Background
The need for a surveillance system is often realized in face of health emergencies not allowing enough time to establish a formal surveillance system. The prohibitive costs involved in establishment and maintenance of surveillance systems quoted in literature along with paucity of detailed cost-effectiveness analyses is a commonly reported barrier to surveillance system establishment. Our study aimed to evaluate an innovative mechanism to extract actionable data from pre-existing records in resource limited settings.

Methods
Pre-intervention values were derived from 3 studies conducted by FELTP fellows on Pakistan Army’s existing disease surveillance system using CDC guidelines for evaluating public health surveillance systems, 2001. Post intervention analysis was done using same methodology 1 year after the establishment of a central data management body called COVID “CMC (Crisis Management Cell)” aimed to enhance surveillance efficiency of routinely collected hospital records. Effectiveness was measured by number of reports generated, cases detected, events reported, clusters identified, and by performing data quality audit. Set-up and operating costs were calculated.

Results
In 1 year, 742 daily and 37 weekly reports were generated and 87 clusters identified. Data guided 32 changes in health policy and practice and system adapted to changing needs 57 times. Chi squared test showed significant improvement in completeness (23% to 97%, p-value < 0.001), sensitivity (20% to 96%, p-value < 0.001) and PPV (90% to 99%, p-value < 0.001) a year after intervention. CMC incurred a total cost of Rs 22,337,600 (USD 136,080) over a period of 1 year and Rs 28,674 (USD 175) per report generated, Rs 563 (USD 3) per case reported and Rs 23 (USD 0.14) per variable recorded which was significantly less than that quoted by other studies.

Conclusion
Utilizing routinely collected hospital records is a cost-effective intermediate
step for countries / health systems in emergency situations as posed by the COVID-19
pandemic.