Building a Structured Field within Clinical Notes in a Migraine Practice: A
Quality Improvement Project
Chapter 1
Introduction
Migraine is a common neurologic condition affecting hundreds of millions of people
around the world with significant sequelae that prevents those affected from living a better life.
Finding a reliable and effective treatment, and eventually a cure, will improve the lives of
millions of people. Improving migraine documentation in the electronic health record (EHR) by
adding a structured field (SF) to better track migraine metrics, such as migraine headache days
(MHD), will help achieve this aim by improving migraine therapy evaluation and facilitating
clinic-based research. Migraine is highly prevalent around the world and the United States of
America as one in six people in the United States suffer with migraine (Burch, Rizzoli, & Loder,
2018) and is the seventh leading disorder worldwide (Global Burden of Disease, 2015).
Therapies for the disorder have been variably effective and often poorly tolerated, leading to
poor medication adherence and forcing patients to cycle through available medications to find
relief. Migraine incurs significant direct and indirect costs for patients and industry and
disproportionately affects women and vulnerable populations in lower socioeconomic statuses
(SES). Effective and consistent documentation of therapeutic outcomes in a MHD SF will help
patients and clinicians better understand therapeutic outcomes improve migraine patients’ quality
of life (QOL).
1
Migraine Therapy
Patients with migraine use prophylactic approaches to prevent migraine attacks and acute
therapies to treat migraine attacks when they occur. Preventive medications are considered
effective if MHD are reduced by 50% or more within three months (Silberstein, 2015). A
systematic review by Hepp, Bloudek, & Varon (2014) found migraine patients had poor
adherence to migraine prophylaxis and that adverse events caused by preventive medications was
the most typical reason for discontinuation. Lipton et al. (2016) found that currently available
acute therapies are commonly ineffective, with 56% of migraine survey participants reporting
poor relief at 2 hours after taking a acute medication, and 53% of participants reporting poor
relief in 24 hours. The same study found that, of the 44% of participants that did find adequate
relief, 25% of them had another migraine attack within 24 hours (Lipton et al., 2016). A survey
study of 1,165 migraine patients from around the world found that the average number of
different medications ever used by patients with episodic migraine was M = 2.9, SD = 3.6, and M
= 3.9, SD = 3.5 in patients with chronic migraine (Blumenfeld et al., 2013). Discontinuing acute
medication can lead to negative consequences. About 80% of patients with migraine who
completely discontinued acute therapies were unlikely to report being able to work or function
normally with a headache (Lipton et al., 2019). Novel migraine therapies targeting the calcitonin
gene-related peptide emerged in 2018 and have revolutionized the migraine field. The calcitonin
gene-related peptide monoclonal antibody (CGRP MAB) drug class demonstrated a reduction of
MHD per month by 50% or more in 50% of patients that received erenumab 140 mg/ml in the
clinical trial setting (Goadsby et al., 2017). Quantifying clinical responses to established and
novel therapies by documenting MHD SF within the EHR will improve medication adherence
and restore hope in a patient with migraine to live a less disabled life.
2
Migraine Cost
Migraine management carries a high monetary and QOL burden. Direct migraine costs
have been estimated to be $9.2 billion per year (Raval & Shah, 2017). Chronic migraine, which
is defined as 15 MHD a month or more by the International Headache Society (2018), is three
times as expensive as episodic migraine (Messali et al., 2016). Indirect costs of migraine include
a 50% reduced work productivity (Munakata et al., 2009), lower reported QOL scores
(Blumenfeld et al., 2011), and higher work absences and short-term disability costs (Gilligan et
al., 2018). Migraine is commonly associated with lower SES, and given the associated expenses,
can be an obstacle that prevents patients with migraine from achieving a higher SES.
Migraine and Vulnerable Populations
Migraine disproportionately affects women and vulnerable populations in lower SES.
Migraine affects women three times as much as men (Vetvik & MacGregor, 2017), which is
around 20% of women in the United States (Burch, Rizzoli, & Loder, 2018). Those in lower
SES, such as the unemployed, those in a lower family income bracket, the elderly, and those with
disabilities report higher migraine frequency compared to their more affluent counterparts
(Burch, Rizzoli, & Loder, 2018). The American Migraine Prevalence and Prevention report
demonstrated that household income predicted migraine risk because a household income of
$90,000 and above reduced migraine risk by 50% (Lipton et al., 2007). Charleston et al. (2018)
found that people living in areas of low SES reported more migraine disability and suggests
increased stress caused by violence, food insecurity and lack of healthcare access as contributing
factors. Befus, Irby, Coeytaux, and Penzien (2018) propose that migraine is a treatable condition
and should be classified as a health equity issue because those who lack access to care
experience worse outcomes. SES of parents directly predicts adult educations level of their
children (Acacio-Claro, Doku, Koivusilta, & Rimpela, 2017) and Charleston and Heisler (2016)
3
developed a theoretical model for Headache Literacy suggesting that lower education levels
negatively affect headache literacy and self-care leading to worse headache outcomes. Finding
effective therapies for migraine can lead to economic advantages for migraine patients and create
a more equitable reality for them to achieve their goals.
The Los Angeles County Department of Public Health’s report of Key Indicators of
Health by Service Planning Area (SPA) (2017) published a snapshot comparing health outcomes
across the 8 SPAs (Appendix A).
The Electronic Health Record and Structured Fields The Institute of
Medicine (2003) recommended healthcare organizations to use an electronic health record
(EHR) to promote patient safety and improve the quality and efficiency of care. The EHR has
since become the standard of care in the United States health system and has improved
trustworthiness between clinician and patient, improve patient safety and increase the efficacy of
care (Ozair, Jamshed, Sharma, & Aggarwal, 2015). However, the structure in which data is
documented into the EHR has not been standardized in clinical practices. Moreover, narrative,
uncodified documentation is difficult to use for quality improvement projects or clinic-based
research (Meyers et al., 2018), thus limiting the benefits achieved with an EHR. Clinicians miss
the EHR’s full potential when narratively documenting clinical notes as if it was a paper chart.
Clinical notes containing SF help providers document important information more
consistently in a way that can be retrieved efficiently. Cicchini et al. (2016) improved cancer
staging documentation from 28% to 60% over 12 months after having cancer staging attending
physicians document in a structured note format that contained cancer staging specific SF.
Neuroradiologists utilizing structured notes have demonstrated improved rates of documenting
important findings in multiple sclerosis diagnosis and achieved a 100% radiologist utilization
4
rate (Dickerson et al., 2017). Future goals utilizing SF within progress notes to track metrics is to
leverage large codified data with biobanking and predict response to therapy for future patients
(Meyers et al., 2018).
Clinician documentation practices in an academic migraine clinic in Southern California
document notes in narrative and structured formats. The subjective section of the clinical note is
documented in a narrative-only form, while current medications, vital signs and laboratory
results are documented in an SF. Documenting subjective data in a narrative form does not
leverage EHR features that help clinicians capture and track subjective metrics, such as MHD
and acute medications doses used a month, migraine severity and duration, among other metrics,
efficiently. Narrative data is free-text, uncodified data that is challenging to use for quality
improvement projects and research (Meyers et al., 2018). A systematic review and meta-analysis
found that effective utilization the EHR can improve the quality of healthcare by increasing
efficiency and adherence to guidelines while decreasing medication errors and adverse drug
(Campanella et al. 2015). Campanella et al. (2015) suggest that clinicians can enhance the benefit
of the EHR by including a decision support system and expanding SFs.
The reliability of subjective data documented in migraine notes depends on the accuracy
of what patients report. Migraine patients can keep track of their MHD as a way to evaluate the
efficacy of migraine therapies and documenting MHD reduces recall bias when reporting MHD
per month (Torelli & Jensen, 2010). Tracking MHD is so critical in migraine research that the
change in MHD a month is a primary end point when conducting randomized controlled trials of
preventive treatments (Tassorelli et al., 2018). Furthermore, a systematic review evaluating
psychological interventions for migraine found that 14 out the 19 studies selected used headache
diaries as their dependent variables (Sullivan, Cousins, & Ridsdale, 2016). Traditionally, MHD
were tracked using paper and pencil approaches, and now electronic methods of tracking MHD
5
are available. Bandarian-Balooch, Martin, McNally, Brunelli, and Mackenzie found that
shortform migraine diaries and electronic migraine diaries had equivalent adherence when
measuring headache variables (2017). A systematic review by Ramsey et al. (2014) found that
adherence rates to paper diaries were between 83% and 95% and adherence to electronic diaries
was 90% in migraine patients.
Provider documentation and patient self-monitoring practices vary, and this variation can
yield inconsistent documentation quality. Variation in documentation styles between providers
has been found to reduce efficacy and safety when documenting in EHRs (Cohen, Friedman,
Ryan, & Richardson, 2019). Linder, Schnipper, and Middleton (2012) found that providers who
dictated their notes had lower quality of clinical documentation compared to providers who
documented in structured notes for the management of patients with coronary artery disease and
diabetes. Solely narrative notes vary in documentation method and completion, can miss key
clinical data, and make it difficult to determine evidence of meeting documentation quality
measures (Edwards, Neri, Volk, Schiff, & Bates, 2014). While narrative sections of clinical notes
are important, limiting the note within an EHR to be solely narrative is a missed opportunity. The
ideal clinical note in migraine medicine has not been established in the literature. Marmura and
Nahas (2010) discuss the advantages and disadvantages of using an EHR to document clinical
notes and suggest that an ideal migraine clinical note would include the clinical plan from the
previous note, a combination of SFs documenting MHD frequency and severity, weight, blood
pressure and QOL indicators plus narrative free-text to describe findings not covered by the SF.
Structured notes increase the quality of documentation and yield a more consistent note (Bink et
al., 2018; Cecchini, Framski, Lazette, Vega, Strait, & Adelson, 2016; Narayanan et al., 2017;
Simon et al., 2018; Simon et al, 2019; Meyers et al., 2018). In fact, the work by Maraganore et
al. (2015) has improved documentation methods and their group is combining the improved note
6
with DNA biobanking, paving the way for more precise medicine that may lead to predicting
responses to therapy in future patients. Rheumatology practices have incorporated Patient
Reported Outcome Measurement Information System where patients can directly document
clinical outcomes that generate into SFs within a clinical note, a strategy that has increased
patient engagement and improved clinical outcomes (Schmajuk & Yazdany, 2017). Incorporating
more SFs into clinical notes will allow providers to maximize the benefit an EHR can provide
and may encourage patients to increase their participation in their healthcare by documenting
directly into their chart.
Documenting subjective data within the clinical notes of a migraine clinic in a
narrativeonly format limits the EHRs ability to demonstrate the outcomes of the care provided
and hinders clinic-based research. Adding an SF within the subjective section of a clinical note to
document MHD per month exercises the EHR’s ability to better demonstrate the impact therapies
have on MHD and increases documentation consistency between providers. This DNP scholarly
QI project aimed to create a novel MHD SF in a migraine clinic’s EHR and to ascertain
feasibility of having patients keep MHD diaries to supply more reliable MHD for the SF. The
MHD per month data would be input into the MHD SF once built. This initial PDSA cycle only
included patients seen by the PI.
Chapter 2
Review of Literature
A review of literature using the search terms, “structured” “data” “field” “discrete”
“electronic medical record or electronic health record” “headache” “migraine” “neurology”
“diary” “calendar” “quality improvement” and “meaningful use” was conducted in various
combinations in the PubMed, Cumulative Index of Nursing and Allied Health Literature
7
(CINAHL), and Web of Science data bases. The filters of humans, English and 10-year limit
were added to the results of combinations from the above search terms. PubMed yielded the
most results and 50 publications were chosen based on the abstracts if the abstracts included key
terms that discussed structured notes. Further exploration of related publications and cited by
publications from chosen abstracts produced an additional 25 publications. Of the 75 abstracts
reviewed, 50 publications were read in detail and 20 were included in the review of literature.
CINAHL produced six results and three of the publications were chosen for this project. Further
research into similar articles and reviewing cited publications and a manual search of relevant
bibliographies also helped uncover 5 pertinent publications.
Simon et al. (2019) built customized stroke-structured clinical documentation support
into the EHR used in their practice. This paper is this author’s DNP scholarly project executed in
a stroke clinic. The aim was to develop notes containing structured and free text fields. SFs
included information on stroke events, such as the National Institutes of Health Stroke Scale,
timing, presenting symptoms and locations of stroke events, residual deficits, stroke subtype,
diagnostic work up and sequalae, as well as laboratory findings. Optional SFs included measures
capturing disease impact and mood changes, such as the Barthel Index, Epworth Sleepiness
Scale, and Short Test of Mental Status. Workflow processes had medical assistants document
patient self-reported items, cognitive testing was documented by registered nurses, and other
pertinent data was documented by stroke neurologists. The research group met every 2 weeks
with programmers to review ease of use and make modifications based on clinician and staff
feedback. As of August 2018, 2,332 patients had been evaluated. Best Practice Alerts were
created as hard stops depending on the scores and intervention for those scores needed to be
documented. Change in patient outcomes has not been evaluated, but this work demonstrates
feasibility of workflow and use of SFs in a stroke clinic.
8
Cecchini et al. (2016) conducted a quality improvement project utilizing the PDSA
quality improvement framework, to improve cancer stage data capture. In this study an EHR was
customized with a structured module that required physicians to enter a cancer stage. Primary
outcomes were measuring the increase in use of the structured staging module. Secondary
outcomes included physician responses to the process and to inquire why the process was not
used. If staging could not be done, free text areas were available to describe the issue. PDSA
cycles guided the intervention and primary outcomes were measured at 4, 8, and 12 months
postintervention. This group found an increase in staging documentation from 28% to 60% after
12 months demonstrating statistically significant improvement (p < 0.001) in rates of cancer
staging. Limitations included lack of timely staging, the exclusion criteria of patients before
2014 and advanced practice participation such as nurse practitioners (NP) or physician assistants.
Meyers et al. (2018) created a note containing structured and narrative fields in order to
improve quality and support practice-based research within a headache clinic. The note is
currently being tested in 14 headache clinics throughout the United States and data on headache
outcomes are pending. This quality improvement project structured General Anxiety Disorder
7item, Center for Epidemiology Studies Depression Scale and the Migraine Disability
Assessment questionnaire, among others. The developers of the tool met every 2 weeks for three
months, although the methodology used to guide this project was mentioned they used principles
from the model of improvement. The workflow created utilized a medical assistant to input
patientreported data so face-to-face time with the neurologist was not changed. Cohort
characteristics were analyzed once 100 patients with clean data were achieved and correlation
between metrics and initial visits were compared. Dashboards measuring trends were created.
Referrals to appropriate providers were carried out when patients screened positive for
depression or anxiety. Data collection continues and will be published in the future. Limitations
9
in this research include lack of tracking number of days a month affected by migraine, doses of
acute medications taken per month or other QOL indicators. The Migraine Disability Assessment
Scale does include this data, but frequent testing with the same tool incurs the risk of test-retest
validity issues. This work does not address provider and staff response to the structured note.
The articles above demonstrate the feasibility of utilizing SFs in multiple clinical
environments. Strategies to reduce undue burden on providers using structured notes were
explored and avoided reducing face-to-face time between patients and providers. The articles
above described methods of avoiding pitfalls associated with structured notes. The Model of
Improvement was helpful in guiding this type of documentation QI project. Adding SFs into a
clinical note within a migraine practice is feasible and allows for the creation of dashboards and
clinic-based research. Gaps in the published studies include failure in measuring the patient’s
ability to account for migraine days using migraine calendars and describing the clinical provider
and staffs experience with the SF.
Chapter 3
Theoretical Models
The Model for Improvement will be used to guide the process of this DNP scholarly QI
project and the Chronic Care Model will provide the theoretical context that guided the
development of this project. The Model for Improvement applies five principles for
improvement; 1. Knowing why you need to improve, 2. Having a feedback mechanism to tell
you if the improvement is happening, 3. Developing an effective change that will result in
improvement, 4. Testing a change before attempting to implement, and 5. Knowing when and
how to make the change permanent (Langley et al., 2009). The model is comprised of two
10
sections, the first of which asks three questions; 1. What are we trying to accomplish, 2. How
will we know that a change is an improvement, and 3. What change can we make that will result
in improvement (Langley et al., 2009)? The second section is comprised of plan-do-study-act
(PDSA) cycles used to test changes, determine if a change is an improvement, alter the change if
needed, then re-evaluate (Langley et al., 2009). Melnyk and Morrison-Beedy (2019) outline the
PDSA cycle where; 1. The plan phase consists of creating a plan once an improvement project
has been identified, 2. The do phase implements the plan created on a small scale, 3. Results of
the do phase are analyzed in the study phase, and 4. Changes based on results are implemented
into a new plan and the cycle starts again at a larger scale. After the fourth PDSA cycle, the
process of permanence is started and the change can spread if applicable (Langley et al., 2009).
This scholarly project is the first round of PDSA on the journey to eventually spread this
documentation style using the MHD a month SF as the standard documentation practice in a
migraine clinic after evaluating its impact on migraine documentation.
The Chronic Care Model (CCM) addresses quality gaps associated with the demands a
chronic condition exerts on a health system. Migraine is a chronic disorder that has acute
episodes and the CCM is an appropriate model to address this condition. The CCM aims to
change a reactive healthcare system to a proactive one, anticipating the needs of patients with
chronic illness by increasing patient self-management, integrating decision support into provider
practice, and providing information technology (Coleman, Austin, Brach, & Wagner, 2009). This
scholarly project utilizes these guiding points in developing the SF to document migraine and
headache days, tailoring migraine documentation to the migraine population.
11
Chapter 4
Methods
This DNP scholarly QI project aimed to create a novel MHD SF in a migraine clinic’s
EHR and to ascertain feasibility of having patients keep MHD diaries to supply more accurate
MHD for the SF. The MHD per month data would be input into the MHD SF once built. This
initial PDSA cycle only included patients seen by the NP Principal Investigator.
The novel COVID-19 pandemic altered this project’s original timeline and methodology
and the PI adjusted the original methodology in order to continue the endeavor. At the beginning
of the QI project the author began building the MHD SF in collaboration with IT. There were a
series of meetings with stakeholders, which included patients, migraine specialists, the clinic
manager, the medical director, the licensed vocational nurses (LVNs), and IT. However, the
pandemic changed the time line once informaticists were pulled by the hospital system to build
COVID-19 smart phrases, shortcuts, and optimization. The author resumed building the MHD
SF on April 22, 2020 in collaboration with informaticists and migraine clinicians.
Originally, patients were going to be asked to keep MHD per month data and present that
data to the LVNs, who would then input that data into the MHD SF, and the field would populate
into the NP’s note. The migraine clinic currently sees migraine patients for follow up via
telemedicine in order to reduce the spread of the novel virus, thus patients are not interacting
with the LVNs at the moment. The PI planned to train the LVN’s to document MHD in the SF.
The Structured Field Staff Satisfaction Questionnaire (SFSSQ), a three question, 1-5
Likert scale questionnaire and a write-in section for additional comments was to be administered
to the LVN and NP staff at the end of the current PDSA cycle (Appendix B). Responses would
help identify areas of improvement for the MHD SF before initiating the second PDSA cycle,
12
which will open patient enrollment opportunity to patients seen by the NP and two fellow
physicians. The SFSSQ was omitted since the MHD SF has not been completed.
The NP Principal Investigator lost patients prepared to participate in this QI project
because some patients lost access to the migraine clinic once they became unemployed and lost
their health insurance. Other patients were lost to follow up when the migraine clinic triaged
patients, maximizing telemedicine follow up and limiting the number of patients coming to the
migraine clinic. Patients typically seen by the NP for procedures were transferred to the attending
physicians schedules for those procedures. Patients typically managed by attending physicians
were scheduled with the NP for telemedicine appointments. Many of the appointments with the
NP became emergent visits trying to keep patients from going to the ER due to migraine.
Prior to the pandemic, the NP Principal Investigator averaged 40 to 50 patients a week.
Clinic visits decreased significantly after the pandemic started, ranging between 10 to 25
consultations per week. Patients with severe migraine are scheduled at a higher frequency, seen
every two to four weeks and patients that do not require therapy changes are seen once every
three to six months.
Revised Methodology
Participants
The influx of patients typically seen by attending physicians into the PI’s schedule
presented an opportunity to enroll participants in the QI project based on whether they either
documented or did not document MHD per month. Convenience sampling included recruiting
migraine patients within an academic migraine telemedicine clinic in Southern California.
Inclusion criteria were: being a patient of the clinic’s NP (PI on this QI project), age 18 years and
older, having a diagnosis of migraine, with or without aura, or chronic migraine, as defined by
the International Classification of Headache Disorders-3 (International Headache Society, 2018),
13
and volunteering to participate in this QI project by filling out an online questionnaire. Exclusion
criteria included patients in the clinic not seen by NP, clinic patients seen by NP that do not have
a diagnosis of migraine with or without aura or chronic migraine, migraine patients with
comorbid chronic daily headache, and migraine patients who chose not to participate.
Instruments
The Migraine Demographic Data Questionnaire (MDDQ) is a two-part questionnaire
developed by the PI to tabulate demographic and migraine characteristics (Appendix C for part
1; Appendix D for part 2). The first part of the MDDQ was completed by patients who met
inclusion criteria and agreed to participate. This part of the MDDQ collected marital status,
highest education level, income range and preferred healthcare language not systematically
documented in the EHR. These variables were codified to facilitate analysis (Appendix E).
Part two of the MDDQ quantified migraine characteristics through a manual chart review
completed by the PI. The following migraine characteristics were collected in the chart review:
age, gender, zip code, type of insurance – commercial or federal/state sponsored, age of migraine
onset, number of preventive and acute medications ever used, and number of preventive and
acute medications currently used. The last documented blood pressure, weight, height, average
MHD per month, and average severity of migraine attacks were also migraine variables
tabulated. These variables were codified to facilitate analysis (Appendix F). All variables of
interest in this QI project have been reported to directly or indirectly influences migraine risk
(Blumenfeld et al., 2013; Burch, Rizzoli, & Loder, 2018; Charleston et al., 2018; Hepp, Bloudek,
& Varon, 2014; Messali et al., 2016; Vetvik & MacGregor, 2017).
Definitions
The International Classification of Headache Disorder, 3rd edition, definition of a
headache day was used which is the “number of days during an observed period of time affected
14
by headache for any part or the whole day” (International Headache Society, 2018), to define
MHD in this work.
Preventive medications are defined as medications “used to reduce the frequency,
duration, or severity of attacks” by Silberstein (2015), and is the definition used for preventive
medications in this work. Standard migraine preventive medication classes include anti-epileptic
(depakote, gabapentin, topiramate), antidepressant (TCA, SSRI, SNRI), betablocker (atenolol,
labetalol, propranolol), calcium channel blocker (verapamil, nicardipine, flunarizine), serotonin
antagonist (cyproheptadine, pizotifen), select nonsteroidal anti-inflammatories (NSAID)
(Celebrex and meloxicam only), and botulinum toxin (Botox, Xeomin, Myobloc) medication
classes (Siberstein, 2015). Atypical preventive medication classes include alpha-adrenergic
agonists (clonidine, tizanidine, guanfacine), ACEI (lisinopril, enalopril, captopril), ARB
(candesartan, telmisartan) and the NMDAR antagonists (memantine) drug classes (Rau &
Dodick, 2019). The CGRP MAB medication class (erenumab, fremanezumab, galcanezumab,
and eptinezumab) is the first medication class to be developed specifically as a migraine
preventive (Ceriani, Wilhour, & Silberstein, 2019). Prior to the CGRP MABs, every medication
used to prevent migraine was developed for other conditions, found to reduce migraine disability
in some patients with migraine, and secondarily adopted as migraine prophylaxis (Charles,
2018). Migraine clinicians have the opportunity to demonstrate the life enhancing effects of the
new therapies and the MHD SF will aid them in doing so.
Acute medications in migraine are defined as, “acute therapy taken to relieve the pain of
migraine attacks” (Becker, 2015) and is the definition used for acute medication in this work.
Acute medications include acetaminophen, NSAID (ibuprofen, indomethacin, nabumetone,
naproxen, diclofenac, ketorolac, etodolac, mefenamic acid, aspirin) , triptans (sumatriptan,
rizatriptan, zolmitriptan, eletriptan, almotriptan, naratriptan, frovatriptan), triptan-NSAID
15
combinations (treximate), ergotamines (methelergonavine, migranal), and medication
combinations that include butalbital (fioricet, fiorinal) or caffeine (excedrin,
aceminophen/caffeine) (Becker, 2015). Antiemetics (ondansetron, prochlorperazine and
metochlopramide) (Lew & Punnapuzha, 2020) and muscle relaxers (cyclobenzaprine, baclofen)
were considered to be acute medications in migraine therapy for the purposes of this work. The
ditan (lasmitidan) and gepant (ubrogepant and rimegepant) (Ceriani, Wilhour, & Silberstein,
2019) medication classes debuted in 2020 and are counted as acute medications in this work.
Medication classes targeting the CGRP pathways seem to exhibit both preventive and acute
properties (Charles, 2019), thus defining where the agents will be classified is important for this
work.
Most of the migraine variables tabulated were found in SFs within the participants’
charts. Three variables tabulated in this QI project, the age of migraine onset, MHD per month,
and the average severity of migraine attacks as measured by the 1-10 numerical scale, were
found in narrative data documented in clinic visits. Variations of each medication dose and route
were counted as one. For example, sumatriptan 100 mg tablets, sumatriptan 20 mg/ml nasal
spray, and sumatriptan 6 mg/ml subcutaneous injections count as one.
Insurance type was categorized into two groups; commercial, such as HMO and PPO
insurances, and state or federally sponsored insurance, such as Medicare, medical, or Tricare.
Blood pressure was measured in mmHg. Weight was measured in pounds, height was
measured in inches, and pain severity was measured on a 0-10 numerical pain scale.
Zip codes were categorized into their respective SPA (Los Angeles County, 2002). SPA
distribution was analyzed and compared between the two migraine groups to explore for
differences.
16
Data Analysis
Stata 15.1 was used to analyze the data collected. Data were checked for valid coding.
The questionnaire responses and chart review data were summarized using descriptive statistics:
mean/standard deviation for continuous data such as age, weight, and height; and
frequency/percent for categorical data such as marital status, education level, and income
category. Chi square (2) tests, Wilcoxen rank-sum (Mann-Whitney U) tests, and two sample t
tests were used to compare the group of patients who recorded their MHD to those who did not
on the demographic and migraine characteristics. A p-value <0.05 was considered statistically
significant.
Chapter 5
Results
The MHD SF remains under construction. Of the 100 patients that met inclusion criteria
during the QI time period, 40 choose not to participate and 8 had a diagnosis of chronic daily
headache as well as chronic migraine. Fifty-two participants were enrolled in the study. A total of
28 participants reported keeping MHD diaries and 24 did not keep diaries. Participants kept track
of MHD in a varieties of ways (Appendix G). Some participants kept a tally of the number of
attacks that occurred a month, others kept a calendar where they documented MHD and
treatments used per attack, and others used an application to help them keep track of migraine
days where they documented duration and most bothersome symptoms. The systematic
documentation of most bothersome symptoms would be useful in the evaluation of therapeutic
approaches.
No difference in migraine characteristics and demographics was found between the two
groups (with alpha equal to 0.5 criterion), except for the number of preventive medications
17
ordered ever, which was higher in the MHD documentation group, p = 0.03. The migraine group
that documented MHD tried more preventive therapies in the past (M 8.7 SD 5.3) versus the
group that did not document (M 5.9 SD 3.5), z = 2.2, p=0.03.
The participants unanimously reported English as their health care language. There were
mostly women (85%), the most frequently reported education level completed was a bachelor’s
degree (44%), almost half of the participants reported a household income less than $75,000 per
year (46%), and almost half reported being married (44%). The average age was 46 years (SD =
18.5), and a majority had commercial insurance (81%).
The Los Angeles County Department of Health published health outcomes for the eight
SPA regions in Los Angeles County (2017) (Appendix A), therefore, categorizing zip codes into
their respective SPA regions gives insight to health factors and general outcomes of people living
in those regions. Analysis comparing the distribution of SPA regions per group did not find
statistically significant differences. When comparing the two groups after excluding zip codes
not in Los Angeles County, no statistical difference were detected. This analysis did reveal the
cohorts most common SPA regions, 2 (18.9%), 4 (16.2%), and 5 (46%), low represented regions,
1 (5.4%), 6 (5.4%) and 8 (8.1%), and regions with no representation, 3 and 7.
Table 1: Means (M) and standard deviations (SD) for continuous migraine variables for all
participants, group that documented MHD and group that did not document MHD.
All participants
M (SD) N = 53
Documented MHD M
(SD) n = 28
Did not document MHD
M (SD) n = 24
Blood Pressure, Systolic 121.1 (15.6) 120.3 (15.8) 122.2 (15.6)
Blood Pressure, Diastolic 77.21 (7.8) 76.5 (8.1) 78 (7.6)
Weight, pounds 155.1 (46.8) 143.9 (35.1) 168 (55.6)
Height, inches 65.1 (3.7) 65.57 (3.8) 64.45 (3.6)
Age (years) at migraine
onset
21.3 (14.5) 21.2 (14.5) 21.46 (14.7)
18
Migraine frequency, days
per month
10 (8.3) 10 (7.7) 10 (9.1)
Migraine severity, 1-10
scale
4.8 (2) 5.1 (1.8) 4.5 (2.2)
All preventives tried 7.5 (4.8) 8.9 (5.3) 5.9 (3.5)*
Current preventives 2.5 (1.7) 2.4 (1.8) 2.5 (1.7)
All acute therapies tried 8.1 (4.5) 8.6 (4.7) 7.4 (2.1)
Current acute therapies. 2.7 (1.8) 2.4 (1.5) 3.1 (2.7)
* t(50) = 2.2, p = 0.03
Table 2: Highest education completed for all participants, group that documented MHD and
group that did not document MHD.
Education,
highest level
completed
All participants
M (SD) N = 53
Documented
MHD
M (SD) n = 28
Did not
document MHD
M (SD) n = 24
High School 14 (26.9%) 7 (25%) 7 (29.2%)
Bachelor’s
degree
22 (42.3%) 13 (46.4%) 9 (37.5%)
Graduate degree 13 (25%) 7 (25%) 6 (25%)
Doctoral degree 1 (1.9%) 0 (0%) 1 (4.2%)
Post-Doctoral
degree
2 (3.9%) 1 (3.6%) 1 (4.2%)
1 = 1.5, df = 4, p = 0.9.
Table 3: Income reported for all participants, group that documented MHD and group that did
not document MHD.
Income,
thousands of
dollars per year
All participants
M (SD) N = 53
Documented
MHD
M (SD) n = 28
Did not
document MHD
M (SD) n = 24
Under 15
2 (3.9%) 1 (3.6%) 1 (4.2%)
15 – 24.999
0 (0%) 0 (0%) 0 (0%)
1 = 3.1, df = 7, p = 0.9, Fishers Exact p = 0.9
19
25 – 34.999 1 (1.9%) 0 (0%) 1 (4.2%)
35 – 49.999 5 (9.6%) 2 (7.1%) 3 (12.5%)
50 – 74.999 7 (13.5%) 3 (10.7%) 4 (16.7%)
75 – 99.999
9 (17.3%) 6 (21.4%) 3 (12.5%)
100 – 149.999 6 (11.5%) 4 (14.3%) 2 (8.3%)
150 – 199.999 10 (19.2%) 6 (21.4%) 4 (16.7%)
200 or more 12 (23.1%) 6 (21.4%) 6 (25%)
Table 4: Marital status reported for all participants, group that documented MHD and group that
did not document MHD.
Marital Status All participants
M (SD) N = 53
Documented
MHD
M (SD) n = 28
Did not
document MHD
M (SD) n = 24
Single 16 (30.8%) 10 (35.7%) 6 (25%)
Married 23 (44.2%) 13 (46.4%) 10 (50%)
Domestic
Partnership
4 (7.7%) 1 (3.6%) 3 (12.5%)
Widowed 2 (3.9%) 2 (7.1%) 0 (0%)
Divorced 6 (11.5%) 2 (7.1%) 4 (16.7%)
Separated
1 (1.9%) 0 (0%) 1 (4.2%)
1 = 5.8, df = 5, p = 0.3, Fishers Exact p = 0.4
Table 5: Gender category listed in chart for all participants, group that documented MHD and
group that did not document MHD.
Gender All participants
M (SD) N = 53
Documented
MHD
M (SD) n = 28
Did not
document MHD
M (SD) n = 24
Female 44 (84.6%) 24 (85.7%) 20 (83.3%)
1 = .06, df = 1, p = 0.8.
20
Male 8 (15.4%) 4 (14.3%) 4 (16.7%)
Other 0 (0%) 0 (0%) 0 (0%)
Table 6: Insurance type listed in chart for all participants, group that documented MHD and
group that did not document MHD.
1 = 1.3, df = 1, p
= 0.3.
Table 7: Last documented migraine severity found in chart for all participants, group that
documented MHD and group that did not document MHD.
Migraine
severity, 1-10
scale.
All participants
M (SD) N = 53
Documented
MHD
M (SD) n = 28
Did not
document MHD
M (SD) n = 24
0 0 (0%) 0 (0%) 0 (0%)
1 0 (0%) 0 (0%) 0 (0%)
2 6 (11.5%) 2 (7.1%) 4 (16.7%)
3 11 (21.2%) 4 (14.3%) 7 (29.2%)
4 10 (19.2%) 6 (21.4%) 4 (16.7%)
5 5 (9.6%) 4 (14.3%) 1 (4.2%)
6 7 (13.5%) 5 (17.9%) 2 (8.3%)
1 = 8.8, df = 7, p = 0.3. The 0-10 pain scale asks patients to rate their pain, 0 meaning no pain, 1
and 2 meaning mild, 3 and 4 meaning tolerable, 5 and 6 meaning very distressing, 7 and 8
meaning very intense, 9 and 10 meaning excruciating/unbearable (Jensen, Turner, Romano, &
Fisher, 1999).
21
Insurance All participants
M (SD) N = 53
Documented
MHD
M (SD) n = 28
Did not
document MHD
M (SD) n = 24
Commercial 42 (80.8%) 21 (75%) 21 (87.5%)
State or
Federally funded
10 (19.2%) 7 (21%) 3 (12.5%)
7 7 (13.5%) 5 (17.9%) 2 (8.3%)
8 5 (9.6%) 1 (3.6%) 4 (16.7%)
9 1 (1.9%) 1 (3.6%) 0 (0%)
10 0 (0%) 0 (0%) 0 (0%)
Table 8: Analysis of zip codes categorized by SPA regions for all participants, group that
documented MHD and group that did not document MHD.
Service Planning
Area
All participants
M (SD) N = 53
Documented
MHD
M (SD) n = 28
Did not
document MHD
M (SD) n = 24
1 2 (3.6%) 0 (0%) 2 (8.3%)
2 7 (13.5%) 5 (17.9%) 2 (8.3%)
3 0 (0%) 0 (0%) 0 (0%)
4 6 (11.5%) 2 (7.1%) 4 (16.7%)
5 17 (32.7%) 10 (35.7%) 7 (29.2%)
6
2 (3.8%) 1 (3.6%) 1 (4.2%)
7 0 (0%) 0 (0%) 0 (0%)
8 3 (5.8%) 2 (7.1%) 1 (4.2%)
9 - Not in Los
Angeles County
15 (28.9%) 8 (28.6%) 7 (29.2%)
1 = 4.6, df = 6, p = 0.6.
Table 9: Analysis of zip codes categorized by SPA regions within Los Angeles County only for
all participants, group that documented MHD and group that did not document MHD.
Service Planning
Area
All Participants
M (N = 37)
Documented
MHD
M (n = 20)
Did not
document MHD
M (n = 17)
1 2 (5.4%) 0 (0%) 2 (11.8%)
2 7 (18.9%) 5 (25%) 2 (11.8%)
3 0 (0%) 0 (0%) 0 (0%)
4 6 (16.2%) 2 (10%) 4 (23.5%)
5 17 (46%) 10 (50%) 7 (41.2%)
6 2 (5.4%) 1 (5%) 1 (5.9%)
1 = 4.6, df = 5, p = 0.5.
22
7 0 (0%) 0 (0%) 0 (0%)
8 3 (8.1%) 2 (10%) 1 (5.9%)
Chapter 6
Discussion
This DNP scholarly project was a documentation practice improvement QI project. It
represents the first PDSA cycle of building and documenting MHDs in a MHD SF. This PDSA
cycle started with patient participants managed by the PI, who is an NP. A MHD SF for the EHR
clinical note was developed and is being built. The feasibility of having patients keep diaries to
supply data needed for the MHD SF was demonstrated with 52 patient participants, 28 who
historically recorded their MHD and 24 who did not. When demographic and migraine
characteristics were compared between the two groups, no statistically significant differences
were found in anthropomorphic characteristics, such as blood pressure, height, and weight, or in
migraine characteristics, such as age of migraine onset, migraine frequency and severity or
number of acute medications tried, or current use of preventive and acute medications. The group
that tracked MHD met statistical significance for having more preventive medications tabulated
from the medication tab than the group that did not track MHDs (M = 8.9, SD = 5.3 v. M = 5.9,
SD = 3.5, respectively). Participants that recorded MHDs may have been more systematic in
their approach to migraine therapy and cycled through preventive medications faster in order to
reduce migraine’s impact, resulting in more preventive medications found on chart review.
Participant characteristics in this migraine clinic reflected what has been described in
migraine literature regarding the gender most affected, and migraine’s cost and impact on QOL.
Most of the participants were women (85%), many cycled through preventive (M = 7.5, SD =
4.8) and acute (M = 8.1, SD = 4.4 ) medications, almost all currently used preventive (M = 2.5,
SD = 1.7) and acute (M = 2.7, SD = 1.8) medications. This echoes the financial impact of
23
migraine therapy, and the migraine days per month frequency (M = 10, SD = 8.3) and severity
(M = 4.8, SD = 2) provides insight into migraine’s impact on the participants QOL. Of note, one
participant who met criteria for preventive therapy had no current preventive medications listed
in the chart, as counted by the methodology used in this QI project, and was an outlier in
adhering to current migraine preventive therapy recommendations (Silberstein, 2018; Charles
2018). Upon further investigation, the participant incorporates non-pharmacologic and lifestyle
approaches to reduce migraine’s impact. This illustrates the need for documenting
nonpharmacologic approaches in a systematic manner within the EHR.
Participants in this study were diverse in SES elements, but overall the participants
reported features belonging to a higher SES. This is evidenced by the reported highest education
level achieved; bachelor’s (42%), graduate (25%), doctoral (2%) and post-doctoral (4%) (totaling
73%), the majority of participants reported earning more than $100,000 per year (54%), most
having commercial insurance (81%), and most having zip codes in more affluent SPA regions, 2
(19%) and 5 (46%) totaling 65%.
A majority of participants with zip codes in Los Angeles County lived in SPA regions 2
(19%), 4 (16%), and 5 (46%) (totaling 81%). Of these three SPA regions, 2 and 5 are reported to
have better health outcomes than 4, as reported by the Los Angeles County Department of Public
Health, Office of Health Assessment and Epidemiology report (2017) (Appendix A). SPA 2 is
reported to have low housing instability, adult obesity and type 2 diabetes and high percentage of
adults exercising in their neighborhood. SPA 5 is reported to have low difficulty accessing
medical care, adult obesity, smoking cigarettes and number of days that poor health limited
physical activity and high percentage of feeling safe from crime, adults exercising and being
insured. In contrast, SPA 4 is reported to have low rates of insured adults, high rates of difficulty
accessing medical care and adults living with depression The disproportionate enrollment of
24
participants from SPA regions 2, 4, and 5 may be due to location of the migraine clinic, which is
in SPA five, or other socioeconomic factors, such as having reliable transportation, access to
healthcare, or having knowledge that specialized care for migraine exists. Low or no
representation from SPA regions 1, 3, 6, 7, and 8 may indicate an unmet need in managing
migraine in these regions.
Limitations
The beta version of the MHD SF developed in this PDSA cycle (Appendix H) will be
completed in collaboration between migraine specialists and informaticists. Published literature
has not quantified changes in therapeutic outcomes in migraine clinics that document migraine
characteristics in SFs. This field of study remains to be explored and findings published.
While the current participant sample is small and lacks the necessary power to draw
definitive conclusions, this QI project has generated new queries to be explored, such as new
migraine characteristics to trend and new medication designations to consider. Given the small
data set from one migraine clinic, and the multiple comparisons, these findings need to be
retested with a larger sample to explore queries generated. Since this project collected, tabulated
and described various migraine characteristics in participants from one migraine clinic, the
findings cannot be generalized to other migraine clinics. The COVID-19 pandemic presented
pandemic scale challenges. The author navigated the project while adjusting to clinical demands
and adjusted the methodology in order to continue the project.
The methodology used to tabulate past and current medications was completely
dependent on the medication listed in the medication history tab of the EHR. Limiting the count
of preventive and acute approaches to the medication history tab did not include non-medication
approaches used to prevent or resolve migraine. Lifestyle behaviors such as regular exercise,
meal intervals, caffeine use and sleep habits are therapeutic staples in migraine management
25
(Charles, 2018), and were not counted as a preventive or acute approach. A secondary analysis of
the United States 2012 National Health Interview Survey by Zhang et al., (2017) found that 3%
of patients with migraine reported using complementary and alternative modalities (CAM), such
as receiving acupuncture and chiropractic treatments, inhaling and applying essential oils, and/or
ingesting herbal supplements specifically for migraine management. Interest in non-invasive,
percutaneous neuro-modulatory devices has increased in the last decade and novel devices have
been approved by the United States Food and Drug Administration. The type of energy and
frequency of energy type vary among neuro-modulatory devices, some deliver single-pulsed
transcranial magnetic stimulation and others emit electrical current stimulation, that have been
shown to reduce migraine disability (Grimsrud & Singh, 2018). Lifestyle, CAM and
neuromodulatory approaches are not systematically documented in this clinic, thus were not
included in the preventive or acute therapy tabulation. Other limitations identified in this QI
project is in methodology used to tabulate different doses and routes of the same medication
because the various routes and doses were counted as one. The PI did not count IV fluid noted in
the medication tab as an acute treatment because it was not clear if the order was for fluid
resuscitation due to migraine, or for surgical indications. The medication count in this QI project
under represents the true number of medications and therapies ever tried and currently used.
Future opportunities
This QI project has inspired ideas for more areas of migraine exploration. Future PDSA
cycles should address the methodologic limitations discussed above; determining the method of
tabulating various doses and routes of medications, categorizing lifestyle, CAM and
neuromodulatory approaches as acute, preventive, or both, and determine if IV fluid should be
counted as an acute medication. Adding a category of medication that is simultaneously
preventive and acute therapy, such as medications modulating the CGRP system, CAM
26
approaches, and neuromodulatory devices considered both preventive and acute treatment would
be of benefit. Another analysis of interest is exploring the relationship between migraine onset
and tracking MHDs.
Future iterations of this QI project can increase the number of SFs built to aid in
systematic documentation of migraine characteristics. Characteristics of interest can include the
date of initial clinic visit, the severity and frequency of migraine attacks during the initial visit,
average duration of migraine attacks, and most bothersome migraine symptoms. The need to
systematically document lifestyle, CAM and neuro-modulatory approaches used to manage
migraine has been demonstrated in this QI project. Other SFs can be created to facilitate
documenting characteristics helpful in evaluating responses to acute treatments such as the
length of time it takes to terminate a migraine attack and side effects experienced from the acute
therapy.
Further analysis of participant characteristics with a larger data set may help illuminate
the relationship between SPA region and migraine disability. It is likely that those with zip codes
from SPA regions with worse health outcomes are disproportionately affected by migraine. The
52 patients in the study sample were not representative of all SPA regions. A better understanding
of patients SPA regions can guide the clinic’s approach in patient recruitment as well as inform
clinicians of the challenges patients with migraine face in their SPA region.
The current use of SFs to document migraine characteristics in this clinic are limited to
past medical history diagnoses, review of systems, medications, blood pressure, heart rate,
weight, height, body mass index, temperature, and pain level on the 1-10 pain scale. This is an
area in documentation practices that is prime for further development. The evolution of this QI
project can eventually include linking therapeutic outcomes to genetic traits, which is currently
being done in the NorthShore health system (Meyers et al., 2018), in order to develop the
27
capability to predict responses to migraine therapy. This form of documentation can also spread
to other medical practices who manage chronic conditions with acute exacerbations, such as
other neurology practices, and even pulmonology and cardiology practices managing asthma,
COPD and CHF, as well as the primary care setting.
Conclusion
Migraine continues to affect hundreds of millions of people around the world despite
advances in migraine pathophysiology understanding and the development of novel therapies.
Finding reliable and effective treatments, and eventually a cure, will improve QOL, reduce the
disability caused by this neurologic condition, and reduce a barrier for patients to achieve their
full potential. Documentation of migraine characteristics is essential in the practice of migraine
medicine and is indispensable in the evaluation of therapies. Improving the structure in which
migraine data is recorded within the EHR will help clinicians leverage one of the EHRs function
in retrieving characteristic documented in SF and trend changes overtime. This QI project began
the process to facilitate the migraine clinicians ability to efficiently document MHDs per month
within a SF in the EHR. This will help support clinical decision making to either continue or
change treatments. Building a SF to document subjective variables adds value to medical
practices that manage chronic health conditions accompanied by acute exacerbation, such as
migraine and epilepsy. Building a SF encourages and facilitates providers to document important
data consistently, increasing the rate of documenting variables of interest, and result in a more
consistent note among providers. A SF enables the EHR to extract variables of interest,
facilitating clinic-based research and demonstrate change of a variable over time. The EHR has
the capability to visually demonstrate the change in variables of interest if they are documented
in SFs.
28
Gaps that remain in using the MHD SF include building the MHD SF, documenting
MHD in the field, tracking changes in MHD over time, evaluating the MHD SF contribution to
improved MHD documentation and explore its impact in aiding clinicians and patients to
determine whether or not to continue with a treatment course or not. This QI project also
explored the published literature utilizing SF and described the process of building a MHD SF in
this migraine clinic.
In addition to beginning the process of developing a novel MHD SF in a migraine clinic,
this QI project successfully described select migraine characteristics of participants who
documented MHDs compared to those who did not. Statistical analysis demonstrated that both
groups were similar, except that the number of preventives tried was higher in the group that
tracked MHDs. While a larger sample size would be needed to better detect differences in
migraine characteristics between the groups created, this QI project demonstrated the feasibility
of describing migraine characteristics of patients in this clinics and is another step in improving
the clinician’s understanding of their patient population. Better understanding of patients may
help clinicians build a therapeutic report with those they serve and custom tailor migraine
treatments that will hopefully lead to better migraine control. Future analysis to better understand
the impact of SPA regions on migraine can help clinicians link the patients environment to health
outcomes. Continuing this project will uncover more areas of improvement to better serve
patients and cure migraine. Documenting notes narratively limits the EHR’s ability to track
changes in migraine characteristics and is not an efficient way to use the EHR. A SF built to
document MHDs in a clinical note bridges the documentation gap and will increase the
efficiency and validity of MHD documentation. Future studies demonstrating migraine outcomes
after implementing this documentation intervention are needed to demonstrate value added.
29
