Scene 1

Is the patient-morbidity data in routine health information systems ready to support the impending National Health Insurance in South Africa?

Nicol E , Hanmer L , Zitho A , Basera W, Mukumbang F, Bradshaw D

Presented at 6th Global Symposium on Health Systems Research, November 8-12, 2020

Scene 2

While the National e-Health Strategy appeals for  standardised , coded clinical records as a component of health information systems to support effective health services implementation,  The potential success of the National Health Insurance (NHI) also necessitates a reliable,  standardised  patient-level health information system that supports Diagnosis-Related Groupers (DRGs) for reimbursements and resource management.  To this end, discharge summaries, an overview of a patient’s  hospitalisation  from admission through discharge containing information relating to the cause of morbidity and treatment following discharge, are critical. South African National Health Act requires that a written DS report be provided to all in-patients at the time of discharge.

Scene 3

Despite the importance of DS as a source of morbidity data, no study in resource-limited settings has examined the availability of standard DS, assessed the completeness of, and agreement between DS and patient-level records.    STUDY AIM

We investigated  the availability of standard discharge summaries and coded clinical data, assessed the quality of inpatient health records, and the congruence between inpatient health records and discharge summaries in public sector hospitals to support the NHI implementation in terms of reimbursement and resource management.

Scene 4

Objective Sample  Method Measure  Data  availability  an d  quality 45 public hospitals ( 7 tertiary, 10 regional and 28 district hospitals )  in 10 NHI Pilot Districts Data was extracted from  5806 routine patient-level records   in public hospitals, from  5 treating depts,  using proportional  sampling   Surgical Medical Paediatrics O  & G Psychiatry  A cross-sectional and observational analytic study was  undertaken  to investigate the availability and quality of coded and codable clinical data in a sample  of  hospitals in NHI pilot districts Record reviews of registers , Patient files   & Discharge summary  ( 9 pre-defined data elements on diagnoses and procedures )   -  Wimsett  et al 2014; WHO, 2003 Accuracy :  comparing the source registers with the patient files, and the patient files with the discharge summaries for patients admitted in  March and July 2015 ;  and also the calculated patient age against reported age  Completeness :  assessing the proportion of  patient medical record  and discharge summaries that have all the expected required data fields completed by a clinical registrar (specialist in training), a general practitioner/medical officer or nursing staff. Data collection  period :   Aug 2016 – April 2019

Scene 5

Results: Availability of discharge summary BY

Scene 6

Results: Coding diagnosis data using ICD-10 by

100
90
80
70
60
50
40
30
20
10
100
EC
100
73
27
GP
70
100
KZNI KZN2
•Yes • No
97
74
26
97
NC
99
NW
41
59

Scene 7

Result: Accuracy of patient clinical records

Register vs Patient folders (n-
Variable
Patient identifier
Attending
physician's signature
Admission diagnosis
Discharge date
Agreement
970
850
952
96 0-980
84
94 6-951
Cohen's
Kappa (K)
0.93
0.70
0.90
0.62
0.92094
088-012
0.89-081
0.59064
-5, 795)
p•value
<O.OOI
<O.OOI
<O.OOI
<O.OOI
Strength of
agreement
Very good
Substantial
Very good
Moderate

Patient folders vs Discharge summary (n=3,767)
Variable
71.3-74 7
0.70-0 74
Patient age
95.5-96 7
0.91-0 93
Patient identifier
Attending physician's
82.9-87 2
0.67-0 75
signature
86.2-88 4
0.86-0 91
Admission diagnosis
78.8-83 9
0.57-0.63
Discharge date
95.2-97.2
0.90-0.94
Discharge diagnosis
Condition on
86.5-90 6
0.73-0 79
discharge
91 .4-95.2
0.84-0 89
Procedure(s)
87.1-895
0.72-0 76
Follow-up plan
Results of
84.0-88 1
0.69-0 74
Investigation
Cl
Agreement (%)
73.0
96.1
85.1
87.3
81.4
88.5
93.3
88.2
86.0
Cohen's
Kappa (K)
0.72
0.92
0.71
0.89
0.60
0.92
0.76
0.86
0.74
0.71
p-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
Strength of
agreement
Substantial
Very good
Substantial
Very good
Moderate
Very good
Substantial
Very good
Substantial
Substantial

Re istervs Patient folders vs Dischar
Variable
Patient identifier
Attending
physician's signature
Admission diagnosis
Cl
Agreement (%)
960
87.6
893
85.7
950-97.0
88 0-896
886-90M
84.8-86.6
Fleiss'
Kappa (K)
0.90
077
077
0.70
0.89-0.92
015-019
074-080
0.68-072
e summa
p-value
<O.OOI
<OOOI
<OOOI
<0.001
n-3,767
Strength of
agreement
Very good
Substantial
Substantial
Moderate

Scene 8

This study provides a situational analysis of the level of South Africa’s preparedness of public sector hospitals for the planned implementation of the National Health Insurance in terms of claims reimbursements. The practice of completing discharge summaries as part of the patient discharge process is suboptimal at all levels of public sector hospitals in all provinces in South Africa. The absence of coded inpatient diagnoses and identified data inaccuracies indicate that existing routine health information systems in public-sector hospitals in the NHI pilot districts are not yet able to support reimbursements and resource management.  Institutional capacity is needed to undertake diagnostic coding, improve data quality and ensure that a standard discharge summary is completed for every discharged inpatient health record.

Is the patient-morbidity data in routine health information systems ready to support the impending National Health Insurance in South Africa?

Page 1 (0s)

Page 2 (20s)

Page 3 (48s)

Page 4 (1m 13s)

Page 5 (1m 54s)

Page 6 (2m 4s)

Page 7 (2m 15s)

Page 8 (3m 15s)