A comparison of computer- and hand-generated clinical dental notes with statutory regulations in record keeping
Applied Clinical Research and Public Health
School of Dentistry
Cardiff CF14 4XYN,
Tel: 02920 742614
Fax: 02920 742614
Introduction: Dental patient records should be of high quality, contain information to allow for good continuity of care and clinical defence (should the need ever arise) and, ideally, facilitate clinical audit. Handwritten dental records have been assessed for their compliance to statutory regulations, but the same cannot be levelled at computer-generated notes. This study aimed to compare and analyse the compliance of both methods of data recording with statutory regulations.
Method: Fifty consecutive sets of handwritten notes and 50 sets of computer-generated notes were audited for compliance with a number of legal requirements and desirable characteristics for dental records and the results compared. The standard set for compliance with all characteristics was 100%.
Results: The computer-generated notes satisfied the set standard for 8 of the 11 legal requirements and three of six desirable characteristics. The handwritten notes satisfied the set standard for 1 of 11 legal requirements and none of the desirable characteristics. A statistical difference (using a 95% confidence interval) between the two methods was observed in 5 of 11 legal characteristics and three of six desirable characteristics, all of which were in favour of computer-generated notes.
Conclusions: Within the limitations of this study, computer-generated notes achieved a much higher compliance rate with the set parameters, making defence in cases of litigation, continuity of care and clinical audit easier and more efficient.
It is a requirement of the General Dental Council (UK) that all dentists produce and manage patient records appropriately and at the time of treatment (1). Similar requirements are in force throughout clinical practice in the United States and Europe. Clinical records are paramount in recording provision of dental care, allow for appropriate continuity of care and provide for correct sequence of treatment (2). This is particularly important in environments where numerous dental practitioners or dental professionals treat patients over a given time frame (3).
Regardless of place of practice, a good standard of record keeping is essential (4). In cases of litigation, this is something regularly advocated by dental protection societies with the aphorism ‘good records facilitate good defence, poor records a poor defence and no records no defence’, being more applicable today than ever (5).
In addition to the above, dental audit projects can rely heavily on what is recorded in dental records (5). Therefore, it is paramount that dental records are of a high standard if this essential aspect of clinical governance is to be carried out and improvements to the standard of patient care made.
The current practice of keeping clinical records is moving from the traditional (handwritten) method to more contemporary (computer-generated) methodologies, and with this change, debate has arisen as to which should be the method of choice in carrying out this essential requirement of a dental practice. Handwritten records have long been audited for their quality and content (2, 5–7), and it has been noted on more than one occasion that the standard of these records can be less than satisfactory (2, 8). A significant shortcoming of handwritten records has been found in the past. Pessian and Beckett (3) noted, during an audit of records written by undergraduates, that 15% lacked an updated medical history and nearly a third omitted patient details and treatment plans. Such findings are not exclusive to hospital records; Morgan (2) found almost 50% of the records of the NHS and private dentists assessed did not contain an updated medical history as well as over 70% of the records not containing full charting.
Whilst there are no uniformly acceptable guidelines to describe what a dental record should contain, there are a number of fundamental characteristics that are frequently described as being desirable; clinical records need to be assigned to specific patients, be dated, be legible, be legibly signed, with a procedural record noted and medical history validated (2, 3, 5, 9, 10).
Computerisation of dental records introduces challenges and requirements upon dental professionals as patient records are subject to the Data Protection Act (1998) and the Freedom of Information Act (2000). They must only be used for the purpose they were collected, and adequate security measures be in place to protect them from unauthorised access with regard to the first Act (11). Both Acts also allow patients to request the information held about them, therefore making good record keeping essential (11, 12). There has been a comparison made of computer-generated and handwritten records in general dental practice in the UK with the computerised records shown to be significantly better in detailing a number of generic criteria such as patient identifiers, medical history, dental charting, periodontal condition, soft tissue examination, treatment provided and treatment plans (4).
Because there is debate over the best method of generating dental records, the aim of the current study was to address the quality of the record, establish and compare the strengths and weaknesses of both methods of data recording and ascertain which resulted in the best clinical record as measured against a number of legally required and some desirable characteristics.
This retrospective study examined entries in one hundred patient records generated by final-year dental students from Cardiff University School of Dentistry. The sample was made up of 50 consecutive sets of handwritten patient records from the School of Dentistry and 50 consecutive sets of ‘computer-generated’ patient records from an outreach centre based in St David’s Community Hospital Cardiff. In this instance, the sample was determined for ease of handling, convenience and suitability for data handling. The computer-generated entries were made in an electronic patient record system (SALUD) developed by the Dublin-based company, Two Ten Health. It was the same students and the same staff who supervised them as they made their entries at both locations. Students had approximately 18 months of experience of working with computer-generated records and around 24 months of clinical experience with handwritten records. The records were assessed by a single clinician who evaluated a number of legal requirements and desirable characteristics contained within the clinical record (Table 1). In all cases, it was the last clinical entry from the collected files that was evaluated. The proportions of records from the Dental School and Hospital which demonstrate compliance with legal and desirable requirements are given together with 95% confidence intervals for the difference in proportions (13).
Table 1. The legal and desirable components of clinical patient records
|Patient ID on each page of last entry||Author’s name in capitals|
|Permanent black ink||Appropriate corrections|
|Medical history checked||Investigation reports|
|Procedure recorded||Plan for next visit|
|No ambiguous abbreviations|| |
|Correct analgesia record|| |
|Advice recorded|| |
|Identifiable signature|| |
|Countersignature identifiable|| |
Table 2 details the proportion of records from each location, which included information compliant with legal requirements of patient records. With reference to the fields evaluated, all were based on 100 records except the analgesia data because not all patients received analgesia on the visit pertaining to the record’s entry. It can be seen that there were a number of differences between the two methods of recording data with respect to a patient’s attendance at clinic. Five of the 11 legal requirements showed a statistically significant difference in the favour of electronic records: medical history checked, correct analgesia record, advice recorded, identifiable signature and an identifiable countersignature. Two requirements showed no difference at all, those being recording of the date and no ambiguous abbreviations, and four legal requirement differences (patient identification on last page of entry, legible entry, use of permanent black ink and procedure recorded) were not statistically significant but showed a small difference in favour of electronic records.
Table 2. Compliance with legal requirements
|Patient ID on each page of last entry||50||46 (92)||50||50 (100)||8.0 (−0.6 to 18.8)|
|Date||50||50 (100)||50||50 (100)||0 (−7.1 to 7.1)|
|Legible||50||46 (92)||50||50 (100)||8.0 (−0.6 to 18.8)|
|Permanent black ink||50||46 (92)||50||50 (100)||8.0 (−0.6 to 18.8)|
|Medical history checked||50||37 (74)||50||50 (100)||26.0 (13.6 to 39.6)|
|Procedure recorded||50||49 (98)||50||50 (100)||2.0 (−5.3 to 10.5)|
|No ambiguous abbreviations||50||49 (98)||50||49 (98)||0 (−8.7 to 8.7)|
|Correct analgesia record||21||14 (66.7)||19||19 (100)||33.3 (10.0 to 54.6)|
|Advice recorded||50||9 (18)||50||21 (42)||24.0 (6.0 to 40.0)|
|Identifiable signature||50||38 (76)||50||50 (100)||24.0 (12.0 to 37.4)|
|Countersignature identifiable||50||27 (54)||50||46 (92)||38.0 (21.1 to 54.0)|
Table 3 details the results with respect to desirable characteristics in a clinical record. It is noted that some fields (appropriate corrections and investigation reports) contained <50 entries. There were a number of differences between the two methods of data recording in fulfilling the desirable requirements. All 50 (100%) computer-generated records showed both the author’s name and designation compared with 36% and 86%, respectively, in the 50 handwritten records. Both of these comparisons were statistically significant. Although the numbers of entries for the desirable characteristic of recording appropriate corrections were very small, there was a significant difference in favour of the computer records. Two characteristics did not show a statistically significant difference but did demonstrate a difference in favour of computer-generated notes, those being recording of presenting complaint and investigation reports. One characteristic showed a difference in favour of the handwritten method, this being plan for next visit; however, it was not statistically significant.
Table 3. Compliance with desirable characteristics
|Author’s name in capitals||50||18 (36)||50||50 (100)||64 (48.4 to 75.9)|
|Author’s designation||50||43 (86)||50||50 (100)||14 (4.0 to 26.2)|
|Presenting complaint||50||19 (38)||50||20 (40)||2 (−16.6 to 20.4)|
|Appropriate corrections||14||2 (14.3)||4||4 (100)||85.7 (30.4 to 96.0)|
|Investigation reports||16||8 (50)||13||10 (77)||27 (−8.0 to 53.5)|
|Plan for next visit||50||43 (86)||50||36 (72)||−14 (−2.1 to 29.4)|
This study was carried out by one assessor to reduce variability and increase reliability. It followed a method that was adapted from previous audits of NHS practice (14) and used criteria that would be required at any visit and omitting requirements of dental records that would apply only to initial consultations. This differs from previous studies (4,14) that examined more generic fields with respect to general record keeping. It was felt appropriate for this study to concentrate on what had previously been described as legally required and desirable characteristics with respect to record keeping and by limiting the fields prevented an unnecessarily complicated protocol or analysis. The sample group used only final-year dental students for both methods of recording the patient notes, thus increasing validity as only final-year students use both methods to write the patient records at Cardiff Dental School and they essentially acted as their own control groups. This study was retrospective, and the students were not aware that such an evaluation of clinical records was planned; this prevented a biased sample being generated by students consciously creating good or bad record entries.
The study compared handwritten records and computer-generated records against a number of clearly identifiable parameters and showed that when dental students use computer software to record a dental record, they adhere to a much closer degree to a set of legal and desirable parameters for clinical records.
Recorded advice and an identifiable staff countersignature were present in 42% and 92% of computer-generated records and 18% and 54% of handwritten records, respectively. It is noteworthy that the identifiable staff countersignature on these computer-generated records was in this case ‘virtual’ as their name is assigned a password that is only known to them with their name appearing on the patient note once the password is inputted. A staff countersignature was absent in 8% of computer-generated notes; if these records had been authorised, 100% of the computer-generated notes would have been noted as having an identifiable staff countersignature. Therefore, of the records that had a countersignature, 100% were legible.
One hundred per cent of the computer-generated records contained patient identification, were legible and were presented in a permanent black ink format compared with 92% of handwritten records containing each of the requirements. Computer-generated records would always contain these requirements as the software will not allow for their exclusion.
Corrections to the dental record were observed in only 14 handwritten and 4 computer-generated records, with an appropriate correction being a single line placed through the mistake, which was noted in only two handwritten but all four computer-generated records. It is appreciated that there is a very small sample size for both types of notes; however, the computer-generated notes will always show the original note with a single line placed through it if it is modified after saving and therefore will always achieve 100% compliance to the study parameters if a correction is made.
The only feature that the handwritten records were shown to have adhered to the study guidelines over the computer-generated ones was the recording of a plan for treatment at the next visit.
It is believed that presenting complaint is required for every entry into the patient record, and whether the patient presented with a complaint or not it should be noted that the question was asked. Although these methods were not performed in every case, the results were very similar for both methods.
Computer-generated records achieved a 100% compliance rate with 8 of the 11 legal requirements vs. handwritten records achieving a 100% compliance rate with just one legal requirement, showing that computer-generated records induce compliance with the inclusion of legal requirements in dental record keeping to a far greater degree than handwritten records. Of the three legal requirements that did not achieve 100% compliance for computer-generated records, two of them, advice recorded and identifiable countersignature, did achieve a difference in favour over handwritten records.
Two of the three legal requirements that did not meet the standard for computer-generated records could quite simply achieve 100% compliance with small changes to the computer software. To address the ‘advice recorded’ requirement, an extra input box could be added to the record, titled ‘Advice given’ which would prompt the user to record the advice they had issued the patient at each appointment, therefore likely to increase compliance and achieve the standard. ‘Countersignature identifiable’ did not achieve a 100% compliance rate as a number of patient records had not been digitally authorised by the supervising clinician; this could be addressed by not allowing the student to log off the computer software unless all new inputs in that session had been authorised by the supervising clinician. These minor improvements could lead to 10 of the 11 legal requirements being fulfilled consistently for every dental record.
The 10 legal requirements for handwritten records that did not meet the set standard may be addressed in one of two ways: in line with addressing ambiguous abbreviations in computer-generated records, retraining or issuing guidance to the students and staff containing information showing exactly what is required of them when generating a handwritten record, this may increase compliance and achieve a 100% compliance rate. Alternatively a printed template could be used for each appointment entry, with specific boxes prompting the student to write the required information, ‘Date’, ‘Medical history change? If yes, note changes’, and so on. If any box were to be left unfilled, the student could see that they had not completed the note in line with what is required of them. This, however, would impose restrictions of the author’s ability to write as much as is required of them and may not ensure 100% compliance.
Improving handwritten records to include some of the desirable characteristics identified could be achieved if the methods suggested above could be implemented for the legal requirements, and as such would not guarantee their inclusion but increase the likelihood of their inclusion. The compliance of computer records for desirable characteristics could be increased by adding extra input boxes for ‘Presenting complaint’, ‘Investigation report’ and ‘Plan for next visit’.
It is a concern that the handwritten records fell below the standard required of them: 26% of clinicians did not check the patients’ medical history or if it had changed since the last visit, 33.3% did not correctly record the analgesia administered and 82% failed to record the advice given to the patient. This shows that the patient is not receiving the high standard of care that the clinician is required to give and leaves the clinician open to the distinct possibility of a successful law suit being brought against them if any discrepancy were to arise. Identifying the student operator and staff member responsible for them would also prove difficult as 24% of students and 46% of staff could not be identified.
Again, with respect to litigation and best practice, the handwritten records show 85.7% of corrections not being deemed as appropriate with some of these being heavily blacked out sentences that cannot be read, which would leave a prosecution lawyer in a litigation case arguing the clinician had something to hide. Investigation reporting was poor with both methods of recording, with only half of the appropriate handwritten notes containing a report and 23% of the applicable computer-generated records not containing one.
Handwritten records have continuously been proven to be ineffective at capturing all the information required of them by assessors (2, 3, 7). Persian and Beckett (3) found even after a repeat audit, patient’s medical histories were not being checked at each visit; at best, the sample group achieved an 85% compliance rate. Smith and Farrington (7) carried out two audits of nine dentists, auditing dental record legibility. The range of legible records found by the first audit was between 24.2% and 100% with the second audit finding a range of 53% to 100%.
The current study shows there is a need to address illegibility of handwritten records, and one such method would be the use of computer software which would ensure 100% legibility with every record. This study has also demonstrated that computer-generated dental records have a better compliance with the statutory regulations. It is believed that changes to increase compliance to 100% would be minimal for computer-generated records, whereas handwritten records would require greater changes and would not necessarily guarantee an improvement in compliance. It appears that computer-generated notes therefore allow for better continuity of care and provide a sound basis for improvements in standard of care by providing a detailed dental record for clinical audit. They also provide for a better legal defence in the event of litigation. It is accepted that computer-generated records created through the use of different softwares to that of SALUD may have produced different results; SALUD, however, is a relatively established piece of computer software, and besides being used in the UK and Ireland, it is also used in Norway, Spain, Hong Kong, Australia and the United States. This study also highlights the value of instituting good clinical practice with respect to record keeping at the undergraduate level, which will translate into effective record keeping in clinical dental practice. Clearly, the effectiveness of different computer-generated records to that studied here would require separate comparison and evaluation.