HIV self testing

[pineapple-cat]

SAS ln 4898.

       * SELF-TESTING;

	eff_self_test_targeting = self_test_targeting;

	w = rand('uniform');	
	if hard_reach=0 or (hard_reach = 1 and w < prob_self_test_hard_reach) then do;

		u_self_test=rand('uniform');
 		if . < np_lasttest <= 0 then u_self_test = u_self_test * eff_self_test_targeting;  
		if newp_lasttest ge 1 then u_self_test=u_self_test/eff_self_test_targeting;  
		if secondary_self_test=1 and eponart=1 then u_self_test=u_self_test/secondary_self_test_targeting;  
		if tested ne 1 and (caldate{t]-max(0,dt_last_self_test) >= 0.25) and u_self_test < rate_self_test then do;
			self_tested=1; 
			dt_last_self_test=caldate{t}; 
		end;
	end;

	v = rand('uniform'); z = rand('uniform');
	if self_tested = 1 and hiv = 1 and z < prob_pos_self_test_conf and v < self_test_sens then do; 
	tested=1; tested_due_to_self_test=1;
	dt_last_test=caldate{t}; ever_tested=1; 	np_lasttest=0; newp_lasttest_tested_this_per=newp_lasttest; newp_lasttest=0;
	end;
	* note this depends on primary infection lasting 3 months - ts1m ;

[mmcleod89]

Open questions about how self testing should impact the probability of getting a clinical test in this / the next time step if the self test is negative.

[mmcleod89]

There is also a question about how to deal with people who have never tested but have had a condomless partner at some point (should they also have an increased rate of testing)

[ValentinaCambiano]

Andrew was saying, not sure whether it was during a moment in which we were having parallel conversations, that he felt there might be reasons why they have never tested. Wondering whether this should not be changed in the conventional testing for the general population but should be considered for self-testing