Extent of and Reasons for Osteoporosis Medication Non-Adherence Among Veterans and Feasibility of a Pilot Text Message Reminder Intervention

Abstract

Purpose:

To evaluate the extent of and reasons for non-adherence to oral bisphosphonates among Veterans and to assess the acceptability and feasibility of a pilot text message reminder application.

Methods:

We surveyed 105 Veterans initiating oral bisphosphonates for osteoporosis/osteopenia within the prior 18 months utilizing a validated self-report measure adapted for osteoporosis. Additionally, we conducted a pilot text message reminder to determine feasibility in 12 Veterans who were initiating or were currently non-adherent to oral bisphosphonates.

Results:

Of the 43 (40.9% response rate) completed surveys, the most common reasons for non-adherence were “I forgot” (37.5%), “I had other medications to take” (20.5%), “my bones are not weak” (18.4%), “I felt well” (18.4%), and “I worried about taking them for the rest of my life” (17.9%). Median MPR for the 49 (46.7%) non-adherent (MPR< 0.80) Veterans was 0.35 (IQR 0.21–0.64). Of Veterans offered a weekly automated text message reminder, 12 (50%) accepted. Nine of these 12 Veterans reported the text message reminders did “very well” at reminding them to take their medication and would recommend the application to other patients/family/friends. The median 6-month MPR for the reminder group was 0.96 (IQR 0.54 – 1.00).

Conclusion:

Half the Veterans in our sample were taking insufficient doses of oral bisphosphonates to attain the full benefit of fracture risk reduction. Reasons for poor adherence included forgetfulness, polypharmacy, and misconceptions about osteoporosis. A pilot text message reminder intervention targeted to one of the most commonly cited reasons was found to be acceptable and feasible among Veterans.

SUMMARY:

We determined the extent of and reasons for non-adherence to oral bisphosphonates among Veterans and conducted a pilot text message reminder application aimed at the most commonly cited reason for non-adherence. The intervention was found to be acceptable and feasible.

Introduction

Medication non-adherence is a major concern for those with chronic medical conditions, such as osteoporosis. Multiple studies have shown treatment of osteoporosis with oral bisphosphonates to be safe and effective, reducing risk of fracture by up to 50%[1, 2]. Nevertheless, there remains a significant gap in implementation and adherence of bisphosphonate therapy in the clinical setting. Patients with medication non-adherence defined by a medication possession ratio (MPR) < 80% have higher rates of fracture than those with MPR ≥ 80%[3]. Importantly, only 40–85% of patients are adherent with an MPR ≥ 80% 1 year after starting oral bisphosphonates [3]. Gaining insight into reasons for non-adherence will guide future intervention studies to improve adherence.

Medication non-adherence is multi-factorial and often behavioral. A validated, self-report measure of non-adherence developed by Voils assesses the extent of and reasons for medication non-adherence for other chronic medical conditions such as hypertension, dyslipidemia, and hepatitis C[4]. The initial validation of this self-report measure showed that “I forgot” was the most commonly cited reason for non-adherence to anti-hypertensive medications (27%)[4]. Several studies have evaluated the effectiveness of various interventions, such as educational materials, counseling sessions, and alarm clock/calendar reminders, targeted to improve adherence and persistence to osteoporosis treatment, with variable and generally modest effectiveness[5, 6]. Few of these studies included men and none assessed a text message reminder.

In the current study, we assessed the extent of and reasons for non-adherence to osteoporosis medications utilizing Voils self-report measure of non-adherence adapted for osteoporosis among older male Veterans with osteoporosis. We hypothesized that forgetfulness would be the most commonly cited reason for poor adherence. A second aim of the current study was to assess the acceptability and feasibility of a text message reminder for adherence to weekly oral bisphosphonates.

Methods

Study Design

The current study was conducted in 2 phases: 1) a cross-sectional survey assessing the extent of and reasons for non-adherence to oral bisphosphonates at the Durham Veterans Affairs (VA) Medical Center; 2) a pilot study of a text message reminder application for Veterans who were newly initiating an oral bisphosphonate or had been non-adherent to a current prescription. The study was approved by the Durham VA Medical Center Institutional Review Board.

Cross-sectional study

Between January 2018 and May 2019, a consecutive series of 105 Veterans age 50 years and older with a diagnosis of osteoporosis or osteopenia were identified through pharmacy records and a Fracture Liaison Service as having been initiated on an oral bisphosphonate at the Durham VA Medical Center.

To assess the extent of non-adherence, MPRs were calculated for all these individuals from the pharmacy records as the sum of days’ supply dispensed divided by the total number of days since their initial prescription. Veterans were defined as adherent if MPR ≥ 0.80 or non-adherent if MPR < 0.80. This cut point was selected based on several prior studies demonstrating that fracture risk increases exponentially below this threshold [7, 8].

The extent of and reasons for non-adherence were assessed through the self-report measure developed by Voils [4] adapted for osteoporosis. These surveys were mailed to the 105 Veterans identified above. If the survey had not been completed and returned after 1 month, each Veteran was called to administer the survey over the phone. This measure has been validated for use in chronic medical conditions, such as hypertension[4], dyslipidemia[9], and hepatitis C virus[10]. It consists of 2 parts, the first part of 3 items assesses the extent of non-adherence to medication over the previous 7 days. Due to the weekly dosing of oral bisphosphonates, this was changed to assess the extent of non-adherence to osteoporosis medication over the previous 3 months. Three extent statements assessed non-adherence: “I took all doses of my osteoporosis medication”(reverse scored), “I missed or skipped at least one dose of my osteoporosis medication”, and “I was not able to take all my osteoporosis medication”. Responses were recorded based on a 5-point scale ranging from “never” (1) to “always” (5). Scores were recorded as the average of the 3 responses. Patients were considered as reporting non-adherence based on a score of >1. The second part of 25 items assesses reasons for non-adherence. Three of the 25 items were adapted to address the specific side effects associated with oral bisphosphonates including heartburn, atypical fractures, and osteonecrosis of the jaw/dental issues. Results of the survey were used descriptively.

Pilot study

Veterans were recruited between June 2019 and September 2019 for the text message reminder application. Eligible Veterans were either: 1) started on oral bisphosphonates within the prior 3 months (new starts); or 2) identified as currently non-adherent based on an active prescription for oral bisphosphonate with a medication possession ratio (MPR) since initiation of < 0.80. MPR was calculated from pharmacy records as described above.Veterans were required to have a text-capable phone or smartphone.

Enrolled Veterans were assisted to sign up for the Script Your Future text message reminder application, and received a once weekly automated text message reminder to take their oral bisphosphonate. Script Your Future is a medication adherence campaign with tools to improve adherence including a free online text-message reminder program at scriptyourfuture.org. Veterans were able to select the day, time, and wording of the text message reminder, and the study staff enrolled them in this free program online by simply entering these selections and their cell phone number. One month after enrollment, participants were contacted to ensure they were receiving the text message reminders. MPRs were calculated at 6 months after enrollment based on pharmacy refill records. At the completion of the study, participants were administered a survey with the questions listed in Table 5. The aim of the survey was to assess usability, utility, barriers, and satisfaction with the application.

Table 5.

Patient satisfaction survey (n = 12)

Statement	n/Na (%)
Did you receive the weekly text messages?
All the messages	10/12 (83.3%)
Most of the messages	1/12 (8.3%)
Some of the messages	0/12
None of the messages	1/12 (8.3%)
How well did you find the text messages reminded you to take your medication?
Very well	9/10 (90%)
Well	0/10
Uncertain	0/10
Poorly	0/10
Very poorly	1/10 (10%)
Were the reminders as useful at the end of the intervention as at the beginning?
Not as useful	2/10 (20%)
As useful	7/10 (70%)
More useful	1/10 (10%)
How likely are you to continue using the text message reminders?
Very likely	8/10 (80%)
Somewhat likely	0/10
Uncertain	1/10 (10%)
Somewhat unlikely	0/10
Very unlikely	1/10 (10%)
How likely are you to recommend using this application to other patients or family/friends as a reminder to take medications?
Very likely	8/10 (80%)
Somewhat likely	1/10 (10%)
Uncertain	1/10 (10%)
Somewhat unlikely	0/10
Very unlikely	0/10

^an is the number of patients reporting the response, N is the number of patients answering the survey question

The main outcome for this portion of the study was acceptability and feasibility of the text messages which were measured by the proportion of Veterans accepting the intervention, reasons for declining the intervention, persistence with the intervention for the duration of the study, and patient experience/satisfaction. Due to small sample size, an exploratory secondary outcome was MPR at 6 months. As a comparison, the median MPR at 3 to 9 months for all Veterans recently started on oral bisphosphonates at the Durham and Richmond VAs between March 2019 – December 2019 was calculated from pharmacy records.

Statistical Analysis

We used descriptive statistics for reasons for non-adherence obtained through survey responses, and to define the feasibility and acceptability of the reminder application. Median MPR for the comparison group was calculated from pharmacy records by the sum of days’ supply divided by the total number of days in the interval. MPR was an exploratory descriptive outcome considering the small sample size and statistical testing was not performed. We used descriptive statistics to define the proportion of Veteran responses in the categories of usability, utility, barriers, and satisfaction.

Results

Among all 105 Veterans prescribed an oral bisphosphonate in the preceding 1 to 17 months, the median MPR was 0.82 (IQR 0.39–1.00) and nearly half (46.7%) had MPR < 0.80. Of the non-adherent Veterans (MPR < 0.80) the median MPR was 0.35 (IQR 0.21 – 0.64). Surveys were sent to all 105 Veterans and 23 (21.9%) were completed and returned by mail. Twenty additional surveys were completed over the phone for a total of 43 (40.9%). Baseline characteristics of Veterans for the survey portion of the study are presented in Table 1. Of Veterans who completed the survey, 34.9% were non-adherent as defined by MPR. However, a higher proportion of Veterans were identified by the survey self-report as non-adherent (53.5%). The most common reason for non-adherence was “I forgot” (37.5%), Table 2. The other most cited reasons for non-adherence were “I had other medications to take” (20.5%), “my bones are not weak” (18.4%), “I felt well” (18.4%), and “I worried about taking them for the rest of my life” (17.9%).

Table 1.

Baseline Characteristics for Patients Surveyed

	Total (n = 105)	Responders (n = 43)
Mean Age (SD)	73 (10)	71 (9.7)
Female, n (%)	23 (21.9%)	11 (25.6%)
Mean MPR (SD)	0.71 (0.3)	0.79 (0.3)
Median MPR (IQR)	0.82 (0.39 – 1.00)	0.98 (0.64 – 1.00)
Adherent (MPR ≥ 0.80) n = 56	1.00
Non-adherent (MPR < 0.80) n = 49	0.35 (0.21 – 0.64)
MPR < 80, n (%)	49 (46.7%)	15 (34.9%)

Table 2.

Survey Results (n = 43)

Statement	n/Na (%)
I forgot, n/N (%)	15/40 (37.5%)
I had other medications to take, n/N (%)	8/39 (20.5%)
I felt well, n/N (%)	7/38 (18.4%)
My bones are not weak, n/N (%)	7/38 (18.4%)
I worried about taking them for the rest of my life, n/N (%)	7/39 (17.9%)
I felt I did not need them, n/N (%)	5/38 (13.2%)
I ran out of medication, n/N (%)	5/38 (13.2%)
I was busy, n/N (%)	5/40 (12.5%)
The time to take them was between my meals, n/N (%)	4/38 (10.5%)
They make me have heartburn, n/N (%)	4/38 (10.5%)
They cost a lot of money, n/N (%)	4/39 (10.3%)
I did not have any symptoms of osteoporosis, n/N (%)	4/40 (10%)
The medication caused side effects, n/N (%)	4/41 (9.8%)
I was traveling, n/N (%)	3/37 (8.1%)
I was afraid the medication would interact with other medication I take, n/N (%)	3/38 (7.9%)
I didn’t want to, n/N (%)	3/38 (7.9%)
I was afraid of becoming dependent on them, n/N (%)	2/39 (5.1%)
I was afraid they may cause fractures, n/N (%)	2/39 (5.1%)
I was afraid they may cause jaw/dental problems, n/N (%)	2/39 (5.1%)
I was supposed to take them too often, n/N (%)	1/38 (2.6%)
I was feeling too ill to take them, n/N (%)	1/38 (2.6%)
I was going on a long car/bus/plane ride, n/N (%)	1/38 (2.6%)
I was with friends or family, n/N (%)	1/40 (2.5%)
I came home late, n/N (%)	0/40
I was in a public place, n/N (%)	0/40

^an is the number of patients reporting this reason for non-adherence, N is the number of patients that answered this question on the survey

Twenty-nine Veterans were identified by the Bone Health Service or the pharmacy records as eligible for the text reminder application because they were new starts or non-adherent as defined by MPR <0.80 in the last 6 months. Of those 29 Veterans, 12 accepted the intervention, 12 declined, and 5 could not be reached. Of the reasons for declining the intervention, most Veterans reported they “did not need a reminder” or were “not interested” in the intervention/participating in the study, Table 3. The 12 Veterans enrolled in the application were mostly male (92%), White (75%), and had a mean age of 70 (7.3) years, Table 4. Ten Veterans had a prior fragility fracture. The 12 patients who declined the intervention had a mean age of 69 (7.0) years, were mostly male (75%), White (58%), and 4 had a prior fragility fracture.

Table 3.

Reasons for declining the text message application (n = 12)

Statement	n/Na (%)
I don’t need a reminder, n (%)	3/12 (25%)
Not interested, n (%)	3/12 (25%)
I don’t like texts/cellphone, n (%)	2/12 (16.7%)
No reason, n (%)	2/12 (16.7%)
Stopped medication due to side effects, n (%)	1/12 (8.3%)
No compensation for study participation, n (%)	1/12 (8.3%)

^an is the number of patients reporting the reason, N is the number of people who declined

Table 4.

Text Message Application Pilot Baseline Characteristics

Characteristic	Pilot Patients N=12
Age
Mean (SD)	69.9 (7.3)
Range	59 – 84
Sex, n (%)
Male	11 (91.7%)
Female	1 (8.3%)
Race, n (%)
White	9 (75%)
Black	2 (16.7%)
Unknown	1 (8.3%)
Duration on oral bisphosphonates, n (%)
New starts (< 3 months)	9 (75%)
Non-adherent (4 – 18 months)	3 (25%)
Prior fracture, n (%)	10 (83.3%)

At the completion of the study all but one Veteran continued to receive the text message reminders. This Veteran was not able to receive the text messages due to his cell-phone capabilities, and had an MPR of 0.34. Another Veteran stopped taking his medication at the recommendation of his provider due to extensive dental work. Although sample size precluded exploration of subgroups, qualitatively we observed 67% of new starts versus 33% with prior nonadherence achieving an MPR ≥ 0.80.

Ten Veterans completed the patient satisfaction survey at the end of the study. One Veteran reported at follow up that he had multiple other medications to take, did not fall, and did not feel he needs the osteoporosis medication. That Veteran’s MPR was 0 and he did not complete the survey. The one Veteran who did not receive any of the text message reminders could therefore not complete the survey. Of the 10 Veterans who completed the survey, 9 reported that the text messages did “very well” at reminding them to take their medication, Table 5. Eight Veterans reported the reminders were “as useful/more useful” at the end of the intervention as at the beginning, and they were “very likely” to continue using the text message reminders after the study concluded. Nine of 10 Veterans reported they were “very likely/somewhat likely” to recommend this application to other patients or family/friends as a reminder to take their medication.

The median MPR for the intervention group at 6 months was 0.96 (IQR 0.54 – 1.00). Seven (58.3%) Veterans had an MPR > 0.80 at 6 months. In comparison, the median 3 to 9-month MPR for all Veterans on oral bisphosphonates at the Durham and Richmond VA Medical Centers between March 2019 – December 2019 was 0.67 (IQR 0.39 – 1.00) with 30 of 67 (44.7%) having MPR ≥ 0.80.

Discussion

Fragility fractures due to osteoporosis are associated with increased morbidity, mortality, and costs. In order to attain the maximum benefit, oral bisphosphonates must be taken correctly and consistently. Adherence to osteoporosis medications and oral bisphosphonates, in particular, is poor[3]. In our sample, nearly half the Veterans had inadequate adherence by MPR and more expressed poor adherence by self-report. Similar to prior studies of nonadherence in women on bisphosphonates, the most common reasons given for inadequate adherence in this sample of mostly male Veterans were in the categories of forgetfulness, polypharmacy, and misconceptions about osteoporosis. These findings will help tailor interventions aimed at improving adherence to osteoporosis medications, such as our pilot text message reminder application, which was easy to implement and well accepted among the participants.

MPR is an indirect measure of medication adherence as it measures medications dispensed and cannot ensure a patient took the medication or took it as prescribed. Other indirect methods of measuring medication adherence include pill count, medication event monitoring systems (MEMS), and self-report.[11] Each method has its own advantages and disadvantages. A strength of this study is the use of different measures of adherence including utilizing VA pharmacy records and a self-report survey. By utilizing these two measures we were able to identify those who would be considered adherent by MPR but who were not taking their medications as prescribed.

Similar to prior studies, one of the most commonly cited reasons for medication nonadherence in our sample was forgetfulness[4]. Among 33 studies identified between 1999 and November 2017 evaluating interventions to improve osteoporosis medication adherence, only 2 studies assessed the impact of reminders on secondary adherence[5, 6]. One was a small nonrandomized study showing improvement in 12 month adherence with use of an alarm clock as a prompt to take a bisphosphonate[12]. The other study was a 3-arm RCT including women with postmenopausal osteoporosis comparing usual care to educational booklets, calendars, and an alarm clock to these interventions plus phone calls from trained providers on prespecified osteoporosis topics. The investigators found no difference in adherence at 12 months between groups [13]. In another more recent review[14], 2 additional studies assessing telephone reminders were identified. One study utilized 10 minute phone calls every 2 months to motivate women to maintain good adherence to an oral bisphosphonate or strontium ranelate and if poor adherence was detected the patient was encouraged to follow up with their primary care provider. Significantly more patients in the intervention group were adherent at 1 year[15]. In the second study of all women with recent fracture attending a Fracture Liaison Service, telephone calls were made at 1, 4, and 12 months to remind patients to take oral alendronate. No difference in medication persistence at 12 months was seen [16]. To our knowledge, a text message reminder to take weekly bisphosphonates has not been studied.

A recent review identified 52 studies assessing medication regimen adherence with text message reminders [17]. Text messages reminders can also be used to improve appointment attendance and to help with behavior changes such as smoking cessation and weight loss. Text message reminders have been studied in other chronic medical conditions such as coronary artery disease, the human immunodeficiency virus, asthma, epilepsy, and diabetes with an overall positive impact on adherence[18]. However, the median age of 39 (age range, 31 – 64) in these prior studies is much lower than in our current study with mean age of 70. With a growing proportion of older patients having access to smartphones, this low-cost intervention may be an attractive option to improve adherence, especially for medications with weekly or less frequent administration schedules. Medication reminders may be helpful in those with memory or cognitive impairment such as older patients or those with mental health conditions. Despite the number of benefits of this type of intervention, there are some drawbacks including potential issues with patient confidentiality, alert fatigue, and the possibility of patients not reading or ignoring the messages.

The other most cited reasons for medication non-adherence identified in the current study were in the categories of polypharmacy (“I had other medications to take”) and misconceptions about osteoporosis (“my bones are not weak”, “I felt well”, and “I worried about taking them for the rest of my life”). Polypharmacy is highly prevalent among older adults, and deprescribing non-essential medications and those with poor risk-benefit may reduce the pill burden described by our patients. To our knowledge, deprescribing interventions which reduce pill burden have not tested whether they also improve adherence with medications for other chronic medical conditions such as osteoporosis. [19]. Addressing misconceptions about osteoporosis is a common target for interventions aimed at improving osteoporosis medication adherence with the use of various patient education and counseling practices but has not been consistently shown to be beneficial[5, 6].

The results of our pilot study show the acceptability and feasibility of the text message reminder application. Although only half of the potential Veterans agreed to participate, none dropped out, and the majority found the intervention beneficial. The intervention was easy to implement: the text message reminders could be set up within a few minutes and required no additional management by study staff. This intervention could be easily incorporated into a Fracture Liaison Service or clinical practice to enhance adherence without additional cost or labor. Further, most Veterans found the text message reminders were as useful at the end of the study as at the beginning, suggesting alert fatigue was not an issue during the 6 to 9-month duration of the study. Future studies to examine the impact on long-term adherence are warranted. It is also unclear from this study if a certain subgroup of patients, such as those reporting forgetfulness as a reason for nonadherence, would benefit more from this type of intervention.

Other limitations of this study include generalizability, selection bias, and response bias. While the sample was obtained through pharmacy records and the Bone Health Service to be a representative sample of all patients at our center who are prescribed bisphosphonates, the study was among a population of Veterans at a single center and may not be generalizable to other populations. Specifically, the pilot text message study was among mostly male Veterans with a prior fragility fracture and may not be generalizable to females or patients who have not fractured. Further, since the self-report surveys assessing extent of and reasons for adherence were completed by less than half of the sample, selection and response bias may be a factor influencing the results. Similarly, selection and social desirability bias may play a role in patients’ perceptions of the pilot text message reminder application as patients accepting the intervention and reporting directly to a study investigator may be more likely to rate it favorably. A randomized controlled trial or nonrandomized study with matched control group may have provided a better estimate of the true effect of this intervention. However due to the nature of the study and the primary aim being acceptability and feasibility, a small sample size was attained, and MPR was compared to the mean MPR among similar patients concurrently in the same population. Measuring adherence with MPR is another limitation of the study as MPR does not reflect whether a patient is actually taking a medication or taking it correctly. Half the Veterans offered the intervention declined, while this may limit the feasibility of the intervention, only 5 of 12 declined due to the nature of the intervention (2 reported not liking texts/cellphones, 3 reported not needing a reminder). The overall small sample size in this study limits the reliability of the conclusions. Future studies should include a larger and more diverse selection of patients to improve generalizability and confirm our findings. Lastly, the duration of the study does not allow for the durability of the effect on adherence to be evaluated. Considerations for future studies include a longer duration of follow up and assessing the effect on clinical outcomes such as fractures or bone density.

Conclusion

Adherence to oral bisphosphonates within the first 18 months is poor, with only half of older, primarily male Veterans taking sufficient doses to attain the full benefit in terms of fracture risk reduction. Reasons cited for poor adherence among our sample included forgetfulness, polypharmacy, and misconceptions about osteoporosis. A pilot text message reminder intervention targeted to one of the most commonly cited reasons for non-adherence to oral bisphosphonates, forgetfulness, was found to be acceptable and feasible among Veterans.