Table 2. Calculated CICT effectiveness values and model-estimated CICT effectiveness before and after CICT protocol change, Philadelphia, Pennsylvania, USA, 2021*.

Characteristic	Period 1, before protocol change	Period 2, after protocol change
Calculated CICT effectiveness values
% Case-patients and contacts isolated because of CICT (range)†	17 (11.7–21.9)	10 (6.7–12.5)
Days from infection to isolation‡	9	8
Model-estimated CICT effectiveness
No. cases averted by CICT	657–968	1,156–1,609
No. hospitalizations averted by CICT	16–24	28–40
% Disease prevalence averted by CICT	8.4–12.0	6.8–9.2
Average staff hours per case averted§	21–30	8–11
Average staff hours per 1% disease prevalence averted¶	1,661–2,358	1,397–1,892

*CICT, case investigation and contact tracing. †Including contacts who later become case-patients. Calculated as follows using the observed performance metrics (Table 1), assumed compliance with isolation and quarantine guidance among cases and contacts (Appendix Table 1), and an assumed k = 1.2: [(% case-patients interviewed × compliance) + k × % contacts identified × (% contacts monitored × compliance + % contacts notified but not monitored × compliance)] / (1 + k), where k is approximated from the effective reproduction number (R_t), because undetected infected contacts will infect R_t additional persons on average. During the evaluation period, the average R_t in Philadelphia was 1.29 during periods 1 and 0.99 during period 2. If the assumed compliance was 100%, the estimated effectiveness could be as high as 26% for period 1 and 15% for period 2. ‡The average length of time from infection to isolation and quarantine between case-patients and contacts who later became case-patients. We assumed a 5-day presymptomatic period. We further assumed that interviewed case-patients and notified contacts began isolation and quarantine the day after their interactions with the health department (Appendix). §Calculated by dividing the total staff hours by the estimated number of cases averted by CICT. Lower value represents a more cost-effective program, given that it requires fewer staff hours to prevent each case. ¶Calculated by dividing the total staff hours by the estimated proportion of disease prevalence averted by CICT. Lower value represents a more cost-effective program, given that it requires fewer staff hours to prevent each percentage of disease prevalence.