Big data drawn from bad
data makes big, bad data
The
abundance of data and the technology used to capture and understand it is
causing powerful shifts in the insurance industry. Insuretech is the term used
to describe the use of technology
innovations designed to promote accuracy and efficiency while reducing costs
from the current insurance industry model. Now nimble insuretech organizations
are disrupting the insurance market by challenging the status quo.
However, organizations can step ahead of the oncoming disruption by following the common-sense steps noted in the article, “How to Avoid Insuretech Disruption”.[1] The first important component of this effort is accurate data. Data should be viewed as a significant organizational asset that is respected, valued, and protected at its collection points and throughout its useful life.
However, organizations can step ahead of the oncoming disruption by following the common-sense steps noted in the article, “How to Avoid Insuretech Disruption”.[1] The first important component of this effort is accurate data. Data should be viewed as a significant organizational asset that is respected, valued, and protected at its collection points and throughout its useful life.
No longer will incomplete and erroneous data be
tolerated because little can be done to actualize the benefits of technology
and prevent disruption without addressing the issue of data integrity. Organizations
can confront their data issues by addressing the essential elements of collection,
correction, and accountability.
Collection
Examine the data. Data is frequently
(usually) incomplete or erroneous. Begin by evaluating the data collected by the organization, identifying its various sources.
Usually, an organization has multiple data sources.
Bill review is a good example. Consider the
fact that the bill review company collects the data from medical provider
bills, often by means of digital scanning, called OCR (optical character
recognition) that converts the hard copy of the bill to digital data. While OCR
is considered fairly accurate, it makes sense to go upstream to evaluate the data
input source.
Medical providers generate the bills and now
most are able to produce and deliver digital bills. That eliminates the need
for scanning, however, for the provider to produce bills, manual data entry is
required. This is where essential information is omitted. Often the treating
location on the bill is listed as a PO Box, while the specialty of the treating
physician or other professionals are omitted, as well as individual NPI numbers
(National Provider Identification).
Correction
Inaccurate and incomplete data from providers ranges from deliberate fraud to simple negligence. Data integrity is the basis for accurate provider performance evaluation and selection, essential to quality care and accurate claims adjudication. Source networks should require accuracy from their provider members or correct the data at their point of data progression. Payers, the purchasers of networks and bill review services, can use their leverage to encourage data correction at these points, as well.
Keep in mind manually entered data, is the least reliable method of data collection. Misspelled and abbreviated names and addresses create duplicate entities in the data, thereby skewing downstream analysis. Because the data entry source is not arms-length for payers, they will need to correct the data themselves.
Claim data entry is an internal manual data entry process for payers. Years ago, typists were tested and scored for how fast they could type a paragraph without errors. Accurate typing performance determined hiring or dismissal. Today’s data entry personnel should be similarly evaluated for accuracy, regardless of their level in the organization. Make accuracy a performance measurement for everyone.
Manually correcting the data is a challenging prospect, but the benefits to downstream analysis are enormous. Programmatic systems are available to merge duplicate records, for instance, and they are accurate to a point. Also, systems can be created to highlight data needing manual attention to facilitate data corrections.
Accountability
Inform and train employees regarding the corporate value of data integrity. Design systems that highlight errors and omissions for work in progress. They can also be used as accountability tools. A simple audit trail will identify the error perpetrators going forward. Moreover, objective feedback is the best way to improvement.
For the last thirty years, the big push in the insurance industry has been collecting data while little emphasis was placed on accuracy. Now the focus has shifted. Predictive analytics relies on historic data to inform so be sure the data is as robust as possible.
MedMetrics,
a WC insuretech company, leverages predictive analytics to provide WC claims
reps knowledge assistance by projecting probable ultimate medical losses,
automatically integrating claims and medical management to contain predicted medical
threats, and offering additional intelligent tools such as provider performance
analysis and diagnostic severity scoring. MedMetrics clients are easily and
affordably more accurate, efficient, and profitable.
[1]
Wolfe, K. How to Avoid Insuretech Disruption. http://medmetrics.blogspot.com/2017/07/how-to-avoid-insuretech-disruption.html
