The documentation of health problems in relation to prescribed medication in people with profound intellectual and multiple disabilities

Background  Persons with profound intellectual and multiple disabilities (PIMD) suffer from a wide range of health problems and use a wide range of different drugs. This study investigated for frequently used medication whether there was a health problem documented in the medical notes for the drug prescribed. Method  Persons with PIMD with an estimated intelligence quotient of 25 and profound or severe motor disorders were studied. Data on health problems were taken from medical notes and prescribing data were obtained from pharmacies. Data covering 1 year were analysed. For four therapeutic areas (anticonvulsants, laxatives, drugs for peptic ulcer and gastro‐oesophageal reflux disease and psycholeptics), we determined whether we could find an indication for prescribed medication. Results  Some 254 persons with PIMD (46% male, 54% female; median age 49 years, range 6–82) from eight residential facilities participated. Some 226 participants (89%) were prescribed medication over the course of 1 year. An indication for the prescribed medication was documented for 92% (n = 130) (95% confidence interval 88–96%) of 141 participants on anticonvulsants, for 68% (n = 112) (61–75%) of 165 participants on laxatives, for 44% (n = 58) (36–52%) of 132 participants on drugs for peptic ulcer and gastro‐oesophageal reflux disease, and for 89% (n = 102) (83–95%) of 115 participants on psycholeptic drugs. Conclusions  The best level of documentation was found for anticonvulsants the worst for drugs for peptic ulcer and gastro‐oesophageal reflux disease. Lack of documenting an indication may be due to off‐label use, inadvertent continuation of no longer indicated medication, inadequate documentation and underdiagnosis. Adequate documentation practices are essential because of the communication problems that are characteristic for persons with PIMD.

Keywords: medical records; mental retardation; multiple disabilities; profound intellectual and multiple disabilities; quality assurance; health care; drug utilisation

Persons with profound intellectual and multiple disabilities (PIMD) have such intellectual disabilities (ID) that existing standardised tests are not applicable for a valid estimation of their level of intellectual capacity and therefore their estimated intelligence quotient is set at 25. These ID are combined with profound or severe motor disabilities and minimal communication skills ([16]). These persons need intensive and continuous support not only because of the severity of their intellectual and motor disabilities but also because they frequently suffer from additional health problems, such as constipation ([5]; [19]), epilepsy ([10]; [1]; [21]), sleep disorders ([17]; [7]), recurrent respiratory infections ([21]) and gastro‐oesophageal reflux disease (GORD) ([4]). Behavioural problems are also frequently found ([2]; [21]). Prescribing medication is the most frequent intervention ([14]).

A range of studies have investigated prescribing patterns of psycholeptic and anticonvulsant drugs in people with ID ([10]; [15]; [11]; [18]). These studies raised concern about the quality of prescribing of psychotropic medication in this group ([15]). For example, [11]) showed that psychotropic medication was often prescribed without being indicated by a diagnosis. There is less research on the quality of prescribing in other pharmacotherapeutic areas. Correctly diagnosing and prescribing is especially important because of the difficulties of people with PIMD to communicate about medical problems, effects and side effects of medication and also because of the magnitude of the medication use. Good documentation of medical treatment is an essential step in this process. We therefore investigated prescribing patterns in PIMD. For frequently used medication, we investigated whether there was a health problem documented in the medical notes for the drug prescribed.

At 10 facilities for people with ID throughout the Netherlands, all persons with PIMD were selected, using the following criteria: (1) being a resident of that facility for at least 5 years; (2) intellectual development up to 24 months (estimated intelligence quotient of 25 points) ([16]); (3) profound or severe motor disorders, leading to incapacity to move oneself ([16]); and (4) written informed consent by parents or legal representatives. Ethic's committee approval was obtained prior to the study.

The data used during this study existed of information on health problems and prescription data. A questionnaire, specially designed for this investigation, was used to collect the data about health problems. Information was collected from the medical notes and if necessary additional information was obtained from the physician or nurse. The following data were collected: (1) age; (2) gender; (3) facility; (4) the cause of the disability; (5) motor disabilities; (6) sensory problems; (7) health problems; and (8) behaviour disturbances.

Prescription data were collected from the pharmacy where the participant was registered. This was possible because Dutch pharmacies keep a prescription record of all their clients. Data included information on the prescribed drug, date of prescription and duration of therapy. The present study was part of a larger project, so data were collected from the previous 5 years. However, in the present study only prescribing data (from the 12 months), preceding the date when data collection from medical notes took place, were used. The data were electronically supplied by the pharmacies. All data were stored in a relational database (version: Apodat 2.8, My SQL of Sun Microsystems, Inc., Santa Clara, USA).

Drugs were classified using the anatomical therapeutic chemical (ATC) classification system. This is an internationally used system controlled by the [20]). It classifies drugs into different groups according to the organ or system on which they act and/or their therapeutic and chemical characteristics. Drugs are classified in groups at five different levels. The drugs are divided into 14 main groups (first level), with one pharmacological/therapeutic subgroup (second level). The third and fourth levels are chemical/pharmacological/therapeutic subgroups and the fifth level is the chemical substance ([20]).

The prevalence of the prescribed medication (ATC‐code level 5) was determined. The following four therapeutic areas were selected for further analysis: drugs for peptic ulcer and GORD, laxatives, anticonvulsants, and psycholeptics. These were chosen because these were frequently prescribed drugs used for chronic treatment. Furthermore, they have only a limited number of indications, so we were able to establish a link between the drugs and specific health problems. For example, laxatives are only indicated to treat constipation and to prevent constipation because of therapy with opioids. The indication of drugs was based on the standard Dutch drug reference guide ([8]). The health problems as noted in the medical notes did not always use the same terminology as in the reference guide; therefore synonyms and more vague descriptions of the indications were also included. Some participants used multiple drugs from the same group, and those cases were counted once.

We investigated whether there was a match between registered health problems and prescribed drugs. There were four possible situations: (1) a person had a registered health problem and was prescribed drugs for this problem; (2) a person had a registered health problem, but no drugs were prescribed for this problem; (3) a person was prescribed drugs, but there was no health problem registered justifying the prescription of this particular drug; and (4) a person had no health problem and was not prescribed drugs either.

The first and the last situation indicated that prescribing was in agreement with the health problems registered in the medical notes. The second situation was more complex therefore we named it doubtful. It could be that non‐ pharmacological measures were used to treat the health problem, the health problem could not be so serious that it did not require pharmacological treatment, or there could be undertreatment. The third situation suggested that prescribing was not in agreement with the registration of health problems, i.e. we could not determine the indication of the prescription. For each therapeutic area, we determined the number of persons in each situation. Furthermore, we determined the percentage of people for whom we could find an indication (situation 1) out of all people who were prescribed medication from this therapeutic area. We also calculated 95% confidence intervals (CI).

Overall, 435 people were eligible for the study and received a letter to participate in the study. Of these, 309 participants responded (71%). Consent was given by 301 participants (eight rejected). Data from 254 (58%) participants from eight residential facilities could be used in the study. A total of 47 participants dropped out because of death or relocation of the participant (n = 18) or incomplete prescribing data (n = 29). The latter persons lived in two facilities that were covered by the same pharmacy. This pharmacy could not deliver the available data because of practical reasons.

They were 46% male (n = 117) and 54% female (n = 137). Median age was 49 (range 6–82). Age distribution was: 5–18 years, 7% (n = 18); 19–37 years, 21% (n = 53); 38–57 years, 49% (n = 124); ≥58 years, 23% (n = 59). Table 1 gives an overview of the predominant diagnosis and additional problems.

1 Overview of predominant diagnosis and additional health problems of the participants (n = 254)

Two‐hundred and twenty‐six participants (89%) were prescribed medication in the course of the 1‐year period. Forty per cent (n = 101) were prescribed five or more medications. On average each participant was prescribed 3.8 (standard deviation = 2.7) medications. Most frequently prescribed were laxatives (n = 165, 65%), anticonvulsants (n = 141, 56%), drugs for peptic ulcer and GORD (n = 132, 52%), antibacterials for systemic use (n = 121, 48%) and psycholeptics (n = 115, 45%). Table 2 shows more details of the four therapeutic groups which were further analysed.

2 Drugs prescribed to the participants (n = 254)

1 *  Sum of medications prescribed for five or less participants.

Overall, 71% of participants suffered from epilepsy and 56% of the total sample were prescribed anticonvulsant medication (Table 3). In 92% (n = 130/141; 95% CI 88–96%) of the cases where a participant was prescribed an anticonvulsant drug, epilepsy was a registered health problem.

3 Prescribing of medication in relation to the documentation of relevant health problems

2 *  Epilepsy was a possible diagnosis for the prescribing of diazepam (N05BA01), midazolam (N05CD08), and clobazam (N05BA09). GORD, gastro‐oesophageal reflux disease.

Constipation was a registered health problem for 60% of the participants and laxatives were prescribed for 65% of the participants (Table 3). Most frequently used laxatives were bisacodyl and macrogol combinations (Table 2). Constipation was a registered health problem in 68% (n = 112/165; 95% CI 61–75%) of the cases when a laxative was prescribed.

Drugs from the ATC‐group A02B can be used for peptic ulcer and GORD and related disorders as well as to prevent the gastrointestinal side effects of non‐steroidal anti‐inflammatory and antirheumatic drugs (NSAIDs). GORD was registered for 62 (24%) of the participants, nine participants (4%) had oesophagitis, two participants (1%) had a diagnosis of peptic ulcer disease, and 30 (12%) participants were prescribed NSAIDs. Fifteen participants had more than one of these health problems. A relevant health problem was registered in 44% (n = 58/132; 95% CI 36–52%) of the participants who were prescribed a drug from A02B (Table 3).

Behavioural disturbances, sleep disorders and/or epilepsy were documented in 70% of the medical notes (Table 3). Overall, 45% (n = 115) of the population were prescribed psycholeptics; of those 89% (n = 102; 95% CI 83–95%) had a related health problem and 11% (n = 13) had not.

In this study, we have investigated the match between the documentation of health problems and prescribed medication for frequently used medication in persons with PIMD. The best level of documentation was found for antiepileptic medication. An indication was documented for over 90% of persons who were prescribed antiepileptic medication. For the other three therapeutic areas, this was lower, ranging from 89% for psycholeptics to only 44% for peptic ulcer and GORD medication. Good clinical practice requires the documentation of an indication for every prescribed medicine.

A number of possible scenarios have to be discussed. First, the medication could have been used for other disorders than the ones we have investigated including off‐label use. For example, anticonvulsants, such as carbamazepine and valproic acid, are sometimes used as mood stabilisers. This may have been the case in three of the 11 persons who used anticonvulsants but did not have epilepsy. Furthermore, laxatives can be used to prevent side effects from opioids, drugs for GORD can be used to prevent side effects of NSAIDs, but none of the participants were prescribed opioids, and we included participants who were prescribed NSAIDs in the analysis.

Second, the health problem may have been present in the past triggering the prescribing of medication. While the condition has long resolved, inadvertently the medication has never been stopped. Medications initially prescribed for a short‐term treatment may unintentionally be used chronically. Such prescribing practices have been found in a recent study in nursing home patients ([9]). This is a particular risk in patients who receive many different drugs (polypharmacy), such as our patients.

Third, an adequate diagnosis has been made, but it has not been documented in the medical notes. This would suggest that documentation practices vary widely according to conditions, i.e. epilepsy is well documented and other conditions, such as GORD, are not carefully documented. It could be that physicians perceive that a 'disease', such as epilepsy, requires a higher level of documentation which is not required for 'conditions', such as GORD.

Finally, medication could have been prescribed on the basis of complaints and symptoms of the patient (which were not adequately documented so we were not able to retrieve them) without attempting to diagnose the condition. This is the most likely explanation for the very‐low level of documentation of indications for drugs used for peptic ulcer and GORD. [4]) showed that GORD is underdiagnosed in people with ID, because the symptoms are non‐specific and the examinations to confirm a diagnosis are incriminating for the patient. In a cross‐sectional study, [3]) found a prevalence of GORD of almost 50% in institutionalised persons with an IQ below 50. This would correspond with our findings that more than half of the persons were prescribed a drug against GORD. However, [4]) warn against the use of drugs against GORD without adequately diagnosing the condition. Although the level of documentation was better for the prescribing of laxatives and psycholeptics, similar problems have been found in both areas. [5]) found that constipation occurred frequently in people with ID (prevalence of 70%) but was not always recognised easily. Behavioural disturbances were documented very frequently, but communicative barriers increase the difficulty of making a diagnosis of mental health problems, such as depression ([21]). Methods for diagnosing sleep problems are time‐consuming and not always with result ([7]), and most of the treatment studies on sleep problems have been case studies with limited controls for internal validity and limited data on generalisability and maintenance ([6]).

Some methodological issues of our study need to be addressed. Although the response rate of 71% (n = 309/435) is acceptable, as stated complete data of only 254 participants were available. The majority of those dropping out were persons belonging to the same pharmacy that could not deliver the available data because of practical reasons. Furthermore, the participants were selected using internationally accepted criteria ([16]). However, a small part of the sample might consist of persons who may not belong to the core group of persons with PIMD because of old age (up to 82) and because they do not suffer from severe motor disabilities (14% of the sample) ([16]). Other characteristics, such as the prevalence of general health problems, are comparable to figures in the literature ([5]; [2]; [21]; [19]; [7]). In our study, we focussed on the mismatch between documentation of indication and prescribing. We did not assess the adequacy of treatment for the cases where a diagnosis has been documented in the medical notes. Adequate documentation is the basis to evaluate the quality of prescribing. Further work is needed on the evidence base to support prescribing in this vulnerable patient group. Similarly, we did not investigate the cases where there was a health problem documented, but no medication prescribed. As we have already highlighted in the methods section there may be various reasons why clients do not have medication prescribed to treat their health problem. On the basis of the data we obtained in the present study, we could not differentiate between cases of undertreatment and adequate use of non‐pharmacological measures.

In contrast to previous studies in other groups with ID we found possible indications for the prescribing of psycholeptic drugs for the majority of clients. [10]) found non‐significant associations between the administration of major tranquilisers and reported behaviour problems in children and adults with PIMD who live at home and attend schools and adult training centres. [11]) found that the use of only 54% of psychotropic drugs was justified by psychiatric diagnosis and/or symptoms in persons with ID ranging from mild to severe. Similar results were found by the Dutch Health Care Inspectorate ([12]). It may be that these studies have lead to improved prescribing of psycholeptic drugs. But the differences may also be due to differences in definitions. We included a broad range of possible indications for the prescribing of psycholeptic drugs. If we only included behavioural disturbances or sleep disorders, excluding epilepsy, only 60% (n = 69) of clients would have had a documented health problem.

The systematic evaluation and monitoring of the relation between medication use and health problems is essential because of the communication problems that are characteristic for this group. Persons with PIMD are totally dependent on their carers to recognise specific behaviour or symptoms. Our study showed it is likely that clients receive medication which is not or no longer necessary, putting them at risk to experience adverse effects because of this. Furthermore, these deficiencies in documentation practice may mean that clients suffer from health problems that are not recognised by carers, and thus are not treated, like behaviour problems caused by GORD. In conclusion, more efforts are needed to adequately diagnose GORD and constipation for effective and safe use of medication. A regular (multidisciplinary) medication review has been successful in other areas to improve the quality of prescribing and patient care ([13]). Adequate documentation practices are essential to communicate about health care for people with PIMD.