Medication Adherence Among Patients with Corneal Diseases

Abstract

Purpose:

Medication non-adherence is a ubiquitous problem. However, the adherence of patients to medications to manage corneal conditions is unknown. A prospective cohort study investigated the patterns of eye drop adherence among patients with corneal conditions.

Methods:

Patients ≥18 years taking prescription eye medications were recruited from an academic-center’s cornea clinic. Data collected included age, gender, total doses of eye medications, and category of primary corneal diagnosis. Participants completed adapted versions of the 12 question Adherence to Refills and Medications Scale (ARMS) and the three question Voils Medication Adherence Scale (VMAS). Survey data was dichotomized as “adherent” and “non-adherent” and subscales reported for reasons of non-adherence. Logistic regression analyses were used to test associations with adherence.

Results:

A total of 199 participants were surveyed February to March 2019 (95% response rate). Participants were aged 19 to 93 with a mean age of 59 (SD 17.8). The percent of participants were considered non-adherent was 72% by ARMS and 33% by VMAS. Older age was associated with higher adherence by ARMS (odds ratio (OR)=1.48, 95% confidence interval (CI): 1.14–1.93, p=0.004) and by VMAS (OR=1.24, CI: 1.04–1.48, p=0.012). Adherence was not significantly associated with race, gender, education, total doses of eye medications, or primary cornea diagnosis.

Conclusion:

Medication adherence was lower than expected, particularly on the ARMS scale that asks more detailed questions. Clinicians should engage in conversations about adherence, especially with younger patients, if they are not seeing an expected clinical response.

The World Health Organization (WHO) approximates that ten million people globally are visually impaired due to corneal disease.¹ Severe corneal diseases, such as microbial keratitis (corneal ulcers), are the fourth leading cause of blindness worldwide.^2–8 Microbial keratitis is considered to be an ocular emergency due to its rapid progression and therefore requires immediate treatment in order to avoid blindness.⁹

According to WHO guidelines, broad-spectrum topical antimicrobials are recommended as the primary therapeutic option in treating suspected corneal ulcers.^9,10 Eyedrops are the most common method to deliver antibiotic to the eye.¹¹ The dosage and frequency of eyedrop administration is dependent on the severity of the ulcer.^9–11 For example, physicians could recommend that eyedrops be administered four to six times a day for less serious infections, and up to every 30–60 minutes for 24–36 hours if the infection is severe.^9–11

Medication non-adherence is a ubiquitous problem in medicine. Although eye drop adherence has been studied in glaucoma,^12–14 very little is known about medication adherence among cornea patients. Before creating interventions to improve medication adherence among patients with corneal disease, it is important to understand the magnitude of the problem of medication adherence among various corneal diseases, as well as factors that may influence whether patients with corneal disease are adherent to their prescribed medication regimen. Among glaucoma patients, Dr. Tsai categorized barriers to medication adherence as follows: situational/environmental factors (49%), medication regimen factors (32%), patient-related factors (16%), and physician-related factors (3%).¹⁵ Additional factors, such as medication side effects, poor health literacy, confusion on when or how to take medications, asymptomatic conditions, forgetfulness, and affordability have also been identified as potential barriers to optimal adherence.^16,17

Medication adherence can greatly impact the efficacy of a treatment regimen. Poor medication adherence often contributes to the discordance between real-world outcomes and results from drug trials, as real-world outcomes are often worse. The purpose of this research is to quantify the prevalence of medication non-adherence among patients with varying levels of severity of corneal disease and identify potential reasons for poor adherence. Additionally, we will identify demographic and clinical characteristics that predict poor self-reported medication adherence among patients with corneal disease.

METHODS

Study Population:

This was a prospective cohort survey study. Participants were recruited from the cornea clinic at the University of Michigan Kellogg Eye Center (KEC). Participants who were a returning patient, taking at least one prescription medication for a corneal condition, and ≥ age 18 were eligible to take the cornea medication adherence survey. The University of Michigan’s Institutional Review Board approved this project as a quality improvement initiative with exemption but approval for publication any resultant findings. Patients were asked whether they preferred to complete the questionnaire on their own or have a team member read the survey out loud to them.

Measures:

Patients completed survey data including race and ethnicity, self-reported education level, household income range, and primary cornea diagnosis extracted from the medical record alongside two instruments that assess self-reported adherence adapted for cornea medications, the Adherence to Refills and Medications Scale (ARMS)¹⁸ and Voils’ Medication Adherence Scale (VMAS) measures.¹⁹ Data were extracted from the electronic health record and included patients’ sex, age, primary corneal diagnosis, total doses of eye medication per day, and total number of eye medications. Primary cornea diagnoses were categorized and listed as either ‘Severe Corneal Conditions’, or ‘Non-severe Corneal Conditions’ by an ophthalmologist (SM) (Supplementary Table 1).

Medication adherence was assessed through adapted versions of the Adherence to Refills and Medications Scale (ARMS)¹⁸ and Voils’ Medication Adherence Scale (VMAS) (Appendices A and B).^19,20 Patients were asked to consider both eye drops and pills for eye conditions when answering the questionnaire. The ARMS is a 13-question instrument with two subscales: adherence to taking the eye medication (ARMS Taking) and adherence to re-filling the eye medication (ARMS Filling). Responses to each question ranged from (1): “none of the time” to (4): “all of the time.” Total ARMS scores ranged from 12 to 48. The ARMS Taking subscale ranged from 8 to 32 and the ARMS Filling subscale ranged from 4 to 16. Lower scores indicated a better adherence score. Scores for the ARMS were dichotomized into adherent, a score of 12 or less, and non-adherent, a score greater than 12, per the scale’s authors’ guidelines.¹⁸ Each of the subscales were also dichotomized into adherent, a score of 4 or less for the ARMS Filling and a score of 8 or less for the ARMS Taking, and non-adherent, a score greater than 4 or 8, respectively. The first VMAS scale contained three questions that ask about how often patients forgot or missed taking their eye medication in the previous seven days. Responses to each question ranged from (1): “none of the time” to (5): “every time.” Scores on the VMAS were tabulated as a mean of the three items.¹⁹ The VMAS scores were dichotomized to adherent, mean score of one, and non-adherent, a mean score greater than one. The second VMAS scale included 18 items that aimed to identify reasons for non-adherence independently from the first three questions. The responses to “I missed my dose because…” ranged from (1): “not at all” to (5): “very much” for each item. VMAS Supplementary questions were categorized into “did not occur” if “Not at all” was chosen and “occurred” if anything other than “Not at all” was chosen.

Statistical Analysis

Demographics of the participant sample were summarized with descriptive statistics including means, standard deviations, frequencies, and percentages. Because less than 2% of participants had any missing values for the ARMS portion of the questionnaire, missing scores in the ARMS were imputed as the participant’s average score across all of the ARMS questions. Groups were compared for differences with Chi-Square tests for all categorical variables and t-tests for all continuous variables. Chi-square tests (or Fisher’s Exact test in instances with small cell counts) were performed.

Logistic regression models were used to examine the associations between the self-reported adherence status, cornea diagnosis, number of cornea medications taken, and sociodemographic variables. Cronbach alpha was used to check for internal consistency of the surveys in the population with corneal disease. Statistical models were adjusted with demographic variables if significant in univariate analysis or hypothesis-driven variables. In the instance where race was significant in univariate analysis, the models were also adjusted for income and education. All analyses were done in SAS 9.4 (Cary, NC).

RESULTS

Data were collected from February to March 2019 on 199 participants. Two participants were excluded from data analysis for not actively taking any prescribed ocular medications at the time of the questionnaire, and one participant was excluded for answering only the first eight questions of the survey, so 196 responses were analyzed. The mean age of participants was 59.3 years (SD=17.8), with a range from 19 to 93 years. Table 1 summarizes the participants’ demographic and clinical characteristics. Of the participants, 52% were female, 80% were Caucasian, 28% had a high school diploma or less, and 18% had household incomes of less than $25,000. Primary corneal diagnosis was categorized into four categories as follows: 1) post-corneal transplant (49%), 2) infectious keratitis (18%), 3) other - non-severe corneal conditions (20%), and 4) other – severe corneal condition (13%),

Table 1:

Demographics and Clinical Information

	Overall (N=196)
Demographics
Gender
Female	101 (51.5%)
Male	95 (48.5%)
Race
White	153 (79.7%)
Non-White	39 (19.9%)
Education Level
High School Diploma or Less	53 (28.2%)
Some College	51 (27.1%)
College Degree	45 (23.9%)
Graduate Degree	39 (20.7%)
Income
Less than $25,000	29 (18.1%)
$26,000-$50,000	42 (26.3%)
$51,000-$100,000	54 (33.8%)
Over $100,000	35 (21.9%)
Age
Mean (SD)	59.25 (17.8)
Clinical Information
Diagnosis
Transplant	96 (49%)
Infectious Keratitis	35 (18%)
Other-Non-Severe	39 (20%)
Other-Severe	26 (13%)
Total Eye Medicine Doses per Day
Mean (SD)	7.8 (7.9)
Number of Different Eye Meds
Mean (SD)	2.3 (1.5)

The average number of unique eye medication for corneal transplant, infectious keratitis, other – not severe, other – severe was 2.3 (SD=1.5), 2.9 (SD=1.1), 2.3 (SD=1.7), and 1.8 (SD=1.2) respectively. The average number of doses of eye medications for transplant, infectious keratitis, other – not severe, and other – severe was 7.2 (SD=6.9), 13.7 (SD=12.1), 5.9 (SD=4.5), 5.4 (SD=3.9) respectively. A majority of participants (71%) were considered non-adherent to their medications by the ARMS scale, while 33% were considered non-adherent as measured by the VMAS. 59% were non-adherent to taking the corneal medication based on the ARMS Taking sub-scale and 43% were non-adherent to corneal medication refills based on the ARMS Filling subscale. ARMS and VMAS agreed 60% of the time on adherence status with 32% of participants labeled non-adherent and 28% of participants labeled adherent by both surveys. Disagreement on classification occurred for 39% of participants with ARMS labeling participants as non-adherent and VMAS labeling them as adherent. Converse disagreement on classification had 1.0% of participants on ARMS labeled as adherent where VMAS labeled them as non-adherent. ARMS had a Cronbach’s alpha, the measure of internal consistency, of 0.78, and VMAS had a Cronbach’s alpha of 0.94.

Adherence to Refills and Medications Scale (ARMS)

Table 2 summarizes adherence status by demographics and clinical information for ARMS. The mean age of adherent participants was 66.8 compared to 56.1 for non-adherent participants (p <0.001). Adherence was significantly different by race with a higher proportion of self-reported non-adherence among non-Whites compared to Whites (p = 0.004). Other demographic and clinical variables were not significantly different between groups (Table 2). Similarly, older age was significant associated with corneal medication adherence on the ARMS Taking Medication sub-scale and on the ARMS Filling Medication subscale even after adjustment in a multivariable model (Supplementary Table 2). Figure 1 depicts adherence status for ARMS and both of its subscales as distributed by patient age.

Table 2:

Self-report of Adherence to Refills and Medications Scale (ARMS), and Voils Medication Adherence Scale (VMAS)

	ARMS			VMAS
	Adherent (N=57)	Non-Adherent (N=139)	P value & Chi-Square	Adherent (N=131)	Non-Adherent (N=65)	P value & Chi-Square
Demographics
Gender	N (column %)	N (column %)		N (column %)	N (column %)
Female	32 (56.1%)	69 (49.6%)	0.41	70 (54.2%)	30 (46.2%)	0.29
Male	25 (43.9%)	70 (50.4%)		60 (45.8%)	35 (53.9%)
Race
White	52 (91.2%)	101 (72.7%)	0.004	100 (76.3%)	53 (81.5%)	0.41
Non-White	5 (8.8%)	38 (27.3%)		31 (23.7%)	12 (18.5%)
Education Level
High School Diploma or Less	15 (26.8%)	38 (28.8%)	0.12	36 (28.6%)	17 (27.4%)	0.14
Some College	17 (30.4%)	34 (25.8%)		34 (27.0%)	17 (27.4%)
College Degree	8 (14.3%)	37 (28.0%)		25 (19.8%)	20 (32.3%)
Graduate Degree	16 (28.6%)	23 (17.4%)		31 (24.6%)	8 (12.9%)
Income
Less than $25,000	8 (17.0%)	21 (18.6%)	0.45	18 (17.3%)	11 (19.6%)	0.61
$26,000-$50,000	10 (21.3%)	32 (28.3%)		29 (27.9%)	13 (23.2%)
$51,000-$100,000	15 (31.9%)	39 (34.5%)		32 (30.8%)	22 (39.3%)
Over $100,000	14 (29.8%)	21 (18.6%)		25 (24.0%)	10 (17.9%)
Age
Mean (SD)	66.8 (12.7)	56.1 (18.6)	< 0.001	61.7 (17.47)	54.4 (18.2)	0.007
Clinical Information
Diagnosis	N (row %)	N (row %)		N (row %)	N (row %)
Transplant	32 (33.3%)	64 (66.6%)		68 (70.8%)	28 (29.2%)
Infectious Keratitis	7 (20.0%)	28 (80%)	0.43	19 (54.3%)	16 (45.7%)	0.29
Other - Severe	6 (23.1%)	20 (76.9%)		19 (73.1%)	7 (26.9%)
Other - Non-Severe	12 (30.8%)	27 (69.2%)		25 (64.1%)	14 (35.9%)
Total Eye Medicine Doses per Day
Mean (SD)	7.5 (9.9)	7.9 (6.9)	0.73	7.6 (8.0)	8.2 (7.7)	0.61
Number of Different Eye Meds
Mean (SD)	2.1 (1.4)	2.4 (1.5)	0.15	2.3 (1.5)	2.4 (1.4)	0.83

SD: Standard Deviation. Chi-square tests were performed on all categorical variables, t-tests were performed on all continuous variable

Figure 1. Age by Survey Adherence Status -.

Boxplots of age by adherence status for Voils Medication Adherence Scale (VMAS), Adherence to Refills and Medication Scale (ARMS) and ARMS Subscales.

In the multivariable logistic model for the ARMS scale, age remained significantly associated with adherence to eye medication with an odds ratio (OR) of 1.48 (95% CI: 1.14 – 1.93, p = 0.004). There were no significant differences in adherence based on race, diagnosis, total medications, education, and income (Table 3). Notably, patients with corneal transplantations reported between 29–66% non-adherence, and patients with infectious keratitis reported between 46–80% non-adherence, depending on the survey measure used. In the ARMS Filling sub-scale, the total number of eye medications was significantly associated with worse adherence to re-filling corneal medications (OR =0.75, 95% CI: 0.57 – 0.97, p = 0.03) and white race was significantly associated with improved adherence to re-fills (OR=2.74, 95% CI: 1.00 – 7.51, p = 0.05) (Table 3).

Table 3.

Logistic Regression Models of Adherence for Adherence to Refills and Medication Scale (ARMS), ARMS Subscales, and Voils Medication Adherence Scale (VMAS)

	ARMS			ARMS Filling
	Odds Ratio	95% CI	P value	Odds Ratio	95% CI	P value
Age (Scaled by decades)	1.48	1.14, 1.93	0.004	1.38	1.10, 1.72	0.006
Race: (Ref: Non-White)
White	2.68	0.68, 10.51	0.16	2.74	1.00, 7.51	0.05
Number of Eye Medications	0.76	0.57, 1.02	0.07	0.75	0.57, 0.97	0.03
Education: (Ref: High School Diploma or Less)
Some College	1.1	0.40, 3.05	0.85	0.66	0.26, 1.72	0.40
College Degree	0.42	0.13, 1.40	0.16	0.77	0.27, 2.16	0.61
Graduate Degree	1.27	0.38, 4.18	0.70	1.48	0.47, 4.69	0.51
Income: (Ref: Less than $25,000)
$26,000-$50,000	0.71	0.21, 2.38	0.58	0.48	0.16, 1.46	0.20
$51,000-$100,000	0.82	0.26, 2.61	0.74	0.43	0.14, 1.29	0.13
Over $100,000	1.53	0.40, 5.81	0.54	0.88	0.24, 3.19	0.85
Diagnosis: (Ref: Other - Not Severe)
Transplant	1.17	0.43, 3.18	0.76	0.80	0.30, 2.09	0.64
Infectious Keratitis	0.68	0.18, 2.56	0.57	1.68	0.52, 5.37	0.38
Other - Severe	0.69	0.16, 2.98	0.62	0.39	0.11, 1.41	0.15
	ARMS Taking			VMAS
	Odds Ratio	95% CI	P value	Odds Ratio	95% CI	P value
Age (Scaled by decades)	1.3	1.04, 1.63	0.022	1.27	1.03, 1.57	0.028
Race: (Ref: Non-White)
White	2.55	0.83, 7.88	0.10	0.66	0.24, 1.79	0.41
Number of Eye Medications	0.91	0.71, 1.17	0.46	1.01	0.78, 1.30	0.96
Education: (Ref: High School Diploma or Less)
Some College	0.92	0.36, 2.32	0.86	0.74	0.29, 1.89	0.53
College Degree	0.49	0.17, 1.41	0.19	0.51	0.19, 1.40	0.20
Graduate Degree	1.15	0.38, 3.46	0.80	1.06	0.34, 3.37	0.53
Income: (Ref: Less than $25,000)
$26,000-$50,000	0.61	0.21, 1.80	0.37	1.34	0.46, 3.82	0.60
$51,000-$100,000	0.52	0.18, 1.51	0.23	0.98	0.35, 2.72	0.96
Over $100,000	1.42	0.42, 4.77	0.57	1.83	0.53, 6.26	0.34
Diagnosis: (Ref: Other - Not Severe)
Transplant	1.03	0.41, 2.63	0.94	1.18	0.47, 2.98	0.73
Infectious Keratitis	0.67	0.21, 2.16	0.50	0.81	0.27, 2.39	0.70
Other - Severe	0.99	0.28, 3.49	0.98	1.14	0.34, 3.91	0.83

CI: Confidence Interval

Voils Medication Adherence Scale (VMAS)

Table 2 summarizes adherence status by demographics and clinical information for each survey for VMAS. The mean age of those who were adherent to taking their medications in the past seven days was 61.7, while the mean age of those who were non-adherent was 54.4 (p = 0.007). There were no other significant differences between groups for any other variables.

Age scaled by decades was significantly associated with adherence to eye medication in the past seven days as measured by the VMAS with an odds ratio of 1.27 (95% CI: 1.03 – 1.57) (p = 0.028), adjusting for diagnosis and number of eye medications (Table 3). No other variables were significant in the adjusted analysis. Figure 1 presents the adherence status for VMAS as distributed by patient age.

VMAS Supplementary Questions

Many reasons were stated for missing a dose in the last seven days by adherence status according to VMAS and ARMS. Of the 190 participants who completed all of the supplemental questions, 57% of patients reported facingno barriers, 8% of patients reported facing one barrier, and 35% of patients reported facing two or more barriers. Of those that were globally classified as non-adherent by ARMS and responded to the VMAS supplemental questions, 39% reported poor adherence because they forgot, 36% missed their doses because they were out of their routine, 29% missed doses because they did not have their medications with them, 27% missed doses because they were asleep, 17% missed doses because they were late, 15% missed doses because they could not meet the requirements of the medication schedule, 13% missed doses because they ran out of medication, and 7% missed doses because the medication caused side effects (Supplementary Table 3).

DISCUSSION

In our sample of 196 patients with corneal conditions, one third to three-quarters of participants self-reported non-adherence on two different measure of self-reported medication adherence. In both measures of self-reported adherence, younger age was associated with a greater risk of non-adherence to corneal medications. The ARMS survey is more detailed and includes 12 questions and also asks about adherence in general. The VMAS survey has three questions and asks about adherence over the past seven days, so we hypothesize that patients may have been more likely to either have better adherence closer to a physician appointment or to have better recall about a single week period leading to a lower reporting of poor adherence on the VMAS. Because social desirability bias often leads to underestimation of medication adherence by self-report, it may be useful in practice to use the more sensitive measure of corneal medication adherence, the ARMS. There were differences in adherence within disease categories. Patients with corneal transplantations reported between 29–66% non-adherence. Given the risk of graft rejection with non-adherence to medications, this was a surprising finding, but parallels much seen in the glaucoma literature with risk associated with chronic medication use for a non-symptomatic condition. Surprisingly, patients with infectious keratitis reported high non-adherence with 46–80% non-adherence, depending on the survey measure used. Given the pain and acute nature of this condition, this area deserves further exploration. Perhaps the high burden of medication dosing for management of keratitis exceeds patient’s capacities especially if patients have fewer resources or are new to taking ocular medications. Looking deeper into the ARMS subscales, more than half of participants reported having two or more barriers to adherence. Participants reported “forgetting medications” or “being asleep” as two of the main reasons for missing doses, and corneal medications for acute conditions are often prescribed around the clock, which appears to be a significant barrier to optimal adherence. Additionally, research has found that when people report “forgetting” as the barrier to adherence, if clinicians dig a little deeper, they may find that it is not just “forgetting,” but that the person also has a problem with the medication.²¹

In this study, younger age was associated with lower odds of adherence by both ARMS subscales and VMAS. Participants that missed a medication dose in the last seven days due to being “out of a routine” were on average younger than those who did not. Park and colleagues found that middle-aged patients were at a higher risk of medication non-adherence due to their busy schedules and more hectic lifestyle.²² Older patients may be more adherent because they monitor their own health, are more vigilant in seeking healthcare, value managing their health, and have prior experience dealing with medications.^23–26 That said, older patients often are affected by multiple chronic illnesses and co-morbidities, are prescribed more medications.²⁷ Older adults have been shown to schedule daily activities around their medication regimen, in contrast to younger patients who try to accommodate their medication regimen into their own daily schedule, not prioritizing medications compared to other tasks.²² Consequently, younger patients allocate and utilize very little cognitive resources to remembering to maintain adherence.²² Reasons for poor medication adherence in young patients include being newly diagnosed,^28–30 lack of knowledge regarding their disease, dealing with maintaining a medication regimen, and the fear of side effects.^26,31,32 Interventions may need to be tailored to the different barriers younger and older people experience to be more salient for each age group.

More specific barriers will need to be explored in qualitative interviews for this patient population. In a 2006 study, barriers to adherence for 48 glaucoma patients was explored and classified into four categories: situational/environmental factors (sociocultural, racial, and ethnic factors) (49%), medication regimen factors (32%), patient-related factors (16%), and provider-related factors (3%).¹⁵ Psychosocial aspects have been found to influence patient’s medication adherence. For instance, a patient who has poor family or peer support, poor understanding of the prescribed medication instructions, or denial of the disease severity may have worse adherence.^15,33 This qualitative work could expand our knowledge of disease-specific issues for patients with corneal conditions.

In this study, participants reported missing doses of medication because of the high cost of medications. A nationwide study of older adults afflicted with chronic illnesses revealed that two-thirds of the patients chose not to tell their physician of their plan to underuse the prescribed medication due to its high cost.³⁴ Physician prescribing other ophthalmic medications often do not ask about medication affordability, likely an issue for this participant population as well.¹⁵ Our prior work indicates that patients with infectious keratitis have to take many medications and undergo many procedures, likely posing a significant financial burden on the patient.³⁵ In other work, cost-related non-adherence (to antihypertensive medications) was significantly higher for younger patients, so perhaps issues of age and cost are intertwined.³⁶

There are some limitations to our study. Data was collected at a single tertiary-care academic cornea clinic; a high portion of surveyed patients completed at least some college; and the patient population was predominately white, reflective of the local area. Second, patients self-reported adherence is subject to recall bias but perhaps biases patients towards reporting higher-than-actual adherence.³⁷ We attempted to offset this occurrence by emphasizing to patients that individual survey results would not be shared with their physicians. Third, the ARMS was validated for chronic illnesses so many not be generalizable to many corneal conditions.³⁸ Even so, there are many advantages to using a questionnaire to assess medication adherence. Questionnaires are inexpensive and a quick method to ask patients the necessary questions and receive real-time feedback.^37,39

Results from this study should emphasize to cornea specialists the need to have conversations with their patients about medication adherence as non-adherence was high for patients with corneal conditions. The ARMS questionnaire may be a good starting instrument to integrate into clinical practice to identify patients at high risk for poor outcomes due to poor adherence. The answers to the questionnaire may facilitate improved discussion about potential barriers the patient is facing to optimal adherence.

Supplementary Material

Supplemental Table 1

Supplemental Table 1: Severe & Non-severe Corneal Conditions- Primary cornea diagnoses categorized and listed as either ‘Severe Corneal Conditions’, or ‘Non-severe Corneal Conditions’ by an ophthalmologist.

Supplemental Table 2

Supplemental Table 2: Adherence to Refills and Medication Scale (ARMS) Subscales- A summary of adherence status by demographics and clinical information for ARMS subscales: adherence of taking the eye medication (ARMS Taking) and adherence to re-filling the eye medication (ARMS Filling).

Supplemental Table 3

Supplemental Table 3: Voils Medication Adherence Scale (VMAS) Supplementary Questions by Adherence Status as Determined by VMAS and Adherence to Refills and Medications Scale (ARMS)- A comparison of the two medication adherence measures, VMAS and ARMS, in determining the non-adherence status for patients.

Appendix A.

Appendix A. Adapted Adherence to Refills and Medications Scale (ARMS) survey for corneal conditions. A 13-item survey aimed to determine the patient’s adherence of taking the eye medication (ARMS Taking) and adherence to re-filling the eye medication (ARMS Filling).

Appendix B.

Appendix B. Adapted Voils’ Medication Adherence Scale (VMAS) survey for corneal conditions. An 18-item survey aimed to identify reasons for the patient’s medication non-adherence

Acronyms:

MK

Microbial Keratitis

HER

Electronic Health Record