine use is not established. PrEP-naive individuals may need additional support in translating a prevention plan into a durable habit.
The study also raises a difficult question about vulnerability to HIV acquisition in those transitioning toward partner-dependent or no protection states. Partner-dependent protection can be effective when partner information is accurate and sustained, especially in the context of viral suppression, but it still relies on trust, disclosure, and stable relationship assumptions. When those assumptions break down, the protection model weakens immediately.
One sentence captures the clinical tension: reliance on partner status is only as strong as the accuracy of that status.
What the cohort can and cannot answer
The findings are confined to the prevention states defined in this cohort and the transitions observed during follow-up. The analysis does not prove causation, and it does not measure every social or relational factor that may steer prevention choices. It also does not tell clinicians exactly which counseling script will prevent a transition to less self-controlled protection.
Even so, the direction of the signal is clear. Most participants stayed where they were, and the earliest prevention choices were linked to later movement between states. For sexual health services, that places unusual weight on the first conversation about HIV prevention, especially for PrEP-naive individuals and people using event-driven regimens. If those conversations fail to establish a stable framework, later prevention can drift toward strategies that leave more room for missed risk.