Comprehend (YC W11) Raises $8.4M Led by Sequoia to Lower Clinical Trial Costs | Hacker News Reader

Comprehend (YC W11) Raises $8.4M Led by Sequoia to Lower Clinical Trial Costs

blogs.wsj.com

119 pointsjakek40 comments

Rick from Comprehend here, we're also hiring. Definitely reach out if you're interested in transforming the medical industry! http://www.comprehend.com/careers

Do you hire non-Americans? (Not that I'm looking at the moment, but I think this should be mentioned on startup job postings).

Yes we do = )

Silly question, but you don't mention security or HIPAA/HITECH anywhere on your site (or in jobs).

Are you guys working on that, or is it just 'assumed' that your stack and cloud capability is compliant?

Subjects enrolled in clinical trials fall under informed consent and are outside HIPPA/HITECH. We actually take steps beyond either of those standards to ensure the data, costing pharma billions to create, is not breached.

Don't most Life Sciences companies already do this internally? I mean, I only have anecdotal evidence, but I worked on and released a couple versions of something for Abbott Labs that sounds eerily similar to what the press release mentions you doing (actually, went a lot further into budget forecasting (including updating and alerting for deviations based on incoming data automatically) and automatic payment calculation and authorization in addition to collaboration, data input, time keeping, and dashboarding for analysts/executives.

In any event, good luck to you. From what I've seen from the inside, newer tech stacks can deal some serious disruption in a very short time frame.

Most companies do something internally, but it's typically slow, expensive, out-of-date, breaks frequently, and is prone to errors.

Congratulations Rick! Great news.

Congrats, cool topic. Impressive stuff.

We're tackling eTMF plus other R&D from a content perspective through SaaS. See you're integrating with Medidata as well.

Greetings from Pleasanton ;)

Veeva? You should stop by next time you're in Palo Alto, would love to meet. Email is in my profile.

I invested in a clinical trials company many years ago (operating an investigative site), and I'm curious what kind of traction you've gotten from the big pharmaceutical companies?

I totally get how it is a compelling story for investors, but I have serious doubts about adoption and use.

It was my experience that data and sound record keeping was a secondary to approvals. Many times paperwork was "corrected" and patients were excluded because of serious adverse events. Unless your model is to help identify exclusionary criteria retroactively?

The industry has been drastically changing over the past decade. 15 years ago, all data collection was paper based. Then came Electronic Data Collection systems. Now, there are cloud systems likes Medidata Rave and Veeva which are doing exceptionally well. This change has created tremendous opportunity for new startups.

This is happening across all enterprise verticals, and this is why there is such an opportunity for enterprise startups. As Warren Hogarth says in the original article "There’s about a trillion dollars of enterprise software", much of which will "transition to [specialized] enterprise software... and create multi-billion dollar companies".

Based on your linkedin profiles neither of you has a life sciences background.

How did you guys come up with this idea?

We both do. I've actually been involved in writing clinical analytics software since 2000, including having been one of the lead developers on software used by the FDA and top pharma. Jud previously had a consulting company in the space performing data integrations.

Rick, one of the biggest costs I hear about in the field is the cost of pre-allocating funds to upcoming trials without knowing how much to pre-allocate. The other is knowing how to profile clinicians and help them to raise their ability to recruit patients. Do you manage to address either of these points?

Finally, channeling 21st century software technology to something useful. One of the most important industries in the world - healthcare - is sadly broken and advances significantly slower than other technological domains. I wish more investments went that way, for our own good...

Healthcare and education are the biggest laggards in technology yet data is super important to their effectiveness. The regulatory minefield (and traditionalist culture) within these industries can only hold back modern technology adoption for so long. Fortunately they seem to be becoming more open to it these days.

They have problems to profit.

My startup is for educational games for kids, and almost noone want to invest in us, because of how the market works, that is basically or you sell, or you don't. There is no middle ground, no organic slow growth...

Happily our seed money is sustaining us, but it is clear that we need to spend in marketing about 5 or 6 times our current budget (That go almost 100% for development).

What part of education is a laggard when it comes to technology? Education is always trying something new when it comes to technology. It seems like a bigger issue with education is that it's an easy target for con artists.

> The regulatory minefield

Ha! The FDA review board for a drug is allowed to own and trade pharmaceutical stocks, receive cash incentives and guarantees of future employment from companies under review, and generally relies on the summary findings of the company to make their decisions.

It is as much of a minefield as the FCC is to the Family Research Council, or congress is to big oil.

I took it that they were speaking to people trying to get in on the action and not the incumbents.

To fix healthcare we need less technology, not more. People are getting cancer and diabetes because we've gone from eating nutrient dense high quality food to industrially processed corn, soy, and wheat-based junk. We have to stop the vilification of red meat, promote the health benefits of grass-fed meat, and try to get factory farms to employ some sort of holistic management [1]. Not only can we feed the world in a healthful manner this way, we can also reverse climate change.

Big pharma isn't interested in curing anything. It's financially beneficial to them to keep us sick. Why help them even more?

[1]: http://www.ted.com/talks/allan_savory_how_to_green_the_world...

Oh, is soy a villain this week?

I'm sorry, but it's so hard to keep up with which wacked out unscientific belief about food I'm supposed to believe from day to day.

I can understand your frustration. The key is not to follow the headlines but do deep research on each claim. For example, intake of dietary saturated fat is simply not associated with heart disease. http://ajcn.nutrition.org/content/early/2010/01/13/ajcn.2009...

Most soy consumption traditionally has been in Asia in small quantities, and after fermentation (soy sauce, miso etc). Vegan widespread usage of soy does not fit this pattern. http://www.westonaprice.org/soy-alert/studies-showing-advers...

To really lower clinical trial costs, one of the biggest line item that can have an impact on the bottom line is patient recruitment. According to stats[1], 80% of trials are delayed due to patient recruitment issues and causes several billions of dollars to industry. Furthermore, different locations, countries have different accural rates and proper planning is required to select the right sites.

IMO, providing a dashboard of trial stats to pharma is helpful but it would not really cut down the majority of costs.

1. http://www.ciscrp.org/professional/facts_pat.html

There are a lot of companies doing this already - Qunitiles, Blue Chip, Acurian, etc. Anything that can streamline the process is potentially beneficial

There are a lot of companies doing this, but it's still a huge problem and I agree w/ chintan that there is a lot of opportunity for innovation.

Hi,

I used to work for a BigPharma designing their in-house clinical trial tracking software. The problem is not so much the platform, but system integration. We had a nice system with d3.js charts plotting Pk trials after results. A scheduling system with nice Ajax calendar component to schedule daily animal dosing. A in-life component that linked up with RFID scanners to track patient vitals/conditions. A nice drag-and-drop web GUI for setting up samples for mass spec analysis. All under a phase-by-phase, wizard step-by-step platform, from study design all the way to compliance reporting. Designing the system from the ground up wasn't difficult, system integration wasn't. We even had gadgetized all of the components so that PIs/lab associates could customize their daily dashboard.

Research groups are across the globe and use different stat packages. No one is willing to give up years of work in Excel spreadsheets, Phoenix, STATA whatever. So have to build tons of import/export tools to accommodate all of the tools.

Different research groups wanted different things. We had stakeholders, e.g., PIs from different research groups come in during all of the design/review meetings. One team wanted their scheduling calendar on 'MS Project' linear style layout, another team wanted it to be on the Outlook style. We ended up building both.

Different research groups had different hardware spec's as well. Throughout the pipeline, we had a nice tool that set up samples to be analyzed for mass spec analysis. Integrating with various hardware formats and different layout's, well-templates with existing in-house assay library database/web-services took a long time.

tl,dr; BigPharma has tons of proprietary assay, compounds database/web-service, third-party statistical packages and mass spec hardware spec's that a external vendor needs to integrate with; good luck.

I wonder how you deal with security concerns. I think since the result of clinical trials can have such a large impact on a pharma co's share price, there have been a number of cases where insiders with access to trial results have given information to financial trading firms who then proceed to exploit the information in the market.

how do you assure your customers that the information put in your system won't leak?

We actually spend a lot of time thinking and dealing with security concerns, both from a software development perspective and from a devops perspective. You're correct that the data is very valuable, so a large percentage of our architecture and design decisions are dictated by security and privacy concerns.

This actually becomes even more complicated when you think about maintaining blinding, aggregate-only blinding, and some of the other features we need to support.

"“We are very interested in vertical software as a service,” he said. “There’s about a trillion dollars of enterprise software, if you’re looking at market capitalization out there today. Some of that will transition to [specialized] enterprise software…and create multi-billion dollar companies.”"

Interesting point. Two points converging really. Enterprise SAAS & Vertical SAAS. There is still lots of room here. Customers willing to pay. Anyone that's trying to come up with "synthetic" startup ideas might start here. More difficult to comprehend verticals are probably least served.

Basecamp for X, where X is some business type that not many people understand.

Isn't most of the cost in clinical trials in manufacturing the drugs, finding and paying investigators, paying for legal services, recruitment of patients, getting informed consent, etc? Does Comprehend help with that stuff?

I am surprised that Medidata solutions( mdsol.com ) hasn't been brought up. They do something very similar with very forward looking products/apps that are already being adopted by big pharma.

Medidata is a data collection system, and Comprehend works out of the box with them. If you're one of their customers you can start looking at your data in Comprehend simply by providing us with your API credentials.

The problem is, many companies use Medidata, Oracle, MERGE, and other vendors. Inside of a single study, let alone across several. That's where Comprehend comes in. We simply help customers get the actionable insights they need, and then help them take those actions.

Is this just another CTMS? Similar to OpenClinica etc?

No, we actually work on top of CTMS or EDC systems. Basically, there are just so many of these types of systems, that without Comprehend companies have no way to answer questions across them. The abundance of these systems, which you allude to, is a big part of the problem that we solve.

Congrats guys! This is amazing stuff!!

Congrats!

Congrats Rick and Jud!

Congrats guys!

Reply on news.ycombinator.com