1 of 45

Data Use Community Meeting - July 21, 2020

HIV/AIDS Treatment Retention Outcomes: Field Perspective on Data Use

Join the discussion on WhatsApp!

2 of 45

Welcome!

Before we begin:

  • This is a safe space for sharing our experiences and learning from each other
  • Please feel free to use the chat to speak amongst yourselves
  • The chat feature can be used to communicate with everyone or only the hosts (listed as “panelists” in this software).

Join the discussion on WhatsApp!

3 of 45

Q&A Feature (Question & Answer)

  • We will also be using the Q&A feature in this meeting:
    • Anyone can ask or answer questions, which will be visible to all participants.
    • Questions and answers can be given anonymously by selecting the “anonymous” box
    • Questions and answers can be upvoted

Join the discussion on WhatsApp!

4 of 45

Introductions

  • Hosts
    • Jennifer Shivers and Paul Biondich
  • Provocateurs
    • Dr. Violet Oramisi, NASCOP
    • Ms. Akipu Ehoche, NDR Data Manager and Statistician
  • Please introduce yourself in the chat:
    • Name
    • Organization
    • Country
    • Role

Join the discussion on WhatsApp!

5 of 45

HIV/AIDS Treatment Retention Outcomes: Experience from Kenya ��

Dr. Violet Oramisi

Strategic Information Team lead

National AIDS and STI Control Program, Kenya

6 of 45

The National Data Warehouse (DWH)��A data repository & analytics platform

As a data repository,

Stores monthly uploaded data from the EMRs in secure data marts located at the NASCOP servers

Has dashboards accessible openly to Program teams

Access controlled access to underlying data

As a data analytic & visualization platform,

Presents program data in interactive dashboards

Users can generate & run custom queries on the database to generate datasets for further analysis.

Has been used for preliminary analysis of MOH & PEPFAR priorities

A data source for HIV case based surveillance

7 of 45

HTS data

EMR

Testing

Care

Laboratory

Pharmacy

HTS app

HTS app

Key:

Data warehouse application interface client tool

DW

Encrypted PKV database

DWH Web Portal

Web Portal

HTS paper

Encrypted HIV clinical data

  • Matching
  • Merging
  • Deduplication

Contains both HIV- and HIV+

Data synchronization

Data entry from paper records

Central NDWH database

Records linking & merging

Data Analysis and Display Portal

High level data flow to DWH

  • Data Collected from EMR using DWAPI
  • Data Transmitted over internet into Staging area
  • ETLs

8 of 45

Overall care & treatment reporting rate to the DWH

8

9 of 45

EMR SitesA total of 1,114 sites with EMRs

eHTS Sites ��Total sites with eHTS 220 

Coverage of PLHIV as at December 2019

52% (548,822) of all ART Patients have records in DWH

72% (835,644) of all ART Patients were seen at a site with an EMR

73% of ART Patients seen at EMR Sites had records in DWH

10 of 45

�TWELVE MONTHS RETENTION: THE 2018 COHORT ANALYSIS

11 of 45

Twelve months HAART outcomes, 2018 (N=110,657)

12 of 45

Retention by county (N=110,657)

12

13 of 45

Overall retention (N=110,657)

13

14 of 45

Retention by Diagnosis to HAART initiation (N=110,657)

14

*Missing date of diagnosis (n=28,851 [26.1%])

15 of 45

HIV/AIDS CARE AND TREATMENT OUTCOMES �15 YEARS OF PEPFAR PROGRAM: THE KENYAN EXPERIENCE�

2004 – 2018

16 of 45

15 year Retention: Treatment outcomes, 2004-2018 (N=1,233,450)

*Outcomes as at 31st Dec 2018

17 of 45

Overall attrition (N= 1,233,450)*

17

*Outcomes as at 31st Dec 2018

Rate: 11.2/100 pyos

Follow up

Time

Beginning

Total

Attrition

Survivor function

0

1233450

0

1.000

3

644129

397289

0.660

6

351100

104871

0.5331

9

158440

0.438

0.429

12

45751

20101

0.3544

15

1766

2796

0.3125

18 of 45

Using a Sampling-Based Approach to Ascertain True Outcomes of Lost to Follow-Up Patients and Reasons for Patient Dropout among HIV Patients Enrolled in HIV Care and Treatment in Kenya

2004 – 2018

19 of 45

Project Objectives

    • To ascertain true outcomes in probability sample of LTFU patients and reasons for loss through intensive tracing in the community
    • To use ascertained outcomes to revise survival and retention estimates in the underlying clinic patient populations
    • To build capacity of facility staff on use of findings to improve data quality service delivery, site-level estimates, and strategies for patient retention
    • To assess the current practices on patient retention and measurement of LTFU

20 of 45

Method

  • Study carried out across 40 sites in Kenya between 1st – 30 September 2019 (included DICES)
  • All facilities in the country that reported a minimum of 450 patients on ART (as of December 31, 2016 according to DHIS2) were eligible for selection
  • The study population included patients who had made at least 1 clinic visit in the last 2 years (1st August 2017 – 30th August 2019) and had ever initiated ART (including those who had stopped ART for whatever reason)
  • LTFUs were pulled from those within Current Clinic Population (CCP) as per the definition and the sampling process followed
  • 150 (max) LTFUs were sampled per site and intensively traced by tracers
  • 3 main tracing strategies included chart /EMR data review, direct tracing of the patients or tracing of the patients via the informants

21 of 45

Overall Tracing Outcomes (n=3,203)

22 of 45

Disengagement Category and Top Five Reasons (n=481)

23 of 45

Silent Transfer Category and Top Five Reasons (n=999)

24 of 45

Summary and Recommendations

  • Individual level information systems enable tracking at granular levels ensuring we don’t leave anyone behind in the journey to 95-95-95
  • It is important to try ascertain the outcomes of clients labelled at LTFU
  • Tracing efforts can be relatively simple and generally at low costs to the facility
  • Majority of the tracing efforts can be done by contacting patients and their informants
  • The new WHO definition of LTFU patients will allow prompt detection and follow-up resulting in better treatment outcomes

25 of 45

Summary relevant to webinar….

  • What kinds of novel data approaches have you seen for evaluating patient retention rates?- Use of EMR and LTFU ascertainment
  • What essential information do you have or wish you had to help you most accurately understand patient retention?-Routine patient feedback
  • What environments have you learned about or seen that amazed you from a HIV patient management perspective? How did these environments leverage data effectively?-Under exploited digital space
  • In the last session, we spoke about the patient experience and how it can be improved to help HIV Lost to Follow-up. How do you use data to evaluate the patient experience?- Empower patients to take charge of own health, patient centric services and interventions

26 of 45

Acknowledgement

  • MOH counterparts at the national and county levels
  • USG: CDC, USAID and DOD
  • UN Family: WHO, UNAIDS, UNICEF
  • Palladium: KenyaHMIS II Project- National HIS Partner
  • University of California San Francisco: Global Health Programs- National Surveillance and M&E Partner
  • All Service delivery partners
  • The patient for trusting us to provide the services and adherence to the guidance provided towards attaining the 95-95-95

27 of 45

Nigeria Data Use Cases July 21, 2020

28 of 45

Background to NDR

  • The NDR is a unified data storage and analysis

facility for the Government of Nigeria with support from PEPFAR

  • The NDR is EMR agnostic with APIs for direct

data push from EMRs

  • It houses de-identified patient level data from EMRs for HIV treatment data
  • There are currently 2,088,467 patient records

on the NDR

29 of 45

Primary EMR Data Sources for the NDR

Nigeria Medical Records System (NMRS)

Lafiya Management Information System (LAMIS)

30 of 45

Data Use Cases

31 of 45

Data Quality Assurance

Timeliness

Weekly NDR Upload Tracker

Consistency

Key variables on client level line-list from NDR against line-list from IPs EMR across randomly selected health facilities.

Validity

Monthly Treatment and Retention

Indicators concurrence check using external source (SAS etc)

Completeness

Monthly NDR error analysis.

Inbuilt data validation checks on NDR platform

Monthly NDR Data Quality Assessment across different dimensions of data quality

*This slide is using Font Awesome

5

32 of 45

Data Use Case: Retention Analysis

  • Estimate the probability of retention at the end of each quarter
  • Just over a quarter of patients are lost after their first visit
  • The median LTFU time is 15 days, and the survival rate at the end of Q2 for all categories is above 90%

33 of 45

Kaplan Meier curve: Cohort analysis of patients that started ART Q4FY19

Figure 1: Kaplan Meier curve showing Time to LTFU by Age group (September 2019)

Figure 2: Kaplan Meier curve showing Time to LTFU by Age group (October 2019)

Figure 3: Kaplan Meier curve showing Time to LTFU by Age group (November 2019)

Figure 4: Kaplan Meier curve showing Time to LTFU by Age group (December 2019)

34 of 45

Data Use Case: Retention Analysis

35 of 45

Data Use Case: Regimen Analysis

36 of 45

Data Use Case: Regimen Analysis - TLD optimization

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

January 2020

May 2020

June 2020

February 2020

TLD

March 2020

Adult 1st Line

April 2020

Adult 2nd Line

37 of 45

Data Use Case: HIV Recent Infection Surveillance

HIV recent infection surveillance aims to achieve epidemic control

  • It provides real-time data about recent infection in order to identify hotspots of current HIV transmission.
  • Helps to streamline prevention efforts

and strategies

  • It identifies individuals with new infections so that treatment can be commenced early, linked to care and eventually break HIV transmission.

38 of 45

Data Use Case: Case-based Surveillance (CBS)

HIV Infection

-1st Positive confidential Confirmatory Test

1st CD4+ T-

Cell Count

1st Viral Load

Test

1st AIDS

defining Lab (<200 CD4+

Cell Count or

OI)

ART Regimen

CD4+and VL monitoring according to regulations

Death

Sentinel events

=

Case-based Surveillance

+

HIV diagnosis

“HIV case-based surveillance, is the ongoing, systematic collection, analysis, interpretation, and dissemination of information about persons in whom HIV and AIDS are diagnosed.”

39 of 45

Data Use Case: Mortality Surveillance

ART Clients

Missed Appointment/LTFU

Dead

Conduct Verbal Autopsy

Analyze VA

Instrument

Determine/Assign Cause of Death and enter into EMR/NDR

Analyze: Determine proportion HIV related mortality et. al

Report

40 of 45

HIV Infection

AIDS

Death

Mortality Surveillance

EMR

Viral Load Data

Care/case management Data

Linkage ART data

HTS data

Recency

Surveillance

95

95 95

Diagnosis Treatment Viral Load

Data integration into Data-to-Care activities:

NDR/CBS

  • Identify, Investigate, Locate, Link/Re-Link

Unique ID

95

95

41 of 45

Acknowledgements

Special thanks to:

  • Federal Ministry of Health of Nigeria
  • PEPFAR
  • PEPFAR implementing partners
  • Staff and patients

The NDR is supported by #NU2GGH001976

42 of 45

THANKS

Do you have any questions?

43 of 45

Field Perspectives on Data Use

  1. What kinds of novel data approaches have you seen for evaluating patient retention rates?
  2. What essential information do you have or wish you had to help you most accurately understand patient retention?
  3. What environments have you learned about or seen that amazed you from a HIV patient management perspective? How did these environments leverage data effectively?
  4. In the last session, we spoke about the patient experience and how it can be improved to help HIV Lost to Follow-up. How do you use data to evaluate the patient experience?

Join the discussion on WhatsApp!

44 of 45

Staying in Contact

Join our mailing list for additional information and to learn about future meetings: https://ohie.org/duc/

Participate in ongoing moderated discussions on Data Use and Lost to Follow-up in our WhatsApp group: qrco.de/Data-Use-Community (or scan the code with your smartphone) to join

Join the discussion on WhatsApp!

45 of 45

Join Us.

The Data Use Community

duc@ohie.org�ohie.org/duc