Family Practice Vol. 20, No. 2, 173-177
© Oxford University Press 2003
Health Services Research |
Quality of recording of data from patients with type 2 diabetes is not a valid indicator of quality of care. A cross-sectional study
Julius Center for Health Science and Primary Care, University Medical Center, Utrecht, The Netherlands.
Correspondence to Alex N Goudswaard; E-mail: lex{at}goudswaard.cx
Goudswaard AN, Lam K, Stolk RP and Rutten GEHM. Quality of recording of data from patients with type 2 diabetes is not a valid indicator of quality of care. A cross-sectional study. Family Practice 2003; 20: 173–177.
Received 23 April 2002; Revised 5 September 2002; Accepted 4 November 2002.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background. The quality of recording of clinical data in diabetes care in general practices is very variable. It has been suggested that better recording leads to improved glycaemic control.
Objectives. The purpose of this study was to assess the completeness of recording by GPs of data from type 2 diabetes patients; to compare recorded and missing data; and to investigate the association between completeness and glycaemic control.
Methods. A cross-sectional survey was carried out in 52 general practices. Medical records were scrutinized for the presence of 11 variables. Examining patients through an active approach completed incomplete records. We compared recorded and unrecorded items. Completeness of recording was determined at both patient and practice levels.
Results. Fifty-two general practices with 1641 type 2 diabetes patients cared for by the GP participated. The frequency of absence of any particular item ranged from 20 to 70%. Weight, systolic blood pressure and HbA1c were slightly lower in patients with those items missing on their files, and more such patients were non-smokers (P < 0.05). The percentage of patients with unrecorded variables that exceeded target values ranged from 39 to 75. Neither at practice level nor at patient level was any association between the completeness of the data recording and HbA1c found.
Conclusion. Records often were incomplete, which hampers a systematic approach to care of diabetic patients. However, the lack of association between completeness of data recording and control of glycaemia indicates that improved recording is not a valid indicator of good quality of care.
Keywords. Diabetes mellitus, glycaemic control, organization of care, primary care, quality of care, recording of data.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The management of diabetes mellitus type 2 in general practice should focus on both glycaemic control and cardiovascular risk factors. Reducing these factors decreases the risk of diabetic complications.1–3 Following treatment protocols and guidelines for diabetes care is likely to improve the outcome of care.4–6 Physician-related factors that are considered to contribute to the quality of diabetes care include special interest in diabetes, use of a recall system and careful recording of clinical data.7–10 Careful recording of clinical data as part of the diabetes care makes its possible to compare past and present status, to review the course of the disease and to justify continuing or changing treatment.10
Previous studies in general practice, however, have found that important clinical data may be poorly recorded, even when physicians have a special interest in management of diabetes.6,11–13 While missing items may indicate that clinical examinations or tests have been omitted, with likely adverse effect on treatment, they may, on the other hand, merely represent failure to record normal values. Therefore, it is not clear if we can consider ‘completeness of data recording’ as a valid indicator of quality of diabetes care.
In this study, at first we assessed the completeness of the medical records of patients with type 2 diabetes treated in general practice. Subsequently, in order to get a full overview, we completed the records as much as possible by questioning and examining the patients. We compared the values of initially recorded data with values ascertained later for initially missing data.
We sought answers to the following questions.
- Which relevant variables are missing most often?
- Do the values of unrecorded data differ from those of recorded data?
- What proportion of missing data is found to exceed target values?
- Is the completeness of data recording associated with improved control of glycaemia?
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Setting and participants
The study was carried out between July 1999 and October 2000. One hundred and ten practices in the Utrecht region were invited, of which 52 were willing to participate. Of these, 27 (52%) were connected to the Utrecht Diabetes Project (UDP), a shared care project providing remote diabetologist support for GPs.5 Of the practices that refused to participate, 55% were involved in the UDP. All practices were computerized and they were asked to generate a list of their known type 2 diabetes patients. Patients treated only by their GP were selected for the study. The medical-ethical committee of the University Medical Center of Utrecht approved the study.
Design
The sentinel items investigated in the patients’ files are given in Table 1
. These items were extracted from the first as well as the updated Guidelines on Diabetes Type 2 from the Dutch College of General Practitioners.14,15 Two research assistants visited the practices and checked the patients’ medical records for the presence or absence of these items. An item was considered as ‘present’ when it was recorded and, in the case of laboratory findings, blood pressure or body weight, had been measured within the last 14 months before the audit. For fasting blood glucose we chose a limit of 4 months. The item ‘annual review’ was considered to be ‘present’ when the patient had been reviewed within the previous 14 months and at least the following five items had been measured and recorded at the time of the review: body weight, blood pressure, total cholesterol, serum creatinine and fasting blood glucose.
|
Every practice later received a list of data that were missing or outdated in their patients’ records, and the GPs were encouraged to complete the data sets by questioning and examining the patients. This was supported by sending invitations to the patients to report to their GPs. Data initially present in patients’ records and data that were missing at the initial survey but obtained later were registered separately as dataset 1 and dataset 2, respectively, using the statistical software program SPSS.
All laboratory data were measured in the GP Lab Corporation of Utrecht. HbA1c was measured with turbidimetric inhibition immunoassay Hitachi 917, Roche (range for normal subjects 4.0–6.0%).
Data analysis
The differences between ‘present’ and ‘missing’ data were calculated by chi-square test and t-test. Completeness of data recording was assessed at both patient and practice levels. Patient scores were calculated by counting the number of items recorded in dataset 1. Practice scores were the mean of patient scores per practice. Both could range from 0 to 11 points. To assess glycaemic control, we used the most recent individual and mean practice HbA1c, respectively. The association between the completeness of data recording and HbA1c concentration was calculated by Pearson’s correlation coefficient.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Fifty-two practices with 67 GPs participated. Thirty-eight practices were single-handed (of which 24 worked independently in a group practice, sharing basic facilities) and 14 were two-doctor or three-doctor practices. A total of 131 000 people were registered with these practices (mean per practice: 2519). At the start of the study, 2140 patients with type 2 diabetes were known (average per practice: 41, crude prevalence 1.6%). Of this group, 77% (1641) were under general practice care. The mean age of these patients was 65.3 years, 25% were aged >75 years and 44% were male. Twenty-two per cent of patients under GP care were being treated with diet only, 66% with oral antihyperglycaemic agents and 12% with insulin.
Missing items
The items most frequently missing were family history of diabetes (70%) and the annual review (60%), while blood pressure and duration of the disease were the most consistently recorded: 80 and 88%, respectively (Table 2).
|
Comparison of recorded and unrecorded data
Table 3
shows the values of the recorded and initially unrecorded variables. Fewer of the patients whose smoking habits were not recorded were smokers. Patients whose body weight was not recorded were less overweight. Patients with missing glycaemic control data had lower HbA1c values but higher fasting blood glucose levels.
|
Patients whose missing data exceeded target values
Table 4
shows that 18–72% of the patients had outcome measures that exceeded the current target values but nevertheless appeared to be unrecorded.
|
Association between completeness of the registration and glycaemic control
The mean completeness patient score was 6.35 (range 0–11), while the mean completeness practice score was 6.30 (range 1.6–10.3). The mean individual HbA1c was 7.1 (SD 1.7) and the mean practice HbA1c was 7.2 (SD 0.6). There was no association between completeness of data recording and glycaemic control: Pearson correlation coefficients were 0.017 (P = 0.51) at patient level and –0.183 (P = 0.20) at practice level (Fig. 1
). One practice with both incomplete registration and a high HbA1c proved to be an outlier, with an apparent correlation between diabetes regulation and registration (Fig. 1). Consequently, when the analyses were performed without this practice, no association was found (Pearson correlation coefficients: –0.004; P = 1.0).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Reliability of results
In this study, we collected the data first by extracting the items directly from the patients’ records, and secondly by an active approach of patients in order to find out the values of variables that were missing or outdated. Except the 3-monthly checks of fasting blood glucose, which were done mostly at each practice, all laboratory tests were performed by the same GP laboratory. It is likely that this procedure gave a reliable and fairly complete insight into both the data that should be collected by GPs and their data-recording habits in a large sample of known type 2 diabetes patients.
Generalizability of findings
The different practice forms in which the participating GPs were organized seemed to be rather typical for urbanized areas such as the Utrecht region.16 Because connection to the UDP may reflect a special interest in diabetes, it is of importance that the percentage of UDP practices was virtually equal in both participating and non-participating practices. In addition, age, sex and treatment of the patients, and the known duration of the disease were comparable with those of other recent investigations in general practice.4,9,17–19 Finally, the prevalence of diabetes in our study (1.6%) corresponded well to the prevalence of patients with known type 2 diabetes in Dutch general practices, which is estimated to be between 1.5 and 2.0%.20 Thus, it is highly likely that our findings were representative of general practice.
Recording of data
Monitoring and careful recording of important clinical data are considered to be a vital part of diabetes care.21 However, in our study, 20–70% of the variables were not recorded. This corresponds to findings in other studies in general practice in the last decade. Dunn et al. found between 18 (fasting blood glucose, blood pressure) and 75% (total cholesterol) of items missing in a study of 37 practices,11 and Hetlevik et al. had comparable results.9 In a large survey in 495 practices with 38 288 patients with diabetes, Khunti et al. found between 12 (blood pressure) and 51% (serum creatinine) of measurements missing.17 On the other hand, compared with a similar study at the end of the 1980s,10 our study showed an improvement in data recording of nearly all items. Nevertheless, the results show that we are still far away from what is considered desirable. Does that matter?
Quality of care
As far as we know, this is the first study in this area that revealed the values of important but unrecorded clinical parameters of diabetic patients. The values for smoking habits, body weight, blood pressure and HbA1c proved to be better in patients in whose records they were not recorded initially. Although these differences were statistically significant, they seem to be too small to have any clinical importance. The discrepancy between the near normal HbA1c value (6.9%) and the abnormal fasting blood glucose (9.3 mmol/l) is difficult to explain, although it is known that fasting blood glucose measures do not always correlate with HbA1c values.22
From a clinical point of view, GPs are likely not to omit recording measures because the values were normal. This finding suggests that not recording measures in patient records could be associated with worse outcome of diabetes care. Indeed, in a substantial number of the patients, the unrecorded variables exceeded the target values advised in the guidelines. Abnormal but unrecorded values deprive the GP of possible indications for starting or adjusting treatment, and may therefore hamper the achievement of optimal diabetes care on an individual level. Poor levels of recording of risk factors for diabetes in itself constitutes suboptimal care. Nevertheless, the lack of association between the completeness of data and glycaemic control definitely means that more careful data recording does not automatically result in better control of patients’ diabetes. This finding attenuates the suggested association between recording of data and outcome of diabetes care. Recording of data should be followed by rigorous efforts to attain recommended targets of care. In conclusion, the quality of data recording is not a valid indicator of the quality of diabetes care.
|
Acknowledgments |
|---|
Thanks are due to Henny Otten and Joyce Hanschen, research assistants, for visiting and supporting the practices, and to Peter Zuithoff and Edwin Martens for performing the statistical analyses. We would also like to thank all the GPs and their diabetic patients who participated in this study.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 O’Connor PJ, Spann SJ, Woolf SH. Care of adults with type II diabetes mellitus. A review of the evidence. J Fam Pract 1998; 47(5 Suppl): S13–S22.[Medline]
2 Stratton IM, Adler AI, Neil HA et al. Association of glycaemia with macrovascular and microvascular complications of type II diabetes (UKPDS 35): prospective observational study. Br Med J 2000; 321: 405–412.
3 UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type II diabetes: UKPDS 38. Br Med J 1998; 317: 703–713.
4 de Sonnaville JJ, Bouma M, Colly LP, Deville W, Wijkel D, Heine RJ. Sustained good glycaemic control in NIDDM patients by implementation of structured care in general practice: 2-year follow-up study. Diabetologia 1997; 40: 1334–1340.[CrossRef][Medline]
5 Rutten GE, Maaijen J, Valkenburg AC, Blankestijn JG, de Valk HW. The Utrecht Diabetes Project: telemedicine support improves GP care in type II diabetes. Diabetic Med 2001; 18: 459–463.[Medline]
6 Chesover D, Tudor-Miles P, Hilton S. Survey and audit of diabetes care in general practice in south London. Br J Gen Pract 1991; 41: 282–285.[Web of Science][Medline]
7 Griffin S. Diabetes care in general practice: meta-analysis of randomised controlled trials. Br Med J 1998; 317: 390–396.
8 Griffin S, Kinmonth AL. Diabetes care: the effectiveness of systems for routine surveillance for people with diabetes. Cochrane Database Syst Rev 2000; 2: CD000541.
9 Hetlevik I, Holmen J, Midthjell K. Treatment of diabetes mellitus—physicians’ adherence to clinical guidelines in Norway. Scand J Primary Health Care 1997; 15: 193–197.[Web of Science][Medline]
10 Rutten G, Van Eijk J, Beek M, van der Velden H. The quality of diabetes registration in eight general practices. Allgemeinmedizin 1990; 19: 68–72.
11 Dunn NR, Bough P. Standards of care of diabetic patients in a typical English community. Br J Gen Pract 1996; 46: 401–405.[Web of Science][Medline]
12 Williams DR, Munroe C, Hospedales CJ, Greenwood RH. A three-year evaluation of the quality of diabetes care in the Norwich community care scheme. Diabetic Med 1990; 7: 74–79.[Web of Science][Medline]
13 Pringle M, Ward P, Chilvers C. Assessment of the completeness and accuracy of computer medical records in four practices committed to recording data on computer. Br J Gen Pract 1995; 45: 537–541.[Web of Science][Medline]
14 Cromme PVM, Mulder JD, Rutten GEHM, Zuidweg J. NHG-Standaard Diabetes Mellitus type II. Huisarts Wet 1989; 32: 15–18.
15 Rutten GEHM, Verhoeven S, Heine RJ, de Grauw WJC, Cromme PVM, Reenders K. NHG-standaard diabetes mellitus type II (first revision). Huisarts Wet 1999; 42: 67–84.
16 Kenens R, Hingstman L. Cijfers uit de registratie van huisartsen. Peiling 2001. Nivel. Utrecht 2001 (www.nivel.nl).
17 Khunti K, Baker R, Rumsey M, Lakhani M. Quality of care of patients with diabetes: collation of data from multi-practice audits of diabetes in primary care. Fam Pract 1999; 16: 54–59.
18 Dunn N, Pickering R. Does good practice organization improve the outcome of care for diabetic patients? Br J Gen Pract 1998; 48: 1237–1240.[Medline]
19 Bouma M, Dekker JH, Van Eijk JT, Schellevis FG, Kriegsman DM, Heine RJ. Metabolic control and morbidity of type II diabetic patients in a general practice network. Fam Pract 1999; 16:402–406.
20 Ruwaard D, Feskens EJM. Suikerziekte. In Volksgezondheid Toekomst Verkenningen 1997. I De gezondheidstoestand: een Actualisering. Bilthoven: Rijksinstituut voor Volksgezondheid en Milieu, 1997: 269–280.
21 O’Connor PJ. Organizing diabetes care: identify, monitor, prioritize, intensify. Diabetes Care 2001; 24: 1515–1516.
22 Bouma M, Dekker JH, de Sonnaville JJ et al. How valid is fasting plasma glucose as a parameter of glycemic control in non-insulin-using patients with type 2 diabetes? Diabetes Care 1999; 22: 904–907.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. S. Tabrizi, A. J. Wilson, P. K. O'Rourke, and E. T. Coyne Patient Perspectives on Consistency of Medical Care With Recommended Care in Type 2 Diabetes Diabetes Care, November 1, 2007; 30(11): 2855 - 2856. [Full Text] [PDF]
|
|
|
|
|
|
C. F. Schaars, P. Denig, W. N. Kasje, R. E. Stewart, B. H.R. Wolffenbuttel, and F. M. Haaijer-Ruskamp Physician, Organizational, and Patient Factors Associated With Suboptimal Blood Pressure Management in Type 2 Diabetic Patients in Primary Care Diabetes Care, January 1, 2004; 27(1): 123 - 128. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Segovia Quality of records and quality of care Fam. Pract., October 1, 2003; 20(5): 612 - 612. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||