This post is the second in our two-part natural language processing series, which will delve into Doctor.com’s proprietary NLP software used to help track and analyze the language throughout all of an organization’s patient comments. You can read the first post here.

Natural language processing isn’t exactly new, but a completely healthcare-centered approach to it with regards to patient reviews definitely is. At Doctor.com, we harness the power of natural language processing technology to help you make sense of all the patient reviews you’ve been collecting and the gold mine of information they contain.

Our unique healthcare-centered approach

Complicated medical jargon, HIPPA concerns, administrative complexities, and patients leaving comments about serious life-or-death medical situations all add staggering complexity to the review process — much different from the reviews for a restaurant or car repair shop, for example.

Sentiment analysis is always complex, but it’s particularly challenging in the healthcare industry when it comes to reviews. Some comments come from government-mandated surveys like CAHPS, which are often gathered on paper and transcribed — sometimes missing words that could not be read or containing other transcription errors. Often comments are full of spelling errors and sentence fragments, as the healthcare industry sees interactions with people from every background.

For example, comments often contain details of a patient’s treatment or their condition that are irrelevant to the sentiment relating to their doctor or the care she is providing. A patient may have cancer, which sounds both frightening and negative, but the care received may be absolutely stellar.

How does NLP help with making sense of patient reviews?

For healthcare systems and hospitals hoping to monitor their reputations, NLP makes possible the streamlining of an incredible amount of data. Just think, for a hospital with 200 doctors — if each provider gets only 20 reviews monthly through internal and external feedback, that’s 4,000 comments every month! With sentiment analysis, NLP can help by automatically identifying the feeling, opinion or belief of a statement, all on a Likert scale from “very positive” to “very negative.” Additionally, the comments can be bucketed into categories, which help to hone in with greater precision on problems and bright spots.

The common sentiment libraries other vendors use for building their NLP engines are not built for healthcare, because they are serving multiple industries. However, this becomes a problem in healthcare because not having an industry-specific tool can cause the results of the analysis to be compromised. You’ve spent good money investing in your patient reviews (through CAHPS surveys and any transparency initiatives) as well as extensive time culling through third-party reviews — make sure you’re extracting as much as value from them as you can.

Visit Doctor.com/enterprise to learn more about solutions.