Maximizing Patient Reported Outcomes through the Internet

Introduction

Pharmaceutical companies use a variety of data sources to guide the process of drug development and commercialization. Traditional methods span focus groups, market research studies, as well as gaining insight through opinion leader advisors. With an estimated 89 million people discussing their health online, patients’ sphere of influence is rapidly increasing and the internet is being recognized as the most substantial source of information about a wide range of medical conditions and their treatments. There is a wealth of information in the virtual world to complement pharmaceutical companies’ traditional research methods as millions of patients broadcast their experiences and opinions online on a daily basis. Although pharmaceutical companies have long been leveraging the web to present information to analysts, patients and prescribers, the current primary practice of driving ‘one-way’ social media brand campaigns, rather than asking patients about their experiences, does not allow for a collective analysis of all the relevant data. These online campaigns do not enable a pharmaceutical company to extract valuable market insights because data is taken out of context and the patient voice is not listened to. As a result, the pharmaceutical industry is progressively turning to new technologies to harness online information on patient reported outcomes and translate it into quantifiable real-world data that can be incorporated into on-going strategies.

The power of the patient voice

The growth and popularity of social media sites can potentially inform the lifecycles of new and existing medicines. Forums enable patients to participate in discussion and share information on their experiences of a particular disease or treatment with other patients. These discussions represent a largely untapped source of real-world data on the detection of specific effects associated with product use, which in turn could define the company’s future manufacturing and commercial strategies. Patients not only want to share their health and drug related concerns with other patients in a public space, but they also want the opportunity to have their say. It is in this manner that unstructured online information develops to create a wealth of valuable insight for pharmaceutical companies.

Patients are less restricted in online discussions than they are in traditional market research practices because they are able to discuss their experiences with others. Equally important, the conversations are unfolding in real-time and are not influenced by clinicians or medical observers. Until recently, social media was merely another platform for patient reported outcomes and, as such, its potential to provide pharmaceutical companies with valuable information was not fully realized because its data could not be effectively analyzed.

Pharmaceutical companies have been aware of the potential uses of this online data and there have been efforts to harness it in the past. Previously, online questionnaires have been used to elicit information about the effect of imatinib dose changes in self-reported disease progression1, patient preferences for reduced toxicities2 and off-label drug use3 to develop patient-reported outcome instruments4. However, by placing the patient voice in the context of backgrounds and motivations by looking at patients as individuals, data can be much more insightful. As new technology has opened the virtual world to patients, collective views have formed the agenda of pharmaceutical companies’ clinical trials and commercial strategies. The internet can illustrate to pharmaceutical companies how patients react to their products and over-the-counter drugs. The challenge is how to effectively collect and organize patient reported outcomes data so that it is directly relevant to pharmaceutical companies and their research.

Issues with traditional methods

Despite the obvious benefits that social media presents for the pharmaceutical industry, there are weaknesses with traditional internet analytical methods. Specifically, most patient-reported information is unstructured and therefore cannot be collectively analyzed through traditional research methods. The complexity of human communication means that standardized methods such as keyword searches will not link individual pieces of information or analyze the background behind patients’ actions. To date, the inability to understand the motivations behind current data limits the value of internet discussion as a meaningful data source for healthcare research.

Possibilities arising from the use of health-focused natural-language processing technology

Recent developments in health-focused natural-language processing technology allow Big Data to be harnessed to provide quantitative and qualitative situational analysis, which can in turn inform a pharmaceutical company’s decisions and focus their future expenditure. This new technology can link patients’ actions, such as changing therapy, to underlying reasons by joining pieces of information together and putting them in context (Figure 1). Thousands of patient reported outcomes, in the very words of the patients, can be analyzed to accurately inform pharmaceutical companies about new discussion topics, as a result changing the way in which they learn about their products. The insights gained from health-focused natural-language processing technology can potentially inform drug development and marketing strategies to produce new medicines as patients talk about what is relevant and important to them. This is especially important today as regulators and payers are continually increasing pressure on pharmaceutical companies to demonstrate the value and benefits of a drug when it goes to market. In addition, advanced language processing of internet discussion can provide quantifiable real-world data on patient reported outcomes and patient needs, in turn facilitating the production of new medicines.

It is essential that clinical expertise is still applied to the analysis of internet discussion in order for the most important themes to be drawn and the most clinically relevant questions to be asked. The algorithms that perform the language analysis must be refined and receive expert feedback to ensure fast, accurate and relevant data analysis. Clinical trial programs and pre-launch product positioning are dependent on the way in which algorithms are used and questions are asked, identifying unmet patient needs. After launch, these patient discussions also generate safety signals that can become part of a company’s pharmacovigilance obligations and can inform risk evaluation and mitigation programs.

Reducing potential for bias

It is widely perceived that the analysis of internet discussion limits reports to patients who are active in forum and chat room conversations. Furthermore, it has only recently been possible to filter posts so that an individual patient is only represented once in a particular data search, which could potentially affect research results. It is significant that not only are all groups of society represented on the internet, but it is also a more natural environment for patient discussion, significantly reducing the risk of bias. It is now possible to mirror observations of clinical trials to apply appropriate inclusion and exclusion criteria, which ensures that the patients are being analyzed for the right reasons. It is even possible to detect specific linguistic patterns and de-duplicate data if multiple posts come from one individual, which would limit the impact of outlying views or deliberately defaming comments.

In contrast to conventional market research and clinical trials, internet data often exist in advance of the development of a study and so it is readily available to analyze. Computer simulated studies can now be conducted to address the questions that market research programs or observational clinical trials can do, but with a substantially shorter timeline and potentially lower costs. The maximum use of health-focused natural-language processing technology saves time and provides companies with valuable pre-existing data.

Conclusions

Traditionally, the internet has been used as a research tool but its impact on pharmaceutical companies has been limited due to the lack of specialist algorithm-based methods and over-reliance on generic technical approaches. The internet is an untapped source of potential intrinsic and insightful real-world data that can be applied to both clinical and commercial strategy. This wealth of information means that advanced online research into patient reported outcomes can be time and cost-efficient as well as focused and prioritized according to the patients’ needs. Patients’ views on drugs will be relevant and, significantly, can be incorporated into clinical and commercial strategy. In order to prioritize patient needs, pharmaceutical companies must listen to and analyze their views in the most effective manner possible. Although it is widely acknowledged that the internet is the most popular place for patient discussion, most pharmaceutical companies are yet to realize how to best extract it. They must reconsider how they interact with their patients online in order to successfully produce patient reported outcomes data that is explicitly relevant to their research and clinical trials. Pharmaceutical companies must harness the power of new technologies if they are to generate a wealth of real-time data on patient needs that will guide future research programs.

1. Call J, Scherzer NJ, Josephy PD, Walentas C. Evaluation of self-reported progression and correlation of imatinib dose to survival in patients with metastatic gastrointestinal stromal tumors: an open cohort study. J Gastrointest Cancer 2012; 41 (1): 60-70.

2. Hauber AB, Gonzalez JM, Sirulnik A, Palacios D, Scherzer N. Patient preferences for reducing toxicities of treatments for gastrointestinalmstromal tumor )GIST). Patient Prefer Adherence 2011; 5: 307-14.

3. Frost J, Okun S, Vaughan T, Heywood J, Wicks P. Patient-reported outcomes as a source of evidence in off-label prescribing: analysis of data from PatientsLikeMe. J Med Internet Res 2011; 13 (1): e6.

4.   Wicks, P, Massagli M, Kulkarni A, Dastani H. Use of an online community to develop patient-reported outcome instruments: the Multiple Sclerosis Treatment Adherence Questionnaire (MS-TAQ). J Med Internet Res 2011; 13 (1): e12