1. Identify the section of the text that contains numbers such as measurements, test results, medication doses, dates, ages, etc. These sections typically contain tables, lists or sentences with numbers and units of measurement. 2. Detect numbers in the identified sections. This can be done using regular expressions to find digits and decimal points. For example, [0-9]+ can be used to find integers and [0-9]+\.[0-9]+ can be used to find decimal numbers. 3. Identify the units associated with the numbers. The units can typically be found in the text immediately before or after the number. Extract the units along with the numbers. 4. Standardize the units to a consistent format. For example, mg, MG and mgs can be standardized to mg. mL and cc can be standardized to mL. 5. Identify any qualifiers or modifiers associated with the numbers such as greater than (>), less than (<), approximately (~), etc. and extract them along with the numbers. 6. Identify the data type for the extracted numbers and units such as weight, height, blood pressure, hematocrit, date, etc. This can be done using the context around the numbers as well as the units. 7. Store the extracted numbers, units and data types for further analysis and processing. 8. Handle exceptions and special cases. For example, numbers spelled out in words need to be converted to digits. Ranges indicated using hyphens or the word 'to' need to be extracted as two separate numbers. 9. Evaluate the accuracy of the extraction using a manually annotated gold standard test set. Precision, recall and F1-score can be calculated to determine accuracy. 10. Improve the extraction based on errors identified in the evaluation. Additional rules and patterns can be added to increase accuracy. The process needs to be tailored for the particular medical domain and record system. But these general principles can help get started with extracting quantitative information from clinical text. Let me know if you have any other questions!