Patients’ Access to Medical Records:Experience with PatCIS

James J. Cimino

Division of General Internal Medicine &

Department of Medical Informatics

Overview

• Motivation

• Architecture

• Functionality

• Recruitment

• Evaluation Techniques

• Results

• Discussion

Motivation

• Understanding on-line health information

• Access to personal health records

• Regulatory requirements are coming

• Commercial sites give patients access to their data

• What will happen to the patient?

• What will happen to the patient-provider relationship?

The Patient Clinical Information System (PatCIS)

• New York Presbyterian Hospital clinical data repository

• Web-based Clinical Information System (WebCIS)

• National Information Infrastructure contract from NLM:

– give patients WebCIS

– see what happens

• Pilot study conducted

Data Entry

Review

Advice

Education

Comments

Help

Logout

Vital Signs Blood Sugar

Data Entry

patcis.cgi

Web ServerWeb Browser

SessionRegistry

UsageLog

Internet

2

3

6

PatCIS Architecture

1

CGI

4

5

Functionality

• Sign on

• Data entry

• Data review

• Education

• Advice

• Infobuttons

• Customization

Recruitment

• Mail physician consent forms to physicians

• Wait for physicians to suggest subjects

• Mail URL for consent form to subjects

• On-line enrollment

• Patient prints, signs and mails consent form

• Physician provides function-specific consent

• Mail user name, password and SecurID card to patients

Evaluation Issues

• Theories and methods from

– cognitive and social sciences

– usability engineering

– studies of human-computer interaction

• Complexity and diversity of users leads to need for multi-method approach to evaluation

– formative - feedback into design

– summative - evaluate final product

Evaluation Challenges

• Varied users who interact from various locations

– Less able to conduct controlled studies

• Current state-of-the-art

– track user actions (e.g. clicks) - tells what they do, but not why

– on-line questionnaires/feedback forms - often not filled in, limited questions

– interviews - users often do not know what they do

Evaluation of Patient Systems

• What type of information do patients want?

• What is the effect on doctor-patient interaction?

• How does system affect patient understanding and communication with their providers?

– understanding their own condition

– decision making and management of their condition

• How to collect useful data from limited number of subjects?

Patient Variability

• Differences in prior knowledge

• Health differences

• Cognitive differences

• Cultural diversity

• Motivational differences

• Varied educational background

PatCIS Evaluation Questions

• What features of PatCIS are most used by patients?

• What features are least used and why?

• Are there usability issues that need to be resolved?

• How does use of PatCIS affect the doctor-patient interaction?

• Can patients comprehend information presented by PatCIS?

• Does use of PatCIS affect decision making and disease management?

1. Video BasedUsability Testing - in laboratory

2. Telephone Interviews (audiotape)

PatientInteract via WWW

3. E-mail(to evaluators)

Patient Information System (PatCIS)

4. Tracking User Actions - System Usage Database (log files)

5. On-line Questionnaire Data (sequenced forms)

“Televaluation” of PatCIS

(Kushniruk, Patel, Patel, Cimino, 2001)

Log File Analysis

sandcar!Fri Oct 27 11:32:22 2000!cim.cpmc.columbia.edu!|patcis^login

sandcar!Fri Oct 27 11:32:24 2000!cim.cpmc.columbia.edu!|patcis^Data Review

sandcar!Fri Oct 27 11:32:28 2000!cim.cpmc.columbia.edu!|patcis^Data Review^Laboratory Detail^lab_detail.cgi

sandcar!Fri Oct 27 11:32:30 2000!cim.cpmc.columbia.edu!|patcis^Data Review^Laboratory Detail^labSum.cgi

sandcar!Fri Oct 27 11:32:35 2000!cim.cpmc.columbia.edu!|patcis^logout

Results

• Functions

• Enrollment

• System usage

• Function usage

• Adverse events

Enrollment

• Mailing to >200 physicians

• 13 physicians returned signed consent forms

• 19 subjects suggested

• 13 enrolled

• 12 used the system over 19 months

• 1 non-CPMC subject enrolled

Baseline Questionnaire data

• Initial patient expectations

“will keep me informed of latest developments”

“will enable me to better understand treatment”

“possible information my doctor may forget”

“timely availability of test results”

• Prior subject education

– Majority of subjects (90 %) had some post-secondary education

– Conditions patients were interested included cholesterol management, diabetes and chronic conditions

– All were prior computer users

System Usage

131 log-on failures

22 wrong user name

51 wrong password

58 wrong Secure ID

33 log-ons without any activity

466 active sessions (261 logged out)

-----

630 log-ons

0

2

4

6

8

10

12

14

16

18 User 1

User 2

User 3

User 4

User 5

User 6

User 7

User 8

User 9

User 10

User 11

User 12

User 13

Log-Ons Failures by User

Active Log-Ons by User

0

5

10

15

20

25

30

35

40

45

50User 1

User 2

User 3

User 4

User 5

User 6

User 7

User 8

User 9

User 10

User 11

User 12

User 13

Average Monthly Log-Ons

0

2

4

6

8

10

12

14

16

U1 U2 U3 U4 U5 U6 U7 U8 U9 U10 U11 U12 U13

0

10

20

30

40

50

60

70User 1

User 2

User 3

User 4

User 5

User 6

User 7

User 8

User 9

User 10

User 11

User 12

User 13

Average Session Time by User

0

50

100

150

200

250User 1

User 2

User 3

User 4

User 5

User 6

User 7

User 8

User 9

User 10

User 11

User 12

User 13

Minutes per Month

Function Usage I

• Data review: 1831 total– 1518 laboratory

• 737 “Laboratory” button 1083 specific reports• 186 “Laboratory Details” button• 249 summaries

– 36 vital signs– 35 diabetes flow sheets– 212 reports (81 radiology, 35 pathology)– 30 Microbiology

Function Usage I

• Data review: 1831 total

I

• Data entry: 73 total– 34 vital signs– 39 diabetes flow sheets

• Education: 53 total• Advice: 6 total

– 5 cholesterol guideline– 1 mammography guideline

• Other:– 10 newsgroups– 83 infobuttons– 2 comments– 10 e-mail to physician– 17 disclaimers– 13 help

Adverse Events

• None reported

Interviews• Conducted (via telephone) summer/fall 2000• Both patients and their physicians interviewed• Audio recorded, transcribed and qualitatively analyzed• Examples of interview probes (for patient interviews)

– General usage• What do you use PatCIS for, why?• What features have you found most (least) useful?

– Usability• Have you had any problems using PatCIS?

– Effects on health care interaction and decision making• Do you feel PatCIS has changed the way you view

health or illness?• Has use of PatCIS changed in any way your

interaction with your health care providers? Explain

Analyses of Interviews with Patients and Physicians

• Subjects felt that use of PatCIS had a direct impact on doctor-patient communication

• For patients

– Improved communication with physician resulting from reviewing their own lab results prior to meeting with their physician – 85%

• For physicians

– Allows for focus on issues of decision making and patient compliance during during doctor-patient interviews (rather than review of data) – 100%

– Decreased workload in reviewing essential data

Excerpts from Patient Interviews

“Communication is less in the way of getting information now, and more in the way of discussing treatment options and agreeing on a course of action, so to me its more efficient than the old way”

“I look for trends in my medical data and if I see something I can contact the doctor to see what’s going on, what we can do, change meds or whatever”

Excerpts from Physician Interviews

“Right now most of the communication takes place during the ten or fifteen minute visit and if I throw a lot of information at the patient about their condition or what I want them to do, its very hard for them to absorb all that.”

“It (PatCIS) gives them a chance to go back and look at things about their health record that they can then ask better questions about in the limited time that we have during the visit. Its another channel of communication”

Additional Findings

• Low usage of certain functions (e.g. advice, educational) may have been due to lack of user knowledge of full system capability – Implications for providing further training

• Usability– Occasional difficulties in viewing graphs– Problems with linkages to outside resources

• Subjects did not have problems in learning how to use PatCIS for reviewing data

• No adverse effects of use of PatCIS were indicated by patients or physicians

Summary

• PatCIS was found to be usable by patients studied

• Patients indicated they understood the information accessed

• Review of data (Lab and reports) was the most important application of PatCIS by patients

• Patients and physicians indicated that PatCIS had improved their level of communication

• Low usage of other functions may have resulted from limited patient knowledge of system capabilities

Discussion

• Architecture supports integration, security and tracking

• Enrollment was disappointing

• Population was highly selected: by MD, by self, by Web

• Two patterns: monthly and daily

• Log-on difficulties overcome

• Laboratories are the most popular

Next Directions

• Diabetes mellitus patients

– Data entry

– Coordination with clinicians

– Targeted educational materials

Conclusion

• Enthusiasm is not universal

• Technical issues were not a problem for our patients

• Privacy is achievable

• Patient understanding of their records was good

• Other features were of less interest

• Patient-physician impact was positive– better D-P communication– education of both patient and doctor– treatment negotiations with common understanding

Conclusions (cont.)

• Our Analyses indicate that use of systems like PatCIS can serve to facilitate processes involved in improved health care delivery through shared workload between doctor and patient, resulting in better communication and negotiation

– Reducing physician workload – patients better informed of day-to-day variations in their own condition

– Making patients more accountable for chronic diseases and management