Here at the Klick Wire we came from digital and we believe in those channels for most patient tasks. However, you have to trust the data when you see them. If you're selling in TV for 2027, this is for you. If not, the HCP stat may be useful.

Health Union surveyed 1,398 patients and 63 HCPs about Direct-to-patient perceptions and usage. The HCPs had a focus on endocrinologists because of the strong alignment of DTP with GLP-1s. 

TV is the top way that patients hear about DTP programs. Unless they actually use the programs, then 69% say they heard about it from their doctor.

For the digital afficionados, notice that social media is the top non-TV channel and may be even more effective for more rare conditions (speculation -Ed.).

While DTP has its place in the healthcare ecosystem, insurance still handles the vast majority of drug sales. Outside of GLP-1s only 4% of patients have used the programs (or average of everyone including them, it's hard to tell from the report wording).

HCPs are way more positive toward DTP programs than might be initially assumed. Here is where we have to remember the selection bias in the small HCP survey. There are mainly GPs and Endos in the mix but even so 80% usage is impressive.

Follow this link to pick up the preview report from Health Union then you can get the full report from the company. 

The full report includes data on 14 different programs. It's worth the discussion for anyone thinking of starting their own DTP program.

A report from Menlo is based on a survey of 5,067 U.S. adults and looks at their use of AI. 

Google Gemini is very close to OpenAI's ChatGPT in this survey. This chart seems to include Google AI Overviews in the Gemini number which explains it's position. 

Here we see a comparison of how many Americans perform a task in general and how many perform it with AI. For example, 71% research health questions and 25% of all Americans use AI to do this.

When it comes to preference, Americans want people first. However, a growing number are beginning to prefer AI, depending on the topic. 

Last week we ran a story on 83% of branded searches getting AI overviews (from Ahrefs). In that story we looked at four random drug names that came to mind and found that all of them had AI Overviews.

This piqued our interest. 

We first did a manual test of 25 brands, looking at whether an AI overview appeared, and if it did was it above or below the sponsored results. Again, all of the brand name searches contained AI overviews and 52% (13) of the results were above the sponsored links. 

We then looked at the links found in various placements on the AI overviews.

This keyed image was used to describe to the AI where we might find the items we wanted it to detect (AIO existence, above or below the sponsored links, citation destinations, etc.).

It turns out that AI's results were as good as our run-to-run variance manually. So we created a set of agents to review the top 100 drugs by global sales and found the average number of times that the company controlled the content being cited.

So, for brand name drug searches there are a lot of citations directing back to the company's information. The process is by no means bulletproof yet, and we end up running the scripts manually in Claude, but if anyone wants a demo of how we did this just reply to your copy of the Wire and the author can set you up.

Trilliant's US healthcare industry report for 2026 is extensive at 141 pages. We were struck by how much of the data has already been reported in the Klick Wire for HCP burnout, AI use by physicians, and in PA management. But, some new items do exist.

Americans' trust in institutions has been falling for years generally, and this data is likely brought down by including hospitals with doctors, but still it's low. Keep in mind that this is only the top answer, not top two boxes and that will naturally depress the numbers.

Pharma at 14% tracks. Sorry, fellow pharma marketing kids. 

The main argument in the report is that costs and results in the US healthcare system are not sustainable and we are set for a "negative-sum game" with increasing usage and decreasing funding base.

In analysis of payer patterns, 32% of written prescriptions went unpaid by insurance from 2018-2024. This can easily cause other costs in the system as patients need more ER time or physical interventions.

At 141 pages chock full of data we just can't do this report justice in the Wire. If you're looking for an excellent, if critical, overview of the US healthcare system this is for you.

Pew Research has data on health influencers and looks at what topics they focus on. Most of the content is on healthy living but 34% is on actual medical content.

The top topics highlight the younger demographic that tunes in for medical podcasts.

Our last call out will be on the various topics and how male and female influencers post about them. The medical topics at the bottom are essentially the same but there are variations in the gendered topics. However, note that even here men talk more about female topics than male ones.

AdTaxi put absolutely no methodology in their 2026 consumer spending report so we're not sure how to position it. For now we will assume there were more than 10 respondents but a survey description would be nice.

When asking consumers about their household budgets, groceries, gas, and housing lead the list. Mid-pack is healthcare.

When asked which expenses they would delay or cancel, healthcare is unsurprisingly far down the list. Perhaps what is surprising is that "beauty / personal care" comes just below it. 

While it's not an overall high expense, it is on consumers' minds. This is likely because, as they say, everyone is a patient eventually.

In a new study researchers looked at 1,695 health-related conversations and analyzed their (anonymized) content. Because of its broad definition, education took the top spot for number of conversations. 

The researchers also looked at the number of turns in the conversations. 

A report released by Global Payments looks at consumers' opinions about purchases made via AI. The 5-year predictions are higher in 2026 than they were in 2025.

These numbers are global, so if they don't match your expectations you may be in the USA. Here, the adoption is lower than in other countries and is flat year-over-year.

Part of the continued draw for agents, regardless of the trust to allow them to buy things, is the time they can save for longer tasks (see our example, above).

ICSC surveyed 1,003 US adults in late July to determine the resiliency of the consumer population. It found that 71% of spending is already locked in less-elastic "essentials." 29% is more discretionary. 

With 41% saying essentials will cost more, and 29% saying the same of nonessentials, it points to a strain on American budgets. As usual, health expenses are mostly shielded from these pressures:

However, it also highlights that any expenses here considered excessive will be subject to scrutiny and perhaps political action.

CB Insights highlights that customer service is a highly fragmented market with the leader, Zendesk, only at 18.8% share.

This is interesting for anyone looking to connect patient support to a more traditional customer service line.

Luth research released some teaser data from a study that combined behavioral metrics and quantitative surveys. The researchers looked at how Americans use AI for health. It's not an all-or-nothing proposition.

In terms of sites used, four of the seven listed can be directly controlled by pharma brands to some extent (circled in red).

YouTube is seeing engagement rates fall across the board. This would tend to be a system change as user behaviors don't change that fast. It might be larger numbers of shorts views resulting in lower engagement due to endless dopamine scroll...

In other news, medical health videos have the second highest level of engagement!