Researching the Real

By Laurie Gelb, April 25, 2011

While the rest of the social Web is constantly redesigning itself based on user context and needs, the health Web lags. Personalization, filtering, sorting, non-linear exploration and other “Web-standard” capabilities on sites like Amazon is lacking as yet.

Why?

One possible reason is that surveys of health Web users commonly manifest a “how much” obsession, neglecting the who, when, where, what, how and why. This creates misleading constructs for action, which we will continue to explore in future posts.

Let’s examine on just one question that Pew asks [not picking on Pew, just that its survey is widely quoted]:

Q32 Overall, who do you think is more helpful when you need... [INSERT FIRST ITEM] – health professionals like doctors and nurses, OR other sources, such as fellow patients, friends and family?  And who is more helpful when you need... [INSERT NEXT ITEM; RANDOMIZE]?  AS NECESSARY: Professional sources like doctors and nurses, OR other sources, such as such as fellow patients, friends and family?  

a. An accurate medical diagnosis

b. Emotional support in dealing with a health issue

c. Practical advice for coping with day-to-day health situations 

d. Information about alternative treatments

e. Information about prescription drugs

f. A quick remedy for an everyday health issue

g. A recommendation for a doctor or specialist 

h. A recommendation for a hospital or other medical facility

1 Professional sources

2 Other sources

3 (VOL) Both equally

8 (DO NOT READ) Don’t know

9 (DO NOT READ) Refused

Beyond the ambiguity ("day to day health situation") and heterogeneity in some of these question items (you might have a different process for researching someone/ somewhere to remove an ingrown toenail vs. a CABG), we can summarize the problem here with two words: false dichotomy. 

How actionable can these answers be, even when tracked over time? The answer items are all binary, they relate to categories rather than actual resources and the unaided “both equally” option is a copout/source of social bias rather than a reality. 

What Pew could be asking:           

Have you or anyone whom you help make health decisions, such as a family member, ever faced  [specific situation]?

[if yes]

How recently did you or someone you care for face [specific situation]?

These decisions never stop [they don't, for someone w/ chronic illness/injury, often neglected in these surveys but also often your high utilizers]

Dealing with that now

Within the last month

A month or two ago

A few months ago

About a year ago

More than a year ago

When situation X most recently arose, from which of the following did you receive information before making a final decision? Please check all that apply. [randomize order w/ selected anchors]

• Your or the patient's physician
• A staff member in that physician's office or clinic
• A brochure or video in a physician's office or clinic
• The Internet
• A magazine or newsletter
• TV or radio program
• Friend or family member who works in health care
• Friend or family member who does not work in health care
• Other (please specify)

Often, we then zoom in on that recent situation and dissect how well the search process (or lack thereof) worked out in terms of needs vs. outcomes.

What else can we ask Web users, that we don’t know the answers to, that we can actually use to design stronger decision support?  An example…when we look at choices that are being made NOW – which we will need to branch into – we can find out:

• Who helps whom (for example, what percentage of the sample is currently influencing (1) health decisions on a child’s behalf (2) health decisions on an adult’s behalf other than themselves. This and the next item will aid subgroup analysis/tracking.
• Is the current decision process around self-care or accepting a professional’s recommendation?
• Where/how do they think they need to research, if anything before making or accepting a choice?
• If the decision process is ending, is it by choice or necessity?
• How well do they think they are succeeding in getting what they need? This is categorical like in real life -- not at all to it's done.
• What else do they need to know, that they do not yet? This can be structured, open-ended or both.
• What are the barriers to getting what they think they need? (e.g. not enough time, not sure how to search, overwhelmed with info, didn't have long enough chat w/ doc, not sure if insurance will cover…)

Does this sound really nitpicking? It’s really no worse than the research process we go through for cereal or paint – just that instead of keying on the purchase process, we are keying on information-seeking. User-driven branching and filtering moves the respondent through quickly. BUT – we should never assume prework. If I walk into CVS like a robot and walk out with my default OTC analgesic, you should know that as well.

It's always a worthy goal to keep your information specific, your verbs active and your sentences short. If you wouldn't say it, apart from legal disclaimers, why write it?