Long-term partner infection #186
Conversation
|
Hi @mmcleod89 there is still quite a bit of work to finish this up but if you could have a look and let me know what you think, that would be great! |
dimitrasal
changed the title
src/hivpy/hiv_status.py
Outdated
| subgroup) | ||
|
|
||
| people_with_ltp = population.get_sub_pop([(col.LONG_TERM_PARTNER, op.eq, True)]) | ||
| population.transform_group([col.VIRAL_SUPPRESSION, col.LTP_STATUS], |
There was a problem hiding this comment.
Instead of col.LTP_STATUS we might want to use a LTP_INFECTION_DATE or something similar to track whether someone is in primary infection.
In general we don't want to transform_group by dates so we can have bool column LTP_PRIMARY_INFECTION which before transform_group is set by LTP_PRIMARY_INFECTION = LTP_INFECTION_DATE < (date - length of primary)
There was a problem hiding this comment.
@mmcleod89 thanks, following your suggestions I transformed this to ltp_primary_infection = ltp_infection_date > (population.date - timedelta(days=90)) assuming that 90 days is the timestep length in agreement with current timestep of ca.3 months in SAS.
Add balancing for concordant couples
src/hivpy/hiv_status.py
Outdated
| def prob_of_new_ltp_already_infected(self, population: Population): | ||
| for sex in [SexType.Male, SexType.Female]: | ||
| for age_group in range(5): | ||
| opposite_sex = population.get_sub_pop([(col.SEX, op.ne, sex), | ||
| (col.AGE_GROUP, op.eq, age_group)]) | ||
| opposite_sex_with_hiv = population.get_sub_pop([(col.SEX, op.ne, sex), | ||
| (col.AGE_GROUP, op.eq, age_group), | ||
| (col.HIV_STATUS, op.eq, True)]) | ||
|
|
||
| if len(opposite_sex) != 0: | ||
| self.prevalence[sex][age_group] = len(opposite_sex_with_hiv) / len(opposite_sex) | ||
|
|
||
| def calculate_new_ltp_infection(sex, age_group, size): | ||
| infected = (rng.uniform(0, 1, size) < 0.5) | (rng.uniform(0, 1, size) < | ||
| self.prevalence[sex][age_group]) | ||
| return infected | ||
|
|
||
| people_with_new_ltp = population.get_sub_pop([(col.LTP_NEW, op.eq, True)]) | ||
|
|
||
| new_ltp_infected = population.transform_group([col.SEX, col.AGE_GROUP], | ||
| calculate_new_ltp_infection, | ||
| use_size=True, | ||
| sub_pop=people_with_new_ltp) | ||
| population.set_present_variable(col.LTP_STATUS, new_ltp_infected, people_with_new_ltp) | ||
|
|
There was a problem hiding this comment.
There are a lot of changes to go through here, but I think the general logic looks sound. There are a few typos to fix, variables to add, and some documentation that I would include, but there don't seem to be any glaring issues that I could find.
Don't forget to add ltp HIV transmission to update_HIV_status when it's done.
src/hivpy/hiv_status.py
Outdated
| # 3% chance that virally suppressed person becomes un-suppressed | ||
| remaining_suppressed = rng.uniform(size=len(viral_suppressed_ltp)) < 0.97 |
There was a problem hiding this comment.
This probability of un-suppression seems like a good idea to turn into a data variable.
There was a problem hiding this comment.
Have put this into a class variable but there isn't an HIV status data file yet so I'd rather create one in a separate PR than make this one bigger
src/hivpy/hiv_status.py
Outdated
| # 2% chance that LTP on ART go off ART | ||
| continuing_ART = rng.uniform(size=len(ltp_on_ART)) < 0.98 |
There was a problem hiding this comment.
Here as well – this would also be good to turn into a data variable.
Co-authored-by: Emily Dubrovska <[email protected]>
Co-authored-by: Emily Dubrovska <[email protected]>
No description provided.