Error in Stats reports after probabilistic analysis
Greetings,
I am conducting a probabilistic simple analysis where rewards are conducted with QALYs (Treeage 2016, Mac OS X 10.11.5). After completing the analysis, I'd like to review statistics related to my Incremental EV (Histograms> Output Distributions> Incremental EV > AvsB). The subsequent histogram is reported just fine, but when I click on Stats Report for that histogram, I am receiving the following error, with no data output:
Could not open the editor: org.eclipse.birt.data.engine.core.DataException: Failed to prepare the query execution for the data set: monteCarloDataSet
Failed to prepare the following query for the data set type (org.eclipse.birt.report.data.oda.jdbc.JdbcSelectDataSet).
[select STATS.STRATEGY, STATS.ATTRIBUTENAME as Attribute, STATS.STATISTIC, STATS.VALUE from (EVSTATS as STATS inner join STATISTICSINDEXES as IX on STATS.STATISTIC = IX.STATISTIC) inner join STRATEGIES AS S on S.strategy=STATS.Strategy where S.INDEX=1_VS_2_NOT and UPPER(STATS.ATTRIBUTENAME)='CETYPE' order by IX.INDEX, ?<>'' ]
org.eclipse.birt.report.data.oda.jdbc.JDBCException: Error preparing SQL statement.
SQL error #1:Table "STRATEGIES" not found; SQL statement:
select STATS.STRATEGY, STATS.ATTRIBUTENAME as Attribute, STATS.STATISTIC, STATS.VALUE from (EVSTATS as STATS inner join STATISTICSINDEXES as IX on STATS.STATISTIC = IX.STATISTIC) inner join STRATEGIES AS S on S.strategy=STATS.Strategy where S.INDEX=1_VS_2_NOT and UPPER(STATS.ATTRIBUTENAME)='CETYPE' order by IX.INDEX, ?<>'' [42102-181]
;
org.h2.jdbc.JdbcSQLException: Table "STRATEGIES" not found; SQL statement:
select STATS.STRATEGY, STATS.ATTRIBUTENAME as Attribute, STATS.STATISTIC, STATS.VALUE from (EVSTATS as STATS inner join STATISTICSINDEXES as IX on STATS.STATISTIC = IX.STATISTIC) inner join STRATEGIES AS S on S.strategy=STATS.Strategy where S.INDEX=1_VS_2_NOT and UPPER(STATS.ATTRIBUTENAME)='CETYPE' order by IX.INDEX, ?<>'' [42102-181
Any thoughts on this error?
Comments
Frank,
This is a known bug in TreeAge Pro. It is scheduled to be fixed in TreeAge Pro 2017, R2 which is scheduled for release in mid July.
Note that you should be able to generate the graph immediately after running the PSA analysis. The bug occurs when you save the PSA analysis output in an RPTX file, then reopen it later to generate that histogram.
As a workaround, you may want to turn on seeding, so every analysis generates identical results. That way if you need to rerun the PSA to get this graph, you can guarantee the same data in the new result set.
Sincerely,
Andrew
P.S., I have added your contact information to the bug to remind me to notify you when the fix is available.
Andrew,
Thanks for the prompt reply. Will await 2017.R2. Beyond the graph itself, is there a way to generate the numeric stat report of the Incremental Effectiveness (Mean IE, SD, CIs etc for the delta between two treatment options)
Thanks,
Frank
That is the only way to generate the graph and aggregated statistics on incremental effectiveness. Note that you can generate this graph and its statistical report if you rerun the PSA analysis and generate the graph before saving and closing the overall PSA results.
Andrew
You could also dump the All Data Report to Excel, create a column for the incremental effectiveness, then analyze those incremental values.
That has been my workaround, dump to .csv, import to Stata, generate the IE, then calculate mean and 95%CIs on the generated IE
Good.
Note that I would not formally call this a "confidence interval". It is more that 95% of the incremental effectiveness values generated by the model fall within a range.
Andrew
Please sign in to leave a comment.