THRio Outline n Data flow and database n

THRio

Outline n Data flow and database n Database matching (linkage) issues

Data flow n Data abstracted from medical charts n n n Specific forms (several) Forms are reviewed on site Forms get to our central office from Health Care Units over the mail Further inconsistencies are spotted by the DB system Double data entry

Database n Flexibility X Complexity n Access + VBA Simpler n User friendly n Not all the features we needed n

Database n SQL Server + Delphi Very flexible n Secure remote access n Simultaneous transactions n Good freeware servers (My. SQL, Postgre. SQL) n Interface with programming languages (S, Perl) n

Database n n Single server running Postgre. SQL (Linux) 5 clients connected through intranet (Windows) n n No Internet access (security concerns) Electronic forms – Delphi Allows logs for every action in the DB Double data entry analyzed within the system

Database n Problems n System getting too complex n n Will suffer major update Auditing takes a long time n Will be done in real time now We are still having bugs in the system that have to be resolved n Not all forms are ready n

Database matching n There are several systems that do not “talk” to each other SINAN – reportable diseases (TB, AIDS) n SIM – Mortality n SICOM – Pharmaceutical database (ARVs) n THRio – Our DB n n We will need to match THRio with all other 3 DBs above

Database matching n Problems There is no unique identifier common for all systems n We use name, gender and DOB as surrogates n n THRio Standardization of names abbreviations n Double data entry n Not enough – names are misspelled n n The other databases – even worse n No QC

Database matching n SICOM Monthly cumulative DB n One Excel spreadsheet for each month n Names are repeated n n No guarantee that the name will remain the same for all spreadsheets Missing DOBs n These problems are more prevalent in older spreadsheets n Hard to reconstruct ARV history for patients n

Database matching n Proposed solution n Compare different approaches Translated SOUNDEX n Reclink – probabilistic linkage n Other algorithms n n n Apply to different examples and get sensitivity/specificity for each one SICOM – it’ll be trickier Sequential matching n Match TB before doing the sequential n

DSMB Issues n TB case definition Very important n Our primary outcome n Special form to get info about Dx and Tx n n Including alternative sources (e. g. “Black book”) Discussed today in THRio meeting n Came up with a classification like this… n

Clinical presentation compatible with TB TB Classification Culture + Culture - Smear Clinical response to Tx AND no alternative Dx Smear + Definite Probable Yes Possible No No TB

DSMB Issues n Adherence evaluation n Sample of patients – standardized survey Adapted from ARV adherence survey n Already validated n n Adherence should be reported in a more timely fashion Originally, data would be collected only after course is complete n Now we will collect it whenever it is available n We’ll still have a delay n

DSMB Issues n Education level n n Years of study (self-reported) Race/ethnicity Population highly admixtured in Brazil n Very hard to objectively classify n Even the Brazilian Census Bureau uses selfdeclared race/ethnicity n Could lead to distortions n