Outcome of Schizophrenia in Later Life: Conceptual Changes and Implications for Treatment and Policy

In tandem with the greying of the general population, there has been a dramatic increase in the number of older adults with schizophrenia (OAS). Since 2000, there has been a doubling of persons aged 55 and over with schizophrenia and they now comprise about one-fourth of all persons with schizophrenia. Globally, the number of persons with schizophrenia aged 60 and over will double between 2014 and 2050 and will reach 10 million. Nevertheless, only 1% of the schizophrenia literature is devoted to older adults.1

Portrait of a sad homeless man.

Schizophrenia typically arises in the second and third decades of life, and roughly three-quarters of people develop schizophrenia before the age of 40. Although persons with schizophrenia have substantially higher mortality rates than the general population, many more are living into older age. A study in the Netherlands found the prevalence estimates of schizophrenia in the adult population at age 60 was .35%, .14%, and .03% for early-onset, late-onset (arising between age 40 to 59), and very late-onset schizophrenia (arising at age 60+), respectively; the overall prevalence was of 0.52%. Thus, the ratio of early-onset to late-onset types was 2:1.2

An understanding of the various outcome measures of schizophrenia in later life is critical for developing rational treatment and public policy decisions. Many of our perspectives on the outcome of schizophrenia in later life are based on studies done in the 20th century when people typically spent many years in institutions. Today, more than 85% of OAS live in various levels of supportive or independent living in the community, 10% reside in nursing homes, and 5% are in hospitals.1 In this article, I will elaborate on 7 essential points that will provide the reader with an appreciation of the evolution in the conceptualization of outcomes in later life and the implications for care. Many of the findings are based on our longitudinal data collected on community-dwelling persons aged 55 and over with EOS (defined as onset before 45) living in New York City.1,2

Point 1. “Outcome” changes historically because of evolving diagnostic criteria, criteria for outcomes, and social factors.

Across the 20th century, there have been wide swings in the clinical outcomes of persons with schizophrenia. In the early part of the century, favorable outcomes were thought to occur in about one-quarter of patients. In the mid-century, optimism grew, and favorable outcomes were thought to approach one-half of patients; however, favorable outcomes diminished to approximately one-third by the end of the century with the release of DSMIII diagnostic criteria and a social climate that stanched funding for severe psychiatric illness and community safety net supports.2 A countervailing perspective emerged in the last quarter of the century from 10 long-term studies, mostly from Europe and the United States, that assessed patients between 20 and 37 years after initial evaluation. Nearly all subjects were in middle or later life. These studies found clinical and social recovery to be about 50% for each category.2 A limitation of these studies was that they provided a single snapshot of late-life clinical status. However, in the past decade, two longitudinal studies of OAS, predominantly outpatients from the United States and the Netherlands, found rates of persistent clinical remission and persistent social recovery to be about 25% for both categories.2 Thus, the optimism of the earlier long-term studies must be tempered by these longitudinal data.

Point 2. A comprehensive view of “outcome” should include symptoms, functionality, a combination of symptoms/functionality (recovery), and successful aging.

Although clinical remission, which comprises positive (e.g., hallucinations, delusions) and negative (e.g., blunted affect, anhedonia, avolition) symptoms, is an essential outcome variable, the DSM5 recognizes other symptom dimensions such as depression and cognitive impairment. Moreover, functionality is an important outcome variable, especially measures such as community integration. The combination of clinical remission and community integration comprises “clinical recovery.” Subjective measures such as self-perceived quality of life are pivotal variables affecting other outcome measures (see below). Finally, physical well-being, although not a psychiatric category, is especially important given the higher mortality rates in OAS, e.g., about twice their age peers. Finally, some investigators have been interested in resilience, coping strategies, and other factors that comprise positive mental health or successful aging.

Point 3. Clinical outcome criteria are largely independent of each other.

Our research found that the various outcome measures—e.g., clinical remission, community integration, depression, and cognitive functioning—only modestly correlated with each other, and the shared variance (overlap) among the variables ranged from 0% to 19%.2 Thus, targeting one category is not likely to have an appreciable impact on other outcome categories. Consequently, each outcome category may require a distinct treatment strategy.

Point 4. “Outcome” is not “quiescent” or “stable” in later life but continues to evolve.

Because of a paucity of longitudinal data in OAS, the literature has often depicted the course in later adulthood as a flat line. Our longitudinal data showed that significant changes (improvement or worsening) in 5 outcome categories (remission, depression, cognition, community integration, and physical health) ranged from 26% to 40% over a 52-month follow-up period.2 Fewer than one-fifth experienced no changes in any outcome category and over three-quarters experienced changes in 2, 3, or 4 categories. The good news is that patients show considerable movement in later life; the bad news is that patients can get worse as well as better. Because of earlier erroneous models of “end-stage” schizophrenia, policy makers have failed to provide sufficient treatment resources to address the fluctuations in symptoms that may occur in later life.

Point 5. “Outcome” is heterogeneous with a variety of combinations that necessitate a personalized approach to care and a more nuanced approach to research.

When the 5 outcome categories described above were dichotomized into “persistently unfavorable or worse over time” versus “persistently favorable or better over time”, we found considerable variation in the clinical profiles of our patients, with fewer than 10% having no favorable outcomes, about 5% having all favorable outcomes, and the rest, having between one and four favorable outcomes.1,2 Thus, even in later life, each person’s outcome is a dynamic multidimensional process with various combinations of features. This finding is consistent with adopting a personalized approach to care.

Point 6. “Recovering” (aka “clinical recovery”) can be assessed empirically and yields a 5-tier taxonomy with varying degrees of recovery that can guide treatment and research.

Clinical recovery, commonly viewed as the ultimate outcome goal, is based on meeting empirical criteria for clinical remission and community integration. It should be distinguished from the consumer model in which recovery is seen as a process of striving to reach one’s full potential. Adhering to the consumer model alone can blur the boundaries between persons who remain disabled and those who have achieved relatively normal mental and social functioning. Our work identified a 5-tier model of clinical recovery: 12% who remained persistently in clinical recovery at both baseline and follow-up (Tier 1); 18% who never attained any components of clinical recovery (Tier 5); and 35% who attained clinical recovery at either baseline, follow-up, or both (Tiers 1 and 2). Overall, 70% of the subjects exhibited a variety of components of clinical recovery at baseline and follow-up (Tiers 2, 3, and 4). Here again, a more nuanced and targeted treatment approach is required to meet the clinical and/or social needs based on a patient’s tier level.2

Point 7. Quality of life (self-perceived well-being) is a pivotal variable and an Important point for Intervention.

Although the focus of this article is on clinical outcomes, among persons with schizophrenia, quality of life (QOL) is considered one of the most crucial indicators of well-being along with symptomatic remission and functional recovery. Our findings in OAS indicate that QOL is not only critical in its own right but that it has significant salubrious effects on positive symptoms, depression, anxiety, and insight. 1,2 Notably, QOL seems to improve in later life. OAS had QOL scores that were only 6% lower than their healthy age peers, whereas they were 28% higher than people with chronic pain, and nearly double those with chronic fatigue syndrome.

In summary, despite the dramatic increase in persons with schizophrenia reaching old age, they have been largely invisible to researchers and policymakers. Historically, the illness course was thought to be static in later life. Recent studies have shown that outcome in later life is dynamic and multidimensional, with components that are largely independent of each other, and individuals have varying combinations of favorable and unfavorable outcome dimensions. More resources must be enlisted to provide the multi-prong, personalized care that is required.

Carl I. Cohen, MD, is SUNY Distinguished Service Professor and Co-Director, Division of Geriatric Psychiatry and Center of Excellence for Alzheimer’s Disease at SUNY Downstate Health Sciences University. You may contact him at carl.cohen@downstate.edu or (718) 270-2003.


1. Cohen, CI et al. Advances in the conceptualization and study of schizophrenia in later life: 2020 Update. Clin Geriatr Med 36:221–236, 2020.

2. Cohen CI, Meesters PD (Eds) Schizophrenia and Psychoses in Later Life: New Perspectives on Treatment, Research, and Policy. London, Cambridge University Press, 2019. Chapters 2 and 8.

Have a Comment?