SSD 2016-20

[QLLZ]

Data pull: HASH: 6772ea9da4f43ddc3da11c494a573af2ca3dd3c8 config

[QLLZ]

Model run: HASH: 6772ea9da4f43ddc3da11c494a573af2ca3dd3c8

[javierps]

All ok except for slight geometrical pattern in NE.

Suggestion: check adj, then accept.

[eclee25]

All diagnostics ok. All of the maps appear to have some geometrical patterns in my view, so it may be that this is something that needs to be dealt with more systemically.

Opinion: Approve

[QLLZ]

Config

Data pull:

HASH: 211913e0eb7d570ed62c48bc1ae9ecd6671a6cbc

[QLLZ]

Failed stan model run log file

[QLLZ]

Rerun on dev_u_combs_fix

HASH: e56580fa5ddab00293e31ff90139351f80ae3f6c

[QLLZ]

country data report

[QLLZ]

Overestimated population, however, incidence 2016-2017 is also super high. Based on the observation patterns in 2017: maybe considering increasing the censoring threshold? Convergency looks good.

[eclee25]

Model diags look good but some OCs need auditing -- OC 20376 has some very large discrepancies -- may want to verify that these data are correct. Population is highly overestimated - adjust in post-processing

Opinion May need to rerun if data are audited or increase censoring threshold

[QLLZ]

Double checked that the cases match with the source doc (probably some reporting issues in those reports)

[javierps]

The model is doing well here I think.

[javierps]

Sep 2023 Production run: all OK.

Suggestion: accept.

[eclee25]

2017 is systematically overestimated, despite the higher od_param 1. It looks like the censored 2017 observations have sCh > than the full annual observations.

Approve.