Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

The impact of methicillin-resistant S. aureus on length of stay, readmissions and costs: a register based case-control study of patients hospitalized in Norway

The impact of methicillin-resistant S. aureus on length of stay, readmissions and costs: a... Background: Patients with methicillin-resistant S. aureus (MRSA) are thought to incur additional costs for hospitals due to longer stay and contact isolation. The aim of this study was to assess the costs associated with MRSA in Norwegian hospitals. Methods: Analyses were based on data fromSouth-Eastern Norway for the year 2012 as registered in the Norwegian Surveillance System for Communicable Diseases and the Norwegian Patient Registry. We used a matched case-control method to compare MRSA diagnosed inpatients with non-MRSA inpatients in terms of length of stay, readmissions within 30 days from discharge, as well as the Diagnosis-Related Group (DRG) based costs. Results: Norwegian patients with MRSA stayed on average 8 days longer in hospital than controls, corresponding to a ratio of mean duration of 2.08 (CI 95%, 1.75–2.47) times longer.A total of 14% of MRSA positive inpatients were readmitted compared to 10% among controls. However, the risk of readmission was not significantly higher for patients with MRSA. DRG based hospital costs were 0.37 (95% CI, 0.19–0.54) times higher among cases than controls, with a mean cost of EUR13,233(SD 26,899) and EUR7198(SD 18,159) respectively. Conclusion: The results of this study indicate that Norwegian patients with MRSA have longer hospital stays, and higher costs than those without MRSA. Keywords: Methicillin-resistant S. aureus, Costs, Inpatient, Diagnosis-related group, Length of stay, Readmission Background Communicable Diseases (MSIS). The proportion of invasive Staphylococcus aureus is considered to be one of the S. aureus isolates in Norway that were methicillin-resistant most common pathogens causing nosocomial infections was only 1.3% in 2012 [3], and Norway is considered to be [1]. The bacterium is a normal inhabitant in the human amongst the European countries with the lowest percentage body, found persistent on the skin or mucosa of 20% of of MRSA in S. aureus clinical isolates. adults [2]. It is however associated with several types of The spread of MRSAin Norwegian hospitals is con- illnesses, most commonly skin and soft tissue infections, trolled through screening routines, which consist of testing although it may also cause more severe infections [1]. patients before admission if they have previously tested In Norway, all diagnosed cases of MRSA are manda- positive or are suspected of having been exposed to MRSA torily notified to the Norwegian Surveillance System for in the 12 months prior. The current MRSA-strategy in Norwegian hospitals consists of isolating suspected and * Correspondence: [email protected] confirmed MRSA positive patients, work restrictions for Norwegian Institute of Public Health, Marcus Thranesgate 2, 0473 Oslo, healthcare personnel who test positive for MRSA, and Norway Department of Antibiotic resistance and Infection control, Norwegian decolonization of carriers [4]. Some studies have explored Institute of Public Health, Oslo, Norway the cost-effectiveness of drugs used to treat MRSA Full list of author information is available at the end of the article © The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 2 of 8 infections and costs related to MRSA screening in Norway the Norwegian Surveillance System for Communicable [5, 6].To our best knowledge, there are currently no stud- Diseases (MSIS). This region represents approximately ies estimating the economic impact of MRSA positive 56% of the Norwegian population, including the capital patients based on LOS, readmissions or DRG costs in city of Oslo, and comprises 18 hospitals. Using the Norwegian hospitals. unique national identification number assigned to each The aim of our study was to estimate the resource use Norwegian citizen, information from both the NPR and associated with MRSA in Norwegian hospitals in terms MSIS databases could be linked in order to extract the of length of stay, readmissions within 30 days and DRG- MRSA status of inpatients that had been tested for based costs, in order to provide a better knowledge base MRSA. Patients who have undergone screening because for infection prevention strategies. they met at least one of the screening criteria, but had a negative MRSA-test, are not registered in MSIS. Ethical Methods approval was obtained from the South-Eastern Regional This was a register-based matched case-control study Committee for Medical and Health Research Ethics with patients admitted to hospital during the year 2012. (REC): 2013/1004/REK. We used data from the South-Eastern Norway health The dataset included individual patients, at all levels of region for the period January 1 to December 31, 2012, care (inpatient, outpatient, day patient), aggregated by epi- provided by the Norwegian Patient Registry (NPR) and sode of care. We identified inpatient stays and defined Fig. 1 Sample selection flowchart. Cases and controls were selected from a merged dataset including all patients who had received health care services in hospitals of the South-Eastern Norway health region (provided by the Norwegian Patient Registry) and all notified cases of methicillin-resistant S. aureus (provided by the Norwegian Surveillance System for Communicable Diseases) during the year 2012. Only patients who were diagnosed MRSA- positive at the earliest eight days prior or during their hospital stay, were included as cases. In total, 82 (86%) cases were matched to four control patients; the rest had less than four controls per case Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 3 of 8 those hospital stays by aggregating all episodes of care Table 1 Description of included patients with continuous dates (Fig. 1). Only the first hospital stay Cases (n = 95) Controls (n = 346) recorded for 2012 was considered in the analysis. MRSA- Variables n (%) n (%) cases were defined as patients diagnosed with MRSA at Male gender 49 (52%) 179 (52%) the earliest eight days prior to hospitalization, or in the Age course of their hospital stay. All patients diagnosed with 0–9 16 (17%) 64 (18%) MRSA in Norway are offered treatment before admission 10–19 3 (3%) 9 (3%) to hospital. This treatment together with microbiological 20–29 5 (5%) 20 (6%) analyses, takes a minimum of nine days. By restricting the case definition to a positive sample within this time frame, 30–39 8 (8%) 29 (8%) we maximized certainty that cases were MRSA positive 40–49 9 (9%) 28 (8%) during the hospital stay. Patients missing DRG codes, 50–59 14 (15%) 44 (13%) MRSA positive patients diagnosed more than eight days 60–69 14 (15%) 56 (16%) prior to hospitalization, and those patients without an 70–79 14 (15%) 51 (15%) exact match were excluded from the analysis. The flow of 80–89 11 (12%) 41 (12%) patient inclusion and exclusion is presented in Fig. 1.Con- trols were selected from the group of patients not > 89 1 (1%) 4 (1%) diagnosed with MRSA. Comorbidity index Cases were randomly matched, one to four controls, 68 (72%) 251 (73%) based on 10-year age groups, gender, hospital, surgery, 1 11 (12%) 39 (11%) whether they were diagnosed with an infection or not dur- 2 13 (14%) 44 (13%) ing the stay and Charlson Comorbidity Index (CCI) [7] 3 2 (2%) 8 (2%) (Table 1). For 13 cases (14%) it was not possible to obtain four controls; among these, one case was matched with 8 1 (1%) 4 (1%) three controls, three with two controls and nine with one Hospitals (anonymized) control patient. Among those patients who were A 20 (21%) 74 (21%) transferred between hospitals, the matching was done B 18 (19%) 70 (20%) based on the hospital where the highest DRG cost C 8 (8%) 32 (9%) weight was registered. We used the DRG codes indicating D 8 (8%) 26 (8%) surgical operations to identify patients who underwent surgery. In addition, we performed analyses of subgroups E 7 (7%) 28 (8%) where we combined ICD-10 codes (registered during F 6 (6%) 23 (7%) a hospital stay) with data from MSIS. The MSIS data G 6 (6%) 22 (6%) include MRSA diagnosis (colonization or infection), H 4 (4%) 16 (5%) sample material and clinical picture at the time of I 4 (4%) 13 (4%) notification (Table 2). J 4 (4%) 13 (4%) The comorbidity index was estimated for each patient according to the method developed for economic evalu- K 3 (3%) 12 (3%) ations by Charlson et al., based on diagnoses as de- L 2 (2%) 8 (2%) scribed by Quan et al. [8, 9]. The Canadian version of M 2 (2%) 2 (1%) the International Coding of Disease 10th revision (ICD- N 1 (1%) 4 (1%) 10) used by Quan et al., was validated against the Nor- O 1 (1%) 2 (1%) wegian ICD-10 codes used in this study. Due to data P 1 (1%) 1 (0%) availability, we used all the ICD-10 codes registered in 2012 to define CCI instead of using previous years’ diag- Infections noses. This included diagnoses registered at both in- No infections 42 (44%) 185 (53%) patient and outpatient episodes of care. Infection 53 (56%) 161 (47%) The outcome variables were length of stay (LOS) of Non-severe infections 31 (33%) 100 (29%) the first hospitalization, readmissions within 30 days Severe infections 15 (16%) 50 (14%) from discharge, and costs based on the highest Procedure Diagnosis-RelatedGroup (DRG)costweightduring the first hospital stay. Estimation of readmissions Surgery 17 (18%) 48 (14%) within 30 days was restricted to patients discharged before December 1, 2012. Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 4 of 8 Table 2 Infections based on ICD-10 codes recorded for cases Results and controls during the first hospital stay The dataset included a total of 948,266 patients. Of Infections Infection types (ICD-10 codes) those, 318,593 were inpatients (Fig. 1). Before matching, 4 patients missing DRG codes and 120 MRSA positive All Intestinal infectious diseases (A04; A08); Tuberculosis (A15); patients who were diagnosed with MRSA more than Sepsis (A41); eight days prior to hospitalization were excluded. Erysipelas (A46); Additionally, 17 patients who met the inclusion criteria Bacterial infection of unspecified site (A49); Human immunodeficiency virus disease (B23); did not have an exact match with other patients and Cytomegaloviral disease (B25); were excluded from the analysis.In total, 95 MRSA posi- Acute and subacute endocarditis (I33); tive patients met the inclusion criteria for cases, and Acute upper respiratory infections (J03; J06); Pneumonia/ other acute lower respiratory were matched with 346 MRSA negative controls. Gender infections (J15; J18; J20; J21); was equally represented in both groups (males Abscess of anal and rectal regions/ peritonitis 52%).Overall, 75% of the study population consisted of (K61; K65); Infections of the skin and subcutaneous tissue children less than 10 years and adults over 50 years of (L02 – L04; L08); age. The majority (84%) had a CCI of <2. In total, 65 Infectious arthropathies (M02); (15%) of the patients underwent surgery while admitted Necrotizing fasciitis (M726); Osteomyelitis (M86); (Table 1). Based on the type of hospital, 44 (10%), 301 Acute tubulo-interstitial nephritis (N10); (68%) and 98 (22%) patients were hospitalized in teach- Urinary tract infection, site not specified (N390); ing hospitals, general hospitals with more than 250 beds, Inflammatory disorders of breast/ infections of breast associated with childbirth (N61; O91) and local hospitals with less than 250 beds, respectively. Infection following a procedure, not elsewhere The median LOS was 6 (IQR 3.00–15.00) days for classified or due to internal fixation device MRSA positive patients and 4 (IQR 2.00–7.00) days for (T814; T846) MRSA negative patients, while the means were 14 (SD Infections possibly caused by S. aureus 17.24) and 6 (SD 9.61) days, respectively.The greatest Non-severe infections Bacterial infection of unspecified site (A49); difference in LOS between cases and controls was ob- Acute tonsillitis, unspecified (J039); Acute upper respiratory infections of multiple served in the subgroup with no infections and the group and unspecified sites (J06); diagnosed with severe S. aureus infections (Table 3). The Abscess of anal and rectal regions (K61) distribution of LOS is shown in Fig. 2. Cutaneous abscess, furuncle and carbuncle (L02); Cellulitis (L03); A higher proportion of cases were readmitted within Other local infections of skin and subcutaneous 30 days compared to controls, but overall there was no tissue, unspecified (L08); statistically significant difference regarding readmissions Urinary tract infection, site not specified (N390); Inflammatory disorders of breast/ Infections of (Table 4). breast associated with childbirth (N61; O91) In total, cases had higher DRG-based costs than con- Severe infections Other sepsis (A41); trols, with a median of EUR5174(IQR 3347–9059) and Acute and subacute endocarditis (I33); EUR3762 (IQR 2405–6669), respectively. In the sub- Bacterial pneumonia, unspecified (J159); group of patients with infections, the median costs were Pneumonia, organism unspecified (J18); Necrotizing fasciitis (M726); almost equal, although among patients with severe infec- Osteomyelitis (M86); tions the mean costs were significantly higher for cases Acute tubulo-interstitial nephritis (N10); than controls (Table 5). The distribution of DRG based Infection following a procedure, not elsewhere classified or due to internal fixation device costs is shown in Fig. 3. (T814; T846) Discussion All statistical analyses were performed in Stata Statis- The results of this study indicate that Norwegian pa- tical Software (StataCorp, 2013). The distribution of tients with MRSA had longer hospital stays, and higher patients in each group were presented as proportions. costs than those without MRSA. We did not find statis- We estimated median in addition to mean for the out- tically significant differences concerning readmission- come variables since the data were not normally distrib- s.Our study is the first in Norway addressing the uted. Given a matching ratio of 1:4 and thus data on economic burden of MRSA in hospitals, and one of few both group and individual level, we measured the out- studies that estimate the costs of MRSA in general, not comes using Multi-level Mixed effect Regression models. only the life threatening MRSA related conditions. Having skewed distributions (Figs. 2 and 3), we analyzed We found that MRSA positive patients stay longer in LOS using negative binomial regression, readmissions the hospital than their controls, and this is statistically with both logistic and linear regression, and costs with significant both for the total study population and sub- log of the linear regression outcome [10]. groups. Previous studies have shown that patients with Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 5 of 8 Fig. 2 Distribution of LOS for cases and controls based on the first hospital stay in 2012. The number of days spent in hospital for the first hospital stay registered in 2012 showed that patients not diagnosed with methicillin-resistant S. aureus were more likely to have shorter hospital stays than patients diagnosed with methicillin-resistant S. aureus. LOS, length of stay severe MRSA infections have longer hospital stays com- delayed examination and treatment of patients, as well pared to patients with severe MSSA infections [11–14]. as poorer follow up (less supervision, worsening of con- A result of particular interest in this study was that we dition due to delayed treatment) [15]. found the largest difference in LOS among patients Readmission within 30 days from discharge is used as without infections. In Norway, this result may be a con- an indicator of quality of hospital care in Norway. We sequence of the comprehensive infection control mea- observed few readmissions overall, and even fewer in sures applied in hospitals, such as single room isolation patients with severe infections. The results showed only of patients with MRSA. Some studies have addressed the small and not statistically significant differences in risk possible negative consequences of single room isolation of readmission among cases and controls. Interestingly, in hospitals; for example isolation may contribute to although our results indicated that isolation adds to Fig. 3 Distribution of DRG based costs for cases and controls during their first hospitalization in 2012. DRG cost weights are the basis for a considerable share (40%) of hospital reimbursement in Norway, and hence, an important proxy for costs per patient, however inaccurate. Due to limited access to micro costing data for patients in this study, DRG cost weights were used to estimate costs per hospital stay, although these costs are potentially an underestimate of the actual economic burden caused by MRSA to hospitals Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 6 of 8 Table 3 Univariate analysis of LOS distributed by infectious condition Length of stay Cases (n = 95) Controls (n = 346) Negative binomial regression Median (IQR) Mean (SD) Median (IQR) Mean (SD) Ratio of mean duration 95% CI p-value All patients (n = 95/441) 6.00 (3.00–15.00) 13.71 (17.24) 4.00 (2.00–7.00) 6.05 (9.61) 2.08 1.75–2.47 <0.05 � No infections (n = 42/227) 9.50 (3.00–23.00) 14.36 (13.87) 3.00 (2.00–6.00) 5.75 (9.03) 2.48 1.91–3.21 <0.05 � Any infection (n = 53/214) 5.00 (3.00–13.00) 13.19 (19.62) 4.00 (2.00–7.00) 6.40 (10.25) 1.58 1.26–1.98 <0.05 - Non-severe infections (n = 31/131) 5.00 (3.00–9.00) 8.94 (19.27) 3.00 (2.00–5.00) 5.18 (10.84) 1.40 1.02–1.92 <0.05 - Severe infections (n = 15/65) 14.00 (5.00–36.00) 21.60 (20.69) 5.00 (4.00–12.00) 8.88 (9.70) 1.86 1.33–2.62 <0.05 IQR Interquartile range, SD Standard deviation, CI Confidence interval prolonged hospital stay, we did not find that isolation 2012 EUR 856 per day [5].In addition, our results showed leads to poorer quality of medical treatment as measured that MRSA positive patients had longer hospital stays than by the risk of readmission. The results must be inter- patients without MRSA. The average cost of a patient per preted cautiously as the sample is quite small. day in Norwegian hospitals in 2012 was estimated to be Median and mean costs were higher for cases than EUR 5368 [17]. controls overall. However, the median costs in the sub- The strengths of this study were the quality and accur- groups of patients with infections, were approximately acy of the dataset utilized for analyses. The dataset in- equal. Claudia Hubner et al. used an average daily reim- cluded the actual LOS for each episode of hospitalization bursement per bed based on the actually claimed G- for all patients in the health region. Norwegian health DRG in the patient record, to estimate opportunity cost registries provide information on all the patients, their of blocked beds. They also accounted for empirical costs unique encrypted identification number allows tracking related to hygienic measures and laboratory use, and them as well as linking data from different health regis- found MRSA-attributable costs of EUR 8673 per case tries. Therefore, despite a small number of cases, it gives a [16].The DRG system is not designed to estimate real complete overview of the diagnosed MRSA cases in the costs of each individual patient, but specific information health region. The Norwegian surveillance system con- regarding the patient’s age, gender, diagnoses and treat- tains all patients who have been diagnosed with MRSA, ments, is used to categorize patients into cost groups. regardless of where the test is taken. However, most pa- We aimed to get cases and controls as similar as pos- tients in Norwegian hospitals are not routinely screened- sible, where MRSA was the only distinguishing factor. for MRSA, which may lead to an underestimation of the Most cases and controls have, consequently ended up in true MRSA rate and undetected MRSA colonization in the same cost groups. None of the cases in this study controls. The distributions of LOS and costs (Figs. 2 and were registered with DRG codes specifically reflecting 3) show distributions heavily skewed to the right, with a the finding of resistant bacteria. mean close to zero. Based on the distribution of the out- Taking into consideration that the DRG system does come variable in the study population, it is likely that not account for MRSA status and the additional costs of MRSA positive patients among controls have ended up infection control measures, such as screening and isola- close to the mean value, and hence would not have signifi- tion of MRSA patients, the results in this study are most cantly influenced the results. However, if any MRSA posi- likely an underestimate of the actual cost differences.The tive controls are outliers, they have consequently masked Norwegian infection control measures are comprehen- a larger difference between cases and controls. sive and costly, and they are applied to both colonized Two other European studies have used DRG cost and infected patients. Jinshuo Li estimated that the cost of weights in addition to other costs [14, 16]. Healthcare resources associated with isolation were approximately resource utilization measurements that are not included Table 4 Univariate analysis of readmissions within 30 days distributed by infectious condition Readmissions within 30 days Cases (n = 74) Controls (n = 299) Logistic regression Linear regression n (%) n (%) OR 95% CI p-value Coef. 95% CI p-value All patients (n = 74/299) 10 (14%) 22 (10%) 1.50 0.65–3.46 0.34 0.04 −0.04 – 0.12 0.34 � No infections (n = 30/144) 4 (13%) 15 (13%) 1.02 0.31–3.32 0.98 0.00 −0.13 – 0.14 0.98 � Any infection (n = 44/155) 6 (14%) 7 (6%) 3.44 0.79–15.00 0.10 0.08 −0.01 – 0.17 0.07 - Non-severe infections (n = 25/83) 4 (16%) 1 (2%) 10.86 1.15–102.77 <0.05 0.14 0.04–0.25 <0.05 - Severe infections (n = 13/57) 2 (15%) 5 (11%) 1.76 0.19–16.67 0.62 0.04 −0.13 – 0.21 0.62 OR odds ratio, CI confidence interval Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 7 of 8 Table 5 Univariate analysis of DRG based costs distributed by infectious condition Reimbursement (EUR) Cases (n = 95) Controls (n = 346) Linear regression with log (reimbursement) as outcome Median (IQR) Mean (SD) Median (IQR) Mean (SD) Coef 95% CI p-value All patients (n = 95/441) 5174 (3347–9059) 13,233 (26,899) 3762 (2405–6669) 7198 (18,159) 0.37 0.19–0.54 <0.05 � No infections (n = 42/227) 6423 (3603–8445) 13,208 (26,850) 3680 (2103–6669) 7797 (22,619) 0.48 0.22–0.74 <0.05 � Any infection (n = 53/214) 4560 (3148–11,178) 13,252 (27,195) 4355 (3148–6326) 6511 (11,033) 0.21 −0.03 – 0.45 0.09 - Non-severe infections (n = 31/131) 3450 (2523–5794) 8065 (18,243) 3347 (2523–5794) 5299 (12,802) 0.09 −0.25 – 0.44 0.60 - Severe infections (n = 15/65) 7600 (5794–21,931) 26,904 (41,802) 7600 (4560–7600) 9300 (7268) 0.47 0.15–0.79 <0.05 IQR Interquartile range, SD Standard deviation, CI Confidence interval, EUR Euro, Exchange rate NOK-EUR7.4805 average for 2012 in the hospital reimbursement DRG, were not available Authors’ contributions AESA contributed to the analysis of results, research and writing of the through the registry data used in our study. For this rea- manuscript. CMJ contributed with research and editing the manuscript. BFdB son, we do not know the total cost and resource use of was a great contributor to the methodology in this study. RW quality this population. Considering that MRSA-positive pa- assured statistical analyses. ISK contributed with supervision of the work and recommendations regarding formulations and statistical analyses. PE tients are subject to very strict infection control routines performed the data linkage of the NPR and MSIS datasets, and prepared the in Norway and have extended LOS compared with data for analysis. He contributed to the analysis of results and manuscript controls, the cost difference between cases and controls writing. He was the main supervisor of the work. All authors agreed on the methods and statistical analyses, read and approved the final manuscript. is potentially greater.Norwegian health registries are intended for administrative purposes and are not de- Ethics approval and consent to participate signed for research, therefore, although we used registry Ethical approval was obtained from the South-Eastern Regional Committee data of high quality in our study, we were limited in our for Medical and Health Research Ethics (REC): 2013/1004/REK. scope of analysis by the measurements available. The Consent for publication study design implies that there are uncontrolled risk fac- Not applicable. tors. We attempted to control for potential confounders thought to be relevant by matching. Competing interests The authors declare that they have no competing interests. Conclusion This study shows that MRSA contributes to longer Publisher’sNote length of stay in hospitals and higher costs based on Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. DRG cost weights. It gives an improved knowledge base with regards to the consequences of resistant bacteria, and Author details may help Norwegian policy makers to make informed Norwegian Institute of Public Health, Marcus Thranesgate 2, 0473 Oslo, Norway. Institute of Health and Society, University of Oslo, Norway, decisions concerning resource allocation, infection pre- Forskningsveien 3A, 0373 Oslo, Norway. Department of Biostatistics, Oslo vention programs and guideline development. However, Centre for Biostatistics and Epidemiology, Institute of Basic MedicalSciences, in order to get a better understanding of the costs related University of Oslo, Sognsvannsveien 9, 0372 Oslo, Norway. Department of Antibiotic resistance and Infection control, Norwegian Institute of Public to resistant bacteria, micro-cost studies are needed. Health, Oslo, Norway. Abbreviations Received: 29 March 2017 Accepted: 27 June 2017 CCI: Charlson comorbidity index; DRG: Diagnosis-related group; ICD- 10: International coding of disease 10th revision; LOS: Length of stay; MRSA: Methicillin-resistant Staphylococcus aureus; MSIS: Norwegian surveillance system for communicable diseases; NPR: Norwegian patient References registry; REC: South-Eastern regional committee for medical and health 1. Lowy FD. Staphylococcus aureus infections. N Engl J Med. 1998;339(8):520–32. research ethics 2. Kluytmans J, van Belkum A, Verbrugh H. Nasal carriage of Staphylococcus aureus: epidemiology, underlying mechanisms, and associated risks. Clin Funding Microbiol Rev. 1997;10(3):505–20. This work was supported by the Norwegian Institute of Public Health, 3. European Centre for Disease Prevention and Control. Antimicrobial Department of Antibiotic Resistance and Infection Prevention. resistance surveillance in Europe 2012. Stockholm: ECDC; 2013. 4. Elstrøm P, Kacelnik O, Bruun T, Iversen BG, Hauge SH, Aavitsland P. Availability of data and materials Meticillin-resistant Staphylococcus aureus in Norway, a low-incidence The data that support the findings of this study are available from the country, 2006-2010. J Hosp Infect. 2012;80(1):36–40. Norwegian Patient Registry (NPR) and Norwegian Surveillance System for 5. Li J, Ulvin K, Biboh H, Kristiansen IS. Cost-effectiveness of supplementing a Communicable Diseases (MSIS), but restrictions apply to the availability of broth-enriched culture test with the Xpert meticillin-resistant Staphylococcus these data, which were used under license for the current study, and so are aureus (MRSA) assay for screening inpatients at high risk of MRSA. J Hosp not publicly available. Data are however available from the authors upon Infect. 2012;82(4):227–33. reasonable request and with permission of the South-Eastern Regional 6. Nguyen TC. [En legemiddeløkonomisk analyse av antibiotikabruk ved Committee for Medical and Health Research Ethics (REC). invasive MRSA-infeksjoner] [Norwegian]. Tromsø: University of Tromsø; 2009. Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 8 of 8 7. Charlson M, Wells MT, Ullman R, King F, Shmukler C. The Charlson comorbidity index can be used prospectively to identify patients who will incur high future costs. PLoS One. 2014;9(12):e112479. 8. Charlson ME, Charlson RE, Peterson JC, Marinopoulos SS, Briggs WM, Hollenberg JP. The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients. JClin Epidemiol. 2008;61(12): 1234–40. 9. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, Saunders LD, Beck CA, Feasby TE, Ghali WA. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005; 43(11):1130–9. 10. Austin PC, Rothwell DM, Tu JV. A comparison of statistical modeling strategies for analyzing length of stay after CABG surgery. Health Serv Outcomes Res Method. 2002;3(2):107–33. 11. Anderson DJ, Kaye KS, Chen LF, Schmader KE, Choi Y, Sloane R, Sexton DJ. Clinical and financial outcomes due to methicillin resistant Staphylococcus aureus surgical site infection: a multi-center matched outcomes study. PLoS One. 2009;4(12):e8305. 12. Cosgrove SE, Qi Y, Kaye KS, Harbarth S, Karchmer AW, Carmeli Y. The impact of methicillin resistance in Staphylococcus aureus bacteremia on patient outcomes: mortality, length of stay, and hospital charges. InfectControl Hosp Epidemiol. 2005;26(2):166–74. 13. de Kraker ME, Davey PG, Grundmann H. Mortality and hospital stay associated with resistant Staphylococcus aureus and Escherichia coli bacteremia: estimating the burden of antibiotic resistance in Europe. PLoS Med. 2011;8(10):e1001104. 14. Macedo-Viñas M, De Angelis G, Rohner P, Safran E, Stewardson A, Fankhauser C, Schrenzel J, Pittet D, Harbarth S. Burden of methicillin- resistant Staphylococcus aureus infections at a Swiss University hospital: excess length of stay and costs. J Hosp Infect. 2013;84(2):132–7. 15. Morgan DJ, Diekema DJ, Sepkowitz K, Perencevich EN. Adverse outcomes associated with contact precautions: a review of the literature. Am J Infect Control. 2009;37(2):85–93. 16. Hübner C, Hüber NO, Hopert K, Maletzki S, Flessa S. Analysis of MRSA- attributed costs of hospitalized patients in Germany. Eur J Clin Microbiol Infect Dis. 2014;33:1817–22. 17. Norwegian Directorate of Health. Samdata spesialisthelsetjenesten 2012. Oslo: Norwegian; 2013. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Antimicrobial Resistance & Infection Control Springer Journals

The impact of methicillin-resistant S. aureus on length of stay, readmissions and costs: a register based case-control study of patients hospitalized in Norway

Download PDF
8 pages

Loading next page...
 
/lp/springer-journals/the-impact-of-methicillin-resistant-s-aureus-on-length-of-stay-26x6gZF7ID

References (17)

Publisher
Springer Journals
Copyright
2017 The Author(s).
eISSN
2047-2994
DOI
10.1186/s13756-017-0232-x
Publisher site
See Article on Publisher Site

Abstract

Background: Patients with methicillin-resistant S. aureus (MRSA) are thought to incur additional costs for hospitals due to longer stay and contact isolation. The aim of this study was to assess the costs associated with MRSA in Norwegian hospitals. Methods: Analyses were based on data fromSouth-Eastern Norway for the year 2012 as registered in the Norwegian Surveillance System for Communicable Diseases and the Norwegian Patient Registry. We used a matched case-control method to compare MRSA diagnosed inpatients with non-MRSA inpatients in terms of length of stay, readmissions within 30 days from discharge, as well as the Diagnosis-Related Group (DRG) based costs. Results: Norwegian patients with MRSA stayed on average 8 days longer in hospital than controls, corresponding to a ratio of mean duration of 2.08 (CI 95%, 1.75–2.47) times longer.A total of 14% of MRSA positive inpatients were readmitted compared to 10% among controls. However, the risk of readmission was not significantly higher for patients with MRSA. DRG based hospital costs were 0.37 (95% CI, 0.19–0.54) times higher among cases than controls, with a mean cost of EUR13,233(SD 26,899) and EUR7198(SD 18,159) respectively. Conclusion: The results of this study indicate that Norwegian patients with MRSA have longer hospital stays, and higher costs than those without MRSA. Keywords: Methicillin-resistant S. aureus, Costs, Inpatient, Diagnosis-related group, Length of stay, Readmission Background Communicable Diseases (MSIS). The proportion of invasive Staphylococcus aureus is considered to be one of the S. aureus isolates in Norway that were methicillin-resistant most common pathogens causing nosocomial infections was only 1.3% in 2012 [3], and Norway is considered to be [1]. The bacterium is a normal inhabitant in the human amongst the European countries with the lowest percentage body, found persistent on the skin or mucosa of 20% of of MRSA in S. aureus clinical isolates. adults [2]. It is however associated with several types of The spread of MRSAin Norwegian hospitals is con- illnesses, most commonly skin and soft tissue infections, trolled through screening routines, which consist of testing although it may also cause more severe infections [1]. patients before admission if they have previously tested In Norway, all diagnosed cases of MRSA are manda- positive or are suspected of having been exposed to MRSA torily notified to the Norwegian Surveillance System for in the 12 months prior. The current MRSA-strategy in Norwegian hospitals consists of isolating suspected and * Correspondence: [email protected] confirmed MRSA positive patients, work restrictions for Norwegian Institute of Public Health, Marcus Thranesgate 2, 0473 Oslo, healthcare personnel who test positive for MRSA, and Norway Department of Antibiotic resistance and Infection control, Norwegian decolonization of carriers [4]. Some studies have explored Institute of Public Health, Oslo, Norway the cost-effectiveness of drugs used to treat MRSA Full list of author information is available at the end of the article © The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 2 of 8 infections and costs related to MRSA screening in Norway the Norwegian Surveillance System for Communicable [5, 6].To our best knowledge, there are currently no stud- Diseases (MSIS). This region represents approximately ies estimating the economic impact of MRSA positive 56% of the Norwegian population, including the capital patients based on LOS, readmissions or DRG costs in city of Oslo, and comprises 18 hospitals. Using the Norwegian hospitals. unique national identification number assigned to each The aim of our study was to estimate the resource use Norwegian citizen, information from both the NPR and associated with MRSA in Norwegian hospitals in terms MSIS databases could be linked in order to extract the of length of stay, readmissions within 30 days and DRG- MRSA status of inpatients that had been tested for based costs, in order to provide a better knowledge base MRSA. Patients who have undergone screening because for infection prevention strategies. they met at least one of the screening criteria, but had a negative MRSA-test, are not registered in MSIS. Ethical Methods approval was obtained from the South-Eastern Regional This was a register-based matched case-control study Committee for Medical and Health Research Ethics with patients admitted to hospital during the year 2012. (REC): 2013/1004/REK. We used data from the South-Eastern Norway health The dataset included individual patients, at all levels of region for the period January 1 to December 31, 2012, care (inpatient, outpatient, day patient), aggregated by epi- provided by the Norwegian Patient Registry (NPR) and sode of care. We identified inpatient stays and defined Fig. 1 Sample selection flowchart. Cases and controls were selected from a merged dataset including all patients who had received health care services in hospitals of the South-Eastern Norway health region (provided by the Norwegian Patient Registry) and all notified cases of methicillin-resistant S. aureus (provided by the Norwegian Surveillance System for Communicable Diseases) during the year 2012. Only patients who were diagnosed MRSA- positive at the earliest eight days prior or during their hospital stay, were included as cases. In total, 82 (86%) cases were matched to four control patients; the rest had less than four controls per case Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 3 of 8 those hospital stays by aggregating all episodes of care Table 1 Description of included patients with continuous dates (Fig. 1). Only the first hospital stay Cases (n = 95) Controls (n = 346) recorded for 2012 was considered in the analysis. MRSA- Variables n (%) n (%) cases were defined as patients diagnosed with MRSA at Male gender 49 (52%) 179 (52%) the earliest eight days prior to hospitalization, or in the Age course of their hospital stay. All patients diagnosed with 0–9 16 (17%) 64 (18%) MRSA in Norway are offered treatment before admission 10–19 3 (3%) 9 (3%) to hospital. This treatment together with microbiological 20–29 5 (5%) 20 (6%) analyses, takes a minimum of nine days. By restricting the case definition to a positive sample within this time frame, 30–39 8 (8%) 29 (8%) we maximized certainty that cases were MRSA positive 40–49 9 (9%) 28 (8%) during the hospital stay. Patients missing DRG codes, 50–59 14 (15%) 44 (13%) MRSA positive patients diagnosed more than eight days 60–69 14 (15%) 56 (16%) prior to hospitalization, and those patients without an 70–79 14 (15%) 51 (15%) exact match were excluded from the analysis. The flow of 80–89 11 (12%) 41 (12%) patient inclusion and exclusion is presented in Fig. 1.Con- trols were selected from the group of patients not > 89 1 (1%) 4 (1%) diagnosed with MRSA. Comorbidity index Cases were randomly matched, one to four controls, 68 (72%) 251 (73%) based on 10-year age groups, gender, hospital, surgery, 1 11 (12%) 39 (11%) whether they were diagnosed with an infection or not dur- 2 13 (14%) 44 (13%) ing the stay and Charlson Comorbidity Index (CCI) [7] 3 2 (2%) 8 (2%) (Table 1). For 13 cases (14%) it was not possible to obtain four controls; among these, one case was matched with 8 1 (1%) 4 (1%) three controls, three with two controls and nine with one Hospitals (anonymized) control patient. Among those patients who were A 20 (21%) 74 (21%) transferred between hospitals, the matching was done B 18 (19%) 70 (20%) based on the hospital where the highest DRG cost C 8 (8%) 32 (9%) weight was registered. We used the DRG codes indicating D 8 (8%) 26 (8%) surgical operations to identify patients who underwent surgery. In addition, we performed analyses of subgroups E 7 (7%) 28 (8%) where we combined ICD-10 codes (registered during F 6 (6%) 23 (7%) a hospital stay) with data from MSIS. The MSIS data G 6 (6%) 22 (6%) include MRSA diagnosis (colonization or infection), H 4 (4%) 16 (5%) sample material and clinical picture at the time of I 4 (4%) 13 (4%) notification (Table 2). J 4 (4%) 13 (4%) The comorbidity index was estimated for each patient according to the method developed for economic evalu- K 3 (3%) 12 (3%) ations by Charlson et al., based on diagnoses as de- L 2 (2%) 8 (2%) scribed by Quan et al. [8, 9]. The Canadian version of M 2 (2%) 2 (1%) the International Coding of Disease 10th revision (ICD- N 1 (1%) 4 (1%) 10) used by Quan et al., was validated against the Nor- O 1 (1%) 2 (1%) wegian ICD-10 codes used in this study. Due to data P 1 (1%) 1 (0%) availability, we used all the ICD-10 codes registered in 2012 to define CCI instead of using previous years’ diag- Infections noses. This included diagnoses registered at both in- No infections 42 (44%) 185 (53%) patient and outpatient episodes of care. Infection 53 (56%) 161 (47%) The outcome variables were length of stay (LOS) of Non-severe infections 31 (33%) 100 (29%) the first hospitalization, readmissions within 30 days Severe infections 15 (16%) 50 (14%) from discharge, and costs based on the highest Procedure Diagnosis-RelatedGroup (DRG)costweightduring the first hospital stay. Estimation of readmissions Surgery 17 (18%) 48 (14%) within 30 days was restricted to patients discharged before December 1, 2012. Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 4 of 8 Table 2 Infections based on ICD-10 codes recorded for cases Results and controls during the first hospital stay The dataset included a total of 948,266 patients. Of Infections Infection types (ICD-10 codes) those, 318,593 were inpatients (Fig. 1). Before matching, 4 patients missing DRG codes and 120 MRSA positive All Intestinal infectious diseases (A04; A08); Tuberculosis (A15); patients who were diagnosed with MRSA more than Sepsis (A41); eight days prior to hospitalization were excluded. Erysipelas (A46); Additionally, 17 patients who met the inclusion criteria Bacterial infection of unspecified site (A49); Human immunodeficiency virus disease (B23); did not have an exact match with other patients and Cytomegaloviral disease (B25); were excluded from the analysis.In total, 95 MRSA posi- Acute and subacute endocarditis (I33); tive patients met the inclusion criteria for cases, and Acute upper respiratory infections (J03; J06); Pneumonia/ other acute lower respiratory were matched with 346 MRSA negative controls. Gender infections (J15; J18; J20; J21); was equally represented in both groups (males Abscess of anal and rectal regions/ peritonitis 52%).Overall, 75% of the study population consisted of (K61; K65); Infections of the skin and subcutaneous tissue children less than 10 years and adults over 50 years of (L02 – L04; L08); age. The majority (84%) had a CCI of <2. In total, 65 Infectious arthropathies (M02); (15%) of the patients underwent surgery while admitted Necrotizing fasciitis (M726); Osteomyelitis (M86); (Table 1). Based on the type of hospital, 44 (10%), 301 Acute tubulo-interstitial nephritis (N10); (68%) and 98 (22%) patients were hospitalized in teach- Urinary tract infection, site not specified (N390); ing hospitals, general hospitals with more than 250 beds, Inflammatory disorders of breast/ infections of breast associated with childbirth (N61; O91) and local hospitals with less than 250 beds, respectively. Infection following a procedure, not elsewhere The median LOS was 6 (IQR 3.00–15.00) days for classified or due to internal fixation device MRSA positive patients and 4 (IQR 2.00–7.00) days for (T814; T846) MRSA negative patients, while the means were 14 (SD Infections possibly caused by S. aureus 17.24) and 6 (SD 9.61) days, respectively.The greatest Non-severe infections Bacterial infection of unspecified site (A49); difference in LOS between cases and controls was ob- Acute tonsillitis, unspecified (J039); Acute upper respiratory infections of multiple served in the subgroup with no infections and the group and unspecified sites (J06); diagnosed with severe S. aureus infections (Table 3). The Abscess of anal and rectal regions (K61) distribution of LOS is shown in Fig. 2. Cutaneous abscess, furuncle and carbuncle (L02); Cellulitis (L03); A higher proportion of cases were readmitted within Other local infections of skin and subcutaneous 30 days compared to controls, but overall there was no tissue, unspecified (L08); statistically significant difference regarding readmissions Urinary tract infection, site not specified (N390); Inflammatory disorders of breast/ Infections of (Table 4). breast associated with childbirth (N61; O91) In total, cases had higher DRG-based costs than con- Severe infections Other sepsis (A41); trols, with a median of EUR5174(IQR 3347–9059) and Acute and subacute endocarditis (I33); EUR3762 (IQR 2405–6669), respectively. In the sub- Bacterial pneumonia, unspecified (J159); group of patients with infections, the median costs were Pneumonia, organism unspecified (J18); Necrotizing fasciitis (M726); almost equal, although among patients with severe infec- Osteomyelitis (M86); tions the mean costs were significantly higher for cases Acute tubulo-interstitial nephritis (N10); than controls (Table 5). The distribution of DRG based Infection following a procedure, not elsewhere classified or due to internal fixation device costs is shown in Fig. 3. (T814; T846) Discussion All statistical analyses were performed in Stata Statis- The results of this study indicate that Norwegian pa- tical Software (StataCorp, 2013). The distribution of tients with MRSA had longer hospital stays, and higher patients in each group were presented as proportions. costs than those without MRSA. We did not find statis- We estimated median in addition to mean for the out- tically significant differences concerning readmission- come variables since the data were not normally distrib- s.Our study is the first in Norway addressing the uted. Given a matching ratio of 1:4 and thus data on economic burden of MRSA in hospitals, and one of few both group and individual level, we measured the out- studies that estimate the costs of MRSA in general, not comes using Multi-level Mixed effect Regression models. only the life threatening MRSA related conditions. Having skewed distributions (Figs. 2 and 3), we analyzed We found that MRSA positive patients stay longer in LOS using negative binomial regression, readmissions the hospital than their controls, and this is statistically with both logistic and linear regression, and costs with significant both for the total study population and sub- log of the linear regression outcome [10]. groups. Previous studies have shown that patients with Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 5 of 8 Fig. 2 Distribution of LOS for cases and controls based on the first hospital stay in 2012. The number of days spent in hospital for the first hospital stay registered in 2012 showed that patients not diagnosed with methicillin-resistant S. aureus were more likely to have shorter hospital stays than patients diagnosed with methicillin-resistant S. aureus. LOS, length of stay severe MRSA infections have longer hospital stays com- delayed examination and treatment of patients, as well pared to patients with severe MSSA infections [11–14]. as poorer follow up (less supervision, worsening of con- A result of particular interest in this study was that we dition due to delayed treatment) [15]. found the largest difference in LOS among patients Readmission within 30 days from discharge is used as without infections. In Norway, this result may be a con- an indicator of quality of hospital care in Norway. We sequence of the comprehensive infection control mea- observed few readmissions overall, and even fewer in sures applied in hospitals, such as single room isolation patients with severe infections. The results showed only of patients with MRSA. Some studies have addressed the small and not statistically significant differences in risk possible negative consequences of single room isolation of readmission among cases and controls. Interestingly, in hospitals; for example isolation may contribute to although our results indicated that isolation adds to Fig. 3 Distribution of DRG based costs for cases and controls during their first hospitalization in 2012. DRG cost weights are the basis for a considerable share (40%) of hospital reimbursement in Norway, and hence, an important proxy for costs per patient, however inaccurate. Due to limited access to micro costing data for patients in this study, DRG cost weights were used to estimate costs per hospital stay, although these costs are potentially an underestimate of the actual economic burden caused by MRSA to hospitals Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 6 of 8 Table 3 Univariate analysis of LOS distributed by infectious condition Length of stay Cases (n = 95) Controls (n = 346) Negative binomial regression Median (IQR) Mean (SD) Median (IQR) Mean (SD) Ratio of mean duration 95% CI p-value All patients (n = 95/441) 6.00 (3.00–15.00) 13.71 (17.24) 4.00 (2.00–7.00) 6.05 (9.61) 2.08 1.75–2.47 <0.05 � No infections (n = 42/227) 9.50 (3.00–23.00) 14.36 (13.87) 3.00 (2.00–6.00) 5.75 (9.03) 2.48 1.91–3.21 <0.05 � Any infection (n = 53/214) 5.00 (3.00–13.00) 13.19 (19.62) 4.00 (2.00–7.00) 6.40 (10.25) 1.58 1.26–1.98 <0.05 - Non-severe infections (n = 31/131) 5.00 (3.00–9.00) 8.94 (19.27) 3.00 (2.00–5.00) 5.18 (10.84) 1.40 1.02–1.92 <0.05 - Severe infections (n = 15/65) 14.00 (5.00–36.00) 21.60 (20.69) 5.00 (4.00–12.00) 8.88 (9.70) 1.86 1.33–2.62 <0.05 IQR Interquartile range, SD Standard deviation, CI Confidence interval prolonged hospital stay, we did not find that isolation 2012 EUR 856 per day [5].In addition, our results showed leads to poorer quality of medical treatment as measured that MRSA positive patients had longer hospital stays than by the risk of readmission. The results must be inter- patients without MRSA. The average cost of a patient per preted cautiously as the sample is quite small. day in Norwegian hospitals in 2012 was estimated to be Median and mean costs were higher for cases than EUR 5368 [17]. controls overall. However, the median costs in the sub- The strengths of this study were the quality and accur- groups of patients with infections, were approximately acy of the dataset utilized for analyses. The dataset in- equal. Claudia Hubner et al. used an average daily reim- cluded the actual LOS for each episode of hospitalization bursement per bed based on the actually claimed G- for all patients in the health region. Norwegian health DRG in the patient record, to estimate opportunity cost registries provide information on all the patients, their of blocked beds. They also accounted for empirical costs unique encrypted identification number allows tracking related to hygienic measures and laboratory use, and them as well as linking data from different health regis- found MRSA-attributable costs of EUR 8673 per case tries. Therefore, despite a small number of cases, it gives a [16].The DRG system is not designed to estimate real complete overview of the diagnosed MRSA cases in the costs of each individual patient, but specific information health region. The Norwegian surveillance system con- regarding the patient’s age, gender, diagnoses and treat- tains all patients who have been diagnosed with MRSA, ments, is used to categorize patients into cost groups. regardless of where the test is taken. However, most pa- We aimed to get cases and controls as similar as pos- tients in Norwegian hospitals are not routinely screened- sible, where MRSA was the only distinguishing factor. for MRSA, which may lead to an underestimation of the Most cases and controls have, consequently ended up in true MRSA rate and undetected MRSA colonization in the same cost groups. None of the cases in this study controls. The distributions of LOS and costs (Figs. 2 and were registered with DRG codes specifically reflecting 3) show distributions heavily skewed to the right, with a the finding of resistant bacteria. mean close to zero. Based on the distribution of the out- Taking into consideration that the DRG system does come variable in the study population, it is likely that not account for MRSA status and the additional costs of MRSA positive patients among controls have ended up infection control measures, such as screening and isola- close to the mean value, and hence would not have signifi- tion of MRSA patients, the results in this study are most cantly influenced the results. However, if any MRSA posi- likely an underestimate of the actual cost differences.The tive controls are outliers, they have consequently masked Norwegian infection control measures are comprehen- a larger difference between cases and controls. sive and costly, and they are applied to both colonized Two other European studies have used DRG cost and infected patients. Jinshuo Li estimated that the cost of weights in addition to other costs [14, 16]. Healthcare resources associated with isolation were approximately resource utilization measurements that are not included Table 4 Univariate analysis of readmissions within 30 days distributed by infectious condition Readmissions within 30 days Cases (n = 74) Controls (n = 299) Logistic regression Linear regression n (%) n (%) OR 95% CI p-value Coef. 95% CI p-value All patients (n = 74/299) 10 (14%) 22 (10%) 1.50 0.65–3.46 0.34 0.04 −0.04 – 0.12 0.34 � No infections (n = 30/144) 4 (13%) 15 (13%) 1.02 0.31–3.32 0.98 0.00 −0.13 – 0.14 0.98 � Any infection (n = 44/155) 6 (14%) 7 (6%) 3.44 0.79–15.00 0.10 0.08 −0.01 – 0.17 0.07 - Non-severe infections (n = 25/83) 4 (16%) 1 (2%) 10.86 1.15–102.77 <0.05 0.14 0.04–0.25 <0.05 - Severe infections (n = 13/57) 2 (15%) 5 (11%) 1.76 0.19–16.67 0.62 0.04 −0.13 – 0.21 0.62 OR odds ratio, CI confidence interval Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 7 of 8 Table 5 Univariate analysis of DRG based costs distributed by infectious condition Reimbursement (EUR) Cases (n = 95) Controls (n = 346) Linear regression with log (reimbursement) as outcome Median (IQR) Mean (SD) Median (IQR) Mean (SD) Coef 95% CI p-value All patients (n = 95/441) 5174 (3347–9059) 13,233 (26,899) 3762 (2405–6669) 7198 (18,159) 0.37 0.19–0.54 <0.05 � No infections (n = 42/227) 6423 (3603–8445) 13,208 (26,850) 3680 (2103–6669) 7797 (22,619) 0.48 0.22–0.74 <0.05 � Any infection (n = 53/214) 4560 (3148–11,178) 13,252 (27,195) 4355 (3148–6326) 6511 (11,033) 0.21 −0.03 – 0.45 0.09 - Non-severe infections (n = 31/131) 3450 (2523–5794) 8065 (18,243) 3347 (2523–5794) 5299 (12,802) 0.09 −0.25 – 0.44 0.60 - Severe infections (n = 15/65) 7600 (5794–21,931) 26,904 (41,802) 7600 (4560–7600) 9300 (7268) 0.47 0.15–0.79 <0.05 IQR Interquartile range, SD Standard deviation, CI Confidence interval, EUR Euro, Exchange rate NOK-EUR7.4805 average for 2012 in the hospital reimbursement DRG, were not available Authors’ contributions AESA contributed to the analysis of results, research and writing of the through the registry data used in our study. For this rea- manuscript. CMJ contributed with research and editing the manuscript. BFdB son, we do not know the total cost and resource use of was a great contributor to the methodology in this study. RW quality this population. Considering that MRSA-positive pa- assured statistical analyses. ISK contributed with supervision of the work and recommendations regarding formulations and statistical analyses. PE tients are subject to very strict infection control routines performed the data linkage of the NPR and MSIS datasets, and prepared the in Norway and have extended LOS compared with data for analysis. He contributed to the analysis of results and manuscript controls, the cost difference between cases and controls writing. He was the main supervisor of the work. All authors agreed on the methods and statistical analyses, read and approved the final manuscript. is potentially greater.Norwegian health registries are intended for administrative purposes and are not de- Ethics approval and consent to participate signed for research, therefore, although we used registry Ethical approval was obtained from the South-Eastern Regional Committee data of high quality in our study, we were limited in our for Medical and Health Research Ethics (REC): 2013/1004/REK. scope of analysis by the measurements available. The Consent for publication study design implies that there are uncontrolled risk fac- Not applicable. tors. We attempted to control for potential confounders thought to be relevant by matching. Competing interests The authors declare that they have no competing interests. Conclusion This study shows that MRSA contributes to longer Publisher’sNote length of stay in hospitals and higher costs based on Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. DRG cost weights. It gives an improved knowledge base with regards to the consequences of resistant bacteria, and Author details may help Norwegian policy makers to make informed Norwegian Institute of Public Health, Marcus Thranesgate 2, 0473 Oslo, Norway. Institute of Health and Society, University of Oslo, Norway, decisions concerning resource allocation, infection pre- Forskningsveien 3A, 0373 Oslo, Norway. Department of Biostatistics, Oslo vention programs and guideline development. However, Centre for Biostatistics and Epidemiology, Institute of Basic MedicalSciences, in order to get a better understanding of the costs related University of Oslo, Sognsvannsveien 9, 0372 Oslo, Norway. Department of Antibiotic resistance and Infection control, Norwegian Institute of Public to resistant bacteria, micro-cost studies are needed. Health, Oslo, Norway. Abbreviations Received: 29 March 2017 Accepted: 27 June 2017 CCI: Charlson comorbidity index; DRG: Diagnosis-related group; ICD- 10: International coding of disease 10th revision; LOS: Length of stay; MRSA: Methicillin-resistant Staphylococcus aureus; MSIS: Norwegian surveillance system for communicable diseases; NPR: Norwegian patient References registry; REC: South-Eastern regional committee for medical and health 1. Lowy FD. Staphylococcus aureus infections. N Engl J Med. 1998;339(8):520–32. research ethics 2. Kluytmans J, van Belkum A, Verbrugh H. Nasal carriage of Staphylococcus aureus: epidemiology, underlying mechanisms, and associated risks. Clin Funding Microbiol Rev. 1997;10(3):505–20. This work was supported by the Norwegian Institute of Public Health, 3. European Centre for Disease Prevention and Control. Antimicrobial Department of Antibiotic Resistance and Infection Prevention. resistance surveillance in Europe 2012. Stockholm: ECDC; 2013. 4. Elstrøm P, Kacelnik O, Bruun T, Iversen BG, Hauge SH, Aavitsland P. Availability of data and materials Meticillin-resistant Staphylococcus aureus in Norway, a low-incidence The data that support the findings of this study are available from the country, 2006-2010. J Hosp Infect. 2012;80(1):36–40. Norwegian Patient Registry (NPR) and Norwegian Surveillance System for 5. Li J, Ulvin K, Biboh H, Kristiansen IS. Cost-effectiveness of supplementing a Communicable Diseases (MSIS), but restrictions apply to the availability of broth-enriched culture test with the Xpert meticillin-resistant Staphylococcus these data, which were used under license for the current study, and so are aureus (MRSA) assay for screening inpatients at high risk of MRSA. J Hosp not publicly available. Data are however available from the authors upon Infect. 2012;82(4):227–33. reasonable request and with permission of the South-Eastern Regional 6. Nguyen TC. [En legemiddeløkonomisk analyse av antibiotikabruk ved Committee for Medical and Health Research Ethics (REC). invasive MRSA-infeksjoner] [Norwegian]. Tromsø: University of Tromsø; 2009. Andreassen et al. Antimicrobial Resistance and Infection Control (2017) 6:74 Page 8 of 8 7. Charlson M, Wells MT, Ullman R, King F, Shmukler C. The Charlson comorbidity index can be used prospectively to identify patients who will incur high future costs. PLoS One. 2014;9(12):e112479. 8. Charlson ME, Charlson RE, Peterson JC, Marinopoulos SS, Briggs WM, Hollenberg JP. The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients. JClin Epidemiol. 2008;61(12): 1234–40. 9. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, Saunders LD, Beck CA, Feasby TE, Ghali WA. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005; 43(11):1130–9. 10. Austin PC, Rothwell DM, Tu JV. A comparison of statistical modeling strategies for analyzing length of stay after CABG surgery. Health Serv Outcomes Res Method. 2002;3(2):107–33. 11. Anderson DJ, Kaye KS, Chen LF, Schmader KE, Choi Y, Sloane R, Sexton DJ. Clinical and financial outcomes due to methicillin resistant Staphylococcus aureus surgical site infection: a multi-center matched outcomes study. PLoS One. 2009;4(12):e8305. 12. Cosgrove SE, Qi Y, Kaye KS, Harbarth S, Karchmer AW, Carmeli Y. The impact of methicillin resistance in Staphylococcus aureus bacteremia on patient outcomes: mortality, length of stay, and hospital charges. InfectControl Hosp Epidemiol. 2005;26(2):166–74. 13. de Kraker ME, Davey PG, Grundmann H. Mortality and hospital stay associated with resistant Staphylococcus aureus and Escherichia coli bacteremia: estimating the burden of antibiotic resistance in Europe. PLoS Med. 2011;8(10):e1001104. 14. Macedo-Viñas M, De Angelis G, Rohner P, Safran E, Stewardson A, Fankhauser C, Schrenzel J, Pittet D, Harbarth S. Burden of methicillin- resistant Staphylococcus aureus infections at a Swiss University hospital: excess length of stay and costs. J Hosp Infect. 2013;84(2):132–7. 15. Morgan DJ, Diekema DJ, Sepkowitz K, Perencevich EN. Adverse outcomes associated with contact precautions: a review of the literature. Am J Infect Control. 2009;37(2):85–93. 16. Hübner C, Hüber NO, Hopert K, Maletzki S, Flessa S. Analysis of MRSA- attributed costs of hospitalized patients in Germany. Eur J Clin Microbiol Infect Dis. 2014;33:1817–22. 17. Norwegian Directorate of Health. Samdata spesialisthelsetjenesten 2012. Oslo: Norwegian; 2013. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit

Journal

Antimicrobial Resistance & Infection Control – Springer Journals

Published: Dec 1, 2017

Keywords: medical microbiology; drug resistance; infectious diseases

There are no references for this article.