(1996; 157 pages)
7.1 Summary of conclusions of the injection practices research
The injection practices research aimed to estimate the extent to which injections are used as a route for the administration of medications (prevalence of injection use); to determine the type and degree of improper and unsafe practices in the process of administration of injections (the medical and hygienical appropriateness) and to gain insight into the question of why injections are so popular. The last objective of the research was to develop a simple and rapid survey methodology for future assessments of inappropriate injection use.
The extent of injection use
The prevalence of injection use was established by household surveys focusing on injection prevalence in households with pre-school children in urban, rural and semi-rural communities in two different districts of the two countries. In both countries, the percentage of households receiving one or more injection in the past two weeks was high, ranging from around four in ten households in Indonesia, to about three in ten in Uganda. Taking the total research population as the denominator, it could be established that, in the Indonesia sample, 10% of all household members in Lebak had been injected in the past two weeks, and in Lombok 12%.
Differentials between the districts, and between urban and rural areas have also been studied. There are no differences in prevalence between the districts in both countries. In Busoga, Uganda, injection use prevalence is significantly higher in the semi-rural and urban areas and lowest in the remote areas, but in Ankole no differences were found. In both districts in Indonesia, no differences between the injection rates of the urban, suburban and rural households could be established.
The differences in prevalence rates between Indonesia and Uganda can have many explanations. For example, disease patterns and family size can differ from one country to another. For this reason, the plan was to also describe injection prevalence for the general population by age category using individuals as the unit of analysis; and to assess the proportion of injection use in specific tracer conditions. The relative frequencies on the type of injections received (i.e. immunizations, therapeutic injections or contraceptives) were also to be described.
The injection prevalence in the population based on age and sex could only be calculated in Indonesia. There are no major differences in injection use in Indonesia based on sex, but the very young constitute a high risk group for receiving injections. Some 20% of the under-fives in the households had received one or more injections in the past two weeks. Since most immunizations occur in this age group, the data should be adjusted for this factor. However, in both districts only a slight percentage of all injections are immunizations (less than 5%). The percentage of children under five receiving an injection is twice as high as the average injection rate for the research population.
Type of injections: the great majority of the injections reported in the two-weeks recall period in both countries had been given for therapeutic reasons (80-90%). Intravenous drips are not very common in the countries under study. Immunizations are infrequently reported (accounting for just 6 to 15% of the households who had received an injection in Uganda and 3 to 4% of household members receiving an injection in Indonesia). Contraceptive injections were only reported in Lebak (1%).
The study also investigated illness-related injection use in the households. The burden of illness is quite high. In both countries, some 70% of households reported having had one or more sick members in the past two weeks. In Uganda, some 40% of the households with ill subjects reported having used injections in the past fortnight. There were no differences between Busoga and Ankole. For Indonesia, data are available on illness episodes and health seeking behaviour of all research subjects. Over one-fifth of all subjects in Lebak and Lombok claimed to have suffered an illness episode in the past two weeks. While illness rates are similar in both districts, the percentage of individuals seeking medical attention is significantly higher in Lombok. In addition, more patients who seek help in health facilities are being injected in Lombok (62% against 59% in Lebak). It can be calculated from the household survey that in Lebak, of all visits to formal health facilities, between 40 to 56% end in one or more injections; in Lombok these percentages are much higher: between 70 and 83%.
There is a marked difference between the two countries with respect to the source of the injections received in the households. The bulk of the injections received in the Indonesian household survey originate from the public sector. Over half of the injected persons receive their injections in the so-called puskesmas (public health centre). The share of the private health services (both formal and informal) is much smaller, but there are differences between the two districts. In Lebak, private practices of nurses accounted for 20% of the injections received, in Lombok this was only 1%. Here, doctor's private practices are a far more important source of injections (20%) than in Lebak (10%). Only a small number of injections in Indonesia is given both at home and by non-medically trained personnel. Providers receive their supplies of injection equipment from the government, but they also supplement their stock through purchases from pharmacies and wholesalers. For this, higher injection fees are often demanded from customers.
In contrast, in Uganda only a minority of the last injection to be received in the household was given in the government health facilities: 35% in Busoga and 23% in Ankole. Private medical practices are far more popular. Most striking, however, is the fact that many injections are given by non-formal providers or at home (by family members). In Busoga, 11% of the last injections were given at a non-formal facility (in Ankole this percentage is 15%) and 17% at home (in Ankole 16%). This reflects the trend of informalization in Uganda where public facilities are often mistrusted and held responsible for the spread of the AIDS epidemic. The domestication of injections is demonstrated by the fact that in Uganda, the majority of households keep injection equipment at home. In the public and NGO sector, the injection equipment usually originates from official (donor) sources. The informal providers and many of their customers buy their equipment over the counter in pharmacies.
The study also investigated injection use at the level of health facilities. In Uganda, prescriptions of 30 consecutive patients in formal health facilities were recorded. Injection rates are fairly high in both regions: an injection is given in between six to seven out of ten treatments. In Indonesia, patients - mainly in public health facilities - were interviewed about their complaints and the treatment given. Of every ten patients treated in Lebak, seven received an injection. In Lombok, the mean injection rate in public health facilities is even higher: almost nine out of ten visits here end with one or more injections being given. Injection rates in Lombok are significantly higher than in Lebak.
The medical and hygienic appropriateness
The medical appropriateness of injection use was researched by means of investigating actual and preferred practices in a few selected tracer conditions, representing uncomplicated, non-severe and self-limiting illnesses. In both countries, high rates of injection use were revealed. Most often injections were combined with oral medication. The presence of actual tracer conditions was recorded in the household questionnaires in both countries and in exit interviews at health facilities (Indonesia). Hypothetical illness cases were presented in the household questionnaires, in focus group discussions (Indonesia) and during in-depth interviews (Uganda).
In Uganda the most common tracer condition is fever. This condition was most often treated with injections - especially when the fever was accompanied by the presence of other symptoms. In both regions over 95% of all injections prescribed were chloroquine, Penicillin Procaine Fortified (PPF) and crystalline penicillin. Many providers prescribe a special combination of injectables (PPF and chloroquine) especially for patients with fever and cough. The popularity of this combination is basically the result of poor diagnostic capacity. In some health facilities, chloroquine is even used to reconstitute PPF or crystalline penicillin, resulting in a potentially dangerous, hypertonic solution. When asked which treatment is preferred should one of the tracer conditions mentioned occur in the household, the replies of the respondents reflect the tendencies described for actual use of injections.
In Indonesia, injections are given in over half of the illness cases in the households in which the tracer conditions were recorded. The highest injection rates in both areas were found in the treatment of skin diseases (some 60%). In the other tracer conditions, injection use varies from 33 to 56%. Injection rates are even higher in the exit interviews at health facilities. In Lombok injection rates are again consistently higher than in Lebak (an average of 85% of all patients with tracer conditions injected at the health facility in Lombok and 64% in Lebak). A review of patient records at health facilities revealed that the most commonly used injectables include antibiotics, vitamins, analgesics and antihistamines. Particularly striking is the popularity of oxytetracycline for the treatment of all tracer conditions. It is the most commonly used antibiotic in Lebak, and it is second only to penicillin in Lombok. When asked which treatment is preferred should a tracer condition occur in the household, the respondents in Lombok state a clear preference for a combination of injections and oral therapy in all tracer conditions. However, in focus group discussions conducted in Lebak, there was no clear cut agreement between participants as to whether injections are always required to effect a cure for most tracer conditions. The only tracer condition for which all mothers agreed that injections are required for a more rapid cure is skin disease.
Hygienic appropriateness of the conditions under which injections are administered was surveyed by means of a combination of observation and interviews in Uganda, and through a questionnaire completed by providers in Indonesia. In both countries the research demonstrates that injections are often unsafe since the minimum hygienic requirements are not being met.
With respect to sterilization, the Ugandan situation is rather unique. Lay person's standards of sterilization diverge from those stipulated by biomedical experts. Besides the general lack of understanding of the concepts of hygiene and sterilization, many households, especially those in the remote and semi-rural areas, consider family control over the needles and syringes and personal knowledge of the users to be a more significant and determining factor in the safety of injection than actual sterilization procedures. Due to the AIDS epidemic the public has come to mistrust communal sterilization procedures and shared needles and syringes, especially in the public health facilities. Personal appropriation of needles and syringes is now very common in Uganda. In the health facilities it was observed that some 60% of the patients bring along their own syringe and needle.
As a result of the widespread practice of keeping injection equipment at home it is rather difficult to meet optimum hygienic standards in Uganda. At the same time, the results of the study also indicate that a high percentage of provider facilities in both regions do not meet the required minimum standards of hygiene at each stage of injection administration. The poor hygienic practices include inadequate sterilization both at home and at provider facilities, picking up the boiled syringes and touching the needles with bare hands, improper disposal of needles and syringes or giving the equipment to the patient to carry home. Moreover, the same equipment is used on multiple patients and the injection site is often not cleaned before injecting. A higher level of training of the health worker was not related to the provision of safer injections. On the whole, NGO health facilities meet the highest hygienic standards.
In Indonesia, the majority of providers interviewed used disposable syringes. However, most disposable syringes are not discarded immediately after use. The majority of providers thus reuse disposable equipment, 'sterilizing' it by a variety of methods. Most injection providers boil the equipment, but usually for less than 20 minutes. There are also signs that the same equipment is sometimes used for more than one patient without first being sterilized although actual practices were not observed in the Indonesian study.
Explaining the popularity of injections
When ill, Ugandans first usually self-medicate with oral therapy. If this brings no relief, or when a fast cure is desired, patients tend to solicit providers for an injection. This preference for injections is guided by local ideas and beliefs of illness and concepts of efficacy of various treatment options. Injections are believed to go directly into the blood stream, unlike oral medications which have to pass the digestive system. Therefore, for diseases that are transmitted through the blood system such as malaria, the administration of injections is preferred. This set of preferences of the general population is strengthened by the profit motive of private providers. Providers often justify their choice for injectables as a way to limit patient non-compliance in oral therapy. The combination of patient beliefs/preferences, and the profit interests of providers results in injections being a therapy which is too frequently administered.
Popular demand for injections is also enormous in Indonesia. In the households and health facilities it was found that over half of all injections had been given on request of the patients or their families. Most patients were given oral medication as well. The majority of users, when asked why this request for injections had been made, stated that injections are preferred because of their "fast action", and also because it is "customary" to receive injections. However, not all injections are given on request. They are also part of a routine treatment procedure in health facilities over which customers have little say. In fact, in focus group discussions, mothers explained that it is usually the nurses who decide whether the children are injected or not. When the providers are asked why they give injections, they usually claim that this is because of patient demand. This vicious circle (health workers give injections because they think patients expect them; patients want injections because health workers give them) and the lack of communication between both parties serves to continue the practice of routine administration of injections. The research confirms that communication between health workers and patients is unsatisfactory. Over half of the patients had not received any explanation from the health worker with regard to their treatment.
The development of a simple and rapid methodology
The two most important strengths of this research project's methodology were the combination of qualitative and quantitative data collection methods; and the flexible research design which allowed for modifications according to local conditions. Formulation of common injection practices indicators and two universal tracer conditions helped to provide cross-country comparable data. All indicators developed in this research, with the exception of the hypothetical tracer conditions, may be considered feasible. For the determination of injection prevalence at the household level, households were visited with a standard questionnaire. It is important that questionnaire design be flexible, as opposed to a blue print design, as this has the advantage of allowing specific questions relevant to the local situation to be included. The two-week recall period to determine injection prevalence proved a good method. All households were selected by means of cluster sampling. There were no problems with non-response.
The research aimed at researching injection prevalence and preference among users and providers. Obtaining consent and collaboration from injection providers, particularly from those in private and illegal practices proved very difficult. In both countries an attempt was made to research a relatively large sample of different providers, using questionnaires, patient record reviews and more qualitative research methods (such as observation) - but the data in both countries are only representative of the formal, public sector.
To measure for which common health problems injections are preferred and used, both universal and local tracer conditions were formulated. Through tracer conditions the medical appropriateness of injection use and treatment preferences can be inferred. The inclusion of country specific tracer conditions has the advantage of covering the most relevant diseases in that country. However, it proved to be rather difficult to find tracer conditions which meet the criteria (a self-limiting ailment for which other forms of treatment are appropriate). For example, for symptoms such as fever or severe vomiting, injections could, in some cases, be medically justified. Therefore it seems necessary to include the degree of severity in the definition of the tracer conditions. For this reason, using tracer conditions in hypothetical illness case presentations in questionnaires proved quite difficult. Another problem encountered in applying the tracer condition method relates to the fact that the selected symptoms often occur in tandem (diarrhoea and vomiting or fever and cough, for example). This makes it necessary that the researchers record all symptoms. Despite some difficulties in applying the method, the defined tracer conditions proved very useful in the injection practices research for the recording of injection use in actual illness cases and for determining the medical appropriateness of the therapy. They seem less suitable for hypothetical illness case presentations in questionnaires.