Tag Archives: research

Advanced Analytics, Part III

This post is also about the front-end part, as a conduit for information delivery to the decision maker. Previous two posts are available, it’s recommended to check out the Introduction into the Big Picture and Ruminations on Conveying, Organization and Segmentation of Information for Executives as users.

Big Data? All Data!

It’s time to pay attention to all data available. Personally I see no reasons to limit to big data. All data matters, most recent data matters more, oldest data matters less. It is possible to visualize plenty of data on relatively small space, which is convenient for delivery onto smartphones and wrist-sized gadgets. The rationale is to depict firm details on the most recent/relevant data, the relevancy is determined by the adopted processes. In SDLC it could be a sprint or iteration; in healthcare it could be a period since current admission. The latest measured value matters a lot, hence must be clearly distinguished on top of the other values within the period. The dynamics during the period also matters, hence should be visualized to convey the dynamics.

Previous periods/cycles do matter, especially for comparison and benchmarking to enable better strategic planning. The firm details on dynamics during past cycles are not so valuable, while deviations into both positive and negative directions are very informative. Decision maker knows how to classify the current cycle exceptions, whether something brand new happened or whether business experienced even more severe deviations in the past, and recall how.

Being inspired some time ago by medical patient summaries by Tufte and Powsner I’ve tried to generalize the concept to be applicable to other non-healthcare industries. So far it fits perfectly, allows customization and flexibility, especially for the optimization of the processes, where people usually use control charts on dashboards. Below is a generalized version of the ‘All Data’ chart as a concept.

all_data

Inverted Pyramid

The principle of Inverted pyramid is partially present there, the pyramid is rotated by 90 degrees. Most important information is within the biggest part of the chart, in the center and on the right. It is rather information than data, because id conveys latest value, dynamics during recent cycle, benchmarking against the normal range, indication of deviations (in qualitative way, using only two categories: somewhat and significant). It’s rationale to stay in the range of 10 with the measurements so that they are remember-able relatively easy.

The next narrow part to the left from the sparkline is partially information and partially data. It’s used for comparison and benchmarking, analysis of exceptions, retrospective analysis. It is absolutely logical to fit there 10 times more data, so that if there is a lack of information in the biggest part, the user is able to dig deeper and obtain significantly more facts and reasons, as measurements of the same thing. Hence phase shift means at least 10x growths. With medical patient summaries the ratio was similar: one-two months between admission and discharge vs. one previous year. But 10x is not a hard ratio, it’s more indicative that we need a kind of phase shift to different data, different level of abstraction.

The leftmost narrow part is actually the all and oldest data. It is additional phase shift, relatively to the middle part, hence imagine additional 10x increase and digging to the different level of abstraction again. Only exceptions marked as min/max are comparable between all parts. Everything else constitutes the inverted pyramid of making the information out of raw data.

Cap of the pyramid: Vital Signs

I think the cap of the information pyramid requires special conceptualization. ‘All Data’ is attractive tool to deliver project/process vital signs for executives and other managers (decision makers), they could be compressed even more. Furthermore, the top five-seven measurements could be stacked and consumed all together. That increases the value of the information synergistically, because some indicators are naturally perceived together as juxtaposition of what is going on.

Specific vital signs for business performance and SDLC process optimization were listed in details in my previous post Advanced Analytics, Part II. Here I will only mention them for your convenience: productivity, predictability, value and value-add, innovation in core competency, human factor and emotional intelligence/engagement. Those are ‘the must’ for executives. They could be stacked as vital signs and consumed as integral big picture.

vistal_signs

Of course we can introduce normal range there, ticks for the time tracking, highlight min/max… The drawing represents the idea of stacking and consumption of executive information of SDLC project/process performance in modern manner. You could critisize or improve it, I’ll be thankful for feedback.

There are two dozens of lower level operational indicators and measurements. Some of them could be naturally conveyed via ‘All Data’ concept, others require other concepts. I am going to address them in next posts. Stay tuned.

Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

[20 Facts about SoftServe]

SoftServe sets up its Research and Development Department in 2008…

Tagged , , , , , , , , , , ,

Mobile EMR, Part II

On 27th of August I’ve published Mobile EMR, Part I. This post is a continuation.

The main output from initial implementation was feedback from users. They just needed even more information. We initially considered mEMR and All Information vs. Big Data. But it happened that some important information was missing from the concept relied on Powsner/Tufte research. Hence we have added more information and now ready to show the results of our research.

First of all, data is still slightly out of sync, so please be tolerant. It is mechanical piece of work and will be resolved as soon as we integrate with hospital’s backend. The charts on the default screen will show the data most suitable for the each diagnosis. This will be covered in Part III when we are ready with data.

Second, quick introduction for redesign of the initial concept. Vital Signs had to go together, because they deliver synergetically more information when seen relatively to each other. Vital Signs are required for all diagnosis. Hence we have designed a special kind of chart for vital signs and hosted it on top of the iPad. Medications happened to be extremely important, so that physician instantly see what meds are used right now, reaction of vital signs, diagnosis and allergy, and significant events. All other charts are specific to the diagnosis and physician should be able to drag’n’drop them as she needs. It is obvious that diabetes is cured differently than Alzheimer. Only one chart has its dedicated place there – EKG. Partially, EKG is connected to the vital signs, but historically (and technically too) the EKG chart is complemently different and should be rendered separately. Below is a snapshot of the new default screen:

Default Screen (with Notes)

Most important notes are filtered as Significant Events and could be viewed exclusively. Actually default screen can start with Significant Events. We just don’t have much data for today’s demo. Below is a screenshot with Significant Events for the same patient.

Default Screen (with Significant Events)

Charts are configurable like apps on iPad. You tap and hold the one, then move to the desired place and release it. All other charts are ordered automatically around it. This is very useful for the physician to work as she prefers. It’s a good opportunity to configure the sets according to diagnosis. Actually we embedded pre-sets, because it is obvious that hypertension disease is cured differently than cut wound. Screenshot below shows some basic charts, but we are working on its usability. More about that in Part III some time.

Charts Configuration

According to Inverted Pyramid , default screen is a cap of the information mountain. When many are hyping around Big Data, we move forward with All Information. Data is a low-level atoms. Users need information from the data. Our mEMR default screen delivers much information. It can deliver all information. It is up to MD to configure the charts that are most informative in her context. MD can dig for additional information on demand. Labs are available on separate view, groupped into the panels. Images (x-rays) are available on separate view too. MD can click onto the tab IMAGERY and switch to the view with image thumbnails, which correspond to MRIs, radiology/x-ray and other types of medical imagery. Clicking on any thumbnail leads to the image zoomed to the entire iPad screen estate. The image becomes zoomable and draggable. We use our BigImage(tm) IP to empower image delivery of any size to any front end. The interaction with the image is according to Apple HIG standard.

Imagery (empowered by BigImage)

I don’t put here a snapshot of the scan. because it looks like standard full screen picture. Additional description and demo of the BigImage(tm) technology is available at SoftServe site http://bigimage.softserveinc.com. If new labs or new PACS are available, then they are pushed to the home screen as red notifications on the tab label (like on MEASUREMENTS tab above) so that physician can notice and click to see them. It is common scenario if some complicated lab required, e.g. tissue research for cancer.

Labs are shown in tabular form. This was confirmed by user testing. We have grouped the labs by the corresponding panels (logical sets of measurements). It is possible to order labs by date in ascending (chronological) and descending (most recent result is first) orders. Snapshot below shows labs in chronological order. Physician can swipe the table to the left (and then right) to see older results.

Labs

Editing is possible via long tap of the widget, until corresponding widget goes into the edit mode. Quick single click will return the widget to preview mode. MD can edit (edit existing, delete existing and assign new) medications, enter significant sign, notes. Audit is automatic, according to HIPAA, time and identity is captured and stored together with edited data.

Continued in Mobile EMR, Part III.

Tagged , , , , , , , , , , , , , , , , , , , , , , , ,

Mobile EMR, Part I

At SoftServe we do some researches, I’m in charge of them. This time I’d like to describe one of our researches. It is called Mobile EMR. EMR stands for Electronic Medical Record. This research is intended to find what will hospitals, agencies, practitioners will use tomorrow. Mobile EMR aka mEMR is a hot intersection of booming Mobility, Big Data, Touch Interface, Cloud trends. Today we research it, tomorrow they will use it. SoftServe works with multiple ISVs which build solutions for hospitals, agencies, hence we have influence on evolution of medical solutions.

Part I will be about how we started and what we get up to date, with conclusion for next steps, which will be described later in Part II.

Beginning

It has started some time ago during UX research for medical data visualization. Then famous government initiative has been launched to transfer healthcare industry to EMR. Mixing that soup we figured out that it is possible to propose really new concept for the EMR for physicians. We found research by Seth M. Powsner and Edward R. Tufte, “Graphical Summary of Patient Status” published in far 1994. They work at Yale University. One of them is MD, another is PhD in statistics and guru of presenting data. It was oriented onto All Data paradigm that Mr.Tufte loves. I love it too. I love the idea of having All Data on single pager. As soon as Apple released iPad we understood that it is perfect one pager to put the EMR onto. In 2011 I attended E.Tufte one-day course, and found a moment to speak about healthcare data visualization. Mr.Tufte confirmed there was nothing done in the industry still! He pointed to the printed copy of mentioned research and proposed to implement it. After that we had one more face-to-face contact with Mr.Tufte on that research, we god some additional clarifications and recommendations (mainly related to low level details such as sparklines). Below is a snapshot of the proposed visualization by Powsner/Tufte:

Seth M. Powsner and Edward R. Tufte, “Graphical Summary of Patient Status”, The Lancet 344 (August 6, 1994), 386-389.

All Data has to be handled by special kind of chart, that shows three periods of data. Rightmost biggest part shows week or 10 days, middle narrow part shows previous year, and leftmost part shows all possible data prior to last year. Having such data presentation we are capable to display all anomalies and trends for the whole period that has data logs.

'All Data' chart

“All Data” chart. Seth M. Powsner and Edward R. Tufte, “Graphical Summary of Patient Status”, The Lancet 344 (August 6, 1994), 386-389.

 

We were aware there was web implementation. It was 100 percent copy of the research (charts part of it). Below is a screenshot from the browser:

Web EMR by KavaChart

First Version

We took our Mobile IP and SDK (such as authentication, cache, cryptography etc reusable blocks; Big Image(tm), SaaS SDK) and built first app for iPad. Obviously we lacked deep domain knowledge, hence first version is not perfect. But idea was to do technology feasibility rather than ready-made solution (because we work with healthcare ISVs who keep deep domain expertise). There were few cycles for visualization and layout of the charts and other UI elements. As a result, we got this “first version”:

mEMR Default Screen

All remarks and proposed improvements will be listed herein a bit later! Right now I’d like to show few more screenshots what we have got. Physicians identify the patients by MRN or by name, if the patient stays at the hospital for some time. Hence, we introduced two-way patient identification: via My Patients list, and via wristband/MRN scan. Below are screenshots: My Patients and Wristband/MRN Scan. First one is My Patients:

mEMR My Patients

This one is Wristband/MRN Scan:

mEMR Wristband Scan

User Testing and Recommendations

“First Version” has been shown to MD from New York large hospital. Impressions were mixed :-O
In general – idea of such mEMR is good. But using it with its current data is not so valuable. I’ve got recommendation for data presentation improvements. They are listed in unsorted order below. The format is the following: subheader represents an issue or proposal for improvement. The text after it describe reasoning and clarifies details.

User Info

In the left top corner next to the photo we show basic patient information. It happened that both DOB and age is required simultaneously. Year serves as additional identification for multiple Johm Smiths. Ok, the information about age should be like 56 y.o. F, where F stands for female. M stands for male. Sex must follow the age, and must be encoded into single letter F or M. It should be next to the patient name. DOB should be on second line. Look below:

Photo    John Smith  62 y.o. M

    DOB 05/24/1950

In the right top corner we shown Discharged. It happened that nobody cares about discharge, while everybody needs Admitted. Hence, right below the MRN we have to add new line. MRN and Admitted record must be kept together. Look below:

MRN: 221881-5

Admitted: 08/24/2012

Those requirements are related to the integrity of  the data, what is kept with what. There is no requirements to UI layout. Current layout was confirmed as good.

Diagnosis

There is a widely used abbreviation for diagnoses. It is OK to use smth like HTN and MDs will understand it. Writing Depression is also good, but often names are long and abbreviation works better. Important information that is mission in our “first version” is Allergy. It is “the must” that complements the diagnosis. We have to allocate some space below Diagnosis and specify there Allergy. If Allergy is present, then it must be very contrast, may be highlighted in red. Look below:

HTN, Depression

Allergy: Aspirin (rash)

Medication

The patient can be currently on some medications. They must be listed on top of our single pager. Current Medication is such important as Diagnosis. We can specify it as a simple list of what the patient takes. Look below:

Metoprolol 50 daily [sorry, I can’t put units here right now]

Norvase 5 mg daily

Paroxetine 20 mg daily

Medication can be labeled as Meds. It should be in the upper area of the screen, as well as improved User Info and Diagnosis. Other info such as hospitalization and clinic visits should be kept within the upper area of the screen. Later sections describe the middle area of the screen.

I will provide typical medications for those 4 diagnosis: HTN, CAD/HLD, CHF, Asthma. Please update patients data to those 4 diagnosis only. This is our restriction for demo purposes. Medications will take max 4 lines on text, each line up to 25-30 chars.

Vital Signs

We shown then as separate charts: chart per temperature, chart per pressure, chart per heart rate. It sucks. The reason is that it does not represent information. What MDs see as information from vital signs? The relation between all signs together! Hence we must pack all 5 vistal signs into single chart so that all peaks or dropdowns are visible relatively to each other. The labeling for vital signs is the following: t for temperature, HR for heart rate, BP for blood pressure, RR for respiratory, SpO2 for something I don’t remember right now. The order of vital signs should be as listed with t on top and SpO2 at the bottom. Chart should be from sparklines, not for dots. All mins and maxs should be visualized properly. Now I understood what Mr.Tufte explained to our UX designer about use of sparklines. We though it was for all charts, but it happened it is only for vital signs. Look here for visualization spec. The chart with vital signs must be on top. Let’s keep it alwas in left top corner. If there is insufficient space, stretch it to the width of two charts.

Notes

They suck. Usually doctors put a lot of secondary info there. Hence there is no place for such irrelevant notes on the Home Screen. All doctors need is an entry point into the Notes. Doctors read them on demand. As a secondary screen opened intentionally. Implement it if possible.

Significant Events

Those are important! Significant Event is some even that is related to the life of the patient. E.g. patient fall and got head injury. In other words, non-typical events according to the diagnosis. If vomiting is common for that diagnosis, then vomiting is not a significant event, hence not worth big attention.

Probably we can redesign the right pane to show Medication on top, then Significant Events on top and in the middle, and has Notes as entry point at the bottom. If Medication takes too much space, then we can group it with Vital Signs (increased horizontally chart) and keep them together on top as a separate section. All other charts and Significant Events should be below them. Personally I like first alternative better – to use current right pane for both Medications and Significant Events.

EKG

Electrocardiology is “the must” on home screen. Add it as a separate chart, as in Tufte’s research and in web demo (see links above). It is OK to keep EKG on same place, hence keep it hear Vital Signs for example, at same horizontal level.

Imaging (aka Radiology)

This is a section of all scans. MDs call them interchangeably Imaging and Radiolody.  They have it as secondary information. Hence we have to put probably 1 or 2 images on the Home Screen, but implement separate screen with thumbnails to all imagery available for that patient. For our demo we can keep few images on Home Page, then we will see. But they (images) should be organised in some group labelled as Imaging.

Labs

Our “first version” doesn’t have Labs data at all. How Labs look like? It is a list of text and numbers. Furthermore, MDs get used to the order how to read the labs (BTW I confirm the same for vital signs). E.g. Labs called Hepatic Panel is shown as a sequence of AST, ALT, Alkaline Phosph. Labs called Coagulation is shown as PT, PTT, INR. It is common that Labs list contains 5+ sets of results (named as Panel here). Superscripts are used to represent K+, Na+, Cl-, CO3-, Ca2+ and so on.

We can allocate the space at the bottom of the Home Screen to visualize Labs and Imaging. Labs should grab more space than Imaging. As I’ve told, having 2 thumbnails as an entry point to Big Image is sufficient. Then physician should click either title Imaging or “…” next to the title, to get to the separate screen with all other images. Re Labs, they could be shown in tab format. Look below:

              date1  date2  date3

Erythr.    4.5    4.3    2.8

HGB    140    138    100

Disease Profiles

It could be useful to have layout presets for chronic diseases like Diabetes etc.

Continued on Mobile EMR, Part II.

Tagged , , , , , , , , , , , , , , , , , , , ,