Watching and asking

Editor’s note: Peter Simpson is principal at Segmedica Inc., a Buffalo, N.Y., research firm. Lynn Ford-Somma is a Segmedica senior project director.

Like other professions, market research is constantly evolving and this is no more evident than in the area of health care research. While traditionally the industry has relied on interrogational research of various standard types to understand motivations and behaviors, in recent years it has become clear that a purely interrogational approach does not always deliver accurate and reliable information about past events or unconscious processes.

New data collection techniques are becoming increasingly available to researchers ranging from cutting-edge non-invasive neuroimaging that has radically revised our understanding of brain processes and emotional functions to expanded use of technology for data collection using the Web, digital recordings, eye-tracking and various devices such as iPads.

Recently, the trend in health care is a move toward more observational methods with the intent to get closer to “reality.” So we have seen an increasing emphasis on ethnography, audio/video recordings of patient visits with physicians and other techniques such as video diaries that presumably bring us closer to the truth.

While this is a significant step in upgrading research from asking about mental processes to observing them, there are also inherent limitations in these approaches. Observation without subsequent interrogation can be misleading. What is seemingly so obvious may mean something entirely different when put into a broader context.

With all the options available, the challenge today is how to use an integrated approach to maximize value and produce impactful results. How might an integrated approach be used in health care research?

  • to better understand the patient journey from diagnosis onward;
  • to understand how treatment decisions are made and in the context of rational concerns such as insurance/cost, social-psychological barriers, etc.;
  • to better understand and potentially identify physician and patient segments;
  • for program development in any number of areas such as compliance/adherence, disease management, patient education or even improving customer service; and
  • for positioning and messaging research.

In bringing together watching and asking, we have three main types of modality to consider: interrogational research - qualitative or quantitative Q&A; observational research - in which we observe people and processes as they interact in the real world; simulation research - in which respondents are placed in a situation which resembles real life.

Traditional interrogational research includes all the standard (and not-so-standard) forms of research that ask respondents to articulate their attitudes and behaviors and occasionally their emotions. Surveys, in-depth interviews and focus groups are the most common forms of this research. This type of research is most effective in understanding rational reasons for making certain decisions. For example, in health care it could be whether the patient has health insurance or whether particular medical services are available in a geographic area. In an integrated approach, this research could be used to understand the contexts in which the physician or patient make decisions as well as identify general attitudes and behavior.

Observational research or traditional ethnography is usually carried out by an anthropologist with a video camera and a notebook observing the minutia of someone’s day. Indeed this approach has very tangible value, for example in understanding how patients manage diet and insulin with diabetes, or the needs of physically-challenged patients. However, advances in technology have expanded our observational options and in many ways enhanced the value of observational research.

Not all ethnography needs to be carried out in person, which has the added value of removing the ethnographer from the equation. Simple and miniature digital recording devices make it possible to, for example, record physician-patient interactions in a medical practice without interference from the researcher, or a patient’s experience at the pharmacy counter. This has the added advantage of expanding the geography that can be covered by traditional ethnography and potentially reducing incurred travel costs. The same trends in miniaturization and mass production now make it feasible to provide respondents with a Webcam to record their own daily activities and keep video or photo diaries.

Nor are all observations visual. For example, patient respondents can be provided with Bluetooth-enabled accelerometers that measure their physical activities and sleep patterns remotely in studies that need to determine the frequency and intensity of patient activity.

Blog monitoring and social media monitoring are also observational. Monitoring this unscripted activity can provide real-time information on what is of most concern to the target respondents, whether they are patients with a rare and potentially serious health condition or members of a health insurer discussing their customer experience.

Simulation research techniques might be considered when there are considerable challenges in observing the specific behaviors of interest. They may be of a very personal nature, or occur infrequently or unpredictably. There may not be time, or budget, to overcome these challenges in field observation.

Simulation research can include setting up situations that resemble real-life experiences. For example, such research might include simulating the patient’s experience in a waiting room by renting a physician’s office, displaying patient education or promotional materials and observing patient behavior. Another type of simulation research might involve using actors to portray either patients or physicians in a simulated patient/physician interaction. This might be accompanied by traditional interrogation techniques to assess the overall experience. Simulation research, while not real, offers an advantage over observational research by providing a more controlled setting and allowing various stimuli to be manipulated to test different reactions.

Combining multiple approaches provides value

While each of these types of research approaches offers value in and of itself, combining multiple approaches provides value beyond the sum of the parts.

Consider a specific example where observation, simulation and interrogation can be combined. In this case the challenge is to understand how physicians manage patients to accept and be compliant with a drug that may generate significant concerns in the patients’ minds.

As a first step a significant number of physician-patient interactions where this treatment possibility is likely to arise might be recorded; let’s say 100 such appointments would provide a sound base. In analyzing the results, the Hawthorne effect (that observing a behavior modifies the behavior observed) would need to be considered. In this case, while the purpose of the study is blinded, physicians and patients know they are being recorded and may be more conscious of their behavior. It is also possible that either the physician or the patient may be more reluctant to have as open and honest communication as they might have otherwise. This is where enhancing the research with traditional interrogation such as a follow-up interview is helpful. This gives an opportunity to ask a physician why they did not offer the apparently most appropriate treatment or discover that the patient who said they would try a new drug has, in fact, no intention of actually doing so.

Having completed this stage, the next step might be to identify more optimal physician-patient conversations. How can this be developed further? This is where simulation research can be appropriate.

In this case physicians are recruited to come to a research facility for an in-person interview. Partway through the interview a case study is presented complete with appropriate clinical information (such as lab results or imaging). When this has been absorbed by the physician a carefully briefed actor is brought in, without prior warning, and the physician is asked to assume they are a patient and to, for example, explain the diagnosis and treatment options. While this simulation approach is not real it has the advantages that the physician respondents have no time to develop strategies to make themselves appear in the best light and, as the actor and the case study are completely controlled, it is possible to run the same scenario or potentially more than one scenario over and over with different physicians.

For health care companies, such as insurance providers, combined modalities could be used to understand member engagement and develop outreach programs. For example, members with newly diagnosed serious conditions could be followed using video or photo journals as they navigate the health care system. This could be combined with traditional interrogational techniques and finally, simulation research could be added to test ways of improving the member experience.

New understanding

The frontiers of understanding the working of the human brain are being rapidly expanded and this new understanding is impacting market research in many ways. Useful discoveries include:

  • The mechanism of memory formation and the role of emotion in laying down long-term memory records.
  • The existence of hardwired heuristics (unconscious decision-making processes) that, for example, apply pressure to the conscious brain to conform to group norms.
  • The separate processing of input by the cerebral cortex (the conscious rational mind) and the cerebellum (the limbic or emotional mind), emphasizing the necessity of understanding - and messaging to - both sets of mental processes.

Psychoprofiling respondents adds another layer of understanding to respondent behaviors. For example, some physicians go to great lengths to ensure that their patients get the treatments they wish to prescribe and devote considerable effort to supporting them in navigating the reimbursement system. Others will be very forthright that this is “not their job” and may ignore the issue completely. Psychoprofiling not only helps understand what leads to these different points of view but also helps to predict which physicians will adopt either position, or something in between.

As another example, consumers/patients who exhibit certain heuristics with respect to obtaining and maintaining personal assets will be very resistant to spending on health care costs and make apparently irrational choices between treatment and retaining income. Overlaying this information enables the research to understand how these factors influence health care behaviors.

Multimodal projects result in incredibly rich data but often in non-standardized forms - video, audio recording, artifacts, still pictures, projective exercise results, psychoprofiles and transcripts. Not only are there multiple forms of data but there tends to be a lot of it! This requires a higher level of analytic sophistication than with traditional single-mode research. While there are software packages for organizing multimedia output, they not in themselves analytical.

People remain vital at this stage. There are aids such as computer linguistic analysis models and psycholinguistic interpreters, but in the end, research is analyzed by people, and agencies need more well-rounded and more versatile staff than ever before to distill the many data down to conclusions and actions.

Analysts, researchers, linguists and ethnographers must collaborate to interpret the output and its significance to the business issues at hand. This type of approach tends to throw up all manner of fascinating output which might have great business value outside of the core purpose of the study, or no value at all beyond being plain interesting. Organizing the conclusions and expressing them succinctly and in appropriate categories with supporting video-audio-photographic evidence requires considerable discipline and teamwork.

Some recommendations

For clients and agencies who are considering a multimodal solution with a mix of interrogation, observational and stimulation research, here are some recommendations.

Clients need to keep in mind that:

  • This type of research takes longer than solely interrogational research.
  • They will have less immediate control over the activities of the researchers and therefore need to exercise appropriate strategic control over the project and a close relationship with the project leaders.
  • The final results will not be determined from any one modality and will not emerge until the analysis is competed. A traditional topline is typically not possible in this type of work.

Agencies need to:

  • Develop staff with a broad worldview and an understanding of multiple techniques, social and even political issues and the basic psychology of the human mind.
  • Develop working practices and investigate ways to integrate data in different forms.
  • Promote strong team ethics.
  • Use rigorous “challenge” and quality-control methods to ensure that the more complicated analysis leads to a clear, unambiguous and, above    all, relevant, result.

Upgrade in the insights

While not a magic bullet and it is not suitable for all projects, observing processes in real or simulated circumstances offers a tremendous upgrade in the insights and actionability of research. Observation needs to be integrated with more traditional interrogational research in a skilled and thoughtful way and its value to the overall project critically assessed. One contribution of significance is to challenge and overturn erroneous conclusions that can be drawn from traditional research because respondents give answers based on what they think should be the case rather than what actually is the case. In this multimodal research approach there are no “ethnography projects” as we have seen in the past, because ethnography and its relatives are means to end not the end themselves.