Abstract
Background
Accurate accounting of antibiotic use is necessary for studies comparing the CF airway microbiota across clinically relevant disease states. While poor adherence to chronic therapies is well described for individuals with CF, use patterns of episodic oral antibiotics are less clear.
Methods
Eleven individuals with CF completed daily questionnaires regarding antibiotic use for a mean of 458 days. Self-report of episodic oral antibiotic use was compared to antibiotic prescription data in the electronic medical record (EMR).
Results
Self-reported use of episodic oral antibiotics differed from EMR data an average of 8.3% of days per subject. The majority of these discrepancies were due to self-reported use of oral antibiotics outside of the EMR-documented dates of antibiotic prescription.
Conclusions
Discrepancies exist between self-reported use of episodic oral antibiotics and EMR data that have implications for studies of the CF airway microbiota.
Keywords: Microbiome, Adherence, Antibiotics
1. Introduction
Culture-independent studies of the cystic fibrosis (CF) airway microbiota have provided new insights into the pathogenesis of CF lung disease and have the potential to identify novel treatment strategies [1]. Notable findings have included the decrease in diversity in the CF airway microbiota associated with advancing age and disease severity [2–4], changes in the microbiota that are associated with the onset of pulmonary exacerbation [5–7], and the response of microbiota to antibiotic treatment [2,8–10]. Because antibiotic use can have a significant short-term impact on culture-independent measures of bacterial community structure [2,8–10], accurate accounting of antibiotic use at the time of sample collection is necessary for meaningful interpretation of CF airway microbiome data. Reliable ascertainment of subjects’ dates of antibiotic use is also necessary to define clinically relevant disease states, including baseline health (B), onset of exacerbation (prior to antibiotic administration) (E), antibiotic treatment (T), and recovery (defined number of days after antibiotic completion) (R) [2].
Most often in studies of the CF airway microbiota, data on the subject's antibiotic use at the time of sample collection are obtained from the electronic medical record (EMR). However, poor adherence rates to chronic oral (e.g., azithromycin) and inhaled (e.g., tobramycin) antibiotics are well described in individuals with CF [11,12]. EMR prescription data, thus, does not accurately reflect actual medication usage of these medications. Less is known about how EMR prescription data reflect use patterns of episodic oral antibiotics prescribed for pulmonary exacerbations.
In this study, we compared the EMR with daily self-reports in a cohort of individuals with CF to determine the accuracy of the EMR for capturing use patterns of episodic oral antibiotics.
2. Methods
2.1. Data collection and analyses
Under approval from the University of Michigan Heath System (UMHS) Institutional Review Board (HUM00037056), we enrolled 15 adults with CF between 2011 and 2014 for a prospective collection of daily sputum samples as well as completion of daily questionnaires. The questionnaires included information on daily antibiotic use (Supplemental data 1). The UMHS EMR was reviewed and information on prescription of episodic oral antibiotics during the time periods of interest was extracted and compared to the questionnaires. Inhaled antibiotics and chronic oral antibiotics (including azithromycin) were excluded from the analysis.
An antibiotic was considered to be prescribed on a given day if so indicated in any area of the EMR, including prescription record, clinic notes, and telephone encounters. The EMR and self-report were considered concordant for a given day if all episodic oral antibiotics were in agreement between the two measures. The EMR and self-report were considered discordant if any discrepancy in episodic oral antibiotics existed between the two measures. Sensitivity, specificity, positive predictive value (PPV), and negative predictive values (NPV) were calculated considering self-report to be the gold standard.
3. Results
3.1. Subject characteristics
One of the 15 patients did not have any oral antibiotic courses during the study period and thus was excluded from further analyses. Of the remaining 14 subjects, three completed <90% of their daily questionnaires and were thus excluded from further analyses. The remaining 11 subjects completed >95% of their daily home questionnaires. Characteristics of these subjects are shown in Table 1.
Table 1.
Subject Characteristics
| Sex | 6 female, 5 male |
|---|---|
| Age in years at start of study, mean (range) | 33.7 (20-45) |
| Baseline % predicted FEV1, mean (range) | 68% (42%-106%) |
| Days of data collection, mean (range) | 458 (238-731) |
| Days of self-reported episodic oral antibiotic use, mean (range) | 100 (9-350) |
3.2. Comparison of EMR to self-report
The EMR and self-report were discordant on an average of 8.3% of days of data collection (median 4.8%, range 0.3-26%). The sensitivity, specificity, PPV, and NPV of the EMR for capturing self-reported antibiotic use are listed in Table 2. Individual values are included in the supplemental materials (Supplemental data 2).
Table 2.
Sensitivity, specificity, PPV, and NPV
| Mean (95% CI) | Range between subjects | |
|---|---|---|
| Sensitivity1 | 73.5% (70.8-76.1%) | 39.7-100% |
| Specificity2 | 95.9% (95.2-96.5%) | 82.5-100% |
| PPV3 | 83.6% (81.1-85.9%) | 56.3-100% |
| NPV4 | 92.7% (91.8-93.5%) | 71.7-100% |
Sensitivity: The probability that the EMR listed an antibiotic prescription when the subject reported having taken the same antibiotic.
Specificity: The probability that the EMR did not list an antibiotic prescription when the subject reported not taking any episodic antibiotics.
PPV: The probability that a subject reported taking an antibiotic on a day on which an antibiotic was prescribed.
NPV: The probability that a subject did not report taking an antibiotic on a day that an antibiotic was not prescribed.
3.3. Characterization of discordant days
Discrepancies between the EMR and self-report on days that antibiotics were prescribed but not taken accounted for 31.7% of the discordant days. The majority of discordant days were due to use of antibiotics on days when no antibiotic was prescribed (Figure). This most often consisted of antibiotic use in the days following the prescription stop date. Subjects also reported starting antibiotics in the days immediately prior to the prescription start date, as well as taking antibiotics on days that were not associated with EMR-documented prescription. An example of self-reported antibiotic use compared to EMR for one subject is included in the supplemental materials (Supplemental data 3).
Figure.
Characterization of discordant days between self-report and EMR
4. Discussion
The degree to which the EMR accurately reflects use patterns of episodic oral antibiotic has implications for studies of the CF airway microbiota in relationship to clinically relevant disease states. In this study, the EMR reflected self-reported use of episodic oral antibiotics the majority of the time. In particular, the EMR had a high specificity and negative predictive value for self-reported antibiotic use. However, every subject in this study had some discordance between the EMR and self-report for episodic oral antibiotic use.
The EMR was less reliable for representing actual antibiotic usage around the time of pulmonary exacerbation. Our subjects self-reported that they occasionally took antibiotics on the days immediately preceding the EMR recorded start date and extended them beyond the EMR recorded end date. These patterns of antibiotic use are particularly relevant for studies in which assignment of clinically important disease states (B,E,T,R) to samples depends upon accurate knowledge of subjects’ antibiotic usage (e.g., onset of exacerbation but prior to antibiotic treatment (E)). Our finding of discordance between EMR and self-report during the dates of episodic antibiotic prescription (i.e., lack of adherence to the prescription) is consistent with prior investigations that have shown poor adherence rates to chronic oral (e.g., azithromycin) and inhaled (e.g., tobramycin) antibiotics [11,12].
Our study was limited to 11 adults treated in a single CF center. Adherence is known to vary by age, and our subjects were older than the adolescent age group having the lowest rates of adherence to medical therapies in one study [13]. Additionally, our subjects had agreed to participate in the study of daily sputum collection and surveys, making it likely that they represent a more adherent subgroup. For our analyses, we considered the subjects’ self-reports to be the gold standard. But, self-report has known limitations and may show inflated adherence in individuals with CF compared to data obtained by user-independent electronic monitoring [14,15]. We made efforts to minimize biases associated with self-report in this study. Our subjects were told on enrollment that their questionnaire responses regarding medication use would not be communicated back to their providers and that accurate responses were critical for a successful study. Although subjects may have been self-motivated to report incorrectly that they were taking prescribed antibiotics, they would have had little to gain from reporting antibiotic usage when it was not prescribed.
Finally, we considered the EMR and self-report to be concordant for a given day if all antibiotics were in agreement between the two measures. Individuals with CF are often prescribed more than one oral antibiotic for a pulmonary exacerbation. Our measures of discordance included days on which a subject may have been adherent to one prescribed antibiotic, but not another (Supplemental data 3), further illustrating the complexity of accounting for accurate antibiotic use patterns in CF.
In summary, we report a comparison of the EMR to self-report for episodic oral antibiotic usage in individuals with CF. These data serve as a cautionary note for studies assessing changes in the CF airway microbiota as they relate to clinically relevant outcomes and disease states. Our data demonstrate the limitations of the EMR for representing antibiotic use, and suggest the added value of patient self-report for studies that rely on accurate knowledge of antibiotic use.
Supplementary Material
Acknowledgements
The authors thank the dedication of the study subjects whose generous participation made this study possible. We also thank Michelle Azar and Lisa Leininger for their assistance with this effort.
Financial support
This work was funded by the Cystic Fibrosis Foundation (LIPUMA1310) and the Nesbitt Program for Cystic Fibrosis Research. Additional support was provided by the NIH/NICHD (K12 HD 028820 to L.J.C).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Caverly LJ, Zhao J, LiPuma JJ. Cystic fibrosis lung microbiome: Opportunities to reconsider management of airway infection. Pediatr Pulmonol. 2015;50(Suppl 4):S31–8. doi: 10.1002/ppul.23243. doi:10.1002/ppul.23243. [DOI] [PubMed] [Google Scholar]
- 2.Zhao J, Schloss PD, Kalikin LM, Carmody LA, Foster BK, Petrosino JF, et al. Decade-long bacterial community dynamics in cystic fibrosis airways. Proc Natl Acad Sci U S A. 2012;109:5809–14. doi: 10.1073/pnas.1120577109. doi:10.1073/pnas.1120577109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Cox MJ, Allgaier M, Taylor B, Baek MS, Huang YJ, Daly RA, et al. Airway microbiota and pathogen abundance in age-stratified cystic fibrosis patients. PLoS One. 2010;5:e11044. doi: 10.1371/journal.pone.0011044. doi:10.1371/journal.pone.0011044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Stokell JR, Gharaibeh RZ, Hamp TJ, Zapata MJ, Fodor AA, Steck TR. Analysis of changes in diversity and abundance of the microbial community in a cystic fibrosis patient over a multiyear period. J Clin Microbiol. 2015;53:237–47. doi: 10.1128/JCM.02555-14. doi:10.1128/jcm.02555-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Carmody LA, Zhao J, Schloss PD, Petrosino JF, Murray S, Young VB, et al. Changes in cystic fibrosis airway microbiota at pulmonary exacerbation. Ann Am Thorac Soc. 2013;10:179–87. doi: 10.1513/AnnalsATS.201211-107OC. doi:10.1513/AnnalsATS.201211-107OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Carmody LA, Zhao J, Kalikin LM, LeBar W, Simon RH, Venkataraman A, et al. The daily dynamics of cystic fibrosis airway microbiota during clinical stability and at exacerbation. Microbiome. 2015;3:12. doi: 10.1186/s40168-015-0074-9. doi:10.1186/s40168-015-0074-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sibley CD, Parkins MD, Rabin HR, Duan K, Norgaard JC, Surette MG. A polymicrobial perspective of pulmonary infections exposes an enigmatic pathogen in cystic fibrosis patients. Proc Natl Acad Sci U S A. 2008;105:15070–5. doi: 10.1073/pnas.0804326105. doi:10.1073/pnas.0804326105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Smith DJ, Badrick AC, Zakrzewski M, Krause L, Bell SC, Anderson GJ, et al. Pyrosequencing reveals transient cystic fibrosis lung microbiome changes with intravenous antibiotics. Eur Respir J. 2014;44:922–30. doi: 10.1183/09031936.00203013. doi:10.1183/09031936.00203013. [DOI] [PubMed] [Google Scholar]
- 9.Stressmann FA, Rogers GB, van der Gast CJ, Marsh P, Vermeer LS, Carroll MP, et al. Long-term cultivation-independent microbial diversity analysis demonstrates that bacterial communities infecting the adult cystic fibrosis lung show stability and resilience. Thorax. 2012;67:867–73. doi: 10.1136/thoraxjnl-2011-200932. doi:10.1136/thoraxjnl-2011-200932. [DOI] [PubMed] [Google Scholar]
- 10.Cuthbertson L, Rogers GB, Walker AW, Oliver A, Green LE, Daniels TW, et al. Respiratory microbiota resistance and resilience to pulmonary exacerbation and subsequent antimicrobial intervention. ISME J. 2015 doi: 10.1038/ismej.2015.198. doi:10.1038/ismej.2015.198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Quittner AL, Zhang J, Marynchenko M, Chopra PA, Signorovitch J, Yushkina Y, et al. Pulmonary medication adherence and health-care use in cystic fibrosis. Chest. 2014;146:142–51. doi: 10.1378/chest.13-1926. doi:10.1378/chest.13-1926. [DOI] [PubMed] [Google Scholar]
- 12.Eakin MN, Bilderback A, Boyle MP, Mogayzel PJ, Riekert KA. Longitudinal association between medication adherence and lung health in people with cystic fibrosis. J Cyst Fibros. 2011;10:258–64. doi: 10.1016/j.jcf.2011.03.005. doi:10.1016/j.jcf.2011.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Masterson TL, Wildman BG, Newberry BH, Omlor GJ. Impact of age and gender on adherence to infection control guidelines and medical regimens in cystic fibrosis. Pediatr Pulmonol. 2011;46:295–301. doi: 10.1002/ppul.21366. doi:10.1002/ppul.21366. [DOI] [PubMed] [Google Scholar]
- 14.Modi AC, Lim CS, Yu N, Geller D, Wagner MH, Quittner AL. A multi-method assessment of treatment adherence for children with cystic fibrosis. J Cyst Fibros. 2006;5:177–85. doi: 10.1016/j.jcf.2006.03.002. doi:10.1016/j.jcf.2006.03.002. [DOI] [PubMed] [Google Scholar]
- 15.Burrows JA, Bunting JP, Masel PJ, Bell SC. Nebulised dornase alpha: adherence in adults with cystic fibrosis. J Cyst Fibros. 2002;1:255–9. doi: 10.1016/s1569-1993(02)00095-4. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.