Results of Field-Testing of EDPS2000

The first field test of the EDPS2000 software was conducted at 3 major hospitals in Bangalore, India, during October 1999 to January 2000. It was based on a study of 10,000 patients at the KIMS Hospital, Victoria Hospital, and Vani Vilas Hospital; the first is a private hospital, while the other two are government hospitals. Outpatients were screened at these hospitals using both the EDPS2000 system and by medical examination by a doctor. In some cases, when required, laboratory tests were conducted either at the EDPS2000 clinic or at the hospital laboratory. Subsequently, the two sets of evaluations were compared and analyzed. The results of this first pilot test are shown in the following table. Overall, the EDPS2000 software performed exceedingly well.
The data presented in the table below are based on the entire study population of 10,000 patients seen over the 4-month period. Based on the ongoing analysis performed during the test period, the diagnostic logic within EDPS2000 system was continually improved. As a result, the data obtained during the last two months of the pilot project were better than the overall results presented in the following table. We shall continue to further enhance the capability and reliability of the EDPS2000 software.

EDPS2000 system will be implemented at PHCs only with government approval. The use of the system and its benefits will be explained to local officials, community leaders, and the general population at village meetings. It must be noted that the system is only the first line of intervention, and all serious cases will be referred to doctors and/or hospitals.

The data entered into the computers will be regarded as sensitive and highly confidential as it will contain the patient's and possibly their family's medical and social history. It is therefore essential that we address the issue of confidentiality properly. To guarantee this the EDPS2000 software will incorporate the same degree of checks and security procedures that are employed in today's financial world.

EDPS2000 Pilot Test Results (10/99 -01/00)

Sl. No.
TYPE OF DISEASE/ COMPLAINT OF PATIENT
CORRECT DIAGNOSIS (provided by EDPS2000, or "sensitivity")
BLANK DIAGNOSIS
(No diagnosis provided by EDPS2000)
WRONG DIAGNOSIS
(a wrong diagnosis provided by EDPS2000, or "false negatives")
TOTAL CASES
1.
MALNUTRITION
ANEMIA
VITAMIN DEFICIENCY
611(90.3%)
24(3.5)
42(6.2)
677
2.
MUSCULOSKELETAL PROBLEMS
199(76.53)
34(13.1)
27(10.4)
260
3.
VIRAL FEVER
COMMON COLD
MALARIA
OTHER INFECTIOUS DISEASES
1106(85.88)
54(4.19)
128(9.94)
1288
4.
SORE THROAT
105(80.77)
9(6.92)
16(12.30)
130
5.
PREGNANCY
277(92.95)
13(4.36)
8(2.68)
298
6.
CHEST PAIN
220(81.48)
20(7.47)
30(11.01)
270
7.
STOMACH PROBLEMS
GASTROENTERITIS
ACID PEPTIC DISEASE
1025(87.08)
72(6.12)
80(6.80)
1177
8.
BACK PROBLEM
CERVICAL SPONDYLOSIS
199(89.64)
5(2.52)
18(8.11)
222
9.
GENITAL PROBLEM
221(96.51)
4(1.75)
4(1.75)
229
10.
EYE PROBLEM
193(94.15)
2(.98)
10(4.88)
205
11.
WOUND
FRACTURE
DOG BITE
447(92.74)
14(2.90)
21(4.36)
482
12.
EAR PROBLEM
183(93.37)
3(1.53)
10(5.10)
196
13.
HYPERTENSION
DIABETES
256(100)
0
0
256
14.
PROBLEMS WITH THE
NERVOUS SYSTEM
105(90.52)
5(4.31)
6(5.17)
116
15.
JAUNDICE
51(100)
0
0
51
16.
BLEEDING FROM ANY SITE: NOSE BLEEDS, HEMORRHOIDS, INJURIES
96(88.07)
4(3.67)
9(8.26)
109
17.
URINARY PROBLEM
209(85.65)
15(6.15)
20(9.08)
244
18.
TOOTH PROBLEM
21(95.45)
0
1(4.54)
22
19.
HEADACHE
295(84.29)
28(8.0)
27(7.71)
350
20.
RESPIRATORY PROBLEMS
ASTHMA
BRONCHITIS
PULMONARY KOCHS
1949(97.55)
2(0.1)
47(2.35)
1998
21.
SKIN DISEASES
1202(84.64)
90(6. 34)
128(9.O1)
1420
TOTAL
8970(89.7)
398 (3.98)
632(6.32)
10000

Notes:
  1. Study population included 30% of children aged 0-14 years.
  2. Non-medical personnel manned 2 of the three sites; medical interns conducted the 3rd site testing
  3. The results are pooled over a four- month period covering three major hospitals outpatient departments. Figures in parentheses are percentages
  4. The software was continuously being modified through the pilot testing. Hence, more recent results are better than overall results shown.
  5. Blank and wrong diagnoses primarily due to: Incorrect entries
    Complicated cases
    Incomplete forms
    Patients sent for lab. tests which did not return

 

[ home | projects | news | volunteer | donate | contact ]