Characteristics of high-cost inpatients with peripheral artery disease

Inpatient treatment of peripheral artery disease (PAD) is more than six times as costly as that of the general inpatient population. Our objective was to describe factors associated with hospital cost for patients admitted for PAD, the characteristics of high-cost patients, and their outcomes including amputations and death.

We performed a retrospective cohort study of admitted patients receiving a procedure for PAD at The Ottawa Hospital between 2007 and 2016. Demographics, comorbidity, inpatient events, and hospital cost data during the index admission were collected. We defined high-cost patients as those whose total costs of index admission were in the tenth percentile and above. Features associated with high-cost status were examined using logistic regression with elastic net regularization. We used generalized linear models to examine overall drivers of cost.

We identified 3084 eligible patients, incurring $72.2 million in hospital costs. The mean cost of the most expensive 10% of patients was $88,076 (standard deviation, $54,720), more than five times the mean cost of $16,217 (standard deviation, $10,322) for nonhigh-cost patients. High-cost patients were more likely to present urgently (odds ratio [OR], 1.63; 95% confidence interval [CI], 1.16-2.25; P < .01). After adjustment for preadmission factors, high-cost patients were more likely to have experienced an adverse patient safety incident (OR, 13.49; 95% CI, 6.97-24.8; P < .01), amputation (OR, 2.79; 95% CI, 1.68-4.49; P <.01), intensive care unit admission (OR, 6.42; 95% CI, 3.62-10.2; P < .01), and disposition barriers requiring alternate level of care status (OR, 10.44; 95% CI, 6.42-15.2; P < .01). The high-cost group was more likely to have received hybrid revascularization (OR, 7.07; 95% CI, 3.34-13.6; P < .01). High-cost patients had higher than predicted in-hospital mortality (18% vs 9.2% predicted) compared with the low-cost group (3.0% vs 2.7%; P < .001), and fewer than half of high-cost patients were discharged home.

Providing hospital care for the top 10% most expensive patients in our cohort was more than five times as costly per patient than providing care for the nonhigh-cost patients. Whereas pre-existing factors may predispose a patient to require expensive care, there are potentially modifiable factors during the admission that could reduce costs of these patients.