Simply put, data analysis is the process of collecting information or data and interpreting it to find useful insights. The data processed reveals patterns that allow us to better understand the habits of members or providers, thus helping teams either take proactive steps to help our clients and their members, or to better assist them and advise them. There are three types of data analytics that are useful to companies – descriptive, predictive, and prescriptive. Essentially, looking at what happened, what could happen and what should happen.

Predictive analytics

Predictive analytics are an advanced form of data forecasting. The key is to collect and analyse enough quality data to be able to employ it in forecasting and modelling, using machine learning tools and techniques. Combining these with highly skilled team members is key to providing clear insights.

Working closely with a strong case management team ensures optimised cost prevention

Historical claims data can be used to forecast future events or trends in patient care, while directional care tools can facilitate the steerage of patients to a level of care that best fits their needs. This practice uses predictive analytics, and it combines methods for benchmarking to ensure a reliable system for referring patients to the best quality, most cost-effective services adapted to the type of care required by the patient (i.e. best patient-provider match). We refer to this as providing the right care, at the right location, at the right time and for the right cost.

Working closely with a strong case management team ensures optimised cost prevention, mitigation and containment before the services even occur, as patients are directed to the best solutions for their needs, whether that is telemedicine, home or hotel visits, walk-in clinics, or hospital settings – based on the patient’s optimal level of care. 

When a member requires on-site care at a clinic or hospital, they can then be referred to the most appropriate location based on a combination of quality and cost using appropriate benchmarking tools. This leads to an overall decrease in healthcare spend for insurers, without negatively impacting the quality of services for their members. Being able to compare costs using historical and industry data is important to be able to estimate costs and steer patients to the best location for their treatment.

How exactly do companies manage to steer each individual patient to the most cost-effective healthcare solution?

Predictive analytics help quite a bit, but we also suggest using descriptive data, which helps data mining and forecast experts in developing the latest directional care tools. Identifying high-cost claimants and high-risk patients through trigger diagnoses is also essential for both quality and cost control. This can be done based on known chronic conditions or trigger diagnoses lists. A health risk score can then be computed based, for example, on claims for behavioural health, substance abuse, ESRD, cancer, transplant, etc., and compliance issues affecting treatment. The score can also identify the risk of readmission post-discharge.

Finally, we suggest leveraging non-medical factors that influence health outcomes. One example would be using Social Determinants of Healthcare (SDOH) codes as a strategy for cost savings, which help reduce inpatient and outpatient utilisation by identifying and connecting members to the services they need, like health, housing, medical transportation, food programmes and financial assistance. Research shows that addressing SDOH can be a key factor in contributing to better care and lower costs.

Analytics and provider profiles

How is it possible establish which healthcare provider is a match for each member?

This is challenging, and there are many variables that need to be tracked. Experience is key since it can influence the outcome of an analysis. In the case of providers, market trends indicate engagement in provider-centric analytics based on peer comparisons that show outlier detections, such as detecting billing from high-risk addresses, developing scoring models to detect abusive billing patterns, establishing a quality rating based on hospital acquired conditions and quality of care issues, which can be tracked and used as renegotiation leverage. A further practice is to add a tracking diversity score, which considers trends and patterns on high-volume claims with emergency room upcoding, weekend work, etc., to facilitate billing the same diagnosis and current procedural terminology codes for all patients. These metrics make it possible to accurately match a healthcare provider to a patient based on the provider’s strengths and the patient’s needs. Ultimately, the right tools, combined with human experience and knowledge, can help provide the best solution, both for members and for providers.

In a nutshell, FWA addresses using a number of different methodologies: by monitoring external databases for expired or invalid provider identifiers, by leveraging regional and international sanctions lists, by identifying providers with invalid, suspicious or shared addresses or phone numbers with known suspect providers, and by flagging services to members that reside at the same address, or that have been approved on holidays, etc. This is a sensitive matter across many regions in the world; predictive analytics – and the systems we use – play an important role in identifying fraud issues, but also in helping better prepare subrogation teams for prompt and decisive action.

Benchmarking

Does benchmarking really work in cost containment strategies? If so, how? What does it do for members?

Yes, it does, but we recommend consulting multiple benchmarks in the US healthcare industry, which is critical to proper healthcare cost assessment and the value of the care provided. Looking at healthcare costs from different angles ensures that the correct benchmark is utilised to make an assessment for each provider, service, and location.

Ultimately, benchmarking helps reduce net paid claims costs – which benefits both the insurer and their members.

Both the provider and payer need to ensure that the reimbursement is reasonable and appropriate based on the services offered. Looking at a percentage of savings or a multiple of Medicare as a sole and unique method of analysing data could be quite misleading if you’re not considering a provider’s billing practices, location, codes, industry standards, provider costs and price inflation. The industry is ever changing, so remaining current and constantly re-evaluating the data driving these benchmarks is important in ensuring reasonable payment for services rendered.

Furthermore, reviewing trends on billing practices in the industry also helps spot areas of concern regarding the value of the care provided to patients (i.e. appropriate treatments, avoiding over-utilisation, coding concerns or errors, medical necessity, etc.). Combining the data with appropriate and recognised guidelines assists in ensuring both value and cost of care are considered in a cost containment strategy.

Ultimately, benchmarking helps reduce net paid claims costs – which benefits both the insurer and their members.