In a world of fleeting fancies, the pursuit of medical coding automation seems akin to the search for the elixir of life; healing all infirmities and maladies – past, present, and future. Medical coding automation is an extremely strenuous, systematic and time-consuming process. If not executed intelligently, most initiatives lead to disappointing ends.

But what accounts for these failures?

Insufficient planning

‘Enough time’ and ‘inadequate time’ are concepts that walk the scalpel’s edge. Thinking through and planning with key stakeholders, is an important precursor to automation. Spending ‘enough’ time in the planning phase is critical to the automation output, on expected lines. Not involving key stakeholders in the initial discussions, and starting off the automation with little understanding of the process nuances, is an ingredient for heartburn.

Domain understanding, exception handling

Any automation in medical coding requires a high level of domain expertise. While it is common to designate a SPOC (single point of contact) for every automation initiative, the individual’s domain expertise alone is often insufficient to articulate domain requirements in a manner understandable to software programmers. Many automation initiatives also fail because of inadequate understanding of the process exceptions. Often times, this results in more manual work than before.

Change resistance

Healthcare being a human capital-intensive business, the success of any automation is in the hands of the people who work around these processes. Our experience has shown that the resistance of team members to changing their existing process is a huge hurdle. The level of change depends on the level of automation, which in turn immediately impacts the way things are done. A great people principle to adopt here is: ‘get hit in the short,’ but ‘reap benefits in the long run’. Middle management can play a huge role in helping the team members overcome their fear of change.

Fill it, shut it, why can’t I forget it?

Believers in the maxim that process automation is a permanent fix, are generally in for rude surprises. The need to monitor and periodically test the output, cannot be adequately stressed. Since the environmental variables always keep changing, those envisaged during the automation process may / may not be applicable to the present situation. For example, a rule set on an ICD-9 code would have been based on rules applicable at the time. During the transition to ICD-10, if the rule is not revisited, the automation actioned is a failure. Constant calibration and tuning is therefore a must.

Unrealistic goals, high expectations

Most expectations around automation, presume fastfood delivery type of results, at the highest quality standards. This works perfectly if the team is given time. Unfortunately, automation programming goals are unrealistic and subject to a million other business pressures. They are often times set by people who’ve chosen to forget reality and pursue quick business results. Any automation requires repetitive learning and unlearning, before it delivers expected results. The biggest gain by waiting, is to arrive at a process that’s predictable, and will deliver similar results at different scales.

Giridharan T S is part of the core team at Quintessence Business Solutions & Services. He has over 18 years’ experience in the US healthcare domain, with proven technical capabilities in conceptualization, design and implementation of cutting edge software that have had a pioneering impact on the industry.