Back pain management: a personalised smartphone-based solution
Author
Prof. Walter Karlen
Universität Ulm
Interview with the principal investigator of this NRP75 project.
What was the aim of your project “Smartphones to monitor and understand the burden of low back pain”?
The goal of this project was to bring together computer scientists, mobile health specialists, and application experts from industry, medicine, and physiotherapy to investigate low back pain (LBP) from a data science perspective.
Results of the project?
We built the smartphone app “Swiss Health Challenge” which allowed to collect, transmit and store anonymised sensor data for research. The app was designed to engage with the user in the form of so-called challenges. It was used in two clinical studies where former patients and persons at risk for LBP participated.
The first study provided new insights about the use of digital tools for treatment support. We complemented the app with an exergame that provided interactive exercise sessions at LBP patients´ homes. The use of this exergaming intervention did not affect the movement patterns, which were assessed at regular intervals at the physiotherapy unit of the University Hospital Zurich. This result confirms that a patient group who is not severely affected by LBP is difficult to monitor and treat. However, the study stands out from others in the use of digital tools for objectively monitoring the adherence in real-life settings. Also, we gained new insights on how fear of movement is related to postural sway in LBP.
In the second study adolescent skiers were monitored to identify physical stress that could potentially lead to back injury. The app was used to monitor ski training with altitude and movement sensors. It could be shown that this is valuable data that can provide far more detailed information about exercise behaviour than self-reports.
Finally, we also developed numerous machine learning and technical contributions to solve challenges with the analysis and challenges of such type of data. These have been described in established engineering and computer science journals and international conferences. The source code of these approaches is openly accessible and can be easily adopted for other health applications.
What are the main messages of the project?
- Technology can promote but not substitute user engagement and motivation.
In both of our studies we used digital technologies to engage with users and attempted to reduce the overall effort needed to accomplish specific tasks. We observed that independent of technology the motivation of the user is essential to achieve these tasks. Additional guidance or more enjoyable interventions were not success factors for better engagement if the users were not inherently motivated. For example, in the randomized controlled trial on digital LBP intervention with exergaming we saw that some participants did not follow the recommended exercising opportunity sufficiently that would likely have led to a change in outcomes. Similarly, monitoring the training efforts of the adolescent skiers was dependent on adhering to carrying the smartphone with them. Additional features and just-in-time reminders could address adherence challenges in the real-world settings but require in-depth understanding of user needs and motivations. Thus, we still see a lot of not yet fully leveraged potential in promoting the use of sensor-based technology for people with LBP. - Access to health data at larger scale, especially prospectively collected outcome data, remains challenging.
Currently, infrastructure and data silos still dominate the Swiss health care landscape. Access to data is challenging and can only be overcome with new approaches of collecting and managing data. Due to high heterogeneity in health systems and regulations, Switzerland faces significant challenges to bring various data sources together. The prospective collection of labelled health data within studies is costly and a lot of resources must go in recruiting and managing regulatory processes of studies. If the data management and infrastructure could be harmonised and processes unified, data collection directly from patients and citizens could overcome these barriers. The health care system would need to be designed for more transparent data flows and a more open culture. However, our experience was that there is an increase of mistrust into governmental approaches and the delegation to private institutions (i.e Swiss ID, SwissCovid App etc). We experienced low recruitment rates and the protocols caused high burden and effort for the study participants. Therefore, simpler access to studies, with open platforms to connect patients with researchers might increase data accesses for deeper health research. - Low back pain remains a complex and poorly understood disease.
Our research could only shine light on small aspects of rehabilitation and roles of digital sensors and intervention for supporting LBP management. Larger, coordinated initiatives will be needed to combine further information from other fields such as genetics, epidemiology, and musculoskeletal rehabilitation.
Does your project have any scientific implications?
The results on adherence during the randomised controlled trial highlight the urgent need to improve adherence to interventions for patients with LBP. If users do not stick to an intervention until possible benefits can be perceived, optimisation of the exercises themselves may be of little help. Therefore, the promotion of adherence, usability and removal of barriers from accessing the intervention should have priority.
Digital tools offer unique opportunities for integrating features such as automated reminders, scheduling options or building other structures that can be leveraged to promote use of interventions. In our randomised controlled trial, the “effort” was frequently mentioned as a reason for drop out. This “effort” always needs to be seen in relation to a perceived benefit. Therefore, it is essential that the intervention character and the potential clinical value of the tools used are highlighted and clearly communicated, while access, usability and motivation are improved through different features.
Highlighting the rationale and purpose to improve LBP may be central to promote adherence, especially for interventions that are presented as a game. The basis for such communication could be established through studies under narrowly defined, optimal conditions. Results obtained under these conditions could then be used as a basis to increase the uptake and motivation also under more realistic, applied conditions in the field.
Does your project have any policy recommendations?
Population wide data collection specific to a particular disease remains challenging within Switzerland. The traditional view of a “Study Centre” that is linked to a hospital or a Canton is no longer suitable for big data research in medicine. Only if all medical centres across Switzerland can connect their data hubs in a collaborative and open way, meaningful data sizes that allow real insights into complex diseases structures will be possible. There is urgent pressure to achieve this at a more global scale as individual centres are unable to compete against the dominant commercial “data collectors” (Big 5). Clarification of legal frameworks such as digital signatures and consenting for medical research, as well as better informed regulatory bodies such as Ethics Commissions will also be needed.
About the team
The multidisciplinary research team driving this project is complemented by Dr Anita Meinke and Dr Patrick Schwab of ETH Zürich, Dr Jaap Svanenburg and Dr Rudolf Knols of University Hospital Zürich, PD Dr Jörg Spörri of University Hospital Balgrist, and Dr Lars Lünenburger of Hocoma AG. Prof. Robert Riener, ETH Zürich provided administrative support towards the end of the project.