help bronch
-------------------------------------------------------------------------------

Title


    bronch -- Calculates severity of illness in bronchiolitis using the
        National Children's Hospital severity of bronchiolitis model,
        (NCH-SOB).



Syntax

        bronch [ work_of_breathing tachycardia age dehydration] {if} [,
                 options]

        bronchi [ work_of_breathing tachycardia age dehydration] {if} [,
                 options]


    options               Description
    -------------------------------------------------------------------------
    Main
      nch                 uses coefficients based on an inpatient derivation
                            dataset, the default
      olhsc               uses coefficients based on an emergency department
                            validation dataset
      printscreen         sends the output to a printer
      admit               calculates the probability of admission
      discharge           calculates the probability of discharge
      probabilities       calculates the probability of a patient falling
                            into another severity category
    -------------------------------------------------------------------------




Description

    bronch calculates the severity of illness for infants presenting with
    bronchiolitis. It relies on four variables age(in months), dehydration
    classified as none(0) mild(1) moderate(2) or severe(3), tachycardia
    (heart rate at or above the 98th centile for age), Increased work of
    breathing defined as (1)for moderate or severe increase in work of
    breathing and (0) otherwise.
  
    This command categorizes bronchiolitis severity as mild, moderate or
    severe using an ordinal logistic regression model.  bronch expects data
    in the wide format i.e. one observation per row and all the required
    variables in that row. When issuing bronch the order in which the
    variables are specified matters. After executing bronch stata issues a
    warning to the user reminding them which variable has been used to
    specify each component in the model. It is the user's to ensure that they
    have correctly specified the variables used to calculate the score.
 
 


Options

        +------+
    ----+ Main +-------------------------------------------------------------

    nch uses coefficients based on an inpatient derivation dataset. This is
        the default.  The coefficients (odds ratio)for this option are as
        follows. Age *** ()per month, Dehydration ***() per step increase
        from none,mild, moderate, severe, Tachycardia ** (), Increased work
        of breathing** ()

    olhsc uses coefficients based on an emergency department validation
        dataset.  The coefficients (odds ratio)for this option are as
        follows. Age *** ()per month, Dehydration ***() per step increase
        from none,mild, moderate, severe, Tachycardia ** (), Increased work
        of breathing** ()

    probabilities calculates the proabability for each observation falling
        within each bronchiolitis severity classification (mild, moderate,
        severe). This option creates three new variables for each observation
        _pmild _pmoderate and _psevere containing the values for the mild
        moderate and severe categories respectively.

    admit calculates the proabability for each observation of not having mild
        disease. This is the probability of admission when using the ohlsc
        option and the probability of being discharged in less than one day
        with the nch option. This option creates a new variable for each
        observation _padmit.

    printscreen sends the output to the screen.

    generate(newvar) creates newvar containing the value for bronchiolitis
        severity.If not specified the default is bronch.


Saved Results

        +---------+
    ----+ Support +----------------------------------------------------------

    _bronch contains the severity of bronchiolitis (i.e. mild, moderate, or
        severe)

    _pmild contains the probability of mild disease

    _pmoderate contains the probability of moderate disease

    _psevere contains the probability of severe disease

    _tachycardia_product contains the product of the coeficient (odds ratio)
        of tachycardia and the record value for each observation

    _age_product contains the product of the coeficient (odds ratio) of age
        and the record value for each observation

    _dehydration_product contains the product of the coeficient (odds ratio)
        of dehydration and the record value for each observation

    _work_of_breathing_product contains the product of the coeficient (odds
        ratio) of work of breathing and the record value for each observation

    _Z contains the summation of the products of the four parameters for each
        observation

    _Z_difference_Zi_k1 contains the data that corresponds to mild disease
        for each observation

    _difference_Zi_k2 contains the data that corresponds to both mild and
        moderate disease for each observation

    _probability_Zi_k2 contains the combined probability of mild and moderate
        disease for each observation

    _logit1_1 contains the probability of mild disease

    _logit2_1 contains the probability of moderate disease

    _logit3_1 contains the probability of severe disease

    _padmit contains the probability of admission using anything greater than
        mild disease as a surrogate marker for a patient that should be
        admitted

    _discharge contains the probability of discharge using mild disease as a
        surrogote marker for a dischargable patient

    _wob_processed contains a simplified version of the mapped work of
        breathing data

    _wob_processed_mod contains the mapped work of breathing data as
        described in the algorithm in reference 1

    _age_processed contains the mapped age data as described in the algorithm
        in reference 1

    _dehyd_processed contains the mapped dehydration data as described in the
        algorithm in reference 1

    _tachy_processed contains the mapped tachycardia data as described in the
        algorithm in reference 1

    _missing_wob contains the number of observation with missing data for
        work of breathing

    _missing_age contains the number of observation with missing data for age

    _missing_tachy contains the number of observation with missing data for
        tachycardia

    _missing_dehydration contains the number of observation with missing data
        for dehydration

    _valid_calculation contains the number of observation with valid data


Remarks

 The NCH-SOB was derived in one childrens hospital where mild was defined as im
> mediate 
 discharge on review of the resident admission by an attending pediatrician, mo
> derate as 
 hospital admission less than and including median length of stay, and severe d
> isease as 
 admission greater than the median length of stay. The model was validated in a
> nother 
 childrens hospital Emergency Department. In the validation 'mild' severity was
>  based on 
 actual discharge from the  pediatric emergency room by a resident with at leas
> t five years
 post graduate training.  'Moderate' disease was defined as hospital admission 
> less than and 
 including median length of stay, and 'severe' disease as admission for greater
>  than the 
 median length of stay. The median length of stay was the same in both derivati
> on and 
 validation phases. 
 
 
 Both models satisfied the proportional odds assumption. Users may select which
>  coefficents 
 to use based on whether they are dealing primarily with inpatient or outpatien
> t data. Full 
 details of the model used in this severity of illness classification tool can 
> be found in
 reference 1.
 

Examples


    . bronchi mild 150 2 mild

    . bronch work_of_breathing tachycardia age dehydration

    . bronch work_of_breathing tachycardia age dehydration , olhsc prob admit

    . bronch work_of_breathing tachycardia age dehydration, by(gender)



Authors

Carl E. Mitchell
Department of Emergency Medicine
Kern Medical Center
Bakersfield, CA 93306


Paul Walsh
Department of Emergency Medicine
Kern Medical Center
Bakersfield, CA 93306


References

 Walsh, P.  Rothenberg, S.J.,  O'Doherty, S., Hoey, H. and  Healy, R. 2004 A va
> lidated clinical
 model to predict the need for admission and length of stay in children with ac
> ute 
 bronchiolitis. European Journal of Emergency Medicine  11:5 265-72