2025/2026 SHIPP update by rtesra · Pull Request #39 · mrc-ide/naomi.utils

rtesra · 2025-11-24T12:51:13Z

Updates to SHIPP calculations after 2025/2026 survey update:

Fix for KP new infections exceeding new infections in a given age/district
Fix for PSE for KPs exceeding size of non-regular partners group
Add survey_year included in generate_shipp_tool call

Updates to SHIPP calculations to separate KPs from elevated risk group:

Adapted shipp_combine_cats_female() to replace 'elevrisk' group with 'sexnonkp and 'kp' group:

Calculate new variables for non-KP risk group (combining sexcohab and sexnonreg)
No need to create new KP group as single female KP group
Corrected denominator error - to check with @krisher1
Calculated low, mod, high and v.high incidence risk groups for:
- Non-KP total: sexcohab + sexnonreg
- Non-KP: One cohabiting partner (sexcohab)
- Non-KP: Non-regular partner
- KP: FSW (sexpaid12m)

Adapted shipp_combine_cats_male() to replace 'elevrisk' group with 'sexnonkp and 'kp' group:

Calculate new variables for non-KP risk group (combining sexcohab and sexnonreg)
Calculate new variables for KP risk group (combining pwid and msm)
Calculated low, mod, high and v.high incidence risk groups for:
- Non-KP total: sexcohab + sexnonreg
- Non-KP: One cohabiting partner (sexcohab)
- Non-KP: Non-regular partner
- KP: PWID + MSM

Add fixes for: -KP new infections exceeding new infections in a given age/district -PSE for KPs exceeding size of non-regular partners group -survey_year included in call

…nsistent formatting

jeffeaton

Thanks very much -- I've done a couple of formatting PRs, mostly for standardizing white space around operators.

See a few questions and comments; nothing substantive.

jeffeaton · 2026-01-05T09:07:26Z

NEWS.md

+  Year set to year of most recent survey with sexual risk behaviour data and to 2018 for countries where
+  most recent sexual risk behaviour survey is older that 2018.


What is the rationale for specifying this as 2018 for earlier surveys rather than the year of the survey? [Not suggesting a change, just a question for interest and clarification.]

jeffeaton · 2026-01-05T09:10:23Z

R/shipp.R

-                                            survey_year = 2018,
+                                            survey_year = 2022,


The NEWS.md refers to default survey year being set to 2018, but here in the code 2018 is changed to 2022. Is that referring to the same thing, or is this different from the changes about default year referenced in the NEWS bullet?

jeffeaton · 2026-01-05T09:13:10Z

R/shipp.R

-                                          survey_year = 2018,
+                                          survey_year = 2022,


Same as previous question about 2018 vs. 2022 in NEWS

jeffeaton · 2026-01-05T09:25:36Z

R/shipp.R

+      incidence_sexnonkp = ((incidence_sexnonreg * susceptible_sexnonreg) +
+                              (incidence_sexcohab * susceptible_sexcohab)) / susceptible_sexnonkp,


Could this be equivalently calculated as

incidence_sexnonkp = infections_sexnonkp / susceptible_sexnonkp

?

Might prefer that for clarity if that's correct

jeffeaton · 2026-01-05T09:27:05Z

R/shipp.R

+      # Non-KP total: sexcohab + sexnonreg
+      pop_low_inc_nonkp = ifelse(incidence_sexnonkp < 0.002, susceptible_sexnonkp, 0.0),
+      pop_mod_inc_nonkp = ifelse(incidence_sexnonkp >= 0.002 & incidence_sexnonkp < 0.005,
+                                 susceptible_sexnonkp, 0.0),
+      pop_high_inc_nonkp = ifelse(incidence_sexnonkp >= 0.005 & incidence_sexnonkp < 0.02,
+                                  susceptible_sexnonkp, 0.0),
+      pop_vhigh_inc_nonkp = ifelse(incidence_sexnonkp >= 0.02, susceptible_sexnonkp, 0),
+
+      # KP: FSW (sexpaid12m)
+      pop_low_inc_kp = ifelse(incidence_sexpaid12m < 0.002, susceptible_sexpaid12m, 0.0),
+      pop_mod_inc_kp = ifelse(incidence_sexpaid12m >= 0.002 & incidence_sexpaid12m < 0.005,
+                                 susceptible_sexpaid12m, 0.0),
+      pop_high_inc_kp = ifelse(incidence_sexpaid12m >= 0.005 & incidence_sexpaid12m < 0.02,
+                                  susceptible_sexpaid12m, 0.0),
+      pop_vhigh_inc_kp = ifelse(incidence_sexpaid12m >= 0.02, susceptible_sexpaid12m, 0.0),
+
+      # Non-KP: One cohabiting partner (sexcohab)
+      pop_low_inc_cohab = ifelse(incidence_sexcohab < 0.002, susceptible_sexcohab, 0.0),
+      pop_mod_inc_cohab = ifelse(incidence_sexcohab >= 0.002 & incidence_sexcohab < 0.005,
+                                 susceptible_sexcohab, 0.0),
+      pop_high_inc_cohab = ifelse(incidence_sexcohab >= 0.005 & incidence_sexcohab < 0.02,
+                                  susceptible_sexcohab, 0.0),
+      pop_vhigh_inc_cohab = ifelse(incidence_sexcohab >= 0.02, susceptible_sexcohab, 0.0),
+
+      # Non-KP: Non-regular partner
+      pop_low_inc_sexnonreg = ifelse(incidence_sexnonreg < 0.002, susceptible_sexnonreg, 0.0),
+      pop_mod_inc_sexnonreg = ifelse(incidence_sexnonreg >= 0.002 & incidence_sexnonreg < 0.005,
+                                susceptible_sexnonreg, 0.0),
+      pop_high_inc_sexnonreg = ifelse(incidence_sexnonreg >= 0.005 & incidence_sexnonreg < 0.02,
+                                 susceptible_sexnonreg, 0.0),
+      pop_vhigh_inc_sexnonreg = ifelse(incidence_sexnonreg >= 0.02, susceptible_sexnonreg, 0.0)


What is the source and importance of the 0.002, 0.005, and 0.02 magic numbers in the code block below? Are those the incidence threshold categories specified by UNAIDS?

Can we add a comment at the top of this block describing what those numbers are?
And perhaps ideally code those once to variables and reuse the variable name throughout the code below.

jeffeaton · 2026-01-05T09:27:52Z

R/shipp.R

+      incidence_sexnonkp = ((incidence_sexnonreg * susceptible_sexnonreg) +
+                              (incidence_sexcohab* susceptible_sexcohab))/susceptible_sexnonkp,


See comment above -- about whether this can be simplified using the infections_sexnonkp created on the previous line.

jeffeaton · 2026-01-05T09:28:32Z

R/shipp.R

+
+      # Create new variables for population sizes by incidence category
+
+      # Non-KP total: sexcohab + sexnonreg


See comment above about changing magic numbers to variable.

Update to sum pop groups by risk category by 5-year age band

…utils into 2025-shipp-update

rtesra · 2026-01-07T12:08:08Z

R/shipp.R

                    prev_sexnonreg = plhiv_sexnonreg/(susceptible_sexnonreg + plhiv_sexnonreg),
                    prev_sexpaid12m = plhiv_sexpaid12m/(susceptible_sexpaid12m + plhiv_sexpaid12m),
-                    rr_sexpaid12m = NA)
+                    rr_sexpaid12m = incidence_sexpaid12m/incidence_sexcohab)


@krisher1 RR calculated here for combined age categories

R/shipp.R

rtesra · 2026-01-07T12:10:41Z

R/shipp.R

@@ -1910,11 +1953,25 @@ shipp_combine_cats_female <- function(age_filter, shipp, naomi_output) {

 shipp_combine_cats_male <- function(age_filter, shipp, naomi_output) {


@krisher1 Updated to sum population groups from 5-year age bands

Updated risk population sums

-Input consensus_est = NA if not using Goals or KP Workbooks for the KP new infections

Update SHIPP code

17da6bc

Add fixes for: -KP new infections exceeding new infections in a given age/district -PSE for KPs exceeding size of non-regular partners group -survey_year included in call

rtesra changed the title ~~Update SHIPP template with 2025 survey update~~ Fixes for 2025 survey update Nov 24, 2025

rtesra added 6 commits November 24, 2025 23:53

bump version

a6f5622

Merge branch 'master' into 2025-shipp-update

e14ed82

fix spelling

e505448

spelling error

282e70a

Separate KP from elevated risk group

a0280eb

replace empty columns in male PSE with placeholder names to ensure co…

cb2f728

…nsistent formatting

rtesra changed the title ~~Fixes for 2025 survey update~~ 2025/2026 SHIPP update Jan 2, 2026

update survey year to read automatically in internal calculations

2a00d64

rtesra requested review from jeffeaton, krisher1 and r-ash January 5, 2026 07:26

jeffeaton added 5 commits January 5, 2026 09:06

Edits to NEWS.md for clarity

4e27ee3

update white space to match @r-ash style preference

a0a6a44

update white space in condition to @r-ash style preference

877a9dc

update condition white space formatting to @r-ash style preference

19959fc

formatting: add white spaces round operators

2eb6ff5

jeffeaton reviewed Jan 5, 2026

View reviewed changes

rtesra added 2 commits January 7, 2026 11:48

Update shipp.R

f46c493

Update to sum pop groups by risk category by 5-year age band

Merge branch '2025-shipp-update' of https://github.com/mrc-ide/naomi.…

0934bed

…utils into 2025-shipp-update

rtesra commented Jan 7, 2026

View reviewed changes

R/shipp.R Show resolved Hide resolved

rtesra commented Jan 7, 2026

View reviewed changes

rtesra and others added 2 commits January 8, 2026 12:16

Update shipp.R

69af37e

Updated risk population sums

Add option to not use Goals

84c9da1

-Input consensus_est = NA if not using Goals or KP Workbooks for the KP new infections

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Comments

2025/2026 SHIPP update#39

2025/2026 SHIPP update#39
rtesra wants to merge 17 commits intomasterfrom
2025-shipp-update

rtesra commented Nov 24, 2025 •

edited

Loading

Uh oh!

jeffeaton left a comment

Uh oh!

jeffeaton Jan 5, 2026

Uh oh!

jeffeaton Jan 5, 2026

Uh oh!

jeffeaton Jan 5, 2026

Uh oh!

jeffeaton Jan 5, 2026

Uh oh!

jeffeaton Jan 5, 2026

Uh oh!

jeffeaton Jan 5, 2026

Uh oh!

jeffeaton Jan 5, 2026

Uh oh!

rtesra Jan 7, 2026

Uh oh!

Uh oh!

rtesra Jan 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

		Year set to year of most recent survey with sexual risk behaviour data and to 2018 for countries where
		most recent sexual risk behaviour survey is older that 2018.

		incidence_sexnonkp = ((incidence_sexnonreg * susceptible_sexnonreg) +
		(incidence_sexcohab * susceptible_sexcohab)) / susceptible_sexnonkp,

		incidence_sexnonkp = ((incidence_sexnonreg * susceptible_sexnonreg) +
		(incidence_sexcohab* susceptible_sexcohab))/susceptible_sexnonkp,


		# Create new variables for population sizes by incidence category

		# Non-KP total: sexcohab + sexnonreg

		@@ -1910,11 +1953,25 @@ shipp_combine_cats_female <- function(age_filter, shipp, naomi_output) {

		shipp_combine_cats_male <- function(age_filter, shipp, naomi_output) {

Comments

Conversation

rtesra commented Nov 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jeffeaton left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

rtesra commented Nov 24, 2025 •

edited

Loading