Skip to main content
NCBI home page
Oxford University Press logoLink to Oxford University Press
. 2020 Feb 4;65(12):1554–1562. doi: 10.1373/clinchem.2019.309575

Inadequate Reporting of Analytical Characteristics of Biomarkers Used in Clinical Research: A Threat to Interpretation and Replication of Study Findings

Qian Sun 1, Kerry J Welsh 1, David E Bruns 2,✉, David B Sacks 1,✉, Zhen Zhao 1,3
PMCID: PMC7055667  PMID: 31672858

Abstract

BACKGROUND

Analytical characteristics of methods to measure biomarkers determine how well the methods measure what they claim to measure. Transparent reporting of analytical characteristics allows readers to assess the validity and generalizability of clinical studies in which biomarkers are used. Our aims were to assess the reporting of analytical characteristics of biomarkers used in clinical research and to evaluate the extent of reported characterization procedures for assay precision.

METHODS

We searched 5 medical journals (Annals of Internal Medicine, JAMA: The Journal of the American Medical Association, The Lancet, The New England Journal of Medicine, and PLOS Medicine) over a 10-year period for the term “biomarker” in the full-text field. We included studies in which biomarkers were used for inclusion/exclusion of study participants, for patient classification, or as a study outcome. We tabulated the frequencies of reporting of 11 key analytical characteristics (such as analytical accuracy of test results) in the included studies.

RESULTS

A total of 544 studies and 1299 biomarker uses met the inclusion criteria. No information on analytical characteristics was reported for 67% of the biomarkers. For 65 biomarkers (3%), ≥4 characteristics were reported (range, 4–8). The manufacturer of the measurement procedure could not be determined for 688 (53%) of the 1299 biomarkers. The extent of assessments of assay imprecision, when reported, did not meet expectations for clinical use of biomarkers.

CONCLUSIONS

Reporting of the analytical performance of biomarker measurements is variable and often absent from published clinical studies. We suggest that readers need fuller reporting of analytical characteristics to interpret study results, assess generalizability of conclusions, and compare results among clinical studies.

Transparency and reproducibility are fundamental to sound clinical research. Nevertheless, there is concern that many published studies lack transparency of reporting and that too many studies represent research waste (1–6). Many initial clinical studies show promising results, but different or even contradictory findings are seen in subsequent studies (7, 8). Several factors contribute to the lack of reproducibility in medical research (6, 9). An underappreciated contributing factor may be inadequate characterization and reporting of methods used to measure biomarkers, which are used in important ways in clinical trials.

The US National Cancer Institute defines a biomarker as “A biological molecule found in blood, other body fluids, or tissues that is a sign of a normal or abnormal process, or of a condition or disease.” Biomarkers, such as cardiac troponin and creatinine, are used widely in clinical research studies. They are used to select patients for study, to confirm diagnoses, and to categorize patients; thus, biomarkers have potential to profoundly influence findings of the studies. Accurate measurement of biomarkers is often challenging, however. The biomarkers may be present in low concentrations (e.g., cardiac troponin), and interfering substances often are present in blood samples.

Analytical characterization of tests addresses the question: How well does the test measure what it claims to measure (10)? Before new tests can be used for patient care in the US, the Food and Drug Administration (FDA)4 requires manufacturers to perform a comprehensive analytical characterization of the test. Some fundamental analytic performance characteristics, such as analytical accuracy of the results, are listed in Table 1 (11, 12). Clinical laboratories in the US must confirm that, in their hands, the analytical performance of each measurement procedure used is similar to that claimed by its manufacturer.

Table 1.

Analytical performance characteristics of tests.

Analytical performance characteristic Description
Accuracy Agreement between measured concentration and the true concentration in the sample.
Cutoff Level of biomarker that is used for the categorization of test results, e.g., as “positive” or “negative.”
Interferences/specificity Assessment of the effects of common interfering substances, such as hemoglobin, bilirubin, and lipids.
Precision/imprecision Agreement between repeated independent measurements of the same sample. Assessed as within-run, total, or unspecified analytical variation (imprecision).
Quality control Quality control includes measurement of quality control materials that contain the analyte of interest, combined with statistical analysis of results and appropriate responses to the information (such as repeat testing of a batch of samples).
Reference interval An interval, between upper and lower limits, of test results that includes a defined proportion (usually 95%) of a reference group of (usually) healthy people.
Reportable range The interval from the lowest to the highest values that can be reliably measured.
Calibration/calibration verificationa Calibration: Establishing a relation between the instrument's measurement and the actual concentration of the substance. Calibration verification: Testing materials of known concentration to ensure accuracy.a
Analytical sensitivityb Limit of detection: Lowest concentration that can be reliably identified as being qualitatively present.
Limit of quantitation: Lowest concentration that can be measured with a defined analytical imprecision, e.g., CV < 20%.
Open in a new tab
b

This usage differs from the classical definition in analytical chemistry but reflects usage in publications in the journals surveyed in this study.

In contrast to the FDA and CLIA (12) regulations that apply to tests used in clinical care settings, it is unclear to what extent the analytical performance of such tests has been characterized or clearly reported in clinical studies. In this study, we evaluated the reporting of analytical characteristics of chemical biomarkers as seen in publications in high-profile medical journals during a 10-year period.

Materials and Methods

SEARCH STRATEGY AND SELECTION CRITERIA

Five journals that publish large clinical trials were selected to obtain a sample of biomarker use in clinical studies, including randomized clinical trials of therapeutic interventions. Specifically, we surveyed publications in the Annals of Internal Medicine, JAMA: the Journal of the American Medical Association (JAMA), The Lancet, The New England Journal of Medicine (NEJM), and PLOS Medicine (PLOS Med). The archives of each journal were searched for the term “biomarker” in the full-text field for 10 years, January 1, 2006 through January 1, 2016.

Studies were selected if a biomarker was used (a) for selection of participants in the study, (b) for classification of patients, (c) as a study outcome, or (d) for any combination of these uses. Articles were excluded if they were not clinical studies in humans, had cohorts with <10 participants, used only immunohistochemical or imaging markers, were meta-analyses of primary studies, were computer modeling studies with no actual biomarker measurements, or had been retracted.

DATA EXTRACTION

For each publication, we ascertained whether the study was multicenter or was conducted at a single site, whether the biomarker measurement procedure was approved (or cleared) by the FDA, whether the biomarker test was granted waived status by CLIA, the molecular type of the biomarker (protein, nucleic acid, lipid/steroid, small molecule/mineral, cell-based analysis, vitamin/trace element, or other), whether the name of the manufacturer of the test was provided, and whether the test was quantitative or qualitative.

The FDA-approval/clearance status of each assay was determined by reference to the FDA website. CLIA-waived status of each test was evaluated based on the table entitled “Tests Granted Waived Status Under CLIA” published on the Centers for Medicare & Medicaid Services website (13).

We evaluated the reporting of 11 key analytical characteristics (Table 1): (1) analytical accuracy or trueness (estimates based on method-comparison or recovery studies were accepted, regardless of whether a primary reference measurement procedure or a certified reference material was used), (2) total (day-to-day) imprecision, (3) within-run imprecision, (4) imprecision not otherwise specified, (5) analytical “sensitivity” (limit of detection and/or limit of quantification), (6) interferences, (7) reportable range of results, (8) reference interval, (9) cutoffs for test positivity or decision limits (decision values), (10) quality control, and (11) calibration/calibration verification (reporting about either one was accepted) (14).

For each biomarker, we counted the number of the 11 analytical characteristics reported (possible range, 0–11). For nonwaived FDA-approved/cleared methods, we evaluated elements 1 through 4 and 7 through 11, as CLIA requires users to verify only these 9 items.

All available sources were used for data extraction, including the full-text journal articles, references therein (but not articles cited by the cited papers), and supplemental protocols/data published with the article. Search results and full-text articles were independently assessed by 2 clinical chemists (QS and KW). All entered data were checked for accuracy by 3 authors.

Results

SEARCH RESULTS AND CHARACTERISTICS OF INCLUDED STUDIES

We identified 778 research articles from the initial search of the full text of the archives of Annals of Internal Medicine, The Lancet, JAMA, NEJM, and PLOS Med (see Fig. 1 in the online Data Supplement). Among these articles, 544 were considered potentially relevant based on our selection criteria, and 1348 biomarkers used in the studies in those articles were considered potentially eligible. After examining the testing methods, 1299 biomarkers were included in the review. Reasons for exclusion of potentially eligible biomarkers are provided in Fig. 1 of the online Data Supplement.

Of the 1299 eligible biomarkers, 1028 (79%) were measured in multicenter studies (see Table 1 in the online Data Supplement). A total of 317 (24%) of the 1299 biomarker measurements were approved or cleared by the FDA, whereas 445 (34%) did not have such clearance or approval. For 537 (41%) biomarker tests, the FDA status was not specified and was not ascertainable because the articles contained insufficient information about the method, e.g., the name of the manufacturer was not stated.

Only 2 (0.2%) of the 1299 biomarker tests were CLIA-waived based on the test and name of the manufacturer, whereas the other 1297 (99.8%) tests were either nonwaived or the status could not be determined because of insufficient description of the method. A quantitative result was reported for 1043 (80%) of the assays, and a qualitative result was reported for 119 (9%) of the tests. The quantitative/qualitative nature of the methods could not be determined for the remaining 137 (11%) because results either were not reported or were reported with words such as “increased.”

The most frequent molecular types of biomarkers (see Table 2 in the online Data Supplement) were proteins (719 reports; 55%), nucleic acids (161 reports; 12%), and lipids/steroids (154 reports; 12%). The most common biomarkers were C-reactive protein (79 reports; 6%), cardiac troponin (63 reports; 5%), and glucose (47 reports; 4%).

OVERALL REPORTING OF ANALYTICAL CHARACTERISTICS OF BIOMARKER TESTS

For each biomarker that was included, we tabulated the number of analytical characteristics reported (possible range, 0–11 items; 0–9 items for FDA-cleared/approved methods). No characteristic was provided for 865 of 1299 (67%) biomarker measurements (Fig. 1). Five or more items (range, 5–8) were reported for only 33 biomarkers (3%) (Fig. 1). The name of the manufacturer of the measurement procedure could not be ascertained for 53% of the biomarkers.

Fig. 1. Distribution of the number of analytic performance characteristics reported for 1299 measurement systems.

Fig. 1.

Open in a new tab

For each biomarker that was included, the number of analytical characteristics reported was tabulated. For example, no characteristic was provided for 865 biomarker measurements.

For each of the 5 journals, and for the full data set, the median reporting rate was zero. The reporting rates for the individual method characteristics ranged from 2% (for interference studies) to 13% (total imprecision) (Fig. 2).

Fig. 2. Reporting rate (%) of analytic performance characteristics in 5 journals.

Fig. 2.

Open in a new tab

For each of the 11 method characteristics, the overall reporting rate and the rates in each journal were tabulated. Annals, Annals of Internal Medicine.

The examined articles sometimes lacked evidence that the authors assessed the analytical performance of the test in their laboratories and that they did so during the period of the study. For some analytical characteristics, >25% of the results that were provided were neither determined in the authors' laboratory nor based on cited papers from the authors of the subject publication (see Table 3 in the online Data Supplement). Rather, the characteristics came either from previous publications by other investigators or from the manufacturers. In addition, 15 references cited in the methods sections failed to contain pertinent information about the analytical characteristics of the methods. For example, 3 of the cited articles were reviews, and 11 others provided no method details.

REPORTING OF METHOD CHARACTERISTICS ACCORDING TO FDA REVIEW STATUS AND QUANTITATIVE/QUALITATIVE NATURE OF METHODS

FDA status of the measurement procedure had no consistent relationship with the frequency of reporting of analytical characteristics. Six of the 9 applicable analytical characteristics were reported more often for the 445 non-FDA-approved/cleared measurement procedures than for the 317 FDA-approved/cleared tests, and the frequencies of reporting were low for both (Fig. 3 here and Table 4 in the online Data Supplement). For example, accuracy (trueness) information was reported for 13% of research use-only tests vs 4% of FDA-cleared/approved tests. Information about 3 other analytical characteristics was provided more often for FDA-approved/cleared than for research use-only methods, and again the frequencies were low (Fig. 3 here and Table 4 in the online Data Supplement).

Fig. 3. Analytical characterization reporting based on (A) FDA approval status and (B) quantitative nature of methods.

Fig. 3.

Open in a new tab

A numerical scale from 0% to 25% was used to obtain the reporting rate. Two characteristics (analytical sensitivity and interferences) were not included in the analysis based on FDA approval status because these items are not required for FDA-approved methods by CLIA. Three characteristics (analytical sensitivity, interferences, and reference interval) were not included in the analysis based on quantitative nature because of the lack of numeric data in qualitative tests. QC, quality control.

Validation of qualitative tests differs from that for quantitative tests (15), but the rates of reporting of analytical characteristics were similar for 1043 quantitative methods and 119 qualitative methods (Fig. 3 here and Table 4 in the online Data Supplement).

ASSESSMENT OF THE EXTENT OF REPORTED CHARACTERIZATIONS OF ANALYTICAL IMPRECISION

As an indicator of the completeness and quality of the reported analytical characterizations, we examined the extent of analytical imprecision studies as described in the articles (see Fig. 2 in the online Data Supplement). A total of 274 studied articles provided within-run, day-to-day, and/or unspecified imprecision results. Of these, 208 (76%) reported only a total CV with no details, such as the tested concentration(s) of the biomarker, and whether the precision study was based on a single day or multiple days of testing. For only 3 biomarker tests (1%), the reports stated that precision studies were performed for at least 20 days. For 43 of 273 biomarkers (16%), the reported precision data were at ≥2 concentrations.

SPECIFIC BIOMARKERS: CARDIAC TROPONIN AND GLUCOSE

Cardiac troponin I results vary greatly among manufacturers' methods (16, 17). Of the 1299 uses of biomarkers included in our analysis, cardiac troponin measurements were the second-most common. Among these 63 uses of troponin, we assessed the reporting frequency of 5 analytical characteristics that we judged to be essential for interpreting troponin results in the context of variable results among available measurement procedures. The frequencies of reporting were all <50%, and only 35% provided the name of the manufacturer (Fig. 4). Among the 41 publications that did not include names of the troponin assay manufacturer, 38 (93%) studied participants at multiple hospitals, raising a concern that assays from different manufacturers were used; a universal cutoff was used in 4 (10%) of those reports.

Fig. 4. Reporting rate of reference interval, accuracy, precision, reportable range, cutoff, and manufacturer information in 63 cardiac troponin measurements.

Fig. 4.

Open in a new tab

Second, we examined reporting of manufacturer information for glucose assays. Of the 1299 uses of biomarkers included in our analysis, glucose ranked just below cardiac troponin in frequency. Analytical characteristics of glucose measurement procedures vary significantly, especially between point-of-care glucose meters and central laboratory methods (18, 19). Among 48 publications, 16 (33%) provided manufacturer information for the glucose assays used. In 5 of the 16 publications (31%), ≥2 methods were used to measure glucose concentrations. Two publications used pooled results from point-of-care glucose meters and laboratory methods.

Discussion

This study evaluated the extent to which published clinical studies reported analytical characteristics of the methods used to measure biomarkers. More than two-thirds of the articles provided no information on the analytical performance of the methods, and half did not identify the manufacturer of the test. Moreover, the extent of reported characterization of assay precision was limited.

Transparent reporting of medical research has received considerable attention (3, 4, 20), but reporting of analytical characteristics of medical tests used in clinical studies has not. Our results suggest that published articles provide insufficient information to answer important questions about the reported clinical studies in which biomarkers are used to select participants in the study, categorize patients, or make diagnoses. These questions include questions of generalizability and internal validity.

GENERALIZABILITY OF TRIAL RESULTS: DIFFERING ANALYTICAL RESULTS AMONG MANUFACTURERS

For many biomarkers, including tests cleared or approved by the FDA, results for a given patient sample differ depending on the manufacturer of the test (21–23). For example, a study in this Journal (24) documented marked differences among manufacturers' tests for cardiac troponin I: Results varied as much as 33-fold for 9 “sensitive-contemporary” assays, 4-fold among 5 point-of-care assays, and 3-fold for 5 high-sensitivity assays (24). Manufacturer-dependent differences occur for many other biomarkers, including thyroid-stimulating hormone, prostate-specific antigen, and human chorionic gonadotropin. Such biomarkers circulate in multiple modified forms, and different manufacturers' methods measure different subsets of those forms. When the manufacturer of the test is not identified, physicians cannot know if the decision cutpoints used in the study are appropriate for the test in their hospitals or at other laboratories where their patients' samples are assayed.

Like newer biomarkers, long-established biomarkers, such as albumin and creatinine, also display marked quantitative differences among manufacturers' products. Urine albumin assays show biases that range from −35% to 34%, resulting in large differences between methods (25). Similarly, serum (and plasma) albumin results differ markedly among manufacturers' methods (26). In the latter study, the proportion of samples that met an albumin goal of 3.5 g/dL for dialysis patients ranged from 35% to 70% depending on the manufacturer of the test (26). This marked difference was seen despite the use of the same chemical method principle (bromocresol green dye-binding) by the manufacturers. Moreover, when a manufacturer offered more than one method to measure serum albumin, the results of the methods did not agree when measuring the same (nonfrozen, human) samples (26).

As for albumin, creatinine results differ markedly among methods (27), and these differences produce erratic results in patients (28).

Inclusion of the names of the manufacturers of biomarkers represents an opportunity to improve the reporting of clinical studies.

REPRODUCIBILITY OF TRIALS: IMPACT OF ANALYTICAL PERFORMANCE

Registration of trials is important for avoiding many causes of wasted research (29), but it cannot replace a clear description of the analytical performance of biomarker measurement procedures during the time of a clinical trial. Lack of information on analytical performance characteristics, such as the limit of detection and imprecision, can lead to apparent failure to reproduce results of an earlier trial.

The limit of detection and limit of quantification of biomarker tests that are used to select patients for trials may affect the compositions of the study populations and alter the apparent effectiveness of the randomized clinical trial's intervention. When a drug is effective in early disease, it will appear effective if patients are selected for study by use of a biomarker test that has a low limit of detection and identifies patients with early disease. By contrast, if the study is repeated with a biomarker assay that is less analytically sensitive and identifies only patients with advanced disease who have high concentrations of the biomarker, the study will include a set of patients that is enriched in late-stage cases of the condition being treated. This population will be less likely to achieve the desired clinical outcome; thus, the studied intervention will appear less effective than in the previous study.

A trial (30) that was among those included in our study appeared to contradict an earlier randomized controlled trial (31). The earlier study had concluded that tight glucose control decreased mortality in critically ill patients. By contrast, the second trial found an increase in mortality with tight glucose control (30). In the earlier study, glucose was measured by a highly accurate and precise method, which was described in the publication. By contrast, in the later (multicenter) study, glucose was measured by a variety of methods and devices, including point-of-care glucose meters (30). The manufacturers' names were not provided, and no information was provided about the analytical performance of the various glucose methods during the trial.

Given the state of the art of glucose meters at the time of the second study, the imprecision of each of the point-of-care devices was certainly large. High imprecision of measurement of glucose is associated with poor glucose control (32–34). Moreover, for some glucose meters, results were falsely high in critically ill patients, whereas for other meters the results were consistently low (35, 36). Despite this variability of results, the same algorithm was used at all centers to determine the insulin infusion rates for patients based on the glucose concentrations reported by the various analyzers. Thus, patients who were monitored with meters that produced high results were at risk of receiving more insulin than was intended for their true blood glucose concentration, with a resulting risk of hypoglycemia. The magnitude of this risk was not apparent to readers because no data were reported on the quality of the glucose measurements during the course of the study. A subsequent independent publication reported poor performance at one trial site, during the trial, of glucose meters used in the multicenter study (37). Inclusion of analytical information in the report of that study might have avoided continuing controversy about the impact of errors in glucose measurements on the study's conclusions (37, 38).

LIMITATIONS

A limitation of the current study is that it included only studies that were published in 5 high-impact journals. In journals with lower impact factors, adherence to reporting guidelines may be less extensive than in journals with higher impact factors (39). Thus, our findings may overestimate the overall reporting rate of analytical characteristics in the clinical research literature. Second, we included only articles that contained the term “biomarker” and may have missed some relevant articles. However, the search was on the full text of all articles published in the journals, and it did identify a large number of published studies. It seems unlikely that use or nonuse of the term biomarker anywhere in the article would dramatically affect generalizability of our findings. Moreover, use of the selection criteria clearly showed that in numerous studies, published in high-impact clinical journals, the reporting of the analytical characteristics of tests was frequently inadequate. Third, some analytical characteristics listed in Table 1 may not be applicable to all analytes. Nonetheless, the finding that 67% of articles provided no information on analytical performance of the tests is troubling.

FACTORS THAT MAY LIMIT THE EXTENT OF REPORTING OF ANALYTICAL CHARACTERISTICS OF BIOMARKER TESTS

The limited reporting of analytical characteristics of biomarker tests may reflect several factors. One is restricted space for publication of methodological details in printed journals. This limitation, however, has been largely overcome by the use of online supplements to print articles (which was supported by each of the 5 journals that were included in this study). Second, researchers may assume that analytical characterization and method optimization are the manufacturers' responsibilities. Our analysis showed, however, that only 24% of biomarker measurements in the examined clinical studies could be ascertained to have been made with FDA-approved/cleared methods, with analytical characterization performed by their manufacturers. For other assays, analytical characterization by their manufacturers is not regulated and may not be transparent, making it important that authors report analytical characterizations for the 76% of the biomarkers included in this report that were not identified as having been cleared or approved by the FDA. Third, guidelines for reporting of clinical studies provide limited guidance on reporting of the analytical characteristics of biomarker measurements used in the studies. The CONSORT guideline, for reporting parallel-group randomized trials, addresses research methodology in detail but provides no guidance on the reporting of analytical characteristics of measurements of biomarkers. Provision of some guidance may represent another opportunity for improvement of reporting of clinical studies.

It is beyond the scope of this study to propose a set of analytical characteristics to be included in a report of a clinical trial, and the set of characteristics likely will depend in part on the specific use of a biomarker in each trial. Based on the experience of performing this study, however, we suggest that a minimal set of characteristics to report might include 3 items: (a) citation of a publication that describes the performance of the method and, if it is commercially available, at least the name of the product and its manufacturer; (b) reference interval or other decision point(s) used; and (c) imprecision, measured during the study, at the concentration (or concentrations) used as decision point(s) in the study and at (nonzero) limits of the reference interval. For studies in which the presence/absence of a detectable (or measurable) concentration of a biomarker is used for decisions, the limit of detection (or limit of quantification as appropriate) of the procedure would be an important addition.

Conclusion

Our analysis identifies shortcomings in the reporting of, and potentially the quality or even the existence of, analytical characterization of biomarker tests used in clinical studies reported in high-impact medical journals. Opportunities for improvement include simple steps, such as identifying the manufacturer when a commercially available measurement procedure is used. Improvement of reporting has the potential to enable investigators and others to (a) better evaluate the internal and external validity of the results of clinical studies, (b) assess the generalizability of the findings, and (c) replicate the results.

Acknowledgments

The authors thank Dr. Timothy Billiar (University of Pittsburgh, Pittsburgh, PA) for his help with study design and manuscript review.

4 Nonstandard abbreviations

FDA

Food and Drug Administration

JAMA

JAMA: The Journal of the American Medical Association

NEJM

The New England Journal of Medicine

PLOS Med

PLOS Medicine.

Footnotes

(see editorial on page 1479)

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 4 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved.

Q. Sun, K. Welsh, and Z. Zhao did the literature search, data extraction, and collection. Z. Zhao and Q. Sun designed the statistical analysis plan and did the analysis. Q. Sun, D.E. Bruns, D.B. Sacks, and Z. Zhao drafted the manuscript. All co-authors edited, revised, and contributed to the intellectual content of the manuscript.

Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:

Employment or Leadership: D.B. Sacks, Clinical Chemistry, AACC.

Consultant or Advisory Role: None declared.

Stock Ownership: None declared.

Honoraria: None declared.

Research Funding: D.B. Sacks, NIH. This work was supported in part by the National Institutes of Health Clinical Center Intramural Research Program.

Expert Testimony: None declared.

Patents: None declared.

Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, preparation of manuscript, or final approval of manuscript.

Contributor Information

David E Bruns, Email: david.sacks2@nih.gov.

David B Sacks, Email: zhz9010@med.cornell.edu.

References

  • 1. Nosek BA, Alter G, Banks GC, Borsboom D, Bowman SD, Breckler SJ, et al. Scientific standards. Promoting an open research culture. Science 2015;348:1422–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Prinz F, Schlange T, Asadullah K. Believe it or not: how much can we rely on published data on potential drug targets? Nat Rev Drug Discov 2011;10:712–81. [DOI] [PubMed] [Google Scholar]
  • 3. Glasziou P, Chalmers I. Research waste is still a scandal—an essay by Paul Glasziou and Iain Chalmers. BMJ 2018;363:k4645. [Google Scholar]
  • 4. Altman DG. The scandal of poor medical research. BMJ 1994;308:283–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Wen H, Wang H-Y, He X, Wu C-I. On the low reproducibility of cancer studies. Natl Sci Rev 2018;5:619–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Baker M. 1,500 scientists lift the lid on reproducibility. Nature 2016;533:452–4. [DOI] [PubMed] [Google Scholar]
  • 7. McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, Clark GM, Statistics Subcommittee of the NCIEWGoCD. Reporting recommendations for tumor MARKer prognostic studies (REMARK). Nat Clin Pract Oncol 2005;2:416–22. [PubMed] [Google Scholar]
  • 8. Begley CG, Ellis LM. Drug development: raise standards for preclinical cancer research. Nature 2012;483:531–3. [DOI] [PubMed] [Google Scholar]
  • 9. Iqbal SA, Wallach JD, Khoury MJ, Schully SD, Ioannidis JP. Reproducible research practices and transparency across the biomedical literature. PLoS Biol 2016;14:e1002333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Lathrop JT. Analytical validation and points for discussion. https://www.fda.gov/downloads/MedicalDevices/NewsEvents/WorkshopsConferences/UCM401486.pdf (Accessed July 2019).
  • 11. Clinical Chemistry. Information for authors. http://clinchem.aaccjnls.org/content/information-authors (Accessed July 2019).
  • 12. Zhao Z, Sacks DB. Assay development in clinical pathology. In: McManus LM, Mitchell RN, editors. Pathobiology of human disease: a dynamic encyclopedia of disease mechanisms. St. Louis (MO): Elsevier; 2014. p. 3194–206. [Google Scholar]
  • 13. Centers for Medicare & Medicaid Services. Tests granted waived status under CLIA. https://www.cms.gov/Regulations-and-Guidance/Legislation/CLIA/downloads/waivetbl.pdf (Accessed July 2019).
  • 14. Centers for Medicare & Medicaid Services. Calibration and calibration verification. https://www.cms.gov/Regulations-and-Guidance/Legislation/CLIA/Downloads/6065bk.pdf (Accessed July 2019).
  • 15. Pereira P. Basic validation of qualitative tests. https://www.westgard.com/validating-qualitative-tests.htm (Accessed July 2019).
  • 16. Apple FS, Sandoval Y, Jaffe AS, Ordonez-Llanos J, Clinical ITF. Cardiac troponin assays: guide to understanding analytical characteristics and their impact on clinical care. Clin Chem 2017;63:73–81. [DOI] [PubMed] [Google Scholar]
  • 17. Chenevier-Gobeaux C, Deweerdt L, Cantero AV, Renaud B, Desmaizieres B, Charpentier S, et al. Multi-centre evaluation of recent troponin assays for the diagnosis of NSTEMI. Pract Lab Med 2018;11:23–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Sacks DB, Arnold M, Bakris GL, Bruns DE, Horvath AR, Kirkman MS, et al. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Clin Chem 2011;57:e1–47. [DOI] [PubMed] [Google Scholar]
  • 19. Fraser CG, Scott MG. Analytical performance requirements for point-of-care testing: glucose as an example. In: Price C, St. John A, Kricka L, editors. Point-of-care testing: needs, opportunity and innovation. Washington (DC): AACC Press; 2010. p. 145–56. [Google Scholar]
  • 20. Glaszious P, Chalmers L. Is 85% of health research really “wasted”? https://blogs.bmj.com/bmj/2016/01/14/paul-glasziou-and-iain-chalmers-is-85-of-health-research-really-wasted/ (Accessed July 2019).
  • 21. Saenger AK, Korpi-Steiner N. Advances in cardiac biomarkers of acute coronary syndrome. Adv Clin Chem 2017;78:1–58. [DOI] [PubMed] [Google Scholar]
  • 22. Labaere D, Smismans A, Van Olmen A, Christiaens P, D'Haens G, Moons V, et al. Comparison of six different calprotectin assays for the assessment of inflammatory bowel disease. United European Gastroenterol J 2014;2:30–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Semjonow A, De Angelis G, Oberpenning F, Schmid HP, Brandt B, Hertle L. The clinical impact of different assays for prostate specific antigen. BJU Int 2000;86:590–7. [DOI] [PubMed] [Google Scholar]
  • 24. Apple FS, Ler R, Murakami MM. Determination of 19 cardiac troponin I and T assay 99th percentile values from a common presumably healthy population. Clin Chem 2012;58:1574–81. [DOI] [PubMed] [Google Scholar]
  • 25. Bachmann LM, Nilsson G, Bruns DE, McQueen MJ, Lieske JC, Zakowski JJ, Miller WG. State of the art for measurement of urine albumin: comparison of routine measurement procedures to isotope dilution tandem mass spectrometry. Clin Chem 2014;60:471–80. [DOI] [PubMed] [Google Scholar]
  • 26. Bachmann LM, Yu M, Boyd JC, Bruns DE, Miller WG. State of harmonization of 24 serum albumin measurement procedures and implications for medical decisions. Clin Chem 2017;63:770–9. [DOI] [PubMed] [Google Scholar]
  • 27. Saenger AK, Lockwood C, Snozek CL, Milz TC, Karon BS, Scott MG, et al. Catecholamine interference in enzymatic creatinine assays. Clin Chem 2009;55:1732–6. [DOI] [PubMed] [Google Scholar]
  • 28. Charifa A, Bunch DR, El-Khoury JM. Erratic serum creatinine concentrations in a cardiac patient. Clin Chem 2018;64:1543–4. [DOI] [PubMed] [Google Scholar]
  • 29. Chan AW, Song F, Vickers A, Jefferson T, Dickersin K, Gotzsche PC, et al. Increasing value and reducing waste: addressing inaccessible research. Lancet 2014;383:257–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. NICE-SUGAR Study Investigators, Finfer S, Chittock DR, Su SY, Blair D, Foster D, et al. Intensive versus conventional glucose control in critically ill patients. N Engl J Med 2009;360:1283–97. [DOI] [PubMed] [Google Scholar]
  • 31. van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001;345:1359–67. [DOI] [PubMed] [Google Scholar]
  • 32. Boyd JC, Bruns DE. Performance requirements for glucose assays in intensive care units. Clin Chem 2014;60:1463–5. [DOI] [PubMed] [Google Scholar]
  • 33. Boyd JC, Bruns DE. Effects of measurement frequency on analytical quality required for glucose measurements in intensive care units: assessments by simulation models. Clin Chem 2014;60:644–50. [DOI] [PubMed] [Google Scholar]
  • 34. Krinsley JS, Bruns DE, Boyd JC. The impact of measurement frequency on the domains of glycemic control in the critically ill—a Monte Carlo simulation. J Diabetes Sci Technol 2015;9:237–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Gijzen K, Moolenaar DLJ, Weusten JJAM, Pluim HJ, Demir AY. Is there a suitable point-of-care glucose meter for tight glycemic control? Evaluation of one home-use and four hospital-use meters in an intensive care unit. Clin Chem Lab Med 2012;50:1985–92. [DOI] [PubMed] [Google Scholar]
  • 36. Finfer S, Wernerman J, Preiser JC, Cass T, Desaive T, Hovorka R, et al. Clinical review: consensus recommendations on measurement of blood glucose and reporting glycemic control in critically ill adults. Crit Care 2013;17:229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Cembrowski GS, Tran DV, Slater-MacLean L, Chin D, Gibney RTN, Jacka M. Could susceptibility to low hematocrit interference have compromised the results of the NICE-SUGAR trial? Clin Chem 2010;56:1193–5. [DOI] [PubMed] [Google Scholar]
  • 38. Scott MG, Bruns DE, Sacks DB. Tight glucose control in critically ill adults. JAMA 2008;300:2725–6. [DOI] [PubMed] [Google Scholar]
  • 39. Samaan Z, Mbuagbaw L, Kosa D, Borg Debono V, Dillenburg R, Zhang S, et al. A systematic scoping review of adherence to reporting guidelines in health care literature. J Multidiscip Healthc 2013;6:169–88. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Clinical Chemistry are provided here courtesy of Oxford University Press

RESOURCES