Family Practice Vol. 20, No. 5, 612
© Oxford University Press 2003
Correspondence |
Quality of records and quality of care
E-mail: csegovia{at}gapse07.sacyl.es
In 2000, we published a study1 very similar to that of Goudswaard et al.2 The aim of our study was also to investigate if the quality of medical records is associated with better metabolic control of diabetic patients of 15 health centres from the province of Segovia, Spain. Specific information items were considered as quality criteria of medical records.
No relationship between the quality of the medical records and HbA1c was found for the whole sample. However, amongst patients treated with diet or oral drugs, HbA1c was 7.40 for medical records that achieved only 
49% of quality criteria, and 6.94 for records that attained 
50% [95% confidence interval (CI) of difference: 0.03–0.9; P = 0.038]. The difference was not attributable to other variables, including possible confounders.
In addition, health centres with better records had better HbA1c compared with those with lower quality for the whole sample of patients, treated or not with insulin (7.15 and 7.53, respectively; P = 0.039). Patients treated with insulin had higher HbA1c than those treated with diet or oral drugs (7.70 and 7.16, respectively; P = 0.0008), and longer evolution (6.9 more years since diagnosis; P < 0.001). Centres with lower quality had more patients treated with insulin (11% more; P = 0.026) and, in these centres, the quality of medical recording improved when the patient was treated with insulin (6.8% better; P = 0.051).
Our conclusion was that the association between good clinical recording and better patient control was not simple. First, this association may happen only if data are recorded at the appropriate moment within the evolution of the illness. If some physicians improve recording when the patient has already got bad control due to the natural history of the illness (or if they miss more data in patients with better control), a finding in our study as well as in the study of Goudswaard et al., we may discover the opposite association, namely, good recording with bad control. Mixing timely recording with late recording masks the relationship of complete and timely recording with better control. Time seems to be a confounder.
Secondly, many quality criteria have no direct causal link with our variable ‘control’, i.e. the level of HbA1c. We cannot expect to control HbA1c by obtaining serum creatinine. In our study, only recording of 24 h glycaemic profiles was associated by itself with better control. Many quality criteria may relate to patient HbA1c, perhaps because a careful professional makes everything better.
Thirdly, we need to clarify the difference between obtaining clinical data and recording them. Not recorded includes at least a part of not made. However, even more subtly, if clinical information has an ‘expected value’,3 this value can be a divided: there is the present value of making good clinical decisions now that we obtain information. In addition, there is the value of having this information available in the future, allowing us to consider the evolution, avoid duplicating procedures and be alert regarding important clinical features.
References
1 Segovia Pérez C, Maín Pérez A, Corral Cuevas L et al. Control metabólico de la diabetes mellitus en relación con la calidad de las historias clínicas. Aten Primaria 2000; 26: 670–676.[Medline]
2 Goudswaard AN, Lam K, Stolk RP, Rutten UEM. Quality of recording of data from patients with type 2 diabetes is not a valid indicator of quality of care. A cross-sectional study. Fam Pract 2003; 20: 173–177.
3 Weinstein MC, Fineberg HV. Clinical Decision Analysis. Philadelphia: WB Saunders; 1998: 146.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||