What’s missing from the electronic health record

Dr. Straub is a third-year Maternal-Fetal Medicine Fellow in the Department of Obstetrics and Gynecology, NorthShore University HealthSystem, Evanston, Illinois.

Dr. Silver is a Maternal-Fetal Medicine Physician, Department Chair of Obstetrics and Gynecology, NorthShore University HealthSystem, Evanston, Illinois, and Clinical Professor and Associate Dean at the University of Chicago Pritzker School of Medicine. 

Neither author has a conflict of interest to disclose with respect to the content of this article. This work was funded through intramural support by the Department of Obstetrics and Gynecology at NorthShore University HealthSystem. The funding source had no role in the preparation, review, or approval of the manuscript, or the decision to submit the manuscript for publication.

A 37-year-old woman without a primary care provider comes to the emergency department (ED) for a recurrent severe yeast infection. An obstetrician-gynecologist is consulted and recommends screening for diabetes. The patient’s serum glucose is found to be 230 mg/dL, prompting a hemoglobin A1C to be sent, which returns as 8.1% after the patient is discharged. The emergency physician calls her home and leaves a message explaining the need for follow-up. However, the telephone encounter is not documented in the electronic health record (EHR) and an appropriate diagnosis (“Rule out diabetes”) is not added to the problem list.

The patient does not follow up as suggested but is seen for 2 subsequent medical visits in the same year; first for an atypical migraine that is treated by a neurologist and second, at an ED after a car accident for which she is discharged without any specific therapy. Both visits are documented using the same EHR.

Two years later at age 39, the patient presents to the ED with vaginal bleeding. She is found to be unexpectedly pregnant. An ultrasound reveals fetal malformations consistent with caudal regression, likely related to her untreated diabetes of at least
2 years’ duration. Her spot glucose is 332 mg/dL and her hemoglobin A1C is 9.6%. She receives counseling regarding the ultrasound findings, aggressive diabetes management as well as outpatient endocrine and obstetrical referrals.

Patient adherence and the EHR

This clinical vignette reflects the reality that only one-quarter to one-third of patients follow up as directed after ED visits.¹ Thus, dependence on patient adherence to care recommendations will continue to result in missed opportunities for both health maintenance and disease prevention. The EHR, which might have provided a safety net in this instance, failed to do so because an otherwise conscientious physician did not document a suspected diagnosis. In this case, the EHR’s utility was significantly limited by incomplete data entry.

How might the EHR have altered the course of events that led to the adverse reproductive outcome in this patient? Had the suspected diagnosis of diabetes been recorded, subsequent caregivers would have had the opportunity to see it in the problem list and to have acted upon it. Incomplete ED documentation effectively thwarted one of the putative “virtues” of this technology, namely, linkage of health information directly to the patient rather than to the provider of care.

Data not entered in the electronic chart or hidden from view are equivalent to a paper record locked in a medical office filing cabinet or a hospital’s health records department. In either circumstance, important patient information is unavailable precisely because it is sequestered by the care provider.

Disease prevention and the EHR

In response to anecdotal evidence as well as evidence supporting the effectiveness of prenatal health promotion,^2,3 we evaluated our own EHR for its potential to help our patients avoid adverse outcomes.

Recognizing the frequency of unplanned pregnancy and the fact that 30% of women who conceive have modifiable risk factors that could be treated to improve pregnancy outcome,³ we created a case-finding algorithm to screen all outpatient encounters from our health system’s unified EHR (EPIC, Verona, WI).

Patient data are routinely transferred each evening into an enterprise-wide data warehouse (EPIC Clarity with Oracle and IBM COGNOS) allowing for subsequent data mining for quality improvement, care innovation and research.

We sought to identify reproductive-age women of child-bearing potential and use their data entries to identify risk factors for adverse maternal or fetal outcome. Child-bearing potential was defined as women of reproductive age lacking a history of either sterilization or hysterectomy, and without documented contraceptive use, while the preconception risk factors chosen included morbid obesity, hypertension, poorly controlled diabetes, anemia, renal insufficiency, teratogen exposure, and alcohol, tobacco and illicit drug use.

The algorithm was designed to mitigate incomplete charting by cross-referencing multiple electronic data fields (problem lists, medical and surgical histories, clinical diagnoses, laboratory results, medication orders, and ICD-9 codes), so that multiple dimensions of the record for each risk factor were queried.⁴

Where the EHR falls short

Although our case-finding strategy showed promise, accurate identification of women of child-bearing potential was problematic because up to 25% of patients were incorrectly classified due to incomplete electronic records.⁴ Poor data quality has been noted by others to confound the EHR^5,6 and administrative databases from insurance claims and birth certificates are also notorious for missing information.⁷

In the domain of funded clinical research, a standard for precise data entry has existed for decades. Agencies including the National Institutes of Health place great emphasis on complete data capture and auditing, often insisting on robust data monitoring committees for just this purpose.

One needs only to reflect on the new norm of electronic banking to realize how important data precision is to our wealth, but, apparently, not yet to our health.

We believe it is time to insist that electronic medical records are assiduously created and carefully maintained so that they are more than just expensive versions of their paper ancestors. We suggest that caregivers need to conceptualize data entered on behalf of patients as of the highest value to their current and future health status, making accurate completion of the EHR an act of professionalism.

It is not a coincidence that Stage I meaningful use criteria include proper utilization of the problem list as an essential element in electronic recordkeeping.⁸ Accurate and comprehensive charting is no different than other measures designed to improve patient health and safety. While the activity may not feel particularly important on its face, our clinical vignette should leave no doubt as to the potential consequences of getting this wrong.

Bridging the EHR documentation chasm

The good news is that we do not have to be alone in this process. We can enlist the help of those office- and hospital-based health professionals who already participate directly in patient care (nurses, physician assistants, and others), asking them to be particularly attentive to complete and accurate EHR documentation as part of their workflow.

We can also leverage our patients’ desires to engage in their own health care management by encouraging them to augment and audit their own records at regular intervals (through secure, Web-based patient portals), as well as when they receive episodic care, similar to our current process for medication reconciliation.

We contend that the possibilities for both disease prevention and health maintenance in the context of an accurate EHR are real, but such potential will be governed by the processes surrounding our information capture. We should all strive to close this documentation chasm promptly so that opportunities to innovate using the EHR and improve our patients’ health can be fully realized.

References

1. Vukmir RB, Kremen R, Ilis GL, Hart DA, Iewa MC, Menegazzi J. Compliance with emergency department referral: the effect of computerized discharge instructions. Ann Emerg Med. 1993;22(5):819–823.

2. Whitworth M, Dowswell T. Routine pre-pregnancy health promotion for improving pregnancy outcomes. Cochrane Review. 2009. Issue 4. CD007536.

3. Centers for Disease Control and Prevention. Recommendations for improving preconception health and health care: United States: a report of the CC/ATSDR Preconception Care Workgroup and the Select Panel on Preconception Care. MMWR Morb Mortal Weekly Rep. 2006; 55(RR06):1–23.

4. Straub H, Adams M, Ng D, Silver RK. Can an electronic health record system be used to provide population-based preconception care? Am J Obstet Gynecol. 2013; 208(1) S297-S298.

5. Hasan S, Padman R. Analyzing the effect of data quality on the accuracy of clinical decision support systems: a computer simulation approach. AMIA Annu Symp Proc. 2006;324–328.

6. Terry AL, Chevendra V, Thind A, Stewart M, Marshall JN, Cejic S. Using your electronic medical record for research: a primer for avoiding pitfalls. Fam Pract. 2010;27:121–126.

7. Grimes DA, Epidemiologic research using administrative databases: Garbage in, garbage out. Obstet Gynecol.2010;116:1018.

8. Zywiak W, Metzger J, Mann M. Meaningful use for eligible professionals: the top ten challenges. 2010. http://www.csc.com/health_services/insights/44165-meaningful_use_for_eligible_professionals_the_top_ten_challenges. Accessed December 9, 2013.