開放資料時代之健康及社會照顧

• THE OPEN DATA ERA IN HEALTH AND SOCIAL CARE
• 此為大綱: (若有人願意畫重點(copy&paste), 翻譯) 
• 可作為 health domain , open data persuasion for government 的藍圖  
• Source : http://thegovlab.org/nhs/    有pdf 

Forward (finish)

• Evidence suggests that giving patients access to their own healthcare records can increase healthy behavior and improve decision-making
• Giving providers access to comparative performance indicators across hospitals and physicians increases cost-efficiency in the healthcare system
• Giving researchers access to clinical data improves medical outcomes
• Giving the public-at-large access to nationwide prescription data and hospital performance increases choice, empowerment and accountability.
• NHS: to increase patient power, save lives, improve quality of treatment
  • Transparency
    • The safe sharing of data and information between clinicians, patients, and the public
  • Participation
    • Supporting patients and citizens to take more control of their health and care and fully engage in the design of local services
  • Interoperability
    • The development of seamless digital records across all care settings, based on open standards

EXECUTIVE SUMMARY(finish)

• Open data:  is publicly available data that can be universally and readily accessed, used, and redistributed free of charge.
• Change: the way governments, nonprofits, and the private sector use data to understand public issues and solve problems in areas as diverse as financial regulation, energy, education, and more.
• Review case studies and published research to highlight the following value for using more open data in healthcare: 
  • Accountability : healthcare organizations
  • Choice: patients make informed choices
  • Efficiency : improving efficiency and cost-effectiveness of healthcare delivery
  • Outcomes : improving treatment outcome
  • Patient satisfaction and customer service : educate patients and family to make healthcare institutions more responsive
  • Economic growth and innovation
• Central Goal of this paper : 
  • To establish a conceptual framework, or logic model, that can be used by researchers and programme managers to design their open data initiatives and then measure their impact
  • Assess risks , challenges and potential economic and social benefits
  • Establish ODLE
• Open data vs Big data
• 
  
• 
  

Introduction(finish)

• 舉例:
• Mastodon C (Open Data Institute in Shoreditch) : 
  • Do: Analysis massive amounts of data on prescription patterns released by the NHS.
  • Results: significant geographic variations across England
  • Difference: expensive prescription drugs being prescribed in some areas when generics would work just as well
  • Impact: if change prescription pattern  => Saves $ 1 billion GPB/year
• Society for Cardiothoracic Surgeons:
  • Do: publishing the results of surgeries done by individual physicians across the UK
  • Found: Surgical outcomes improved rapidly
  • Impact : recording/reporting/publishing outcomes => improve quality of care
• NHS: 
  • Do: friends and family” study : ask patient/family to recommend hospital or care center.  Then publish the ratings
  • Difference : no rating data => rating data available
  • Impact: Give consumers more information and choice , improving care provider quality
  • 
    
• UK experience  :  
  • Healthcare is one of the sectors with the potential to be most radically transformed by greater data accessibility.
  • the NHS has invested substantial resources (financial and intellectual) into developing a comprehensive open data strategy
  • Information concerning GP prescribing, mortality rates, waiting times, and a wide variety of other data has been released to the public and researchers
  • An incipient ecology of innovation, analysis and research has begun to emerge as a result of such initiatives
• Philosophical (or ethical) argument for open data
  • 1. information whose collection is paid for by taxpayers should be available to those taxpayers
  • 2. personal information about individuals (e.g., medical data) should likewise be accessible to those individuals
  • impact: 
    •  it facilitates third parties to use it in ways that create new analyses, visualizations, and mash ups
    • Solving complex challenges requires many people with diverse skills and talents to work together
    • When experts of all kinds have access to open data, it becomes a catalyst for creative problem solving and community innovation
    • By making data open, you enable others to bring fresh perspectives, insights, and additional resources to your data, and that’s when it can become really valuable.
    • the ambitious plan to shift an entire nation’s bureaucracy to more evidence-based decision-making

I. The open data era (OD as a driver of innovation)(finish)

 Health Data and the Open Data Revolution

 The NHS and Open Data: Where We Stand 

• Trend of big data and open data is related, but not identical !
• Bid data : 
  • data explosion : In late 2012, it was estimated that there existed some 2.8 zettabytes (2.8 trillion GB) of data; this was expected to grow to 40ZB by 2020
  • value: €32 billion in 2010, with a 7% annual growth rate (EU)
• Open data : 
  • Open data is publicly available data that can be universally and readily accessed, used, and redistributed free of charge.
  • Open data is released in ways that protects private, personal or proprietary information
  • It is structured for usability and computability.
• Openness : 
  • easier to find data
  • present usable format
  • streamline terminology
  • readable, comprehensible by individual
  • published without proprietary condition
  • available online in a downloadable format
• Application:
  • After non-personally identifiable (through anonymization or pseudonymization)
    • land records
    • welfare schemes
    • political operations
    • fundraising
  • Without restrictions of copyright or other mechanisms of control
• Belief: The data transparency might prove to be a particularly revolutionary tool when it is applied to health information
• Challenge : concerning privacy and identity
• Health data classification in NHS : by accessibility ( from open to closed)
  • Open Health Data
    • Open Government License (UK)
    • available from data.gov.uk hscic.gov.uk & appendix V
  • Restricted Health Data
    • anonymized personal or proprietary information by NHS
  • Published Health Data
    • prepared for presentation (doc, pdf )
    • original data are not included 

• Data is the 21st century’s new raw material
• Data value 
  • holding governments to account
  • driving choice and improvements in public services
  • inspiring innovation and enterprise that spurs social and economic growth.
• The core of the government’s healthcare strategy: 
  • White Paper on Open Data and The Power of Information
  • The passage of the Health and Social Care Act of 2012
• Hallmark of the UK’s data transparency policies : Third-party innovation
  • data-driven businesses , Open Data Institute
• NHS Open Data Milestone: 
  • 1980 : Collect data as early as 1980
  • 2005 : NHS Information Centre as the authoritative repository of NHS health information
  • 2007 : "NHS Choices" launched : 
    • empowering patients to make decisions about their healthcare based on comparative data
    • Users were also able to share their experiences of using NHS services 
    • 這個健保署一定不敢做, 因為和財團利益衝突
  • 2010 : after election : give priority to data transparency
  • 2012 : 
    • White Paper on Open Data and The Power of Information 
    • The passage of the Health and Social Care Act of 2012
    • The Power of Information : 10-year-plan on IT strategy
  • 2013: CareConnect Pilot 
    • A phone and web service that allows citizens to interact with the NHS, get information, and provide real-time data through feedback and other mechanisms
  • 2014: MOU with " US Department of Health and Human Services"
• Looking forward :
  • Releasing more open data
    • 2013 publishing surgery outcome data for 10 major specialties
  • Expanding datasets to measure health outcomes
    • Transform :Hospital Episode Statistics (HES) => Care Episode Statistics (CES)
    • Linking hospital datasets to
      • primary care data, community health services data,clinical audit data, cancer registry data, social care data, mental health data, genomic data
  • Making data actionable for healthcare consumers
• All initiatives is to help the NHS achieve greater transparency and improve health outcomes

II. Potential and limitations of open data (the arguments in favor of using OD in a healthcare setting)

• For more transparency and accountable governance

Value Propositions for Using Open Health Data

• Proposition : OD can ensure health regulator and provider meeting standards
• Evidence : New York State Cardiac Surgery Reporting System ; Annals of Internal Medicine
• Increasing data transparency appears to have a particularly powerful role on the behavior of regulators.
• Clinical auditing can lead to changes in practice and greater adherence to published recommendations or guidelines
  • exp: prescription quality, patient management,reductions in patient harms,resource allocation
• Taxonomy: 
  • Political: Improved transparency and accountability in terms of reaching desired outcomes in public health.
  • Economic: Improved accountability and transparency in spending, cost, waste, and fraud
  • Participatory/Citizen Involvement: Increased public participation in healthcare services
  • Organizational/Clinical: Higher success rates with regard to internal NHS objectives on patient outcomes
• Example: 
  • Dr Foster Intelligence : benchmarking hospitals and healthcare commissioners. For instance,it used mortality datasets to improve accountability at the hospital level, which ultimately identified hospitals in the UK with abnormal high mortality rates and poor clinical practices as the causes. 
  • Better Procurement, Better Value, Better Care
  • IMS Health : uses open health and other data to provide a set of commercial health intelligence tools that allow their customers to assess healthcare performance and impact.
  • Dollars for Docs : whether their healthcare professional has received drug company money

• Proposition : To allow patients to seek out the services that best meet their individual healthcare needs ; increased access to health data, in turn driving patient satisfaction, quality improvements and the enhancement of standards in many areas within the NHS.
• Evidence: 
  • Medical information on the Internet plays a role in influencing patients’ decisions,in some cases an even bigger role than that played by their doctors. PMID:14517108
  • US: nearly half (48.6%) went to the Internet first for information on cancer (compared to 10.9% who went to their doctor) Ref:45
  • NHS: availability of medical data plays a significant role in influencing how patients seek care. Ref:47
  • Open data can be particularly helpful in aiding patients in the access of information that will allow them to evaluate different treatment programmes. Open health data in this area should support the guidelines outlined in the NHS Choice Framework. 
  • Results from a study assessing the effectiveness of choice programmes showed that quality assessments will become more important in influencing patient decisions and choices as quality-related information is made more readily available.
  • NHS’s way: facilitating shared decision making between patients and their physicians. 
  • Study pro  : 37% patients using NHS Choice Website: decrease in GP consultations & savings of approximately £44 million a year.
  • Study cons: information on internet may not always lead patients to make better (or more informed) medical decisions
  • Future ? : Shared decision-making. Such approaches combine data-driven and information-armed patients with the clinical judgment of doctors
• Taxonomy : Choice has at least two dimensions:
  • Variety: Increasing the amount of information available to patients for their decision-making
  • Quality: Increasing the quality, completeness, and timeliness of information provided to patients and the broader public to determine whether treatment is needed and/or available
• Examples
  • NHS Choices : combination of decision tools, open data on service providers and, increasingly, patient feedback to enable patients to compare and choose treatments and services
  • WiserTogether(US) : gathers open data through its Wiser Health Platform from consumers and doctors who have encountered similar medical issues and creates a list of best options based on clinical efficacy, financial considerations and treatment preferences. "choose the right care at the right time" " evidence-based, cost-effective treatment decisions" 
  • iTriage (US): provides patients with the information they need to make decisions about their symptoms, possible treatment methods, nearby care options, and medical providers. Content is produced by an in-house clinical team of physicians, as well as through a survey of published literature
  • MedWatcher : Using data from the Food and Drug Administration, MedWatcher maintains a database of adverse side effects reported from different drugs
  • Healthgrades : based upon open patient safety data, it compared healthcare providers is it likely to survive their hospitalization or encounter the least risk of major complications.

• Proposition : more cost-effective, lowering costs, reducing fraud, increasing productivity, accelerate data-sharing practices and communication strategies, and will improve care quality
• Evidence : 
  • GP prescriptions for generic and patented statins in England
  • Greater use of open data can also help control fraud : find cases of over-billing or fraudulent billing
• Taxonomy: 
  • Operational: Improving the speed at which patients are processed, treated, and discharged.
  • Economic: Lowering the costs associated with each hospital or service, for example by comparing open data across regional and local contexts.
  • Communication and Technology: Identifying spots in the healthcare system where communication needs to be improved both internally and externally (i.e., by reaching the broader public).
  • Treatment: Determining which treatments work effectively, and for which populations
  • Resource allocation: Determining where and when services are needed most.
• Examples: 
  • The Camden Health Metrics Explorer(US): open datasets linked to health exchange information and claims data, providing actionable metrics in real time; identify the patients who are heavy users of the healthcare system
  • Aidin (US): lists and assesses specialized- care providers to simplify the patient-discharge process, and is aimed at preventing overspending. Aidin integrates into the discharge planning workflow to free social workers from administrative tasks and re-center their time around patients
  • CDEC Open Health Data Platform:  facilitate innovation in the field of data analysis, data visualization, service design, and other web and app development enabling innovation using linked health data

• Proposition : drives competition between healthcare professionals => spur quality improvements and innovation
• Evidence: 
  • Early evidence: mixed => Now : increased data transparency can lead to improved outcomes
  • New York State Cardiac Surgery Reporting System
  • Germany and Sweden study
  • Preventive healthcare : allows individuals to compare their health status with demographic information contained in public databases (NIH , CDC)

Potential Challenges and Barriers of Open data

• Culture and Institutional Barriers
• Privacy
• Standards and Interoperability
• Good Analysis

III. Toward a conceptual framework for open data in healthcare

• The use of certain kinds of inputs and data (INPUTS: OPEN DATA), by certain kinds of users (USERS), for certain kinds of activities (ACTIVITIES), will achieve certain outputs (OUTPUTS) and outcomes (INDICATORS) that indicate impact (IMPACT). Specific methodologies (METHODOLOGIES) will be used to collect and measure indicators, helping to assess impact
• 

• Inputs : what open data set was used ?
  • Internal Data ( more used by researchers, agencies, watchdog organizations )
    • Financial Data : cost and spending 
    • Administrative Data : 
    • Statistical/Diagnostic Data : statistics on chronic disease rates, preventable infections, and a wide range of other topics. HSCIC.gov.uk’ included data quality, hospital care, illnesses and conditions, mental health, patient experience, prescribing, primary care services, public health, social care, workforce data.
    • Audits
  • STREAMS OF HEALTH DATA
    • Open Health Data on Practitioners: Examples include doctors, administrators, midwives, nurses, etc.
    • Open Health Data on the System/Infrastructure: Examples include hospitals, clinics, equipment, and waiting times, etc.
    • Open Health Data on Patients: Examples include diagnoses, prescribing rates, mortality rates, treatment options, etc.
• Users : Who is using the data ? 
  • Different users of data may have different criteria for determining the success of an activity or programme.
  • Internal Users : who work within the health system
  • External Users : from a position that is partly or completely external to the NHS
• Activities: How is data used ?
  • Open data will be used for different purposes, depending on the position and intention of the user accessing the information
  • The primary uses of open data will be used to improve management practices within the NHS, as well as to improve policies geared towards public services.
  • The internal user intends to improve efficiency within the NHS; the external user seeks to increase accountability.
• Outputs: What was the value produced by the open data initiative(activity)?
  • For instance, an administrator using open data to assess the rate of preventable in- hospital disease would consider the output to be the difference between the expected rate of preventable disease and the actual reported result. 
  • Number of additional patients served
  • Number of pounds saved
  • Number of patients using data tools for decision making
  • Percent decrease contraction of preventable in-hospital disease
• Indicators : To what end ?
• Methodologies to Measure impact : According to to what measure?
• Impact : To add what value to the NHS and the UK as a whole ?

IV. Building an open data learning environment (ODLE)(finish)

As much as the NHS might intend or desire to open all its data, there do of course exist practical constraints. An ODLE can help prioritize which data to make open, and in what

way.

• 我不太知道怎麼取捨QAQ pathway的部分我先全貼，再想想怎刪修比較好
• 沒關係, 我覺得看完比較知道怎樣跟gov 溝通和push; 以後有人要找參考資料, 這邊可以當base

 1) BUILD AN OPEN DATA LEARNING CAPACITY AND CULTURE WITHIN THE NHS.

Broad, comparative view will make it easier to determine what works and what doesn’t, and to use cross-programme insights to build a more effective institutional strategy. Measuring impact should not be an afterthought; it must be considered up front at the design stage of every project. 

• Develop an NHS Open Data “Do and Learn Tank.” Unlike a traditional think tank, this would be structured to help launch new projects using lessons from existing programmes and initiatives and evaluate their results.
• Set up an NHS Data Geek Squad to create a corps of volunteer data geeks and researchers from Britain’s best universities to work with open health data. This could be modeled on Datakind in the U.S., or on Code for America (although Code for America focuses its work city by city rather than by datasets).
• Connect research organizations already existing within the NHS, such as the NIHR.ac.uk, that are already examining the impact of open data.149
• Use prizes to stimulate new solutions to public problems using open health data, as the UK and U.S. governments have already begun doing.
• Develop campaigns designed to promote data sharing, and raise awareness about the release of open datasets, their locations online, and their potential uses by the public.
• Set up an open health data academy that trains people to use open health data and measure its impact, using online learning, project-focused instruction, and mentoring.
• Fund the creation of NHS open health data fellowships for students and graduates with compelling ideas and practical ways to implement them.
• Develop an open health data mentor network to encourage and train new recruits and younger members of the NHS in the use of open data. 

 2) ENGAGE THE PUBLIC IN DEFINING METRICS

The public also has valuable insights and lessons to contribute to this proposed ODLE. In addition, since all indicators and uses of data are inextricably linked with priorities and values, open data policies should reflect a larger social and political perspective. The NHS should therefore ensure that direct stakeholders (e.g., patients and providers) are included in decisions about open data and how it is used. 

• Develop a citizens’ open health data panel (similar to Hackney’s online citizen panel) to review metrics on a regular basis.
• Facilitate user-led design exercises to better understand how open data can support stakeholders’ work and ultimately improve people’s lives.
• Design and implement online mechanisms such as ratings and feedback tools to gauge public opinion and solicit insights from citizens.
• Create a “Metrics Bank” of important indicators, with input from stakeholders, researchers, and experts in the field who have studied the relevant literature.
• Prioritize the creation of an inventory of datasets by departments within the NHS, and invite experts to curate the datasets released to the public.

 3) STAY FOCUSED ON WHAT REALLY MATTERS.

Too often, indicators used to measure open data only quantify the level of use (e.g., the number of datasets used or money saved). The metrics that really matter, however, are those that are focused on improving people’s health and lives. 

• Develop NHS Open Data Stories that will allow stakeholders to share how open data improved people’s lives in real time.
• Use surveys, social media, and sentiment analysis to learn what dimensions of healthcare improvement are most important to the public, and ensure that success metrics and indicators capture those priorities.

 4) DEVELOP A COMMON ASSESSMENT FRAMEWORK

It is important to coordinate with other organizations, sectors and countries that are also designing analytical frameworks for the use of open data across sectors. 

• Set up an annual meeting/listserv/monthly hangouts on open health data research to trade best practices and ideas.
• Create a directory (perhaps in wiki format) of other assessment frameworks across countries and sectors. Such a directory would also include a list of key contacts and organizations.
• Use online and offline meet-ups and other approaches to create a culture that encourages knowledge sharing and collaboration with other organizations

 5) STAY FLEXIBLE AND ADAPTIVE IN MEASURING IMPACT.

NHS will need to find a balance between a [centralized measurement function] (which sets measurement standards, aggregates data across programmes, and monitors data

quality) and [a more diffused structure] that empowers users at different levels of the organization.

• Hold regular “What Works Camps” that connect various users and researchers.
• Ensure that all metrics (especially those to measure use and impact) are categorized by different levels and can be analyzed separately.
• Continue ongoing research into new and existing approaches to impact measurement. This research effort should be considered a core part of the NHS open data programme.

 6) SHARE WHAT IS LEARNED (INCLUDING FAILURES) WITH EVERYONE

• Develop visualizations of how open data has made an impact--for example, through maps that show changes in healthcare quality, efficiency, or cost in different geographical areas.
• Develop an Open Health Data 10 – a listing of the ten most impactful uses of open health data.
• Consider wikis, seminars and other means for stakeholders within the NHS to share stories and experiences with open data.

 7) BUILD A RESEARCH NETWORK AND EXPERT NETWORK

Build capacity quickly and broker debate around methodology and findings. 

• Hold an annual summit to exchange what has been learned, along with regular meet-ups.
• Set up an NHS Open Data listserv.
• Consider establishing an external advisory or consulting board of experts to whom the NHS can turn for advice and guidance. 

 8) DEVELOP OPEN HEALTH DATA ECOLOGY MAP

Including a dictionary of all open health data sets in the NHS along with the variety of uses and users. This map should include visualizations and other graphical representations to fully represent how open data is being accessed and used within the NHS.

• Develop a conceptual map of users, uses and impact.
• Develop a site inventorying all available datasets within the NHS, making it easy for users to access datasets from a variety of online spaces and departments.
• Develop a crowdsourced resource of innovative start-ups and new applications that use open health data.

 9) PUBLISH, INTEGRATE AND FINE-TUNE THE OPEN DATA CONCEPTUAL FRAMEWORK

 In order to bridge the gap between theory and practice, it is essential that various stakeholders participate in refining the framework. The framework must itself be responsive and flexible; its design must constantly be refined and implemented using the same level of transparency as embodied by the concept of open data.

• Develop an interactive version of the conceptual framework that can be annotated.
• Create an expert, online advisory network to vet and review the conceptual framework
• Create channels for feedback and review by various stakeholders.
• Research and evaluate similar frameworks used in other sectors or countries and build on insights or lessons learned.
• Build on the Memorandum of Understanding between the United States’ HHS, and the NHS, and include impact assessment as a joint activity undertaken by these two countries.

 10) ENGAGE STAKEHOLDERS IN SHAPING THE OPEN HEALTH DATA PROGRAMME

The NHS should engage key stakeholders on a regular basis to determine which datasets have the highest priority for them; what new datasets should be released as open data; and which open data collections are particularly easy or difficult to use. 

• Set up a wiki, forum, or combination of online tools for stakeholders to provide this feedback.
• Develop a subcommittee of the Open Data User Group to focus on health data specifically.
• Hold roundtables with different groups of stakeholders--health-related businesses, advocacy groups, and patient groups--to help shape government policy on the release of open health data.

V. Conclusion