HEALTH NEEDS ASSESSMENT METHODS

A variety of methods have been used to assess health-care needs, and as stated by McEwen et al., “there is no single best method of assessing health needs—different issues and questions require different methods and approaches and degrees of detail and different combinations of professionals to be involved” (63). Accuracy of results is likely to be achieved if data from one source are checked against data from at least two other sources (70). Applying a “composite” of methods—preferably a combination of quantitative and qualitative methods (70)—to gain more informative data in HNA allows a more precise estimate of a population's health status than when a single measure is applied (29).

To perform HNA surveys nationwide or regionally, it is preferable to organize activities systemically, a principal established in the Scottish Needs Assessment Programme (SNAP; 63). A SNAP core group was organized with seven subgroups to cover specific health-care topics; this strategy created new opportunities for collaboration with public health services, collated data that were not readily available for local use, facilitated surveys, and used the information provided by the policy-makers. Findings of other research confirm that using such an organized approach produces a standard environment in which comparison would be possible (62). However, such strict organization may confine the researchers to specific topics and neglect other needs.

Several structured review articles dealing with HNA tools are available, dealing mainly with mental health subjects (30;33). The criteria in the review by Gilbody et al. (33) were so strict that no single-needs assessment fulfilled them. The study by Evans et al. (30) was also problematic. First, “needs” and “health status” were mixed up, and they referred to appraisals of health-related quality of life from time to time; a confusing approach adopted by others (27). Second, “needs” and “outcome” were used interchangeably, and tools were criticized for not being tested for their responsiveness, an important factor because needs assessment should be on-going rather than a simple cross-sectional “snapshot.” Third, a somewhat primitive circular model has been proposed to describe the relationship of needs, care, and outcome. A more realistic model should incorporate needs analysis, outcome assessment, an acknowledgment that some needs would be met and others remain unfulfilled, a higher level of need might be encountered, and unmet needs must be continually scrutinized until resolved (Figure 1). This model does not rule out potential coexistence of “lower needs” when “higher needs” are being encountered.

Needs, care and outcome relationship.

Selection Criteria

Our purpose was to collect useful data about health needs assessment tools; only surveys that targeted “health needs assessment” in the community or within conditions were selected. A key word search was conducted in each database separately, and original papers were reviewed to identify study characteristics, including the kind of tool, sample size, method of administration, and psychometric properties.

Review articles (provided that no new needs assessment instrument was introduced), papers published in non-English journals, surveys on children (two papers), and oral health (one paper) were excluded. Articles that did not include details of any needs surveys tools were discarded.

Findings

The search revealed thirty-one tools in fifty-two publications that described the development of a new “health needs assessment” tool or the deployment of a readily available instrument. Some of these studies attempted to produce a definition of need (17;64;89) or had a critical view on current definitions (7;31). These reports, summarized in Table 1, can be categorized into four main groups in terms of method of administration: “interview,” “self-administered questionnaire,” “community rapid appraisal,” and “epidemiological rapid assessment.”

Interview

Qualitative research is usually performed by direct observation, consensus, analysis of texts or documents, case study, and interview (80). Interviewing, usually defined as a continuum of standardization (68) comprising structured questions (mostly with a structured questionnaire), semistructured questions (4;8), depth questions (unstructured but dependent on the interviewee's responses; 18), and focus group discussion (51), is well established, especially in sociology. Denzin categorized the interviews as standard scheduled, nonschedule standardized, and nonstandardized (28).

Interviewing, a commonplace methodology for qualitative research may help to define dimensions that a quantitative method such as self-administered questionnaires would aim to measure. Qualitative methods also describe a population's intelligible language that definitely helps quantitative methods perform well. Face-to-face, telephone interview, focus group discussion, or check lists are often used to conduct HNA surveys; these methods are usually followed by other quantitative approaches such as self-administered questionnaires (51). A qualitative approach is supposed to facilitate quantitative methodology in different ways (4;77), although meticulous quantitative research followed by a thorough content analysis may provide a more accurate picture of unmet health needs (60;83). Moreover, including extra facilities such as a vignette (11) within, for example, a focus group may provide a valuable insight into expressed beliefs.

The use of HNA tools is burgeoning, especially among those dealing with mental disorders. Conceptualization of needs by the Medical Research Council, Social Psychiatry Unit (99) led to various research work, including development of the MRC Needs for Care Assessment Schedule (NFACS; 17), The Camberwell Assessment of Need (CAN; 78), and the Cardinal Needs Schedule (61). The NFACS comprises three elements of identifying patients' needs, lists interventions, and finally recognizes the unmet needs to reorder the cycle (17). NFACS focused on health professionals rather than patients and carers, which was corrected to some degree in the CAN, as a tool for research and clinical administration in severe mental illness (30), with good validity and inter-rater reliability (78). However, it was unclear how different scores from several parties might be combined. Two problems have been identified: the small range offered by its three-point scale may increase the floor–ceiling effect and incompatibility can affect test–retest reliability. Versions of CAN have been developed for children and the elderly and computerized and foreign language versions are available (78). The CAN for the Elderly (CANE) has good validity and reliability in those areas in which it has been developed, covering perceptions of patients needs by patients, carers, and professionals (81); it has been administered in primary care as well (93).

The Salford Needs Assessment Schedule for Adolescents (S.NASA), which includes an interview schedule for client and carer (53), and the Patient Needs Assessment Tool (PNAT), an interviewer-rated scale to detect cancer patients' needs (26) are examples of structured interviews in which validity and reliability can be measured.

The semistructured interview is a useful way of evaluating needs in mental health and palliative care services (55;96); however, the limitations of an interview should be given serious consideration. The Southampton Needs Assessment Questionnaire (SNAQ), which has been used in a face-to-face interview survey to assess the needs of the disabled, is a semistructured tool with acceptable validity and internal reliability (48). Checklists, that require a health-worker to conduct a detailed survey to measure psychometric properties, have to be carefully developed to avoid misleading results that might fail to identify unmet needs. The Needs Assessment Checklist is a very detailed tool comprising 216 items (42), used for stroke patients with established psychometric features (14).

Although an interview has noticeable advantages in terms of a direct relationship to clients (which avoids any misunderstanding and ensures a high response rate), it does require trained interviewers, standard questions, precise transcription, and adequate time and resources. In addition, organizing meetings may present various difficulties, as interviewees are often patients with major physical or psychological problems. Furthermore, obtaining simple information such as date of birth is not necessarily straightforward (57), so reproducibility in interviews can be less than perfect.

Self-Administered Questionnaire

Self-administered questionnaires are low-cost; easily distributed; less intrusive than interviews, as patients complete the tool at home at their own time, stop and start at will, and generate less stress when information is disclosed; tend to have a large sample size and short data collection period; and their validity and reliability (especially inter-rater) can be measured. However, developing these instruments does require precise wording (avoidance of jargon and double-negative and double-barreled statements) and may confine researchers to limited interaction with patients and consequently may lead to misunderstandings, blank or incomplete answers, and a relatively superficial investigation.

Using standard tools to assess health-care needs is a focus of researchers' attention, but there are certain drawbacks; for example, some health needs may remain unidentified (55) or misunderstandings are possible. To overcome such limitations, most HNA researchers prefer a standard approach of performing an initial qualitative (interview, focus group, and so on) followed by a postal survey (7). By contrast, Mesters et al. (65), however, conducted both approaches at the same time, which seems somewhat futile.

A postal survey can provide valuable data, especially about chronic illnesses in the community and perceived needs about existing and potential services. Postal HNA surveys are not new (19); however, they are rapidly gaining in popularity (20;52;67). Most are in the area of mental health (41; see Table 1); few deal with local health services and have been excluded from this review due to lack of publicity (85;90).

Disease-specific instruments have been developed; for example, the Northwick Park Dependency Score (NPDS) for HNA in hospital, and its community rehabilitation equivalent have been administered in various studies (91) and takes outcome expenditure into account to evaluate cost-effectiveness of rehabilitation (72).

INTERMED (39) is a well-validated methodology that gathers data from various biological, psychological, social, and health-care systems over the time (88). Self-administered instruments can provide a short-cut in the investigation of service shortcomings (32).

The Supportive Care Needs Survey has been developed exclusively for cancer patients from its “mother” version, the Cancer Patient Needs Questionnaire (CPNQ), consisting of fifty-two variables in a five-point scoring format (31); given its simplicity, patients had no difficulties completing it, and the tool showed acceptable validity and reliability (16;84). To assess HNA in multiple sclerosis patients, their carers, and the professionals' perception of Services and Unmet Needs (the SUN), a specific tool has been used in different European countries (46). The Needs Evaluation Questionnaire, originally developed with twenty-three items to ascertain HNA in oncology inpatients, has acceptable validity and reliability (89). “Needs assessment” has a cross-sectional nature that is valuable for certain circumstances, although test–retest technique is far inferior to measuring internal reliability or inter-rater assessment, particularly when different methods are considered. Hence, some of the HNA reports (31;89) are contentious.

A Combined Approach

Qualitative methods describe dimensions of a population's health needs and, provided they are followed by a thorough content analysis, help quantitative surveys to perform well (80). Despite the ability of qualitative methods to discover unmet needs, lack of generalizability jeopardizes their conclusions. Therefore, some researchers prefer to conduct both at the same time or at different stages (4;24;77;97). For example, to assess pharmaceutical needs in primary care, a three-stage set was used: semistructured interviews with clienteles and providers followed by a postal survey and finally an open forum meeting to present findings and prioritize services. This strategy may make a formal cooperative mechanism for both consumers and providers (97), although validity and reliability of the instrument remain to be proven.

Community Rapid Appraisal

This method seeks to gain community perspectives of local health and social needs in a finite period of time and to translate these findings into action. The development of rapid appraisal methods during the 1980s was fundamental to the “Health for all 2000” framework (94) in response to the time-consuming and the rather structured and rigid nature of traditional epidemiological and questionnaire surveys. These appraisal methods showed that, when conducted well, they provide valuable, reliable, and timely information on health status, knowledge, attitude, and behaviors (100). Data are generally collected from interviews with local key individuals, written records, and direct observation. Informants are not selected randomly but “purposefully,” that is, seeking a range of people who are in the best position to understand the issues. Including relevant interviewees and summary health data from primary and secondary care can incorporate professional insights (69). This method provides specific insight over the population's needs rather than merely epidemiological data (71), especially about deprived areas (74), although it retains the limitations of qualitative methods.

Whatever method is used for appraisals, to ensure that people's genuine needs are fully transcribed, the emphasis is on qualitative techniques of interviewing and listening to people. In community appraisal, the assessors support and facilitate the community's understanding and action rather than just recording information. Actively involving people in assessing their heath-care needs is likely to be an appropriate approach for community appraisal; this was the aim of the Waterfront Health Action Project (WHAP) (92). The Patient Liaison Group (PLG) is a practical attempt to encourage people to participate systematically in identifying health-care needs and medical decision making (98).

Although team-working results in rapid completion of surveys and is considered an “asset rather than a liability” (12), differences in surveyors' perceptions and views may cause confused expectations within the group, distorted group effectiveness, intragroup conflict, and attribution of minority perspectives to the majority view of the group (50). Emerging appraisal methods such as Q methodology, which classifies individuals according to their subjective feelings about a topic using a mixed quantitative/qualitative statistical technique (9), has been used recently in needs assessment (23) but requires further investigation of its reliability and validity in this field.

Epidemiological Rapid Assessment

This method is based on traditional measures of incidence and prevalence and can be used for rapid health assessment after disasters and in emergency medicine. In such cases, a rapid needs assessment demonstrates priorities where resources should be spent to satisfy the immediate needs of sufferers. Modified cluster-sampling methods have often been used to conduct needs assessment surveys to explore food, water, sanitation, and health-care needs after an exceptionally violent hurricane in the Dominican Republic in 1998 (1), similar to that in Florida (35) and after the recent earthquake in Iran (2). However, HNA techniques have not been formally used.

In another case, a survey carried out to determine the health-care needs of the Cambodian refugees comprised two components: a rapid epidemiological appraisal and then an interview-based needs assessment. The first step was a tent-by-tent search to determine who required hospital admission—the mortality rate fell from 9.1 to 0.7-deaths' 10,000 population'day (34).

RECOMMENDATIONS

Health needs is a deceptively tricky concept in which a wide range of issues must be taken into account. Evaluation of these needs, therefore, should be as comprehensive as possible; nevertheless, to obtain sensible results for routine use in clinical practice, it is essential to select a concise, valid, and reliable tool or to develop some standard measure. Attention to needs surveys is much less than to outcome assessment such as the measurement of health-related quality of life, although less evidence exists about the psychometric requirements in the literature. Triangulating HNA data with other valid sources of information, such as clinical data, health-related quality of life measurement, or satisfaction surveys, is fundamental to ascertain the validity and responsiveness of a questionnaire. Obviously, there is considerable needs assessment experience in routine health-care services buried within local delivery points; however, more published data are needed in this field to help researchers publicize their findings and encourage policy-makers to meet genuine health needs.

CONTACT INFORMATION

Mohsen Asadi-Lari, MD, PhD (msxma@nottingham.ac.uk), Division of Cardiovascular Medicine, University of Nottingham, Derby Road, Nottingham NG7 2UH, UK; Senior Expert, Ministry of Health & Medical Education, Iran, Tehran 11344, Iran

David Gray, DM, MPH, Reader in Medicine, Honorary Consultant Cardiologist (d.gray@nottingham.ac.uk), Division of Cardiovascular Medicine, University of Nottingham, Derby Road, Nottingham NG7 2UH, UK; Division of Cardiovascular Medicine, Queen's Medical Center, University Hospital, Nottingham NG7 2UH, UK