Economic evaluation of laboratory testing strategies for hospital-associated Clostridium difficile infection

Abstract

Clostridium difficile infection (CDI) is the most common cause of infectious diarrhea in health care settings, and for patients presumed to have CDI, their isolation while awaiting laboratory results is costly. Newer rapid tests for CDI may reduce this burden, but the economic consequences of different testing algorithms remain unexplored. We used decision analysis from the hospital perspective to compare multiple CDI testing algorithms for adult inpatients with suspected CDI, assuming patient management according to laboratory results. CDI testing strategies included combinations of on-demand PCR (odPCR), batch PCR, lateral-flow diagnostics, plate-reader enzyme immunoassay, and direct tissue culture cytotoxicity. In the reference scenario, algorithms incorporating rapid testing were cost-effective relative to nonrapid algorithms. For every 10,000 symptomatic adults, relative to a strategy of treating nobody, lateral-flow glutamate dehydrogenase (GDH)/odPCR generated 831 true-positive results and cost $1,600 per additional true-positive case treated. Stand-alone odPCR was more effective and more expensive, identifying 174 additional true-positive cases at $6,900 per additional case treated. All other testing strategies were dominated by (i.e., more costly and less effective than) stand-alone odPCR or odPCR preceded by lateral-flow screening. A cost-benefit analysis (including estimated costs of missed cases) favored stand-alone odPCR in most settings but favored odPCR preceded by lateral-flow testing if a missed CDI case resulted in less than $5,000 of extended hospital stay costs and <2 transmissions, if lateral-flow GDH diagnostic sensitivity was >93%, or if the symptomatic carrier proportion among the toxigenic culture-positive cases was >80%. These results can aid guideline developers and laboratory directors who are considering rapid testing algorithms for diagnosing CDI.

FIG 1

Net cost per patient tested by the extra cost of an additional missed case. The net cost was calculated as the cost of initial testing, treatment, and isolation, plus the extra cost of additional missed cases (consisting of transmission of disease, extended hospital stay, and eventual CDI diagnosis and treatment) for each missed case occurring in the given testing strategy. The colored lines represent different testing strategies (rapid algorithms are dashed), with shaded areas representing 5th to 95th percentiles from the corresponding Monte Carlo simulations. Strategies with the lowest cost per patient tested (i.e., lower along the graph at any given vertical “slice”) are preferred. The incremental cost-effectiveness of each nondominated strategy (relative to the next best strategy) from Table 3 can be seen on the x axis, where the minimal cost lines intersect ($1,600 and $6,900). Using our model, the most likely value for the extra cost of a missed case is represented by the black line (median, $13,848), with the uncertainty range shaded gray (5th and 95th percentiles). GDH, glutamate dehydrogenase; EIA, enzyme immunoassay; Tox, C. difficile toxin A/B; odPCR, on-demand PCR.

FIG 2

Two-way sensitivity analysis: missed-case extended stay ($) versus transmission. The color in each area of this diagram depicts the economically preferred CDI testing strategy according to the cost of an extended hospital stay for a missed case on the y axis and the number of secondary transmissions per missed case on the x axis; the intensity of color reflects the savings per patient tested over the next best option. Here, the savings for lateral-flow GDH/odPCR are calculated relative to stand-alone odPCR, as there is significant uncertainty overlap between the two lateral-flow strategies. Three strategies are preferred at different magnitudes of the extended hospital stay (or other morbidities not specifically modeled) and transmission resulting from a missed case: stand-alone odPCR (red), lateral-flow GDH/odPCR (blue), and treat-none (green). Where lateral-flow GDH/odPCR testing is preferred, the next best option (stand-alone odPCR) is up to $80 more costly per patient tested, but where stand-alone odPCR is preferred, the next best option is up to $210 more costly per patient tested. CDI, C. difficile infection; GDH, glutamate dehydrogenase; Tox, C. difficile toxin A/B; odPCR, on-demand PCR.