Every year I write about the projects and trends which keep me up at night. Here's my list for FY12:
1. Workforce recruitment/retention - $27 billion in stimulus funds from HITECH have increased demand for experienced IT staff to implement and support electronic health records. In many ways, it's a mini "dot com" boom for healthcare IT experts. This makes recruiting and retaining qualified staff even harder. Tomorrow, I'm meeting with a consulting team to formulate an FY12 workforce strategy.
2. 5010/ICD10 - 5010 describes a set of X12 standards used for administrative transactions (benefits/authorization. referral authorization, claims). Payers and providers must support 5010 by January 1, 2012 or risk disruption of the revenue cycle. BIDMC completed all its 5010 work and is now in final testing with every payer. Most payer and provider stakeholders will meet the deadline, but significant resources have been pulled from other projects. ICD-10 implementation is required by October 1, 2013 and I've written about those challenges. Billions will be spent, many healthcare IT projects will be deferred for the next 2 years, and the end result will be no cost savings (coding costs are likely to increase 50%), no quality improvement, no increased safety, and no efficiency gains. If we complete the ICD-10 project on time, no one will notice, but customers will all be angry at the IT department (and the CIO) for the work on other projects that was deferred.
3. Vendor Product Quality - over the past year, I've had several bad experiences with infrastructure and application vendors which delivered products that did not have the reliability, security, or performance promised. Why?
* the pace of innovation is so fast, that time for quality assurance is diminished.
* the economy has stressed companies and they are focused on making as many sales as fast as they can while controlling development and support costs
* the end result is less satisfied customers.
4. Storage growth - BIDMC is approaching 2 petabytes of clinical data and Harvard Medical School has exceeded 3 petabytes of research data. Balancing transactional performance, reliability, and the cost of storage at petabyte scale is very challenging. In FY12, I'll continue to introduce new storage technologies and management tools (including chargebacks to sustain the cost of storage growth), with the hope that I'll be able to keep up with demand.
5. Analytics/Business Intelligence - We have petabytes of data but users want information, knowledge and wisdom. In FY12, I have a five part strategy to support analytics:
a. For comprehensive ad hoc analytics with quality assurance, BIDMC has an expert data mining team which can explore any clinical or financial data while reviewing the accuracy of the underlying data.
b. For less rigorous ad hoc analytics, we have a self service query tool which enables users to explore data themselves. However, such self service queries are unlikely to filter out data of questionable quality (men having babies, women having prostate procedures etc)
c. For reports that are run frequently, we're developing a new set of business intelligence tools that use Microsoft SQL Analysis and Reporting services.
d. For our Blue Cross Alternative Quality Contract and our Accountable Care organization activities we have a Quality Data Center that aggregates clinical data. We also have an all payer claims data warehouse.
e. For unstructured data we plan to experiment with a new generation of natural language processing tools
6. Healthcare Reform and Mergers - Accountable care organizations require substantial IT investments and new processes to coordinate/manage care across the community. Smaller organizations may be unable to implement all that is needed so they may seek mergers with a larger organization. Today I oversee the IT requirements for 2 hospitals and multiple clinician groups. Over the next year, I believe the number of affiliated organizations requiring IT support will increase, requiring new investments in healthcare information exchange and analytics.
7. Mobile devices/consumer IT - We have 1000+ iPads accessing our web-based applications today. Clinicians are mobile people and need to view results, enter orders, and communicate with team members at the bedside. Increasingly they'll want to use their own devices, which creates support and security challenges.
8. Governance - Governance is essential to maintain satisfaction in budget constrained times. In all IT organizations, the supply of resources is fixed, but demand is infinite. There’s an direct relationship among project scope, project timing, and project resources. Governance and communication are the best tools to reduce scope, limit demand, and keep satisfaction at reasonable levels. Here’s how BIDMC does it.
9. Compliance/Regulatory Pressures - At the same time that Meaningful Use empowers care coordination, population health, and public health with health information exchange, the penalties for privacy breaches have increased. Thousands of new regulations have been enacted in the US during my 15 years as CIO. With nearly 25% of my IT organization working on some aspect of compliance, the amount of time left for innovation is diminishing.
10. Security - The internet has become a swamp with nearly 50% of internet devices infected with some type of malware. The most basic freedoms we've enjoyed on the internet - the ability to visit any site, experiment with new applications, and share media with friends, is now a threat to the privacy of business data. The next year will require us to rethink how consumer computing and business computing can co-exist.
That's an intimidating list. However, there is always hope and we'll make progress in all these areas over the next year. By FY15, we will not even remember the events of FY12. Accelerating challenges come with the territory of being an IT leader. Keep smiling.
Monday, 31 October 2011
Friday, 28 October 2011
Cool Technology of the Week
On November 18, I'm giving a lecture about technologies that educators can use to mentor their students.
My experience running Harvard's Mycourses taught me that social networking for student/faculty interaction works very well. Blogs, wikis, chat rooms, and interactive simulations are useful, but structured community question/answer requires a more powerful tool.
I recently heard about Piazza, a Silicon Valley startup that supports over 900 school campuses and ten thousands of students with a free online collaboration platform.
It has received investment from several venture capitalists, no doubt because it has attracted a large number of devoted users who spend hours per day using the site.
It provides a faculty platform for managing queues of questions, triaging crowd sourced answers to questions, and entering answers.
Also, there's a student platform for reading answers and sharing ideas with other students.
A social networking platform for students and faculty that empowers students to master difficult concepts together. That's cool!
My experience running Harvard's Mycourses taught me that social networking for student/faculty interaction works very well. Blogs, wikis, chat rooms, and interactive simulations are useful, but structured community question/answer requires a more powerful tool.
I recently heard about Piazza, a Silicon Valley startup that supports over 900 school campuses and ten thousands of students with a free online collaboration platform.
It has received investment from several venture capitalists, no doubt because it has attracted a large number of devoted users who spend hours per day using the site.
It provides a faculty platform for managing queues of questions, triaging crowd sourced answers to questions, and entering answers.
Also, there's a student platform for reading answers and sharing ideas with other students.
A social networking platform for students and faculty that empowers students to master difficult concepts together. That's cool!
Thursday, 27 October 2011
A Personal Experience with 4G LTE
Steve Berry, BIDMC's Director of Academic and Research Computing, wrote this guest blog entry, about his experience with 4G LTE:
"I've had the Verizon MiFi 4G LTE card for 6 months.
In a 4G network area, the access and speed is incredible. It is so good that you can exceed the 5GB monthly service cap in 4-5 hours!
Once you limit your new found flexibility, it is like having your Ethernet access everywhere. It even worked well while driving long distances.
In a limited 4G network (like the commuter rail line between Boston and Worcester), it is very problematic. When the 4G signal is low and a 3G is available, the device switches modes. This drops the existing connection and 3G takes 20-30 seconds to activate. If you happen to come back into a 4G area (signal level above the 3G value), it switches modes again taking another 30 seconds to establish an active link. In the 1 hour Worcester to Boston route, I used to lose 3G connectivity twice and lose about 2 minutes while the SSLVPN and MiFi resynchronized. Since moving to the 4G, I lose 30 minutes of the 1 hour commute due to continuous mode changes.
There is no ability to force the unit into 3G or 4G only mode. There is no standard for setting up a new connection before dropping an existing connection. Recent firmware updates have not improved performance.
If you have not yet upgraded to 4G LTE, it's important to first study your usage patterns and the 3G/4G coverage in your area. Moving through a mixed area is actually the worst of both worlds.
On my recent vacation, my daughter found that the iPad easily linked to the MiFi (all LTE in Tampa). She happily watched her Netflix videos for several days and I was none the wiser until a month later when the monthly bill came and I was being charged for 25GB of data activity. I had no idea what caused the overage until I reviewed the dates!"
Thanks for your insights Steve.
My iPhone 4S arrives soon and should not have this problem because it does not yet support LTE. By the time Apple releases an LTE phone, 4G wireless should be more ubiquitous in Massachusetts.
"I've had the Verizon MiFi 4G LTE card for 6 months.
In a 4G network area, the access and speed is incredible. It is so good that you can exceed the 5GB monthly service cap in 4-5 hours!
Once you limit your new found flexibility, it is like having your Ethernet access everywhere. It even worked well while driving long distances.
In a limited 4G network (like the commuter rail line between Boston and Worcester), it is very problematic. When the 4G signal is low and a 3G is available, the device switches modes. This drops the existing connection and 3G takes 20-30 seconds to activate. If you happen to come back into a 4G area (signal level above the 3G value), it switches modes again taking another 30 seconds to establish an active link. In the 1 hour Worcester to Boston route, I used to lose 3G connectivity twice and lose about 2 minutes while the SSLVPN and MiFi resynchronized. Since moving to the 4G, I lose 30 minutes of the 1 hour commute due to continuous mode changes.
There is no ability to force the unit into 3G or 4G only mode. There is no standard for setting up a new connection before dropping an existing connection. Recent firmware updates have not improved performance.
If you have not yet upgraded to 4G LTE, it's important to first study your usage patterns and the 3G/4G coverage in your area. Moving through a mixed area is actually the worst of both worlds.
On my recent vacation, my daughter found that the iPad easily linked to the MiFi (all LTE in Tampa). She happily watched her Netflix videos for several days and I was none the wiser until a month later when the monthly bill came and I was being charged for 25GB of data activity. I had no idea what caused the overage until I reviewed the dates!"
Thanks for your insights Steve.
My iPhone 4S arrives soon and should not have this problem because it does not yet support LTE. By the time Apple releases an LTE phone, 4G wireless should be more ubiquitous in Massachusetts.
Wednesday, 26 October 2011
It Takes a Village
Little did I know that at the meeting, I was chatting with the future National Coordinator for HIT, the future CTO of HHS, the future CTO of the US, and an influential policy thinker at the Center for American Progress.
Since that meeting, I've stayed in touch with them to exchange ideas, seek their advice, and share lessons learned.
In 2009, I became co-chair of the HIT Standards Committee. Little did I know that the HIT Standards Committee would become the most functional, most productive, and hardest working federal advisory committee in the Obama administration. Its experts have helped me enhance IT capabilities in all my technology roles.
In 2010, I worked with Brian Biles and Steven Morrison of the Center for Strategic and International Studies on Japanese healthcare IT policy. Little did I know that the work would become a foundation for earthquake/tsunami recovery IT planning. Brian and Steve inspired several trips to Japan and meetings with numerous government, academic, and industry leaders.
In 2011, I began working with Rick Shoup, Manu Tandon, and Micky Tripathi on Healthcare Information Exchange planning for Massachusetts. Little did I know that together they would create a unified Healthcare Information Exchange strategy for the Commonwealth that integrates public sector and private sector priorities with multiple funding streams into a single, extraordinary work plan. It has become one of my favorite projects.
On Friday, I'm co-leading a design session for public key infrastructure (PKI) in Massachusetts. I called my friend Dixie Baker at SAIC, my friend Arien Malec at RelayHealth, and my colleagues in government to share their experiences creating a trust fabric for large groups. Massachusetts will succeed by seeking the wisdom of others.
When I was young, I thought I had to be smart enough to solve every problem myself. In today's world, I'm convinced the best way to make a difference is surrounding yourself with people who are smarter than you. The best solutions take a village.
I've said that my tombstone will hopefully read "he made a difference". After the past few years of working with smart people, I'm convinced it would be better as "he was part of a village that changed the world."
Tuesday, 25 October 2011
The Healthpad Panel at AMIA
Yesterday I was in Washington DC at the American Medical Informatics Association annual meeting in Washington to join a panel with Dr. Henry Feldman, Dr. Larry Nathanson, and Janet Meyers RN discussing the use of tablet computers in medicine - Session S36 "Tablets in Healthcare: No Just for Pills Anymore"
Dr. Feldman began the presentation with great showmanship, pulling his iPad from a bucket of water, illustrating how the FrogSkin protects even a submerged iPad. He discussed his iPhone sterilization experiments. As a busy hospitalist, he explained the value of the iPad in providing "everywhere" computing - access to healthcare records, provider order entry, and clinical documentation applications at the bedside. He illustrated patient education materials that he uses to consent patients and explain care plans. He noted that the iPad's touch screen keyboard is good enough and that he has never used his physical keyboard dock.
Dr. Nathanson described his early adoption of the iPad to manage workflow in the Emergency Department. Over the past year we've worked on a streamlined web-based provider order entry application from the Emergency Department which brings one click ordering on the iPad to all our ED docs. Clinical documentation is fully functional on the iPad. Dr. Nathanson cleans his iPad with alcohol wipes and notes that he drops it at least once shift, with no discernible damage to date.
Janet Meyers RN presented an overview of Airstrip technologies, noting that they make perinatal telemetry available anywhere anytime to clinicians.
I presented an enterprise view of mobile technologies, highlighting they importance of infrastructure, mobile-enabled applications, an organizational culture which fosters adoption of new technology, and a willingness to invest in enhanced security protections to address the security risks of supporting consumer technologies on hospital network.
I'm a great champion of mobile devices and I truly believe the future of all clinician workflow is mobile, but that enthusiasm has to be tempered by the risks of commingling "Angry Birds" with clinical lookup on the same device. I highlighted the need for applications such as Good Technologies which separate the memory space/storage of work related applications from consumer applications. Only through the use of advanced intrusion detection/prevention, restrictive firewalls, web content filtering, web application firewalls, and security education can we keep mobile devices safe enough for use with clinical applications.
A great discussion with the important take home lesson - tablet/pad-based computing in healthcare is the future, but it must be implemented and managed prudently.
Dr. Feldman began the presentation with great showmanship, pulling his iPad from a bucket of water, illustrating how the FrogSkin protects even a submerged iPad. He discussed his iPhone sterilization experiments. As a busy hospitalist, he explained the value of the iPad in providing "everywhere" computing - access to healthcare records, provider order entry, and clinical documentation applications at the bedside. He illustrated patient education materials that he uses to consent patients and explain care plans. He noted that the iPad's touch screen keyboard is good enough and that he has never used his physical keyboard dock.
Dr. Nathanson described his early adoption of the iPad to manage workflow in the Emergency Department. Over the past year we've worked on a streamlined web-based provider order entry application from the Emergency Department which brings one click ordering on the iPad to all our ED docs. Clinical documentation is fully functional on the iPad. Dr. Nathanson cleans his iPad with alcohol wipes and notes that he drops it at least once shift, with no discernible damage to date.
Janet Meyers RN presented an overview of Airstrip technologies, noting that they make perinatal telemetry available anywhere anytime to clinicians.
I presented an enterprise view of mobile technologies, highlighting they importance of infrastructure, mobile-enabled applications, an organizational culture which fosters adoption of new technology, and a willingness to invest in enhanced security protections to address the security risks of supporting consumer technologies on hospital network.
I'm a great champion of mobile devices and I truly believe the future of all clinician workflow is mobile, but that enthusiasm has to be tempered by the risks of commingling "Angry Birds" with clinical lookup on the same device. I highlighted the need for applications such as Good Technologies which separate the memory space/storage of work related applications from consumer applications. Only through the use of advanced intrusion detection/prevention, restrictive firewalls, web content filtering, web application firewalls, and security education can we keep mobile devices safe enough for use with clinical applications.
A great discussion with the important take home lesson - tablet/pad-based computing in healthcare is the future, but it must be implemented and managed prudently.
Monday, 24 October 2011
The October HIT Standards Committee Meeting
The October HIT Standards Committee meeting included recommendations for privacy/security certification criteria, a discussion of next steps for the Standards and Interoperability framework priorities, and a review of the comments received from the Metadata Advanced Notice of Proposed Rulemaking.
We began with a discussion of the challenge of writing certification criteria for privacy and security, since the security of EHRs should depend primarily upon infrastructure assurances (networks, servers, storage, client devices, operating systems) and specialized security services. The EHR itself should provide only those security services which are specific to protecting the confidentiality, integrity, and availability of the electronic health information it manages.
The certification criteria that the Privacy and Security Workgroup developed are all “addressable”. To meet the criteria, each Complete EHR or EHR Module submitted for certification needs to either:
– Implement the required security functionality within the Complete EHR or EHR Module(s) submitted for certification or
– Assign the function to a third-party security component or service, and demonstrate how the certified EHR product, integrated with its third-party components and services, meets the criterion
We discussed the important topic of securing data at rest and recommended encryption for data on end-user devices controlled by EHR. However, we recognized that encryption of data in data centers is a risk management decision and out of scope for certification criteria.
We discussed audit trails and recognized that applications collect audit data in different formats using different architectures. The real value of an audit trail is the events it captures, not the format it stores them in. We selected the ASTM E2147-01 standard which specifies auditable events, leaving implementation details to each vendor. At some future time, it may be useful to standardize audit trail formats, but for now, there is limited value in imposing a standard audit trail format and architecture on existing products.
The HITSC approved the privacy and security certification criteria recommendations by consensus, with one small clarification of SHA-1/SHA-2 encryption requirements.
Next, we discussed the Standards and Interoperability Framework efforts on the NwHIN Exchange transport standards and transitions of care as well as a brief discussion of future work on radiology image exchange standards.
We agreed that additional testimony is needed from implementers of NwHIN exchange to understand their experiences with each component of the Exchange specification:
NHIN Messaging Platform Specification
NHIN Web Services Registry Specification
NHIN Authorization Framework Specification
NHIN Patient Discovery Specification
NHIN Query for Documents Specification
NHIN Retrieve Documents Specification
NHIN Access Consent Policies Specification
NHIN Health Information Event Messaging (HIEM) Specification
NHIN Document Submission Specification
NHIN Administrative Distribution Specification
We're developing a set of questions for implementers and will seek broad input from those in trenches who have coded or operated NwHIN Exchange environments.
We had a rich discussion about the consolidated CDA project. Consolidated CDA enhances and further constrains CCD/C32. Wes Rishel made the following comments:
Many standards experts who have been actively working on the Consolidated CDA project feel that it is a major accomplishment by HL7 to consolidate the specifications into a single document, organize them so that consistent XML structures are used for common data items in multiple document types, and to include well-specified data element names.
Many feel that the consolidated CDA alone will prevent as much as 50% of the programming errors found in C32 testing and that disagreements on the interpretations of the specifications will be far more easily resolved.
Programming, testing and resolution will all be enhanced again when the data element names are used in less highly nested XML and when Green CDA becomes accepted as the "over the wire" format.
HITSC will continue work on Consolidated CDA as it represents an important step forward for Transition of Care summaries.
Finally, we discussed the comments received about the Advanced Notice of Proposed Rulemaking. This input will be incorporated into the Notice of Proposed Rulemaking.
September and October were landmark meetings for the HITSC, with completion of the standards and certification criteria needed for Meaningful Use Stage 2. What's next? Based on my discussion with HITSC experts, I believe our work ahead includes:
Content
*Continued refinement of the Consolidated CDA implementation guides and tools to enhance semantic interoperability including consistent use of business names in "Green" over-the-wire standards.
*Standardizing DICOM image objects for image sharing and investigating other possible approaches. We'll review image transfer standards, image viewing standards, and image reporting standards.
*Simplifying the specification for quality measures to enhance consistency of implementation.
*Query Health - distributed queries that send questions to data instead of requiring consolidation of the data
Vocabulary
*Extending the quality measurement vocabularies to clinical summaries
*Finalizing a standardized lab ordering compendium
Transport
*Specifying how the metadata ANPRM be integrated into health exchange architectures
*Supporting additional NwHIN standards development (hearings about Exchange specification complexity, review/oversight of the S&I Framework projects on simplification of Exchange specifications). Further defining secure RESTful transport standards.
*Accelerating provider directory pilots (Microdata, RESTful query/response that separates the transaction layer from the schema) and rapidly disseminating lessons learned.
I look forward to our November meeting.
We began with a discussion of the challenge of writing certification criteria for privacy and security, since the security of EHRs should depend primarily upon infrastructure assurances (networks, servers, storage, client devices, operating systems) and specialized security services. The EHR itself should provide only those security services which are specific to protecting the confidentiality, integrity, and availability of the electronic health information it manages.
The certification criteria that the Privacy and Security Workgroup developed are all “addressable”. To meet the criteria, each Complete EHR or EHR Module submitted for certification needs to either:
– Implement the required security functionality within the Complete EHR or EHR Module(s) submitted for certification or
– Assign the function to a third-party security component or service, and demonstrate how the certified EHR product, integrated with its third-party components and services, meets the criterion
We discussed the important topic of securing data at rest and recommended encryption for data on end-user devices controlled by EHR. However, we recognized that encryption of data in data centers is a risk management decision and out of scope for certification criteria.
We discussed audit trails and recognized that applications collect audit data in different formats using different architectures. The real value of an audit trail is the events it captures, not the format it stores them in. We selected the ASTM E2147-01 standard which specifies auditable events, leaving implementation details to each vendor. At some future time, it may be useful to standardize audit trail formats, but for now, there is limited value in imposing a standard audit trail format and architecture on existing products.
The HITSC approved the privacy and security certification criteria recommendations by consensus, with one small clarification of SHA-1/SHA-2 encryption requirements.
Next, we discussed the Standards and Interoperability Framework efforts on the NwHIN Exchange transport standards and transitions of care as well as a brief discussion of future work on radiology image exchange standards.
We agreed that additional testimony is needed from implementers of NwHIN exchange to understand their experiences with each component of the Exchange specification:
NHIN Messaging Platform Specification
NHIN Web Services Registry Specification
NHIN Authorization Framework Specification
NHIN Patient Discovery Specification
NHIN Query for Documents Specification
NHIN Retrieve Documents Specification
NHIN Access Consent Policies Specification
NHIN Health Information Event Messaging (HIEM) Specification
NHIN Document Submission Specification
NHIN Administrative Distribution Specification
We're developing a set of questions for implementers and will seek broad input from those in trenches who have coded or operated NwHIN Exchange environments.
We had a rich discussion about the consolidated CDA project. Consolidated CDA enhances and further constrains CCD/C32. Wes Rishel made the following comments:
Many standards experts who have been actively working on the Consolidated CDA project feel that it is a major accomplishment by HL7 to consolidate the specifications into a single document, organize them so that consistent XML structures are used for common data items in multiple document types, and to include well-specified data element names.
Many feel that the consolidated CDA alone will prevent as much as 50% of the programming errors found in C32 testing and that disagreements on the interpretations of the specifications will be far more easily resolved.
Programming, testing and resolution will all be enhanced again when the data element names are used in less highly nested XML and when Green CDA becomes accepted as the "over the wire" format.
HITSC will continue work on Consolidated CDA as it represents an important step forward for Transition of Care summaries.
Finally, we discussed the comments received about the Advanced Notice of Proposed Rulemaking. This input will be incorporated into the Notice of Proposed Rulemaking.
September and October were landmark meetings for the HITSC, with completion of the standards and certification criteria needed for Meaningful Use Stage 2. What's next? Based on my discussion with HITSC experts, I believe our work ahead includes:
Content
*Continued refinement of the Consolidated CDA implementation guides and tools to enhance semantic interoperability including consistent use of business names in "Green" over-the-wire standards.
*Standardizing DICOM image objects for image sharing and investigating other possible approaches. We'll review image transfer standards, image viewing standards, and image reporting standards.
*Simplifying the specification for quality measures to enhance consistency of implementation.
*Query Health - distributed queries that send questions to data instead of requiring consolidation of the data
Vocabulary
*Extending the quality measurement vocabularies to clinical summaries
*Finalizing a standardized lab ordering compendium
Transport
*Specifying how the metadata ANPRM be integrated into health exchange architectures
*Supporting additional NwHIN standards development (hearings about Exchange specification complexity, review/oversight of the S&I Framework projects on simplification of Exchange specifications). Further defining secure RESTful transport standards.
*Accelerating provider directory pilots (Microdata, RESTful query/response that separates the transaction layer from the schema) and rapidly disseminating lessons learned.
I look forward to our November meeting.
Friday, 21 October 2011
Cool Technology of the Week
In the aftermath of the Blackberry outage last week, many people have asked me about iPhone 4S and Droid devices. As I wrote about yesterday, I chose the iPhone 4S to achieve a consistent user experience between my phone and my Macbook Air laptop.
Many of my staff use Android-based devices and are very happy with application availability and phone performance.
I polled my staff and asked them to identify the best Droids available today. Here's what they said:
1. Droid Charge by Samsung
2. Droid Bionic by Motorola
3. Droid Incredible 2 by HTC
Also worth mentioning is the Galaxy S II by Samsung
I welcome your comments about these 4 devices. My 500 staff have spoken and believe these are the best Android phones on the market - that's cool!
Many of my staff use Android-based devices and are very happy with application availability and phone performance.
I polled my staff and asked them to identify the best Droids available today. Here's what they said:
1. Droid Charge by Samsung
2. Droid Bionic by Motorola
3. Droid Incredible 2 by HTC
Also worth mentioning is the Galaxy S II by Samsung
I welcome your comments about these 4 devices. My 500 staff have spoken and believe these are the best Android phones on the market - that's cool!
Subscribe to:
Posts (Atom)
