Primetime Medical Software

Learn More

Optimizing Clinical Documentation

« Back to: Articles

One of our consultants recently wrote an article that appeared in the Group Practice Journal. The full text is below:

by Joe Weber

Clinical documentation is the core of an electronic health record (EHR). It’s also the major obstacle to acceptance. Physicians love to dictate and then let a transcriptionist turn their voice into a document. That’s what they’ve been doing for years, and they’re not the least bit interested in making a change. However, this documentation approach results in a dangerous delay in report availability, as well as a substantial transcription expense.

EHRs pride themselves on selective access to data, and free text is not really data. So that’s another point against dictation. Thus, most EHRs have been rolled out with an alternative documentation method, and it happens to be at the very opposite pole from dictation. Physicians are expected to perform direct data entry into an EHR--via pointing and clicking. Not surprisingly, they’ve generally balked at this, asserting that patient conditions can not be fully placed into little boxes. They also claim that direct data entry will take them considerably more time, and most evaluated experiences have borne out this claim. EHR vendors have had limited marketing success to date, in large part because physicians are so concerned about becoming less productive. And we can’t blame physicians for being concerned about that. So what’s the answer?

There is a hybrid approach to clinical documentation that leverages the strengths of available technologies, actually increases physician productivity, and provides other benefits as well. This hybrid approach utilizes three components:

1. Patient-delivered history
2. Physician-friendly EHR
3. Speech recognition

The Patient’s Contribution

Patients are the least expensive, most highly incented, and most underutilized resource in the entire health care system. The time has come to start using them (or allowing them) to contribute to their own care and documentation.

The hybrid approach would leverage this valuable resource. Whenever patients perceive a medical problem, they would go to a Web site, enter their problem, and then answer all the questions that medical science would like to have answered as the subjective contribution to the eventual clinical encounter--i.e., symptoms and other relevant medical history. Unlike traditional paper history forms, the set of questions presented would depend upon the patient’s answers. Some conditions will require just a small amount of questioning. Complex conditions will result in the channeling of questions more extensively, in order to drill down to the necessary data. This requires a knowledge base of tens of thousands of questions, compiled by intelligent brute force, presented through rather sophisticated, patient-friendly software with smart-branching technology. The history-taking Web site can either be a dedicated site for this purpose or it can be incorporated into the physician’s own Web site, likely tied to the appointment-request process.

Starting with Warner Slack’s landmark article back in 1972,1 we’ve known that computer-based patient interviews provide better histories than relying entirely on time-pressured physician/patient interactions. There is no way that any physician can learn, remember, and recall all the knowledge necessary to elicit all the patient answers that medical science knows are necessary for an accurate assessment. It is time we implemented the workflow efficiencies that have proven to have substantial health care benefits.

Once all the questions are answered, the positive and negative responses are summarized and organized. They can be displayed for the patient. But they are also delivered to the physician’s EHR, populating the subjective portion of the database with these structured/codified results. That will provide the physicians with a big head start when they eventually get to the encounter. But there’s now the option to engage in another process prior to that encounter, and it’s potentially the most powerful weapon we have to rationalize and streamline health care delivery.

That process is triage. Since we have a comprehensive and organized description of the patient’s problem(s), we can immediately take some efficiency-oriented actions. The urgency of the patient’s needs can be assessed and the required encounter time estimated, so that the appointment can be appropriately scheduled. For some conditions, self care can be prescribed, either in lieu of the encounter or in advance of the encounter. If the patient’s condition clearly requires a specialist, the gatekeeper visit can be bypassed. Plus, the indicated lab tests and/or X-rays can be requisitioned, so that they are available at the time of the encounter. Triage, as was stated above, can have a profound impact on the overall cost-effectiveness of health care.

The Physician Enters Clinical Data

Now let’s fast forward to the encounter. The history part of an encounter, although it varies considerably, averages about half the total encounter time. That means that we can chop encounter times in half and see twice as many patients a day. Well, not really. It’s highly likely that the physician will want to acquire elaboration on some of the results presented by the patient-delivered history. But this will consume far less time than taking a history from scratch, asking all those questions that mostly just get answered in the negative. So probably two to three minutes will be saved at each encounter. That means several more appointments each day and/or more quality time with the patient.

Now the physician moves into the objective, assessment, and plan. The EHR will be able to handle most of this, since these data fit rather well into the little boxes. But there needs to be considerable flexibility in the manner in which data can be entered: templates, macros, pick lists, etc. The entry method can be mouse, pen, or voice. The device can be desktop PC, laptop, Tablet PC, or PDA—potentially wireless. This flexibility will accommodate the preferences of individual physicians, but will also enable the use of whatever approach works most efficiently for the various data entry requirements.

Handling the Free Text

Not all the data fit perfectly into the little boxes. There needs to be an “escape hatch,” allowing the physician the freedom of expression and semantic richness that only free text can provide. Dictation is the best course for this narrative input. The most cost-effective mode is for speech recognition technology to immediately display the text on the screen, positioning the physician to correct any mistakes by voice or keyboard, and then instantly authenticate. If this is unpalatable for the physician, the draft report can be edited by someone else, or it can even be conventionally transcribed. But, in any case, we are now talking about maybe 10 percent of what would have been dictated if the entire report was done this way. That is an enormous reduction in either the cost for transcription/editing or in the time required for the dictator to make corrections.

Conclusion

The bottom line is substantially increased speed and productivity. The entire report is finalized in far less physician time than either the current standard of complete dictation or the full point-and-click entry of a typical EHR. Plus there are other benefits. Users of this approach can achieve the following:

• See more patients each day
• Reduce dictation and data entry time
• Eliminate transcription expense
• Populate EHR data fields automatically
• Optimize and support appropriate levels of billing
• Enhance patient communication and satisfaction

Additionally, all the key clinical data are entered in structured/codified form, and there is consistency in the symptom-based inputs for each chief complaint. This enables the kind of data-mining and clinical practice research that will ultimately advance the science of medicine. What we learn from analyzing these data can then be infused into an empirically developed clinical guidance system at the point of care, advancing the practice of medicine, and enabling health care to be all it can be. In the meantime, however, we can immediately and dramatically increase physician productivity and the cost effectiveness of patient care.

References

1. W.V. Slack and C.W. Slack. 1972. Patient-Computer Dialogue. New England Journal of Medicine 286:1304-1309.

Joe Weber is CEO of Narratek, Inc., provider of clinical documentation solutions, including speech recognition, dictation, and transcription. He can be reached at joeweber@alum.mit.edu.