Pharma forecasting: The changing role of patients
Milos Graonic, senior vice president, global leader of Nielsen BASES Pharmaceutical Practice, on how to mine patient data for forecasting insights
The intersection of the Internet, social media, and mobile has created a new dynamic in healthcare. Today’s patients have more access to information and to each other and can thus exert more influence on their treatments and the healthcare system at large.
And yet, some pharma companies continue to operate as if this shift were merely theoretical, without tangible impacts on their brands or their bottom lines. As a result, forecasters are often asked to use the same old models to predict what Milos Graonic, senior vice president, global leader of Nielsen BASES Pharmaceutical Practice, describes as a fundamentally altered landscape.
“In the past, the patient was only used in that first step of a [forecasting] framework, pretty much as an object,” Graonic told eyeforpharma’s recent Pharma Forecasting Excellence conference in Berlin. Moving forward, it will be critical to determine “who the patient is and what can happen to that patient” in order to create accurate forecasts.
He estimates that of 20 forecasting requests that come across his company’s desk, all 20 have a physician arm of research. Only four of them request patient research. “We are completely underestimating what’s going on,” Graonic said. There’s “a perfect storm” gathering on the horizon, unless forecasters devise a framework to incorporate patient influence into new product forecasts.
Traditional forecasting models rely heavily on physicians. Forecasters examine the epidemiology of the disease, the size of the market, the competition, the price of reimbursements and other constraints and weigh these against physician demand.
But what happens in a world where patients begin to engage doctors, question them, and ultimately exert influence over the medicine or treatments they prescribe? New statistics from Nielson suggest that 15 percent of patient visits include a request for a specific drug. Forty-one percent of those requests are granted.
Finally, the research suggests that 30 percent of overall prescriptions written are heavily influenced by patients. “Physician demand no longer equals market demand,” Graonic said. The need for a model that weighs physician demand against patient needs is therefore paramount.
Understanding patient needs and their influence on demand requires forecasters to look beyond standard patient flow and anonymous patient level data. Thankfully, finding and listening to the patient voice has never been easier.
Mining patient conversations on social media sites and patient forums can reveal what patients respond positively and negatively to with certain products on the market, and how their alignment differs from that of physicians. Forecasters can also mine patient searches on search engines like Google to gather insights into what types of information patients are pursuing.
When forecasting flu medicines, for instance, it can be difficult to find concrete data about where to focus a forecast. Companies like Google track flu trends by monitoring and aggregating when searchers click on information about the “flu.” That trend data turns out to be an excellent predictor for where the flu will migrate, “because if your neighbor or your wife or your kids start sneezing and coughing you’re quickly [online searching],” Graonic said.
“It’s one of the best predictors,” and it’s a valuable technique that can be replicated for other product forecasts.
Proprietary patient data
Forecasters can also conduct primary research to better understand what will drive patients toward a medicine and how that will influence price, reimbursement, and sales. Graonic gave an example of a forecast he conducted where his team interviewed physicians around the world, presented them with specific patient cases, and asked them the likely outcomes.
One product they predicted would receive 33 percent adoption from patients. Similar interviews conducted with patients around the world determined that, in actuality, that same product is hardly ever used. “The physician-only model was dramatically off,” Graonic said.
Graonic recommends patient surveys and interviews as a valuable way to secure proprietary patient data. He also recommends simulating patient-physician dialogue.
Nielsen uses a process called ‘e-talk’ whereby the company interviews physicians, presenting a series of scenarios with patients and asking how these various scenarios would influence their prescribing behavior. Then they conduct the same process with patients, measuring when patients pose the questions they’re most interested in and to which physician answers they respond enthusiastically.
“We [need] to figure out how these other pieces that people are discussing out there on social networks, how much they can influence patients and a physician and where the dialogue can take our client in the future,” said Graonic.
For an overview of eyeforpharma’s forecasting coverage, see Highlights from eyeforpharma’s Forecasting coverage.
For more on patients, see Special report: Patient's Week 2011.