Not another database – lesson
learned in the development of
sustainable information systems
for HIV-related programs in
Indonesia
Suzanne Blogg
∗ Ministry of Health decentralised
∗ Some NGOs receive more support than others
∗ Some provinces more wealthy than others
∗ M&E staff often do not have qualifications
∗ High staff turnover
∗ Most do not have anti-virus subscriptions on the computers available and some use their personal computers
∗ Some staff have limited computer skills
∗ Software is often not licensed
∗ Access to internet can be limited in some areas
∗ Monitoring data – outputs
∗ Number of trainings conducted
∗ Number of people reached by programs
∗ Number of condoms, needles syringes distributed
∗ Number of people referred to health services by outreach workers
∗ Evaluation data – outcomes, impact
∗ Percentage people using condoms with casual partners or with clients
∗ Percentage injecting drug users not sharing needles and syringes
∗ Quality of services, client satisfaction with services
∗ Impact of programs on individuals
∗ Methadone registration so people can access methadone in more than one clinic
∗ Information about Needle Syringe Programs provided in Community Health Centres and by NGOs
∗ Information about NGO programs for sex workers, gay men and injecting drug users
∗ HIV treatment (ART) provided in hospitals
∗ Services provided to inmates of prisons and detention centres
∗ Minimum data
∗ Clear definitions of key indicators
∗ Only include indicators service providers can collect
∗ Excel templates
∗ One sheet for each year
∗ Disaggregated by month and gender
∗ Annual refresher training in Excel for staff in services
What has worked for Monitoring
data
∗ Complex Linux systems with passwords
∗ Monitoring databases with built in behaviour surveys with
many indicators/questions
∗ Ongoing maintenance with IT support budgets limited ∗ Regular changes to indicators being collected or changes
to lists
∗ Reliant on government departments to maintain and supervise
∗ Poor access to internet
What has not worked for Monitoring
databases
Issues Excel Spreadsheet Database
Changes to indicators Easy Difficult
IT support Minimal, low skill High level of skill
Workload Minimal Data entry for all occasions of service
Internet access Once a month Regular Budgets allocated for IT
support, virus checkers and access to internet
Minimal IT support and
maintenance can be expensive
Timeliness Usually good Dependent on adequate number of staff to enter data, usually not good Data Analysis and Reports Simple Complex but can be fast if
reports defined
Excel vs Database for Monitoring
Data
Information System thinking needed
rather than ‘database’
Data entered on form or in registration book Data entered into computer Data sent via internet Data analysed and
results disseminated Service provided Excel spreadsheet or Database Service users Service providers Manager, Funders Aggregated or Individual, monthly or daily
∗ Train-the trainer approach (with most talented partner
staff from each province)
∗ Excel data entry templates with data validation ∗ Excel pivot tables for analysis
∗ Refresher workshops with service providers every year for socialising survey approach, data entry, analysis (Excel) and presentation (PowerPoint)
∗ Provincial report writing workshops with Trainers (MS
Word)
∗ Ownership of the information which is then used for
planning
∗ Many databases last a few years (at most) without intensive support
∗ Licenses often expire when funders stop supporting programs
∗ IT support is often costly and rarely planned for in long term budgets
∗ Bugs are often not sorted out before expansion
∗ Reports are tagged on the end of the database design without much deliberation
∗ Methadone registration system (HCPI)– gone
∗ Needle syringe program database in health centres (IHPCP) – gone
∗ NGO program information (FHI) - gone
∗ Jakarta AIDS Information System (Jakarta Gvt)– gone
∗ SIHA – System Information HIV and AIDS – expanding with many issues (GF)
∗ Many program designs specify the need for a database rather than an information system
∗ Programs often have to develop a database with an IT expert (e.g. Global Fund - AIDS and TB programs),
even if not practical in the field
∗ Some databases have less than 50% compliance from service providers but the data available are reported
∗ There is an assumption that databases are superior to spreadsheets
What has not worked for Evaluation
data
∗ Software that requires an expensive license as most service providers do not have ongoing funds
∗ Each service/university/ government department have staff familiar with different statistical software
packages, e.g. STATA, SPSS
∗ Staff use statistical tests inappropriately because the tests are there
∗ Databases can be very useful tools when appropriately resourced
∗ In countries with limited resources, spreadsheets often work well for collecting monitoring data
∗ Sustainable information systems need to be specified in program designs rather than databases which are often expensive,
impractical in the field and unsustainable in the long run
∗ Any information system needs local ownership and local use of data to improve quality and to have an impact on service
provision
∗ Any system required ongoing supervision and support