Diabetes is estimated to be responsible for 3.96 million adult deaths per year at global level. By 2025, the number of people with diabetes in India is estimated to rise to 70 million. In spite of well-defined treatment for type 2 diabetes, in majority of the people, disease is poorly controlled. Hence controlling the disease is a major issue to prevent complications, increase the life expectancy and improve the quality of life. To find out the factors associated with poor control of diabetes, a systematic review and meta-analysis was carried out. The data source was Published and unpublished studies from 1980 to October 2010. Two of the authors applied selection criteria to get the relevant studies. Glycated hemoglobin level was the measure for outcome. Meta-analysis was performed by pooling the results of selected studies. Mantel – Haenszel Odds Ratios, standardized mean differences and 95% Confidence Intervals of estimates were calculated for Meta-analysis. The results were presented using forest plot. Meta-analysis showed that, elderly people and males had better control on diabetes. Presence of coronary heart disease and non-adherence to diet, exercise, medication and glucose monitoring are the factors associated with poor control of diabetes. Neuropathy, retinopathy, renal failure and neurological disorders are the complications of poor control. In spite of our sincere attempt to consolidate all studies, which give evidence of factors responsible for poor control of diabetes, we did not find this as the primary objective in many well-conducted studies. Hence, this area requires more attention of diabetes researchers.

Health care management, Poor control of diabetes, Meta-analysis

Diabetes mellitus is one of the most common chronic diseases, among adults. Globally, the number of adults with diabetes in 2010 was estimated to be 285 million, with prevalence of 6.4%. By 2030, the estimated number will increase to 439 million with prevalence of 7.7% (1). Number of deaths in adult due to diabetes is estimated to be 3.96 million per year and mortality rate of diabetes in all ages is 6.8%, at global level (2).

In India, currently, there are 40 million people with diabetes. By 2025 this number is estimated to rise to 70 million. This means that every fifth diabetic in the world would be an Indian (3). Type 2 diabetes is also a growing cause of disability and premature death, mainly due to cardiovascular disease and other chronic complications (4). Lifestyle and nutritional status etc influence the prevalence of glucose intolerance and complications of diabetes (5).

Prevention, early identification and systematic follow up of treatment are the basic strategies for controlling the disease. In spite of well-defined treatment for type 2 diabetes, in majority of the people, disease is poorly controlled with existing therapies (6-7). Hence it would be interesting to identify the factors associated with poor control of diabetes. There are number of studies which have attempted to address this question. However, as far as our knowledge is concerned, no attempt has been made to consolidate the data in the form of a systematic review. Hence, we carried out this systematic review.

All studies conducted on people with type 2 diabetes from 1980 to October 2010 irrespective of region or languages were included. This includes case control, cohort and cross-sectional studies which made an attempt to address the factors, responsible for poor control of diabetes. The studies conducted on people with type 1 and gestational diabetes was excluded. Glycated hemoglobin level (HbA1c) was the measure for outcome with two categories. HbA1C > 7% is considered as poor control and Hb1c < 7% as good control (8).

A comprehensive search was done by using the keywords, “causes of uncontrolled AND type 2 diabetes”, “factors of poor control AND type 2 diabetes”, “factors of uncontrolled AND type 2 diabetes”, “poor control AND type 2 diabetes” and “uncontrolled AND type 2 diabetes”.

The journals and databases used are Diabetes (1980 to 2010), Diabetes Care (1980 to 2010), Diabetes Educator (1980 to 2010), Diabetes and Metabolism (1990 to 2010), International Journal of Diabetes in Developing Countries (2001 to June 2010), Indian Journal of Medical Research (2003 to march 2010), Journal of Diabetology (2010), New England Journal of Medicine (1980 to 2010), Public Library of Science and Pub Med (1980 to October 2010). References of the included studies were also searched further.

A two-phase strategy was adopted for selection of studies. In the first phase, titles and abstracts obtained with preliminary search were scrutinized for inclusion. In the second phase, full manuscripts of all the studies qualified in phase one, were obtained. Selection criteria were applied to each of these studies by two of the authors. Valid studies were included for final data extraction, based on pre-designed proforma.

Methodological quality of the selected studies was assessed with the help of Critical Appraisal of Evidence Effectiveness tool (case- control). This tool has been endorsed by the Joanna Briggs Institute (JBI) (9). Internal consistency was tested, using Cronbach alpha and it was found to be 0.728.

From the selected studies, patient, treatment and disease related factors of diabetes were extracted. For Meta-analysis, Odds Ratio (O.R.) with 95% Confidence Interval (C.I.) and mean with Standard Deviation (S.D.) were retrieved. If O.R. and C.I were not reported, those measures were calculated from the available data. O.R.s of the selected studies was combined and Mantel- Haenszel Odds Ratio (M.H. O.R.) was calculated with poor control of diabetes as reference. For continuous variables standardized mean difference with 95% C.I. was also calculated. The results were presented using forest plot with fixed effect model. Chi- square statistics with P value < 0.10 and I2 statistics > 65% were used to test the heterogeneity, among the selected studies (10). Meta-analysis was performed using the Review Manager Software (Rev Man 5) from the Cochrane library (11).

Over all 7,501 studies were identified from the initial search, of which 7,458 studies were excluded and 43 studies were retrieved, in phase one. Out of this, 22 studies were excluded from the review because of either not having satisfactory inclusion criteria or insufficient information. Remaining 21 studies were included in the review of which 10 studies qualified for Meta-analysis in phase two. (figure.1)

Age, gender, body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), smoking, depression, level of knowledge on diabetes, calcium channel blockers, acetyl salicylic acid and creatinine were included in patient related factors. Control of diabetes was poor among younger adults (< 60 years) compared to elders (M.H. O.R. = 1.61, 95% C.I. = 1.11 to 2.33). Males (M.H. O.R. = 0.80, 95% C.I. = 0.72 to 0.88) had better diabetes control compared to females. Habit of smoking (M.H. O.R. = 0.89, 95% C.I. = 0.75 to 1.06) and presence of depression (M.H. O.R. = 0.93, 95% C.I. = 0.69 to 1.26) had no association with poor control. There was a difference in mean BMI of poorly controlled and well controlled diabetics (standardized mean difference is 0.47 with 95% C.I. is 0.38 to 0.55). Increase in SBP and DBP was not associated with poor control of diabetes. Meta-analysis results of patient related factors are shown in figure 2 and table 1.

Duration of diabetes, diabetic complications, hypertension, Cornorary Heart Disease (CHD), retinopathy, neuropathy, nephropathy, foot problems, fatty liver, renal failure and neurological disorders were included in disease related factors. Presence of CHD (M.H. O.R.= 1.48, 95% C.I. = 1.17 to 1.87), neuropathy (M.H O.R = 1.63, 95% C.I. = 1.19 to 2.22), retinopathy (M.H. O.R. = 1.61, 95% C.I. = 1.25 to 2.10), renal failure (M.H. O.R. = 4.14, 95% CI. = 1.88 to 9.12), neurological disorders (M.H. O.R. = 3.11, 95% C.I. = 2.06 to 4.70) were associated with poor control of diabetes. Duration of diabetes (M.H. O.R. = 0.72, 95% C.I. = 0.49 to 1.07), medication compliance, fatty liver as well as foot problems were not associated with poor control of diabetes. Meta-analysis results of disease related factors are shown in figure 3.

Insulin, oral drugs, adherence to diet and exercise, antihypertensive drugs, glibendamide, metformin, adherence to glucose monitoring and adherence to taking medication were included in treatment related factors. Surprisingly, adherence to diet (M.H. O.R. = 6.22, 95% CI. = 3.58 to 10.82), adherence to exercise (M.H. O.R. = 1.43, 95% CI. = 1.10 to 1.85) and intake of oral drugs (M.H. O.R. = 4.32, 95% CI. = 2.42 to 7.71) were associated with poor control of diabetes. Insulin (M.H. O.R. = 1.02, 95% CI. = 0.71 to 1.46) and metformin (M.H. O.R. = 1.55, 95% CI. = 0.57 to 4.24) were not associated with poor control of diabetes. Meta-analysis results of treatment related factors are shown in Figure 4.

We have conducted this systematic review to summarize the factors associated with poor control of diabetes. Ten studies qualified for Meta-analysis. Life style modification is one of the major determinants of diabetes control. In our review elderly patients having (> 60 years), males and having normal BMI patients had better control on diabetes. Probably, younger diabetics did not care about the disease control. Usually, the females take the disease only as a second priority as compared to males.

Presence of diseases like coronary heart disease, neuropathy, retinopathy, renal failure and neurological disorders was associated with poor control of diabetes. This shows the importance of diabetes control to prevent complications. Foot problems and fatty liver were not related to poor control of diabetes. Probably there could be other factors that are responsible for poor control of diabetes. With the use of insulin, the control of diabetes improves. Metformin reduces insulin resistance, thereby improving diabetes control. Surprisingly, poorly controlled patients were more adhered to diet, exercise, medication and regular glucose monitoring. One of the reasons could be that once these patients notice that their diabetes is poorly controlled, they are more likely to get adhered to the good behavior.

In spite of our sincere attempt to consolidate all studies, which provide evidence for the factors responsible for poor control of diabetes, we could not find this as primary objectives in many well-conducted studies. However, our experience with literature review showed that this area requires more attention of diabetes researchers.