A total of 270 PAs were analyzed for 193 unique patients, resulting in 137 PAs for the intervention group and 133 PAs for the historical comparison group. Twenty-three classes of medications were analyzed for the intervention group and 20 for the historical control group, with anticonvulsants and gastrointestinal agents being the most common in both. PAs analyzed were for patients aged 1 to 30 years (median, 13 years) and 3 to 47 years (median, 13 years) for the intervention and historical comparison groups, respectively.

Within the intervention group, 119 PAs (87%) had a positive outcome and 18 PAs (13%) had a negative outcome (Figure). In the historical comparison group, 33 PAs (25%) had a positive outcome and 100 PAs (75%) had a negative outcome (Figure). The intervention group was 3.5 times more likely to have a positive outcome vs the historical comparison group (χ2 = 106; P  .00001).10 Overall, the relative risk of having a positive outcome was 3.5 (95% CI, 2.6-4.7) (Table 3).

Outreach was conducted on 3 separate days, and the impact of each outreach day was evaluated based on the percentage of PAs submitted. The total outreach attempts for days 1, 4, and 7 were 137, 104, and 101, respectively. The percentages of PAs submitted following outreach on days 1, 4, and 7 were 12% (n = 16), 28% (n = 29), and 48% (n = 48), respectively. Of the 119 PAs with a positive outcome in the intervention group, 93 (78%) were associated with a new PA submission and 26 (22%) had provider rationale for a lack of a new PA submission. Among the 93 new PA submissions, 16 PAs (17%) were submitted after 1 attempt, 29 PAs (31%) were submitted after 2 attempts, and 48 PAs (52%) were submitted after 3 attempts.

PAs within the intervention group were further analyzed based on types of outreach and outcomes. The majority of PAs (75%) resulted in successful outreach and a new PA submission, whereas 9% of the PAs resulted in successful outreach and no new PA submission. Seven percent of PAs had successful outreach and a provider rationale for no new PA submission, whereas 4% of PAs had successful outreach, but the patient was no longer active at the practice site. Additionally, 4% of PAs had modified outreach and a new PA submission and 1% of PAs had unsuccessful outreach (Table 2). Among these new PA submissions, 98% were subsequently approved. On average, PAs were submitted a mean (SD) of 3.5 days (10.7 days) prior to the expiration date in the intervention group vs 13.0 days (12.0 days) after the expiration date in the historical comparison group (t = –7.50; P  .00001).


This pre-post evaluation indicated that the proactive intervention resulted in a 3.5-fold increase in new PA submissions and/or provider response (ie, positive outcome) compared with a historical comparison group. Among the intervention group, the majority (78%) of outreach attempts resulted in a new PA submission, of which 98% of PAs were subsequently approved.

The majority of new PAs submitted in the intervention group were after outreach days 4 and 7, illustrating that multiple forms of outreach (eg, phone call, fax) and multiple attempts may be necessary in prompting new PA submissions and/or provider response. Although telephonic outreach may be more labor intensive compared with other types of outreach (eg, fax, mail), it can facilitate dialogue between the provider and the health plan. Through these conversations, we were able to understand the provider’s perspective in discontinuing the medication, document any rationale for delaying the PA submission (eg, awaiting laboratory assessments), and determine whether patients were no longer under the care of the provider. In our initial telephonic outreach, we confirmed the provider’s current fax number. If fax outreach was required, we sent the appropriate patient health information (PHI) to the provider. Because fax numbers are updated in our system only when a new PA is received, conducting the telephonic outreach prior to the fax outreach was important to prevent PHI breaches and follow Health Insurance Portability and Accountability Act best practices. Incorporating multiple attempts and forms of outreach can convey the urgency of the message for an upcoming potential disruption in therapy if a new PA is not submitted.

Although payers advise providers to track and actively assess their patients’ PAs, many offices are unable to do so effectively, resulting in treatment disruptions for patients. By implementing a proactive program, health plans can reduce the administrative burden of providers tracking PAs and allow them to engage in direct patient care. Not only did the intervention prompt new PA submissions, it may have also prompted providers to reassess the necessity of the medication. When conducting outreach, we received feedback that the outreach prompted the provider to assess the patient’s continued need for the medication.

Some MassHealth patients are enrolled in the CCM program based on their need for continuous skilled nursing hours. These patients are “at risk” because they are medically complex and frequently utilize PAs; however, this intervention program is not limited to serving only this population. This program can be implemented for other at-risk populations and stratified by age (eg, pediatric patients), patients utilizing high-cost specialty medications (eg, hepatitis C medications), or patients exceeding a specific PA threshold (eg, patients with more than 5 active PAs). Expanding this program to additional populations could aid both providers and health plans in minimizing treatment delays.

Poor medication adherence contributes to a variety of undesirable effects for the patient, including worsening of disease, and can increase medical costs for both patients and the health care system. A lack of an active PA on file to fill a medication is one risk factor contributing to medication nonadherence. The proactive intervention was associated with new PA submissions before the PA expiration date, whereas in the historical comparison group, new PA submissions were submitted after the PA expiration date. This highlights the impact of the intervention in both increasing overall PA submissions and influencing the timeliness of PA submission, thus serving as an additional opportunity to minimize potential medication access delays.

Further studies could examine proactive PA interventions for additional at-risk populations. Additionally, studies may also directly compare various forms of outreach (eg, mail, telephone, fax, email), frequency of outreach (eg, weekly, twice weekly), and the total number of attempts to determine the most effective approaches. Lastly, further studies can evaluate the intervention’s impact on total medical expenses (eg, hospitalizations, ED visits).


Due to potential eligibility changes for CCM patients across years, one limitation is that the PAs included in the intervention and historical comparison groups were not for the same patients. For example, a patient who required a PA in the historical comparison group may not have met the inclusion criteria for the intervention group based on their eligibility for CCM and the methodology. This accounted for the differences in number of PAs analyzed among both groups. Additionally, providers were inconsistent between the intervention and historical comparison groups, which potentially could result in differences in provider practices and responses to outreach.

Because the historical comparison group utilized retrospective data, the reason for no new PA submissions (eg, clinical rationale for discontinuing medication, patient lacking active relationship with provider) could not be determined. With this information, there may have been an increase in positive outcomes among the PAs in the historical comparison group.

Within the intervention group, multiple sets of outreaches had state and federal holidays (eg, Christmas, New Year’s Day) on which provider’s offices were closed, thus increasing outreach attempts and potentially delaying PA submissions. Additionally, prior to outreach, the status of patient-provider relationship for the PAs could not be determined, thus resulting in additional outreach attempts and negative outcomes. Furthermore, outreach was not conducted for 6 PAs due to unverifiable phone and fax numbers for the provider.


This proactive intervention to providers regarding expiring PAs in medically complex patients in a state Medicaid population supports the use of an outreach program to increase new PA submissions and preserve continuity of care. Proactive outreach resulted in significantly more PA submissions and a reduction in time to PA submission, both of which aid in decreasing the time during which patients do not have access to medications. Our findings provide important information for payers to guide the development of outreach programs to enhance continuity of care among a target population. Considerations identified include types of effective interventions (eg, telephonic, fax), frequency of interventions (eg, weekly, biweekly), total number of interventions, and potential patient populations in which to conduct the intervention. Implementing an outreach program could benefit providers in reducing administrative burden, health plans in monitoring medications, and patients in ensuring access to necessary medications.


The authors would like to acknowledge the following members of the Clinical Pharmacy Services Special Populations Pharmacist Team of UMass Chan Medical School–Commonwealth Medicine for conducting provider outreach: Anna Libman, RPh; Cynthia Vahey, RPh; and Hanh Hoang, PharmD, RPh.

Author Affiliations: UMass Chan Medical School–Commonwealth Medicine (SV, SNT, TCP, KCB, MT, MP, CJA, LM, KMC), Shrewsbury, MA; Office of Clinical Affairs, MassHealth (KL), Quincy, MA.

Source of Funding: None.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (SV, SNT, TCP, KCB, MT, MP, CJA, LM, KMC, KL); acquisition of data (SV, LM); analysis and interpretation of data (SV, SNT, TCP, KCB, MT, MP, CJA, KMC, KL); drafting of the manuscript (SV, SNT, TCP, KCB, MT, MP, CJA, KMC, KL); critical revision of the manuscript for important intellectual content (SV, SNT, TCP, KCB, MT, MP, CJA, KMC, KL); statistical analysis (SV, KMC); provision of patients or study materials (SV, SNT, MT, LM); administrative, technical, or logistic support (SV, SNT, LM); and supervision (SV, SNT, TCP, MP, LM).

Address Correspondence to: Stephanie N. Tran, PharmD, BCPS, UMass Chan Medical School–Commonwealth Medicine, 333 South St, Shrewsbury, MA 01545-7807. Email:


1. Continuity of care. American Academy of Family Physicians. 2015. Accessed March 23, 2021.

2. Gulliford M, Naithani S, Morgan M. What is ‘continuity of care’? J Health Serv Res Policy. 2006;11(4):248-250. doi:10.1258/135581906778476490

3. ASHP Continuity of Care Task Force. Continuity of care in medication management: review of issues and considerations of pharmacy. Am J Health Syst Pharm. 2005;62(16):1714-1720. doi:10.2146/ajhp050257

4. Prior authorization practice resources. American Medical Association. Accessed March 23, 2021.

5. New England Healthcare Institute. Thinking outside the pillbox: a system-wide approach to improving patient medication adherence for chronic disease. Network for Excellence in Health Innovation. August 2009. Accessed March 23, 2021.

6. Cutler RL, Fernandez-Llimos F, Frommer M, Benrimoj C, Garcia-Cardenas V. Economic impact of medication non-adherence by disease groups: a systematic review. BMJ Open. 2018;8(1):e016982. doi:10.1136/bmjopen-2017-016982

7. Raper JL, Willig JH, Lin HY, et al. Uncompensated medical provider costs associated with prior authorization for prescription medications in an HIV clinic. Clin Infect Dis. 2010;51(6):718-724. doi:10.1086/655890

8. Boston Globe: Community Case Management program provides caregivers for families of medically complex children. Commonwealth Medicine. February 10, 2014. Accessed March 23, 2021.

9. Community Case Management. Commonwealth Medicine. January 28, 2018. Accessed March 23, 2021.

10. Chi-square calculator. Social Science Statistics. Accessed March 23, 2021.

11. Tenny S, Hoffman MR. Relative risk. In: StatPearls. StatPearls Publishing; 2020. Accessed March 23, 2021.

12. T-test calculator for two independent means. Social Science Statistics. Accessed March 23, 2021.