TB R

1
TB RR System's strong points

• Standardized system implemented widely
• Collection of follow-up information on each
  patient over long course of treatment
• Data sources succinct, cleverly designed for
  tabulation, cross-referencing
• Definitions that are clear, serve both epi and
  clinical purposes
• Outcomes based on mutually exclusive irrevocable
  categories that are assessed by cohort
• Works as a completely paper-based system up to
  central level in developing countries
• Known benchmarks (e.g. proportion cases smear )

2
TB records treatment card
3
TB records patient register
Tool of DHO (not all of the data in TC)
1 row 1 course of treatment Does not "track" a
change in regimen (re-reg)
Handy use of multiple col. (probably ART pgm
cannot afford this luxury)
Ordered by date of reg.
Transfer-out outcome is really a subset of
unknowns (transfer-out cases with outcome
unknown). It is responsibility of the initial
registration (only) to report the outcome. The
receiving unit registers for mgmt/monitoring, but
record is disregarded when making cohort reports.

• Default outcome does not exactly say what
  happened to the patient and his episode of TB.
  Defaulters may or may not be lost patients. Or
  they may be known to be dead. Still, the default
  outcome is irrevocable. It is a point of closure
  on the current regimen.

4
TB records lab register
5
TB quarterly reports
6
TB quarterly reports
7
Implementation of district TB register2
scenarios
Centralized registration
De-centralized registration
Patient presents centrally (1), is examined and
diagnosed (2), registered (3) and takes observed
treatment at same site or (preferrable) at
peripheral facility near his home where treatment
card is kept (4).
Patient presents to peripheral facility and is
examined (1), sputum is sent to one of several
labs (2), patient begins treatment with treatment
card (3) at a later date, district supervisor
visits facility and registers case (4),
transcribing from treatment card.
8
Some notes about TB system

• Drug planning is relatively simple (few
  deviations from defined regimen) Can estimate
  monthly/quarterly needs from the case-finding
  report.
• Duration of follow-up relatively limited (6-9
  mo), and limited pieces of follow-up info
  (follow-up smears). Patients may travel to
  higher level registration/diagnostic center for
  follow-up smears.
• Outcomes are mutually exclusive and irrevocable
  events (e.g., once defaulted, end of story for
  that "case" because end of the road for the use
  of that regimen.)

9
Some notes about TB system (2)
Each arrow represents a patient's treatment
across time, with some outcome reached
Q3
Q2
Q1
Q4

• Cohort analysis common time period of initiating
  treatment (versus a common period when outcome is
  reached highly variable).
• Useful to monitoring trends in programme
  performance.
• Involves a long delay in assessment (to allow
  everyone to have a chance to finish their
  regimen), but works nicely in context of SCC (std
  duration).
• Cohort exclusions are few, so cohort N is stable

10
Issues for ART monitoring

• Substitutions and switches in ARV regimens are
  not uncommon.
• Implications (probably)
• Default (interruption) outcome may not have same
  significance / usefulness.
• Definition and significance of "failure"?
• Need frequent reports for drug planning (monthly)
• So register (and reporting) must be facility
  based.

11
Issues for ART monitoring (2)

• Treatment duration is forever only irrevocable
  outcome for the patient is death.
• Implication/Issues
• Might not make sense to monitor only outcome of
  the original regimen course (where 2nd regimen
  entails re-registration on new line in the
  register book). Rather, it might make sense to
  monitor outcomes of patients (up to a certain
  point 2 yrs?), and keep all info on one line in
  the register (no re-registration).
• Probably cannot expect initial registration unit
  to follow transferred patients outcomes
  quarterly for 2 yrs, so
• transferring could be a legitimate and final
  outcome, but then
• cohort N will become fuzzy due to transfer-in
  cases this would be a sacrifice in transparency
  at lower level OK at higher levels)

12
Issues for ART monitoring (3)

• Substitutions/switches are important
  status/events to monitor.
• Implication
• Substitution/switch (1st/2nd-line) events could
  be viewed as part of the outcome definition,
  e.g.,
• those alive and still on original regimen
• those alive and on substituted regimen
• those alive and on switched 2nd line regimen.

13
ART outcome analyses are tricky

• Possibilities "How many patients stopped
  treatment"
• In the most recent month? Cross-Sectional
• From Jan-Mar of this year? Cross-Sectional
• Ever (since beginning of pgm)? Cumulative
• As of their 6 month of treatment? Survival
  analysis
• Example of cohort analysis
• "How many patients entering ART during a given
  time period had a "stopped treatment" outcome one
  year after the close of that time period."
• Hand-outs suggest a minimum way to get
  cross-sectional and cohort data of interest to
  programmes using a paper-based system. (More
  data available in the register for computerized
  analyses).

This is a "transient" (not irrevocable) status
/ event / outcome.