Assessing the outcomes of implantable cardioverter defibrillator treatment in a real world setting: results from hospital record data.
BMC Health Serv Res 2013;
13:100. [PMID:
23496994 PMCID:
PMC3602059 DOI:
10.1186/1472-6963-13-100]
[Citation(s) in RCA: 6] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/28/2012] [Accepted: 03/08/2013] [Indexed: 11/10/2022] Open
Abstract
BACKGROUND
A plethora of clinical studies have assessed the benefits of implantable cardioverter defibrillators (ICDs) and supported their use in clinical practice. However, evidence on the safety and efficacy of ICDs appears insufficient to support expansion of their use in clinical practice, and more information on their impact in real life settings is warranted. This paper aims to investigate the impact of ICDs using a large administrative dataset reflecting actual clinical practice.
METHODS
Data were obtained from the hospital discharge database of the Friuli Venezia Giulia region in Italy containing patient-level information on 169,488 cases. Data on mortality outside hospital were obtained from regional sources. Exact matching method was used to estimate the outcomes associated with ICDs: mortality, length of stay, re-hospitalization and regional expenditure. The method was applied in two steps. First, patients with ICDs were matched with those without using the following: age class (by 5 years), gender, year of admission, type of admission (day hospital vs. ordinary) and primary diagnosis. In the second step, matching included also Charlson Comorbidities Index. Exact matching average treatment effect on the treated (ATT) was used as a main measure of impact.
RESULTS
Compared with matched controls, treatment with ICDs was associated with lower mortality (absolute risk reduction 10.6% at 1 year and 8.3% at 2 and 8.4% at 3 years, p < 0.001 and hazard ratio 0.80, p < 0.001), greater regional expenditure at index hospitalization (ATT: €9459.64, p < 0.001) and during follow up (ATT: €1707.29, p < 0.001) and higher re-hospitalization rate (ATT: 0.53, p < 0.001). No significant difference was found for length of stay (9.07 vs. 8.86 days). The results were maintained after more restrictive matching was applied.
CONCLUSIONS
Assessing the impact of innovative, expensive medical technologies on the basis of real world data is warranted, especially when there are barriers to implementation. Hospital administrative datasets can be of great value when a technology such as the ICD is implemented in a relatively small sample of patients, to allow use of exact matching techniques.
Collapse