Quality of Diabetes Care in Patients with Schizophrenia: A Case-Control Study.

Background Patients with schizophrenia are at least twice as likely to develop diabetes mellitus compared to the general population. This is of signicance in Qatar given the high prevalence of obesity and diabetes. Furthermore, the lifespan of people with schizophrenia and diabetes is shortened by approximately15 years due to long-term microvascular and macrovascular complications. High quality diabetes care can signicantly reduce morbidity and mortality. We assessed the level of diabetes care delivered to patients with schizophrenia and diabetes compared to those with diabetes alone. Methods We performed a retrospective chart review of patients with diabetes mellitus with (n=73) and without (n=73) schizophrenia. Demographic information and electronic medical records were reviewed for 6-month and 1-year American Diabetic Association standards of diabetes care. Optimal diabetes care was dened as having completed glycated hemoglobin (HbA1c), lipid prole and retinal examination within 12 months.


Abstract
Background Patients with schizophrenia are at least twice as likely to develop diabetes mellitus compared to the general population. This is of signi cance in Qatar given the high prevalence of obesity and diabetes. Furthermore, the lifespan of people with schizophrenia and diabetes is shortened by approximately15 years due to long-term microvascular and macrovascular complications. High quality diabetes care can signi cantly reduce morbidity and mortality. We assessed the level of diabetes care delivered to patients with schizophrenia and diabetes compared to those with diabetes alone.

Methods
We performed a retrospective chart review of patients with diabetes mellitus with (n=73) and without (n=73) schizophrenia. Demographic information and electronic medical records were reviewed for 6-month and 1-year American Diabetic Association standards of diabetes care. Optimal diabetes care was de ned as having completed glycated hemoglobin (HbA1c), lipid pro le and retinal examination within 12 months.

Results
Optimal diabetes care was signi cantly lower in patients with schizophrenia and diabetes compared to diabetes alone [26.0% (n=19/73) vs 52.1% (n=38/73), p=0.002]. Patients with diabetes and schizophrenia were also signi cantly less likely to have had body mass index recorded within six months (p=0.008) and HbA1c (p=0.006), lipid pro le (p=0.015), estimated glomerular ltration rate (eGFR) (p=0.001) and order for retinal screening (p=0.004) over 12 months. After adjusting for multiple comparisons, only assessment of eGFR (p=0.01) and order for retinal screening (p=0.04) remained signi cant.

Conclusion
Patients with schizophrenia and diabetes in Qatar receive sub-optimal diabetes care compared to those with diabetes alone.

Background
Diabetes Mellitus (DM) has an estimated global prevalence of 8.5% in the adult population (1). Qatar, Kuwait, and Saudi Arabia in the Middle East and North Africa Region (MENA) are three of the top 10 countries with the highest diabetes prevalence worldwide (2). In Qatar, 70.1% of Qatari nationals are clinically overweight ( Body Mass Index (BMI) ≥ 25 kg/m 2 ), and 41.4% obese (BMI ≥ 30 kg/m 2 ) (3). The prevalence of diabetes in people with schizophrenia is estimated to be 2 to 3 times higher compared to the general population (4), with over four-fold increased odds of comorbid diabetes in patients with non-affective psychosis and a family history of type II DM (5). Recently, a retrospective cohort study from Scotland has shown greater diabetes incidence in women compared to men with schizophrenia (RR 2.40 [95% Con dence Interval (CI) 2.01, 2.85] and 1.63 [1.38, 1.94]), respectively) (6). The high prevalence of diabetes in people with schizophrenia generally re ects multiple factors including a more sedentary lifestyle, higher prevalence of obesity, poor diet, metabolic side effects of antipsychotic medication and possibly intrinsic factors including genetic predisposition and in ammation (7). Nielsen et al showed that second generation antipsychotics were associated with an increased risk of diabetes (8). However, a study in Qatar did not nd a higher prevalence of metabolic syndrome in patients receiving antipsychotics (9). Dysglycemia is also associated with the severity of schizophrenia, particularly negative symptoms and cognitive impairment (10). All-cause mortality for people with schizophrenia is signi cantly increased, with a recent longitudinal study in the United States showing an all-cause standardized mortality ratio of 3.7 (95% CI, 3.7-3.7) (11). The lifespan of patients with schizophrenia is shortened by approximately 15 years with the largest single cause of death being cardiovascular disease (12). Provision of high-quality diabetes care, particularly the management of cardiovascular risk factors, could reduce this elevated morbidity and mortality and healthcare costs for providers.
Delivering optimal diabetes care in patients with schizophrenia can be challenging as patient, provider, treatment and system-level factors may act as barriers to following evidence-based guidelines (13). Several studies have reported that patients with DM and serious mental illness (SMI) are less likely to receive optimal diabetes care (13)(14)(15)(16). In Canada, Kurdyak et al (2017) showed that schizophrenia was associated with a reduced likelihood of optimal diabetes care (HbA1c, lipid pro le and retinal exam) over the preceding 2 years [adjusted odds ratio (OR) (95% CI): 0.64 (0.61-0.67)] and an increased likelihood of diabetes-related emergency department (ED) visits (adjusted OR (95% CI): 1.34 (1.28-1.41)) and hospitalization (adjusted OR (95% CI): 1.36 (1.28-1.43)) (13). These results con rm those of an earlier study showing an increased rate of diabetes-related emergency department visits in people with schizophrenia (Hazard Ratio = 1.74, 95% CI (1.42-2.12)) (17). Indeed, the non-treatment rates for diabetes in the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study in patients with schizophrenia was 45.3% (18). However, a study by Whyte et al found the quality of diabetes care in the United Kingdom comparable between patients with schizophrenia or bipolar disorder with comorbid diabetes and those with diabetes alone (19). Engagement with mental health services and adherence to antipsychotic treatment is associated with signi cant reduction in diabetes related hospitalization (20). An association has also been found between more frequent outpatient psychiatry (OR = 1.28, 95% CI = 1.20-1.37) and primary care (OR = 2.10, 95% CI = 1.85-2.39) visits and optimal diabetes testing (21).
The quality of diabetes care received by people with schizophrenia in Qatar has not been investigated previously. This is of relevance in Qatar given the high prevalence of obesity and diabetes. Our aim was to determine the quality of diabetes care in patients with schizophrenia and DM compared to patients with diabetes alone.

Methods
We identi ed two cohorts of patients on the CERNER Electronic Medical Records (EMRs) system in outpatient clinics in Hamad Medical Corporation (HMC), Doha, Qatar: those with DM and schizophrenia (study group) and those with DM and no serious psychiatric illness (control group). Cohorts were limited to subjects 18 years of age and older, both genders, all nationalities and with type 2 diabetes. Diagnoses of both DM and schizophrenia were based on documented clinical diagnosis on EMRs and subjects were matched for the service followed for diabetes care. Standards of diabetes care were retrospectively recorded for the year 2018, and patients were only included if they were diagnosed with both DM and schizophrenia prior to 2018. Patients were assigned subject numbers and selected using a random number generator. Based on recommendations by the American Diabetes Association (ADA) (22), we identi ed the completion of the following investigations/measurements/ examinations as measures of optimal diabetes care: HbA1c, BMI and blood pressure over six months (July through December 2018) and HbA1c, BMI, blood pressure, lipid pro le, retinal examination, foot examination, serum creatinine and eGFR over 12 months (January through December of 2018). In addition, we recorded the total number of medical and psychiatric admissions over 12 months and referrals to the dietician and diabetes educators.

Sample Size Calculation
A previous study showed that the rate of HbA1c testing in patients with schizophrenia and DM was signi cantly lower compared to patients with diabetes alone (48% vs 71%, p = .0001) (23). Using the same proportions, with an alpha = 0.05, power = 0.9, and sample size distribution ratio of 1:1, the calculated sample size in each group was 73, with a total sample size of 146 (24).

Statistical Analysis
Statistical analysis was conducted using the Statistical Package for the Social Sciences (IBM-SPSS, version 23.0, IBM Corp, Armonk, NY, USA). For categorical variables, we calculated the frequency. For continuous variables, we calculated the mean and standard deviation (SD). The differences between the study and control groups were analyzed using Pearson's chisquared test for categorical variables (and in case of non-validity, the Fishers' exact test), and t-test for continuous variables. Optimal diabetes care was de ned as having completed all three of HbA1c, lipid pro le, and eye examination within 12 months, adapted from a previous study by Kurdyak et al (13). We assessed all the key ADA measures of diabetes care. For multiple comparisons, p values were adjusted using the Holm Bonferroni method.

Sociodemographic and clinical characteristics of the groups
The study assessed participants with schizophrenia and diabetes (n = 73) versus those with diabetes only (n = 73). Participants with diabetes and schizophrenia were signi cantly younger compared to the control group (45.1 ± 13.0 vs 57.9 ± 11.1; p < 0.001). There was no signi cant difference for gender or nationality between groups ( Table 1). The number of psychiatric and medical admissions combined did not differ signi cantly between groups.

Standards of optimal diabetes care
Five standards were signi cantly less likely to be met in patients with diabetes and schizophrenia compared to diabetes alone: BMI recorded within six months (p = 0.008), and HbA1c (p = 0.006), lipid pro le (p = 0.015), eGFR (p = 0.001) and order for retinal screening (p = 0.004) over 12 months. However, after adjusting for multiple comparisons, only assessment of eGFR (p = 0.01) and order for retinal screening (p = 0.04) remained signi cant ( Table 2). Using the three-item de nition of optimal diabetes care (HbA1c, lipid pro le and eye examination over 12 months), patients with diabetes and schizophrenia were less likely to receive optimal diabetes care compared to those with diabetes alone [26.0% (n = 19/73) vs 52.1% (n = 38/73), p = 0.002].

Discussion
In this study patients with diabetes and schizophrenia were half as likely to receive optimal diabetes care compared to those with diabetes alone using the three-item de nition (HbA1c, lipid pro le and retinal examination over 12 months). A Canadian study using the same de nition over two years reported very similar ndings (14). More recently, a nationwide Danish study found that patients with schizophrenia were less likely to receive high quality diabetes care (de ned as ful lling 80% of recommended measures) (14).
Suboptimal diabetes care in our study group may be attributed to a lack of integrated physical and mental health services and di culty sharing information and expertise amongst health care specialties (25). Negative symptoms in schizophrenia including lack of motivation, goal-directed activity, and disorganization (26) might affect patient's adherence with treatment and clinic appointments, and are associated with poorer mental health outcomes. Schizophrenia is also frequently comorbid with depression and substance misuse (28,29), which can further hinder engagement with diabetes care. Socioeconomic status is unlikely to affect the quality of care in Qatar, as health care services are mostly free of charge for citizens and provided with minimal charge for residents of the country. Stigma against mental illness may affect quality of care and act as a barrier to health care access (30). In Qatar, 40.6% of surveyed individuals perceive patients with psychiatric disorders as mentally challenged and 48.3% believe mental illness is a punishment from God (31).
Patients with diabetes alone were signi cantly older compared to those with both diabetes and schizophrenia, which could re ect the earlier onset of type 2 diabetes due to lifestyle and medications used for schizophrenia (12). However, the level of obesity (BMI) and glycemic control (HbA1c) was comparable between the two groups. A recent study in Qatar also found no difference in the prevalence of metabolic syndrome in patients receiving antipsychotics (9). The high prevalence of diabetes and obesity in the Qatar population could explain a lack of further effect of antipsychotics or schizophrenia on these outcomes (2,3).
We propose several strategies to reduce the inequalities reported. Educational awareness programs should be instated and encouraged to address mental health stigma and negative stereotypes. Further work is required to increase awareness amongst HCPs regarding the increased morbidity and mortality in patients with schizophrenia. Communication between different services providing health care for patients with schizophrenia is essential to ensuring access to health care services and optimizing management of medical comorbidities. Education to prevent or limit obesity and diabetes needs to be promoted among patients with severe mental illness to help decrease the prevalence of diabetes and ensure that those diagnosed with DM receive high quality care. These recommendations are in line with the recently published blueprint for protecting physical health in people with mental illness (32).

Strengths And Limitations
This population-based study included randomly assigned patients from both genders and all nationalities followed in outpatient clinics. Data to identify and control for the duration of DM and schizophrenia in our cohorts was limited as the EMRs system was started in 2015. Additionally, important demographic information such as smoking, marital and socioeconomic status were missing or inconsistently documented. Nevertheless, lab results, orders and documentation pertaining to diabetes care were easily extractible, and the criteria for optimal diabetes care were clearly de ned. As this is a case-control study, the incidence and relative risk of suboptimal diabetes care in patients with schizophrenia could not be calculated.

Conclusion
This study found that patients with schizophrenia and DM in Qatar are less likely to receive optimal diabetes care compared to patients with diabetes alone, particularly in relation to the microvascular complications with a decreased frequency of eGFR completion and order for retinal screening. These data encourage further population cohort studies to assess adherence with treatment and the burden of diabetes-related medical comorbidities and complications in patients with schizophrenia.