The State and Profile of Open Source Software Projects in health and medical informatics

Balaji Janamanchi, Evangelos Katsamakas, Wullianallur Raghupathi, Wei GaoInternational Journal of Medical Informatics 2009 · DOI

TL;DR

This paper provides a comprehensive analysis of open-source software (OSS) projects in health and medical informatics, revealing a growing and diverse ecosystem. It highlights the increasing sponsorship of these projects by organizations, the prevalence of restrictive licenses, and the distinct profiles of OSS applications tailored to the healthcare domain, suggesting OSS is a viable and evolving approach to developing health IT solutions.

What Problem Does This Paper Address?

The field of health and medical informatics has seen a surge in interest in open-source software (OSS) as a potential solution for lowering costs, fostering innovation, and improving interoperability. However, despite this growing attention, there was a significant lack of systematic research detailing the landscape of OSS projects specifically within this domain. Researchers and practitioners lacked a clear understanding of the types of health and medical informatics OSS projects available, their development patterns, their characteristics, and their intended audiences. This paper aims to fill this knowledge gap by providing a detailed profile and analysis of these projects.

Methods

The researchers conducted a systematic search of the SourceForge repository, a large platform for open-source projects, between May 1st and May 15th, 2007. They used keywords related to health, medical, and bioinformatics to identify relevant projects. After an initial search yielded 607 projects, they refined the sample by excluding indirectly related projects and those not considered typical of health and medical informatics, resulting in a final sample of 174 projects.

To extract data for key variables, they developed a Java-based parser. The study then employed descriptive statistical analysis, including the generation of visually descriptive statistics and charts, to analyze the profiles of these OSS projects. Key variables analyzed included:

  • Sponsorship: Whether a project was initiated or actively supported by a healthcare organization or a firm providing health-related software.
  • Project Rank: A metric from SourceForge reflecting traffic, communication, and development activity.
  • License Type: Categorized as restrictive (e.g., GPL, LGPL) or non-restrictive (e.g., BSD).
  • Downloads: A proxy for project usage and popularity.
  • Development Status: The stage of the software development lifecycle (e.g., pre-planning, alpha, beta, mature).
  • Technological Environment: Including programming language, operating system, and database environment.

Datasets

The primary dataset for this study was derived from SourceForge.net. The researchers collected data on 174 open-source projects related to health and medical informatics identified on the platform during a specific two-week period in May 2007. No new benchmarks were introduced; the study focused on analyzing existing project data available on SourceForge.

Key Insights and Ideas

  • Growing Organizational Interest: The significant proportion of sponsored projects (40%) indicates a rising interest and investment in OSS by healthcare organizations, suggesting a move beyond just academic or individual developer interest.
  • Restrictive Licenses as a Trend: The prevalence of restrictive licenses (nearly two-thirds of projects) is a key finding. This suggests a strategic choice by development teams to ensure projects remain open-source and to incentivize continued contributions by protecting against unscrupulous commercialization.
  • Distinct Healthcare OSS Profile: The analysis reveals that OSS in health and medical informatics has unique characteristics compared to OSS in other domains, with specific application areas and audience focuses.
  • Sponsorship Matters for Success: The study hints that sponsorship has a significant impact on project success metrics, likely by providing resources and signaling commitment.
  • Developer-Centric Infrastructure: SourceForge, as the data source, inherently favors projects focused on development infrastructure, meaning the study might reflect developers' perspectives more than end-users'.

Key Results

  • Sponsorship: 64% of the projects had a sponsor, indicating growing organizational engagement.
  • License Type: Approximately two-thirds of the projects used restrictive licenses.
  • Project Diversity: A wide range of OSS applications were found, spanning bio-informatics, hospital information systems, data formats, and medical science applications.
  • Target Audience: Projects targeting "industry, scientific organizations and research" and "developers" were prominent. Advanced end-user projects were concentrated in bio-informatics, databases, and medical science.
  • Programming Languages: Java was the most popular language, used by 58% of projects, followed by a combination of languages.
  • Database Environment: A majority of projects (60 out of 73 with specified environments) used an open-source database environment.
  • Top Projects: The top-ranked project by downloads was OsiriX, a 3D DICOM medical viewer. Many of the top projects were sponsored and in a "production/stable" or "mature" development status.

Limitations and Future Work

Limitations:

  • Snapshot in Time: The data was collected on a specific date, making it a snapshot and precluding long-term trend analysis. Projects are frequently added or deleted from repositories.
  • SourceForge Bias: The study relied solely on SourceForge, potentially underrepresenting the total number of health and medical informatics OSS applications, as many projects may be hosted elsewhere or have their own developer sites.
  • Data Quality: Project listings on SourceForge can be outdated or contain missing information.
  • Developer-Centric Focus: The reliance on SourceForge as an infrastructure for development may skew the results towards a developer's perspective rather than an end-user's.
  • Unspecified Variables: A significant percentage of projects had unspecified data for variables like "development status" and "operating system."
  • Compliance Not Assessed: The study did not assess compliance with regulations like HIPAA or standards like HL7.

Future Work:

The authors suggest several avenues for future research:

  • OSS Acceptance and Diffusion: Investigating how OSS is adopted and spreads within healthcare organizations.
  • Impact on Cost, Efficiency, and Quality: Quantifying the actual impact of OSS on healthcare delivery.
  • User Perspective: Conducting surveys of user organizations to gather their insights and experiences.
  • Technological and Cultural Barriers: Understanding the challenges to OSS adoption in typical US healthcare practices.
  • Interoperability Standards: Exploring the role of standards in facilitating OSS adoption.
  • Predictors of OSS Success: Identifying factors that contribute to the success of OSS projects in this domain.

Summary updated

BibTeX
@article{janamanchi2009state,
  author = {Balaji Janamanchi and Evangelos Katsamakas and Wullianallur Raghupathi and Wei Gao},
  title = {{The State and Profile of Open Source Software Projects in health and medical informatics}},
  journal = {International Journal of Medical Informatics},
  year = {2009},
  doi = {10.1016/j.ijmedinf.2009.02.006},
}
Figure