Biostatistics, bioinformatics and other methodological support in clinical research projects

Inventory of needs and existing resources as well as a model for structured support via Clinicum.

Report from the Clinicum project group, Karolinska Institutet, January 2022.
By Sandra Eloranta, Docent, Biostatisticians
.

Report summary

In May 2021, Clinicum’s project group was formed with the task to investigate existing support structures for clinical research, and to develop a model that may be used to shape a new organization for structured research  support, easier access to health care data and promotion of collaboration at KI and in Region Stockholm. This report describes the unmet needs for methodological support, the mapping of existing support structures at KI and with the care providers in the Region, as well as the proposed model for methodological support via Clinicum. The project group’s work has included: 

  1. To map similar initiatives abroad and in Sweden.
  2. To survey the self-assessed need for support in biostatistics, bioinformatics and data management among clinical researchers at KI and within the Region.
  3. To take inventory of the most important existing operations that currently provide methodological support at KI and at the various healthcare providers.
  4. Identify the needs of clinical researchers who currently lack satisfactory solutions.
  5. Propose principles for structured method support through a future Clinicum organization.

The results of a survey answered by over 500 researchers showed a great need for support that can be filled through strategic collaborations but also to a large extent through consultations, supervision and training and new  recruitment. The results of the inventory of existing support functions showed that the competence that is requested by researchers is, to some extent, already existing in our organizations, but the availability and existing support is fragmented and lacking entirely for certain identified advanced methodological areas. A survey of international and national initiatives (Harvard University, Mayo Clinic, University of Copenhagen, Michigan Institute for Clinical and Health Research and Sahlgrenska Academy, University of Gothenburg) showed a number of well-functioning guiding principles for methodological support at medical faculties. 

The proposed model aims to supplement the existing analytics support in existing core facilities and other support functions with a virtually organized network of methods specialists who can provide advice for early phase projects, or consultation of another type that is  clearly limited in scope (eg review of abstract, manuscripts, help with reference comments, etc.). The subsidised consultations are also planned to provide information about other methodological support functions that are  available for researchers who need additional help, for example with hands-on analyses. The model provided via Clinicum should be centrally funded to lower the thresholds for researchers to use it. The proposed model is also  scalable in terms of scope and specialist areas. Areas that may be included in addition to biostatistics, bioinformatics and AI are advice when ordering health data or regarding practical data management. 

Below is summarized the unmet needs that form the basis for the presented model for method support, as well as the project group’s planned work for 2022, which aims to enable a launch of the model within the Clinicum organization. As part of this work, the project team will investigate practical details of how the model for methodological support should be further developed and designed, including budget and infrastructure. A final report regarding methodological support, as well as a proposal for the establishment of the Clinicum organization is planned for 2022. 

Inventory of needs, and planned continued work

Needs identified in 2021:

Consultations

  • Support in biostatistics can be found at Campus Solna, Flemingsberg, Södersjukhuset and Danderyd Hospital, but to different levels of subsidy. 
  • Support in bioinformatics is available via NBIS (SciLifelab) but also via CBB (for researchers in Huddinge).
  • Support in data management is available to a limited extent via Biostatistics Core in Solna and via NBIS. 
  • AI/machine learning support is missing.
  • Subsidized advice services available regardless of organisational affiliation are missing.
  • Procurement for external statistical services is missing.

Research environment

  • For clinical researchers:
    • Forum for scientific discussion, further education in statistics and support in guidance on methodological issues need to be developed.
    • Knowledge of existing method activities and support features can be improved.
  • For methods experts:
    • In several places, methodological experts often work alone or in small groups and without much context linked to their own discipline, which is negative from a recruitment and retention point of view.
    • Clear career ladders and professional development plans are lacking.

Activitites planned for 2022:

Consultations

  • Define the target audience for support.
  • Investigate possible method profiles for the nodes. 
  • Develop a model for organizing local and central support.
  • Investigate how increased collaboration between nodes can be achieved.
  • Identify method support contact persons in each node.
  • Details needs in terms of specialist competencies and required FTEs.
  • Suggest a budget for method support. 
  • Identify key performance indicators for evaluation.

Research environment

  • For clinical researchers:
    • Kick-off for Clinicum and show case of activities offering methodological support arranged in February.
    • Method seminars planned at CBB and in Solna.
    • Increased communication that makes Clinicum and current activities visible.
  • For methods experts:
    • Regular network meetings for biostatisticians, bioinformaticians, data scientists are planned jointly with representatives from several KI departments.
    • Multiple training activities are planned.