Solomon 2012.

Methods	Randomized controlled trial
Participants	The study location was Pennsylvania, USA 1050 participants were randomized to the intervention group and 1047 participants were randomized to the control group The inclusion criteria were 65 years or more in age; enrolled in PACE for at least 12 months, reside in a non‐institutional setting, and do not have a designated power of attorney
Interventions	Intervention: TELEPHONE‐BASED COUNSELING Subjects in both groups started receiving educational material regarding osteoporosis 30 days after randomization. In addition the intervention group subjects received motivational intervening based counseling administered by a health educator. 10 sessions were targeted per subject Control: CONTROL Approximately 30 days after randomization, all subjects received mailings regarding osteoporosis. During the study, all subjects received 7 informational mailings covering topics such as exercise, fall prevention, and recommended calcium intake
Outcomes	The measures of adherence were self reported medication adherence assessment and pharmacy fill record. Self reported medication adherence assessment were done through phone calls at 1, 3, 7, 11, 17, 23, 31, 39, and 52 weeks. Pharmacy fill records were collected from the pharmacy. They contain information about the name and dosage of the medication, the date dispensed, and the number of days supplied. Using this the days with available osteoporosis medication from the day supply field was calculated. 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 The patient outcomes were self reported fractures, self reported falls, and general health. All of these outcome are self reported. Fractures assessment was done at 12 months, falls assessment at weeks 7 and 31 and general health assessments were done at weeks 0, 11, and 39
Notes	The large discrepancy between groups in missing data is unexplained and might be suggestive of logistic issues for intervention delivery. Also, note that the level of intervention participation is not reported. Depression and satisfaction are not reported although they have been mentioned as secondary outcomes. General health as a patient outcome is mentioned in the supplementary but not in the methods section of primary but results reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization of subjects occurred centrally using a random number generator and was stratified by gender, allowing recruitment of similar numbers of men and women in both treatment arms (pg 137 (s1))
Allocation concealment (selection bias)	Low risk	Randomization of subjects occurred centrally. Subjects were assigned into treatment arm "A" or "B" (intervention or control) based on a randomization schedule generated by a random number program. Only the study co‐ordinator was aware of intervention and control assignment. All study investigators and biostatisticians remained blind to treatment assignment (pg 139 (s1))
Selective reporting (reporting bias)	High risk	They failed to report the health care utilization data from Medicare ‐ these were an indication of fracture rates(pg 142 (s1))
Other bias	Unclear risk	"Our follow‐up may have been inadequate to detect a change in fracture rate attributable to a modest change in medication use." Large imbalance between groups in dropouts. 16 in control versus 204 in intervention; no reason given. (pg 481‐2) "The use of routinely collected data presents important methodologic challenges. We relied on pharmacy claims data to identify recent initiators of an osteoporosis medication regimen. These data needed to be processed, and then we allowed potential subjects to opt out of recruitment. This meant that, by the time subjects were recruited and received their first intervention call, a median of 113 days had passed since they filled their first prescription. Although data shown in Figure 3 suggest that the intervention's effect increased during the study, the lag period may have limited the benefit of our intervention because many subjects had already discontinued use of their osteoporosis medication at the time of their first telephone call." (pg 481)
Blinding of outcome assessment (detection bias) Adherence measure	Low risk	(PRIMARY) PHARMACY FILL RECORD ‐ Only the study co‐ordinator is aware of intervention and control assignment. All study investigators and biostatisticians remain blind to treatment assignment (pg 138). "The investigators assessing and analyzing the outcomes are blinded to the treatment assignment." (s1)
Blinding of outcome assessment (detection bias) Patient outcome	Low risk	(PRIMARY) SELF REPORTED FALLS, GENERAL HEALTH ‐ "The investigators assessing and analyzing the outcomes are blinded to the treatment assignment." (pg 138)
Blinding of participants (performance bias) Adherence measure	Low risk	(PRIMARY) PHARMACY FILL RECORD ‐ pg 139 indicates that patients were blinded. Also, because all subjects in both arms received enhanced care, they were not aware of their treatment arm allocation (478, primary)
Blinding of participants (performance bias) Patient outcome	Low risk	(PRIMARY) SELF REPORTED FALLS, GENERAL HEALTH ‐ All subjects in both arms received enhanced care, therefore they were not aware of their treatment arm allocation (478, primary); pg 138‐9 says that subjects were also blinded
Blinding of personnel (performance bias) Adherence measure	Unclear risk	(PRIMARY) PHARMACY FILL RECORD ‐ Depending on who we consider to be the key study personnel the risk of bias will vary. If the study co‐ordinator, who was aware of intervention and control assignment, is the key personal then there will be a high risk of bias; on the other hand, if the investigators and biostatisticians are considered key personnel ‐ then there would be a low risk of bias
Blinding of personnel (performance bias) Patient outcome	Unclear risk	(PRIMARY) SELF REPORTED FALLS, GENERAL HEALTH ‐ No information regarding other personnel other than the study co‐ordinator
Incomplete outcome data (attrition bias) Adherence measure	Low risk	(PRIMARY) PHARMACY FILL RECORD ‐ ITT analysis is used and as the author notes, "...these are old people who died. We had more complete death data for intervention participants than for controls. Controls who died were considered loss to followup but not drop outs."
Incomplete outcome data (attrition bias) Patient outcome	Low risk	(PRIMARY) SELF REPORTED FALLS, GENERAL HEALTH ‐ imbalance across the groups. ITT analysis done