AAPHP News Volume 1 Issue 16 
News Items

1. Agenda for meeting (especially useful for those who don't want to call in or stay for the whole AAPHPSaturday meeting) Item 1.

2. Spread the public health word, be tapped at APHA see item 2

3. HIPAA Should zip code level data be confidential - If not see Item 3

4. Flu vaccines are available for high risk patients see Item 4

5. Kim Buttery highly recommends this site: http://www.iom.edu/IOM/IOMHome.nsf/Pages/october+17 

on IOM Annual Meeting Quality of Health and Health Care


AAPHP News is sent to members whenever we receive several items of potential interest. Send information for this newsletter to the editor at vmdato@pitt.edu .

Feel free to forward this newsletter to physicians who may be interested in joining. NOTE: New member special join now for 2001 at 2001 dues ($30.00 for resident/retired/reduced income $60.00 for active physicians) and receive the end of 2000 free. A membership application form can be found on our web page http://www.aaphp.org . Pay by web or mail If you pay by mail use this address: AAPHP, PMB#1720, P.O. Box 2430, Pensacola, Florida 32513-2430). 


Item 1 Agenda Call in number 360-923-2997 SATURDAY,11TH NOVEMBER,2000
MAINE ROOM, MARRIOTT COPLEY PLACE, BOSTON
4:30 P.M-6:00 P.M
* Welcome 4:30 to 4:35 Call to order by President
* Housekeeping - dues and registration fees will be collected at the dinner break
* For Individuals who call in. If you have a questions or comment please tap one of the keys on your phone and we will recognize you
* Introductions 4:35 to 4:45
* Members present introduce themselves
* President report and review of mission 4:45 to 4:55
* Brief reports of officers
* Acting Executive Director - Bulletin, 4:55-5:00
* Vice President - 5:00 to 5:05
* Secretary- 5:05 to 5:10
* Treasurer's Report 5:10 to 5:15
* Web Master 5:15 to 5:20
* AMA Delegate 5:20 to 5:30
* Action Matrix Part 1 5:30 to 6:00
* Job Market 5:30 to 5:45
* Public Health Infrastructure 5:45 to 5:55
* Membership Recruitment 5:55 to 6:00

Dinner Break 6 P.M. to 7P.M

7:00 P.M. to 8:30 P.M.

* Action Matrix Part 2
* Selection of indiviuduals to nominate for RRC, ABPM, and AMA Outcomes measurement workgroup - 7:00 to 7:10
* Accreditation , performance standards and licensure 7:10 to 7:20
* ATPM grant applications 7:20 to 7:30
* Resolutions 7:30 to 8:00
* Bulletin8:00 to 8:10
* Tobacco 8:10 to 8:20
* Other new business. 8:20 to 8:30

Item 2 - I received this information from Joseph Friedman of the The Health TV Channel, Inc. Anyone who will be at APHA and has a favorite topic can participate in an educational taping. See the information below:

>From the Health TV Channel, Inc.
The Health TV Channel will be taping during the APHA conference. We are particularly interested in inviting individuals to participate in our programming by presenting talks for our cameras, or by creating moderated discussion groups ala "Larry King." Our staff is creating a schedule and have hour long blocks from 8am to 5 pm Monday through Thursday. If you would like to present or create a discussion group, please respond to scheduling@healthtvchannel.org

We will have three cameras setup in the exhibit / entrance area. We may tape a "scientific session" if the presenter can come to us. Generally we're looking for informative content geared to workforce professionals or the general public. We are more able to host a "roundtable" on a topic. There are no expenses or fees involved. The final product is not sold, but aired as part of The Health TV Channel. We will confirm a time for the presenters / interviewees to meet at our "studio" and begin to tape. We need to evaluate the topics before confirming. Please submit your topic

suggestion to scheduling@healthtvchannel.org .

The Health TV Channel is developing a national public health cable network. These tapes will air for general information and will not be used commercially. We cover quite a bit of Public Health content which looks much like CSPAN. The tapes will not be edited for content. We ask all presenters to sign a "Talent Release." Leslie Gulden, our Production Coordinator, is happy to answer inquiries. You may reach her at scheduling@healthtvchannel.org  There will also be a s
sign up sheet at the conference.
Joseph Friedman, Executive Director
The Health TV Channel, Inc.
A non - profit organization
3820 Lake Otis Parkway
Anchorage, Alaska 99508
907 770 6200,  907 563 8453(FAX)

Item 3. Evidently ZIP CODE is being considered a confidential data element Below are a variety of emails that came through the Public Health Consortium. I am told that there may still be time to prevent this. Dr. Bill Braithwaite at email address: bbraithwaite@osaspe.dhhs.org

 

HIPPA- From: Elixhauser, Anne [mailto:AElixhau@AHRQ.GOV]
Sent: Tuesday, October 24, 2000 9:31 AM
To: PH-CONSORTIUM-L@LIST.NIH.GOV
Subject: [PH-CONSORTIUM-L] Patient ZIP code and county will be lost in privacy regs on discl osure of de-identified data
The most recent version of the privacy regulations state that 5-digit patient ZIP code and patient county must be removed or encrypted before release of data. (3-digit ZIP code would still be releasable.)
These two variables are key to current geo-coding that allows us to link area-level variables to person-level records. For example, we would lose the ability to link hospital discharge records to information on the median income of the ZIP code in which the patient resides. This means some critical information we currently have on SES would be lost. Similarly, we could no longer link discharge records to information on medical resources in the patient's county from data sources like the Area Resource File.
As I understand it, the decisions about these regulations will be made in the next week. I haven't seen any discussion among Consortium members on this issue. Is this something that the Consortium should take a stand on? Perhaps we could request that the decision on de-identified data be delayed until the ramifications of this decision are better understood?


Anne Elixhauser, Ph.D., Agency for Healthcare Research and Quality (AHRQ)
2101 East Jefferson St., Suite 605
Rockville, MD 20852

Subj: Re: [PH-CONSORTIUM-L] Patient ZIP code and county will be lost in privacy regs on disclosure of de-identified data
Date: 10/24/00 12:10:53 PM Eastern Daylight Time
From: MFitzmau@AHRQ.GOV (Fitzmaurice, Michael)
Sender: PH-CONSORTIUM-L@LIST.NIH.GOV (Main Listserv for discussions relating to public health data standards)
Reply-to: PH-CONSORTIUM-L@LIST.NIH.GOV (Main Listserv for discussions relating to public health data standards)
To: PH-CONSORTIUM-L@LIST.NIH.GOV

Steve,
It is my understanding based on the NPRM of Nov 3, 1999, that first covered entities (health providers, health plans, clearinghouses) may disclose individually identifiable health information to researchers without the individuals authorization if the project has been approved by an institutional review board or a privacy board.

Second, if the IIHI has 19 variables stripped from it, the resulting data set is considered de-identified and no longer falls under the domain of the HIPAA privacy standard rule. One of those variables includes zip code and county code. Public comments suggest a second-best solution--keeping zip code at the 3-digit level. This may be better than nothing but loses the granularity of a 5-digit zip code and this loss of granularity means the loss of measured variation in socioeconomic status and other variables.

Further, many of the SES variables are not easily obtained or aggregated at the 3-digit level.

There are implications for what data researchers can obtain from covered entities, for what data organizations who are business associates of covered entities can disclose, and for the richness of public use files that may result. Remember that nothing in the privacy rule forces covered entities to disclose IIHI (except to the individual, or to HHS for fraud and abuse); the covered entity may still say "No, I will not disclose this information to you." Also remember that the HIPAA rules can be changed once a year and

that there are two years before implementation. Solidarity on these issues can make a difference. Is something better than nothing? Is so much lost as to make use of a 3-digit zip code valueless? What about the loss of county codes in IIHI that becomes de-identified? Would it affect your normal practices?

I think these are some of the questions that Anne is raising for your discussion.

Mike

J. Michael Fitzmaurice, Ph.D., FACMI
Senior Science Advisor for Information Technology
Immediate Office of the Director
Agency for Healthcare Research and Quality
2101 East Jefferson St, Suite 600
Rockville, Maryland USA 20852
P: 1-301-594-3938,  F: 1-301-594-2168  E: Mfitzmau@ahrq.gov

Subj: Re: [PH-CONSORTIUM-L] Patient ZIP code and county will be lost in privacy regs on discl osure of de-identified data
Date: 10/24/00 12:30:35 PM Eastern Daylight Time
From: MFitzmau@AHRQ.GOV (Fitzmaurice, Michael)
Sender: PH-CONSORTIUM-L@LIST.NIH.GOV (Main Listserv for discussions relating to public health data standards)
Reply-to: PH-CONSORTIUM-L@LIST.NIH.GOV (Main Listserv for discussions relating to public health data standards)

d)(1) Standard: use or disclosure of de-identified protected health information. The requirements of this subpart do not apply to protected health information that a covered entity has de-identified, provided, however, that:

(i) Disclosure of a key or other device designed to enable coded or otherwise de-identified information to be re-identified constitutes disclosure of protected health information; and

(ii) If a covered entity re-identifies de-identified information, it may use or disclose such re-identified information only in accordance with this subpart.

(2) Implementation specifications. (i) A covered entity may use protected health information to create

de-identified information by removing, coding, encrypting, or otherwise eliminating or concealing the information that makes such information individually identifiable.

(ii) Information is presumed not to be individually identifiable (de-identified), if:

(A) The following identifiers have been removed or otherwise concealed
(1) Name;
(2) Address, including street address, city, county, zip code, and equivalent geocodes;
(3) Names of relatives;
(4) Name of employers;
(5) Birth date;
(6) Telephone numbers;
(7) Fax numbers;
(8) Electronic mail addresses;
(9) Social security number;
(10) Medical record number;
(11) Health plan beneficiary number;
(12) Account number;
(13) Certificate/license number;
(14) Any vehicle or other device serial number;
(15) Web Universal Resource Locator (URL);
(16) Internet Protocol (IP) address number;
(17) Finger or voice prints;
(18) Photographic images; and
(19) Any other unique identifying number, characteristic, or code that the covered entity has reason to believe may be available to an anticipated recipient of the information; and

(B) The covered entity has no reason to believe that any anticipated recipient of such information could use the information, alone or in combination with other information, to identify an individual.

(iii) Notwithstanding paragraph

(d)(2)(ii) of this section, entities with appropriate statistical experience and expertise may treat information as de-identified, if they include information listed in paragraph (d)(2)(ii) of this section and they determine that the probability of identifying individuals with such identifying information retained is very low, or may remove additional information, if they have a reasonable basis to believe such additional information could be used to identify an individual.Subj:

To: PH-CONSORTIUM-L@LIST.NIH.GOV

This message is a follow-up to e-mail discussions earlier this week on the privacy regulations being promulgated under HIPAA. On this listserv, I had raised the question about the possibility of losing patient ZIP code and patient county in administrative databases for research.

First, thanks very much for your responses to my initial e-mail. I forwarded all the messages I received on this subject to Nicole Lurie (Principal Deputy Assistant Secretary for Health, Office of Public Health and Science, Office of the Secretary of the Department of Health and Human Services), who has been investigating this issue with us.

Through Dr. Lurie, I just received communication from DHHS (Bill Braithwaite) that the following description of the situation is correct: (By the way, I mention HCUP below. For those who don't know, HCUP is the Healthcare Cost and Utilization Project, an administrative data collection and dissemination partnership between AHRQ and State data organizations and hospital associations -- http://www.ahcpr.gov/data/hcup/. NAHDO plays a key role in HCUP.)


State data collection, HCUP-related activities, and research are covered under the following exclusions (from the a list of "Proposed Disclosures Allowed Without Patient Consent" provided by Bill Braithwaite):
2 - "public health activities"
4 - "health oversight activities"
9 - "research purposes"
11 - "specialized government function" and maybe:
1 - "required by state law" (in states with laws about what specific variables are to be collected) Data projects like HCUP could disseminate identifiable data (e.g., patient ZIP and county) for research purposes if there was IRB or human subjects approval for the project. However, the exception stated in the last paragraph of the section of regulation we've been studying (Privacy NPRM [Nov. 3, 1999] Federal Register) provides alternative avenues for data release:

"entities with appropriate statistical experience and expertise may treat information as de-identified, if they include information listed in paragraph (d)(2)(ii) of this section [the list of identifiable data elements] and they determine that the probability of identifying individuals with such identifying information retained is very low, or may remove additional information, if they have a reasonable basis to believe such additional information could be used to identify an individual."

For example, we might be able to release patient ZIP code if we reported patient age in 5-year increments rather than in yearly increments. HCUP will be soon be embarking on a study to determine the security trade-offs between variables such as these. This looks like research and data dissemination activities such as those conducted by state data organizations and HCUP would be covered, but we would just have to get stricter in our release of certain data elements (ZIP code and county, specifically), in the absence of other adjustments in the data.

Hope this helps -- Anne

Thanks to Anne for bring this up. NAHDO is very concerned with this recommendation to classify zip as an identifiable data element, that this may diminish the utility of publicly-available information to our constituencies.  For almost 20 years, many state health data agencies have released zip code with their publicly-available/de-identified discharge data sets. With proper access controls, comprehensive data use agreements, and aggregation of patient demographic data, these agencies have proven it is possible to promote health research and public information resources without compromising patient confidentiality. Most of these agencies prohibit users from linking the information with other data for the purposes of attempting to identify an individual (fines/penalties are the result).


FYI---Excerpts from NAHDO's comments on the Privacy NPRM regarding the de-identifiable data.2. Creation of De-identified Information (Section 164.506(d)) NAHDO recognizes this provision as one providing "safe harbor" for releasing data, but is very concerned about the inclusion of "city, county, zip code, and equivalent geocodes" as "individually identifiable" information. Not only is this provision counter to the concept of the "chain of trust" referred to in the proposed regulations, but it has the potential to hinder the following activities conducted routinely by public users of statewide discharge data: · Community assessment · Sub-population analyses · Market studies Zip code and other geocodes are vital data elements in de-identified data sets and patient privacy is not compromised if other measures are taken (encrypting, truncating, aggregating, and/or removing sensitive data elements). Preservation of useful de-identified public use files is important to state health data agencies and the provider community. In addition to the loss of zip codes in public use files, removal of geocode data elements creates a hardship for the covered entities and other private sector users. The results could be: · Diminishing the value of public use files which are a cornerstone of public-private partnerships at the local level · Increases in the number and types of requests for research-level and other sensitive data elements Additionally, information needs evolve over time and removing data from the public domain concerns NAHDO. NAHDO and its members are experts in establishing methods and standards for masking patient identity in public use files. NAHDO recognizes there is always a risk in releasing health information, even de-identified information, and that no method is fail-safe.  The risks of any data release must be balanced against the societal or public benefit expected to be gained by such releases. Recommendations: a) Permit for flexibility in defining the standard of de-identified data to accommodate local needs. If HHS insists on prescribing a standard, then these presumptive standards should include geocoding at the county, city, or zip-code level--when accompanied by other standards of de-identification as appropriate for the local community. b) Expand the "types of entities" with statistical experience and expertise to make a judgment for determining standards of "identifiability" to include state health data agencies, public health agencies, and other health data organizations with epidemiology/statistical/health services research staff". c) NAHDO offers to assist the Secretary in establishing guidance and standards for determining the identifiability of health information.

Zip code increases the policy and market relevance of the data to broad audiences. Health data agencies have long weighed the data requestor's needs with their statutory limitations and due diligence as stewards of the data. NAHDO would support Anne's suggestion that a request that the decision on de-identified data be delayed until the ramifications of this decision are better understood? Other thoughts?you can address comments to:

U.S. Department of Health and Human Services, Assistant Secretary for Planning and Evaluation
Attention: Privacy-P
Hubert H. Humphrey Building, Room G-322A
200 Independence Avenue SW,  Washington DC 20201

and send (or email to): NAHDO, Denise Love, 391 Chipeta Way, Suite G Salt Lake City, Utah 84108 dlove@nahdo.org

I am hoping someone from the Consortium Listserve will provide an opinion on responding to HHS at this point. My "unofficial" guess is that the rules have been revised and sent to Donna Shalala for signing. My other guess is that there may be another comment period in which we will review the rules and submit comments again--and that we won't see any final rules for at least a year. But I may be way off.

ITEM 4. From the Pennsylvania Medical Society STAT mail system. PennsylvaniaMedicalSociety@pamedsoc.org

Priority given to orders for high-risk patients CDC contracts for additional flu vaccine doses Nine million doses of the influenza vaccine are now available for providers to vaccinate high-risk persons. The Centers for Disease Control and Prevention asked Aventis-Pasteur, Inc., to manufacture additional doses, and Aventis will give first priority to orders from providers who plan to vaccinate high-risk patients.

Physicians may apply for vaccine orders now through Dec. 1 by calling Aventis' Fluzone(r) hotline at (800) 720-8972. You will be asked to provide a fax number, or you may leave your name and address to have an application mailed. You may also get an application on-line. Go to www.vaccineshoppe.com  http://www.vaccineshoppe.com/  and click on the Fluzone(r) Application Form link.

Completed application forms can be faxed to (888) 889-7129. Orders will not be taken by telephone. If you have trouble contacting Aventis, call the Society's Business Resource Center at (877) BRC-2425, and the application will be faxed to you. Wholesale distributors can apply to purchase vaccine

starting Dec. 4, if doses remain available. The application asks for the number of patients you are considering vaccinating for the flu, the number of those patients you consider to be high-risk for complications or a person in close contact with a high-risk patient, and the number of requested doses. Aventis will notify applicants before mid-December if their order will be filled. Delivery of the vaccine is expected to begin Dec. 12 and end by early January 2001.

The CDC expects that approximately 75 million doses of flu vaccine will be distributed for the 2000-2001 flu season. This includes the 9 million doses contracted by CDC. The vaccine fell into short supply earlier this year when four licensed manufacturers had difficulties growing one of the three flu strains. Aventis reported that customers placed so many orders for the vaccine this spring that it was sold out of its 26 million doses by May.

For the 1999-2000 influenza season, approximately 77 million doses of vaccine were distributed, and 3 million were returned to the manufacturers. Vaccine availability tracked on Web site

The Centers for Disease Control and Prevention National Immunization Program has developed an "Influenza Vaccine Availability" Web site that provides information about the availability of the vaccine from manufacturers and wholesale distributors and lists state health departments that may have

information about vaccine availability among local providers. The Web site is updated weekly.

Access the site at http://www.cdc.gov/nip/flu-vac-supply  .

Updated ACIP recommendations for influenza vaccine for the 2000-2001 season and other influenza-related information is available at http://www.cdc.gov/ncidod/diseases/flu/fluvirus.htm  .

Or contact the NIP by e-mail, nipinfo@cdc.gov , or by telephone (800) 232-2522.

9 - November 10, 2000