This study assessed a wide scope of diabetes care at Jimma University Specialized Hospital in Ethiopia using information from chart reviews and patients. Glycaemic control and blood pressure control were far below any recommended standards and attempts to prevent, detect early and manage chronic complications of diabetes were alarmingly poor. The mean FBS of 171.1 ± 63.6 mg/dl is better than the 190 ± 89.6 mg/dl in Addis Ababa [18
] however; it is far higher than the recommendations in the developed world [6
]. The majority of patients (73.1%) had FBS above the target level of 130 mg/dl as compared with 79% having >120 mg/dl in previous study [18
] indicating that glycaemic control in Ethiopia is in dire need of being addressed. Similar to most studies in the country [18
], no patient had HbA1c determination in this study because it is not available in public health sector in the country.
Over 99% of the patients were on pharmacologic treatment for their diabetes at the time of study with the delay after diagnosis for drug treatment being 3 months. Reasons may be lack of routine medical check-up and lack of knowledge of diabetic symptoms which resulted in patients presenting only when they were overtly sick.
Over 55% of the patients needed insulin and about 33% of type 2 patients have become insulin requiring during the course of their diabetes. This can be explained by the secondary insulin failure as the disease progresses since most patients were over 5 years post diagnosis. A similar trend has been documented in UK Prospective Diabetes Study (UKPDS) 33 in which case a significant proportion of patients on sulfonylurea subsequently required insulin due to severe hyperglycemia [7
]. The daily insulin requirement was significantly higher in type 1 patients (52.4 Vs 45.0IU/Kg/d, p = 0.0001) which can be explained by the pathophysiology of the disease.
Patients taking a single oral agent were found to have a better glycaemic control than those taking insulin or combination OGLAs (p = 0.002). Possible explanation for this is the duration since diagnosis of diabetes as 72.5% patients taking single OGLA had diabetes for less than 5 years. In contrast, over 60% of patients requiring insulin and 56.6% of patients taking combination OGLAs had DM for over 5 years. This implies that good control in the single OGLA was due to the early disease course than the effect of the treatment given. A good illustration for this hypothesis is UKPDS 33 finding which showed progressive increments in fasting plasma glucose levels and doses of insulin and OGLA. During the follow-up period in that study, the median insulin dose increased by about 64% from 22IU at 3 years to 36IU at 12 years [7
Similarly, patients taking lower doses of oral agents had a far better FBS level than those taking higher doses. More than 90% of patients taking glibenclamide > 20 mg/d and metformin > 1000 mg/d had persistently high FBS over the last 3+ visits. The mean FBS for them was 203.3 mg/dl and 65% of them had diabetes diagnosis for over 5 years. Surprisingly, over 80% of them did not have any modification in their glycaemic management over the last three visits. It is probably important here that patients taking such a high dose of oral agents are no more responsive to them and may sooner need insulin therapy . Overall, a high FBS over the last 3 visits did not attract attention of treating physicians. In about 70% of the patients with high blood sugar, no modification in treatment regimen was done. Possible reasons are lack of awareness, time constraint, lack of adequate human power, and most importantly lack of appropriate guidelines and diabetes education for both care givers and patients.
Access for SMBG remains to be very low as it has been in previous study [18
] (5.5 vs. 5.0%). However, access for blood glucose determination at the hospital was not found as a constraint as 98.5% of the patients had it done during each of the last three visits. Blood glucose determination is free at the hospital, but patients need to buy the glucometer and the strips for SMBG.
Morbidities and mortalities in patients having coexistence of hypertension and DM are immense [29
]. Due to this fact, the target BP level for diabetes is consistently dropping and has become lower than the target level of the general population [6
]. In previous studies in Ethiopia, hypertension was found to be an associated morbidity in 19.9% of diabetic patients as in Lester FT 1988 [31
] and 34% Feleke Y 2005 [18
]. However, the proportion of patients with hypertension in this study is much higher than the national findings and figures in the western world [29
]. In this study, 44% patients have already been diagnosed with hypertension at the time of the study. Overall, 64.1% of the patients had systolic BP ≥130 mmHg and/or diastolic≥80 mmHg, and 24.6% had >140 and/or ≥ 90 mmHg.
About 90% of patients diagnosed to have hypertension were on pharmacologic therapy, the majority of which were taking ACE inhibitors, which abides with the general recommendation [6
]. However, only 6.9% of patients currently taking antihypertensive medications had target BP. This figure, which does not look better than placebo effect, might have been due to under dosage, poor adherence to medications and lifestyle management, and less concern by health care providers.
The same as in the glycemic management, even though 100% of the patients had BP measurement on each visits; it was rarely used for patient management. Among patients not considered to have hypertension, 41.0% had mean BP of hypertensive ranges over the last 3 visits. Despite having BP of hypertensive range over the last three visits, 79.2% of the patients did not have modification to their hypertension regimen.
The proportion of patients who have been evaluated for diabetes related morbidities is very low. Less than 30% had RFT done and about 35% of patients did not have urinalysis within the last 5 years. Less than 5% of the patients had ECG, echocardiography, lipid test, or ultrasound of the kidneys. Alarmingly a significant proportion of tested patients had abnormal findings. Most of the imaging studies would have been expensive and inaccessible in the hospital setting; however, urine dipstick for albumin test is cheap and readily available at the hospital.
Far from recommended practice less than 10% of patients had ever had an evaluation for diabetic neuropathy. Those who had been examined were those with disabling symptoms as 90.3% of those evaluated had evidences for peripheral sensory polyneuropathy. Similarly, only 42.9% of patients in this study ever had a recommended yearly eye evaluation and the majority were evaluated for their symptoms rather than as a routine screening follow-up.
Diabetic ketoacidosis was found to be the commonest cause of hospital admission. Peripheral neuropathy and retinopathy were the most typical chronic complications identified. These findings are compatible with previous studies in Ethiopia [19
] and other countries in Africa [22
]. Similarly, the chief complaints for most patients were eye and sensory related. A significant proportion of them also had polysymptoms of diabetes that indicates poor sugar control.
Lifestyle management is an important component of diabetic care [6
] and intensive nutrition treatment besides the conventional pharmacotherapy has been shown to improve both glycaemic control and anthropometric measures [33
]. In this regard, diabetes nurse educators and diabetes dietitian play an important role in diabetic care. However, no emphasis has been given to diabetes health education at the clinic and in Ethiopia in general. To date there are no diabetes nurse educators and diabetes dietitian in the country. Those rendering health service for diabetes patients at the hospital had no special training for diabetes care and most of them were medical interns who were naïve not only to diabetes care but also to the general medical practice.
Demographic backgrounds, type and duration of DM since diagnosis, and duration of pharmacologic treatment of hyperglycemia were not found to affect level of glycaemic control. Similarly, the influences of good glycemic and hypertension control on morbidity were not observed in this study probably due to two major reasons. Firstly, most of the patients did not have adequate glycaemic and BP control. Secondly, documentation for diabetes related morbidities was found for only 1/3rd of patients.
Most alarming is that there is very poor record keeping system and widespread trend of not documenting physical findings, laboratory results, and reason for changing managements. Despite this, the majority of the patients reported that they were satisfied with the care they were given at the clinic indicating that patients have inappropriate expectation about diabetes care. Poor awareness among patients about the extent and components of diabetes care affects not only their expectations but importantly the quality of services they obtain and the outcome of diabetes. In this study it may be mentioned here that patients might have been satisfied with the free laboratory services (for blood glucose) and medications, which were probably their main expectations of the care.
The strengths of our study was the high patient response rate, good cooperation from hospital staff, and the fact that it was one of the few studies on diabetes care in Ethiopia.
There are potential shortcomings in our study that require comment. The major limitation was poor chart keeping that might have limited us from getting full information about chronic complications of diabetes. Another limitation was the cross-sectional study design that is not adequate to assess most of the chronic complications of diabetes. The classification of diabetes to type 1 and type 2 was based solely on history and age of the patient. In this study, the proportion of type1 was 35.6% which is much higher than findings in the western world of 5 -10% [6
]. We thus think that the proportion of type 1 diabetes may be lower than our finding with appropriate antibody study.