Need more statins or not

Need More Statins or Not

2013 ACC/AHA Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults

Don’t Give More Patients Statins
Source: NYT by John Abramson and Rita Redberg

"ON Tuesday, the American Heart Association and the American College of Cardiology issued new cholesterol guidelines that essentially declared, in one fell swoop, that millions of healthy Americans should immediately start taking pills — namely statins — for undefined health “benefits.” " ...

New ACC/AHA/NHLBI Guidance on Lifestyle for CVD Prevention
Source: Medscape by Shelley Wood

ACC-AHA cardiovascular prevention guidelines drop cholesterol treatment goals
Source: Cardiology News Digital Network  by Elizabeth Mechcatie

Risk Calculator for Cholesterol Appears Flawed
Source: NYT by Gina Kolata
"Last week, the nation’s leading heart organizations released a sweeping new set of guidelines for lowering cholesterol, along with an online calculator meant to help doctors assess risks and treatment options. But, in a major embarrassment to the health groups, the calculator appears to greatly overestimate risk, so much so that it could mistakenly suggest that millions more people are candidates for statin drugs." ...

Speaking Stata of the Stata Journal

Speaking Stata of the Stata Journal

Blog: Stata Tips of The Stata Journal and Others

This column of the Stata Journal is focusing on how to improve your fluency in Stata

1. Loops, again and again [20(4)]
2. Is a variable constant?
3. More ways for rowwise
4. Concatenating values over observations
5. Some simple devices to ease the spaghetti problem
6. The last day of the month
7. How best to generate indicator or dummy variables
8. Seven steps for vexatious string variables
9. From rounding to binning
10. Logarithmic binning and labeling
11. Tables as lists: The groups command
12. Letter values as selected quantiles
13. Shading zones on time series and other plots
14. Multiple bar charts in table form
15. Truth, falsity, indication, and negation
16. A set of utilities for managing missing values 
17. Species of origin [15(2)]
18. Design plots for graphical summary of a response given factors
19. Self and others
20. Trimming to taste
21. Creating and varying box plots: Correction
22. Matrices as look-up tables
23. Axis practice, or what goes where on a graph
24. Transforming the time axis
25. Output to order
26. Fun and fluency with functions
27. Compared with ...
28. MMXI and all that: Handling Roman numerals within Stata
29. Graphing subsets
30. The limits of sample skewness and kurtosis
31. findname: Finding variables
32. statsby [subsets, total]: The statsby strategy
33. graph twoway: Paired, parallel, or profile plots for changes, correlations, and other comparisons
34. graph box: Creating and varying box plots; Creating and varying box plots: Correction
35. qsbayesi, qsbayes: I. J. Good and quasi–Bayes smoothing of categorical frequencies
36. egen, rowsort, rowranks: Rowwise
37. distinct: Distinct observations
38. corrci, corrcii: Correlation with confidence, or Fisher's z revisited
39. labmask, seqvar: Between tables and graphs
40. Spineplots and their kin
41. egen, by: Counting groups, especially panels
42. stem, scatter, stemplot: Turning over a new leaf
43. egen, by: Identifying spells
44. count: Making it count
45. In praise of trigonometric predictors
46. cycleplot, sliceplot: Graphs for all seasons
47. Time of day
48. Smoothing in various directions
49. qplot, displot: The protean quantile plot
50. Density probability plots
51. modeldiag: Graphing model diagnostics
52. Graphing agreement and disagreement
53. Graphing categorical and compositional data
54. Graphing distributions
55. matrix list, list, tabdisp, tabcount, groups: Problems with tables, Part II
56. tabulate, table, tabstat, tabdisp, list: Problems with tables, Part I
57. for, foreach, forvalues, levels: Problems with lists
58. egen, foreach, forvalues, reshape: On structure and shape: the case of multiple responses
59. egen: On getting functions to do the work
60. On numbers and strings
61. foreach, forvalues, for: How to face lists with fortitude
62. _n, _N: How to move step by: step
63. How to repeat yourself without going mad [2001;1:(1)]

Measures of effect size in Stata 13

Measures of Effect Size in Stata 13
Soruce: the Stata Blog
"Today I want to talk about effect sizes such as Cohen’s d, Hedges’s g, Glass’s Δ, η2, and ω2. Effects sizes concern rescaling parameter estimates to make them easier to interpret, especially in terms of practical significance.

Many researchers in psychology and education advocate reporting of effect sizes, professional organizations such as the American Psychological Association (APA) and the American Educational Research Association (AERA) strongly recommend their reporting, and professional journals such as the Journal of Experimental Psychology: Applied and Educational and Psychological Measurement require that they be reported.

Anyway, today I want to show you

What effect sizes are.
How to calculate effect sizes and their confidence intervals in Stata.
How to calculate bootstrap confidence intervals for those effect sizes.
How to use Stata’s effect-size calculator.

...". Read full text here

11 Tips to Cut Your Cholesterol Fast

True

11 Tips to Cut Your Cholesterol Fast

Source: WebMD By Laura J. Martin, MD

 

1. Set a target.

2. Consider medication.

3. Get moving.

4. Avoid saturated fat.

5. Eat more fiber.

6. Go fish.

7. Drink up.

8. Drink green.

9. Eat nuts.

10. Switch spreads.

11. Don't smoke.

 

Full text here.

 

Asians Eat Weird Things

Asians Eat Weird Things 

Debating obesity the disease

Debating obesity the disease
Source: MedScape

Follow the Money?

Fat Equals Sick: Is This About the Money?
Getting Paid for Treating Obesity, Now That It's a Disease
AMA: Diagnosis by Majority

Is It an Addiction? Can Obesity Be an Addiction?

Also Weighing in on the Disease Debate Obesity: It's a Risk! It's a Symptom! It's a Disease!
Obesity as a Disease? 'It's Academic'

News Behind the Perspectives AMA Declares Obesity a Disease
Obesity Disease Classification Will Help With Treatment, Docs Say

How to turn on/off and diagnose the ANC (active noise cancellation) system of 2013 Honda Pilot

How to turn off/on the ANC (active noise cancellation) system of 2013 Honda Pilot

1. Turn the ignition switch on the I (radio) position.
2. Turn the radio on and check it operates normally, and turn the radio off.
3. Press and hold the preset buttons #1, #6, and volumn (VOL) push button at the same time.
4. when you see 'DIAG' appears on the display, you can release the buttons
5. Press the #1 preset button, you see 'ANC ON' on the display.
6. Press the #1 preset button again to turn the ANC OFF and hear a humming/booming noise coming from the speakers for about 1 minutes.
7. You can press the #1 button repeatedly for ON and OFF, but the humming/booming noise will not come up repeatedly except repeating step 1, 3-6 again.

This is a similar approach for Acura

Data visulization - a tutorial

Data visulization - a tutorial
Source: Tyler Rinker's blog

This is a nice video about how to visualize your data more effectively.

You can find the whole blog with slides and data here: The Mechanics of Data Visualization.

How to test proportionality assumption in survival analysis using SAS

How to test proportionality assumption in survival analysis using SAS

 

There are several approaches to test proportionality assumption in survival analysis:

• Graphical Approach is to plot Log–Log survival function by researched major predictor. If the two line is parallel without cross each other, the assumption is considering confirmed.

PROC LIFETEST DATA=ONE; METHOD=KM PLOTS=(S,LLS);
     TIME SURVTIME*EVENT(0);
     STRATA RISK0;
RUN;

PROC PHREG DATA=ONE;
     MODEL SURVTIME*EVENT(0)=RISK1 AGE SEX;
    BASELINE OUT=LLSOUT LOGLOGS=LOGLOGS;
RUN;
PROC GPLOT DATA=LLSOUT;
     PLOT LOGLOGS*SURVTIME=RISK1;
RUN;

• Or, you can include an interaction term of a predictor and follow up time in the model. If this interaction term is not significant, there is no violation of assumption. For example:

PROC PHREG DATA=ONE;

MODEL SURVTIME*EVENT(0)=RISK1 TIMEX;
        TIMEX=RISK1*(LOG(SURVTIME)-LOG(average followup time));
          * someone also uses simple forms: TIMEX=RISK1*LOG(SURVTIME);
          * or even: TIMEX=RISK1*SURVTIME; (not good enough); 
     PROPORTIONALITY_TEST: TIMEX;
RUN;

• Possibly, the easiest and powerful approach after SAS 9.2 is to use ASSESS statement, for example:

ODS GRAPHICS ON;

PROC PHREG DATA=ONE;

MODEL SURVTIME*EVENT(0)=RISK1;
     ASSESS PH/resample;
RUN;

ODS GRAPHICS OFF;

 

What’s New of PROC PHREG of SAS 9.2 (pdf)

 

American Medical Association declared obesity a “disease.”

American Medical Association declared obesity a “disease.”

"RESOLVED, That our American Medical Association recognize obesity as a disease state with 24 multiple pathophysiological aspects requiring a range of interventions to advance obesity 25 treatment and prevention. (New HOD Policy - Resolution 420)" - 06/16/2013

• New York Times (2013). A.M.A. Recognizes Obesity as a Disease
• Time.com (2013). The Best Cure for Obesity? Personal Responsibility.
• Boston.com (2013). Has obesity been mislabeled as a disease? Why doctors don’t mind.
• Bays (2013). Obesity, adiposity, and dyslipidemia: A consensus statement from the National Lipid Association.

Risk prediction and model comparison

 Risk prediction and model comparison
 
The potential approaches of prediction and comparison:

• Relative risk/hazard ratio/odds ratio, P-value.
• Sensitivity/specificity
• Area under ROC (receiver operating characteristics) curve (AUC)/Harrell's c statistic
• Somers' D (Kendall's Tau): the mathmatic conversion of Somers'D and c statistic are: [c statistic = D/2+0.5] or [D = (c - 0.5) x 2]. (SAS: PROC FREQ)
• NRI (net reclassification improvement)
• IDI (integrated discrimination improvement).
• K-fold cross-validation - Wikipedia
• ...

References:

• Pencina (2011). Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers (Comment).
• Pencina (2013). Understanding increments in model performance metrics.
• Pencina (2008). Evaluating the added predictive ability of a new marker: From area under the ROC curve to reclassification and beyond.
• Measuring the Accuracy of Prediction Models - An International Symposium (2008): IDI, NRI and different c statistics (pdf)
• Hilden (2013). A note on the evaluation of novel biomarkers: do not rely on integrated discrimination improvement and net reclassification index.
• Liu (2012). Evaluating a New Risk Marker’s Predictive Contribution in Survival
• Mühlenbruch (2013). Assessing improvement in disease prediction using net reclassification improvement: impact of risk cut-offs and number of risk categories (Commentary).
• Cassell (2007). Don't be loopy: re-Sampling and simulation the SAS way (pdf).
• Kerr (2011). Evaluating the incremental value of new biomarkers with integrated discrimination improvement.

Software:

• Pepe Lab has some homemade software for biomarker/risk evaluation using Stata, SAS, R, SPSS, and even FORTRAN.
• Stata: to install 'Risk Prediction Package' (predcurve and incrisk), in a Stata session type: ".net from http://labs.fhcrc.org/pepe/stata/" and follow the instruction.To update the risk_prediction package at a later time, in Stata type: ".adoupdate risk_prediction, update".
• UCR posted SAS, Stata, and R codes for NRI & IDI on the website.
• Stata: How can I get a Somers' D after logistic regression in Stata?

 

rule of halves of diabetes

Rule of halves of diabetes
Source: DAWNStudy Diabetic Attitudes Wishes and Needs

Multicollinearity Issue

Multicollinearity Issue

Multicollinearity happens when two or more predictor/independent variables/regressors are highly correlated. I have been discussed about this issue many times by colleagues and journal reviewers. Paul Allison has a blog of some rules of thumb: When Can You Safely Ignore Multicollinearity? Wikipedia also has a article about this issue. It's true this is issue theoretically, but based on my experience in public health of chronic diseases, if the selection of predictors based on the logic/knowledge behind the model but not dump everything in one model, this should not be an issue.

Do We Really Need Zero-Inflated Models?

Do We Really Need Zero-Inflated Models?
Source: Statistical Horizon blog by Paul Allison
 
"... Of course, there are certainly situations where a zero-inflated model makes sense from the point of view of theory or common sense. For example, if the dependent variable is number of children ever born to a sample of 50-year-old women, it is reasonable to suppose that some women are biologically sterile. For these women, no variation on the predictor variables (whatever they might be) could change the expected number of children.

So next time you’re thinking about fitting a zero-inflated regression model, first consider whether a conventional negative binomial model might be good enough. Having a lot of zeros doesn’t necessarily mean that you need a zero-inflated model."

Read full text here
 
This question has haunted me for a while, thank Dr. Allison answered this question in such a layman-kind way. I like his book "Survival Analysis Using SAS: A Practical Guide" much; I don't have his book "Logistic Regression Using SAS: Theory and Application". Hope this logistic regression related book is in the same style.

More Blog on Statistical Horizon Blog

My favorite quotes of Magic School Bus

My favorite quotes of Magic School Bus (1994-1998)

It was my happy time watching Magic School Bus with my kids around 2000. The series is informative and educative, and the last but the best is a mindset of exploring and innovation. There are several quotes I don't want to forget.

• Ms. Frizzle, "Take chances, make mistakes, get messy."
• Ms. Frizzle, "Never say 'never.'"

These are also on the top of list of Wikiqote

Diease or not disease, a diabetic question

Diease or not disease, a diabetic question

Last a few days, I read two diabetic related articles, which both are not my cup of tea; but I am really appreciating their thoughts and contrast.

• The Lie That's Killing Us: Pre-Diabetes (Riva Greenberg) on Huffpost defines the Type 2 diabetes as a diesase into four stages, which includes pre-diabetes as the stage 1.
• No Such Thing As Type 2 Diabetes? (Alice G. Walton) on Forbes is on the another side. The original article is published on Lancet (Is type 2 diabetes a category error? by Edwin AM Gale)

Both of authors are creative and thoughtful. Riva Greenberg is much more on the side of novelty/grassroot side (without enough evidence, to me) and Edwin Gale is much more on the side of classic/ivy tower (with typical theory/evidence, to me). I enjoy reading both articles.

Dance with my father - Luther Vandross

Dance With My Father, In Memory of My Fathers

 

"Dance With My Father"

Back when I was a child, before life removed all the innocence
 My father would lift me high and dance with my mother and me and then
 Spin me around ‘til I fell asleep
 Then up the stairs he would carry me
 And I knew for sure I was loved
 If I could get another chance, another walk, another dance with him
 I’d play a song that would never, ever end
 How I’d love, love, love
 To dance with my father again
 When I and my mother would disagree
 To get my way, I would run from her to him
 He’d make me laugh just to comfort me
 Then finally make me do just what my mama said
 Later that night when I was asleep
 He left a dollar under my sheet
 Never dreamed that he would be gone from me
 If I could steal one final glance, one final step, one final dance with him
 I’d play a song that would never, ever end
‘Cause I’d love, love, love
 To dance with my father again
 Sometimes I’d listen outside her door
 And I’d hear how my mother cried for him
 I pray for her even more than me
 I pray for her even more than me
 I know I’m praying for much too much
 But could you send back the only man she loved
 I know you don’t do it usually
 But dear Lord she’s dying
 To dance with my father again
 Every night I fall asleep and this is all I ever dream

Impact Factor Distortions

Impact Factor Distortions by Bruce Alberts
Source: Science

"This Editorial coincides with the release of the San Francisco declaration on research Assessment (DORA), the outcome of a gathering of concerned scientists at the December 2012 meeting of the American Society for Cell Biology. To correct distortions in the evaluation of scientific research, DORA aims to stop the use of the "journal impact factor" in judging an individual scientist's work. The Declaration states that the impact factor must not be used as "a surrogate measure of the quality of individual research articles, to assess an individual scientist's contributions, or in hiring, promotion, or funding decisions." DORA also provides a list of specific actions, targeted at improving the way scientific publications are assessed, to be taken by funding agencies, institutions, publishers, researchers, and the organizations that supply metrics. These recommendations have thus far been endorsed by more than 150 leading scientists and 75 scientific organizations, including the American Association for the Advancement of Science (the publisher of Science). Here are some reasons why: "

How to make compost

How to make compost

• A very helpful Youtube video from howdini.com. Scott's teaching how to make compost using cheap chicken wire.

• Garden Organic gives the quite detailed information about How to Make Compost.

slope index of inequality (SII) and the relative index of inequality (RII).

Slope Index of Inequality (SII) and Relative Index of Inequality (RII)

The slope index of inequality (SII) and the relative index of inequality (RII) are measures of health inequality (Mackenbach, 1997). The SII and RII can be calculated through regression analysis on an indicator of the cumulative relative position of each group with respect to a socioeconomic variable account for both the socioeconomic status (SES) of the groups and the size of the population.


The groups of SES are ordered by decreasing socioeconomic status from higher to lower. Each socioeconomic category is given a score called ridit score, which reflects the average cumulative frequency of the group, a midpoint of the range in the cumulative distribution introduced by Bross (1958). For example, if the highest educated women include 20% of the population, the range of women in this category is from 0.00 to 0.20 and assigned a ridit score of 0.10 (= 0.20/2 = [0 + 0.2]/2), and if the next level of educated women include 50% of the population from .20 to 0.70, the corresponding ridit score is 0.45 (= 0.20 + 0.5/2 = [0.20 + 0.70]/2) and so forth. 
 

SII (rate difference) is the slope of the regression line (b) estimated by the weighted least squares method and represents the change in measured outcome/event when the position of the SES changes by one unit.

RII (rate ratios) can be which can be estimated in two ways: one way is to divide the SII by the mean level of population health or by the frequency of the health problem in the population; the other way is to divide the predicted value of the regression at the highest point (range=1) by the predicted value of the regression at the lowest point (range=0). The second method for the RII is calculated by log-linear—or logistic—regression after the logarithmic—or logit—transformation of the dependent variable. One way to facilitate the interpretation of this measure of the second method may be to express the RII as a percentage by subtracting 1 from it and multiplying the result by 100 (see Table 1 below) (Regidor, 2004).

Other References:

• Software: Health Disparity Measures in HD*Calc - NCI
• NCI (2005). Methods for Measuring Cancer Disparities
• Keppel & CDC (2005). Methodological issues in measuring health disparities
• Cheng (2008). Methods and Software for Estimating Health Disparities: The Case of Children's Oral Health
• Hayes (2002). How to calculate confidence limits for the relative index of inequality?
• Schwarz (2006). The Contributions of Diseases to Increasing Educational Mortality Differential in Austria
• OpenMichigan. Measuring Health Disparities
• BMJ (2017). Say what you mean, mean what you say: inequality and inequity
• Klein (2010): Defining and measuring disparities, inequities, and inequalities in the Healthy People initiative
• Arcaya (2015). Inequalities in health: definitions, concepts, and theories

Calculation of SII and RII using aggregated data (modified from the paper of Regidor, 2004)

SII of the year 1990 = 104.4, the beta coefficient (b) of a linear regression of Ridit90 and Mort90.

RII of the year 1990 can be calculated as SII/(average mortality rate) = 104/134 = 0.7761 (Measures of Health Disparity - NCI (pdf))

Calculation of SII and RII using individual data - Stata 12 (modified from the paper of Ernstsen, 2012)

ID	Men	Survey	Smoking	Ridit	Age
1	1	1	0	0.34	53
2	0	1	1	0.55	48
3	0	2	0	0.34	55
4	1	3	1	0.34	40

Estimating Relative Index of Inequality (RII) for smoking in men at the first survey:

.glm i.smoking ridit age if survey==1 & men==1, fam(bin) link(log) nolog eform

Estimating Slope Index of Inequality (SII) for smoking in men at the first survey:

.glm i.smoking ridit age if survey==1 & men==1, fam(bin) link(identity)

Estimating trend in RII and SII over time for men:

.glm i.smoking c.ridit##survey age if men==1, fam(bin) link(log) nolog eform
.glm i.smoking c.ridit##survey age if men==1, fam(bin) link(identity)

Estimating gender differences in RII and SII at each survey:

.glm i.smoking c.ridit##men age if survey==1, fam(bin) link(log) nolog
.glm i.smoking c.ridit##men age if survey==1, fam(bin) link(identity)

Estimating if RII and SII changed differently over time in men and women:

.glm i.smoking c.ridit##men##survey age, fam(bin) link (log) nolog
.glm i.smoking c.ridit##men##survey age, fam(bin) link (identity) nolog