CHAS IT error: A timeline of how it unfolded

CHAS IT error: A timeline of how it unfolded

CHAS website
The Community Health Assist Scheme (CHAS) website.

SINGAPORE: An error in the means-test system managed by IT firm NCS led to inaccuracies in the subsidies computed for healthcare schemes for Singaporeans, the Ministry of Health (MOH) said on Saturday (Feb 16).

The inaccuracies affected 7,700 people who applied for or renewed their Community Health Assist Scheme (CHAS) between Sep 18 and Oct 10. 

A majority of them, 6,400, received higher subsidies due to the error, but will not need to return the additional subsidies, MOH said. The others will be reimbursed.

MOH and NCS gave a breakdown of when the error was discovered and the events that followed.

Sep 16, 2018: The means-test system was migrated from one Government data centre to another.

Sep 18, 2018: First run of means-test for CHAS was performed. There was a software deployment error in one server. One of the files deployed was not the correct version.

Sep 24, 2018: The first means-test discrepancy was detected by CHAS processing team and reported to NCS. The NCS support applications support team investigated the issue, but was unable to pinpoint a root cause.

Oct 10 - Nov 29, 2018: The NCS deployment team investigated an unrelated slow performance issue and discovered a software version issue which had resulted in the failure of the means-test system to retrieve requisite income information in order to compute results. The issue was fixed immediately.

However, unknown to the team, while the fix prevented more means-test results being affected, it did not resolve the incorrect means-test results generated before the fix. Five additional cases were detected during this period.

Nov 29, 2018: After extensive investigations, NCS confirmed that the root cause of the means-test discrepancies was due to the software version issue which was earlier fixed. 

NCS developed scripts to extract the means-tests results that were computed on the affected server between Sep 18 and Oct 10. The firm also developed programmes to simulate the correct subsidies based on the initial set of income data to estimate the number of individuals and schemes affected.

Dec 5, 2018: Thousands of individuals and schemes besides CHAS were determined to be affected.

Jan 18, 2019: Measures were implemented to contain the impact on non-CHAS schemes.

Feb 10, 2019: The number of people affected, 7,700, was determined.

Feb 16, 2019: Subsidy tiers of affected individuals were fully corrected, and affected individuals will be progressively informed by scheme administrators and service providers. Reimbursements to be started.

Source: CNA/zl