Deductibles and co-payments are meant to make patient’s think twice about seeing physicians, making them pay a bit more to use the insurance benefits they already pay – skin in the game. A new study in Annals of Internal Medicine looks at how well that approach may be working.
The researchers made use of a database containing payment information for medical care, restricting their interest to employer-sponsored commercial insurance – the type you get as a “benefit” from work. They identified two groups of patients with diabetes, one whose deductible had changed during the study period from low (less than $500) to high (greater than $1,000)  and the second group whose deductible had remained low. Employers made the changes in deductibility, so this was not a choice made by patients. After matching for sociodemographic factors , there were roughly 34,000 individuals with a higher deductible matched with approximately 295,000 whose deductible remained low.
They reported on the time intervals from the change in insurance deductibility to 1) identification of a symptom of diabetes’ vascular complications (heart attacks, stroke, and vascular disease of the legs), 2) diagnostic testing and 3)treatment – termed “events.” Because it was a retrospective review they could use the time it required half of the population to experience an event as their aggregate measure. While these intervals in the year before the change in deductibility were identical for both groups, having a higher deductible resulted in delays in care.
- 1.5 months longer to report symptoms
- 1.9 months longer to undergo diagnostic testing
- 3.1 months longer to proceed with treatment
Delays in care were all measured from an index time, the day the higher deductibles went into effect. In clinical care, after reporting symptoms, you proceed to diagnostic testing, and treatment based on their findings. And calculating a delay from the index time to the “report of symptoms” is a real difference. But then any further delays should be calculated from subsequent events - otherwise, you keep "adding in" that initial delay. The researchers calculated differences solely from the index event, and I am not sure whether this was because they have no data linking event intervals for individual patients or that they simply chose to look at the population in aggregate. In, my view, that population based approach mightr overstated delays, so I have recalculated those delays from that clinical perspective.
The insights of a closer look
Oddly enough, the first event for coronary artery and cerebrovascular disease, in both high and low deductible groups, was a diagnostic test, anywhere from 3 to 10 months before a report of symptoms. So I would infer that these were screening tests, since they were not prompted temporally by a diagnostic concern. And in recalculating the time from testing to reporting symptoms was three months for the high-deductible group and 3.3 months for the low deductible group. Subsequent times to treatment were mixed, those with high deductibles underwent their heart care 1.4 to 2.4 months later, but underwent their cerebrovascular care 0.5 to 4.2 months earlier.
For the vascular disease of the legs, that pattern of symptoms, testing, treatment was again perturbed - diagnostic testing followed treatment. But by definition diagnostic tests precede treatment, so I would assume that the researchers were more likely looking at post-treatment surveillance. In any instance, the time from reporting of symptoms to treatment was 2.3 months for the high deductible patients and 4.7 months for the low deductible. Again, a high deductible did not seem to impede care.
As much as I want patients to have skin in the game and to use health care resources appropriately, this study does not demonstrate a real delay in care for high deductible patients except for treating coronary artery disease. That makes it difficult for me to agree with the researcher's conclusions, that high deductibles impede the care of patients with diabetes. More importantly, and as the researchers point out, there is no detection of adverse outcomes. We need to determine whether high deductibles influence patient behavior, for good or bad; but this paper doesn’t begin to answer that question. It does demonstrate that high deductibles effectively shift cost from insurance companies back upon their beneficiaries – you and me.
 According to the Kaiser foundation, the average for a single individual is about $1,500
 Participants were matched by ethnicity, employer characteristics, geographic region, age, gender, classified by their morbidity or the current state of health, and census-derived estimates of their income and education
Source: High Deductible Insurance and Delay in Care for the Macrovascular Complications of Diabetes Annals of Internal Medicine DOI: 10.7326/M17-3365