Notes

Outline

IT for Healthcare: Why is it taking so long? NIST Workshop Barriers to Implementation RS Johannes, MD, MS

Story of Scurvy Once terribly important: 100 of 160 of Vasco de  Gama’s crew died of it on the Cape of Good Hope voyage. James Lancaster:  Gave lemon juice to seamen on one of four ships on voyage to India.  Most remained healthy but he lost 110 of 278 seamen on other three ships. James Lind: British naval physician.  Usually credited with having done the first clinical trial.  Scurvy patients were assigned 1 of 5 treatment groups.  He produced objective evidence that the citrus fruit subgroup improved. British Navy adopts citrus fruits for long voyages. British Merchant Marine adopts same policy.

The History  of Scurvy

Personal Context 1978: Chief Resident, Osler Medical Service, Johns Hopkins Hospital. 1979: Implemented HOPS adapted from Bill Stead’s System at Duke University. 1990: Left Johns Hopkins to join a startup dedicated to solving the structured medical document problem. 1994 & 1997: Successfully competed for 2 NIST ATP Grants.

NIST ATP Program Windows based component for “posting codified patient data to a repository via automation of the process of clinical note writing.” MedEncode project was piloted at Brigham & Women’s hospital. 70% of clinical abstracts submitted to AGA since 1995 have used data from GIstation Distribution rights were sold to Olympus America Over 350 (~10% of endoscopy centers) were installed

Olympus ImageManager

Recent Context Joined Cardinal MediQual in August of 2000 MediQual is a Cardinal company that does health services and outcomes research for acute care settings Pioneered one of the first successful methodologies to perform risk adjustment for reporting performance measures in acute care settings The methods involve use of clinical data to attain physician acceptance and higher precision Support Pennsylvania Health Care Cost Containment Council (PHC4)

PHC4 Has 15 years of publically reporting performance measures across all hospitals in Pennsylvania. Is the only state with such a track record. Uses the clinically based MediQual methodology. Relies on Key Clinical Findings (KCFs): Ejection Fraction Presence on admission of pleural effusions or ascites Glasgow Coma score Vital signs Laboratory data

The Importance of Clinical Data “Administrative data cannot account for differences in severity of illness with any degree of clinical subtlety.” Mark R. Chassin NEJM, August 30, 2001 “Risk models based upon administrative data can lead to substantial misclassification of hospitals when compared with models based upon higher-quality data.” Harlan M Krumholz JAMA, March 13, 2002

PHC4 Has 15 years of publically reporting performance measures across all hospitals in Pennsylvania. Is the only state with such a track record. Uses the clinically based MediQual methodology. Relies on Key Clinical Findings (KCFs): Ejection Fraction Presence on admission of pleural effusions or ascites Glasgow Coma score Vital signs Laboratory data

Clinical Data Remains Hard to Come by UB92 data is available electronically An electronic feed for laboratory data has been available for several years.  To date, less than 20% of hospitals avail them selves of it.  The rest abstract. Pharmacy data is just now beginning to be available electronically and the heterogeneity of pharmacy system data is shocking. No hospital provides vital signs electronically. Other KCFs are abstracted from the medical record.

Barriers to Implementation 1998 American College of Gastroenterology Meeting Timing of EMRs I had a four year old set of slides I also reviewed several articles on the subject

The European Experience

Gerry O’Conner & NNE

Everett M. Rogers “S-shaped Curve”

The European Experience

Origins of the “S-shaped Curve”

The innovation-decision process

Who Drives This Process Not the innovators! Early adopters Great peer respect Must be consistently correct Early majority A full 1/3 of eventual users Much peer group interaction Seldom opinion leaders Often have a long deliberation period

The Roger’s Model

Homophily Communication between like individuals Much communication is like this since it is often highly effective communication Can act as an “invisible barrier to the flow of innovations” Given a high degree of homophily the “in the know” interact primarily with themselves and impede the flow of information

State of the Art This process has and will continue to take more time than one might wish Rapid growth in this area is possible Key barriers remain Absence of an universal patient identifier Continued separation of clinical and administrative systems Continued separation of hospital and ambulatory systems Clear elucidation and understanding of benefits

Other Important Points Clement McDonald: JAMIA 1997 Importance of standards Too many different separate systems with different data structures Diversity of kinds of sites of care Optimistic about Kassebaum-Kennedy Bill Kevin Johnson: Arch Pediat Adolesc Med 2002 Knapp categorization (Situational; Cognitive; Liability; Knowledge) Highlighted knowledge and attitudinal barriers IT costs Unproven return on investment

Is Now the Time? Yes…but, You can not all of what you might want today You will be an early adopter You must set realistic expectations and goals You will find yourself ahead of the curve in 2-3 years You will have setbacks You must document the benefits You should share the results of documented benefits You must have strong internal leadership and institutional will

Perhaps the Timing is Good Market is looking for a growth sector Medical costs continue to rise Calls for data are coming from everywhere Importance of clinical research is both being demonstrated and called into question HIPAA is around the corner NIST ATP has a proven record in technology transfer

It Will Be Worth the Wait!