The OPTIMA Trial is a 12-month randomized blinded controlled trial. The investigators assessing and analyzing the outcomes are blinded to the treatment assignment. The subjects in both arms, intervention and control, receive mailed educational materials. Subjects in the intervention arm also receive telephone counseling using a Motivational Interviewing approach. Because subjects in both arms receive enhanced care and do not know whether they are in the group receiving the more intensive regimen, subjects and investigators are blinded to treatment assignment. In this respect, the trial is double-blind. Randomization of subjects occurs centrally using a random number generator and is stratified by gender, allowing recruitment of similar numbers of men and women in both treatment arms.
Study Population and Recruitment Procedure
All subjects are Medicare enrollees and participate in Pennsylvania’s Pharmaceutical Assistance Contract for the Elderly (PACE), managed by the Pennsylvania Department of Aging. Persons who meet the annual income criteria and who are 65 years or older, receive prescription medications after paying a modest co-payment ($6–15 per drug per month). On a monthly basis, PACE identifies potentially eligible subjects who filled a new prescription to treat osteoporosis, including alendronate, calcitonin, ibandronate, raloxifene, risedronate, teriparatide, and zoledronic acid. A new prescription is defined as the cardholder’s first claim for an osteoporosis medication within the past 12 months. In addition, eligible subjects must be enrolled in PACE for at least 12 months, reside in a non-institutional setting, and not have a designated power of attorney.
Potentially eligible subjects receive an invitation letter giving them the opportunity to opt-out of any further contact by returning a letter or calling a toll-free telephone number. If no opt-out is received within two weeks, we attempt telephone contact for recruitment on at least three separate occasions at different times of the day. Potentially eligible subjects successfully reached by telephone are explained the goals of the study and asked to participate. Some potentially eligible subjects cannot be successfully contacted by telephone after three attempts or cannot communicate by telephone (severe hard of hearing or non-English speakers). After consent is received by telephone, subjects are assigned into treatment arm “A” or “B” (intervention or control) based on a randomization schedule generated by a random number program. Only the study coordinator is aware of intervention and control assignment. All study investigators and biostatisticians remain blind to treatment assignment.
Potentially eligible subjects sort into four categories: opt-out by letter or telephone call; unable to reach by mail or telephone; refuse participation; and consent to participate in the trial. We illustrate assembly of our study population (see ) and compare basic demographic and pharmacy data from these four categories for the initial seven months of recruitment (see ).
Baseline Characteristics of All Potentially Eligible Subjects During the First Seven Months of the OPTIMA Trial
During the first 7 months of recruitment, we have screened 3,638 potentially eligible subjects. After an initial mailing, 1,115 (30.6%) opted out of telephone recruitment and 1,019 (28.0%) could not be successfully contacted. Of the remaining, 879 (24.2%) consented to participate and were randomized while 625 (17.2%) refused participation. Characteristics of all potentially eligible subjects are described in . Women comprise over 90% of all groups, mean ages range from 77–80 years old, and the majority in all groups was white. The distribution of osteoporosis medications was comparable across groups and the median number of different prescription drugs used in the prior year was 8–10.
The entire study protocol was reviewed by the Partners Healthcare Institutional Review Board and deemed exempt based on the fact that the trial aimed to improve quality for beneficiaries of a federally-funded health care program (Medicare) and was performed in conjunction with a state-run pharmacy program (PACE).
Subjects in both the intervention and control arms receive mailed education specially designed for this trial (see Supplementary File
). The educational material comprises seven topics: basic information about osteoporosis and fractures; the appropriate use of osteoporosis medications; how to talk with your doctor about medications; fall prevention through home safety; sufficient calcium and vitamin D through diet and supplements; exercises to improve balance; and bone mineral density testing. These educational materials are written at a sixth-grade reading level, use 14-point font, are limited to less than two pages; and incorporate color graphics to make them attractive. All subjects in the intervention and control arms receive materials on one topic at a time every four to eight weeks. The schedule of topics mailed to the intervention arm is timed to coincide with the educational topics discussed by the counselors (see below).
In addition to this mailed material, subjects randomized to the intervention arm also receive one-on-one counseling by telephone. Four of the seven counselors are certified health educators and the others are health professionals with experience in patient counseling. The counselors have received 6–8 hours of education about osteoporosis and then periodic updates of new information. They are supported by a clinician (DHS) and are instructed to refrain from giving medical advice. They are instructed to answer informational questions about osteoporosis and medications. However, they are instructed to refer questions about treatment decisions back to the subject’s prescribing physician. In addition, they participated in workshops on Motivational Interviewing. The initial Motivational Interviewing workshop was a half-day interactive session with an experienced trainer. Counselors also participate in biweekly conference calls with the investigative team to discuss issues related to the use of Motivational Interviewing and questions arising from the subjects. In addition, Motivational Interviewing supervision is provided to the counselors every six months through the review of audiotaped subject telephone calls in a one-on-one discussion with a supervisor experienced in Motivational Interviewing.
During the 12-month trial, counselors are scheduled to make ten calls to subjects in the intervention arm. The first call aims to explain the counseling program, to build rapport with subjects, and to answer client questions. The remaining nine calls occur at set intervals and have specific themes noted in . Each call entails discussion of medication adherence -- exploring barriers to adherence, offering suggestions for overcoming barriers, and answering specific questions that subjects pose. Suggested scripts for each telephone call are supplied through a web-based counseling tool developed specifically for the trial.
List of Primary, Secondary and Exploratory Outcomes and Planned Analyses
For example, the counselor may be prompted to say, “would you mind telling me a bit about how you take your osteoporosis medication?” or “could you please share with me your experiences when taking your osteoporosis medication?” The counselor may also incorporate a simple reflection strategy which includes restating what the subjects reported such as, “when you take your Fosamax (alendronate), you sometimes feel sick to your stomach.” Another strategy used is reflection of meaning. Using this approach the counselor integrates the perspective of the subject into their refection. For example, “your osteoporosis makes you anxious about falling and having a hip fracture.” These comments are made in an open-ended manner allowing the subject to elaborate and problem-solve without instruction from the counselor. In addition, these strategies help maintain a client-focused approach to discussions and illuminate the subject’s underlying beliefs and concerns which may affect adherence.
The primary outcome for this trial is medication adherence at twelve months, the end of the study period (see ). Medication compliance will be measured as the medication possession ratio (MPR), calculated as the percentage of days in which the subject has an osteoporosis medication available for use during the follow-up period. Thus, the denominator will be the number of days of follow-up during the 12 month study period with 365 days being the maximum. Follow-up begins with the first successful telephone call and is censored at the first of any of the following events: loss of PACE eligibility; transfer to a nursing home; or death. For the first 100 days of a nursing home admission, patients’ medications are reimbursed by Medicare, not PACE. Since Medicare drug files are not available in a timely fashion, these data cannot be used for endpoint assessment and thus, follow-up is censored at nursing home admission.
The numerator in the MPR calculation is the number of days with available osteoporosis medication. Available medication is defined as the number of days of the medication dispensed by the subject’s pharmacy. When people fill a prescription, pharmacies submit a claim containing the name and dosage of the medication, the date dispensed, and the number of days supplied. As part of our collaboration with PACE on this intervention, we have access to pharmacy dispensing information for all enrollees. Thus, we are able to calculate the days with available osteoporosis medication from the day supply field. Since zoledronic acid is an annual medication, the MPR will be 100% by definition.
The secondary outcomes to be examined are listed in . Since a key question is the durability of the twelve-month intervention, we will assess the MPR at 18 and 24 months after enrollment. In addition to compliance, we are interested in persistence to therapy, with persistence defined as the time until a prolonged period without treatment. A prolonged period without treatment is defined as at least 90 days without available medication. Persistence will be assessed at 12, 18, and 24 months of follow-up. These definitions are consistent with current recommendations.(Cramer)
The most relevant clinical outcome for this medication adherence trial is fracture. We will assess fracture rates at the hip, wrist, humerus, pelvis, clavicle, tibia, ribs and spine. These will be assessed using health care utilization data from Medicare. However, because Medicare health care claims data require a 12–18 month lag period in their availability, we will also obtain self-report for these fractures at the study conclusion. Self-report of fractures has been found to be valid.23, 24
Secondary outcomes include nursing home admissions at 12 months, overall health care utilization at 12 months, and survival at 12 months. Each of these outcomes may relate to osteoporosis medication adherence and the telephone counseling. Since the relationship of these outcomes with the intervention is less direct, we consider these exploratory. Nursing home admissions will be considered as a dichotomous variable as well as cumulative days in nursing. Nursing home admission information will be extracted from Medicare utilization data. Overall health care resource use will include acute care hospitalizations, nursing home stays, rehabilitation stays, physician and emergency department visits, all medication dispensings, laboratory, and radiology use. Finally, mortality will be assessed from social security files.
Statistical Analyses and Power Considerations
The primary analysis will use an intention to treat (ITT) design. Thus, all subjects will be analyzed according to their randomization group, regardless of whether they participate fully in the intervention. A secondary analysis will include information about a subject’s level of participation in the intervention, represented by the number of telephone counseling sessions in which the subject participated. Because small imbalances in subject characteristics may obscure the effect of the intervention, we will include important baseline variables as covariates in a linear regression model to calculate the primary outcome. The linear regression will include the continuous MPR as the dependent variable with treatment assignment as the variable of interest. The baseline covariates will include age, gender, osteoporosis treatment, frequency of treatment (daily, weekly, monthly), and race.
We have calculated the required sample size for this trial based on the following assumptions. First, we estimate that an absolute improvement of 10% in osteoporosis medication MPR was likely the smallest clinically relevant difference.9
Second, we assumed that the control group would have an MPR of 50%. This is slightly higher than prior analyses suggest, but we anticipate that enrolled subjects may be more motivated than a typical population.7
Based on these assumptions, we will require 1050 subjects in each arm to have at least a 90% power to rule out a two-sided Type I error of 5%. This number of subjects provides less than 80% power to detect a 15% relative reduction in fractures in the intervention arm compared with the control arm (9% of 1050 = 95 versus 10.5% of 1050 = 110).
In addition to the primary outcomes, we are conducting a cost-effectiveness analysis side-by-side with the trial. This analysis aims to examine the economic implications of the intervention and its effect on adherence and fractures. Thus, we are collecting information about the cost of the intervention and the economic implications of different levels of adherence. As well, through a survey of a subset of subjects, we are collecting information about quality of life and the non-medical costs of osteoporosis.