Payment Restrictions for Prescription Drugs under Medicaid

A cap of three prescriptions per month sounds reasonable until you learn what it did to the sickest patients in the program. For Medicaid recipients who were already managing multiple chronic conditions — like the elderly woman on insulin, a diuretic, and a heart medication — that limit cut their medication fills nearly in half, dropping by forty-six percent. And here is the question that drives everything else: did the state actually save money, or did it just move the cost somewhere else? In nineteen eighty-one, New Hampshire's legislature imposed exactly that limit — a maximum of three reimbursed prescriptions per patient per month, starting September first. It wasn't an outlier policy. Soumerai, Avorn, Ross-Degnan, and Gortmaker report that twenty-seven Medicaid programs had already implemented some form of patient-level payment restriction, driven in large part by budget pressures following the Omnibus Budget Reconciliation Act of nineteen eighty-one. What made New Hampshire unusual was timing: the cap was in place for roughly a year and then replaced, in August nineteen eighty-two, by a one-dollar copayment per prescription. That clean, dated policy switch — cap on, cap off, copayment in — gave the researchers something rare: a natural experiment in real-world benefit design. To measure what happened, the team pulled forty-eight months of computerized Medicaid claims from New Hampshire and compared them against a parallel series from New Jersey. New Jersey had comparable reimbursement characteristics, similar pharmacy dispensing fees, no restrictive formulary, and demographic features that made it a reasonable stand-in for what New Hampshire's prescribing trends would have looked like without the policy change. The analytic approach was time-series regression — modeling month-by-month prescription counts before, during, and after the cap, with the New Jersey series as a control to filter out any nationwide shifts in prescribing behavior. The study population in New Hampshire included ten thousand seven hundred thirty-four continuously enrolled, non-institutionalized patients. Within that group, eight hundred sixty patients — about eight percent — were classified as multiple-drug recipients, meaning they averaged five or more prescriptions per month before the cap. This small subgroup accounted for forty-seven percent of all prescriptions in the baseline year. They were predominantly female, with a mean age of fifty-six, nearly half of them over sixty, and the majority eligible for Medicaid because of old-age assistance or disability. These were the patients with the most to lose. The cap hit immediately and hard. Across the full cohort, the number of prescriptions filled dropped thirty percent — from one point one to zero point seven seven prescriptions per patient per month — and stayed there throughout the eleven-month cap period. New Jersey showed no comparable movement, holding steady at about one point two prescriptions per patient per month. The policy effect was real and it was specific to New Hampshire. For the multiple-drug recipients, the drop was nearly twice as steep. Their average fell from five point two to two point eight prescriptions per person per month — that forty-six percent collapse, sustained month after month. And it wasn't just marginal or low-value drugs that disappeared. Drugs the team classified as having limited clinical efficacy did fall the most, by fifty-eight percent. But essential medications fell too, and by amounts that should give any clinician pause. Insulin prescriptions dropped twenty-eight percent, from eleven point six to eight point four per one hundred patients per month. Furosemide, a loop diuretic used in heart failure and edema, dropped thirty percent. Thiazide diuretics dropped twenty-eight percent. Digoxin, used for heart conditions, dropped forty-five percent. Across the category of effective essential medications, the decline was twenty-eight percent — from sixty-seven to forty-eight point six prescriptions per one hundred patients per month. These aren't comfort medications. These are drugs that, if stopped abruptly, can destabilize chronic disease. And patients were not finding ways around the limit. The regulation technically allowed larger quantities per prescription — up to a three-month supply — which meant staggered prescribing could in principle allow more than three drugs. But average prescription size among multiple-drug recipients rose by only about thirteen percent during the cap, and showed no progressive increase over time. The team looked closely at insulin: among seventy-nine continuously enrolled insulin users, average monthly dose fell twenty-eight percent. Nearly half experienced decreases greater than twenty-five percent, and one quarter experienced decreases greater than fifty percent. Only two patients temporarily stopped receiving insulin through Medicaid altogether — the rest simply took less. Out-of-pocket substitution was negligible. In a pharmacy-level sample of ten high-use patients, combined Medicaid and out-of-pocket use still fell from four point nine to three point one prescriptions per patient per month — a thirty-seven percent drop even counting private purchases. Nine of those ten patients increased their out-of-pocket spending by an average of one point thirty-six dollars per month, against a mean loss of about twenty dollars per month in Medicaid reimbursements. The authors note that twenty dollars a month may represent more than half of all remaining disposable income for a typical low-income elderly person. The arithmetic of substitution simply didn't work. Then the cap came off. In August nineteen eighty-two, New Hampshire replaced the three-prescription limit with a one-dollar copayment per prescription — no numerical ceiling, just a small cost at the point of purchase. Prescriptions for most medications rose back to just below precap levels. Among the multiple-drug recipients, the average rebounded toward four point seven prescriptions per person per month by the end of the study period. For the broader population of other outpatients, the cap had caused a seventeen percent reduction; during the copayment period, their rate climbed back to approximately precap levels. The budget comparison sharpens the whole argument. The cap saved New Hampshire roughly seven hundred eighty thousand dollars per year in drug reimbursements. The copayment, under conservative assumptions, saved about four hundred thousand dollars per year. Soumerai and colleagues summarize this as comparable savings of four hundred thousand to eight hundred thousand dollars annually. So the two instruments produced similar fiscal outcomes for the program. But the cap cut essential medications for the sickest patients by nearly half, while the copayment left prescription use broadly intact. Same savings target. Profoundly different effect on the people who could least afford the difference. The paper is direct about what its data cannot tell you. Prescription claims show what was reimbursed. They do not show what happened next — whether patients whose insulin doses fell by half ended up in the emergency department, whether reduced diuretic use led to hospitalizations, whether the money Medicaid saved on drugs was spent five times over on inpatient care. That downstream question was, and to a meaningful degree remains, unanswered. What the data do reveal clearly is who bore the cost of these savings. The multiple-drug recipients — eight percent of continuously enrolled patients, accounting for nearly half of all prescriptions — were predominantly elderly women eligible because of disability or old-age assistance. The cap, by design, hit hardest whoever used the most medications. And whoever uses the most medications is, by definition, whoever is sickest. A quota-style restriction that ignores clinical need doesn't just reduce spending uniformly — it concentrates the burden on patients with the fewest resources to compensate and the highest medical stakes if they go without. The broader lesson is worth holding onto. The New Hampshire switch — from cap to copayment, with a dated transition and a parallel comparison state — is the kind of real-world experiment that rarely appears in policy research. What Soumerai and colleagues uncovered is something that budget projections routinely miss: when a blunt administrative limit is applied, patients and prescribers do not reliably adapt. They don't substitute, they don't find workarounds, and they don't compensate with out-of-pocket spending. Many simply go without. The tension between containing drug costs and maintaining access to essential medications isn't a solved problem. Every generation of health policy has its version of the cap — some mechanism that looks reasonable in a spreadsheet and lands differently on the ground. What this study established, with unusual methodological rigor for its era, is that the mechanism matters enormously, and that the patients most harmed are often the ones least visible in aggregate cost figures. The question of whether those drug savings came at the price of worse health, or higher costs somewhere else in the system, is the one that still needs answering. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.