Journal Club Discussion

1
Journal Club Discussion

• Syed I. Nabi, MD
• Resident, MHRI
• February 01, 2006

2
Clinical effect of computer generated physician
reminders in health screening in primary health
care- a controlled CLINICAL TRIALof preventive
services among the elderly.
Eva Toth-Pal, Gunnar H. Nisson, Anna-Karin
Furhoff International Journal of Medical
Informatics 2004 73, 695-703
3
Background

• Several studies have documented large
  discrepancies between evidence-based guidelines
  and what is done in real practice
• Number of factors are attributed including
  organizational problems, reimbursement issues,
  attitudes of physicians, patients etc
• Literature has showed preventive services can be
  improved by addressing these issues
• Use of illness visits to physician as an
  opportunity for preventive service is widely
  recommended and applied
• Computer based reminders as a tool for promotion
  of these services could be effective if
  integrated as part of EMR (Electronic Medical
  Record)
• In Sweden where this study is being conducted,
  electronic patient record is used in all patient
  practices and hospitals
• No established screening schedules for elderly
  people in Sweden!, therefore extent of proper
  screening is still unknown

4
Null Hypothesis
USING COMPUTER GENERATED SCREENING METHODS WILL
HAVE NO EFFECT ON PREVENTION OF DIABETES,
HYPERTENSION, ANEMIA, COBALAMIN DEFICIENCY AND
HYPOTHYROIDISM.
5
Introduction

• Pilot Study
• Computer Generated on screen physician reminder
  program for health screening of elderly
  individuals
• PHC (Primary Health Center ) have an integrated
  EPR (Electronic Patient Records) system
• Integration plays a key role in this study

6
Material Methods

• Setting
• 4 PHC (Primary Health Center) serving 32,000
• 1 center used for computer generated screening
• 5 GP (General Practitioner) 1 Resident
• 3 centers used as control
• 12 GPs and 2 Residents
• GPs Family Medicine doctors 5-15 yrs experience
• No differences in age experience
• GPs of only 2 center had little clue of what was
  going on. Other 2 had no clue.

7
Inclusion Criteria

• Age of 70 years or Older
• Visit to GP during study period, excluding
  emergency visits

8
Exclusion Criteria

• Pretty much everyone else was excluded.
• Most people in this room would have been excluded

9
Material Methods contd

• 914 pts. Fulfilled inclusion criteria _at_ 1 PHC
• 602 actual PARTICIPANT in actual health screen
• 224 excluded as GP didnt have enough time
• 24 demented thus not able to consent
• 12 severe illness at time of visit
• 51 who refused altogether
• 1989 pts. _at_ other 3 PHC fulfilled inclusion
  criteria and acted as CONTROLs

10
Material Methods contd

• 20 MONTH of BASELINE DATA
• April 91- December 92
• To see how commonly screening test done _at_ the
  respective centers

20 MONTH INTERVENTION PERIOD April 93- December 94
20 MONTH INTERVENTION-FREE PERIOD February
95-September 96
FOLLOW UP
11
Intervention???

• Five Intervention areas and related tests were
  chosen
• Hypertension, blood pressure
• Diabetes, Blood glucose
• Anemia, Hemoglobin level
• Cobalamin deficiency, Serum Cobalamin level
• Hypothyroidism, serum thyrotropin level
• Mixture of manual and lab procedures and new
  screening area (Cobalamin)
• Exclusively used computerized methods throughout
  the study

12
EPR/Reminder system

• GPs given brief introduction prior to study
• Program voluntariy triggered by GP _at_ time of
  encounter while reading pt. record
• Computer then adjusted the list of five screening
  test to the individual recommended (screening)
  test individualized to the patient

13
Evaluation and Outcome

• Results evaluated using MQL (Medical Query
  Language) from all 4 centers
• MQL Evaluated each test separately for both
  initial results and result follow up
• Resulted in list of patients for whom the test
  would have been recommended as a screening test
• EPR of these patients analyzed divided into
• Had Undergone the test
• Had a Pathological test
• Had a New Diagnosis
• Appropriate pharmacological treatment initiated
  (cobalamin/hypothyroidism)

14
Evaluation and Outcome contd

• Baseline data analyzed for similar screening
  tests and diagnosis and compared
• STATISTICS
• Confidence Interval of 95 used for comparison
• If value of 0 outside, observed difference
  regardede as being statistically significant with
  a p value lt0.05 (5)

15
Results as expected ?
no acute abnormality identified
16
Groups

• Participants 602 pts 65 Women, 70-74 (38)
• Controls 1989 pts 64 Women, 70-74 (40)

17
Results Salient Features

• Number of Pre-study (baseline)/Controls/
  participants were different (TABLE 1)
• A diagnosis of Deficiency of Cobalamin was less
  in participants 3.5 compared to controls 6.5
• Based on analysis comparison from Baseline data
• Statistically significant plt0.05 (-4.9 -1.2)
• of pts who had undergone (screening) test was
  significantly lower _at_ the PHC 1 (TABLE 2)
  (comparision is with Baseline Study groups) for
  4/5 tests with lowest rate for Cobalamin (11.1)

18
TABLE 1
19
(No Transcript)
20
Table 3

• Presents results from Intervention and Follow up
  period
• HTN more tested 13, more pathology 12, new
  diagnosis comparative
• ANEMIA more tested 12.9, new diagnosis
  comparative
• DIABETES more tested 26.2, new diagnosis lt0.8
  _at_ f/u. 62.9 had even fasted!!!
• B-12 more tested 74.5, abnormal results,
  diagnosis, Rx was ALSO MORE FREQUENT
• HYPOTHYROIDISM more tested 60.5, diagnosis
  comparative

21
(No Transcript)
22
Discussion

• Statistical significant increase in amount of
  testing (lab manual) in established and new
  screening methods
• However the clinical outcome was seen affecting
  only Cobalamin deficiency
• CGS effective for new screening methods being
  developed

23
Discussion Strengths

• COMPUTER SYSTEM
• Universality of the software _at_ all the centers
• Integrated Medical record systems system
• Familiarity/ease of use for person involved
• Breadth of Database
• Swedens health system model
• Minimal time spend gathering information
• Screening was integrated to doctors routine
  rather than running special protocols minimizing
  bias
• Long Periods of Intervention and Intervention
  Free period

24
Discussion Weakness

• Not a Randomized Control
• Article states that would have caused influencing
  physician and thus altered the setting and could
  have affected working conditions
• Randomizing Patients- cause of bias? No
  elaboration
• Large group of patient excluded for various
  reasons Demented, GPs with not enough time
• Could not justify us residents to convince Clinic
  Director to invest more in computer technology
  just for the sole purpose of screening
• HTN, DM, Hypothyroid, Anemia are routinely
  screened in elderly universally in Developed
  countries, hence CGS system might be beneficial
  for less commonly screened conditions

25

• Special thanks to
• Milan Mathew, MD