Healthcare today and in the future will be driven by data, analytics and digital transformation to support decision making, planning, commissioning and risk management.
Electronic health records are generated from multiple sources; appointment booking systems to physiological measurement systems, from real-time vital signs monitoring systems to imaging systems, all consolidating into a rich longitudinal data. The longitudinal data is high volume but low value in its raw state. Analysis unlocks this value by providing context—which is Information. Information enables the right questions to ask, what and where to look. Advance analytics provides insights—insight unveils the ‘why’. Why are the numbers what they are? Understanding the ‘why’ informs recommendations. Recommendation unlike data is distilled, low volume, high value, should be relevant, actionable and measurable.
Data driven digital transformation enables visibility and utilisation of the data, information, insight and recommendation.
Garbage in, garbage out
Analytics unlocks the value of data. The value is in the benefit it provides to the patients, the healthcare service providers, Clinical Commissioning Groups (CCGs) and regulatory bodies. But without good quality data, the associated analytics and resulting decisions made are most likely to be flawed. Data analytics is only as good as the data feeding it.
So, it is essential to get your data right and getting it right as early as possible, ideally during the data creation process. But what we see in practice is the cleansing of data within the analytics process often consumes over half of the process time. Repeated resolution of data issues during the analytics process through data cleansing and transformation tasks is therefore inefficient. Business needs change overtime; though the resulting analytics work often points to the same data sources. If data is corrected at source, it will save time and this is achieved through good data management.
At VIQTOR DAVIS we have developed a methodology to drive good data management called the ‘Data Delta’, which is articulated in our book ‘Crossing The Data Delta’ and available as a free download from our website. The data delta refers to the gap between the data you have and the businesspeople who need information from the data. To cross the data delta 6 principles are applied:
- Data must be governed and owned. We have seen scenarios in hospitals where data could be collected with no data governance oversight. This may be as a result of suboptimal appreciation for data governance amongst some staff or simply, awareness issues.
- Data must be described accurately and consensually. In some projects we have found clients with little or no documentation for their data assets and processes.
- Data quality must be defined, measured and managed. “You can’t improve what you don’t measure”. Crossing the data delta promotes continuous improvement.
- Principle of access must be established and enforced.
- How data is used and shared must be agreed and enforced using the Caldicott principles as baseline.
- The data warehousing, business intelligence and analytics domain is one of the 14 domains of the data delta methodology and is a potential medium for unintended data leaks or secondary incurred data quality depreciation if not properly implemented and controlled.
Analytics has a mutual relationship with the other 14 domains—It is enabled by the other domains while the other domains could benefit from analytics feedback. That is why VQD Analytics has developed a Data Delta informed method and process for analytics that encourages tech and technique from across all domains to be embedded into an analytics pipeline that delivers truly trustworthy insights in a time efficient way. We call the method Responsible Enterprise Analytics (REA).
Everybody! Including patients, primary and secondary care providers, CCGs, CQC, NHS England, NHS Improvement, NHS Resolution, pharmaceuticals, the Department of Health, clinical research organisations and beyond.
How do they benefit? Generally, benefits include cost efficiency and quality outcomes. Specifically, this could be:
- Improved treatment outcomes through data driven pathway redesign across value-based care continuums.
- Improved overall patient experience by reducing wait-times, personalising patient care and getting it right the first time.
- Cost and time efficient clinical trials through effective candidate selection processes.
- Safe staffing through effective demand and capacity management.
- Increased cost efficiencies through appropriate cost-value and forecasting models
- Improved Patient safety through the reduction of hospital acquired conditions such as MRSA and C. difficile.
Though some healthcare organisations are somewhere midway in their journey to analytics maturity, generally we can see a strategic push in the direction of achieving full maturity capabilities. There is a change in paradigm—such as: from presentation of numbers and trends to narratives describing why the numbers are what they are and predicting the likelihood of a future outcome, diagnosis or event. This provide capabilities for a step change from reactive to proactive decision making. The relevance of this type of capability is increasingly becoming ubiquitous across the healthcare ecosystem from clinical to operational management functions. Broadly speaking, healthcare analytics is currently focused on these two functions.
Clinical analytics directly supports the delivery of effective, cost efficient clinical outcomes for patients. This includes preventive methods and managing patients’ conditions during and after diagnosis or treatment. Use cases include:
- Genomics analytics for precision medicine, therapies and discovery of new drugs. This is one of many handles nudging us towards personalised care.
- The use of analytics to assist in spotting easily missed features in specimens of medical imaging, speeding up the diagnosis process and patient’s journey through the healthcare system.
- Proactive clinical intervention for preventive care using analytics to provide risk stratification of patients based on bio-markers and comorbidities. This improves the potentials for better treatment success rate and overall cost reduction.
- A patient condition could suddenly deteriorate. This could be due to infections such as sepsis, MRSA or C. difficile while in hospital or triggered by an existing condition. Analytics can help predict deterioration in a patient’s condition long before symptoms become visible.
Operational management analytics directly supports continuous efficiency improvements in how the various moving functional parts within the organisation work together in unison to achieve set goals, objectives and targets such as the Accident and Emergency 4 hours wait, 18 weeks RTT, Outpatients appointment-day wait, First to Follow-up ratio and excess Length of Stay. Use cases include:
- 30-day Emergency Hospital Readmissions: This could be used as a proxy measure for poor patient outcomes and could incur penalties on the service provider. Identifying the cohort of patients with a high risk of readmission could enable a targeted and tailored pathway for the cohort with the aim of reducing the chances of readmission.
- Did Not Attend (DNA): In England, has remained above 8% since 2008. Though this may vary at hospital level, it cost the tax payer millions of pounds (£) per year for both First and Follow-up Attendance DNAs. The wasted time slots could have been allocated to other patients on the waiting list. Also, patients could put themselves at risk by not attending their appointments as this could expose them to further complications with their condition. Knowing patients with a high likelihood of missing their appointments could help in curating specific reminder messages to patients or arrange an alternative medium for consultation such as by telephone.
- Staff Turnover: The NHS employs over a million staff. There has been continuous growth in the FTE since May 2013. However, staff turnover rate is a challenge. Retention is advantageous especially for business continuity, and reduced acquisition cost and training time. Analytics can help in identifying determinants behind staff attrition, knowing more about why individuals may intend to leave.
- Demand Pressures: Apart from the winter pressures there are everyday patterns in patients demand on the NHS resources. For example demand patterns on admission beds through A&E based on factors such as time, patients’ demography, requested tests, existing or diagnosed conditions. Demand patterns on diagnostic resources through direct access referrals. Understanding the intricacies of the demand, its nature and pattern will help in the proactive management of capacity to ensure efficient patient flow, safe staffing, optimum resource utilization, cost efficiencies and reduce suboptimal outcomes.
- Clinical Trials: Could cost millions of pounds (£) and about 10 years to introduce a new drug into the market. Though it cost this much and takes so long for a new drug, clinical trial failure rates are high and this as a result of a number of issues such as the candidate selection and monitoring process. Analytics can help in accelerating the selection process, identifying the most appropriate cohort of patients for a trial.
- Patient Adherence: Taking medicine as prescribed is important for successful treatment of temporary conditions, managing chronic conditions and overall well-being. Analytics can identifying patients with a high likelihood of not adhering to their therapy. This kind of information provides the opportunity to proactively trigger procedures to ensure adherence.
- Vaccination and Screening Uptake: Patients who miss out on their vaccination may remain vulnerable to serious or even fatal infections that are vaccine-preventable. Screening such as cervical screening test are critical to identifying cancer in its early stages, which directly correlates to the chances of survival. Analytics can point out patients at risk of missing their vaccination or screening test. This can help in enabling the targeting of these patients with appropriate intervention.
Ethics, patient-centred care and analytics
Analytics is powerful and can be used for good but can also cause harm if not properly regulated. Currently there is no clear government legislation sufficiently covering all aspects of analytics. Existing regulations such as the European Union’s General Data Protection Regulation (GDPR) requires service providers to be transparent in how data is collected, used and what for, but this is open to interpretation and as a result, exposed to potential ethical issues.
Today we have all become human sensors, generating data both patients and staff alike. Patients sign-in on arrival generating arrivals data and clinicians balancing their time between patient care and data capture during a consultation—this can be quite challenging! Both patient care and data capture are fundamentally expected to be patient centred and ethical, that is, providing good data-driven decisions and excellent patient care.
Analytics is sometimes thought to deliver just the facts—but this is not always true—as there may be all kinds of biases from data capture, analytics through to the decisions based on them. Biases if not properly vetted and managed can put to risk the patient centred principle of choice or the ethics principle of respect.
There is the issue of transfer of responsibility. This is where a decision maker solely and unquestionably relies on an analytics service or outcome. There is the tendency that this kind of decision maker would shifts accountability away from themselves to the analytics service. This could impact on the ethics principle of do no harm or the patient centred principle of compassion.
At VIQTOR DAVIS we address these ethics and patient centred imperatives through our FACT framework. It measures 4 key drivers: Fairness, accountability, confidentiality and transparency.
We introduce control-specifications agreed with stakeholders through a comprehensive assessment to reduce risks before deployment. Our analytic products are ethical and aligned to the principles of patient centred care. We promote our analytic products as tools to argument domain experts in their decision making not as masters. We foster transparency and clearly defined points of human intervention. We also promote the importance of industry self-regulation and support the established of fit for purpose government regulation covering all aspects of analytics.
Our current focus within healthcare is in the Operational Management domain, where we have developed capabilities to deliver accelerated and reliable insights through sustainable, ethical, trustworthy and patient centred solutions. Talk to us today about how we can help.