Table 2 Summary of results: association between prescription drug insurance/cost-sharing and drug use, health services use, and health

Drug use: general population

 Lack of insurance and higher cost-sharing were associated with lower drug use [19, 22,23,24, 27, 29, 30, 37, 39, 40, 51];

  – Own-price η ≈ − 0.1 to − 0.6 [29];

  – Own-price η ≈ − 0.6 to − 0.8; based on aggregate data [24];

  – Own-price η ≈ − 0.2 to − 0.6; based on individual/household data [24];

  – Own-price η ≈ − 0.2 to − 0.6 [23, 30];

  – Own-price η ≈ − 0.2 [22];

  – Own-price η ≈ − 0.1 to − 0.4 [19];

 Restriction to reimbursement was associated with decreased drug use, either immediately after policy change or long-term [26];

 Among Canadians, the introduction of or increases in drug cost-sharing was associated with either no change or lower use (essential and non-essential) [16];

 The magnitude of association between cost-sharing and drug use depended on drug class, [23, 27, 30, 51] condition of patients, [23] and patient population [37];

 Lack of drug insurance and higher cost-sharing were associated with lower medication adherence and a higher risk of cost-related nonadherence [19, 21, 32, 36, 41, 46, 51];

  – Overall, a $10 increase was associated with a 3.8% decrease in adherence [32];

  – Publicly insured patients with copayments had higher odds of reporting nonadherence relative to those without copayments [36];

 Duration of coverage and type of coverage modified the magnitude of the association between cost-sharing and adherence [21];

 Drug insurance restrictions were associated with lower drug use and adherence [39];

 Essential drugs:

  – With increased cost-sharing, both essential and non-essential drug use was decreased but the decrease was larger for nonessential drugs [29];

  – Mixed evidence that higher cost-sharing was associated with lower use of essential drugs, relative to nonessential; reductions in the use of non-essential drugs were usually slightly larger [23, 25];

 Drug type: generic drugs, preferred brand-name drugs, over-the-counter drugs:

  – Limited evidence that increased generic-brand cost-sharing differential was associated with changes in patterns of drug use [19];

  – Limited evidence that increased cost-sharing for prescription drugs was associated with higher use of over-the-counter drugs [19];

  – Increases in drug cost-sharing for non-preferred brand-name drugs was associated with lower use of non-preferred brand-name drugs and higher use of preferred brand-name drugs [19];

  – Statin users required to make a copayment were more likely than others to be nonadherent [17];

  – Among statin users ≥65 years, higher copayment/cost (not necessarily solely drug cost-sharing) increased the likelihood of nonadherence and discontinuation [47].

Drug use: older adults, seniors

 Among older adults, lack of insurance and higher cost-sharing were associated with lower drug use, increased nonadherence and discontinuation [24, 27, 29, 47];

 Older people were not found to be more sensitive to price than the general population [24];

  – Own-price η, older adults ≈ − 0.1 to − 0.6 [24];

 In the US Medicare population, drug insurance was associated with higher drug use and decreased a risk of cost-related nonadherence [15, 21];

  – The inception of Medicare Part D was associated with an increase in drug use (6 to 13%) [23, 31];

  – Entry into Medicare Part D coverage gaps was associated with lower drug use (9 to 16%) [31];

  – Among US Medicare population in long-term care, drug insurance was associated with lower use of drugs that carry safety concerns, but overall drug utilization may have been unaffected [35];

 Among seniors, there was mixed evidence that higher cost-sharing was associated with lower drug use [20];

 The magnitude of association between cost-sharing and drug use depended on drug class, [15, 28] condition of patients, [15] and, patient population [31];

 Essential drugs: Among seniors, higher cost-sharing was associated with lower use of essential drugs [15, 18, 28];

 Generic drugs: Entry into Medicare Part D coverage gaps was associated with increased use of generic drugs (20%) [7, 28, 51].

Drug use: socioeconomic status, chronically ill

 Among the poor and chronically ill, higher cost-sharing was associated with lower drug use [17, 23, 27];

  – Own-price η, poor ≈ − 0.05 to − 0.4; based on aggregate data [24];

  – Own-price η, poor ≈ − 0.03 to − 0.2; based on individual/household data [24];

  – Own-price η, poor/chronically ill ≈ − 0.3 to − 0.5 [17];

 Vulnerable populations were more responsive to cost-sharing than non-vulnerable population [37];

 Among individuals with cardiovascular-related chronic disease, drug insurance was associated with increased adherence and persistence to medications [38, 42];

 Among individual with hypertension lower drug cost-sharing was associated with hypertension treatment [34];

 Statin users required to make a copayment were more likely than others to be nonadherent [33];

 Higher cost-sharing was associated with lower use of specialty drugs indicated for rheumatoid arthritis (RA), multiple sclerosis (MS), and cancer [43].

Health services use: general population

 Limiting (expanding) drug insurance was associated with an increase (decrease) in the use of health services (emergency department visits, emergency mental health service, hospitalizations, psychiatric hospitalizations, nursing home admissions [39];

 Higher levels of prescription drug cost-sharing were associated with lower use of health services:

  – Outpatient visits [24, 32, 37, 50];

  – Preventative services [32];

  – Emergency department visits [24, 25, 32, 37, 50];

  – Emergency mental health services [24];

  – Hospitalizations [24, 25, 50];

  – Nursing home admissions [24, 37];

 Higher levels of prescription drug cost-sharing were not associated (or the association was unclear) with lower use of health services:

  – Outpatient visits [19, 23, 25, 26];

  – Home health visits [19];

  – Emergency department visits [19, 23, 26, 40];

  – Hospitalizations [19, 23];

 Among Canadians, it was unclear if cost-related nonadherence was associated with lower health services use (hospitalizations, emergency department visits) [46].

Health services use: socioeconomic status, chronically ill, children

 Higher levels of prescription drug cost-sharing were associated with lower use of health services:

  – Outpatient visits [27];

  – Emergency department visits [17, 23, 25];

  – Hospitalizations [17, 23, 25, 27];

  – Nursing home admissions [17];

 Among individuals with heart failure or diabetes mellitus, higher levels of prescription drug cost-sharing were generally not associated with lower use of health services (outpatient visits, emergency department visits, hospitalizations, or laboratory/diagnostic tests) [45];

 Among individuals affected by US federal and state generic drug policies, government insurance plans with high-cost sharing on generic drugs were associated with less use of health services among children [51].

Health: general population

 Evidence on the association between prescription drug cost-sharing and health suggested that higher drug cost-sharing generally lowered health status [24, 32, 50];

 Evidence on the association between prescription drug cost-sharing and health was limited and/or unclear [16, 19, 23, 25, 26, 29, 30, 46];

 Evidence on the association between prescription drug insurance and health was limited, but generally indicated a positive association [39].

Health: older adults, seniors

 Among seniors, evidence on the association between prescription drug cost-sharing and health was limited and/or unclear [20, 35].

Health: socioeconomic status, chronically ill

 Some evidence that higher cost-sharing was associated with poorer health among the poor and chronically ill [17, 23, 29, 30].