• Non ci sono risultati.

Healthcare process management in Python: a use case

N/A
N/A
Protected

Academic year: 2021

Condividi "Healthcare process management in Python: a use case"

Copied!
56
0
0

Testo completo

(1)

Healthcare process

management in Python:

a use case

(2)

CRS4@Polaris Park, Sardinia

(3)

Agenda

• what field do we work in?

• what's the project?

• why (and how) Python?

• did it work?

• whither?

(4)

what field do we work in?

(5)

Healthcare Flows

(6)

distributed

medicine

(7)

semantic and computational

management of biomedical

and heterogeneous data

(8)

clinical systems

integration and traceability

in health processes

(9)

To Err is Human

Building a Safer Health System

(Institute of Medicine, 2000)

(10)

44 to 98

thousands

people die each year in US as a result of preventable medical errors

Institute Of Medicine, To Err Is Human: Building a Safer Health System, 2000

(11)

Motorcycle acc Homicide AIDS/HIV Breast Cancer Medical Errors

44,000 41,000

20,000 17,000

3,600

Death causes in the U.S.

44.000

Institute Of Medicine, To Err Is Human: Building a Safer Health System, 2000

(12)

15%

16%

R. Amarsingham, M.D et al. Clinical Information Technologies and Inpatient Outcomes - A multiple hospital study; Archives

hospitals with automated

clinical information systems had less mortality

less post-operative

complications

(13)

1 billion €

Osservatorio ICT in Sanità, ICT in Sanità: l'innovazione in cerca d'autore, 2011

healthcare ICT yearly budget

in Italy is about

(14)

Billing

Scheduling

Laboratory

Comm w/hospitals

Claims status

0% 25% 50% 75% 100%

What do physicians use computers for?

Modern Physician / PricewosterhouseCoopers survey of executive options on key information systems issues; Modern

(15)

Bad Health Informatics can kill

(16)

HL7

• global authority

• interoperability

• HL7 v2 / v3

• documents, and more

(17)

IHE

Abbott

Siemens IBM

Stanford

Harvard

CRS4

Inpeco TIH

Carl Zeiss

US Social Security adm AGFA

...and 438 more

(18)

what's the project?

(19)

60% - 70%

of critical decision making

ToybertME, Chevret S, Cassinat B, Schlageter, Forsman; Why is the laboratory an afterthought for managed care

laboratory services leverage

(20)

Pre-Analytical

60% Analytical

10%

Post-analytical 30%

Errors in laboratory medicine

J. Kalra, Medical errors: impact on clinical laboratories and other critical areas; Clinical Biochemistry 37 (2004) 1052-1062

(21)

Laboratory test

exam

insertion ? Results

(22)

Laboratory test

exam

insertion ? Results

phlebotomy

(23)

What can go wrong

• typing errors

• patient misidentification (and swapping)

• wrong test order entry

• wrong tube type / number

• problems in associating samples with patients

• ...

(24)

tubes and labels

(25)

Our scenarios

• new patient and exam insertion

• query mode in phlebotomy room

• query mode in ward room

• request mode at GP's office

(26)
(27)

why (and how) Python?

(28)

why?

(29)

1. readability

not everyone involved is a developer;

code audits should be easy to do

(30)

2. agility

lots of ever-changing standards, laws and requirements

(31)

3. portability

must support different operating systems and

environments

(32)

4. completeness

lots of different ICT needs in a project of this scope

(33)

5. ease of deployment

as self-consistent as possible;

as few external packages as possible

(34)

Our requirements

• 1. readability

• 2. agility

• 3. portability

• 4. completeness

• 5. ease of deployment

(35)

Python's best features

• 1. readability

• 2. agility

• 3. portability

• 4. completeness

• 5. ease of deployment

(36)

how?

(37)

"it's not Java"

(38)

shut up and show them the prototype

First Law of Python advocacy

(39)

you are probably

already using Python

Zeroth Law of Python advocacy

(40)

did it work?

(41)
(42)

1.hardware drivers

(43)

Managing hardware with Python

• PySerial, python-usb, ctypes...

• use struct!

• see lightning talk @ 18:30 :-)

• hardware CLI

• fast scripting for hardware testing

• build second/third level APIs

(44)

2. asynchronous messages and

networking

(45)

One word: Twisted

• one of Python's killer apps

• write new protocols in minutes

once you finally get it :-)

• non blocking, asynchronous

• look, ma: no Apache!

• very robust and quite scalable

• XMLRPC, SOAP, SSL come for free

(46)

3. applications

(47)

Why web apps

• no deployment or client maintainance

• support several usage and business models

• modern interfaces: RIA, AJAX, etc.

• Django of course!

• MVC, reusable components

• very useful even outside the web

• loose coupling always a good idea

(48)

admin's application

(49)

phlebotomist's

(50)

IHE certification

• our system was certified as IHE-compliant during Connectathon Europe 2011

• LB, LIP, PDQ-S, PDQ-C

(51)

AO G. Brotzu

(52)

whither?

(53)

Python in Healthcare SIG

• established on march 2011

• last edit of wiki on 2011-03-17

• mailing list

• 18 messages in 4 months

• last message in April

(54)

No Country for Old Snakes...

• HL7 implementations

• we only have v2...

• ...and it's quite limited

• everyone uses Java

• Mirth

• made in Java

• scriptable in Javascript

(55)

...or is it?

• we must work together

• Healthcare ICT is important

• philosophically

• and economically -in fact we're hiring ;-)

• show them prototypes; use Trojan horses

• have a good list of use cases

(56)

questions, thoughts, suggestions

Federico Caboni

federico.caboni@crs4.it

http://www.crs4.it

Riferimenti

Documenti correlati

Leonardi, University of Catania, Italy *Correspondence: Nikolaos Tzortzakis nikolaos.tzortzakis@cut.ac.cy Silvana Nicola silvana.nicola@unito.it Dimitrios Savvas dsavvas@aua.gr

[r]

We have reported the multiwavelength properties of CID-42, which shows three unusual features: two close sources embed- ded in the same galaxy, resolved in the optical ACS image,

In this uncommon case report, we describe the case of a 75-year-old woman with a clinical history of recent deep vein thrombosis of the right leg, due to mass effect of an abscess

Rapid identification of bacteria in positive blood culture broths by matrix-assisted laser desorption ionization- time of flight mass spectrometry. Stevenson LG, Drake SK, Shea

Centoventi pazien- ti candidati a colonscopia diagnostica sono stati randomizzati per esse- re sottoposti all'esame utilizzando o la convenzionale luce bianca o un colonscopio con

Clinical Governance, “a framework through which the NHS organizations are accountable to continue to impro- ve the quality of the service and safeguarding high stan- dards of care

By analyzing urine albumin, a 1 -microglobulin, immunoglobulin G and a 2 -macroglobulin together with total protein, prerenal, glomerular, tubular and postrenal causes of