The medico-professional quality of GP consultations assessed by analysing patient records

Abstract

Objective. To assess the medico-professional quality of consultations by analysing textual data from patient records. Design. Qualitative analyse of textual data. Setting. Four primary health care centres using electronic patient records (EPR) in Finland. Subjects. EPR and paired questionnaires of 175 consultations filled in by GPs and their patients independently. Main outcome measures. Medico-professional quality of consultations, quality of care of acute respiratory infections, and hypertension. Results. The medico-professional quality of the consultations was quite good. However, 9% of the records could not been assessed at all because of missing or poor documentation and 9% were assessed as poor. The treatment of acute respiratory infections and hypertension is not in line with current care guidelines. Smoking habits or other health behaviour or lifestyle factors were seldom recorded. Conclusions. The medico-professional quality of the consultation was quite good. Quality improvement is needed in the treatment of acute respiratory infections and hypertension. User-friendly EPR systems would improve the content of patient records.

Key Words::

• Acute respiratory infections
• consultation
• electronic patient records
• general practice
• hypertension
• quality of health care

Most studies on EPR have investigated structured rather than textual data of patient records.

• Textual assessment of patient records implies that the medico-professional quality of consultations is quite good.

• Textual assessment of patient records shows that quality improvement is needed in the treatment of acute respiratory infections and hypertension.

• User-friendly EPR systems improve the content of patient records.

Introduction

Electronic patient records (EPR) have generally replaced handwritten records, and it is generally felt that EPR increase the availability of high-quality clinical information. Although EPR may contain more information than paper-based records [1], they are often incomplete [2,3]. Additionally, any association between completeness of data recording and good outcomes of care or the quality of the consultations has not been found [2,3,4]. Both patients and GPs are generally satisfied with EPR but, in practice, some systems are so complicated and laborious and that hamper their use by GPs [3,5,6]. Although EPR have been widely used within Finnish primary health care [7], EPR have been shown neither to be very accurate nor do they meet all the legal requirements [3,8].

Most previous studies on EPR have investigated structured data rather than textual data. “Soft” parameters, such as lifestyle, have only seldom been studied [9]. Although the textual content of EPR is extensive and the diagnoses are usually coded well, most of the data are not coded, which limits data retrieval [10]. On the other hand, GPs cannot be expected to code and collect stringently structured data as a part of their busy routines [11]. Most patient records are descriptive [10] and it is not easy to use such information for research, quality improvement, statistics, and management, as has been shown [12].

The aim of this study was to assess the medico-professional quality of the textual data in patient records and to compare the assessment by us as a third party with the assessments of GPs and patients.

Material and methods

The study material consists of 175 patient record entries and paired questionnaires of the same consultations filled in by GPs and their patients independently. The material was selected from total material consisting of 2191 consultations. We chose 86 consultations of the highest quality and 89 consultations of the lowest quality. Quality was assessed as described previously [3,13]. Briefly, the sum of the score consisted of ratings regarding the physician's professional skills, communication, and consultation conditions, and of the economic quality and duration of the consultation. Patients and GPs rated each consultation independently of each other with the same set of variables and scores. The selection of the material has been described in detail previously [3,13]. Briefly, approximately 5% of the highest quality consultations and 5% of the lowest quality consultations were chosen.

The patient records were analysed by two independent researchers (MK, A-LK). The documents were assessed and grouped by focusing on their medico-professional quality as poor, moderate, or good. “Good” was defined as consultations where the physician followed current care guidelines or whose performance was up to the standards of an experienced GP. “Poor” was defined as consultations where the physician did not follow the current care guidelines or the professional behaviour was not clear. “Moderate” fell in between. Both researchers reviewed the papers independently twice, after which divergent assessments were decided by consensus.

The agreement between the medico-professional quality assessments by the researchers with the physicians and with the patients was measured with the κ-coefficient and the association between the medico-professional quality of the medical records and categorical variables was analysed with the chi-squared test. The SAS statistical software, version 8.01 (SAS Institute Inc., Cary, NC, USA) was used for the statistical data processing. A two-sided p-value of less than 0.05 was considered statistically significant.

The objective of the selected consultations dealt with two common problems, acute respiratory infections (ARI) and hypertension (HT). These diagnoses were selected because they are common and treatment practices were established, even though nationally approved current care guidelines were available only for ARI in 2000. [14,15]. In the ARI group symptoms of cough, fever, or sore throat mentioned in the medical records or mentioned by the patient were triggers for inclusion of the record in the study. In the HT group the triggers were documented blood pressure readings, mention of hypertension, or a hypertension-related diagnosis, e.g. diabetes or coronary artery disease. Documentation of these triggers within the year preceding the current consultation was inclusive. Also, a mention of these diagnoses as chronic diseases was considered as inclusive for this study. The content of the records were compared with the current care practices.

Results

The medico-professional quality was good in 81 (46%) of the consultations, moderate in 63 (36%), and poor in 16 (9%). The rest (15 [9%]) could not been assessed because of missing or very poor documentation. Our assessments of the GPs and the patients’ experiences on the total quality of these consultations were in agreement [3,13] (p = 0.01), but not on the opinions concerning medico-professional quality. Our assessments diverged significantly from the assessments of the GPs (κ = 0.17) and of the patients (κ = 0.02). There was no association between our assessments of the medico-professional quality and the physician's gender, age, or speciality or of the patient's age, duration of the consultation, or number of problems. Neither did the matter of whether the GP was the patient's personal physician or not affect the association with the perceived medico-professional quality as assessed by us. There was association (p < 0.0001) between the medico-professional quality and the technical quality of the patient records [3].

There were 16 consultations on acute respiratory infections; seven concerned children. Altogether eight patients were prescribed an antibiotic, but in only two cases were both the patient's status and the reason for medication (i.e. diagnosis) marked in the medical records. The patients with cough (n = 8) were prescribed either an antibiotic or some other pharmaceutical in six cases. The smoking habits of the patients or the parents were not recorded.

Altogether 87 records of the HT group were analysed (Table I). In 55 cases, the patient had hypertension, diabetes mellitus, abnormal serum lipid levels, or arteriosclerosis. In the remaining cases the blood pressure (BP) should have measured for other indications (n = 10) or it was measured without any specific indication (n = 22). Altogether 11 of these patients had BP values which indicated mild hypertension, five of them values that indicated moderate hypertension, and one had severe hypertension. According to the medical records, only 10 patients with an abnormal BP were treated or advised appropriately. Only 24% (n = 13) of the patients with arteriosclerosis, diabetes mellitus, abnormal serum lipid levels, or hypertension had their BP followed up and treated appropriately, 12 had mild hypertension, 15 had moderate hypertension, and four had severe hypertension.

Table I. Indications for blood pressure measurement and levels of hypertension.

Indication for blood pressure measurement	n	Severe hypertension	Moderate hypertension	Mild hypertension	Normal blood pressure	No value
Hypertension	26	3	8	6	4	5
Hypertension and/or arteriosclerosis	9			3	4	2
Hypertension and/or diabetes	10		4	1	2	3
Hypertension and/or hypercholesterolemia	10	1	1	4	3	1
Other indications	10		1		5	4
“Healthy” patients	22	1	4	11	6	0
Intervention (medical treatment or advice)		2	8

Notes: Severe hypertension: systolic BP ≥ 180 or ≥ diastolic BP 110. Moderate hypertension: systolic BP 160–179 or diastolic BP 100–109. Mild hypertension: systolic BP 140–159 or diastolic BP 90–99. Normal blood pressure: < 140/90, for diabetics < 130/85.

The other risk factors for arteriosclerosis were seldom mentioned: smoking habits were marked in three records (in five cases this information was found in the records of earlier visits) and factors concerning lifestyle (nutrition or physical activity) were found in 13 records (in seven cases in the records of earlier visits).

Discussion

The medico-professional quality of the consultations assessed by analysing patient records was good in almost half (46%) of the instances. However, 9% of the records could not be assessed at all because of missing or poor documentation and 9% of the consultations were assessed as being poor. The treatment of acute respiratory infections and hypertension needs to improve in view of the finding that GPs did not follow the current care guidelines and practices adequately. Recordings of smoking habits or other lifestyle factors were rare.

The medico-professional quality of the consultations was assessed by reading and analysing the textual data in the patient records. Our qualitative method forced us to use quite a small sample of our total material. To analyse a mixture of typical consultations we selected approximately 5% of the highest quality consultations and 5% of the lowest quality consultations from our total material. We are aware that this selection and limited material cause statistical problems, but statistical analyses are not the main focus of qualitative analyses. Qualitative analyses rather aim at increasing our understanding of human behaviour and the reasons that govern such behaviour [16].

Our method was slow and uneconomical, but offered information we could not have found by processing data automatically. Most of the information is concealed in volumes of running text [9], and a lack of coding and structured data is a momentous barrier to practical and economical exploitation of clinical data. Our material was limited, but offered a realistic view into how GPs perform clinically. We used paper prints of EPR, which may have left out some material [3] and this might have affected the results, but this is also the real world for GPs: the patient's medical history is divided into separate files and may easily be missed, because it is difficult to relocate. Nevertheless, despite this caveat we consider that there was usually enough information to make the right assessments and draw the appropriate conclusions.

Although poor and incomplete patient records are problematic for researchers analysing the quality of care, the problem is more serious for the GPs who need to make their decisions without adequate current information and information on the patient's history.

The agreement of our assessments with the GPs’ or the patients’ opinions on medico-professional quality was poor. The problems of the doctor–patient relationship may affect the patients’ assessments in this setting [17]. It is also possible that the GPs’ experiences of their medico-professional performance were affected by communication. The main reason for a lack of association could be poor documentation. More research is needed into how GPs experience their medico-professional performance and their adherence to the care guidelines.

The treatment of acute respiratory infections and hypertension did not follow established practices or the nationally approved current care guidelines in Finland, which is not a novel observation [18,19]. Although antibiotic prescribing should be reduced [20,21], many patients received an antibiotic without any specific reason. Also cough medicines were prescribed, though there is no evidence for their effectiveness [22]. Both GPs and the patients were, however, happy with these consultations and with the GP's medico-professional performance. Similar findings have been reported previously [18]. The lack of continuity in care, an insecure doctor–patient relationship, time constraints, and pressures from the patient on the physician may all contribute to activities in violation of the guidelines [23]. On the other hand, maintaining a good doctor–patient relationship is not the primary reason for prescribing antibiotics – physicians are truly uncertain as to which individual patients do benefit from antibiotics [18]. Also experts experience difficulties in finding quality indicators which are relevant for patient health benefit or which focus on the diagnostic process of respiratory infections [24].

There is also need for improved care of hypertension: the BP reading was not documented in 20% of the records of patients with a risk of arteriosclerosis. Only one-fourth of the patients with hypertension had appropriate treatment for hypertension, as is already known from Finland [19] and Sweden [25]. The GPs seldom reacted to high blood pressure or discussed smoking cessation, healthy nutrition, active physical exercise, or other preventive factors. Similar results have been reported in Denmark concerning treatment of dyslipidemia [26].

Although GPs are familiar with the current care guidelines [27,28], there are still barriers to their implementation [25]. With the help of efficient and user-friendly EPR some of these barriers could be avoided, e.g. the system could remind, warn, or guide GPs in making decisions. If a physician deviates from the recommendations of the current care guidelines, this needs to be recorded, together with the reason for doing so, in the patient records.

EPR are widely used in Finland and the use of them is increasing in many countries [3,6,29,30]. EPR could be much more useful for research, management, and quality improvement, but this demands that they are user-friendly for the main users, the physicians [30]. If a physician is not able to extract the patient's history and clinical information effortlessly, and gets no feedback on his/her medical performance, the likelihood of quality improvement in everyday clinical work is low. Much improvement remains to be made to make EPR systems user-friendly and interactive.

Acknowledgements

The authors would like to thank Robert Paul for language checking.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.