Use of electronic medical record templates improves quality of care for patients with infantile spasms

Abstract

Background:

Infantile spasms (IS) is a neurologic disorder of childhood where time to treatment may affect long-term outcomes. Due to the clinical complexity of IS, care can be delayed.

Objective:

To determine if the use of electronic medical record templates (EMRTs) improved care quality in patients treated for IS.

Method:

Records of patients newly diagnosed with IS were retrospectively reviewed both before and after creation of an EMRT for the workup and treatment of IS. Quality of care measures reviewed included delays in treatment plan, medication administration, obtaining neurodiagnostic studies and discharge. The need for repeat neurodiagnostic studies was also assessed. Resident physicians were surveyed regarding template ease of use and functionality.

Results:

Of 17 patients with IS, 7 received template-based care and 10 did not. Patients in the non-template group had more delays in treatment (p = 0.010), delay in medication administration (p = 0.10), delay in diagnostic studies (p = 0.01) and delay in discharge (p = 0.39). Neurodiagnostic studies needed to be repeated in 5 out of 10 patients in the non-template group and none of the 7 patients in the template group (p = 0.04). Surveyed resident physicians reported improved coordination in care, avoidance of delays in discharge and improved ability to predict side effects of treatment with template use.

Conclusion:

In a single centre, the use of protocolised EMRTs decreased treatment delays and the need for repeated invasive procedures in patients with newly diagnosed IS and was reported as easy to use by resident physicians.

Implications:

The use of protocolised EMRTs may improve the quality of patient care in IS and other rare diseases.

Keywords

electronic health records quality improvement electronic clinical documentation quality of healthcare healthcare quality assessment patient outcome assessment medical informatics infantile spasms protocol compliance checklist

Introduction

With the introduction of electronic medical records (EMRs), documentation and clinical workflow have changed dramatically. Electronic medical record templates (EMRTs) are free text elements ranging from single words to checklists or multiple paragraphs that can be placed into clinical documentation to instruct or inform healthcare workers. Additionally, patients can benefit from EMRTs in the form of generation of after visit summaries and instructions specific to their clinical issue by their healthcare providers. Initially, EMRTs were utilised as checklists with prior studies demonstrating variable but generally positive efficacy in EMR systems (Kargul et al., 2013; Mehta et al., 2016). The customisation of these systems has been useful for some medical providers in reducing the burden of documentation although this has not been assessed in rare diseases (Gardner and Pearce, 2013).

Early studies have shown potential clinical care improvements with the use of EMRTs (Sonoo et al., 2016). Additionally, studies across various settings demonstrated increased compliance with guidelines (Rodriguez-Torres et al., 2017), non-verbal communication (Talmon et al., 2013), reduction in medical error (Lin et al., 2014) and improved documentation with the use of EMRTs (Rodriguez-Torres et al., 2017; Scheuner et al., 2017; Sonoo et al., 2016). Although prior studies have demonstrated modest improvement in these quality measures, there is a need for further evidence regarding whether standardised EMRTs improve the care patients receive (Reimschissel et al., 2017).

This study sought to retrospectively identify improvements in a set of clinical care quality measures following the introduction of an EMRT for patients admitted with new diagnoses of infantile spasms (IS). Additionally, this study sought to examine resident physician opinions on the use and utility of these templates.

Method

Template creation

No standardised guideline exists for the inpatient evaluation and treatment of IS. Our institution utilises an algorithm created by our epilepsy and genetics teams for the evaluation and management of IS and this was used as the basis for our EMRT (see Appendix 1). The EMRT could be generated by any physician computer user and populated directly into a patient note within our EMR system. The wording generated an assessment and plan in flowsheet format in accordance with institutional practices, but did not function as a checklist or order set. This EMRT was made available by open access for physicians to reference and use from January 2016 onward at our institution, although use of the EMRT by providers was optional.

IS patient record review and definitions

A retrospective chart review was performed following approval by the Stanford University institutional review board that permitted waivers of consent. All patients had to have been evaluated and treated for IS at our institution (an urban, quaternary academic children’s hospital in the United States) and have had inpatient evaluations. Inclusion criteria were patient records containing diagnosis codes for IS, according to the International Classification of Disease; Version 9 (Center for Disease Control, 2010). Exclusion criteria included atypical presentations of IS, evidence of other epileptic encephalopathies of childhood, or age less than 2 months or over 3 years. Patients were treated according to the same institutional protocol both before and after the introduction of the EMRT and records were included from both before and after the introduction of the EMRT. Records were retrospectively reviewed from January 2010 to November 2017 for use of standardised EMRT (any clinical notes written by physicians) and measures of clinical quality of care defined as: delay in pharmacotherapy, delay in diagnostic testing (magnetic resonance imaging (MRI), lumbar puncture or serum/urine labs), need for repeat diagnostic testing due to medical error and delay in discharge. All cases were reviewed by one author for data acquisition and abstraction and then verified by a second author.

Two types of medical teams work at our institution: primary (inpatient-based medical teams responsible for admission, hospital orders and day-to-day care, and discharge) and consultant (inpatient teams offering clinical recommendations and evaluations of patients but not directly responsible for orders or hospital-based care). Both primary and consultant team service rounds are delivered to patients once per day; thus, 24 hours between ordering and initiation of treatment, study obtainment (e.g. having an MRI or having phlebotomy performed to obtain blood for testing), or discharge was defined as a “delay” in care. Delay in treatment plan and pharmacotherapy was defined as administration of medication ≥24 hours after the initial order was placed for any reason. Delay in diagnostic testing was defined as specimen collection ≥24 hours after the initial order was placed. Delay in discharge was defined as ≥24 hours of delay following having an order written for discharge, or as continued hospitalisation for repeat lab testing, lab results or lack of medication delivery. Documented nursing, laboratory or pharmacy issues out of the control of physician care team were excluded from the calculation of delay when available. Further details of the measures and definitions are contained in the “Infantile spasms template” in Appendix 1 and the “Definitions for IS data review” in Appendix 2.

Resident physician survey

Paediatric, adult neurology and child neurology resident physicians at our institution were asked to participate in a survey following a scheduled educational conference. The paper-based survey had been previously reviewed by physicians in the divisions of epilepsy and genetics at our institution and was piloted by graduating resident physicians for assessing validity prior to use. The data in this pilot were not included in this study. This survey was administered on two dates (1 November 2017 and 8 November 2017) by one investigator. The survey was administered after an educational conference on both occasions and after resident physicians had been asked to participate. Surveys were completed in the same conference room as the educational conference preceding it. The survey asked respondents for demographics, recall of components of the IS protocol and their perceptions regarding ease of use and functionality of the EMRT and included dichotomous, multiple-choice and open-ended questions (survey shown in Appendix 3). All responses were scored manually on two separate occasions to ensure reliable extraction of data by one investigator and were tallied based on the number of responses for each question and answer (raw score). Another author later verified these findings. This information was expressed as a percent of the total number of persons answering the question. Open-ended questions were reviewed for quality improvement of the EMRT but not included in analysis. Incomplete surveys were excluded. Residents qualified as having used the EMRT if they utilised it in any direct patient care capacity (e.g. supervisory resident physician or consultant).

Analysis

A Fisher’s exact test (SPSS version 25) assessed for significance between group differences for patients who had the EMRT used during inpatient care and those who did not. Survey data were scored as a percentage of respondents answering in the affirmative. Certain survey data (question 3) were not scored and were used for institutional quality improvement only. Cohen’s κ coefficient was used to calculate inter-rater reliability in data abstraction.

Results

A total of 21 patients were identified by medical record review and 17 were identified for inclusion in this study. Demographic data are displayed in Table 1. Excluded patients had atypical presentations of IS (n = 2) or syndromes more consistent with other epileptic encephalopathies (n = 2). All but one patient in the non-template group was treated prior to the introduction of the EMRT and this was 2 months after introduction. The most common aetiologies for IS were unknown (cryptogenic) (n = 9) and genetic disorders (SCN2A mutation (n = 1), FOXG1 mutation (n = 1), KCNT1 mutation (n = 1), tuberous sclerosis (n = 1)). There was one case each of Sturge–Weber syndrome and adenylosuccinate lyase deficiency.

Table 1.

Patient demographics and template versus non-template differences in care.

	Template used, n (%)	No template used, n (%)
Demographics	n = 7	n = 10
Median age at diagnosis	7 months	6 months
Male sex	5 (71)	7 (70)
Aetiology of IS
Unknown (cryptogenic)	4 (57)	5 (50)
Genetic	2 (29)	3 (30)
Other	1 (14)	2 (20)
Intervention
ACTH	4 (57)	2 (20)
Prednisone	2 (29)	7 (70)
Vigabatrin	1 (14)	1 (10)
Clinical course			p-Value^a	95% CI^b
Delay in treatment plan	0 (0)	7 (70)	0.01	0.18–0.67
Delay in medication administration	0 (0)	4 (40)	0.10	0.07–0.50
ACTH administration		2 (50)
Prednisone administration		1 (25)
Vigabatrin administration		1 (25)
Delay in neurodiagnostic study obtainment	1 (14)	8 (80)	0.01	0.28–0.77
MRI and lumbar puncture	1 (100)	4 (50)
MRI only	0 (0)	2 (25)
Lumbar puncture only	0 (0)	2 (25)
Repeat neurodiagnostic study needed	0 (0)	5 (50)	0.04	0.10–0.56
Lumbar puncture		3 (60)
MRI		2 (40)
Delay in discharge	1 (14)	4 (40)	0.34	0.10–0.56
ACTH administration	0 (0)	2 (50)
Prednisone administration	0 (0)	1 (25)
Prior delay in medication administration	0 (0)	3 (75)
Prior delay in neurodiagnostic study	1 (100)	4 (100)
Need for repeat neurodiagnostic study	0 (0)	2 (50)
Any clinical delay or repeat study	3 (6)	24 (48)	<0.01	1.37–2.38

MRI: magnetic resonance imaging; IS: infantile spasms.

^a Statistical significance defined as p < 0.05.

^b Binomial 95th percentile confidence interval (negative clinical outcome = positive or yes, positive clinical outcome = negative or no).

There were statistically significant differences between the non-template and template groups (Table 1 and Figure 1). There were significantly more delays in treatment plan (p = 0.01), delays in obtaining neurodiagnostic studies (p = 0.01) and the need for repeat neurodiagnostic studies (p = 0.04). Of patients with delays in medication administration, two patients (50%) involved adrenocorticotropin hormone (ACTH) and one with anti-epileptic drug only. There were no statistically significant differences in delays in medication (p = 0.10) nor delays in discharge (p = 0.34) between groups.

Figure 1.

Comparison of outcome measures in patients in the non-template and template groups.

Additionally, 29 resident physicians were surveyed on the use of the EMRT, although one survey was removed because it was incompletely filled out (Table 2). A total of 16 (57%) residents had used the EMRT at some point in their training (either as primary, supervisory or consultant resident), with all respondents reporting that the EMRT was useful and reporting an average time of less than 1 minute required accessing it. Of resident physicians who had not previously utilised the EMRT, eight (67%) had not taken care of a patient with IS and four (33%) had not rotated on an inpatient paediatric rotation at our institution (all were PGY1 level residents).

Table 2.

Resident physician demographics and responses to template usage.

	Responses, n (%)
Demographics (n = 28)
Residency programme
Paediatrics	22 (79)
Child Neurology	4 (14)
Adult Neurology	2 (7)
Age (years)
25–30	18 (64)
30–35	10 (36)
>35	0 (0)
Year of training
PGY1	7 (25)
PGY2	7 (25)
PGY3	11 (39)
PGY4	2 (7)
PGY5+	1 (4)
Aware of IS protocol	5 (18)
Have used IS template	16 (57)
Feelings on template use (n = 16)
Average time to retrieve template	<1 minute
Improved treatment plan comprehension	16 (100)
Improved medication ordering accuracy	16 (100)
Reduced need to call consultants	15 (94)
Improved avoidance of lab error	15 (94)
Improved comprehension of side effects of treatment	14 (88)
Improved avoidance of discharge delays	14 (88)
Improved coordination of care	13 (82)
Overall feelings on template use (n = 28)
Believe templates help care for complex neurologic disease	28 (100)
Would like to see more templates for complex disease states	28 (100)

IS: infantile spasms.

The κ coefficient for abstraction of all data was 1.0 between two raters, conferring 100% agreement in abstraction of data from chart review and surveys.

Discussion

In this cohort of children admitted for newly diagnosed IS, EMRT use was associated with significant reductions in delays in treatment and in obtaining neurodiagnostic studies. Reducing delay is critical in the treatment of IS as prompt, appropriate treatment is associated with improved outcomes (Chiang et al., 2013; Go et al., 2012). The EMRT was associated with near total elimination of improperly ordered or executed neurodiagnostic studies (such as lumbar punctures or MRIs under anaesthesia), thereby avoiding the risk, discomfort and cost associated with having to repeat each procedure. Uniformly, resident physicians found the EMRT helpful and requested more like it be made for other complex diseases.

This study is unique in that it is one of only a handful of studies to demonstrate improvements in quality measures with the use of EMRTs, specifically in a rare disease. Our study builds on existing literature to demonstrate the potential utility of EMRTs, specifically for complex and rare diseases (Hye et al., 2014; Reimschissel et al., 2017). These types of diseases may be of interest in that a practice/exposure effect may be less evident in improved adherence to guidelines, specifically among trainees, given the infrequency at which they are encountered. Finally, with the recent institution of quality guidelines in child neurology care pathways in IS, the results of this study may become even more relevant (Patel et al., 2018).

This study had several limitations. First, it was a single centre retrospective study and therefore limited by both post hoc assessment of the utility of our primary intervention (EMRT use) and small numbers, driven by the relative infrequency of admissions for IS. It is thus difficult to know whether these findings are generalisable to a larger patient population. Ideally, this study would have been strengthened by a multicentre design, although institutions in the United States utilise different workups and treatments for IS, making this less feasible. Our study utilised an arbitrary cut-off of 24 hours to define delays in care based on the frequency of inpatient rounding in our hospital (once daily) which may not be reflective of care at other institutions and affects generalisability of the definition of delays in care in this study; a shorter time range may have detected even more errors. Utilisation of a single data abstractor with knowledge of IS and verification by another physician was made to minimise inconsistencies in data acquisition, improving the objectivity of the data. Finally, we note that by comparing patients largely before and after introduction of the EMRT, we cannot exclude the possibility that the improvements we found are in part due to practice changes over time rather than due to the EMRT alone. With regard to resident physician perceptions, our survey was limited in that it was a one-time administration of recall of measures of a multifaceted treatment plan for a complex disease state. Further, only 57% of resident physicians surveyed had used our IS EMRT in the past, which limited our data collection abilities. This was likely secondary to the rarity of IS even at tertiary teaching hospitals. Finally, a skew of responses towards those having only been exposed to the EMRT is unmistakable given the introduction time of the template to our survey being administered.

Even with limitations, this report presents dramatic improvements in the quality of care for patients with a rare neurologic disorder. Although a low number of patients were included in this study (a reflection of a single centre study on a rare disease), the authors feel the results of this study are applicable not just to IS but to other rare disorders where a complex, time sensitive, inpatient workup is necessary. Further, the minimal time to acquire our EMRT (<1 minute) was a notable factor driving ease of use. Residents found the EMRT improved efficiency and were in favour of its use although this was not assessed in our study. Our intervention is cost free, easy to use and improves the resident physician educational experience in treating disorders such as IS. In an era of increasing medical protocolisation and setting of quality standards, use of EMRT is an enhancement to clinical care for physician providers.

Conclusion

Use of an EMRT during admission for newly diagnosed IS was associated with fewer delays in treatment, improved acquisition of diagnostic studies and reduced need to repeat invasive diagnostic testing. The EMRT was also reported as convenient and easy to use by all resident physicians. We believe the results of this study may be generalisable even in centres with different IS diagnostic and treatment protocols.

Footnotes

Acknowledgements

The authors would like to thank the Department of Child Neurology at Stanford University for their support in the production of this manuscript.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Jonathan D Santoro, MD

Appendix 1. Infantile spasms template

@NAME@ is a @AGE@ @SEX@ with a history of *** who presented to *** with episodes of flexor spasms (brief, repeated episodes of back arching and arm extension) consistent with infantile spasms (IS). The patient has characteristic findings on his/her electroencephalogram (EEG) of IS (hypsarrhythmia). There were also clinical spasms noted on video and appreciated upon physical and neurologic examination, which is also notable for ***. These are sufficient for a diagnosis of IS; the triad of IS, hypsarrhythmia on EEG and developmental delay supports a diagnosis of West syndrome ***. The causes of IS and West syndrome are multiple, often genetic. There are multiple treatment options, although data are most supportive of the efficacy of high-dose steroid therapy for IS. The goal of treatment of IS is to quickly halt the spasms and prevent the progression of the disease. High-dose ACTH can be successful at stopping the progression of spasms, although sometimes spasms can be refractory to treatment.

Appendix 2. Definitions for IS data review

Aetiology of IS: defined as the most likely cause of infantile spasms (IS).

Cryptogenic: no known or identified aetiology and normal development prior to that point.

Genetic: diagnosis of a genetic condition that has been previously associated with IS or West syndrome.

Tuberous sclerosis was included in this group

Structural: a lesion on Magnetic resonance imaging (MRI) that may or may not be associated with an underlying condition (i.e. Sturge–Weber syndrome, hemimegancephaly or evidence of TORCH infection).

As no patients with tuberous sclerosis had evidence of tuber formation on neuroimaging in this study, they were not included as “structural” aetiologies although we could not rule out microstructural changes for epileptiform activity.

Standardised electronic template usage.

Defined as use of the established template after 1 January 2016 within a history and physical examination, progress note or independent note within 24 hours of identification of hypsarrythmia on electroencephalogram (EEG).

IS quality measures.

Delay in treatment plan.

Defined as any delay in administration of medication or obtainment of a neurodiagnostic study due to lack of obtaining prior necessary medical documentation, consent or appropriate lab work.

Delay ≥24 hours was considered a delay in treatment plan.

Delay in medication administration.

Defined as any delay in medication administration (prednisone, ACTH or anti-epileptic medications) that was not related to prior need for medical documentation, consent or appropriate lab work (e.g. medication was not available or was ordered incorrectly).

Family refusal of medication was not included as a delay.

Unstable medical status was not included as a reason for delay (e.g. patient hypertensive prior to medication initiation).

Delay ≥24 hours was considered a delay in medication administration.

Delay in neurodiagnostic study.

Defined as any delay in routine obtainment of EEG, MRI or lumbar puncture (LP) due to issues with prior need for medical documentation, consent, inpatient scheduling difficulties or safety issues/concerns.

Delays in ability to coordinate anaesthesia for MRI/LP were included in this group.

Delay ≥24 hours was considered a delay in neurodiagnostic study obtainment.

Need for repeat testing.

Defined as the need to repeat any neurodiagnostic study (EEG, MRI or LP) or serum/urine lab more than one time.

Included accidental omission of test on earlier recommendations, unintentional test ordering omission by primary team or improper collection of sample (LP or MRI) on first attempt.

Delay in discharges.

Defined as any delay in discharge caused by incomplete lab workup, need for repeat inpatient neurodiagnostic study (MRI, LP, EEG or any combination), inability to obtain medications needed for treatment (ACTH, prednisolone or anti-epileptic medication) or inability to obtain necessary outpatient follow-up.

Delay ≥24 hours was considered a delay in discharge.

Additionally defined as any time a discharge order was written in the electronic medical record but the patient was discharged ≥24 hours later than the order for any non-medical reason (with exceptions below):

Definition of medical reasons for delay in discharge included:

Any medication side effects causing alteration in regimen (e.g. hypertension associated with ACTH).

Changes in clinical status prior to discharge requiring additional inpatient observation (e.g. increase in epileptic spasms immediately prior to discharge).

Non-medical reasons for delay in discharge that were excluded from being defined as “delay”:

Parental discomfort with plan (e.g. despite appropriate training, parents unable to administer injections).

Weather issues or safety concerns that necessitated additional hospital days (e.g. snowstorm prevented family being able to get home).

Social concerns raised by medical team or family causing delay in discharge (e.g. patient is in child protective services custody).

Appendix 3. Resident physician survey

References

Center for Disease Control (2010) International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM. Available at: http://www.cdc.gov/nchs/icd/icd9cm.htm (accessed 10 January 2018).

Chiang

Read-Brown

, et al. (2013) Evaluation of electronic health record implementation in ophthalmology at an academic medical center (an American Ophthalmology Society thesis). Transactions of the American Ophthalmological Society 111: 70–92.

Gardner

Pearce

(2013) Customization of electronic medical record templates to improve end-user satisfaction. CIN: Computers, Informatics, Nursing 31: 115–121.

Mackay

Weiss

, et al. (2012) Evidence-based guideline update: medical treatment of infantile spasms. Report of the Guideline Development Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology 78(24): 1974–1980.

Hye

Smith

Wong

, et al. (2014) Leveraging the electronic medical record to implement an abdominal aortic aneurysm screening program. Journal of Vascular Surgery 59(6): 1535–1542.

Kargul

Wright

Knight

, et al. (2013) The hybrid progress note: semiautomating daily progress notes to achieve high-quality documentation and improve provider efficiency. American Journal of Medical Quality 28(1): 25–32.

Lin

Powell

Kagetsu

(2014) Efficacy of a checklist-style structured radiology reporting template in reducing resident misses on cervical spine computed tomography examinations. Journal of Digital Imaging 27: 588–593.

Mehta

Radhakrishnan

Warring

, et al. (2016) The use of evidence-based, problem-oriented templates as a clinical decision support in an inpatient electronic health record system. Applied Clinical Informatics 7(3): 790–802.

Patel

Berg

Billinghurst

, et al. (2018) Quality improvement in neurology: child neurology quality measure set. Executive summary. Neurology 90(2): 67–73.

10.

Reimschissel

Dela Cruz

Gonzalez

, et al. (2017) Immunotherapy toxicities: a new electronic documentation template to improve patient care. Clinical Journal of Oncology Nursing 21(2): 41–44.

11.

Rodriguez-Torres

Huang

Mihlstin

, et al. (2017) The effect of electronic health record software design on resident documentation and compliance with evidence-based medicine. PLoS One 12(9): e0185052.

12.

Scheuner

Peredo

Tangney

, et al. (2017) Electronic health record interventions at the point of care improve documentation of care processes and decrease orders for genetic tests commonly ordered by nongeneticists. Genetics in Medicine 19(1): 112–120.

13.

Sonoo

Iwai

Inokuchi

, et al. (2016) Embedded-structure template for electronic records affects patient note quality and management for emergency head injury patients: an observational pre and post comparison quality improvement study. Medicine (Baltimore) 95(40): e5105.

14.

Talmon

Horn

Wedel

, et al. (2013) How well do we communicate?: a comparison of intraoperative diagnoses listed in pathology reports and operative notes. American Journal of Clinical Pathology 140(5): 651–657.