Well, it's a better title than "Duration of Stool Colonization in Patients Infected with Extended-Spectrum Beta-Lactamase-Producing Escherichia coli and Klebsiella pneumoniae." What the authors were looking at is how long E. coli and K pneumoniae (another opportunistic commensal that is found in people at low densities) that carry resistance genes that encode enzymes known as ESBLs (extended-spectrum beta-lactamases).
ESBLs confer resistance to many of the antibiotics that start with "cef-" (or "ceph-") or that end in "-cillin" making these drugs ineffective at treating these infections.
On average, a patient who had been in the hospital for a week or longer would shed (i.e., poop) ESBL-producing bacteria for 98 days. If patient were given antibiotics to take home, they would shed ESBL-producing bacteria for 154 days (the median for those not given antibiotics was 56 days).
The good news is that this rate of strain turnover (the length of time that a bacterial strain resides in a person) isn't longer than what we usually observe for most E. coli in healthy people (a few weeks to several months). The bad news is that this isn't shorter than what we see in healthy people either. That means the survival of these resistant strains might be largely dependent on features of their ecology other than antibiotic resistance. Once these E. coli leave the hospital, they might be as fit as any other E. coli, which means that responsible use of antibiotics, while absolutely necessarily, doesn't act very quickly.
Cited article: Apisarnthanarak A., T.C. Bailey, and V.J. Fraser. 2008. Duration of Stool Colonization in Patients Infected with Extended-Spectrum Beta-Lactamase-Producing Escherichia coli and Klebsiella pneumoniae. Clinical Infectious Diseases 46:1322-1323.