A VERSATILE COMPUTER PROGRAM FOR AUDIT AND QUALITY IMPROVEMENT IN DIABETES CARE
Osama Hamdy, M.D.*; E. Andrew Balas, Ph.D., M.D.**
*Department of Internal medicine, School of Medicine, University of Mansourah, Mansourah, Egypt. **Health Services Management, School of Medicine, University of Missouri-Columbia,U.S.A.
INTRODUCTION
Present trends in the auditing of diabetes care and medical processing methods are dependent on several programs involving the repetition of standard data entry procedure.This method relies on subjective analyses and manual comparisons of medical information, resulting in inefficient medical procedures and subsequent treatments.
Computer programmers have advocated the introduction of a more practical process of accessing medical records. The main difficulties in most of the programs used in organizing diabetes care are; the time consumed in using the keyboard when entering patient data, the orientation of computer processing, irrelevent information provided to the diabetiologist, lack of communication between different stations within the diabetes centre and limited statistical capabilities. No single program fulfills the requirements of both the diabetologist and researchers.The introduction of a more resourceful screening procedure would help in planning and retrieving data from other stations when patients are transferred to different sections of the diabetes centre, thus allowing for alterationns and modifications of subsequent treatment.
The initiation of a more flexible audit program was thought to be better raised and designed by the diabetologists themselves, especially those having computing and statistical backgrounds. The model program aimed at providing a more efficient approach in abstracting and structuring a diverse collection of information in order to administer better quality of care. The diabetes database program Ver.2 was designed to achieve two main purposes; the first was to organize the care process offered to the diabetics in a given center; the second was to perform statistical evaluations of the pooled data in the database with the aim of conducting a survey or epidemiological studies on a large collection of patients in a community.
SYSTEM DESCRIPTION
A database program was developed to incorporate all patient data from the registration point to the various clinical observation stations. Each screening station had the capability of creating reports and running statistical processes. All screening stations were linked through a LAN network with a main file server with a 486DX2 66MHz microprocessor PC and a 600 megabyte hard disk storage capacity. Each terminal had the capabilities of retrieving data stored at any other station. In order to assist patient identification and avoid duplication, several screens could access the data simultaneously. A mouse was used to facilitate data input, enabling the user of any terminal to select his or her choice of entry, thus reducing the time consumed in entering data through a keyboard. The linked screens were designed for four different users within a diabetic center (receptionist, nurse, diabetlogist, dietician). In addition, two reporting screens were available to those requiring summary information for the evaluation of patient care.
REGISTRATION:-
The registration screen was located at the reception station and monitored the medical records of the patients examined. The data requirements of this screening form were patient's name, medical record number (MRN), age, sex, address and telephone number (if available), date of examination, date of last annual review and the name of the last diabetologist in charge of the patient. Patients' records were indexed according to the patient's first name, middle name, last name or medical record number. To avoid duplication of registration, a table of those patients with coinciding data was presented at the time of data entry. Scanning through the total database was made possible by means of simulated video buttons (for play, pause, fast forward, rewind).
DIABETES NURSE:-
The second screen organized the routine work of the diabetes nurse during the follow-up of patients. The screening form contained some data retrieved from the reception form The patient's weight and height were entered and the computer then automatically calculated the body mass index, the ideal body weight and the excess weight in overweight and obese patients. During the following visit, the initial weight entry was maintained along with the date of its entry, and the current body weight was recorded.Also the patient's blood pressure, number of hypoglycaemic attacks per month and the degree of patient's compliance to diabetes medication was recorded as either fair, good or excellent. The previous and current readings of the fasting blood sugar, postprandial blood sugar, glycosylated hemoglobin and fructosamine were entered with the date of the recent reading. Results of urine tests were entered with the recorded levels of glucose, ketones and protein. Home monitoring of urine tests and blood tests were also reported and the number of procedures completed per month was recorded. The type of therapy was also stated and the date of the next appointment was noted.
DIABETOLOGIST:-
The regular follow-up by the diabetologist was supported by a screen encompassing all previously entered data. Previously inputted data of interest was easily accessible, providing information on the patient's name, age, weight, MRN, last annual review date, last visit date, and previous and current FBS, PPBS and HbAlc. The age at diagnosis and duration of diabetes were also given. Fluctuation in weight from the last visit and the difference between the current and ideal weights were calculated. The treatment policy was specified with the dose of oral hypoglycaemics, types of combination, if any, the insulin dosage, times of injections and methods of its delivery., Quantity of each combination and number of injections were also recorded and the computer calculated the total dose injected per day. This screen also showed the differences between the past and current diabetes monitoring tests for blood sugar,HbAlc and fructosamine and assisted the diabetologist in the decision-making of the treatment persued.
Two screens were designed to support the annual check-up of diabetic patients. Symptoms such as neuropathic pain, parathesia, impotence, amenorrhea, claudication, angina were entered Feet examination results were given with special emphasis on the presence or absence of deformity, ulcers and ischaemia and were catogerised as either normal, abnormal or not done. State of peripheral circulation with special emphasis on the palpation of the dorsalis pedis and posterior tibial pulsation were included. Amputation of the toe, forefoot or below knee was stated if performed. Pinprick sensations and vibration perception weregiven as an indication of the peripheral nerves status. Skin inspection of the injection site and diabetic dermopathy was considered Any concommitant drugs such as hypolipidaemics, antianginal or antihypertensives were stated. These screens provided information and advice on dietary advice, diabetes education, foot care and any other recommendations. The clinics attended by the patients were given which mainly included the cardiology, opthalmology, nephrology, vascular, chiropody and dietician clinics.Therefore the diabetologist was able to enter his or her opinion on the results of the annual review and to provide recommendations for the following year.
Two report screens provided summarised reports of investigations of each patient's status. Data supporting these reports were retrieved from previous screens. One screen gave the results of investigations made on the patient's total cholesterol, serum triglycerides, high density lipoprotein-cholesterol (HDL), low density lipoprotein-cholesterol (LDL) and the HDL/TC. Renal profiles that included serum creatinine, BUN, albustix protien, 24 urinary albumin excretion and microalbuminuria were also stated. The screens also provided the results of the ECG and chest x-rays.
NUTRITIONAL COUNSELING:-
A dietician was assigned a terminal to provide specifications for daily caloric requirement, mode of dieting, patient compliance to diet, dietary instructions and the date of the next review. The dietician's referrals were backed up by data from previous screening stations. The mode of dieting was a choice between diabetic, diabetic reduction, diabetic high fibre, diabetic low salt, diabetic low protien and diabetic high protien diets. Compliance with the diets was signified by either excellent, average, good or poor. Instructions given were either written and verbal, verbal only or written only.
SYSTEM APPLICATION
The program was in clinical application in two major European diabetes centers, the London Diabetes and Lipid Center, ( 115, Harley Street, London, U.K.) and the Diabetes Night Clinic, (1, Duden Square, Berlin, Germany). 1145 diabetic patients were registered in the database over a period of one year The pooled data were statistically evaluated in order to test the validity of the modified program. All patients were given regular and annual review dates after passing through the initial five stations namely: the reception station, the diabetes nurse's station, the the diabetologist regular review station, the dietician station and the investigation station. To achieve the first target, Ver.1 of the program was released in early 1991 and was tested for a period of six months to study the versatility of program usage by those diabetologists with little or no knowledge in computer management. Later in 1991, Ver.2 of the program was released with additional screens in accordance with the demand of the diabetic centers in which the program was being tested. These extra screens provided several modifications that solved the problems in input process discovered in the first version.9
The database was used to register 1145 patients over a period of one year All records were kept as complete as possible and patients were informed of their next regular and annual appointment. At the initiation of the program the average time engaged in filling one screen was 1 minute and 38+47 seconds for the non-expert, and was reduced by the end of the year's operation to just 46+24 seconds for each screen The following tables show examples of the data used to build an outcome audit or to aid in an epidemiological survey while being pooled by the computer statistical modules.
Table 1: Data collected from the reception registration form
Table 2: Data collected from the diabetes nurse form
Table 3: Data collected from the diabetologist regular review form
Table 4:Data collected from the dietician form
Table 5: Data collected from the investigations form
Table 6: Example of complex choice of certain grouped criteria
The statistical module of the database provided vital information of each patient's data. The reception and nurse's screens tested the efficiency of the follow-up program on the data related to the patients' gender ratio and ages. Any group of one or more criteria could be isolated for further studies.The screening program also enabled the diabetologist to analyse specific criterias and percentages and to review the follow-ups of examined patients.The diabetologist annual review station was not thoroughly tested as the patients were registered over just one year and there was insufficient time to review them for another year. However a medical report was produced after each regular review and was attached to the patient's file.
DISCUSSION
The application of this particular database program provides significant features and greater flexibility in the analysis of diabetes. For example we can isolate the data of female patients to above the age of 45 years, to study among them, the abnormalities in lipid patterns. The proficiency of the system can also be assessed by the percentage of patients attempting the regular follow-up at the appointed time.The degree of diabetic control and the incidence of complications can be assesed in obese type II diabetics under oral hypoglycaemics. Many epidemiological values can be calculated such as the mean age of diabetics in the locality, the mean duration, the percentage of obese type II, the prevalance of hypertension and the incidence of proteinuria. The trends of therapy in a given community can also be given as the percentage of patients using a certain procedure of therapy.
The trend of the center concerning the appointment time would also help in making use of the resources and facilities available in relation to the number of patients treated. Such valuable data would help in upgrading the present service offered. The regular review data incorporates the percentage of each type of diabetes which could help in finding more facilities for each diabetic type.
The outcome of each patient's health can also be understood by the findings of the FBS, PPBS, HbAlc and fluctuations in weight. The trends in insulation injections in insulin treated diabetics can be evaluated by the number of injections and the method of injection. The dietician can scrutinize the compliance of patients to dietary advice and its relation to the degree of diabetic control. This would help in fostering the emphasis on diet education and its outcome on diabetic control. It also gives information on the patients recieving dietetic advice and the form of such advice whether it is verbal, written or both. Evaluations of the incidence of diabetics with hyperlipidaemia and its relation with certain groups of patients, certain age groups or other associations can be made. The different stages of diabetic nephropathy and the incidence of ECG ischaemic changes can also be evaluated
Observations of complex epidemiological studies, with several criteria, can be achieved giving a critical analysis of diabetics in the community; observations that would otherwise be very difficult to obtain without an extensive preplanned design or a comprehensive computing database. For example, an epidemiological study was made on a sample of 1145 diabetics registered over a period of one year to analyse the outcome of the patient's health and to assist in planning the subsequent process of service within the scope of the results obtained. The data was analysed using the statistical module and from this data we concluded that the degree of diabetic control in this community was below average. 90% of type I and 81% of type II still had PPBG values above 160mg/dl after one year of therapy. The modification of HbAlc was 1.8%but 97% of the patients still had HbAlc above 5.8%. Many factors played a critical role in the explanation of this poor outcome. 22% of the patients were above 60 years of age. This group of patients were assumed to be less adherent to the medical treatment and showed poor compliance to the diabetic education and dietary instructions. Most of the patients in the community were obese type II (97%) and inspite of the dietary instruction that was given to 62% of them, poor compliance was recorded in 46%and excellent compliance was recorded in only 12%. The remaining diabetics showed either average or good compliance. This could be reflected by the adjustment in body weight where average weight loss was just 3.1kg over a one year duration.Another reason was the poor compliance of patients to medications, where only 51.2% of the patients registered, showed good compliance to the level of diabetes and medications. Also the low incidence of patients attempting the regular review (41%) explained, in part, the lack of modification in their treatment by the recent blood glucose values and HbAlc. Most of the patients of type I were treated with only two injections per day. 24% ofthem were treated with more than two injections per day, which is concradictory to the recent recommendation of the insulin way of therapy. The incidence of hypercholesterolaemia (45%), hypertiglycerdiamia (72%) and hypertension (54%) were relatively high This explains the high incidence of patients with ECG ischaemic changes (19%) in this group which only include 24% of patients with duration greater than 15 years. Based on these satistical indices and in an attempt to make use of the resources and improve the quality of care, we recommend a greater emphasis on the patient's follow-up program and also the initiation of a recall system through which those patients who failed to follow up at the center for a period of one month, be recalled by phone or mail, so as to enhance the follow-up rate which was just 41%. More emphasis must be given to instruction and follow-ups by the dietician responsible for encouraging weight reduction of the obese type II diabetics, as only 62% were given dietary instruction and most of these patients showed poor compliance to the recommended diets. The average regular review was at 9.6 weeks, which may need to be shortened to enable proper modification of therapy with respect to the last recorded values of blood sugar and HbAlc. Most of the hyperlipidaemic patients were not recieving hypolipidaenic drugs and that explains the high incidence of coronary heart disease in the group studied. Also more importance to therapy must be stressed.
An extended survey of this group of patients for another year would add to the information on completion of the annual review forms in order to report the consultations required and the state of co-operation between the different specialities concerned with diabetes management. The real remote outcome could be evaluated by estimations of the rate of occurrence of diabetic complications over a given period, where any modification in the process of care could be evaluated as an outcome in the occurrence of such complications that mainly depended on the degree of diabetic control.
After much comprehensive, practical testing , the program is versatile for epidemiological studies needed to evaluate diabetes in a given community and in the assistance of comparind trends between different centers or even countries. It also helps in resource planning to achieve better process modifications.