The Clinical History and Costs Associated With Delayed Diagnosis of Bipolar Disorder

Paul E. Stang, PhD; Cathy Frank, MD; Anupama Kalsekar, MS; Marianne Ulcickas Yood, DSc, MPH; Karen Wells, BS; Steven Burch, PhD

In This Article


This is the first study of bipolar disorder in a managed care setting that has attempted to address the broader issues of delays in recognition of the disorder and the potential costs associated with these delays. We found, as have others,[6,7,8,9] that bipolar disorder is costly and that those patients with bipolar disorder incur 2 to 4 times the health care costs as those without bipolar disorder. Although some numerical variation exists across studies, the differences are related to numerous factors including the definition of disease cohort and cost versus charges. For example, the work by Simon and Unützer[8] reflects actual costs whereas the remainder of the studies used charges, which are higher.[6,7,9] However, although inherent variations and limitations in these data sources make it difficult to precisely calculate the actual dollar value associated with bipolar disorder, relative comparisons conducted within each study are valid. Our study shows that differences in cost and utilization differ by time from initial mental health encounter to bipolar disorder diagnosis; however, we do not show substantive changes in mean monthly costs before and after the diagnosis of bipolar disorder.

The financial impact of bipolar disorder has been studied in a number of ways. The most cited study of the cost of bipolar disorder in the United States, conducted by Wyatt and colleagues,[6] indicates that 17% of the $45 billion in annual costs of bipolar disease are direct costs while the remaining costs are indirect costs. Other studies, using claims analyses, have shown that patients with bipolar depression incur 4 times the outpatient and twice the inpatient costs as those with bipolar mania.[7] Subsequent work by Simon and Unützer[8] at the staff-model Group Health Cooperative of Puget Sound found that the average annual treatment costs for a patient with bipolar disorder was $3416, an amount exceeding that of depression ($2570), diabetes ($3083), and the general medical outpatient population ($1462). In addition, the costs were concentrated: 5% of patients with bipolar disorder accounted for 40% of the specialty mental health and substance abuse services. This estimate was eclipsed by that of Bryant-Comstock and colleagues,[9] who reported mean annual total costs of $7663.

Some investigators have explored the time to bipolar diagnosis based on previous mental health diagnoses or time in the healthcare system as a contributor to the overall costs of the disorder. In the 2000 survey of the National Depressive and Manic-Depressive Association (NDMDA),[10] more than one third of respondents with bipolar disorder were found to have sought professional help within 1 year of the onset of symptoms; however, 69% were misdiagnosed, most frequently with unipolar depression. Patients who were misdiagnosed consulted a mean of 4 physicians before receiving the correct diagnosis and one third waited 10 years or more before receiving an accurate diagnosis. Despite having underreported manic symptoms, more than 50% of patients believe that their physicians' poor understanding of bipolar disorder prevented a correct diagnosis from being made earlier.[10]

The literature contains evidence that bipolar disorder is an underdiagnosed condition. In a recent clinical series from 1 center, 37% to 40% of patients with a hospital discharge diagnosis of bipolar disorder had been considered to have unipolar depression prior to admission.[11,12] These findings are extremely pertinent to the current effort, because prescribing activating antidepressants, such as some of the selective serotonin reuptake inhibitors, may actually exaggerate the mania symptoms associated with bipolar disorder and lead to adverse outcomes, including switching and rapid cycling.[13,14,15,16] Retrospective analysis by Perugi and colleagues[17] has shown that rapid cycling (in addition to suicide and psychotic symptoms) is more likely to manifest in those whose bipolar illness includes depression at onset than in those with manic or mixed state onset. This finding has led to uncovering a link between antidepressant medication use and the induction of rapid cycling because of the higher rates of antidepressant use in those with depressive onset.

These potential delays in diagnosis result in indirect costs as well as actual costs to the healthcare system. Birnbaum and colleagues,[18] in their analysis of a cohort of patients on antidepressants, found that patients with unrecognized bipolar disorder incurred significantly higher mean monthly medical costs in the 12 months following initiation of antidepressant treatment when compared with patients with recognized bipolar disorder ($1179 vs $801) and nonbipolar patients taking antidepressants ($585). Monthly indirect costs from these employer data were also higher in the patients with unrecognized bipolar disorder ($570 vs $514 in recognized patients with bipolar disorder and $335 in patients without bipolar disorder on antidepressants).

Our study examined treatment patterns of antidepressants and utilization before and after the index bipolar diagnosis, and several issues emerged. First, using our definition and assumptions, our results show that although one third of patients with bipolar disorder were identified within 6 months of their initial mental health encounter, another third (31%) went for at least 4 years from their initial mental health encounter until the initial bipolar diagnosis ( Table 1 ). This finding was of interest because it suggested that, in some cases, either some clinicians were better than others in recognizing and referring patients or that some cases were far more difficult to diagnose. In either case, assuming that bipolar disorder was present at the initial presentation with depressive symptoms, the data indicate that barriers still exist in our ability to identify bipolar disorder in the general patient population setting and to distinguish it from unipolar depression. This latter point was confirmed by the number of antidepressants and the time that patients remain on antidepressants. Our data show that as time between the initial mental health encounter and bipolar disorder diagnosis increases, so do the number and proportion of patients who receive multiple trials of different antidepressants.

The possibility of unrecognized bipolar disorder being treated as depression is also suggested by Birnbaum and colleagues,[18] who found a significant proportion of patients with unrecognized bipolar disorder among their cohort of 9009 patients treated for depression. In addition, Russell and colleagues[19] found that 51% of patients with a bipolar diagnosis were initially diagnosed with depression. Russell's work further shows the economic impact, which accumulates with each additional antidepressant medication regimen change. Subsequent analyses will examine the extent to which these patients are treated in primary care as having depression, are treated as such, and, after treatment fails, are referred to a psychiatry specialist, who finally makes the diagnosis of bipolar disorder. Of note, 82.4% of the patients with bipolar disorder in this study were seen in the psychiatry department at their index visit.

Our analysis also examined the cost implications of the time lapse before patients were diagnosed with bipolar disorder and found that these delays may result in excess costs both during the time of the delay and after diagnosis. This finding seems to be consistent with analyses from other cohorts[18,19] reporting that additional costs accrue in those patients with undetected bipolar disorder. This finding may reflect treatment refractoriness in the post-bipolar diagnosis period or more intensive treatment, and also suggests that more aggressive recognition and treatment can reduce healthcare costs.

Study Limitations

There are several potential limitations to the use and interpretation of these data. First, this analysis is based on claims and encounter data, relying on coded diagnoses and encounters. Using claims-based data to measure disease may reflect misdiagnosis, bias, and underdiagnosis. In this analysis we assumed that the diagnosis of bipolar disorder is correct and that, for the purposes of this analysis, the patient may have been able to be detected earlier in their course of disease given their previous mental health diagnoses. We feel comfortable with a coded diagnosis of bipolar disorder, however, because Unützer and colleagues[20] found in their medical record review validation of bipolar disorder cases that the false-positive rate was less than 10% for an outpatient diagnosis of bipolar disorder. More than 74% of those in our bipolar cohort had a second bipolar coded encounter within 1 year of their initial bipolar diagnosis. It is possible that some patients with bipolar disorder could receive healthcare services outside of the health system studied; given that these patients are enrolled in a prepaid plan, however, there is a strong disincentive to obtain such services outside the health plan. Most of the patients with bipolar disorder in this study had been enrolled in the health plan for a substantial time before their bipolar diagnosis; 44.6% of them had been enrolled for 10 or more years.

These data represent the payer viewpoint and, in many ways, characterize a population based on economic impact from a payer perspective. One must interpret financial information carefully. We did not attempt to attribute individual visits as related or not related to underlying mental health illness and, thus, we can make no direct statements about what costs are directly attributable to bipolar disorder. Furthermore, we required the comparison cohort patients to have made at least 1 visit within 60 days of the index date because patients with bipolar disorder were defined by a coded visit. As mentioned previously, this requirement most likely resulted in an analysis with the most conservative comparison of cost because, by definition, we selected a comparison population that may have higher healthcare utilization than the general population in the healthcare system. Therefore, any gaps in healthcare utilization identified in this study are probably underestimates of the actual differences. Last, although high utilization rates were observed among patients with bipolar disorder, we found that costs did not differ before and after bipolar disorder diagnosis. This finding suggests that, in some cases, the index diagnosis may not have been a new diagnosis.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.