Use of Diagnosis Codes to Find Blood Transfusion Adverse Events in Electronic Health Records

Roselie A. Bright, Susan J. Bright-Ponte, Lee Anne Palmer, Summer K. Rankin, Sergey Blok
doi: https://doi.org/10.1101/2020.12.30.20218610

ABSTRACT

Background Electronic health records (EHRs) and big data tools offer the opportunity for surveillance of adverse events (patient harm associated with medical care). We chose the case of transfusion adverse events (TAEs) and potential TAEs (PTAEs) because 1.) real dates were obscured in the study data, and 2.) there was emerging recognition of new types during the study data period.

Objective We aimed to use the structured data in electronic health records (EHRs) to find TAEs and PTAEs among adults.

Methods We used 49,331 adult admissions involving critical care at a major teaching hospital, 2001-2012, in the MIMIC-III EHRs database. We formed a T (defined as packed red blood cells, platelets, or plasma) group of 21,443 admissions vs. 25,468 comparison (C) admissions. The ICD-9-CM diagnosis codes were compared for T vs. C, described, and tested with statistical tools.

Results TAEs such as transfusion associated circulatory overload (TACO; 12 T cases; rate ratio (RR) 15.61; 95% CI 2.49 to 98) were found. There were also PTAEs similar to TAEs, such as fluid overload disorder (361 T admissions; RR 2.24; 95% CI 1.88 to 2.65), similar to TACO. Some diagnoses could have been sequelae of TAEs, including nontraumatic compartment syndrome of abdomen (52 T cases; RR 6.76; 95% CI 3.40 to 14.9) possibly being a consequence of TACO.

Conclusions Surveillance for diagnosis codes that could be TAE sequelae or unrecognized TAE might be useful supplements to existing medical product adverse event programs.

INTRODUCTION

FDA relies on spontaneous reports of patient harm (known as adverse events (AE)) from manufacturers, healthcare providers, and the general public to learn of AEs that may need regulatory action [1]. Published deficiencies of these reports [2-10] include varying representativeness of harm and problems. Now that electronic healthcare records (EHRs) are very common [11], we have an opportunity to leverage them as sources of AE information [4, 12-14].

Challenges to the automated use of diagnosis codes for surveillance include:

  • ICD [International Classification of Diseases] codes may be “invalid, insensitive or non-specific” [15].

  • EHRs are “…subject to access restrictions…” [2].

  • “…[N]ot all events and outcomes are consistently captured …” [13].

The most directly useful program would analyze EHRs [1, 6] in real time. The data available to us (critical care admissions) did not allow calendar time analyses, so we compared blood transfusion admissions (T) vs. comparison admissions (C) because:

  • An earlier version of the dataset showed a higher risk of near-term mortality for red blood cell transfusion patients, compared to non-transfused patients [16].

  • During the time of the dataset, new transfusion AE (TAE) types, transfusion related acute lung injury (TRALI) and transfusion associated cardiac overload (TACO), were being recognized among transfusion experts and prompted new guidelines to reduce the use of blood transfusion [17].

  • During the dataset time, far fewer reports were coming to FDA than would have been expected, considering the level of professional concern [18-20].

The focus of this study is on diagnosis codes in electronic health records (EHR) because it is a relatively available source of information in EHRs and are readily computable. Many studies have looked for predetermined unwelcome conditions; for example, emergency department visits due to opioid overdose [21], motor vehicle crash injury [22], and serious infections [23]. We studied the utility of this method in this setting so that we can compare the results to using another surveillance method with the same data.

The objective of this study was to observe the utility of diagnosis codes to find TAEs that could be further evaluated (outside this study’s scope) with a formal analysis to demonstrate causality.

METHODS

We used EHRs for critical care admissions within an adult hospital, Beth Israel Deaconess Medical Center, Boston, MA. The Massachusetts Institute of Technology (MIT) worked with the hospital to process EHRs, including unstructured notes, from 2001 to 2012 into a privacy-protected dataset, Medical Information Mart for Intensive Care III (MIMIC-III) [24, 25], that is publicly available to those meeting human subjects research requirements. The research was designated not human subjects research by the FDA Institutional Review Board under Code of Federal Regulations Title 45 Part 46 [26].

We removed admissions with patient age < 16 years on the date of admission, as well as admissions without any notes, from the total of 58,976 admissions, resulting in 49,284 admissions.

The transfused group (T) of 21,443 admissions had at least one transfusion (red blood cells, platelets, or plasma) “input code” in the “Inputs” table or International Statistical Classification of Diseases and Related Health Problems (ICD) 9-CM Volume 3 procedure code in the “Procedures” table, listed in Appendix Table A1 (Codes for transfusion and exclusion groups). Among the remaining patients, we identified 2,373 admissions whose only indication of possible transfusion are the codes listed in Appendix Table A1 and excluded them from further study. The other 25,468 admissions formed the comparison (C) group.

We counted the numbers of admissions in T and C for all ICD9 diagnoses [27] that occurred at least as often in T as in C, using Python packages SciPy Stats [28] and Pandas [29]. Most of the codes were present in the MIMIC-III dictionary of ICD9 diagnosis codes. For those that were not, names were found on a public website and enclosed in brackets [30]. When the choice of appropriate code name was not clear, we examined the procedure codes for the relevant admissions, and if necessary, the discharge notes to adjudicate the correct name.

We manually reviewed the information that was available for admissions that received an explicit TAE code despite being assigned to C.

For each diagnosis code we calculated the rate difference (RD), rate ratio (RR), their normal approximate 95% lower and upper confidence limits (95% LCL and 95% UCL), and p values [31]. When any of the rate numerators had a value of 5 or less, we calculated the Fishers Exact Test (FET) p value with Python and then derived the confidence limits [32].

Highly related codes were presented adjacent to each other in Tables 2 through 9 and subtotaled. All 524 diagnoses in the Tables met one of the following criteria [33-86]:

  • Known TAE

  • Hypercoagulation, embolism, or anemia diagnosis

  • RR > 2

  • RR > 1.5 and actual (T minus C) counts over 100

  • The FET p value is < 0.05 (calculated when either of the T or C counts < 6)

TAE codes in our study were compared to TAE codes in published literature with a Fisher’s Exact Test calculator [87].

Our presentation of exact p values as low as “<0.0001” allows the reader to calculate their own adjustment of alpha based on their clinical interpretation of the interdependence of the diagnosis codes. If all 524 presented diagnoses are completely dependent, the alpha level to use is 0.0000979 [88]. Specific adjusted alpha levels assuming complete dependence are given with the results in tables organized by TAE-oriented system: iatrogenic (Table 2), circulation (3), bleeding and anemia (4), immune, neoplasm, and infection (5), lungs and oral (6), fluids regulation (7), digestion and hormones (8), and connective, nervous and reproductive (9).

We used published studies of TAEs in Medicare data [19, 20, 89, 90] to calculate expected TAE rates in MIMIC-III.

RESULTS

Table 1 shows the explanations for all admissions that were assigned to the C group and had an explicit TAE diagnosis code. For six of the twelve admissions, the notes say transfusion occurred. For four, the notes say there was transfusion just before admission to the hospital or CCU. For the remaining two, there was no mention of transfusion in the notes, so the TAE diagnosis codes were not explained.

View this table:
Table 1. Explanations for all admissions that were assigned to the comparison group and had an explicit transfusion adverse event (TAE) diagnosis code (International Statistical Classification of Diseases and Related Health Problems 9-CM).

Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in the notes.

View this table:
Table 2. Analysis of T (transfusion) and C (comparison) groups for iatrogenic diagnosis codes.

If all diagnoses in the table are dependent, alpha is 0.0057. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

The remaining tables have 524 diagnosis codes. The following TAE codes were found in the T group:

  • ICD9 code 27661, Transfusion associated circulatory overload, in Table 7

  • ICD9 code 27951, Acute graft-versus-host disease, in Table 5, which could apply to transplants or transfusion, in Table 5

  • ICD9 code 78066, Febrile nonhemolytic transfusion reaction, in Table 5

  • ICD9 code 9998, Other and unspecified transfusion reaction not elsewhere classified, in Table 2

  • ICD9 code 99989, Other transfusion reaction, in Table 2

Most of them were at least one of: predated admission (53), known reasons or associated with reasons to transfuse (315), known risk factors for TAE (12), sequelae of TAE (120), clinically very similar to known TAE (120), or known TAE (13).

For two explicit TAE codes, codes for similar conditions were also statistically significantly more common in T.

  • For febrile nonhemolytic transfusion reaction in Table 5, 78061 (Fever presenting with conditions classified elsewhere) was also much more common in T than C and was assigned much more often (196 admissions vs 3).

  • For transfusion associated circulatory overload in Table 7, 2766 (fluid overload disorder) was also much more common in T than C and was assigned much more often (571 admissions vs 13).

View this table:
Table 3. Analysis of T (transfusion) and C (comparison) groups for circulation diagnosis codes.

If all diagnoses in the table are dependent, alpha is 0.00069. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

View this table:
Table 4. Analysis of T (transfusion) and C (comparison) groups for bleeding and anemia diagnosis codes.

If all diagnoses in the table are dependent, alpha is 0.00026. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

View this table:
Table 5. Analysis of T (transfusion) and C (comparison) groups for immune, neoplasm, and infection diagnosis codes.

If all diagnoses in the table are dependent, alpha is 0.00049. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

View this table:
Table 6. Analysis of T (transfusion) and C (comparison) groups for lungs and oral diagnosis codes.

If all diagnoses in the table are dependent, alpha is 0.0034. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

View this table:
Table 7. Analysis of T (transfusion) and C (comparison) groups for diagnosis codes that relate to fluids regulation.

If all diagnoses in the table are dependent, alpha is 0.0032. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

View this table:
Table 8. Analysis of T (transfusion) and C (comparison) groups for digestion and hormones diagnosis codes.

If all diagnoses in the table are dependent, alpha is 0.00071. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

View this table:
Table 9. Analysis of T (transfusion) and C (comparison) groups for connective, nervous, and reproductive diagnosis codes.

If all diagnoses in the table are dependent, alpha is 0.0023. TAE: transfusion adverse event. Bracketed diagnosis code names were not in the MIMIC-III database, therefore were looked up [30] and confirmed in some notes.

There were codes for conditions similar to TAE that were statistically significantly more common in T:

  • Some of the hypercoagulation and emboli disorders in Table 3 and bleeding in Table 4 may have been alloimmune thrombocytopenic purpura [35].

  • Some of the anemia diagnoses in Table 4 may have been hemolytic transfusion reactions [37, 40].

  • Some of the infections in Table 5 may have been transmitted via transfusion 43, 75-79].

  • Codes in Table 6 that might have signified TRALI [65] were:

    • 5119 (Unspecified pleural effusion) [80]

    • 5118 (Other specified forms of pleural effusion except tuberculous) [80]

    • 9973 (Respiratory complications not elsewhere classified) [49, 72]

    • If transfusion was associated with the trauma or surgery [72]

      • ▪ 5185 (Pulmonary insufficiency following trauma and surgery)

      • ▪ 51851 (Acute respiratory failure following trauma and surgery)

      • ▪ 51852 (Other pulmonary insufficiency, not elsewhere classified, following trauma and surgery)

Many other codes may have been consequences of TAE, for example:

  • Kidney injury [44, 70] in Table 7, such as:

    • 5845 (Acute kidney failure with lesion of tubular necrosis)

    • 5848 (Acute kidney failure with other specified pathological lesion in kidney)

  • Esophagus injury or gastritis in Table 8 possibly due to intubation, such as:

    • 53010 (Esophagitis, unspecified) [43, 49, 86]

    • 53540 (Other specified gastritis, without mention of hemorrhage) [43]

  • Pressure ulcer [81] due to extended bed rest in Table 9:

    • 70703 (Pressure ulcer, lower back)

    • 70711 (Ulcer of thigh)

As one would expect, many of the diagnosis codes that were statistically significantly more common in T than C were, or could have been indicators of, reasons to transfuse (including preexisting conditions:

Some codes might have been preexisting conditions that raise the risk or severity of TR, for example:

  • 28262 (Hb-SS disease with crisis) [65] in Table 4

  • 2690 (Deficiency of vitamin K) [37] in Table 8

Some codes may represent both reasons to transfuse and consequences of transfusion, for example:

There are many codes that we were not able to explain. Most were vague, while some were specific, for example:

  • Benign neoplasms in Table 5

  • 5933 (Stricture or kinking of ureter) in Table 7

  • 73342 (Aseptic necrosis of head and neck of femur) in Table 9.

Many medical care AEs occur at higher frequency in hospital critical care [91, 92].

As would be expected, many diagnosis codes that were associated with transfusion were reasons to transfuse or predated the admission.

A relatively small number were explicit TAE codes. Some codes were for conditions that are the same as explicit TAE codes, but without the attribution and many of them may have been TAE. The strategy of looking for codes for conditions similar to each other has been used to analyze adverse event reports [93]. One interpretation of this phenomenon is knowledge that it is TAE but reluctance to code as such; another could be lack of awareness that the condition was a TAE, and a remaining possibility is that the potential TAE occurred before transfusion.

Our findings on two TAE codes were statistically significantly less than would have been expected from Medicare studies of the frequencies of TRALI and febrile non-haemolytic transfusion reaction (ICD9 code 78066) [20, 90] (see Table 10). Our results were not statistically significantly different for TACO and posttransfusion purpura (ICD9 code 28741) [19, 89].

View this table:
Table 10. Comparison of MIMIC-III diagnoses of TAEs (transfusion adverse events) with published Medicare data from varying time periods 2007-2012.

The expected number applies the rate found in Medicare data to the study T (transfusion) group.

Our results showed that allowing additional diagnoses that could indicate TAE could be useful for widening the net for detection of signals of TAEs.

Some diagnoses may have been consequent to TAEs.

A number of diagnoses were not explained.

Our investigation of TAE codes in C demonstrated that codes alone may not provide the full clinical picture that is in the notes. Other studies have reported that while codes can provide more information than AE reports [4], the codes themselves provide an incomplete picture [12]. Our main measures of RD and RR could be underestimates of the risks of T if the loss of information due to reliance on codes is evenly distributed across the data or otherwise biased.

The analysis of codes is merely to generate signals of AE. All signals require further investigation to assess causality because diagnosis codes generally do not provide the sequence of events before, and never during, hospitalization or assess competing explanations of the associations.

CONCLUSIONS

This study successfully demonstrated simple surveillance of patient harm by using diagnosis codes in a large healthcare database, as a supplement to current post-marketing surveillance methods. In routine practice, automation of the interpretation phase could point human monitors to the codes that warrant further investigation by clinical experts.

The limitations of full reliance on diagnosis codes motivate exploring automated methods of using the notes for surveillance of AEs.

We look forward to accessible methods for public health analysts to observe patterns in patient care and experience in real time, by detecting both anticipated and unanticipated signals in electronic health records that indicate potential safety concerns with medical care. Such a method could supplement existing safety surveillance methods.

Data Availability

Original data (MIMIC-III) are available from https://mimic.physionet.org/about/mimic/ after receiving their permission.

Financial support

The research was done with FDA support and under contract HHSF223201510027B between FDA and Booz Allen Hamilton Inc.

Conflicts of interest

None of the authors have other relevant financial interests.

Disclaimer

The opinions are those of the authors and do not represent official policy of either the FDA or Booz Allen Hamilton.

ACKNOWLEDGEMENTS

We thank enthusiastic support by our FDA and Booz Allen Hamilton supervisors, Department of Health and Human Services innovation programs (Ignite Accelerator and Data Science CoLab), and Alistair Johnson, DPhil, of the MIMIC-III program, Massachusetts Institute of Technology. George Plopper, PhD, Booz Allen Hamilton, provided project and consultation support. Many FDA colleagues offered ideas and feedback regarding the selection of the case of blood transfusion and the final paper. All authors had access to the data. Drs. Bright, Bright-Ponte, and Palmer are responsible for the study topic, design, and interpretation. Drs. Rankin and Bright formed the study groups. Drs. Bright and Blok are responsible for data analysis.

Appendix to

Bright et al, “Use of Diagnosis Codes for Blood Transfusion Adverse Events in Electronic Health Records”

View this table:
Table A1. Codes for definitions of T (transfused) and Excluded groups.

See main paper for methods.

ABBREVIATIONS USED MORE THAN ONCE

95% LCL
normal approximate 95% lower confidence limits 95%
UCL
normal approximate 95% upper confidence limits
AE
adverse events
C
comparison group of admissions
EHRs
electronic health records
FDA
Food and Drug Administration
FET
Fisher’s exact test
GI
gastrointestinal
ICD9
International Statistical Classification of Diseases and Related Health Problems 9-CM
MIMIC-III
Medical Information Mart for Intensive Care III
NOS
Not otherwise specified
PTAE
potential transfusion adverse event
RD
rate difference
RR
rate ratio
T
group of transfusion admissions
TACO
transfusion associated circulatory overload
TAE
transfusion adverse event
TRALI
transfusion related acute lung injury

REFERENCES

  1. 1.↵
    Brewer T, Colditz GA. Postmarketing surveillance and adverse drug reactions: current perspectives and future needs. JAMA. 1999 Mar 3;281(9):824–9. PMID:10071004. DOI: 10.1001/jama.281.9.824.
    OpenUrlCrossRefPubMedWeb of Science
  2. 2.↵
    Scott HD, Thacher-Renshaw A, Rosenbaum SE, et al. Physician reporting of adverse drug reactions: Results of the Rhode Island Adverse Drug Reaction Reporting Project. JAMA 1990;263:1785–1788. PMID:2313850.
    OpenUrlCrossRefPubMedWeb of Science
  3. 3.
    Hoang T, Liu J, Pratt N, et al. Authenticity and credibility aware detection of adverse drug events from social media. Int J Med Inform. 2018 Dec;120:101–115. PMID:30409335. DOI:10.1016/j.ijmedinf.2018.09.002.
    OpenUrlCrossRefPubMed
  4. 4.↵
    Bright RA, Nelson RC. Automated support for pharmacovigilance: a proposed system. Pharmacoepidemiol and Drug Safety. 2002;11(2):121–125. PMID: 11998536. DOI:10.1002/pds.684.
    OpenUrlCrossRefPubMed
  5. 5.
    Samore MH, Evans RS, Lassen A, et al. Surveillance of medical device-related hazards and adverse events in hospitalized patients. JAMA 2004;291:325–34. PMID: 14734595. DOI:10.1001/jama.291.3.325.
    OpenUrlCrossRefPubMedWeb of Science
  6. 6.↵
    Bright RA. Strategy for surveillance of adverse drug events. Food Drug Law J 2007;62(3):605–615. PMID:17915403.
    OpenUrlPubMed
  7. 7.
    Classen D, Li M, Miller S, Ladner D. An electronic health record-based real-time analytics program for patient safety surveillance and improvement. Health Aff (Millwood). 2018Nov;37(11):1805–1812. PMID: 30395491. DOI:10.1377/hlthaff.2018.0728.
    OpenUrlCrossRefPubMed
  8. 8.
    Wang L, Rastegar-Mojarad M, Ji Z, et al. Detecting pharmacovigilance signals combining electronic medical records with spontaneous reports: a case study of conventional disease-modifying antirheumatic drugs for rheumatoid arthritis. Front Pharmacol. 2018 Aug 7;9:875. PMID:30131701. DOI:10.3389/fphar.2018.00875.
    OpenUrlCrossRefPubMed
  9. 9.
    Alghamdi AA, Keers RN, Sutherland A, Ashcroft DM. Prevalence and nature of medication errors and preventable adverse drug events in paediatric and neonatal intensive care settings: a systematic review. Drug Saf. 2019Dec;42(12):1423–1436. PMID:31410745. DOI:10.1007/s40264-019-00856-9.
    OpenUrlCrossRefPubMed
  10. 10.↵
    Molina FJ, Rivera PT, Cardona A, et al. Adverse events in critical care: Search and active detection through the trigger tool. World J Crit Care Med. 2018Feb4;7(1):9–15. PMID:29430403. DOI:10.5492/wjccm.v7.i1.9.
    OpenUrlCrossRefPubMed
  11. 11.↵
    Report to Congress: Update on the adoption of health information technology and related efforts to facilitate the electronic use and exchange of health information. Office of the National Coordinator for Health Information Technology, US Department of Health and Human Services. 2016 Feb. https://www.healthit.gov/sites/default/files/Attachment_1_-_2-26-16_RTC_Health_IT_Progress.pdf.
  12. 12.↵
    Melton GB, Hripcsak G. Automated detection of adverse events using natural language processing of discharge summaries. J Am Med Inform Assoc. 2005;12:448–457. PMID:15802475. DOI:10.1197/jamia.M1794.
    OpenUrlCrossRefPubMed
  13. 13.↵
    Classen D, Li M, Miller S, Ladner D. An electronic health record-based real-time analytics program for patient safety surveillance and improvement. Health Aff (Millwood). 2018Nov;37(11):1805–1812. PMID:30395491. DOI:10.1377/hlthaff.2018.0728.
    OpenUrlCrossRefPubMed
  14. 14.↵
    Patadia VK, Schuemie MJ, Coloma PM, et al. Can electronic health records databases complement spontaneous reporting system databases? A historical-reconstruction of the association of rofecoxib and acute myocardial infarction. Front Pharmacol. 2018Jun 6;9:594. PMID:29928230. DOI:10.3389/fphar.2018.00594.
    OpenUrlCrossRefPubMed
  15. 15.↵
    Zhou L, Siddiqui T, Seliger SL, et al. Text preprocessing for improving hypoglycemia detection from clinical notes - A case study of patients with diabetes. Int J Med Inform. 2019 Sep;129:374–380. PMID:31445280. DOI:10.1016/j.ijmedinf.2019.06.020.
    OpenUrlCrossRefPubMed
  16. 16.↵
    Dejam A, Malley BE, Feng M, et al. The effect of age and clinical circumstances on the outcome of red blood cell transfusion in critically ill patients. Crit Care 2014;18(4):487. PMID:25175389. DOI:10.1186/s13054-014-0487-z.
    OpenUrlCrossRefPubMed
  17. 17.↵
    Carson JL, Grossman BJ, Kleinman, S, et al. Red blood cell transfusion: a clinical practice guideline from the AABB. Ann Intern Med. 2012; 57:49–58. PMID:22751760. DOI:10.7326/0003-4819-157-1-201206190-00429.
    OpenUrlCrossRefPubMed
  18. 18.↵
    Holness L, Knippen MA, Simmons L, Lachenbruch PA. Fatalities caused by TRALI. Transfus Med Rev. 2004;18(3):184–188. PMID:15248168. DOI:10.1016/j.tmrv.2004.03.004.
    OpenUrlCrossRefPubMedWeb of Science
  19. 19.↵
    Menis M, Anderson SA, Forshee RA, et al. Transfusion-associated circulatory overload (TACO) and potential risk factors among the inpatient US elderly as recorded in Medicare administrative databases during 2011. Vox Sang. 2014Feb;106(2):144–52. PMID:23848234. DOI:10.1111/vox.12070.
    OpenUrlCrossRefPubMed
  20. 20.↵
    Menis M, Anderson SA, Forshee RA, et al. Transfusion-related acute lung injury and potential risk factors among the inpatient US elderly as recorded in Medicare claims data, during 2007 through 2011. Transfusion. 2014;54:2182–2193. PMID:24673344 DOI:10.1111/trf.12626.
    OpenUrlCrossRefPubMed
  21. 21.↵
    Reardon JM, Harmon KJ, Schult GC, Staton CA, Waller AE. Use of diagnosis codes for detection of clinically significant opioid poisoning in the emergency department: A retrospective analysis of a surveillance case definition. BMC Emerg Med. 2016 Feb 8;16:11. PMID:26856978. DOI:10.1186/s12873-016-0075-4.
    OpenUrlCrossRefPubMed
  22. 22.↵
    Hargrove J, Waller A. Motor vehicle crash case definitions and how they impact injury surveillance. N C Med J. 2018 Nov-Dec;79(6):351–357. PMID: 30397080. DOI:10.18043/ncm.79.6.351.
    OpenUrlAbstract/FREE Full Text
  23. 23.↵
    Wiese AD, Griffin MR, Stein CM, et al. Validation of discharge diagnosis codes to identify serious infections among middle age and older adults. BMJ Open. 2018Jun19;8(6):e020857. PMID:29921683. DOI:10.1136/bmjopen-2017-020857.
    OpenUrlAbstract/FREE Full Text
  24. 24.↵
    Johnson AEW, Pollard TJ, Shen L, et al. MIMIC-III, a freely accessible critical care database. Sci Data. 2016;3:160035. PMID: 27219127. DOI:10.1038/sdata.2016.35.
    OpenUrlCrossRefPubMed
  25. 25.↵
    MIMIC-III Critical Care Database. https://mimic.physionet.org/about/mimic/.
  26. 26.↵
    Code of Federal Regulations Title 45 Part 46 Protection of Human Subjects, Subpart A—Basic HHS Policy for Protection of Human Research Subjects, §46.101 (b) (4). 2000 Oct 1. https://www.govinfo.gov/content/pkg/CFR-2000-title45-vol1/pdf/CFR-2000-title45-vol1-part46.pdf.
  27. 27.↵
    Slee VN. The International Classification of Diseases: Ninth Revision (ICD-9). Ann Intern Med. 1978;88(3):424–426. PMID:629506. DOI:10.7326/0003-4819-88-3-424.
    OpenUrlCrossRefPubMedWeb of Science
  28. 28.↵
    Statistical functions (scipy.stats). The SciPy Community. 2019. https://docs.scipy.org/doc/scipy/reference/stats.html.
  29. 29.↵
    Pandas. Python Software Foundation. https://pypi.org/project/pandas/.
  30. 30.↵
    The Web’s Free ICD-9-CM Medical Coding Reference. CD9Data.com. http://www.icd9data.com/.
  31. 31.↵
    Comparison of two rates. MedCalc Software Ltd. 2020. https://www.scistat.com/statisticaltests/rate_comparison.php.
  32. 32.↵
    Altman DG, Bland JM. How to obtain the confidence interval from a P value. BMJ 2011;343–344. PMID:21824904. DOI:10.1136/bmj.d2090.
    OpenUrlFREE Full Text
  33. 33.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Stanworth SJ, Tinmouth AT. Plasma transfusion and use of albumin. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4 th ed. Blackwell Publishing Ltd. 2009;287-297. ISBN:978-1-444-35886-5.
  34. 34.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Omi EC, Gamelli RL. Transfusion therapy in the care of trauma and burn patients. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;589-603. ISBN:978-1-444-35886-5.
  35. 35.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Arnold DM, Smith JW, Warkentin TE. Management of immune thrombocytopenia. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;344-375. ISBN:978-1-444-35886-5.
  36. 36.
    1. Juffermans NP,
    2. Walsh TS
    Juffermans NP, Walsh TS. Introduction. IN: Juffermans NP, Walsh TS, eds. Transfusion in the Intensive Care Unit. Springer, NY, 2015;1. DOI:10.1007/978-3-319-08735-1_1.
    OpenUrlCrossRef
  37. 37.↵
    1. Harmening D
    Kennedy MS, Wu HM. Transfusion therapy and transfusion in transplantation. IN: Harmening D, ed. Modern Blood Banking and Transfusion Practices, 5th ed. F.A. Davis Company, Philadelphia, 2005;303-321. ISBN:9780803612488.
  38. 38.
    Carvalho B, Quiney NF. ‘Near-miss’ hyperkalaemic cardiac arrest associated with rapid blood transfusion. Anaesthesia. 1999Nov;54(11):1094–6. PMID:10540099. DOI:10.1046/j.1365-2044.1999.01109.x.
    OpenUrlCrossRefPubMedWeb of Science
  39. 39.
    Heafner T, Bews K, Kalra M, et al. Transfusion timing and postoperative myocardial infarction and death in patients undergoing common vascular procedures. Ann Vasc Surg. 2019. PMID:31626929. DOI:10.1016/j.avsg.2019.08.090.
    OpenUrlCrossRefPubMed
  40. 40.↵
    Schmidt AE, Refaai MA, Blumberg N. Platelet transfusion and thrombosis: More questions than answers. Semin Thromb Hemost. 2016 Mar;42(2):118–24. PMID:26716501. DOI:10.1055/s-0035-1564833.
    OpenUrlCrossRefPubMed
  41. 41.
    Chang S-L, Huang Y-L, Lee M-C, et al. Association of varicose veins with incident venous thromboembolism and peripheral artery disease. JAMA. 2018; 319(8):807–817. PMID:29486040. DOI:10.1001/jama.2018.0246.
    OpenUrlCrossRefPubMed
  42. 42.
    Rossetti A, Buchs NC, Breguet R, et al. Transarterial embolization in acute colonic bleeding: review of 11 years of experience and long-term results. Int J Colorectal Dis. 2013Jun;28(6):777–82. PMID:23208010. DOI:10.1007/s00384-012-1621-5.
    OpenUrlCrossRefPubMed
  43. 43.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Vamvakas EC, Bordin JO, Blajchman MA. Immunomodularity and proinflammatory effects of allogeneic blood transfusion. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;699-717. ISBN:978-1-444-35886-5.
  44. 44.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Davenport RD. Hemolytic transfusion reactions. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;811-825. ISBN:978-1-444-35886-5.
  45. 45.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Raife TJ, Rose JS, Lentz SR. Bleeding from acquired coagulation defects and antithrombotic therapy. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;376-390. ISBN:978-1-444-35886-5.
  46. 46.
    Debus ES, Mueller-Huelsbeck S, Koelbel T, Larena-Avellaneda A. Intestinal ischemia. Int J Colorectal Dis. 2011 Sep;26(9):1087–97. PMID: 21541663 DOI:10.1007/s00384-011-1196-6.
    OpenUrlCrossRefPubMed
  47. 47.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Marks PW. Thrombocytopenia and platelet transfusion. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;199-210. ISBN:978-1-444-35886-5.
  48. 48.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Carson JL, Hébert P. Anemia and red blood cell transfusion. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4 th ed. Blackwell Publishing Ltd. 2009;131-148. ISBN:978-1-444-35886-5.
  49. 49.↵
    Halaweh O, Usmani SZ, Abouasaleh R. Barrett’s esophagitis related bronchoesophageal fistula - the diagnostic value of persistent air leak in the ventilated setting. Mcgill J Med 2007; 10(2):96–98. PMID: 18523545. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2323478/.
    OpenUrlPubMed
  50. 50.
    Shears M, Alhazzani W, Marshall JC, et al. Stress ulcer prophylaxis in critical illness: a Canadian survey. Can J Anaesth. 2016Jun; 63(6):718–724. PMID:26911559. DOI:10.1007/s12630-016-0612-3.
    OpenUrlCrossRefPubMed
  51. 51.
    Rashid F, Aber A, Iftikhar SY. A review on gastric diverticulum. World J Emerg Surg. 2012;7:1. PMID: 22257431. DOI:10.1186/1749-7922-7-1.
    OpenUrlCrossRefPubMed
  52. 52.
    Warkentin TE, Moore JC, Anand SS, et al. Gastrointestinal bleeding, angiodysplasia, cardiovascular disease, and acquired von Willebrand syndrome. Transfus Med Rev. 2003;17(4):272–286. PMID:14571395. DOI: 10.1016/S0887-7963(03)00037-3.
    OpenUrlCrossRefPubMedWeb of Science
  53. 53.
    Ivascu FA, Janczyk RJ, Bair HA, Bendick PJ, Howells GA. Spontaneous retroperitoneal hemorrhage. Am J Surg. 2005;189(3):345–347. PMID:15792766. DOI:10.1016/j.amjsurg.2004.11.020.
    OpenUrlCrossRefPubMed
  54. 54.
    Wang CY, Won CW, Shieh MJ. Aggressive colonoscopic approaches to lower intestinal bleeding. Gastroenterol Jpn. 1991;26:125–128. PMID:1884944. DOI:10.1007/BF02779280.
    OpenUrlCrossRefPubMed
  55. 55.
    1. Jameson JL,
    2. Fauci AS,
    3. Kasper DL, et al.
    Bacon BR. Cirrhosis and its complications. In: Jameson JL, Fauci AS, Kasper DL, et al., eds. Harrison’s Principles of Internal Medicine, 20th ed. McGraw-Hill Medical. ISBN:978-1-259-64404-7. Accessed 2019. https://accessmedicine.mhmedical.com/content.aspx?bookid=2129&sectionid=192283819.
  56. 56.
    Briglia AE, Anania FA. Hepatorenal syndrome: Definition, pathophysiology, and intervention. Crit Care Clin. 2002 Apr;18(2):345–73. PMID:12053838. DOI:10.1016/s0749-0704(01)00003-3.
    OpenUrlCrossRefPubMed
  57. 57.
    Morgan DJ, Kellerman R. Epistaxis: evaluation and treatment. Prim Care. 2014 Mar;41(1):63–73. PMID:24439881. DOI: 10.1016/j.pop.2013.10.007.
    OpenUrlCrossRefPubMed
  58. 58.
    Ittrich H, Bockhorn M, Klose H, Simon M. The diagnosis and treatment of hemoptysis.Dtsch Arztebl Int. 2017Jun; 114(21):371–381. PMID:28625277. DOI: 10.3238/arztebl.2017.0371.
    OpenUrlCrossRefPubMed
  59. 59.
    Gourgiotis S, Villias C, Germanos S, et al. Acute limb compartment syndrome: A review. J Surg Education. 2007; 64(3):178–186. PMID:17574182. DOI:10.1016/j.jsurg.2007.03.006.
    OpenUrlCrossRefPubMed
  60. 60.
    Sosa G, Gandham N, Landeras V, et al. Abdominal compartment syndrome. Dis Mon. 2019 Jan;65(1):5–19. PMID:30454823. DOI:10.1016/j.disamonth.2018.04.003.
    OpenUrlCrossRefPubMed
  61. 61.
    Selvan K, Edriss H, Sigler M, Nugent KM. Complications and resource utilization associated with mechanical ventilation in a medical intensive care unit in 2013. J Intensive Care Med. 2017 Feb; 32(2):146–150. PMID:26474803. DOI: 10.1177/0885066615612413.
    OpenUrlCrossRefPubMed
  62. 62.
    Rabinstein AA. Vascular myelopathies. Continuum (Minneap Minn). 2015 Feb;21(1 Spinal Cord Disorders):7–83. PMID:25651218. DOI:10.1212/01.CON.0000461085.79241.e0.
    OpenUrlCrossRefPubMed
  63. 63.
    Radford MJ, Johnson RA, Daggett WM Jr, et al. Ventricular septal rupture: a review of clinical and physiologic features and an analysis of survival. Circulation. 1981;64(3):545–53. PMID:7020978. DOI: 10.1161/01.cir.64.3.545.
    OpenUrlAbstract/FREE Full Text
  64. 64.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Sheth S. Transfusional iron overload. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;858-869. ISBN:978-1-444-35886-5.
  65. 65.↵
    Sharma S, Sharma P, Tyler LN. Transfusion of blood and blood products: indications and complications. Am Fam Physician. 2011 Mar 15;83(6):719–24. PMID:21404983. https://www.aafp.org/afp/2011/0315/p719.html.
    OpenUrlPubMed
  66. 66.
    1. Harmening D
    Harmening DM, Steffey NB, Prihoda LA, Green REB. Autoimmune hemolytic anemias. IN: Harmening D, ed. Modern Blood Banking and Transfusion Practices, 5th ed. F.A. Davis Company, Philadelphia, 2005;396-430. ISBN:9780803612488.
  67. 67.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Duffy TP. Autoimmune hemolytic anemia and paroxysmal nocturnal hemoglobinuria. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;321-343. ISBN:978-1-444-35886-5.
  68. 68.
    Gibson C, Berliner N. How we evaluate and treat neutropenia in adults. Blood. 2014Aug;124(8):1251–8. PMID:24869938. DOI:10.1182/blood-2014-02-482612.
    OpenUrlAbstract/FREE Full Text
  69. 69.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Hellstern P. Composition and mechanisms of blood coagulation and fibrinolysis. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;247-259. ISBN:978-1-444-35886-5.
  70. 70.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Tormey CA, Edward L. Snyder EL. Transfusion support for the oncology patient. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;482-497. ISBN:978-1-444-35886-5.
  71. 71.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Pomper GJ. Febrile, allergic, and nonimmune transfusion reactions. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;826-846. ISBN:978-1-444-35886-5.
  72. 72.↵
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Wallis JP, Sachs UJH. Transfusion-related acute lung injury. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;870-884. ISBN:978-1-444-35886-5.
  73. 73.
    Vianna NJ. Lymphoreticular Malignancies: Epidemiologic and Related Aspects. Medical & Technical Publishing Co. Ltd, Lancaster, UK, 1975; ISBN:839108133.
  74. 74.
    Muslimani A, Chisti MM, Nadeau L, et al. How we treat tumor lysis syndrome. Oncology J 2011;25(4):369–75. PMID:21618960. https://www.cancernetwork.com/view/how-we-treat-tumor-lysis-syndrome.
    OpenUrlPubMed
  75. 75.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Park YA, Brecher ME. Bacterial contamination of blood products. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;771-790. ISBN:978-1-444-35886-5.
  76. 76.↵
    1. Harmening D
    Cook LO, Wise SC, Larison PJ. Adverse effects of blood transfusion. IN: Harmening D, ed. Modern Blood Banking and Transfusion Practices, 5th ed. F.A. Davis Company, Philadelphia, 2005;336-358. ISBN:9780803612488.
  77. 77.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Alter HJ, Esteban-Mur JI. Transfusion-transmitted hepatitis. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;718-745. ISBN:978-1-444-35886-5.
  78. 78.
    1. Simon TL,
    2. Snyder EL,
    3. Solheim BG, et al.
    Barbara JAJ, Dow BC. Retroviruses and other viruses. IN: Simon TL, Snyder EL, Solheim BG, et al., eds. Rossi’s Principles of Transfusion Medicine, 4th ed. Blackwell Publishing Ltd. 2009;746-759. ISBN:978-1-444-35886-5.
  79. 79.
    1. Harmening D
    Williams EF, Jarreau PC, Zitzmann MP. Transfusion-transmitted disease. IN: Harmening D, ed. Modern Blood Banking and Transfusion Practices, 5th ed. F.A. Davis Company, Philadelphia, 2005;359-382. ISBN:9780803612488.
  80. 80.↵
    Finley C, Clifton J, FitzGerald JM, Yee J. Empyema: An increasing concern in Canada. Can Respir J. 2008 Mar; 15(2): 85–89. PMID:18354748. DOI:10.1155/2008/975312.
    OpenUrlCrossRefPubMedWeb of Science
  81. 81.↵
    Kalanuria AA, Ziai W, Mirski M. Ventilator-associated pneumonia in the ICU. Crit Care. 2014; 18(2): 208. PMID:25029020. DOI:10.1186/cc13775.
    OpenUrlCrossRefPubMed
  82. 82.
    Wang J-M, Lim H_K. Necrotizing fasciitis: eight-year experience and literature review. Braz J Infect Dis. 2014;18 (2):137–143. PMID:24275377. DOI:10.1016/j.bjid.2013.08.003.
    OpenUrlCrossRefPubMed
  83. 83.
    Cross MR, Greenwald MF, Dahhan A. Esophageal perforation and acute bacterial mediastinitis: Other causes of chest pain that can be easily missed. Medicine (Baltimore). 2015 Aug; 94(32): e1232. PMID:26266352. DOI:10.1097/MD.0000000000001232.
    OpenUrlCrossRefPubMed
  84. 84.
    Cheville AL, McGarvey CL, et al. Lymphedema management. Semin Radiat Oncol. 2003 Jul;13(3):290–301. PMID:12903017. DOI:10.1016/S1053-4296(03)00035-3.
    OpenUrlCrossRefPubMedWeb of Science
  85. 85.
    Karaoren G, Bakan N, Tomruk SG, et al. Bilaterally symmetrical lower extremity compartment syndrome following massive transfusion. Case Rep Orthop. 2016;2016:2718421. PMID:26885421. DOI: 10.1155/2016/2718421.
    OpenUrlCrossRefPubMed
  86. 86.↵
    Mutlu GM, Mutlu EA, Factor P. GI complications in patients receiving mechanical ventilation. Chest. 2001 Apr;119(4):1222–41. PMID:11296191. DOI:10.1378/chest.119.4.1222.
    OpenUrlCrossRefPubMedWeb of Science
  87. 87.↵
    Stangroom J. Easy Fisher exact test calculator. Social Science Statistics. 2020. https://www.socscistatistics.com/tests/fisher/default2.aspx.
  88. 88.↵
    Dunn OJ. Confidence intervals for the means of dependent, normally distributed variables. J Am Stat Assoc. 1959;54:613–21. https://www.jstor.org/stable/pdf/2282541.pdf.
    OpenUrlCrossRefWeb of Science
  89. 89.↵
    Menis M, Forshee RA, Anderson SA, et al. Posttransfusion purpura occurrence and potential risk factors among the inpatient US elderly, as recorded in large Medicare databases during 2011 through 2012. Transfusion 2015:55:284–295. PMID:25065878. DOI:10.1111/trf.12782.
    OpenUrlCrossRefPubMed
  90. 90.↵
    Menis M, Forshee RA, Anderson SA, et al. Febrile non-haemolytic transfusion reaction occurrence and potential risk factors among the U.S. elderly transfused in the inpatient setting, as recorded in Medicare databases during 2011-2012. Vox Sang. 2015;108:251–261. PMID:25470076. DOI:10.1111/vox.12215.
    OpenUrlCrossRefPubMed
  91. 91.↵
    Bates DW, Cullen DJ, Laird N. et al. Incidence of adverse drug events and potential adverse drug events: implications for prevention. JAMA.1995;274:29–34. PMID:7791255. DOI:10.1001/jama.1995.03530010043033.
    OpenUrlCrossRefPubMedWeb of Science
  92. 92.↵
    Kane-Gill SL, Kirisci L, Verrico MM, Rothschild JM. Analysis of risk factors for adverse drug events in critically ill patients. Crit Care Med. 2012;40(3):823–828. PMID:22036859. DOI:10.1097/CCM.0b013e318236f473.
    OpenUrlCrossRefPubMedWeb of Science
  93. 93.↵
    Rivkees SA, Szarfman A. Dissimilar hepatotoxicity profiles of propylthiouracil and methimazole in children. J Clin Endocrinol Metab. 2010Jul;95(7):3260–3267. PMID:20427502. DOI:10.1210/jc.2009-2546.
    OpenUrlCrossRefPubMed
Back to top
Posted January 04, 2021.
Download PDF
Data/Code
Email
Use of Diagnosis Codes to Find Blood Transfusion Adverse Events in Electronic Health Records
Roselie A. Bright, Susan J. Bright-Ponte, Lee Anne Palmer, Summer K. Rankin, Sergey Blok
medRxiv 2020.12.30.20218610; doi: https://doi.org/10.1101/2020.12.30.20218610
Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo
Citation Tools

Subject Area