read

Leveraging Patient-Focused Drug Development by Social Media Patient Listening Studies

By Semalytix
Philipp Cimiano (Chief Technology Officer)
Matthias Hartung (Chief Research Officer)
Jens Höwelkröger (Product Manager)
Janik Jaskolski (Chief Executive Officer)

 

Patients are experts in their own experience of a disease or condition. They are also the end consumers of medical products. Therefore, the collection of patient experience data is a key activity in patient-focused drug development, as it provides an opportunity to inform and enhance decision making in pharmaceutical companies, regulatory agencies, and payers to better address patients’ needs.

According to the FDA´s definition[1], patient experience data consists of “data intended to provide information about patients’ experiences with a disease or condition, including the impact of a disease or condition, or a related therapy or clinical investigation on patients’ lives and patient preferences with respect to treatment of their disease or condition.”

These patient experiences include:

  • the symptoms of their condition
  • the impact of the condition on their functioning and quality of life
  • their experience with treatments
  • input on which outcomes are important to them
  • patient preferences for outcomes and treatments

Patient experience data can be used to identify unmet needs in coping with a disease: What matters most to patients, and which clinical outcomes have the greatest potential to improve their daily lives? From this perspective, patient experience data can inform clinical trial design, trial endpoint selection, regulatory reviews including benefit-risk assessments as well as labeling claims. 

In their series of draft guidelines, the FDA has set out to provide recommendations on how stakeholders can collect and submit patient experience data and other relevant information from patients and caregivers. The purpose of the first guideline is to discuss sampling methods for collecting information on the patient experience that is representative of the intended population and can inform the development and evaluation of medical products throughout the medical product lifecycle.

As with any research method, it is vital to define the target population, the research objective and question, as well as the methods used to interpret the data before starting the data collection.  The target population is defined as the group of patients for which experience data shall be collected. The FDA recommends that, ideally, experience directly reported by patients should be considered.

Regarding the methods for collecting and analyzing patient experience data, the FDA distinguishes qualitative, quantitative and mixed research methods.  Qualitative methods generally serve to generate in-depth information about the experiences, perspectives, and feelings of patients, their relatives and caregivers in their own words. In contrast, quantitative methods aim at discovering statistical patterns in the data in terms of frequencies, distributions and correlations between continuous and categorical variables, which are summarized and visualized using tables or graphs.  Mixed methods combine qualitative and quantitative methods. Besides traditional research approaches including interviews, focus groups and questionnaires, the FDA also explicitly mentions the option of collecting data from social media, particularly in the form of „spontaneous online communities and social media data“.

In response to these preliminary guidelines on how to collect patient experience data, Semalytix has developed Pharos® Pharma Analytics as a platform to conduct patient listening studies based on a mixed-methods observational approach that extracts patient experience data from social media (fora, blogs, Twitter, etc.) using cutting edge technology from artificial intelligence, machine learning and natural language processing.

Pharos® relies on a knowledge-based iterative bootstrapping technique to automatically find fora where patients suffering from a given condition are active. To avoid the early introduction of biases, the data from selected fora is collected in its entirety. Subsequently, the data is automatically analyzed using proprietary Semalytix technology in order to extract patient-reported insights regarding i) the impact of the condition on their functioning and quality of life, ii) disease burdens, as well as iii) treatment burdens and preferences.  

The main advantage of the method is that it overcomes biases related to sampling as well as the research method, and the data collection approach adopted. The FDA distinguishes probabilistic sampling approaches on the one hand (including simple random sampling, stratified sampling, and cluster sampling) as well as non-probabilistic sampling methods on the other hand (including convenience sampling, purposive sampling, quota sampling, and respondent-driven sampling).

The data collection method applied by Semalytix can be described as a type of purposive sampling in the sense that it allows for selecting a target population with specific characteristics. By downloading the data from complete fora without any further filtering or selection, the method minimizes sampling bias and does not distort the distribution of the characteristics of the whole target population as represented in a given forum. The method of listening to unsolicited patient experience as voluntarily and openly shared on the Web overcomes biases inherent in other data collection methods based on interviews or focus groups. However, while this method is a fast way of gathering patient experience and getting deep patient experience insights due to a high level of self-exposure, the use of spontaneous online communications comes with its own biases.

Pharos provides automated classification of users of a forum into relevant author groups (patients, relatives, family members, caregivers, etc.). The patient experiences are summarized in easy-to-understand and user-interactive dashboards that allow drill-downs into specific aspects of the patient experience and reality.

With this method, Semalytix can compile anonymized experience data from patient populations of interest for any indication comprising the trajectories of thousands of patients.  Users of the platform can drill down and filter the data according to demographic variables, geography, and time period, effectively “zooming” into a target study population with specific characteristics.

The method developed by Semalytix combines the benefits of quantitative and qualitative methods.  By summarizing the patient experience in interactive dashboards, it provides an in-depth understanding of the (unmet) needs, burdens and treatment experiences of patients, in their own words. This level of analytical and qualitative depth, while at the same time striving for representativity in terms of population size but also in terms of ensuring concept saturation, is not provided by any existing method.  At the same time, the number of patients reporting a certain need, or complaining about a certain symptom, or mentioning an issue affecting their quality of life can be quantified and correlated with other variables of key interest, thereby also providing quantification for the qualitative findings.

If the results of patient experience gathering methods are used to substantiate the design of clinical studies and/or their endpoints, then transparent documentation of  the interpretation of the data supporting these decisions is crucial to meet regulators’ requirements and to justify the inclusion of endpoints or the design of a PRO research instrument. Semalytix is preparing for these requirements and is ensuring that the results of analytical methods are documented to allow for conclusions and interpretations to be traceable down to every single data point.  Semalytix ensures the highest standards regarding human subject protection. Semalytix only processes data that has been overtly and publicly made available by subjects and follows a GDPR-compliant procedure for data collection and analysis.

At Semalytix, we are looking forward to the further specification of the FDA guidelines, and we are already partnering with pharmaceutical companies to anticipate the final guidelines and put methodologies in place that are in line with the regulators’ requirements.

 

[1] https://www.fda.gov/media/113653/download

 

Tags: Semalytix, AI, Healthcare, Patient Experience