Haystacks: How to Improve EHR and the Healthcare System

I woke up today to a request on LinkedIn from someone at Harvard for a healthcare startup that seemed on its surface anything but innovative. The only interesting thing about it was the word “Harvard.” Another “me too” wellness program. I declined and went on to click on a story in the NY Times about Tim Cook, and how he’s shaping Apple and maybe moving into some more healthcare applications.

I flashed back to 5 years ago (2009).

I was a resident and could see so many problems with the healthcare system, including the attempts at creating a good Electronic Health Record (EHR). I knew I could do better, and sketched out some ideas. After creating my first iphone app, I pulled in a friend and colleague Sanjai Rao (also a psychiatrist who’s interested in technology and innovative thinking). We created a business plan and even mockups for a new startup called Haystacks.

It was bold. It was a big idea.

It solved many of the issues that EHR was banging its head against, and still is. It would improve lives for patients and for doctors.

We met with some other startups who told us to quit, because it couldn’t be done. It was too big and we were nobodies.

That didn’t stop us from believing we had good ideas. It did make us realize that it would take years of focusing only on this (with our limited resources) to develop a company. Giving up the rest of our careers and lives for this idea.

We didn’t stick with it, because other interests took precedent. I was focused on being a great therapist and to develop my work as a writer. It would be hard to do those and run a startup.

So we stuck it in a drawer.

Jumping to present day, things have only mildly improved. EHRs have begun to incorporate some of these ideas, but not all, and not enough to solve the ongoing issues. So I decided the ideas should be put out there. Maybe someone with more resources and manpower can utilize these ideas in their companies, and actually improve the products out there.  Maybe they’ll just inspire to think outside the usual paradigms.   I could of course also be naive and uniformed, and this could all be in the works already.  But just in case…

If you think this might be useful or interesting to someone in healthcare or tech, PLEASE SHARE IT.

Today being Father’s Day, I’m dedicating this idea to my father. He is an engineer turned entrepreneur, and has inspired me to think about problems as creative opportunities to come up with elegant solutions. Work Smarter, Not Harder.


So onto Haystacks:

jump to bullet point summary–

An Electronic Health Record is a digital version of the patient information. It’s essentially a digital version of the old paper chart your doctor kept in a file drawer. And that has essentially been all it is used for.

Everyone (including the president) talked about EHR saving time and money, and it hasn’t really done either. Doctors end up spending more time typing up notes and clicking hundreds of boxes and popup windows, rather than handwriting the notes. It’s not why doctors got into medicine. This also takes away time from being with the patient. Patients feel shortchanged on time with their doctor, feeling like just another widget being pumped out of the factory.

So EHR doesn’t save time.

Patients also move between doctors, and often have difficulty getting ahold of their records from their last doctor or hospital visit. The hospital or medical office is holding these records, even though technically the patient “owns” their own health records. This is a massive barrier to integrating care across the system. Even if the patient can get ahold of the right people at the hospital, there’s a weeklong delay to get ahold of the records. That’s not particularly useful if you had to go to an emergency room or an urgent care that wasn’t part of your original clinic.

And even then the records that get sent aren’t often useful. They’re blurry faxes of page upon page of useless information (I got a 900 page medical record sent to me recently, much of which was worthless), or a summary that often leaves out important information.

Patients just want to know their doctor (wherever) can get access to their information to give them good care, and that the information is safe.

So EHR isn’t useful to doctors, isn’t portable as it should be between providers, and still leaves massive gaps in communication between providers.

These gaps in care also mean patients get repeated tests which are expensive, and some even have health consequences (like X-rays).

EHR hasn’t really solved most of these issues, because every medical practice uses a different system.

So EHR doesn’t help the portability of records. Much.

There are some better EHRs out there, but they’re usually designed for large hospital systems. And they’re expensive. Which doesn’t help when a patient shows up to a random emergency room, or a new doctor.

To summarize: 

The Problems:

  • Current Medical Records systems are fragmented, inefficient, difficult to access, costly, and paradoxically take up more time for the providers than using paper charts
  • Current systems have some useful components, but lack many of the real benefits of making a medical record system electronic
  • Most currently available systems are designed for hospitals or large practices
  • What EHRs seem to help with now is portability. Doctors can pull up information on a computer anywhere, but only for their home clinic. Well, not anywhere, but closer to it.
  • Medical errors are a major cause of death, and much of it is due to inefficiency and lack of information to the people that need it when they need it.


Who cares about these issues– 


  • Want to know their records are safe and accessible to future physicians
  • The current standard of transfer of care is abbreviated summaries or blurry delayed faxes


  • Want ease of use, improvement in level of care, low cost, better efficiency
  • Little investment in keeping own medical records system
  • EHRs can cost thousands of dollars for a practice
  • Small providers can’t afford this, and can’t communicate with the bigwigs
  • If it could make them better at their job, that’s ideal


The Main Goals of an EHR could be:

  • An integrated cloud-based EHR system
  • Multiple points of access (phone, tablet, browser)
  • Appealing features to physicians and patients
  • Minimal set-up time
  • Intuitive interface with minimal training required
  • Minimal installation costs for small physician practices


HAYSTACKS — a Multi-tiered EHR – from individual to systemic 

  • Mobile Record Entry and Storage (smartphone based)
    • Patient enters own basic data
    • Begins a grassroots system and concept of patient ownership of their record
    • Available wherever they go (ED’s, other doctors)
    • Available during an emergency
    • Can have online backup (Centralized storage), accessible for a registered, verified physician.
    • Every access of a patient record is logged.
  • Tablet
    • Data entry and Retrieval of notes
    • A quick entry system of note-taking to improve efficiency for physicians and other providers
    • Involves sliders and radio buttons to quantify symptoms and their history –
      • Ultimately trackable and usable
    • Can add a text entry
    • Can incorporate dictation features for certain elements
    • An intelligent user interface that prompts questions based on presenting symptoms
    • On-the-go data entry and retrieval of information to access and update a chart and note along the way
    • Quantifies data as variables that improves tracking and patient outcome
    • Historically tracked data (blood pressure, pain, medication dose) are viewable in graphed form at multiple scales [pinch-zoom], overlaid, allowing for simplified scanning of records
    • Integrates with centralized record system as a whole
  • Browser-based EHR
    • Includes tablet features for viewing records and data entry
    • Could use citrix or other means to access browser-based records
    • Set up customized templates for record viewing and record entry
      • Example: An emergency room physician may have a preferred workup for vomiting that he can make auto-populated.
    • The system can begin with workup and management recommendations, further able to be edited by provider
    • Access point for an emergency physician to view centralized records without having our EHR as their primary EHR
    • Patients may then request data entry by ED physician into centralized EHR so their PCP can have access, furthering a buy-in by the healthcare system and integration
  • “Cloud-based” database of all records
    • A centralized backup of records from all other entry (mobile, tablet, browser)
    • Bypasses bureaucratic hurdles of records departments, blurry faxes, wait times, and useless information
    • “Bank-level security”
    • Storage of records from each site/clinic/hospital, accessible for any registered physician
    • Indexed records including all keywords, akin to a “spotlight” on your records
    • Scanning feature of old records into pdf +/or OCR to make documents searchable
    • An offline mode
      • A clinic can download the cloud data, alter it during a visit, then transmit the update back to the main server after visit is complete
  • Additional modules (necessary, less unique)
    • Billing
    • e-Prescriptions
    • Scheduling
    • Patient reminders using push notification
    • Premium features for physicians/practices
    • Importing ability from other EHRs

With modern software technology, none of the current issues are necessary. It would require some leaps in thinking and practice, of course.

Current EHRs are often designed by software engineers, who don’t really understand how physicians or other healthcare providers think and work. Physicians, all too often, just want to be left alone. They are creatures of habit. They want to do things the way they learned to do them. They have difficulty imagining how an EHR could be helpful. When asked what they want, many physicians respond “let me do what I do now [writing a paragraph on a piece of paper] on a computer.” So many EHRs are more or less exactly that. Blank pages where doctors type information. They may be able to pull up lab results and information like that, but that’s the extent of the “innovation.”

So the solutions in the idea of Haystacks comes in the form of multiple tiers. It’s not one solution, but many solutions that work together. There’s the top tier (a backbone or system level), then a desktop/browser tier, a tablet tier, and even a mobile phone level. Some of these have started to be developed in modern companies.

The backbone of Haystacks would be a centralized server, essentially a cloud EHR for the whole country and beyond. If the patient owns their own record, then they should be able to choose where their information is stored. Right now every medical practice or hospital keeps their own records. There’s little exchange of information between sites. Centralizing it all removes the blurry faxes and the long delays. Now of course there would have to be an incentive for organizations to adopt such a system, and that might take many forms, from the ease of the UI to universal accessibility to the promise of actually making data useful to physicians. This would involve bank-level security, but make records available to any licensed physician who registers with the system. Patients could know who has ever accessed their records, including a timestamp for every access. This would allow a patient to walk into any clinic or hospital and basically take their whole medical record with them, as it’s all available on the cloud to their new doctor. Even moreso any emergency room physician could access outside records so that they know the entire background of a person. In an emergency, there’s no time to request outside information.

Now of course there would need to be innovations, and this is about making this specific EHR useful to physicians, in a way that makes their job easier. And this improvement is in the use of data. Physicians really want specific pieces of information, and much of this can be quantified. Besides lab results, there’s reason for each visit, symptoms at that visit, medications taken and doses. There’s a large amount of quantifiable data that physicians look for. It should be organized in a way that is useful for physicians, such as overlapping timelines that could be scrolled through and zoomed in/out, such as a timeline for types of visit, med dose at that time, and lab results. Having the useful data quickly accessible and organized, rather than having to comb through hundreds of verbose notes is the real goldmine for EHR. Once the data is digitized, it needs to be organized and indexed. Much like “spotlight” on a mac, each person’s medical record should be indexed and easily searchable. Information should be findable quickly. Even old notes could be scanned into pdf format and converted with OCR (or hand-converted into typed notes) to get useful data points from each medical visit. This is all systems-level aspects of Haystacks.

Every level of Haystacks offers unique innovations, fitting unmet needs within that level of healthcare, and leveraging the available technology and convenience to improve data entry, use, and patient care.

The better and more efficiently information is entered and used, the more time and money are saved. More importantly, that car accident victim that doesn’t remember their medication list, will be less likely to die from a medical error if all their records are easily available in the emergency room. This could be the way to make that happen.

Now of course I could write at some length about this project, and am happy to do so. For a Father’s day post about my dreams for how to clean up one area of the healthcare system, this post is a good start. I may clean it up in the future, if people are so interested.

Haystacks: How to Improve EHR and the Healthcare System
Tagged on:                                     

Paul Puri

Dr. Puri is a board certified psychiatrist, in private practice in Los Angeles. He practices multiple forms of psychotherapy, including hypnosis, in addition to managing medications. He attended medical school at University of Rochester, and specialty training at University of California, San Diego. He is currently on the Vol Clinical Faculty at UCLA. In his non-clinical time he writes TV pilots, and designs iPhone apps.