CIV 2011-15

[QLLZ]

Data pull: HASH: 6772ea9da4f43ddc3da11c494a573af2ca3dd3c8 config

[QLLZ]

Model run: HASH: 6772ea9da4f43ddc3da11c494a573af2ca3dd3c8

[QLLZ]

Country data report

[eclee25]

Convergence is pretty good, observation and grid fits seem good, some Rhat issues but nothing major. GAM input and modeled estimates look similar.

Opinion: Approve

[QLLZ]

Config

Data pull & Model run:

HASH: 211913e0eb7d570ed62c48bc1ae9ecd6671a6cbc

[QLLZ]

failed country data report log file

[QLLZ]

country data report

[QLLZ]

cases are overestimated in 2012 (this may be due to the conflicted observations in 2012). There's a hole in hte population raster (which seems to be a national park, this might be the reason). Others look good except for std_dev_w in figure 26.

Suggestion: double checked OCs for 2012 (OC306 and OC901) in figure 23.

[QLLZ]

WHO annual report was under estimated (OC 306). The other source was promed-mail (a local cholera outbreak was reported).

[eclee25]

2012 censored observation seems to push the 2012 annual estimate higher -- and this seems valid after looking at the OCs. Agree with Qulu's assessment here. sd_w traceplot has some instability but does not seem to have major impact on w traceplots. Other diagnostics look okay. Standard model is probably okay but there is a case that could be made to use a min_subnat model.

Opinion: Temp Approve (come back after min_subnat settings are decided)

[eclee25]

Unstable rho and sd_w that trickles down to w traceplots, minimal subnational data. High sd_w value. Try no-mixture sd_w prior setting

[QLLZ]

Data pull, model run:

HASH: 05f598b95b5d42ac0049963f8a0b7e04d6f1dbd7

[javierps]

Rho convergence issues, std_w and ws OK. Rhats OK.

Suggestion: accept.

[eclee25]

Agree with J. 2012 estimate seems high relative to WHO, but there is a censored observation that supports the higher estimate. Approve.