Changes in version 0.1.0                        

Formula corrections (change computed values)

Several index formulas were verified against their original publications
and corrected. These fixes change the numeric output of the affected
indices, so results from 0.1.0 are not directly comparable to 0.0.1 for
these columns.

  - VAI_Men_inv / VAI_Women_inv: fixed the denominator grouping so waist
    is divided by the full (constant + slope * BMI) term, and switched
    triglycerides and HDL-C to mmol/L as required by the original
    constants (Amato 2010, doi:10.2337/dc09-1825). Previous values were
    substantially off.
  - LAP_Men_inv / LAP_Women_inv: triglycerides are now used in mmol/L
    (Kahn 2005, doi:10.1186/1471-2261-5-26); the previous mg/dL values
    were inflated by ~88.6x.
  - McAuley_index: triglycerides are now used in mmol/L (McAuley 2001,
    doi:10.2337/diacare.24.3.460).
  - Avignon_Sim: now computed as (0.137 * Si0 + Si120) / 2, the weighted
    mean defined by Avignon 1999 (doi:10.1038/sj.ijo.0800864). The
    previous expression returned an invalid, dataset-wide collapsed
    value.
  - Matsuda_ISI: the mean glucose and insulin terms are now computed in
    the same converted units (mg/dL and µU/mL) as the fasting terms,
    fixing a unit mismatch (Matsuda & DeFronzo 1999,
    doi:10.2337/diacare.22.9.1462).
  - Avignon_Sim: now uses the data-driven weighting w = mean(Si120) /
    mean(Si0), i.e. (w * Si0 + Si120) / 2, matching the source
    publication's Table 2. Note this index therefore depends on the full
    sample.
  - BigttSi: corrected the 30-minute insulin coefficient from 0.000565
    to 0.000556 (Hansen 2007, doi:10.2337/dc06-1240).
  - Homa_IR_inv: now uses glucose in mmol/L, giving the standard HOMA-IR
    (Matthews 1985, doi:10.1007/BF00280883); the previous mg/dL values
    were ~18x too large.

Maintenance

  - Homa_IR_inv and HOMA_IR_rev_inv are both retained as separate
    columns; after the units fix they return the same standard HOMA-IR
    value (computed via mmol/L / 22.5 and mg/dL / 405 respectively).
  - Removed library() calls from package source; imports are handled via
    NAMESPACE. Dropped the unused tidyr dependency.
  - Documentation now states the correct triglyceride (* 88.57) and
    HDL-C (* 38.67) unit-conversion factors, and corrects a reversed
    insulin conversion comment.

                 Changes in version 0.0.1 (2024-04-04)                  

  - Initial release of the InsuSensCalc, featuring the isi_calculator
    function.

New Features

  - isi_calculator Function: A comprehensive tool for calculating
    surrogate insulin sensitivity indices based on various measurements,
    including fasting, Oral Glucose Tolerance Test (OGTT), and lipid
    (adipose) values. This function supports a wide range of indices
    calculations, making it a versatile tool for research in metabolic
    health and diabetes.

Capabilities

  - Calculates indices using fasting glucose and insulin levels,
    including:
    
      - Fasting Insulin Sensitivity
      - HOMA-IR (and its inverse)
      - QUICKI
      - And several other fasting-related indices.

  - Incorporates OGTT (0 min, 30 min, 120 min post-glucose load) values
    for:
    
      - Gutt Index
      - Matsuda Index
      - Insulin Sensitivity Index at 120 min
      - And more, adapting calculations based on available time points.

  - Utilizes lipid (adipo) measurements like triglycerides, free fatty
    acids and HDL cholesterol for indices such as:
    
      - Visceral Adiposity Index (VAI) for Men and Women (inversed)
      - Lipid Accumulation Product (LAP)
      - ATIRI_inv

  - Utilizes tracer and dxa based data measurements like rate of
    glycerol, palmitate and dxa based fat mass for indices such as:
    
      - LIRI_inv
      - Lipo_inv
      - TyG Index (inversed)
      - And other adipose-related indices.

Flexible Input

  - The function accepts a dataframe containing the necessary variables
    for calculation and a character vector specifying the categories of
    indices to calculate ("fasting", "ogtt", "adipo"), allowing users to
    customize the scope of their analysis.

User-friendly

  - Includes comprehensive documentation and examples to facilitate easy
    use and integration into research workflows.

This release lays the foundation for robust and flexible insulin
sensitivity analysis within the R ecosystem, catering to a wide array of
research needs in metabolic health.