As an epidemiologist trained in health systems research, I spend much of my time with quantitative data—numbers, statistics, writing code in SAS or R. However, my favorite projects are those which are a jumble of data and require making sense of a heap of numbers before we can even begin the analysis. My background is in global health, specifically in using service numbers collected by governments to document the number of diagnoses made at community health centers and hospitals throughout the country. In those databases, it is common to have missing data for a few months—maybe even a year—at a time.
Drawing on this background, my favorite ICH projects are those which are rich in data, but only when we do some digging to understand what we have. The challenges have been varied. We’ve worked with:
- Data from a 14-year period with each year stored in a different database;
- Data collected by different people over time and documented in different ways;
- Changing variable names and definitions;
- Surveys with low response rates;
- Organizational data with changing metrics and variability across departments;
- Data stored in Access, Excel, SQL, online databases, and on paper;
- Changing electronic medical records databases.
Each data challenge has its own unique solution and determining how to use the data is part of the fun. My work at ICH has involved large and small databases, each of which has its own compelling story to tell.
Last year, we assisted the Mental Health First Aid program at Cambridge Public Health Department in managing their registration data. Mental Health First Aid educates community members about how to support people who are experiencing a mental health or substance-use related crisis. The program had registration data for hundreds of trainees, but data were stored in numerous files, there was no way to automatically know when people were due for a refresher course, and reporting on numbers trained was an onerous challenge. Our support included:
- Developing an online registration form to streamline data collection
- Building a registration database with fields which automatically flag participants who did not complete the initial training or who are eligible for refresher training
- Creating reports to make monthly, quarterly, and annual reporting a streamlined process.
By creating a user-friendly database that meets the needs of the program, we were able to support the data collection process and pave the way for future evaluation work.