Wednesday, October 23, 2013

Reading a paper and Doing Literature Review

Most of the research works are based on papers, both reading papers and writing papers. The efficient way of reading a paper is always personalized by tons of experience. There is one paper about how to read a paper [1], which gives 3 passes for a understanding a paper. As mentioned in [1], the three passes are,
  • The first pass is a quick scan to get a bird's-eye view of the paper. You can also decide whether you need to do any more passes. 
  • In the second pass, read the paper with greater care, but ignore details such as proofs. In the second pass, read the paper with greater care, but ignore details such as proofs. Focus on understanding the figures, make a note of the import references
  • The key to the third pass is to attempt to virtually re-implement the paper: that is, making the same assumptions as the authors, re-create the work. This pass needs great attention to detail.
Doing literature review can be divided into 3 steps as well.
  • First, use an academic search engine and some well-chosen keywords to fi nd three to fi ve recent highly-cited papers in the area. Do one pass to each of the paper to understand the summary of the area. If you can find a good review, done for the first step.
  • Second step, fi nd shared citations and repeated author names in the bibliography. These are the key papers and researchers in that area. Download the key papers and set them aside. Then go to the websites of the key researchers and see where they've published recently.
  • The third step is to go to the website for these top conferences and look through their recent proceedings. Identify recent high-quality related work quickly. These papers, along with the ones you set aside earlier, constitute the first version of your survey. Make two passes through these papers. 

The paper also mentioned some related references:
A website that covers the entire spectrum of research skills [2].
If you are reading a paper to do a review, you should also read [3]. 
If you're planning to write a technical paper, you should refer both to [4] and [5]. 

Also, a downloadable `review matrix' that simplifies paper reviewing using the three-pass approach for papers in experimental psychology [6]. 







Friday, April 26, 2013

Prepare for the coming of Era of BIG DATA


“Big Data” is a popular topic that has been gaining attention from the high-performance computing niche of the information technology market. Big Data is an enormous amount of data from which it is extremely difficult to manage and glean information. Big Data provides both challenges and opportunities for quantitative analysts to develop improved predictive and descriptive models [1].


SAS launched SAS High-Performance Data Mining in December 2011 to enable you to analyze more data faster than ever before possible. Based on the SAS High-Performance Analytics model for distributed in-memory processing, SAS High-Performance Data Mining is delivered SAS software that uses Teradata or EMC Greenplum hardware. A subset of SAS High-Performance Analytics, SAS High-Performance Data Mining has truly revolutionized the model building and model-scoring processes. The massively parallel in-memory algorithms enable organizations to derive highly accurate and timely data mining models in minutes, not hours or days, to make better-informed business decisions [1].

SAS has developed many HP procedures, such as,


  • HPDMDB Summarize data
  • HPDS2 Parallel execution of DS2
  • HPFOREST Random forest
  • HPLOGISTIC Logistic regression
  • HPNEURAL Neural network modeling
  • HPNLIN Nonlinear regression
  • HPREDUCE Unsupervised variable selection
  • HPREG Regression
  • HPBIN Variable Binning
  • HPSAMPLE Sampling and data partitioning
  • HPIMPUTE Imputation
  • HPSEVERITY Severity models
  • HPCOUNTREG Regression of count variables
  • HPSUMMARY Summarize data
  • HPLMIXED Mixed linear models
  • HPATEST Test operational status of system



Hopefully, these procedures will be available in the late 2013, and will be available to academic field too.


Reference:
1. A New Age of Data Mining in the High-Performance World
http://support.sas.com/resources/papers/proceedings12/137-2012.pdf

2. SAS workshop in UCLA.

Wednesday, April 24, 2013

U.S. and California Pursue Coordinated Care for Dual Eligibles: What's Next for Cal-MediConnect?


You may always wonder what is the future for Dual Eligibles under the Affordable Care Act. Is it too good to be true or eventually doing harm to the benefits of Medicare benefits? We may seek answer from the series of lectures/seminars from the UCLA health forum on April 24, 2013.  

The UCLA Jonathan and Karin Fielding School of Public Health is pleased to host The Health Forum at UCLA (FSPH), a series of regularly scheduled free public programs featuring health leaders discussing critical issues in public health.

 
"U.S. and California Pursue Coordinated Care for Dual Eligibles:
What's Next for Cal-MediConnect?"
California has over 450,000 people eligible for both Medicare and Medi-Cal coverage, also known as "dual eligibles." Dual eligible beneficiaries face difficulties navigating a currently fragmented medical system to find the services they need to treat serious and chronic medical conditions. Recently, California received Federal approval for a three-year demonstration program, Cal-MediConnect, to improve coordinated health care delivery and quality of services to dual eligibles which will be integrated in eight counties, including Los Angeles County. In this health forum we will discuss what changes in health service delivery dual eligibles should expect to see in managed care plans and network delivery systems.
When: This Wednesday, April 24th, 6:00-8:00pm
            6:00-6:30pm: Coffee and Reception
            6:30-7:15pm: Panel Presentations
            7:15-8:00pm: Question and Answer Session
 
Where: Neuroscience Research Building Auditorium, UCLA 
             635 Charles E. Young Drive South, Los Angeles
 
Panelists include: 
Richard Chambers
President
Molina Healthcare of California
Hector Flores, MD
Co-Director
WWMC Family Medicine Residency Program
Jane Ogle
Deputy Director of Health Care Delivery Systems 
California Department of Health Care Services
  
Moderated by: 

Frank Matricardi, DrPH, MPH
Principal
Executive Healthcare Consulting
To RSVP, please click HERE.

Thursday, February 14, 2013

Seminar Series: The Future of Private Medical Practice


 
"The Future of Private Medical Practice"
The number of private medical practices is decreasing as more and more physicians choose to leave their practices to join larger health care systems. Playing a role in this shift are low government reimbursements, costly technological improvements and health care reform initiatives designed to encourage coordinated care. In this health forum we will discuss what's next for private medical practices and how it will impact the doctor-patient relationship.
When: Wednesday, February 27th, 6:00-8:00pm
            6:00-6:30pm: Coffee and Reception
            6:30-7:15pm: Panel Presentations
            7:15-8:00pm: Question and Answer Session
 
Where: Neuroscience Research Building Auditorium, UCLA 
             635 Charles E. Young Drive South, Los Angeles
 
Panelists include: 
Patricia Kapur
Executive Vice President
UCLA Health System
Allan Lew
Attending Physician, Division of Cardiology
Cedars-Sinai Medical Center
Iris Weil
CEO
Longdale Medical Management
  
Moderated by: 

Donald Crane
CEO
California Association of Physician Groups
To RSVP, please click HERE.

Tuesday, February 12, 2013

ESA WINTER SPEAKER SERIES


Dr. Lisa Smith
Epidemiologist at LACPHD & adjunct faculty at UCLA

Dr. Mark Danese
President of Outcomes Insights, Inc.

Dr. Cathy Critchlow
Executive Director and Head for the Center of Observational Research at Amgen

Andrew Dagis
City of Hope National Medical Center & President of Southern California Chapter, American Statistical Association

Tuesday February 19  6:30 - 8:30 PM
Kerckhoff Grand Salon


There will be drinks and appetizers!
Please come dressed business casual

Wednesday, January 23, 2013

Free Seminar on 1/30 6-8pm in UCLA.




"How Will the Mental Health System Change 
Under the Affordable Care Act?"
More than 40 million children and adults in the United States live with mental illness. The Affordable Care Act will require coverage for individuals with these illnesses and extend benefits to at least 10 million previously uninsured Americans suffering from mental disorders. How will the health system, which is already straining to provide adequate treatment, provide quality mental health services to this greatly expanded pool of patients? In this health forum we will discuss what is being done around the country and across California to adapt our mental health system.
When: Wednesday, January 30th, 6:00-8:00pm
            6:00-6:30pm: Coffee and Reception
            6:30-7:15pm: Panel Presentations
            7:15-8:00pm: Question and Answer Session
 
Where: Neuroscience Research Building Auditorium, UCLA 
             635 Charles E. Young Drive South, Los Angeles
 
Panelists include: 
Jeffrey Ring
Director of Behavioral Sciences and Cultural Medicine
WMMC Family Medicine Residency Program
Molly Brassil *
Associate Director
California Mental Health Directors Association
 
* Due to a scheduling conflict, Molly Brassil will take the place of Patricia Ryan, who was originally scheduled to appear.
Kenneth Wells
Director of Health Services Research Center
Jane and Terry Semel Institute for Neuroscience and Human Behavior
  
Moderated by: 

Andrew Leuchter
Professor of Psychiatry and Biobehavioral Sciences
David Geffen School of Medicine at UCLA
To RSVP, please click HERE.
To forward this invitation to a friend or colleague, please click HERE.

Monday, December 17, 2012

EHR ranking according to KLAS


Below are the news according to Medscape.
Hundreds of companies sell electronic health record (EHR) systems, but just 2 of them dominated the annual year-end rankings published by research firm KLAS, suggesting the tremors of a long-awaited industry shakeout.
The KLAS findings partly reflect the accelerating shift of physicians from independent practice to hospital employment, experts told Medscape Medical News. KLAS named Epic the number 1 overall software vendor and the number 1 overall physician practice vendor.
Epic received the company's Best of KLAS Award for EHRs for medical practices with from 11 to 75 physicians and those with more than 75 physicians. It also took top honors in the category of inpatient EHRs. Meanwhile, Athenahealth had the number 1 EHR for practices with 1 to 10 physicians.
CureMD ranks as the number-one vendor of electronic health record (EHR) programs accessed online solely with a Web browser, according to a new study by research firm KLAS. Close behind in second place is Practice Fusion, which has the only free system on the market. Athenahealth and Medical Informatics Engineering tied for third place in the comparison of these inexpensive, easily implemented systems and their vendors.
The remaining 6 EHR vendors in the KLAS survey are:
  • MedPlus/Quest Diagnostics
  • Sevocity
  • OptumInsight (formerly Ingenix)
  • AdvancedMD
  • Waiting Room Solutions
  • Bizmatics







Thursday, October 11, 2012

Healthcare Reform: The Employer Perspective

The UCLA Jonathan and Karin Fielding School of Public Health is pleased to host The Health Forum at UCLA (FSPH), a series of regularly scheduled free public programs featuring health leaders discussing critical issues in public health.
"Healthcare Reform: The Employer Perspective"
The Affordable Care Act has the potential to affect the health insurance choices and responsibilities of employers as healthcare coverage is expanded to include roughly 32 of the 54 million uninsured Americans. The implications of the reforms will vary depending on the employer's size. Some small businesses, which historically have faced multiple barriers to offering affordable health insurance to their employees, will now receive tax credits to assist in purchasing coverage, while larger employers will face new requirements to contribute to the cost of their employees' health insurance coverage. In this health forum we will discuss the impact of the Affordable Care Act on both small and large businesses.
When: Wednesday, October 24th, 6:00-8:00pm
            6:00-6:30pm: Coffee and Reception
            6:30-7:15pm: Panel Presentations
            7:15-8:00pm: Question and Answer Session
 
Where: Neuroscience Research Building Auditorium, UCLA 
             635 Charles E. Young Drive South, Los Angeles
 
Panelists include:
 
John Arensmeyer
Chief Executive Officer
Small Business Majority
Jim Scilacci
EVP, CFO & Treasurer
Edison International
Lucien Wulsin
Director
Insure the Uninsured Project
 
 
Moderated by:
 
Amir Hemmat
President & CEO
        SABEResPODER         
To RSVP, please click HERE.
To forward this invitation to a friend or colleague, please click HERE.

Monday, October 8, 2012

ROC curve in BioStatistics

ROC (Receiver Operating Characteristic) Curve is used to evaluate the accuracy of a test to discriminate disease cases from normal cases, i.e. the performance of diagnosis. Also, it can be used to compare two or more tests [1]. ROC curve is to plot the Sensitivity (true positive rate) against the 1-Specificity (false positive rate). Each point on the curve represents a sensitivity/specificity pair and results in a particular decision threshold. The Area Under the Curve (AUC) is a measure to describe how well the test can distinguish the disease and normal.

When you consider the test in two populations, one with disease and the other normal, you can rarely observe a perfect separation between the two groups. Assume the distribution of the test results is normal, these two test results will overlap as [1],


For every possible cut-off point or criterion value you select to discriminate between the two populations, there will be some cases with the disease correctly classified as positive (TP = True Positive fraction), but some cases with the disease will be classified negative (FN = False Negative fraction). On the other hand, some cases without the disease will be correctly classified as negative (TN = True Negative fraction), but some cases without the disease will be classified as positive (FP = False Positive fraction).


There are some statistics based on TN, TP, FN, FP as below [1],

 

The different fractions (TP, FP, TN, FN) are represented in the following table.
 Disease      
TestPresentn Absentn Total
PositiveTrue Positive (TP)a False Positive (FP)c a + c
NegativeFalse Negative (FN)b True Negative (TN)d b + d
Total a + b  c + d  

  • Sensitivity: probability that a test result will be positive when the disease is present (true positive rate, expressed as a percentage).
    = a / (a+b)
  • Specificity: probability that a test result will be negative when the disease is not present (true negative rate, expressed as a percentage).
    = d / (c+d)
  • Positive likelihood ratio: ratio between the probability of a positive test result given the presence of the disease and the probability of a positive test result given the absence of the disease, i.e.
    = True positive rate / False positive rate = Sensitivity / (1-Specificity)
  • Negative likelihood ratio: ratio between the probability of a negative test result given the presence of the disease and the probability of a negative test result given the absence of the disease, i.e.
    = False negative rate / True negative rate = (1-Sensitivity) / Specificity
  • Positive predictive value: probability that the disease is present when the test is positive (expressed as a percentage).
    = a / (a+c)
  • Negative predictive value: probability that the disease is not present when the test is negative (expressed as a percentage).
    = d  / (b+d)




When you select a higher criterion value, the false positive fraction will decrease with increased specificity but on the other hand the true positive fraction and sensitivity will decrease; Meanwhile, when you select a lower criterion value, then the true positive fraction and sensitivity will increase. On the other hand the false positive fraction will also increase, and therefore the true negative fraction and specificity will decrease.

The ROC curve

In a Receiver Operating Characteristic (ROC) curve the true positive rate (Sensitivity) is plotted in function of the false positive rate (100-Specificity) for different cut-off points. Each point on the ROC curve represents a sensitivity/specificity pair corresponding to a particular decision threshold. A test with perfect discrimination (no overlap in the two distributions) has a ROC curve that passes through the upper left corner (100% sensitivity, 100% specificity). Therefore the closer the ROC curve is to the upper left corner, the higher the overall accuracy of the test [2].







An ROC curve demonstrates several things [3]:
  1. It shows the trade-off between sensitivity and specificity (any increase in sensitivity will be accompanied by a decrease in specificity).
  2. The closer the curve follows the left-hand border and then the top border of the ROC space, the more accurate the test.
  3. The closer the curve comes to the 45-degree diagonal of the ROC space, the less accurate the test.
  4. The slope of the tangent line at a cut point gives the likelihood ratio (LR) for that value of the test. 
  5. The area under the curve is a measure of test accuracy. 



ROC curve in SAS
The ROC plot in SAS program is embedded in proc logistic procedure as following [3],


 ods graphics on; 
 ods html; 
 proc logistic data=mydata plots(only)=(roc); 
    model Y=marker; 
 run; 
 ods html close; 
 ods graphics off; 




ROC curve in R

You can use the function "roccurve" to plot ROC curve in R. Also, you can implement this function in SAS by SUBMIT.

You can use R in SAS by the code SUBMIT [3]. The SUBMIT statement takes options, and it is the option R on the SUBMIT statement which indicates that code is to be directed to R.  Thus,  a clearer indication of the usage of submit block code would be:


 submit; 
 <SAS data step or procedure code to be executed> 
 endsubmit; 
 submit / R;  <R code to be executed> 
 endsubmit; 

Now, it is likely that R code is being submitted from a SAS session because the user is performing data manipulation (and perhaps some analyses) in SAS, but R has some functions for data analysis which are not available in SAS.  Thus, data exist in the SAS session, and must be passed to the R session.  The submit block directs code to an R session, but the user also needs to exchange data between SAS and R – often in both directions.






References:
1. http://www.medcalc.org/manual/roc-curves.php
2. Zweig MH, Campbell G (1993) Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clinical Chemistry 39:561-577. [Abstract]
3. ROC curve in SAS: http://labs.fhcrc.org/pepe/dabs/ROC_Curve_Plotting_in_SAS_9_2.pdf

Wednesday, September 12, 2012

UCLA Health Disparities Symposium (Free)

Mark your calendar for this free event

The 13th Annual Health Disparities Symposium 



Stress, Health, and Health Equity




Thursday, October 11, 2012 • 8 a.m. – 4 p.m. 
This is a FREE Symposium 

Center at Cathedral Plaza • 

555 West Temple Street 
Los Angeles, CA 90012


To Register: 


Registration is required
Breakfast, lunch and parking will be provided




Issues to be addressed: 
• What is stress? 
• Sources of stress: socio-economic, environmental, workplace, interpersonal 
• How does stress affect health and health behaviors?  
• Stress and health across the lifespan

• What is the role of stress in health disparities? 

• Coping with stress




Friday, September 7, 2012

The Health Forum for DUAL ELIGIBLES

We should re-post this Free admission 2-hr Forum at UCLA on 9/19/2012.


"Covering and Caring for Dual Eligibles"

There are nearly 9 million so called "dual eligibles" -- low income seniors and younger people with disabilities -- who qualify for both Medicaid and Medicare benefits. They are among the country’s sickest and poorest populations and frequently encounter great difficulty navigating the complex and uncoordinated systems of the two beneficiary programs. The Affordable Care Act contains provisions to improve the quality of care delivered to dual eligibles while better controlling costs. The panel will discuss the implementation and challenges involved in managing care for dual eligibles.

When: Wednesday, September 19th, 6:30-8:30pm
6:30-7:00pm: Coffee and Reception
7:00-7:45pm: Panel Presentations
7:45-8:30pm: Question and Answer Session
Where: Neuroscience Research Building Auditorium, UCLA, 635 Charles E. Young Drive South, Los Angeles
Panelists include:
Maya Altman, MPP
Chief Executive Officer
Health Plan of San Mateo
Sandra Shewry, MSW/MPH
Director, State Health Policy
California Healthcare Foundation
Kathleen Wilber, PhD
Professor, Gerontology
University of Southern California
Moderated by:
Bruce Chernof, MD
Chief Executive Officer
                        The SCAN Foundation                       
To RSVP, please click HERE.

Thursday, September 6, 2012

HIPPA and PHI

Under the US Health Insurance Portability and Accountability Act (HIPAA),  protected health information (PHI) that is linked based on the following list of 18 identifiers must be removed,
(Source: http://www.ucdmc.ucdavis.edu/compliance/guidance/privacy/deident.html),


  1. Names;
  2. All geographic subdivisions smaller than a State, including street address, city, county, precinct, zip code, and their equivalent geocodes, except for the initial three digits of a zip code if, according to the current publicly available data from the Bureau of the Census:
    1. The geographic unit formed by combining all zip codes with the same three initial digits contains more than 20,000 people; and
    2. The initial three digits of a zip code for all such geographic units containing 20,000 or fewer people is changed to 000.
    3. Currently, 036, 059, 063, 102, 203, 556, 592, 790, 821, 823, 830, 831, 878, 879, 884, 890, and 893 are all recorded as "000".
  3. All elements of dates (except year) for dates directly related to an individual, including birth date, admission date, discharge date, date of death; and all ages over 89 and all elements of dates (including year) indicative of such age, except that such ages and elements may be aggregated into a single category of age 90 or older;
  4. Telephone numbers;
  5. Fax numbers;
  6. Electronic mail addresses;
  7. Social security numbers;
  8. Medical record numbers;
  9. Health plan beneficiary numbers;
  10. Account numbers;
  11. Certificate/license numbers;
  12. Vehicle identifiers and serial numbers, including license plate numbers;
  13. Device identifiers and serial numbers;
  14. Web Universal Resource Locators (URLs);
  15. Internet Protocol (IP) address numbers;
  16. Biometric identifiers, including finger and voice prints;
  17. Full face photographic images and any comparable images; and
  18. Any other unique identifying number, characteristic, or code, except as permitted by the re-identification rules, below; and



Summary of UCLA Health System's HIPAA Policies

1. Protection of Health Information

UCLA Health System Workforce members may not disclose, share or otherwise use any individually identifiable health information except for treatment, payment, and health care operations (referred to as "TPO") unless expressly authorized by the patient or as otherwise permitted by law.  Patients also have the right to request that UCLA restrict how their PHI is used or disclosed.

2. Classification of PHI Information

All information contained in patient medical and billing records is confidential regardless of format.  These confidentiality protections extend not only to the patient's medical record, but also to information from the record.  In addition, special laws govern the disclosure of mental health, substance abuse, and HIV test result information. 

3. Notice of Privacy Practices

The Privacy Rule requires UCLA Health System to give each patient detailed information about UCLA Health System's privacy practices, in the form of the University's "Notice of Privacy Practices" (see "Other Forms and Documents").  All uses and disclosures of PHI by UCLA Health System and its workforce members must be consistent with the Notice of Privacy Practices.

4. Authorization to Use PHI

The Privacy Rule requires providers to obtain a written authorization from an individual before using or disclosing a patient's PHI for purposes other than for TPO, unless otherwise authorized by law.

5. Patient Access to PHI

The Privacy Rule gives an individual (or that person's personal representative) the right of access to inspect and obtain a copy of the individual's own PHI.  Providers may deny an individual access to his or her information under certain circumstances only if specified procedures are followed.
All requests for information from medical records should be referred to or coordinated with UCLA Health Information Management Services.

6. UCLA Health System Employee (Workforce) Responsibilities to Maintain Confidentiality of PHI

All members of the UCLA Health System workforce are responsible for maintaining the security and confidentiality of PHI on behalf of UCLA Health System patients. 
  • Minimum necessary: When using or disclosing PHI, a provider must make reasonable efforts to limit PHI to the minimum necessary to accomplish the intended use, disclosure, or request.
  • Employee access: All members of the UCLA Health System workforce should only read and use PHI as necessary for their job functions.

7. Release of PHI to Third Parties

In light of the specific accounting and disclosure requirements imposed by HIPAA, all copying of medical records for release to third parties or agencies must be completed by, or coordinated with, UCLA Health Information Management Services. 

8. Privacy Requirements Relating to Research

Research is not considered to be a part of TPO under the Privacy Rule, except for certain studies related to health care operations, such as research that is also considered quality assurance and utilization management activities.  Consequently, the use or disclosure of PHI for research purposes generally requires either: (1) a written authorization from the individual whose information is collected or (2) a waiver of authorization from UCLA's IRB. The IRB is responsible for reviewing and approving the authorization form that is used for research.
The Privacy Rule permits the use and disclosure of a limited data set of information for research purposes, without patient authorization, provided certain requirements are met, including entering into a Data Use Agreement with the recipient of the information.
Health Information that does not identify an individual ("de-identified information") is generally not considered PHI and may be disclosed without the patient's authorization.  In order to de-identify PHI, UCLA Health System must remove all 18 of the HIPAA identifiers specified in the HIPAA Privacy Rule.

9. Disclosures to Business Associates

The Privacy Rule requires UCLA Health System to enter into a confidentiality agreement with certain third parties when UCLA Health System shares PHI with the third party (e.g., non-health care providers) for TPO purposes.  This is called a business associate agreement ("BAA"). A business associate relationship exists when an individual or entity, acting on behalf of UCLA Health System, assists in the performance of a function or activity involving the use or disclosure of UCLA Health System's PHI.  The UCLA Purchasing Departmentsare responsible for completing the University's HIPAA-compliant business associate agreement with outside vendors that provide goods or services to UCLA Health System. The UCLA Health System's form BAA can be found on the UCLA Health System Office of Compliance Services website. 

10. Marketing and Fundraising

In general, PHI may not be disclosed for marketing purposes without the patient's authorization. PHI includes demographic information, without any accompanying diagnosis or treatment information.  An authorization must be obtained from the patient even to use the patient's address or phone number for marketing.
In addition, all fundraising materials sent to an individual must describe how the individual can opt out of receiving further fundraising communications.

11. Media Inquiries

Both California law and the Privacy Rule restrict the amount of information that may be provided to the media without the patient's authorization.  In general, UCLA Health System may release the condition and location of an inpatient, outpatient, or emergency patient, but only if the inquiry specifically contains the patient's name, and only if the patient has not requested that the information is withheld from disclosure.  No information can be given if a request does not include the patient's name or if the patient has requested that information be withheld.
A patient's condition may only be described in general terms that does not communicate specific medical information about the individual.  For example, the following general terms are acceptable: "undetermined," "good," "fair," "serious," "critical," or "deceased."

12. Safeguards to Protect PHI

Reasonable safeguards (physical, electronic and administrative) are to be used at all times to ensure that confidential information is not disclosed to individuals who are not authorized to receive the information and to minimize incidental disclosures of PHI.  Examples of safeguards (such as locking medical and billing records at the end of the day, not sharing passwords, etc.) can be found on the UCLA Health System Office of Compliance Services website, and in UCLA Health System policies, such as Policy HS 9401.

13. UCLA Health System Workforce Training and Education

The Privacy Rule requires that providers train their "workforce" on privacy policies and procedures at a level appropriate for the workforce members to carry out their roles and responsibilities.  All members of the UCLA Health System workforce will be provided with essential instruction regarding Privacy Rule requirements and additional training specific to their job responsibilities.

14. Unauthorized Release and Disclosure

The unauthorized release of PHI is a violation of law, with potential civil and/or criminal penalties and fines.  In addition, workforce members who are found to have violated the law and/or UCLA Health System policies may be subject to disciplinary action, up to and including termination.  Workforce members should immediately report any unauthorized release or disclosure of PHI to the Privacy and Information Security Offices and their supervisor.








Best Consulting Skills

The September issue of ASA news post the article <Consulting Best Practices>. We all believe that there are more important necessary characteristics/skills of a good statistical consultant beyond the technical expertise. The most important characteristics, suggested by practitioners and clients, are the interpersonal skills, the ability to and interest in listening carefully and fully to the client, an appreciation for the client's concerns. Below is my summary notes based on the article.

The purpose of statistical consulting can be as simple as helping understand a perceived problem or as complex as implementing a solution. All the procedures required better communicate effectively at and with all levels of an organization.  “Seek first to understand, then to be understood. Communication is the most important skill in life.” The article list the skills,


  • Professionalism
  • Time management
  • Judgment
  • Team player
  • Good communication
  • Good listening
  • Roles and responsibilities
  • Involving other consultants
  • Reputation


  • Skill in structuring task
  • Technical knowledge/skills
  • Industry experience
  • Commitment to clients
  • Getting along with clients

Suggestion: An important element of good communication between the client and consultant is defining the boundaries of the work. It’s as important that you indicate what you won’t do as it is to describe what you are committing to do. Of course, once you get into the project, you may see something that wasn’t obvious from the outset. You can always provide optional steps that would extend the range of the original agreement and that may require additional resources.In almost any consulting task, assumptions must be made, as few projects will be so obvious that no assumptions are required. It will be to your credit to clearly identify and state these assumptions and, to the extent possible, emphasize their implications. Knowing what is too much effort to expend for the next improvement step is a valuable attribute of a good statistical consultant.

Some of the tools that can help the communication could be agendas, flowcharts, checklists, and minutes from meetings. Following are several important steps in ensuring meetings are efficient:
  • Include the right people
  • Set an agenda
  • Stick to the agenda
  • Follow up the meeting with minutes
Taking and distributing minutes that summarize key points of agreement and disagreement—including action items and next steps—and identify the responsible person and a tentative due date helps move from the meeting to problem resolution.

The most efficient solution may never work if the people who need to bring it about do not benefit in some way from the change. Even more basic, many have an understandable motivation to continue using tried-and-true methods. It may be important to ensure the staff is involved in the development of the new procedure and ‘buys in’ to the new methods.

Finally, never forget that ethical issues undergird all our work. Following accepted ethical practices is a necessary requirement for any statistical consultant.


We will definitely need to learn new knowledge to break the boundaries to serve as the statistical professionals.

Tuesday, September 4, 2012

Oral Health Topics in NHANES 2009-2010

In this post, we discuss some results related to oral health questionnaire for the most recent national data source NHANES 2009-2010. The oral health question were asked to those sampled subjects who are 30+ years old by the trained dentists. The reference page is http://www.cdc.gov/nchs/nhanes/nhanes2009-2010/OHQ_F.htm. Data can be downloaded through: Oral Health (Data [XPT - 366 KB]) (April, 2012 last updated)

The data will represent 177,284,326 size in the population. The related oral health questions are,
Item LabelSAS Label
SEQNRespondent sequence number
OHQ835Do you think you might have gum disease?
OHQ845Rate the health of your teeth and gums
OHQ850Ever had treatment for gum disease?
OHQ855Any teeth became loose without an injury
OHQ860Ever been told of bone loss around teeth
OHQ865Noticed a tooth that doesn't look right
OHQ870How many days use dental floss/device
OHQ875Days used mouthwash for dental problem

SEQN exists for every data set. This is the unique ID for every subject. When using the data to represent the  National Statistics, we need the following variables from the demographic data,
WTINT2YRFull Sample 2 Year Interview Weight
WTMEC2YRFull Sample 2 Year MEC Exam Weight
SDMVPSUMasked Variance Pseudo-PSU
SDMVSTRAMasked Variance Pseudo-Stratum
 The difference between the two weights is whether you will use the MEC exam data. If you need to use that data, the weight should be WTMEC2YR because of the additional missing values in the MEC exam data. If only use the interview questions, we need to use the following macro,


%macro wtfreq(indata,var);
proc surveyfreq data=&indata nosummary missing; 
 tables &var;
strata SDMVSTRA;
cluster SDMVPSU;
weight WTINT2YR;
run;
%mend wtfreq;

%wtfreq(indata, OHQ835);


Do you think you might have gum disease?
OHQ835 Frequency
Weighted
Frequency
Std Dev of
Wgt Freq
Percent
Std Err of
Percent
.  Missing 343 8325221 2196334 4.6960 1.3475
1 Yes 793 26011822 2366676 14.6724 0.7829
2 No 3977 140950470 9104115 79.5053 1.1760
7 Refused 1 20453 20453 0.0115 0.0114
9 Don't know 63 1976361 267363 1.1148 0.1587
Total 5177 177284326 10219592 100.000

The weighted frequency in the table represents the frequency in the population. In the above table, we can reach the conclusion that about 15% of the US population think why might have the gum disease.

Rate the health of your teeth and gums
OHQ845 Frequency
Weighted
Frequency
Std Dev of
Wgt Freq
Percent
Std Err of
Percent
. Missing 343 8325221 2196334 4.6960 1.3475
1 Excellent 545 22985973 2373436 12.9656 0.9643
2 Very good 922 39003006 2929386 22.0003 0.5978
3 Good 1719 61155864 4043994 34.4959 0.8821
4 Fair 1088 30471374 2057740 17.1879 0.7934
5 Poor 551 15177430 1455809 8.5611 0.5190
7 Refused 1 18371 18371 0.0104 0.0100
9 Don't know 8 147086 55579 0.0830 0.0317
Total 5177 177284326 10219592 100.000

We may reach the conclusion based on the above data like, nearly 8.6% of the population rate their health of teeth and gum as poor and only 35% rate them as excellent or very good.

You can make similar conclusion based on the below output form the data.
Ever had treatment for gum disease?
OHQ850 Frequency
Weighted
Frequency
Std Dev of
Wgt Freq
Percent
Std Err of
Percent
. Missing 343 8325221 2196334 4.6960 1.3475
1 Yes 1044 33739681 3079672 19.0314 1.5191
2 No 3769 134555226 9543820 75.8980 1.4850
7 Refused 1 18371 18371 0.0104 0.0100
9 Don't know 20 645827 226540 0.3643 0.1258
Total 5177 177284326 10219592 100.000



Any teeth became loose without an injury
OHQ855 Frequency
Weighted
Frequency
Std Dev of
Wgt Freq
Percent
Std Err of
Percent
. Missing 343 8325221 2196334 4.6960 1.3475
1 Yes 763 20719884 2078602 11.6874 0.8389
2 No 4067 148160937 9570586 83.5725 1.0564
9 Don't know 4 78285 22173 0.0442 0.0136
Total 5177 177284326 10219592 100.000



Ever been told of bone loss around teeth
OHQ860 Frequency Weighted
Frequency
Std Dev of
Wgt Freq
Percent Std Err of
Percent
. Missing 343 8325221 2196334 4.6960 1.3475
1 Yes 542 19298087 1399093 10.8854 0.5542
2 No 4257 148736540 10048259 83.8972 1.4345
9 Don't know 35 924478 174418 0.5215 0.0849
Total 5177 177284326 10219592 100.000



Noticed a tooth that doesn't look right
OHQ865 Frequency Weighted
Frequency
Std Dev of
Wgt Freq
Percent Std Err of
Percent
. Missing 343 8325221 2196334 4.6960 1.3475
1 Yes 778 20769060 1342029 11.7151 0.5357
2 No 4050 148083053 10084759 83.5286 1.2755
9 Don't know 6 106991 42415 0.0604 0.0242
Total 5177 177284326 10219592 100.000



How many days use dental floss/device of last 7 days
OHQ870 Frequency Weighted
Frequency
Std Dev of
Wgt Freq
Percent Std Err of
Percent
. 343 8325221 2196334 4.6960 1.3475
0 1808 54076766 3593763 30.5028 1.1700
1 313 12490459 1134092 7.0454 0.6062
2 436 16727807 1164811 9.4356 0.5075
3 368 14448550 1459178 8.1499 0.5492
4 270 11047692 1258998 6.2316 0.4305
5 175 7549817 926752 4.2586 0.4287
6 62 2842765 369731 1.6035 0.2194
7 1395 49673858 4245024 28.0193 1.0970
77 4 50979 27228 0.0288 0.0148
99 3 50411 34112 0.0284 0.0191
Total 5177 177284326 10219592 100.000



Days used mouthwash for dental problem of last 7 days
OHQ875 Frequency Weighted
Frequency
Std Dev of
Wgt Freq
Percent Std Err of
Percent
. 343 8325221 2196334 4.6960 1.3475
0 2109 79796493 6331924 45.0105 1.4097
1 197 7135899 701857 4.0251 0.4390
2 304 10222261 916427 5.7660 0.3767
3 282 8965504 987258 5.0571 0.4088
4 178 7018847 887664 3.9591 0.4123
5 97 3797136 567744 2.1418 0.3437
6 39 1591065 377335 0.8975 0.2086
7 1620 50190107 3809085 28.3105 1.3582
77 3 38978 24440 0.0220 0.0132
99 5 202815 108603 0.1144 0.0630
Total 5177 177284326 10219592 100.000