|
ORIGINAL RESEARCH ARTICLE |
https://doi.org/10.5005/jp-journals-10054-0216
|
A Comparative Analysis between HaemurEx (Clinical Chemistry Analyzer) and Standard Methodologies for Determining Its Efficacy for Estimation of Various Blood Parameters Like Glucose, Cholesterol, Triglyceride, Creatinine, Direct Bilirubin, and Total Bilirubin
1–5Department of Biochemistry, JSS Medical College, JSS Academy of Higher Education and Research, Mysuru, Karnataka, India
6,7Department of Product Development, Arogyam Medisoft Solution Pvt. Ltd, Kolkata, West Bengal, India
Corresponding Author: Amrita Mukherjee, Department of Product Development, Arogyam Medisoft Solution Pvt. Ltd, Kolkata, West Bengal, India, Phone: +91 9038785693, e-mail: amrita.mukherjee@arogyammedisoft.com
How to cite this article: Shivashankar KK, Kempegowda SN, Devaraju A, et al. A Comparative Analysis between HaemurEx (Clinical Chemistry Analyzer) and Standard Methodologies for Determining Its Efficacy for Estimation of Various Blood Parameters Like Glucose, Cholesterol, Triglyceride, Creatinine, Direct Bilirubin, and Total Bilirubin. Indian J Med Biochem 2023;27(1):1–8.
Source of support: Nil
Conflict of interest: Dr Akila Prashant is associated as the Zonal Representatives of this journal and this manuscript was subjected to this journal’s standard review procedures, with this peer review handled independently of this editorial board member and her research group.
Patient consent statement: The author(s) have obtained written informed consent from the patient for publication of the case report details and related images.
Received on: 18 September 2023; Accepted on: 04 November 2023; Published on: 05 December 2023
ABSTRACT
Background: The rapidly increasing cases of noncommunicable diseases (NCDs) are required to be managed to maintain the sustainable development of society. HaemurEx, developed by Arogyam Medisoft Solution, is a battery-operated, lightweight, photometric clinical chemistry analyzer with the ability to transmit data to a remote health platform. It can evaluate different blood and urine biomarkers related to various NCDs and women’s health at the community level. This novel device can advance universal access to healthcare by enhancing the availability, accessibility, and affordability of testing the blood and urine biomarkers in primary health setup and thus can improve screening and monitoring of NCD conditions. The present study compares the results obtained from HaemurEx with the standard methodologies for patient care to determine its efficacy and accuracy.
Materials and methods: The amount of glucose, total cholesterol, triglyceride, total bilirubin, direct bilirubin, and creatinine were estimated in 40 blood samples collected from the inpatient and outpatient facilities at JSS Medical College, Mysuru, Karnataka, India using the standard methodologies. The same samples were analyzed on the same day using HaemurEx. The results obtained from both methods were compared to validate the performance characteristics of HaemurEx for its operation.
Results: The comparison between the results revealed the accuracy of HaemurEx for the above-mentioned parameters to be in the range of 85–97%.
Conclusion: HaemurEx as a screening tool has the potential to significantly impact the diagnosis and treatment of different NCDs and diseases related to women in low-resource areas.
Keywords: Blood analysis, Clinical chemistry analyzer, HaemurEx, Noncommunicable disease, Remote healthcare.
INTRODUCTION
Providing efficient, accessible, and affordable healthcare solutions to the majority of the population has become one of the most challenging areas in the current world. The limitations of physical healthcare system infrastructure and supplies in remote areas in a country like India coupled with the increased cost of transportation to the health centers have made the scenario much worse than imagined and have created a large gap in accessing basic health care.1
Noncommunicable diseases (NCDs) have become major concerns for the last few decades. As per the World Health Organization (WHO), cardiovascular diseases account for most of the NCD deaths (17.9 million people annually), followed by cancers (9.3 million), respiratory diseases (4.1 million), and diabetes (1.5 million). Most importantly, these diseases cause the death of almost 41 million people each year, equivalent to 71% of all deaths globally.2 Moreover, as per the WHO, each year, more than 15 million people die from NCDs between the ages of 30 and 69 years, whereas 85% of these “premature” deaths occur in low- and middle-income countries. In the low-resource areas, treatment, and management of NCDs can cost a major portion of the income, especially for people living on a daily wage. Thus, the situation leads to a threatened progress toward the sustainable development of society. As per the WHO, an important way to control NCDs is early detection, screening, treatment, and providing access to healthcare for people in need.
In this context, NCD management interventions like telehealth platforms connected to low-cost in vitro diagnostic devices (IVDs) including portable clinical chemistry analyzers can be of great importance and improve easy access to preventive healthcare. These devices can be used in various doctors’ offices, hospitals, and patients’ homes and can give quick feedback on many types of medical tests. It can be very beneficial in rural areas for the early detection of various NCD conditions such as diabetes mellitus, renal impairment, chronic nephritis, obstruction of the urinary tract, muscle dystrophy, various renal diseases, gout, hypolipoproteinemia, liver function, biliary function, intestinal absorption, thyroid function, adrenal disease, coronary artery disease, and cirrhosis of the liver.3–7 These types of devices can also be very useful for detecting and managing other underlying conditions, indicated through the level of different blood and urine biomarkers or cellular materials in the blood.8,9
Arogyam Medisoft Solution Pvt. Ltd, Kolkata, West Bengal, India has developed HaemurEx (Fig. 1), a photometric clinical chemistry analyzer, with the ability to transmit information to a remote server that provides digital health care to patients in low-resource areas. Arogyam Medisoft Solution Pvt. Ltd, Kolkata, West Bengal, India is registered under Startup India and incubated at Startup Incubation and Innovation Centre (SIIC), Indian Institute of Technology, Kanpur, Uttar Pradesh, India. HaemurEx is less than 500 gm of weight, runs on an inbuilt battery, and uses a smart device interface. HaemurEx, a patented class A IVD, is based on microfluidics, machine learning, image processing, and Internet of Things (IoT) technologies. The device has passed the testing for safety requirements for electrical equipment for measuring, control, and laboratory use as per part 1, general requirement of EN 61010-1:2010+A1:2019 from a National Accreditation Board for Testing and Calibration Laboratories (NABL)-accredited lab. It is designed and manufactured in the European CE, ISO 13485, ICMED 13485, and National Small Industries Corporation (NSIC)-certified facility in India. It has also received Central Drugs Standard Control Organisation (CDSCO) clearance from the Government of India. The device has an insertion slot through which a tray containing the blood or urine sample exposed to approved reagents can be inserted. In its current design, there are two different kinds of trays, one is for analyzing blood parameters and another is for analyzing urine parameters. It can test:
Fig. 1: HaemurEx—clinical chemistry analyzer
Hemoglobin, bilirubin (total, direct, and indirect), alanine transaminase [serum glutamate pyruvate transaminase (SGPT)], aspartate transaminase [serum glutamic-oxaloacetic transaminase (SGOT)], total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein cholesterol (LDL), triglyceride, albumin, creatinine, urea, uric acid, glucose, hemoglobin A1C (HbA1C), C-reactive protein (CRP), sodium (Na), potassium (K), calcium (Ca), and magnesium (Mg) in blood;
Glucose, protein, ketone, bilirubin, leukocytes, nitrite, urobilinogen, leukocytes, specific gravity, and pH in urine.
HaemurEx can transmit test results along with patient and sample identifiers using a secured HTTP protocol to a remote cloud server. It can also eliminate the risk of error due to manual transcription of the test result through its in-built integration with a cloud-based laboratory information management system. Its cloud server is hosted on the European Union General Data Protection Regulation (EU GDPR) and Health Insurance Portability and Accountability Act (HIPPA)-compliant Google Cloud Platform10,11 and it does not store any information on the device.
The accuracy of the device has been validated at the School of Tropical Medicine, Kolkata, through an interim validation procedure, has been found in the range of 82–93%.12
The present study deals with the comparison between the data obtained from the standard comparator device which is Integrated Roche Cobas 6000-c501 component (Roche Diagnostics GmbH, Mannheim, Germany) with HaemurEx for the estimation of various blood parameters like glucose, cholesterol, triglyceride, creatinine, total bilirubin and direct bilirubin by determining the accuracy, sensitivity, specificity for each comparison along with their correlation coefficient at JSS Medical College, Mysuru, Karnataka, India. This study aims to validate the performance characteristics of HaemurEx for its operational qualification.
MATERIALS AND METHODS
A total of 40 patients of any sex, admitted at the inpatient and outpatient facility of the Department of Biochemistry, JSS Medical College, Mysuru, Karnataka, India who otherwise have been advised blood examination by the treating physician as part of their current treatment plan were included in the study.
The sample size was determined in such a manner so that the study represents an adequate population.13 This sample size represents approximately 1 billion population with 95% confidence level and 15% margin of error.14
Calibration of HaemurEx Device
For each of the parameters, quality controls (QCs) of HaemurEx were ensured for each of these tests by calibrator solutions (blank/internal standard/control). The internal standards/calibrators were procured from Anamol Laboratories Pvt. Ltd, Palghar, Maharashtra, India vide license number 25-KD/599 dated 4 November 2016. Following recommendations from the United States Food and Drug Administration (US FDA) guidance for the industry for bioanalytical method validation, released in May 2018 had been used for this purpose.15 Four QCs, including lower limit of quantitation (LLOQ), low (L, defined as three times the LLOQ), mid (M, defined as mid-range), and high (H, defined as high range) from 10 replicates in 5 runs were used to establish accuracy and precision. The following acceptance criteria were considered to meet the requirement of the calibration curve:
Accuracy: Within run and between runs, ±15% of nominal concentrations; except ±20% at LLOQ.
Precision: Within run and between runs, ±15% CV (coefficient of variation); except ±20% CV at LLOQ.
Estimation of the Different Blood Parameters by the Comparator Device
After completing the initial QC by Arogyam Medisoft Solution, the device was delivered to the Department of Biochemistry, JSS Medical College, Mysuru, Karnataka, India for the project and the following comparison study was carried out independently by three researchers of JSS Medical College, Mysuru, Karnataka, India. Blood samples were collected from 40 patients for estimation of the target biomarkers following the standard analytical methodologies (Table 1) in the Department of Biochemistry at JSS Medical College, Mysuru, Karnataka, India for 1 month without performing any further calibration.16–22 All the measurements were performed using Roche dedicated reagents except direct bilirubin which was done using Doumas Roche direct bilirubin.
| S. No. | Parameter | Unit of measurement | Standard methodology |
|---|---|---|---|
| 1 | Glucose | mg/dL | Hexokinase method17,18 |
| 2 | Cholesterol | mg/dL | CHOD–POD method19,20 |
| 3 | Triglyceride | mg/dL | Enzymatic end point21 |
| 4 | Creatinine | mg/dL | Enzymatic IFCC-IDMS standardized method22 |
| 5 | Total bilirubin | mg/dL | Diazonium ion blanked (Roche) method23 |
| 6 | Direct bilirubin | mg/dL | Diazotization method23 |
Estimation of the Different Blood Parameters by HaemurEx
The serums collected from the same 40 samples were used to analyze the amount of the target biomarkers using HaemurEx on the same day when they were tested through standard methodology. The procedure for testing through HaemurEx is described below. About 1 mL of venous blood sample was collected in a clean and dry vial and the serum was isolated by centrifugation. About 100 µL of serum sample must be separated. Serum samples must be used for testing within 2–3 hours of blood collection. About 1000 µL (1 mL) of the respective reagent for each biomarker was taken in a tube and 10 µL serum sample was added to it and then it was mixed. The reagents for each target biomarker were manufactured by Anamol Laboratories Pvt. Ltd, Maharashtra, India, and were provided with HaemurEx. The sample and reagent mixture were then kept at the required incubation temperature for the stipulated times (Table 2) and 500 µL (0.5 mL) of the reaction mixture was then taken in a 0.5 mL tube using a micropipette. The 0.5 mL tube containing the reaction mixture was put in the microtray provided with HaemurEx facing the clear side. The microtray was then inserted in a particular slot in the device to take readings. The result is reflected on the digital surface of the device after clicking on the measure button.
| S. No. | Parameter | Time of incubation | Temperature of incubation | Type of reaction |
|---|---|---|---|---|
| 1 | Glucose | 10 minutes | 37°C | End-point |
| 2 | Triglyceride | 10 minutes | 37°C | End-point |
| 3 | Cholesterol | 10 minutes | 37°C | End-point |
| 4 | Total bilirubin | 5 minutes | 20°C–25°C | End-point |
| 5 | Direct bilirubin | 5 minutes | 20°C–25°C | End-point |
| 6 | Creatinine | 30 seconds | 20°C–25°C | Kinetics |
Results obtained from the Integrated Roche Cobas 6000 (c501 component) automated analyzer were compared with the results received from the HaemurEx device.
Statistical Analysis
Data collected in this comparative analysis were compiled and analyzed using the following statistical methods. The bland-Altman plot was used for studying the mean difference and constructing limits of agreement between the results from the comparator device and the results received from the HaemurEx apparatus. The plot would show the mean bias ± standard deviation to establish the agreement between the two methods.23 Receiver operating characteristic (ROC) plot was also used for determining agreement between the results from the comparator device (reference data) and the results from the HaemurEx apparatus (test data).
Statistical analyses were performed using MedCalc for Windows (MedCalc Software, Ostend, Belgium). The correlation factor was determined from the Bland–Altman plot. Bland–Altman plot method is used to estimate agreement between the results from the comparator equipment and the results received from the HaemurEx device. The US FDA guidance for the industry for bioanalytical method validation, released in May 201815 is used as a guideline for validation acceptance criteria of specificity, sensitivity, and accuracy. The following acceptance criteria would be considered to meet requirement of the specificity and sensitivity in the clinically significant region of the measurement:
Accuracy, ±20% of nominal concentration
Precision, ±20% CV
RESULTS
A total of 40 blood samples were collected from the patients admitted in the inpatient and the outpatient facility of the Biochemistry department of JSS Medical College, Mysuru, Karnataka, India for the estimation of six blood parameters. The details of the age and sex distribution of the patients for different parameters are given in Table 3.
| S. No. | Parameter | Age | Sex |
|---|---|---|---|
| 1 | Glucose | 4 days to 88 years | Number of males: 20 |
| Number of females: 20 | |||
| 2 | Total cholesterol | 16 years to 80 years | Number of males: 11 |
| Number of females: 29 | |||
| 3 | Triglycerides | 16 years to 80 years | Number of males: 28 |
| Number of females: 12 | |||
| 4 | Direct bilirubin | 44 days to 85 years | Number of males: 19 |
| Number of females: 21 | |||
| 5 | Total bilirubin | 2 days to 80 years | Number of males: 22 |
| Number of females: 18 | |||
| 6 | Creatinine | 18 years to 82 years | Number of males: 25 |
| Number of females: 15 |
Glucose, total cholesterol, triglycerides, creatinine, total bilirubin, and direct bilirubin were measured, and the data obtained from the standard device and HaemurEx were compared. The sensitivity, specificity, and accuracy (in terms of percent deviation) for each comparison were determined along with their correlation coefficient (Fig. 2). The data obtained from the comparator device and that of the HaemurEx device were found to be highly correlated for all the parameters (Table 4). The specificity, sensitivity, and accuracy were also calculated for HaemurEx as compared with the gold standard device for all the parameters (Table 4). The HaemurEx device has shown high performance as evident from the ROC curve and the AUC values (Fig. 2). The measuring ranges and linearity of the test systems for the different parameters are mentioned in Table 5.
| S. No. | Parameter | Number of samples | Specificity (%) | Sensitivity (%) | Accuracy (%) | Correlation coefficient (r) |
|---|---|---|---|---|---|---|
| 1 | Glucose | 40 | 100 | 72.73 | 92.5 | 0.9767 |
| 2 | Total cholesterol | 40 | 93.75 | 75 | 90 | 0.7859 |
| 3 | Triglyceride | 40 | 95.24 | 73.68 | 85 | 0.8842 |
| 4 | Total bilirubin | 39 | 100 | 87.50 | 97.44 | 0.9966 |
| 5 | Direct bilirubin | 39 | 94.29 | 75 | 92.31 | 0.9937 |
| 6 | Creatinine | 40 | 90 | 80 | 87.50 | 0.8421 |
| S. No. | Parameter | Measuring range | Linearity |
|---|---|---|---|
| 1 | Glucose | 20–500 mg/dL | 20–500 mg/dL |
| 2 | Total cholesterol | 10–1000 mg/dL | 85–500 mg/dL |
| 3 | Triglyceride | 11–1600 mg/dL | 35–800 mg/dL |
| 4 | Total bilirubin | 0.15–20 mg/dL | 0.15–20 mg/dL |
| 5 | Direct bilirubin | Up to 0.5 mg/dL | 0.15–20 mg/dL |
| 6 | Creatinine | 0.6–1.4 mg/dL | 0.3–30 mg/dL |
Figs 2A to L: Bland–Altman plot and ROC curve for agreement analysis between different parameters measured through HaemurEx and standard protocol with 95% standard interval
DISCUSSION
The present study has established a high level of accuracy, sensitivity, and specificity of HaemurEx for six biomarkers of blood (glucose, triglyceride, cholesterol, creatinine, total bilirubin, and direct bilirubin). The accuracy of the tests has been found in the range of 85–97.4%. In this context, as per the recommendations from the US FDA guidance for industry for bioanalytical method validation, released in May 2018, the acceptance criteria for incurred sample reanalysis (ISR) is that 67% of the samples should be ±20% of the mean.15
Based on the previous study,12 HaemurEx has been integrated with a small 37°C incubator, which can also be battery operated to make it more convenient and useful in rural areas where getting continuous electricity is a challenge.
In the future, similar studies will be conducted to evaluate the performance of HaemurEx by collecting blood samples through the capillary method. Further studies should be done to check for other parameters like SGOT, SGPT, HDL, and LDL which would make HaemurEx more effective in the low resource areas, where the availability of technical personnel is less,24 for the detection and prevention of the disease conditions early.
CONCLUSION
The ease of use, portability, no dependence on external power supply, and calibration-independent nature of HaemurEx allow it to be a good candidate for a point-of-care testing instrument in rural healthcare and homecare. Coupled with new age technology of tele-enabled digital electrocardiogram (ECG) and other devices connected to a digital health platform, this novel system can be deployed in all primary health centers and subcenters and outreach clinics in combination with provisioning of physician consultation and essential medicine and can advance universal access to healthcare in countries like India.
Clinical Significance
The ability to test various blood and urine biomarkers specifically in rural healthcare can be beneficial for the successful management of various NCDs. This also helps in detecting and managing other underlying symptoms, indicated through the level of these biomarkers within a short time.
The high accuracy and sensitivity of HaemurEx lead to the early diagnosis of diseases leading to early treatment decisions and monitoring resulting in a fewer number of deaths especially in rural areas. HaemurEx helps the diagnostic technicians to perform many tests within a short time with high accuracy which helps in performance optimization, quality management, and training.
ETHICAL APPROVAL
The study protocol and related documents were reviewed and approved by the Institutional Ethics Committee of JSS Medical College; Mysuru, Karnataka, India vide letter No. JSSMC/IEC/18.02.2022/21 NCT/2021-22 dated 3 March 2022.
REFERENCES
1. Banerjee A, Deaton A, Duflo E. Health care delivery in rural Rajasthan. Econ Polit Weekly 2004;39(09):944–949. Available from: https://www.jstor.org/stable/4414706.
2. World Health Organization. Non-communicable diseases. Available at: https://www.who.int/news-room/fact-sheets/detail/noncommunicable-diseases. Accessed on: July 2022.
3. Gubala V, Harris LF, Ricco AJ, et al. Point of care diagnostics: Status and future. Anal Chem 2012;84(2):487–515. DOI: 10.1021/ac2030199.
4. Altinier S, Zaninotto M, Mion M, et al. Point-of-care testing of cardiac markers: Results from an experience in an emergency department. Clin Chim Acta 2001;311(1):67–72. DOI: 10.1016/s0009-8981(01)00562-9.
5. Louie RF, Tang Z, Shelby DG, et al. Point-of-care testing: Millennium technology for critical care. Lab Med 2000;31(7):402–408. DOI: 10.1309/0Y5F-B7NP-5Y67-GW7T.
6. St-Louis P. Status of point-of-care testing: Promise, realities, and possibilities. Clin Biochem 2000;33(6):427–440. DOI: 10.1016/s0009-9120(00)00138-7.
7. Kuwa K, Nakayama T, Hoshino T, et al. Relationships of glucose concentrations in capillary whole blood, venous whole blood and venous plasma. Clin Chim Acta 2001;307(1–2):187–192. DOI: 10.1016/s0009-8981(01)00426-0.
8. Klonoff DC. Continuous glucose monitoring: Roadmap for 21st century diabetes therapy. Diabetes Care 2005;28(5):1231–1239. DOI: 10.2337/diacare.28.5.1231.
9. Shephard MDS, Mazzachi BC, Shephard AK. Comparative performance of two point-of-care analysers for lipid testing. Clin Lab 2007;53(9–12):561–566. PMID: 18257461.
10. The European Parliament and the Council of the European Union. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation) (Text with EEA relevance), Official Journal of the European Union 2016. Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32016R0679. Accessed on: November 2020.
11. Google Cloud Platform. HIPAA compliance on Google Cloud platform. Available at: https://cloud.google.com/security/compliance/hipaa. Accessed on: July 2022.
12. Sen S, Chakraborty P, Mukherjee A. Analysis of glucose, cholesterol, and triglyceride in blood by a novel point-of care device: Validation and interim analysis. Indian J Med Biochem 2021;25(1):14–18. DOI: 10.5005/jp-journals-10054-0177.
13. Naing NN. Determination of sample size. Mal J Med Sci 2003;10:84–86.
14. www.surveymonkey.com. Sample size calculator. Availabt at: https://www.surveymonkey.com/mp/sample-size-calculator/. Accessed on: July 2022.
15. The US Department of Health and Human Services; Food and Drug Administration Center for Drug Evaluation and Research (CDER); Center for Veterinary Medicine (CVM). Bioanalytical method validation, guidance for industry. 2018. Available at: https://www.fda.gov/files/drugs/published/Bioanalytical-Method-Validation-Guidance-for-Industry.pdf. Accessed on: July 2022.
16. Bondar JL, Mead DC. Evaluation of glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides in the hexokinase method for determining glucose in serum. Clin Chem 1974; 20(5):586–590.PMID: 4363766.
17. Neese JW, Duncan P, Bayse D, et al. Development and evaluation of a hexokinase/glucose-6-phosphate dehydrogenase procedure for use as a national glucose reference method. Center for Disease Control DHEW Publication No. (CDC) 77–8330. Public Health Service, Atlanta, GA; 1976.
18. Meiattini F, Prencipe L, Bardelli F, et al. The 4-hydroxybenzoate/4-aminophenazone chromogenic system used in the enzymic determination of serum cholesterol. Clin Chem 1978; 24(12):2161–2165.PMID: 719864.
19. Abhijit KR, Kusuma KS, Vasudha KC, et al. Estimation of total cholesterol from a single dried blood spot: New application for an age old technique. Ind J Appl Res 2016;6(2):130–135. DOI: 10.36106/ijar.
20. Heather M, Barbour HM. Enzymatic determination of cholesterol and triglyceride with the Abbott bichromatic analyser. Ann din Biochem 1977;14(1):22–28. DOI: 10.1177/000456327701400104.
21. Myers GL, Miller WG, Coresh J, et al. Recommendations for improving serum creatinine measurement: A report from the Laboratory Working Group of the National Kidney Disease Education Program. Clin Chem 2006;52(1):5–18. DOI: 10.1373/clinchem.2005.0525144.
22. Suzuki Y, Sakagishi Y. Determination of Serum Bilirubin by the Diazo Method Using the Diazotized 3-Nitroaniline Reacting Readily with the Photoproducts of Bilirubin. Jpn J Clin Chem 1994;23:158–163. DOI: https://doi.org/10.14921/jscc1971b.23.2_158.
23. Doğan NÖ. Bland–Altman analysis: A paradigm to understand correlation and agreement, Turk J Emergency Med 2018;18(4):139–141. DOI: 10.1016/j.tjem.2018.09.001.
24. Rao M, Rao KD, Kumar AKS, et al. Human resources for health in India. Lancet 2011;377(9765):587–598. DOI: 10.1016/S0140-6736(10)61888-0.
________________________
© The Author(s). 2023 Open Access. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted use, distribution, and non-commercial reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.