The Newcastle PBC score and modeling outcomes of PBC Scoring

Newcastle Variceal Risk in PBC Score Calculator

The Newcastle Risk Score:

1 / 1 + exp -(9.186 + (0.001 * alkaline phosphatase in IU) - (0.178 * (albumin in g/L)) - (0.015 * platelets x 10^9)))

After entering serum albumin in g/L, platelet count (10^9) and alkaline phosphatase in IU click the Calculate button. Your Risk Score will then be computed and displayed in the "Predicted risk of Varices =" text box

Inputs:          
  Enter Your Serum Albumin:   (g/dL)  
  Enter Your Platelets :   (x10^9)  
  Enter Your Alkaline Phosphatase :   (IU)  
  Your reference Alkaline phosphatase range†:   to (IU)  
         
Output:          
  Predicted risk of Varices =   %  

†Chuang-Stein C. Summarizing laboratory data with different reference ranges in multi-center clinical trials. Drug Information Journal 1992; 26(1): 77-84

Newcastle Variceal Risk in PBC Score Calculator Splenomegaly

The Newcastle Risk Score:

1 / 1 + exp -(6.385 - (0.138 * (albumin in g/L)) - (0.012 * platelets x 10^9) + (2.013 * if splenomegaly was present on USS) ))

After entering serum albumin in g/L, platelet count (10^9) and alkaline phosphatase in IU click the Calculate button. Your Risk Score will then be computed and displayed in the "Predicted risk of Varices =" text box

Inputs:          
  Enter Your Serum Albumin:   (g/dL)  
  Enter Your Platelets :   (x10^9)  
  Was splenomegaly was present on USS :   No   Yes  
         
Output:          
  Predicted risk of Varices =   %  

Modeling the outcomes of PBC scoring

This web page contains a simple javascript model that can be used to explore the implications of PBC scoring over a four year time frame. The model has two outcomes of interest; the number of OGDs and the number of variceal bleeds.

The model was designed to capture the impacts of differnt treatemet strategies and as such makes a number of simplifying assumptiosn. These include assuming that all events of interest occur at one point in time during each year, for example: the development of varices, their detection by OGD and their treatement by beta-blockers or EBL. It is also assumed that no patients enter or leave the cohort during the models time frame which effects the relative proportion of varices free patients and the realtive efficacy of screenign by OGD. These simplifications do not detract from the models ability to compare the realtive effects of screening and treatment strategies, but do limit its ability to accuratley reflect absolute outcomes of the group.

Please be aware that face vaildity of the model is affected by input parameter choice combinations. For instance, if the probablity of bleeding whilst on bete-blockers or following EBL is near that of bleeding from untreated varices then any advantage of treament may be minimal or negative.
This web page requires javascript to function and dispaly correctly, and it appears that this is not enabled in this browser.

Model parameters

The model relies on reported prevalence of varices, liklehood of bleeding form varices and tolerance of beta blockers .

This is where choices are made between the data sources for use as model parameters.

  Parameter Data value and source     Parameter Data value and source
1. Prevalence of varices:   6. Varices bleed when on beta blockers:
2. Bleed from varices:   7. EBL Band ligation successful:
3. Develop varices:   8. Complication of EBL Bleeding:
4. Bleed during OGD:   9. Bleed post successful EBL:
5. Tolerate beta blockers:        
 
PBC Scoring option

Results

Here are the outputs based on a hypothetical population of 1,000 patients

     Year 1 Year 2Year 3Year 4Total
Number variceal of bleeds
 Do Nothing       
 OGD All 36 months       
 OGD All 24 months       
 OGD All 12 months       
 PBC Score 36 months       
 PBC Score 24 months       
 PBC Score 12 months       
 
OGDs
 Do Nothing         
 OGD All 36 months       
 OGD All 24 months       
 OGD All 12 months       
 PBC Score 36 months       
 PBC Score 24 months       
 PBC Score 12 months       
 

Data from paper authored by Imran Patanwala, Peter McMeekin, Ruth Walters, Julia Newton, Mark Hudson and Dave Jones Last Update: March 2013

Javascript and Webpage designed by Dr. Peter McMeekin